SOURCES IN RECREATIONAL MATHEMATICS

AN ANNOTATED BIBLIOGRAPHY

 

EIGHTH PRELIMINARY EDITION

 

DAVID SINGMASTER

 

Contact via the Puzzle Museum

http://puzzlemuseum.com

Last updated on 19 March 2004.

 

          This is a copy of the current version from my source files.  I had intended to reorganise the material before producing a Word version, but have decided to produce this version for G4G6 and to renumber it as the Eighth Preliminary Edition.

 

         

 

 

          If I have perchance omitted anything more or less proper or necessary, I beg indulgence, since there is no one who is blameless and utterly provident in all things.  [Fibonacci, translated by Grimm.])

 

CONTENTS        

 

INTRODUCTION 16

NATURE OF THIS WORK 16

SIMILAR WORKS 17

COVERAGE 18

STATUS OF THE PROJECT 19

TECHNICAL NOTES 21

ACKNOWLEDGMENTS 22

ABBREVIATIONS 25

DIACRITICAL MARKS AND NOTATION 25

SOME OTHER RECURRING REFERENCES 89

1.       BIOGRAPHICAL MATERIAL  --  in chronological order 91

2.       GENERAL PUZZLE COLLECTIONS AND SURVEYS 99

3.       GENERAL HISTORICAL AND BIBLIOGRAPHICAL MATERIAL 100

3.A.       GENERAL HISTORICAL MATERIAL 100

3.B.       BIBLIOGRAPHICAL MATERIAL 100

4.       MATHEMATICAL GAMES 105

4.A.       GENERAL THEORY AND NIM‑LIKE GAMES 105

4.A.1.    ONE PILE GAME 106

4.A.1.a.        THE 31 GAME 109

4.A.2.        SYMMETRY ARGUMENTS 110

4.A.3.    KAYLES 111

4.A.4.    NIM 112

4.A.5.    GENERAL THEORY 113

4.B.       PARTICULAR GAMES 115

4.B.1.    TIC‑TAC‑TOE  =  NOUGHTS AND CROSSES 115

4.B.1.a     IN HIGHER DIMENSIONS 123

4.B.2.    HEX 124

4.B.3.    DOTS AND BOXES 125

4.B.4.    SPROUTS 126

4.B.5.    OVID'S GAME AND NINE MEN'S MORRIS 126

4.B.6.    PHUTBALL 132

4.B.7.    BRIDG‑IT 132

4.B.8.    CHOMP 133

4.B.9.    SNAKES AND LADDERS 133

4.B.10.      MU TORERE 135

4.B.11.      MASTERMIND, ETC. 136

4.B.12. RITHMOMACHIA  =  THE PHILOSOPHERS' GAME 136

4.B.13.      MANCALA GAMES 137

4.B.14.DOMINOES, ETC. 138

4.B.15.      SVOYI KOSIRI 138

5.COMBINATORIAL RECREATIONS 139

5.A.       THE 15 PUZZLE, ETC. 139

GENERAL 139

EARLY ALPHABETIC VERSIONS 139

LOYD 140

THE 15 PUZZLE 140

5.A.1.    NON‑SQUARE PIECES 147

5.A.2.    THREE DIMENSIONAL VERSIONS 148

5.A.3.    ROLLING PIECE PUZZLES 149

5.A.4.    PANEX PUZZLE 150

5.B.       CROSSING PROBLEMS 150

5.B.1.    LOWERING FROM TOWER PROBLEM 159

5.B.2.    CROSSING A BRIDGE WITH A TORCH 160

5.C.       FALSE COINS WITH A BALANCE 160

5.C.1RANKING COINS WITH A BALANCE 163

5.D. MEASURING PROBLEMS 163

5.D.1.    JUGS & BOTTLES 163

5.D.2.    RULER WITH MINIMAL NUMBER OF MARKS 169

5.D.3     FALSE COINS WITH A WEIGHING SCALE 170

5.D.4.    TIMING WITH HOURGLASSES 170

5.D.5.    MEASURE HALF A BARREL 170

5.E.       EULER CIRCUITS AND MAZES 171

5.E.1.    MAZES 175

5.E.2.    MEMORY WHEELS  =  CHAIN CODES 180

5.E.2.a.    PANTACTIC SQUARES 181

5.F.    HAMILTONIAN CIRCUITS 182

5.F.1.     KNIGHT'S TOURS AND PATHS 182

5.F.2.     OTHER HAMILTONIAN CIRCUITS 189

5.F.3.     KNIGHT'S TOURS IN HIGHER DIMENSIONS 190

5.F.4.     OTHER CIRCUITS IN AND ON A CUBE 190

5.G.       CONNECTION PROBLEMS 191

5.G.1.    GAS, WATER AND ELECTRICITY 191

5.H.       COLOURED SQUARES AND CUBES, ETC. 192

5.H.1.    INSTANT INSANITY  =  THE TANTALIZER 192

5.H.2.    MACMAHON PIECES 192

5.H.3.    PATH FORMING PUZZLES 194

5.H.4.    OTHER AND GENERAL 195

5.I.     LATIN SQUARES AND EULER SQUARES 197

5.I.1.     EIGHT QUEENS PROBLEM 199

5.I.2.     COLOURING CHESSBOARD WITH NO REPEATS IN A LINE 202

5.J.    SQUARED SQUARES, ETC. 202

5.J.1.     MRS PERKINS'S QUILT 205

5.J.2.     CUBING THE CUBE 205

5.J.3.     TILING A SQUARE OF SIDE 70 WITH SQUARES OF SIDES  1, 2, ..., 24 205

5.K.       DERANGEMENTS 205

5.K.1.    DERANGED BOXES OF  A,  B  AND  A & B 206

5.K.2.    OTHER LOGIC PUZZLES BASED ON DERANGEMENTS 207

5.K.3.    CAYLEY'S MOUSETRAP 207

5.L.       MÉNAGE PROBLEM 207

5.M.      SIX PEOPLE AT A PARTY  --  RAMSEY THEORY 208

5.N.       JEEP OR EXPLORER'S PROBLEM 209

5.O.       TAIT'S COUNTER PUZZLE:  BBBBWWWW  TO  WBWBWBWB 212

5.P.    GENERAL MOVING PIECE PUZZLES 214

5.P.1.    SHUNTING PUZZLES 214

5.P.2.    TAQUIN 216

5.Q.       NUMBER OF REGIONS DETERMINED BY  N  LINES OR PLANES 216

5.Q.1.      NUMBER OF INTERSECTIONS DETERMINED BY  N  LINES 217

5.R.       JUMPING PIECE GAMES 217

5.R.1.    PEG SOLITAIRE 217

5.R.1.a.    TRIANGULAR VERSION 220

5.R.1.b.    OTHER SHAPES 221

5.R.2.    FROGS AND TOADS:  BBB_WWW  TO  WWW_BBB 222

5.R.3.    FORE AND AFT  --  3 BY 3  SQUARES MEETING AT A CORNER 224

5.R.4.    REVERSING FROGS AND TOADS:  _12...n  TO  _n...21 , ETC. 225

5.R.5.    FOX AND GEESE, ETC. 225

5.R.6.    OCTAGRAM PUZZLE 228

5.R.7.    PASSING OVER COUNTERS 230

5.S.    CHAIN CUTTING AND REJOINING 233

5.S.1.     USING CHAIN LINKS TO PAY FOR A ROOM 234

5.T.       DIVIDING A CAKE FAIRLY 234

5.U.       PIGEONHOLE RECREATIONS 235

5.V.       THINK‑A‑DOT, ETC. 236

5.W.      MAKING THREE PIECES OF TOAST 237

5.W.1.       BOILING EGGS 237

5.X.       COUNTING FIGURES IN A PATTERN 237

5.X.1.    COUNTING TRIANGLES 237

5.X.2.    COUNTING RECTANGLES OR SQUARES 240

5.X.3.    COUNTING HEXAGONS 241

5.X.4.    COUNTING CIRCLES 242

5.Y.       NUMBER OF ROUTES IN A LATTICE 242

5.Z.       CHESSBOARD PLACING PROBLEMS 245

5.Z.1.    KINGS 246

5.Z.2.    QUEENS 246

5.Z.3.    BISHOPS 247

5.Z.4.    KNIGHTS 248

5.Z.5.    ROOKS 248

5.Z.6.    MIXTURES 248

5.AA.   CARD SHUFFLING 249

5.AB.     FOLDING A STRIP OF STAMPS 251

5.AC.      PROPERTIES OF THE SEVEN BAR DIGITAL DISPLAY 251

5.AD.    STACKING A DECK TO PRODUCE A SPECIAL EFFECT 252

5.AE.     REVERSING CUPS 252

5.AF.     SPOTTING DICE 252

5.AG.    RUBIK'S CUBE AND SIMILAR PUZZLES 253

5.AG.1.     RUBIK'S CUBE 253

5.AG.2.     HUNGARIAN RINGS, ETC. 253

6.       GEOMETRIC RECREATIONS 254

6.A.       PI 254

6.B.       STRAIGHT LINE LINKAGES 255

6.C.       CURVES OF CONSTANT WIDTH 257

6.D.       FLEXAGONS 257

6.E.       FLEXATUBE 259

6.F.    POLYOMINOES, ETC. 259

6.F.1.        OTHER CHESSBOARD DISSECTIONS 268

6.F.2.     COVERING DELETED CHESSBOARD WITH DOMINOES 270

6.F.3.     DISSECTING A CROSS INTO  Zs  AND  Ls 270

6.F.4.     QUADRISECT AN  L‑TROMINO, ETC. 272

6.F.5.     OTHER DISSECTIONS INTO POLYOMINOES 274

6.G.       SOMA CUBE 275

6.G.1.    OTHER CUBE DISSECTIONS 275

6.G.2.    DISSECTION OF  63  INTO  33,  43  AND  53,  ETC. 276

6.G.3.    DISSECTION OF A DIE INTO NINE  1 x 1 x 3 277

6.G.4.    USE OF OTHER POLYHEDRAL PIECES 277

6.H.       PICK'S THEOREM 278

6.I.     SYLVESTER'S PROBLEM OF COLLINEAR POINTS 278

6.J.    FOUR BUGS AND OTHER PURSUIT PROBLEMS 279

6.K.       DUDENEY'S SQUARE TO TRIANGLE DISSECTION 281

6.L.       CROSSED LADDERS 281

6.L.1.        LADDER OVER BOX 283

6.M.      SPIDER & FLY PROBLEMS 283

6.N.       DISSECTION OF A  1 x 1 x 2  BLOCK TO A CUBE 284

6.O.       PASSING A CUBE THROUGH AN EQUAL OR SMALLER CUBE. PRINCE RUPERT'S PROBLEM 285

6.P.    GEOMETRICAL VANISHING 286

6.P.1.    PARADOXICAL DISSECTIONS OF THE CHESSBOARD BASED ON FIBONACCI NUMBERS 286

6.P.2.    OTHER TYPES 288

6.Q.       KNOTTING A STRIP TO MAKE A REGULAR PENTAGON 290

6.R.       GEOMETRIC FALLACIES 291

6.R.1.    EVERY TRIANGLE IS ISOSCELES 292

6.R.2.    A RIGHT ANGLE IS OBTUSE 292

6.R.3.    LINES APPROACHING BUT NOT MEETING 292

6.R.4.    OTHERS 293

6.S.    TANGRAMS, ET AL. 293

GENERAL HISTORIES. 293

SPECIFIC ITEMS 295

6.S.1.     LOCULUS OF ARCHIMEDES 304

6.S.2.     OTHER SETS OF PIECES 305

6.T.       NO THREE IN A LINE PROBLEM 306

6.U.       TILING 307

6.U.1.    PENROSE PIECES 307

6.U.2.    PACKING BRICKS IN BOXES 308

6.V.       SILHOUETTE AND VIEWING PUZZLES 308

6.W.      BURR PUZZLES 311

6.W.1.       THREE PIECE BURR 311

6.W.2.       SIX PIECE BURR  =  CHINESE CROSS 312

6.W.3.       THREE PIECE BURR WITH IDENTICAL PIECES 314

6.W.4.       DIAGONAL SIX PIECE BURR  =  TRICK STAR 314

6.W.5.       SIX PIECE BURR WITH IDENTICAL PIECES 315

6.W.6.       ALTEKRUSE PUZZLE 315

6.W.7.  OTHER BURRS 316

6.X.       ROTATING RINGS OF POLYHEDRA 316

6.Y.       ROPE ROUND THE EARTH 317

6.Z.       LANGLEY'S ADVENTITIOUS ANGLES 319

6.AA.    NETS OF POLYHEDRA 320

6.AB.     SELF‑RISING POLYHEDRA 321

6.AC.    CONWAY'S LIFE 321

6.AD.    ISOPERIMETRIC PROBLEMS 322

6.AD.1.     LARGEST PARCEL ONE CAN POST 323

6.AE.     6"  HOLE THROUGH SPHERE LEAVES CONSTANT VOLUME 324

6.AF.     WHAT COLOUR WAS THE BEAR? 325

6.AG.    MOVING AROUND A CORNER 327

6.AH.    TETHERED GOAT 328

6.AI.      TRICK JOINTS 329

6.AJ.     GEOMETRIC ILLUSIONS 330

6.AJ.1       TWO PRONGED TRIDENT 333

6.AJ.2.      TRIBAR AND IMPOSSIBLE STAIRCASE 334

6.AJ.3.      CAFÉ WALL ILLUSION 335

6.AJ.4.      STEREOGRAMS 336

6.AJ.5.      IMPOSSIBLE CRATE. 336

6.AK.    POLYGONAL PATH COVERING  N x N LATTICE OF POINTS, QUEEN'S TOURS, ETC. 336

6.AL.     STEINER‑LEHMUS THEOREM 339

6.AM.       MORLEY'S THEOREM 340

6.AN.    VOLUME OF THE INTERSECTION OF TWO CYLINDERS 341

6.AO.    CONFIGURATION PROBLEMS 341

6.AO.1.     PLACE FOUR POINTS EQUIDISTANTLY  =  MAKE FOUR TRIANGLES WITH SIX MATCHSTICKS 349

6.AO.2.     PLACE AN EVEN NUMBER ON EACH LINE 351

6.AP      . DISSECTIONS OF A TETRAHEDRON 352

6.AP.1.     TWO PIECES 352

6.AP.2.     FOUR PIECES 352

6.AQ.    DISSECTIONS OF A CROSS,  T  OR  H 352

6.AR.    QUADRISECTED SQUARE PUZZLE 354

6.AS.     DISSECTION OF SQUARES INTO A SQUARE 354

6.AS.1.      TWENTY 1, 2, Ö5 TRIANGLES MAKE A SQUARE OR FIVE EQUAL SQUARES TO A SQUARE 355

6.AS.1.a.      GREEK CROSS TO A SQUARE 358

6.AS.1.b.      OTHER GREEK CROSS DISSECTIONS 358

6.AS.2.      TWO (ADJACENT) SQUARES TO A SQUARE 359

6.AS.2.a.      TWO EQUAL SQUARES TO A SQUARE 361

6.AS.3.      THREE EQUAL SQUARES TO A SQUARE 361

6.AS.3.a.      THREE EQUAL 'SQUARES' TO A HEXAGON 362

6.AS.4.      EIGHT EQUAL SQUARES TO A SQUARE 362

6.AS.5.      RECTANGLE TO A SQUARE OR OTHER RECTANGLE 362

6.AT.     POLYHEDRA AND TESSELLATIONS 363

6.AT.1.     REGULAR POLYHEDRA 363

6.AT.2      STAR AND STELLATED POLYHEDRA 367

6.AT.3.     ARCHIMEDEAN POLYHEDRA 369

6.AT.4.     UNIFORM POLYHEDRA 374

6.AT.5.     REGULAR‑FACED POLYHEDRA 374

6.AT.6.     TESSELLATIONS 374

6.AT.6.a.     TESSELLATING WITH CONGRUENT FIGURES 375

6.AT.7.     PLAITING OF POLYHEDRA 376

6.AT.8.     DÜRER'S OCTAHEDRON 376

6.AT.9.     OTHER POLYHEDRA 377

6.AU.    THREE RABBITS, DEAD DOGS AND TRICK MULES 377

CHINA 378

OTHER ASIA 381

PADERBORN 383

MEDIEVAL EUROPE 385

MODERN VERSIONS OF THE THREE RABBITS PUZZLE 393

DEAD DOGS 394

TRICK MULES 398

6.AV.    CUTTING UP IN FEWEST CUTS 399

6.AW.       DIVISION INTO CONGRUENT PIECES 399

6.AW.1.    MITRE PUZZLE 399

6.AW.2.    REP‑TILES 400

6.AW.3.    DIVIDING A SQUARE INTO CONGRUENT PARTS 401

6.AW.4.    DIVIDING AN  L-TROMINO INTO CONGRUENT PARTS 402

6.AX.     THE PACKER'S SECRET 402

6.AY.     DISSECT  3A x 2B  TO MAKE  2A x 3B,  ETC. 402

6.AY.1.     O'BEIRNE'S STEPS 405

6.AY.2.     SWISS FLAG PUZZLE 405

6.AZ.     BALL PYRAMID PUZZLES 406

6.BA.     CUTTING A CARD SO ONE CAN PASS THROUGH IT 407

6.BB.     DOUBLING AN AREA WITHOUT CHANGING ITS HEIGHT OR WIDTH 407

6.BC.     HOFFMAN'S CUBE 408

6.BD.  BRIDGE A MOAT WITH PLANKS 408

6.BE.     REVERSE A TRIANGULAR ARRAY OF TEN CIRCLES 410

6.BF.     PYTHAGOREAN RECREATIONS 410

6.BF.1       THE BROKEN BAMBOO 411

6.BF.2.      SLIDING SPEAR  =  LEANING REED 412

6.BF.3.      WELL BETWEEN TWO TOWERS 413

6.BF.4.      RAIL BUCKLING. 416

6.BF.5.      TRAVELLING ON SIDES OF A RIGHT TRIANGLE. 417

6.BG.    QUADRISECT A PAPER SQUARE WITH ONE CUT 417

6.BH.    MOIRÉ PATTERNS 417

6.BI.  VENN DIAGRAMS FOR  N  SETS 418

6.BJ.     3D DISSECTION PUZZLES 420

6.BK     SUPERELLIPSE 420

6.BL.     TAN-1 ⅓ + TAN-1 ½  =  TAN-1 1, ETC. 421

6.BM.       DISSECT CIRCLE INTO TWO HOLLOW OVALS 421

6.BN.     ROUND PEG IN SQUARE HOLE OR VICE VERSA 422

6.BO.    BUTTERFLY PROBLEM 423

6.BP.     EARLY MATCHSTICK PUZZLES 423

6.BQ.    COVERING A DISC WITH DISCS 424

6.BR.     WHAT IS A GENERAL TRIANGLE? 424

6.BS.     FORM SIX COINS INTO A HEXAGON 425

6.BT.     PLACING OBJECTS IN CONTACT 425

6.BU.     CONSTRUCTION OF  N-GONS 426

6.BV.     GEOMETRIC CONSTRUCTIONS 428

6.BW.       DISTANCES TO CORNERS OF A SQUARE 429

7.       ARITHMETIC  &  NUMBER‑THEORETIC RECREATIONS 430

7.A.       FIBONACCI NUMBERS 430

7.B.       JOSEPHUS OR SURVIVOR PROBLEM 433

7.C.       EGYPTIAN FRACTIONS 448

7.D.       THE FIRST DIGIT PROBLEM 449

7.E.       MONKEY AND COCONUTS PROBLEMS 449

7.E.1.    VERSIONS WITH ALL GETTING THE SAME 463

7.F.        ILLEGAL OPERATIONS GIVING CORRECT RESULT 465

7.G.       INHERITANCE PROBLEMS 465

7.G.1.    HALF + THIRD + NINTH, ETC. 465

7.G.2.    POSTHUMOUS TWINS, ETC. 471

7.H.       DIVISION AND SHARING PROBLEMS  --  CISTERN PROBLEMS 474

7.H.1.    WITH GROWTH  --  NEWTON'S CATTLE PROBLEM 493

7.H.2.    DIVISION OF CASKS 495

7.H.3.    SHARING UNEQUAL RESOURCES  --  PROBLEM OF THE PANDECTS 497

7.H.4.    EACH DOUBLES OTHERS' MONEY TO MAKE ALL EQUAL, ETC. 500

7.H.5.    SHARING COST OF STAIRS, ETC. 503

7.H.6.    SHARING A GRINDSTONE 504

7.H.7.    DIGGING PART OF A WELL. 505

7.I.        FOUR FOURS, ETC. 507

7.I.1.     LARGEST NUMBER USING FOUR ONES, ETC. 514

7.J.        SALARY PUZZLE 516

7.K.       CONGRUENCES 517

7.K.1.    CASTING OUT NINES 517

7.L.       GEOMETRIC PROGRESSIONS 520

7.L.1.    1 + 7 + 49 + ...  &  ST. IVES 523

7.L.2.    1 + 2 + 4 + ... 525

7.L.2.a.    CHESSBOARD PROBLEM 526

7.L.2.b.    HORSESHOE NAILS PROBLEM 530

7.L.2.c.    USE OF 1, 2, 4, ... AS WEIGHTS, ETC. 531

7.L.3.    1 + 3 + 9 + ... AND OTHER SYSTEMS OF WEIGHTS 533

7.M.      BINARY SYSTEM AND BINARY RECREATIONS 535

7.M.1.       CHINESE RINGS 536

7.M.2.       TOWER OF HANOI 540

7.M.2.a.       TOWER OF HANOI WITH MORE PEGS 544

7.M.3.       GRAY CODE 544

7.M.4.       BINARY DIVINATION 545

7.M.4.a.       TERNARY DIVINATION 547

7.M.4.b.   OTHER DIVINATIONS USING BINARY OR TERNARY 547

7.M.5.  LOONY LOOP  =  GORDIAN KNOT 551

7.M.6.       BINARY BUTTON GAMES 551

7.N.       MAGIC SQUARES 554

SURVEYS 554

POSSIBLE EARLY REFERENCES 555

7.N.1     MAGIC CUBES 569

7.N.2.    MAGIC TRIANGLES 571

7.N.3.    ANTI‑MAGIC SQUARES AND TRIANGLES 573

7.N.4.    MAGIC KNIGHT'S TOUR 574

7.N.5.    OTHER MAGIC SHAPES 575

7.O.       MAGIC HEXAGON 578

7.O.1.    OTHER MAGIC HEXAGONS 579

7.P.    DIOPHANTINE RECREATIONS 582

7.P.1.    HUNDRED FOWLS AND OTHER LINEAR PROBLEMS 582

7.P.2.    CHINESE REMAINDER THEOREM 602

7.P.3.    ARCHIMEDES' CATTLE PROBLEM 608

7.P.4.    PRESENT OF GEMS 609

7.P.5     . SELLING DIFFERENT AMOUNTS 'AT SAME PRICES' YIELDING THE SAME 610

7.P.6.    CONJUNCTION OF PLANETS, ETC. 616

7.P.7.    ROBBING AND RESTORING 617

7.Q.       BLIND ABBESS AND HER NUNS -- REARRANGEMENT ALONG SIDES OF A  3 x 3  SQUARE TO CONSERVE SIDE TOTALS 619

7.Q.1.    REARRANGEMENT ON A CROSS 622

7.Q.2.    REARRANGE A CROSS OF SIX TO MAKE TWO LINES OF FOUR, ETC. 623

7.R.       "IF I HAD ONE FROM YOU, I'D HAVE TWICE YOU" 624

7.R.1.    MEN FIND A PURSE AND 'BLOOM OF THYMARIDAS' 631

7.R.2.    "IF I HAD  1/3  OF YOUR MONEY, I COULD BUY THE HORSE" 638

7.R.3.    SISTERS AND BROTHERS 648

7.R.4.    "IF I SOLD YOUR EGGS AT MY PRICE, I'D GET ...." 649

7.S.    DILUTION AND MIXING PROBLEMS 650

7.S.1.     DISHONEST BUTLER DRINKING SOME AND REPLACING WITH WATER 650

7.S.2.     WATER IN WINE VERSUS WINE IN WATER 651

7.T.       FOUR NUMBER GAME 652

7.U.       POSTAGE STAMP PROBLEM 653

7.V.       XY  =  YX  AND ITERATED EXPONENTIALS 653

7.W.      CARD PILING OVER A CLIFF 654

7.X.       HOW OLD IS ANN? AND OTHER AGE PROBLEMS 655

7.Y.       COMBINING AMOUNTS AND PRICES INCOHERENTLY 666

7.Y.1.    REVERSAL OF AVERAGES PARADOX 668

7.Y.2.    UNFAIR DIVISION 669

7.Z.       MISSING DOLLAR AND OTHER ERRONEOUS ACCOUNTING 669

7.AA.    NEGATIVE DIGITS 671

7.AA.1.     NEGATIVE BASES, ETC. 672

7.AB.     PERFECT NUMBERS, ETC. 673

7.AC.    CRYPTARITHMS, ALPHAMETICS AND SKELETON ARITHMETIC 675

7.AC.1.     CRYPTARITHMS:  SEND + MORE  =  MONEY,  ETC. 676

7.AC.2.     SKELETON ARITHMETIC:  SOLITARY SEVEN, ETC. 678

7.AC.3.     PAN‑DIGITAL SUMS 681

7.AC.3.a      INSERTION OF SIGNS TO MAKE 100, ETC. 682

7.AC.4.     PAN‑DIGITAL PRODUCTS 683

7.AC.5.     PAN‑DIGITAL FRACTIONS 685

7.AC.6.     OTHER PAN‑DIGITAL AND SIMILAR PROBLEMS 686

7.AC.7.     SELF-DESCRIPTIVE NUMBERS, PANGRAMS, ETC. 689

7.AD. SELLING, BUYING AND SELLING SAME ITEM 690

7.AD.1.     PAWNING MONEY 691

7.AE.     USE OF COUNTERFEIT BILL OR FORGED CHEQUE 691

7.AF.     ARITHMETIC PROGRESSIONS 693

7.AF.1.      COLLECTING STONES 693

7.AF.2.      CLOCK STRIKING 696

7.AG.    2592 697

7.AH.    MULTIPLYING BY REVERSING 697

7.AH.1.     OTHER REVERSAL PROBLEMS 698

7.AI.      IMPOSSIBLE EXCHANGE RATES 699

7.AJ.     MULTIPLYING BY SHIFTING 699

7.AJ.1.      MULTIPLYING BY APPENDING DIGITS 701

7.AK.    LAZY WORKER 701

7.AL.     IF  A  IS  B,  WHAT IS  C? 704

7.AM.       CROSSNUMBER PUZZLES 707

7.AN.    THREE ODDS MAKE AN EVEN, ETC. 709

7.AO.    DIVINATION OF A PERMUTATION 711

7.AP.     KNOWING SUM VS KNOWING PRODUCT 716

7.AQ.    NUMBERS IN ALPHABETIC ORDER 719

7.AR.    1089 719

7.AS.     CIGARETTE BUTTS 722

7.AT.     BOOKWORM'S DISTANCE 722

7.AU.    NUMBER OF CUTS TO MAKE  N  PIECES 723

7.AV.    HOW LONG TO STRIKE TWELVE? 724

7.AW.       28/7  =  13 724

7.AX.     SUM  =  PRODUCT,  ETC. 725

7.AY.     SUM OF POWERS OF DIGITS 725

7.AZ.     DIVINATION OF A PAIR OF CARDS FROM ITS ROWS 727

7.BA.     CYCLE OF NUMBERS WITH EACH CLOSER TO TEN THAN THE PREVIOUS 729

7.BB.     ITERATED FUNCTIONS OF INTEGERS 729

7.BC.     UNUSUAL DIFFICULTY IN GIVING CHANGE 730

8.       PROBABILITY RECREATIONS 731

8.A.       BUFFON'S NEEDLE PROBLEM 731

8.B.       BIRTHDAY PROBLEM 732

8.D.       ATTEMPTS TO MODIFY BOY‑GIRL RATIO 734

8.E.       ST. PETERSBURG PARADOX 734

8.F.    PROBLEM OF POINTS 735

8.G.       PROBABILITY THAT THREE LENGTHS FORM A TRIANGLE 737

8.H.       PROBABILITY PARADOXES 738

8.H.1.    BERTRAND'S BOX PARADOX 738

8.H.2.    BERTRAND'S CHORD PARADOX 738

8.I.     TAKING THE NEXT TRAIN 738

8.J.    CLOCK PATIENCE OR SOLITAIRE 738

8.K.       SUCKER BETS 739

8.L.       NONTRANSITIVE GAMES 740

9.       LOGICAL RECREATIONS 741

9.A.       ALL CRETANS ARE LIARS, ETC. 741

9.B.       SMITH -- JONES -- ROBINSON PROBLEM 744

9.C.       FORTY UNFAITHFUL WIVES 744

9.D.       SPOTS ON FOREHEADS 744

9.E.       STRANGE FAMILIES 746

GENERAL STUDIES OF KINSHIP RELATIONS 746

DECEASED WIFE'S SISTER, ETC. 747

GENERAL FAMILY RIDDLES 748

9.E.1.    THAT MAN'S FATHER IS MY FATHER'S SON, ETC. 755

9.E.2.    IDENTICAL SIBLINGS WHO ARE NOT TWINS 759

9.F.    THE UNEXPECTED HANGING 759

9.G.       TRUTHTELLERS AND LIARS 760

9.H.       PRISONER'S DILEMMA 763

9.I.     HEMPEL'S RAVEN PARADOX 763

9.J.    FALLEN SIGNPOST 763

9.K.       CARROLL'S BARBER PARADOX 764

10.     PHYSICAL RECREATIONS 764

10.A.     OVERTAKING AND MEETING PROBLEMS 765

10.A.1.      CIRCLING AN ARMY 780

10.A.2.      NUMBER OF BUSES MET 781

10.A.3.      TIMES FROM MEETING TO FINISH GIVEN 782

10.A.4.      THE EARLY COMMUTER 783

10.A.5.      HEAD START PROBLEMS 784

10.A.6.      DOUBLE CROSSING PROBLEMS 785

10.A.7.      TRAINS PASSING 785

10.A.8.      TOO SLOW, TOO FAST 785

10.B.     FLY BETWEEN TRAINS 786

10.C.     LEWIS CARROLL'S MONKEY PROBLEM 787

10.D.     MIRROR PROBLEMS 789

10.D.1       MIRROR REVERSAL PARADOX 789

10.D.2.      OTHER MIRROR PROBLEMS 790

10.D.3       MAGIC MIRRORS 790

10.E.      WHEEL PARADOXES 791

10.E.1.      ARISTOTLE'S WHEEL PARADOX 791

10.E.2.      ONE WHEEL ROLLING AROUND ANOTHER 792

10.E.3.      HUNTER AND SQUIRREL 792

10.E.4.      RAILWAY WHEELS PARADOX 793

10.F.      FLOATING BODY PROBLEMS 794

10.H.     SNAIL CLIMBING OUT OF WELL 797

10.I.      LIMITED MEANS OF TRANSPORT -- TWO MEN AND A BIKE, ETC. 802

10.J.      RESISTOR NETWORKS 803

10.K.     PROBLEM OF THE DATE LINE 803

10.L.     FALLING DOWN A HOLE THROUGH THE EARTH 806

10.M.    CELTS  =  RATTLEBACKS 808

10.M.1.     TIPPEE TOPS 809

10.N.     SHIP'S LADDER IN RISING TIDE 810

10.O.     ERRONEOUS AVERAGING OF VELOCITIES 810

10.P.      FALSE BALANCE 811

10.Q.     PUSH A BICYCLE PEDAL 812

10.R.     CLOCK HAND PROBLEMS 812

10.S.      WALKING IN THE RAIN 815

10.T.      CENTRIFUGAL PUZZLES 816

10.U.     SHORTEST ROUTE VIA A WALL, ETC. 816

10.V.     PICK UP PUZZLES  =  PLUCK IT 817

10.W.    PUZZLE VESSELS 817

10.X.     HOW FAR DOES A PHONOGRAPH NEEDLE TRAVEL? 823

10.Y.     DOUBLE CONE ROLLING UPHILL 824

10.Z.      THE WOBBLER 824

10.AA.      NON-REGULAR DICE 825

10.AB.      BICYCLE TRACK PROBLEMS. 828

10.AC.      ROBERVAL'S BALANCE. 830

10.AD.      POUND OF FEATHERS. 830

10.AE.      JUGGLING OVER A BRIDGE 830

11.     TOPOLOGICAL RECREATIONS 831

11.A.     SCISSORS ON STRING 831

11.B.     TWO PEOPLE JOINED BY ROPES AT WRISTS 832

11.C.     TWO BALLS ON STRING THROUGH LEATHER HOLE AND STRAP  =  CHERRIES PUZZLE 833

11.D.     SOLOMON'S SEAL 835

11.E.      LOYD'S PENCIL PUZZLE 837

11.F.      THE IMPERIAL SCALE 838

11.G.     TRICK PURSES 838

11.H.     REMOVING WAISTCOAT WITHOUT REMOVING COAT 839

11.H.1.      REMOVING LOOP FROM ARM 840

11.I.      HEART AND BALL PUZZLE AND OTHER LOOP PUZZLES 840

11.J.      MÖBIUS STRIP 843

11.K.     WIRE PUZZLES 846

11.K.1.      RING AND SPRING PUZZLE 848

11.K.2.      STRING AND SPRING PUZZLE 849

11.K.3.      MAGIC CHAIN  =  TUMBLE RINGS 849

11.K.4.      PUZZLE RINGS 849

11.K.5.      RING MAZES 850

11.K.6.      INTERLOCKED NAILS, HOOKS, HORNS, ETC. 850

11.K.7.      HORSESHOES PUZZLE 851

11.K.8.      THE CAUGHT HEART 851

11.L.     JACOB'S LADDER AND OTHER HINGING DEVICES 852

11.M.    PUZZLE BOXES 853

11.N.     THREE KNIVES MAKE A SUPPORT 855

11.O.     BORROMEAN RINGS 857

11.P.      THE LONELY MONK 859

11.Q.     TURNING AN INNER TUBE INSIDE OUT 859

11.R.     STRING FIGURES 860

11.S.      PUZZLE KNIVES 861

 

 

                              INTRODUCTION

 

                    NATURE OF THIS WORK

 

          Recreational mathematics is as old as mathematics itself.  Recreational problems already occur in the oldest extant sources -- the Rhind Papyrus and Old Babylonian tablets.  The Rhind Papyrus has an example of a purely recreational problem -- Problem 79 is like the "As I was going to St. Ives" nursery rhyme.  The Babylonians give fairly standard practical problems with a recreational context -- a man knows the area plus the difference of the length and width of his field, a measurement which no surveyor would ever make!  There is even some prehistoric mathematics which could not have been practical -- numerous 'carved stone balls' have been found in eastern Scotland, dating from the Neolithic period and they include rounded forms of all the regular polyhedra and some less regular ones.  Since these early times, recreations have been a feature of mathematics, both as pure recreations and as pedagogic tools.  In this work, I use recreational in a fairly broad sense, but I tend to omit the more straightforward problems and concentrate on those which 'stimulate the curiosity' (as Montucla says). 

          In addition, recreational mathematics is certainly as diffuse as mathematics.  Every main culture and many minor ones have contributed to the history.  A glance at the Common References below, or at almost any topic in the text, will reveal the diversity of sources which are relevant to this study.  Much information arises from material outside the purview of the ordinary historian of mathematics -- e.g.  patents;  articles in newspapers, popular magazines and minor journals;  instruction leaflets;  actual artifacts  and even oral tradition.

          Consequently, it is very difficult to determine the history of any recreational topic and the history given in popular books is often extremely dubious or even simply fanciful.  For example, Nim, Tangrams, and Magic Squares are often traced back to China of about  2000 BC.  The oldest known reference to Nim is in America in 1903.  Tangrams appear in China and Europe at essentially the same time, about 1800, though there are related puzzles in 18C Japan and in the Hellenistic world.  Magic Squares seem to be genuinely a Chinese invention, but go back to perhaps a few centuries BC and are not clearly described until about 80AD.  Because of the lack of a history of the field, results are frequently rediscovered.

          When I began this bibliography in 1982, I had the the idea of producing a book (or books) of the original sources, translated into English, so people could read the original material.  This bibliography began as the table of contents of such a book.  I thought that this would be an easy project, but it has become increasingly apparent that the history of most recreations is hardly known.  I have recently realised that mathematical recreations are really the folklore of mathematics and that the historical problems are similar to those of folklore.  One might even say that mathematical recreations are the urban myths or the jokes or the campfire stories of mathematics.  Consequently I decided that an annotated bibliography was the first necessity to make the history clearer.  This bibliography alone has grown into a book, something like Dickson's History of the Theory of Numbers.  Like that work, the present work divides the subject into a number of topics and treats them chronologically. 

          I have printed six preliminary editions of this work, with slightly varying titles.  The first version of 4 Jul 1986 had 224 topics and was spaced out so entries would not be spread over two pages and to give room for page numbers.  This stretched the text from 110pp to 129pp and was printed for the Strens Memorial Conference at the Univ. of Calgary in Jul/Aug 1986.  I no longer worry about page breaks.  The following editions had:  250 topics on 152 pages;  290 topics on 192 pages;  307 topics on 223 pages;  357 topics on 311 pages  and 392 topics on 456 pages.  The seventh edition was never printed, but was a continually changing computer file.  It had about 419 topics (as of 20 Oct 95) and 587 pages, as of 20 Oct 1995.  I then carried out the conversion to proportional spacing and this reduced the total length from 587 to 488 pages, a reduction of 16.87% which is conveniently estimated as 1/6.  This reduction was fairly consistent throughout the conversion process. 

          This eighth edition is being prepared for the Gathering for Gardner 6 in March 2004.  The text is 818 pages as of 18 Mar 2004.  There are about 457 topics as of 18 Mar 2004.

          A fuller description of this project in 1984-1985 is given in my article Some early sources in recreational mathematics, in: C. Hay et al., eds.; Mathematics from Manuscript to Print; Oxford Univ. Press, 1988, pp. 195‑208.  A more recent description is in my article: Recreational mathematics; in: Encyclopedia of the History and Philosophy of the Mathematical Sciences; ed. by I. Grattan-Guinness; Routledge & Kegan Paul, 1993; pp. 1568-1575.

          Below I compare this work with Dickson and similar works and discuss the coverage of this work.

 

                    SIMILAR WORKS

 

          As already mentioned, the work which the present most resembles is Dickson's History of the Theory of Numbers.

 

                    The history of science can be made entirely impartial, and perhaps that is what it should be, by merely recording who did what, and leaving all "evaluations" to those who like them.  To my knowledge there is only one history of a scientific subject (Dickson's, of the Theory of Numbers) which has been written in this coldblooded, scientific way.  The complete success of that unique example -- admitted by all who ever have occasion to use such a history in their work -- seems to indicate that historians who draw morals should have their own morals drawn.

 

                    E. T. Bell.  The Search for Truth.  George Allen & Unwin, London, 1935, p. 131.

 

          Dickson attempted to be exhaustive and certainly is pretty much so.  Since his time, many older sources have been published, but their number-theoretic content is limited and most of Dickson's topics do not go back that far, so it remains the authoritative work in its field.

 

          The best previous book covering the history of recreational mathematics is the second edition of Wilhelm Ahrens's Mathematische Unterhaltungen und Spiele in two volumes.  Although it is a book on recreations, it includes extensive histories of most of the topics covered, far more than in any other recreational book.  He also gives a good index and a bibliography of 762 items, often with some bibliographical notes.  I will indicate the appropriate pages at the beginning of any topic that Ahrens covers.  This has been out of print for many years but Teubner has some plans to reissue it.

          Another similar book is the 4th edition of J. Tropfke's Geschichte der Elementarmathematik, revised by Vogel, Reich and Gericke.  This is quite exhaustive, but is concerned with older problems and sources.  It presents the material on a topic as a history with references to the sources, but it doesn't detail what is in each of the sources.  Sadly, only one volume, on arithmetic and algebra, appeared before Vogel's death.  A second volume, on geometry, is being prepared.  For any topic covered in Tropfke, it should be consulted for further references to early material which I have not seen, particularly material not available in any western language.  I cite the appropriate pages of Tropfke at the beginning of any topic covered by Tropfke.

          Another book in the field is W. L. Schaaf's Bibliography of Recreational Mathematics, in four volumes.  This is a quite exhaustive bibliography of recent articles, but it is not chronological, is without annotation and is somewhat less classified than the present work.  Nonetheless it is a valuable guide to recent material.

 

          Collecting books on magic has been popular for many years and quite notable collections and bibliographies have been made.  Magic overlaps recreational mathematics, particularly in older books, and I have now added references to items listed in the bibliographies of Christopher, Clarke & Blind, Hall, Heyl, Toole Stott and Volkmann & Tummers -- details of these works are given in the list of Common References below.  There is a notable collection of Harry Price at Senate House, University of London, and a catalogue was printed in 1929 & 1935 -- see HPL in Common References.

          Another related bibliography is Santi's Bibliografia della Enigmistica, which is primarily about word puzzles, riddles, etc., but has some overlap with recreational mathematics -- again see the entry in the list of Common References.  I have not finished working through this.

          Other relevant bibliographies are listed in Section 3.B.

 

                    COVERAGE

 

          In selecting topics, I tend to avoid classical number theory and classical geometry.  These are both pretty well known.  Dickson's History of the Theory of Numbers and Leveque's and Guy's Reviews in Number Theory cover number theory quite well.  I also tend to avoid simple exercises, e.g. in the rule of three, in 'aha' or 'heap' problems, in the Pythagorean theorem (though I have now included 6.BF) or in two linear equations in two unknowns, though these often have fanciful settings which are intended to make them amusing and some of these are included -- see  7.R,  7.X,  7.AX.  I also leave out most divination (or 'think of a number') techniques (but a little is covered in 7.M.4.b) and most arithmetic fallacies.  I also leave out Conway's approach to mathematical games -- this is extensively covered by Winning Ways and Frankel's Bibliography.

          The classification of topics is still ad‑hoc and will eventually get rationalised -- but it is hard to sort things until you know what they are!  At present I have only grouped them under the general headings: Biography, General, History & Bibliography, Games, Combinatorics, Geometry, Arithmetic, Probability, Logic, Physics, Topology.  Even the order of these should be amended.  The General section should be subsumed under the History & Bibliography.  Geometry and Arithmetic need to be subdivided.

          I have recently realised that some general topics are spread over several sections in different parts.  E.g. fallacies are covered in 6.P, 6.R, 6.AD, 6.AW.1, 6.AY, 7.F, 7.Y, 7.Z, 7.AD, 7.AI, 7.AL, 7.AN, most of 8, 10.D, 10.E, 10.O.  Perhaps I will produce an index to such topics.  I try to make appropriate cross-references.

          Some topics are so extensive that I include introductory or classifactory material at the beginning.  I often give a notation for the problems being considered.  I give brief explanations of those problems which are not well known or are not described in the notation or the early references.  There may be a section index.  I have started to include references to comprehensive surveys of a given topic -- these are sometimes given at the beginning.

          Recreational problems are repeated so often that it is impossible to include all their occurrences.  I try to be exhaustive with early material, but once a problem passes into mathematical and general circulation, I only include references which show new aspects of the problem or show how the problem is transmitted in time and/or space.  However, the point at which I start leaving out items may vary with time and generally slowly increases as I learn more about a topic.  I include numerous variants and developments on problems, especially when the actual origin is obscure.

          When I began, I made minimal annotations, often nothing at all.  In rereading sections, particularly when adding more material, I have often added annotations, but I have not done this for all the early entries yet.

          Recently added topics often may exist in standard sources that I have not reread recently, so the references for such topics often have gaps -- I constantly discover that Loyd or Dudeney or Ahrens or Lucas or Fibonacci has covered such a topic but I have forgotten this -- e.g. looking through Dudeney recently, I added about 15 entries.  New sections are often so noted to indicate that they may not be as complete as other sections.

          Some of the sources cited are lengthy and I originally added notes as to which parts might be usable in a book of readings -- these notes have now been mostly deleted, but I may have missed a few.

 

                    STATUS OF THE PROJECT

 

          I would like to think that I am about 75% of the way through the relevant material.  However, I recently did a rough measurement of the material in my study -- there is about 8 feet of read but unprocessed material and about 35 feet of unread material, not counting several boxes of unread Rubik Cube material and several feet of semi-read material on my desk and table.  I recently bought two bookshelves just to hold unread material.  Perhaps half of this material is relevant to this work.

          In particular, the unread material includes several works of Folkerts and Sesiano on medieval MSS, a substantial amount of photocopies from Schott, Schwenter and Dudeney (400 columns), some 2000 pages of photocopies recently made at Keele, some 500 pages of photocopies from Martin Gardner's files, as well as a number of letters.  Marcel Gillen has made extracts of all US, German and EURO patents and German registered designs on puzzles -- 26 volumes, occupying about two feet on my shelves.  I have recently acquired an almost complete set of Scripta Mathematica (but I have previously read about half of it), Schwenter-Harsdörffer's Deliciæ Physico‑Mathematicae, Schott's Joco-Seriorum and Murray's History of Board Games Other Than Chess.  I have recently acquired the early issues of Eureka, but there are later issues that I have not yet read and they persist in not sending the current copies I have paid for!

          I have not yet seen some of the earlier 19C material which I have seen referred to and I suspect there is much more to be found.  I have examined some 18C & 19C arithmetic and algebra books looking for problem sections -- these are often given the pleasant name of Promiscuous Problems.  There are so many of these that a reference to one of them probably indicates that the problem appears in many other similar books that I have not examined.  My examination is primarily based on those books which I happen to have acquired.  There are a few 15-17C books which I have not yet examined, notably those included at the end of the last paragraph.

          In working on this material, it has become clear that there were two particularly interesting and productive eras in the 19C.  In the fifteen years from 1857, there appeared about a dozen books in the US and the UK: The Magician's Own Book (1857);  Parlour Pastime, by "Uncle George" (1857);  The Sociable (1858);  The Boy's Own Toymaker, by Landells (1858);  The Book of 500 Curious Puzzles (1859);  The Secret Out (1859);  Indoor and Outdoor Games for Boys and Girls (c1859);  The Boy's Own Conjuring Book (1860);  The Illustrated Boy's Own Treasury (1860, but see below);  The Parlor Magician (1863);  The Art of Amusing, by Bellew (1866);  Parlour Pastimes (1868);  Hanky Panky (1872);  Within Doors, by Elliott (1872);  Magic No Mystery (1876), just to name those that I know.  Most of these are of uncertain authorship and went through several editions and versions.  The Magician's Own Book, The Book of 500 Curious Puzzles, The Secret Out, The Sociable, The Parlor Magician, Hanky Panky, and Magic No Mystery seem to be by the same author(s).  I have recently had a chance to look at a number of previously unseen versions at Sotheby's and at Edward Hordern's and I find that sometimes two editions of the same title are essentially completely different!  This is particularly true for US and UK editions.  Many of the later UK editions say 'By the author of Magician's Own Book etc., translated and edited by W. H. Cremer Jr.'  From the TPs, it appears that they were written by Wiljalba Frikell (1818‑1903) and then translated into English.  However, BMC and NUC generally attribute the earlier US editions to George Arnold (1834-1865), and some catalogue entries explicitly say the Frikell versions are later editions, so it may be that Frikell produced later editions in some other language (French or German ??) and these were translated by Cremer.  On the other hand, the UK ed of The Secret Out says it is based on Le Magicien des Salons.  This is probably Le Magicien des Salons ou le Diable Couleur de Rose, for which I have several references, with different authors! --  J. M. Gassier, 1814;  M. [Louis Apollinarie Christien Emmanuel] Comte, 1829;  Richard (pseud. of A. O. Delarue), 1857 and earlier.  There was a German translation of this.  Some of these are at HPL but ??NYS.  Items with similar names are:  Le Magicien de Société, Delarue, Paris, c1860?  and  Le Manuel des Sorciers (various Paris editions from 178?-1825, cf in Common References).  It seems that this era was inspired by these earlier French books.  To add to the confusion, an advertisement for the UK ed. of Magician's Own Book (1871?) says it is translated from Le Magicien des Salons which has long been a standard in France and Germany.  Toole Stott opines that Frikell had nothing to do with these books -- as a celebrated conjuror of the times, his name was simply attached to the books.  Toole Stott also doubts whether Le Magicien des Salons exists -- but it now seems pretty clear that it does, though it may not have been the direct source for any of these works, but see below.

          Christopher 242 cites the following article on this series.

          Charles L. Rulfs.  Origins of some conjuring works.  Magicol 24 (May 1971) 3-5.  He discusses the various books, saying that Cremer essentially pirated the Dick & Fitzgerald productions.  He says The Magician's Own Book draws on Wyman's Handbook (1850, ??NYS), Endless Amusement, Parlour Magic (by W. Clarke?, 1830s, ??NYS), Brewster's Natural Magic (??NYS).  He says The Secret Out is largely taken, illustrations and all, from Blismon de Douai's Manuel du Magicien (1849, ??NYS) and Richard & Delion's Magicien des salons ou le diable couleur de rose (1857 and earlier, ??NYS).

          Christopher 622 says Harold Adrian Smith [Dick and Fitzgerald Publishers; Books at Brown 34 (1987) 108-114] has studied this series and concludes that Williams was the author of Magician's Own Book, assisted by Wyman.  Actually Smith simply asserts: "The book was undoubtedly [sic] written by H. L. Williams, a "hack writer" of the period, assisted by John Wyman in the technical details."  He gives no explanation for his assertion.  He later says he doubts whether Cremer ever wrote anything.  He suggests The Secret Out book is taken from DeLion.  He states that The Boy's Own Conjuring Book is a London pirate edition.

          Several of the other items are anonymous and there was a tremendous amount of copying going on -- problems are often reproduced verbatim with the same diagram or sometimes with minor changes.  In some cases, the same error is repeated in five different books!  I have just discovered some earlier appearances of the same material in The Family Friend, a periodical which ran in six series from 1849 to 1921 and which I have not yet tracked down further.  However, vol. 1-3 of 1849‑1850 and the volume for Jul‑Dec 1859 contain a number of the problems which appear repeatedly and identically in the above cited books.  Toole Stott 407 is an edition of The Illustrated Boy's Own Treasury of c1847 but the BM copy was destroyed in the war and the other two copies cited are in the US.  If this date is correct, then this book is a forerunner of all the others and a major connection between Boy's Own Book and Magician's Own Book.  I would be most grateful to anyone who can help sort out this material -- e.g. with photocopies of these or similar books or magazines.

          The other interesting era was about 1900.  In English, this was largely created or inspired by Sam Loyd and Henry Dudeney.  Much of this material first appeared in magazines and newspapers.  I have seen much less than half of Loyd's and Dudeney's work and very little of similar earlier material (but see below).  Consequently problems due to Loyd or Dudeney may seem to first appear in the works of Ball (1892, et seq.), Hoffmann (1893) and Pearson (1907).  Further examination of Loyd's and Dudeney's material will be needed to clarify the origin and development of many problems.  Though both started puzzle columns about 1896, they must have been producing material for a decade or more previously which does not seem to be known.  I have just obtained photocopies of 401 columns by Dudeney in the Weekly Dispatch of 1897-1903, but have not had time to study them.  Will Shortz and Angela Newing have been studying Loyd and Dudeney respectively and turning up their material.

          The works of Lucas (1882‑1895), Schubert (1890s) and Ahrens (1900‑1918) were the main items on the Continent and they interacted with the English language writers.  Ahrens was the most historical of these and his book is one of the foundations of the present work.  All of these also wrote in newspapers and magazines and I have not seen all their material.

 

          I would be happy to hear from anyone with ideas or suggestions for this bibliography.  I would be delighted to hear from anyone who can locate missing information or who can provide copies of awkward material.  I am particularly short of information about recreations in the Arabic period.  I prepared a separate file, 'Queries and Problems in the History of Recreational Mathematics', which is about 23 pages, and has recently been updated.  I have also prepared three smaller letters of queries about Middle Eastern, Oriental and Russian sources and these are generally more up-to-date.

 

                    TECHNICAL NOTES

 

          I have prepared a CD containing this and much else of my material.  I divided Sources into four files when I used floppy discs as it was too big to fit on one disc, and I have not yet changed this.  The files are:  1: Introductory material and list of abbreviations/references;  2: Sections 1 - 6;  3: Section 7;  4: Sections 8 - 11.  It is convenient to have the first file separate from the main material, but I might combine the other three files.  (I have tried to send it by email in the past, but this document is very large (currently c4.1MB and the Word version will be longer) and most people who requested it by email found that it overflowed their mailbox and created chaos in their system -- this situation has changed a bit with larger memories and improved transmission speeds.)

          This file started on a DEC-10, then was transferred to a VAX.  It is now on my PC using Script Professional, the development of LocoScript on the Amstrad.  Even in its earliest forms, this provided an easy and comprehensive set of diacritical marks, which are still not all available nor easy to use in WordPerfect or Word (except perhaps by using macros and/or overstriking??).  It also provides multiple cut and paste buffers and easy formatting, though I have learned how to overcome these deficiencies in Word.

          Script provides an ASCII output, but this uses IBM extended ASCII which has 8-bit codes.  Not all computers will accept or print such characters and sometimes they are converted into printer control codes causing considerable confusion.  I have a program that converts these codes to 7-bits -- e.g. accents and umlauts are removed.  However, ASCII loses a great deal of the information, such as sub- and superscripts, so this is not a terribly useful format. 

          Script also provides WordStar and "Revisable-Form-Text DCA" output, but neither of these seems to be very successful (DCA is better than WordStar).  Script later added a WordPerfect exporting facility.  This works well, though some (fairly rare) characters and diacritical marks are lost and the output requires some reformatting.  (Nob Yoshigahara reports that Japanese WordPerfect turns all the extended ASCII characters into Kanji characters!) 

          Reading the WordPerfect output in Word (you may need to install this facility) gives a good approximation to my text, but in  Courier 10pt.  Selecting All  and changing to  Times New Roman 12pt  gives an better approximation.  (Some files use a smaller font of 10pt and I may have done some into 9pt.)  You have to change this in the Header separately, using  View  Header and Footer.  The page layout is awkward as my page numbering header gets put into the text, leaving a large gap at the top.  I go into  Page Setup  and set the  Paper Size  to A4 and the  Top, Bottom and Header Margins  to 15mm and the  Left and Right Margins  to 25mm.  (It has taken me some time to work this out and some earlier files may have other settings.)  However, I find that lines are a bit too close together and underlines and some diacritical marks are lost, so one needs to also go into  Format  Paragraph  Spacing -- Line Spacing  and choose  At least  and  12pt  (or 10pt).  I use hanging indentation in most of the main material and this feature is not preserved in this conversion.  By selecting a relevant section and going into  Format  Paragraph  Indentation -- Special  and selecting  Hanging,  it should automatically select  10.6mm  which corresponds to my automatic spacing of five characters in 12pt.  Further, I use second level hanging indentation in quite a number of places.  You need to create a style which is the basic style with the left hand margin at  10.6mm (or 10 or 11 mm).  When second level indenting is needed, select the desired section and apply this style to it. 

          However, this still leaves some problems.  I use em dashes a bit, i.e.  –,  which gets converted into an underline,  _.  In Word, this is obtained by use of  CTRL  and the  -  sign on the numeric keypad.  One can use the find and replace feature, EXCEPT that a number of other characters are also converted into underlines.  In particular, Cyrillic characters are all converted into underlines.  This is not insuperable as I always(?) give a transliteration of Cyrillic (using the current Mathematical Reviews system) and one can reconstruct the original Cyrillic from it.  I notice that the Cyrillic characters are larger than roman characters and hence may overlap.  One can amend this by selecting the Cyrillic text and going into  Format  Font  Character Spacing  Spacing  and choosing  Expanded  By  2 pt (or thereabout).  But a number of characters with unusual diacritical marks are also converted to underlines or converted to the unmarked character and not all of these are available in Word.  E.g.  ĭ,  which is the transliteration of  й  becomes just  i.  I am slowly forming a Word file containing the Word versions of entries having the Cyrillic or other odd characters, and I will include this file on my CD, named  CYRILLIC.DOC.  For diacritical marks not supported by Word, I use an approximation and/or an explanation. 

          It is very tedious to convert the underlines back to em dashes, so I will convert every em dash to a double hyphen --. 

          Finally, I have made a number of diagrams by simple typing without proportional spacing and Word does not permit changing font spacing in mid-line and ignores spaces before a right-alignment instruction.  The latter problem can be overcome by using hard spaces and the former problem is less of a problem, and I think it can be overcome.

 

          Later versions of Script support Hewlett-Packard DeskJets and I am now on my second generation of these, so the 7th and future editions will be better printed (if they ever are!).  However, this required considerable reformatting as the text looks best in proportional spacing (PS) and I found I had to check every table and every mathematical formula and diagram.  Also, to set off letters used as mathematical symbols within text, I find PS requires two spaces on each side of the letter -- i.e.  I refer to  x  rather than to x.  (I find this easier to do than to convert to italics.)  I also sometimes set off numbers with two spaces, though I wasn't consistent in doing this at the beginning of my reformatting.  The conversion to proportional spacing reduced the total length from  587  to  488  pages, a reduction of  16.87%  which is conveniently estimated as  1/6.  The percentage of reduction was fairly consistent throughout the conversion process.

          The printing of Greek characters went amiss in the second part of the 6th Preliminary Edition, apparently due to the printer setting having been changed without my noticing -- this happens if an odd character gets sent to the printer, usually in DOS when trying to use or print a corrupted file, and there is no indication of it.  I was never able to reproduce the effect!

          The conversion to (Loco)Script provided many improved features compared to my earlier DEC versions.  I am using an A4 page  (8¼  by  11⅔  inches) rather than an  8½  by  11 inch page, which gives  60  lines of text per page, four more or  7%  more than when using the DEC or VAX.

 

                    NEW SECTIONS IN THIS EDITION.

 

[SIXTH EDITION: 1: Fibonacci, 1: Montucla;  3.B;  4.A.1.a, 4.B.9, 4.B.10, 4.B.11, 4.B.12;  5.R.1.a, 5.W.1, 5.AA, 5.AB;  6.AS.1.b, 6.AS.2.a, 6.AS.5, 6.AW.4, 6.BP, 6.BQ, 6.BR;  7.I.1, 7.Y.2, 7.AY, 7.AZ; 7.BA;  8.I, 8.J;  9.E.2, 9.K;  10.A.4, 10.A.5, 10.U, 10.V, 10.W;  11.K.6, 11.K.7, 11.K.8.]

 

In the last edition, I had 8.K instead of 8.J in the list of New Sections and in the Contents.

 

1: Pacioli, Carroll, Perelman;  4.B.13, 4.B.14, 4.B.15;  5.B.2, 5.H.3 (the previous 5.H.3 has been renumbered 5.H.4), 5.K.3, 5.R.1.b, 5.X.4, 5.AC, 5.AD, 5.AE, 5.AF, 5.AG.1, 5.AG.2;  6.AJ.4, 6.AJ.5, 6.AS.3.a, 6.AT.8, 6.AT.9, 6.AY.2, 6.BF.4, 6.BF.5, 6.BS, 6.BT, 6.BU, 6.BV, 6.BW;  7.H.6, 7.H.7 (formerly part of 7.H.5), 7.M.4.a, 7.M.4.b, 7.M.6, 7.R.4, 7.AC.3.a, 7.AC.7, 7.AH.1, 7.AJ.1, 7.BB, 7.BC;  8.K, 8.L;  10.A.6, 10.A.7, 10.A.8, 10.D has become 10.D.1, 10.D.2, 10.D.3, 10.E.4, 10.X, 10.Y, 10.Z, 10.AA, 10.AB, 10.AC, 10.AD, 10.AE;  11.N, 11.O, 11.P, 11.Q, 11.R, 11.S.  (65 new sections)

 

 

                    ACKNOWLEDGMENTS

 

          I am immensely indebted to many mathematicians, historians, puzzlers, bookdealers and others who have studied particular topics, as will be apparent.

          I have had assistance from so many sources that I have probably forgotten some, but I would like to give thanks here to the following, and beg forgiveness from anyone inadvertently omitted -- if you remind me, I will make amendment.  In some cases, I simply haven't got to your letter yet!  Also I have had letters from people whose only identification is an undecipherable signature and phone messages from people whose name and phone number are unintelligible. 

          Sadly, a few of these have died since I corresponded with them and I have indicated those known to me with †.

 

André Allard,  Eric J. Aiton†,  Sue Andrew,  Hugh Ap Simon,  Gino Arrighi,  Marcia Ascher,  Mohammad Bagheri,  Banca Commerciale Italiana,  Gerd Baron,  Chris Base,  Rainier [Ray] Bathke,  John Beasley,  Michael Behrend,  Jörg Bewersdorff,  Norman L. Biggs,  C. [Chris] J. Bouwkamp,  Jean Brette,  John Brillhart,  Paul J. Campbell,  Cassa di Risparmio di Firenze,  Henry Cattan,  Marianna Clark,  Stewart Coffin,  Alan & Philippa Collins,  John H. Conway,  H. S. M. Coxeter,  James Dalgety,  Ann E. L. Davis,  Yvonne Dold,  Underwood Dudley,  Anthony W. F. Edwards,  John Ergatoudis,  John Fauvel†,  Sandro Ferace,  Judith V. Field,  Irving Finkel,  Graham Flegg,  Menso Folkerts,  David Fowler,  Aviezri S. Fraenkel,  Raffaella Franci,  Gregory N. Frederickson,  Michael Freude,  Walter W. Funkenbusch,  Nora Gädeke,  Martin Gardner,  Marcel Gillen,  Leonard J. Gordon,  Ron Gow,  Ivor Grattan‑Guinness,  Christine Insley Green,  Jennifer Greenleaves (Manco),  Tom Greeves,  H. [Rik] J. M. van Grol,  Branko Grünbaum,  Richard K. Guy,  John Hadley,  Peter Hajek,  Diana Hall,  Joan Hammontree,  Anton Hanegraaf†,  Martin Hansen,  Jacques Haubrich,  Cynthia Hay,  Takao Hayashi,  Robert L. Helmbold,  Hanno Hentrich,  Richard I. Hess,  Christopher Holtom,  Edward Hordern†,  Peter Hosek,  Konrad Jacobs,  Anatoli Kalinin,  Bill Kalush,  Michael Keller,  Edward S. Kennedy,  Sarah Key (The Haunted Bookshop),  Eberhard Knobloch,  Don Knuth,  Bob Koeppel,  Joseph D. E. Konhauser,  David E. Kullman,  Mogens Esrom Larsen,  Jim Lavis (Doxa (Oxford)),  John  Leech†,  Elisabeth Lefevre,  C. Legel,  Derrick [Dick] H. Lehmer†,  Emma Lehmer,  Leisure Dynamics,  Hendrik W. Lenstra,  Alan L. Mackay,  Andrzej Makowski,  John Malkevitch,  Giovanni Manco,  Tatiana Matveeva,  Ann Maury,  Max Maven,  Jim McArdle,  Patricia McCulloch,  Peter McMullen,  Leroy F. Meyers†,  D. P. Miles,  Marvin Miller,  Nobuo Miura,  William O. J. Moser,  Barbara Moss,  Angela Newing,  Jennie Newman,  Tom and Greta O'Beirne††,  Owen O'Shea,  Parker Brothers,  Alan Parr,  Jean J. Pedersen,  Luigi Pepe,  William Poundstone,  Helen Powlesland,  Oliver Pretzel,  Walter Purkert,  Robert A. Rankin†,  Eleanor Robson,  David J. A. Ross,  Lee Sallows,  Christopher Sansbury,  Sol Saul,  William L. Schaaf,  Doris Schattschneider,  Jaap Scherphuis,  Heribert Schmitz,  Š. Schwabik,  Eileen Scott†,  Al Seckel,  Jacques Sesiano,  Claude E. Shannon†,  John Sheehan,  A. Sherratt,  Will Shortz,  Kripa Shankar Shukla,  George L. Sicherman,  Deborah Singmaster,  Man‑Kit Siu,  Gerald [Jerry] K. Slocum,  Cedric A. B. Smith† (and Sue Povey & Jim Mallet at the Galton Laboratory for letting me have some of Cedric's books),  Jurgen Stigter,  Arthur H. Stone,  Mel Stover†,  Michael Stueben,  Shigeo Takagi†,  Michael Tanoff,  Gary J. Tee,  Andrew Topsfield,  George Tyson†,  Dario Uri,  Warren Van Egmond,  Carlo Viola,  Kurt Vogel†,  Anthony Watkinson,  Chris Weeks,  Maurice Wilkes,  John Winterbottom,  John Withers,  Nob. Yoshigahara,  Claudia Zaslavsky.

 

 

          I would also like to thank the following libraries and museums which I have used:

 

University of Aberdeen;  University of Bristol;   Buckleys Shop Museum, Battle, East Sussex;  University of Calgary;  University of Cambridge;  Marsh's Library, Dublin; 

 

FLORENCE:

          Biblioteca Nazionale;  Biblioteca Riccardiana;

 

University of Keele -- The Turner Collection(†) and its librarian Martin Phillips;

Karl‑Marx‑Universität, Leipzig: Universität Bibliothek and Sektion Mathematik Bibliothek,

          especially Frau Letzel at the latter;

 

LONDON:

          Birkbeck College;  British Library (at Bloomsbury and then at St. Pancras; also at Colindale);  The London Library;  School of Oriental and African Studies, especially Miss Y. Yasumara, the Art Librarian;  Senate House, particularly the Harry Price Library;  South Bank University;  Southwark Public Library;  University College London, especially the Graves Collection and the Rare Book Librarians Jill Furlong, Susan Stead and their staff;  Warburg Institute; 

 

MUNICH:

          Deutsches Museum; Institut für Geschichte der Naturwissenschaften;

 

NEW YORK:

          Brooklyn Public Library;  City College of New York;  Columbia University;

 

Newark Public Library, Newark, New Jersey;

University of Newcastle upon Tyne -- The Wallis Collection and its librarian Lesley Gordon;

 

OXFORD:

          Ashmolean Museum;  The Bodleian Library;  Museum of the History of Science, and its librarian Tony Simcock; 

 

University of Reading;  University of St. Andrews; 

 

SIENA:

          Biblioteca Comunale degli Intronati;  Dipartimento di Matematica, Università di Siena;

 

University of Southampton;  Mathematical Institute, Warsaw.

 

 

          I would like especially to thank the following.

 

          Interlibrary Loans (especially Brenda Spooner) at South Bank University and the British Library Lending Division for obtaining many strange items for me.

          Richard Guy, Bill Sands and the Strens bequest for a most useful week at the Strens/Guy Collection at Calgary in early 1986 and for organizing the Strens Memorial Meeting in summer 1986 and for printing the first preliminary edition of these Sources.

          Gerd Lassner, Uwe Quasthoff and the Naturwissenschaftlich‑Theoretisches Zentrum of the Karl‑Marx‑Universität, for a very useful visit to Leipzig in 1988.

          South Bank University Computer Centre for the computer resources for the early stages of this project, and especially Ann Keen for finding this file when it was lost.

          My School for printing these preliminary editions.

          Martin Gardner for kindly allowing me to excavate through his library and files.

          James Dalgety, Edward Hordern, Bill Kalush, Chris Lewin, Tom Rodgers and Will Shortz for allowing me to rummage through their libraries.

          John Beasley, Edward Hordern, Bill Kalush, Will Shortz and Jerry Slocum for numerous photocopies and copies from their collections.

          Menso Folkerts, Richard Lorch, Michael Segre and the Institut für Geschichte der Naturwissenschaft, Munich, for a most useful visit in Sep 1994 and for producing a copy of Catel.

          Raffaella Franci and the Dipartimento di Matematica and the Centro Studi della Matematica Medioevale at Università di Siena for a most useful visit in Sep 1994.

          Takao Hayashi for much material from Japan and India.

          My wife for organizing a joint trip to Newcastle in Sep 1997 where I made use of the Wallis Collection.

 

          Finally, I would like to thank a large number of publishers, distributors, bookdealers and even authors who have provided copies of the books and documents upon which much of this work is based.  Bookdealers have often let me examine books in their shops.  Their help is greatly appreciated.  There are too many of these to record here, but I would like to mention Fred Whitehart (†1999), England's leading dealer in secondhand scientific books for many years who had a real interest in mathematics.

 

 

                             

 

                              ABBREVIATIONS

 

                    DIACRITICAL MARKS AND NOTATION

 

          Before converting to LocoScript, I used various conventions, given below, to represent diacritical marks.  Each symbol (except  ') occurred after the letter it referred to.  I have now converted these and all mathematical conventions into correct symbols, so far as possible, but I may have missed some, so I am keeping this information for the present.

          Common entries using such marks are given later in this section and only the abbreviated or simplified form is used later -- e.g. I use Problemes for Bachet's work rather than Problèmes.  (Though this may change??)

          Initially, I did not record all diacritical marks, so some may be missing though I have checked almost all items.  I may omit diacritical marks which are very peculiar. 

          Transliterations of Arabic, Sanskrit, Chinese, etc. are often given in very different forms.  See Smith, History, vol. 1, pp. xvii-xxii for a discussion of the problems.  The use of  ^  and  ˉ  seems interchangeable and I have used  ^  when different versions use both  ^  and  ˉ , except when quoting a title or passage when I use the author's form.  [Smith, following Suter, uses  ^  for Arabic, but uses  ˉ  for Indian.  Murray uses  ˉ  for both.  Wieber uses  ˉ  for Arabic.  Van der Linde uses  ´  for Arabic.  Datta & Singh use  ^  for Indian.]

          There are two breathing marks in Arabic -- ayn    and alif/hamzah    -- but originally I didn't have two forms easily available, so both were represented by  '.  I have now converted almost all of these to    and  ’.  These don't seem to be as distinct in the printing as on my screen.

          French practice in accenting capitals is variable and titles are often in capitals, so expected marks may be missing.  Also, older printing may differ from modern usage -- e.g. I have seen: Liège and Liége;  Problèmes, Problêmes and Problémes.  When available, I have transcribed the material as printed without trying to insert marks, but many places insert the marks according to modern French spelling.

          Greek and Cyrillic titles are now given in the original with an English transliteration (using the Amer. Math. Soc. transliteration for Cyrillic).

          I usually ignore the older usage of  v  for  u  and  i  for  j,  so that I give  mathematiqve  as  mathematique  and  xiij  as  xiii.

          I used  a1, a2, ..., ai,  etc. for subscripted variables, though I also sometimes used  a(1), a(2), ..., a(i),  etc.  Superscripts or exponents were indicated by use of  ^,  e.g.  2^3  is  8.  These have been converted to ordinary sub- and superscript usage, but  ^  may be used when the superscript is complicated -- e.g. for  2^ai  or  9^(99). 

          Greek letters were generally spelled out in capitals or marked with square brackets, e.g. PI, [pi], PHI, but these have probably all been converted. 

          My word processor does not produce binomial coefficients easily, so I use  BC(n, k)  for  n!/k!(n‑k)! 

          Many problems have solutions which are sets of fractions with the same denominator and I abbreviate  a/z,  b/z,  c/z  as  (a, b, c)/z.  Notations for particular problems are explained at the beginning of the topic. 

          Rather than attempting to italicise letters used as symbols, I generally set them off by double-spaces on each side -- see examples above.  Other mathematical notations may be improvised as necessary and should be obvious.

          Recall that the symbols below occurred after the letter they referred to, except for  ' .

 

"         denoted umlaut or diaeresis in general, e.g.:  ä, ë, ï, ö, ü.

/         was used after a letter for accent acute,  ́, after  l  for  ł  in Polish, and after  o  for  ø  in Scandinavian.

\         denoted accent grave,  ̀.

^        denoted the circumflex,  ^,  in Czech, etc.; the overbar (macron)  ˉ   or  ^  for a long vowel in Sanskrit, Hindu, etc.; and the overbar used to indicate omission in medieval MSS.

@       denoted the cedilla (French  ç  and Arabic  ş)  and the ogonek or Polish hook (Polish  ą).

.         denoted the underdot in  ḥ, ḳ, ṇ, ṛ, ṣ,  ,  in Sanskrit, Hindu, Arabic.  These are sometimes written with a following  h  -- e.g.  k  may also be written  kh  and I may sometimes have used this.  (It is difficult to search for  . , etc., so not all of these may be converted.)  This mark vanishes when converted to WordPerfect.

*        denoted the overdot for  ġ, , ṅ,  in Sanskrit, Hindu, Arabic.  This vanishes over  m  and  n  in WordPerfect.

~        denoted the Spanish tilde  ~  and the caron or hachek  ˇ,  in  ğ, š.  The breve is a curved version,  ˘,  of the same symbol and is essentially indistinguishable from the caron.  It occurs in Russian  й,  which is translitereated as  ĭ.

_        denoted the underbar in  , j, ṯ (I cannot find a  j  with an underbar in Arial).  This mark vanishes in WordPerfect.

'         denotes breathing marks in Arabic, etc.  There are actually two forms of this -- ayn    and alif/hamzah    -- but I didn't have two forms easily available and originally entered both as apostrophe  '  .  These normally occur between letters and I placed the  '  in the same space.  I have converted most of these.

 

          Commonly occurring words with diacritical marks are: Académie, arithmétique, bibliothèque, Birkhäuser, café, carré, école, Erdös, für, géomètre, géométrie, Göttingen, Hanoï -- in French only, ‑ième, littéraire, mathématique, mémoire, ménage, misère, Möbius, moiré, numérique, Pétersbourg, probabilités, problème (I have seen problêmes??), Rätsel, récréation, Sändig, siècle, société, Thébault, théorie, über, umfüllung.

 

          I have used  ??  to indicate uncertainty and points where further work needs to be done.  The following symbols after  ??  indicate the action to be done.

          *                  check for diacritical marks, etc.

          NX               no Xerox or other copy

          NYS             not yet seen

          NYR  not yet read

          o/o               on order

          SP                check spelling

          Other comments may be given.

 

 

          ABBREVIATIONS OF JOURNALS AND SERIES. 

See:  AMM,  CFF,  CM,  CMJ,  Family Friend,  G&P,  G&PJ,  HM,  JRM,  MG,  MiS,  MM,  MS, MTg,  MTr,  M500,  OPM,  RMM,  SA,  SM,  SSM  in Common References below.

 

          ABBREVIATIONS OF PUBLISHERS. 

See:  AMS,  C&W,  CUP,  Loeb Classical Library,  MA,  MAA,  NCTM,  OUP  in Common References below.

 

          ABBREVIATIONS OF MONTHS.  All months are given by their first three letters in English:  Jan,  Feb,  ....

 

          PUBLISHERS' LOCATIONS.  The following publisher's locations will not be cited each time.  Other examples may occur and can be found in the file PUBLOC.

 

AMS (American Mathematical Society),  Providence, Rhode Island, USA.

Chelsea Publishing,  NY, USA. 

CUP (Cambridge University Press),  Cambridge, UK. 

Dover,  NY, USA.

Freeman,  San Francisco, then NY, USA.

Harvard University Press,  Cambridge, Massachusetts, USA.

MA (Mathematical Association),  Leicester, UK.

MAA (Mathematical Association of America),  Washington, DC, USA.

NCTM (National Council of Teachers of Mathematics),  Reston, Virginia, USA.

Nelson,  London, UK. 

OUP (Oxford University Press),  Oxford, UK (and also NY, USA).

Penguin,  Harmondsworth, UK. 

Simon & Schuster,  NY, USA.

 

 

 

                              COMMON REFERENCES.

 

          NOTES.  When referring to items below, I will usually include the earliest reasonable date, even though the citation may be to a much later edition.  For example, I would say "Canterbury Puzzles, 1907", even though I am citing problem numbers or pages from the 1958 Dover reprint of the 1919 edition.  Sometimes the earlier editions are hard to come by and I have sometimes found that the earlier edition has different pagination -- in that case I will (eventually) make the necessary changes.

          Edition information in parentheses indicates items or editions that I have not seen, though I don't always do this when the later version is a reprint or facsimile.

 

 

Abbaco.                        See:  Pseudo-dell'Abbaco.

Abbot Albert. Abbot Albert von Stade.  Annales Stadenses.  c1240.  Ed. by J. M. Lappenberg.  In: Monumenta Germaniae Historica, ed. G. H. Pertz, Scriptorum t. XVI, Imp. Bibliopolii Aulici Hahniani, Hannover, 1859 (= Hiersemann, Leipzig, 1925), pp. 271‑359.  (There are  13  recreational problems on pp. 332‑335.)  [Vogel, on p. 22 of his edition of the Columbia Algorism, dates this as 1179, but Tropfke gives 1240, which is more in line with Lappenberg's notes on variants of the text.  The material of interest, and several other miscellaneous sections, is inserted at the year 1152 of the Annales, so perhaps Vogel was misled by this.]  I have prepared an annotated translation of this: The problems of Abbot Albert (c1240).  I have numbered the problems and will cite this problem number.

Abraham.                      R. M. Abraham.  Diversions and Pastimes.  Constable, London, 1933  = Dover, 1964 (slightly amended and with different pagination, later retitled: Tricks and Amusements with Coins, Cards, String, Paper and Matches).  I will cite the Constable pages (and the Dover pages in parentheses).

Ackermann.                  Alfred S. E. Ackermann.  Scientific Paradoxes and Problems and Their Solutions.  The Old Westminster Press, London, 1925.

D. Adams.  New Arithmetic.  1835. 

                                    Daniel Adams (1773-1864).  ADAMS NEW ARITHMETIC.  Arithmetic, in which the principles of operating by numbers are analytically explained, and synthetically applied; thus combining the advantages to be derived both from the inductive and synthetic mode of instructing:  The whole made familiar by a great variety of useful and interesting examples, calculated at once to engage the pupil in the study, and to give him a full knowledge of figures in their application to all the practical purposes of life.  Designed for the use of schools and academies in the United States.  J. Prentiss, Keene, New Hampshire, 1836, boarded.  1-262 pp + 2pp publisher's ads, apparently inserted backward.  [Halwas 1-6 lists 1st ed as 1835, then has 1837, 1838, 1839, 1842, c1850.]  This is a reworking of The Scholar's Arithmetic of 1801.

D. Adams.  Scholar's Arithmetic.  1801.

                                    Daniel Adams (1773-1864).  The Scholar's Arithmetic;  or,  Federal Accountant:  Containing.  I. Common arithmetic, ....  II. Examples and Answers with Blank Spaces, ....  III. To each Rule, a Supplement, comprehending, 1. Questions ....  2. Exercises.  IV. Federal Money, ....  V. Interest cast in Federal Money, ....  VI. Demonstration by engravings ....  VII. Forms of Notes, ....  The Whole in a Form and Method altogether New, for the Ease of the Master and the greater Progress of the Scholar.  Adams & Wilder, Leominster, Massachusetts, 1801; 2nd ed, 1802.  3rd ed ??.  4th ed, by Prentiss, 1807; 6th ed, 1810; 10th ed, 1816; Stereotype Edition, Revised and Corrected, with Additions, 1819, 1820, 1824; John Prentiss, Keene, New Hampshire, 1825.  [Halwas 8-14.]  I have the 1825, whose Preface is for the 10th ed of 1816, so is probably identical to that ed.  The Preface says he has now made some revisions.  The only change of interest to us is that he has added answers to some problems.  So I will cite this as 1801 though I will be giving pages from the 1825 ed.  The book was thoroughly reworked as Adams New Arithmetic, 1835.

M. Adams.  Indoor Games.  1912.

                                    Morley Adams, ed.  The Boy's Own Book of Indoor Games and Recreations.  "The Boy's Own Paper" Office, London, 1912; 2nd ptg, The Religious Tract Society, London (same address), 1913.  [This is a major revision of: G. A. Hutchison, ed.; Indoor Games and Recreations; The Boy's Own Bookshelf; New ed., Religious Tract Society, London, 1891 (possibly earlier) -- see 5.A.]

M. Adams.  Puzzle Book.  1939.

                                    Morley Adams.  The Morley Adams Puzzle Book.  Faber & Faber, London, 1939.

M. Adams.  Puzzles That Everyone Can Do.  1931.

                                    Morley Adams.  Puzzles that Everyone Can Do.  Grant Richards, London, 1931, boarded.

AGM.           Abhandlungen zur Geschichte der Mathematischen Wissenschaften mit Einschluss ihrer Anwendungen.  Begründet von Moritz Cantor.  Teubner, Leipzig.  The first ten volumes were Supplements to Zeitschrift für Math. u. Physik, had a slightly different title and are often bound in with the journal volume.

Ahrens, Wilhelm Ernst Martin Georg (1872-1927).  See:  A&N,  MUS,  3.B,  7.N.

al‑Karkhi.                      Aboû Beqr Mohammed Ben Alhaçen Alkarkhî [= al‑Karagi  = al‑Karajī].  Untitled MS called Kitāb al-Fakhrī (or just Alfakhrî) (The Book Dedicated to Fakhr al-Din).  c1010.  MS 952, Supp. Arabe de la Bibliothèque Impériale, Paris.  Edited into French by Franz Woepcke as: Extrait du Fakhrî.  L'Imprimerie Impériale, Paris, 1853; reprinted by Georg Olms Verlag, Hildesheim, 1982.  My page citations will be to Woepcke.  Woepcke often refers to Diophantos, but his numbering gets ahead of Heath's.

Alberti.  1747.               Giuseppe Antonio (or Giusepp-Antonio) Alberti (1715-1768).  I Giochi Numerici Fatti Arcani Palesati da Giuseppe Antonio Alberti.  Bartolomeo Borghi, Bologna, 1747, 1749.  Venice, 1780, 1788(?).  4th ed., adornata di figure, Giuseppe Orlandelli for Francesco di Niccolo' Pezzana, Venice, 1795  (reprinted: Arnaud, Florence, 1979), 1813.  Adornata di 16 figure, Michele Morelli, Naples, 1814.  As: Li Giuochi Numerici Manifestati, Edizione adorna di Figure in rame, Giuseppe Molinari, Venice, 1815.

                    The editions have almost identical content, but different paginations.  I have compared several editions and seen little difference.  The 1747 ed. has a dedication which is dropped in the 2nd ed. which also omits the last paragraph of the Prefazione.  I only saw one other point where a few words were changed.  I will give pages of 1747 (followed by 1795 in parenthesis).  Much of Alberti, including almost all the material of interest to us and many of the plates, is translated from vol. 4 of the 1723 ed. of Ozanam. 

                    (Serge Plantureux's 1993 catalogue describes a 1747-1749 ed. with Appendice al Trattato de' Giochi Numerici (1749, 72 pp) & Osservazioni all'Appendice de' Giochi Numerici (38 pp), ??NYS.  The copy in the Honeyman Collection had the Appendice.  Christopher 3 has the Osservazioni.  The Appendice is described by Riccardi as a severe criticism of Alberti, attributed to Giovanni Antonio Andrea Castelvetri and published by Lelio dall Volpe, Bologna, 1749.  The Osservazioni are Alberti's response.)

 

Alcuin (c735-804).

                                    Propositiones Alcuini doctoris Caroli Magni Imperatoris ad acuendos juvenes.  9C. 

                    IN: B. Flacci Albini seu Alcuini, Abbatis et Caroli Magni Imperatoris Magistri.  Opera Omnia: Operum pars octava: Opera dubia.  Ed. D. Frobenius, Ratisbon, 1777, Tomus secundus, volumen secundum, pp. 440‑448.  ??NYS.  Revised and republished by J.‑P. Migne as: Patrologiae Cursus Completus: Patrologiae Latinae, Tomus 101, Paris, 1863, columns 1143‑1160.

                    A different version appears in: Venerabilis Bedae, Anglo‑Saxonis Presbyteri.  Opera Omnia: Pars Prima, Sectio II -- Dubia et Spuria: De Arithmeticus propositionibus.  Tomus 1, Basel, 1563.  (Rara, 131, says there were earlier editions: Paris, 1521 (part), 1544‑1545 (all), 1554, all ??NYS.)  Revised and republished by J.‑P. Migne as: Patrologiae Cursus Completus: Patrologiae Latinae, Tomus 90, Paris, 1904, columns 665‑672.  Incipiunt aliae propositiones ad acuendos juvenes is col. 667‑672.  A version of this occurs in Ens' Thaumaturgus Mathematicus of 1636 -- cf under Etten.

                    The Alcuin has 53 numbered problems with answers.  The Bede has 3 extra problems, but the problems are not numbered, there are only 31 1/2 answers and there are several transcription errors.  The editor has used the Bede to rectify the Alcuin.

                    There is a recent critical edition of the text by Folkerts -- Die älteste mathematische Aufgabensammlung in lateinischer Spräche: Die Alkuin zugeschriebenen Propositiones ad Acuendos Iuvenes; Denkschriften der Österreichischen Akademie der Wissenschaften, Mathematische‑naturwissenschaftliche Klasse 116:6 (1978) 13‑80.  (Also separately published by Springer, Vienna, 1978.  The critical part is somewhat revised as: Die Alkuin zugeschriebenen "Propositiones ad Acuendos Iuvenes"; IN: Science in Western and Eastern Civilization in Carolingian Times, ed. by P. L. Butzer & D. Lohrmann; Birkhäuser, Basel, 1993, pp. 273-281.)  He finds that the earliest text is late 9C and is quite close to the first edition cited above.  He uses the same numbers for the problems as above and numbers the extra Bede problems as 11a, 11b, 33a.  I use Folkerts for the numbering and the titles of problems.

                    John Hadley kindly translated Alcuin for me some years ago and made some amendments when Folkerts' edition appeared.  I annotated it and it appeared as: Problems to Sharpen the Young, MG 76 (No. 475) (Mar 1992) 102-126.  A slightly corrected and updated edition, containing some material omitted from the MG version, is available as Technical Report SBU-CISM-95-18, School of Computing, Information Systems, and Mathematics, South Bank University, Oct 1995, 28pp.

                    Menso Folkerts and Helmuth Gericke have produced a German edition: Die Alkuin zugeschriebenen Propositiones ad Acuendos Juvenes (Aufgabe zur Schärfung des Geistes der Jugend); IN: Science in Western and Eastern Civilization in Carolingian Times, ed. by P. L. Butzer & D. Lohrmann; Birkhäuser, Basel, 1993, pp. 283-362.

                    See also:  David Singmaster.  The history of some of Alcuin's Propositiones.  IN: Charlemagne and his Heritage  1200 Years of Civilization and Science in Europe: Vol. 2  Mathematical Arts; ed. by P. L. Butzer, H. Th. Jongen & W. Oberschelp; Brepols, Turnhout, 1998, pp. 11‑29.

 

AM.  1917.                   H. E. Dudeney.  Amusements in Mathematics.  Nelson, 1917.  (There were reprintings in 1919, 1920, 1924, 1925, 1927, 1928, 1930, 1932, 1935, 1938, 1939, 1941, 1943, 1946, 1947, 1949, 1951, but it seems that the date wasn't given before 1941?)  = Dover, 1958.

AMM.                          American Mathematical Monthly.

AMS.                           American Mathematical Society.

Les Amusemens.  1749.

                                    Les Amusemens Mathématiques Precedés Des Elémens d'Arithmétique, d'Algébre & de Géométrie nécessaires pour l'intelligence des Problêmes.  André‑Joseph Panckoucke, Lille, 1749.  Often listed with Panckoucke as author (e.g. by the NUC, the BNC and Poggendorff), but the book gives no such indication.  Sometimes spelled Amusements.  There were 1769 and 1799 editions.

Apianus.  Kauffmanss Rechnung.  1527. 

                                    Petrus Apianus (= Peter Apian or Bienewitz or Bennewitz) (1495‑1552).  Eyn Newe Unnd wolgegründte underweysung aller Kauffmanss Rechnung in dreyen Büchern / mit schönen Regeln uň [NOTE:  ň  denotes an  n  with an overbar.] fragstucken begriffen.  Sunderlich was fortl unnd behendigkait in der Welschē Practica uň Tolletn gebraucht wirdt / des gleychen fürmalss wider in Teützscher noch in Welscher sprach nie gedrückt.  durch Petrum Apianū von Leyssnick / d Astronomei zů Ingolstat Ordinariū / verfertiget.  Georgius Apianus, Ingolstadt, (1527), facsimile, with the TP of the 1544 ed. and 2pp of publication details added at the end, Polygon-Verlag, Buxheim-Eichstätt, 1995, with 8pp commentary leaflet by Wolfgang Kaunzner.  (The TP of this has the first known printed version of Pascal's Triangle.)  Smith, Rara, pp. 155-157.  (The  d  is an odd symbol, a bit like a  δ  or an  8,  which is used regularly for  der  both as a single word and as the ending of a word, e.g.  and  for  ander.)  Smith notes that Apianus follows Rudolff (1526) very closely.

AR.  c1450.                  Frater Friedrich Gerhart (attrib.).  Latin & German MSS, c1450, known as Algorismus Ratisbonensis.  Transcribed and edited from 6 MSS by Kurt Vogel as: Die Practica des Algorismus Ratisbonensis; C. H. Beck'sche Verlagsbuchhandlung, Munich, 1954.  (Kindly sent by Prof. Vogel.)  Vogel classifies the problems and gives general comments on the mathematics on pp. 155‑189.  He gives detailed historical notes on pp. 203‑232.  When appropriate, I will cite these pages before the specific problems.  He says (on p. 206) that almost all of Munich 14684 (see below) is included in AR.

Arnold, George.  See:  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Hanky Panky.

Arrighi, Gino.  See:  Benedetto da Firenze,  Calandri,  Pseudo-Dell'Abbaco,  della Francesca,  Gherardi,  Lucca 1754,  P. M. Calandri.

Aryabhata.                    Āryabhata (I)) [NOTE:  ţ  denotes a  t  with a dot under it and  ş  denotes an  s  with a dot under it.]  (476-  ). Āryabhatīya.  499.  Critically edited and translated into English by Kripa Shankar Shukla, with K. V. Sarma.  Indian National Science Academy, New Delhi, 1976.  (Volume 1 of a three volume series -- the other two volumes are commentaries, of which Vol. 2 includes the commentary Āryabhatīya-Bhāşya, written by Bhaskara I in 629.  Aryabhata rarely gives numerical examples, so Bhaskara I provided them and these were later used by other Indian writers such as Chaturveda, 860.  The other commentaries are later and of less interest to us.  Prof. Shukla has sent a photocopy of an introductory booklet, which is an abbreviated version of the introductory material of Vol. 1, with some extensions relating Aryabhata to other writers.)  The material is organized into verses.  There is an older translation by Walter Eugene Clark as:  The  Âryabhaţîya  of  Âryabhaţa; Univ. of Chicago Press, Chicago, 1930.  (There was an Aryabhata II, c950, but he only occurs in 7.K.1.)

A&N.                           Wilhelm Ahrens.  Altes und Neues aus der Unterhaltungsmathematik.  Springer, Berlin, 1918.

 

Bachet, Claude‑Gaspar (1581-1638).  See:  Problemes.

Bachet-Labosne.  See:  Problemes.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].

                                    Philosophical Recreations, or, Winter Amusements.  Thomas Hughes, London, nd [1820].  [BCB 18-19; OCB, pp. 180 & 197.  Heyl 22-23.  Toole Stott 75-77.  Christopher 54-56.  Wallis 34 BAD, 35 BAD.  These give dates of 1820, 1822, 1828.]  HPL [Badcock] RBC has three versions with slightly different imprints on the title pages, possibly the three dates mentioned.

                    Wallis 34 BAD has this bound after the copy of:  John Badcock; Domestic Amusements, or Philosophical Recreations ...; T. Hughes, London, nd [1823], and it is lacking its Frontispiece and TP -- cf in 6.BH.  HPL [Badcock] has both books, including the folding Frontispieces.  The earlier does not give an author, but its Preface is signed J. B. and the later book does give his name and calls itself a sequel to the earlier.  Toole Stott 75-80 clearly describes both works.  Some of the material is used in Endless Amusement II. 

Baker.  Well Spring of Sciences.  1562?

                                    Humfrey Baker (fl. 1557-1587).  The Well Sprynge of Sciences Which teacheth the perfect worke and practise of Arithmeticke both in whole numbers and fractions, with such easye and compendious instruction into the sayde arte, ....  Rouland Hall for James Rowbotham, London, 1562.  [Smith, Rara, p. 327, says it was written in 1562 but wasn't actually printed until 1568, but a dealer says the 1st ed. was 1564 and there was a 4th ed. in 1574, which I have examined.]  Apparently much revised and extended, (1580).  Reprinted, with title: The Wel [sic] Spring of Sciences: Which teacheth the perfect worke and practise of Arithmetike; Thomas Purfoote, London, 1591.  I have seen Thomas Purfoot, London, 1612, which is essentially identical to 1591.  I have also seen:  Christopher Meredith, London, 1646;  Christopher Meredith, London, 1650;  R. & W. L. for Andrew Kemb, London, 1655;  which are all the same, but differently paged than the 1591.  I have also seen Baker's Arithmetick, ed. by Henry Phillippes, Edward Thomas, London, 1670, which has different pagination and some additional problems compared to the 1646/1655 ed.  [Smith, Rara, 327-330 & 537, says it was rewritten in 1580, but there is little difference between the 1580 and the many later editions, so the 1591 ed. is probably close to the 1580 ed.  The copy of the 1562 in the Graves collection ends on f. 160r, but an owner has written a query as to whether the book is complete.  Neither Smith nor De Morgan seems to have seen a 1562 so they don't give a number of pages for it.  (STC records no copies of the 1562, 1564, 1576, 1584, 1607 editions, but there was a 1576 by [T. Purfoote], apparently the 5th ed., of c500pp, in the Honeyman Collection.)  Almost all the problems of interest occur on ff. 189r-198r of the 1591 ed. and hence are not in the Graves copy of the 1562 ed., but H&S 61 refers to one of these problems as being in Baker, 1568.  The 1574 ends at fol. 200 (misprinted as 19?, where the ? is an undecipherable blob) and Chapter 16, which is headed:  The 16 Chapter treateth of sportes and pastime, done by number,  is on ff. 189r-200v, and contains just a few recreations, as in Recorde.  So I will date the book as 1562?, but most of the later material as 1580?.  The problems of 7.AF.1 and 10.A may be in Graves copy of the 1562 ed. -- ??check.  I will cite the 1580?, 1646 and 1670 editions, e.g.  1580?: ff. 192r 193r; 1646: pp. 302-304; 1670: pp. 344-345.]  Bill Kalush has recently sent a CD with 1574, 1580, 1591, 1598, 1602, 1607, 1612, 1617, 1650, 1655 on it -- ??NYR.

Bakhshali MS.               The Bakhshālī Manuscript, c7C.  This MS was found in May 1881 near the village of Bakhshālī, in the Yusufzāī district of the Peshawer division, then at the northwestern frontier of India, but apparently now in Pakistan.  This is discussed in several places, such as the following, but a complete translation has only recently appeared.  David Pingree says it is 10C, but his student Hayashi opts for 7C which seems pretty reasonable and I will adopt c7C.

                    1.             A. F. Rudolf Hoernle.  Extract of his report in some journal of the previous year.  The Indian Antiquary 12 (Mar 1883) 89-90.  A preliminary report, saying it was found near Bakhshâlî in the Yusufzai District in the Panjâb.

                    2.             A. F. Rudolf Hoernle.  On the Bakhshālī Manuscript.  Berichte des VII. Internationalen Orientalisten‑Congresses, Wien, 1886.  Alfred Hölder, Vienna, 1889.  Arische Section, p. 127-147 plus three folding plates.  Cf next item.  I will cite this as Hoernle, 1886.

                    3.             A. F. Rudolf Hoernle.  The Bakhshali manuscript.  The Indian Antiquary 17 (Feb 1888) 33‑48 & Plate I opp. p. 46;  275‑279 & Plates II & III opp. pp. 276 & 277.  This is essentially a reprint of the previous item, with a few changes or corrections, but considerable additional material.  He dates it c4C.  I will cite this as Hoernle, 1888.

                    4.             G. R. Kaye.  The Bakhshālī Manuscript – A Study in Medieval Mathematics.  Archæeological Survey of India – New Imperial Series XLIII: I-III, with parts I & II as one volume, (1927‑1933).  (Facsimile reprint in two volumes, Cosmo Publications, New Delhi, 1981 – this is a rather poor facsimile, but all the text is preserved.  I have a letter detailing the changes between the original and this 'facsimile'.)  I will only cite Part I – Introduction, which includes a discussion of the text.  Part II is a discussion of the script, transliteration of the text and pictures of the entire MS.  Part III apparently was intended to deal with the language used, but Kaye died before completing this and the published Part III consists of only a rearranged version of the MS with footnotes explaining the mathematics.  Gupta, below, cites part III, as Kaye III and I will reproduce these citations.  He dates it c12C.

                    5.             B. Datta.  The Bakhshâlî mathematics.  Bull. Calcutta Math. Soc. 21 (1929) 1‑60.  This is largely devoted to dating of the MS and of its contents.  He asserts that the MS is a copy of a commentary on some lost work of 4C or 5C (?).

                    6.             R. C. Gupta.  Some equalization problems from the Bakhshālī manuscript.  Indian Journal of the History of Science 21 (1986) 51-61.  Notes that Hoernle gave the MS to the Bodleian Library in 1902, where it remains, with shelf mark MS. Sansk. d.14.  He follows Datta in believing that this is a commentary on a early work, though the MS is 9C, as stated by Hoernle.  He gives many problems from Kaye III, sometimes restoring them, and he discusses them in more detail than the previous works.

                    7.             Takao Hayashi.  The Bakhshālī Manuscript  An ancient Indian mathematical treatise.  Egbert Forsten, Groningen, Netherlands, 1995.  (Based on his PhD Dissertation in History of Mathematics, Brown University, May 1985, 774pp.)  A complete edition and translation with extensive discussion of the context of the problems.  He dates it as 7C. 

Ball, Walter William Rouse (1850-1925).  See:  Ball‑FitzPatrick;  MRE.

Ball‑FitzPatrick.

                                    French translation of MRE by J. Fitz‑Patrick, published by Hermann, Paris.

                    1st ed., Récréations et Problèmes Mathématiques des Temps Anciens & Modernes.  From the 3rd ed, 1896, of MRE, 'Revue et augmentée par l'auteur'.  1898.  The Note says 'M. Ball ... a bien voulu apporter à la troisième édition anglaise des additions et des modifications importantes.'  352pp.

                    2nd ed., Récréations et Problèmes Mathématiques des Temps Anciens et Modernes.  From the 4th ed, 1905, of MRE, 'et enrichie de nombreuses additions'.

              As three volumes, 1907‑09.  [I have vol. 1, 1907, which is 356pp.  Pp. 327‑355 is a note by A. Hermann, Comptabilité d'une persone qui dépense plus que son revenu.  I have not yet seen the other volumes to compare with the 1926 reprint, but Strens's notes in his copy indicate that they are identical.]

              Reprinted in one vol., Gabay, Paris, 1992, 544pp.

              Reprinted, 1926-1927.  The only copies that I have seen are bound as one volume, but with separate pagination.  My copy has left out the title pages of vols. 2 & 3.  The copy in the Strens Collection has these title pages, but its vol. II is 1908.  The 1926 vol. 1 says Nouvelle édition française, but the 1927 vol. 3 says Deuxième édition française. 

                    [Vol. 1 is 326pp, omitting the note by Hermann.  Vol. 2 is 363pp (pp. 322‑355 is a historical note on the cubic, based on Cossali (1797)).  Vol. 3 is 363pp including: Notes diverses de M. Aubry, pp. 137‑206 (or 340? -- the Table des Matières and the page set up do not make it clear if Aubry's Notes end on p. 206);  Note de M. Fitz‑Patrick, La géométrie par le pliage et découpage du papier, pp. 341‑360;  A. Margossian, De l'ordonnance des nombres dans les carrés magiques impairs, pp. 1‑60 (pp. 61-64 is a Note on the same subject, presumably part of Margossian's material);  Capt. Reinhart, some geometric notes, pp. 130-136.]

Barnard.  50 Observer Brain-Twisters.  1962.

                                    Douglas St. Paul Barnard.  Fifty Observer Brain‑Twisters  A Book of Mathematical and Reasoning Problems.  Faber, 1962.  US ed.:  A Book of Mathematical and Reasoning Problems: Fifty Brain Twisters;  Van Nostrand, 1962.  The editions have identical pagination.

Bartl.  c1920.                János Bartl.  Preis-Verzeichnis von Bartl's Akademie für moderne magische Kunst.  Hamburg, c1920.  Reprinted by Olms Verlag, Zürich, 1983, as: Zauberkatalog Bartl.  References are to the section: Vexier- und Geduldspiele, pp. 305‑312.

Bartoli.  Memoriale.  c1420.

                                    Francesco Bartoli (  -1425).  Memoriale (= Notebook) containing some 30 mathematical problems copied during 1400?-1425.  Ms 1 F 54 of the Archives départementales du Vaucluse, France.  ??NYS -- described and quoted in:  Jacques Sesiano; Les problèmes mathématiques du Memoriale de F. Bartoli; Physis 26:1 (1984) 129-150.

BC.                              Binomial Coefficient, i.e.  BC(n, k) = n!/k!(n-k)!.

BCB.  See: Hall, BCB.

BDM.  See under DSB.

Bede, The Venerable (c672-735).  (Now St. Bede.)  See:  Alcuin.

Benedetto da Firenze.  c1465.

                                    Benedetto da Firenze.  Trattato d'Abacho.  c1465.  This was a popular treatise and Van Egmond's Catalog 356 lists 18 copies under Benedetto.  Six show  B  as author, one has Benedetto, one has Benedetto da Firenze, one has  Po Ma  and one has Filipo Chalandri, so it seems Benedetto is the most likely author.  The MSS date from c1465 to c1525 and contain 9 to 25 chapters. 

                    The version in Cod. Acq. e doni 154, Biblioteca Medicea Laurenziana, Florence, c1480.  has been transcribed and edited by Gino Arrighi as:  Pier Maria Calandri; Tractato d'Abbacho; Domus Galilaeana, Pisa, 1974.  The incipit names  Po Ma  as author.  Cf Van Egmond's Catalog 96.  This version has 23 chapters.

Benson.  1904.              J. K. Benson.  The Book of Indoor Games for Young People of All Ages.  C. Arthur Pearson, London, 1904.  [This copies a lot from Hoffmann (or a common ancestor?).] 

                    Much of the material of Indoor Games is repeated in: J. K. Benson, ed.; The Pearson Puzzle Book; C. Arthur Pearson, London, nd [1921 -- BMC].  This is not in BMC or NUC under Benson -- but I have seen an ad listing this as by Mr. X and it is listed under Mr. X in BMC.  Puzzle Book pp. 1-96  =  Indoor Games pp. 189-257;  Puzzle Book pp. 109-114  =  Indoor Games pp. 258-262.  The only different material in Puzzle Book is pp. 97-108.  Neither book refers to the other.  Cf Mr. X in Section 4.A.1

Berkeley & Rowland.  Card Tricks & Puzzles.  1892.

                                    "Berkeley" [Peel, Walter H.]  &  Rowland, T. B.  Card Tricks and Puzzles.  The Club  Series, George Bell & Sons, London, 1892 -- according to BMC, but my copy is 1897.  Card Puzzles, etc., pp. 1-74 is by Berkeley; Arithmetical Puzzles, pp. 75-120 is by Rowland.

Berlekamp, Elwyn R. (1940-  )  See:  Winning Ways.

Bestelmeier.  1801-1803.

                                    G. H. [Georg Hieronimus] Bestelmeier.  Magazin von verschiedenen Kunst‑ und andern nützlichen Sachen ....  [Toy catalogues.]  Nuremberg, 1801‑1803.

                    Eight issues and cumulative classified index reprinted by Olms, Zurich, 1979.  Issue VII is 1801; the others are 'neue verbesserte Auflage', 1803.  This includes items numbered 1 through 1111.

                    Selections, with English translations.  Daniel S. Jacoby, ed.  The Amazing Catalogue of the Esteemed Firm of George Hieronimus Bestelmeier.  Selected Excerpts from Editions of 1793 and 1807.  [A comment inside makes me wonder if 1793-1807 is meant??]  Merrimack Publishing Corp., NY, 1971, 82pp.  The numeration is the same as in the Olms edition, but the Jacoby continues to item 1321.  Obviously these later items come from the 1807 edition, but we cannot tell if they might date from 1805, say, nor whether all the earlier items go back to 1793.  Jerry Slocum uses Jacoby in his Compendium and has kindly provided photocopies of Jacoby's pp. 70-82 containing all the items after 1111 and some examples of the earlier items.  Jacoby does not translate the texts, but just provides English labels for each picture and these labels are sometimes unconnected with the text. 

                    Many of Bestelmeier's items are taken from Catel; Kunst-Cabinet; 1790.  Sometimes the figure is identical (often reversed) or is a poor copy.  Texts are often copied verbatim, or slightly modified, but often abbreviated.  E.g. Catel often explains the puzzle and this part is frequently omitted in Bestelmeier.  Bestelmeier was the successor to Catel, qv.  The booklet by Slocum & Gebhardt (qv under Catel) gives precise datings for the various parts of these catalogues, but I have not yet entered these details.

Bhaskara I.  629.

                                    Bhāskara I.  Āryabhaţīya-Bhāşya.  [NOTE:  ţ  denotes a  t  with a dot under it and  ş  denotes an  s  with a dot under it.]   629.  Critically edited, including an English Appendix of the numerical examples used, by Kripa Shankar Shukla.  Indian National Science Academy, New Delhi, 1976.  (Vol. 2 of a three volume series devoted to the Āryabhaţīya (499) of Aryabhata (476-  ), qv.)  Bhaskara I repeats and exposits Aryabhata verse by verse, but Aryabhata rarely gives numerical examples, so Bhaskara I provided them and these were later used by other Indian writers such as Chaturveda, 860.  His earlier Maha-Bhaskariya (Mahā‑Bhāskarīya) of c629 is cited in 7.P.2.  Shukla's Appendix is sometimes brief, but sometimes very detailed, e.g. on the 26 examples of Chinese remainder problems.

Bhaskara II (1114-c1185).

                                    Bhâskara II (1114-c1185, see Colebrooke).

Biggs, Norman L.  See:  BLW.

Bijaganita.                     Bîjaganita of Bhaskara II, 1150 (see Colebrooke).

The Bile Beans Puzzle Book.  1933. 

                                    Bile Beans (C. E. Fulford, Ltd., Leeds, England).  The Bile Beans Puzzle Book.  1933.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

                                    Claude Birtwistle.  Mathematical Puzzles and Perplexities  How to Make the Most of Them.  George Allen & Unwin, London, 1971.

Birtwistle.  Calculator Puzzle Book.  1978.

                                    Claude Birtwistle.  The Calculator Puzzle Book.  Paperfronts (Elliot Right Way Books), Kingswood, Surrey, 1978.  (There is a US ed. by Bell, NY, 1978.)

BL(LD).                       British Library (Lending Division).

Blasius.  1513.               Johannis (or Joannes) Martinus Blasius (later denoted Sileceus or Sciliceus).  Liber Arithmetice Practice Astrologis Phisicis et Calculatioribus admodum utilis.  Thomas Kees for Joannis Parui & Joannis Lambert (in colophon; TP has Jehanlambert), Paris, 1513.  Facsimile by Heffer Scientific Reprint, Cambridge, 1960.  See Smith, Rara, pp. 95-97.  The Glaisher article in 7.P.5 [Messenger of Mathematics 53 (1923-24) 1‑131] discusses this book and says he only knows one example of it, which he has in front of him, so I suspect this facsimile is from that copy.  See Rara 95-97.  The Honeyman Collection had a copy, saying it was printed for J. Petit and J. Lambert and that copy had Petit's device on the TP while the TP shown in Rara has Lambert's device, which is as in this facsimile.  There was a reprinting in 1514 and extended editions in 1519 (ed. by Oronce Finé) and 1526 (ed. by T. Rhaetus) [Honeyman Collection, nos. 350-352].

BLC.                            British Library Catalogue, replacing BMC, in progress since 1970s. 

BLC-Ø                         Indicates that I could not find the item in the BLC.

BLW.  1976. Norman L. Biggs,  E. Keith Lloyd  &  Robin J. Wilson.  Graph Theory 1736‑1936.  OUP, 1976.

Blyth.  Match-Stick Magic.  1921.

                                    Will Blyth.  Match-Stick Magic.  C. Arthur Pearson, London, 1921, reprinted 1923, 1939.

BM(C).                         British Museum (Catalogue (of books) to 1955.  c1963). 

BMC65.                       Supplement to the above Catalogue for 1956‑1965.  c1968.

BN(C).                         Bibliothèque National, Paris.  (Catalogue, 1897-1981.)

Bodleian.                      The Bodleian Library, University of Oxford, or its catalogue.

Bonnycastle.  Algebra.  1782

                                    John Bonnycastle (??-1821).  An Introduction to Algebra, with Notes and Observations; designed for the Use of Schools, and Other Places of Public Education.  1782.  The first nine editions appeared "without any material alterations".  In 1815, he produced a 10th ed., "an entire revision of the work" which "may be considered as a concise abridgment" of his two volume Treatise on Algebra, 1813, (2nd ed. in 1820).  The 1815 ed. had an Appendix: On the application of Algebra to Geometry.  I have a copy of the 7th ed., 1805, printed for J. Johnson, London, and it is identical to the 2nd ed. of 1788 except for a problem in the final section of Miscellaneous Questions.  However, the 9th ed. of 1812 has page numbers advanced by 10 except toward the end of the book.  I also have the 13th ed. of 1824, printed for J. Nunn and 11 other publishers, London, 1824.  This version has an Addenda: A New Method of resolving Numerical Equations, by his son Charles Bonnycastle (1797-1840), but is otherwise identical to the 10th ed. of 1815.  The earlier text was expanded by about 10% in 1815, so a number of problems only occur in later editions.  I will cite these later problems as 1815 and will cite the earlier problems as 1782.  [Halwas 36-38 gives some US editions.]

 

Book of 500 Puzzles.  1859.

                                    The Book of 500 Curious Puzzles: Containing a Large Collection of Entertaining Paradoxes, Perplexing Deceptions in Numbers, and Amusing Tricks in Geometry.  By the author of "The Sociable," "The Secret Out," "The Magician's Own Book," "Parlor Games," and " Parlor Theatricals," etc.  Illustrated with a great Variety of Engravings.  Dick & Fitzgerald, NY, 1859.  Compiled from The Sociable (qv) and Magician's Own Book.  Pp. 1-2 are the TP and its reverse.  Pp. 3‑36, are identical to pp. 285-318 of The Sociable; pp. 37-54 are identical to pp. 199-216 of Magician's Own Book and pp. 55-116 are identical to pp. 241-302 of Magician's Own Book.  [Toole Stott 103 lists it as anonymous.  NUC, under Frikell, says to see title.  NUC, under Book, also has an 1882 ed, compiled by William B. Dick.  Christopher 129.  C&B lists it under Cremer.]

 

                    The authorship of this and the other books cited -- The Sociable, The Secret Out, The Magician's Own Book, Parlor Games, and Parlor Theatricals, etc. -- is confused.  BMC & NUC generally assign them to George Arnold (1834-1865) or Wiljalba (or Gustave) Frikell (1818 (or 1816) - 1903), sometimes with Frikell as UK editor of Arnold's US version -- but several UK versions say they are translated and edited by W. H. Cremer Jr, and one even cites an earlier French book (though the given title may not exist!, but cf Manuel des Sorciers, 1825) -- see the discussion under Status of The Project, in the Introduction, above.  The names of Frank Cahill, Henry Llewellyn Williams and Gustave Frikell (Jr.) are sometimes associated with versions of these as authors or coauthors.  The Preface of The Sociable says that most of the Parlor Theatricals are by Frank Cahill and George Arnold -- this may indicate they had little to do with the parts that interest us.  Toole Stott 640 opines that this reference led Harry Price to ascribe these books to these authors.

                    A publisher's ad in the back says: "The above five books are compiled from the "Sociable" and "Magician's Own."", referring to: The Parlor Magician [Toole Stott 543, 544]; Book of Riddles and Five Hundred Home Amusements [Toole Stott 107, 951]; Book of Fireside Games [possibly Toole Stott 300??]; Parlor Theatricals; The Book of 500 Curious Puzzles.  However, [Toole Stott 951] is another version of The Book of Riddles and Five Hundred Home Amusements "by the author of "Fireside Games" [Toole Stott 300], "The Parlor Magic" [perhaps Toole Stott 543, 544], "Parlor Tricks with Cards" [Toole Stott 1056 lists this as by Frikell, "abridged from The Secret Out" (see also 547, 1142)], ..."; Dick & Fitzgerald, 1986 [sic, but must mean 1886??]. 

                    See Magician's Own Book for more about the authorship.

                    See also:  Boy's Own Book,  Boy's Own Conjuring Book,  Illustrated Boy's Own Treasury,  Indoor and Outdoor,  Landells: Boy's Own Toy-Maker,  The Secret Out,  Hanky Panky,  The Sociable.

 

Book of Merry Riddles.  1629?

                                    The Book of Merry Riddles.  London, 1629.  [Santi 235.] 

          Several reprints.  Also known as Prettie Riddles.

                    A Booke of Merry Riddles; Robert Bird, London, 1631.  [Mark Bryant; Dictionary of Riddles; Routledge, 1990, p. 100.]

                    Booke of Merry Riddles; John Stafford & W. G., London, 1660.

                    Reprint of the 1629 in: J. O. Halliwell; The literature of the sixteenth and seventeenth centuries; London, 1851, pp. 67‑102.  [Santi 235.]

                    Reprint of the 1660 in: J. O. Halliwell; The Booke of Merry Riddles, together with proper questions, and witty proverbs, to make pleasant pastime.  Now first reprinted from the unique edition printed at London in the year 1660.  For the author, London, 1866.  This was a printing of 25 copies.  There is a copy at UCL and a MS note at the end says 15 copies were destroyed on 9 Apr 1866, signed: J. O. H., with Number 9 written below.  [Santi 307.]  I have seen this, but some of the riddles are quoted by other authors and I will date all items as 1629? until I examine other material.

                    Reprint of the 1629 in: Alois Brandl; Shakespeares Book of Merry Riddles und die anderen Räthselbücher seiner Zeit; Jahrbuch der deutschen Shakespeare-Gesellschaft 42 (1906) 1-64 (with the 1631 ed on pp. 53-63).  ??NYR.  [Santi 235 & 237.]

Borghi.  Arithmetica.  1484.

                                    Pietro Borghi  = Piero Borgo or Borgi (?? - ³1494).  Qui comenza la nobel opera de arithmethica ne la qual se tracta tute cosse amercantia pertinente facta & compilata p Piero borgi da veniesia.  Erhard Ratdolt, Venice, 1484.  2 + 116 numbered ff.  This is the second commercial arithmetic printed in Italy and was reprinted many times.  See Rara 16-22.  This edition was reproduced in facsimile, with notes by Kurt Elfering, as: Piero Borghi; Arithmetica  Venedig 1484; Graphos, Munich, 1964;  in: Veröffentlichungen des Forschungsinstituts des Deutschen Museums für die Geschichte der Naturwissenschaften und der Technik, Reihe C -- Quellentexte und Übersetzunge, Nr. 2, 1965.

                      The 3rd ed of 1491 had a title: Libro dabacho.  From the 4th ed of 1501, the title was Libro de Abacho, so this is sometimes used as the title for the first editions also.  Rara indicates that the printing was revised to 100 numbered ff by the 4th ed. of 1491.  I have examined a 1509 ed. by Jacomo Pentio, Venice, ??NX.  This has 100 numbered ff, but the last three ff contain additional material, though Rara doesn't mention this until the 11th ed of 1540.  H&S discusses a problem and the folio in the 1540 ed is the same as in the 1509 ed.  The locations of interest in the 1509 ed. are c18ff before the corresponding locations of the 1484.  Van Egmond's Catalog 293-297 lists 13 Venetian editions from 1484 to 1567.

                    It has been conjectured that this was a pseudonym of Luca Pacioli, but there is no evidence for this [R. Emmett Taylor; No Royal Road  Luca Pacioli and His Times; Univ. of North Carolina Press, Chapel Hill, 1942, pp. 60 & 349].

                    See also:  D. E. Smith; The first great commercial arithmetic; Isis 8 (1926) 41-49.

Bourdon.  Algèbre.  7th ed., 1834.

                                    Louis Pierre Marie Bourdon (1779-1854).  Élémens d'Algèbre.  7th ed., Bachelier, Paris, 1834.  (1st ed, 1817; 5th, 1828; 6th, 1831; 8th, 1837; 1840.  Undated preface in the 7th ed. describes many changes, so I will cite this as 1834, though much of the material would have occurred earlier.)

 

Boy's Own Book.  1828.

                                    William Clarke, ed.  The Boy's Own Book.  The bibliography of this book is extremely complex -- by 1880, it was described as having gone through scores of editions.  My  The Bibliography of Some Recreational Mathematics Books  has 11 pages listing 76 English (40 UK, 37 US, 1 Paris) versions and a Danish version, implying 88 English (50 UK, 37 US, 1 Paris) versions, and 10 (or 11) related versions, and giving a detailed comparison of the versions that I have seen.  Because of the multiplicity of versions, I have cited it by title rather than by the original editor's name, which is not in any of the books (except the modern facsimile) though this attribution seems to be generally accepted.  I have examined the following versions, sometimes in partial photocopies or imperfect copies. 

                    Vizetelly, Branston and Co., London, 1828, 448pp.;  2nd ed., 1828, 462pp.;  3rd ed., 1829, 464pp (has an inserted advertisement sheet);  6th ed??, c1830, 462pp?? (my copy lacks TP, pp. 417-418, 431-436, 461-462);  9th ed., 1834, 462pp.  Longman, Brown & Co., London, 24th ed., 1846, 462pp.  [The latter five are identical, except for a bit in the Prelude (and the extra sheet in 3rd ed), so I will just cite the first of these as 1828‑2.  It seems that all editions from the 2nd of 1828 through the 29th of 1848, 462pp. are actually identical except for a bit of the Prelude (and the advertisement sheet in the 3rd ed.)]

                    First American Edition.  Munroe & Francis, Boston  &  Charles S. Francis, NY, 1829, 316pp.  Facsimile by Applewood Books, Bedford, Massachusetts, nd [1998?].  This is essentially an abridgement of the 2nd ed of 1828, copying the Prelude and adding  "So far the London Preface.  The American publishers have omitted a few articles, entirely useless on this side of the Atlantic, ...."  The type is reset, giving some reduction in pages.  A number of the woodcuts have been omitted.  The section title pages are omitted.  Singing Birds, Silkworms, White Mice, Bantams, Magnetism, Aerostatics, Chess and Artificial Fireworks are omitted.  Angling, Rabbits, Pigeons, Optics are reduced.  Rosamond's Bower is omitted from Paradoxes and Puzzles.  Surprisingly, The Riddler is increased in size.  The 2pp Contents is omitted and an 8pp Index is added.

                    Baudry's European Library  &  Stassin & Xavier,  Paris, 1843, 448pp.  [The existence of a Paris edition was previously unknown to the vendor and myself, but it is Heyl 354 and he cites Library of Congress.  It is very different than the English and US editions, listing J. L. Williams as author.  Even when the topic is the same, the text, and often the topic's name, are completely rewritten.  See my  The Bibliography of Some Recreational Mathematics Books  for details -- in it I have found it generally necessary to treat this book separately from all other editions.  I will cite it as 1843 (Paris).  Much of this, including almost all of the material of interest is copied exactly in  Anon: Boy's Treasury, 1844,  qv, and in translated form in  de Savigny, Livre des Écoliers, 1846,  qv.  The problem of finding the number of permutations of the letters of the alphabet assumes 24 letters, which makes me wonder if these books are based on some earlier French work.  Heyl 355 is probably the same book, with slight variations in the title, by Dean and Munday, London, c1845.]

                    David Bogue, London, 1855, 611pp.  [It seems that this version first appears in 1849 and continues through about 1859, when two sections were appended.]

                    [W. Kent (late D. Bogue), London, 1859, 624pp??.  For almost all material of interest, this is identical to the 1855 ed, so I will rarely (if ever?) cite it.]

                    [Lockwood & Co., London, 1861, 624pp.  Identical to the 1859 ed., so I will not cite it.]

                    Lockwood & Co., London, 1868, 696pp.

                    [Lockwood & Co., London, 1870, 716pp.  Identical to 1868 with 20pp of Appendices, so page numbers for material of interest are the same as in 1868, so I will not cite it.]

                    [Crosby Lockwood & Co., London, 1880, 726pp.  Identical to 1870, but having the Appendices and 20 more pages incorporated into a new section.  For almost all material of interest, the page numbers are 30 ahead of the 1868 & 1870 page numbers, so I will not cite it except when the page numbers are not as expected.]

                    [5th (US?) ed., Worthington, NY, 1881, 362pp.  For almost all material of interest, this is identical to the 1829 (US) ed., so I will rarely (if ever?) cite it.]

          I will cite pages with edition dates and edition numbers or locations if needed (e.g.  1828-2: 410  or  1829 (US): 216).  See also:  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Illustrated Boy's Own Treasury.

                    Anonymous.  The Riddler; A Collection of Puzzles, Charades, Rebusses, Conundrums, Enigmas, Anagrams, &c. for the Amusement of Little Folks.  S. Babcock, New Haven, Connecticut, 1835.  22pp.  My copy has leaf 11/12 half missing and leaf 17/18 missing; NUC & Toole Stott 1392 say it should be 24pp, so presumably leaf 23/24 is also missing here.  [Toole Stott 1392 has The Riddler: or, Fire-Side Recreations; a collection ..., 1838, also listed in NUC.]  Paradoxes and Puzzles section consists of the introduction and 11 puzzles copied almost exactly from the Paradoxes and Puzzles section of Boy's Own Book, 2nd ed. of 1828 and this material is all in the first American edition of 1829.  Other material is charades, etc. and is all in both these versions of Boy's Own Book.  Shortz states that this is the first American book with puzzles -- but there were at least five American versions of Boy's Own Book before this and all the material in The Riddler, except some woodcuts, is taken from Boy's Own Book, so this pamphlet seems to be a pirate version.  NUC also lists a 1838 version.

 

Boy's Own Conjuring Book.  1860.

                                    The Boy's Own Conjuring Book: Being a Complete Hand-book of Parlour Magic; and Containing over One Thousand Optical, Chemical, Mechanical, Magnetical, and Magical Experiments, Amusing Transmutations, Astonishing Sleights and Subtleties, Celebrated Card Deceptions, Ingenious Tricks with Numbers, Curious and Entertaining Puzzles, Charades, Enigmas, Rebuses, etc., etc., etc.  Illustrated with nearly two hundred engravings.  Intended as a source of amusement for one thousand and one evenings.  Dick and Fitzgerald, NY, 1860.  384pp.  [Toole Stott 115, corrected, lists this as (1859), and under 114, describes it as an extended edition of The Magician's Own Book -- indeed the running head of the book is The Magician's Own Book! -- but see below.  Toole Stott 481 cites a 1910 letter from Harris B. Dick, of the publishers Dick & Fitzgerald.  He describes The Boy's Own Conjuring Book as a reprint of Magician's Own Book "evidently gotten up and printed in London, but singularly enough it had printed in the book on the title-page -- New York, Dick & Fitzgerald."  Indeed, all the monetary terms are converted into British.  Harold Adrian Smith [Dick and Fitzgerald Publishers; Books at Brown 34 (1987) 108-114] states that this is a London pirate edition.  BMC has 384pp, c1860.  NUC has a 384pp version, nd.  Christopher 145-149 are five versions from 1859 and 1860, though none has the blue cover of my copy.  Christopher 145 says the 1859 versions were printed by Milner & Sowerby, Halifax, and describes it as an extraction from Magician's Own Book, but see below.  Christopher 148 cites Smith's article.]  I also have a slightly different version with identical contents except omitting the date and frontispiece, but with a quite different binding, probably Christopher 149.  [NUC lists 334pp, nd; 416pp, nd and 416pp, 1860.  Toole Stott 114 is a 416pp version, 1861.  Toole Stott 959 is a 534pp version, 1861.  C&B cite a New York, 1859 with 416pp, a New York, nd, 334pp and London, c1850 (surely too early?).]

                    I have now compared this with The Magician's Own Book of 1857 and it is essentially a minor reworking of that book.  The Magician's Own Book has 17 chapters and an answers chapter and a miscellaneous chapter of items which are almost all also listed in the Contents under earlier sections.  All together, there are some 635 items.  The Boy's Own Conjuring Book copies about 455 of these items essentially directly, completely omitting the chapters on Electricity, Galvanism, Magnetism, Geometry, Art, Secret Writing and Strength, and almost completely omitting the chapter on Acoustics.  Of the 488 items in the other chapters, 453 are copied into the Boy's Own Conjuring Book, and this has in addition two of the acoustic problems, 125 new miscellaneous problems and 38pp of charades, riddles, etc.  (The later UK edition of Magician's Own Book is very different from the US edition.)  Many of the problems are identical to the Boy's Own Book or the Illustrated Boy's Own Treasury.   See also:  Book of 500 Puzzles,  Boy's Own Book,  Illustrated Boy's Own Treasury,  Landells: Boy's Own Toy‑Maker.

 

Boy's Treasury.  1844.

                                    Anonymous.  The Boy's Treasury of Sports, Pastimes, and Recreations.  With four hundred engravings.  By Samuel Williams.  [The phrasing on the TP could be read as saying Williams is the author, but the NUC entry shows he was clearly listed as the designer in later editions and his name appears on the Frontispiece.]  D. Bogue, London, 1844.  Despite the similarity of title, this is quite different from Illustrated Boy's Own Treasury and the similar books of c1860.  [Toole Stott 116.  Toole Stott 117 is another ed., 1847, 'considerably extended'.  Toole Stott gives US editions: 959; 960; 118; 199 & 961-965 are 1st, 1847; 2nd, 1847; 3rd, 1848; 6 versions of the 4th, 1850, 1848, 1849, 1852, 1854, 1848.  Hall, BCB 37 is a US ed. of 1850 = Toole Stott 119.  Christopher 151 is a US version of 1850?  NUC lists 9 versions, all included in Toole Stott.  Toole Stott cites some BM copies, but I haven't found this in the BMC.  A section of this, with some additional material, was reissued as Games of Skill and Conjuring: ..., in 1860, 1861, 1862, 1865, 1870 -- see Toole Stott 312-317.]

                    I have now found that much of this, including all the material of interest, is taken directly from the 1843 Paris edition of  Boy's Own Book,  qv, by J. L. Williams, including many of the illustrations - indeed they have the same Frontispiece, with S. Williams' name on it.

BR.  c1305.                  Greek MS, c1305, Codex Par. Suppl. Gr. 387, fol. 118v‑140v.  Transcribed, translated and annotated by Kurt Vogel as: Ein Byzantinisches Rechenbuch des frühen 14.Jahrhunderts; Wiener Byzantinistische Studien, Band VI; Hermann Böhlaus Nachf., Wien, 1968.  I will cite problem numbers and pages from this -- Vogel gives analysis of the methods on pp. 149‑153 and historical comments on pp. 154‑160, but I will not cite these.

Brahmagupta, c628.  See:  Brahma‑sphuta‑siddhanta;  Colebrooke.

Brahma‑sphuta‑siddhanta.

                                    Bráhma‑sphuta‑siddhânta of Brahmagupta, 628 (see Colebrooke).  He only states rules, which are sometimes obscure.  It appears from Colebrooke, p. v, and Datta (op. cit. under Bakhshali, p. 10), that almost all the illustrative examples and all the solutions are due to Chaturveda Prthudakasvâmî in 860.  Brahmagupta's rules are sometimes so general that one would not recognise their relevance to these examples and I have often not cited Brahmagupta.  E.g. cistern problems are given as examples to Brahmagupta's verse on how to add and subtract fractions.  (See also Datta & Singh, I, p. 248.)  Some of these comments are taken from Bhaskara I in 629.

Brush.          Hubert Phillips.  Brush Up Your Wits.  Dent, London, 1936.

BSHM.                         British Society for the History of Mathematics.  The produce a useful Newsletter.

Buteo.  Logistica.  1559.

                                    Johannes Buteo (= Jean Borrel, c1485-c1560 or c1492-1572).  Ioan. Buteonis Logistica, quæ & Arithmetica vulgò dicitur in libros quinque digesta: quorum index summatim habetur in tergo.  Gulielmus Rovillius, Lyons, 1559.  Most of the material is in books IV and V.  H&S cites some problems in the 1560 ed with the same pages as in the 1559 ed, so these editions are presumably identical.  See Rara 292-294.

 

c.                                 circa, e.g. c1300.  Also  c=  means "approximately equal", though  @  will be used in mathematical contexts.

C.                                 Century, e.g. 13C, -5C.

Calandri.  Arimethrica.  1491.

                                    Philippo Calandri.  Untitled.  Frontispiece is labelled "Pictagoras arithmetrice introductor".  Text begins: "Philippi Calandri ad nobilem et studiosus Julianum Laurentii Medicē de arimethrica opusculū."  Lorenzo de Morgiani & Giovanni Thedesco da Maganza, Florence, 1491.  Van Egmond's Catalog 298-299.  The Graves collection has two copies dated 1491, one with the folio number  c iiii  misprinted as  b iiii - cf Van Egmond for other differences in this unique variant.  There was a reprint by Bernardo Zucchetta, Florence, 1518 -- ??NYS but mentioned:  in a handwritten note in one of the Graves copies of the 1491 (giving Bernardo Zucchecta, 1517);  in Smith, Rara, p. 48 (giving Bernardo Zuchetta, 1518);  in Riccardi [I, col. 208-209] (giving Bernardo Zuchecta, 1515)  and in Van Egmond's Catalog 299.  "It is the first printed Italian arithmetic with illustrations accompanying problems, ...." (Smith, Rara, pp. 46‑49).  There are about 50 of these illustrations, which appear to be woodcuts, but they are quite small, about 25mm (1") square, and the same picture is sometimes repeated for a related but inappropriate problem.  Rara reproduces some of these, slightly reduced.  Riccardi [I, col. 208-209] says there may have been a 1490 ed. by Bernardo Zuchecta, but Van Egmond did not find any example.

Calandri.  Aritmetica.  c1485.

                                    Filippo Calandri.  Aritmetica.  c1485 [according to Van Egmond's Catalog 158-159].  Italian MS in Codex 2669, Biblioteca Riccardiana di Firenze.  Edited by Gino Arrighi, Edizioni Cassa di Risparmio di Firenze, Florence, 1969.  2 vol.: colour facsimile; transcription of the text.  Copies of the facsimile were exhausted about 1980 and repeated requests to the Cassa di Risparmio have not produced a reprint, though they usually send a copy of the text volume every time I write!  I have now (1996) acquired a example of the 2 vol. set and I find that copies of the text volume which are not part of a set have 8 colour plates inserted, but these are not in the copy in the set. 

                    I cite folios from the facsimile volume and pages from the text volume.  These are in direct correspondence with the original except for those pages with full page illustrations.  The original begins with a blank side with a Frontispiece verso, then 9 sheets (18 pp.) of full page tables, then two blank sheets.  The numbered folios then begin and go through 110.  Ff. 1r - 32r  are  pp. 3 - 65  of the text.  F. 32v is a full page calculation which is not in the text.  Then  ff. 33r - 110r  are  pp. 66 - 220  of the text.  F. 110v is a full page illustration omitted in the text.  The first 80 folio numbers are in elaborate Roman numerals centred at the head of the page.  (These are sometimes unusually written -- e.g. XXIIIIII.)  The later folios were not originally numbered and were later numbered in the top right corner using Hindu-Arabic numerals.

                    In Sep 1994, I examined the original MS, though it is on restricted access.  The original colours are rather more luminous than in the facsimile, but the facsimile is a first class job.  The history of this codex is obscure.  It is said to have belonged to Piero di Lorenzo dei Medici and it may be the book catalogued in the library of Francesco Pandolfini, c1515, as 'uno libretto ... di Filippo Calandri in arithmetica'.  The Riccardi family collected continuously from their rise in the mid 15C until the library was acquired by the city in 1813.  A number of items from the Pandolfini catalogue can be identified as being in the Riccardiana.  Van Egmond's dating may be early as some claim this was produced for Giuliano de' Medici, who was born in 1479. 

Calandri.  Raccolta.  c1495.

                                    Filippo Calandri.  Una Raccolta di Ragioni.  In: Cod. L.VI.45, Biblioteca Comunale di Siena.  Ed. by D. Santini.  Quaderni del Centro Studi della Matematica Medioevale, No. 4, Univ. di Siena, 1982.  Van Egmond's Catalog 193 identifies this as ff. 75r-111v of the codex, titles it Ragone Varie and gives a date of c1495.

Calandri.  See also:  Benedetto da Firenze,  P. M. Calandri.

Cardan.  Ars Magna.  1545.

                                    Jerome Cardan  = Girolamo Cardano  = Hieronymus Cardanus (1501‑1576).  Artis Magnae sive de Regulis Algebraicis  Liber Unus.  Joh. Petreium, Nuremberg, 1545, ??NYS  Included in Vol. IV of the Opera Omnia, Joannis Antonius Huguetan & Marcus Antonius Ravaud, Lyon, 1663, and often reprinted, e.g. in 1967.  NEVER CITED??

Cardan.  Practica Arithmetice.  1539.

                                    Jerome Cardan  = Girolamo Cardano  = Hieronymus Cardanus (1501‑1576).  Practica Arithmetice, & Mensurandi Singularis.  (Or: Practica Arithmeticae Generalis Omnium Copiosissima & Utilissima, in the 1663 ed.)  Bernardini Calusci, Milan, 1539.  Included in Vol. IV of the Opera Omnia, 1663, see above.  Some of the section numbers are omitted in the Opera Omnia and have to be intuited.  I will give the folios from the 1539 ed. followed by the pages of the 1663 ed., e.g. ff. T.iiii.v-T.v.r (p. 113).

Cardan.  De Rerum Varietate.  1557.

                                    Jerome Cardan  = Girolamo Cardano  = Hieronymus Cardanus (1501‑1576).  De Rerum Varietate.  Henricus Petrus, Basel, 1557; 2nd ed., 1557; 5th ed., 1581, ??NYS.  Included in Vol. III of the Opera Omnia, 1663, see above.

Cardan.  De Subtilitate.  1550.

                                    Jerome Cardan  = Girolamo Cardano  = Hieronymus Cardanus (1501‑1576).  De Subtilitate Libri XXI.  J. Petreium, Nuremberg, 1550; Basel, 1553; 6th ed., 1560; and five other 16C editions, part ??NYS.  Included in Vol. III of the Opera Omnia, 1663, see above.  French ed. by Richard Leblanc, Paris, 1556, 1584, titled: Les Livres d'Hieronymus Cardanus: De la Subtilité et subtiles Inventions, ensemble les causes occultes et les raisons d'icelles; 9th ed., 1611.  I have seen a note that the 1582 ed. by Henricus Petrus, Basel, was augmented by a riposte to attacks by Scaliger with further illustrations.

Carlile.  Collection.  1793.

                                    Richard Carlile.  A Collection of One Hundred and Twenty Useful and Entertaining Arithmetical, Mathematical, Algebraical, and Paradoxical Questions: With the Method of Working Each.  Printed by T. Brice for the author, Exeter, 1793.  Wallis 227 CAR, ??NX.  Includes a number of straightforward problems covered here, but I have only entered the more unusual examples.

Carroll-Collingwood.  1899.

                                    The Lewis Carroll Picture Book.  Stuart Dodgson Collingwood, ed.  T. Fisher Unwin, London, 1899.  = Diversions and Digressions of Lewis Carroll, Dover, 1961.  = The Unknown Lewis Carroll, Dover, 1961(?).  Reprint, in reduced format, Collins, c1910.  The pagination of the main text is the same in the original and in both Dover reprints, but is quite different than the Collins.  I will indicate the Collins pages separately.  The later Dover has 42 additional photographs.

Carroll-Gardner.  c1890?  or  1996

                                    Martin Gardner.  The Universe in a Handkerchief.  Lewis Carroll's Mathematical Recreations, Games, Puzzles and Word Plays.  Copernicus (Springer, NY), 1996.  As with Carroll-Wakeling, Carroll material will be dated as 1890?, but there is much material by Gardner which is dated 1996.

Carroll-Wakeling.  c1890?

                                    Lewis Carroll's Games and Puzzles.  Newly Compiled and Edited by Edward Wakeling.  Dover and the Lewis Carroll Birthplace Trust, 1992.  This is mostly assembled from various manuscript sheets of Carroll's containing problems which he intended to assemble into a puzzle book.  Wakeling has examined a great deal of such material, including a mass of Carroll's notes to Bartholomew Price (1818‑1898) who was Sedleian Professor of Natural Philosophy at Oxford in 1853-1898.  Price was at Pembroke College, becoming the Master, adjacent to Carroll's Christ Church.  He had tutored Carroll (1833‑1898) and they were close friends and in continual contact until their deaths, both in 1898.  However, few of the papers are dated and they are simply loose sheets with no indication of being in order, so there is no way to date the undated sheets and I have given a fairly arbitrary date of c1890? for these, though Carroll was more active before then rather than after.  Some items are taken from Carroll's youthful magazines or his correspondence and hence are more precisely dated.  The correspondence is more fully given in Carroll-Collingwood.

                    In response to an inquiry, Wakeling wrote on 28 May 2003 and said that some of the Carroll-Price notes were typewritten 'probably using Dodgson's Hammond typewriter, purchased in 1888.'  This gives a somewhat more precise dating than my c1890? and I will give:  1888 to 1898  for such items, unless there is other evidence.

Carroll-Wakeling II.  c1890?

                                    Rediscovered Lewis Carroll Puzzles.  Newly Compiled and Edited by Edward Wakeling.  Dover, 1995.  See the notes to Carroll-Wakeling, above.

Cassell's.  1881.

                                    Cassell's Book of In‑Door Amusements, Card Games, and Fireside Fun.  Cassell, Peter, Gilpin & Co., London, 1881; Cassell, London, 1973.  217pp [probably + 1p + 6pp Index] (pp. 1-8 are preliminary matter).  [There was a companion volume: Cassell's Book of Sports and Pastimes.  In 1887, the two were  combined, with the spine titled Cassell's Book of Outdoor Sports and Indoor Amusements.  The front cover says Out Door Sports, the back cover says Indoor Amusements, while the title page says Cassell's Book of Sports and Pastimes.  It contains all the main text of Book of In‑Door Amusements, ..., advanced by 744 pages.  From at least 1896, Card Games and Parlour Magic were completely revised and later there were a few other small changes.  The title varies slightly.  Manson (qv) is a 1911 revision and extension to 340pp of main text.]

Catel.  Kunst-Cabinet.  1790.

                                    Peter Friedrich Catel.  Mathematisches und physikalisches Kunst-Cabinet, dem Unterrichte und der Belustigung der Jugend gewidmet.  Nebst einer zweckmässigen Beschreibung der Stücke, und Anzeige der Preise, für welche sie beim Verfassser dieses Werks P. F. Catel in Berlin zu bekommen sind.  [I.e. this is a catalogue of items for sale by post!]  Lagarde und Friedrich, Berlin & Libau, 1790.  [MUS #113.]  P. iv says he started his business in 1780.

                    There is a smaller Vol. 2, with the same title, except 'beim Verfasser dieses Werkes P. F. Catel' is replaced by 'in der P. F. Catelschen Handlung', and the publisher is F. L. Lagarde, Berlin, 1793. 

                    My thanks to M. Folkerts for getting a copy of the example in the Deutsches Museum made for me.

                    All citations are to vol. 1 unless specified.

                    Many of Bestelmeier's items are taken from Catel.  Sometimes the figure is identical (often reversed) or is a poor copy.  Texts are often copied verbatim, or slightly modified, but usually abbreviated.  E.g. Catel often explains the puzzle and this part is frequently omitted in Bestelmeier.  Bestelmeier was the successor to Catel.  Dieter Gebhardt has searched for the various editions and associated price lists of the Catel and Bestelmeier catalogues in German libraries and he and Jerry Slocum have published the details in:  Jerry Slocum  &  Dieter Gebhardt.  Puzzles from Catel's Cabinet and Bestelmeier's Magazine  1785 to 1823.  English translations of excerpts from the German Catel-Katalog and Bestelmeier-Katalog.  Intro. by David Singmaster.  History of Puzzles Series.  The Slocum Puzzle Foundation, PO Box 1635, Beverly Hills, California, 90213, USA, 1997.  I have not yet made detailed entries from this which gives precise dates for the various parts of these catalogues.

CFF.                             Cubism for Fun.  This is the Newsletter of the Nederlandse Kubus Club (NKC) (Dutch Cubists Club) which has been in English since the mid 1980s. 

Chambers -- see: Fireside Amusements.

Charades, Enigmas, and Riddles.  1859.

                                    Charades, Enigmas, and Riddles.  Collected by A Cantab. [BLC gives no author.  "A Cantab." was a common pseudonym.  One such author of about the right time and nature was George Haslehurst.]  (Cambridge, 1859). 

                    3rd ed., J. Hall and Son, Cambridge, 1860, HB.  Half-title, 6 + 96pp. 

                    4th ed., Bell & Daldy, London, 1862.  8 preliminaries (i = half-title; FP facing iii = TP; v-viii = Introduction; errata slip; two facing plates illustrating a charade for Harrowgate [sic] Waters), 1-160, 32pp publisher's ads, dated Jan 1863; (my copy is lacking pp. 63-64).  The three plates are signed J.R.J.  This is a substantial expansion of the 3rd ed.

                    I also have photocopy of part of the 5th ed., Bell and Daldy, London, 1865, and this shows it was even larger than the 4th ed, but most of the problems of interest have the same or similar problem numbers in the three editions that I have seen.  I will cite them as in the following example.  1860: prob. 28, pp. 59 & 63; 1862: prob. 29, pp. 135 & 141; 1865: prob. 573, pp. 107 & 154.

Chaturveda. Chaturveda Pŗthudakasvâmî [NOTE:  ŗ  denotes an  r  with an underdot.].  Commentator on the Brahma‑sphuta‑siddhanta (qv), 860.  Some of these comments are taken from Bhaskara I in 629.  Shukla calls him Pŗthūdaka, but Colebrooke cites him as Ch.

Chessics.                       Chessics.  The Journal of Generalised Chess.  Produced by G. P. Jelliss, 5 Biddulph Street, Leicester, LE2 1BH.  No. 1 (Mar 1976) -- No. 29 & 30 (1987).  Succeeded by G&PJ.

Child.  Girl's Own Book.

                                    Mrs. L. Maria Child [= Mrs. Child  = Lydia Maria Francis, later Child].  The Girl's Own Book.  The bibliography of this book is confused.  According to the Opies [The Singing Game, p. 481], the first edition was Boston, 1831 and there was a London 4th ed. of 1832, based on the 2nd US ed.  However the earliest edition in the BMC is a 6th ed. of 1833.  I have examined and taken some notes from the 3rd ed., Thomas Tegg, London, 1832 -- unfortunately I didn't have time to go through the entire book so I may have missed some items of interest.  I have also examined the following.

                    Clark Austin & Co., NY, nd [back of original TP says it was copyrighted by Carter, Hendee, & Babcock in Massachusetts in 1833];  facsimile by Applewood Books, Bedford, Massachusetts, nd [new copy bought in 1998 indicates it is 4th ptg, so c1990].  The facsimile is from a copy at Old Sturbridge Village.  The back of the modern TP says the book was first published in 1834 and the Cataloguing-in-Publication data says it was originally published by Carter, Hendee and Babcock in 1834.  However, the earliest version in the NUC is Clark, Austin, 1833.  I am confused but it seems likely that Carter, Hendee and Babcock was the original publisher in Boston in 1831 and that that this facsimile is likely to be from 1833 or an 1834 reprint of the same.  The pagination is different than in the 1832 London edition I have seen.

                    The Tenth Edition, with Great Additions.  By Mrs. Child.  Embellished with 144 Wood Cuts.  Thomas Tegg, London (& three other copublishers), 1839.  12 + 307 pp + 1p publisher's ad.  Has Preface to the Second Edition but no other prefaces.  This Preface is identical to that in the 1833 NY ed, except that it omits the final P.S. of season's greetings.  The 1833 NY essentially has the same text, but they have different settings and different illustrations with some consequent rearrangement of sections.  However the main difference is that the NY ed omits 41pp of stories.  There are a number of minor differences which lead to the NY ed having 9 extra pages of material.

                    The Eleventh Edition, with Great Additions.  By Mrs. Child.  Embellished with 124 Wood Cuts.  Thomas Tegg, London (& three other copublishers), 1842.  12 + 363 pp + 1p publisher's ad.  The Preface is identical to that in the 10th ed, but omits 'to the Second Edition' after Preface.  90 pp of games and 40 pp of enigmas, charades, rebuses, etc. have been added; 56 pp of stories have been dropped.

                    The Girl's Own Book of Amusements, Studies and Employments.  New Edition.  Considerably enlarged and modernized by Mrs. L. Valentine, and others.  William Tegg, London, 1876.  This differs considerably from the previous editions.

                    I will cite the above by the dates  1832, 1833, 1839, 1842, 1876.

                    Various sources list:  13th ed., 1844 [BMC, Toole Stott 831];  Clark Austin, NY, 1845 [NUC];  16th ed., 1853 [BMC];  17th ed. by Madame de Chatelain, 1856 [BMC, NUC, Toole Stott 832];  18th ed. by Madame de Chatelain, 1858 [BMC, Toole Stott 833];  1858 [Osborne Collection (at Univ. of Toronto)];  rev. by Mrs. R. Valentine, 1861 [BMC, Osborne Collection];  rev. by Mrs. R. Valentine, 1862 [BMC, NUC];  rev. by Mrs. R. Valentine, 1864 [BMC];  rev. by Mrs. R. Valentine, 1867 [BMC];  enlarged by Mrs. L. Valentine, 1868 [NUC];  enlarged by Mrs. L. Valentine, 1869 [BMC];  enlarged by Mrs. L. Valentine, 1873 [NUC];  enlarged by Mrs. L. Valentine, 1875 [NUC];  enlarged by Mrs. L. Valentine, 1876 [BMC];

                    Heyl gives the following under the title  The Little Girl's Own Book:  Carter, Hendee and Co., Boston, 1834;  American Stationers Co, John B. Russell, Boston, 1837;  Edward Kearney, NY, 1847;  NY, 1849.

                     I think there were at least 33 editions.  See my  The Bibliography of Some Recreational Mathematics Books  for more details.  Cf Fireside Amusements, below, which is largely copied from Child.

Chiu Chang Suan Ching.  c-150?

                                    Chiu Chang Suan Ching (Nine Chapters on the Mathematical Art).  (Also called Chiu Chang Suan Shu and variously transliterated.  The pinyin is Jiŭ Zhāng Suàn Shù.)  c‑150?  German translation by K. Vogel; Neun Bücher arithmetischer Technik; Vieweg, Braunschweig, 1968.  My citations will be to chapter and problem, and to the pages in Vogel.  (Needham said, in 1958, that Wang Ling was translating this, but it doesn't seem to have happened.)  Some of the material dates from the early Han Dynasty or earlier, say c-200, but Chap. 4 & 9, the most original of all, have no indication of so early a date.  A text of c50 describes the contents of all the chapters and Høyrup suggests that Chap. 4 & 9 and the final assembly of the book should be dated to the [early] 1C.

Christopher.  1994.

                                    Maurine Brooks Christopher & George P. Hansen.  The Milbourne Christopher Library.  Magic, Mind Reading, Psychic Research, Spiritualism and the Occult  1589-1900.  Mike Coveney's Magic Words, Pasadena, 1994.  1118 entries.  References are to item numbers.

Christopher II.  1998.

                                    Maurine Brooks Christopher & George P. Hansen.  The Milbourne Christopher Library -- II.  Magic, Mind Reading, Psychic Research, Spiritualism and the Occult  1589-1900.  Mike Coveney's Magic Words, Pasadena, 1998.  3067 entries.  References are to item numbers.  Recently received, ??NYR.

Chuquet.  1484.             Nicolas Chuquet.  Problèmes numériques faisant suite et servant d'application au Triparty en la science des nombres de Nicolas Chuquet Parisien.  MS No. 1346 du Fonds Français de la Bibliothèque Nationale, 1484, ff. 148r-210r.  Published in an abbreviated version as: Aristide Marre; Appendice au Triparty en la science des nombres de Nicolas Chuquet Parisien; Bulletino di bibliografia e di storia delle scienze matematiche e fisiche 14 (1881) 413‑460.  (The first part of the MS was published by Marre; ibid. 13 (1880) 593-814; ??NYS)  Marre generally transcribes the text of the problem, but just gives the answer without any of the text of the solution.  I will cite problems by number.  There are 166 problems.  (Much of this was used in his student's book: Estienne de la Roche; Larismethique novellement composee par maistre Estienne de la roche dict Villefrāche; Lyons, 1520, ??NYS.  (Rara 128‑130).) 

                    FHM        Graham Flegg,  Cynthia Hay  &  Barbara Moss.  Nicolas Chuquet, Renaissance Mathematician.  A study with extensive translation of Chuquet's mathematical manuscript completed in 1484.  Reidel, Dordrecht, 1985.  This studies the entire MS, of which the above Appendice is only the second quarter.  It often gives a full English translation of the text of the problem and the solution, but it may summarize or skip when there are many similar problems.  The problems in the first part of the MS are not numbered in FHM.  I will cite this as FHM xxx, where xxx is the page number, with 'English in FHM xxx' when the problem is explicitly translated.

Clark.  Mental Nuts.  1897, 1904, 1916.

A book of Old Time Catch or Trick Problems  Regular old Puzzlers that kept your Grandad up at night.  Copyright, 1897, by S. E. Clark, Philadelphia.  Flood & Conklin Co.  Makers of Fine Varnishes, Newark, N.J.  100 problems and answers.  32pp + covers.

A book of 100 Catch or Trick Problems  Their simplicity invites attack, while their cunningly contrived relations call forth our best thought and reasoning.  Copyright, 1897, by S. E. Clark, Philadelphia.  Revised 1904 Edition.  Waltham Watches, Waltham, Massachusetts.  This was an promotional item and jewellers would have their address printed on the cover.  My example has:  With the compliments of  J. H. Allen  Jeweler [sic]  Shelbina, Mo.  Thanks to Jerry Slocum for this.   In fact there are only 95 problems; numbers 68, 75, 76, 78, 84 are skipped.  32pp + covers.

Revised Edition 1916, with no specific company mentioned.  Enlarged PHOTOCOPY from Robert L. Helmbold.  100 numbered problems, but some figures inserted after no. 75 are the solutions to a problem in the other editions and I have counted this as a problem (no. 75A), making 101 problems.  28pp + covers.

                     The editions are considerably different.   Only 40 problems occur in all three editions.  There are 50 problems common to 1897 and 1904, 42 common to 1897 and 1916 and 71 common to 1904 and 1916, though this counting is a bit confused by the fact that problems are sometimes combined or expanded or partly omitted, etc.  Solutions are brief.  It includes a number of early examples or distinct variants, which is remarkable for a promotional item.  I have entered 36 of the 1897 problems plus 13 of the 1904 problems not in 1897 and 7 of the 1916 problems not in 1897 or 1904.  Many others are standard examples of topics covered in this work, but are not sufficiently early to be worth entering. 

                     I originally had the 1904 ed and cited the 1904 problems as 1897 on the grounds that editions of this period do not change much, but having now seen the 1897 and 1916 eds, I realise that the editions are very different, so I will cite the actual dates.  Since only the 1897 version is paginated, I will just cite problem numbers; the solutions are at the back.

Clarke, William.  See:  Boy's Own Book.

CM.                             Crux Mathematicorum (originally titled Eureka until 4:3)

CMJ.                            The College Mathematics Journal.  Before the early 1980s, this was the Two Year College Mathematics Journal.

Colebrooke.  1817.

                                    Henry Thomas Colebrooke (1765-1837), trans.  Algebra, with Arithmetic and Mensuration from the Sanscrit of Brahmegupta and Bháscara.  John Murray, London, 1817.  Contains Lîlâvatî and Bîjaganita of Bhâskara II (1150) and Chapters XII (Arithmetic) and XIII (Algebra) of the Bráhma‑sphuta‑siddhânta of Brahmagupta (628).  There have been several reprints, including Sändig, Wiesbaden, 1973.  (Edward Strachey produced a version: Bija Ganita: or the Algebra of the Hindus; W. Glendinning, London, 1813; by translating a Persian translation of 1634/5.)

Collins.  Book of Puzzles.  1927.

                                    A. Frederick Collins.  The Book of Puzzles.  D. Appleton and Co., NY, 1927. 

Collins.  Fun with Figures.  1928.

                                    A. Frederick Collins.  Fun with Figures.  D. Appleton and Co., NY, 1928. 

Columbia Algorism.  c1350.

                                    Anonymous Italian MS, c1350 [according to Van Egmond's Catalog 253‑254], Columbia X511 .A1 3.  Transcribed and edited by K. Vogel; Ein italienisches Rechenbuch aus dem 14.Jahrhundert; Veröffentlichungen des Forschungsinstituts des Deutschen Museums für die Geschichte der Naturwissenschaften und der Technik, Reihe C, Quellentexte und Übersetzungen, Nr. 33, Munich, 1977.  My page references will be to this edition.  Van Egmond says it has a title in a later hand: Rascioni de Algorismo.

               The Algorism is discussed at length in Elizabeth B. Cowley; An Italian mathematical manuscript; Vassar Medieval Studies, New Haven, 1923, pp. 379‑405.

Conway, John Horton. (1937-  ).  See:  Winning Ways.

Cowley, Elizabeth B.  See:  Columbia Algorism.

CP.  1907.                    H. E. Dudeney.  Canterbury Puzzles.  (1907);  2nd ed. "with some fuller solutions and additional notes", Nelson, 1919;  4th ed.  = Dover, 1958.  (I have found no difference between the 2nd and 4th editions, except Dover has added an extra note on British coins and stamps.  I now have a 1st ed, which has different page numbers, but I have not yet added them.)

CR                               Comptes Rendus des Séances de l'Académie des Sciences, Paris.

Crambrook.  1843.        W. H. M. Crambrook.  Crambrook's Catalogue of Mathematical & Mechanical Puzzles Deceptions and Magical Curiosities, contained in the Necromantic Tent, Royal Adelaide Gallery, West Strand, London.  ...  To which is added, a Complete Exposé [of] the Baneful Arts by which unwary Youth too often become the prey of professed gamesters.  And ... an extract from The Anatomy of Gambling.  Second Edition, Corrected & Enlarged.  T. C. Savill, 107 St. Martin's Lane, 1843.  23pp.  Photocopy provided by Slocum.  [According to: Edwin A. Dawes; The Great Illusionists; Chartwell Books, Secaucus, New Jersey, 1979, p. 138, this is the first known magical catalogue.  It has a list of about 100 puzzles on pp. 3-5, with the rest devoted to magic tricks.  Unfortunately there are no pictures.  Comparison with Hoffmann helped identify some of the puzzles, but I can not identify many of them.  I have marked almost all these entries with ?? or check??, but the only way one can check is if actual examples or an illustrated catalogue turn up.  Some of the names are so distinctive that it seems certain that the item does fit where I have cited it; others are rather speculative.  There are several names which may turn up with more investigation.  Toole Stott 190 says there should be 48pp, though the later pages may be the added material on gambling.]

Cremer, William Henry, Jr.  See under:  Book of 500 Puzzles,  Hanky Panky,  Magician's Own Book.

CUP.                            Cambridge University Press.

Cyclopedia.  1914.

                                    Sam Loyd's Cyclopedia of 5,000 Puzzles, Tricks and Conundrums (ed. by Sam Loyd Jr).  Lamb Publishing, 1914  = Pinnacle or Corwin, 1976.  This is a reprint of Loyd's "Our Puzzle Magazine", a quarterly which started in June 1907 and ran till 1908.  See OPM for details.

C&B.  1920. Sidney W. Clarke  &  Adolphe Blind.  The Bibliography of Conjuring  And Kindred Deceptions.  George Johnson, London, 1920.  Facsimile by Martino Fine Books, Mansfield Centre, Connecticut, nd [obtained new in 1998].

C&W.          Chatto & Windus, London.

 

Datta & Singh.              Bibhutibhusan Datta & Avadhesh Narayan Singh.  History of Hindu Mathematics.  Combined edition of Parts I (1935) and II (1938), Asia Publishing House, Bombay, 1962.  NOTE: This book makes some contentious assertions.  Readers are referred to the following reviews.

                    O. Neugebauer.  Quellen und Studien zur Geschichte der Mathematik 3B (1936) 263-271.

                    S. Gandz.  Isis 25 (1936) 478-488.

Datta, B.  See:  Bakhshali MS;  Datta & Singh.

De Morgan (1806-1871).  See:  Rara.

De Viribus.  See:  Pacioli.

dell'Abbaco.  See:  Pseudo-dell'Abbaco.

Depew.  Cokesbury Game Book.

                                    Arthur M. Depew.  The Cokesbury Game Book.  Abingdon-Cokesbury Press, NY & Nashville, 1939.  [The back of the TP says it is copyright by Whitmore & Smith -- ??  The Acknowledgements say material has been assembled from various sources and colleagues who have been collecting and writing over the previous thirty years.]

Dickson.                       Leonard Eugene Dickson (1874-1954).  History of the Theory of Numbers, 3 vols.  Carnegie Institution of Washington, Publication 256, 1919-1923;  facsimile reprint by Chelsea, 1952.

Dilworth.  Schoolmaster's Assistant.

                                    Thomas Dilworth.  The Schoolmaster's Assistant:  Being a Compendium of Arithmetic, both Practical and Theoretical.  (1743);  11th ed., Henry Kent, London, 1762 (partly reproduced by Scott, Foresman, 1938.)  20th ed., Richard & Henry Causton, London, 1780.  De Morgan suggests the 1st ed. was 1744 or 1745, but the testimonials are dated as early as Jan 1743, so I will assume 1743.  Comparison of a 1762 ed. (Wallis 321 DIL) with my 1780 ed. shows the 1780 ed. is identical to the 1762 ed., except the section on exchange is much expanded, so the page numbers of all material of interest are increased by 12pp.  I will cite the pages of the 1762 ed., but give the date as 1743.  [Wallis also has:  14th ed., 1767;  15th ed., 1768;  1783;  22nd ed., 1785;  1791;  24th ed., 1792;  1793;  33rd ed., 179-;  1799;  1800;  1804.]  [Halwas 149‑162 are some US editions.]

Diophantos.  c250.

                                    Diophantos.  Arithmetica.  c250.  In: T. L. Heath; Diophantos of Alexandria; 2nd ed., (OUP, 1910); Dover, 1964.  Note: Bachet edited a Greek and Latin version of Diophantos in 1620, which inserted 45 problems from the Greek Anthology at the end of Book V.  (It was in Fermat's copy of this work that Fermat wrote the famous marginal note now called his Last Theorem; Fermat's son published an edition with his father's annotations in 1670, but the original copy was lost in a fire.)

DNB.                           Leslie Stephen, ed.  The Dictionary of National Biography.  Smith, Elder and Co., London, 1885‑1901 in 22 volumes.  OUP took it over in 1917.  Decennial Supplements were added. 

                    Compact Edition, with Supplement amalgamating the six decennial supplements to 1960, OUP, 1975.  The Compact Ed. shows the original volumes and pages so I will cite them in ( ), followed by the pages in the Compact Ed.

Dodson.  Math. Repository.  (1747?); 1775.

                                    James Dodson.  The Mathematical Repository.  Containing Analytical Solutions of near Five Hundred Questions, mostly selected from Scarce and Valuable Authors.  Designed As Examples to Mac-Laurin's and other Elementary Books of Algebra; And To conduct Beginners to the more difficult Properties of Numbers.  2nd ed., J. Nourse, London, 1775, HB.  (I have now acquired vols. II & III (1753 & 1755), but these are largely concerned with annuities, etc., except the beginning of vol. II has a section on indeterminate equations, entered in 7.P.1.  From references in these volumes, it seems that the 1775 ed. of volume I is pretty close to the first ed. of c1747, but has been a little rearranged, so I have redated the entries as above.)

Doubleday -- n.  1969, etc.

                                    Eric Doubleday.  Test Your Wits, Vols. 1 - 5.  Ace Publishing, NY, 1969; 1971; 1972; 1969[sic]; (1969), revised 1973.  [Vols. 1 - 3 are good collections, with a number of novel variations of standard problems.  Vols. 4 & 5 are vol. 1 split into two parts and much padded by putting each answer on a separate page!  The books refer to Doubleday as puzzle setter for a London newspaper and one of the best known setters in the English speaking world.  However, none of the older puzzle setters/editors in England have ever heard of him and there is no book by him in the British Library Catalogue.  Surprisingly, there is also no book by him in the Library of Congress Catalogue!  I am beginning to think the author is a deception, but the first three books are better than scissors and paste hack work.]

DSB.                            Dictionary of Scientific Biography.  Ed. by Charles C. Gillespie for the American Council of Learned Societies.  Scribner's, NY, 1970-1977, in 18 volumes.  I will give the volume and the pages. 

                    The mathematical material has been reprinted in four volumes as:  Biographical Dictionary of Mathematicians  Reference Biographies from the Dictionary of Scientific Biography.  Scribner's, NY, 1990?  This has new pagination, continuous through the four volumes.  If I don't have the DSB details, I will cite this as BDM.

Dudeney, Henry Ernest (1857-1930).  See:  AM,  CP,  MP,  PCP,  536.  I also cite his columns or contributions in The Captain, Cassell's Magazine, Daily Mail, London Magazine, The Nineteenth Century, The Royal Magazine, Strand Magazine, Tit-Bits, The Tribune, The Weekly Dispatch.

 

Eadon.  Repository.  1794.

                                    John Eadon.  The Arithmetical and Mathematical Repository, Being a New Improved System of Practical Arithmetic, in all its Branches; Designed for the Use of Schools, Academies, Counting-Houses, and Also for the Benefit of private Persons who have not the Assistance of a Teacher.  In Four Volumes.  Volume 1.  In Three Books.  Printed for the author, and sold by G. G. and J. Robinson, Paternoster Row, London, 1794.

 

EB                               Encyclopædia Britannica.  I tend to use my 1971 ed.

Endless Amusement I.  c1818.

                                    Anonymous.  Endless Amusement; A Collection of Nearly 400 Entertaining Experiments  In various Branches of Science; ..., All the Popular Tricks and Changes of the Cards, ....  3rd ed.(?), Thomas Boys, London, nd [1825].  Frontispiece & TP are missing, but James Dalgety has inserted a photocopy of the TP of the 3rd ed.  [BMC lists  2nd ed., 1819?;  3rd ed., 1825?  BMC65 lists 1st ed. by Thorp & Burch, c1818;  2nd ed., c1820.  Toole Scott 255-267 lists  3rd ed., c1820;  4th ed., 1822;  5th ed., 1830;  6th ed., 1834;  7th ed., 1839.  Hall, BCB 116-123 are:  1st ed., c1815;  2nd ed., c1820;  3rd ed., c1820;  3rd ed., Philadelphia, 1822;  4th ed., c1825;  5th ed., c1830;  6th ed., 1834;  Philadelphia, 1847.  Heyl 110-115, 121 are  1819;  2nd & 3rd ed., M. Carey & Sons, Philadelphia, 1821 & 1822;  3rd ed, C. Tilt, London, 1825;  Borradaile, NY, 1831;  Henry Washburne & Thomas Tegg, London, 1839;  Lea & Blanchard, Philadelphia, 1847.  Almost all of these are listed as 2 + 216 pp, so the editions are probably all the same as the 1st ed., except that Hall notes that the 1st ed. title is slightly different: Endless Amusement; A Collection of Upwards of 400 Entertaining and Astonishing Experiments.  Among a Variety of other Subjects, are Amusements in Arithmetic, Mechanics, Hydraulics ....  All the Popular Tricks and Changes of the Cards, ...,  and Heyl gives a similar title for the 1825 ed. and the 1831 NY ed. has some variations.  Christopher 330-338 are  2nd ed., Philadelphia, 1821;  3rd ed., c1820;  3rd ed., Philadelphia, 1822;  4th ed., c1822;  4th ed., c1822 (slightly different to preceding;  5th ed., c1830;  6th ed., 1834;  7th ed., 1839;  Philadelphia, 1847.  C&B list it under Thorp and Burch, the publishers, with no dates.]  [There is a Recreations in Science, ..., by the author of Endless Amusement, 1830.]

Endless Amusement II.  1826?

                                    Anonymous.  A Sequel to the Endless Amusement, Containing Nearly Four Hundred Interesting Experiments, In various Branches of Science, ..., to Which are Added, Recreations with cards, and a Collection of Ingenious Problems.  Thomas Boys, London, nd [1826?].  Pp. 203-216 are missing, but James Dalgety has inserted photocopies.  [BMC lists one ed., 1826?  Hall, BCB 252 gives c1825.  Heyl says this refers to Thomas Boys ... and Thorp and Burch, London (1825).  Toole Stott 623 gives 1825.]

                                    = Anonymous.  The Endless Amusement.  New Series Containing Nearly Four Hundred Interesting Experiments, ... (as above).  Thomas Tegg & Sons, London, 1837.  Angela Newing has provided a photocopy of the interesting parts of this and it is virtually identical to the 1826? ed., though it has been reset, resulting in an extra word fitting on some lines, and it has rather poorer pictures.  One problem has been replaced by another.  [Heyl 122.]  21 problems, including the replacement problem, are copied in The New Sphinx.

                                    = Anonymous.  A Companion to the Endless Amusement.  James Gilbert, London, 1831.  [Toole Stott 172 says this is a reprint of A Sequel ..., from the same type, with new TP, and this is clear from examination of the example Wallis 187.5 COM.  Heyl 66 dates it as c1820?]

                                    Some of the material is taken from Badcock.

 

van Etten/Leurechon.  1624.

                                    Recreation Mathematicque.

                    The bibliography of this book is very complicated.  I have now made a separate bibliography of this, augmented by many contributions from Voignier, which is now (Aug 2001) 19pp, listing 50 French editions, 5 English editions, 4 Latin editions and 8 (or 9) Dutch editions -- a total of 67 (or 68) editions from 1624 to 1706, though at least 10 of the French editions may be 'ghosts'.  This is part of my  The Bibliography of Some Recreational Mathematics Books.

                    This book has traditionally been attributed (since 1643) to Père Jean Leurechon, SJ (c1591-1670), who was probably van Etten's university teacher, but the book specifically names van Etten and there seems to be very little real evidence for Leurechon's authorship.  Trevor Hall's booklet and chapter are a substantial study of this question and he concludes that there is no real reason to doubt van Etten's authorship, though he may well have had help or inspiration from his teacher.  Hall has also shown that van Etten and his uncle, the book's dedicatee, were real people.  (Toole Stott 429‑431 dismisses Hall's work as a result of completely misunderstanding it!)  However, the book was amended, revised and translated many times, so that versions may occur under the following names:  Hendrik van Etten;  Jean Leurechon;  D.H.P.E.M.  = Denis (or Didier) Henrion, professeur en mathématique (or Professeur ès Mathematiques), a pseudonym of Clément Cyriaque de Mangin, who also called himself Pierre Hérigone;  Claude Mydorge;  Caspar (or Gaspar) Ens;  Wynant van Westen;  William Oughtred;  William Leake; not to mention Anonymous and versions of the title -- I have found it under Recreations or Récréations and under Vermakelijkheden.  E.g. Lucas, RM1, 239-240, has 7 entries for this book under five different authors and twice under Récréations.

                    Jacques Voignier; Who was the author of "Recreation Mathematique" (1624)?; The Perennial Mystics #9 (1991) 5-48 (& 1-2 which are the cover and its reverse).  [This journal is edited and published by James Hagy, 2373 Arbeleda Lane, Northbrook, Illinois, 60062, USA.]  This is the second serious study of this book.  He points out evidence for Leurechon's connection with the book, which makes it seem more likely, but definite evidence is still lacking, so I am suggesting that it may have been some kind of joint production and I will change my references it to  van Etten/Leurechon.  The work of Hall and Voignier form the basis of the following discussion, supplemented by the standard catalogues and personal inspection of about a half of the French and English editions -- generally after 1630.

 

                    Henrik van Etten.  Recreation Mathematicque.  Composee de Plusieurs Problemes Plaisants et Facetieux.  En faict d'Arithmeticque Geometrie, Mechanicque, Opticque, & autres parties de ces belles sciences.  Jean Appier Hanzelet, Pont‑a‑Mousson, 1624 [taken from facsimile of the 1626 ed.].  155pp., ??NYS.

                    2nd ed., 1626, ibid.  = recent facsimile with no details, but with 'Pont à Mousson  13 ‑ 10 ‑ 54' written inside the back cover.  [An apparently identical copy at the Museum of the History of Science, Oxford, has a small insert saying it was reissued by La Compagnie de Pont-à-Mousson, printed by l'Imprimerie Berger-Levrault, nd.].  91 problems on xiv + 144 = 158pp.  [The extra pages include questions V and VI of problem 91 -- these questions occur in no other edition, except probably in the 1629 reissue in Pont-à-Mousson.] 

 

                    After the two Hanzelet editions, there were three editions in Paris in 1626, by Rolet Boutonné (2nd ed.), by Antoine Robinot (2nd ed.) and by Jean Moreau & Guillaume Loyson (3rd ed.).  Boutonné and Robinot were closely associated and their output was interchangeable.  Their 2nd eds. appear to be essentially the 1624 ed.  The Moreau & Loyson has Notes added to the problems and 8pp. of Additions.  This was the first to put the illustrations as woodcuts in the text rather than using copperplates for five separate sheets of 8 figures.  The Notes are signed D.A.L.G., but are due to Claude Mydorge.  (NUC indicates the Robinot had further comments signed D.H.P.E.M., later identified as Denis (or Didier) Henrion Professeur En Mathématique (though Henrion is a pseudonym of Clément Cyriaque de Mangin!) -- but this seems to be a confusion.)  In the next few years, editions appeared in Paris, Rouen and Lyon.  In 1627, Boutonné issued a '4th ed.' with "Nottes sur les recreations mathematiques ... Par D. H. P. E. M." and the D.A.L.G. notes were omitted.  In 1627, Claude Rigaud & Claude Obert, Lyon, issued a version with 9pp of Additions as in the 1626 Moreau & Loyson.

                    In 1628, Charles Osmont, Rouen, issued a version in three parts: Récréations mathématiques ... 1re et 2de partie.  La 3e partie contient un recueil de plusieurs gentilles et récréatives inventions de feux d'artifice ....  Part 1 was van Etten's 91 problems, with questions V & VI of prob. 91 omitted, omitting the D.A.L.G. notes and the Additions.  Part 2 had 45 new problems, often attributed to Mydorge and/or Henrion, but they had no connection with this and the authorship of these problems is unknown, though Voignier suggests the printer, Osmont.  Part 3 is an independent treatise on fireworks which Hall attributes to Hanzelet.  This edition was reissued in Rouen by various publishers in 1628, 1629, 1630, 1634, 1638 and in Lyon in 1642-1643, 1653, 1656, 1658, 1669, 1680.

                    In 1630, Boutonné and Robinot (their printing is indistinguishable and volumes often have parts from both of them, indeed the Privilege is issued to them jointly) issued an extended version in four parts, titled Examen du Livre des Recreations Mathematiques, stated to be by Mydorge.  Part 1 is van Etten's 91 problems, with parts V & VI of prob. 91 omitted, with many problems being followed by an Examen signed D.A.L.G.  These are by Mydorge and are a revision of his material in the 1626 Moreau & Loyson.  Part 2 has its own TP and had 45 new problems, taken from the 1628 Rouen ed.  Part 3 has its own TP, but doesn't state the publisher, and is the independent treatise on fireworks, also taken from the 1628 Rouen ed.  Part 4 again has its own TP and is Nottes [sic] sur les Recreation Mathematiques by D.H.P.E.M. and are additions to 27 of van Etten's problems, taken or extended from the Notes in the 1627 4th ed.  The book is also described as 3 parts with the Nottes, but Parts 2 and 3 are consecutively paged, leading to some descriptions of the book as being in 3 parts.  The parts were probably issued separately as they sometimes are catalogued separately and different copies of the whole work often have a mixture of the Boutonné and Robinot printings.  This most extended form was reissued by various publishers in Paris: 1634(??), 1638, 1639,  and in Rouen: 1639 (two publishers), 1643 (two publishers ??), 1648?, 1649?

                    In 1659, Cardin Besonge, Paris, issued  Les Récréations mathématiques, avec l'examen de ses problèmes en arithmétique, géométrie, ....  Premièrement reveu par D. Henrion, depuis par M. Mydorge, et tout nouvellement corrigé et augmenté, 5e et dernière édition.  The Nottes are incorporated in the text (or perhaps omitted??).  The entire text is consecutively page-numbered.  Reissued in Paris: 1660, 1661 and in Rouen as the 6th ed.: 1660?, 1663?, 1664, 1669 (seven publishers!).

 

                    The 1630 Paris ed. and the 1626 ed. have the same problem numbers for the first 91 problems, as do almost all French editions.  I will cite the problem number and the pages of the 1626 ed.  I will add reference to the 1630 Paris ed., when the latter has additional information.  Only one of the additional problems in part 2 (prob. 7) is of any interest to us, but several of Henrion's Nottes give corrections, extensions, additional references and even additional problems.  I didn't find any of the D.A.L.G. notes of any interest.

 

     English editions.

                    Mathematicall Recreations.  Or a Collection of sundrie [1653 has: many] Problemes, extracted out of the Ancient and Moderne Philosophers, as secrets in nature, and experiments in Arithmeticke, Geometrie, Cosmographie, Horologographie, Astronomie, Navigation, Musicke, Opticks, Architecture, Staticke, Machanicks, Chimestrie, Waterworkes, Fireworks, &c.  Not vulgarly made manifest untill this time: Fit for Schollers, Students, and Gentlemen, that desire to know the Philosophicall cause of many admirable Conclusions.  Usefull for others, to acuate and stirre them up to the search of further knowledge; and serviceable to all for many excellent things, both for pleasure and Recreation.  Most of which were written first in Greeke and Latine, lately compiled in French, by Henry Van Etten Gent.  And now delivered in the English tongue, with the Examinations, Corrections, and Augmentations.  Printed by T. Cotes for Richard Hawkins, London, 1633.  328pp.  (This ed. is 'excessively rare'.) 

                    Reissued by William Leake, London: 2nd ed., 1653; (1667(??)); 1674.  344pp.  [Hall, OCB, says the 2nd ed. is similar to the 1633 edition, but with an extra 16pp description (of 1636) of some dials by Oughtred (which led to the book or the translation often being attributed to Oughtred).  Hall also states that the English editions are based on the Rouen ed. of 1628.  Sadly some interesting problems were omitted in the English, leading to confusion in plate numbers.  However, I have just noticed that Prob. 63 is about two pages longer than the corresponding Prob. 70 of the French editions.]  The 1633, 1653 and 1674 editions are identical except for the additional English material in the later editions.  I will add citations to the English editions in parentheses.  I now have an imperfect copy of the 1674 ed, covers missing and lacking the Frontispiece and pp. 273-282 and later material.  Heyl 311 is a 1753 ed., which must be an error for 1653.

                    W. Leybourn, qv, takes several sections directly from the English editions.

 

     Latin editions.

                    In 1628(??), Caspar (or Gaspar) Ens made a Latin translation but added some other material, e.g. 49 problems from Alcuin.  I have only studied the 1636 ed. carefully.

                    Thaumaturgus Mathematicus, Id est, Admirabilium Effectorum e Mathematicarum Disciplinarum Fontibus Profluentium Sylloge.  Casparo Ens L. Collectore & Interprete.  1628.  [Taken from 1636 TP.  MUS #30 says this is only a translation of van Etten.  There is some doubt whether the 1628 edition exists!]

                    Reissued in 1636 and 1651.  It has 89 of van Etten's problems (omitting 38 & 46) and adding 25 new problems, with some numbering errors so the last is numbered 113.  This is followed by 55 problems of Alcuin, using the Bede version of 56 problems, but omitting 18.

                    Thaumaturgus mathematicus Gasparo Ens lectore collectore, & interprete, Nunc denuò Correctus, & Auctus.  Apollonius Zambonus, Venice, 1706.  113 problems + 49 from Alcuin (check??).  There are some differences between this and the 1636 ed.

                    [MUS #30 gives Köln, 1651, and further editions.]

 

     Dutch editions.

                    In 1641, Wynant van Westen translated van Etten into Dutch.  The title is: Het eerste [- derde] deel van de Mathematische vermaecklyckheden.  Te samen ghevoeght van verscheyden ghenuchlijcke ende boertige werckstucken, soo uyt arithmetica, geometria, astronomie, geographia, cosmographia, musica, physica, optica, catoptrica, architectonica, sciotetica, als uyt andere ongehoorde mysterien meer.  Ghetranslateert uyt het fransch in nederduytsche tale: ende verrijckt, vermeerdert, ende verbetert met verscheyden observatien ende annotatien, dienende tot onderrichtinge van eenige duystere questien, ende mis-slaghen in den franschen druck.  Door Wynant van Westen .... op nieus oversien verbetert.  Jacob van Biesen, Arnhem, 1641.  3 parts with separate title pages and pagination, perhaps in 3 vols, but later in 1 vol. 

                    This was reissued:  Van Biesen, Arnhem, 1641, ??, 1644, 1662, 1671-72; Lootsman and Jacobsz, Amsterdam, 1673.  I haven't examined any of these.

 

Euler.  Algebra.  1770.

                                    Leonard Euler (1707-1783).  Vollständig Anleitung zur Algebra.  Royal Academy of Sciences at Petersburg, 1770.  [A Russian translation appeared in 1768.]  Translated into French by John III Bernoulli, with additions by Bernoulli and La Grange (pp. 463-593 here), 1774.  Translated from French into English as Element of Algebra, with further notes, by Rev. John Hewlett, with a Memoir of Euler by Francis Horner [Horner actually did the translation; Hewlett edited it.], (1797),  5th ed., Longman, Orme, and Co., London, 1840.  Reprinted, with Introduction by C. Truesdell (1972), omitting 4 pp of Horner, Springer, NY, 1984 [hidden on back of title page].  I will cite part, section, chapter, article and the pages from the Springer ed.  (Part II has no sections.)  Unfortunately these numbers seem to have little connection with other editions.  [Though most of the recreational material in Euler is much older than Euler, I have included it as a representative 18C text.]  [Halwas 175-176 are some US editions -- the 1818 edition was the first example of a translated algebra in the US.]

 

Family Friend.               The Family Friend.  This was a magazine founded by Robert Kemp Philp in 1849.  The dating is awkward -- vol. 1 is dated 1850 on the cover, but the Preface is dated 15 Nov 1849 and refers to the success of the past year, when it appeared monthly.  It also says the magazine will henceforth appear twice a month with two volumes per year, due on the first of June and December.  The Gardening section of vol. 1 goes from Jan to Dec.  The Preface of Vol. 2 is dated 10 Jun 1850 and its gardening section covers Jan - Jun.  The Preface of Vol. 3 is dated 15 Dec 1850 and its Gardening section goes Jul - Dec.  BMC shows Philp left in 1852 and the magazine continued with two volumes per year through a fifth series, ending in 1867, then restarted with one volume a year from 1870 until 1921.  I have vols. 1 - 3 & the second half of 1858, which is dated 1858-9, but appears to be Jul-Dec.  None of the text is signed.  At the back of volumes are included answers to correspondents.  The puzzles are often identical to those in The Magician's Own Book or The Illustrated Boy's Own Treasury, etc., but are considerably earlier.

FHM.                           Graham Flegg,  Cynthia Hay  &  Barbara Moss -- see under Chuquet.

Fibonacci.                     Leonardo Pisano, called Fibonacci (c1170->1240).  Liber Abbaci.  (1202); 2nd ed., 1228.  In: Scritti di Leonardo Pisano; vol. I, ed. and pub. by B. Boncompagni; Tipografia delle Scienze Matematiche e Fisiche, Rome, 1857.  The title pages give 'abbaci', but Boncompagni's text begins 'Incipit liber Abaci ... Anno MCCII.', while the c1275 MS starts 'Incipit abbacus'.  Both forms are used, sometimes even in the same article -- e.g. Loria's biographical article, see in Section 1.

                    Richard E. Grimm was working on a critical edition of this and he kindly gave me some details.  There are 15 known MSS, all of the 1228 2nd ed.  Six of these consist of 1½ to 3 chapters only; five of the others lack Chapter 10 and the second half of Chapter 9; one lacks Chapter 10 and one lacks much of Chapter 15, leaving two essentially complete texts.  The last four MSS mentioned are the most important:  Siena L.IV.20, c1275, lacking much of Chap. 15, "the oldest and best";  Siena L.IV.21, 1463 [Grimm said c1465 -- there are dates up through 1464 in interest calculations, but the Incipit specifically says 1463], which includes much other material from later writers, so it is at least double the size of L.IV.20;  Vatican Palatino #1343, end of 13C, lacking Chap. 10;  Florence Bibl. Naz. Conventi Soppressi C. 1. 2616, early 14C, "handsome but frequently badly faded" so "that a later hand found it necessary to rewrite what he saw there."  When I examined it in Sep 1994, the black ink was indeed sometimes badly faded but the numbers were in a clear red -- perhaps these are what was rewritten??  L.IV.20 has the beginning sentence ending "et correctus ab eodem a MCCXXVIII", but Grimm says all the others are also of the 1228 ed even if they do not carry this addition or the extra initial dedication.  Sadly, I heard in Aug 1998 that Grimm had Alzheimer's disease and was in a nursing home.  Inquiry has revealed no trace of the photocopies of all the Liber Abbaci MSS which he said he had obtained and in summer 2000 I heard he had died.

                    Boncompagni used only one MS, then denoted Codex Magliabechiana, C. I, 2616, Badia Fiorentina, no. 73, now Conventi Soppressi, C. I. 2616, the badly faded fourth MS described above. 

                    In Sep 1994 and Mar 1998, I examined Siena L.IV.20 and 21 and Conv. Soppr. C.1.2616.  I have slides of the Incipit & Fibonacci numbers from all of these and some other material.

                    The dates of 1202 and 1228 are based on the Pisan calendar.

     Fibonacci-Sigler.  Liber Abaci.  Translated by Laurence E. Sigler as:  Fibonacci's Liber Abaci  A Translation into Modern English of Leonardo Pisano's Book of Calculation.  Springer, New York, 2002.  I have added page references to this, denoted  S,  after the Boncompagni pages, e.g.  pp. 397-398 (S: 543-544).  I have given Sigler's English wherever I previously had just quoted the Latin. 

Fibonacci.  Flos  and  Epistola.

                                    Leonardo Pisano, called Fibonacci.  MS of c1225 which begins "Incipit flos Leonardi bigolli pisani ...", Biblioteca Ambrosiana, Milan, E. 75.  In: Scritti di Leonardo Pisano, vol. II, ed. and pub. by B. Boncompagni, Rome, 1862, pp. 227-252.

                    Part of the MS has a separate heading: "Epistola suprascripsit Leonardi ad Magistrum Theodorum phylosophum domini Imperatoris" and is sometimes considered a separate work.  It occupies pp. 247-252 of the printed version.  For an English description, see:  A. F. Horadam; Fibonacci's mathematical letter to Master Theodorus; Fibonacci Quarterly 29 (1991) 103-107.

                    Italian translation (including the Epistola) and commentary: E. Picutti; Il 'Flos' di Leonardo Pisano; Physis 25 (1983) 293-387.

Fireside Amusements.  1850.

                                    Fireside Amusements.  Chambers's Library for Young People.  William and Robert Chambers, Edinburgh, 1850, 188pp.  The BMC has this under Fireside Amusements and refers to Chambers for the Library, which was 19 vols, 1848-1851.  Pp. 187+ are missing in the copy I have seen, but it seems that just one page of solutions is missing -- the NUC gives 188pp.  The NUC lists a 1870 reprint.

                    [The BMC lists an 1880 ed with 159pp, part of Chambers's Juvenile Library, NYS.]

                    Fireside Amusements  A Book of Indoor Games.  W. & R. Chambers, London and Edinburgh, nd, 128pp.  The BMC lists this as 1890[1889].  Though laid out entirely differently, almost all the material is taken from the 1850 ed.  I will cite both editions.

                    Much of the material of interest is taken from Child: Girl's Own Book.

Folkerts.  Aufgabensammlungen.  13-15C.

                                    Menso Folkerts.  Mathematische Aufgabensammlungen aus dem ausgehenden Mittelalter.  Sudhoffs Archiv 55 (1971) 58-75.  He examines 33 anonymous Latin manuscript problem collections from 13-15 C in Oxford, London, Berlin, Munich, Vienna and Erfurt and catalogues the problems therein.  Of these, only Munich 14684 is published (cf below).  He notes that many more such sources exist.  His catalogue covers 14 of my topics.  I will not try to cite the individual MSS, since many of the topics occur in over a dozen of them.  I will simply say he has  n  sources, though some of the sources have several examples.

Folkerts, Menso.  See:  Alcuin.

della Francesca.  Trattato.  c1480.

                                    Piero della Francesca (1412-1492).  Trattato d'Abaco.  Italian MS in Codex Ashburnhamiano 359* [291*] - 280 in the Biblioteca Mediceo-Laurenziana, Florence.  c1480 [according to Van Egmond's Catalog 84, based on watermarks in the paper which date from 1470 to 1500, but Davis, below, p. 16, says c1450].  Transcribed and annotated by Gino Arrighi, Testimonianze di Storia della Scienze 6, Domus Galilæana, Pisa, 1970.  Arrighi uses  c.  (for carta) instead of  f.  (for folio), but I will use  f.  for consistency with other usage, followed by the pages in Arrighi in ( ).  Arrighi reproduces many of the diagrams, but he doesn't say anything about whether he has included all of them.  This MS appears to be that which was in the possession of Piero's descendents until 1835 when it was reported as having disappeared.  Guglielmo Libri, the noted historian of mathematics, who was also a shady bookdealer, transcribed part of this MS in vol. 3 of his Histoire de la Mathématique en Italie in 1840 as an anonymous work, then sold it to Lord Ashburnham in 1847 (recorded in his collection in 1881) whose collection was bought for the Laurentian Library in 1884.  There are three different catalogue numbers - I use the format used in Van Egmond's Catalog.  The MS had passed out of common knowledge until it was rediscovered in the Laurentian Library in 1917 by Girolamo Mancini who recognised the handwriting as Piero's.

                    This work and Piero's Libellus de Quinque Corporibus Regularibus are the subject of a long standing plagiarism argument.  Giorgio Vasari [Le Vite de' più eccellenti pittori, scultori e architetti; 1550;  The Essential Vasari, ed. by Betty Burroughs from the 1850 translation of Mrs. Jonathan Foster, Unwin Books, London, 1962] states:  "... Piero della Francesca, who was a master of perspective and mathematics but who first went blind and then died before his books were known to the public.  Fra Luca di Borgo, who should have cherished the memory of his master and teacher, Piero, did his best, on the contrary, to obliterate his name, taking to himself all the honour by publishing as his own work that of that good old man.  ...  Maestro Luca di Borgo caused the works of his master, Piero della Francesca, to be printed as his own after Piero died."  The mathematical works of Piero were unknown until they were rediscovered in 1850/1880 and 1917.  Examination shows that Pacioli certainly used 105 problems, many unusual, from Piero in the Summa.  But he does praise Piero in the Epistola (f. 2r) of the Summa, as "the monarch of painting of our times".  It has been suggested that Pacioli had a large hand in the writing of Piero's works and hence was just reusing his own material and he frequently expands on it.  However, there is no evidence that Pacioli was ever a student of Piero.  Entire books have been written on the question, so I will not try to say any more.  See:  Margaret Daly Davis; Piero della Francesca's Mathematical Treatises  The "Trattato d'abaco" and "Libellus de quinque corporibus regularibus"; Longo Editore, Ravenna, 1977,  for detailed comparisons and the work of R. E. Taylor in Section 1: Pacioli.  Davis identifies  139  problems in the Libellus, of which  85  (= 61%)  are taken from the Trattato.  Davis notes that Pacioli's Summa, Part II, ff. 68v ‑ 73v, prob. 1-56, are essentially identical to della Francesca's Trattato, ff. 105r ‑ 120r.  See also section 6.AT.3 where the Libellus and the Pacioli & da Vinci: De Divina Proportione are discussed.

                    The work is discussed and 42 problems are given in English in:  S. A. Jayawardene;  The 'Trattato d'Abaco' of Piero della Francesca;  IN: Cecil H. Clough, ed.;  Cultural Aspects of the Italian Renaissance  Essays in Honour of Paul Oskar Kristeller;  Manchester Univ. Press, Manchester, nd [1976?];  pp. 229-243.  I will note 'English in Jayawardene.' when relevant.

Frikell, Wiljalba (1818 (or 1816) - 1903).  (The given name Gustave sometimes occurs -- I thought Gustave might be a son of Wiljalba, but the son was named Adalbert (  -1889) and his name was pirated by a clumsy imposter in England.)

                                    See the discussion at:  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Hanky Panky,  Magician's Own Book,  The Secret Out.  Frikell was a noted conjuror of the time and his name has been associated with the UK versions of these books, but there is no evidence he had anything to do with them.  The Art of Amusing, by Frank Bellew, Hotten, London, 1866?, op. cit. in 5.E, has a note on the back of the TP saying The Secret Out is a companion volume, just issued, by Hermann Frikell.  C&B, under  Williams, Henry Llewellyn ("W. Frikell")  lists:  Hanky Panky;  Magician's Own Book, London & New York;  (Magic No Mystery);  The Secret Out  and says to also see Cremer.

 

Gamow & Stern.  1958.

                                    George Gamow & Marvin Stern.  Puzzle‑Math.  Macmillan, London, 1958.

 

Gardner.                       Martin Gardner (1914-  ).  Many references are to both his SA column, cited by (month & year), e.g. SA (Mar 1982), and to the appearance of the column as a chapter in one of his books, abbreviated as shown below.  In general, I will only give the chapter reference as the various editions and translations are differently paginated.  Answers, comments and extensions appeared in succeeding issues of SA, usually in Gardner's column, but sometimes in the Letters.  All this material is collected in the book chapter, sometimes by rewriting of the article, sometimes as notes or an Addendum at the end of the chapter.  Since many years usually passed before the book version, the Addenda often contain material that never appeared in SA, as well as references to work done as a result of the SA article.  I have not tried to enter all of Gardner's references here, so anyone interested in a topic that Gardner has considered should consult the book version of Gardner's column.  Currently some of the earlier books are being reissued in new editions, with further extensions and updating.  See also the next entry.

                    For years from at least 1950, SA appeared in two volumes per year, each of six issues.  In year  1950 + n,  vol. 182 + 2n  covers  Jan-Jun  and  vol. 183 + 2n  covers  Jul-Dec.

     1st Book                  The Scientific American Book of Mathematical Puzzles and Diversions.  Simon & Schuster, 1959.

               UK version: Mathematical Puzzles and Diversions from Scientific American.  Bell, London, 1961; Penguin (without the words 'from Scientific American'), 1965.

     2nd Book The Second Scientific American Book of Mathematical Puzzles and Diversions.  Simon & Schuster, 1961.

               UK version: More Mathematical Puzzles and Diversions from Scientific American.  Bell, London, 1963.  Penguin (without the words 'from Scientific American'), 1966.  (The UK versions omit Chapter 20: "The Mysterious Dr. Matrix".  The dust wrapper of the HB has a sentence referring to this chapter which has been blacked out.  ??)

     New MD                 Martin Gardner's New Mathematical Diversions from Scientific American.  Simon & Schuster, 1966.

     Unexpected              The Unexpected Hanging and Other Mathematical Diversions.  Simon & Schuster, 1969.

               UK version: Further Mathematical Diversions.  Allen & Unwin, London, 1970; Penguin, 1977.

     6th Book                  Martin Gardner's Sixth Book of Mathematical Games from Scientific American.  Freeman, 1971.

     Carnival                   Mathematical Carnival.  Knopf, NY, 1975; Penguin, 1978.

     Magic Show             Mathematical Magic Show.  Random House, NY, 1978.

     Circus                      Mathematical Circus.  Knopf, NY, 1979.

     Wheels                    Wheels, Life and Other Mathematical Amusements.  Freeman, 1983.

     Knotted                   Knotted Doughnuts and Other Mathematical Entertainments. Freeman, 1986.

     Time Travel             Time Travel and Other Mathematical Bewilderments.  Freeman, 1988.

     Penrose Tiles           Penrose Tiles to Trapdoor Ciphers.  Freeman, 1989.

     Fractal                     Fractal Music, Hypercards and More ....  Freeman, 1992.

     Last                         The Last Recreations  Hydras, Eggs, and Other Mathematical Mystifications.  Copernicus (Springer), NY, 1997.  ??NYR.

     Magic Numbers        The Magic Numbers of Dr. Matrix.  Prometheus, Buffalo, 1985.  Chaps. 1-18 previously appeared as: The Incredible Dr. Matrix; Scribner's, NY, 1976.  Chaps. 1‑7 & 9 previously appeared as: The Numerology of Dr. Matrix; Simon & Schuster, NY, 1967.  In contrast to his other books above, the answers and comments occur at the end of this book instead of following the original articles.

     Workout                  A Gardner's Workout  Training the Mind and Entertaining the Spirit.  A. K. Peters, Natick, Massachusetts, 2001.  This comprises 41 chapters of articles written after his retirement from SA.

Gardner.  MM&M.  1956.

                                    Martin Gardner.  Mathematics, Magic and Mystery.  Dover, NY, 1956.

 

General Trattato.  1556.

                                    Nicolo Tartaglia (c1499-1557).  (La Prima Parte del) General Trattato di Numeri et Misure.  Curtio Troiano, Venice, 1556.  (Modern Italian spells his given name as Niccolò, but it appears as Nicolo on the title page.)  Six parts actually appeared in 1556-1560.  All references are to Part 1.  Unless otherwise specified, reference is to Book 16 (of Part 1), but I also have references to Books 12 and 17.  CAUTION -- the running head in Book 17 says Libro Decimosesto for several pages before changing to Libro Decimosettimo.  Since it is hard to find the beginnings of books, this can cause confusion.  See Rara 275-279; Van Egmond's Catalog 345-346.

                    In 1578, Guillaume Gosselin produced an annotated translation of parts 1 & 2 into French as: L'Arithmetique de Nicolas Tartaglia -- cf Van Egmond's Catalog 347.

Ghaligai.  Practica D'Arithmetica.  1521.

                                    Francesco Ghaligai.  Practica D'Arithmetica di Francesco Ghaligai Fiorentino.  Nuovamente Rivista, & con somma Diligenza Ristampata.  I Giunti, Florence, 1552.  Smith, Rara, says that this is identical to the first (Latin?) edition by Bernardo Zucchetta, Florence, 1521, except that edition was titled Summa De Arithmetica, so I will date the entries as 1521.  See Rara 132; Van Egmond's Catalog 316-317.

Gherardi.  Libro di ragioni  and  Liber habaci.  1328 & c1310.

                                    Paolo Gherardi.  Two Italian MSS in Codici Magliabechiani Classe XI, no. 87 & 88 in Bib. Naz. di Firenze.  Van Egmond's Catalog 115-116.  The first is dated 1327 (but see below).  The second is undated, but clearly of a similar date which I originally denoted 1327? - see below.  Transcribed by Gino Arrighi; Collana di Storia della Scienza e della Tecnica, No. 2; Maria Pacini Fazzi, Lucca, 1987.  See also:  Warren Van Egmond; The earliest vernacular treatment of algebra: the Libro di ragioni of Paolo Gerardi (1328); Physis 20 (1978) 155-189.  Van Egmond notes that the date of 30 Jan 1327 is in our year 1328 and uses this in his Catalog.  He doubts whether Liber habaci is actually by Gherardi and his Catalog assigns no author to it, so I will put Gherardi? as author.  He dates it to c1310.  His paper is concerned with the quadratic and cubic equations and hence of little interest to us.

Good, Arthur.  See:  Tom Tit.

Gori.  Libro di arimetricha.  1571.

                                    Dionigi Gori.  Libro di arimetricha.  1571.  Italian MS in Biblioteca Comunale di Siena, L. IV. 23.  ??NYS.  Extensively quoted and discussed in: R. Franci & L. Toti Rigatelli; Introduzione all'Aritmetica Mercantile del Medioevo e del Rinascimento; Istituto di Matematica dell'Università di Siena, nd [1980?].  (Later published by Quattroventi, Urbino, 1981.)  (I will quote Gori's folios and also give the pages of this Introduzione.)  Van Egmond's Catalog 191-192.

Graves.                         The Graves Collection of early mathematical books at University College London (UCL).

Guy, Richard Kenneth (1916-  ).  See:  Winning Ways.

G4Gn                           Gathering for Gardner n, held in Atlanta.  1: Jan 1993;  2: Jan 1996;  3: Jan 1998;  4: Feb 2000;  5: Apr 2002.

G&P.                           Games & Puzzles.  The first version ran from 1972 through 1981.  The second series started in Apr 1994 and finished with No. 16 in Jul 1995.

G&PJ.                          Games and Puzzles Journal.  Successor to Chessics.  Ran through 12 issues, Sep 1987 -- Dec 1989, then restarted intermittently in May 1996.

 

Haldeman-Julius.  1937.

                                    E. Haldeman-Julius.  Problems, Puzzles and Brain-teasers.  Haldeman-Julius Publications, Girard, Kansas, 1937.  Facsimile (I believe) presented by Bob Koeppel at IPP13, 1993.

Hall.  BCB.  1957.

                                    Trevor H. Hall.  A Bibliography of Books on Conjuring in English from 1580 to 1850.  (Carl Waring Jones, Minneapolis, 1957); Palmyra Press, Lepton, W. Yorks., 1957.  323 entries.  I will cite item numbers.  A Supplement is in Hall, OCB.  See Heyl for a list of items not in BCB.

Hall.  OCB.  1972.

                                    Trevor H. Hall.  Old Conjuring Books.  Duckworth, London, 1972.  This covers books in English up through 1850 and it includes a Supplement to his BCB and should be checked for further information on items in BCB.  This contains 39 new items and additional notes to 36 previous items.  New items are given interpolated item numbers, e.g. 24.5.  OCB also includes a slightly revised version of his booklet on van Etten, see Section 1 below.

Halwas.                        Robin Halwas, Ltd.  List XV.  American Mathematical Textbooks  1760-1850.  Catalogue of 511 items being sold as a collection.  London, 1997, 144pp.  Quite a number of English works and a few French works had US editions which are detailed in this.

Hanky Panky.  1872.

                                    Hanky Panky  A Book of Easy and Difficult Conjuring Tricks  Edited by W. H. Cremer, Jun.  (John Camden Hotten, London, 1872 [BMC & Toole Stott 193, listed under Cremer, while C&B, under Cremer, give London, 1872]; Hotten was succeeded by Chatto & Windus c1873 and they produced several editions [NUC has 1872, Toole Stott 1017 & Shortz have 1875].)  My copy says: A new edition with 250 practical illustrations.  John Grant, Edinburgh, nd [Toole Stott 1016 gives 1874; NUC gives 1875?  Christopher 235, under Cremer, is c1890.  NUC says it is also attributed to Henry Llewellyn Williams, but their entry under Williams says 'supposed author'.  C&B also list it under Williams.  (This has been attributed to Frikell, but Toole Stott doubts that Frikell had anything to do with this.  I may put this under Cremer.)

HB.XI.22.  1488.

                                    Stuttgart Landesbibliothek German MS HB.XI.22, 1488.  Brief description by  E. Rath; Über einen deutschen Algorismus aus dem Jahr 1488; Bibl. Math. (3) 14 (1913‑14) 244‑248.

Heath, Sir Thomas L.  See:  Diophantos;  HGM.

Heyl.  1963.                  Edgar Heyl (1911-1993).  A Contribution to Conjuring Bibliography.  English Language  1580 to 1850.  Edgar Heyl conjuring books, Baltimore, 1963.  Facsimile edition of 100 copies by Maurizio Martino Fine Books, PO Box 373, Mansfield Center, Connecticut, 06250, nd [1998?].  360 entries + appendix of 14 more, almost all not in Hall, BCB.

HGM.  1921. Sir Thomas L. Heath.  A History of Greek Mathematics, 2 vols.  (OUP, 1921); corrected reprint, Dover, 1981.

HM.                             Historia Mathematica.

Hoernle, A. F. Rudolf.  See:  Bakhshali MS.

Hoffmann.  1893.          Professor Louis Hoffmann [pseudonym of Angelo John Lewis (1839‑1919)].  Puzzles Old and New.  Warne, London, 1893.  Reprinted with Foreword by L. E. Hordern; Martin Breese, London, 1988. 

                    In 1984, Hordern published a limited edition (15 copies) of "The Hordern Collection of Hoffmann Puzzles  1850‑1920", which gives colour photos of examples from his collection and the appropriate text.  I often cite these pictures as they often differ from those in the following item, with the heading Hordern Collection.  Generally, the next item gives more specific dating and/or older examples.

                    In 1993, Hordern produced a corrected edition of all of Hoffmann as: Hoffmann's Puzzles Old & New; published by himself.  This has colour photos of all puzzles for which known examples exist.  I will cite this as Hoffmann-Hordern.  The 1893 edition gives solutions for each chapter in a following chapter, but both of Hordern's illustrated versions give each solution immediately after the problem, with colour picture nearby.  A small section on Elementary Properties of Numbers is omitted from the 1993 edition.

                    See also:  Tom Tit.

Honeyman Collection.

                                    The Honeyman Collection of Scientific Books and Manuscripts.  Sold by Sotheby's [Sotheby Parke Bernet], 1978-1981.  Seven volumes -- details given in Section 3.B.

Hordern, L. Edward (1941-2000).  See under Hoffmann and in 5.A.

HPL.                            The Harry Price Library, Senate House, University of London  OR  its catalogues. 

                                    Harry Price (1881-1948).  Short-Title Catalogue and Supplementary Catalogue of Works on Psychical Research, Spiritualism, Magic, Psychology, Legerdemain and Other Methods of Deception, Charlatanism, Witchcraft, and technical Works for the Scientific Investigation of Alleged Abnormal Phenomena from Circa 1450 A.D. to 1935 A.D.  Compiled by Harry Price.  (The first part was originally  ... to 1929 A. D.; Proc. National Lab. of Psychical Research 1:2 (1929); National Laboratory of Psychical Research, London, 1929.  The second part was originally: Short-Title Catalogue of the Research Library, for 1472 A.D. to the Present Day; Bull. Univ. of London Council for Psychical Investigation 1 (1935); Univ. of London Council for Psychical Investigation, 1935.)  New Introduction by R. W. Rieber and Andy Whitehead  With an appendix entitled "The St. Louis Magnet" (which originally appeared in 1845) by T. J. McNair and J. F. Slafter.  Da Capo Press (Plenum), NY, 1982.  NOTE: The works listed here are now in the Harry Price Library, though the editors have added 36 items in their Introduction.

Hummerston.  Fun, Mirth & Mystery.  1924.

                                    R. A. Hummerston.  The Book of Fun, Mirth & Mystery  A feast of delightful entertainment, including games, tricks, puzzles and solutions, "how to makes," and various other means of amusement.  Pearson, London, 1924.

Hunt.  1631  &  1651.

                                    Nich. Hunt.  Newe Recreations or The Mindes release and solacing.  Aug. Math. for Luke Fawne, 1631.

                                    Nich. Hunt.  New Recreations or A Rare and Exquisite Invention.  J. M. for Luke Fawn, London, 1651.  This edition contains a few more pages and several problems of interest and is differently paginated.  I will give both page numbers for problems in both editions.  Bill Kalush has sent both texts on a CD.

Hutton.  A Course of Mathematics.  1798?

                                    Charles Hutton (1737-1823).  A Course of Mathematics.  Composed for the Use of the Royal Military Academy.  (In 2 vols, plus a third, 1798-1811.)  A New Edition, entirely Remodelled.  By William Ramsay, B. A., Trinity College, Cambridge.  T. T. & J. Tegg, London, and Richard Griffin & Co., Glasgow, 1833 (in one volume).  8 + 822 pp.

Hutton-Rutherford.  A Course of Mathematics.  1841?

                                    Charles Hutton.  A Course of Mathematics,  Composed for the Use of The Royal Military Academy.  By Charles Hutton, LL.D., F.R.S., Late Professor of Mathematics in that Institution.  A new and carefully corrected Edition, Entirely Re-modelled, and Adapted to the Course of Instruction Now Pursued in the Royal Military Academy.  By William Rutherford, F.R.A.S.  Royal Military Academy.  William Tegg, London, 1857 [Preface dated Nov 1840, so probably identical or nearly identical to the 1841 ed].  8 + 895 pp.

                    [Two volume versions:  1798/1801;  3rd, 1800/1801;  4th, 1803/1804;  5th, [1810, NUC gives 1806- and 1807];  6th, [1810-1811 -- NUC].  Three volume versions -- apparently the early forms were just the earlier 2 volumes with an additional third vol:  6th, 1811;  1813;  7th, 1819-1820;  8th, 1824;  9th, 1827/1828.  10th ed by Olinthus Gregory, in 3 vols., 1827-1831.  New ed by William Ramsay in one vol., 1833; 1838.  11th ed by Gregory in 2 vols, 1836-1837.  12th ed, revised by Thomas Stephens Davies, 2 vols., 1841-1843.  Ed by William Rutherford in one vol, 1841; 1843; 1846; 1849; 1851; 1853; 1857; 1860. 

                    This is a pretty straightforward text, but it well illustrates the situation in early 19C England, outside Oxford and Cambridge.  Almost all the material of interest is in the first two sections: Arithmetic and Algebra and is identical in these two editions.  I imagine most of these problems appeared in the first edition, so I will date this as 1798?, citing pages as 1833 and 1857.  However the 1857 has an additional three pages on Practical Questions in Arithmetic which has 44 problems, some of which are recreational, and another new problem.  Assuming the 1857 is essentially the same as the 1841, I will cite this as 1841?.]

                    See also:  Ozanam‑Hutton.

H&S.  1927. Vera Sanford.  The History and Significance of Certain Standard Problems in Algebra.  (Teachers College, Columbia University, NY, Contributions to Education, No. 251, 1927)  = AMS Press, NY, 1972.

 

Illustrated Boy's Own Treasury.  c1847.

                                    The Illustrated Boy's Own Treasury of I. - Science, II. - Drawing, III. - Painting, IV. - Constructive Wonders, V. - Rural Affairs, VI. - Wild and Domesticated Animals, Outdoor Sports & Indoor Pastimes forming a Complete Repository of Home Amusements & Healthful Recreations embellished with five hundred descriptive engravings.  (John & Robert Maxwell, London, c1847 [Toole Stott 407]).  Ward and Lock, 1860 [Toole Stott 1091].  (3rd ed, for the Proprietors, 1865? [Toole Stott 408].)  [Toole Stott's descriptions make it seem that these editions are identical.]  See also:  Boy's Own Book,  Boy's Own Conjuring Book.  [Many of the problems are the same as in the other two books, but the illustrations here occasionally omit some labels, so these must be errors in copying from some earlier source.  If the c1847 date is correct, then this considerably changes the chronology of these problems, with this book being the major known intermediate between Boy's Own Book and Magician's Own Book.  I will hold off making these changes until I see the c1847 ed -- this may take some time as the BM copy was lost in the war and the other two copies cited are in the US.  Hall, BCB 187 is: The Illustrated Boys' Own Treasury of Indoor Pastimes; Robson, London, c1845.  This may be related to this book.]

Indoor & Outdoor.  c1859.

                                    Indoor and Outdoor Games for Boys and Girls: Comprising Parlour Pastimes, Charades, Riddles, Fireside Games, Chess, Draughts, &c, &c.  With a Great Variety of Athletic Sports, Parlour Magic, Exercises for Ingenuity, and Much That is Curious, Entertaining, and Instructive.  James Blackwood, London, nd [c1859].  This is a combination of two earlier books, comprising two separately paginated parts.  The earlier books are Parlour Pastime (1857 -- qv) and Games for All Seasons [Toole Stott 311 & BMC give 1858].  There is a later version of the first part -- Parlour Pastimes, qv, which the BMC dates as 1868.  Also there is another version of the combined ed "with additions by Oliver Optic", as Sports and Pastimes for Indoors and Out, G. W. Cottrell, Boston, 1863 [Toole Stott 1186, which identifies Optic as William Taylor Adams].  Both BMC and NUC say Indoor & Outdoor is by George Frederick Pardon.  Hence the problems in the first part will be cited as: Parlour Pastime, 1857  = Indoor & Outdoor, c1859, Part 1  = Parlour Pastimes, 1868.  Many of the problems are identical to Book of 500 Puzzles.

IPPn                             n-th International Puzzle Party.  10 = London, 1989;  13 = Amsterdam, 1993;  16 = Luxembourg, 1996;  19 = London, 1999;  20 = Los Angeles, 2000;  22 = Antwerp, 2002.  This are the ones I have attended, but some material has appeared at other IPPs.

 

Jackson.  Rational Amusement.  1821.

                                    John Jackson.  Rational Amusement for Winter Evenings; or, A Collection of above 200 Curious and Interesting Puzzles and Paradoxes relating to Arithmetic, Geometry, Geography, &c.  With Their Solutions, and Four Plates.  Designed Chiefly for Young Persons.  By John Jackson, Private Teacher of the Mathematics.  London: Sold by J. and A. Arch, Cornhill; and by Barry & Son, High‑Street; and P. Rose; Bristol.  1821.  [Other copies, apparently otherwise identical, say: London: Sold by Longman, Hurst, Rees, Orme, and Brown; G. and W. B. Whittaker; and Harvey and Darton.  And Barry and Son, High-Street, Bristol.  1821.  [Heyl 185.  Toole Stott 413.]  Will Shortz says this is the first English-language book devoted to non-word puzzles.]

JRM.                            Journal of Recreational Mathematics.

 

Kanchusen.  Wakoku Chiekurabe.  1727.

                                    Tagaya Kanchusen [pseud. of Fuwa Senkuro].  Wakoku Chiekurabe [Japanese Wisdom Competition -- in Japanese].  2 vols, 1727, 12 & 29 pp.  PHOTOCOPY from Shigeo Takagi's copy sent by Naoaki Takashima.  Edited into modern Japanese, with commentary on Kanchusen, by Shigeo Takagi, 1991, 42pp, present from Takagi.  Translated into English by Hiroko Dean, 1999, 15pp plus annotations on the 42pp.  Takagi and Takashima are working on a translation and annotation into modern Japanese.  We intend to produce an English version from Dean's translation with commentary on the puzzles.  I will cite pages from Takagi's edition.  (Partly reproduced in Akira Hirayama; Tôzai Sûgaku Monogatari [Mathematical Stories from East and West]; (1973), 3rd ed., 1981, p. 208, ??NYS, from which it has been reproduced in the exhibition Horizons Mathématiques at La Villette, Paris, and elsewhere.)

Kaye, George R.  See:  Bakhshali MS.

King.  Best 100.  1927

                                    Tom King.  The Best 100 Puzzles.  W. Foulsham, London, nd [1927, according to BMC -- my copy says 'Wartime reprint'.]  A selection of these are reproduced in a booklet: Foulsham's Games and Puzzles Book; W. Foulsham, London, nd [c1930].  I will indicate this by  = Foulsham's, no. & pp.

Knott, Cargill G.  See under:  Tom Tit.

 

Labosne.  See under:  Problemes.

Ladies' Diary.  See under: T. Leybourn.

Landells.  Boy's Own Toy-Maker.  1859.

                                    E[benezer] Landells.  The Boy's Own Toy-Maker: A Practical Illustrated Guide to the Useful Employment of Leisure Hours.  Griffith & Farran, London, 1859(1858); Shepard, Clark & Brown, Boston, 1859; Griffith & Farran, 3rd ed., 1860; D. Appleton, NY, 1860; Griffith & Farran, 6th ed., 1863 [Toole Stott 1286‑1290].  [BMC has 1859(1858) and a longer 10th ed, 1881.  NUC has the latter four of the versions given by Toole Stott.]  [The Preface to the Second Edition, reproduced in the 3rd ed., says it appeared just two months after the first edition.  Toole Stott indicates that all the versions he cites are identical.  I have 3rd ed., 1860.  Shortz has Appleton, 1860.]  The date 1859(1858) indicates that the book appeared in late 1858 to catch the Christmas trade, but was postdated 1859 to seem current for the whole of 1859, so I will date this as 1858.  The 2nd ed. must be 1859.  Comparison shows that the section on Practical Puzzles is essentially an exact subset of the material in Boy's Own Conjuring Book.

Leeming.  1946.            Joseph Leeming.  Fun with Puzzles.  (Lippincott, Philadelphia, 1946); Comet Books (Pocket Books), NY, 1949.

Lemon.  1890.               "Don Lemon" [= "The Sphinx"  = Eli Lemon Sheldon], selector.  Everybody's Illustrated Book of Puzzles.  Saxon & Co., London, 1890 (with 1891 on the back cover) and 1892.  The 1892 ed. omits the text on the back cover and adds some pages of publisher's advertisements, but is otherwise identical.  794 problems, about 100 being mathematical, on 125pp.  This looks like a UK reprint of a US book, but the NUC only lists London editions, so perhaps it is just selected from US publications.  Some of the problems are attributed to Golden Days, Good Housekeeping, St. Nicholas, etc.

                    I also have an undated edition which says 'Selected by the Sphinx'.  This has 744 problems on 122pp, about 45% of which come from the other edition.  The NUC dates this as 1895.  I will refer to this edition as:  Sphinx.  1895.

Leopold.  At Ease!  1943.

                                    Jules Leopold.  At Ease!  Ill. by Warren King.  Whittlesey House (McGraw‑Hill), 1943.  [This appears to be largely drawn from Yank, The Army Weekly, over the previous few years.] 

Leske.  Illustriertes Spielbuch für Mädchen.  1864?

                                    Marie Leske.  Illustriertes Spielbuch für Mädchen   Unterhaltende und anregende Belustigungen, Spiele and Beschäftigungen für Körper und Geist, im Zimmer sowie im Freien.  (1864; 19th ed., 1904);  20th ed., Otto Spamer, Leipzig, 1907.  There is no indication of any updating in the Foreword to the 19th ed. which is included here, and it was common to describe new printings as new editions, so I will date this as 1864?  This book is jammed with material of all sorts, including lots of rebuses, riddles and puzzles.  It is a bit like Boy's Own Book.  My copy is lacking pp. 159-160 and 211‑212.

Leurechon, Jean (c1591-1670).  See:  van Etten.

T. Leybourn.

                                    Thomas Leybourn, ed.  The Mathematical Questions, proposed in the Ladies' Diary, and Their Original Answers, Together with some New Solutions, from its commencement in the year 1704 to 1816.  4 vols., J. Mawman, London, and two co-publishers, 1817.  I have only examined vols. I & II so far.  The problems are proposed each year with solutions in the following year.  Leybourn puts the solutions just after the problem and numbers almost all the problems consecutively, though I don't know if these numbers are in the Ladies' Diary.  Problems do not have any names and sometimes have pseudonyms or vague names, e.g. Mr. Deare.  I will give the names of the proposer and solver(s), followed by  Ladies' Diary  and the two years involved, then  = T. Leybourn, his volume and pages and his question number.  E.g.  Chr. Mason, proposer; Rob. Fearnside, solver.  Ladies' Diary, 1732-33  =  T. Leybourn, I: 223, quest. 168.

W. Leybourn.  Pleasure with Profit.  1694.

                                    William Leybourn.  Pleasure with Profit: Consisting of Recreations of Divers Kinds, viz. Numerical, Geometrical, Mechanical, Statical, Astronomical, Horometrical, Cryptographical, Magnetical, Automatical, Chymical, and Historical.  Published to Recreate Ingenious Spirits; and to induce them to make farther scrutiny into these (and the like) Sublime Sciences.  And  To divert them from following such Vices, to which Youth (in this age) are so much Inclin'd.  To this work is also Annext, A Treatise of Algebra, ..., by R. Sault.  Richard Baldwin and John Dunton, London, 1694.  The text consists of several parts, labelled Tract. I, Tract. II, ..., which are separately paginated.  All material is from Tract. I unless otherwise specified.  Several sections are taken from the English editions of van Etten.  [Santi 371.]

Li & Du.  1987.             Li Yan & Du Shiran.  Chinese Mathematics: A Concise History.  (In Chinese: Commercial Press, Hong Kong, c1965.)  English translation by John Crossley & Anthony W.‑C. Lun.  OUP, 1987.

Libbrecht.  1973.           Ulrich Libbrecht.  Chinese Mathematics in the Thirteenth Century.  MIT Press, Cambridge, Mass., 1973.

Lilavati.  1150.              Lîlâvatî of Bhaskara II, 1150 (see Colebrooke).

Lloyd, E. Keith.  See:  BLW.

Loeb Classical Library.

                                    Published by Harvard Univ. Press, or Putnam's, NY, & Heinemann, London.

Loyd, Sam (1841-1911)  (&  Sam Loyd Jr. (1873-1934).  See:  Cyclopedia,  MPSL,  OPM,  SLAHP.

Lucas, Édouard (1842-1891).  See:  RM and the following.

Lucas.  L'Arithmétique Amusante.  1895.

                                    Édouard Lucas.  L'Arithmétique Amusante.  Ed. by H. Delannoy, C.‑A. Laisant & E. Lemoine.  (Gauthier-Villars, Paris, 1895.)  = Blanchard, Paris, 1974.

Lucca 1754.  c1330.

                                    Scuola Lucchese.  Libro d'abaco.  c1390.  Dal Codice 1754 (sec. XIV) della Biblioteca Statale di Lucca.  Edited by Gino Arrighi.  Cassa di Risparmio di Lucca, 1973.  Arrighi gives folio numbers and I will cite these and the pages of his edition.  Arrighi has c1390, but Van Egmond's Catalog 163-164 gives c1330.

 

MA.                             Mathematical Association (UK).

MAA.           Mathematical Association of America.

Magician's Own Book.  1857.

                                    The Magician's Own Book, or The Whole Art of Conjuring.  Being a Complete Hand-Book of Parlor Magic, and Containing over One Thousand Optical, Chemical, Mechanical, Magnetical, and Magical Experiments, Amusing Transmutations, Astonishing Sleights and Subtleties, Celebrated Card Deceptions, Ingenious Tricks with Numbers, Curious and Entertaining Puzzles, Together with All the Most Noted Tricks of Modern Performers.  The Whole Illustrated with over 500 Wood Cuts, and Intended as a Source of Amusement for One Thousand and One Evenings.  Dick & Fitzgerald, NY, ©1857.  12 + 362 pp. + 10 pp. publisher's ads.  My thanks to Jerry Slocum for providing a copy of this.  [Toole Stott 481 lists this as anonymous and entirely different from the UK ed.  He cites a 1910 letter from Harris B. Dick who says H. L. Williams may have edited it, but both Dick's father and John Wyman may also have had a hand in it.  Toole Stott 929, 930, 1378, 931 lists Dick & Fitzgerald, 1862, 1866, 1868, 1870, all apparently identical to the 1857.  929-930 are listed under Arnold and he there cites Cushing's Anonyms as saying the book is by Arnold and Cahill.  Christopher 622-625 are all Dick & Fitzgerald; 622-623 are 1st ed., 624-625 are reprints of about the same time and my copy seems most likely to be 625.  C&B, under Cremer, say "It is believed that they were all written by H. L. Williams, a prolific hack writer of the period."  Christopher 622 says Harold Adrian Smith [Dick and Fitzgerald Publishers; Books at Brown 34 (1987) 108-114] has studied this book and concludes that Williams was the author, assisted by Wyman.  Actually Smith simply asserts: "The book was undoubedly [sic] written by H. L. Williams, a "hack writer" of the period, assisted by John Wyman in the technical details."  He gives no explanation for his assertion, but it may be based on C&B.  NUC lists this as by George Arnold (1834-1865) and Frank Cahill, under both Arnold and Cahill.  C&B list it under Cremer, attributed to Arnold & Cahill, but they give a date of 1851, which must be a transcription error.  C&B also list it under Magician's, from New York, 1857, and under Williams, but as London, 1857.]  See the discussion under Status of the Project in the Introduction for the sources of the material.

                    Boy's Own Conjuring Book, qv, appears to be a UK pirate edition largely drawn from this.  Book of 500 Puzzles copies about 80 pages of this.  See the comments under Book of 500 Puzzles.  A fair number of the problems are identical to or similar to the Boy's Own Book and a woodcut, a poem and the introduction to a section are taken directly from Boy's Own Book.  Otherwise I had thought that this book was the source for the spate of puzzle books in the following 15 years, but I have found that some of the identical puzzles appeared in The Family Friend c1850. 

Magician's Own Book (UK version).  1871.

                                    This is quite different than the previous book.

                                    The Magician's Own Book.  By the Author of "The Secret Out," "The Modern Conjuror," &c.  Edited by W. H. Cremer, Jun.  Containing Ample Instructions for Recreations in Chemistry, Acoustics, Pneumatics, Legerdemain, Prestidigitation, Electricity (with and without apparatus).  (In the middle of the page is an illustration of a wizard in white on red.)  Performances with Cups and Balls, Eggs, Hats, Flowers, Coin, Books, Cards, Keys, Rings, Birds, Boxes, Bottles, Handkerchiefs, Glasses, Dice, Knives, &c., &c.  With 200 Practical Illustrations.  John Camden Hotten, nd [1871].  This has a two page list of Very Important New Books at the beginning on pp. i-ii.  This lists Magician's Own Book as by the Author of "The Secret Out" and The Secret Out as by the Author of the "Magician's Own Book".  But a further note says "Under the title of "Le Magicien des Salons" the first has long been a standard Magic Book with all French and German Professors of the Art." -- see the discussion under Status of the Project in the Introduction, above.  This list is followed by a half-title, p. iii, whose reverse (p. iv) has the printer's colophon, then a blank page v, backed by a Frontispiece, p. vi, comprising Figures 105 & 110 from the text.  The TP is p. vii,, backed by a blank p. viii.  The Preliminary on pp. ix-x has the address  Piccadilly  at the end and states this is "an Entirely New Edition" and is by the same author as The Secret Out.  It refers to Cremer's display of Toys of the World at the recent International Exhibition (possibly the 1862??) and to [Frank Bellew's] "The Art of Amusing" (of 1866 and published by Hotten in 1870) and Clara Bellew's "The Merry Circle".  Contents are given on pp. xi‑xii and then the text on pp. 13-326.  [Toole Stott 194 lists this under Cremer and says there are 30pp of publisher's catalogue at the end, dated 1872 -- I didn't record these details.  Christopher 239 lists this as "An entirely new edition" with 200 illustrations, 325 pp. + 30 pp. publisher's ads.]

                    I have also seen a John Grant, Edinburgh, ed. which omits pp. i-vi, and has a simplified title page, saying it is A New Edition, and drops the address  Piccadilly.  Otherwise the text appears to be identical.  It has no date but Toole Stott 1015 gives the date 1871.  C&B, under Cremer, have London, 1871 with no indication of the New York ed.  C&B also list it under Williams.

                    I have a Chatto & Windus ed., which omits pp. i-vi, and whose TP has only slight changes from the Hotten TP, saying it is A New Edition, but the text appears to be identical, except for dropping the address  Piccadilly  from the end of the Preliminary.  It is dated 1890, with separately paginated publisher's catalogue of 32pp. dated Apr 1893. 

Mahavira.  850.             Mahāvīrā(cārya).  Gaņita‑sāra‑sangraha [NOTE:  ņ  denotes an  n  with a dot under it and  ń  denotes an  n  with a dot over it.] (= Gaņita‑sāra‑samgraha [The  m  should have a dot over it.] = Ganitasar Samgrha).  850.  Translated by M. Rańgācārya.  Government Press, Madras, 1912.  The sections in this are verses.  I will refer to the integral part of the first verse of the problem.  E.g. where he uses  121½ ‑ 123,  I will use  v. 121.  [This work is described by David Eugene Smith; The Ganita-Sara-Sangraha of Mahāvīrācārya; Bibliotheca Mathematica (3) (1908/09) 106-110.  In: [G. R. Kaye;  A brief bibliography of Hindu mathematics;  J. Asiatic Society of Bengal (NS) 7:10 (Nov 1911) 679-686], this work is cited as 1908 with the note: "This is really an advance copy of a work not yet actually published, kindly supplied to me by the author."  See the entry under Pearson, 1907, in 7.E.]

Mair.  1765?.

                                    John Mair.  Arithmetic, Rational and Practical.  Wherein  The Properties of Numbers are clearly pointed out, the Theory of the science deduced from first principles, the methods of Operation demonstratively explained, and the whole reduced to Practice in a great variety of useful Rules.  Consisting of Three Parts, viz.  I. Vulgar Arithmetic.  II. Decimal Arithmetic.  III. Practical Arithmetic.  A. Kincaid & J. Bell, Edinburgh, in three vols, 1765-1766 (Turner G1.14/1-3);  2nd ed., A. Kincaid & W. Creech, and J. Bell, Edinburgh, 1772;  3rd ed, John Bell and William Creech, Edinburgh, 1777.  I have the 3rd ed and have seen the 2nd ed.  All the material of interest is in part 3, which first appeared as vol. 3 in 1765 and books of this era often had little or no change between editions, so I will date entries as 1765?

Manson.  Indoor Amusements.  1911.

                                    J. A. Manson, compiler.  Indoor Amusements.  Cassell & Co., London, 1911.  FP + 8pp + 340pp + 8pp Index (341-348).  This is an extension of Cassell's Book of In-Door Amusements ..., expanding the earlier 209pp of main text to 340pp.  This is partly due to using larger type, getting 47 lines per page instead of 54.  The material which was in Cassell's is generally unchanged.

Manuel des Sorciers.  1825.

                                    Manuel des Sorciers  ou  Cours de Récréations Physiques, Mathématiques, Tours de Cartes et de Gibecière, suive  Des Petits Jeux de Société, et de Leurs Pénitences.  (Conort, Paris, 178?;  2nd ed, Metier & Levacher, Paris, 1802 [Christopher 642, C&B];  4th ed, Ferra Jeune, Paris, 1815 [C&B];  5th ed, Ferra Jeune, 1820 [C&B]);  6th ed, Augmentée d'une Notice sur la Magie noire, Ferra Jeune, Paris, 1825 [Christopher 643, C&B, HPL].  There are a great many French books with similar titles from this era.  They seem to be the predecessors of Magician's Own Book, etc. -- cf the discussion under Status of the Project in the Introduction and under Book of 500 Puzzles and Magician's Own Book.  This is the first that I have found and examined carefully, at HPL.  It is likely that most of the material dates back to the first ed of 178?, but until I see some earlier editions, I'll date it as 1825,  Gaidoz, in Section 7.B, cites the 2nd ed, but the material is on a different page than in the 1825.

Marinoni, Augusto.  See:  Pacioli.  De Viribus.  c1500.

McKay.  At Home Tonight.  1940.

                                    Herbert McKay.  At Home Tonight.  OUP, 1940.  Section V: Puzzles and problems, pp. 63-88.

McKay.  Party Night.  1940.

                                    Herbert McKay.  Party Night.  OUP, 1940.  Sections on Dinner-Table Tricks, pp. 134-171;  Some Tricks in English, pp. 174-175;  Arithmetical Catches and Puzzles, pp. 176-184.

Metrodorus.                  c510.  In: The Greek Anthology, W. R. Paton, trans.  Loeb Classical Library, 1916‑1918.  Vol. 5, Book 14.  This contains 44 mathematical problems, most of which are attributed to Metrodorus, though he is clearly simply a compiler and some may be much older.  I have cited the pages of the English translation -- the Greek is on the previous page.  Paton gives English answers, but they are not in the Greek. 

                    The Greek Anthology is the modern name for a combination of two anthologies which were based on earlier compilations dating back to the time of Alexander (-4C), e.g. the Garlands of Meleager (c-95) and Philippus of Thessalonika (c40) and the Cycle of Agathias (c570).  In the late 9C or early 10C, Konstantinos Kephalas assembled these into one collection, but distributed them according to type and then added several other collections -- this was somewhat revised c980.  In 1301, Maximus Planudes re-edited Kephalas' anthology, omitting much and adding some (Paton believes that Planudes' source was missing a book).  The Planudean version replaced Kephalas's version and was printed in 1484.  However a copy of Kephalas's version of c980 was discovered at Heidelberg in the Palatine Library (hence the anthology is sometimes described as Palatine) in 1606 and modern versions now use this version as the first 15 books and put all of Planudes' additions as Book 16, the Planudean Appendix.  The Anthology comprises some 4000 poems.  Modern scholars view Paton as obsolete, but I don't know of any later versions of the Anthology which include the mathematical problems -- e.g. they are not even mentioned in Peter Jay; The Greek Anthology; Allen Lane, 1973; Penguin, 1981.

                    See also:  David Singmaster; Puzzles from the Greek Anthology;  Math. Spectrum 17:1 (1984/85) 11-15  for a survey of these problems.

Meyer.  Big Fun Book.  1940.

                                    Jerome S. Meyer.  The Big Fun Book.  Greenberg : Publisher, NY, 1940.

MG.                             Mathematical Gazette.

Mikami.  1913.              Yoshio Mikami.  The Development of Mathematics in China and Japan.  Teubner, Leipzig, 1913; reprinted, Chelsea, 1961?  See also:  Smith & Mikami.

Minguet.  1733.

                                    Pablo Minguet (or Minguét) è (or é or e or y) Yról (or Irol) (  -1801?).  Engaños à Ojos Vistas, y Diversion de Trabajos Mundanos, Fundada en Licitos Juegos de Manos, que contiene todas las diferencias de los Cubiletes, y otras habilidades muy curiosas, demostradas con diferentes Láminas, para que los pueda hacer facilmente qualquier entretenido.  Pedro Joseph Alonso y Padilla, Madrid, nd [1733].  Frontispiece + 12 + 218 pp.  Imprimaturs or licenses dated 3 Nov 1733, 10 Nov 1733 & 12 Dec 1733.  [NUC.  Christopher 672 for a 18 + 110 pp version.  C&B gives versions with 18 + 110 pp and with 12 + 218 pp.  HPL has 4 versions -- the one I examined was 12 + 218 pp.  The BM has a version, but it is not clear which.]

                    The early history of this book is confused.  The first edition may have only had 18 + 110 pp (or 105 pp ??), then was apparently frequently reprinted, without changing the dates, sometimes with additions.  Or it may be that both versions appeared in 1733.  However, my copy is identical to one in HPL which is catalogued as 1733 and Palau (a list of Spanish book sales, c1955) lists six sales of the 12 + 218 pp version, dated 1733, and only one sale of the 18 + 110 pp version, dated 1733.  Christopher dates the 12 + 218 pp versions to c1760. 

                    I now have discovered 26 editions of this work, and 2 and 9 editions of two derivative works, but I have only seen a few versions.  If anyone has access to a copy, I would like a photocopy of the TP and other publishing details and of the Indice.

                    3rd ed, Domingo Fernandez de Arrojo, Madrid, 1755, 18 + 157 + 3 pp.  [This includes the same material as 1733, but it is reset with smaller type and much rearranged and includes some new material.  It seems to be the same as the 1766.  Palau gives this as 18 + 150 + 10 pp and says there was a cheap edition of 14 + 171 + 10 pp.  NUC gives 14 + 171 +11 pp, with publisher Domingo Fernandez.  HPL has two 'entirely different' editions of 1755 -- the one I examined was 18 + 157 + 3 pp.  BM has a 1755, but it is not clear which.]

                    3rd ed, Dionisio Hernández, Madrid, 1755?  [Palau, with nd.]

                    Antonio del Valle, Madrid, 1756.  171 + 20 (or 40??) pp.  [Palau.]

                    Revised.  Pedro J. Alfonso y Padilla, Madrid, c1760, 14 + 218 pp.  [Christopher 673 & 674.]

                    Idem.  Añadido en esta quarta impresion, 18 enigmas, 6 quisicosas muy curiosas.  4th ptg, D. Gabriel Ramirez, Madrid, 1766, 16 + 150 + 10 pp.  [BM; Palau; Christopher 675.]

                    Palau says he has seen a catalogue with a 1793 ed, but thinks this is a printing error for either 1733 or 1893.

                    Palau lists an 1804 reprint of the 1733, 14 + 218 pp.

                    Sierra y Martí, Barcelona, 1820, 1 + 224 pp.  [Identical to 1733, 12 + 218 pp, except the text has been reset, spelling modernized, the front matter updated and the text starts on p. 10, with everything the same as the 1733, except page numbers are increased by 9, and the index is reduced in size.  HPL; Palau; Christopher 676.]

                    Neuva Edicion corregida y aumentada por D. J. M. de L.  Juan Francisco Piferrer, Barcelona, 1822, 1 + 224 pp.  [This is quite differently arranged than the 1733, 12 + 218 pp, and 1820 eds, and contains some extra material.  Palau, with  1 + 240 pp; NUC; Shortz; Christopher 677.]. 

                    Retitled: Juegos de Manos ó sea Arte de Hacer Diabluras, y Juegos de Prendas.  Que contiene varias demonstraciones de mágia, fantasmagoria, sombras y otros entretenimientos de diversion para tertulias y sociedades caseras.  ilustrado con láminas  Por D. Pablo Minguet, y aumentado considerablemente en esta nueva edicion con infinidad de juegos nuevos, y con laminas intercaladas en el texto.  D. Manuel Saurí, Barcelona, 1847, 189 + 10 pp.  [Palau; Christopher 678.]

                    Juegos de manos; ó sea, Arte de hacer diabluras ...ilustrado con 60 grabados....  New ed.  Simon Blanquet, Mexico, nd [1856].  7 + 498 pp.  [NUC.]

                    Palau cites an 1857 reprint of the 1847, presumably the 2nd Saurí ed..

                    Title varied: Juegos de Manos ó sea Arte de Hacer Diabluras.  Contiene: juegos de prendas, de naipes, varias demonstraciones de majia, fantasmagoria, sombras y otros entretenimientos de diversion, para tertulias y sociedades caseras.  Por D. Pablo Minguet.  Tercera Edicion  Aumentada con gran número de Juegos nuevos, y grabados intercalados en el texto.  Manuel Saurí, Barcelona, 1864.  1 + 213 pp.  [From TP of 1993 facsimile.  This omits one problem and some discussion that was in 1733 and adds 22 new problems, but I see some of these already appeared in 1755 and 1822.  Palau, noting that this is the 3rd ed from Saurí and mentioning a recent facsimile; HPL.]

                    Palau cites an 1875 Barcelona reprint as 185pp, but has no publisher's name, probably the 4th Saurí ed.

                    Title varied: Juegos de manos; ó sea, El arte de hacer diabluras, contiene 150 clases de juegos, de prendas, de naipes, varias demonstraciones de mágia, fantasmagoria, sombras y otros entretenimientos de diversion, para tertulias y sociedades caseras.  5. ed., aumentada con gran numero de juegos nuevos y 70 grabados intercalados en el texto.  Manuel Saurí, Barcelona, 1876.  192pp.  [NUC.]

                    Palau confusingly cites Sauri reprints of 192pp:  8th ed, 1885;  8th ed, 1888;  9th ed, 1888.  [Christopher 679 & 680 are the latter two.  Perhaps there is an error in the dates, e.g. the 1885 might really be the 7th ed.]

                    Title varied: Juegos de Manos o sea Arte de hacer Diabluras, contiene 150 clases de juegos, de prendas de naipes, varias demonstraciones de magia, fantasmagorias, sombras y otros entretenimientos de mucha diversion, para tertulias y sociedades caseras.  Décima edición aumentada con juegos nuevos, y 70 grabados intercalados en el texto.  Juan Tarroll y Cla., Barcelona, 1893.  192pp.  [Palau.]

                    Palau lists 11th ed [from Sauri].  Sauri y Sabater, Barcelona, 1896.  192pp.  [NUC lists 2 copies of the 11th ed as Sauri y Sabater, 1897, 192pp & 189pp.]

                    12th ed [from Sauri].  Sauri, Barcelona, 1906.  190pp.  [Christopher II 2102.  HPL with nd.]

                    Facsimile of the 1864 ed, with Presentación por Joan Brossa: De la brujería blanca.  Editorial Alta Fulla, Barcelona, 1981.  11pp new material + II-VIII + 9-213 pp of facsimile (II is FP; III is TP; V-VIII is Al Lector; 9-12 is a poetic introductory Relacion; 203-213 is Indice).

                    2nd ed (i.e. printing) of the 1981 facsimile of the 1864 ed, with a new cover, 1993.  [DBS.]

                    Pp. 1-25 is a fairly direct translation of the 1723/1725 Ozanam, vol. IV, pp. 393‑406.  A number of other pictures and texts also are taken from Ozanam.  I will give the date as 1733, though the expansion may not have occurred until c1755.  I will cite the 1755, 1822 and 1864 pages in parentheses, e.g.  Pp. 158-159 (1755: 114-115; 1822: 175-176; 1864: 151).

MiS.                             Mathematics in School.

Mittenzwey.  1880.

                                    Louis Mittenzwey.  Mathematische Kurzweil.  Julius Klinkhardt, Leipzig.  1880;  2nd ed., 1883;  3rd ed., 1895;  4th ed., 1904;  5th ed., 1907;  6th ed., 1912;  7th ed., 1918.  I will give the date as 1880 or 1895?  I now have copies of the 1st, 4th, 5th and 7th eds.  In working through the other editions, I have seen many more items than I had previously recorded and I now think this is one of the most important 19C puzzle books.

                    [Ahrens, MUS #363 lists the first ed. being 1879, apparently taking the date of the Foreword.  He also has 3rd ed., 1903, but V&T (and another reference) cite 3rd ed., 1895.  I think this is a misinterpretation of the last Vorrede in the 4th ed., which is for the 3rd and 4th ed. and dated 1903.  Trey Kazee has obtained a copy of the 2nd ed.  The first two editions have the author's given name and have 300 problems; the 4th, 5th and 7th have 333.  The Vorrede to the 2nd ed in the 2nd ed. says it has no major changes.  The Vorrede to the 3rd & 4th eds in the 4th ed says the 3rd ed was extended and the 4th ed. has replaced some problems, so it seems that the 3rd probably had 333 problems.  Compared to the 1st ed, the 4th drops 9 problems and adds 42.  The Vorrede to the 3rd, 4th and 5th eds in the 5th ed also says some problems have been replaced, but I have not discovered any differences between the 4th and 5th eds except one minor amendment, some resetting which changes the page breaks on pp. 1-10 and 36-37 and a few line breaks.  I suspect that the 4th and 5th eds are very similar to the 3rd.  The 5th and 7th eds. are very similar, but the 7th is reset with smaller type and occupies five fewer pages.  There are a few amendments and reorderings and four problems (36, 92, 122, 137) have been replaced, but a number of simple misprints persist through all editions.  I will give pages in 1st, 3rd?, and 7th eds, e.g.  Prob. 193 & 194, pp. 36 & 89;  1895?: 218 & 219, pp. 41 & 91;  1917: 218 & 219, pp. 37 & 87.]

MM.                             Mathematics Magazine.

Montucla, Jean Étienne (1725-1799).  See:  Ozanam;  Ozanam‑Montucla.

MP.  1926.                   H. E. Dudeney.  Modern Puzzles.  C. Arthur Pearson, London, 1926; new ed., nd [1936?].  (Almost all of this is in 536.)

MPSL1.  1959  &  MPSL2.  1960.

                                    Mathematical Puzzles of Sam Loyd, vol. 1 & 2, ed. by Martin Gardner, Dover, 1959, 1960.  (These contain about 2/3 of the mathematical problems in the Cyclopedia, often with additional material by Gardner.)

MRE.                           W. W. Rouse Ball (1850-1925).  Mathematical Recreations and Essays.  (First three editions titled Mathematical Recreations and Problems  of Past and Present Times.)  I have compiled an 8 page detailed comparison of the contents of all editions as part of my  The Bibliography of Some Recreational Mathematics Books.

                    1st ed., Feb 1892, 240pp.

                    2nd ed., May 1892 ('No material changes').

                    3rd ed., 1896, 276pp.

                    4th ed., 1905, 388pp.  [In the 4th ed., it says  'First Edition, Feb. 1892.  Reprinted, May 1892.  Second Edition, 1896.  Reprinted, 1905.'  However, it calls itself the 4th ed. and the 3rd ed. calls itself 3rd and there are substantial changes in the 4th ed.]

                    5th ed., 1911, 492pp.

                    6th ed., 1914, 492pp.

                    7th ed., 1917, 492pp.

                    8th ed., 1919, 492pp.

                    9th ed., 1920, 492pp.

                    10th ed., 1922, 366pp.

                    11th ed., 1939, revised by H. S. M. Coxeter, 418pp. 

          Macmillan (for 1st to 11th ed.).

                    12th ed., 1974, revised by H. S. M. Coxeter, 428pp, U. of Toronto Press.

                    13th ed., 1987, revised by H. S. M. Coxeter, 428pp, Dover.

          A few of the editions are actually reprintings of the previous edition:  2 is essentially the

                    same as 1,  6 as 5,  9 as 8,  13 as 12.  Consequently I will rarely, if ever, cite

                    editions 2, 6, 9, 13.  For each topic occurring in Ball, I have examined all the

                    editions, so the absence of a reference indicates the topic does not occur in that

                    edition -- unless it is buried in some highly unlikely place.

          See also:  Ball‑FitzPatrick.

MS.                              Mathematical Spectrum (Sheffield, UK).

MTg.                            Mathematics Teaching (UK).

MTr.                            Mathematics Teacher (US).

Munich 14684.  14C.

                                    Munich Codex Lat. 14684.  14C.  Ff. 30‑33.  Published by M. Curtze; Mathematisch‑historische Miscellan: 6 -- Arithmetische Scherzaufgaben aus dem 14 Jahrhundert; Bibliotheca Math. (2) 9 (1895) 77‑88.  34 problems.  Curtze gives brief notes in German.  Curtze says these problems also appear in Codex Amplonianus Qu. 345, ff. 16‑16', c1325, ??NYS.  Munich 14684 comes from the same monastery (St. Emmeran) as AR and AR incorporates much of it.  In a later paper, Curtze says this is 13C -- ??NYR.  Cf Folkerts, Aufgabensammlungen.

Murray.  1913.              Harold James Ruthven Murray.  A History of Chess.  (OUP, 1913); reprinted by Benjamin Press, Northampton, Massachusetts, nd [c1986].

MUS.  1910 & 1918.

                                    Wilhelm Ernst Martin Georg Ahrens (1872-1927).  Mathematische Unterhaltungen und Spiele.  2nd ed., 2 vols., 1910, 1918, Teubner, Leipzig.  [The first ed. was 1901 in one volume.  There is a 3rd ed. of Vol 1, 1921, but it is a reprint of the 2nd ed. with just 2 pages of extra notes and the typographical corrections made.]  I will tend to cite this, as e.g. MUS 1 153-155.  MUS #n  denotes item  n  in the substantial Literarischer Index, MUS 2 375-431. 

                    [MUS II vii lists 22 items which he had been unable to see and which he suspected might not exist.  I have seen the following items from the list:  86 (= Les Amusemens, above),  102 (Hooper, listed in Some Other Recurring References, below; cf Section 4.A.1),  145 (Jackson, see above),  212 (author's name is Horatio Nelson Robinson);  item 32 exists as an English edition of van Etten, but the citation gives the publisher as though he was the author;  I have seen a copy of item 82 advertised for sale.  A number of items (135, 152, 164, 221, 223, 277, 297, 317) are cited by Wölffing - op. cit. in Section 3.B.]

Muscarello.  1478.

                                    Pietro Paolo Muscarello.  Algorismus.  MS of 1478.  Published in 2 vols.: I -- facsimile; II -- transcription with notes and commentaries; Banca Commerciale Italiana, Milan, 1972.  I will cite the folios (given in vol. I) and the pages of the transcribed version in vol. II.  Van Egmond's Catalog 275-276.

M500.          This is the actual name of the journal of the M500 Society, the Open University student mathematics society.

 

NCTM.                        National Council of Teachers of Mathematics.

nd.                               no date.  Estimated dates may follow in  [   ].

Needham.  1958.           Joseph Needham (1900-1995).  Science and Civilization in China, Vol. 3.  CUP, 1958.  (Occasional references may be made to other volumes: Vol. 2, 1956;  Vol. 5, Part IV, 1980.)

The New Sphinx.  c1840.

                                    Anonymous.  The New Sphinx  An elegant Collection of upwards of 500  Enigmas  Charades  Rebusses  Logogriphes  Anagrams  Conundrums  &c. &c.  To which are added, a Number of Ingenious Problems.  T. Tegg & Son, London, nd, HB with folding frontispiece.  [Vendor suggests it is 1840s.  Shortz has 4th ed, by Gye & Baine and says it is c1840.  He also has 'a new(?) edition', by T. Tegg, and says it is c1843.  He says the chapter of geometrical problems and brainteasers was new in the 4th ed.  Heyl 238 is a 7th ed., London, 18??, referring to HPL, where I find it in the Supplement.]  The chapter of problems has 27 problems, of which 21 are copied from Endless Amusement II, 1837 ed., 20 of which come from the 1826? ed.

E. P. Northrop.  Riddles in Mathematics.  1944.

                                    Eugene P. Northrop.  Riddles in Mathematics.  Van Nostrand, 1944;  English Universities Press, 1945;  revised ed., Penguin, 1961.  The Van Nostrand ed has the main text on pp. 1-262.  The EUP ed. has it on pp. 1-242.  The Penguin ed. has it on pp. 11-240.  The revision seems to consist of only a few additional notes.  I will cite the dates and pages, e.g.  1944: 209-211 & 239;  1945: 195‑197 & 222;  1961: 197‑198 & 222.

H. D. Northrop.  Popular Pastimes.  1901.

                                    Henry Davenport Northrop.  Popular Pastimes  for  Amusement and Instruction  being a  Standard Work on Games, Plays, Magic and  Natural Phenomena  Suitable for All Occasions  containing  Parlor Games; Charming Tableaux; Tricks of Magic; Charades  and Conundrums; Curious Puzzles; Phrenology and  Mind Reading; Palmistry, or How to Read the  Hand; Humorous and Pathetic Recitations,  Dialogues, Etc., Etc.  including  The Delightful Art of Entertaining  The Whole Forming a  Charming Treasury of Pastimes for the Home, Public Schools  and Academies, Lodges, Social Gatherings,  Sunday Schools, Etc., Etc.  Frank S. Brant, Philadelphia, 1901.  [Not in any of my bibliographies.  Vendor says only two copies in NUC, none in BL.]

NUC.                           National Union Catalogue  Pre‑1956 Imprints.  Library of Congress, USA.  c1960.  ??check details

Nuts to Crack.               Nuts to Crack, Part nn.  Or, Enigmatical Repository; containing near mmm Hieroglyphics, Enigmas, Conundrums, Curious Puzzles, and Other Ingenious Devices.  R. Macdonald, 30 Great Sutton Street, Clerkenwell, London.  These are single broadsheets.  The publisher's details are often trimmed from the bottom of the sheet.  At least 25 annual parts appeared, from Part I of 1832, but the year is not always given.  Answer books -- The Nutcrackers -- also appeared and the publishers kept the old sheets available for some years.  This series is very rare -- Will Shortz and James Dalgety have the only examples known to me.  I have photocopy of almost all from Dalgety and Shortz, ??NYR.  mmm is either 250 or 200 and the problems are individually numbered in each part.  I will cite problems as, e.g. Nuts to Crack I (1832), no. 200.

NX.                              No copy.  Usually prefixed by  ??  as a flag for further action.

NYR.                           Not yet read -- i.e. I have a copy which I have not yet studied.  Usually prefixed by  ??  as a flag for further action.

NYS.                            Not yet seen.  Usually prefixed by  ??  as a flag for further action.

 

OCB.  See:  Hall, OCB.

OED.                            Oxford English Dictionary.  (As: New English Dictionary, OUP, 1884‑1928), reprinted with supplements, OUP, 1933 and in various formats since.

o/o.                              On order.

OPM.  1907-1908.        Our Puzzle Magazine.  Produced by Sam Loyd.  The pages were unnumbered.  The Magazine was reprinted as the Cyclopedia as follows, but some pages of the Magazine were omitted and the answers to each magazine were normally in the next one.

           Vol. 1, No. 1 (Jun 1907)  =  Cyclopedia pp. 7‑70.

           Vol. 1, No. 2 (Oct 1907)  =  Cyclopedia pp. 71‑121.

           Vol. 1, No. 3 (Jan 1908)  =  Cyclopedia pp. 122‑178.

           Vol. 1, No. 4 (Apr 1908)  =  Cyclopedia pp. 179‑234.

                     Since the Cyclopedia goes to p. 339, there appear to have been two further issues which have not been seen by anyone??

           (The above data were provided by Jerry Slocum.)

OUP.                            Oxford University Press.

 

Ozanam.  1694.

                                    The bibliography of this book is a little complicated.  I have prepared a more detailed 7 pp. version covering the 19 (or 20) French and 10 English editions, from 1694 to 1854, as well as 15 related versions -- as part of my  The Bibliography of Some Recreational Mathematics Books.

                                    Jacques Ozanam (1640-1717).  Recreations Mathematiques et Physiques, qui contiennent Plusieurs Problémes [sic] utiles & agreables, d'Arithmetique, de Geometrie, d'Optique, de Gnomonique, de Cosmographie, de Mecanique, de Pyrotechnie, & de Physique.  Avec un Traité nouveau des Horloges Elementaires.  2 vols., Jombert, Paris, 1694, ??NYS.  [Title taken from my 1696 ed.]

                    [BNC.  NUC -- but NUC lists a 1693 3rd ed. from Amsterdam, which appears to be a misreading for 1698.  MUS II 380 says 1st ed. was Récréations Mathématiques, Paris, 1694, 2 vols.  Serge Plantureux's 1993 catalogue describes a 1694 edition in 2 vols., by Jombert and notes that it is the original edition, that the privilege is dated 11 Jan 1692, but that it was not printed until 30 Apr 1694.  The dating of the privilege may account for some references to the first edition being 1692 -- e.g. the Preface to the 1778 Ozanam-Montucla ed. says the first ed. was 1692, but Hutton changes this to 1694.  MRE, 1st ed, 1892, pp. 3‑4, says the the 1st ed. was 2 volumes, Amsterdam, 1696, but this was amended in his 4th ed., 1905.

                    This first appeared in two volumes, but later versions were sometimes in one volume.  I have references to versions in Paris: 1694, 1696, 1697, 1698, possibly 1700? and apparently 1720; and in Amsterdam: 1696, 1697, 1698, 1700.  There is no indication of any textual changes in these, except that the pages are numbered consecutively in later versions.  Plantureaux describes a 1696 edition as: Paris, Jombert (mais Hollande), so there seems to have been some piracy going on.  I have the 1696 ed.  I will assume that the 1696 is essentially identical in content to the 1694, though in the second volume, the page numbers may be different and there is some confusion of plate numbering.

                    The Traité des Horologes élémentaires, which appears in the 1694 ed., is a translation of Domenico Martinelli's Horologi elementari.  NUC says this was separately paginated in 1694, but it occupies pp. 473-583 of the 1696 ed and pp. 301-482 of vol. 3 of the 1725 ed.]

 

                    About 1723, the work was revised into 4 vols., sometimes described as 3 vols. and a supplement.  MUS #52 gives 1720, 1723, 1724, 1725 and says the dates vary in the literature.  The 1725 has privilege dated 1720, but I haven't found any catalogue entry for a 1720 ed. of this revision, so it may be a spurious date based on the privilege. 

     Nouv. ed.

                    Recreations Mathematiques et Physiques, qui contiennent Plusieurs Problêmes [sic] d'Arithmétique, de Géométrie, de Musique, d'Optique, de Gnomonique, de Cosmographie, de Mécanique, de Pyrotechnie, & de Physique.  Avec un Traité des Horloges Elementaires.  Par feu [misprinted Parfeu in vol. 1] M. Ozanam, de l'Académie Royale des Sciences, & Professeur en Mathematique.  Nouvelle edition, Revûë, corrigée & augmentée.

               Vol. 4 has different title page.

                    Recreations Mathematiques et Physiques, ou l'on traite Des Phosphores Naturels & Artificiels, & des Lampes Perpetuelles.  Dissertation Physique & Chimique.  Avec l'Explication des Tours de Gibeciere, de Gobelets, & autres récréatifs & divertissans.  Nouvelle edition, Revûë, corrigée & augmentée.

          Claude Jombert, Paris, 1723.  [Taken from my 1725 ed.]

                    [Ball and Glaisher [op. cit. in 7.P.5, p. 119] both cite a 1723 ed. as though they had seen it, but there is no BMC entry for this date -- perhaps there is a copy at Cambridge??.  I have seen one volume in an exhibition which was 1723.  MUS #52 says it was edited by Grandin.  NUC -- "The editor is said to be one Grandin."  I have a brief 1899 reference to this ed.

                    I have 1725, which is apparently a reprint of the 1723.  The privilege/approbation is dated 16 May 1722 in Vol. 3 and Vol. 4 and also 28 Apr and 15 May 1720 in Vol. 3.  NUC says this is 1723 with new title pages.  BNC has: Nouvelle édition ... augmentée [par Grandin], 4 vols, Paris, 1725.  It was reprinted in 1735, 1737?, 1741, 1750/1749, 1770.

                    The text and plates of the 1725 and 1735 eds. seem identical, though some of the accessory material -- lists of corrections and of plates -- has been omitted and other has been rearranged.  I have seen two versions of the 1735 -- one has the plates inserted in the text, the other has them at the end as ordinary pages, while my 1725 has them at the end on folding pages.  Most of the 1725 plates are identical to the 1696 plates, but there were a number of additions and reorderings.  The 1725 plates have their 1696 plate numbers and 1725 page references at the top with new, more sequential, plate numbers at the bottom.  The 1725 text sidenotes refer to the plate numbers at the top, while the 1735 and later sidenotes refer to the bottom numbers.  (However some of the new illustrations in vol. 4 are not described in the text and this makes me wonder if there was an earlier version with these new plates??)  I will give the 1725 top plate numbers, followed by the bottom numbers in  ( )  -- e.g. plate 12 (14).  The 1741 and the 1750/1749 eds. are essentially identical to the 1735 ed.

                    Ball, MRE, and MUS #52 say the 1750 and/or the 1770 ed. were revised by Montucla, but all other sources say his revision was 1778.  Indeed Montucla was only 25 in 1750.  Inspection of 1750 copies in the Turner Collection and at the Institute für Geschichte der Naturwissenschaft shows the 1750 is identical to my 1725 ed. except for some accents and a new publisher.]

     English versions

                    Recreations Mathematical and Physical; Laying down, and Solving Many Profitable and Delightful Problems of Arithmetick, Geometry, Opticks, Gnomonicks, Cosmography, Mechanicks, Physicks, and Pyrotechny.  By Monsieur Ozanam, Professor of the Mathematicks at Paris.  Done into English, and illustrated with very Many cuts.  R. Bonwick, et al., London, 1708. 

                    [Pp. 130-191 are omitted, but there is no gap in the text and the Contents also shows these pages are lacking.  Ball, MRE.  BMC.  NUC.  Hall, OCB, p. 165.  Hall, BCB 216.  UCL.  Toole Stott 520, noting the gap.  Bodleian.  This is a pretty direct translation of the 1696 French ed. or an early simple revision.  Prob. 18-20 of Cosmographie have been omitted.  C&B list this and say there were three later editions, though they then list the 1756 and 1790 editions.]

     2nd ed.

                    Recreations for gentlemen and ladies: or, Ingenious Amusements.   Being Curious and diverting sports and pastimes, natural and artificial.  With Many Inventions, pleasant Tricks on the Cards and Dice, Experiments, artificial Fireworks, and other Curiosities, affording variety of entertainments.  James Hoey, Dublin, 1756.  [Taken from Toole Stott entry.]

                    [NUC.  Hall, BCB 212.  Hall, OCB, p. 165, but giving the 1790 title.  Toole Stott 518, but he has Hoez, which seems to be either a misreading or a miswriting??]

     3rd ed. 

                    Recreations for Gentlemen and Ladies; being Ingenious Sports and Pastimes.  Containing Many curious Inventions, pleasant Tricks on Cards and Dice; Arithmetical Sports; new Games; Rules for Assuredly winning at all Games, whether of Cards or Dice; Recreative Fire-works; Tricks to promote Diversion in Company, and other Curiosities....  James Hoey, Dublin, 1759.  [Taken from Hall, BCB 213.]

                    [Ball, MRE.  Hall, BCB 213.  Hall, OCB, p. 165.  Bodleian.] 

     4th ed.

                    Recreations for gentlemen and ladies; being ingenious sports and pastimes: containing Many curious Inventions, pleasant Tricks on Cards and Dice; Arithmetical Sports; new Games; Rules for Assuredly winning at all Games, whether of Cards or Dice; Recreative Fire-works; Tricks to promote Diversion in Company, and other Curiosities....  The fourth edition.  Peter Hoey, Dublin, 1790.  [Taken from Toole Stott entry.]

                    [BMC calls it the 4th ed. [abridged].  NUC.  Hall, BCB 214.  Hall, OCB, p. 165.  Toole Stott 519.  Bodleian.]

 

                    I will cite these and the following editions by date, though the varying problem numbers and volume numbers will make this a bit unwieldy.  All references to the 4 volume versions are to vol. I unless specified otherwise.  See the entry in 4.A.1 as an example.  Note that I give the problem number first because this is usually the same in the 1790, 1803 and 1814 editions, and often in the 1840.  Figure numbers will also been given with the problem number, although the 1840 has the figures in the text with different numbers.  Additions will be entered at the relevant point.

Ozanam‑Hutton.  1803.

                                    Recreations in Mathematics and Natural Philosophy: ...  first composed by M. Ozanam, of the Royal Academy of Sciences, &c. Lately recomposed, and greatly enlarged, in a new Edition, by the celebrated M. Montucla.  And now translated into English, and improved with many Additions and Observations, by Charles Hutton, ....  4 vols., T. Davison for G. Kearsley, London, 1803. 

                    [This is a pretty direct translation of the 1790 Ozanam-Montucla ed. with a few changes, and some notes and extra material, generally in sections which do not interest us.  Only one or two plates have been changed.  Erroneous problem numbers have been retained.

                    There was an 1814 ed. by Longman, Hurst, Rees, Orme and Brown, London.  The texts are basically identical, but the 1814 has been reset to occupy about 15% fewer pages, the problem numbers have been corrected, a few corrections/additions have been made and the plates do not fold out.]

Ozanam‑Montucla.  1778.

                                    J. E. Montucla (1725-1799)'s revision of Ozanam. 

                    Récréations Mathématiques et Physiques, Qui contiennent les Problêmes & les Questions les plus remarquables, & les plus propres à piquer la curiosité, tant des Mathématiques que de la Physique; le tout traité d'une maniere à la portée des Lecteurs qui ont seulement quelques connoissances légeres de ces Sciences.  Par feu M. Ozanam, de l'Académie royale des Sciences, &c.  Nouvelle Edition, totalement refondue & considérablement augmentée par M. de C. G. F.  Claude Antoine Jombert, fils aîné, Paris, 1778, 4 vols.  Approbation & privilege dated 5 Aug 1775 & 4 Sep 1775. 

                    [BMC says this is "par M. de C. G. F. [i.e. M. de Chanla, géomètre forézien, pseudonym of J. E. Montucla.]"  BMC, under Montucla, says M. de Chanla is a pseudonym of Montucla.  BNC, under Montucla, describes Montucla as Éd.  Lucas, RM1, p. 242 lists this under Chaula "attribué à Montucla".  MUS #52 has Chaula.  NUC has Chanla.  Montucla's connection with the book was so little known that the 1778 version was sent to him in his role as Mathematical Censor and he made some additions to it before approving it.  Hutton says the 'last edition' (presumably 1790) bears Montucla's initials.  'Forézien' means from Feurs or the Forez region.  Reprinted in 1785-1786 and 1790 (see below).

                    This is a considerable reworking of the earlier versions.  In particular, the interesting material on conjuring and mechanical puzzles in Vol. IV has been omitted.  There are occasional misnumberings of problems.  I recall the plates are folding at the end of each volume, but I didn't note this specifically.]

     Nouv. ed.

                    "Nouvelle édition, totalement refondue et considérablement augmentée par M. de M***" [BMC adds: [i.e. Jean Étienne Montucla]].  Firmin Didot, Paris, 1790. 

                    [NUC lists this as a reissue of 1778 with new TPs.  I have this ed.]

Ozanam‑Riddle.  1840.

                                    (Edward Riddle (1788-1854)'s revision of Ozanam‑Hutton.) 

                    Recreations in mathematics and natural philosophy: translated from Montucla's edition of Ozanam, by Charles Hutton, LL.D. F.R.S. &c.  A new and revised edition, with numerous additions, and illustrated with upwards of four hundred woodcuts,  By Edward Riddle, Master of the Mathematical School, Royal Hospital, Greenwich.  Thomas Tegg, London, 1840.  [Taken from my copy.  MUS #130 asserts this is by a C. Biddle, but this must be due to be a misreading or misprint.]

     Another ed.

                    Recreations in science and natural philosophy: Dr. Hutton's translation of Montucla's edition of Ozanam.  The present new edition of this celebrated work is revised By Edward Riddle, Master of the Mathematical School, Royal Hospital, Greenwich, who has corrected it to the present era, and made numerous additions.  This Edition is also Illustrated by upwards of Four Hundred Woodcuts.  Thomas Tegg, London, 1844.

     Another ed.

                    Recreations in Science and Natural Philosophy.  Dr. Hutton's translation of Montucla's edition of Ozanam.  New Edition, revised and corrected, with numerous additions, by Edward Riddle, Master of the Mathematical School, Royal Hospital, Greenwich.  Illustrated by upwards of Four Hundred Woodcuts.  William Tegg, London, 1851.

     [Reprinted by William Tegg, 1854.]

 

                    All the Riddle printings seem to be from the same plates -- the date of the Prefatory Note and the Erratum given on p. xiv are the same in 1840 and 1851.  Marcia Ascher cites an 1844 edition from Nuttall & Hodgson, but this is the printer.

                    These have the figures in the text, but otherwise seem to be little different than Ozanam-Hutton.  I will generally cite it just as 1840.

 

P. M. Calandri.  See:  Benedetto da Firenze.

Pacioli.  De Viribus.  c1500.

                                    Luca Pacioli (or Paciuolo) (c1445-1517).  De Viribus Quantitatis.  c1500.  Italian MS in Codex 250, Biblioteca Universitaria di Bologna.  Santi 3 dates it as 1498 and describes the later parts which deal with riddles, etc., but include some magic, etc.  However, Pacioli petitioned for a privilege to print this in 1508 and Part 2, Chap. CXXIX, ff 228r-228v, has a date of 1509, so he may have been working on the MS for many years.  Van Egmond's Catalog 55-56 is not as helpful as usual. 

                    Part 1: Delle forze numerali cioe di Arithmetica is described in: A. Agostini; Il "De viribus quantitatis" di Luca Pacioli; Periodico di Matematiche (4) 4 (1924) 165‑192 (also separately published with pp. 1-28).  Agostini's descriptions are sometimes quite brief -- unless one knows the problem already, it is often difficult to figure out what is intended.  Further, he sometimes gives only one case from Pacioli, while Pacioli does the general situation and all the cases.  All references are to the problem numbers in this part, unless specified otherwise.  I will use problem numbers and names as in the MS at the problem -- these often differ considerably from the numbers and names given in the index at the beginning of the MS.  There are 81 problems in part 1, but the Index lists 120 -- in a few cases, the Index name clearly indicates a problem similar to an actual problem and I will mention this.  There are microfilms at the Warburg Institute (currently misplaced) and at Munich and Siena.  I printed Part 1 and some other relevant material from the Warburg microfilm, some 125pp.  When I read this, I saw that the material near the last pages I had copied would be of interest, but when I went back to copy this material, the microfilm had been misplaced.

                    Dario Uri has photographed the entire MS and enhanced the images and put them all on a CD.  This has 614 images, including the insides of the covers.  This is often more legible than the microfilm, but the folio numbers are often faint, sometimes illegible.  It was not until the transcription (below) became available that I could read the material just after this point and see that there were more interesting problems, but I found it difficult to read the Italian (many words are run together and/or archaic) and the diagrams referred to are lacking.  Dario was able to carry on and found the Chinese Rings and about a dozen other interesting items.  He has put some material up on his website:  http://digilander.libero.it/maior2000/.  This includes the indexes and a number of the most interesting items, with his comments and diagrams of later examples of the puzzles.  I have now gone through all of the text and found a number of problems of interest for this bibliography.  Nonetheless, quite a number of problems, some clearly of interest, remain obscure.

                    Transcription by Maria Garlaschi Peirani, with Preface and editing by Augusto Marinoni.  Ente Raccolta Vinciana, Milano, 1997.  (The publisher did not reply to a letter, but Bill Kalush kindly obtained a copy for me.  Dario Uri says it can be bought from:  Libreria Pecorini, 48 foro Buonaparte, Milano; tel: 02 8646 0660; fax: 02 7200 1462; web: www.pecorini.com.)  I will cite the text as Peirani and any references to Marinoni's work as Marinoni.  The transcription is not exactly literal in that Peirani has expanded abbreviations and inserted punctuation, etc.  Also, Peirani seems to have worked from the microfilm or a poor copy as she sometimes says the manuscript has an incorrect form which she corrects, but Dario Uri's version clearly shows the MS has the correct form.  Peirani uses the problem numbers and names in the MS (see comment above about these differing from those in the Index), but with some amendments.  I will probably give problem names as in the MS, with some of Peirani's amendments.  I will give translations of the names, but some of them are pretty uncertain and some have defeated me completely.  Marinoni, pp. VIII‑IX indicates the MS was written in about 1496-1509.

                    The title is a bit cryptic, but I think the best English version is: On the Powers of Numbers, but Pacioli [f. Ir] has 'forze numerali' and R. E. Taylor [op. cit. in Section 1, pp. 307 & 339] follows Agostini and uses: On the Forces of Quantity.

Pacioli.  Summa.  1494.

                                    Luca Pacioli (or Paciuolo or Paccioli) (??-c1509).  Sūma de Arithmetica Geometria Proportioni & Proportionalita.  Paganino de Paganini, Venice, 1494 -- cf Van Egmond's Catalog 325-326;  facsimile printed by Istituto Poligrafico e Zecca dello Stato, Rome, for Fondazione Piero della Francesca, Comune di Sansepolcro, 1994, with descriptive booklet edited by Enrico Giusti.  There was a second ed., Paganino de Paganinis, Toscolano, 1523, but the main text seems identical, except for corrections (and errors) and somewhat different usage of initial letters and colour.  For extensive studies of this book, see the works by Narducci, Taylor, Davis and Rankin given in Section 1.  Davis identifies material taken from Piero della Francesca's works.  Taylor says 99 copies of the 1494 and 36 copies of the 1523 are known.  The text is in two parts -- the second part is geometry and is separately paged.  All page references will be to the first part unless specified as Part II.  Problems are numbered at the right hand edge of the last line of the previous problem.  See Rara 54-59. 

                    Davis notes that Pacioli's Summa, Part II, ff. 68v - 73v, prob. 1-56, are essentially identical to della Francesca's Trattato, ff. 105r - 120r.

Panckoucke, André Joseph (1700-1753).  See:  Les Amusemens.

Pardon, George Frederick (1824-1884).  See:  Indoor & Outdoor;  Parlour Pastime;  Parlour Pastimes.

Parlour Pastime.  1857.

                                    Parlour Pastime for the Young: Consisting of Pantomime and Dialogue Charades, Fire-side Games, Riddles, Enigmas, Charades, Conundrums, Arithmetical and Mechanical Puzzles, Parlour Magic, etc. etc.  Edited by Uncle George [NUC says this is George Frederick Pardon].  James Blackwood, London, 1857 [Toole Stott 545; Christopher 724].  This was combined with Games for All Seasons [Toole Stott 311; Christopher 723] into Indoor & Outdoor, c1859, qv.  There is a later edition, Parlour Pastimes, 1868, qv.  Hence the problems will be cited as: Parlour Pastime, 1857  = Indoor & Outdoor, c1859, Part 1  = Parlour Pastimes, 1868.  Many of the problems are identical to Book of 500 Puzzles.

Parlour Pastimes.  1868?

                                    Parlour Pastimes: A Repertoire of Acting Charades, Fire-side Games Enigmas, Riddles, Charades, Conundrums, Arithmetical and Mechanical Puzzles, Parlour Magic, etc., etc.  James Blackwood, London, nd.  [BMC, NUC and Toole Stott 1136 date this 1868 and say it is by George Frederick Pardon.  Toole Stott 1136 indicates that By G. F. P. is on the TP, but it is not in my example.  Toole Stott 1137 is 1870, a slightly smaller ed.]  It is an expanded version of Parlour Pastime, with the material of interest to us being directly copied, though the page layout varies slightly.  The running head of this is actually Parlour Pastime.  Hence the problems will be cited as: Parlour Pastime, 1857  = Indoor & Outdoor, c1859, Part 1  = Parlour Pastimes, 1868.  Many of the problems are identical to Book of 500 Puzzles.

PCP.  1932.                  H. E. Dudeney.  Puzzles and Curious Problems.  Nelson, 1932; revised ed., nd [1936?].  (Almost all of this is in 536.)  There are almost no changes in the revised ed., except that problem 175 and its solution have been corrected -- it is a cross number puzzle and the text was for a different diagram.  See: 7.AM.

Peano.  Giochi.  1924.

                                    Giuseppe Peano (1858‑1932).  Giochi di Aritmetica e Problemi Interessanti.  G. B. Paravia, Torino, nd [1924 and later reprints].  (Thanks to Luigi Pepe for a photocopy of this.)

Pearson.  1907.             A. Cyril Pearson.  The Twentieth Century Standard Puzzle Book.  Routledge, London, nd [1907].  Three parts in one volume, separately paginated.  The parts were also published separately.  Each part has several numbered sequences of problems.

Peck & Snyder.  1886.

                                    Price List of Out & Indoor Sports & Pastimes.  Peck & Snyder, 126‑130 Nassau Street, N. Y., 1886.  Reprinted, with some explanatory material, in the American Historical Catalogue Collection, Pyne Press, Princeton, 1971.  Unpaginated -- I have numbered the pages starting with 1 as the original cover.

Peirani, Maria Garlaschi.  See:  Pacioli.  De Viribus.  c1500.

Perelman.  FFF.  1934.

                                    Yakov Isidorovich Perelman [Я. И. Перелман] (1882-1942).  Figures for Fun.  Живая Математика [Zhivaya Matematika], Наука [Nauka], Moscow.  The books give no indication of the original dates, but Tatiana Matveeva has kindly searched the Russian State Library and found it was originally published by Гос. техн.-теор. издат., Leningrad-Moscow, 1934.  Schaaf I 9 cites Recreational Arithmetic, 6th ed., Leningrad, 1935 and Sphinx 5 (1935) 96 reviews the 5th ed. of L'Arithmétique Récréative, Leningrad, 1934 -- both presumably the same book ??

                    Translated by G. Ivanov‑Mumjiev.  Foreign Languages Publishing House, Moscow, 1957, 120 sections. 

                    (2nd ed., 1973 -- used for MCBF, below, apparently the same as the 3rd ed.) 

                    Translated by G. Ivanov‑Mumjiev.  3rd ed., MIR, 1979, 123 sections.  The 3rd ed. drops 3 sections and adds 6 sections and has some amended English.

Perelman.  FMP.  1984.

                              Yakov Isidorovich Perelman [Я. И. Перелман].  Fun with Maths and Physics.  MIR, Moscow, 1984.  [This is a translation of Занимательныи Задачи и Оыты [Zanimatel'nye Zadachi i Oryty], Детская Литературы [Detskaya Literatura], Moscow.)  (There was a Занимательныи Задачи, with 4th ed. in 1935 – no earlier version in Russian State Library.  This title originally published by Деттиз, 1959.)  Compiled by I. I. Pruskov.  Translated by Alexander Repyev.  This is a compilation from several of his books from 1913 to c1942 (when he died).  I have not yet seen the earlier books (??NYS), but if this is mainly based on the earlier book of the similar title, this would date the material to c1935? and I will use this date.

Perelman.  MCBF.  c1980?

                                    Yakov Isidorovich Perelman [Я. И. Перелман].  Mathematics Can Be Fun.  3rd ed., MIR, 1985.  This consists of the following, originally separate, works, but with the second part having its page and problem numbers continued from the first part.  Both these works exist in many other editions and translations.

                    Figures for Fun, translated from Живая Математика [Zhivaya Matematika], Наука [Nauka], Moscow, translated 1973 -- with 123 sections.  Translator not specified, but presumably G. Ivanov-Mumjiev, as in FFF above.

                    Algebra Can be Fun, translated from Занимательная Алгебра [Zanimatel'naya Algebra] (3rd ed. was published by ОНТИ, Leningrad-Moscow, 1937 -- The Russian State Library apparently has no earlier edition), edited and supplemented by V. Boltyansky, Наука [Nauka], Moscow, 1976, translated by G. Yankovsky, 1976. 

                    References to this will be to material not in FFF, so will be dated 1937.

Phillips, Hubert (1891-1964).  See:  Brush;  Week‑End.

Pike.  Arithmetic.  1788.

                                    Nicolas Pike (1743-1819).  A New and Complete System of Arithmetic, composed for the Use of the Citizens of the United States.  John Mycall, Newbury-Port, Massachusetts, 1788.  (I have a 2nd ed., 1797, [Halwas 318], ??NYR.  This went through at least 5 editions and then at least six variants, often abridged for schools [Halwas 318-326].)

Poggendorff. J. C. Poggendorff.  Biographisch-Literarisches Handwörterbuch zur Geschichte der Exacten Wissenschaften enthaltend Nachweisungen über Lebensverhältnisse und Leistungen von Mathematikern, Astronomen, Physikern, Chemikern, Mineralogen, Geologen usw aller Völker und Zeiten.  Johann Ambrosius Barth, Leipzig.  Facsimile by Maurizio Martino, Storrs-Mansfield (later Mansfield Center), Connecticut.

                    Vols. I (A - L) and II (M - Z).  (1863), nd [bought in 1996].

                    Vol. III (1858-1883), edited by B. W. Feddersen & A. J. von Oettingen;

                                    Parts I (A - L) and II (M - Z).  (1898), nd [bought in 1998].

                    Vol. IV (1883-1904), edited by Arthur von Oettingen;

                                    Parts I (A - L) and II (M - Z).  (1904), nd [bought in 1998].

                    [Vol. V covers 1904-1922.  Vol. VI covers 1922-1949.  One can get I-VI on microfiche.  Vol. VIIa covers 1932-1953 and apparently comprises 5 volumes.  There is also a VIIa Supplement which gives material supplementary to vols. I-VI.]

Prévost.  Clever and Pleasant Inventions.  (1584), 1998.

                                    J. Prévost.  (La Première Partie des Subtiles et Plaisantes Inventions, Contenant Plusieurs Jeux de Récréation.  Antoine Bastide, Lyons, 1584.  ??NYS.)  Translated by Sharon King as:  Clever and Pleasant Inventions  Part One  Containing Numerous Games of Recreations and Feats of Agility, by Which One May Discover the Trickery of Jugglers and Charlatans.  Hermetic Press, Seattle, 1998.  [No Second Part ever appeared.  Hall, OCB, pp. 43, 100 & 113.]  This is apparently the first book primarily devoted to conjuring.  Only five copies of the original are known.  There was a facsimile in 1987.  My thanks to Bill Kalush for bringing this work to my attention.

Price, Harry (1881-1948).  See: HPL.

Problemes.  1612.

                                    Claude‑Gaspar Bachet (c1587-1638).  Problèmes plaisans & délectables qui se font par les nombres.  (1st ed., 1612);  2nd ed., 1624, P. Rigaud, Lyon (for 1 & 2 ed.);  revised by A. Labosne, (3rd ed., 1874; 4th ed., 1879); 5th ed., 1884, Gauthier‑Villars, Paris (for 3, 4, 5 ed.).)  5th ed. reprinted by Blanchard, Paris, 1959 et al., with a Frontispiece portrait and an introduction by J. Itard, based on the article by Collet and Itard cited in 1 below. 

                    I have now obtained a photocopy of the 2nd ed. and have examined a 1st ed.  I had believed that Bachet added 10 problems in the 2nd ed., but the additional section of 10 problems, beginning  "S'Ensuivent quelques autres ..." is already in the 1612 1st ed.  In the 1612, there are two problems  V, but in 1624, these are made into two parts of prob. V.  However, he does extend the initial section of 22 problems to 25 problems, inserting the new material as problems 3, 16 and 21.  Prob. 16 (1612) = 18 (1624) has additional material.  Also, Bachet greatly expands his preliminary material on the properties of numbers from 14 to 52 pages, but Labosne drops this.  Otherwise, the material seems identical and the main text seems pretty much identical with the fifth edition except that orthography is modernised -- e.g. plaisans becomes plaisants, mesme becomes même, luy becomes lui, etc.  I have now compared the 3rd ed with the 5th ed and I could find no differences between them -- though I didn't check every word.  Labosne adds a Supplement of 15 problems, four Notes and a table of contents.  Labosne's Préface given in the 5th ed. is for the 3rd ed.  I will cite problem numbers and pages from the 1st ed., 1612; 2nd ed., 1624 and the 5th ed., 1884 (1959 reprint), e.g.  Prob. XIX, 1612, 99-103.  Prob. XXII, 1624: 170-173;  1884: 115‑117.  I will generally not give problem titles as they usually run to several lines.  I will denote Labosne's supplementary problems as Bachet-Labosne, 1874.

                    I have seen a 4th ed. by Gauthier‑Villars, 1905, no editor named, containing only 37 of the 50 problems in the 5th ed.  A contemporary review by E. Lampe (Fortschritte der Math. 36 (1905) 300‑301) was also mystified by this edition.  C&B list this ed.

Pŗthudakasvâmî or Pŗthūdaka.  See: Chaturveda.

Pseudo-dell'Abbaco.  c1440.

                                    This is attributed to Paolo dell'Abbaco (sometimes called Dagomari) (c1281-1367).  Trattato d'Aritmetica.  (c1370, according to Arrighi, but see below).  Codex Magliabechiano XI, 86 at Biblioteca Nazionale di Firenze.  Edited by Gino Arrighi, Domus Galilaeana, Pisa, 1964.  Arrighi gives some black & white reproductions of illustrations.  I examined this MS in Sep 1994 and found the illustrations are often lightly coloured and that Arrighi's illustrations were probably made from poorish photocopies -- the writing on the opposite side shows through much more in several of his illustrations than it does in the originals.  I have colour slides of 11 pages. 

                    Warren Van Egmond [New light on Paolo dell'Abbaco; Annali dell'Istituto e Museo di Storia della Scienza di Firenze 2:2 (1977) 3‑21 and Van Egmond's Catalog 114-115] asserts this MS is a c1440 compilation, based on watermark evidence, and doubts that it is due to dell'Abbaco, giving the author as pseudo-dell'Abbaco. 

                    Smith, Rara, 435‑440 describes a different MS at Columbia, headed 'Trattato d'Abbaco, d'Astronomia e di segreti naturali e medicinali', which he dates c1339.  Van Egmond, above, gives the title of this as 'Trattato di tutta l'arte dell'abacho', but Van Egmond's Catalog 254-255 describes it as Plimpton 167, a codex containing two works.  The first is the dell'Abbaco: Trattato di tutta l'arte dell'abacho; the second is Rinaldo da Villanova: Medichamento Generale, which has the title Trattato d'Abbaco, d'Astronomia e di segreti naturali e medicinali added in a later hand.  The first includes the Regoluzze which is sometimes cited separately.  This is quite a different book than the c1440 Trattato.  There is a c1513 version at Bologna, MS B 2433, which is dated 1339 -- Dario Uri has sent a CD of images of it.  See the entry for dell'Abbaco in 7.E.

Putnam.  Puzzle Fun.  1978.

                                    Graham R. Putnam, ed.  Puzzle Fun.  Fun Incorporated, np [Chicago?], 1978.

 

Rara.  1970.                  David Eugene Smith.  Rara Arithmetica.  (1908; with some addenda, 1910; Addenda, 1939, published both separately and with the 1910 ed.); 4th ed., combining the original with both Addenda and with De Morgan's Arithmetical Books of 1847 and a new combined index, Chelsea, 1970.  References are to the main entry of this.  Check the index for references to the Addenda and to De Morgan.

Rational Recreations.  1824.

                                    Rational Recreations.  Midsummer MDCCCXXIV.  Knight and Lacey, London.  This is a six part work, but is bound together -- perhaps the parts were issued monthly.  The parts are consecutively paginated.  [Toole Stott 590.  Toole Stott 591 is 2nd ed, 1825 and 592 is 3rd ed, 1825, copublished in Dublin.  Hall BCB 235, 236 are 1824 and 1825.  C&B.  HPL.  Not in Christopher.  I have examined the BL copy.]

Recorde.  First Part.  1543.

Recorde.  Second Part.  1552.

Recorde-Mellis.  Third Part.  1582.

                                    Recorde (or Record), Robert (1510?-1558).  The Grounde of Artes Teaching the worke and practice of Arithmetike.  ....  The dating of this book is uncertain.  Smith, Rara, p. 526 records a 1540 edition.  An edition by  Reynold Wolff, London, at the Bodleian (Douce R.301) has generally been dated as 1542 and there is a facsimile by  Theatrum Orbis Terrarum, Amsterdam  &  Da Capo Press, NY,  1969, with the date 1542.  There were reprints in 1543 and 1549.  However, the DSB entry for Recorde ignores the Smith 1540 edition (presumably because it has not been confirmed) and says the 1542 is now dated as 1550?, making the 1543 the first edition.  This edition only contained material on whole numbers. 

                    In 1552, Recorde added a Second Part dealing with fractions, so the earlier material will be called the First Part. 

                    At some stage, John Dee augmented it, but it appears he simply made some revisions and additions to the existing text without adding new topics.  The Dee material was added in 1590 (Smith) or 1573 (De Morgan). 

                    In 1582, John Mellis added a Third Part, mostly on rules of calculation, published by J. Harison & H. Bynneman, London. 

                    By 1640, the title was changed to:  Record's Arithmetick, or, The Ground of Arts; Teaching The perfect Work and Practice of ARITHMETICK, both in whole Numbers and Fractions, after a more easie and exact form then in former time hath been set forth.  Afterwards augmented by Mr. JOHN DEE.  And since enlarged with a third part of RULES of PRACTICE, abridged into a briefer method then hitherto ..., by JOHN MELLIS. 

                    By 1648, more material was added by Robert Hartwell.  I have a 1668 edition which has a little more material by Thomas Willsford, but these latter two extensions are of no interest to us.  It is clear that the text was increased by accretion, with only minor revisions of Recorde's text, which is generally preserved in gothic (= black-letter) type, and this is indicated by Smith.  So the presence of a problem in Part One or Part Two or Part Three of the 1668 ed almost certainly indicates its presence in the first version of these parts.  This is certainly true for Part One, as I have the facsimile to compare, and it is confirmed by brief examination of a 1582 ed, though at the time, I was looking at Part Three and did not know of the material in Part Two.  I will cite the pages from my 1662 ed and the side notes which are titles of the problems, and pages of any earlier editions that I have seen.

Riccardi.                       Pietro Riccardi.  Biblioteca Matematica Italiana dalla Origine della Stampa ai Primi Anni del Secolo XIX.  G. G. Görlich, Milan, 1952, 2 vols.  This work appeared in several parts and supplements in the late 19C and early 20C, mostly published by the Società Tipografica Modense, Modena, 1878-1893.  For details, see in Section 3.B.

Riddle, Edward (1788-1854).  See:  Ozanam‑Riddle.

The Riddler.  See under Boy's Own Book.

Riese.  Coss.  1524.

                                    Adam Riese (c1489-1559).  Die Coss.  German MS of 1524 found at Marienberg in 1855.  Described and abstracted in Programm der Progymnasial‑ und Realschulanstalt zu Annaberg 1860.  Reprinted in 1892.  My reference to this comes from Johannes Lehmann; Rechnen und Raten; Volk und Wissen, Berlin (DDR), 1987, pp. 7‑14, esp. p. 13.  I have since seen the Glaisher paper, op. cit. in 7.G.1 under Widman, esp. p. 37.  Glaisher and Lehmann cite: Bruno Berlet; Adam Riese, sein Leben und seine Art zu rechnen; Die Coss von Adam Riese; Leipzig & Frankfurt, 1892.  Glaisher notes that this was a pamphlet.  BLLD provided a copy from Biblioth. Regia Berolinen. G., but it was lacking a title page.  It seems to have the title: Zur Feier des vierhundertsten Geburtsjahres von Adam Riese.  It was printed by Königl. Universitätsdruckerei von H. Stürtz in Würzburg.  The Vorwort is dated 1892, but only signed 'Der Verfasser' and his name does not appear anywhere except on the spine of the library's cover.  The booklet has two parts.

                    Adam Riese, sein Leben, seine Rechenbücher und seine Art zu rechnen (from the Programm for 1855), pp. 1‑26.  This is a discussion of Riese's Rechnung, but it also mentions some material from his 'grosse Rechenbuch' titled Rechenung nach der lenge, auf den Linihen und Feder, written in 1525 but not published until 1550.  Glaisher, loc. cit., p. 43, says he sees no authority for the date of 1525 and assumes it was written c1550.  (I have recently obtained a 1976 reprint of this work, ??NYR)

                    Die Coss von Adam Riese (with Abdruck der Coss) (from the Programm for 1860), pp. 27‑62.  This gives many numbered problems -- I will cite the problem number and pages from this.

                    There is a recent facsimile of the MS which I have just received -- ??NYR.

Riese.  Rechnung.  1522.

                                    Adam Riese (c1489-1559).  Rechnung auff der Linien unnd Federn ...  Erfurt, 1522.  I have two reprinted editions.  See Rara 138-143.

                    Christian Egenolph, Frankfurt, 1544.  (Riese's text is dated 1525.  There is a supplement on gauging by Erhard Helm, dated 1544.)  Facsimile by Th. Schäfer, Hannover, 1978.

                    Christian Egenolff's Erben, Frankfurt, 1574.  (Riese's text is dated 1525 and appears to be the same text as above, but reset.  The Supplement has further material.)  Facsimile by Th. Schäfer, Hannover, 1987.

Ripley's Puzzles and Games.  1966.

                                    Ripley's Believe It or Not!  Puzzles and Games.  Essandess Special Edition (Simon & Schuster), New York, 1966.  Much of the material occurred in the various Ripley's Believe it or Not! books.  Most of the material is well known, but there are a number of unusual variations and some interesting incomplete assertions and mistakes!

RM.                             (François-) Édouard (-Anatole) Lucas (1842-1891).  Récréations mathématiques.  (Gauthier‑Villars, Paris, 4 vols, 1882, 1883, 1893, 1894, 2nd eds. of vol. 1, 1891, vol. 2, 1893)  = Blanchard, Paris, (1960), 1975‑1977, using 2nd ed. of vol. 1 and 1st ed. of vol. 2 (however, there seem to be very few differences in the editions).  I will cite the Blanchard reprint volumes as RM1, etc.  (Dates are as given in Harkin, op. cit. in 1 below, and on the books, but I have seen other dates cited.) 

                    Lucas; L'Arithmétique Amusante, 1895, Note IV, pp. 210-260 gives various fragments of material for further volumes which were found after Lucas's untimely death.  His draft Tables of Contents for volumes 5 and 6 are given on p. 210, but no material exists for most of the chapters.

RMM.                          Recreational Mathematics Magazine.  Nos. 1 - 14 (Feb 1961 - Jan/Feb 1964).  Quite a bit of the material in this was abstracted and sometimes extended in: Joseph S. Madachy; Mathematics on Vacation; Scribner's, NY, 1966; somewhat corrected as: Madachy's Mathematical Recreations; Dover, 1979.

Rocha.  Libro Dabaco.  1541.  SEE:  Tagliente.  Libro de Abaco.  (1515).  1541.

Rohrbough.   Lynn Rohrbough edited a series of 20 booklets, called Handy Series, Kits A-J and M‑V, for Cooperative Recreation Service, Delaware, Ohio, during at least 1925-1941.  Most of these were reprinted and revised several times.  Two of these are of especial interest to us and are listed below.  Several others are cited a few times.

Rohrbough.  Brain Resters and Testers.  c1935.

                                    Lynn Rohrbough, ed.  Brain Resters and Testers.  Handy Series, Kit M, Cooperative Recreation Service, Delaware, Ohio, nd [c1935].

Rohrbough.  Puzzle Craft.  1932.

                                    Lynn Rohrbough, ed.  Puzzle Craft  Plans for Making and Solving 40 Puzzles in Wire, Wood and String.  Handy Series, Kit U, Cooperative Recreation Service, Delaware, Ohio, (1930), 1932. 

                    I have another version of this, unfortunately undated, but apparently later, so I have dated it as 1940s?  Jerry Slocum located this by tracking down Rohrbough's successors.  The 1932 version has 39 puzzles in its index, but 4 more that were not indexed and some Notes which are listed in the index of this version, making 44 items in all.  13 items are omitted and replaced by 5 in the present version, giving 36 indexed items.  The outside of the back cover of the 1932 version shows a number of puzzles, including several not described in either version of the booklet. 

Rudin.  1936.                Jacob Philip Rudin.  So You Like Puzzles!  Frederick A. Stokes Co., NY, 1936.

 

SA.                               Scientific American, usually Martin Gardner's Mathematical Games column.  For years from at least 1950, SA appeared in two volumes per year, each of six issues.  In year  1950 + n,  vol. 182 + 2n  covers  Jan-Jun  and  vol. 183 + 2n  covers  Jul-Dec.  See also under Gardner.

Sanford.  1930.             Vera Sanford.  A Short History of Mathematics.  Houghton Mifflin, Boston, 1930 & 1958.  See also:  H&S.

Santi.  1952.  Aldo Santi.  Bibliografia della Enigmistica.  Sansoni Antiquariato, Florence, 1952.  2541 entries, often citing several editions and versions, arranged chronologically from 1479 onward.  My thanks to Dario Uri for providing this.  I will site item numbers.  I have only entered part of the information so far.

de Savigny.  Livre des Écoliers.   1846.

                                    M. l'Abbé de Savigny.  Le Livre des Écoliers  Illustré de 400 vignettes.  Jeux. -- Récréations.  Exercises. -- Arts utiles et d'agrément.  Amusements de la science.  Gustave Havard, Paris, nd [dealer has written in 1846], HB.  [This is almost entirely the same as  Boy's Own Book, 1843 (Paris).  Each has a few sections the other does not.  de Savigny's illustrations seem to have been copied by a new hand, generally simplifying a little.  Many of the copies have been done in reverse and this leads to one erroneous chessboard.  However, there is one diagram in Boy's Own Book, 1843 (Paris), which looks like it was badly copied, so it is possible that both these books are based on an earlier book.]

Schaaf.  1955-1978.

                                    William L. Schaaf.  A Bibliography of Recreational Mathematics.  Vol. 1, (1955, 1958, 1963); 4th ed., 1970.  Vol. 2, 1970.  Vol. 3, 1973.  Vol. 4, 1978.  National Council of Teachers of Mathematics, Washington, DC.  See Schaaf & Singmaster in Section 3.B for a Supplement to these.

Schott.  1674.                Gaspare Schott.  Cursus Mathematicus.  Joannis Arnoldi Cholini, Frankfurt, 1674.  The material of interest is in Liber II, Caput VI: De Arithmetica Divinatoria, pp. 57-60, and in Liber XXVI: Algebra, Pars III: De Exercitatione Algebraicam, Caput I, II, IV, VI, pp. 551-563, and Pars V: Exercitationes Algebraicae, pp. 570-571.

Schwenter.  1636.

                                    Daniel Schwenter (1585-1636).  Deliciæ Physico‑Mathematicae.  Oder Mathemat‑ und Philosophische Erquickstunden, Darinnen Sechshundert Drey und Sechsig, Schöne, Liebliche und Annehmliche Kunststücklein, Auffgaben und Fragen, auf; der Rechenkunst, Landtmessen, Perspectiv, Naturkündigung und andern Wissenschafften genomēn, begriffen seindt, Wiesolche uf der andern seiten dieses blats ordentlich nacheinander verzeichnet worden: Allen Kunstliebenden zu Ehren, Nutz, Ergössung des Gemüths und sonderbahren Wolgefallen am tag gegeben Durch M. Danielem Schwenterum.  Jeremiæ Dümlers, Nuremberg, 1636.  [Note: the text is in elaborate Gothic type with additional curlicues so that it is not always easy to tell what letter is intended!]  Probably edited for the press by Georg Philip Harsdörffer.

                    Extended to three volumes by Harsdörffer in 1651 & 1653, with vol. 1 being a reprint of the 1636 vol.  Vol. 2 & 3 have titles as follows.

                    Delitiæ Mathematicæ et Physicæ  Der Mathematischen und Philosophischen Erquickstunden  Zweiter Theil:  Bestehend in Fünffhundert nutzlichen und lustigen Kunstfragen / nachsinnigen Aufgaben / und derselben grundrichtigen Erklärungen / Auss Athanasio Kirchero Petro Bettino, Marion Mersennio, Renato des Cartes, Orontio Fineo, Marino Gethaldo, Cornelio Drebbelio, Alexandron Tassoni, Sanctorio Sanctorii, Marco Marco, und vielen anderen Mathematicis und Physicis zusammen getragen durch Georg Philip Harsdörffern.  Jeremia Dümlern, Nürnberg, 1651.

                    Delitiæ Philosophicæ et Mathematicæ  Der Philosophischen und Mathematischen  Erquickstunden / Dritter Theil:  Bestehend in Fünffhundert nutzlichen und lustigen Kunstfragen / und derselben gründlichen Erklärung: Mit vielen nothwendigen Figuren / so wol in Kupffer als Holz / gezieret.  Und  Aus allen neuen berühmten Philosophis und Mathematicis, mit grossem Fleiss zusammen getragen.  Durch  Georg Philip Harsdörffern.  Wolffgang dess Jüngern und Joh. Andreas Endtern, Nürnberg, 1653.

                    This 3 vol. version was reprinted in 1677 and 1692.  Modern facsimile of the 3 vol. version edited by Jörg Jochen Berns, Keip Verlag, Frankfurt Am Main, 1991, HB.  See also MUS II 325-326.  Schott described this as a German translation of van Etten/Leurechon, but this is quite wrong.  V&T, p. 152, say it is 'partially derived from' van Etten/Leurechon.  C&B list just the 1636, under Schwenterum.  P. 549 of vol. 1 is misprinted 249 which indicates that it was the first issue of the 1st ed.

                    I have not yet entered all the items from this.

The Secret Out.  1859.

                                    The Secret Out; or, One Thousand Tricks with Cards, And Other Recreations.  Illustrated with over three hundred engravings.  And containing  Clear and comprehensive explanations how to perform with ease, all the curious card deceptions, and slight of hand tricks extant.  With an endless variety of entertaining experiments in drawing room or white magic, including the celebrated science of second sight.  Together with a choice collection of intricate and puzzling questions, amusements in chance, natural magic, etc., etc., etc.  By the Author of "The Sociable, or, One Thousand and One Home Amusements," "The Magician's Own Book," etc, etc.  Dick & Fitzgerald, NY, 1859.

                    [Toole Stott 191, listing all versions under Cremer.  C&B, under Frikell, have New York, 1859.  H. A. Smith dates this as 1869.]

The Secret Out (UK).  c1860.

                                    The Secret Out or, One Thousand Tricks in Drawing-room or White Magic, with an Endless Variety of Entertaining Experiments.  By the author of "The Magician's Own Book."  Translated and edited by W. H. Cremer, Junr.  With three hundred illustrations. 

                    I have seen several editions.  [Toole Stott lists all versions under Cremer.]

                    C&W (based on the John Camden Hotten, London, 1871?) (with ads from Sep 1886 at back and inscription dated 12 Oct 1887 on flyleaf).  [NUC; Toole Stott 192; C&B, under Cremer, have London & New York, 1871, and under Frikell, have London, 1870.] 

                    C&W, nd [1871? -- NUC lists several dates; Toole Stott 1013 is 1870; Christopher 242 is 1878?]. 

                    John Grant, Edinburgh, nd [1872 -- Toole Stott 1014, no ads]. 

                                    All these copies have identical green covers with five magic tricks on the cover.

                    [Toole Stott 192 discusses the authorship, saying that Wiljalba (or Gustave) Frikell is named on the TP of some editions, but that most of the tricks are taken from the US ed of The Magician's Own Book.  In the US ed, The Author acknowledges his indebtedness to The Sociable and The Magician's Own Book 'and many other works of similar character and value', but claims 'that the greater portion of it [i.e. the book] is entirely original.'  In the Preliminary to the UK ed he says he is indebted to  '"Le Magicien des Salons," revised by references to  Decremps, Servière, Leopold, Besson, Kircher, Hildebrandt, Ozanam, &c., &c.'  though an 1874 ad by C&W indicates that it is translated from Le Magicien des Salons.  (This may be Le Magicien de Société, Delarue, Paris, c1860, but see Rulfs, below.)  The back of the TP of Bellew's The Art of Amusing, Hotten, 1866?, says The Secret Out is a companion volume, just issued, by Hermann Frikell.  BMC & Toole Stott say it is also attributed to Henry L. Williams.  Toole Stott 481 cites a 1910 letter from Harris B. Dick, of the publishers Dick & Fitzgerald, who thinks their version of The Secret Out "was a reprint of an English book by W. H. Cremer" -- but there seems to be no record of a UK ed before the US one.  NUC says an 1871 ed. gives author as Gustave Frikell.  Christopher 240-242 are two copies from Dick & Fitzgerald, c1859, and a C&W, 1878?  He repeats most of the above comments from Toole Stott and 242 cites the Rulfs article mentioned under Magician's Own Book, above.  Rulfs says The Secret Out is largely taken, illustrations and all, from Blismon de Douai's Manuel du Magicien (1849) and Richard & Delion's Magicien des salons ou le diable couleur de rose (1857 and earlier).  H. A. Smith [op. cit. under Magician's Own Book] says the first US ed is 1869 (this must be a misprint or misreading -- though the date is a little hard to read in my copy, it is clearly 1859) and the UK eds are basically a condensed version with a few additions.  He suggests the book is taken from DeLion.  He doubts whether Cremer ever wrote anything.  C&B, under Gustave Frikell, say it is a translation of Richard & Delion.  C&B, under Herrman Frikell, list London, 1870.  C&B, under Secret, list New York, nd.  C&B also list it under Williams, as London, 1871.]

                    [Toole Stott 1056 is [Frikell, Wiljalba]; Parlor Tricks with Cards, ...; By the Author of Book of Riddles and 500 Home Amusements, etc.; Dick & Fitzgerald, 1860?; which is described as "abridged from The Secret Out.  Toole Stott 547 and 1142 are two versions of 1863, but without the description of the author and hence listed anonymously.]

                    The US and UK editions are fairly different.  The US ed has 382 sections, of which 157 (41%) are used in the UK ed.  The UK ed has 323 sections, so 51% of it is taken from the US ed.  The US ed seems like a magic book, with chapters on Scientific Amusements and Miscellaneous Tricks.  The UK ed has much less magic and tricks, adding other general tricks and a lot more scientific tricks.  The illustrations for the common sections are not quite identical -- one was probably copied from the other.  The amount taken from The Magician's Own Book and The Sociable is fairly small, perhaps 10% from each, in either edition.

Shortz.                          Will Shortz's library or his catalogues thereof, called "Puzzleana".  The most recent I have is:  May 1992, 88pp with 1175 entries in 26 categories, with indexes of authors and anonymous titles.  Some entries cover multiple items.  In Jan 1995, he produced a 19pp Supplement extending to a total of 1451 entries.

SIHGM.  1939-1941.

                                    Ivor Bulmer Thomas.  Selections Illustrating the History of Greek Mathematics.  Loeb Classical Library, 1939‑1941, 2 vols.  I will give volume and pages as in  SIHGM I 308‑309.

Simpson.  Algebra.  1745.

                                    Thomas Simpson (1710-1761).  A Treatise of Algebra; Wherein the Fundamental Principles Are fully and clearly demonstrated, and applied to the Solution of a great Variety of Problems.  To which is added, The Construction of a great Number of Geometrical Problems; with the Method of resolving the same Numerically.  John Nourse, London, 1745. 

                    I also have the 7th ed., with the slightly different title:  A Treatise of Algebra.  Wherein the Principles Are Demonstrated, And Applied In many useful and interesting Enquiries, and in the Resolution of a great Variety of Problems of different Kinds.  To which is added, The Geometrical Construction of a great Number of Linear and Plane Problems; with the Method of resolving the same Numerically.  The Seventh Edition, carefully Revised.  F. Wingrave, Successor to Mr. Nourse, London, 1800. 

                    I have now seen a 6th ed., F. Wingrave, 1790 and it appears identical to the 7th ed.  Both have an Author's Preface to the Second Edition. 

                    I will give the 1745 details with the 1790/1800 details in parenthesis like (1790: ...).

SLAHP.  1928.             Sam Loyd Jr. (1873-1934)  Sam Loyd and His Puzzles.  An Autobiographical Review.  Barse & Co., NY, 1928.  (This contains somewhat more original material than I expected, but he claims that he devised a lot of his father's puzzles given in the Cyclopedia, OPM, and even earlier.)

Slocum.  Compendium.  1977.

                                    Jerry (= G. K.) Slocum.  Compendium of Mechanical Puzzles from Catalogues.  Published by the author, Beverly Hills, 1977, 57pp.  This is a compendium of illustrations and descriptions from 30 catalogues.  The earliest ones are Bestelmeier, 1793, 1807 (Jacoby edition); The Youth's Companion, 1875; Montgomery Ward, 1886, 1889, 1903, 1930; Peck & Snyder, 1886; Joseph Bland, c1890.  Others of some interest are: Gamage's, 1913, c1915, c1928; Johnson Smith, 1919, 1935, 1937, 1938, 1942.

Slocum & Gebhardt -- see under Catel.

SM.                              Scripta Mathematica.

Smith, David Eugene (1860‑1944).  See:  Rara  and the following four items.

Smith.  History.  1923.

                                    David Eugene Smith.  History of Mathematics.  Two vols.  (Ginn, NY, 1923)  = Dover, 1958.

Smith.  Number Stories.  1919.

                                    David Eugene Smith.  Number Stories of Long Ago.  NCTM, (1919), reprinted 1968?  Chaps. IX and X.

Smith.  Source Book.  1929.

                                    David Eugene Smith.  A Source Book in Mathematics.  Two vols.  (1929)  = Dover, 1959.

Smith & Mikami.  1914.

                                    David Eugene Smith  &  Yoshio Mikami.  A History of Japanese Mathematics.  Open Court, Chicago, 1914.

The Sociable.  1858.

                                    The Sociable; or, One Thousand and One Home Amusements.  Containing Acting Proverbs; Dramatic Charades; Acting Charades, or Drawing-room Pantomimes; Musical Burlesques; Tableaux Vivants; Parlor Games; Games of Action; Forfeits; Science in Sport, and Parlor Magic; and a Choice Collection of Curious Mental and Mechanical Puzzles; &c,&c.  Illustrated with nearly three hundred engravings and diagrams, the whole being a fund of never-ending entertainment.  By the author of "The Magician's Own Book."  (Dick & Fitzgerald, NY, 1858 [Toole Stott 640 lists this as anonymous; C&B list it under the title, with no author]); G. G. Evans, Philadelphia, nd, but the back of the title gives  ©1858 by Dick & Fitzgerald, NY, so this copy seems to be a reprint of the 1858 book.  Cf. Book of 500 Puzzles for discussion of possible authorship.  The Preface here says that most of the Parlor Theatricals are by Frank Cahill and George Arnold -- Toole Stott opines that this reference led Harry Price to ascribe this and the related books to these authors.  My thanks to Jerry Slocum for providing a copy of this.

                    The entire section Puzzles and Curious Paradoxes, pp. 285-318, is identical to the same section in Book of 500 Puzzles, pp. 3-36.  Rulfs (see under Status of the Project in the Introduction) says this draws on the same sources as Magician's Own Book, with more taken from Endless Amusement and Parlour Magic. 

                    See also:  Book of 500 Puzzles;  Boy's Own Conjuring Book;  Illustrated Boy's Own Treasury;  Indoor and Outdoor;  Landells: Boy's Own Toy-Maker;  The Secret Out;  Hanky Panky.

SP.                               Prefixed by  ??  is a flag to check spelling.

"The Sphinx".  See:  Lemon.

Sridhara.  c900.             Śrīdharācārya.  Patiganita (= Pâţîgaņita [NOTE:  ţ, ņ  denote  t, n  with underdot.]).  c900.  Transcribed and translated by Kripa Shankar Shukla.  Lucknow Univ., Lucknow, 1959.  The text is divided into verses and examples, separately numbered by the editor.  I will cite verse (v.) and example (ex.) and the page of the English text.  The editor has appended answers on pp. 93‑96, some of which were given by an unknown commentator.  (I have seen this dated 8C, which would put it before Mahavira -- ??)

SSM.                            School Science and Mathematics.

Struik.  Source Book.  1969.

                                    D. J. Struik (1894-  ), ed.  A Source Book in Mathematics 1200‑1800.  Harvard Univ. Press, 1969.

Sullivan.  Unusual.  1943 & 1947.

                                    Orville A. Sullivan.  Problems involving unusual situations.  SM 9 (1943) 114‑118 & 13 (1947) 102‑104.  (Previously listed in Section 2 below.)

Suter.  1900-1902.         Heinrich Suter.  Die Mathematiker und Astronomen der Araber und Ihre Werke.  (AGM 10 (1900) & 14 (1902)); reprinted by APA -- Academic Publishers Associated, Amsterdam, 1981.  See also:  H. P. J. Renaud; Additions et corrections à Suter "Die Mathematiker und Astronomen der Araber"; Isis 18 (1932) 166-183.

S&B.  1986.  Jerry (= G. K.) Slocum & Jack Botermans.  Puzzles Old & New -- How to Make and Solve Them.  Univ. of Washington Press, Seattle, 1986.  Slocum had produced a detailed index for this and has extended it to a joint index with New Book of Puzzles.

 

Tabari.  Miftāh al-mu‘āmalāt.  c1075.

                                    Mohammed ibn Ayyūb Ţabarī [NOTE:  Ţ  denotes a  T  with a underdot.].  Miftāh al-mu‘āmalāt.  c1075.  Ed. by Mohammed [the  h  should have an underdot] Amin Riyāhi [the  h  should have an underdot], Teheran, 1970.  ??NYS -- frequently cited and sometimes quoted by Tropfke and others.

Tagliente.  Libro de Abaco.  (1515).  1541.

                                    Girolamo [& Giannantonio] Tagliente.  Libro de abaco che insegnia a fare ogni raxone marcadantile & apertegare le terre con larte di la giometria & altre nobilissime raxone straordinarie cō la tarifa come raspondeno li pexi & monete de molte terre del mondo con la inclita citta de. Venetia.  El qual Libro se chiama Texauro universale ...; Venice, 1515.  See Rara 114-116, 495, 511‑512, which seems to confuse this with another work whose title starts:  Opera che insegna ....  Riccardi lists  27  editions of this and three editions of the other work.  He says Boncompagni has studied this work and found that Giovanni Rocha made some corrections and that several editions have only his name, so it is sometimes catalogued under Rocha -- cf below.  Smith mentions  25  editions under Tagliente and says Riccardi mentions  11  others.  Van Egmond's Catalog 334-344 lists 31 editions to 1586.  It is clear that this was a major book of its time.  I have briefly looked at a few examples and they seem to have the same material, though the woodcuts were often changed.

                    I have examined:  Giovanne Rocha;  Libro Dabaco che insegna a fare ogni ragione mercadãtile: & a ptegare le terre cõ larte di la geometria: & altre nobilissime ragione straordinarie cõ la Tariffa come respõdeno li pesi & mone de molte terre del mõdo con la inclita di Vinegia.  El qual Libro se chiama Thesauro universale.  Venturino Rossinello, Venice, 1541.  Smith, Rara, p. 529, only records a 1550 ed. printed by Giovanni Padovano in Venice.  The Crawford Collection has 1544 & 1550.  Not under Rocha in Riccardi.  I found this in the Turner Collection at Keele as A4.32, but it is not under Rocha in Hill's Catalogue (listed in 3.B), but is under Tagliente.  It has nice woodcut(?) illustrations in the text -- see Rara 512 for an example.  Having first found this book on the shelf at the Turner Collection, I originally thought it was by Rocha and hence an extraordinarily rare book.  I am grateful to Bill Kalush for identifying this as a version of Tagliente and for pointing out its importance.  Cf Van Egmond's Catalogue 338, item 15.

Tartaglia, Nicolo (or Niccolò) (c1506-1559).  See:  General Trattato.

Thomas, Ivor B.  See:  SIHGM.

Tissandier.  Récréations Scientifiques.  1880.

                                    Gaston Tissandier.  Les Récréations Scientifiques ou L'Enseignement par les Jeux.  G. Masson, Paris, (1880);  2nd ed., 1881;  3rd ed., 1883;  (4th ed., 1884);  5th ed., 1888;  (6th ed., 1893;  7th ed., 1894.)  I have seen 3rd ed., 1883, and I have 2nd ed., 1881, & 5th ed., 1888.  Tissandier was editor of La Nature and the articles often have fine illustrations by L. Poyet and others, frequently copied elsewhere.  See Tom Tit for more of Poyet's work.  I have seen the date of the first two editions as 1881 & 1882, but the Avertissement of the 2nd ed. says the 1st ed. was Nov 1880 and the Avertissement is dated Apr 1881.  [C&B only list 1881.]

Tissandier.  Popular Scientific Recreations.  1881.

                                    Translation and enlargement of the above.  Ward Lock, London, ([1881, 1882]); New enlarged ed., nd [1890, 1891].  [Not clear which French edition the translation is based on.  The new enlarged ed. contains a Supplement on pp. 775‑876 which includes material which is in the 5th French ed. of 1888, but not in the 3rd French ed. of 1883, so it seems the the main text is c1885 and the supplement is c1890.  C&B list a London edition, nd, 780pp.]  The index refers to a puzzle of knots and cords on p. 775 which is not present.  Most of the Supplement appeared as a series of Scientific Amusements in (Beeton's) Boy's Own Magazine from 1889 -- I have vol. 3 (1889) which has the first 10 articles, comprising 62 pages and 67 problems. 

                    Some of the material appeared in: Marvels of Invention and Scientific Puzzles.  Being  A Popular Account of Many Useful and Interesting Inventions and Discoveries.  Ward, Lock, & Co., nd [c1890].  My copy has no authors listed, but Jerry Slocum has a copy with Tissandier and Firth on the TP, though it is difficult to see what Firth could have done to warrant his inclusion.  This consists of Chapters 56-60, pp. 726-774, and Chap. 32, pp. 448-465, of Popular Scientific Recreations, set on smaller pages, plus a few extra items:  An economical mouse-trap (pp. 57-58);  Flying bridges (pp. 64-66);  Performing fleas (pp. 77-79);  Knots and cords  and  A Curious Toy (pp. 83-85).  This last must be what was on pp. 775-776 of the earlier edition of Popular Scientific Recreations and the other material seems to have come from some edition of that work or from the articles in (Beeton's) Boy's Own Magazine.  I won't bother to cite this version.

Tom Tit.  1890-1893?

                                    Arthur Good [= "Tom Tit"].  La Science Amusante.  3 vols., Larousse, Paris, 1890, 1892, 1893?  [I have seen the dates given as 1889, 1891, 1893, but the Introductions are dated as I have given, the last being Dec 1893.]  The material originally appeared in the magazine L'Illustration with classic engravings by Poyet which have been often reproduced, e.g. in:  Beeton's Boy's Own Magazine;  The Boy's Own Paper;  Kolumbus‑Eier (1890, 1976) -- translated as: Columbus' Egg (1978).  Arthur Good's name is clearly English and I have wondered if the articles were written originally in French or if they were translated from his English.

                    I have the three volume set and five one volume selections/adaptations.  I will give references to these by the initials shown below.

C.                 François Caradec, ed.  La Science Amusante -- 100 Experiences de Physique.  Les Editions 1900 [no accent], Paris, 1989.  [This consists of all of Vol. 1, reordered, but otherwise little changed.  Caradec's Preface gives a little about the author.  The illustrations are a bit dark.]

H.                 Magic at Home.  A Book of Amusing Science.  Annotated translation of La Science Amusante, vol. 1, by Prof. Hoffmann.  Cassell, 1891.  Cf. VBM below.

VBM.           The Victorian Book of Magic Illustrated or Professor Hoffman's [sic] Curious & Innocent Diversions for Parlour & Refined Gatherings.  Selected [from the above] & with a note to readers by C. Raymond Reynolds.  (Stephen Greene Press, Japan, 1969); Hugh Evelyn, London, nd [c1970].  [This has 26 of the items in Vol. 1.  Illustrations are small but good.]

K.                 Tom Tit.  Scientific Amusements.  Selected and translated by Cargill G. Knott.   Nelson, nd [1918].  [This contains 178 items, mostly from Vols. 2 & 3, but Knott has added a few others, possibly taken from Tom Tit's later articles?  Knott has also extended some items.  Sadly, the illustrations were poorly redrawn for this edition.]

R&A.           David Roberts & Cliff Andrew, eds.  100 Amazing Magic Tricks.  Cape, London, 1977.  [Selections from all three volumes.  Although this refers to the original books and L'Illustration, it avoids mentioning the original author's name!  Illustrations are very good.]

[I have also seen an Italian translation.  There was a US ed., trans. by Camden Curwen & Robert Waters; Magical Experiments or Science in Play; (© Worthington, 1892); David McKay, Philadelphia, 1894, 329 pp. [Christopher 398.]]

Todhunter.  Algebra, 5th ed.  1870.

                                    Isaac Todhunter (1820-1884).  Algebra  For the Use of Colleges and Schools.  With Numerous Examples.  Macmillan, (1858; 5th ed, 1870); new edition, 1879, HB.  [1st ed was 1858; 2nd, 1861; 5th, 1870.  The Preface in my 1879 copy is dated 1870 and says the work has been carefully revised, with two chapters and 300 miscellaneous examples added, so it was quite different than previous editions and I will date citations as 1870.]

Tonstall.  De Arte Supputandi.  1522.

                                    Cuthbert Tonstall [often spelled Tunstall] (c1474-1559).  De Arte Supputandi Libri Quatuor.  ([With Quattuor] Richard Pynson, London, 1522 -- the first arithmetic printed in England, with TP engraved by Holbein.)  I have seen:  Robert Stephan, Paris, 1538.  Though the TP and pagination are different, Smith, Rara, gives no indication that the 1538 text is any different than the 1522, so I will cite this as 1522.  Most citations are to Book III, whose problems are numbered.  See Rara 132-136.

Toole Stott.  1976-1978.

                                    Raymond Toole Stott.  A Bibliography of English Conjuring  1581‑1876.  2 vols., published by the author, Derby, 1976, 1978; distributed by Harpur & Sons, Derby.  1414 entries.  References are to the item number.

TP                               Title Page.

Tropfke.  1980.             Johannes Tropfke, revised by Kurt Vogel, Karin Reich and Helmuth Gericke.  Geschichte der Elementarmathematik.  4th ed., Vol. 1: Arithmetik und Algebra.  De Gruyter, Berlin, 1980.  (Prof. Folkerts says (1994) that vol. 2 is being edited.)

                    [The 1st ed. was De Gruyter?, Leipzig, 1902, 2 vols.  2nd ed., De Gruyter, Berlin & Leipzig, 1921-1924, 7 vols.  3rd ed., De Gruyter, Berlin & Leipzig, 1930-1940, vols. 1-4 (the MSS of the remaining volumes were destroyed in 1945).]

Tunstall, Cuthbert -- see:  Tonstall, Cuthbert.

Turner.                         The Turner Collection, formerly at University of Keele.  Sadly this collection was secretly sold by the University in 1998 and has now been dispersed.  A useful catalogue was prepared:  Susan Hill; Catalogue of the Turner Collection of the History of Mathematics Held in the Library of the University of Keele; University Library, Keele, 1982.

 

UCL.                            University College London or its Library, which includes the Graves Collection, cf Graves.

Uncle George.  See:  Parlour Pastime.

Unger.  Arithmetische Unterhaltungen.  1832.

                                    Ephraim Salomon Unger.  Arithmetische Unterhaltungen, bestehend in einer systematisch geordneten Sammlung von mehr als 900 algebraischen Aufgaben, verbunden mit einer Anleitung, diese Aufgaben mittelst der einfachsten Regeln der Arithmetick zu lösen.  (Erfurt, 1832, 4 + 253 pp., ??NYS);  2nd ed., Keysersche Buchhandlung, Erfurt, 1838, 10 + 268pp.  MUS 166 only mentions the 1838 ed. but a copy of the 1st ed. was advertised by Sändig in Jun 1997.  I haven't seen any other reference to this work.  In general, this book believes in beating problems to death -- each type of problem is done several times.  E.g. there are 10 problems of the Chinese Remainder Theorem with two moduli.  Hence I will generally not describe all the problems.

 

van Etten.  See:  Etten, above.

Van Egmond's Catalog.

                                    Warren Van Egmond.  Practical Mathematics in the Italian Renaissance: A Catalog of Italian Abbacus Manuscripts and Printed Books to 1600.  Supp. to Annali dell'Istituto e Museo di Storia della Scienza (1980), fasc. 1.  = Istituto e Museo di Storia della Scienza, Monografia N. 4.  Florence, 1980.  I have consulted this for (almost?) all MSS cited in these Sources and have made a number of changes of dates and even authorship based on it.  Like any such catalog, there are some omissions and errors, but it is by far the most authoritative listing of the material and I have adopted his dates and attributions -- cf Benedetto da Firenze, P. M. Calandri, dell'Abbaco, pseudo-dell'Abbaco. 

Vinot.  1860.

                                    Joseph Vinot.  Récréations Mathématiques  Nouveau Recueil de Questions Curieuses et Utiles Extraites des Auteurs Anciens et Modernes.  Larousse & Boyer, Paris, 1860, (3rd ed., Larousse, Paris, 1893;  1898;  1902;  6th ed., nd [1911]).  I have found no difference between the 1st and 6th eds -- indeed I have found a simple typographical error repeated.

Vogel, Kurt (1888-1985).  See:  AR;  BR;  Chiu Chang Suan Ching;  Columbia Algorism;  Tropfke;  items in 7.K under al‑Khwârizmî;  items in 7.P.7 and 7.R.2 under Fibonacci.

Vyse.  Tutor's Guide.  1771?

                                    Charles Vyse (fl. 1770-1815).  The Tutor's Guide, being a Complete System of Arithmetic; with Various Branches in the Mathematics.  In Six Parts, ....  To which is added, An Appendix, Containing Different Forms of Acquittances, Bills of Exchange, &c. &c.  The whole being designed for the Use of Schools, ....  The Eighth Edition, corrected and improved, with Additions.  G. G. J. and J. Robinson, London, 1793, HB.  12 + 324 pp.  [1st ed. was 1770 or 1771;  2nd ed., 1772;  4th ed., 1779;  6th ed., 1785;  1790;  8th ed., 1793;  1799;  12th ed., 1804;  13th ed., 1807;  14th ed., 1810;  15th ed., 1815;  1817;  16th ed., 1821.]  Since books of this nature rarely had major changes, I will date this as 1771? until I see further editions.

                                    Charles Vyse (fl. 1770-1815).  The Tutor's Guide, being a Complete System of Arithmetic; with Various Branches in the Mathematics.  In Six Parts, ....  To which is added, An Appendix, Containing Different Forms of Acquittances, Bills of Exchange, &c. &c.  The whole being designed for the Use of Schools, ....  10th ed., ed. by J. Warburton.  S. Hamilton for G. G. and J. Robinson, London, 1799.  12 + 335 pp.  [Dedication removed; Preface by the Editor and references to the Key added.  Though the text is reset with one or two more lines per page, the text seems to be preserved, though the editor has added substantial footnotes in places.]

                                    The Key to the Tutor's Guide; or the Arithmetician's Repository: containing Solutions to the Questions, &c. in the Tutor's Guide, with references to the pages where they stand.  To which are added Some Useful Rules, &c.  Likewise An Appendix; showing the Combination of Quantities; The different Ways they may be varied; with the method of filling the magic squares, &c.  ....  Eighth edition; Carefully revised, corrected, and augmented.  G. & J. Robinson, London, 1802.  8 + 370 pp + 2pp publisher's ads.  [1st ed., 1773;  3rd ed., 1779;  4th ed., 1785;  7th ed., 1799;  8th ed., 1802;  9th ed., 1807;  11th ed., 1818.] 

                    Despite the claim in the title of the Key, the references to the answers are in the text of the 10th ed at the beginning of each set of exercises.  There is no mention of the Key in the 8th ed of the book.  I have the 8th and 9th eds. of the Key and have not seen any difference between them -- indeed I have found a common misprint -- and they give the answers to the 10th ed. of the book, on the pages cited in the book, so I assume they are essentially identical to the 7th ed. of the Key. 

V&T, 1952.                  Volkmann, Kurt  &  Tummers, Louis.  Bibliographie de la Prestidigitation  Tome I  Allemagne et Autriche.  Cercle Belge d'Illusionnisme, Bruxelles, 1952.  With a 2 page list of Bibliographies of Conjuring.

 

Walkingame.  Tutor's Assistant.

                                    Francis Walkingame.  The Tutor's Assistant; Being a Compendium of Arithmetic, and a Complete Question-Book.  ...  To which are added, A new and very short Method of extracting the Cube Root, and a General Table for the ready calculating the Interest ....  The Fifteenth Edition.  Printed for the Author, London (Great Russel-Street, Bloomsbury), 1777.  [First ed. was 1751.  There were about a hundred editions in all.  See Wallis's article on this book in MG 47 (No. 361) (1963) 199-208.]

                                    The Tutor's Assistant.  Being a Compendium of Arithmetic, and a Complete Question-Book.  ...  To which are added, A new and very short Method of extracting the Cube Root, and a General Table for the ready calculating the Interest ....  The Twentieth Edition.  Printed for the Author, London (Kensington), 1784.  Appears to be identical to the 15th ed., except for resetting and some small changes, both corrections and errors, so I won't cite this separately.

                                    The Tutor's Assistant, being a Compendium of Arithmetic, and Complete Question-Book; ...; to which is added, A Compendium of Book-keeping, by Single Entry, by Isaac Fisher.  Thomas Richardson, Derby; Simpkin, Marshall & Co., London, 1835.  [The first Derby edition by Fisher was 1826.] 

                                    The Tutor's Companion; or, Complete Practical Arithmetic.  To which is added A Complete Course of Mental Arithmetic, ..., by Isaac Butler.  Webb, Millington, and Co., London, 1860.

                    The editions are pretty similar, but the interesting collection of problems at the end is much shortened in the 1835 ed. and a few are omitted in the 1860 ed.  Consequently I will assume the problems date from 1751, unless they vary in some important way.  I also have a Key to Walkingame from 1840, but it does not correlate with any of the editions I have!

Wallis.                          The Wallis Collection of early English mathematics books.  Gathered by Peter J. Wallis and left to the University of Newcastle upon Tyne in 1992.  A typical catalogue entry is  Wallis 227 CAR  and there are special sections for Newtoniana and Record(e).

Week‑End.  1932.         Hubert Phillips.  The Week‑End Problems Book.  Nonesuch Press, London, 1932.

Wehman.  New Book of 200 Puzzles.  1908.

                                    Wehman Bros.'  New Book of  200 Puzzles.  Wehman Bros., 126 Park Row, NY, 1908.  Largely copied from various 19C works: Boy's Own Book, Magician's Own Book, The Sociable, sometimes with typographical omissions.  I only count 130 puzzles!  There seems to have been a Johnson Smith reprint at some time.

Wells.  1698. Edward Wells.  Elementa Arithmeticæ Numerosæ et Speciosæ.  In Usum Juventutis Academicæ.  At the Sheldonian Theatre (i.e. OUP), Oxford, 1698.  (My copy was previously in the Turner Collection, Turner D1.1.)  All the material cited is in Appendix Posterior: Viz. Problemata sive Quæstiones ad exercendas Regulas Arithmeticæ.

Western Puzzle Works.  1926.

                                    Western Puzzle Works, 979 Marshall Avenue, St. Paul, Minnesota.  1926 Puzzle Catalogue.  Photocopy provided by Slocum.  Unpaginated, 8pp.

Williams.  Home Entertainments.  1914.

                                    Archibald & F. M. Williams.  Home Entertainments.  The Hobby Books, ed. by Archibald Williams.  Nelson, London, nd [1914 -- BMC].

Williams, Henry Llewellyn, Jr. (1842-??).  Books by Frikell?, particularly Magician's Own Book, are often attributed to him.  [C&B, under  Williams, Henry Llewellyn ("W. Frikell")  lists:  Hanky Panky;  Magician's Own Book, London & New York;  (Magic No Mystery);  The Secret Out  and says to also see Cremer.]

Williams, J. L.  See:  Boy's Own Book, 1843 (Paris) edition, which lists him as author.

Wilson, Robin J.  See:  BLW.

Wingate/Kersey.  1678?.

                                    Edmund Wingate (1596-1656).  Mr. Wingate's Arithmetick; Containing A Plain and Familiar Method For Attaining the Knowledge and Practice of Common Arithmetick.  (1629.)  The Seventh(?) Edition, very much Enlarged.  First Composed by Edmund Wingate, late of Gray's-Inn, Esquire.  Afterwards, upon Mr. Wingate's Request, Enlarged in his Life-time: Also since his Decease carefully Revised, and much Improved; ... By John Kersey, late Teacher of the Mathematicks.  My copy is lacking the TP and pp. 345+, but it appears to be identical to The Tenth Edition; J. Philips, J. Taylor & J. Knapton, London, 1699 -- though reset, it has the same pagination throughout except for the Dedication.  A Librarian's note suggests my earlier version is the 7th ed. of 1678.  This would be the fourth and last of Kersey's versions.  Kersey began editing from the second or third (1658) ed. and did four versions, the last in 1678. 

                    The material of interest is almost all in Chapter 10 of Kersey's Appendix: A Collection of subtil Questions to exercise all the parts of Vulgar Arithmetick; to which also are added various practical Questions about the mensuration of Superficial Figures and Solids, with the Gauging of Vessels, pp. 475-527, 75 questions.  There are a few further items in Chapter 11: Sports and Pastimes, pp. 528-544, 7 problems.  Chapter 11 is not clearly marked as being by Kersey in these copies, but is so marked in later editions and it is pretty clear that the entire Appendix is due to Kersey.  At the end, he says he has taken the problems of Chap. 11 from Bachet's Problemes.

                    I have seen a 14th ed. of 1720 which has the same text, reset and repaginated, with some supplementary material by George Shelley.  I have also seen a 19th ed. of 1760 which has been considerably reorganized by James Dodson.  The two chapters of puzzle problems have become Chapters XLIII and XLIV and the material has been changed, generally omitting some problems of interest and only adding two.

Winning Ways.  1982.   

                                    Elwyn R. Berlekamp, John H. Conway & Richard K. Guy.  Winning Ways for Your Mathematical Plays: Vol. 1: Games in General; Vol. 2: Games in Particular.  2 vols, Academic Press, NY, 1982.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.

                                    Simon Witgeest.  Het Natuurlyk Tover-Boek, Of't Nieuw Speel-Toneel Der Konsten.  Verhandelende over de agt hondert natuurlijke Tover-Konsten. so uyt de de Gogel-tas, als Kaartspelen, Mathematische Konsten, en meer andered diergelijke aerdigheden, die tot vermaek, en tijtkorting verstrecken.  Mitsgaders een Tractaet van alderley Waterverwen, en verligteryen;  Als oock Een verhandelinge van veelderley Blanketsels.  Om verscheyde wel-ruykende Wateren, Poederen en Balsemen, als ook kostelijke beeydselen, om het Aensicht, Hals en Handen, wit en sagt te maecken, door Simon Witgeest.  Jan ten Hoorn, Amsterdam, 1686.  ??NYS -- some photocopies sent by Jerry Slocum.  'Boek' is 'Boeck' on the frontispiece and running heads.  This is a much expanded and retitled 3rd edition of Witgeest's 1679 work.  The new material is stated to already be in the 2nd ed. of 1682. 

                    [There were many later editions:  1695;  1698;  Ten Hoorn, 1701; G. de Groot Keur, Amsterdam, 1725 (10th ed.), 1739;  1749;  1760;  Amsterdam, 1773;  1815  [Christopher 1098-1099, C&B, HPL].  It was translated as: Naturliches Zauber-Buch (or Zauber=Buch) oder neuer Spiel-Platz der Künste; Hoffmanns sel Wittw. & Engelbert Streck, Nürnberg, 1702.  There were later editions (all in Nürnberg?) of 1713 (in 1 or 2 vols); Hoffmann, Nürnberg, 1718; 1730; 1739; 1740; 1745; 1753; Johann Adam Stein, Nürnberg, 1755; 1760-1762; 1763; 1766; 1781; 1786; 1798 and a Lindau reprint of 1978  [Christopher 1092-1097, C&B, V&T], all apparently based on the 1682 Dutch ed.]

Witgeest.  Het Nieuw Toneel der Konsten.  1679.

                                    Simon Witgeest.  Het Nieuw Toneel der Konsten, Bestaande uyt Sesderley Stukken: het eerste, handelt van alderley aardige Speeltjes en Klugjes: het tweede, van de Verligt-konst in 't Verwen en Schilderen: het derde, van het Etzen en Plaat-shijden: het vierde, van de Glas-konst: het vijfde, heest eenige aardige remedien tegen alderley Ziekten: het sesde, is van de Vuur-werken.  Uyt verscheyde Autheuren by een vergadert, door S. Witgeest, Middel-borger.  Jan ten Hoorn, Amsterdam, 1679;  facsimile with epilogue by John Landwehr, A. W. Sijthoff's Uitgeversmaatschappij N. V., Leiden, 1967 (present from Bill Kalush). 

                    There were many later editions, but Nanco Bordewijk has examined these and discovered that the 3rd ed. of 1686 (I can't recall if he saw the 1682 ed.) was so extensively revised and extended as to constitute a new book, and it has the different title given in the previous entry.  (Other sources indicate these revisions are already in the 2nd ed. of 1682.)  Landwehr has written a bibliographical article on this book -- ??NYR.

Wood.  Oddities.           Clement Wood.  A Book of Mathematical Oddities.  Little Blue Book 1210.  Haldeman-Julius, Girard, Kansas, nd [1927].

 

Young Man's Book.  1839.

                                    Anonymous.  The Young Man's Book of Amusement.  Containing the Most Interesting and Instructive Experiments in Various Branches of Science.  To Which is Added All the Popular Tricks and Changes in Cards; and the Art of Making Fire Works.  William Milner, Halifax, 1844, HB.  2 + 384 pp + folding plate (originally a frontispiece).  My copy has a number of annotations as though in preparation for another edition.  [Hall BCB 322.  Toole Stott 751.  BCB 320-323 are  1839, 1840, 1844, 1848.  Heyl 358-360 are  1846, 1846 (Milner & Sowerby ??), 1850.  Toole Stott 749-752, 1216 are  1839, 1840, 1844, 1846, 1850.  Christopher 1111-1113 are 1839, 1846, 1859 (Milner & Sowerby).  All these are apparently the same except for the publisher's name change.]

Young World.  c1960.

                                    Young World Productions.  Tricks and Teasers.  303  Gags  Games  Tongue Twisters  Problems  Tricks.  Young World Productions, London, nd [inscribed 1965 on first page, so probably c1960; BLC-Ø].

 

536.                              H. E. Dudeney.  536 Puzzles and Curious Problems.  Ed. by M. Gardner.  Scribner's, NY, 1967.  (This consists of almost all the puzzles from Modern Puzzles (MP) and Puzzles and Curious Problems (PCP).)  [There is also a Fontana, London, 1970, edition in two volumes: Puzzles and Curious Problems (258 problems); More Puzzles and Curious Problems (261 problems).]

 

??                                 indicates uncertainty and points where further work needs to be done. 

                                   BC, e.g.  ‑330  is  330 BC  and  ‑5C  is  5th century BC.

¹                                  Inequality or incongruence (mod m).  (My word processor does not have an incongruence sign.  I may change this in Word using an Arial character.)

Ø                                 Nothing, used after catalogues, etc., to indicate that I have looked in that catalogue and found no entry.  E.g.  BLC-Ø.

 

 

          SOME OTHER RECURRING REFERENCES

 

          Details of these items are given under the first reference in Sources.  Later references often cite the first reference.  I tend to make entries below when I use the item, but sometimes I have entered the entry long after the first usage of the item, and I haven't searched the text for other (perhaps entered long ago) appearances of these items.

 

    For:                                            See Sections:

 

Anon: Home Book ..., 1941                4.B.1, 4.B.3, 5.U, 6.AO, 6.BF.4, 7.B, 7.AT, 9.E.1

Anon: Treatise, 1850                          7.H, 7.P.6, 7.S, 7.X, 10.A, 10.R

 

Allen, 1991                                       5.B, 5.I.1, 5.N, 6.L, 7.E, 7.I, 7.P.1, 7.R.3, 7.AC.6, 7.AH, 7.AL, 9.B, 10.A.4

Always: More Puzzles to Puzzle You, 1967

                                                        6.BF.4,

Always: Puzzles for Puzzlers, 1971

                                                        5.D.2, 5.D.5, 7.G.1, 7.AC.1, 9.J, 10.I, 10.K

Always: Puzzles to Puzzle You, 1965

                                                        5.K.2, 5.W.1, 5.X.2, 7.AC.3, 7.AC.6, 7.AS,

Always: Puzzling You Again, 1969

                                                        5.C, 6.BD, 7.AH,

Ananias of Shirak, c640                     7.E, 7.H, 10.A

André, 1876                                      7.H, 7.H.1, 7.S.1, 7.S.2, 7.AF, 7.AF.2

August, 1939                                     5.X.1, 6.BE, 7.I, 7.X, 7.Z, 7.AL, 7.AN, 7.AT, 7.AV, 10.H

Badcock, 1823                                  6.BH, 7.H.3, 7.P.5, 7.Q

Bagley: Paradox Pie, 1944                  6.BN, 7.Z, 7.AI, 7.AW, 10.F, 10.Q, 10.S

Bagley: Puzzle Pie, 1944                    5.D.5, 6.O, 6.P.1, 6.P.2, 6.R.1, 6.R.2, 6.Y, 6.AF, 6.AI, 7.AV, 10.L

Bath, 1959                                        5.C, 7.G.1, 7.I, 7.P.5, 7.AC.3, 7.AC.6, 7.AM

Bellew, 1866                                     5.E, 6.AO.1, 6.AQ

Berloquin, 1981                                 5.N, 7.H.5, 7.N.3, 10.R

Black, 1952 [1946?]                          5.T, 6.F.2, 9.D, 9.F

Bourdon, 1834                                  7.E.1, 7.H, 7.P.1, 7.S, 7.X, 7.AF.1, 7.AK, 10.A, 10.R

Brandreth Puzzle Book [1895]            5.B, 5.B.1, 5.O, 6.AW.1, 6.AY.1, 7.B, 7.G.1

Bullen, 1789                                      7.G.1, 7.H, 7.H.5, 7.L.2.b, 7.S.1, 7.AF.1

Bullivant, 1910                                  5.S, 6.T, 6.AK

[Chambers], 1866?                            7.H, 7.L.2.a, 7.L.2.b, 7.Y, 7.AF.2

Chang Chhiu‑Chien -- see Zhang Qiujian

Colenso, Algebra, 1849                      7.P.1, 10.G

Colenso, Arithmetic, 1853                  7.H, 7.X, 10.G, 10.R

Devi, 1976                                        5.D.1, 5.X.1, 7.E, 7.P.1, 7.AC.3, 7.AC.6, 7.AE, 10.A.3; 10.K

Dresner, 1962                                   5.B.1, 5.C, 5.D.4, 5.K.1, 5.W

Dudeney: World's best puzzles, 1908

                                                        2, 5.P.1, 5.S, 6.P.1, 6.S, 6.AI, 6.AO, 6.AW

Elliott, 1872                                      6.V, 6.AQ, 6.AV, 6.AZ, 11.B, 11.C, 11.D

Filipiak, 1942                                    5.H.1, 6.W.1, 6.W.2, 6.AK

Fisher, 1968                                      6.E, 6.P.2, 7.M.4, 7.AI

Fisher, 1973                                      1, 5.E, 7.S.2, 10.L

Fourrey, Cur. Geom., 1907                6.S.1, 8.G  (also 6.R.1)

Fourrey, Rec. Arith., 1899                  4.A.1, 5.B, 5.P.1, 5.U, 7.N.1

Fuss, 1843                                        5.F.1

Goldston, nd [1910?]                         6.AK, 11.E

Gomme, 1894/98                               4.B.1, 5.R.5  (Also 4.A.3)

Greenblatt, 1965                                6.U.2, 6.AE, 7.AG

Heald, 1941                                      7.Z, 10.E.3, 10.G, 10.0

Hooper, 1774                                    4.A.1, 5.AA, 6.F, 6.P.2, 7.B, 7.AO, 7.AZ

Hutton, 1804                                     7.G.2, 7.H, 7.X, 7.AF.1, 7.AK, 10.A, 10.R

Kamp, 1877                                      5.B, 5.D.1, 5.E, 5.R.7, 7.B, 7.L, 7.Q

Kraitchik, MJ, 1930                           4.A.2, 5.J, 7.E, 7.G.1, 7.H.3, 7.AR, 10.B, 10.P

Kraitchik, MR, 1943                          4.A.2, 5.J, 6.M, 7.H.2, 7.H.3, 9.D, 9.G, 10.P

Laisant, 1906                                    6.P.1, 7.AR, 10.A.2, 10.B, 10.H, 10.I

Larte de labbacho, 1478                     See: Treviso Arithmetic.

Von der Lasa, 1897                           5.F.1, 7.B  (Also 4.B.1, 4.B.5)

Licks, 1917                                       5.A, 6.R.4, 6.AG, 7.P.3, 7.S.2, 7.AC, 7.AD

Van der Linde, 1874                          5.F.1, 7.N  (Also 4.B.1, 4.B.5)

Littlewood, 1953                               5.C, 5.W, 6.J, 8.B, 9.C, 9.D

Lucas, Théorie, 1891                         5.L, 5.Z.5, 5.AB

Madachy, 1966                                 5.O, 6.D, 6.X, 7.N.3, 7.N.4, 7.AC.3, 7.BB

Meyer, 1965                                     7.I, 7.AC.4, 7.AH, 7.AR, 7.AX

Milne, 1881                                       7.E, 7.H, 7.R, 7.X, 7.Y, 10.A, 10.A.3, 10.G, 10.R

W. O. J. Moser, 1981                        6.I, 6.T

Nordmann, 1927                               4.A.4, 5.G.1, 6.AR, 7.AC.3, 7.AR, 11.C, 11.E

Papyrus Rhind, c‑1650                       7.C, 7.G.1, 7.L, 7.S.1

Phillips: Playtime Omnibus, 1933        6.AF, 7.S.2, 7.AC.1, 7.AD.1, 7.AE, 9.D

Phillips: Question Time, 1937             5.U, 7.E, 7.AG, 9.G

Ransom, 1955                                   6.M, 7.F, 7.X, 7.AC.2, 8.B, 10.A.1, 10.B, 10.I

Smith: Origin, 1917                            3A, 7.G.2, 7.H, 10.A

Steinhaus: 1938,1950,1960,1969         5.C.1, 6.E, 6.G.1, 6.H, 6.AB

Strutt, 1791?, 1801                            4.B.1, 5.R.1, 5.R.5  (Also 4.A.3)

Strutt-Cox, 1903                                4.B.1, 5.R.1, 5.R.5  (Also 4.A.3)

Trenchant, 1566                                7.L.2.a, 7.S.1 (also 5.B, 5.D.1, 7.E, 7.S.1, 7.AF.1)

Treviso Arithmetic, 1478                    7.H, 7.K.1, 7.AL, 10.A

Trigg: Quickies, 1967                         5.Q, 6.AE, 6.AN, 7.N.3, 7.W

Wagner, Rechenbuch, 1483                7.G.1, 7.G.2, 7.H, 7.AK, 10.A

Wecker, (1582), 1660                        7.L.3, 7.AO, 10.P, 11.I, 11.N

A. C. White, 1913                             1, 5.I.1, 6.T, 6.AK, 7.X, 11.E

Widman(n), 1489                              7.G.1, 7.H, 7.L.2, 7.P.1, 7.P.5, (7.AL)

Williams & Savage, 1940                   7.P.5, 7.X, 7.AC.2, 7.AM, 7.AP, 8.I, 10.E.2

Wolff, 1937                                      7.R.3, 7.S.2, 7.AC.1, 7.AE, 9.E, 9.E.1, 10.O

Workman, 1902                                7.H.1, 7.H.4, 7.J, 7.S.2, 7.AC.2, 10.G

Wyatt, 1928                                      5.H.1, 6.V, 6.W.1, 6.W.2, 6.AI

Mr. X, 1903, 1911                             4.A.1, 5.B, 5.P.1, 5.S, 6.AF, 6.AU, 7.H.3, 7.I, 7.J, 7.M.4, 9.E, 9.J, 10.H

Yang Hui, 1275                                 7.N, 7.P.1, 7.P.2, 10.A

Zhang Qiujian, 468                             7.E, 7.L, 7.P.1, 7.P.6, 10.A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.       BIOGRAPHICAL MATERIAL  --  in chronological order

 

 

          ALCUIN  (c735‑804)

 

Phillip Drennon Thomas.  Alcuin of York.  DSB I, 104‑105.

Robert Adamson.  Alcuin, or Albinus.  DNB, (I, 239‑240), 20.

Andrew Fleming West.  Alcuin and the Rise of the Christian Schools.  (The Great Educators -- III.)  Heinemann, 1893.  The only book on Alcuin that I found which deals with the Propositiones.

Stephen Allott.  Alcuin of York  c. A.D. 732 to 804  --  his life and letters.  William Sessions, York, 1974.

 

          FIBONACCI [LEONARDO PISANO]  (c1170->1240)

 

          See also the entries for Fibonacci in Common References.

 

Fibonacci.  (1202 -- first paragraph);  1228 -- second paragraph, on p. 1.  In this paragraph he narrates almost everything we know about him.  [In the second ed., he inserted a dedication as the first paragraph.]

                    The paragraph ends with the notable sentence which I have used as a motto for this work.  "Si quid forte minus aut plus iusto vel necessario intermisi, mihi deprecor indulgeatur, cum nemo sit qui vitio careat et in omnibus undique sit circumspectus."  (If I have perchance omitted anything more or less proper or necessary, I beg indulgence, since there is no one who is blameless and utterly provident in all things.  [Grimm's translation.])

Richard E. Grimm.  The autobiography of Leonardo Pisano.  Fibonacci Quarterly 11:1 (Feb 1973) 99-104.  He has collated six MSS of the autobiographical paragraph and presents his critical version of it, with English translation and notes.  Sigler, below, gives another translation.  I give Grimm's translation, omitting his notes.

 

                    After my father's appointment by his homeland as state official in the customs house of Bugia for the Pisan merchants who thronged to it, he took charge; and, in view of its future usefulness and convenience, had me in my boyhood come to him and there wanted me to devote myself to and be instructed in the study of calculation for some days.  There, following my introduction, as a consequence of marvelous instruction in the art, to the nine digits of the Hindus, the knowledge of the art very much appealed to me before all others, and for it I realized that all its aspects were studied in Egypt, Syria, Greece, Sicily, and Provence, with their varying methods; and at these places thereafter, while on business, I pursued my study in depth and learned the give-and-take of disputation.  But all this even, and the algorism, as well as the art of Pythagoras I considered as almost a mistake in respect to the method of the Hindus.  Therefore, embracing more stringently that method of the Hindus, and taking stricter pains in its study, while adding certain things from my own understanding and inserting also certain things from the niceties of Euclid's geometric art, I have striven to compose this book in its entirety as understandably as I could, dividing it into fifteen chapters.  Almost everything which I have introduced I have displayed with exact proof, in order that those further seeking this knowledge, with its pre-eminent method, might be instructed, and further, in order that the Latin people might not be discovered to be without it, as they have been up to now.  If I have perchance omitted anything more or less proper or necessary, I beg indulgence, since there is no one who is blameless and utterly provident in all things.

 

F. Bonaini.  Memoria unica sincrona di Leonardo Fibonacci novamente scoperta.  Giornale Storico degli Archivi Toscani 1:4 (Oct-Dec 1857) 239-246.  This reports the discovery of a 1241 memorial of the Comune of Pisa, which I reproduce as it is not well known.  This grants Leonardo an annual honorarium of 20 pounds.  In 1867, a plaque bearing this inscription and an appropriate heading was placed in the atrium of the Archivio di Stato in Pisa.

 

                    "Considerantes nostre civitatis et civium honorem atque profectum, qui eis, tam per doctrinam quam per sedula obsequia discreti et sapientis viri magistri Leonardi Bigolli, in abbacandis estimationibus et rationibus civitatis eiusque officialium et aliis quoties expedit, conferunter; ut eidem Leonardo, merito dilectionis et gratie, atque scientie sue prerogativa, in recompensationem laboris sui quem substinet in audiendis et consolidandis estimationibus et rationibus supradictis, a Comuni et camerariis publicis, de Comuni et pro Comuni, mercede sive salario suo, annis singulis, libre xx denariorum et amisceria consueta dari debeant (ipseque pisano Comuni et eius officialibus in abbacatione de cetero more solito serviat), presenti constitutione firmamus."

 

                    A translation follows, but it can probably be improved.  My thanks to Steph Maury Gannon for many improvements over my initial version.

 

                    Considering the honour and progress of our city and its citizens that is brought to them through both the knowledge and the diligent application of the discreet and wise Maestro Leonardo Bigallo in the art of calculation for valuations and accounts for the city and its officials and others, as often as necessary; we declare by this present decree that there shall be given to the same Leonardo, from the Comune and on behalf of the Comune, by reason of affection and gratitude, and for his excellence in science, in recompense for the labour which he has done in auditing and consolidating the above mentioned valuations and accounts for the Comune and the public bodies, as his wages or salary, 20 pounds in money each year and his usual fees (the same Pisano shall continue to render his usual services to the Comune and its officials in the art of calculation etc.).

 

                    Bonaini also quotes a 1506 reference to Lionardo Fibonacci.

 

Mario Lazzarini.  Leonardo Fibonacci  Le sue Opere e la sua Famiglia.  Bolletino di Bibliografia e Storia delle Scienze Matematiche  6 (1903) 98‑102  &  7 (1904) 1-7.  Traces the family to late 11C, saying Leonardo's father was Guglielmo and his grandfather was probably Bonaccio.  He estimates the birth date as c1170.  He describes a contract of 28 Aug 1226 in which Leonardo Bigollo, his father, Guglielmo, and his brother, Bonaccingo, buy a piece of land from a relative.  This land included a tower and other buildings, outside the city, near S. Pietro in Vincoli.  [G. Milanesi; Documento inedito intorno a Leonardo Fibonacci; Rome, 1867 -- ??NYS].  Says nothing is known of the 1202 ed of Liber Abbaci.  Quotes the above memorial.

R. B. McClenon.  Leonardo of Pisa and his liber quadratorum.  AMM 26:1 (Jan 1919) 1-8.

Gino Loria.  Leonardo Fibonacci.  Gli Scienziati Italiana dall'inizio del medio evo ai nostri giorni.  Ed. by Aldo Mieli.  (Dott. Attilio Nardecchia Editore, Rome, 1921;)  Casa Editrice Leonardo da Vinci, Rome, 1923.  Vol. 1, pp. 4-12.  This reproduces much of the material in Lazzarini and the opening biographical paragraph of Liber Abaci. 

Ettore Bortolotti.  Article on Fibonacci in:  Enciclopedia Italiana.  G. Treccani, Rome, 1949 (reprint of 1932 ed.).

Charles King.  Leonardo Fibonacci.  Fibonacci Quarterly 1:4 (Dec 1963) 15-19.

Gino Arrighi, ed.  Leonardo Fibonacci: La Practica di Geometria -- Volgarizzata da Cristofano di Gherardo di Dino, cittadino pisano.  Dal Codice 2186 della Biblioteca Riccardiana di Firenze.  Domus Galilaeana, Pisa, 1966.  The Frontispiece is the mythical portrait of Fibonacci, taken from I Benefattori dell'Umanità, vol. VI; Ducci, Florence, 1850.  (Smith, History II 214 says it is a "Modern engraving.  The portrait is not based on authentic sources".)  P. 15 shows the plaque erected in the Archivio di Stato di Pisa in 1855 which reproduces the above memorial with an appropriate heading, but Arrighi has no discussion of it.  P. 19 is a photo of the statue in Pisa and p. 16 describes its commissioning in 1859.

Joseph and Francis Gies.  Leonard of Pisa and the New Mathematics of the Middle Ages.  Crowell, NY, 1969.  This is a book for school students and contains a number of dubious statements and several false statements.

Kurt Vogel.  Fibonacci, Leonardo, or Leonardo of Pisa.  DSB IV, 604-613.

A. F. Horadam.  Eight hundred years young.  Australian Mathematics Teacher 31 (1975) 123‑134.  Good survey of Fibonacci's life & work.  Gives English of a few problems.  This is available on Kimberling's website - see below.

Ettore Picutti.  Leonardo Pisano.  Le Scienze 164 (Apr 1982) ??NYS.  = Le Scienze, Quaderni; 1984, pp. 30-39.  (Le Scienze is a magazine;  the Quaderni are collections of articles into books.)  Mostly concerned with the Liber Quadratorum, but surveys Fibonacci's life and work.  Says he was born around 1170.  Includes photo of the plaque in the Archivo di Stato di Pisa.

Leonardo Pisano Fibonacci.  Liber quadratorum, 1225.  Translated and edited by L. E. Sigler as:  The Book of Squares; Academic Press, NY, 1987.  Introduction: A brief biography of Leonardo Pisano (Fibonacci) [1170 - post 1240], pp. xv-xx.  This is the best recent biography, summarizing Picutti's article.  Says he was born in 1170 and his father's name was Guilielmo -- cf Loria above.  Gives another translation of the biographical paragraph of the Liber Abbaci.

A. F. Horadam & J. Lahr.  Letter to the Editor.  Fibonacci Quarterly 28:1 (Feb 1990) 90.  The authors volunteer to act as coordinators for work on the life and work of Fibonacci.  Addresses:  A. F. Horadam, Mathematics etc., Univ. of New England, Armidale, New South Wales, 2351, Australia;  J. Lahr, 14 rue des Sept Arpents, L‑1139 Luxembourg, Luxembourg.

Thomas Koshy.  Fibonacci and Lucas Numbers with Applications.  Wiley-Interscience, Wiley, 2001.  Claims to be 'the first attempt to compile a definitive history and authoritative analysis' of the Fibonacci numbers, but the history is generally second-hand and marred with a substantial number of errors,  The mathematical work is extensive, covering many topics not organised before, and is better done, but there are more errors than one would like.

Laurence E. Sigler.  Translation of Liber Abaci as:  Fibonacci's Liber Abaci  A Translation into Modern English of Leonardo Pisano's Book of Calculation.  Springer, 2002.

Clark Kimberling's site web includes biographical material on Fibonacci and other similar number theorists.  http://cedar.evansville.edu/~ck6/bstud/fibo.html .

Ron Knott has a huge website on Fibonacci numbers and their applications, with material on many related topics, e.g. continued fractions, π, etc. with some history.  www.ee.surrey.ac.uk/personal/r.knott/fibonacci/fibnat.html .

 

          Luca PACIOLI (c1445-1517)

 

S. A. Jayawardene.  Luca Pacioli.  BDM 4, 1897-1900.

 

Bernardino Baldi (Catagallina) (1553-1617).  Vita di Pacioli.  (1589, first published in his Cronica de Mathematici of 1707.)  Reprinted in: Bollettino di bibliografia e di storia delle scienze matematiche e fisiche 12 (1879) 421-427.  ??NYS -- cited by Taylor, p. 338.

Enrico Narducci.  Intorno a due edizioni della "Summa de arithmetica" di Fra Luca Pacioli.  Rome, 1863.  ??NYS -- cited by Riccardi [Biblioteca Matematica Italiana, 1952]

D. Ivano Ricci.  Luca Pacioli, l'uomo e lo scienziato.  San Sepolcro, 1940.  ??NYS -- cited in BDM.

R. Emmett Taylor.  No Royal Road  Luca Pacioli and His Times.  Univ. of North Carolina Press, Chapel Hill, 1942.  BDM describes this as lively but unreliable.

Ettore Bortolotti.  La Storia della Matematica nella Università di Bologna.  Nicola Zanichelli Editore, Bologna, 1947.  Chap. I, § V, pp. 27-33: Luca Pacioli.

Margaret Daly Davis.  Piero della Francesca's Mathematical Treatises  The "Trattato d'abaco" and "Libellus de quinque corporibus regularibus".  Longo Editore, Ravenna, 1977.  This discusses Piero's reuse of his own material and Pacioli's reuse of Piero's material.

Fenella K. C. Rankin.  The Arithmetic and Algebra of Luca Pacioli.  PhD thesis, Univ. of London, 1992 (copy at the Warburg Institute), ??NYR.

Enrico Giusti, ed.  Descriptive booklet accompanying the 1994 facsimile of the Summa -- qv in Common References.

Edward A. Fennell.  Figures in Proportion: Art, Science and the Business Renaissance.  The contribution of Luca Pacioli to culture and commerce in the High Renaissance.  Catalogue for the exhibition, The Institute of Chartered Accountants in England and Wales, London, 1994.

 

          Claude-Gaspar BACHET de Méziriac  (1581‑1638)

 

C.‑G. Collet & J. Itard.  Un mathématicien humaniste -- Claude‑Gaspar Bachet de Méziriac (1581‑1638).  Revue d'Histoire des Sciences et leurs Applications 1 (1947) 26‑50. 

J. Itard.  Avant-propos.  IN:  Bachet; Problemes; 1959 reprint, pp. v‑viii.  Based on the previous article. 

There is a Frontispiece portrait in the reprint.

Underwood Dudley.  The first recreational mathematics book.  JRM 3 (1970) 164‑169.  On Bachet's Problemes.

William Schaaf.  Bachet de Méziriac, Claude‑Gaspar.  DSB I, 367‑368.

 

          Jean LEURECHON (c1591‑1670)  and  Henrik VAN ETTEN

 

A. Deblaye.  Étude sur la récréation mathématique du P. Jean Leurechon, Jésuite.  Mémoires de la Société Philotechnique de Pont-à-Mousson 1 (1874) 171-183.  [MUS #314.  Schaaf.  Hall, OCB, pp. 86, 88 & 114, says the only known copy of this journal is at Harvard, which has kindly supplied me with a photocopy of this article.  Hall indicates the article is in vol. II and says it is 12 pages, but only cites pp. 171 & 174.]  This simply assumes Leurechon is the author and gives a summary of his life.  The essential content is described by Hall.

G. Eneström.  Girard Desargues und D.A.L.G.  Biblioteca Mathematica (3) 14 (1914) 253‑258.  D.A.L.G. was an annotator of van Etten's book in c1630.  Although D.A.L.G. was used by Mydorge on one of his other books, it had been conjectured that this stood for  Des Argues Lyonnais Girard (or Géomètre).  Eneström can find no real evidence for this and feels that Mydorge is the most likely person.

Trevor H. Hall.  Mathematicall Recreations.  An Exercise in Seventeenth Century Bibliography.  Leeds Studies in Bibliography and Textual Criticism, No. 1.  The Bibliography Room, School of English, University of Leeds, 1969, 38pp.  Pp. 18‑38 discuss the question of authorship and Hall feels that van Etten probably was the author and that there is very little evidence for Leurechon being the author.  Much of the mathematical content is in Bachet's Problemes and may have been copied from it or some common source.  [This booklet is reproduced as pp. 83-119 of Hall, OCB, with the title page of the 1633 first English edition reproduced as plate 5, opp. p. 112.  Some changes have been made in the form of references since OCB is a big book, but the only other substantial change is that he spells the name of the dedicatee of the book as Verreyken rather than Verreycken.]

William Schaaf.  Leurechon, Jean.  DSB VIII, 271‑272.

Jacques Voignier.  Who was the author of "Recreation Mathematique" (1624)?  The Perennial Mystics #9 (1991) 5-48 (& 1-2 which are the cover and its reverse).  [This journal is edited and published by James Hagy, 2373 Arbeleda Lane, Northbrook, Illinois, 60062, USA.]  Presents some indirect evidence for Leurechon's authorship.

 

          Jacques OZANAM (1640‑1717)

 

On the flyleaf of J. E. Hofmann's copy of the 1696 edition of Ozanam's Recreations is a pencil portrait labelled Ozanam -- the only one I know of.  This copy is at the Institut für Geschichte der Naturwissenschaft in Munich.  Hofmann published the picture -- see below.

Charles Hutton.  A Mathematical and Philosophical Dictionary.  1795-1796.  Vol. II, pp. 184-185.  ??NYS  [Hall, OCB, p. 166.]

Charles Hutton.  On the life and writings of Ozanam, the first author of these Mathematical Recreations.  Ozanam-Hutton.  Vol. I.  1803: xiii-xv;  1814: ix-xi.

William L. Schaaf.  Jacques Ozanam on mathematics ....  MTr 50 (1957) 385-389.  Mostly based on Hutton.  Includes a sketchy bibliography of Ozanam's works, generally ignoring the Recreations.

Joseph Ehrenfried Hofmann.  Leibniz und Ozanams Problem, drei Zahlen so zu bestimmen, dass ihre Summe eine Quadratzahl und ihre Quadratsumme eine Biquadratzahl ergibt.  Studia Leibnitiana 1:2 (1969) 103-126.  Outlines Ozanam's life, gives a bibliography of his works and reproduces the above-mentioned drawing as a plate opp. p. 124.  (My thanks to Menso Folkerts for this information and a copy of Hofmann's article.)

William L. Schaaf.  Ozanam, Jacques.  DSB X, 263‑265.

 

          Jean Étienne MONTUCLA  (1725-1799)

 

Charles Hutton.  Some account of the life and writings of Montucla.  Ozanam‑Hutton.  Vol. I.  1803: viii-xii;  1814: iv-viii.

Charles Hutton.  A Philosophical and Mathematical Dictionary.  2nd ed. of the Dictionary cited under Ozanam, 1815,  Vol. II, pp. 63-64.  ??NYS.  According to Hall, OCB, p. 167, this is not in the 1795-1796 ed. and is a reworking of the previous item.

 

          Lewis CARROLL  (1832-1898)

 

          Pseudonym of Charles Lutwidge Dodgson.  There is so much written on Carroll that I will only give references to his specifically recreational work and some basic references.

 

The Diaries of Lewis Carroll.  Edited by Roger Lancelyn Green.  (OUP, 1954);  2 vols, Greenwood Publishers, Westport, Connecticut, 1971, HB.

Lewis Carroll's Diaries  The private journals of Charles Lutwidge Dodgson (Lewis Carroll)  The first complete version of the nine surviving volumes with notes and annotations by Edward Wakeling.  Introduction by Roger Lancelyn-Green.  The Lewis Carroll Society, Publications Unit, Luton, Bedfordshire.  [There were 13 journals, but 4 are lost.]

                    Vol. 1.  Journal 2, Jan-Sep 1855.  1993, 158pp.

                    Vol. 2.  Journal 4, Jan-Dec 1856.  1994, 158pp.

                    Vol. 3.  Journal 5, Jan 1857 - Apr 1858.  1995, 199pp.

                    Vol. 4.  Journal 8, May 1862 - Sep 1864 and a reconstruction of the four missing

                              years, 1858-1862.  1997, 399pp.

                    Vol. 5.  Journal 9, Sep 1864 - Jan 1868, including the Russian Journal. 

                              1999, 416pp.

                    Vol. 6.  Journal 10, Apr 1868 - Dec 1876.  2001, 552pp.

                    Vol. 7.  Journal 11, Jan 1877 - Jun 1883.  2003, 606pp.

The Letters of Lewis Carroll.  Edited by Morton N. Cohen  with the assistance of Roger Lancelyn Green.  Volume One  ca.1837 - 1885;  Volume Two  1886 - 1898.  Macmillan London, 1979.

 

Stuart Dodgson Collingwood.  The Life and Letters of Lewis Carroll.  T. Fisher Unwin, London, 1898.

Stuart Dodgson Collingwood, ed.  The Lewis Carroll Picture Book.  T. Fisher Unwin, London, 1899.   = Diversions and Digressions of Lewis Carroll, Dover, 1961.  = The Unknown Lewis Carroll, Dover, 1961(?).  Reprint, in reduced format, Collins, c1910.  The pagination of the main text is the same in the 1899 and in both Dover reprints, but is quite different than the Collins.  Cited as: Carroll-Collingwood, qv in Common References.

R. B. Braithwaite.  Lewis Carroll as logician.  MG 16 (No. 219) (Jul 1932) 174-178.  He notes that Carroll assumed that a universal statement implied the existence of an object satisfying the antecedent, e.g. 'all unicorns are blue' would imply the existence of unicorns, contrary to modern convention.

Derek Hudson.  Lewis Carroll -- An Illustrated Biography.  Constable, 1954;  new illustrated ed., 1976.

Warren Weaver.  Lewis Carroll: Mathematician.  SA 194:4 (Apr 1956) 116‑128.  +  Letters and response.  SA 194:6 (Jun 1956) 19-22.

Martin Gardner.  The Annotated Alice.  C. N. Potter, NY, 1960.  Penguin, 1965; 2nd ed., 1971.  Revised as: More Annotated Alice, 1990, qv.

Martin Gardner.  The Annotated Snark.  Bramhall House, 1962.  Penguin, 1967;  revised, 1973 & 1974.

John Fisher.  The Magic of Lewis Carroll.  Nelson, 1973.  Penguin, 1975.

Morton N. Cohen, ed.  The Selected Letters of Lewis Carroll.  Papermac (Macmillan), 1982.

Martin Gardner.  More Annotated Alice.  [Extension of The Annotated Alice.]  Random House, 1990. 

Edward Wakeling.  Lewis Carroll's Games and Puzzles.  Dover and the Lewis Carroll Birthplace Trust, 1992.  Cited as Carroll-Wakeling, qv in Common References.

Francine F. Abeles, ed.  The Pamphlets of Lewis Carroll -- Vol. 2: The Mathematical Pamphlets of Charles Lutwidge Dodgson and Related Pieces.  Lewis Carroll Society of North America, distributed by University Press of Virginia, Charlottesville, 1994.

Edward Wakeling.  Rediscovered Lewis Carroll Puzzles.  Dover, 1995.  Cited as Carroll‑Wakeling II, qv in Common References.

Martin Gardner.  The Universe in a Handkerchief.  Lewis Carroll's Mathematical Recreations, Games, Puzzles and Word Plays.  Copernicus (Springer, NY), 1996.  Cited as Carroll‑Gardner, qv in Common References.

Martin Gardner.  The Annotated Alice: The Definitive Edition.  1999.  [A combined version of The Annotated Alice and More Annotated Alice.]

 

          Professor Louis HOFFMANN  (1839‑1919)

 

          Pseudonym of Angelo John Lewis.

 

Joseph Foster.  Men-at-the-Bar: A biographical Hand-List of the Members of the Various Inns of Court, including Her Majesty's Judges, etc.  2nd ed, the author, 1885.  P. 277 is the entry for Lewis.  Born in London, eldest son of John Lewis.  Graduated from Wadham College, Oxford.  Entered Lincoln's Inn as a student in 1858, called to the bar there in 1861.  Married Mary Ann Avery in 1864.  Author of Manual of Indian Penal Code and Manual of Indian Civil Procedure.  Address: 12 Crescent Place, Mornington Crescent, London, NW.  (My thanks to the Library of Lincoln's Inn for this information.)

Anonymous.  Professor Hoffmann.  Mahatma 4:1 (Jul 1900) 377-378.  A brief note, with photograph, stating that he is Mr. Angelo Lewis, M.A. and Barrister-at-Law.

Will Goldston.  Will Goldston's Who's Who in Magic.  My version is included in a compendium called: Tricks that Mystify; Will Goldston, London, nd [1934-NUC].  Pp. 106-107.  Says he was a barrister, retired to Hastings about 1903 and died in 1917.

Who Was Who, 1916-1928, p. 627.  This says he attended North London Collegiate School and that he only practised law until 1876.  He was on the staff of the Saturday Review and a contributor to many journals.  Won the £100 prize offered by Youth's Companion (Boston) for best short story for boys.  Lists 36 books by him and 9 card games he invented.  Address: Manningford, Upper Bolebrooke Road, Bexhill-on-Sea.  (My thanks to the Library of Lincoln's Inn for this information.)

J. B. Findlay & Thomas A. Sawyer.  Professor Hoffmann: A Study.  Published by Thomas A. Sawyer, Tustin, California, 1977.  A short book, 12 + 67 pp, with two portraits (one from Mahatma) and 27pp of bibliography.  He was born at 3 Crescent Place, Mornington Crescent, London.  He was a barrister and wrote two books on Indian law.

Charles Reynolds.  Introduction  --  to the reprint of Hoffmann's Modern Magic, Dover, 1978, pp. v‑xiv.  This says Lewis was a barrister, which is mentioned in another reprint of a Hoffmann book and in S. H. Sharpe's translation of Ponsin on Conjuring.

Edward Hordern.  Foreword to this edition.  In:  Hoffmann's Puzzles Old and New (see under Common References), 1988 reprint, pp. v‑vi.  This says he was the Reverend Lewis, but this is corrected in Hoffmann-Hordern to saying he was a barrister.

Hoffmann-Hordern, p. viii, is a version of the photograph in Mahatma.

Hall, OCB, p. 189, gives Hoffmann's address as Ireton Lodge, Cromwell Ave., N. -- presumably the Cromwell Ave. in Highgate.

Toole Stott 386 gives a little information about Hoffmann and Modern Magic, including an address in Mornington Crescent in 1877.

No DNB or DSB entry -- I have suggested a DNB entry.

 

          Sam LOYD  (1841‑1911)  and  Sam LOYD JR.  (1873‑1934)

 

[W. R. Henry.]  Samuel Loyd.  [Biography.]  Dubuque Chess Journal, No. 66 (Aug-Sep 1875) 361-365.  ??NX -- o/o (11 Jul 91).

Loyd.  US Design 4793 -- Design for Puzzle-Blocks.  11 April 1871.  These are solid pieces, but unfortunately the drawing did not come with this, so I am not clear what they are.  ??Need drawing -- o/o (11 Jul 91).

Anonymous & Sam Loyd.  Loyd's puzzles (Introductory column).  Brooklyn Daily Eagle (22 Mar 1896) 23.  Says he lives at 153 Halsey St., Brooklyn.

L. D. Broughton Jr.  Samuel Loyd.  [A Biography.]  Lasker's Chess Magazine 1:2 (Dec 1904) 83-85.  About his chess problems with a mention of some of his puzzles.

G. G. Bain.  The prince of puzzle‑makers.  An interview with Sam Loyd.  Strand Magazine 34 (No. 204) (Dec 1907) 771‑777.  Solutions of Sam Loyd's puzzles.  Ibid. 35 (No. 205) (Jan 1908) 110.

Walter Prichard Eaton.  My fifty years in puzzleland -- Sam Loyd and his ten thousand brain‑teasers.  The Delineator (New York) (April 1911) 274 & 328.  Drawn portrait of Loyd, age 69.

Anon.  Puzzle inventor dead.  New-York Daily Tribune (12 Apr 1911) 7.  Says he died at his house, 153 Halsey St.  "He declared no one had ever succeeded in solving [the "Disappearing Chinaman"]."  Says he is survived by a son and two daughters (!! -- has anyone ever tracked the daughters and their descendents??).

Anon.  Sam Loyd, puzzle man, dies.  New York Times (12 Apr 1911) 13.  Says he was for some time editor of The Sanitary Engineer and a shrewd operator on Wall Street.

Anon.  Sam Loyd.  SA (22 Apr 1911) 40-41??  Says he was for some years chess editor of SA and was puzzle editor of Woman's Home Companion when he died.

W. P. Eaton.  Sam Loyd.  The American Magazine 72 (May 1911) 50, 51, 53.  Abridged version of Eaton's earlier article.  Photo of Loyd on p. 50.

P. J. Doyle.  Letter to the Chess column.  The Sunday Call [Newark, NJ] (21 May 1911), section III, p. 10.

A. C. White.  Sam Loyd and His Chess Problems.  Whitehead and Miller, Leeds, UK, 1913;  corrected, Dover, 1962.

Alain C. White.  Supplement to Sam Loyd and His Chess Problems.  Good Companion Chess Problem Club, Philadelphia, vol. I, nos. 11-12 (Aug 1914), 12pp.  This is mostly corrections of the chess problems, but adds a few family details with a picture of the Loyd Homestead and Grist Mill in Moylan, Pennsylvania.

Alain C. White.  Reminiscences of Sam Loyd's family.  The Problem [Pittsburgh] (28 Mar 1914) 2, 3, 6, 7.

Louis C. Karpinski.  Loyd, Samuel.  Dictionary of American Biography, Scribner's, NY, vol. XI, 1933, pp. 479‑480.

Loyd Jr.  SLAHP.  1928.  Preface gives some details of his life, making little mention of his father, "who was a famous mathematician and chess player".  He claims to have created over 10,000 puzzles. There are some vague biographical details on pp. 1‑22, e.g. 'Father conducted a printing establishment.'  'My "Missing Chinaman Puzzle"'.  (It may have been some such assertion that led me to estimate his birthdate as 1865, but I now see it is well known to be 1873.)

Anonymous.  Sam Loyd dead; puzzle creator.  New York Times (25 Feb 1934).  Obituary of Sam Loyd Jr.  Says he resided at 153 Halsey St., Brooklyn -- the same address as his father -- see the Brooklyn Daily Eagle article of 1896, above.  He worked from a studio at 246 Fulton St., Brooklyn.  It says Jr. invented 'How Old is Ann?'.

Clark Kinnaird.  Encyclopedia of Puzzles and Pastimes.  Grosset & Dunlap, NY, 1946.  Pp. 263‑267: Sam Loyd.  Asserts that Loyd Jr. invented 'How Old is Ann?'

Gardner.  Sam Loyd: America's greatest puzzlist.  SA (Aug 1957)  c= First Book, Chap. 9.

Gardner.  Advertising premiums.  SA (Nov 1971)  c= Wheels, chap. 12.

Will Shortz is working on a biography.

No DSB entry.

 

          François Anatole Édouard LUCAS  (1842‑1891)

 

Jeux Scientifiques de Ed. Lucas.  Advertisement by Chambon & Baye (14 rue Etienne-Marcel, Paris) for the 1re Serie of six games.   Cosmos.  Revue des Sciences et Leurs Applications 39 (NS No. 254) (7 Dec 1889) no page number on my photocopy.

B. Bailly [name not given, but supplied by Hinz].  Article on Lucas's puzzles.  Cosmos.  Revue des Sciences et Leurs Applications.  NS, 39 (No. 259) (11 Jan 1890) 156-159.  NEED 156‑157.

Nécrologie:  Édouard Lucas.  La Nature 19 (1891) II, 302.

Obituary notice: "La Nature announces the death of Prof. Edouard Lucas ...."  Nature 44 (15 Oct 1891) 574-575.

Duncan Harkin.  On the mathematical work of François‑Édouard‑Anatole Lucas.  L'Enseignement Math. (2) 3 (1957) 276‑288.  Pp. 282‑288 is a bibliography of 184 items.  I have found many Lucas publication not listed here and have started a new Bibliography -- see below.

P. J. Campbell.  Lucas' solution to the non‑attacking rooks problem.  JRM 9 (1976/77) 195‑200.  Gives life of Lucas.

A photo of Lucas is available from Bibliothèque Nationale, Service Photographique, 58 rue Richelieu, F‑75084 Paris Cedex 02, France.  Quote Cote du Document Ln27 . 43345 and Cote du Cliche 83 A 51772.  (??*)  I have obtained a copy, about 55 x 85 mm, with the photo in an oval surround.  It looks like a carte-de-visite, but has  Édouard LUCAS (1842-1891).  --  Phot. Zagel.  underneath.  (Thanks to H. W. Lenstra for the information.)

Norman T. Gridgeman.  Lucas, François‑Édouard‑Anatole.  DSB VIII, 531‑532.

Susanna S. Epp.  Discrete Mathematics with Applications.  Wadsworth, Belmont, Calif., 1990, p. 477 gives a small photo of Lucas which looks nothing like the photo from the BN.  I have since received a note from Epp via Paul Campbell that a wrong photo was used in the first edition, but this was corrected in later editions.

Alain Zalmanski.  Edouard Lucas  Quand l'arithmétique devient amusante.  Jouer Jeux Mathématiques 3 (Jul/Sep 1991) 5.  Brief notice of his life and work.

Andreas M. Hinz.  Pascal's triangle and the Tower of Hanoi.  AMM 99 (1992) 538-544.  Sketches Lucas' life and work, giving details that are not in the above items.

David Singmaster.  The publications of Édouard Lucas.  Draft version, 14pp, 1998.  I discovered many items in Dickson's History of the Theory of Numbers and elsewhere which are not given by Harkin (cf above).  This has 248 items, though many of these are multiple items so the actual count is perhaps 275.  However, Dickson does not give article titles, and may not give the pages of the entire article, so the same article may be cited more than once, at different pages.  I hope to fill in the missing information at some time.

 

          Hermann Cäsar Hannibal SCHUBERT  (1848-1911)

 

Acta Mathematica  1882-1912.  Table Générale des Tomes 1-35.  1913.  P. 169.  Portrait of Schubert.

Werner Burau.  Schubert, Hermann Cäsar Hannibal.  DSB XII, 227‑229.

 

          Walter William Rouse BALL  (1850‑1925)

 

Anon.  Obituary:  Mr. Rouse Ball.  The Times (6 Apr 1925) 16. 

Anon.  Funeral notice:  Mr. W. W. R. Ball.  The Times (9 Apr 1925) 13.

(Lord) Phillimore.  Letter:  Mr. Rouse Ball.  The Times (9 Apr 1925) 15.

"An old pupil".  The late Mr. Rouse Ball.  The Times (13 Apr 1925) 12.

J. J. Thomson.  W. W. Rouse Ball.  The Cambridge Review (24 Apr 1925) 341-342.

Anon.  Obituary of W. W. Rouse Ball.  Nature 115 (23 May 1925) 808‑809. 

Anon.  The late Mr. W. W. Rouse Ball.  The Trinity Magazine (Jun 1925) 53-54.

Anon.  Entry in Who's Who, 1925, p. 127.

Anon.  Wills and bequests:  Mr. Walter William Rouse Ball.  The Times (7 Sep 1925) 15.

E. T. Whittaker.  Obituary.  W. W. Rouse Ball.  Math. Gaz. 12 (No. 178) (Oct 1925) 449-454, with photo opp. p. 449.

F. Cajori.  Walter William Rouse Ball.  Isis 8 (1926) 321‑324.  Photo on plate 15, opp. p. 321.  Copy of Ball's 1924 Xmas card on p. 324.

J. A. Venn.  Alumni Cantabrigienses.  Part II:  From 1752 to 1900.  Vol. I, p. 136.  CUP, 1940.

David Singmaster.  Walter William Rouse Ball (1850-1925).  6pp handout for 1st UK Meeting on the History of Recreational Mathematics, 24 Oct 1992.  Plus extended biographical (6pp) and bibliographical (8pp) notes which repeat some of the material in the handout.

No DNB or DSB entry -- however I have offered to write a DNB entry.  I have since seen the proposed list of names for the next edition and Ball is already on it.

 

          Henry Ernest DUDENEY  (1857‑1930)

 

Anon. & Dudeney.  A chat with the puzzle king.  The Captain 2 (Dec? 1899) 314‑320, with photo.  Partly an interview.  Includes photos of Littlewick Meadow.

Anon.  Solutions to "Sphinx's puzzles".  The Captain 2:6 (Mar 1900) 598‑599  &  3:1 (Apr 1900) 89.

Anon.  Master of the breakfast table problem.  Daily Mail (1 Feb 1905) 7.  An interview with Dudeney in which he gives the better version of his spider and fly problem.

Fenn Sherie.  The Puzzle King: An Interview with Henry E. Dudeney.  Strand Magazine 71 (Apr 1926) 398‑4O4.

Alice Dudeney.  Preface to PCP, dated Dec 1931, pp. vii‑x.  The date of his death is erroneously given as 1931.

Gardner.  Henry Ernest Dudeney: England's greatest puzzlist.  SA (Jun 1958)  c= Second Book, chap. 3.

Angela Newing.  The Life and Work of H. E. Dudeney.  MS 21 (1988/89) 37‑44.

Angela Newing is working on a biography.

No DNB or DSB entry.  I have suggested a DNB entry.

 

          Wilhelm Ernst Martin Georg AHRENS  (1872‑1927)

 

Wilhelm Lorey.  Wilhelm Ahrens zum Gedächtnis.  Archiv für Geschichte der Mathematik, der Naturwissenschaften und der Technik 10 (1927/28) 328‑333.  Photo on p. 328.

O. Staude.  Dem Andenken an Dr. Wilhelm Ahrens.  Jahresbericht DMV 37 (1928) 286-287.

No DSB entry.

 

          Yakov Isidorovich PERELMAN [Я. И. Перелман]  (1882-1942)

 

Perelman.  FMP.  1984.  P. 2 (opp. TP) is a sketch of his life and the history of the book.  There is a small drawing of Perelman at the top of the page.

Patricio Barros.  Website -- Yakov I. Perelman [in Spanish]: www.geocities.com/yakov_perelman/index.html.  This includes a four page biography, in collaboration with Antonio Bravo, and two photos.

 

          Hubert PHILLIPS  (1891-1964)

 

Hubert Phillips.  Journey to Nowhere.  A Discursive Autobiography.  Macgibbon & Kee, London, 1960.  ??NYR

No DNB entry -- I have suggested one.

 

 

2.       GENERAL PUZZLE COLLECTIONS AND SURVEYS

 

H. E. Dudeney.  Great puzzle crazes.  London Magazine 13?? (Nov 1904) 478‑482.  Fifteen Puzzle.  Pigs in Clover, Answers, Pick-me-up (spiral ramp) and other dexterity puzzles.  Get Off the Earth.  Conjurer's Medal (ring maze).  Chinese Rings.  Chinese Cross (six piece burr).  Puzzle rings.  Solitaire.  The Mathematician's Puzzle (square, circle, triangle).  Imperial Scale.  Heart and Balls.

H. E. Dudeney.  Puzzles from games.  Strand Magazine 35 (No. 207) (Mar 1908) 339‑344.  Solutions.  Ibid. 35 (No. 208) (Apr 1908) 455‑458.

H. E. Dudeney.  Some much‑discussed puzzles.  Strand Magazine 35 (No. 209) (May 1908) 580‑584.  Solutions.  Ibid. 35 (No. 210) (Jun 1908) 696.

H. E. Dudeney.  The world's best puzzles.  Strand Magazine 36 (No. 216) (Dec 1908) 779‑787.  Solutions.  Ibid. 37 (No. 217) (Jan 1909) 113‑116.

H. E. Dudeney.  The psychology of puzzle crazes.  The Nineteenth Century 100:6 (Dec 1926) 868‑879.  Repeats much of his 1904 article.

Sam Loyd Jr.  Are you good at solving puzzles?  The American Magazine (Sep 1931) 61‑63, 133‑137.

Orville A. Sullivan.  Problems involving unusual situations.  SM 9 (1943) 114‑118  &  13 (1947) 102‑104.

 

 

3.       GENERAL HISTORICAL AND BIBLIOGRAPHICAL MATERIAL

 

          I have tried to divide this material into historical and bibliographical parts, but the two overlap considerably. 

 

3.A.   GENERAL HISTORICAL MATERIAL

 

Raffaella Franci.  Giochi matematici in trattati d'abaco del medioevo e del rinascimento.  Atti del Convegno Nazionale sui Giochi Creative, Siena, 11-14 Jun 1981.  Tipografia Senese for GIOCREA (Società Italiana Giochi Creativi), 1981.  Pp. 18-43.  Describes and quotes many typical problems.  17 references, several previously unknown to me.

Heinrich Hermelink.  Arabische Unterhaltungsmathematik als Spiegel Jahrtausendealter Kulturbeziehungen zwischen Ost und West.  Janus 65 (1978) 105-117, with English summary.  An English translation appeared as:  Arabic recreational mathematics as a mirror of age-old cultural relations between Eastern and Western civilizations; in:  Ahmad Y. Al-Hassan, Ghada Karmi & Nizar Namnum, eds.; Proceedings of the First International Symposium for the History of Arabic Science, April 1976 -- Vol. Two: Papers in European Languages; Institute for the History of Arabic Science, Aleppo, 1978, pp. 44-52.  (There are a few translation and typographical errors, which make it clear that the English version is a translation of the German.)

D. E. Smith.  On the origin of certain typical problems.  AMM 24 (1917) 64‑71.  (This is mostly contained in his History, vol. II, pp. 536‑548.)

 

          3.B.   BIBLIOGRAPHICAL MATERIAL

 

          Many of the items cited in the Common References have extensive bibliographies.  In particular:  BLC;  BMC;  BNC;  DNB;  DSB;  Halwas;  NUC;  Schaaf;  Smith & De Morgan: Rara;  Suter  are basic bibliographical sources.  Datta & Singh;  Dickson;  Heath: HGM;  Murray;  Sanford: H&S  &  Short History;  Smith: History  &  Source Book;  Struik;  Tropfke  are histories with extensive bibliographical references.  AR;  BR  are editions of early texts with substantial bibliographical material.  Ahrens: MUS;  Ball: MRE;  Berlekamp, Conway & Guy: Winning Ways;  Gardner;  Lucas: RM  are recreational books with some useful bibliographical material.  Of these, the material in Ahrens is by far the most useful.  The magic bibliographies of Christopher, Clarke & Blind, Hall, Heyl, Price (see HPL), Toole Stott and Volkmann & Tummers have considerable overlap with the present material, particularly for older books, though Hall, Heyl and Toole Stott restrict themselves to English material, while Volkmann & Tummers only considers German.  Santi is also very useful.  Below I give some additional bibliographical material which may be useful, arranged in author order.

 

Anonymous.  Mathematical bibliography.  SSM 48 (1948) 757‑760.  Covers recreations.

 

Wilhelm Ahrens.  Mathematische Spiele.  Section I G 1 of Encyklopadie der Math. Wiss., Vol. I, part 2, Teubner, Leipzig, 1900‑1904, pp. 1080‑1093.

Raymond Clare Archibald.  Notes on some minor English mathematical serials.  MG 14 (1928-29) 379-400.

Elliott M. Avedon  &  Brian Sutton‑Smith.  The Study of Games.  (Wiley, NY, 1971);  Krieger, Huntington, NY, 1979.

Anthony S. M. Dickins.  A Catalogue of Fairy Chess Books and Opuscules Donated to Cambridge University Library, 1972‑1973, by Anthony Dickins M.A.  Third ed., Q Press, Kew Gardens, UK, 1983.

Underwood Dudley.  An annotated list of recreational mathematics books.  JRM 2:1 (Jan 1969) 13-20.  61 titles, in English and in print at the time.

Aviezri S. Fraenkel.  Selected Bibliography on Combinatorial Games and Some Related Material.  There have been several versions with slightly varying titles.  The most recent printed version is:  400 items, 28 pp., including 4 pp of text, Sep 1990.  Technical Report CS90‑23, Weizmann Institute of Science, Rehovot, Israel.  = Proc. Symp. Appl. Math. 43 (1991) 191-226.  Fraenkel has since produced Update 1 to this which lists 430 items on 31pp, Aug 1992;  and Update 2, 480 items on 33pp, with 5 pp of text, accidentally dated Aug 1992 at the top but produced in Feb 1994.  On 22 Nov 1994, it became a dynamic survey on the Electronic J. Combinatorics and can be accessed from:

                    http://ejc.math.gatech.edu:8080/journal/surveys/index.html.

          It can also be accessed via anonymous ftp from  ftp.wisdom.weizmann.ac.il.  After logging in, do  cd pub/fraenkel  and then  get  one of the following three compressed files:  games.tex.z;  games.dvi.z;  games.ps.z.

Martin P. Gaffney  &  Lynn Arthur Steen.  Annotated Bibliography of Expository Writing in the Mathematical Sciences.  MAA, 1976.

JoAnne S. Growney.  Mathematics and the arts -- A bibliography.  Humanistic Mathematics Network Journal 8 (1993) 22-36.  General references.  Aesthetic standards for mathematics and other arts.  Biographies/autobiographies of mathematicians.  Mathematics and display of information (including mapmaking).  Mathematics and humor.  Mathematics and literature (fiction and fantasy).  Mathematics and music.  Mathematics and poetry.  Mathematics and the visual arts. 

JoAnne S. Growney.  Mathematics in Literature and Poetry.  Humanistic Mathematics Network Journal 10 (Aug 1994) 25-30.  Short survey.  3 pages of annotated references to 29 authors, some of several books.

R. C. Gupta.  A bibliography of selected book [sic] on history of mathematics.  The Mathematics Education 23 (1989) 21-29.

Trevor H. Hall.  Mathematicall Recreations.  Op. cit. in 1.  This is primarily concerned with the history of the book by van Etten.  [This booklet is revised as pp. 83-119 of Hall, OCB -- see Section 1.]

Catherine Perry Hargrave.  A History of Playing Cards and a Bibliography of Cards and Gaming.  (Houghton Mifflin, Boston, 1930);  Dover, 1966.

Susan Hill.  Catalogue of the Turner Collection of the History of Mathematics Held in the Library of the University of Keele.  University Library, Keele, 1982.  (Sadly this collection was secretly sold by Keele University in 1998 and has now been dispersed.)

Honeyman Collection -- see: Sotheby's.

Horblit Collection -- see: Sotheby's and H. P. Kraus.

Else Høyrup.  Books about Mathematics.  Roskilde Univ. Center, PO Box 260, DK‑4000, Roskilde, Denmark, 1979.

D. O. Koehler.  Mathematics and literature.  MM 55 (1982) 81-95.  64 references.  See Utz for some further material.

H. P. Kraus (16 East 46th Street, New York, 10017).  The History of Science including Navigation.

Catalogue 168.  A First Selection of Books from the Library of Harrison D. Horblit.  Nd [c1976].

Catalogue 169.  A Further Selection of Books, 1641-1700 (Wing Period) from the Library of Harrison D. Horblit.  Nd [c1976].

Catalogue 171.  Another Selection of Books from the Library of Harrison D. Horblit.  Nd [c1976].

These are the continuations of the catalogues issued by Sotheby's, qv.

John S. Lew.  Mathematical references in literature.  Humanistic Mathematics Network Journal 7 (1992) 26-47.

Antonius van der Linde.  Das erst Jartausend [sic] der Schachlitteratur -- (850‑1880).  (1880);  Facsimile reprint by Caissa Limited Editions, Yorklyn, Delaware, 1979, HB.

Andy Liu.  Appendix III:  A selected bibliography on popular mathematics.  Delta-k 27:3 (Apr 1989)  --  Special issue:  Mathematics for Gifted Students, 55-83.

Édouard Lucas.  Récréations mathématiques, vol 1 (i.e. RM1), pp. 237-248 is an Index Bibliographique.

Felix Müller.  Führer durch die mathematische Literature mit besonderer Berücksichtigung der historisch wichtigen Schriften.  Abhandlungen zur Geschichte der Mathematik 27 (1903).

Charles W. Newhall.  "Recreations" in secondary mathematics.  SSM 15 (1915) 277‑293.

Mathematical Association.  259 London Road, Leicester, LE2 3BE.

Catalogue of Books and Pamphlets in the Library.  No details, [c1912], 19pp, bound in at end of Mathematical Gazette, vol. 6 (1911‑1912).

A First List of Books & Pamphlets in the Library of the Mathematical Association -- Books and Pamphlets acquired before 1924.  Bell, London, 1926.

A Second List of Books & Pamphlets in the Library of the Mathematical Association -- Books and Pamphlets acquired during 1924 and 1925.  Bell, London, 1929.

A Third List of Books & Pamphlets in the Library of the Mathematical Association -- Books and Pamphlets added from 1926 to 1929.  Bell, London, 1930.

A Fourth List of Books & Pamphlets in the Library of the Mathematical Association -- Books and Pamphlets added from 1930 to 1935.  Bell, London, 1936. 

           Lists 1‑4 edited by E. H. Neville.

Books and Periodicals in the Library of the Mathematical Association.  Ed. by R. L. Goodstein.  MA, 1962.  Includes the four previous lists and additions through 1961.

SEE ALSO:  Riley; Rollett; F. R. Watson.

Stanley Rabinowitz.  Index to Mathematical Problems  1980-1984.  MathPro Press, Westford, Massachusetts, 1992.

Cecil B. Read  &  James K. Bidwell.

Selected articles dealing with the history of elementary Mathematics.  SSM 76 (1976) 477-483.

Periodical articles dealing with the history of advanced mathematics -- Parts I & II.  SSM 76 (1976) 581-598  &  687-703.

Rudolf H. Rheinhardt.  Bibliography on Whist and Playing Cards.  From:  Whist Scores and Card-table Talk, Chicago, 1887.  Reprinted by L. & P. Parris, Llandrindod Wells, nd [1980s].

Pietro Riccardi.  Biblioteca Matematica Italiana dalla Origine della Stampa ai Primi Anni del Secolo XIX.  G. G. Görlich, Milan, 1952, 2 vols.  This work appeared in several parts and supplements in the late 19C and early 20C, mostly published by the Società Tipografica Modense, Modena, 1878-1893.  Because it appeared in parts, the contents of early copies are variable and even the reprints may vary.  The contents of this set are as follows.

          I.        20pp prelims   +  Col. 1 - 656 (Abaco - Kirchoffer).  [= original Vol. I.]

                    Col. 1 - 676 (La Cometa - Zuzzeri)  +  2pp correzioni.  [= original Vol. II.]

          II.      4pp titles and reverses.  Correzioni ed Aggiunte.  [= original Appendice.]

                    Serie I.a  Col. 1 - 78  +  1½pp  Continuazione delle Correzioni (note that these

                              have  Pag.  when they mean  Col.).

                    Serie II.a.  Col. 81 - 156.

                    Serie III.a.  Col. 157 - 192  +  Aggiunte al Catalogo delle Opere di sovente citate,

                              col. 193-194  +  1p Continuazione delle Correzioni (note that these have 

                              Pag.  when they mean  Col.). 

                    Serie IV.a.  Col. 197 - 208  +  Seconda Aggiunta al Catalogo delle Opere più di

                              sovente citate, col. 209 - 212  +  Continuazione delle Correzioni in

                              col. 211-212.

                    Serie V.a.  Col. 1 - 180.

                    Serie VI.a.  Col. 179 - 200.

                              Serie V & VI must have been published as one volume as Serie V ends

                                        halfway down a page and then Serie VI begins on the same page.

                    Serie VII.a.  2pp introductory note by Ettore Bortolotti in 1928 saying that this

                              material was left as a manuscript by Riccardi and never previously

                              published  +  Col. 1 - 106.

                    Indice Alfabetico, of authors, covering the original material and all seven Series

                              of Correzioni ed Aggiunte, in 34 unnumbered columns.

                    Parte Seconda.  Classificazione per materie delle opere nella Parte I.  18pp

                              (including a chronological table)  +  subject index, pp. 1 - 294.

                    Catalogo  Delle opere più di sovente citate, col. 1 - 54.

                    [I have seen an early version which had the following parts:  Vol. I, 1893, col. 1‑656;  Vol. II, 1873, col. 1-676;  Appendice, 1878-1880-1893, col. 1-228.  Appendice, nd, col. 1-212.  Serie V, col. 1-228.  Parte 2, Vol. 1, 1880, pp. 1-294.  Renner Katalog 87 describes it as 5 in 2 vols.]

A. W. Riley.  School Library Mathematics List -- Supplement No. 1.  MA, 1973. 

          SEE ALSO:  Rollett.

Tom Rodgers.  Catalog of his collection of books on recreational mathematics, etc.  The author, Atlanta, May 1991, 40pp.

Leo F. Rogers.  Finding Out in the History of Mathematics.  Produced by the author, London, c1985, 52pp.

A. P. Rollett.  School Library Mathematics List.  Bell, London, for MA, 1966. 

          SEE ALSO:  Riley.

Charles L. Rulfs.  Origins of some conjuring works.  Magicol 24 (May 1971) 3-5.

José A. Sánchez Pérez.  Las Matematicas en la Biblioteca del Escorial.  Imprenta de Estanislao Maestre, Madrid, 1929.

William L. Schaaf. 

List of works on recreational mathematics.  SM 10 (1944) 193-200. 

           PLUS:  A. Gloden; Additions to Schaaf's "List of works on mathematical recreations"; SM 13 (1947) 127.

A Bibliography of Recreational Mathematics.  Op. cit. in Common References, 4 vols., 1955-1978.  In these volumes he gives several lists of relevant books.

           Books for the periods 1900-1925 and 1925-c1956 are given as Sections 1.1 (pp. 2-3) and 1.2 (pp. 4-12) in Vol. 1. 

           Chapter 9, pp. 144-148, of Vol. 1, is a Supplement, generally covering c1954-c1962, but with some older items. 

           In Vol. 2, 1970, the Appendix, pp. 181-191, extends to c1969, including some older items and repeating a few from the Supplement of Vol. 1. 

           Appendix A of Vol. 3, 1973, pp. 111-113, adds some more items up through 1972. 

           Appendix A, pp. 134-137, of Vol. 4, 1978, extends up through 1977.

 

  The following VESTPOCKET BIBLIOGRAPHIES are extensions of the material

           in his Bibliographies.

No. 1:  Pythagoras and rational triangles;  Geoboards and lattices.  JRM 16:2 (1983-84) 81-88.

No. 2:  Combinatorics;  Gambling and sports.  JRM 16:3 (1983-84) 170-181.

No. 3:  Tessellations and polyominoes;  Art and music.  JRM 16:4 (1983-84) 268‑280.

No. 4:  Recreational miscellany.  JRM 17:1 (1984-85) 22-31.

No. 5:  Polyhedra;  Topology;  Map coloring.  JRM 17:2 (1984-85) 95-105.

No. 6:  Sundry algebraic notes.  JRM 17:3 (1984-85) 195-203.

No. 7:  Sundry geometric notes.  JRM 18:1 (1985-86) 36-44.

No. 8:  Probability;  Gambling.  JRM 18:2 (1985-86) 101-109.

No. 9:  Games and puzzles.  JRM 18:3 (1985-86) 161-167.

No. 10:          Recreational mathematics;  Logical puzzles;  Expository mathematics.  JRM 18:4 (1985-86) 241-246.

No. 11:          Logic, Artificial intelligence, and Mathematical foundations.  JRM 19:1 (1987) 3-9.

No. 12:          Magic squares and cubes;  Latin squares;  Mystic arrays and Number patterns.  JRM 19:2 (1987) 81-86.

 

The High School Mathematics Library.  NCTM, (1960, 1963, 1967, 1970, 1973);  6th ed., 1976;  7th ed., 1982;  8th ed., 1987. 

SEE ALSO:  Wheeler; Wheeler & Hardgrove.

Early Books on Magic Squares.  JRM 16:1 (1983-84) 1-6.

William L. Schaaf  &  David Singmaster.  Books on Recreational Mathematics.  A Supplement to the Lists in William L. Schaaf's A Bibliography of Recreational Mathematics.  Collected by William L. Schaaf;  typed and annotated by David Singmaster.  School of Computing, Information Systems and Mathematics, South Bank University, London, SE1 0AA.  18pp, Dec 1992 and revised several times afterwards.

Peter Schreiber.

Mathematik und belletristik [1.] & 2. Teil.  Mitteilungen der Mathematischen Gesellschaft der Deutschen Demokratischer Republik.  (1986), no. 4, 57-71  &  (1988), no. 1-2, 55-61.  Good on German works relating mathematics and arts.

Mathematiker als Memoirenschreiber.  Alpha (Berlin) (1991), no. 4, no page numbers on copy received from author.  Extends previous work.

S. N. Sen.  Scientific works in Sanskrit, translated into foreign languages and vice‑versa in the 18th and 19th century A.D.  Indian J. History of Science 7 (1972) 44‑70.

Will Shortz.  Puzzleana [catalogue of his puzzle books].  Produced by the author.  14 editions have appeared.  The latest is:  May 1992, 88pp with 1175 entries in 26 categories, with indexes of authors and anonymous titles.  Some entries cover multiple items.  In Jan 1995, he produced a 19pp Supplement extending to a total of 1451 entries.

David Singmaster.

The Bibliography of Some Recreational Mathematics Books.  School of Computing, Information Systems

           and Mathematics, South Bank Univ.

13 Nov 1994, 39pp.  Technical Report SBU-CISM-94-09.

2nd ed., Aug 1995, 41pp.  Technical Report SBU-CISM-95-08. 

3rd ed., Jun 1996, 42pp.  Technical Report SBU-CISM-96-12.

4th ed., Jun 1998, 44pp.  Technical Report SBU-CISM-98-02.

           (Current version is 61pp.)

Books on Recreational Mathematics.  School of Computing, Information Systems and

           Mathematics, South Bank Univ., until 1996.

21 Jan 1991.  Approx. 2951 items on 120pp, ringbound.

30 Jan 1992.  Approx. 3314 items on 138pp, ringbound.

10 Jan 1993.  Approx. 3606 items on 95pp, ringbound.

10 Dec 1994.  Approx. 4303 items plus 67 Old Books on 110pp.  Technical

           Report SBU‑CISM-94-11.

10 Oct 1996.  Approx. 4842 items plus 84 Old Books on 127pp.  Technical

           Report SBU-CISM-96-17.

24 May 1999.  Approx. 6015 items plus 133 Old Books on 166pp.  Technical

           Report SBU-CISM-99-14.

26 Feb 2002.  Approx. 7185 items plus 192 Old Books plus Supplement of

           Calculating Devices, on 220pp. thermal bound.

22 Nov 2003.  Approx. 7811 items plus 202 Old Books plus Supplement of

           Calculating Devices, on 244pp. thermal bound.

Index to Martin Gardner's Columns and Cross Reference to His Books.  (Oct 1993.)  Slightly revised as:  Technical Report SBU-CISM-95-09; School of Computing, Information Systems, and Mathematics; South Bank University, London, Aug 1995, 22pp.  (Current version is 23pp and Don Knuth has sent 9pp of additional material and I will combine these at some time.)

Harold Adrian Smith.  Dick and Fitzgerald Publishers.  Books at Brown 34 (1987) 108-114. 

Sotheby's [Sotheby Parke Bernet]. 

Catalogue of the J. B. Findlay Collection  Books and Periodicals on Conjuring and the Allied Arts.  Part I: A-O  5-6 Jul 1979.  Part II: P-Z  plus:  Mimeographed Books and Instructions; Flick Books  Catalogues of Apparatus and Tricks  Autograph Letters, Manuscripts, and Typescripts  4-5 Oct 1979.  Part III: Posters and Playbills  3-4 Jul 1980.  Each with estimates and results lists.

The Celebrated Library of Harrison D. Horblit Esq.  Early Science  Navigation & Travel  Including Americana  with a few medical books.  Part I  A - C  10/11 Jun 1974.  Part II  D - G  11 Nov 1974.  HB.  The sale was then cancelled and the library was sold to E. P. Kraus, qv, who issued three further catalogues, c1976.

The Honeyman Collection of Scientific Books and Manuscripts.  Seven volumes, each

           with estimates and results booklets.

Part I: Printed Books A-B, 30-31 Oct 1978.

Part II:  Printed Books C-E, 30 Apr - 1 May 1979.

Part III: Manuscripts and Autograph Letters of the 12th to the 20th Centuries.

Part IV: Printed Books F-J, 5-6 Nov 1979.

Part V: Printed Books K-M, 12-13 May 1980.

Part VI: Printed Books N-Sa, 10-11 Nov 1980.

Part VII:  Printed Books Sc-Z and Addenda, 19-20 May 1981.

Lynn A. Steen, ed. 

Library Recommendations for Undergraduate Mathematics.  MAA Reports No. 4, 1992.

Two-Year College Mathematics Library Recommendations.  MAA Reports No. 5, 1992.

Strens/Guy Collection.  Author/Title Listing.  Univ. of Calgary.  Preliminary Catalogue, 319 pp., July 1986.  [The original has a lot of blank space.  I have a computer version which is reduced to 67pp.]

Eva Germaine Rimington Taylor.  The Mathematical Practitioners of Tudor & Stuart England  1485-1714.  CUP for the Institute of Navigation, 1970.

Eva Germaine Rimington Taylor.  The Mathematical Practitioners of Hanoverian England  1714‑1840.  CUP for the Institute of Navigation, 1966. 

                    PLUS:  Kate Bostock, Susan Hurt & Michael Hart; An Index to the Mathematical Practitioners of Hanoverian England  1714-1840; Harriet Wynter Ltd., London, 1980.

W. R. Utz.  Letter:  Mathematics in literature.  MM 55 (1982) 249‑250.  Utz has sent his 3pp original more detailed version along with 4pp of further citations.  This extends Koehler's article.

George Walker.  The Art of Chess-Play: A New Treatise on the Game of Chess.  4th ed., Sherwood, Gilbert & Piper, London, 1846.  Appendix:  Bibliographical Catalogue of the chief printed books, writers, and miscellaneous articles on chess, up to the present time, pp. 339-375.

Frank R. [Joe] Watson, ed.  Booklists.  MA.

Puzzles, Problems, Games and Mathematical Recreations.  16pp, 1980.

Selections from the Recommended Books.  18pp, 1980.

Full List of Recommended Books.  105pp, 1984.

Margariete Montague Wheeler.  Mathematics Library -- Elementary and Junior High School.  5th ed., NCTM, 1986. 

          SEE ALSO:  Schaaf; Wheeler & Hardgrove.

Margariete Montague Wheeler  &  Clarence Ethel Hardgrove.  Mathematics Library -- Elementary and Junior High School.  NCTM, (1960;  1968;  1973);  4th ed., 1978. 

          SEE ALSO:  Schaaf; Wheeler.

Ernst Wölffing.  Mathematischer Bücherschatz.  Systematisches Verzeichnis der wichtigsten deutschen und ausländischen Lehrbücher und Monographien des 19. Jahrhunderts auf dem Gebiete der mathematischen Wissenschaften.  I:  Reine Mathematik;  (II:  Angewandte Mathematik never appeared).  AGM 16, part I (1903).

 

 

4.       MATHEMATICAL GAMES

 

Aviezri S. Fraenkel.  Selected Bibliography on Combinatorial Games and Some Related Material.  Op. cit. in 3.B.

 

4.A.   GENERAL THEORY AND NIM‑LIKE GAMES

 

          Conway's extension of this theory is well described in Winning Ways and later work is listed in Fraenkel's Bibliography -- see section 3.B & 4 -- so I will not cover such material here.

 

          4.A.1.          ONE PILE GAME

 

          See MUS I 145-147.

          (a, b)  denotes the game where one can take  1, 2, ...,  or  a  away from one pile, starting with  b  in the pile, with the last player winning.  The version  (10, 100)  is sometimes called Piquet des Cavaliers or Piquet à Cheval, a name which initially perplexed me.  Piquet is one of the older card games, being well known to Rabelais (1534) and was known in the 16C as Cent (or Saunt or Saint) because of its goal of 100 points.  See:  David Parlett; (Oxford Guide to Card Games, 1990  =)  A History of Card Games; OUP, 1991, pp. 24 & 175-181.  The connection with horses undoubtedly indicates that  (10, 100)  was viewed as a game which could be played without cards, while riding -- see Les Amusemens, Decremps.

 

                        INDEX

(  3,   13)                Dudeney, Stong

(  3,   15)                Mittenzwey, Hoffmann, Mr. X, Dudeney, Blyth,

(  3,   17)                Fourrey,

(  3,   21)                Blyth, Hummerston,

(  4,   15)                Mittenzwey,

(  6,   30)                Pacioli, Leske, Mittenzwey, Ducret,

(  6,   31)                Baker,

(  6,   50)                Ball-FitzPatrick,

(  6,   52)                Rational Recreations

(  6,   57)                Hummerston,

(  7,   40)                Mittenzwey,

(  7,   41)                Sprague,

(  7,   45)                Mittenzwey,

(  7,   50)                Decremps,

(  7,   60)                Fourrey,

(  8, 100)                Bachet, Carroll,

(  9, 100)                Bachet, Ozanam, Alberti

(10, 100)                 Bachet, Henrion, Ozanam, Alberti, Les Amusemens, Hooper, Decremps,

                                        Badcock, Jackson, Rational Recreations, Manuel des Sorciers,

                                        Boy's Own Book, Nuts to Crack, Young Man's Book, Carroll,

                                        Magician's Own Book, Book of 500 Puzzles, Secret Out,

                                        Boy's Own Conjuring Book, Vinot, Riecke, Fourrey, Ducret, Devant,

(10, 120)                 Bachet,

(12, 134)                 Decremps,

          General case:  Bachet, Ozanam, Alberti, Decremps, Boy's Own Book, Young Man's Book, Vinot, Mittenzwey, (others ?? check)

          Versions with limited numbers of each value or using a die -- see 4.A.1.a.

          Version where an odd number in total has to be taken:  Dudeney, Grossman & Kramer, Sprague.

          Versions with last player losing:  Mittenzwey,

 

Pacioli.  De Viribus.  c1500.  Ff. 73v - 76v.  XXXIIII effecto afinire qualunch' numero na'ze al compagno anon prendere piu de un termi(n)ato .n. (34th effect to finish whatever number is before the company, not taking more than a limiting number) = Peirani 109‑112.  Phrases it as an addition problem.  Considers  (6, 30)  and the general problem.

David Parlett.  (Originally:  The Oxford Guide to Card Games; OUP, 1990);  reissued as:  A History of Card Games.  Penguin, 1991, pp. 174-175.  "Early references to 'les luettes', said to have been played by Anne de Bretagne and Archduke Philip the Fair in 1503, and by Gargantua in 1534, seem to suggest a game of the Nim family (removing numbers of objects from rows and columns)."

Cardan.  Practica Arithmetice.  1539.  Chap. 61, section 18, ff. T.iiii.v - T.v.r (p. 113).  "Ludi mentales".  One has  1, 3, 6  and the other has  2, 4, 5;  or one has  1, 3, 5, 8, 9  and the other has  2, 4, 6, 7, 10;  one one wants to make  100.  "Sunt magnæ inventionis, & ego inveni æquitando & sine aliquo auxilio cum socio potes ludere & memorium exercere ...."

Baker.  Well Spring of Sciences.  1562?  Prob. 5: To play at 31 with Numbers, 1670: pp. 353‑354.  ??NX.  (6, 31).

Bachet.  Problemes.  1612.  Prob. XIX: 1612, 99-103.  Prob. XXII, 1624: 170-173;  1884: 115‑117.  Phrases it as an addition problem.  First considers  (10, 100),  then  (10, 120),  (8, 100),  (9, 100),  and the general case.  Labosne omits the demonstration.

Dennis Henrion.  Nottes to van Etten.  1630.  Pp. 19-20.  (10, 100)  as an addition problem, citing Bachet.

Ozanam.  1694.  Prob. 21, 1696: 71-72;  1708: 63‑64.  Prob. 25, 1725: 182‑184.  Prob. 14, 1778: 162-164;  1803: 163-164;  1814: 143-145.  Prob. 13, 1840: 73-74.  Phrases it as an addition problem.  Considers  (10, 100)  and  (9, 100)  and remarks on the general case.

Alberti.  1747.  Due persone essendo convenuto ..., pp. 105‑108 (66‑67).  This is a slight recasting of Ozanam.

Les Amusemens.  1749.  Prob. 10, p. 130: Le Piquet des Cavaliers.  (10, 100)  in additive form.  "Deux amis voyagent à cheval, l'un propose à l'autre un cent de Piquet sans carte." 

William Hooper.  Rational Recreations, In which the Principles of Numbers and Natural Philosophy Are clearly and copiously elucidated, by a series of Easy, Entertaining, Interesting Experiments.  Among which are All those commonly performed with the cards. [Taken from my 2nd ed.]  4 vols., L. Davis et al., London, 1774;  2nd ed., corrected, L. Davis et al., London, 1783-1782 (vol. 1 says 1783, the others say 1782; BMC gives 1783-82);  3rd ed., corrected, 1787;  4th ed., corrected, B. Law et al., London, 1794.  [Hall, BCB 180-184  &  Toole Stott 389-392.  Hall says the first four eds. have identical pagination.  I have not seen any difference in the first four editions, except as noted in Section 6.P.2.  Hall, OCB, p. 155.  Heyl 177 notes the different datings of the 2nd ed,  Hall, BCB 184 and Toole Stott 393 is a 2 vol. 4th ed., corrected, London, 1802.  Toole Stott 394 is a 2 vol. ed. from Perth, 1801.  I have a note that there was an 1816 ed, but I have no details.  Since all relevant material seems the same in all volumes, I will cite this as 1774.]  Vol. 1, recreation VIII: The magical century.  (10, 100)  in additive form.  Mentions other versions and the general rule.

                    I don't see any connection between this and Rational Recreations, 1824.

Henri Decremps.  Codicile de Jérôme Sharp, Professeur de Physique amusante;  Où l'on trouve parmi plusieurs Tours dont il n'est point parlé dans son Testament, diverses récréations relatives aux Sciences & Beaux-Arts;  Pour servir de troisième suite  À La Magie Blanche Dévoilée.  Lesclapart, Paris, 1788.  Chap. XXVII, pp. 177-184: Principes mathématiques sur le piquet à cheval, ou l'art de gagner son diner en se promenant.  Does  (10, 100) in additive form, then discusses the general method, illustrating with  (7, 50)  and  (12, 134).

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 33-34, no. 48: A curious recreation with a hundred numbers, usually called the magical century.  (10, 100)  as an additive problem where each person starts with  50  counters.  Discusses general case, but doesn't notice that the limitation to  50  counters each considerably changes the game!

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles, no. 47, pp. 11 & 64.  Additive form of  (10, 100).

Rational Recreations.  1824.  Exercise 12(?), pp. 57-58.  As in Badcock.  Then says it can be generalised and gives  (6, 52).

Manuel des Sorciers.  1825.  Pp. 57-58, art. 30: Le piquet sans cartes.  ??NX  (10, 100) done subtractively.

The Boy's Own Book.

The certain game.  1828: 177;  1828-2: 236;  1829 (US): 104;  1855: 386‑387;  1868: 427.

The magical century.  1828: 180;  1828-2: 236‑237;  1829 (US): 104-105;  1855: 391‑392. 

Both are additive phrasings of  (10, 100).  The latter mentions using other numbers and how to win then.

Nuts to Crack V (1836), no. 70.  An arithmetical problem.  (10, 100).

Young Man's Book.  1839.  Pp. 294-295.  A curious Recreation with a Hundred Numbers, usually called the Magical Century.  Almost identical to Boy's Own Book.

Lewis Carroll. 

Diary entry for 5 Feb 1856.  In Carroll-Gardner, pp. 42-43.  (10, 100).  Wakeling's note in the Diaries indicates he is not familiar with this game.

Diary entry for 24 Oct 1872.  Says he has written out the rules for Arithmetical Croquet, a game he recently invented.  Roger Lancelyn Green's abridged version of the Diaries, 1954, prints a MS version dated 22 Apr 1889.  Carroll-Wakeling, prob. 38, pp. 52-53 and Carroll-Gardner, pp. 39 & 42 reprint this, but Gardner has a misprinted date of 1899.  Basically  (8, 100),  but passing the values  10, 20, ...,  requires special moves and one may have to go backward.  Also, when a move is made, some moves are then barred for the next player.  Overall, the rules are typically Carrollian-baroque.

Magician's Own Book.  1857.

The certain game, p. 243.  As in Boy's Own Book.

The magical century, pp. 244-245.  As in Boy's Own Book.

Book of 500 Puzzles.  1859.

The certain game, p. 57.  As in Boy's Own Book.

The magical century, pp. 58-59.  As in Boy's Own Book.

The Secret Out.  1859.  Piquet on horseback, pp. 397-398 (UK: 130‑131) -- additive  (10, 100)  unclearly explained.

Boy's Own Conjuring Book.  1860.

The certain game, pp. 213‑214.  As in Boy's Own Book.

Magical century, pp. 215.  As in Boy's Own Book.

Vinot.  1860.  Art. XI: Un cent de piquet sans cartes, pp. 19-20.  (10. 100).  Says the idea can be generalised, giving  (7, 52)  as an example.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 563-III, pp. 247: Wer von 30 Rechenpfennigen den letzen wegnimmt, hat gewonnen.  (6, 30).

F. J. P. Riecke.  Mathematische Unterhaltungen.  3 vols., Karl Aue, Stuttgart, 1867, 1868 & 1873;  reprint in one vol., Sändig, Wiesbaden, 1973.  Vol. 3, art 22.2, p. 44.  Additive form of  (10, 100).

Mittenzwey.  1880.  Probs. 286-287, pp. 52 & 101-102;  1895?: 315-317, pp. 56 & 103-104;  1917: 315-317, pp. 51 & 98.  

(6, 30), last player wins.

(4, 15), last player loses, the solution discusses other cases: (7, 40), (7, 45) and indicates the general solution.

(added in 1895?) (3, 15), last player loses.

Hoffmann.  1893.  Chap VII, no. 19: The fifteen matches puzzle, pp. 292 & 300‑301 = Hoffmann-Hordern, p. 197.  (3, 15).  c= Benson, 1904, The fifteen match puzzle, pp. 241‑242.

Ball-FitzPatrick.  1st ed., 1898.  Deuxième exemple, pp. 29-30.  (6, 50).

E. Fourrey.  Récréations Arithmétiques.  (Nony, Paris, 1899;  2nd ed., 1901);  3rd ed., Vuibert & Nony, Paris, 1904;  (4th ed., 1907);  8th ed., Librairie Vuibert, Paris, 1947.  [The 3rd and 8th eds are identical except for the title page, so presumably are identical to the 1st ed.]  Sections 65‑66: Le jeu du piquet à cheval, pp. 48‑49.  Additive forms of  (10, 100)  and  (7, 60).  Then gives subtractive form for a pile of matches for  (3, 17).

Étienne Ducret.  Récréations Mathématiques.  Garnier Frères, Paris, nd [not in BN, but a similar book, nouv. ed., is 1892].  Pp. 102‑104: Le piquet à cheval.  Additive version of  (10, 100)  with some explanation of the use of the term piquet.  Discusses  (6, 30).

Mr. X [possibly J. K. Benson -- see entry for Benson in Abbreviations].  His Pages.  The Royal Magazine 9:3 (Jan 1903) 298-299.  A good game for two.  (3, 15) as a subtraction game.

David Devant.  Tricks for Everyone.  Clever Conjuring with Everyday Objects.  C. Arthur Pearson, London, 1910.  A counting race, pp. 52-53.  (10, 100).

Dudeney.  AM.  1917.  Prob. 392: The pebble game, pp. 117 & 240.  (3, 15)  &  (3, 13)  with the object being to take an odd number in total.  For 15, first player wins;  for 13, second player wins.  (Barnard (50 Telegraph ..., 1985) gives the case  (3, 13).)

Blyth.  Match-Stick Magic.  1921. 

Fifteen matchstick game, pp. 87-88.  (3, 15).

Majority matchstick game, p. 88.  (3, 21).

Hummerston.  Fun, Mirth & Mystery.  1924. 

Two second-sight tricks (no. 2), p. 84.  (6, 57),  last player losing.

A match mystery, p. 99.  (3, 21),  last player losing.

H. D. Grossman & David Kramer.  A new match-game.  AMM 52 (1945) 441‑443.  Cites Dudeney and says Games Digest (April 1938) also gave a version, but without solution.  Gives a general solution whether one wants to take an odd total or an even total.

C. L. Stong.  The Amateur Scientist.  Ill. by Roger Hayward.  S&S, 1960.  How to design a "Pircuit" or Puzzle circuit, pp. 388-394.  On pp. 388-391, Harry Rudloe describes a relay circuit for playing the subtractive form of  (3, 13), which he calls the "battle of numbers" game.

Ronald Sprague.  Unterhaltsame Mathematik.  Vieweg, Braunschweig, 1961.  Translated by T. H. O'Beirne as:  Recreations in Mathematics, Blackie, London, 1963.  Problem 24: "Ungerade" gewinnt, pp. 16 & 44‑45.  (= 'Odd' is the winner, pp. 18 & 53‑55.)  (7, 41)  with the winner being the one who takes an odd number in total.  Solves  (7, b)  and states the structure for  (a, b). 

                    I also have some other recent references to this problem.  Lewis (1983) gives a general solution which seems to be wrong. 

 

          4.A.1.a.       THE 31 GAME

 

          Numerical variations:  Badcock, Gibson, McKay.

          Die versions:  Secret Out (UK), Loyd, Mott-Smith, Murphy.

 

Baker.  Well Spring of Sciences.  1562?  Prob. 5: To play at 31 with Numbers, 1670: pp. 353‑354.  ??NX.  (6, 31).  ??CHECK if this has the limited use of numbers.

John Fisher.  Never Give a Sucker an Even Break.  (1976);  Sphere Books, London, 1978.  Thirty-one, pp. 102-104.  (6, 31)  additively, but played with just 4 of each value, the 24 cards of ranks  1 -- 6,  and the first to exceed 31 loses.  He says it is played extensively in Australia and often referred to as "The Australian Gambling Game of 31".  Cites the 19C gambling expert Jonathan Harrington Green who says it was invented by Charles James Fox (1749‑1806).  Gives some analysis.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 33-34, no. 48: A curious recreation with a hundred numbers, usually called the magical century.  (10, 100)  as an additive problem where each person starts with  50  counters.  Discusses general case, but doesn't notice that the limitation to  50  counters each considerably changes the game!

Nuts to Crack V (1836), no. 71.  (6, 31) additively, with four of each value.  "Set down on a slate, four rows of figures, thus:--  ...  You agree to rub out one figure alternately, to see who shall first make the number thirty-one."

Magician's Own Book.  1857.  Art. 31: The trick of thirty‑one, pp. 70‑71.  (6, 31)  additively, but played with just 4 of each value -- e.g. the 24 cards of ranks  1 -- 6.  The author advises you not to play it for money with "sporting men" and says it it due to Mr. Fox.  Cf Fisher.  = Boy's Own Conjuring Book; 1860; Art. 29: The trick of thirty‑one, pp. 78‑79.  = The Secret Out; 1859, pp. 65-66, which adds a footnote that the trick is taken from the book One Hundred Gambler Tricks with Cards by J. H. Green, reformed gambler, published by Dick & Fitzgerald.

The Secret Out (UK), c1860.  To throw thirty‑one with a die before your antagonist, p. 7.  This is incomprehensible, but is probably the version discussed by Mott-Smith.

Edward S. Sackett.  US Patent 275,526 -- Game.  Filed: 9 Dec 1882;  patented: 10 Apr 1883.  1p + 1p diagrams.  Frame of six rows holding four blocks which can be slid from one side to the other to play the 31 game, though other numbers of rows, blocks and goal may be used.  Gives an example of a play, but doesn't go into the strategy at all.

Larry Freeman.  Yesterday's Games.  Taken from "an 1880 text" of games.  (American edition by H. Chadwick.)  Century House, Watkins Glen, NY, 1970.  P. 107: Thirty-one.  (6, 31)  with 4 of each value -- as in Magician's Own Book.

Algernon Bray.  Letter:  "31" game.  Knowledge 3 (4 May 1883) 268, item 806.  "... has lately made its appearance in New York, ...."  Seems to have no idea as how to win.

Loyd.  Problem 38: The twenty‑five up puzzle.  Tit‑Bits 32 (12 Jun  &  3 Jul 1897) 193  &  258.  = Cyclopedia.  1914.  The dice game, pp. 243 & 372.  = SLAHP: How games originate, pp. 73 & 114.  The first play is arbitrary.  The second play is by throwing a die.  Further values are obtained by rolling the die by a quarter turn.

Ball-FitzPatrick.  1st ed., 1898.  Généralization récente de cette question, pp. 30-31.  (6, 50)  with each number usable at most 3 times.  Some analysis.

Ball.  MRE, 4th ed., 1905, p. 20.  Some analysis of  (6, 50)  where each player can play a value at most 3 times -- as in Ball-FitzPatrick, but with the additional sentence:  "I have never seen this extension described in print ...."  He also mentions playing with values limited to two times.  In the 5th ed., 1911, pp. 19-21, he elaborates his analysis.

Dudeney.  CP.  1907.  Prob. 79: The thirty-one game, pp. 125-127 & 224.  Says it used to be popular with card-sharpers at racecourses, etc.  States the first player can win if he starts with 1, 2 or 5, but the analysis of cases 1 and 2 is complicated.  This occurs as No. 459: The thirty-one puzzle, Weekly Dispatch (17 Aug 1902) 13 & (31 Aug 1902) 13, but he leaves the case of opening move 2 to the reader, but I don't see the answer given in the next few columns.

Devant.  Tricks for Everyone.  Op. cit. in 4.A.1.  1910.  The thirty-one trick, pp. 53-54.  Says to get to  3, 10, 17, 24.

Hummerston.  Fun, Mirth & Mystery.  1924.  Thirty-one -- a game of skill, pp. 95-96.  This uses a layout of four copies of the numbers  1, 2, 3, 4, 5, 6  with one copy of  20  in a  5 x 5  square with the  20  in the centre.  Says to get to  3, 10, 17, 24,  but that this will lose to an experienced player.

Loyd Jr.  SLAHP.  1928.  The "31 Puzzle Game", pp. 3 & 87.  Loyd Jr says that as a boy, he often had to play it against all comers with a $50 prize to anyone who could beat 'Loyd's boy'.  This is the game that Loyd Sr called 'Blind Luck', but I haven't found it in the Cyclopedia.  States the first player wins with 1, 2 or 5, but only sketches the case for opening with 5.  I have seen an example of Blind Luck -- it has four each of the numbers 1 - 6 arranged around a frame containing a horseshoe with 13 in it.

McKay.  Party Night.  1940.  The 21 race, pp. 166.  Using the numbers  1, 2, 3, 4,  at most four times, achieve 21.  Says to get  1, 6, 11, 16.  He doesn't realise that the sucker can be mislead into playing first with a 1 and losing!  Says that with  1, ..., 5  at most four times, one wants to achieve 26 and that with  1, ..., 6  at most four times, one wants to achieve 31.  Gives just the key numbers each time.

Geoffrey Mott-Smith.  Mathematical Puzzles for Beginners and Enthusiasts.  (Blakiston, 1946);  revised 2nd ed., Dover, 1954.

Prob. 179: The thirty-one game, pp. 117-119  &  231-232.  As in Dudeney.

Prob. 180: Thirty-one with dice, p. 119  &  232-233.  Throw a die, then make quarter turns to produce a total of 31.  Analysis based on digital roots (i.e. remainders (mod 9)).  First player wins if the die comes up 4, otherwise the second player can win.  He doesn't treat any other totals.

"Willane".  Willane's Wizardry.  Academy of Recorded Crafts, Arts and Sciences, Croydon, 1947.  "Trente et un", pp. 56-57.  Says he doesn't know any name for this.  Get 31 using 4 each of the cards  A, 2, ..., 6.  Says first player loses easily if he starts with  4, 5, 6  (not true according to Dudeney) and that gamblers dupe the sucker by starting with 3 and winning enough that the sucker thinks he can win by starting with 3.  But if he starts with a 1 or 2, then the second player must play low and hope for a break.

Walter B. Gibson.  Fell's Guide to Papercraft Tricks, Games and Puzzles.  Frederick Fell, NY, 1963.  Pp. 54-55: First to fifty.  First describes  (50, 6),  but then adds a version with slips of paper:  eight marked 1 and seven marked with  2, 3, 4, 5, 6  and you secretly extract a 6 slip when the other player starts.

Harold Newman.  The 31 Game.  JRM 23:3 (1991) 205-209.  Extended analysis.  Confirms Dudeney.  Only cites Dudeney & Mott-Smith.

Bernard Murphy.  The rotating die game.  Plus 27 (Summer 1994) 14-16.  Analyses the die version as described by Mott-Smith and finds the set,  S(n),  of winning moves for achieving a count of  n  by the first player, is periodic with period 9 from  n = 8,  i.e.  S(n+9) = S(n)  for  n ³ 8.  There is no first player winning move if and only if  n  is a multiple of 9.  [I have confirmed this independently.]

Ken de Courcy.  The Australian Gambling Game of 31.  Supreme Magic Publication, Bideford, Devon, nd [1980s?].  Brief description of the game and some indications of how to win.  He then plays the game with face-down cards!  However, he insures that the cards by him are one of of each rank and he knows where they are.

 

          4.A.2.          SYMMETRY ARGUMENTS

 

Loyd??  Problem 43: The daisy game.  Tit‑Bits 32 (17 Jul  &  7 Aug 1897) 291  &  349.  (= Cyclopedia.  1914.  A daisy puzzle game, pp. 85 & 350.  c= MPSL2, prob. 57, pp. 40‑41 & 140.  c= SLAHP: The daisy game, pp. 42 & 99.)  Circular version of Kayles with 13 objects.  Solution uses a symmetry argument -- but the Tit‑Bits solution was written by Dudeney.

Dudeney.  Problem 500: The cigar puzzle.  Weekly Dispatch (7 Jun, 21 Jun, 5 Jul, 1903) all p. 16.  (= AM, prob. 398, pp. 119, 242.)  Symmetry in placement game, using cigars on a table.

Loyd.  Cyclopedia.  1914.  The great Columbus problem, pp. 169 & 361.  (= MPSL1, prob. 65, pp. 62 & 144.  = SLAHP: When men laid eggs, pp. 75 & 115.)  Placing eggs on a table.

Maurice Kraitchik.  La Mathématique des Jeux.  Stevens, Bruxelles, 1930.  Section XII, prob. 1, p. 296.  (= Mathematical Recreations; Allen & Unwin, London, 1943; Problem 1, pp. 13‑14.)  Child plays black and white against two chess players and guarantees to win one game.  [MJ cites L'Echiquier (1925) 84, 151.]

                    CAUTION.  The 2nd edition of Math. des Jeux, 1953, is a translation of Mathematical Recreations and hence omits much of the earlier edition.

Leopold.  At Ease!  1943.  Chess wizardry in two minutes, pp. 105‑106.  Same as Kraitchik.

 

          4.A.3.          KAYLES

 

          This has objects in a line or a circle and one can remove one object or two adjacent objects (or more adjacent objects in a generalized version of the game).  This derives from earlier games with an array of pins at which one throws a ball or stick.

          Murray 442 cites Act 17 of Edward IV, c.3 (1477):  "Diversez novelx ymagines jeuez appellez Cloishe Kayles ..."  This outlawed such games.  A 14C picture is given in [J. A. R. Pimlott; Recreations; Studio Vista, 1968, plate 9, from BM Royal MS 10 E IV f.99] showing a  3 x 3  array of pins.  A version is shown in Pieter Bruegel's painting "Children's Games" of 1560 with balls being thrown at a row of pins by a wall, in the back right of the scene.  Versions of the game are given in the works of Strutt and Gomme cited in 4.B.1.  Gomme II 115‑116 discusses it under Roly‑poly, citing Strutt and some other sources.  Strutt 270‑271 (= Strutt-Cox 219-220) calls it "Kayles, written also cayles and keiles, derived from the French word quilles".  He has redrawings of two 14C engravings (neither that in Pimlott) showing lines of pins at which one throws a stick (= plate opp. 220 in Strutt-Cox).  He also says Closh or Cloish seems to be the same game and cites prohibitions of it in c1478 et seq.  Loggats was analogous and was prohibited under Henry VIII and is mentioned in Hamlet.

 

14C MS in the British Museum, Royal Library, No. 2, B. vii.  Reproduced in Strutt, p. 271.  Shows a monk(?) standing by a line of eight conical pins and another monk(?) throwing a stick at the pins.

Anonymous.  Games of the 16th Century.  The Rockliff New Project Series.  Devised by Arthur B. Allen.  The Spacious Days of Queen Elizabeth.  Background Book No. 5.  Rockliff Publishing, London, ©1950, 4th ptg.  The Background Books seem to be consecutively paginated as this booklet is paginated 129-152.  Pp. 133-134 describes loggats, quoting Hamlet and an unknown poet of 1611.  P. 137 is a photograph of the above 14C illustration.  The caption is "Skittles, or "Kayals", and Throwing a Whirling Stick".

van Etten.  1624.  Prob. 72 (misnumbered 58) (65), pp 68‑69 (97‑98): Du jeu des quilles (Of the play at Keyles or Nine-Pins).  Describes the game as a kind of ninepins.

Loyd.  Problem 43: The daisy game.  Tit‑Bits 32 (17 Jul  &  7 Aug 1897) 291  &  349.  (= Cyclopedia.  1914.  A daisy puzzle game, pp. 85 & 350.  c= MPSL2, prob. 57, pp. 40‑41 & 140.  c= SLAHP: The daisy game, pp. 42 & 99.)  Circular version of Kayles with 13 objects.  See also 4.A.2.

Dudeney.  Sharpshooters puzzle.  Problem 430.  Weekly Dispatch (26 Jan, 9 Feb, 1902) both p. 13.  Simple version of Kayles.

Ball.  MRE, 4th ed., 1905, pp. 19-20.  Cites Loyd in Tit‑Bits.  Gives the general version:  place  p  counters in a circle and one can take not more than  m  adjacent ones.

Dudeney.  CP.  1907.  Prob. 73: The game of Kayles, pp. 118‑119 & 220.  Kayles with 13 objects.

Loyd.  Cyclopedia.  1914.  Rip van Winkle puzzle, pp. 232 & 369‑370.  (c= MPSL2, prob. 6, pp. 5 & 122.)  Linear version with 13 pins and the second knocked down.  Gardner asserts that Dudeney invented Kayles, but it seems to be an abstraction from the old form of the game.

Rohrbough.  Puzzle Craft, later version, 1940s?.  Daisy Game, p. 22.  Kayles with 13 petals of a daisy.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 45, pp. 48 & 95.  Circular kayles with five objects.

Doubleday - 2.  1971.  Take your pick, pp. 63-65.  This is Kayles with a row of 10, but he says the first player can only take one.

 

          4.A.4.          NIM

 

          Nim is the game with a number of piles and a player can take any number from one of the piles.  Normally the last one to play wins.

 

David Parlett.  (Originally:  The Oxford Guide to Card Games; OUP, 1990);  reissued as:  A History of Card Games.  Penguin, 1991.  Pp. 174-175.  "Early references to 'les luettes', said to have been played by Anne de Bretagne and Archduke Philip the Fair in 1503, and by Gargantua in 1534, seem to suggest a game of the Nim family (removing numbers of objects from rows and columns)."

Charles L. Bouton.  Nim: a game with a complete mathematical theory.  Annals of Math. (2) 3 (1901/02) 35‑39.  He says Nim is played at American colleges and "has been called Fan‑Tan, but as it is not the Chinese game of that name, the name in the title is proposed for it."  He says Paul E. More showed him the misère (= last player loses)  version in 1899, so it seems that Bouton did not actually invent the game himself.

Ahrens.  "Nim", ein amerikanisches Spiel mit mathematischer Theorie.  Naturwissenschaftliche Wochenschrift  17:22 (2 Mar 1902) 258‑260.  He says that Bouton has admitted that he had confused Nim and Fan‑Tan.  Fan‑Tan is a Chinese game where you bet on the number of counters (mod 4) in someone's hand.  Parker, Ancient Ceylon, op. cit. in 4.B.1, pp. 570-571, describes a similar game, based on odd and even, as popular in Ceylon and "certainly one of the earliest of all games".

                    For more about Fan-Tan, see the following.

Stewart Culin.  Chess and playing cards.  Catalogue of games and implements for divination exhibited by the United States National Museum in connection with the Department of Archæology and Paleontology of the University of Pennsylvania at the Cotton States and International Exposition, Atlanta, Georgia, 1895.  IN: Report of the U. S. National Museum, year ending June 30, 1896.  Government Printing Office, Washington, 1898, HB, pp. 665-942.  [There is a reprint by Ayer Co., Salem, Mass., c1990.]  Fan-Tan (= Fán t‘án = repeatedly spreading out) is described on pp. 891 & 896, with discussion of related games on pp. 889-902.

Alan S. C. Ross.  Note 2334:  The name of the game of Nim.  MG 37 (No. 320) (May 1953) 119‑120.  Conjectures Bouton formed the word 'nim' from the German 'nimm'.  Gives some discussion of Fan‑Tan and quotes MUS I 72.

J. L. Walsh.  Letter:  The name of the game of Nim.  MG 37 (No. 322) (Dec 1953) 290.  Relates that Bouton said that he had chosen the word from the German 'nimm' and dropped one 'm'.

W. A. Wythoff.  A modification of the game of Nim.  Nieuw Archief voor Wiskunde (Groningen) (2) 7 (1907) 199‑202.  He considers a Nim game with two piles allows the extra move of taking the same amount from both piles.  [Is there a version with more piles where one can take any number from one pile or equal amounts from two piles??  See Barnard, below for a three pile version.]

Ahrens.  MUS I.  1910.  III.3.VII: Nim, pp. 72‑88.  Notes that Nim is not the same as Fan‑Tan, has been known in Germany for decades and is played in China.  Gives a thorough discussion of the theory of Nim and of an equivalent game and of Wythoff's game.

E. H. Moore.  A generalization of the game called Nim.  Annals of Math.  (2) 11 (1910) 93‑94.  He considers a Nim game with  n  piles and one is allowed to take any number from at most  k  piles.

Ball.  MRE, 5th ed., 1911, p. 21.  Sketches the game of Nim and its theory.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 13: The last match, pp. 10-11.  Thirty matches divided at random into three heaps.  Last player loses.  Explanation of how to win is rather cryptic:  "you must try and take away ... sufficient ... to leave the matches in the two or three heaps remaining, paired in ones, twos, fours, etc., in respect of each other."

Loyd Jr.  SLAHP.  1928.  A tricky game, pp. 47 & 102.  Nim (3, 4, 8).

Emanuel Lasker.  Brettspiele der Völker.  1931.  See comments in 4.A.5.  Jörg Bewersdorff [email of 6 Jun 1999] says that Lasker considered a three person Nim and found an equilibrium for it -- see: Jörg Bewersdorff; Glück, Logik und Bluff  Mathematik im Spiel -- Methoden, Ergebnisse und Grenzen; Vieweg, 1998, Section 2.3  Ein Spiel zu dritt, pp. 110-115. 

Lynn Rohrbough, ed.  Fun in Small Spaces.  Handy Series, Kit Q, Cooperative Recreation Service, Delaware, Ohio, nd [c1935].  Take Last, p. 10.  Last player loses Nim (3, 5, 7).

Rohrbough.  Puzzle Craft.  1932.

Japanese Corn Game, p. 6 (= p. 6 of 1940s?).  Last player loses Nim (1, 2, 3, 4, 5).

Japanese Corn Game, p. 23.    Last player loses Nim (3, 5, 7).

René de Possel.  Sur la Théorie Mathématique des Jeux de Hasard et de Réflexion.  Actualités Scientifiques et Industrielles 436.  Hermann, Paris, 1936.  Gives the theory of Nim and also the misère version.

Depew.  Cokesbury Game Book.  1939.  Make him take it, pp. 187-188.  Nim (3, 4, 5),  last player loses.

Edward U. Condon, Gereld L. Tawney & Willard A. Derr.  US Patent 2,215,544 -- Machine to Play Game of Nim.  Filed: 26 Apr 1940;  patented: 24 Sep 1940.  10pp + 11pp diagrams.

E. U. Condon.  The Nimatron.  AMM 49 (1942) 330‑332.  Has photo of the machine.

Benedict Nixon & Len Johnson.  Letters to the Notes & Queries Column.  The Guardian (4 Dec 1989) 27.  Reprinted in:  Notes & Queries, Vol. 1; Fourth Estate, London, 1990, pp. 14-15.  These describe the Ferranti Nimrod machine for playing Nim at the Festival of Britain, 1951.  Johnson says it played  Nim (3, 5, 6)  with a maximum move of  3.  The Catalogue of the Exhibition of Science shows this as taking place in the Science Museum.

H. S. M. Coxeter.  The golden section, phyllotaxis, and Wythoff's game.  SM 19 (1953) 135‑143.  Sketches history and interconnections.

H. S. M. Coxeter.  Introduction to Geometry.  Wiley, 1961.  Chap. 11: The golden section and phyllotaxis, pp. 160-172.  Extends his 1953 material.

A. P. Domoryad.  Mathematical Games and Pastimes.  (Moscow, 1961).  Translated by Halina Moss.  Pergamon, Oxford, 1963.  Chap. 10: Games with piles of objects, pp. 61‑70.  On p. 62, he asserts that Wythoff's game is 'the Chinese national game tsyanshidzi ("picking stones")'.  However M.‑K. Siu cannot recognise such a Chinese game, unless it refers to a form of jacks, which has no obvious connection with Wythoff's game or other Nim games.  He says there is a Chinese character, 'nian', which is pronounced 'nim' in Cantonese and means to pick up or take things.

N. L. Haddock.  Note 2973:  A note on the game of Nim.  MG 45 (No. 353) (Oct 1961) 245‑246.  Wonders if the game of Nim is related to Mancala games.

T. H. O'Beirne.  Puzzles and Paradoxes.  OUP, 1965.  Section on misère version of Wythoff's game, p. 133.  Richard Guy (letter of 27 Feb 1985) says this is one of O'Beirne's few mistakes -- cf next entry.

Winning Ways.  1982.  P. 407 says Wythoff's game is also called Chinese Nim or Tsyan‑shizi.  No reference given.  See comment under Domoryad above.  This says many authors have done this incorrectly.

D. St. P. Barnard.  50 Daily Telegraph Brain‑Twisters.  Javelin Books, Poole, Dorset, 1985.  Prob. 30: All buttoned up, pp. 49‑50, 91 & 115.  He suggests three pile game where one can take any number from one pile or an equal number from any two or all three piles.  [See my note to Wythoff, above.]

Matthias Mala.  Schnelle Spiele.  Hugendubel, Munich, 1988.  San Shan, p. 66.  This describes a nim-like game named San Shan and says it was played in ancient China.

Jagannath V. Badami.  Musings on Arithmetical Numbers  Plus Delightful Magic Squares.  Published by the author, Bangalore, India, nd [Preface dated 9 Sep 1999].  Section 4.16: The game of Nim, pp. 124-125.  This is a rather confused description of one pile games (21, 5) and (41, 5), but he refers to solving them by (mentally) dividing the pile into piles.  This makes me think of combining the two games, i.e. playing Nim with several piles but with a limit on the number one can take in a move.

 

          4.A.5.          GENERAL THEORY

 

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  Ff. 134-144 are:  Essay 10 Part 5.  See 4.B.1 for more details.  At the top of f. 134.r, he has added a note:  "This is probably my earliest Note on Games of Skill.  I do not recollect the date.  3 March 1865".  He then describes Tit Tat To and makes some simple analysis, but he never uses a name for it. 

Charles Babbage.  Notebooks -- unpublished collection of MSS in the BM as Add. MS 37205.  ??NX.  See 4.B.1 for more details.  On f. 304, he starts on analysis of games.  Ff. 310‑383 are almost entirely devoted to Tit-Tat-To, with some general discussions.  F. 321.r, 10 Sep 1860, is the beginning of a summary of his work on games of skill in general.  F. 324-333, Oct 1844, studies "General laws for all games of Skill between two players" and draws flow charts showing the basic recursive analysis of a game tree (ff. 325.v & 325.r).  On f. 332, he counts the number of positions in Tit Tat To as  9! + 8! + ... + 1!  =  409,113.  F. 333 has an idea of the tree structure of a game.

John M. Dubbey.  The Mathematical Work of Charles Babbage.  CUP, 1978, pp. 96‑97 & 125‑130.  See 4.B.1 for more details.  He discusses the above Babbage material.  On p. 127, Dubbey has:  "The basic problem is one that appears not to have been previously considered in the history of mathematics."  Dubbey, on p. 129, says:  "This analysis ... must count as the first recorded stochastic process in the history of mathematics."  However, it is really a deterministic two-person game.

E. Zermelo.  Über eine Anwendung der Mengenlehre auf die Theorie des Schachspiels.  Proc. 5th ICM (1912), CUP, 1913,  vol. II, 501‑504.  Gives general idea of first and second person games.

Ahrens.  A&N.  1918.  P. 154, note.  Says that each particular Dots and Boxes board, with rational play, has a definite outcome.

W. Rivier.  Archives des Sciences Physiques et Naturelles (Nov/Dec 1921).  ??NYS -- cited by Rivier (1935) who says that the later article is a new and simpler version of this one.

H. Steinhaus.  Difinicje potrzebne do teorji gry i pościgu (Definitions for a theory of games and pursuit).  Myśl Akademicka (Lwów) 1:1 (Dec 1925) 13‑14 (in Polish).  Translated, with an introduction by Kuhn and a letter from Steinhaus in:  Naval Research Logistics Quarterly 7 (1960) 105‑108.

Dénès König.  Über eine Schlussweise aus dem Endlichen ins Unendliche.  Mitteilungen der Universitä Szeged 3 (1927) 121-130.  ??NYS -- cited by Rivier (1935).  Kalmár cites it to the same Acta as his article. 

László Kalmár.  Zur Theorie der abstracten Spiele.  Acta Litt. Sci. Regia Univ. Hungaricae Francisco‑Josephine (Szeged) 4 (1927) 62‑85.  Says there is a gap in Zermelo which has been mended by König.  Lengthy approach, but clearly gets the idea of first and second person games.

Max Euwe.  Proc. Koninklijke Akadamie van Wetenschappen te Amsterdam 32:5 (1929).  ??NYS -- cited by Rivier (1935).

Emanuel Lasker.  Brettspiele der Völker.  Rätsel‑ und mathematische Spiele.  A. Scherl, Berlin, 1931, pp. 170‑203.  Studies the one pile game  (100, 5)  and the sum of two one‑pile games:  (100, 5) + (50, 3).  Discusses Nimm, "an old Chinese game according to Ahrens" and says the solver is unknown.  Gives Lasker's Nim -- one can take any amount from a pile or split it in two -- and several other variants.  Notes that  2nd person + 2nd person  is  2nd person  while  2nd person + 1st person  is  1st person.  Gives the idea of equivalent positions.  Studies three (and more) person games, assuming the pay‑offs are all different.  Studies some probabilistic games.  Jörg Bewersdorff [email of 6 Jun 1999] observes that Lasker's analysis of his Nim got very close to the idea of the Sprague-Grundy number.  See: Jörg Bewersdorff; Glück, Logik und Bluff  Mathematik im Spiel -- Methoden, Ergebnisse und Grenzen; Vieweg, 1998, Section 2.5  Lasker-Nim: Gewinn auf verborgenem Weg, pp. 118-124. 

W. Rivier.  Une theorie mathématique des jeux de combinaisions.  Comptes-Rendus du Premier Congrès International de Récréation Mathématique, Bruxelles, 1935.  Sphinx, Bruxelles, 1935, pp. 106‑113.  A revised and simplified version of his 1921 article.  He cites and briefly discusses Zermelo, König and Euwe.  He seems to be classifying games as first player or second player.

René de Possel.  Sur la Théorie Mathématique des Jeux de Hasard et de Réflexion.  Actualités Scientifiques et Industrielles 436.  Hermann, Paris, 1936.  Gives the theory of Nim and also the misère version.  Shows that any combinatorial game is a win, loss or draw and describes the nature of first and second person positions.  He then goes on to consider games with chance and/or bluffing, based on von Neumann's 1927 paper.

R. Sprague.  Über mathematische Kampfspiele.  Tôhoku Math. J. 41 (1935/36) 438‑444.

P. M. Grundy.  Mathematics and games.  Eureka 2 (1939) 6‑8.  Reprinted, ibid. 27 (1964) 9‑11.  These two papers develop the Sprague-Grundy Number of a game.

D. W. Davies.  A theory of chess and noughts and crosses.  Penguin Science News 16 (Jun 1950) 40-64.  Sketches general ideas of tree structure, Sprague-Grundy number, rational play, etc.

H. Steinhaus.  Games, an informal talk.  AMM 72 (1965) 457‑468.  Discusses Zermelo and says he wasn't aware of Zermelo in 1925.  Gives Mycielski's formulation and proof via de Morgan's laws.  Goes into pursuit and infinite games and their relation to the Axiom of Choice.

H. Steinhaus.  (Proof that a game without ties has a strategy.)  In:  M. Kac; Hugo Steinhaus -- a reminiscence and a tribute; AMM 81 (1974) 572‑581.  Repeats idea of his 1965 talk.

 

          4.B.   PARTICULAR GAMES

 

          See 5.M for Sim and 5.R.5 for Fox and Geese, etc.

          Most of the board games described here are classic and have been extensively described and illustrated in the various standard books on board games, particularly the works of Robert C. Bell, especially his Board and Table Games from Many Civilizations;  OUP, vol. I, 1960, vol. II, 1969;  combined and revised ed., Dover, 1979 and the older work of Edward G. Falkener; Games Ancient and Oriental and How to Play Them;  Longmans, Green, 1892;  Dover, 1961.  The works by Culin (see 4.A.4, 4.B.5 and 4.B.9) are often useful.  Several general works on games are cited in 4.B.1 and 4.B.5 -- I have read Murray's History of Board Games Other than Chess, but not yet entered the material.  Note that many of these works are more concerned with the game than with its history and have a tendency to exaggerate the ages of games by assuming, e.g. that a  3 x 3  board must have been used for Tic-Tac-Toe.  I will not try to duplicate the descriptions by Bell, Falkener and others, but will try to outline the earliest history, especially when it is at variance with common belief.  The most detailed mathematical analyses are generally in Winning Ways.

 

          4.B.1. TIC‑TAC‑TOE  =  NOUGHTS AND CROSSES

 

          Popular belief is that the game is ancient and universal -- e.g. see Brandreth, 1976.  However the game appears to have evolved from earlier three‑in‑a‑row games, e.g. Nine Holes or Three Men's Morris, in the early 19C.  See also the historical material in 4.B.5.  The game is not mentioned in Strutt nor most other 19C books on games, not even in Kate Greenaway's Book of Games (1889), nor in Halliwell's section on slate games (op. cit. in 7.L.1, 1849, pp. 103-104), but there may be an 1875 description in Strutt-Cox of 1903.  Babbage refers to it in his unpublished MSS of c1820 as a children's game, but without giving it a name.  In 1842, he calls it Tit Tat To and he uses slight variations on this name in his extended studies of the game -- see below.  The OED's earliest references are:  1849 for Tip‑tap‑toe;  1855 for Tit‑tat‑toe;  1861 for Oughts and Crosses.  However, the first two entries may be referring to some other game -- e.g. the entries for Tick‑tack‑toe for 1884 & 1899 are clearly to the game that Gomme calls Tit‑tat‑toe.  Von der Lasa cites a 1838-39 Swedish book for Tripp, Trapp, Trull.  Van der Linde (1874, op. cit. in 5.F.1) gives Tik, Tak, Tol as the Dutch name.  Using the works of Strutt, Gomme, Strutt-Cox, Fiske, Murray, the OED and some personal communications, I have compiled a separate index of  121  variant names which refer to  5  basic games, with a few variants and a few unknown games.  The Murray and Parker material is given first, as it deals generally with the ancient history.  Then I list several standard sources and then summarize their content.  Other material follows that.  Fiske says that van der Linde and von der Lasa (see 5.F.1) mention early appearances of Morris games, but rather briefly and I don't always have that material.

          The usual  #  shape board will be so indicated.  If one is setting down pieces, then the board is often drawn as a 'crossed square', i.e. a square with its horizontal and vertical midlines drawn, and one plays on the intersections.  Fiske 127 says this form is common in Germany, but unknown in England and the US.  In addition, the diagonals are often drawn, producing a 'doubly crossed square'.  The squares are sometime drawn as circles giving a 'crossed circle' and a 'doubly crossed circle', though it is hard to identify the corners in a crossed circle.  The  3 x 3  array of dots sometimes occurs.  The standard  #  pattern is sometimes surrounded by a square producing a '3 x 3  chessboard'.

          Fiske 129 says the English play with  O  and  +,  while the Swedes play with  O  and  1.  My experience is that English and Americans play with  O  and  X.  One English friend said that where she grew up, it was called 'Exeter's Nose' as a deliberate corruption of 'Xs and Os'.

          The first clear references to the standard game of Noughts and Crosses are Babbage (1820) and the items discussed under Tic-tac-toe below.  Further clear references are:  Cassell's, Berg, A wrangler ..., Dudeney, White and everything entered below after White.

          Misère version: Gardner (1957); Scotts (1975);

 

Murray mentions Morris, which he generally calls Merels, many times.  Besides the many specific references mentioned below and in 4.B.5, he shows, on p. 614, under Nine Holes and Three Men's Morris, a number of  3 x 3  diagrams. 

                    Kurna, Egypt, (-14C) -- a double crossed square and a double crossed circle -- see Parker below.

                    Ptolemaic Egypt (in the BM, no. 14315) -- a square with  #  drawn inside.  See below where I describe this, from a recent exhibition, as just a  #  board.

                    Ceylon -- a doubly crossed square -- see Parker below.

                    Rome and Pompeii -- doubly crossed circles. 

          Under Nine Holes, he says a piece can be moved to any vacant point;  under Three Men's Morris, he says a man can only be moved along a marked line to an adjacent point, i.e. horizontally, vertically or along a main diagonal.

                    Under Nine Holes, he shows the  #  board for English Noughts and Crosses.  He specifically notes that the pieces do not move.  His only other mention of this board is for a Swedish game called Tripp, Trapp, Trull, but he does not state that the pieces do not move.  He gives no other examples of the  #  board nor of non‑moving pieces.

                    He also mentions Five (or Six) Men's Morris, of which little is known.  On p. 133, he mentions a  3 x 3  "board of nine points used for a game essentially identical with the 'three men's merels', which has existed in China from at least the time of the Liang dynasty (A.D. 502‑557).  The 'Swei shu' (first half of the 7th c.) gives the names of twenty books on this game."

H. Parker.  Ancient Ceylon.  ??, London, 1909;  Asian Educational Services, New Delhi, 1981.  Nerenchi keliya, pp. 577‑580 & 644.  There is a crossed square with small holes at the intersections at the Temple of Kurna, Upper Egypt, ‑14C.  [Rohrbough, loc. cit. in 4.B.5, says this temple was started by Ramses I and completed by Seti in -1336/-1333, citing J. Royal Asiatic Soc. (1783) 17.]  On p. 644, he shows 34 mason's diagrams from Kurna, which include  #,  #  in a circle, crossed square with small holes at the intersections, doubly crossed square, doubly crossed circle.  He cites Bell, Arch. Survey of Ceylon, Third Progress Report, p. 5 note, for for a doubly crossed square in Ceylon, c1C, but Noughts and Crosses is not found in the interior of Ceylon.  The doubly crossed square was used in 18C Ireland.  On pp. 643-665, he discusses appearances of the crossed square and doubly crossed circle as designs or characters and claims they have mystic significance.  On p. 662, he lists many early appearances of the  #  pattern.

Murray 440, note 63, includes a reference to Soutendam; Keurboek van Delft; Delft, c1425, f. 78 (or p. 78?); who says games of subtlety are allowed, e.g. ... ticktacken.  There is no indication if this may be our game and the OED indicates that such names were used for backgammon back to 1558.  The OED doesn't cite:  W. Shakespeare; Measure for Measure, c1604.  Act I, scene ii, line 180 (or 196):  "foolishly lost at a game of ticktack".  Later it was more common as Tric-trac. 

Murray 746 notes a Welsh game Gwyddbwyll mentioned in the Mabinogion (14C).  The name is cognate with the Irish Fidchell and may be a Three Men's Morris, but the game was already forgotten by the 15C.

 

                    STANDARD SOURCES ON GAMES

 

Joseph Strutt.  The Sports and Pastimes of the People of England.  (With title starting:  Glig‑Gamena Angel-Ðeod., or the Sports ...; J. White, London, 1791, 1801, 1810).  A new edition, with a copious index, by William Hone.  Tegg, London, 1830, 1831, 1833, 1834, 1838, 1841, 1850, 1855, 1875, 1876, 1891.  [The 1830 ed. has a preface, omitted in 1833, stating that the 1810 ed. is the same as the 1801 ed. and that Hone has only changed it by adding the Index and incorporating some footnotes into the text.]  [Hall, BCB 263-266 are:  1801,  1810,  1830,  1831.  Toole Stott 647-656 are:  1791;  1801;  1810;  1828-1830 in 10 monthly parts with Index by Hone; 1830;  1830;  1833;  1838;  1841;  1876, an expanded ed, ed by Hone.  Heyl 300-302 gives  1830;  1838;  1850.  Toole Stott 653 says the sheets were remaindered to Hone, who omitted the first 8pp and issued it in  1833,  1834,  1838,  1841.  I have seen an 1855 ed.  C&B list 1801, 1810, 1830, 1903.  BMC has 1801, 1810, 1830, 1833, 1834, 1838, 1841, 1875, 1876, 1898.]

                    Strutt-Cox.  The Sports and Pastimes of the People of England.  By Joseph Strutt.  1801.  A new edition, much enlarged and corrected by J. Charles Cox.  Methuen, 1903.  The Preface sketches Strutt's life and says this is based on the 'original' 1801 in quarto, with separate plates which were often hand coloured, but not consistently, while the 1810 reissue had them all done in a terra‑cotta shade.  Hone reissued it in octavo in 1830 with the plates replaced by woodcuts in the text and this was reissued in 1837, 1841 and 1875.  (From above we see that there were other reissues.)  "Mr. Strutt has been left for the most part to speak in his own characteristic fashion ....  A few obvious mistakes and rash conclusions have been corrected, ... certain unimportant omissions have been made.  ...  Nearly a third of the book is new."  Reprinted in 1969 and in the 1960s?

J. T. Micklethwaite.  On the indoor games of school boys in the middle ages.  Archaeological Journal 49 (Dec 1892) 319-328.  Describes various  3 x 3  boards and games on them, including Nine Holes and "tick, tack, toe; or oughts and crosses, which I suppose still survives wherever slate and pencil are used as implements of education", Three Men's Morris and also Nine Men's Morris, Fox and Geese, etc.

Alice B. Gomme.  The Traditional Games of England, Scotland, and Ireland.  2 vols., David Nutt, London, 1894 & 1898.  Reprinted in one vol., Thames & Hudson, London, 1984.

Willard Fiske.  Chess in Iceland and in Icelandic Literature with Historical Notes on Other Table-Games.  The Florentine Typographical Society, Florence, 1905.  Esp. pp. 97-156 of the Stray Notes.  P. 122 lists a number of works on ancient games.

 

          These and the OED have several entries on Noughts and Crosses and Tic‑tac‑toe and many on related games, which are summarised below.  Gomme often cites or quotes Strutt.  The OED often gives the same quotes as Gomme.  Gomme's references are highly abbreviated but full details of the sources can usually be found in the OED.

          (Nine Men's) Morris, where Morris is spelled about 30 different ways, e.g. Marl, Merelles, Mill, Miracles, Morals, and Nine Men's may be given as, e.g. Nine‑peg, Nine Penny, Nine Pin.  Also known as Peg Morris and Shepherd's Mill.  Gomme I 80 & 414‑419 and Strutt 317‑318 (c= Strutt-Cox 256-258 & plate opp. 246, which adds reference to Micklethwaite) are the main entries.  See 4.B.5 for material more specifically on this game.

          Nine Holes, also known as Bubble‑justice, Bumble‑puppy, Crates, and possibly Troll‑madam, Troule‑in‑Madame.  Gomme I 413‑414 and Strutt 274‑275 & 384 (c= Strutt‑Cox 222-223 & 304) are the main entries.  Twelve Holes is similar [Gomme II 321 gives a quote from 1611].  There seem to be cases where Nine Men's Morris was used in referring to Nine Holes [Gomme I 414‑419].  There are two forms of the game:  one form has holes in an upright board that one must roll a ball or marble through;  the other form has holes in the ground, usually in a  3  x  3  array, that one must roll balls into.  Unfortunately, none of the references implies that one has to get three in a row -- see Every Little Boys Book for a version where this is certainly not the case.  There are references going back to 1572 for Crates (but mentioning eleven holes) [Gomme I 81  &  II 309] and 1573 [OED] for Nine Holes.  Botermans et al.; The World of Games; op. cit. in 4.B.5; 1989; p. 213, shows a 17C engraving by Ménian showing Le Jeu de Troumadame as having a board with holes in it, held vertically on a table and one must roll marbles through the holes.  They say it is nowadays known as 'bridge'.

          Three Men's Morris.  This is less common, but occurs in several variant spellings corresponding to the variants of Nine Men's Morris, including, e.g. Three‑penny Morris, Tremerel.  The game is played on a  3  x  3  board and each player has three men.  After making three plays each, consisting of setting men on the cells, further play consists of picking up one of your own men and placing it on a vacant cell, with the object of getting three in a row.  There are several versions of this game, depending on which cells one may play to, but the descriptions given rarely make this clear.  [Gomme I 414‑419] quotes from F. Douce; Illustrations of Shakespeare and of Ancient Manners; 1807, i.184.  "In the French merelles each party had three counters only, which were to be placed in a line to win the game.  It appears to have been the tremerel mentioned in an old fabliau.  See Le Grand, Fabliaux et Contes, ii.208.  Dr. Hyde thinks the morris, or merrils, was known during the time that the Normans continued in possession of England, and that the name was afterwards corrupted into three men's morals, or nine men's morals."  [Hyde.  Hist. Nederluddi [sic], p. 202.]  In practice, the board is often or usually drawn as a crossed square.  If one can move along all winning lines, then it would be natural to draw a doubly crossed square.  See under Alfonso MS (1283) in 4.B.5 for versions called marro, tres en raya and riga di tre.  Again, much of the material on this game is in 4.B.5.

          Five‑penny Morris.  None of the references make it clear, but this seems to be (a form of) Three Men's Morris.  Gomme I 122 and the OED [under Morrell] quote:  W. Hawkins; Apollo Shroving (a play of 1627), act III, scene iv, pp. 48-49. 

          "..., Ovid hath honour'd my exercises.  He describes in verse our boyes play.

          Twise three stones, set in a crossed square where he wins the game 

          That can set his three along in a row, 

          And that is fippeny morrell I trow." 

Most of the references (and myself) are perplexed by the reference to five, though the fact that one has at most five moves in Tic‑tac‑toe might have something to do with it??  Since Three Men's Morris is less well known, some writers have assumed Five‑penny Morris was Nine Men's Morris and others have called all such games by the same name.  A few lines later, Hawkins has: "I challenge him at all games from blowpoint upward to football, and so on to mumchance, and ticketacke.  ... rather than sit out, I will give Apollo three of the nine at Ticketacke, ..."

          Corsicrown [Gomme I 80] seems to be a version of Three Men's Morris, but using seven of the nine cells, omitting two opposite side cells.  Gomme quotes from J. Mactaggart; The Scottish Gallovidian Encyclopedia; (1871 or possibly 1824?):  "each has three men .... there are seven points for these men to move about on, six on the edges of the square and one at the centre."

          Tic‑tac‑toe.  The earliest clearly described versions are given in Babbage (with no name given), c1820, and Gomme I 311, under Kit‑cat‑cannio, where she quotes from:  Edward Moor; Suffolk Words and Phrases; 1823 (This word does not occur in the OED).  Gomme also gives entries for Noughts and Crosses [I 420‑421] and Tip‑tap‑toe [II 295‑296] with variants Tick‑tack‑toe and Tit‑tat‑toe.  In 1842-1865, Babbage uses Tit Tat To and slight variants.  Under Tip‑tap‑toe, Gomme says the players make squares and crosses and that a tie game is a score for Old Nick or Old Tom.  (When I was young, we called it Cat's Game, and this is an old Scottish term [James T. R. Ritchie; The Singing Street  Scottish Children's Games, Rhymes and Sayings; (O&B, 1964); Mercat Press, Edinburgh, 2000, p. 61].)  She quotes regional glossaries for Tip‑tap‑toe (1877), Tit‑tat‑toe (1866 & 1888), Tick‑tack‑toe (1892).  The OED entry for Oughts and Crosses seems to be this game and gives an 1861 quote.  Von der Lasa cites a 1838-39 Swedish book for Tripp, Trapp, Trull.  Van der Linde (1874, op. cit. in 5.F.1) gives Tik, Tak, Tol as the Dutch name. 

          Tit‑tat‑toe [Gomme II 296‑298].  This is a game using a slate marked with a circle and numbered sectors.  The player closes his eyes and taps three times with a pencil and tries to land on a good sector.  Gomme gives the verse:

                    Tit, tat, toe, my first go,

                    Three jolly butcher boys all in a row

                    Stick one up, stick one down,

                    Stick one in the old man's ground.

But cf Games and Sports for Young Boys, 1859, below.

          The OED entries under Tick‑tack, Tip‑tap and Tit give a number of variant spellings and several quotations, which are often clearly to this game, but are sometimes unclear.  Also some forms seem to refer to backgammon.

          In her 'Memoir on the study of children's games' [Gomme II 472‑473], Gomme gives a somewhat Victorian explanation of the origin of Old Nick as the winner of a tie game as stemming from "the primitive custom of assigning a certain proportion of the crops or pieces of land to the devil, or other earth spirit."

 

Franco Agostini & Nicola Alberto De Carlo.  Intelligence Games.  (As:  Giochi della Intelligenza; Mondadori, Milan, 1985.)  Simon & Schuster, NY, 1987.  P. 81 says examples of boards were discovered in the lowest level of Troy and in the Bronze Age tombs in Co. Wicklow, Ireland.  Their description is a bit vague but indicates that the Italian version of Tic-tac-toe is actually Three Men's Morris.

Anonymous.  Play the game.  Guardian Education section (21 Sep 1993) 18-19.  Shows a stone board with the  #  incised on it 'from Bet Shamesh, Israel, 2000 BC'.  This might be the same as the first board below??

A small exhibition of board games organized by Irving Finkel at the British Museum, 1991, displayed the following.

Stone slab with the usual  #  Tick-Tac-Toe board incised on it, but really  a  4 x 3  board.  With nine stone men.  From Giza, >-850.  BM items EA 14315 & 14309, donated by W. M. Flinders Petrie.  Now on display in Room 63, Case C.

Stone Nine Holes board from the Temple of Artemis, Ephesus, 2C-4C.  Item BM GR 1873.5.5.150.  This is a  3  x  3  array of depressions.  Now on display in Room 69, Case 9.

Robbie Bell & Michael Cornelius.  Board Games Round the World.  CUP, 1988.  P. 6 states that the crossed square board has been found at Kurna (c-1400) and at the Ptolemaic temple at Komombo (c-300).  They state that Three Men's Morris is the game mentioned by Ovid in Ars Amatoria.  They say that it was known to the Chinese at the time of Confucius (c-500) under the name of Yih, but is now known as  Luk tsut k'i.  They also say the game is also known as Nine Holes -- which seems wrong to me.

The Spanish Treatise on Chess-Play written by order of King Alfonso the Sage in the year 1283.  [= Libro de Acedrex, Dados e Tablas of Alfonso El Sabio, generally known as the Alfonso MS.]  MS in Royal Library of the Escorial (j.T.6. fol).  Complete reproduction in 194 Phototypic Plates.  2 vols.,  Karl W. Hirsemann, Leipzig, 1913.  (There was also an edition by Arnald Steiger, Geneva, 1941.)  See 4.B.5 for more details of this work.  Vol. 2, f. 93v, p. CLXXXVI, shows a doubly crossed square board.  ??NX -- need to study text.

Pieter Bruegel (the Elder).  Children's Games.  Painting dated 1560 at the Kunsthistorisches Museum, Vienna.  In the right background, children are playing a game involving throwing balls into holes in the ground, but the holes appear to be in a straight line.

Anonymous.  Games of the 16th Century.  1950.  Op. cit. in 4.A.3.  P. 134 describes nine-holes, quoting an unknown poet of 1611: "To play at loggats, Nine-holes, or Ten-pinnes".  The author doesn't specify what positions the balls are to be rolled into.  P. 152 describes Troll-my-dames or Troule-in-madame: "they may have in the end of a bench eleven holes made, into which to troll pummets, or bowls of lead, ...."

William Wordsworth.  The Prelude, Book 1.  Completed 1805, published 1850.  Lines 509‑513.

                    At evening, when with pencil, and smooth slate

                    In square divisions parcelled out and all

                    With crosses and with cyphers scribbled o'er,

                    We schemed and puzzled, head opposed to head

                    In strife too humble to be named in verse.

          It is not clear if this is referring to Noughts and Crosses.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  F. 4r is part of the Table of Contents.  It shows Noughts and Crosses games played on the  #  board and on a  4 x 4  board adjacent to entry 4: The Mill.  Ff. 124-146 are:  Essay 10 -- Of questions requiring the invention of new modes of analysis.  On f. 128.r, he refers to a game in which "the relative positions of three of the marks is the object of inquiry."  Though the reference is incomplete, a Noughts and Crosses game is drawn on the facing page, f. 127.v.  Ff. 134-144 are:  Essay 10 Part 5.  At the top of f. 134.r, he has added a note:  "This is probably my earliest Note on Games of Skill.  I do not recollect the date.  3 March 1865".  The Essay begins:  "Amongst the simplest of those games requiring any degree of skill which amuse our early years is one which is played at in the following manner."  He then describes the game in detail and makes some simple analysis, but he never uses a name for it. 

Charles Babbage.  Notebooks -- unpublished collection of MSS in the BM as Add. MS 37205.  ??NX.  On f. 304, he starts on analysis of games.  Ff. 310-383 are almost entirely devoted to Tit-Tat-To, with some general discussions.  Most of this material comprises a few sheets of working, carefully dated, sometimes amended and with the date of the amendment.  A number of sheets describe parts of the automaton that he was planning to build which would play the game, but no such machine was built until 1949.  The sheets are not always in strict chronological order.

                    F. 310.r is the first discussion of the game, called Tit Tat To, dated 17 Sep 1842.  On F. 312.r, 20 Sep 1843, he says he has "Reduced the 3024 cases D to 199 which include many Duplicates by Symmetry."  F. 321.r, 10 Sep 1860, is the beginning of a summary of his work on games of skill in general.  He refers to Tit-tat-too.  F. 322.r continues and he says:  "I have found no game of skill more simple that that which children often play and which they call Tit‑tat-to."  F. 324-333, Oct 1844, studies "General laws for all games of Skill between two players" and draws flow charts showing the basic recursive analysis of a game tree (ff. 325.v & 325.r).  On f. 332, he counts the number of positions as  9! + 8! + ... + 1!  =  409,113.  F. 333 has an idea of the tree structure of a game.  On ff. 337-338, 8 Sep 1848, he has Tit-tat too.  On ff. 347.r-347.v, he suggests Nine Men's Morris boards in triangular and pentagonal shapes and does various counting on the different shapes.  On ff. 348-349, 26 Oct 1859, he uses Tit-Tat-To.

John M. Dubbey.  The Mathematical Work of Charles Babbage.  CUP, 1978, pp. 96‑97 & 125‑130.  He discusses the above Babbage material.  On p. 127, Dubbey has:  "After a surprisingly lengthy explanation of the rules, he attempts a mathematical formulation.  The basic problem is one that appears not to have been previously considered in the history of mathematics."  Babbage represents the game using roots of unity.  Dubbey, on p. 129, says:  "This analysis ... must count as the first recorded stochastic process in the history of mathematics."  However, it is really a deterministic two-person game.

Games and Sports for Young Boys.  Routledge, nd [1859 - BLC].  P. 70, under Rhymes and Calls:  "In the game of Tit-tat-toe, which is played by very young boys with slate and pencil, this jingle is used:--

                    Tit, tat, toe, my first go:

                    Three jolly butcher boys all in a row;

                    Stick one up, stick one down.

                    Stick one on the old man's crown."

Baron Tassilo von Heydebrand und von der Lasa.  Ueber die griechischen und römischen Spiele, welche einige ähnlichkeit mit dem Schach hatten.  Deutsche Schachzeitung (1863) 162-172, 198-199, 225-234, 257-264.  ??NYS -- described on Fiske 121-122 & 137, who says van der Linde I 40-47 copies much of it.  Von der Lasa asserts that the Parva Tabella of Ovid is Kleine Mühle (Three Men's Morris).  He says the game is called Tripp, Trapp, Trull in the Swedish book Hand-Bibliothek för Sällkapsnöjen, of 1839, vol. II, p. 65 (or 57) -- ??NYS.  Van der Linde says that the Dutch name is Tik, Tak, Tol.  Fiske notes that both of these refer to Noughts and Crosses, but it is unclear if von der Lasa or van der Linde recognised the difference between Three Men's Morris and Noughts and Crosses.

C. Babbage.  Passages from the Life of a Philosopher.  1864.  Chapter XXXIV -- section on Games of Skill, pp. 465‑471.  (= pp. 152‑156 in:  Charles Babbage and His Calculating Engines, Dover, 1961.)  Partial analysis.  He calls it tit‑tat‑to.

The Play Room: or, In-door Games for Boys and Girls.  Dick & Fitzgerald(?), 1866.  [Reprinted as: How to Amuse an Evening Party.  Dick & Fitzgerald, NY, 1869.]  ??NX -- the 1869 was seen at Shortz's.  P. 22: Tit-tat-to.  Uses  O  and  +.  "This is a game that small boys enjoy, and some big ones who won't own it."

Anonymous.  Every Little Boy's Book  A Complete Cyclopædia of in and outdoor games with and without toys, domestic pets, conjuring, shows, riddles, etc.  With two hundred and fifty illustrations.  Routledge, London, nd.  HPL gives c1850, but the text is clearly derived from Every Boy's Book, whose first edition was 1856.  But the main part of the text considered here is not in the 1856 ed of Every Boy's Book (with J. G. Wood as unnamed editor), but is in the 8th ed of 1868 (published for Christmas 1867), which was the first seriously revised edition, with Edmund Routledge as editor.  So this may be c1868.  This is the first published use of the term Noughts and Crosses found so far -- the OED's 1861 quote is to Oughts and Crosses..

                    Pp. 46-47: Slate games: Noughts and crosses.  "This is a capital game, and one which every school-boy truly enjoys."  Though the example shown is a draw, there is no mention of the fact that the game should always be a tie.

                    Pp. 85-86: Nine-holes.  This has nine holes in a row and each player has a hole.  The ball is rolled to them and the person in whose hole it lands must run and pick up the ball and try to hit one of the others who are running away.  So this has nothing to do with our games or other forms of Nine Holes.

                    P. 106: Nine-holes or Bridge-board.  This has nine holes in an upright board and the object is get one's marbles through the holes.  (This material is in the 1856 ed. of Every Boy's Book.)

Correspondent to Notes and Queries (1875)  ??NYS -- quoted by Strutt-Cox 257.  Describes a game called Three Mans' Marriage [sic] in Derbyshire which seems to be Noughts and Crosses played on a crossed square board.  Pieces are not described as moving, but in the next description of a Nine Men's Morris, they are specifically described as moving.  However, the use of a crossed square board may indicate that diagonals were not considered.

Cassell's.  1881.  Slate Games: Noughts and Crosses, or Tit‑Tat‑To, p. 84, with cross reference under Tit-Tat-To, p. 87.  = Manson, 1911, pp. 202-203 & 208.

Albert Norman.  Ungdomens Bok [Book for Youth] (in Swedish).  2nd ed., Stockholm, 1883.  Vol. I, p. 162++.  ??NYS -- quoted and described in Fiske 134-136.  Description of Tripp, Trapp, Trull, with winning cry:  "Tripp, trapp, trull, min qvarn är full."  (Qvarn = mill.)

Lucas.  RM2, 1883.  Pp. 73-99.  Analysis of Three Men's Morris, on a board with the main diagonals drawn, with moves of only one square along a winning line.  He shows this is a first person game.  If the first player is not permitted to play in the centre, then it is a tie game.  No mention of Tic-Tac-Toe.

Albert Ellery Berg, ed.  The Universal Self‑Instructor.  Thomas Kelly, NY, 1883.  Tit‑tat‑to, p. 379.  Brief description.

Mark Twain.  The Adventures of Huckleberry Finn.  1884.  Chap. XXXIV, about half-way through.  "It's as simple as tit-tat-toe, three-in-a-row, ..., Huck Finn."

"A wrangler and late master at Harrow school."  The science of naughts and crosses.  Boy's Own Paper 10:  (No. 498) (28 Jul 1888) 702‑703;  (No. 499) (4 Aug 1888) 717;  (No. 500) (11 Aug 1888) 735;  (No. 501) (18 Aug 1888) 743.  Exhaustive analysis, including odds of second player making a correct response to each opening.  For first move in:  middle, side, corner, the odds of a correct response are:  1/2,  1/2,  1/8.  He implies that the analysis is not widely known.

"Tom Wilson".  Illustred Spelbok (in Swedish).  Nd [late 1880s??].  ??NYS -- described by Fiske 136-137.  This gives Tripp, Trapp, Trull as a Three Men's Morris game on the crossed square, with moves "according to one way of playing, to whatever points they please, but according to another, only to the nearest point along the lines on which the pieces stand.  This last method is always employed when the board has, in addition to the right lines, or lines joining the middles of the exterior lines, also diagonals connecting the angles".  He then describes a drawn version using the  #  board and  0  and  +  (or  1  and  2  in the North) which seems to be genuinely Noughts and Crosses.  Fiske says the book seems to be based on an early edition of the Encyclopédie des Jeux or a similar book, so it is uncertain how much the above represents the current Swedish game.  Fiske was unable to determine the author's real name, though he was still living in Stockholm at the time.

Il Libro del Giuochi.  Florence, 1894.  ??NYS -- described in Fiske, pp. 109-110.  Gives doubly crossed square board and mentions a Three Men's Morris game.

T. de Moulidars.  Grande Encyclopédie des Jeux.  Montgredien or Librairie Illustree, Paris, nd.  ??NYS -- Fiske 115 (in 1905) says it appeared 'not very long ago' and that Gelli seems to be based on it.  Fiske quotes the clear description of Three Men's Morris as Marelle Simple, using a doubly crossed square, saying that pieces move to adjoining cells, following a line, and that the first player should win if he plays in the centre.  Fiske notes that Noughts and Crosses is not mentioned.

J. Gelli. Come Posso Divertirmi?  Milan, 1900.  ??NYS -- described in Fiske 107.  Fiske quotes the description of Three Men's Morris as Mulinello Semplice, essentially a translation from Moulidars.

Dudeney.  CP.  1907.  Prob. 109: Noughts and crosses, pp. 156 & 248.  (c= MP, prob. 202: Noughts and crosses, pp. 89 & 175‑176.  = 536, prob. 471: Tic tac toe, pp. 185 & 390‑392.  Asserts the game is a tie, but gives only a sketchy analysis.  MP gives a reasonably exhaustive analysis.  Looks at Ovid's game.

A. C. White.  Tit‑tat‑toe.  British Chess Magazine (Jul 1919) 217‑220.  Attempt at a complete analysis, but has a mistake.  See Gardner, SA (Mar 1957)  = 1st Book, chap. 4.

D'Arcy Wentworth Thompson.  Science and the Classics.  OUP, 1940.  Section V  Games and Playthings, pp. 148-165.  On p. 160, he quotes Ovid and says it is Noughts and Crosses, or in Ireland, Tip-top-castle.

The Home Book of Quizzes, Games and Jokes.  Blue Ribbon Books, NY, 1941.  This is excerpted from several books -- this material is most likely taken from:  Clement Wood & Gloria Goddard; Complete Book of Games; same publisher, nd [late 1930s].  P. 150: Tit-tat-toe, noughts and crosses.  Brief description of the game on the  #  board.  "To win requires great ingenuity."

G. E. Felton  &  R. H. Macmillan.  Noughts and Crosses.  Eureka 11 (Jan 1949) 5-9.  Mentions Dudeney's work on the  3 x 3  board and asks for generalizations.  Mentions pegotty = go-bang.  Then studies the  4 x 4 x 4  game -- see 4.B.1.a.  Adds some remarks on pegotty, citing Falkener, Lucas and Tarry.

Stanley Byard.  Robots which play games.  Penguin Science News 16 (Jun 1950) 65-77.  On p. 73, he says D. W. Davies, of the National Physical Laboratory, has built, and exhibited to the Royal Society in May 1949, an electro-mechanical noughts and crosses machine.  A photo of the machine is plate 16.  He also mentions Babbage's interest in such a machine and an 1874 paper to the US National Academy by a Dr. Rogers -- ??NYS.

P. C. Parks.  Building a noughts and crosses machine.  Eureka 14 (Oct 1951) 15-17.  Cites Babbage, Rogers, Davies, Byard.  Parks built a simple machine with wire and tin cans in 1950 at a cost of about £6.  Says G. Eastell of Thetford, Norfolk, built a machine using sixty valves for the Festival of Britain.

Gardner.  Ticktacktoe.  SA (Mar 1957)  c= 1st Book, chap. 4.  Quotes Wordsworth, discusses Three Men's Morris (citing Ovid) and its variants (including versions on  4 x 4  and  5 x 5  boards), the misère version (the person who makes three in a row loses), three and  n  dimensional forms (giving L. Moser's result on the number of winning lines on a  kn  board), go-moku, Babbage's proposed machine, A. C. White's article.  Addendum mentions the Opies' assertion that the name comes from the rhyme starting  "Tit, tat, toe,  My first go,".

C. L. Stong.  The Amateur Scientist.  Ill. by Roger Hayward.  S&S, 1960.  A ticktacktoe machine, pp. 384-385.  Noel Elliott gives a brief description of a relay logic machine to play the game.

Donald Michie.  Trial and error.  Penguin Science Survey 2 (1961) 129-145.  ??NYS.  Describes his matchbox and bead learning machine, MENACE (Matchbox educable noughts and crosses engine), for the game.

Gardner.  A matchbox game-learning machine.  SA (Mar 1962)  c= Unexpected, chap. 8.  Describes Michie's MENACE.  Says it used 300 matchboxes.  Gardner adapts it to Hexapawn, which is much simpler, requiring only 24 matchboxes.  Discusses other games playable by 'computers'.  Addendum discusses results sent in by readers including other games.

Barnard.  50 Observer Brain-Twisters.  1962.  Prob. 34: Noughts and crosses, pp. 39‑40, 64 & 93‑94.  Asserts there are  1884  final winning positions.  He doesn't consider equivalence by symmetry and he allows either player to start.

Donald Michie  &  R. A. Chambers.  Boxes: an experiment in adaptive control.  Machine Intelligence 2 (1968) 136-152.  Discusses MENACE, with photo of the pile of boxes.  Says there are 288 boxes, but doesn't explain exactly how he found them.  Chambers has implemented MENACE as a general game-learning computer program using adaptive control techniques designed by Michie.  Results for various games are given. 

S. Sivasankaranarayana Pillai.  A pastime common among South Indian school children.  In:  P. K. Srinivasan, ed.; Ramanujan Memorial Volumes:  1: Ramanujan -- Letters and Reminiscences;  2: Ramanujan -- An Inspiration;  Muthialpet High School, Number Friends Society, Old Boys' Committee, Madras, 1968.  Vol. 2, pp. 81-85.  [Taken from Mathematics Student, but no date or details given -- ??]  Shows ordinary tic-tac-toe is a draw and considers trying to get  t  in a row on an  n x n  board.  Shows that  n = t ³ 3  is a draw and that if  t ³ n + 1 - Ö(n/6),  then the game is a draw.

L. A. Graham.  The Surprise Attack in Mathematical Problems.  Dover, 1968.  Tic-tac-toe for gamblers, prob. 8, pp. 19-22.  Proposed by F. E. Clark, solutions by Robert A. Harrington & Robert E. Corby.  Find the probability of the first player winning if the game is played at random.  Two detailed analyses shows that the probabilities for  first player,  second player,  tie  are  (737,  363,  160)/1260.

[Henry] Joseph & Lenore Scott.  Quiz Bizz.  Puzzles for Everyone -- Vol. 6.  Ace Books (Charter Communications), NY, 1975.  P. 143: Ha-ho-ha.  Misère version of noughts and crosses proposed.  No discussion.

Gyles Brandreth.  Pencil and Paper Games and Puzzles.  Carousel, 1976.  Noughts and Crosses, pp. 11-12.  Asserts "It's been played all around the world for hundreds, if not thousands, of years ...."  I've included it as a typical example of popular belief about the game.  = Pencil & Paper Puzzle Games; Watermill Press, Mahwah, New Jersey, 1989, Tic-Tac-Toe, pp. 11‑12.

Winning Ways.  1982.  Pp. 667-680.  Complete and careful analysis, including various uncommon traps.  Several equivalent games.  Discusses extensions of board size and dimension.

Sheila Anne Barry.  The World's Best Travel Games.  Sterling, NY, 1987.  Tic-tac-toe squared, pp. 88-89.  Get  3  in a row on the  4 x 4  board.  Also considers Tic-tac-toe-toe -- get  4  in a row on  5 x 5  board.

George Markovsky.  Numerical tic-tac-toe -- I.  JRM 22:2 (1990) 114-123.  Describes and studies two versions where the moves are numbered 1, 2, ....  One is due to Ron Graham, the other to P. H. Nygaard and Markowsky sketches the histories.

Ira Rosenholtz.  Solving some variations on a variant of tic-tac-toe using invariant subsets.  JRM 25:2 (1993) 128-135.  The basic variant is to avoid making three in a row on a  4 x 4  board.  By playing symmetrically, the second player avoids losing and 252 of the 256 centrally symmetric positions give a win for the second player.  Extends analysis to  2n x 2n,  5 x 5,  4 x 4 x 4,  etc.

Bernhard Wiezorke.  Sliding caution.  CFF 32 (Aug 1993) 24-25  &  33 (Feb 1994) 32.  This describes a sliding piece puzzle on the doubly crossed square board -- see under 5.A. 

See: Yuri I. Averbakh; Board games and real events; 1995; in 5.R.5, for a possible connection.

 

          4.B.1.a                   IN HIGHER DIMENSIONS

 

C. Planck.  Four‑fold magics.  Part 2 of chap. XIV, pp. 363‑375, of W. S. Andrews, et al.; Magic Squares and Cubes; 2nd ed., Open Court, 1917;  Dover, 1960.  On p. 370, he notes that the number of  m‑dimensional directions through a cell of the  n‑dimensional board is the  m‑th term of the binomial expansion of  ½(1+2)n.

Maurice Wilkes says he played  3-D noughts and crosses at Cambridge in the late 1930s, but the game was to get the most lines on a  3 x 3 x 3  board.  I recall seeing a commercial version, called Plato?, of this in 1970. 

Cedric Smith says he played 3-D and 4-D versions at Cambridge in the early 1940s.

Arthur Stone (letter to me of 9 Aug 1985) says '3 and 4 dimensional forms of tic-tac-toe produced by Brooks, Smith, Tutte and myself', but it's not quite clear if they invented these.  Tutte became expert on the 43 board and thought it was a first person game.  They only played the 54 game once - it took a long time.

Funkenbusch & Eagle, National Mathematics Mag. (1944) ??NYR.

G. E. Felton  &  R. H. Macmillan.  Noughts and crosses.  Eureka 11 (1949) 5‑9.  They say they first met the  4 x 4 x 4  game at Cambridge in 1940 and they give some analysis of it, with tactics and problems.

William Funkenbusch & Edwin Eagle.  Hyper‑spacial tit‑tat‑toe or tit‑tat‑toe in four dimensions.  National Mathematics Magazine 19:3 (Dec 1944) 119‑122.  ??NYR

A. L. Rubinoff, proposer;  L. Moser, solver.  Problem E773 -- Noughts and crosses.  AMM 54 (1947) 281  &  55 (1948) 99.  Number of winning lines on a  kn  board is  {(k+2)n ‑ kn}/2.  Putting  k = 1  gives Planck's result.

L. Buxton.  Four dimensions for the fourth form.  MG 26 (1964) 38‑39.  3 x 3 x 3  and  3 x 3 x 3 x 3  games are obviously first person, but he proposes playing for most lines and with the centre blocked on the  3 x 3 x 3 x 3  board.  Suggests  3n  and  4 x 4 x 4  games.

Anon.  Puzzle page: Noughts and crosses.  MTg 33 (1965) 35.  Says practice shows that the  4 x 4 x 4  game is a draw.  [I only ever had one drawn game!]  Conjectures  nn  is first player and  (n+1)n  is a draw.

Roland Silver.  The group of automorphisms of the game of 3‑dimensional ticktacktoe.  AMM 74 (1967) 247‑254.  Finds the group of permutations of cells that preserve winning lines is generated by the rigid motions of the cube and certain 'eviscerations'.  [It is believed that this is true for the  kn  board, but I don't know of a simple proof.]

Ross Honsberger.  Mathematical Morsels.  MAA, 1978.  Prob. 13: X's and O's, p. 26.  Obtains L. Moser's result.

Kathleen Ollerenshaw.  Presidential Address: The magic of mathematics.  Bull. Inst. Math. Appl. 15:1 (Jan 1979) 2-12.  P. 6 discusses my rediscovery of L. Moser's 1948 result.

Paul Taylor.  Counting lines and planes in generalised noughts and crosses.  MG 63 (No. 424) (Jun 1979) 77-82.  Determines the number  pr(k)  of  r-sections of a  kn  board by means of a recurrence   pr(k)  =  [pr-1(k+2) - pr-1(k)]/2r   which generalises L. Moser's 1948 result.  He then gets an explicit sum for it.  Studies some other relationships.  This work was done while he was a sixth form student.

Oren Patashnik.  Qubic:  4 x 4 x 4  tic‑tac‑toe.  MM 53 (1980) 202‑216.  Computer assisted proof that  4 x 4 x 4  game is a first player win.

Winning Ways.  1982.  Pp. 673-679, esp. 678-679.  Discusses getting  k  in a row on a  n x n  board.  Discusses  43  game (Tic-Toc-Tac-Toe) and  kn  game.

Victor Serebriakoff.  A Mensa Puzzle Book.  Muller, London, 1982.  (Later combined with A Second Mensa Puzzle Book, 1985, Muller, London, as: The Mensa Puzzle Book, Treasure Press, London, 1991.)  Chapter 7: Conceptual conflict in multi-dimensional space, pp. 80-94 (1991: 98-112) & answers on pp. 99, 100, 106 & 131 (1991: 115, 116, 122 & 147).  He considers various higher dimensional noughts and crosses on the  33,  34  and  35  boards.  He finds that there are  49  winning lines on the  33  and he finds how to determine the number of  d-facets on an  n-cube  as the coefficients in the expansion of  (2x + 1)n.  He also considers games where one has to complete a  3 x 3  plane to win and gives a problem:  OXO three hypercube planes, p. 91 (1991: 109) & Answer 29, p. 106 (1991: 122)  which asks for the number of planes in the hypercube  34.  The answer says there are  123  of them, but in 1985 I found  154  and the general formula for the number of  d-sections of a  kn  board.  When I wrote to Serebriakoff, he responded that he could not follow the mathematics and that "I arrived at the figures ... from a simple formula published in one of Art [sic] Gardner's books which checked out as far as I could take it.  Several other mathematicians have looked through it and not disagreed."  I wrote for a reference to Gardner but never had a response.  I presented my work to the British Mathematical Colloquium at Cambridge on 2 Apr 1985 and discovered that the results were known -- I had found the explicit sum given by Taylor above, but not the recurrence.

 

          4.B.2. HEX

 

David Fielker sent some pages from a Danish book on games, but the TP is not present in his copies, so we don't have details.  This says that Hein introduced the game in a lecture to students at the Institute for Theoretical Physics (now the Niels Bohr Institute) in Copenhagen in 1942.  After its appearance in Politiken, specially printed pads for playing the game were sold, and a game board was marketed in the US as Hex in 1952.

Piet Hein.  Article or column in Politiken (Copenhagen) (26 Dec 1942).  ??NYR, but the diagrams show a board of hexagons.

Gardner (1957) and others have related that the game was independently invented by John Nash at Princeton in 1948-1949.  Gardner had considerable correspondence after his article which I have examined.  The key point is that one of Niels Bohr's sons, who had known the game in Copenhagen, was a visitor at the Institute for Advanced Study at the time and showed it to friends.  I concluded that it was likely that some idea of the game had permeated to Nash who had forgotten this and later recalled and extensively developed the idea, thinking it was new to him.  I met Harold Kuhn in 1998, who was a student with Nash at the time and he has no doubt that Nash invented the idea.  In particular, Nash started with the triangular lattice, i.e. the dual of Hein's board, for some time before realising the convenience of the hexagonal lattice.  Nash came to Princeton as a graduate student in autumn 1948 and had invented the game by the spring of 1949.  Kuhn says he observed Nash developing the ideas and recognising the connections with the Jordan Curve Theorem, etc.  Kuhn also says that there was not much connection between students at Princeton and at the Institute and relates that von Neumann saw the game at Princeton and asked what it was, indicating that it was not well known at the Institute.  In view of this, it seems most likely that Nash's invention was independent, but I know from my own experience that it can be difficult to remember the sources of one's ideas -- a casual remark about a hexagonal game could have re-emerged weeks or months later when Nash was studying games, as the idea of looking at hexagonal boards in some form, from which the game would be re-invented.  Sylvester was notorious for publishing ideas which he had actually refereed or edited some years earlier, but had completely forgotten the earlier sources.  In situations like Hex, we will never know exactly what happened -- even if we were present at the time, it is difficult to know what is going on in the mind of the protagonist and the protagonist himself may not know what subconscious connections his mind is making.  Even if we could discover that Nash had been told something about a hexagonal game, we cannot tell how his mind dealt with this information and we cannot assume this was what inspired his work.  In other words, even a time machine will not settle such historical questions -- we need something that displays the conscious and the unconscious workings of a person's mind.

Parker Brothers.  Literature on Hex, 1952.  ??NYS or NYR.

Claude E. Shannon.  Computers and automata.  Proc. Institute of Radio Engineers 41 (Oct 1953) 1234‑1241.  Describes his Hex machine on p. 1237.

M. Gardner.  The game of Hex.  SA (Jul 1957) = 1st Book, chap. 8.  Description of Shannon's 8 by 7 'Hoax' machine, pp. 81‑82, and its second person strategy, p. 79.

Anatole Beck, Michael N. Bleicher & Donald W. Crowe.  Excursions into Mathematics.  Worth Publishers, NY, 1969.  Chap. 5: Games (by Beck), Section 3: The game of Hex, pp. 327-339 (with photo of Hein on p. 328).  Says it has been attributed to Hein and Nash.  At Yale in 1952, they played on a  14 x 14  board.  Shows it is a first player win, invoking the Jordan Curve Theorem

David Gale.  The game of Hex and the Brouwer fixed-point theorem.  AMM 86:10 (Dec 1979) 818-827.  Shows that the non-existence of ties (Hex Theorem) is equivalent to the Brouwer Fixed-Point Theorem in two and in  n  dimensions.  Says the use of the Jordan Curve Theorem is unnecessary.

Winning Ways.  1982.  Pp. 679-680 sketches the game and the strategy stealing argument which is attributed to Nash.

C. E. Shannon.  Photo of his Hoax machine sent to me in 1983.

Cameron Browne.  Hex Strategy: Making the Right Connections.  A. K. Peters, Natick, Massachusetts, 2000.

 

          4.B.3. DOTS AND BOXES

 

Lucas.  Le jeu de l'École Polytechnique.  RM2, 1883, pp. 90‑91.  He gives a brief description, starting: "Depuis quelques années, les élèves de l'École Polytechnique ont imaginé un nouveaux jeu de combinaison assez original."  He clearly describes drawing the edges of the game board and that the completer of a box gets to go again.  He concludes: "Ce jeu nous a paru assez curieux pour en donner ici la description; mais, jusqu'a présent, nous ne connaissons pas encore d'observations ni de remarques assez importantes pour en dire davantage."

Lucas.  Nouveaux jeux scientifiques de M. Édouard Lucas.  La Nature 17 (1889) 301‑303.  Clearly describes a game version of La Pipopipette on p. 302, picture on p. 301, "... un nouveau jeu ... dédié aux élèves de l'école Polytechnique."  This is dots and boxes with the outer edges already drawn in.

Lucas.  L'Arithmétique Amusante.  1895.  Note III:  Les jeux scientifiques de Lucas, pp. 203‑209 -- includes his booklet:  La Pipopipette, Nouveau jeu de combinaisons, Dédié aux élèves de l'École Polytechnique, Par un Antique de la promotion de 1861, (1889), on pp. 204‑208.  On p. 207, he says the game was devised by several of his former pupils at the École Polytechnique.  On p. 37, he remarks that "Pipo est la désignation abrégée de Polytechnique, par les élèves de l'X, ...."

Robert Marquard & Georg Frieckert.  German Patent 108,830 -- Gesellschaftsspiel.  Patented: 15 Jun 1899.  1p + 1p diagrams.  8 x 8  array of boxes on a board with slots for inserting edges.  No indication that the player who completes a box gets to play again.  They have some squares with values but also allow all squares to have equal value.

C. Ganse.  The dot game.  Ladies' Home Journal (Jun 1903) 41.  Describes the game and states that one who makes a box gets to go again.

Loyd.  The boxer's puzzle.  Cyclopedia, 1914, pp. 104 & 352.  = MPSL1, prob. 91, pp. 88‑89 & 152‑153.  c= SLAHP: Oriental tit‑tat‑toe, pp. 28 & 92‑93.  Loyd doesn't start with the boundaries drawn.  He asserts it is 'from the East'.

Ahrens.  A&N.  1918.  Chap. XIV: Pipopipette, pp. 147‑155, describes it in more detail than Lucas does.  He says the game appeared recently.

Blyth.  Match-Stick Magic.  1921.  Boxes, pp. 84-85.  "The above game is familiar to most boys and girls ...."  No indication that the completer of a box gets to play again.

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.  Pp. 84-85: Die Käsekiste.  Describes a version for two or more players.  The first player must start at a corner and players must always connect to previously drawn lines.  A player who completes a box gets to play again.

Meyer.  Big Fun Book.  1940.  Boxes, p. 661.  Brief description, somewhat vaguely stating that a player who completes a box can play again.

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  P. 151: Dots and squares.  Clearly says the completer gets to play again.  "The game calls for great ingenuity."

"Zodiastar".  Fun with Matches and Match Boxes.  (Cover says:  Match Tricks  From the 1880s to the 1940s.)  Universal Publications, London, nd [late 1940s?].  The game of boxes, pp. 48-49.  Starts by laying out four matches in a square and players put down matches which must touch the previous matches.  Completing a box gives another play.  No indication that matches must be on lattice lines, but perhaps this is intended.

Readers' Research Department.  RMM 2 (Apr 1961) 38‑41, 3 (Jun 1961) 51‑52, 4 (Aug 1961) 52‑55.  On pp. 40‑41 of No. 2, it says that Martin Gardner suggests seeking the best strategy.  Editor notes there are two versions of the rules -- where the one who makes a box gets an extra turn, and where he doesn't -- and that the game can be played on other arrays.  On p. 51 of No. 3, there is a symmetry analysis of the no‑extra‑turn game on a board with an odd number of squares.  On pp. 52‑54 of No. 4, there is some analysis of the extra‑turn case on a board with an odd number of boxes.

Everett V. Jackson.  Dots and cubes.  JRM 6:4 (Fall 1973) 273‑279.  Studies 3‑dimensional game where a play is a square in the cubical lattice.

Gyles Brandreth.  Pencil and Paper Games and Puzzles.  Carousel, 1976.  Worm, pp. 18-19.  This is a sort of 'anti-boxes' -- one draws segments on the lattice forming a path without any cycles -- last player wins.  = Pencil & Paper Puzzle Games; Watermill Press, Mahwah, New Jersey, 1989, pp. 18-19.

Winning Ways.  1982.  Chap. 16: Dots-and-Boxes, pp. 507-550

David B. Lewis.  Eureka!  Perigee (Putnam), NY, 1983.  Pp. 44‑45 suggests playing on the triangular lattice.

Sheila Anne Barry.  The World's Best Travel Games.  Sterling, NY, 1987.

Eternal triangles, pp. 80-81.  Gives the game on the triangular lattice.

Snakes, pp. 81-82.  Same as Brandreth's Worm.  I think 'snake' would be a better title as only one path is drawn.

 

          4.B.4. SPROUTS

 

M. Gardner.  SA (Jul 1967) = Carnival, chap. 1.  Describes Michael Stewart Paterson and John Horton Conway's invention of the game on 21 Feb 1967 at tea time in the Department common room at Cambridge.  The idea of adding a spot was due to Paterson and they agreed the credit for the game should be 60% Paterson to 40% Conway.

Gyles Brandreth.  Pencil and Paper Games and Puzzles.  Carousel, 1976.  Sprouts, p. 13.  "... actually born in Cambridge about ten years ago."  c= Pencil & Paper Puzzle Games; Watermill Press, Mahwah, New Jersey, 1989, p. 13: "... was invented about ten years ago."

Winning Ways.  1982.  Sprouts, pp. 564-570 & 573.  Says the game was "introduced by M. S. Paterson and J. H. Conway some time ago".  Also describes Brussels Sprouts and Stars-and-Stripes.  An answer for Brussels Sprouts and some references are on p. 573.

Sheila Anne Barry.  The World's Best Travel Games.  Sterling, NY, 1987.  Sprouts, pp. 95-97.

Karl-Heinz Koch.  Pencil & Paper Games.  (As: Spiele mit Papier und Bleistift, no details); translated by Elisabeth E. Reinersmann.  Sterling, NY, 1992.  Sprouts, pp. 36-37, says it was invented by J. H. Conway & M. S. Paterson on 21 Feb 1976 [sic -- misprint of 1967] during their five o'clock tea hour.

 

          4.B.5. OVID'S GAME AND NINE MEN'S MORRIS

 

          See also 4.B.1 for historical material.

          The classic Nine Men's Morris board consists of three concentric squares with their midpoints joined by four lines.  The corners are sometimes also joined by another four diagonal lines, but this seems to be used with twelve men per side and is sometimes called Twelve Men's Morris -- see 1891 below.  Fiske 108 says this is common in America but infrequent in Europe, though on 127 he says both forms were known in England before 1600, and both were carried to the US, though the Nine form is probably older.

 

Murray 615 discusses Nine Men's Morris.  He cites Kurna, Egypt (‑14C), medieval Spain (Alquerque de Nueve), the Gokstad ship and the steps of the Acropolis of Athens.  He says the board sometimes has diagonals added and then is played with 9, 11 or 12 pieces.

Dudeney.  AM. 1917.  Introduction to Moving Counter Problems, pp. 58-59.  This gives a brief survey, mentioning a number of details that I have not seen elsewhere, e.g. its occurrence in Poland and on the Amazon.  Says the board was found on a Roman tile at Silchester and on the steps of the Acropolis in Athens among other sites.

J. A. Cuddon.  The Macmillan Dictionary of Sports and Games.  Macmillan, London, 1980.  Pp. 563‑564.  Discusses the history.  Says there is a c‑1400 board cut in stone at Kurna, Egypt and similar boards were made in years 9 to 21 at Mihintale, Ceylon.  Says Ars Amatoria may be describing Three Men's Morris and Tristia may be describing a kind of Tic‑tac‑toe.  Cites numerous medieval descriptions and variants.

Claudia Zaslavsky.  Tic Tac Toe and Other Three‑in‑a‑Row Games from Ancient Egypt to the Modern Computer.  Crowell, NY, 1982.  This is really a book for children and there are no references for the historical statements.  I have found most of them elsewhere,  and the author has kindly send me a list of source books, but I have not yet tracked down the following items -- ??.

                    There is an English court record of 1699 of punishment for playing Nine Holes in church.

                    There is a Nine Men's Morris board on a stone on the temple of Seti I (presumably this is at Kurna).  There is a picture in the 13C Spanish 'Book of Games' (presumably the Alfonso MS -- see below) of children playing Alquerque de Tres (c= Three Men's Morris).  A 14C inventory of the Duc de Berry lists tables for Mérelles (=? Nine Men's Morris) (see Fiske 113-115 below) and a book by Petrarch shows two apes playing the game.

 

H. Parker.  Ancient Ceylon.  Loc. cit. in 4.B.1.  Nine Men's Morris board in the Temple of Kurna, Egypt, ‑14C.  [Rohrbough, below, says this temple was started by Ramses I and completed by Seti in -1336/-1333, citing J. Royal Asiatic Soc. (1783) 17.]  Two diagrams for Nine Men's Morris are cut into the great flight of steps at Mihintale, Ceylon and these are dated c1C.  He cites Bell; Arch. Survey of Ceylon, Third Progress Report, p. 5 note, for another diagram of similar age.

Jack Botermans, Tony Burrett, Pieter van Delft & Carla van Spluntern.  The World of Games.  (In Dutch, 1987);  Facts on File, NY, 1989. 

                    P. 35 describes Yih, a form of Three Men's Morris, played on a doubly crossed square with a man moving "one step along any line".  A note adds that only the French have a rule forbidding the first player to play in the centre, which makes the game more challenging and is recommended.

                    Pp. 103-107 is the beginning of a section:  Games of alignment and configuration and discusses various games, but rather vaguely and without references.  They mention Al-Qurna, Mihintale, Gokstad and some other early sites.  They say Yih was described by Confucius, was played c-500 and is "the game, that we now know as tic-tac-toe, or three men's morris."  They describe Noughts and Crosses in the usual way.  They then distinguish Tic-Tac-Toe, saying "In Britain it is generally known as three men's morris ...." and say it is the same as Yih, "which was known in ancient Egypt".  They say "Ovid mentions tic-tac-toe" in Ars Amatoria, that several Roman boards have survived and that it was very popular in 14C England with several boards for this and Nine Men's Morris cut into cloister seats.  They then describe Three-in-a-Row, which allows pieces to move one step in any direction, as a game played in Egypt.  They then describe Five or Six Men's Morris, Nine Men's Morris, Twelve Men's Morris and Nine Men's Morris with Dice, with nice 13C & 15C illustration of Nine Men's Morris.

Bell & Cornelius.  Board Games Round the World.  Op. cit. in 4.B.1.  1988.  Pp. 6-8.  They discuss the crossed square board -- see 4.B.1 -- and describe Three Men's Morris with moves only along the lines to an adjacent vacant point.  They then describe Achi, from Ghana, on the doubly crossed square with the same rules.  They then describe Six Men's Morris which was apparently popular in medieval Europe but became obsolete by c1600.

Ovid.  Ars Amatoria.  -1.  II, 203-208  &  III, 353-366.  Translated by J. H. Mozley; Loeb Classical Library, 1929, pp. 80-81 & 142-145.  Translated by B. P. Moore, 1935, used in A. D. Melville; Ovid  The Love Poems; OUP, 1990, pp. 113, 137, 229 & 241.

                    II, 203-208 are three couplets apparently referring to three games: two dice games and Ludus Latrunculorum.  Mozley's prose translation is:

           "If she be gaming, and throwing with her hand the ivory dice, do you throw amiss and move your throws amiss; or if is the large dice you are throwing, let no forfeit follow if she lose; see that the ruinous dogs often fall to you; or if the piece be marching under the semblance of a robbers' band, let your warrior fall before his glassy foe."

          'Dogs' is the worst throw in Roman dice games.

                    Moore's verse translation of 207-208 is:

           "And when the raiding chessmen take the field,  Your champion to his crystal foe must yield." 

          Melville's note says the original has 'bandits' and says the game is Ludus Latrunculorum.

                    III, 357-360 is probably a reference to the same game since 'robbers' occurs again, though translated as brigands by Mozley, and again it immediately follows a reference to throwing dice.  Mozley's translation of 353-366 is:

           "I am ashamed to advise in little things, that she should know the throws of the dice, and thy powers, O flung counter.  Now let her throw three dice, and now reflect which side she may fitly join in her cunning, and which challenge,  Let her cautiously and not foolishly play the battle of the brigands, when one piece falls before his double foe and the warrior caught without his mate fights on, and the enemy retraces many a time the path he has begun.  And let smooth balls be flung into the open net, nor must any ball be moved save that which you will take out.  There is a sort of game confined by subtle method into as many lines as the slippery year has months: a small board has three counters on either side, whereon to join your pieces together is to conquer."

          Moore's translation of 357-360 is:

           "To guide with wary skill the chessmen's fight,  When foemen twain o'erpower the single knight,  And caught without his queen the king must face  The foe and oft his eager steps retrace". 

          This is clearly not a morris game -- Mozley's note above and the next entry make it clear it is Ludus Latrunculorum, which had a number of forms.  Mozley's note on pp. 142-143 refers to Tristia II, 478 and cites a number of other references for Ludus Latrunculorum.

                    Moore's translation of 363-366 is:

           "A game there is marked out in slender zones  As many as the fleeting year has moons;  A smaller board with three a side is manned,  And victory's his who first aligns his band." 

          Mozley's notes and Melville's notes say the first two lines refer to the Roman game of Ludus Duodecim Scriptorum -- the Twelve Line Game -- which is the ancestor of Backgammon.  Mozley says the game in the latter two lines is mentioned in Tristia, "but we have no information about it."  Melville says it is "a 'position' game, something like Nine Men's Morris" and cites R. C. Bell's article on 'Board and tile games' in the Encyclopaedia Britannica, 15th ed., Macropaedia ii.1152‑1153, ??NYS.

Ovid.  Tristia.  c10.  II, 471‑484.  Translated by A. L. Wheeler.  Loeb Classical Library, 1945, pp. 88‑91.  This mentions several games and the text parallels that of Ars Amatoria III. 

           "Others have written of the arts of playing at dice -- this was no light sin in the eyes of our ancestors -- what is the value of the tali, with what throw one can make the highest point, avoiding the ruinous dogs; how the tessera is counted, and when the opponent is challenged, how it is fitting to throw, how to move according to the throws; how the variegated soldier steals to the attack along the straight path when the piece between two enemies is lost, and how he understands warfare by pursuit and how to recall the man before him and to retreat in safety not without escort; how a small board is provided with three men on a side and victory lies in keeping one's men abreast; and the other games -- I will not describe them all -- which are wont to waste that precious thing, our time." 

          A note says some see a reference to Ludus Duodecim Scriptorum at the beginning of this.  The next note says the next text refers to Ludus Latrunculorum, a game on a squared board with 30 men on a side, with at least two kinds of men.  The note for the last game says "This game seems to have resembled a game of draughts played with few men." and refers to Ars Amatoria and the German Mühlespiel, which he describes as 'a sort of draughts', but which is Nine Men's Morris.

R. G. Austin.  Roman board games -- I & II.  Greece and Rome  4 (No. 10) (Oct 1934) 24‑34  &  4 (No. 11) (Feb 1935) 76-82.  Claims the Ovid references are to Ludus Latrunculorum (a kind of Draughts?), Ludus Duodecim Scriptorum (later Tabula, an ancestor of Backgammon) and (Ars Amatoria.iii.365-366) a kind of Three Men's Morris.  In the last, he shows a doubly crossed  3 x 3  board, but it is not clear which rule he adopts for the later movement of pieces, but he says: "the first player is always able to force a win if he places his first man on the centre point, and this suggests that the dice may have been used to determine priority of play, although there is no evidence of this."  He says no Roman name for this game has survived.  He discusses various known artifacts for all the game, citing several Roman  8 x 8  boards found in Britain.  He gives an informal bibliography with comments as to the value of the works.

D'Arcy Wentworth Thompson.  Science and the Classics.  OUP, 1940.  Section V  Games and Playthings, pp. 148-165.  On p. 160, he quotes Ovid, Ars Amatoria.iii.365-366 and says it is Noughts and Crosses, or in Ireland, Tip-top-castle.

The British Museum has a Nine Men's Morris board from the Temple of Artemis, Ephesus, 2C-4C.  Item BM GR 1872,8-3,44.  This was in a small exhibition of board games in 1990.  I didn't see it on display in late 1996.

Murray, p. 189.  There was an Arabic game called Qirq, which Murray identifies with Morris.  "Fourteen was a game played with small stones on a wooden board which had three rows of holes (al‑Qâbûnî)."  Abû‑Hanîfa [the  H  should have a dot under it], c750, held that Fourteen was illegal and Qirq was held illegal by writers soon afterward.  On p. 194, Murray gives a 10C passage mentioning Qirq being played at Mecca.

Fiske 255 cites the Kitāb al Aghāni, c960, for a reference to qirkat, i.e. morris boards.

Paul B. Du Chaillu.  The Viking Age.  Two vols., John Murray, London, 1889.  Vol. II, p.168, fig. 992 -- Fragments of wood from Gokstad ship.  Shows a partial board for Nine Men's Morris found in the Gokstad ship burial.  There is no description of this illustration and there is only a vague indication that this is 10C, but other sources say it is c900.

Gutorm Gjessing.  The Viking Ship Finds.  Revised ed., Universitets Oldsaksamling, Oslo, 1957.  P. 8:  "... there are two boards which were used for two kinds of games; on one side figures appear for use in a game which is frequently played even now (known as "Mølle")."

Thorlief Sjøvold.  The Viking Ships in Oslo.  Universitets Oldsaksamling, Oslo, 1979.  P. 54:  "... a gaming board with one antler gaming piece, ...."

In medieval Europe, the game is called Ludus Marellorum or Merellorum or just Marelli or Merelli or Merels, meaning the game of counters.  Murray 399 says the connection with Qirq is unclear.  However, medieval Spain played various games called Alquerque, which is obviously derived from Qirq.  Alquerque de Nueve seems to be Nine Men's Morris.  However, in Italy and in medieval France, Marelle or Merels could mean Alquerque (de Doze), a draughts‑like game with 12 men on a side played on a  5 x 5  board (Murray 615).  Also Marro, Marella can refer to Draughts which seems to originate in Europe somewhat before 1400.

Stewart Culin.  Korean Games, with Notes on the Corresponding Games of China and Japan.  University of Pennsylvania, Philadelphia, 1895.  Reprinted as:  Games of the Orient; Tuttle, Rutland, Vermont, 1958.  Reprinted under the original title, Dover and The Brooklyn Museum, 1991.  P. 102, section 80: Kon-tjil -- merrells.  This is the usual Nine Men's Morris.  The Chinese name is Sám-k'i (Three Chess).  "I am told by a Chinese merchant that this game was invented by Chao Kw'ang-yin (917-975), founder of the Sung dynasty."  This is the only indication of an oriental source that I have seen.

Gerhard Leopold.  Skulptierte Werkstücke in der Krypta der Wipertikirche zu Quedlinburg.  IN: Friedrich Möbius & Ernst Schubert, eds.; Skulptur des Mittelalters; Hermann Böhlaus Nachfolger, Weimar, 1987, pp. 27-43; esp. pp. 37 & 43.  Describes and gives photos of several Nine-Men's-Morris boards carved on a pillar of the crypt of the Wipertikirche, Quedlinburg, Sachsen-Anhalt, probably from the 10/11 C.

Richard de Fournivall.  De Vetula.  13C.  This describes various games, including Merels.  Indeed the French title is:  Ci parle du gieu des Merelles ....  ??NYS -- cited by Murray, pp. 439, 507, 520, 628.  Murray 620 cites several MSS and publications of the text.

"Bonus Socius" [Nicolas de Nicolaï?].  This is a collection of chess problems, compiled c1275, which exists in many manuscript forms and languages.  See 5.F.1 for more details of these MSS.  See Murray 618‑642.  On pp. 619‑624 & 627, he mentions several MSS which include 23, 24, 25 or 28 Merels problems.  On p. 621, he cites "Merelles a Neuf" from 14C.  Fiske 104 & 110-111 discusses some MSS of this collection.

The Spanish Treatise on Chess-Play written by order of King Alfonso the Sage in the year 1283.  [= Libro de Acedrex, Dados e Tablas of Alfonso El Sabio, generally known as the Alfonso MS.]  MS in Royal Library of the Escorial (j.T.6. fol).  Complete reproduction in 194 Phototropic Plates.  2 vols.,  Karl W. Horseman, Leipzig, 1913.  (See in 4.A.1 for another ed.)  This is a collection of chess problems produced for Alfonso X, the Wise, King of Castile (Castilla).  Vol. 2, ff. 92v‑93r, pp. CLXXXIV‑CLXXXV, shows Nine Men's Morris boards.  ??NX -- need to study text.  See:  Murray 568‑573;  van der Linde I 137 & 279 ??NYS  &  Quellenstudien 73 & 277‑278, ??NYS (both cited by Fiske 98);  van der Lasa 116, ??NYS (cited by Fiske 99). 

                    Fiske 98-99 says that the MS also mentions Alquerque, Cercar de Liebre and Alquerque de Neuve (with 12 men against one).  Fiske 253-255 gives a more detailed study of the MS based on a transcript.  He also quotes a communication citing al Querque or al Kirk in Kazirmirski's Arabic dictionary and in the Kitāb al Aghāni, c960.

                    José Brunet y Bellet.  El Ajedrez.  Barcelona, 1890.  ??NYS -- described by Fiske 98.  This has a chapter on the Alfonso MS and refers to Alquerque de Doce, saying that it is known as Tres en Raya in Castilian and Marro in Catalan (Fiske 102 says this word is no longer used in Spanish).  Brunet notes that there are five miniatures pertaining to alquerque.  Fiske says that all this information leaves us uncertain as to what the games were.  Fiske says Brunet's chapter has an appendix dealing with Carrera's 1617 discussion of 'line games' and describing Riga di Tre as the same as Marro or Tres en Raya as a form of Three Men's Morris

Murray gives many brief references to the game, which I will note here simply by his page number and the date of the item. 

                    438‑439 (12C);  446 (14C);

                    449 (c1400 -- 'un marrelier', i.e. a Merels board);

                    431 (c1430);  447 (1491);  446 (1538).

Anon.  Romance of Alexander.  1338.  (Bodleian Library, Mss Bodl. 264).  ??NYS.  Nice illustration clearly showing Nine Men's Morris board.  I. Disraeli (Amenities of Literature, vol. I, p. 86) also cites British Museum, Bib. Reg. 15, E.6 as a prose MS version with illustrations.  Prof. D. J. A. Ross tells me there is nothing in the text corresponding to the illustrations and that the Bodleian text was edited by M. R. James, c1920, ??NYS.  Illustration reproduced in:  A. C. Horth; 101 Games to Make and Play; Batsford, London, (1943;  2nd ed., 1944);  3rd ed., 1946; plate VI facing p. 44, in B&W.  Also in:  Pia Hsiao et al.; Games You Make and Play; Macdonald and Jane's, London, 1975, p. 7, in colour.

Fiske 113-115 gives a number of quotations from medieval French sources as far back as mid 14C, including an inventory of the Duc de Berry in 1416 listing two boards.  Fiske notes that the game has given rise to several French phrases.  He quotes a 1412 source calling it Ludus Sanct Mederici or Jeu Saint Marry and also mentions references in city statutes of 1404 and 1414.

MS, Montpellier, Faculty of Medicine, H279 (Fonts de Boulier, E.93).  14C.  This is a version of the Bonus Socius collection.  Described in Murray 623-624, denoted M, and in van der Linde I 301, denoted K.  Lucas, RM2, 1883, pp. 98-99 mentions it and RM4, 1894, Quatrième Récréation: Le jeu des mérelles au XIIIe siècle, pp. 67-85 discusses it extensively.  This includes 28 Merels problems which are given and analysed by Lucas.  Lucas dates the MS to the 13C.

Household accounts of Edward IV, c1470.  ??NYS -- see Murray 617.  Record of purchase of "two foxis and 46 hounds" to form two sets of "marelles".

Civis Bononiae [Citizen of Bologna].  This is a collection of chess problems compiled c1475, which exists in several MSS.  See Murray 643‑703.  It has 48 or 53 merels problems.  On p. 644, 'merelleorum' is quoted.

A Hundred Sons.  Chinese scroll of Ming period (1368-1644).  18C copy in BM.  ??NYS -- extensively reproduced and described in:  Marguerite Fawdry; Chinese Childhood; Pollock's Toy Theatres, London, 1977.  On p. 12 of Fawdry is a scene, apparently from the scroll, in which some children appear to be playing on a Twelve Men's Morris board.

Elaborate boards from Germany (c1530) and Venice (16C) survive in the National Museum, Munich and in South Kensington (Murray 757‑758).  Murray shows the first in B&W facing p. 757.

William Shakespeare.  A Midsummer Night's Dream.  c1610.  Act II, scene I, lines 98-100:  "The nine men's morris is fill'd up with mud,  And the quaint mazes in the wanton green  For lack of tread are indistinguishable."  Fiske 126 opines that the latter two lines may indicate that the board was made in the turf, though he admits that they may refer just to dancers' tracks, but to me it clearly refers to turf mazes.

J. C. Bulenger.  De Ludis Privatis ac Domesticus Veterum.  Lyons, 1627.  ??NYS  Fiske 115 & 119 quote his description of and philological note on Madrellas (Three Men's Morris).

Paul Fleming (1609-1640).  In one of his lyrics, he has Mühlen.  ??NYS -- quoted by Fiske 132, who says this is the first German mention of Morris. 

Fiske 133 gives the earliest Russian reference to Morris as 1675.

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  Historia Triodii, pp. 202-214, is on morris games.  (Described in Fiske 118-124, who says there is further material in the Elenchus at the end of the volume -- ??NYS)  Hyde asserts that the game was well known to the Romans, though he cannot find a Roman name for it!  He cites and discusses Bulenger, but disagrees with his philology.  Gives lots of names for the game, ranging as far as Russian and Armenian.  He gives both the Nine and Twelve Men's Morris boards on p. 210, but he has not found the Twelve board in Eastern works.  On p. 211, he gives the doubly crossed square board with a title in Chinese characters, pronounced 'Che-lo', meaning 'six places', and having three white and three black men already placed along two sides.  He says the Irish name is Cashlan Gherra (Short Castle) and that the name Copped Crown is common in Cumberland and Westmorland.  He then describes playing the Twelve Man and Nine Man games, and then he considers the game on the doubly crossed square board.  He seems to say there are different rules as to how one can move.  ??need to study the Latin in detail.  This is said to throw light on the Ovid passages.  Hyde believes the game was well known to the Romans and hence must be much older.  Fiske remarks that this is history by guesswork.

Murray 383 describes Russian chess.  He says Amelung identifies the Russian game "saki with Hölzchenspiel (?merels)".  Saki is mentioned on this page as being played at the Tsar's court, c1675.

Archiv der Spiele.  3 volumes, Berlin, 1819-1821.  Vol. 2 (1820) 21-27.  ??NYS  Described and quoted by Fiske 129-132.  This only describes the crossed square and the Nine Men's Morris boards.  It says that the Three Men's Morris on the crossed square board is a tie, i.e. continues without end, but it is not clear how the pieces are allowed to move.  Fiske says this gives the most complete explanation he knows of the rules for Nine Men's Morris.

Charles Babbage.  Notebooks -- unpublished collection of MSS in the BM as Add. MS 37205.  ??NX.  For more details, see 4.B.1.  On ff. 347.r-347.v, 8 Sep 1848, he suggests Nine Men's Morris boards in triangular and pentagonal shapes and does various counting on the different shapes.

The Family Friend (1856) 57.  Puzzle 17. -- Two and a Bushel.  Shows the standard  #  board.  "This very simple and amusing games, -- which we do not remember to have seen described in any book of games, -- is played, like draughts, by two persons with counters.  Each player must have three, ...  and the game is won when one of the players succeeds in placing his three men in a row; ...."  There is no specification of how the men move.  The word 'bushel' occurs in some old descriptions of Three Men's Morris and Nine Men's Morris as the name of the central area.

The Sociable.  1858.  Merelles: or, nine men's morris, pp. 279-280.  Brief description, notable for the use of Merelles in an English book.

Von der Lasa.  Ueber die griechischen und römischen Spiele, welche einige ähnlichkeit mit dem Schach hatten.  Deutsche Schachzeitung (1863) 162-172, 198-199, 225-234, 257‑264.  ??NYS -- described on Fiske 121-122 & 137, who says van der Linde I 40-47 copies much of it.  He asserts that the Parva Tabella of Ovid is Kleine Mühle (Three Men's Morris).  Von der Lasa says the game is called Tripp, Trapp, Trull in the Swedish book Hand-Bibliothek för Sällkapsnöjen, of 1839, vol. II, p. 65 (or 57??).  Van der Linde says that the Dutch name is Tik, Tak, Tol.  Fiske notes that both of these refer to Noughts and Crosses, but it is unclear if von der Lasa or van der Linde recognised the difference between Three Men's Morris and Noughts and Crosses.

Albert Norman.  Ungdomens Bok [Book for Youth] (in Swedish).  2nd ed., Stockholm, 1883.  Vol. I, p. 162++.  ??NYS -- quoted and described in Fiske 134-136.  Plays Nine Men's Morris on a Twelve Men's Morris board.

Webster's Dictionary.  1891.  ??NYS -- Fiske 118 quotes a definition (not clear which) which includes "twelve men's morris".  Fiske says:  "Here we have almost the only, and certainly the first mention of the game by its most common New England name, "twelve men's morris," and also the only hint we have found in print that the more complicated of the morris boards -- with the diagonal lines ... -- is used with twelve men, instead of nine, on each side."  Fiske 127 says the name only appears in American dictionaries.

Dudeney.  CP.  1907.  Prob. 110: Ovid's game, pp. 156‑157 & 248.  Says the game "is distinctly mentioned in the works of Ovid."  He gives Three Men's Morris, with moves to adjacent cells horizontally or diagonally, and says it is a first player win.

Blyth.  Match-Stick Magic.  1921.  Black versus white, pp. 79-80.  4 x 4  board with four men each.  But the men must be initially placed  WBWB  in the first row and  BWBW  in the last row.  They can move one square "in any direction" and the object is to get four in a row of your colour.

Games and Tricks -- to make the Party "Go".  Supplement to "Pearson's Weekly", Nov. 7th, no year indicated [1930s??].  A matchstick game, p. 11.  On a  4 x 4  board, place eight men,  WBWB  on the top row and  BWBW  on the bottom row.  Players alternately move one of their men by one square in any direction -- the object is to make four in a line.

Lynn Rohrbough, ed.  Ancient Games.  Handy Series, Kit N, Cooperative Recreation Service, Delaware, Ohio, (1938), 1939. 

Morris was Player [sic] 3,300 Years Ago, p. 27.  Says the temple of Kurna was started by Ramses I and completed by Seti in -1336/-1333, citing J. Royal Asiatic Soc. (1783) 17.

Three Men's Morris, p. 27.  After placing their three men, players 'then move trying to get three men in a row.'  Contributor says he played it in Cardiff more than 50 years ago.

Winning Ways.  1982.  Pp. 672-673.  Says Ovid's Game is conjectured to be Three Men's Morris.  The current version allows moves by one square orthogonally and is a first person win if the first person plays in the centre.  If the first player cannot play in the centre, it is a draw.  They use Three Men's Morris for the case with one step moves along winning lines, i.e. orthogonally or along main diagonals.  An American Indian game, Hopscotch, permits one step moves orthogonally or diagonally (along any diagonal).  A French game, Les Pendus, allows any move to a vacant cell.  All of these are draws, even allowing the first player to play in the centre.  They briefly describe Six and Nine Men's Morris.

Ralph Gasser  &  J. Nievergelt.  Es ist entscheiden: Das Muehle-Spiel ist unentscheiden.  Informatik Spektrum 17 (1994) 314-317.  ??NYS -- cited by Jörg Bewersdorff [email of 6 Jun 1999].

L. V. Allis.  Beating the World Champion -- The state of the art in computer game playing.  IN: Alexander J. de Voogt, ed.; New Approaches to Board Games Research: Asian Origins and Future Perspectives; International Institute for Asian Studies, Leiden, 1995; pp. 155-175.  On p. 163, he states that Ralph Gasser showed that Nine Men's Morris is a draw in Oct 1993, but the only reference is to a letter from Gasser.

Ralph Gasser.  Solving Nine Men's Morris.  IN: Games of No Chance; ed. by Richard Nowakowski; CUP, 1996, pp. 101-113.  ??NYS -- cited by Bewersdorff [loc. cit.] and described in  William Hartston; What mathematicians get up to; The Independent Long Weekend (29 Mar 1997) 2.  Demonstrates that Nine Men's Morris is a draw.  Gasser's abstract: "We describe the combination of two search methods used to solve Nine Men's Morris.  An improved analysis algorithm computes endgame databases comprising about 1010 states.  An 18-ply alpha-beta search the used these databases to prove that the value of the initial position is a draw.  Nine Men's Morris is the first non-trivial game to be solved that does not seem to benefit from knowledge-based methods."  I'm not sure about the last statement --  4 x 4 x 4  noughts and crosses (see 4.B.1.a) and Connect-4 were solved in 1980 and 1988, though the first was a computer aided proof and the original brute force solution of Connect-4 by James Allen in Sep 1988 was improved to a knowledge-based approach by L. V. Allis by Aug 1989.  The five-in-a-row version of Connect-4 was shown to be a first person win in 1993.  Bewersdorff [email of 11 Jun 1999] clarifies this by observing that draw here means a game that continues forever -- one cannot come to a stalemate where neither side can move.

 

          4.B.6. PHUTBALL

 

Winning Ways.  1982.  Philosopher's football, pp. 688‑691.  In 1985, Guy said this was the only published description of the game.

 

          4.B.7. BRIDG‑IT

 

          This is best viewed as played on a  n x n  array of squares.  The  n(n+1)  vertical edges belong to one player, say red, while the  n(n+1)  horizontal edges belong to black.  Players alternate marking a square with a line of their colour between edges of their colour.  A square cannot be marked twice.  The object is to complete a path across the board.  In practice, the edges are replaced by coloured dots which are joined by lines.  As with Hex, there can be no ties and there must be a first person strategy.

 

M. Gardner.  SA (Oct 1958) c= 2nd Book, Chap. 7.  Introduces David Gales's game, later called Bridg‑it.  Addendum in the book notes that it is identical to Shannon's 'Bird Cage' game of 1951 and that it was marketed as Bridg‑it in 1960.

M. Gardner.  SA (Jul 1961) c= New MD, Chap. 18.  Describes Oliver Gross's simple strategy for the first player to win.  Addendum in the book gives references to other solutions and mentions.

M. Gardner.  SA (Jan 1973) c= Knotted, Chap. 9.  Article says Bridg‑it was still on the market.

Winning Ways.  1982.  Pp. 680-682.  Covers Bridg-it and Shannon Switching Game.

In Oct 2000, I bought a second-hand copy of a  5 x 5  version called Connections, attributed to Tom McNamara, but with no date. 

 

          4.B.8. CHOMP

 

Fred Schuh.  Spel van delers (Game of divisors).  Nieuw Tijdschrift vor Wiskunde 39 (1951‑52) 299‑304. ??NYS -- cited by Gardner, below.

M. Gardner.  SA (Jan 1973) c= Knotted, Chap. 9.  Gives David Gale's description of his game and results on it.  Addendum in Knotted points out that it is equivalent to Schuh's game and gives other references.

David Gale.  A curious Nim-type game.  AMM 81 (1974) 876-879.  Describes the game and the basic results.  Wonders if the winning move is unique.  Considers three dimensional and infinite forms.  A note added in proof refers to Gardner's article, says two programmers have consequently found that the  8 x 10  game has two winning first moves and mentions Schuh's game.

Winning Ways.  1982.  Pp. 598-600.  Brief description with extensive table of good moves.  Cites an earlier paper of Gale and Stewart which does not deal with this game.

 

          4.B.9. SNAKES AND LADDERS

 

          I have included this because it has an interesting history and because I found a nice way to express it as a kind of Markov process or random walk, and this gives an expression for the average time the game lasts.  I then found that the paper by Daykin et al. gives most of these ideas.

          The game has two or three rules for finishing.

          A.       One finishes by going exactly to the last square, or beyond it.

          B.       One finishes by going exactly to the last square.  If one throws too much, then one stands still.

          C.       One finishes by going exactly to the last square.  If one throws too much, one must count back from the last square.  E.g., if there are 100 squares and one is at 98 and one throws 6, then one counts:  99, 100, 99, 98, 97, 96  and winds up on 96.  (I learned this from a neighbour's child but have only seen it in one place -- in the first Culin item below.)

          Games of this generic form are often called promotion games.  If one considers the game with no snakes or ladders, then it is a straightforward race game, and these date back to Egyptian and Babylonian times, if not earlier.

          In fact, the same theory applies to random walks of various sorts, e.g. random walks of pieces on a chessboard, where the ending is arrival exactly at the desired square.

 

In the British Museum, Room 52, Case 24 has a Babylonian ceramic board (WA 1991-7.20,I) for a kind of snakes and ladders from c-1000.  The label says this game was popular during the second and first millennia BC.

Sheng-kuan t'u [The game of promotion].  7C.  Chinese game.  This is described in:  Nagao Tatsuzo; Shina Minzoku-shi [Manners and Customs of the Chinese]; Tokyo, 1940-1942, perhaps vol. 2, p. 707, ??NYS  This is cited in:  Marguerite Fawdry; Chinese Childhood; Pollock's Toy Theatres, London, 1977, p. 183, where the game is described as "played on a board or plan representing an official career from the lowest to the highest grade, according to the imperial examination system.  It is a kind of Snakes and Ladders, played with four dice; the object of each player being to secure promotion over the others."

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De ludo promotionis mandarinorum,  pp. 70-101 -- ??NX.  This is a long description of  Shing quon tu,  a game on a board of 98 spaces, each of which has a specific description which Hyde gives.  There is a folding plate showing the Chinese board, but the copy in the Graves collection is too fragile to photocopy.  I did not see any date given for the game.

Stewart Culin.  Chinese Games with Dice and Dominoes.  From the Report of the U. S. National Museum for 1893, pp. 489‑537.  Pp. 502-507 describes several versions of the Japanese Sugoroku (Double Sixes) which is a generic name for games using dice to determine moves, including backgammon and simple race games, as well as Snakes and Ladders games.  One version has ending in the form C.  Then says  Shing Kún T’ò  (The Game of the Promotion of Officials)  is described by Hyde as The Game of the Promotion of the Mandarins and gives an extended description of it.  There is a legend that the game was invented when the Emperor Kienlung (1736-1796) heard a candidate playing dice and the candidate was summoned to explain.  He made up a story about the game, saying that it was a way for him and his friends to learn the different ranks of the civil service.  He was sent off to bring back the game and then made up a board overnight.  However Hyde had described the game a century before this date.  It seems that this is not really a Snakes and Ladders game as the moves are determined by the throw of the dice and the position -- there are no interconnections between cells.  But Culin notes that the game is complicated by being played for money or counters which permit bribery and rewards, etc.

Culin.  Chess and Playing Cards.  Op. cit. in 4.A.4.  1898. 

                    Pp. 820-822 & plates 24 & 25 between 821 & 822.  Says  Shing Kún T’ò  (The Game of the Promotion of Officials)  is described by Hyde as  The Game of the Promotion of the Mandarins  and refers to the above for an extended description.  Describes the Korean version:  Tjyong-Kyeng-To  (The Game of Dignitaries)  and several others from Korea and Tibet, with 108, 144, 169 and 64 squares. 

                    Pp. 840-842 & plate 28, opp. p. 841 describes  Chong ün Ch’au  (Game of the Chief of the Literati)  as 'in many respects analogous' to  Shing Kún T’ò  and the Japanese game Sugoroku (Double Sixes) -- in several versions.  Then mentions modern western versions --  Jeu de L'Oie,  Giuoco dell'Oca,  Juego de la Oca,  Snake Game.  Pp. 843-848 is a table listing 122 versions of the game in the University of Pennsylvania Museum of Archaeology and Paleontology.  These are in 11 languages, varying from 22 to 409 squares.

Bell & Cornelius.  Board Games Round the World.  Op. cit. in 4.B.1.  1988.  Snakes and Ladders and the Chinese Promotion Game, pp. 65‑67.  They describe the Hindu version of Snakes and Ladders, called  Moksha-patamu.  Then they discuss  Shing Kun t'o  (Promotion of the Mandarins),  which was played in the Ming (1368-1616) with four or more players racing on a board with 98 spaces and throwing 6 dice to see how many equal faces appeared.  They describe numerous modern variants.

Deepak Shimkhada.  A preliminary study of the game of Karma in India, Nepal, and Tibet.  Artibus Asiae 44 (1983) 4.  ??NYS - cited in Belloli et al, p. 68.

Andrew Topsfield.  The Indian game of snakes and ladders.  Artibus Asiae 46:3 (1985) 203‑214 + 14 figures.  Basically a catalogue of extant Indian boards.  He says the game is called  Gyān caupad [the  d  should have an underdot]  or  Gyān chaupar  in Hindi.  He states that  Moksha-patamu  sounds like it is Telugu and that this name appeared in Grunfield's Games of the World (1975) with no reference to a source and that Bell has repeated this.  Game boards were drawn or painted on paper or cloth and hence were perishable.  The oldest extant version is believed to be an 84 square board of 1735, in the Museum of Indology, Jaipur.  There were Hindu, Jain, Muslim and Tibetan versions representing a kind of Pilgrim's Progress, finally arriving at God or Heaven or Nirvana.  The number of squares varies from 72 to 360. 

                    He gives many references and further details.  An Indian version of the game was described by F. E. Pargiter; An Indian game: Heaven or Hell; J. Royal Asiatic Soc. (1916) 539-542, ??NYS.  He cites the version by Ayres (and Love's reproduction of it -- see below) as the first English version.  He cites several other late 19C versions.

 

F. H. Ayres.  [Snakes and ladders game.]  No. 200682 Regd.  Example in the Bethnal Green Museum, Misc. 8 - 1974.  Reproduced in:  Brian Love; Play The Game; Michael Joseph, London, 1978; Snakes & Ladders 1, pp. 22-23.  This is the earliest known English version of the game, with  100  cells in a spiral and  5  snakes and  5  ladders.

 

N. W. Bazely  &  P. J. Davis.  Accuracy of Monte Carlo methods in computing finite Markov chains.  J. of Res. of the Nat. Bureau of Standards -- Mathematics and Mathematical Physics 64B:4 (Oct-Dec 1960) 211-215.  ??NYS -- cited by Davis & Chinn and Bewersdorff.  Bewersdorff [email of 6 Jun 1999] brought these items to my attention and says it is an analysis based on absorbing Markov chains.

D. E. Daykin, J. E. Jeacocke & D. G. Neal.  Markov chains and snakes and ladders.  MG 51 (No. 378) (Dec 1967) 313-317.  Shows that the game can be modelled as a Markov process and works out the expected length of play for one player  (47.98  moves) or two players  (27.44  moves) on a particular board with finishing rule A.

Philip J. Davis  &  William G. Chinn.  3.1416 and All That.  S&S, 1969, ??NYS;  2nd ed, Birkhäuser, 1985, chap. 23 (by Davis): "Mr. Milton, Mr. Bradley -- meet Andrey Andreyevich Markov", pp. 164-171.  Simply describes how to set up the Markov chain transition matrix for a game with  100  cells and ending B.  Doesn't give any results. 

Lewis Carroll.  Board game for one.  In: Lewis Carroll's Bedside Book; ed. by Gyles Brandreth (under the pseud. Edgar Cuthwellis); Methuen, 1979, pp. 19-21.  ??look for source; not in Carroll-Wakeling, Carroll-Wakeling II or Carroll-Gardner.  Board of 27 cells with pictures in the odd cells.  If you land on any odd cell, except the last one, you have to return to square 1.  "Sleep is almost certain to have overwhelmed the player before he reaches the final square."  Ending A is probably intended.  (The average duration of this game should be computable.)

S. C. Althoen, L. King & K. Schilling.  How long is a game of snakes and ladders?  MG 77 (No. 478) (Mar 1993) 71-76.  Similar analysis to Daykin, Jeacocke & Neal, using finishing rule B, getting  39.2  moves.  They also use a simulation to find the number of moves is about  39.1.

David Singmaster.  Letter [on Snakes and ladders].  MG 79 (No. 485) (Jul 1995) 396-397.  In response to Althoen et al.  Discusses history, other ending rules and wonders how the length depends on the number of snakes and ladders.

Irving L. Finkel.  Notes on two Tibetan dice games.  IN: Alexander J. de Voogt, ed.; New Approaches to Board Games Research: Asian Origins and Future Perspectives; International Institute for Asian Studies, Leiden, 1995; pp. 24-47.  Part II: The Tibetan 'Game of Liberation', pp. 34-47, discusses promotion games with many references to the literature and describes a particular game.

Jörg Bewersdorff.  Glück, Logik und Bluff  Mathematik im Spiel -- Methoden, Ergebnisse und Grenzen.  Vieweg, 1998.  Das Leiterspiel, pp. 67-68  &  Das Leiterspiel als Markow‑Kette.  Discusses setting up the Markov chain, citing Bazley & Davis, with the same board as in Davis & Chinn, then states that the average duration is  39.224  moves.

Jay Belloli, ed.  The Universe  A Convergence of Art, Music, and Science.  [Catalogue for a group of exhibitions and concerts in Pasadena and San Marino, Sep 2000 - Jun 2001.]  Armory Center for the Arts, Pasadena, 2001.  P. 68 has a discussion of the Jain versions of the game, called 'gyanbazi', with a colour plate of a 19C example with a  9 x 9  board with three extra cells.

 

          4.B.10.        MU TORERE

 

          This is a Maori game which can be found in several books on board games.  I have included it because it has been completely analysed.  There are eight (or  2n)  points around a central area.  Each player has four (or  n)  markers, originally placed on consecutive points.  One can move from a point to an adjacent point or to the centre, or one can move from the centre to a point, provided the position moved to is empty.  The first player who cannot move is the loser.  To prevent the game becoming trivial, it is necessary to require that the first two (or one) moves of each player involve his end pieces, though other restrictions are sometimes given.

 

Marcia Ascher.  Mu Torere: An analysis of a Maori game.  MM 60 (1987) 90-100.  Analyses the game with  2n  points.  For  n = 1,  there are 6 inequivalent positions (where equivalence is by rotation or reflection of the board) and play is trivially cyclic.  For  n = 2,  there are  12  inequivalent positions, but there are no winning positions.  For  n = 3,  there are  30  inequivalent positions, some of which are wins, but the game is a tie.  Obtains the number of positions for general  n.  For the traditional version with  n = 4,  there are  92  inequivalent positions, some of which are wins, but the game is a tie, though this is not at all obvious to an inexperienced player.  In 1856, it was reported that no foreigner could win against a Maori.  For  n = 5,  there are  272  inequivalent positions, but the game is a easy win for the first player -- the constraints on first moves need to be revised.  Ascher gives references to the ethnographic literature for descriptions of the game.

Marcia Ascher.  Ethnomathematics.  Brooks/Cole Publishing, Pacific Grove, California, 1991.  Sections 4.4-4.7, pp. 95-109  &  Notes 4-7, pp. 118-119.  Amplified version of her MM article.

 

          4.B.11.        MASTERMIND, ETC.

 

          There were a number of earlier guessing games of the Mastermind type before the popular version devised by Marco Meirovitz in 1973 -- see:  Reddi.  One of these was the English Bulls and Cows, but I haven't seen anything written on this and it doesn't appear in Bell, Falkener or Gomme.  Since 1975 there have been several books on the game and a number of papers on optimal strategies.  I include a few of the latter. 

          NOTATION.  If there are  h  holes and  c  choices at each hole, then I abbreviate this as  ch.

 

A. K. Austin.  How do You play 'Master Mind'.  MTg 71 (Jun 1975) 46-47.  How to state the rules correctly.

S. S. Reddi.  A game of permutations.  JRM 8:1 (1975) 8-11.  Mastermind type guessing of a permutation of  1,2,3,4  can win in  5  guesses.

Donald E. Knuth.  The computer as Master Mind.  JRM 9:1 (1976-77) 1-6.  64 can be won in  5  guesses.

Robert W. Irving.  Towards an optimum Mastermind strategy,  JRM 11:2 (1978-79) 81-87.  Knuth's algorithm takes an average of  5804/1296 = 4.478  guesses.  The author presents a better strategy that takes an average of  5662/1296 = 4.369  guesses, but requires six guesses in one case.  A simple adaptation eliminates this, but increases the average number of guesses to  5664/1296 = 4.370.  An intelligent setter will choose a pattern with a single repetition, for which the average number of guesses is  3151/720 = 4.376.

A. K. Austin.  Strategies for Mastermind.  G&P 71 (Winter 1978) 14-16.  Presents Knuth's results and some other work.

Merrill M. Flood.  Mastermind strategy.  JRM 18:3 (1985-86) 194-202.  Cites five earlier papers on strategy, including Knuth and Irving.  He considers it as a two-person game and considers the setter's strategy.  He has several further papers in JRM developing his ideas.

Antonio M. Lopez, Jr.  A PROLOG Mastermind program.  JRM 23:2 (1991) 81-93.  Cites Knuth, Irving, Flood and two other papers on strategy.

Kenji Koyama and Tony W. Lai.  An optimal Mastermind strategy.  JRM 25:4 (1994) 251‑256.  Using exhaustive search, they find the strategy that minimizes the expected number of guesses, getting expected number  5625/1296 = 4.340.  However, the worst case in this problem requires  6  guesses.  By a slight adjustment, they find the optimal strategy with worst case requiring  5  guesses and its expected number of guesses is  5626/1296 = 4.341.  10 references to previous work, not including all of the above.

Jörg Bewersdorff.  Glück, Logik und Bluff  Mathematik im Spiel -- Methoden, Ergebnisse und Grenzen.  Vieweg, 1998.  Section 2.15  Mastermind: Auf Nummer sicher, pp. 227-234  &  Section 3.13  Mastermind: Farbcodes und Minimax, pp. 316-319.  Surveys the work on finding optimal strategies.  Then studies Mastermind as a two-person game.  Finds the minimax strategy for the 32 game and describes Flood's approach.

 

4.B.12. RITHMOMACHIA  =  THE PHILOSOPHERS' GAME

 

          I have generally not tried to include board games in any comprehensive manner, but I have recently seen some general material on this which will be useful to anyone interested in the game.  The game is one of the older and more mathematical of board games, dating from c1000, but generally abandoned about the end of the 16C along with the Neo-Pythagorean number theory of Boethius on which the game was based.

 

Arno Borst.  Das mittelalterliche Zahlenkampfspiel.  Sitzungsberichten der Heidelberger Akademie der Wissenschaften, Philosophisch-historische Klasse 5 (1986) Supplemente.  Available separately:  Carl Winter, Heidelberg, 1986.  Edits the surviving manuscripts on the game.  ??NYS -- cited by Stigter & Folkerts.

Detlef Illmer, Nora Gädeke, Elisabeth Henge, Helen Pfeiffer & Monika Spicker-Beck.  Rhythmomachia.  Hugendubel, Munich, 1987.

Jurgen Stigter.  Emanuel Lasker: A Bibliography  AND  Rithmomachia, the Philosophers' Game: a reference list.  Corrected, 1988 with annotations to 1989, 1 + 15 + 16pp preprint available from the author, Molslaan 168, NL‑2611 CZ Delft, Netherlands.  Bibliography of the game.

Jurgen Stigter.  The history and rules of Rithmomachia, the Philosophers' Game.  14pp preprint available from the author, as above.

Menso Folkerts.  'Rithmimachia'.  In:  Die deutsche Litteratur des Mittelalters: Verfasserlexikon; 2nd ed., De Gruyter, Berlin, 1990; vol. 8, pp. 86-94.  Sketches history and describes the 10 oldest texts.

Menso Folkerts.  Die Rithmachia des Werinher von Tegernsee.  In:  Vestigia Mathematica, ed. by M. Folkerts & J. P. Hogendijk, Rodopi, Amsterdam, 1993, pp. 107-142.  Discusses Werinher's work (12C), preserved in one MS of c1200, and gives an edition of it.

 

          4.B.13.        MANCALA GAMES

 

          This is a very broad field and I will only mention a few early items.  Four row mancala games are played in south and east Africa.  Three row games are played in Ethiopia and adjacent parts of Somaliland.  Two row games are played everywhere else in Africa, the Middle East and south and south-east Asia.  See the standard books by R. C. Bell and Falkener for many examples.  Many general books mention the game, but I only know a few specific books on the game -- these are listed first below.

          One article says that game boards have been found in the pyramids of Khamit (-1580) and there are numerous old boards carved in rocks in several parts of Africa.

          An anonymous article, by a member of the Oware Society in London, [Wanted: skill, speed, strategy; West Africa (16-22 Sep 1996) 1486-1487] lists the following names for variants of the game: Aditoe (Volta region of Ghana), Awaoley (Côte d'Ivoire), Ayo (Nigeria), Chongkak (Johore), Choro (Sudan), Congclak (Indonesia), Dakon (Philippines), Guitihi (Kenya), Kiarabu (Zanzibar), Madji (Benin), Mancala (Egypt), Mankaleh (Syria), Mbau (Angola), Mongola (Congo), Naranji (Sri Lanka), Qai (Haiti), Ware (Burkina Faso), Wari (Timbuktu), Warri (Antigua),

 

Stewart Culin.  Mancala, The National Game of Africa.  IN: US National Museum Annual Report 1894, Washington, 1896, pp. 595-607.

Chief A. O. Odeleye.  Ayo  A Popular Yoruba Game.  University Press Ltd., Ibadan, Nigeria, 1979.  No history.

Laurence Russ.  Mancala Games.  Reference Publications, Algonac, Michigan, 1984.  Photocopy from Russ, 1995.

Kofi Tall.  Oware  The Abapa Version.  Kofi Tall Enterprise, Kumasi, Ghana, 1991.

Salimata Doumbia  &  J. C. Pil.  Les Jeux de Cauris.  Institut de Recherches Mathématiques, 08 BP 2030, Abidjan 08, Côte d'Ivoire, 1992.

Pascal Reysset  &  François Pingaud.  L'Awélé.  Le jeu des semailles africaines.  2nd ed., Chiron, Paris, 1995 (bought in Dec 1994).  Not much history.

François Pingaud.  L'awélé  jeu de strategie africain.  Bornemann, 1996.

Alexander J. de Voogt.  Mancala  Board Games.  British Museum Press, 1997.  ??NYR.

Larry (= Laurence) Russ.  The Complete Mancala Games Book  How to Play the World's Oldest Board Games.  Foreword by Alex de Voogt.  Marlowe & Co., NY, 2000.  His 1984  book is described as an earlier edition of this.

 

William Flinders Petrie.  Objects of Daily Use.  (1929);  Aris & Phillips, London??, 1974.  P. 55 & plate XLVII.  ??NYS -- described with plate reproduced in Bell, below.  Shows and describes a  3 x 14  board from Memphis, ancient Egypt, but with no date given, but Bell indicates that the context implies it is probably earlier than ‑1500.  Petrie calls it 'The game of forty-two and pool' because of the 42 holes and a large hole on the side, apparently for storing pieces either during play or between games.

R. C. Bell.  Games to Play.  Michael Joseph (Penguin), 1988.  Chap. 4, pp. 54-61, Mancala games.  On pp. 54-55, he shows the ancient Egyptian board from Petrie and his own photo of a  3 x 6  board cut into the roof of a temple at Deir-el-Medina, probably about ‑87.

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De Ludo Mancala, pp. 226-232.  Have X of part of this.

R. H. Macmillan.  Wari.  Eureka 13 (Oct 1950) 12.  2 x 6  board with each cup having four to start.  Says it is played on the Gold Coast.

Vernon A. Eagle.  On some newly described mancala games from Yunnan province, China, and the definition of a genus in the family of mancala games.  IN: Alexander J. de Voogt, ed.; New Approaches to Board Games Research: Asian Origins and Future Perspectives; International Institute for Asian Studies, Leiden, 1995; pp. 48-62.  Discusses the game in general, with many references.  Attempts a classification in general.  Describes six forms found in Yunnan.

Ulrich Schädler.  Mancala in Roman Asia Minor?  Board Games Studies  International Journal for the Study of Board Games  1 (1998) 10-25.  Notes that mancala could have been played on a flat board of two parallel rows of squares, i.e. something like a  2 x n  chessboard, but that archaeologists have tended to view such patterns as boards for race games, etc.  Describes 52 examples from Asia Minor.  Some general discussion of Greek and Roman games.

John Romein  &  Henri E. Bal (Vrije Universiteit, Amsterdam).  New computer cluster solves 3500-year old game.  Posted on  www.alphagalileo.org  on 29 Aug 2002.  They show that Awari is a tie game.  They determined all  889,063,398,406  possible positions and stored them in a 778 GByte database.  They then used a 144 processor cluster to analyse the graph, which 'only' took 51 hours. 

 

          4.B.14.DOMINOES, ETC.       

 

R. C. Bell.  Games to Play.  1988.  Op. cit. in 4.B.13.  P. 136 gives some history.  The Académie Français adopted the word for both the pieces and the game in 1790 and it was generally thought that they were an 18C invention.  However, a domino was found on the Mary Rose, which sank in 1545, and a record of Henry VIII (reigned 1509-1547) losing £450 at dominoes has been found. 

Bell, p. 131, describes the modern variant Tri-Ominos which are triangular pieces with values at the corners.  They were marketed c1970 and marked  © Pressman Toy Corporation, NY.

Hexadoms are hexagonal pieces with numbers on the edges -- opposite edges have the same numbers.  These were also marketed in the early 1970s -- I have a set made by Louis Marx, Swansea, but there is no date on it.

 

          4.B.15.        SVOYI KOSIRI

 

Anonymous [R. S.  &  J. M. B[rew ?]].  Svoyi kosiri is an easy game.  Eureka 16 (Oct 1953) 8‑12.  This is an intriguing game of pure strategy commonly played in Russia and introduced to Cambridge by Besicovitch.  It translates roughly as 'One's own trumps'.  There are two players and the hands are exposed, with one's spades and clubs being the same as the other's hearts and diamonds.  At Cambridge, the cards below 6 are removed, leaving 36 cards in the deck.  The article doesn't explain how trumps are chosen, but if one has spades as trumps, then the other has hearts as trumps!  Players alternate playing to a central discard pile.  A player can take the pile and start a new pile with any card, or he can 'cover' the top card and then play any card on that.  'Covering' is done by playing a higher card of the same suit or one of the player's own trumps -- if this cannot be done, e.g. if the ace of the player's own trumps has been played, the player has to take the pile.  The object is to get rid of all one's cards.

 

5.COMBINATORIAL RECREATIONS

 

          7.AZ is actually combinatorial rather than arithmetical and I may shift it.

 

          5.A.   THE 15 PUZZLE, ETC.

 

Pictorial versions:  The Premier (1880),  Lemon (1890),  Stein (1898),  King (1927).

Double-sided versions:  The Premier (1880),  Brown (1891).

Relation to Magic Squares:  Loyd (1896),  Cremer (1880),  Tissandier (1880 & 1880?),  Cassell's (1881),  Hutchison (1891).

Making a magic square with the Fifteen Puzzle:  Dudeney (1898),  Anon & Dudeney (1899),  Loyd (1914),  Dudeney (1917),  Gordon (1988).  See also:  Ollerenshaw & Bondi in 7.N.

 

GENERAL

 

Peter Hajek.  1995 report of his 1992 visit to the Museum of Money, Montevideo, Uruguay, with later pictures by Jaime Poniachik.  In this Museum is a metal chest made in England in 1870 for the National State Bank of Uruguay.  The front has a  7 x 7  array of metal squares with bolt heads.  These have to be slid in a 12 move sequence to reveal the three keyholes for opening the chest.  This opens up a whole new possible background for the 15 Puzzle -- can anyone provide details of other such sliding devices?

 

S&B, pp. 126‑129, shows several versions of the puzzle.

L. Edward Hordern.  Sliding Piece Puzzles.  OUP, 1986.  Chap. 2: History of the sliding block puzzle, pp. 18‑30.  This is the most extensive survey of the history.  He concludes that Loyd did not invent the general puzzle where the 15 pieces are placed at random, which became popular in 1879(?).  Loyd may have invented the 14‑15 version or he may have offered the $1000 prize for it, but there is no evidence of when (1881??) or where.  However, see the entries for Loyd's Tit‑Bits article and Dudeney's 1904 article which seem to add weight to Loyd's claims.  Most of the puzzles considered here are described by Hordern and have code numbers beginning with a letter, e.g.  E23,  which I will give. 

                    I contributed a note about computer techniques of solving such puzzles and hoping that programmers would attack them as computer power increased. 

In 1993-1995, I produced four Sliding Block Puzzle Circulars, totalling 24 pages (since reformatted to 21), largely devoted to reporting on computer solutions of puzzles in Hordern.  Since then, a large number of solution programs have appeared and many more puzzles have appeared.  The best place to look is on Nick Baxter's Sliding Block Home Page:  http://www.johnrausch.com/slidingblockpuzzles/index.html .

 

EARLY ALPHABETIC VERSIONS

 

Embossing Co.  Puzzle labelled "No. 2 Patent Embossed puzzle of Fifteen and Magic Sixteen.  Manufactured by the Embossing Co.  Patented Oct 24 1865".  Illustrated in S&B, p. 127.  Examples are in the collections of Slocum and Hordern.  Hordern, p. 25, says that searching has not turned up such a patent.

Edward F. [but drawing gives E.] Gilbert.  US Patent 91,737 -- Alphabetical Instruction Puzzle.  Patented 22 Jun 1869.  1p + 1p diagrams.  Described by Hordern, p. 26.  This is not really a puzzle -- it has the sliding block concept, but along several tracks and with many blank spaces.  I recall a similar toy from c1950.

Ernest U. Kinsey.  US Patent 207,124 -- Puzzle-Blocks.  Applied: 22 Nov 1877;  patented: 20 Aug 1878.  2pp + 1p diagrams.  Described by Hordern, p. 27.  6 x 6  square sliding block puzzle with one vacant space and tongue & grooving to prevent falling out.  Has letters to spell words.  He suggests use of triangular and diamond‑shaped pieces.  This seems to be the most likely origin of the Fifteen Puzzle craze.

Montgomery Ward & Co.  Catalogue.  1889.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 34.  Spelling Boards.  Like Gilbert's idea, but a more compact layout.

 

 

                    LOYD

 

Loyd prize puzzle: One hundred pounds.  Tit-Bits (14 Oct 1893) 25  &  (18 Nov 1893) 111.  Loyd is described as "author of "Fifteen Puzzle," ...."

Loyd.  Tit‑Bits 31 (24 Oct 1896) 57.  Loyd asserts he developed the 15 puzzle from a  4 x 4  magic square.  "[The fifteen block puzzle] had such a phenomenal run some twenty years ago. ...  There was one of the periodical revivals of the ancient Hindu "magic square" problem, and it occurred to me to utilize a set of movable blocks, numbered consecutively from 1 to 16, the conditions being to remove one of them and slide the others around until a magic square was formed.  The "Fifteen Block Puzzle" was at once developed and became a craze.

                    I give it as originally promulgated in 1872 ..." and he shows it with the 15 and 14 interchanged.  "The puzzle was never patented" so someone used round blocks instead of square ones.  He says he would solve such puzzles by turning over the 6 and the 9.  "Sphinx" [= Dudeney] says he well remembers the sensation and hopes "Mr. Loyd is duly penitent."

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  "... the "Fifteen Puzzle" that in 1872 and 1873 was sold by millions, ....  When this puzzle was brought out by its inventor, Mr. Sam Loyd, ... he thought so little of it that he did not even take any steps to protect his idea, and never derived a penny profit from it....  We have recently tried all over the metropolis to obtain a single example of the puzzle, without success."  Dudeney says the puzzle came with 16 pieces and you removed the 16.  He also says he recently could not find a single example in London.

Loyd.  The 14‑15 puzzle in puzzleland.  Cyclopedia, 1914, pp. 235 & 371 (= MPSL1, prob. 21, pp. 19‑20 & 128).  He says he introduced it 'in the early seventies'.  One problem asks to move from the wrong position to a magic square with sum = 30  (i.e. the blank is counted as  0).  This is c= SLAHP, pp. 17‑18 & 89.

G. G. Bain.  Op. cit. in 1, 1907.  Story of Loyd being unable to patent it.

Anonymous & Sam Loyd.  Loyd's puzzles, op. cit. in 1, 1896.  Loyd "owns up to the great sin of having invented the "15 block puzzle"", but doesn't refer to the patent story or the date.

W. P. Eaton.  Loc. cit. in 1, 1911.  Loyd refers to it as the 'Fifteen block' puzzle, but doesn't say he couldn't patent it.

Loyd Jr.  SLAHP.  1928.  Pp. 1‑3 & 87.  "It was in the early 80's, ... that the world‑disturbing "14‑15 Puzzle" flashed across the horizon, and the Loyds were among its earliest victims."  He gives many of the stories in the Cyclopedia and two of the same problems.  He doesn't mention the patent story.

 

                    THE 15 PUZZLE

 

W. W. Johnson.  Notes on the 15‑Puzzle -- I.  Amer. J. Math. 2 (1879) 397‑399.

W. E. Story.  Notes on the 15‑Puzzle -- II.  Ibid., 399‑404.

J. J. Sylvester.  Editorial comment.  Ibid., 404.

          (This issue may have been delayed to early 1880??  Johnson & Story are not terribly readable, but Sylvester is interesting, asserting that this is the first time that the parity of a permutation has become a popular concept.)

Anonymous.  Untitled editorial.  New York Times (23 Feb 1880) 4.  "... just now the chief amusement of the New York mind, ... a mental epidemic ....  In a month from now, the whole population of North America will be at it, and when the 15 puzzle crosses the seas, it is sure to become an English mania."

Anonymous.  EUREKA!  The Popular but Perplexing Problem Solved at Last.  "THIRTEEN -- FOURTEEN -- FIFTEEN"  New York Herald (28 Feb 1880) 8.  ""Fifteen" is a puzzle of seeming simplicity, but is constructed with diabolical cunning.  At first sight the victim feels little or no interest; but if he stops for a single moment to try it, or to look at any one else who is trying it, the mania strikes him.  ...  As to the last two numbers, it depends entirely upon the way in which the blocks happen to fall in the first place ....  Two or three enterprising gamblers took up the puzzle and for a time made an excellent living....  The subject was brought up in the Academy of Sciences by the veteran scientist Dr. P. H. Vander Weyde", who showed it could not be solved.  The Herald reporter discovered that the problem is solvable if one turns the board 90o, i.e. runs the numbers down instead of across, and Vander Weyde was impressed.  The article implies the puzzle had already been widely known for some time.

Mary T. Foote.  US Patent 227,159 -- Game apparatus.  Filed: 4 Mar 1880;  patented: 4 May 1880.  1p + 1p diagrams.  The patent is for a box with sliding numbered blocks for teaching the multiplication tables.  Lines 57-63:  "I am aware that it is not novel to produce a game apparatus in which blocks are to be mixed and then replaced by a series of moves; also, that it is not novel to number such blocks, as in the "game of 15," so called, where the fifteen numbers are first mixed and then moved into place."

Persifor Frazer Jr.  Three methods and forty‑eight solutions of the Fifteen Problem.  Proc. Amer. Philos. Soc. 18 (1878‑1880) 505‑510.  Meeting of 5 Mar 1880.  Rather cryptic presentation of some possible patterns.  Asserts his 26 Feb article in the Bulletin (??NYS -- ??where -- Philadelphia??) was the first "solution for the 13, 15, 14 case".

J. A. Wales.  15 - 14 - 13 -- The Great Presidential Puzzle.  Puck 7 (No. 158) (17 Mar 1880) back cover.

Anonymous.  Editorial:  "Fifteen".  New York Times (22 Mar 1880) 4.  "No pestilence has ever visited this or any other country which has spread with the awful celerity of what is popularly called the "Fifteen Puzzle."  It is only a few months ago that it made its appearance in Boston, and it has now spread over the entire country."  Asserts that an unregenerate Southern sympathiser has introduced it into the White House and thereby disrupted a meeting of President Hayes' cabinet.

Sch. [H. Schubert].  The Boss Puzzle.  Hamburgischer Correspondent (= Staats- und Gelehrte Zeitung des Hamburgischen unpartheyeischen Correspondent) No. 82 (6 Apr 1880) 11, with response on  87 (11 Apr 1880) 12 (Sprechsaal).  Gives a fairly careful description of odd and even permutations and shows the puzzle is solvable if and only if it is in an even permutation.  The response is signed X and says that when the problem is insoluble, just turn the box by 90o to see another side of the problem!

Gebr. Spiro, Hofliefer (Court supplier), Jungfernsteig 3(?--hard to read), Hamburg.  Hamburgischer Correspondent (= Staats- und Gelehrte Zeitung des Hamburgischen unpartheyeischen Correspondent) No. 88 (13 Apr 1880) 7.  Advertises Boss Puzzles:  "Kaiser-Spiel 50Pf.  Bismarck-Spiel  50 Pf.  Spiel der 15 u. 16, 50 Pf.  Spiel der 16 separat, 15 Pf.  System und Lösung, 20 Pf."

G. W. Warren.  Letter:  Clew to the Fifteen Puzzle.  The Nation 30 (No. 774) (29 Apr 1880) 326.

Anon.  Shavings.  The London Figaro (1 May 1880) 12.  "The "15 Puzzle," which has for some months past been making a sensation in New York equal to that aroused by "H. M. S. Pinafore" last year, has at length reached this country, and bids fair to become the rage here also."  (Complete item!)

George Augustus Sala.  Echoes of the Week.  Illustrated London News 76 (No. 2138) (22 May 1880) 491.

Mary T. Foote.  US Patent 227,159 -- Game Apparatus.  Applied: 4 Mar 1880;  patented: 4 May 1880.  1p + 1p diagrams.  Described in Hordern, p. 27.  3 x 12  puzzles based on multiplication tables.  Refers to the "game of 15" and Kinsey.

Arthur Black.  ??  Brighton Herald (22 May 1880).  ??NYS -- mentioned by Black in a letter to Knowledge 1 (2 Dec 1881) 100.

Anonymous.  Our latest gift to England.  From the London Figaro.  New York Times (11 Jun 1880) 2(?).  ??page

The Premier.  First (?) double‑sided version, with pictures of Gladstone and Beaconsfield, apparently produced for the 1880 UK election.  Described in Hordern, pp. 32‑33 & plate I.

Ahrens.  MUS II 227.  1918.  Story of Reichstag being distracted in 1880.

P. G. Tait.  Note on the Theory of the "15 Puzzle".  Proc. Roy. Soc. Edin. 10 (1880) 664‑665.  Brief but valid analysis.  Mentions Johnson & Story.  First mention of the possibility of a 3D version.

T. P. Kirkman.  Question 6489  and  Note on the solution of the 15‑puzzle in question 6489.  Mathematical Questions with their Solutions from the Educational Times 34 (1880) 113‑114  &  35 (1881) 29‑30.  The question considers the  n x n  problem.  The note is rather cryptic.  (No use??)

Messrs. Cremer (210 Regent St. and 27 New Bond St., London).  Brilliant Melancholia.  Albrecht Durer's Game of the Thirty Four and "Boss" Game of the Fifteen.  1880.  Small booklet, 16pp + covers, apparently instructions to fit in a box with pieces numbered 1 to 16 to be used for making magic squares as well as for the 15 puzzle.  Explains that only half the positions of the 15 puzzle are obtainable and describes them by examples.  (Photo in The Hordern Collection of Hoffmann Puzzles, p. 74, and in Hordern, op. cit. above, plate IV.)  Possibly written by "Cavendish" (Henry Jones).

H. Schubert.  Theoretische Entscheidung über das Boss‑Puzzle Spiel.  2nd ed., Hamburg, 1880.  ??NYS  (MUS, II, p. 227)

Gaston Tissandier.  Les carrés magiques -- à propos du "Taquin," jeu mathématique.  La Nature 8 (No. 371) (10 Jul 1880) 81‑82.  Simple description of the puzzle called 'Taquin' which came from America and has had a very great success for several weeks.  Says it had 16 squares and was usable as a sliding piece puzzle or a magic square puzzle.  Cites Frénicle's 880 magic squares of order 4.

Anon. & C. Henry.  Gaz. Anecdotique Littéraire, Artistique et Bibliographique.  (Pub. by G. d'Heylli, Paris)  Year 5, t. II, 1880, pp. 58‑59 & 87‑92.  ??NYS

Piarron de Mondésir.  Le dernier mot du taquin.  La Nature 8 (No. 382) (25 Sep 1880) 284‑285.  Simple description of parity decision for the 15 puzzle.  Says 'la Presse illustrée' offered 500 francs for achieving the standard pattern from a random pattern, but it was impossible, or rather it was possible in only half the cases.

Jasper W. Snowdon.  The "Fifteen" Puzzle.  Leisure Hour 29 (1880) 493‑495.

Gwen White.  Antique Toys.  Batsford, London, 1971;  reprinted by Chancellor Press, London, nd [1982?].  On p. 118, she says:  "The French game of Taquin was played in 1880, in which 15 pieces had to be moved into 16 compartments in as few moves as possible; the word 'taquin' means 'a teaser'."  She gives no references.

Tissandier.  Récréations Scientifiques.  1880? 

                    2nd ed., 1881 -- unlabelled section, pp. 143-153.   As: Le taquin et les carrés magiques;  seen in 1883 ed., ??NX;  1888: pp. 208-215.  Adapted from the 1880 La Nature articles of Tissandier and de Mondésir.  1881 says it came from America -- 'récemment une nouvelle apparition', but this is dropped in 1888 -- otherwise the two versions are the same.

                    Translated in Popular Scientific Recreations, nd [c1890], pp. 731‑735.  Text says "Mathematical games, ..., have recently obtained a new addition ....  ... from America, ...."  The references to contemporary reactions are deleted and the translation is confused.  E.g. the newspaper is now just "a French paper" and the English says the problem is impossible in nine cases out of ten!

Lucas.  Récréations scientifiques sur l'arithmétique et sur la géométrie de situation.  Sixième récréation: Sur le jeu du taquin ou du casse‑tête américain.  Revue scientifique de France et de l'étranger (3) 27 (1881) 783‑788.  c= Le jeu du taquin, RM1, 1882, pp. 189‑211.  Revue says that Sylvester told him that it was invented 18 months ago by an American deaf‑mute.  RM1 says "vers la fin de 1878".  Cf Schubert, 1895.

Cassell's.  1881.  Pp. 96‑97: American puzzles "15" and "34".  = Manson, pp. 246-248.  Says "articles ... have appeared in many periodicals, but no one has ... publish[ed] a solution." Then sketches the parity concept and its application. 

Richard A. Proctor.  The fifteen puzzle.  Gentlemen's Magazine 250 (No. 1801) (1881) 30‑45.

"Boss".  Letter:  The fifteen puzzle.  Knowledge 1 (11 Nov 1881) 37-38, item 13.  This magazine was edited by Proctor.  The letter starts:  "I am told that in a magazine article which appeared some time ago, you have attempted to show that there are positions in the Fifteen Puzzle from which the won position can never be obtained."  I suspect the letter was produced by Proctor.  The response is signed Ed. and begins:  "I thought the Fifteen Puzzle was dead, and hoped I had had some share in killing the time-absorbing monster."  Notes that many people get to the position starting  blank, 1, 2, 3  and view this as a win.  Sketches parity argument and suggests "Boss" work on the  3 x 3  or  3 x 2  or even the  2 x 2  version.

Editorial comment.  The fifteen puzzle.  Knowledge 1 (25 Nov 1881) 79.  "I supposed every one knew the Fifteen Puzzle."  Proceeds to explain, obviously in response to readers who didn't know it.

Arthur Black.  Letter:  The fifteen puzzle.  Knowledge 1 (2 Dec 1881) 100, item 80.  Sketches a proof which he says he published in the Brighton Herald of 22 May 1880.

"Yawnups".  Letter:  The fifteen puzzle.  Knowledge 1 (30 Dec 1881) 185.  Solution from the 15-14 position obtained by turning the box.  Editorial comment says the solution uses 102 moves and the editor gets an easy solution in 57 moves.  Adds that a 60 move solution has been received.

Arthur Black.  Letter:  The fifteen puzzle.  Knowledge 1 (13 Jan 1882) 230.  Finds a solution from the 15-14 position in 39 moves by turning the box and asserts no shorter solution is possible.  Says he also gave this in the Brighton Herald in May 1880.  An addition says J. Watson has provided a similar solution, which takes 38 moves??

A. B.  Letter:  The fifteen puzzle.  Knowledge 2 (20 Oct 1882) 345, item 598.  Finds a box-turning solution in 39 moves.

C. J. Malmsten.  Göteborg Handl 1882, p. 75.  ??NYS -- cited by Ahrens in his Encyklopadie article, op. cit. in 3.B, 1904.

Anonymous.  Enquire Within upon Everything.  Houlston and Sons, London.  This was a popular book with editions almost every year -- I don't know when the following material was added.  Section 2591: Boss; or the Fifteen Puzzle, p. 363.  Place the pieces 'indifferently' in the box.  Half the positions are unsolvable.  Cites Cavendish for the solution by turning the box 90o but notes this only works with round pieces.  Goes on to The thirty-four puzzle, citing Dürer.  I found this material in the 66th ed., 862nd thousand, of 1883, but I didn't find the material in the 86th ed of 1892. 

Letters received and short answers.  Knowledge 4 (16 Nov 1883) 310.  'Impossible'.

P. G. Tait.  Listing's Topologie.  Philosophical Mag. (5) 17 (No. 103) (Jan 1884) 30‑46  &  plate opp. p. 80.  Section 11, p. 39.  Simple but cryptic solution.

Letters received and short answers.  Letter from W. S. B. asks how to solve the problem when the last row has 13, 14, 15 [sic!];  Answer by Ed. points out the misprint and says the easiest solution is to remove the 15 and put it after the 14, or to invert the  6  and  9.  Knowledge  6 (No. 159) (14 Nov 1884) 412  &  6 (No. 160) (21 Nov 1884) 429.

Don Lemon.  Everybody's Pocket Cyclopedia ....  Saxon & Co., London, (1888), revised 8th ed., 1890.  P. 137: The fifteen puzzle.  Brief description, with pieces placed randomly in the box -- "to get the last three into order is often a puzzle indeed".

John D. Champlin  &  Arthur E. Bostwick.  The Young Folk's Cyclopedia of Games and Sports.  1890.  ??NYS  Cited in Rohrbough; Brain Resters and Testers; c1935; Fifteen Puzzle, p. 20.  Describes idea of parity of number of exchanges.  [Another reference provided more details of Champlin & Bostwick.]

Lemon.  1890.  A trick puzzle, no. 202, pp. 31 & 105 (= Sphinx, no. 422, pp. 60 & 112).  15 puzzle with lines on the pieces to arrange as "a representation of a president with only one eye".  The solution is a spelling of the word 'president'.  Attributed to Golden Days -- ??.  After The Premier puzzle of c1880, this is the second suggestion of using a picture and the first publication of the idea that I have seen.

G. A. Hutchison, ed.  Indoor Games and Recreations.  The Boy's Own Bookshelf.  (1888);  New ed., Religious Tract Society, London, 1891.  (See M. Adams; Indoor Games for a much revised version, but which doesn't contain this material.)  Chap. 19: The American Puzzles., pp. 240‑241.  "These puzzles, known as the 'Thirty‑four Game' and the 'Fifteen Game,' on their introduction amongst us some years ago ...."  "The '15' puzzle would appear to have been, on its coming to England a few years ago, strictly a new introduction ...."  He sketches the parity concept.  [NOTE.  I have seen a reference to the editor as Hutchinson, but the book definitely omits the first  n.]

Daniel V. Brown.  US Patent 471,941 -- Puzzle.  Applied: 23 Apr 1891;  patented: 29 Mar 1892, 2pp + 1p diagrams.  Double-sided 16 block puzzle to spell George Washington on one side and Benjamin Harrison on the other.  No sliding involved.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  American fifteen puzzle, pp. 105-107.  "The Fifteen Puzzle was introduced by a shrewd American some ten years ago, ...."  Refers to Tait's 1880 paper.  Says half the positions are impossible, but solves them by turning the box 90o or by inverting the 6 and the 9.

Hoffmann.  1893.  Chap IV, no. 69: The "Fifteen" or "Boss" puzzles, pp. 161‑162 & 217‑218 = Hoffmann-Hordern, pp. 142-144, with photo of five early examples, two or three of which also are thirty-four puzzles.  (Hordern Collection, p. 74, has a photo of a version by Cremer, cf above.)  "This, like a good many of the best puzzles, hails from America, where, some years ago, it had an extraordinary vogue, which a little later spread to this country, the British public growing nearly as excited over the mystic "Fifteen" as they did at a later date over the less innocent "Missing Word" competitions."  He distinguishes between the ordinary Fifteen where one puts the pieces in at random, and the Boss or Master puzzle which has the 14 and 15 reversed.  "Notwithstanding the enormous amount of energy that has been expended over the "Fifteen" Puzzle, no absolute rule for its solution has yet been discovered and it appears to be now generally agreed by mathematicians that out of the vast number of haphazard positions ... about half admit [of solution].  To test whether ... the following rule has been suggested."  He then says to count the parity of the number of transpositions.

Hoffmann.  1893.  Chap. IV, no. 70: The peg‑away puzzle, pp. 163 & 218 = Hoffmann‑Hordern, p. 145.  This is a  3 x 3  version of the Fifteen puzzle, made by Perry & Co.  Start with a random pattern and get to standard form.  "The possibility of success in solving this puzzle appears to be governed by precisely the same rule as the "Fifteen" Puzzle."  Hoffmann-Hordern has no photo of this -- do any examples exist??

H. Schubert.  Zwölf Geduldspiele.  Dümmler, Berlin, 1895.  [Taken from his columns in Naturwissenschaftlichen Wochenschrift, 1891-1894.]  Chap. VII: Boss-Puzzle oder Fünfzehner-Spiel, pp. 75-94??  Pp. 75-77 sketches the history, saying it was called "Jeu du Taquin" (Neck-Spiel) in France and was popular in 1879-1880 in Germany.  Cites Johnson & Story and his own 1880 booklet.  Gives the story of a deaf and dumb American inventing it in Dec 1878, saying "Sylvester communicated this at the annual meeting of the Association Française pour l'Avancement des Sciences at Reims".  Cf Lucas, 1881.  [There is a second edition, Teubner??, Leipzig, 1899, ??NYS.  However this material is almost identical to the beginning of Chap. 15 in Schubert's Mathematische Mussestunden, 3rd ed., Göschen, Leipzig, 1909, vol. 2.  The later version omits only some of the Hamburg details of 1879-1880.  Hence the 2nd ed. of Zwölf Geduldspiele is probably very close to these versions.]

Dudeney.  Problem 49: The Victoria Cross puzzle.  Tit‑Bits 32 (4  &  25 Sep 1897) 421  &  475.  = AM, 1917, prob. 218, pp. 60 & 194.  B7.  3 x 3  board with letters Victoria going clockwise around the edges, leaving the middle empty, and starting with  V  in a corner.  Slide to get Victoria starting at an edge cell, in the fewest moves.  Does it in 18 moves, by interchanging the  i's  and  says there are 6 such solutions. 

Dudeney.  Problem 65: The Spanish dungeon.  Tit‑Bits 33 (1 Jan  &  5 Feb 1898) 257  &  355.  = AM, 1917, prob. 403, pp. 122-123 & 244.  B14.  Convert 15 Puzzle, with pieces in correct order, into a magic square.  Does it in 37 moves.

Conrad F. Stein.  US Design 29,649 -- Design for a Game-Board.  Applied: 29 Sep 1898;  patented: 8 Nov 1898  as No. 692,242.  1p + 1p diagrams.  This appears to be a  3 x 4  puzzle with a picture of a city with a Spanish flag on a tower.  Apparently the object is to move an American flag to the tower.

Anon. & Dudeney.  A chat with the Puzzle King.  The Captain 2 (Dec? 1899) 314-320;  2:6 (Mar 1900) 598-599  &  3:1 (Apr 1900) 89.  The eight fat boys.  3 x 3  square with pieces:  1 2 3;  4 X 5;  6 7 8  to be shifted into a magic square.  Two solutions in 19 moves.  Cf Dudeney, 1917.

Addison Coe.  US Patent 785,665 -- Puzzle or Game Apparatus.  Applied: 17 Nov 1904;  patented: 1 Mar 1905.  4pp + 3pp diagrams.  Gives a  3 x 5  flat version and a 3‑dimensional version -- cf 5.A.2.

Dudeney.  AM.  1917. 

Prob. 401: Eight jolly gaol birds, pp. 122 & 243.  E23.  Same as 'The eight fat boys' (see Anon. & Dudeney, 1899) with the additional condition that one person refuses to move, which occurs in one of the two previous solutions.

Prob. 403: The Spanish dungeon, pp. 122-123 & 244.  = Tit-Bits prob. 65 (1898).  B14.

Prob. 404: The Siberian dungeons, pp. 123 & 244.  B16.  2 x 8  array with prisoners  1, 2, ..., 8  in top row and  9, 10, ..., 16  in bottom row.  Two extra rows of 4 above the right hand end (i.e. above  5, 6, 7, 8)  are empty.  Slide the prisoners into a magic square.  Gives a solution in 14 moves, due to G. Wotherspoon, which they feel is minimal.  This allows long moves -- e.g. the first move moves  8  up two and left 3.

"H. E. Licks" [pseud. of Mansfield Merriman].  Recreations in Mathematics.  Van Nostrand, NY, 1917.  Art. 28, pp. 20‑21.  'About the year 1880 ... invented in 1878 by a deaf and dumb man....' 

                    [From sometime in the 1980s, I suspected the author's name was a pseudonym.  On pp. 132-138, he discusses the Diaphote Hoax, from a Pennsylvania daily newspaper of 10 Feb 1880, which features the following people: H. E. Licks, M. E. Kannick, A. D. A. Biatic, L. M. Niscate.  The diaphote was essentially a television.  He says this report was picked up by the New York Times and the New York World.  An email from Col. George L. Sicherman on 5 Jun 2000 agrees that the name is false and suggested that the author was "the eminent statistician Mansfield Merriman" who wrote the article on The Cattle Problem of Archimedes in Popular Science Monthly (Nov 1905), which is abridged on pp. 33-39 of the book, but omitting the author's name.  Sichermann added that Merriman was one of the authors of Pillsbury's List.  William Hartston says this was an extraordinary list of some 30 words which Pillsbury, who did memory feats, was able to commit to memory quite rapidly.  Sicherman continued to investigate Merriman and got Prof. Andri Lange interested and Lange corresponded with a James A. McLennan, author of a history of the physics department at Lehigh University where Merriman had been.  McLennan found Merriman's obituary from the American Society of Civil Engineers which states that Merriman used H. E. Licks as a pseudonym.  [Email from Sicherman on 25 Feb 2002.]]

Stephen Leacock.  Model Memoirs and Other Sketches from Simple to Serious.  John Lane, The Bodley Head, 1939, p. 300.  "But this puzzle stuff, as I say, is as old as human thought.  As soon as mankind began to have brains they must have loved to exercise them for exercise' sake.  The 'jig-saw' puzzles come from China where they had them four thousand years ago.  So did the famous 'sixteen puzzle' (fifteen movable squares and one empty space) over which we racked our brains in the middle eighties."

G. Kowalewski.  Boss‑Puzzle und verwandte Spiele.  K. F. Kohler Verlag, Leipzig, 1921 (reprinted 1939).  Gives solution of general polygonal versions, i.e. on a graph with a Hamilton circuit and one or more diagonals.

Hummerston.  Fun, Mirth & Mystery.  1924.

 

          1     2                              

             9                                   Push, pp. 22 & 25.  This is played on the board

  3    10    11     4              shown at the left with its orthogonal lines, like

    12            13                 3, 10, 11, 4,  and its diagonal lines, like

  5    14    15     6              1, 9, 11, 13, 6.  10, 15  and  11, 14  are not

            16                        connected, so this is an octagram.  Take  16

          7     8                     numbered counters and place them at random on

                                        the board and remove counter 16.  Move the pieces

                    to their correct locations.  He asserts that 'unlike the original ["Sixteen" Puzzle],

                    no position can be set up in "Push" that cannot be solved'.

The six bulls puzzle, Puzzle no. 34, pp. 90 & 177.  This uses the  2 x 3 + 1                                      0    

         board shown at the right, where the  0  is the blank space.  Exchange                                  1  2  3

         3 and 6  and  4 and 5.  He does it in  20  moves.  [This is Hordern's                                   4  5  6

          B3, first known from 1977 under the name Bull Pen, but is a variant of

            Hordern's B2, first known from 1973.]

Q. E. D. -- The sergeant's problem, Puzzle no. 40, pp. 106 & 178.  Take a  2 x 3  board, with the centre of one long side blank.  Interchange the men along one short side.  He does this in  17  moves, but the blank is not in its initial position nor are the other men.  [This is Hordern's B1, first known from Loyd's Cyclopedia, 1914.]

King.  Best 100.  1927.  No. 26, p. 15.  = Foulsham's, no. 9, pp. 7 & 10.  "An entertaining variation ... is to draw, and colour, if you like, a small picture; then cut it into sixteen squares and discard the lower right hand square."

G. Kowalewski.  Alte und neue mathematische Spiele.  Teubner, Leipzig, 1930, pp. 61‑81.  Gives solution of general polygonal versions.

Dudeney.  PCP.  1932.  The Angelica puzzle, prob. 253, pp. 76 & 167.  = 435, prob. 378, pp. 136 & 340.  B8.  3 x 3  problem -- convert:  A C I   L E G   N A X   to   A N G   E L I   C A X.   Requires interchanging the  As.  Solution in  36  moves.  In the answer in 435, Gardner notes that it can be done in  30  moves.

H. V. Mallison.  Note 1454:  An array of squares.  MG 24 (No. 259) (May 1940) 119‑121.  Discusses 15 Puzzle and says any legal position can be achieved in at most about 150 moves.  But if one fixes cells  6, 7, 11,  then a simple problem requires about 900 moves.

McKay.  At Home Tonight.  1940.  Prob. 44: Changing the square, pp. 73 & 88.  In the usual formation, colour the pieces alternately blue and red, as on a chessboard, with the blank at the lower right position 16 being a missing red, so there are 7 reds.  Move so the colours are still alternating but the blank is at the lower left, i.e. position 13.  Takes 15 moves.

Sherley Ellis Stotts.  US Patent 3,208,753 -- Shiftable Block Puzzle Game.  Filed: 7 Oct 1963;  patented: 28 Sep 1965.  4pp + 2pp diagrams.  Described in Hordern, pp. 152-153, F10‑12.  Rectangular pieces of different sizes.  One can also turn a piece.

Gardner.  SA (Feb 1964)  = 6th Book, chap. 7.  Surveys sliding-block puzzles with non-square pieces and notes there is no theory for them.  Describes a number of early versions and the minimum number of moves for solution, generally done by hand and then confirmed by computer.  Pennant Puzzle, C19;  L'Âne Rouge, C27d;  Line Up the Quinties, C4;  Ma's Puzzle, D1;  a form of Stotts' Baby Tiger Puzzle, F10.

Gardner.  SA (Mar & Jun 1965)  c= 6th Book, chap. 20.  Prob. 9: The eight-block puzzle.  B5.  3 x 3  problem -- convert:   8 7 6   5 4 3   2 1 X   to   1 2 3   4 5 6   7 8 X.  Compares it with Dudeney's Angelica puzzle (1932, B8) but says it can be done if fewer than  36  moves.  Many readers found solutions in  30  moves;  two even found all  10  minimal solutions by hand!  Says Schofield (see next entry) has been working on this and gives the results below, but this did not quite resolve Gardner's problem.  William F. Dempster, at Lawrence Radiation Laboratory, programmed a IBM 7094 to find all solutions, getting  10  solutions in  30  moves;  112  in  32  moves  and  512  in  34  moves.  Notes it is unknown if any problem with the blank in a side or corner requires more than  30  moves.  (The description of Schofield's work seems a bit incorrect in the SA solution, and is changed in the book.)

Peter D. A. Schofield.  Complete solution of the 'Eight‑Puzzle'.  Machine Intelligence 1 (1967) 125‑133.  This is the  3 x 3  version of the 15 Puzzle, with the blank space in the centre.  Works with the corner twists which take the blank around a  2 x 2  corner in four moves.  Shows that the 5-puzzle, which is the  3 x 2  version, has every position reachable in at most  20  moves, from which he shows that an upper bound for the 8-puzzle is  48  moves.  Since the blank is in the middle, the  8!/2 = 20160  possible positions fall into  2572  equivalence classes.  He also considers having inverse permutations being equivalent, which reduces to  1439  classes, but this was too awkward to implement.  An ATLAS program found that the maximum number of moves required was  30  and  60  positions of  12  classes required this maximum number, but no example is given -- but see previous entry.

A. L. Davies.  Rotating the fifteen puzzle.  MG 54 (No. 389) (Oct 1970) 237‑240.  Studies versions where the numbers are printed diagonally so one can make a 90o turn of the puzzle.  Then any pattern can be brought to one of two 'natural' patterns.  He then asks when this is true for an  m x n  board and obtains a complicated solution.  For an  n x n  board,  n  must be divisible by  4.

R. M. Wilson.  Graph puzzles, homotopy and the alternating group.  J. Combinatorial Thy., Ser. B, 16 (1974) 86‑96.  Shows that a sliding block puzzle, on any graph of  n + 1  points which is non‑separable and not a cycle, has at least  An  as its group -- except for one case on 7 points.

Alan G. & Dagmar R. Henney.  Systematic solutions of the famous 15‑14 puzzles.  Pi Mu Epsilon J. 6 (1976) 197‑201.  They develop a test‑value which significantly prunes the search tree.  Kraitchik gave a problem which took him  114  moves -- the authors show the best solution has  58  moves!

David Levy.  Computer Gamesmanship.  Century Publishing, London, 1983.  [Most of the material appeared in Personal Computer World, 1980‑1981.]  Pp. 16‑29 discusses 8‑puzzle and uses the Henney's test‑value as an evaluation function.  Cites Schofield.

Nigel Landon & Charles Snape.  A Way with Maths.  CUP, 1984.  Cube moving, pp. 23 & 46.  Consider a 9-puzzle in the usual arrangement:  1 2 3,  4 5 6,  7 8 x.  Move the  1  to the blank position in the minimal number of moves, ignoring what happens to the other pieces.  Generalise.  Their answer only says  13  is minimal for the  3 x 3  board. 

                    My student Tom Henley asked me the  m x n  problem in 1993 and gave a conjectural minimum, which I have corrected to:  if  m = n,  then it can be done in  8m ‑ 11  moves;  but if  n < m,  then it can be done in  6m + 2n - 13  moves, using a straightforward method.  However, I don't see how to show this is minimal, though it seems pretty clear that it must be.  I call this a one-piece problem.  See also Ransom, 1993.

Len Gordon.  Sliding the 15‑1 [sic, but 15‑14 must have been meant] puzzle to magic squares.  G&PJ 4 (Mar 1988) 56.  Reports on computer search to find minimal moves from either ordinary or 15‑14 forms to a magic square.  However, he starts with the blank before the  1,  i.e. as a  0  rather than a  16.

Leonard J. Gordon.  The 16‑15 puzzle or trapezeloyd.  G&PJ 10 (1989) 164.  Introduces his puzzle which has a trapezoidal shape with a triangular wedge in the 2nd and 3rd row so the last row can hold 5 pieces, while the other rows hold four pieces.  Reversing the last two pieces can be done in  85  moves, but this may not be minimal.

George T. Gilbert & Loren C. Larson.  A sliding block problem.  CMJ 23:4 (Sep 1992) 315‑319.  Essentially the same results as obtained by R. Wilson (1974).  Guy points this out in 24:4 (Sep 1993) 355-356.

P. H. R. [Peter H. Ransom].  Adam's move.  Mathematical Pie 128 (Spring 1993) 1017  &  Notes, p. 3.  Considers the one piece problem of Langdon & Snape, 1984.  Solution says the minimal solution on a  n x n  board is  8n - 11,  but doesn't give the answer for the  m x n  board. 

Bernhard Wiezorke.  Sliding caution.  CFF 32 (Aug 1993) 24-25  &  33 (Feb 1994) 32.  In 1986, the German games company ASS (Altenburg Stralsunder Spielkarten AG) produced a game called Vorsicht (= Caution).  Basically this is a  3 x 3  board considered as a doubly crossed square.  It has pieces marked with  +  or  x.  The  +  pieces can only move orthogonally;  the  x  pieces can only move diagonally.  The pieces are coloured and eight are placed on the board to be played as a sliding piece puzzle from given starts to given ends.  The diagonal moves are awkward to make and Wiezorke suggests the board be spread out enough for diagonal moves to be made.  A note at the end says he has received two similar games made by Y. A. D. Games in Israel.

Bala Ravikumar.  The Missing Link and the Top-Spin.  Report TR94-228, Department of Computer Science and Statistics, University of Rhode Island, Jan 1994.  The Missing Link is a cylindrical form of the Fifteen Puzzle, with four layers and four pieces in each layer.  The middle two layers are rigidly joined, but that makes little difference in solving the puzzle.  After outlining the relevant group theory and solving the Fifteen Puzzle, he shows the state space of the Missing Link is  S15.

Richard E. Korf  &  Ariel Felner.  Disjoint pattern database heuristics.  Artificial Intelligence 134 (2002) 9-22.  Discusses heuristic methods of solving the Fifteen Puzzle, Rubik's Cube, etc.  The authors applied their method to 1000 random positions of the Fifteen Puzzle.  The optimal solution length averaged 52.522 and the average time required was 27 msec.  They also did 50 random positions of the Twenty-Four Puzzle and found an average optimal solution length of 100.78, with average time being two days on a 440MHz machine.

 

          5.A.1.          NON‑SQUARE PIECES

 

S&B, pp. 130‑133, show many versions.

See Kinsey, 1878, above, for mention of triangular and diamond‑shaped pieces.

Henry Walton.  US Patent 516,035 -- Puzzle.  Applied: 14 Mar 1893;  patented: 6 Mar 1894.  1p + 1p diagrams.  Described in Hordern, pp. 27 & 68‑69, C1.  4 x 4  area with five  1 x 2  &  two  2 x 1  pieces.

Lorman P. Shriver.  US Patent 526,544 -- Puzzle.  Applied: 28 Jun 1894;  patented: 25 Sep 1894, 2pp + 1p diagrams.  Described in Hordern, p. 27.  4 x 5  area with two  2 x 1  &  15  1 x 1  pieces.  Because there is only one vacant space, the rectangles can only move lengthwise and so this is a dull puzzle.

Frank E. Moss.  US Patent 668,386 -- Puzzle.  Applied: 8 Jun 1900;  patented: 19 Feb 1901.  2pp + 1p diagrams.  Described in Hordern, pp. 27‑28 & 75, C14.  4 x 4  area with six  1 x 1,  two  1 x 2  &  two  2 x 1  pieces, allowing sideways movement of the rectangles.

William H. E. Wehner.  US Patent 771,514 -- Game Apparatus.  Applied: 15 Feb 1904;  patented: 4 Oct 1904.  2pp + 1p diagrams.  First to use  L-shaped pieces.  Described in Hordern, pp. 28 & 107, D5.

Lewis W. Hardy.  US Patent 1,017,752 -- Puzzle.  Applied: 14 Dec 1907;  patented: 20 Feb 1912.  3pp + 1p diagrams.  Described in Hordern, pp. 29 & 89‑90, C43-45.  4 x 5  area with one  2 x 2,  two  1 x 2,  three  2 x 1  &  four  1 x 1  pieces.

L. W. Hardy.  Pennant Puzzle.  Copyright 1909.  Made by OK Novelty Co., Chicago.  No known patent.  Described in Gardner, SA (Feb 1964) = 6th Book, chap. 7 and in Hordern, pp. 28‑29 & 78-79, C19.  4 x 5  area with one  2 x 2,  two  1 x 2,  four  2 x 1,  two  1 x 1  pieces.

Nob Yoshigahara designed Rush Hour in the late 1970s and it was produced in Japan as Tokyo Parking Lot.  Binary Arts introduced it to the US in 1996 and it became very popular.

Winning Ways.  1982.  Pp. 769-777: A trio of sliding block puzzles.  This covers Dad's Puzzler (c19, with piece 8 moved two places to the right), The Donkey (C27d, with all the central pieces moved down one position) and The Century (C42), showing how one can examine partial problems which allow one to consider many positions the same and much reduce the number of positions to be studied.  This allows the graph to be written on a large sheet and solutions to be readily found.

Andrew N. Walker.  Checkmate and other sliding-block puzzles.  Mathematics Preprint Series, University of Nottingham, no. 95-32, 1995, 8pp + covers.  Describes a version by W. G. H. [Wil] Strijbos made by Pussycat.  4 x 4  with an extra position below the left column.  Pieces are alternately black and white and have a black king, a white king and a white rook on them and the object is to produce checkmate, but all positions must be legal in chess, except that the black and white markings do not have to be correct in the intermediate positions.  However, one soon finds that one tile is fixed in place and two other tiles are joined together.  He discusses general computer solving techniques and finds there are five optimal solutions in 68 moves.  He then discusses other problems, citing Winning Ways, Hordern and my Sliding Block Puzzle Circulars.  He gives the UNIX shell scripts that he used.

Ivars Peterson.  Simple puzzles can give computers an unexpectedly strenuous workout.  Science News 162:7 (17 Aug 2002) 6pp PO from their website, http:''sciencenews.org .  Reports on recent work by Gary W. Flake & Eric B. Baum that Nob Yoshigahara's Rush Hour puzzle is PSPACE complete, but is not polynomial time.  Robert A. Hearn and Erik D. Demaine have verified and extended this, showing other sliding block puzzles are PSPACE complete, including the case where all pieces are dominoes and can slide sideways as well as front and back.

 

          5.A.2.          THREE DIMENSIONAL VERSIONS

 

See Hordern, pp. 27, 156‑160 & plates IX & X.

 

P. G. Tait.  Note on the Theory of the "15 Puzzle".  Proc. Roy. Soc. Edin. 10 (1880) 664‑665.  "... conceivable, but scarcely realisable ..."

Charles I. Rice.  US Patent 416,344 -- Puzzle.  Applied: 9 Sep 1889;  patented: 3 Dec 1889.  2pp + 1p diagrams.  Described in Hordern, pp. 27 & 157‑158, G2.  2 x 2 x 2  version with peepholes in the faces.

Ball.  MRE, 1st ed., 1892, p. 78.  Mentions possibility.

Hoffmann.  1893.  Chap. X, No. 1: The John Bull political puzzle, pp. 331 & 357-358  = Hoffmann-Hordern, pp. 215-216.  A  3 x 3  board in the form of a cylinder, with an extra cell attached to one bottom cell.  Pieces can move back and forth around each level, but the connections from one level to the next are all parallel to one of the diagonals -- though this isn't really a complication compared to having vertical connections.  The pieces have two markings: three colours and three letters.  When they are randomly placed on the board, you have to move them so they form a pair of orthogonal  3 x 3  Latin squares.  Fortunately there are such arrangements which differ by an odd permutation, so the puzzle can be solved from any random starting point.  Two examples done.  Says the game is produced by Jaques & Son.

Addison Coe.  US Patent 785,665 -- Puzzle or Game Apparatus.  Applied: 17 Nov 1904;  patented: 1 Mar 1905.  4pp + 3pp diagrams.  Mentioned in Hordern, pp. 158‑159, G3.  Gives a  3 x 5  flat version and a  3 x 3 x 3  cubical version with  3 x 3  arrays of holes in the six faces (in order to push the pieces) and a  3 x 5  cylindrical version.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  The nine disks, pp. 29-34 & 60.  Same as Hoffmann except pieces have colour and shape.

Guy thinks Hein patented Bloxbox, but I have not found any US patent of it -- ??CHECK.

Gardner.  SA (Feb 1973).  First mention of Hein's Bloxbox.

Daniel Kosarek.  US Patent 3,845,959 -- Three-Dimensional Block Puzzle.  Filed: 14 Nov 1973;  patented: 5 Nov 1974.  3pp + 1p diagrams (+ 1p abstract).  Mentioned in Hordern, pp. 158‑159, G3.  3 x 3 x 3  box with  3 x 3  array of portholes on each face.  Mentions  4 x 4 x 4  and larger versions.

Gabriel Nagorny.  US Patent 4,428,581 -- Tri-dimensional Puzzle.  Filed: 16 Jun 1981;  patented: 31 Jan 1984.  Cover page + 3pp + 3pp diagrams.  Three dimensional sliding cube puzzles with central pieces joined together.  A  3 x 3 x 3  version was made in Hungary and marketed as a Varikon Box.  Inventor's address is in France and he cites earlier French applications of 19 Jun 1980 and 19 Nov 1980.  He also describes a  3 x 4 x 4  version with the central areas of each face joined to a  1 x 2 x 2  block in the middle. 

 

          5.A.3.          ROLLING PIECE PUZZLES

 

          Here one has a set of solid pieces in a tray and one tilts or rolls a piece into the blank space.

 

Thomas Henry Ward.

UK Patent 2,870 -- Apparatus for Playing Puzzle or Educational Games.  Provisional: 8 Jun 1883;  Complete as: An Improved Apparatus to be Employed in Playing Puzzle or Educational Games, 6 Dec 1883.  3pp + 1p diagrams.

US Patent 287,352 -- Game Apparatus.  Applied: 13 Sep 1883;  patented: 23 Oct 1883.  1p + 1p diagrams.  Hexagonal board of 19 triangles with 18 tetrahedra to tilt.

George Mitchell  &  George Springfield.  UK Patent 6867 -- A novel puzzle, and improvements in the construction of apparatus therefor.  Applied: 16 Mar 1897;  accepted: 5 Jun 1897.  2pp + 1p diagrams.  Rolling cubes puzzle, where the cube faces are hollowed and fit onto domes in the tray.  Basic form has four cubes in a row with two extra spaces above the middle cubes, but other forms are shown.

Sven Bergling invented the rolling ball labyrinth puzzle/game and they began to be produced in 1946.  [Kenneth Wells; Wooden Puzzles and Games; David & Charles, Newton Abbot, 1983, p. 114.]

Ronald Sprague.  Unterhaltsame Mathematik.  Vieweg, Braunschweig, 1961.  Translated by T. H. O'Beirne as:  Recreations in Mathematics, Blackie, London, 1963.  Problem 3: Schwere Kiste, pp. 3-4 & 22-23  (= Heavy boxes, pp. 4-5 & 25-26).  Three problems with 5 boxes some of which are so heavy that one has to tilt or roll them.

Gardner.  SA (Dec 1963).  = Sixth Book, chap. 8.  Gives Sprague's first problem.

Gardner.  SA (Nov 1965).  c= Carnival, chap. 9.  Prob. 1: The red-faced cube.  Two problems of John Harris involving one cube with one red face rolling on a chessboard.  Gardner says that the field is new and that only Harris has made any investigations of the problem.  The book chapter cites Harris's 1974 article, below, and a 1971 board game called Relate with each player having four coloured cubes on a  4 x 4  board.

Charles W. Trigg.  Tetrahedron rolled onto a plane.  JRM 3:2 (Apr 1970) 82-87.  A tetrahedron rolled on the plane forms the triangular lattice with each cell corresponding to a face of the tetrahedron.  He also considers rolling on a mirror image tetrahedron and rolling octahedra.

John Harris.  Single vacancy rolling cube problems.  JRM 7:3 (1974) 220-224.  This seems to be the first appearance of the problem with one vacant space.  He considers cubes rolling on a chessboard.  Any even permutation of the pieces with the blank left in place is easily obtained.  From the simple observation that each roll is an odd permutation of the pieces and an odd rotation of the faces of a cube, he shows that the parity of the rotation of a cube is the same as the parity of the number of spaces it has moved.  He shows that any such rotation can be achieved on a  2 x 3  board.  Rotating one cube 120o about a diagonal takes 32 moves.  If the blank is allowed to move, the the parity of the permutation of the pieces is the parity of the number of spaces the blank moves, but each cube still has to have the parity of its rotation the same as the parity of the number of spaces it has moved.  If the identical pieces are treated as indistinguishable, the parity of the permutation is only shown by the location of the blank space.  He suggests the use of ridges on the board so that the cube will roll automatically -- this was later used in commercial versions.  He gives a number of problems with different colourings of the cubes.

Gardner.  SA (Mar 1975).  = Time Travel, chap. 9.  Prob. 8: Rolling cubes.  This is the first of Harris's problems.  Computer analysis has found that it can be done in fewer moves than Harris had.  Gardner also reports on the last of Harris's problems, which has also been resolved by computer.

A  3 x 3  array with 8 coloured cubes was available from Taiwan in the early 1980s.  It was called Color Cube Mental Game -- I called it 'Rolling Cubes'.  The cubes had thick faces, producing grooved edges which fit into ridges in the bottom of the plastic frame, causing automatic rolling quite nicely.  I wonder if this was inspired by Harris's article.

John Ewing & Czes Kośniowski.  Puzzle it Out -- Cubes, Groups and Puzzles.  CUP, 1982.  The 8 Cubes Puzzle, pp. 58-59.  Analysis of the Rolling Cubes puzzle.  The authors show how to rotate a single cube about a diagonal in  36  moves. 

 

          5.A.4.          PANEX PUZZLE

 

          Invented by Toshio Akanuma (??SP).  Manufactured by Tricks Co., Japan, in 1983.  Described in Hordern, pp. 144-145 & 220, E35, and in S&B, p. 135.  This looks like a Tower of Hanoi (cf 7.M.2) with two differently coloured piles of 10 pieces on the outside two tracks of three tracks of height 12 joined like a letter E.  This is made as a sliding block puzzle, but with blockages -- a piece cannot slide down a track further than its original position.

 

Mark Manasse,  Danny Sleator  &  Victor K. Wei.  Some Results on the Panex Puzzle.  Preprint sent by Jerry Slocum, 23pp, nd [1983, but S&B gives 1985].  For piles of size  n,  the minimum number of moves,  T(n),  to move one pile to the centre track is determined by means of a 2nd order, non-homogeneous recurrence which has different forms for odd and even  n.  Compensating for this leads to a 2nd order non‑homogeneous recurrence, giving  T(10) = 4875  and  T(n) ~  C(1 + Ö2)n.  This solution doesn't ever move the other pile.  The minimum number of moves,  X(n),  to exchange the piles is bounded above and below and determined exactly for  £ 6  by computer search.  X(5) = 343,  compared to bounds of  320  and  343.  X(6) = 881,  compared to the bounds of  796  and  881.  For  n = 7,  the bounds are  1944  and  2189,  For  n = 10,  the bounds are  27,564  and  31,537.  The larger bounds are considered as probably correct.

Christoph Hausammann.  US Patent 5,261,668 -- Logic Game.  Filed: 6 Aug 1992;  patented: 16 Nov 1993.  1p abstract + 2pp text + 3pp diagrams.  Essentially identical to Panex.

Vladimir Dubrovsky.  Nesting Puzzles -- Part I: Moving oriental towers.  Quantum 6:3 (Jan/Feb 1996) 53-59 & 49-51.  Says Panex was produced by the Japanese Magic Company in the early 1980s.  Discusses it and cites S&B for the bounds given above.  Sketches a number of standard configurations and problems, leading to "Problem 9.  Write out a complete solution to the Panex puzzle."  He says his method is about 1700 moves longer than the upper bound given above.

Nick Baxter.  Recent results for the Panex Puzzle.  4pp handout at G4G5, 2002.  Describes the puzzle and its history.  David Bagley wrote a program to implement the Manasse, Sleator & Wei methods.  On 7 Feb 2002, this confirmed the conjecture that  X(7) = 2189.  On 26 Mar 2002, it obtained  X(8) = 5359,  compared to bounds of  4716  and  5359.  It is estimated that the cases  n = 9 and 10  will take 10 and 1200 years!  If Moore's Law on the increase of computing power continues for another 20 years, the latter answer may be available by then.  He gives a simplified version of the algorithm for the upper bound, which gets  31,544  for  n = 10.  He has a Panex page:  www.baxterweb.com/puzzles/panex/  and will be publishing an edited and annotated version of the Manasse, Sleator & Wei paper on it.

 

5.B.   CROSSING PROBLEMS

 

See MUS I 1-13, Tropfke 658 and also 5.N.

 

Wolf, goat and cabbages:  Alcuin,  Abbot Albert,  Columbia Algorism,  Munich 14684,  Folkerts,  Chuquet,  Pacioli,  Tartaglia,  van Etten,  Merry Riddles,  Ozanam,  Dilworth,  Wingate/Dodson,  Jackson,  Endless Amusement II,  Boy's Own Book,  Nuts to Crack,  Taylor; The Riddler,  Child,  Fireside Amusements,  Magician's Own Book,  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Secret Out (UK),  Mittenzwey,  Carroll 1873,  Kamp,  Carroll 1878,  Berg,  Lemon,  Hoffmann,  Brandreth Puzzle Book,  Carroll 1899,  King,  Voggenreiter,  Stein,  Stong,  Zaslavsky,  Ascher,  Weismantel (a film), 

     Verse version:  Taylor,

Version with only one pair of incompatibles:  Voggenreiter

Extension to four items:  Gori,  Phillips,  M. Adams,  Gibbs,  Ascher

Adults and children:  Alcuin,  Kamp,  Hoffmann,  Parker?,  Voggenreiter,  Gibbs

Three jealous husbands:  Alcuin,  Abbot Albert,  Columbia Algorism,  Munich 14684,  Folkerts,  Rara,  Chuquet,  Pacioli,  Cardan,  Tartaglia,  H&S ‑ Trenchant,  Gori,  Bachet,  van Etten,  Wingate/Kersey,  Ozanam,  Minguét,  Dilworth,  Les Amusemens,  Wingate/Dodson,  Jackson,  Endless Amusement II,  Nuts to Crack,  Young Man's Book,  Family Friend,  Magician's Own Book,  The Sociable,  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Vinot,  Secret Out (UK),  Lemon,  Hoffmann,  Fourrey,  H. D. Northrop,  Mr. X,  Loyd,  Williams,  Clark,  Goodstein,  O'Beirne,  Doubleday,  Allen, 

     Verse mnemonic:  Abbot Albert,  Munich 14684,

     Verse solution:  Ozanam,  Vinot,

Four or more jealous husbands:  Pacioli,  Filicaia,  Tartaglia,  Bachet,  Delannoy,  Ball,  Carroll-Collingwood,  Dudeney,  O'Beirne

Jealous husbands, with island in river:  De Fontenay,  Dudeney,  Ball,  Loyd,  Dudeney,  Pressman & Singmaster

Missionaries and cannibals:  Jackson,  Mittenzwey,  Cassell's,  Lemon,  Pocock,  Hoffmann,  Brandreth Puzzle Book,  H. D. Northrop,  Schubert,  Arbiter,  H&S,  Abraham,  Bile Beans, Goodstein,  Beyer,  O'Beirne,  Pressman & Singmaster.

     With only one cannibal who can row:  Brandreth Puzzle Book,  Abraham,  Beyer.

Bigger boats:  Pacioli,  Filicaia?,  Bachet(-Labosne),  Delannoy,  Ball,  Dudeney,  Abraham?,  Goodstein,  Kaplan,  O'Beirne,

 

Alcuin.  9C.

Prob. 17: Propositio de tribus fratribus singulas habentibus sorores.  3 couples, rather earthily expressed.

Prob. 18: Propositio de lupo et capra et fasciculo cauli.  Wolf, goat, cabbages.

Prob. 19: Propositio de viro et muliere ponderantibus plaustrum.  Man, wife and two small children.

Prob. 20: Propositio de ericiis.  Rewording of Prob. 19.

Ahrens.  MUS II 315‑318, cites many sources, mostly from folklore and riddle collections, with one from the 12C and several from the 14C.  ??NYS.

Abbot Albert.  c1240.

Prob. 5, p. 333.  Wolf, goat & cabbages.

Prob. 6, p. 334.  3 couples, with verse mnemonic.

Columbia Algorism.  c1350. 

No. 122, pp. 130‑131 & 191: wolf, goat, bundle of greens.  See also Cowley 402 & plate opposite.  P. 191 and the Cowley plate are reproductions of the text with a crude but delightful illustration.  P. 130 gives a small sketch of the illustration.  I have a colour slide from the MS.

No. 124, p. 132: 3 couples.  See also Cowley 403 & plate opposite.  The plate shows another crude but delightful illustration.  I have a colour slide from the MS.

Munich 14684.  14C. 

Prob. XXVI, pp. 82‑83: 3 couples, with verse mnemonic.

Prob. XXVII, p. 83: wolf, goat, cabbage.

Folkerts.  Aufgabensammlungen.  13-15C.  11 sources with wolf, goat, cabbage.  12 sources with three jealous couples.

Rara, 459‑465, cites two Florentine MSS of c1460 which include 'the jealous husbands'.  ??NYS.

Chuquet.  1484. 

Prob. 163: wolf, goat & cabbages.  FHM 233 says that a 12C MS claims that every boy of five knows this problem.

Prob. 164: 3 couples.  FHM 233.

Pacioli.  De Viribus.  c1500. 

Ff. 103v - 105v.  LXI. C(apitolo). de .3. mariti et .3. mogli gelosi (About 3 husbands and 3 wives).  = Peirani 146-148.  3  couples.  Says that  4  or  5  couples requires a  3  person boat.

F. IIIv.  = Peirani 6.  The Index lists the above as Problem 66 and lists a Problem 65: Del modo a salvare la capra el capriolo dal lupo al passar de un fiume ch' non siano devorati (How to save the goat and the kid from the wolf in crossing a river so they are not eaten).

Piero di Nicolao d'Antonio da Filicaia.  Libro dicto giuochi mathematici.  Early 16C -- ??NYS, mentioned in Franci, op. cit. in 3.A.  Franci, p. 23, says Pacioli and Filicaia deal with the case of four or five couples and that Pacioli considers bigger boats, but I'm not clear if Filicaia also does so.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 73, f. FF.v.v (p. 157).  (The 73 is not printed in the Opera Omnia).  Three jealous husbands.

Tartaglia.  General Trattato, 1556, art. 141‑143, p. 257r‑ 257v. 

Art. 141: wolf, goat and cabbages.

Art. 142: three couples.

Art. 143: four couples -- erroneously -- see Bachet.

H&S 51 says 3 couples occurs in Trenchant (1566), ??NYS.

Gori.  Libro di arimetricha.  1571.

Ff. 71r‑71v (p. 77).  3  couples. 

F. 80v (p. 77).  Dog, wolf, sheep, horse to cross river in boat which holds  2,  but each cannot abide his neighbours in the given list, so each cannot be alone with such a neighbour.

Bachet.  Problemes.  1612.  Addl. prob. IV: Trois maris jaloux ...,  1612: 140-142;  1624: 212‑215;  1884: 148‑153.  Three couples;  four couples -- notes that Tartaglia is wrong by showing that one can never get five persons on the far side.  Labosne gives a solution with a  3  person boat and does  n  couples with an  n‑1  person boat.

van Etten.  1624. 

Prob. 14: Des trois maistres & trois valets, p. 14.  3  men and  3  valets.  (The men hate the other valets and will beat them if given a chance.)  (Not in English editions.)

Prob. 15: Du loup, de la chevre & du chou, pp. 14‑15.  Wolf, goat & cabbages.  (Not in English editions.)

Book of Merry Riddles.  1629  72 Riddle, pp. 43-44.  "Over a water I must passe, and I must carry a lamb, a woolfe, and a bottle of hay if I carry any more than one at once my boat will sink."  Tony Augarde; The Oxford Guide to Word Games; OUP, 1984; p. 6 says wolf, goat, cabbage appears in the 1629 ed.

Wingate/Kersey.  1678?.  Prob. 6., p. 543.  Three jealous couples.  Cf 1760 ed.

Ozanam.  1725.

Prob. 2, 1725: 3‑4.  Prob. 18, 1778: 171;  1803: 171;  1814: 150.  Prob. 17, 1840: 77.  Wolf, goat and cabbage. 

Prob. 3, 1725: 4‑5.  Prob. 19, 1778: 171-172;  1803: 171-172;  1814: 150-151.  Prob. 18, 1840: 77.  Jealous husbands.  Latin verse solution.  He also discusses three masters and valets:  "none of the the masters can endure the valets of the other two; so that if any one of them were left with any of the other two valets, in the absence of his master, he would infallibly cane him."

Minguet.  1733.  Pp. 158-159 (1755: 114-115; 1822: 175-176; 1864: 151).  Three jealous couples.

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.

Problem 6: Fox, goose and peck of corn.  = D. Adams; Scholar's Arithmetic; 1801, p. 200, no. 8.

Problem 7: Three jealous husbands.  (Dilworth cites Wingate for this -- but this is in Kersey's additions -- cf Wingate/Kersey, 1678? above.)  = D. Adams; Scholar's Arithmetic; 1801, p. 200, no. 9.

Les Amusemens.  1749.  Prob. 14, p. 136: Les Maris jaloux.  Solution is incorrect and has been corrected by hand in my copy.

Edmund Wingate (1596-1656).  A Plain and Familiar Method for Attaining the Knowledge and Practice of Common Arithmetic.  ....  19th ed., previous ed. by John Kersey (1616-1677) and George Shell(e)y, now by James Dodson.  C. Hitch and L. Hawes, et al., 1760. 

                    Art. 749.  Prob. VI.  P. 379.  Three jealous husbands.  As in 1678? ed.

                    Art. 750.  Prob. VII.  P. 379.  Fox, goose and corn.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 7,  pp. 2 & 52.  Fox, goose and corn.  One solution.

No. 13, pp. 4 & 54.  Three jealous husbands.

No. 21, pp. 5 & 56.  Three masters and servants, where the servants will murder the masters if they outnumber them -- i.e. missionaries and cannibals.  First appearance of this type.

Endless Amusement II.  1826? 

Prob. 17, pp. 198-199.  Wolf, goat and cabbage.

Prob. 25, pp. 201-202.  Three jealous husbands.

Boy's Own Book.  The wolf, the goat and the cabbages.  1828: 418‑419;  1828-2: 423;  1829 (US): 214;  1855: 570;  1868: 670.

Nuts to Crack III (1834).

No. 209.  Fox, goose and peck of corn.

No. 214.  Three jealous husbands.

The Riddler.  1835.  The wolf, the goat and the cabbages, pp. 5-6.  Identical to Boy's Own Book.

Young Man's Book.  1839.  Pp. 39-40.  Three jealous Husbands ..., identical to Wingate/Kersey.

Child.  Girl's Own Book.  1842: Enigma 49, pp. 237-238;  1876: Enigma 40, p. 200.  Fox, goose and corn.  Says it takes four trips instead of three -- but the solution has  7  crossings.

Walter Taylor.  The Indian Juvenile Arithmetic, or Mental Calculator; to which is added an appendix, containing arithmetical recreations and amusements for leisure hours ....  For the author at the American Press, Bombay, 1849.  [Quaritch catalogue 1224, Jun 1996, says their copy has a note in French that Ramanujan learned arithmetic from this and that it is not in BMC nor NUC.  Graves 14.c.35.]  P. 211, No. 8.  Wolf, goat and cabbage in verse!  No solution.

                    Upon a river's brink I stand, it is both deep and wide;

                    With a wolf, a goat, and cabbage, to take to the other side.

                    Tho' only one each time can find, room in my little boat;

                    I must not leave the goat and wolf, not the cabbage and the goat.

                    Lest one should eat the other up, -- now how can it be done --

                    How can I take them safe across without the loss of one?

Fireside Amusements.  1850: No. 24, pp. 111 & 181;  1890: No. 24, p. 100.  Fox, goose and basket of corn.

Family Friend 3 (1850) 344 & 351.  Enigmas, charades, etc. -- No. 17: The three jealous husbands.

Magician's Own Book.  1857.

The three jealous husbands, p. 251. 

The fox, goose, and corn, p. 253. 

The Sociable.  1858.  Prob. 33: The three gentlemen and their servants, pp. 296 & 314-315.  "None of the gentlemen shall be left in company with any of the servants, except when his own servant is present" -- so this is like the Jealous Husbands.  = Book of 500 Puzzles, 1859, prob. 33, pp. 14 & 32-33.  = Illustrated Boy's Own Treasury, 1860, prob. 11, pp. 427-428 & 431.

Book of 500 Puzzles.  1859.

Prob. 33: The three gentlemen and their servants, pp. 14 & 32-33.  As in The Sociable.

The three jealous husbands, p. 65.

The fox, goose and corn, p. 67.

Both identical to Magician's Own Book.

Boy's Own Conjuring Book.  1860.

The three jealous husbands, pp. 222‑223. 

The fox, goose, and corn, pp. 225. 

Both identical to Magician's Own Book.

Vinot.  1860.  Art. XXXVII: Les trois maris jaloux, pp. 56-57.  Three jealous husbands, with verse solution taken from Ozanam.

The Secret Out (UK).  c1860.  A comical dilemma, p. 27.  Wolf, goat and cabbage.  Varies it as fox, goose and corn and then as gentlemen and servants, which is jealous husbands, rather than the same problem.

Lewis Carroll.  Letter of 15 Mar 1873 to Helen Feilden.  Pp. 212-215 (Collins: 154-155).  Fox, goose and bag of corn.  "I rashly proposed to her to try the puzzle (I daresay you know it) of "the fox, and goose, and bag of corn.""  Cf Carroll-Collingwood, pp. 212-215 (Collins: 154-155); Carroll-Wakeling, prob. 28, pp. 36-37 and Carroll-Gardner, p. 51.  Cf Carroll, 1878.  Wakeling writes that this does not appear elsewhere in Carroll.

Bachet-Labosne.  1874.  For details, see Bachet, 1612. 

Jens Kamp.  Danske Folkeminder, Aeventyr, Folksagen, Gaader, Rim og Folketro, Samlede fra Folkemende.  R. Neilsen, Odense, 1877.  Marcia Ascher has kindly sent me a photocopy of the relevant material with a translation by Viggo Andressen.

No. 18, pp. 326‑327: Fox, lamb and cabbage.

No. 19, p. 327: Husband, wife and two half‑size sons.

Lewis Carroll.  Letter of 22 Jan 1878 to Jessie Sinclair.  Fox, goose and bag of corn.  Cf Carroll-Collingwood, pp. 205-207 (Collins: 150); Carroll-Wakeling, prob. 26: The fox, the goose and the bag of corn, pp. 34 & 72.  Cf Carroll. 1872.

Mittenzwey.  1880.  Prob. 227-228, pp. 42 & 92;  1895?: 254-255, pp. 46 & 94;  1917: 254‑255, pp. 42 & 90.  Bear, goat and cabbage, mentioning second solution;  three kings and three servants, where the servants will rob the kings if they outnumber them, i.e. like missionaries and cannibals.

Cassell's.  1881.  P. 105: The dishonest servants.  The servants are rogues who will murder masters if they outnumber them, so this is equivalent to the missionaries and cannibals version.

Lucas.  RM1.  1882.  Pp. 1-18 is a general discussion of the problem.

De Fontenay.  Unknown source and date -- 1882??  Described in RM1, 1882, pp. 15‑18 (check 1st ed.??).  n > 3  couples,  2  person boat, island in river, can be done in  8n ‑ 8  passages.  Lucas says this was suggested at the Congrès de l'Association française pour l'avancement des sciences at Montpellier in 1879, ??NYS.  (De Fontenay is unclear -- sometimes he permits bank to bank crossings, other times he only permits bank to island crossings.  His argument really gives  8n - 6  if bank to bank crossings are prohibited.  See Pressman & Singmaster, below, for clarification.)

Albert Ellery Berg, ed.  Op. cit. in 4.B.1.  1883.  P. 377: Fox, goose & peck of corn.

Lemon.  1890.

Gentlemen and their servants, no. 101, pp. 17‑18 & 101.  This is the same as missionaries and cannibals.

The three jealous husbands, no. 151, pp. 24 & 103 (= Sphinx, no. 478, pp. 66 & 114.)  The solution mentions Alcuin.

Crossing the river, no. 450, pp. 59 & 114.  English travellers and native servants  = missionaries and cannibals.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 136, no. 14.  Fox, goose and corn.  No solution.

Herbert Llewelyn Pocock.  UK Patent 15,358 -- Improvements in Toy Puzzles.  Applied: 29 Sep 1890;  complete specification: 29 Jun 1891;  accepted: 22 Aug 1891.  2pp + 1p diagrams.  Three whites and three blacks and the blacks must never outnumber the whites, i.e. same as missionaries and cannibals.  He describes the puzzle as "well known".

Delannoy.  Described in RM1, 1891, Note 1:  Sur le jeu des traversées, pp. 221‑222.  ??check 1882 ed.  Shows  n  couples can cross in an  x  person boat in  N  trips, for  n, x, N  =  2, 2, 5;  3, 2, 11;  4, 3, 9;  5, 3, 11;  n > 5, 4, 2n ‑ 3.  (He has  2n ‑ 1  by mistake.  Simple modification shows we also have  5, 4, 7;  6, 5, 9;  7, 6, 5;  8, 7, 7;  n > 8, n ‑ 1, 5.)

Ball.  MRE, 1st ed., 1892, pp. 45‑47, says Lucas posed the problem of minimizing  x  for a given  n  and quotes the Delannoy solution (with erroneous  2n ‑ 1)  and also gives De Fontenay's version and solution.  (He spells it De Fonteney as does his French translator, though Ahrens gives De Fontenay and the famous abbey in Burgundy is Fontenay -- ??)

The Ballybunnion and Listowel Railway in County Kerry, Ireland, was a late 19C railway using the Lartigue monorail system.  This had a single rail, about three feet off the ground, with a carriage hanging over both sides of the rail.  The principle job of the conductor/guard to make sure the passengers and goods were equally distributed on both sides.  Kerry legend asserts that a piano had to be sent on this railway and there were not enough passengers or goods to balance it.  So a cow was sent on the other side.  At the far end, the piano was unloaded and replaced with two large calves and the carriage sent back.  The cow was then unloaded and one calf moved to the other side, so the carriage could be sent back to the far end and everyone was happy.

Hoffmann.  1893.  Chap. IV, pp. 157‑158 & 211‑213 = Hoffmann-Hordern, pp. 136-138, with photos.

No. 56: The three travellers.  Masters and servants, equivalent to missionaries and cannibals.  Solution says Jaques & Son make a puzzle version with six figures, three white and three black.  Photos in Hoffmann-Hordern, pp. 136 & 137 -- the latter shows Caught in the Rain, 1880-1905, where Preacher, Deacon, Janitor and their wives have to get somewhere using one umbrella.

No. 57: The wolf, the goat, and the cabbages.  Photo on p. 136 of La Chevre et le Chou. with box, by Watilliaux, 1874-1895.  Hordern Collection, p. 72, and S&B, p. 134, show the same puzzle.)

No. 58: The three jealous husbands.

No. 59: The captain and his company.  This is Alcuin's prop. 19 with many adults.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895]. 

P. 7: The wolf, the goat and the cabbages.  Identical to Hoffmann No. 57, with nice colour picture.  No solution.

P. 9: The missionaries' and cannibals' puzzle.  Usual form, with nice colour picture, but only one cannibal can row.  No solution.  This seems to be the first to use the context of missionaries and cannibals and the first to restrict the number of rowers.

Lucas.  L'Arithmétique Amusante.  1895.  Les vilains maris jaloux, pp. 125-144  &  Note II, pp. 198-202. 

Prob. XXXVI: La traversée des trois ménages, pp. 125-130.  3  couples.  Gives Bachet's 1624 reasoning for the essentially unique solution -- but attributes it to 1613.

Prob. XXXVII: La traversée des quatre ménages, pp. 130-132.  4  couples in a  3  person boat done in  9  crossings.

L'erreur de Tartaglia, pp. 133-134.  Discusses Tartaglia's error and Bachet's notice of it and gives an easy proof that  4  couples cannot be done with a  2  person boat.

Prob. XXXVIII: La station dans une île, pp. 135-140.  4  couples,  2  person boat, with an island.  Gives De Fontenay's solution in 24 crossings.

Prob. XXXIX: La traversée des cinq ménages, pp. 141-143.  5  couples,  3  person boat in  11  crossings.

Énoncé général du problème des traversées, pp. 143-144.  n  couples,  x  person boat, can be done in  N  crossings as given by Delannoy above.  He corrects  2n - 1  to  2n - 3  here.

Note II:  Sur les traversées, pp. 198-202.  Gives Tarry's version with an island and with  n  men having harems of size  m,  where the women are obviously unable to row.  He gives solutions in various cases.  For the ordinary case, i.e.  m = 1,  he finds a solution for  4  couples in  21  moves, using the basic ferrying technique that Pressman and Singmaster found to be optimal, but the beginning and end take longer because the women cannot row.  He says this gives a solution for  n  couples in  4n + 5  crossings.  He then considers the case of  n - 1  couples and a ménage with  m  wives and finds a solution in  8n + 2m + 7  crossings.  I now see that this solution has the same defects as those in Pressman & Singmaster, qv.

Ball.  MRE, 3rd ed., 1896, pp. 61‑64, repeats 1st ed., but adds that Tarry has suggested the problem for harems -- see above.

Dudeney.  Problem 68: Two rural puzzles.  Tit‑Bits 33 (5 Feb  &  5 Mar 1898) 355  &  432.  Three men with sacks of treasure and a boat that will hold just two men or a man and a sack, with additional restrictions on who can be trusted with how much.  Solution in  13  crossings.

Carroll-Collingwood.  1899.  P. 317 (Collins: 231 or 232 (missing in my copy))  Cf Carroll-Wakeling II, prob. 10: Crossing the river, pp. 17 & 66.  Four couples -- only posed, no solution.  Wakeling gives a solution, but this is incorrect.  After one wife is taken across, he has another couple coming across and from Bachet onward, this is considered improper as the man could get out of the boat and attack the first, undefended, wife.

E. Fourrey.  Op. cit. in 4.A.1, 1899.  Section 211: Les trois maîtres et les trois valets.  Says a master cannot leave his valet with the other masters for fear that they will intimidate him into revealing the master's secrets.  Hence this is the same as the jealous couples.

H. D. Northrop.  Popular Pastimes.  1901.

No. 5: The three gentlemen and their servants, pp. 67 & 72.  = The Sociable.

No. 12: The dishonest servants, pp. 68 & 73.  "... the servants on either side of the river should not outnumber the masters", so this is the same as missionaries and cannibals.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:2 (Jun 1903) 140-141.  A matrimonial difficulty.  Three couples.  No answer given.

Dudeney.  Problem 523.  Weekly Dispatch (15  &  29 Nov 1903), both p. 10,  (= AM, prob. 375, pp. 113 & 236‑237).  5  couples in a  3  person boat.

Johannes Bolte.  Der Mann mit der Ziege, dem Wolf und dem Kohle.  Zeitschrift des Vereins für Volkskunde 13 (1903) 95-96 & 311.  The first part is unaware of Alcuin and Albert. He gives a 12C Latin solution:  It capra, fertur olus, redit hec, lupus it, capra transit [from Wattenbach; Neuen Archiv für ältere deutsche Geschichtskunde 2 (1877) 402, from Vorauer MS 111, ??NYS]  and a 14C solution:  O natat, L sequitur, redit O, C navigat ultra, / Nauta recurrit ad O, bisque natavit ovis  (= ovis, lupus, ovis, caulis, ovis) [from Mone; Anzeiger für Kunde der deutschen Vorzeit 45 (No. 105) (1838), from Reims MS 743, ??NYS].  Cites Kamp and several other versions, some using a fox, a sheep, or a lamb.  The addendum cites and quotes Alcuin and Albert as well as relatively recent French and Italian versions.

H. Parker.  Ancient Ceylon.  Op. cit. in 4.B.1.  1909.  Crossing the river, p. 623. 

A King, a Queen, a washerman and a washerwoman have to cross a river in a boat that holds two.  However the King and Queen cannot be left on a bank with the low caste persons, though they can be rowed by the washerperson of the same sex.  Solution in  7  crossings.

Ferry-man must transport three leopards and three goats in a boat which holds himself and two others.  If leopards ever outnumber goats, then the goats get eaten.  So this is like missionaries and cannibals, but with a ferry-man.  Solution in 9 crossings.

H. Schubert.  Mathematische Mussestunde.  Vol. 2, 3rd ed., Göschen, Leipzig, 1909.  Pp. 160‑162: Der drei Herren und der drei Sklaven.  (Same as missionaries and cannibals.)

Arbiter Co. (Philadelphia).  1910.  Capital and Labor Puzzle.  Shown in S&B, p. 134.  Equivalent to missionaries and cannibals.

Ball.  MRE, 5th ed., 1911, pp. 71-73, repeats 3rd ed., but omits the details of De Fonteney's solution in  8(n-1)  crossings.

Loyd.  Cyclopedia, 1914. 

Summer tourists, pp. 207 & 366.  3  couples,  2  person boat, with additional complications -- the women cannot row and there have been some arguments.  Solution in  17  crossings.

The four elopements, pp. 266 & 375.  4  couples,  2  person boat, with an island and the stronger constraint that no man is to get into the boat alone if there is a girl alone on either the island or the other shore.  "The [problem] presents so many complications that the best or shortest answer seems to have been overlooked by mathematicians and writers on the subject."  "Contrary to published answers, ... the feat can be performed in  17  trips, instead of  24."

Ball.  MRE, 6th ed., 1914, pp. 71-73, repeats 5th ed., but adds that  6n ‑ 7  trips suffices for  n  couples with an island, though he gives no reference.

Williams.  Home Entertainments.  1914.  Alcuin's riddle, pp. 125-126.  "This will be recognized as perhaps the most ancient British riddle in existence, though there are several others conceived on the same lines."  Three jealous couples. 

Clark.  Mental Nuts.  1916, no. 67.  The men and their wives.  "... no man shall be left alone with another's wife."

Dudeney.  AM.  1917.  Prob. 376: The four elopements, pp. 113 & 237.  4  couples,  2  person boat, with island, can be done in  17  trips and that this cannot be improved.  This is the same solution as given by Loyd.  (See Pressman and Singmaster, below.)

Ball.  MRE, 8th ed., 1919, pp. 71-73 repeats 6th ed. and adds a citation to Dudeney's AM prob. 376 for the solution in  6n ‑ 7  trips for  n  couples.

Hummerston.  Fun, Mirth & Mystery.  1924.  Crossing the river puzzles, Puzzle no, 52, pp. 128 & 180.  'Puzzles of this type ... interested people who lived more than a thousand years ago'.

No. 1: The eight travellers.  Six men and two boys who weigh half as much.

No. 2: White and black.  = Missionaries and cannibals.

No. 3: The fox, the goose, and the corn.

No. 4: the jealous husbands.

H&S, 1927, p. 51 says missionaries and cannibals is 'a modern variant'.

King.  Best 100.  1927.  No. 10, pp. 10 & 40.  Dog, goose and corn.

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.

P. 106: Der Wolf, die Ziege und der Kohlkopf.  Usual wolf, goat, cabbage.

Pp. 106-107: Die 100 Pfund-Familie.  Parents weigh 100 pounds; the two children weigh 100 pounds together.

P. 107: Der Landjäger and die Strolche [The policeman and the vagabonds].  Two of the vagabonds hate each other so much that they cannot be left together.  As far as I recall, this formulation is novel and I was surprised to realise that it is essentially equivalent to the wolf, goat and cabbage version.

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 41, pp. 22 & 194.  Rowing explorer with 4 natives:  A, B, C, D,  who cannot abide their neighbours in this list.  A  can row.  They get across in seven trips.

Abraham.  1933.  Prob. 54 -- The missionaries at the ferry, pp. 18 & 54 (14 & 115).  3  missionaries and  3  cannibals.  Doesn't specify boat size, but says 'only one cannibal can row'.  1933 solution says 'eight double journeys', 1964 says 'seven crossings'.  This seems to assume the boat holds  3.  (For a  2  man boat, it takes  11  crossings with one missionary and two cannibals who can row or  13  crossings with one missionary and one cannibal who can row.)

The Bile Beans Puzzle Book.  1933.  No. 34: Missionaries & cannibals.  Three of each but only one of each can row.  Done in 13 crossings.

Phillips.  Brush.  1936.  Prob. L.2: Crossing the Limpopo, pp. 39‑40 & 98.  Same as in Week‑End, 1932.

M. Adams.  Puzzle Book.  1939.  Prob. C.63: Going to the dance, pp. 139 & 178.  Same as Week‑End, 1932, phrased as travelling to a dance on a motorcycle which carries one passenger.

R. L. Goodstein.  Note 1778:  Ferry puzzle.  MG 28 (No. 282) (1944) 202‑204.  Gives a graphical way of representing such problems and considers  m  soldiers and  m  cannibals with an  n  person boat,  3  jealous husbands and how many rowers are required.

David Stein.  Party and Indoor Games.  P. M. Productions, London, nd [c1950?].  P. 98, prob. 5: Man with cat, parrot and bag of seeds.

C. L. Stong.  The Amateur Scientist.  Ill. by Roger Hayward.  S&S, 1960. 

A puzzle-solving machine, pp. 377-384.  Describes how Paul Bezold made a logic machine from relays to solve the fox, goose, corn problem.

How to design a "Pircuit" or Puzzle circuit, pp. 388-394.  On pp. 391-394, Harry Rudloe describes relay circuits for solving the three jealous couples problem, which he attributes to Tartaglia, and the missionaries and cannibals problem.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Mentor (New American Library), NY, 1961.  [John Fauvel sent some pages from a different printing which has much different page numbers than my copy.]  "River crossing" problems, pp. 168‑171.  Discusses various forms of the problem and adds a problem with two parents weighing  160,  two children weighing  80  and a dog weighing  12,  with a boat holding  160.

E. A. Beyer, proposer;  editorial solution.  River‑crossing dilemma.  RMM 4 (Aug 1961) 46  &  5 (Oct 1961) 59.  Explorers and natives (= missionaries and cannibals), with all the explorers and one native who can row.  Solves in  13  crossings, but doesn't note that only one rowing explorer is needed.  (See note at Abraham, 1933, above.)

Philip Kaplan.  Posers.  (Harper & Row, 1963);  Macfadden Books, 1964.  Prob. 36, pp. 41 & 91.  5  men and a  3  person boat on one side,  5  women on the other side.  One man and one woman can row.  Men are not allowed to outnumber women on either side nor in the boat.  Exchange the men and the women in  7  crossings.

T. H. O'Beirne.  Puzzles and Paradoxes, 1965, op. cit. in 4.A.4, chap. 1, One more river to cross, pp. 1‑19.  Shows  2n ‑ 1  couples (or  2n ‑ 1  each of missionaries and cannibals ?) can cross in a  n  person boat in  11  trips.  2n ‑ 2  can cross in  9  trips.  He also considers variants on Gori's second version.

Doubleday - 2.  1971.  Family outing, pp. 49-50.  Three couples, but one man has quarrelled with the other men and his wife has quarrelled with the other women, so this man and wife cannot go in the boat nor be left on a bank with others of their sex.  Further men cannot be outnumbered by women on either bank.  Gives a solution in  9  crossings, but I find the conditions unworkable -- e.g. the initial position is prohibited!

Claudia Zaslavsky.  Africa Counts.  Prindle, Weber & Schmidt, Boston, 1973.  Pp. 109‑110 says that leopard, goat and pile of cassava leaves is popular with the Kpelle children of Liberia.  However, Ascher's Ethnomathematics (see below), p. 120, notes that this is based on an ambiguous description and that an earlier report of a Kpelle version has the form described below.

Ball.  MRE, 12th ed., 1974, p. 119, corrects Delannoy's  2n ‑ 1  to  2n ‑ 3  and corrects De Fontenay's  8n ‑ 8  to  8n ‑ 6,  but still gives the solution for  n = 4  with  24  crossings.

W. Gibbs.  Pebble Puzzles -- A Source Book of Simple Puzzles and Problems.  Curriculum Development Unit, Solomon Islands, 1982.  ??NYS, o/o??.  Excerpted in:  Norman K. Lowe, ed.; Games and Toys in the Teaching of Science and Technology; Science and Technology Education, Document Series No. 29, UNESCO, Paris, 1988, pp. 54‑57.  On pp. 56‑57 is a series of river crossing problems.  E.g. get people of weights  1, 2, 3  across with a boat that holds a weight of at most  3.  Also people numbered  1, 2, 3, 4, 5  such that no two consecutive people can be in the boat or left together.

In about 1986, James Dalgety designed interactive puzzles for Techniquest in Cardiff.  Their version has a Welshman with a dragon, a sheep and a leek!

Ian Pressman & David Singmaster.  Solutions of two river crossing problems: The jealous husbands and the missionaries and the cannibals.  Extended Preprint, April 1988, 14pp.  MG 73 (No. 464) (Jun 1989) 73‑81.  (The preprint contains historical and other detail omitted from the article as well as some further information.)  Observes that De Fontenay seems to be excluding bank to bank crossings and that Lucas' presentation is cryptic.  Shows that De Fontenay's method should be  8n ‑ 6  crossings for  n > 3  and that this is minimal.  If bank to bank crossings are permitted, as by Loyd and Dudeney, a computer search revealed a solution with  16  crossings for  n = 4,  using an ingenious move that Dudeney could well have ignored.  For  n > 4,  there is a simple solution in  4n + 1  crossings, and these numbers are minimal.  [When this was written, I had forgotten that Loyd had done the problem for  4  couples in  17  moves, which changes the history somewhat.  However, I now see that Loyd was copying from Dudeney's Weekly Dispatch problem 270 of 23 Apr 1899 & 11 Jun 1899.  Loyd states what appears to be a stronger constraint but all the methods in our article do obey the stronger constraint.  However, one could make the constraints stronger -- e.g. our solutions have a husband taking the boat from bank to bank while his wife and another wife are on the island -- the solution of Loyd & Dudeney avoids this and may be minimal in this case --??.]

                    For the missionaries and cannibals problem, the  16  crossing solution reduces to  15  and gives a general solution in  4n ‑ 1  crossings, which is shown to be minimal.  If bank to bank crossings are not permitted, then De Fontenay's amended  8n ‑ 6  solution is still optimal.

Marcia Ascher.  A river‑crossing problem in cultural perspective.  MM 63 (1990) 26‑28.  Describes many appearances in folklore of many cultures.  Discusses African variants of the wolf, goat and cabbage problem in which the man can take two of the items in the boat.  This is much easier, requiring only three crossings, but some versions say that the man cannot control the items in the boat, so he cannot have the wolf and goat or the goat and cabbage in the boat with him.  This still only takes three crossings.  Various forms of these problems are mentioned:  fox, fowl and corn;  tiger, sheep and reeds;  jackal, goat and hay;  caged cheetah, fowl and rice;  leopard, goat and leaves -- see below for more details.

                    She also discusses an Ila (Zambia) version with leopard, goat, rat and corn which is unsolvable!

Marcia Ascher.  Ethnomathematics.  Op. cit. in 4.B.10.  1991.  Section 4.8, pp. 109-116  &  Note 8, pp. 119-121.  Good survey of the problem and numerous references to the folklore and ethnographic literature.  Amplifies the above article.  A version like the Wolf, goat and cabbage is found in the Cape Verde Islands, in Cameroon and in Ethiopia.  The African version is found as far apart as Algeria and Zanzibar, but with some variations.  An Algerian version with jackal, goat and hay allows one to carry any two in the boat, but an inefficient solution is presented first.  A Kpelle (Liberia) version with cheetah, fowl and rice adds that the man cannot keep control while rowing so he cannot take the fowl with either the cheetah or the rice in the boat.  A Zanzibar version with leopard, goat and leaves adds instead that no two items can be left on either bank together.  (A similar version occurs among African-Americans on the Sea Islands of South Carolina.)  Ascher notes that Zaslavsky's description is based on an ambiguous report of the Kpelle version and probably should be like the Algerian or Kpelle version just described. 

Liz Allen.  Brain Sharpeners.  New English Library (Hodder & Stoughton), London, 1991.  Crossing the river, pp. 62 & 125.  Three mothers and three sons.  The sons are unwilling to be left with strange mothers, so this is a rephrasing of the jealous husbands.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995.  Chap. 1, probs. 4-6: The knights and the pages; More knights and pages; Yet more knights and pages: no man is an island, pp. 4-5 & 100-102.  Equivalent to the jealous couples.  Prob. 4 is three couples, solved in 11 crossings.  Prob. 5 is four couples -- "There is no solution unless one of the four pages is sacrificed.  (In medieval times, this was not a problem.)"  Prob. 6 is four couples with an island in the river, solved in general by moving all pages to the island, then having the pages go back and accompany his knight to other side, then return to the island.  After the last knight is moved, the pages then move from the island to the other side.  This takes  7n - 6  steps in general.  It satisfies the jealousy conditions used by Pressman & Singmaster, but not those of Loyd & Dudeney.

John P. Ashley.  Arithmetickle.  Arithmetic Curiosities, Challenges, Games and Groaners for all Ages.  Keystone Agencies, Radnor, Ohio, 1997.  P. 16: The missionaries and the pirates.  Politically correct rephrasing of the missionaries and the cannibals version.  All the missionaries, but only one pirate, can row.  Solves in 13 crossings.

Prof. Dr. Robert Weismantel, Otto-von-Guericke-Universität Magdeburg, Fakultät für Mathematik, PSF 3120, D-39016 Magdeburg, Germany; tel: 0391/67-18745; email: weismantel@imo.math.uni-magdeburg.de; has produced a 45 min. film: "Der Wolf, die Ziege und Kohlköpfe  Transportprobleme von Karl dem Grossen bis heute", suitable for the final years of school.

 

          5.B.1. LOWERING FROM TOWER PROBLEM

 

          The problem is for a collection of people (and objects or animals) to lower themselves from a window using a rope over a pulley, with baskets at each end.  The complication is that the baskets cannot contain very different weights, i.e. there is a maximum difference in the weights, otherwise they go too fast.  This is often attributed to Carroll.

 

Carroll-Collingwood.  1899.  P. 318 (Collins: 232-233 (232 is lacking in my copy)).  = Carroll-Wakeling II, prob. 4: The captive queen, pp. 8 & 65-66.  3  people of weights  195, 165, 90  and a weight of  75,  with difference at most  15.  He also gives a more complex form.  No solutions.  Although the text clearly says  165, the prevalence of the exact same problem with  165  replaced by  105  makes me wonder if this was a misprint??  Wakeling says there is no explicit evidence that Carroll invented this, and neither book assigns a date, but Carroll seems a more original source than the following and he was more active before 1890 than after. 

                    An addition is given in both books:  add three animals, weighing  60, 45, 30. 

Lemon.  1890.  The prisoners in the tower, no. 497, pp. 65 & 116.  c= Sphinx, The escape, no. 113, pp. 19 & 100‑101.  Three people of weights  195, 105, 90  with a weight of  75.  The difference in weights cannot be more than  15.

Hoffmann.  1893.  Chap. IV, no. 28: The captives in the tower, pp. 150 & 196 = Hoffmann‑Hordern, p. 123.  Same as Lemon.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 3: The captives in the tower.  Same as Lemon.  Identical to Hoffmann.  With colour picture.  No solution. 

Loyd.  The fire escape puzzle.  Cyclopedia, 1914, pp. 71 & 348.  c= MPSL2, prob. 140, pp. 98‑99 & 165.  = SLAHP: Saving the family, pp. 59 & 108.  Simplified form of Carroll's problem.  Man, wife, baby & dog, weighing a total of  390.

Williams.  Home Entertainments.  1914.  The escaping prisoners, pp. 126-127.  Same as Lemon.

Rudin.  1936.  No. 92, pp. 31-32 & 94.  Same as Lemon.

Haldeman-Julius.  1937.  No. 150: Fairy tale, pp. 17 & 28.  Same as Lemon, except the largest weight is printed as 196, possibly an error.

Kinnaird.  Op. cit. in 1 -- Loyd.  1946.  Pp. 388‑389 & 394.  Same as Lemon.

Simon Dresner.  Science World Book of Brain Teasers.  Scholastic Book Services, NY, 1962.  Prob. 61: Escape from the tower, pp. 29 & 99‑100.  Same as Lemon.

Robert Harbin [pseud. of Ned Williams].  Party Lines.  Oldbourne, London, 1963.  Escape, p. 29.  As in Lemon.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Allen & Unwin, London, 1968.  No. 60: The tower escape, pp. 78 & 118.  Same as Carroll.  Answer in  15  stages.  He cites Carroll, noting that Carroll did not give a solution and he asks if a shorter solution can be found.

F. Geoffrey Hartswick.  In:  H. O. Ripley & F. G. Hartswick; Detectograms and Other Puzzles; Scholastic Book Services, NY, 1969.  No. 15: Stolen treasure puzzle, pp. 54‑55 & 87.  Same as Lemon.

 

          5.B.2. CROSSING A BRIDGE WITH A TORCH

 

          New section.

          Four people have to get across a bridge which is dark and needs to be lit with the torch.  The torch can serve for at most two people and the gap is too wide to throw the torch across, so the torch has to be carried back and forth.  The various people are of different ages and require  5, 10, 20, 25  minutes to cross and when two cross, they have to go at the speed of the slower.  But the torch (= flashlight) battery will only last an hour.  Can it be done?  I heard this about 1997, when it was claimed to be used by Microsoft in interviewing candidates.  I never found any history of it, until I recently found a discussion on Torsten Sillke's site: Crossing the bridge in an hour (www.mathematik.uni‑bielefeld.de/~sillke/PUZZLES/crossing-bridge), starting in Jun 1997 and last updated in Sep 2001.  This cites the 1981 source and the other references below.  Denote the problem with speeds  a, b, c, d  and total time  t  by 

(a, b, c, d; t),  etc.  t  is sometimes given, sometimes not.

 

Saul X. Levmore  &  Elizabeth Early Cook.  Super Strategies for Puzzles and Games.  Doubleday, 1981, p. 3 -- ??NYS.  (5, 10, 20, 25; 60),  as in the introduction to this section..

Heinrich Hemme.  Das Problem des Zwölf-Elfs.  Vandenhoeck & Ruprecht, 1998.  Prob. 81: Die Flucht, pp. 40 & 105-106, citing a web posting by Gunther Lientschnig on 4 Dec 1996.  (2, 4, 8, 10; t).

Dick Hess.  Puzzles from Around the World.  Apr 1997.  Prob. 107: The Bridge. 

          (1, 2, 5, 10; 17).  Poses versions with more people:  (1, 3, 4, 6, 8, 9; 31)  and, with a three-person bridge,  (1, 2, 6, 7, 8, 19, 10; 25).

Quantum (May/Jun 1997) 13.  Brainteaser B 205: Family planning.  Problem (1, 3, 8, 10; 20).

Karen Lingel.  Email of 17 Sep 1997 to  rec.puzzles.  Careful analysis, showing that the 'trick' solution is better than the 'direct' solution if and only if  a + c  >  2b.  [Indeed,  a + c - 2b  is the time saved by the 'trick' solution.]  She cites  (2, 3, 5, 8; 19)  and  (2, 2, 3, 3; 11)  to Sillke and  (1, 3, 6, 8, 12; 30),  from an undated website.  Expressing the solution for more people seems to remain an open question.

 

          5.C.   FALSE COINS WITH A BALANCE

 

          See 5.D.3 for use of a weighing scale.

          There are several related forms of this problem.  Almost all of the items below deal with 12 coins with one false, either heavy or light, and its generalizations, but some other forms occur, including the following.

           8 coins, £1 light:  Schell, Dresner

          26 coins, £1 light:  Schell

           8 coins,  1 light:  Bath (1959)

           9 coins,  1 light:  Karapetoff, Meyer (1946), Meyer (1948), M. Adams, Rice

 

          I have been sent an article by Jack Sieburg; Problem Solving by Computer Logic; Data Processing Magazine, but the date is cut off -- ??

 

E. D. Schell, proposer;  M. Dernham, solver.  Problem E651 -- Weighed and found wanting.  AMM 52:1 (Jan 1945) 42  &  7 (Aug/Sep 1945) 397.  8 coins, at most one light -- determine the light one in two weighings.

Benjamin L. Schwartz.  Letter:  Truth about false coins.  MM 51 (1978) 254.  States that Schell told Michael Goldberg in 1945 that he had originated the problem.

Emil D. Schell.  Letter of 17 Jul 1978 to Paul J. Campbell.  Says he did NOT originate the problem, nor did he submit the version published.  He first heard of it from Walter W. Jacobs about Thanksgiving 1944 in the form of finding at most one light coin among 26 good coins in three weighings.  He submitted this to the AMM, with a note disclaiming originality.  The AMM problem editor published the simpler version described above, under Schell's name.  Schell says he has heard Eilenberg describe the puzzle as being earlier than Sep 1939.  Campbell wrote Eilenberg, but had no response.

                    Schell's letter is making it appear that the problem derives from the use of  1, 3, 9, ...  as weights.  This usage leads one to discover that a light coin can be found in  3n  coins using  n  weighings.  This is the problem mentioned by Karapetoff.  If there is at most one light coin, then  n  weighings will determine it among  3n ‑ 1  coins, which is the form described by Schell.  The problem seems to have been almost immediately converted into the case with one false coin, either heavy or light.

Walter W. Jacobs.  Letter of 15 Aug 1978 to Paul J. Campbell.  Says he heard of the problem in 1943 (not 1944) and will try to contact the two people who might have told it to him.  However, Campbell has had no further word.

 

V. Karapetoff.  The nine coin problem and the mathematics of sorting.  SM 11 (1945) 186‑187.  Discusses 9 coins, one light, and asks for a mathematical approach to the general problem.  (?? -- Cites AMM 52, p. 314, but I cannot find anything relevant in the whole volume, except the Schell problem.  Try again??)

Dwight A. Stewart, proposer;  D. B. Parkinson & Lester H. Green, solvers.  The counterfeit coin.  In:  L. A. Graham, ed.; Ingenious Mathematical Problems and Methods; Dover, 1959; pp. 37‑38 & 196‑198.  12 coins.  First appeared in Oct 1945.  Original only asks for the counterfeit, but second solver shows how to tell if it is heavy or light.

R. L. Goodstein.  Note 1845:  Find the penny.  MG 29 (No. 287) (Dec 1945) 227‑229.  Non‑optimal solution of general problem.

Editorial Note.  Note 1930:  Addenda to Note 1845.  Ibid. 30 (No. 291) (Oct 1946) 231. Comments on how to extend to optimal solution.

Howard D. Grossman.  The twelve‑coin problem.  SM 11:3/4 (Sep/Dec 1945) 360‑361.  Finds counterfeit and extends to 36 coins.

Lothrop Withington, Jr.  Another solution of the 12‑coin problem.  Ibid., 361‑362.  Finds also whether heavy or light.

Donald Eves, proposer;  E. D. Schell & Joseph Rosenbaum, solvers.  Problem E712 -- The extended coin problem.  AMM 53:3 (Mar 1946) 156  &  54:1 (Jan 1947) 46‑48.  12 coins.

Jerome S. Meyer.  Puzzle Paradise.  Crown, NY, 1946.  Prob. 132: The nine pearls, pp. 94 & 132.  Nine pearls, one light, in two weighings.

N. J. Fine, proposer & solver.  Problem 4203 -- The generalized coin problem.  AMM 53:5 (May 1946) 278  &  54:8 (Oct 1947) 489‑491.  General problem.

H. D. Grossman.  Generalization of the twelve‑coin problem.  SM 12 (1946) 291‑292.  Discusses Goodstein's results.

F. J. Dyson.  Note 1931:  The Problem of the Pennies.  MG 30 (No. 291) (Oct 1946) 231‑234.  General solution.

C. A. B. Smith.  The Counterfeit Coin Problem.  MG 31 (No. 293) (Feb 1947) 31‑39.

C. W. Raine.  Another approach to the twelve‑coin problem.  SM 14 (1948) 66‑67.  12 coins only.

K. Itkin.  A generalization of the twelve‑coin problem.  SM 14 (1948) 67‑68.  General solution.

Howard D. Grossman.  Ternary epitaph on coin problems.  SM 14 (1948) 69‑71.  Ternary solution of Dyson & Smith.

Jerome S. Meyer.  Fun-to-do.  A Book of Home Entertainment.  Dutton, NY, 1948.  Prob. 40: Nine pearls, pp. 41 & 188.  Nine pearls, one light, in two weighings.

Blanche Descartes [pseud. of Cedric A. B. Smith].  The twelve coin problem.  Eureka 13 (Oct 1950) 7 & 20.  Proposal and solution in verse.

J. S. Robertson.  Those twelve coins again.  SM 16 (1950) 111‑115.  Article indicates there will be a continuation, but Schaaf I 32 doesn't cite it and I haven't found it yet.

E. V. Newberry.  Note 2342:  The penny problem.  MG 37 (No. 320) (May 1953) 130.  Says he has made a rug showing the 120 coins problems and makes comments similar to Littlewood's, below.

J. E. Littlewood.  A Mathematician's Miscellany.  Methuen, London, 1953;  reprinted with minor corrections, 1957 (& 1960).  [All the material cited is also in the later version:  Littlewood's Miscellany, ed. by B. Bollobás, CUP, 1986, but on different pages.  Since the 1953 ed. is scarce, I will also cite the 1986 pages in (  ).]  Pp. 9 & 135 (31 & 114).  "It was said that the 'weighing‑pennies' problem wasted 10,000 scientist‑hours of war‑work, and that there was a proposal to drop it over Germany."

John Paul Adams.  We Dare You to Solve This!  Berkley Publishing, NY, nd [1957?].  [This is apparently a collection of problems used in newspapers.  The copyright is given as 1955, 1956, 1957.]  Prob. 18: Weighty problem, pp. 13 & 46.  9 equal diamonds but one is light, to be found in 2 weighings.

Hubert Phillips.  Something to Think About.  Revised ed., Max Parrish, London, 1958.  Foreword, p. 6  &  prob. 115: Twelve coins, pp. 81  &  127‑128.  Foreword says prob. 115 has been added to this edition and "was in oral circulation during the war.  So far as I know, it has only appeared in print in the Law Journal, where I published both the problem and its solution."  This may be an early appearance, so I should try and track this down.  ??NYS

Dan Pedoe.  The Gentle Art of Mathematics.  (English Universities Press, 1958);  Pelican (Penguin), 1963.  P. 30:  "We now come to a problem which is said to have been planted over here during the war by enemy agents, since Operational Research spent so many man‑hours on its solution."

Philip E. Bath.  Fun with Figures.  The Epworth Press, London, 1959.  No. 7: No weights -- no guessing, pp. 8 & 40.  8  balls, including one light, to be determined in two weighings.  Method actually works for  £ 1  light.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 9.  Five boxes of sugar, but some has been taken from one box and put in another.  Determine which in least number of weighings.  Does by weighing each division of  A, B, C, D  into two pairs.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  The "False Coin" problem, pp. 178-182.  Sketches history and solution.

Simon Dresner.  Science World Book of Brain Teasers.  1962.  Op. cit. in 5.B.1.  Prob. 46: Dud reckoning, pp. 21 & 94.  Find one light among eight in two weighings.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 55, pp. 57 & 98.  Six identical appearing coins, three of which are identically heavy.  In two weighings, identify two of the heavy coins.

Charlie Rice.  Challenge!  Hallmark Editions, Kansas City, Missouri, 1968.  Prob. 7, pp. 22 & 54-55.  9 pearls, one light.

Jonathan Always.  Puzzling You Again.  Tandem, London, 1969.  Prob. 86: Light‑weight contest, pp. 51‑52 & 106‑107.  27 weights of sizes 1, 2, ..., 27, except one is light.  Find it in 3 weighings.  He divides into 9 sets of three having equal weights.  Using two weighings, one locates the light weight in a set of three and then weighing two of these with good weights reveals the light one.  [3 weights 1, 2, 3 cannot be done in one weighing, but 9 weights 1, 2, ..., 9 can be done in two weighings.]

Robert H. Thouless.  The 12‑balls problem as an illustration of the application of information theory.  MG 54 (No. 389) (Oct 1970) 246‑249.  Uses information theory to show that the solution process is essentially determined.

Ron Denyer.  Letter.  G&P, No. 37 (Jun 1975) 23.  Asks for a mnemonic for the 12 coins puzzles.  He notes that one can use three predetermined weighings and find the coin from the three answers.

Basil Mager & E. Asher.  Letters:  Coining a mnemonic.  G&P, No. 40 (Sep 1975) 26.  One mnemonic for a variable method, another for a predetermined method.

N. J. Maclean.  Letter:  The twelve coins.  G&P, No. 45 (Feb 1976) 28-29.  Exposits a ternary method for predetermined weighings for  (3n-3)/2  in  n  weighings.  Each weighing determines one ternary digit and the resulting ternary number gives both the coin and whether it is heavy or light.

Tim Sole.  The Ticket to Heaven and Other Superior Puzzles.  Penguin, 1988.  Weighty problems -- (iii), pp. 124 & 147.  Nine equal pies, except someone has removed some filling from one and inserted it in a pie, possibly the same one.  Determine which, if any, are the heavy and light ones in 4 balancings.

Calvin T. Long.  Magic in base 3.  MG 76 (No. 477) (Nov 1992) 371-376.  Good exposition of the base 3 method for 12 coins.

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Problems for an equal-arm balance, pp. 137-141. 

                    1.  Six balls, two of each of three colours.  One of each colour is lighter than normal and all light weights are equal.  Determine the light balls in three weighings.

                    2.  Five balls, three normal, one heavy, one light, with the differences being equal, i.e. the heavy and the light weigh as much as two normals.  Determine the heavy and light in three weighings.

                    3.  Same problem with nine balls and seven normals, done in four weighings.

5.C.1RANKING COINS WITH A BALANCE

 

          If one weighs only one coin against another, this is the problem of sorting except that we don't actually put the objects in order.  If one weighs pairs, etc., this is a more complex  problem.

 

J. Schreier.  Mathesis Polska 7 (1932) 154‑160.  ??NYS -- cited by Steinhaus.

Hugo Steinhaus.  Mathematical Snapshots.  Not in Stechert, NY, 1938, ed.  OUP, NY:  1950: pp. 36‑40 & 258;  1960: pp. 51‑55 & 322;  1969 (1983): pp. 53‑56 & 300.  Shows  n  objects can be ranked in  M(n) = 1 + kn ‑ 2k  steps where  k = 1 + [log2 n].  Gets  M(5) = 8.

Lester R. Ford Jr. & Selmer M. Johnson.  A tournament problem.  AMM 66:5 (May 1959) 387‑389.  Note that  élog2 n!ù = L(n)  is a lower bound from information theory.  Obtain a better upper bound than Steinhaus, denoted  U(n),  which is too complex to state here.  For convenience, I give the table of these values here.

 

                      n                1   2   3   4   5     6     7     8     9   10   11   12   13

                    M(n)             0   1   3   5   8   11   14   17   21   25   29   33   37

                    U(n)             0   1   3   5   7   10   13   16   19   22   26   30   34

                    L(n)              0   1   3   5   7   10   13   16   19   22   26   29   33

 

          U(n) = L(n)  also holds at  n = 20 and 21.

 

Roland Sprague.  Unterhaltsame Mathematik.  Op. cit. in 4.A.1.  1961.  Prob. 22: Ein noch ungelöstes Problem, pp. 16 & 42‑43.  (= A still unsolved problem, pp. 17 & 48‑49.)  Sketches Steinhaus's method, then does 5 objects in 7 steps.  Gives the lower bound  L(n)  and says the case  n = 12  is still unsolved.

Kobon Fujimura, proposer;  editorial comment.  Another balance scale problem.  RMM 10 (Aug 1962) 34  &  11 (Oct 1962) 42.  Eight coins of different weights and a balance.  How many weighings are needed to rank the coins?  In No. 11, it says the solution will appear in No. 13, but it doesn't appear there or in the last issue, No. 14.  It also doesn't appear in the proposer's Tokyo Puzzles.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.  No. 6: In the balance, pp. 18 & 85-86.  Rank five balls in order in seven weighings.

John Cameron.  Establishing a pecking order.  MG 55 (No. 394) (Dec 1971) 391‑395.  Reduces Steinhaus's  M(n)  by  1  for  n ³ 5,  but this is not as good as Ford & Johnson.

W. Antony Broomhead.  Letter:  Progress in congress?  MG 56 (No. 398) (Dec 1972) 331.  Comments on Cameron's article and says Cameron can be improved.  States the values  U(9)  and  U(10),  but says he doesn't know how to do  9  in  19  steps.  Cites Sprague for numerical values, but these don't appear in Sprague -- so Broomhead presumably computed  L(9)  and  L(10).  He gets  10  in  23  steps, which is better than Cameron.

Stanley Collings.  Letter:  More progress in congress.  MG 57 (No. 401) (Oct 1973) 212‑213.  Notes the ambiguity in Broomhead's reference to Sprague.  Improves Cameron by  1  (or more??) for  n ³ 10,  but still not as good as Ford & Johnson.

L. J. Upton, proposer;  Leroy J. Myers, solver.  Problem 1138.  CM 12 (1986) 79  &  13 (1987) 230‑231.  Rank coins weighing  1, 2, 3, 4  with a balance in four weighings.

 

5.D. MEASURING PROBLEMS

 

          5.D.1.          JUGS & BOTTLES

 

          See MUS I 105-124, Tropfke 659.

          NOTATION:  I-(a, b, c)  means we have three jugs of sizes  a, b, c  with  a  full and we want to divide  a  in half using  b  and  c.  We normally assume  a ³ b ³ c  and  GCD(a, b, c) = 1.  Halving  a  is clearly impossible if  GCD(b, c)  does not divide  a/2  or if  b+c < a/2,  unless one has a further jug or one can drink some.  If  a ³ b+c ³ a/2  and  GCD(b, c)  divides  a/2,  then the problem is solvable.

          More generally, the question is to determine what amounts can be produced, i.e. given  a, b, c  as above, can one measure out an amount  d?  We denote this by  II-(a, b, c; d).  Since this also produces  a-d,  we can assume that  d £ a/2.  Then we must have  d £ b+c  for a solution.  When  ³ b+c ³ d,  the condition  GCD(b, c) ½ d  guarantees that  d  can be produced.  This also holds for  a = b+c‑1  and  a = b+c‑2.  The simplest impossible cases are  I‑(4, 4, 3) = II-(4, 4, 3; 2)  and  II‑(5, 5, 3; 1).  Case I‑(a, b, c)  is the same as  II-(a, b, c; a/2).

          If  a  is a large source, e.g. a stream or a big barrel, we have the problem of measuring  d  using  b  and  c  without any constraint on  a  and we denote this  II-(¥, b, c; d).  However, the solution may not use the infiniteness of the source and such a problem may be the same as  II‑(b+c, b, c; d).

          The general situation when  a < b+c  is more complex and really requires us to consider the most general three jug problem:  III‑(A; a, b, c; d)  means we have three jugs of sizes  a, b, c,  containing a total amount of liquid  A  (in some initial configuration) and we wish to measure out  d.  In our previous problems, we had  A = a.  Clearly we must have  a+b+c ³ A.  Again, producing  d  also produces  A-d,  so we can assume  d £ A/2.  By considering the amounts of empty space in the containers, the problem  III-(A; a, b, c; d)  is isomorphic to III‑(a+b+c‑A; a, b, c; d')  for several possible d'.

         

          NOTES.  I have been re-examining this problem and I am not sure if I have reached a final interpretation and formulation.  Also, I have recently changed to the above notation and I may have made some errors in so doing.  I have long had the problem in my list of projects for students, but no one looked at it until 1995-1996 when Nahid Erfani chose it.  She has examined many cases and we have have discovered a number of properties which I do not recall seeing.  E.g. in case  I-(a,b,c)  with  a ³ b ³ c  and  GCD(b,c) = 1,  there are two ways to obtain  a/2.  If we start by pouring into  b,  it takes  b + c - 1  pourings; if we start by pouring into  c,  it takes  b + c  pourings; so it is always best to start pouring into the larger jug.  A number of situations  II-(a,b,c;d)  are solvable for all values of  d,  except  a/2.  E.g.  II‑(a,b,c;a/2)  with  b+c > a  and  c > a/2  is unsolvable.

 

          From about the mid 19C, I have not recorded simple problems.

 

I-(  8,  5,  3):                         almost all the entries below

I-(10,  6,  4):                         Pacioli,  Court

I-(10,  7,  3):                         Yoshida

I-(12,  7,  5):                         Pacioli,  van Etten/Henrion,  Ozanam,  Bestelmeier,  Jackson,  Manuel des Sorciers,  Boy's Own Conjuring Book

I-(12,  8,  4):                         Pacioli

I-(12,  8,  5):                         Bachet,  Arago

I-(16,  9,  7):                         Bachet-Labosne

I-(16,11,  6):                         Bachet-Labosne

I-(16,12,  7):                         Bachet-Labosne

I-(20,13,  9):                         Bachet-Labosne

I-(42,27,12):                          Bachet-Labosne

 

II-(10,3,2;6)                          Leacock

   = II(10,3,2;4)

II-(11,4,3;9):                         McKay

   = II(11,4,3;2)

II-( ¥,5,3;1):                         Wood,  Serebriakoff,  Diagram Group

II-( ¥,5,3;4):                         Chuquet,  Wood,  Fireside Amusements, 

II-( ¥,7,4;5):                         Meyer,  Stein,  Brandes

II-( ¥,8,5;11):                       Young World, 

III-(20;19,13,7;10):                Devi

 

General problem, usually form I, sometimes form II:  Bachet‑Labosne,  Schubert,  Ahrens,  Cowley,  Tweedie,  Grossman,  Buker,  Goodstein,  Browne,  Scott,  Currie,  Sawyer,  Court,  O'Beirne,  Lawrence,  McDiarmid & Alfonsin.

Versions with 4 or more jugs:  Tartaglia,  Anon: Problems drive (1958),  Anon (1961),  O'Beirne.

Impossible versions:  Pacioli,  Bachet,  Anon: Problems drive (1958).

 

Abbot Albert.  c1240.  Prob. 4, p. 333.  I-(8,5,3)  -- one solution.

Columbia Algorism.  c1350.  Chap. 123:  I-(8,5,3).  Cowley 402‑403 & plate opposite 403.  The plate shows the text and three jars.  I have a colour slide of the three jars from the MS.

Munich 14684.  14C.  Prob. XVIII  &  XXIX, pp. 80  &  83.  I-(8,5,3).

Folkerts.  Aufgabensammlungen.  13-15C.  16 sources with  I-(8,5,3).

Pseudo-dell'Abbaco.  c1440.  Prob. 66, p.62.  I-(8,5,3)  -- one solution.  "This problem is of little utility ...."  I have a colour slide of this.

Chuquet.  1484.  Prob. 165.  Measure 4 from a cask using 5 and 3.  You can pour back into the cask, i.e. this is  II-(¥,5,3;4).  FHM 233 calls this the tavern-keeper's problem.

HB.XI.22.  1488.  P. 55 (= Rath 248).  Same as Abbot Albert.

Pacioli.  De Viribus.  c1500. 

Ff. 97r - 97v.  LIII. C(apitolo). apartire una botte de vino fra doi (To divide a bottle of wine between two).  = Peirani 137-138.  I-(8,5,3).  One solution.

Ff. 97v - 98v.  LIIII. C(apitolo). a partire unaltra botte fra doi (to divide another bottle between two).  = Peirani 138-139.  I-(12,7,5).  Dario Uri points out that the solution is confused and he repeats himself so it takes him  18  pourings instead of the usual  11.  He then says one can divide  18  among three brothers who have containers of sizes  5, 6, 7,  which he does by filling the  6  and then the problem is reduced to the previous problem.  [He could do it rather more easily by pouring the  6  into the  7  and then refilling the  6!]

Ff. 98v - 99r.  LV. (Capitolo) de doi altri sotili divisioni. de botti co'me se dira (Of two other subtle divisions of bottles as described).  = Peirani 139-140.  I‑(10,6,4)  and  I-(12,8,4).  Pacioli suggests giving these to idiots.

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 20, ff. 64v-65r.  I‑(8,5,3).  One solution. 

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 33, f. DD.iiii.v (p. 145).  I-(8,5,3).  Gives one solution and says one can go the other way.

H&S 51 says  I-(8,5,3)  case is also in Trenchant (1566).  ??NYS

Tartaglia.  General Trattato, 1556, art. 132 & 133, p. 255v‑256r.

Art. 132:  I-(8,5,3).

Art. 133: divide  24  in thirds, using  5, 11, 13.

Buteo.  Logistica.  1559.  Prob. 73, pp. 282-283.  I-(8,5,3).

Gori.  Libro di arimetricha.  1571.  Ff. 71r‑71v (p. 76).  I-(8,5,3).

Bachet.  Problemes.  1612.  Addl. prob. III: Deux bons compagnons ont 8 pintes de vin à partager entre eux également, ...,  1612: 134-139;  1624: 206-211;  1884: 138‑147.  I‑(8,5,3)  -- both solutions;  I-(12,8,5)  (omitted by Labosne).  Labosne adds  I‑(16,9,7);  I‑(16,11,6);  I‑(42,27,12);  I-(20,13,9);  I-(16,12,7)  (an impossible case!)  and discusses general case.  (This seems to be the first discussion of the general case.)

van Etten.  1624.  Prob. 9 (9), pp. 11  &  fig. opp. p. 1 (pp. 22‑23).  I‑(8,5,3)  -- one solution.  Henrion's Nottes, 1630, pp. 11‑13, gives the second solution and poses and solves  I‑(12,7,5).

Hunt.  1631 (1651).  P. 270 (262).  I-(8,5,3).  One solution.

Yoshida (Shichibei) Kōyū (= Mitsuyoshi Yoshida) (1598-1672).  Jinkō‑ki.  2nd ed., 1634 or 1641??.  ??NYS  The recreational problems are discussed in Kazuo Shimodaira; Recreative Problems on "Jingōki", a 15 pp booklet sent by Shigeo Takagi.  [This has no details, but Takagi says it is a paper that Shimodaira read at the 15th International Conference for the History of Science, Edinburgh, Aug 1977 and that it appeared in Japanese Studies in the History of Science 16 (1977) 95-103.  I suspect this is a copy of a preprint.]  This gives both Jingōki and Jinkōki as English versions of the title and says the recreational problems did not appear in the first edition, 4 vols., 1627, but did appear in the second edition of 5 vols. (which may be the first use of coloured wood cuts in Japan), with the recreational problems occurring in vol. 5.  He doesn't give a date, but Mikami, p. 179, indicates that it is 1634, with further editions in 1641, 1675, though an earlier work by Mikami (1910) says 2nd ed. is 1641.  Yoshida (or Suminokura) is the family name.  Shimodaira refers to the current year as the 350th anniversary of the edition and says copies of it were published then.  I have a recent transcription of some of Yoshida into modern Japanese and a more recent translation into English, ??NYR, but I don't know if it is the work mentioned by Shimodaira.

                    Shimodaira discusses a jug problem on p. 14:  I-(10,7,3)  -- solution in 10 moves.  Shimodaira thinks Yoshida heard about such puzzles from European contacts, but without numerical values, then made up the numbers.  I certainly can see no other example of these numbers.  The recent transcription includes this material as prob. 7 on pp. 69-70.

Wingate/Kersey.  1678?.  Prob. 7, pp. 543-544.  I-(8,5,3).  Says there is a second way to do it.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 38, p. 308.  I-(8,5,3).

Ozanam.  1694. 

Prob. 36, 1696: 91-92;  1708: 82‑83.  Prob. 42, 1725: 238‑240.  Prob. 21, 1778: 175‑177;  1803: 174-176;  1814: 153-154.  Prob. 20, 1840: 79.  I-(8,5,3)  -- both solutions.

Prob. 43, 1725: 240‑241.  Prob. 22, 1778: 177-178;  1803: 176-177;  1814: 154-155.  Prob. 21, 1840: 79‑80.  I-(12,7,5)  -- one solution.

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.  Problem 8.  I-(8,5,3).  (Dilworth cites Wingate for this -- cf in 5.B.)  = D. Adams; Scholar's Arithmetic; 1801, p. 200, no. 10.

Les Amusemens.  1749.  Prob. 17, p. 139: Partages égaux avec des Vases inégaux.  I-(8,5,3)  -- both solutions.

Bestelmeier.  1801.  Item 416: Die 3 Maas‑Gefäss.  I-(12,7,5).

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 48-49, no. 75: How to part an eight gallon bottle of wine, equally between two persons, using only two other bottles, one of five gallons, and the other of three.  Gives both solutions.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 14, pp.  4 & 54.  I-( 8,5,3).  One solution.

No. 52, pp. 12 & 67.  I-(12,7,5).  One solution.

Rational Recreations.  1824.  Exer. 10, p. 55.  I-(8,5,3)  one way.

Manuel des Sorciers.  1825.  ??NX 

Pp. 55-56, art. 27-28.  I-(8,5,3)  two ways.

P. 56, art. 29.  I-(12,7,5).

Endless Amusement II.  1826?  Prob. 7, pp. 193-194.  I-(8,5,3).  One solution.  = New Sphinx, c1840, p. 133.

Nuts to Crack III (1834), no. 212.  I-(8,5,3).  8 gallons of spirits.

Young Man's Book.  1839.  Pp. 43-44.  I-(8,5,3).  Identical to Wingate/Kersey. 

The New Sphinx.  c1840.  P. 133.  I-(8,5,3).  One solution. 

Boy's Own Book.  1843 (Paris): 436 & 441, no. 7.   The can of ale:  1855: 395;  1868: 432.  I‑(8,5,3).  One solution.  The 1843 (Paris) reads as though the owners of the 3 and 5 kegs both want to get 4, which would be a problem for the owner of the 3.  = Boy's Treasury, 1844, pp. 425 & 429.

Fireside Amusements.  1850.  Prob. 9, pp. 132 & 184.  II-(¥,5,3;4).  One solution. 

Arago.  [Biographie de] Poisson (16 Dec 1850).  Oeuvres, Gide & Baudry, Paris, vol. 2, 1854, pp. 593‑???  P. 596 gives the story of Poisson's being fascinated by the problem  I‑(12,8,5).  "Poisson résolut à l'instant cette question et d'autres dont on lui donna l'énoncé.  Il venait de trouver sa véritable vocation."  No solution given by Arago.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 8, pp. 174-175 (1868: 185-186).  I-(8,5,3).  Milkmaid with eight quarts of milk.

Magician's Own Book.  1857.

P. 223-224: Dividing the beer:  I-(8,5,3).

P. 224: The difficult case of wine:  I-(12,7,5).

Pp. 235-236: The two travellers:  I-(8,5,3)  posed in verse.

Each problem gives just one solution.

Boy's Own Conjuring Book.  1860. 

P. 193: Dividing the beer:  I-(8,5,3).

P. 194: The difficult case of wine:  I-(12,7,5).

Pp. 202‑203: The two travellers:  I-(8,5,3)  posed in verse.

Each problem gives just one solution.

Illustrated Boy's Own Treasury.  1860.  Prob. 21, pp. 428-429 & 433.  I‑(8,5,3).  "A man coming from the Lochrin distillery with an 8-pint jar full of spirits, ...."

Vinot.  1860.  Art. XXXVIII: Les cadeaux difficiles, pp. 57-58.  I-(8,5,3).  Two solutions.

The Secret Out (UK).  c1860.  To divide equally eight pints of wine ..., pp. 12-13.

Bachet-Labosne.  1874.  For details, see Bachet, 1612.  Labosne adds a consideration of the general case which seems to be the first such.

Kamp.  Op. cit. in 5.B.  1877.  No. 17, p. 326:  I-(8,5,3).

Mittenzwey.  1880.  Prob. 106, pp. 22 & 73-74;  1895?: 123, pp. 26 & 75-76;  1917: 123, pp. 24 & 73-74.  I-(8,5,3).  One solution.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 135, no. 1.  I-(8,5,3).  No solution.

Loyd.  Problem 11: "Two thieves of Damascus".  Tit‑Bits 31 (19 Dec 1896  &  16 Jan 1897) 211  &  287.  Thieves found with  2 & 2  quarts in pails of size  3 & 5.  They claim the merchant measured the amounts out from a fresh hogshead.  Solution is that this could only be done if the merchant drained the hogshead, which is unreasonable!

Loyd.  Problem 13: The Oriental problem.  Tit‑Bits 31 (19 Jan,  30 Jan  &  6 Feb 1897) 269,  325  &  343.  = Cyclopedia, 1914, pp. 188 & 364: The merchant of Bagdad.  Complex problem with hogshead of water, barrel of honey, three 10 gallon jugs to be filled with 3 gallons of water, of honey and of half and half honey & water.  There are a 2 and a 4 gallon measure and also 13 camels to receive 3 gallons of water each.  Solution takes  521  steps.  6 Feb reports solutions in  516  and  513  steps.  Cyclopedia gives solution in  506  steps.

Dudeney.  The host's puzzle.  London Magazine 8 (No. 46) (May 1902) 370  &  8 (No. 47) (Jun 1902) 481‑482 (= CP, prob. 6, pp. 28‑29 & 166‑167).  Use  5 and 3  to obtain  1 and 1  from a cask.  One must drink some!

H. Schubert.  Mathematische Mussestunden, 3rd ed., Göschen, Leipzig, 1907.  Vol. 1, chap. 6, Umfüllungs‑Aufgaben, pp. 48‑56.  Studies general case and obtains some results.  (The material appeared earlier in Zwölf Geduldspiele, 1895, op. cit. in 5.A, Chap. IX, pp. 110-119.  The 13th ed. (De Gruyter, Berlin, 1967), Chap. 9, pp. 62‑70, seems to be a bit more general (??re-read).)

Ahrens.  MUS I, 1910, chap. 4, Umfüllungsaufgaben, pp. 105‑124.  Pp. 106‑107 is Arago's story of Poisson and this problem.  He also extends and corrects Schubert's work.

Dudeney.  Perplexities: No. 141: New measuring puzzle.  Strand Magazine 45 (Jun 1913) 710  &  46 (Jul 1913) 110.  (= AM, prob. 365, pp. 110 & 235.)  Two 10 quart vessels of wine with 5 and 4 quart measures.  He wants 3 quarts in each measure.  (Dudeney gives numerous other versions in AM.)

Loyd.  Cyclopedia.  1914.  Milkman's puzzle, pp. 52 & 345.  (= MPSL2, prob. 23, pp. 17 & 127‑128  = SLAHP: Honest John, the milkman, pp. 21 & 90.)  Milkman has two full 40 quart containers and two customers with 5 and 4 quart pails, but both want 2 quarts.  (Loyd Jr. says  "I first published [this] in 1900...")

Williams.  Home Entertainments.  1914.  The measures puzzle, p. 125.  I-(8,5,3).

Hummerston.  Fun, Mirth & Mystery.  1924.  A shortage of milk, Puzzle no. 75, pp. 164 & 183.  I-(8,5,3),  one solution.

Elizabeth B. Cowley.  Note on a linear diophantine equation.  AMM 33 (1926) 379‑381.  Presents a technique for resolving  I-(a,b,c),  which gives the result when  a = b+c.  If  a < b+c,  she only seems to determine whether the method gets to a point with  A  empty and neither  B  nor  C  full and it is not clear to me that this implies impossibility.  She mentions a graphical method of Laisant (Assoc. Franç. Avance. Sci, 1887, pp. 218-235)  ??NYS.

Wood.  Oddities.  1927. 

Prob. 15: A problem in pints, pp. 16-17.  Small cask and measures of size 5 and 3, measure out 1  in each measure.  Starts by filling the 5 and the 3 and then emptying the cask, so this becomes a variant of  II-(¥,5,3;1).

Prob. 26: The water-boy's problem, pp. 28-29.  II-(¥;,5,3;4). 

Ernest K. Chapin.  Scientific Problems and Puzzles.  In:  S. Loyd Jr.;  Tricks and Puzzles, Vol. 1 (only volume to appear);  Experimenter Publishing Co., NY, nd [1927]  and  Answers to Sam Loyd's Tricks and Puzzles, nd [1927].  [This book is a selection of pages from the Cyclopedia, supplemented with about 20 pages by Chapin and some other material.]  P. 89 & Answers p. 8.  You have a tablet that has to be dissolved in    quarts of water, though you only need 5 quarts of the resulting mixture.  You have 3 and 5 quart measures and a tap.

Stephen Leacock.  Model Memoirs and Other Sketches from Simple to Serious.  John Lane, The Bodley Head, 1939, p. 298.  "He's trying to think how a farmer with a ten-gallon can and a three-gallon can and a two-gallon can, manages to measure out six gallons of milk."  II-(10,3,2;6) = II-(10,3,2;4).

M. C. K. Tweedie.  A graphical method of solving Tartaglian measuring puzzles.  MG 23 (1939) 278‑282.  The elegant solution method using triangular coordinates.

H. D. Grossman.  A generalization of the water‑fetching puzzle.  AMM 47 (1940) 374‑375.  Shows  II-(¥,b,c;d)  with  GCD(b,c) = 1  is solvable.

McKay.  Party Night.  1940. 

No. 18, p. 179.  II-(11,4,3;9).

No. 19, pp. 179-180.  I-(8,5,3).

Meyer.  Big Fun Book.  1940.  No. 10, pp. 165 & 753.  II-(¥,7,4,5).

W. E. Buker, proposer.  Problem E451.  AMM 48 (1941) 65.  ??NX.  General problem of what amounts are obtainable using three jugs, one full to start with, i.e.  I-(a,b,c).  See Browne, Scott, Currie below.

Eric Goodstein.  Note 153:  The measuring problem.  MG 25 (No. 263) (Feb 1941) 49‑51.  Shows  II-(¥,b,c;d)  with  GCD(b,c) = 1  is solvable.

D. H. Browne & Editors.  Partial solution of Problem E451.  AMM 49 (1942) 125‑127.

W. Scott.  Partial solution of E451 -- The generalized water‑fetching puzzle.  AMM 51 (1944) 592.  Counterexample to conjecture in previous entry.

J. C. Currie.  Partial solution of Problem E451.  AMM 53 (1946) 36‑40.  Technical and not complete.

W. W. Sawyer.  On a well known puzzle.  SM 16 (1950) 107‑110.  Shows that  I-(b+c,b,c)  is solvable if  b & c  are relatively prime.

David Stein.  Party and Indoor Games.  Op. cit. in 5.B.  c1950.  Prob. 13, pp. 79‑80.  Obtain 5 from a spring using measures 7 and 4, i.e.  II-(¥,7,4,5).

Anonymous.  Problems drive, 1958.  Eureka 21 (Oct 1958) 14-16 & 30.  No. 8.  Given an infinite source, use:   6, 10, 15  to obtain  1, 6, 7  simultaneously;  4, 6, 9, 12  to obtain  1, 2, 3, 4  simultaneously;  6, 9, 12, 15, 21  to obtain  1, 3, 6, 8, 9  simultaneously.  Answer simply says the first two are possible (the second being easy) and the third is impossible.

Young World.  c1960.  P. 58: The 11 pint problem.  II-(¥,8,5;11).  This is the same as  II‑(13,8,5;11)  or  II-(13,8,5,2).

Anonymous.  Moonshine sharing.  RMM 2 (Apr 1961) 31  &  3 (Jun 1961) 46.  Divide  24  in thirds using cylindrical containers holding  10, 11, 13.  Solution in No. 3 uses the cylindricity of a container to get it half full.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  "Pouring" problems -- The "robot" method.  General description of the problem.  Attributes Tweedie's triangular 'bouncing ball' method to Perelman, with no reference.  Does  I‑(8,5,3)  two ways, also  I-(12,7,5)  and  I-(16,9,7),  then considers type II questions.  Considers the problem with  II-(10,6,4;d)  and extends to  II-(a,6,4;d)  for  a > 10,  leaving it to the reader to "try to formulate some rule about the results."  He then considers  II‑(7,6,4;d),  noting that the parallelogram has a corner trimmed off.  Then considers  II-(12,9,7;d)  and  II-(9,6,3;d).

Lloyd Jim Steiger.  Letter.  RMM 4 (Aug 1961) 62.  Solves the RMM 2 problem by putting the 10 inside the 13 to measure 3.

Irving & Peggy Adler.  The Adler Book of Puzzles and Riddles.  Or Sam Loyd Up-To-Date.  John Day, NY, 1962.  Pp. 32 & 46.  Farmer has two full 10-gallon cans.  Girls come with 5-quart and 4-quart cans and each wants 2 quarts.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 80, pp. 81 & 109.  Tavern has a barrel with 15 pints of beer.  Two customers, with 3 pint and 5 pint jugs appear and ask for 1 pint in each jug.  Bartender finds it necessary to drink the other 13 pints!

T. H. O'Beirne.  Puzzles and Paradoxes.  OUP, 1965.  Chap. 4: Jug and bottle department, pp. 49‑75.  This gives an extensive discussion of Tweedie's method and various extensions to four containers, a barrel of unknown size, etc.

P. M. Lawrence.  An algebraic approach to some pouring problems.  MG 56 (No. 395) (Feb 1972) 13‑14.  Shows  II-(¥,b,c,d)  with  d £ b+c  and  GCD(b,c) = 1  is possible and extends to more jugs.

Louis Grant Brandes.  The Math. Wizard.  revised ed., J. Weston Walch, Portland, Maine, 1975.  Prob. 5: Getting five gallons of water:  II‑(¥,7,4,5).

Shakuntala Devi.  Puzzles to Puzzle You.  Orient Paperbacks (Vision Press), Delhi, 1976.

Prob. 53: The three containers, pp. 57 & 110.  III-(20;19,13,7;10).  Solution in 15 steps.  Looking at the triangular coordinates diagram of this, one sees that it is actually isomorphic to  II-(19,13,7;10)  and this can be seen by considering the amounts of empty space in the containers.

Prob. 132: Mr. Portchester's problem, pp. 82 & 132.  Same as Dudeney (1913).

Victor Serebriakoff.  A Mensa Puzzle Book.  Muller, London, 1982.  (Later combined with A Second Mensa Puzzle Book, 1985, Muller, London, as: The Mensa Puzzle Book, Treasure Press, London, 1991.)  Problem T.16: Pouring puttonos, part b, pp. 19-20 (1991: 37-38) & Answer 19, pp. 102-103 (1991: 118-119).  II-( ¥,5,3;1).

The Diagram Group.  The Family Book of Puzzles.  The Leisure Circle Ltd., Wembley, Middlesex, 1984.  Problem 161, with Solution at the back of the book.  II-(¥,5,3;1), which can be done as  II-(8,5,3;1).

D. St. P. Barnard.  50 Daily Telegraph Brain Twisters.  1985.  Op. cit. in 4.A.4.  Prob. 4: Measure for measure, pp. 15, 79‑80, 103.  Given 10 pints of milk, an 8 pint bowl, a jug and a flask.  He describes how he divides the milk in halves and you must deduce the size of the jug and the flask.

Colin J. H. McDiarmid  &  Jorge Ramirez Alfonsin.  Sharing jugs of wine.  Discrete Mathematics 125 (1994) 279-287.  Solves  I-(b+c,b,c)  and discusses the problem of getting from one state of the problem to another in a given number of steps, showing that  GCD(b,c) = 1  guarantees the graph is connected. indeed essentially cyclic.  Considers  GCD(b,c) ¹ 1.  Notes that the work done easily extends to  a > b + c.  Says the second author's PhD at Oxford, 1993, deals with more cases.

John P. Ashley.  Arithmetickle.  Arithmetic Curiosities, Challenges, Games and Groaners for all Ages.  Keystone Agencies, Radnor, Ohio, 1997.  P. 11: The spoon and the bottle.  Given a 160 ml bottle and a 30 ml spoon, measure 230 ml into a bucket.

 

          5.D.2.          RULER WITH MINIMAL NUMBER OF MARKS

 

Dudeney.  Problem 518: The damaged measure.  Strand Mag. (Sep 1920)  ??NX.  Wants a minimal ruler for 33 inches total length.  (=? MP 180)

Dudeney.  Problem 530: The six cottagers.  Strand Mag. (Jan 1921)  ??NX.  Wants 6 points on a circle to give all arc distances  1, 2, ..., 20.  (=? MP 181)

Percy Alexander MacMahon.  The prime numbers of measurement on a scale.  Proc. Camb. Philos. Soc. 21 (1922‑23) 651‑654.  He considers the infinite case, i.e.  a(0) = 0,  a(i+1) = a(i) + least integer which is not yet measurable.  This gives:  0, 1, 3, 7, 12, 20, 30, 44, ....

Dudeney.  MP.  1926. 

Prob. 180: The damaged measure, pp. 77 & 167.  (= 536, prob. 453, pp. 173, 383‑384.)  Mark a ruler of length 33 with 8 (internal) marks.  Gives 16 solutions.

Prob. 181: The six cottagers, pp. 77‑78 & 167.  = 536, prob. 454, pp. 174 & 384.

A. Brauer.  A problem of additive number theory and its application in electrical engineering.  J. Elisha Mitchell Sci. Soc. 61 (1945) 55‑56.  Problem arises in designing a resistance box.

Л. Редеи  & А. Реньи [L. Redei & A. Ren'i (Rényi)].  О представленин чисел  1, 2, ..., N  лосредством разностей [O predstavlenin chisel  1, 2, ... , N  losredstvom raznosteĭ (On the representation of  1, 2, ..., N  by differences)].  Мат. Сборник [Mat. Sbornik] 66 (NS 24) (1949) 385‑389.

Anonymous.  An unsolved problem.  Eureka 11 (Jan 1949) 11 & 30.  Place as few marks as possible to permit measuring integers up to  n.  For  n = 13,  an example is:  0, 1, 2, 6, 10, 13.  Mentions some general results for a circle.

John Leech.  On the representation of  1, 2, ..., n  by differences.  J. London Math. Soc. 31 (1956) 160‑169.  Improves Redei & Rényi's results.  Gives best examples for small  n.

Anon.  Puzzle column: What's your potential?  MTg 19 (1962) 35  &  20 (1962) 43.  Problem posed in terms of transformer outputs -- can we arrange  6  outputs to give every integral voltage up through  15?  Problem also asks for the general case.  Solution asserts, without real proof, that the optimum occurs with  0, 1, 4, 7, 10, ..., n‑11, n‑8, n‑5, n‑2  or its complement.

T. H. O'Beirne.  Puzzles and Paradoxes.  OUP, 1965.  Chap. 6 discusses several versions of the problem.

Gardner.  SA (Jan 1965) c= Magic Numbers, chap. 6.  Describes  1, 2, 6, 10  on a ruler  13  long.  Says  3  marks are sufficient on  9  and  4  marks on  12  and asks for proof of the latter and for the maximum number of distances that  3  marks on  12  can produce.  How can you mark a ruler  36  long?  Says Dudeney, MP prob. 180, believed that  9  marks were needed for a ruler longer than  33,  but Leech managed to show  8  was sufficient up to  36. 

C. J. Cooke.  Differences.  MTg 47 (1969) 16.  Says the problem in MTg 19 (1962) appears in H. L. Dorwart's The Geometry of Incidence (1966) related to perfect difference sets but with an erroneous definition which is corrected by references to H. J. Ryser's Combinatorial Mathematics.  However, this doesn't prove the assertions made in MTg 20.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, 1971.  Prob. 22: Starting and stopping, pp. 18 & 66.  Circular track, 1900 yards around.  How can one place marker posts so every multiple of 100 yards up to 1900 can be run.  Answer: at  0, 1, 3, 9, 15.

Gardner.  SA (Mar 1972) = Wheels, Chap. 15.

 

          5.D.3 FALSE COINS WITH A WEIGHING SCALE

 

H. S. Shapiro, proposer;  N. J. Fine, solver.  Problem E1399 -- Counterfeit coins.  AMM 67 (1960) 82  &  697‑698.  Genuines weigh  10,  counterfeits weigh  9.  Given  5  coins and a scale, how many weighings are needed to find the counterfeits?  Answer is  4.  Fine conjectures that the ratio of weighings to coins decreases to  0.

Kobon Fujimura & J. A. H. Hunter, proposers;  editorial solution.  There's always a way.  RMM 6 (Dec 1961) 47  &  7 (Feb 1962) 53.  (c= Fujimura's The Tokyo Puzzles (Muller, London, 1979), prob. 29: Pachinko balls, pp. 35 & 131.)  Six coins, one false.  Determine which is false and whether it is heavy or light in three weighings on a scale.  In fact one also finds the actual weights.

K. Fujimura, proposer;  editorial solution.  The 15‑coin puzzle.  RMM 9 (Jun 1962)  &  10 (Aug 1962) 40‑41.  Same problem with fifteen coins and four weighings.

 

          5.D.4.          TIMING WITH HOURGLASSES

 

          I have just started these and they are undoubtedly older than the examples here.  I don't recall ever seeing a general approach to these problems.

 

Simon Dresner.  Science World Book of Brain Teasers.  1962.  Op. cit. in 5.B.1.  Prob. 17: Two‑minute eggs, pp. 9 & 87.  Time  2  minutes with  3 & 5  minute timers.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.  No. 21: The sands of time, pp. 35 & 93.  Time  9  minutes with  4 & 7  minute timers.

David B. Lewis.  Eureka!  Perigee (Putnam), NY, 1983.  Pp. 73‑74.  Time  9  minutes with  4 & 7  minute timers.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995.  Chap. 1, prob. 8: Grandfather's breakfast, pp. 6 & 102.  Time  15  minutes with  7 & 11  minute timers.

 

          5.D.5.          MEASURE HALF A BARREL

 

          I have just started this and there must be much older examples.

 

Benson.  1904.  The water‑glass puzzle, p. 254.

Dudeney.  AM.  1917.  Prob. 364: The barrel puzzle, pp. 109-110 & 235.

King.  Best 100.  1927.  No. 1, pp. 7 & 38.

Collins.  Fun with Figures.  1928.  The dairymaid's problem, pp. 29-30.

William A. Bagley.  Puzzle Pie.  Vawser & Wiles, London, nd [BMC gives 1944].  [There is a revised edition, but it only affects material on angle trisection.]  No. 14: 'Arf an' 'arf, p. 15.

Anon.  The Little Puzzle Book.  Peter Pauper Press, Mount Vernon, NY, 1955.  P. 52: The cider barrel.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 87: But me no butts, pp. 42 & 88.

Richard I. Hess.  Email Christmas message to NOBNET, 24 Nov 2000.  Solution sent by Nick Baxter on the same day.  You have aquaria (assumed cuboidal) which hold 7 and 12 gallons and a water supply.  The 12 gallon aquarium has dots accurately placed in the centre of each side face.  How many steps are required to get 8 gallons into the 12 gallon aquarium?  Fill the 12 gallon aquarium and tilt it on one corner so the water level passes through the centres of the two opposite faces.  This leaves 8 gallons!  Nick says this is two steps.

 

5.E.    EULER CIRCUITS AND MAZES

 

          Euler circuits have been used in primitive art, often as symbols of the passage of the soul to the land of the dead.  [MTg 110 (Mar 1985) 55] shows examples from Angola and New Hebrides.  See Ascher (1988 & 1991) for many other examples from other cultures.

 

                                        ┌────┬─────────┬────┐

                                                        

                                        ├────┴────┬────┴────┤

                                                         

                                        └─────────┴─────────┘

 

          Above is the 'five‑brick pattern'.  See:  Clausen,  Listing,  Kamp,  White,  Dudeney,  Loyd Jr,  Ripley,  Meyer,  Leeming,  Adams,  Anon.,  Ascher.  Prior to Loyd Jr, the problem asked for the edges to be drawn in three paths, but about 1920 the problem changed to drawing a path across every wall.

          Trick solutions:  Tom Tit,  Dudeney (1913),  Houdini,  Loyd Jr,  Ripley,  Meyer,  Leeming,  Adams,  Gibson,  Anon. (1986).

          Non-crossing Euler circuits:  Endless Amusement II,  Bellew,  Carroll 1869,  Mittenzwey,  Bile Beans,  Meyer,  Gardner (1964),  Willson,  Scott,  Singmaster.

          Kn  denotes the complete graph on  n  vertices.

 

 

Matthäus Merian the Elder.  Engraved map of Königsberg.  Bernhard Wiezorke has sent me a coloured reproduction of this, dated as 1641.  He used an B&W version in his article: Puzzles und Brainteasers; OR News, Ausgabe 13 (Nov 2001) 52-54.  BLW use a B&W version on their dust jacket and on p. 2 which they attribute to M. Zeiller; Topographia Prussiae et Pomerelliae; Frankfurt, c1650.  I have seen this in a facsimile of the Cosmographica due to Merian in the volume on Brandenburg and Pomerania, but it was not coloured.  There seem to be at least two versions of this picture --??CHECK.

L. Euler.  Solutio problematis ad geometriam situs pertinentis.  (Comm. Acad. Sci. Petropol. 8 (1736(1741)) 128‑140.)  = Opera Omnia (1) 7 (1923) 1‑10.  English version:  Seven Bridges of Königsberg  is in:  BLW, 3‑8;  SA 189 (Jul 1953) 66‑70;  World of Mathematics, vol. 1, 573‑580;  Struik, Source Book, 183‑187.

My late colleague Jeremy Wyndham became interested in the seven bridges problem and made inquiries which turned up several maps of Königsberg and a list of all the bridges and their dates of construction (though there is some ambiguity about one bridge).  The first bridge was built in 1286 and until the seventh bridge of 1542, an Euler path was always possible.  No further bridge was built until a railway bridge in 1865 which led to Saalschütz's 1876 paper -- see below.  In 1905 and later, several more bridges were added, reaching a maximum of ten bridges in 1926 (with 4512 paths from the island), then one was removed in 1933.  Then a road bridge was added, but it is so far out that it does not show on any map I have seen.  Bombing and fighting in 1944-1945 apparently destroyed all the bridges and the Russians have rebuilt six or seven of them.  I have computed the number of paths in each case -- from 1865 until 1935 or 1944, there were always Euler paths.

L. Poinsot.  Sur les polygones et les polyèdres.  J. École Polytech. 4 (Cah. 10) (1810) 16‑48.  Pp. 28‑33 give Euler paths on  K2n+1  and Euler's criterion.  Discusses square with diagonals.

Endless Amusement II.  1826?  Prob. 34, p. 211.  Pattern of two overlapping squares has a non-crossing Euler circuit.

Th. Clausen.  De linearum tertii ordinis propietatibus.  Astronomische Nachrichten 21 (No. 494) (1844), col. 209‑216.  At the very end, he gives the five‑brick pattern and says that its edges cannot be drawn in three paths.

J. B. Listing.  Vorstudien zur Topologie.  Göttinger Studien 1 (1847) 811‑875.  ??NYR.  Gives five brick pattern as in Clausen.

??  Nouv.  Ann. Math. 8 (1849?) 74.  ??NYS.  Lucas says this poses the problem of finding the number of linear arrangements of a set of dominoes.  [For a double  N  set,  N = 2n,  this is  (2n+1)(n+1)  times the number of circular arrangements, which is  n2n+1  times the number of Euler circuits on  K2n+1.]

É. Coupy.  Solution d'un problème appartenant a la géométrie de situation, par Euler.  Nouv. Ann. Math. 10 (1851) 106‑119.  Translation of Euler.  Translator's note on p. 119 applies it to the bridges of Paris.

The Sociable.  1858.  Prob. 7: Puzzle pleasure garden, pp. 288 & 303.  Large maze-like garden and one is to pass over every path just once -- phrased in verse.  = Book of 500 Puzzles, 1859, prob. 7, pp. 6 & 21.  = Illustrated Boy's Own Treasury, 1860, prob. 49, pp. 405 & 443.  In fact, if one goes straight across every intersection, one finds the path, so this is really almost a unicursal problem.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 587, pp. 297 & 410: Ariadnerätsel.  Three diagrams to trace with single lines.  No attempt to avoid crossings.

Frank Bellew.  The Art of Amusing.  Carleton, NY (&  Sampson Low & Co., London), 1866 [C&B list a 1871];  John Camden Hotten, London, nd [BMC & NUC say 1870] and John Grant, Edinburgh, nd [c1870 or 1866?], with slightly different pagination.  1866: pp. 269-270;  1870: p. 266.  Two overlapping squares have a non-crossing Euler circuit.

Lewis Carroll.  Letter of 22 Aug 1869 to Isabel Standen.  Taken from: Stuart Dodgson Collingwood; The Life and Letters of Lewis Carroll; T. Fisher Unwin, London, (Dec 1898), 2nd ed., Jan 1899, p. 370: "Have you succeeded in drawing the three squares?"  On pp. 369-370, the recipient is identified as Isabel Standen and she is writing Collingwood, apparently sending him the letter.  Collingwood interpolates:  "This puzzle was, by the way, a great favourite of his; the problem is to draw three interlaced squares without going over the same lines twice, or taking the pen off the paper".  But no diagram is given. 

                    Dudeney; Some much‑discussed puzzles; op. cit. in 2; 1908, quotes Collingwood, gives the diagram and continues: "This is sometimes ascribed to him [i.e. Carroll] as its originator, but I have found it in a little book published in 1835."  This was probably a printing of Endless Amusement II, qv above and in Common References, though this has two interlaced squares.  John Fisher; The Magic of Lewis Carroll; op. cit. in 1; pp. 58‑59, says Carroll would ask for a non‑crossing Euler circuit, but this is not clearly stated in Collingwood.  Cf Carroll-Wakeling, prob. 29: The three squares, pp. 38 & 72, which clearly states that a non-crossing circuit is wanted and notes that there is more than one solution.  Cf Gardner (1964).  Carroll-Gardner, pp. 52-53.

Mittenzwey.  1880.  Prob. 269-279, pp. 47-48 & 98-100;  1895?: 298-308, pp. 51-52 & 100‑102;  1917: 298-308, pp. 46-48 & 95-97.  Straightforward unicursal patterns.  The first is  K5,  but one of the diagonals was missing in my copy of the 1st ed. -- the path is not to use two consecutive outer edges.  The third is the 'envelope' pattern.  The fourth is three overlapping squares, where the two outer squares just touch in the middle.  The last is a simple maze with no dead ends and the path is not to cross itself.  See also the entry for Mittenzwey in 5.E.1, below.

M. Reiss.  Évaluation du nombre de combinaisons desquelles les 28 dés d'un jeu de dominos sont susceptibles d'après la règle de ce jeu.  Annali di Matematica Pura ed Applicata (2) 5 (1871) 63‑120.  Determines the number of linear arrangements of a double‑6 set of dominoes, which gives the number of Euler circuits on  K7.

L. Saalschütz.  [Report of a lecture.]  Schriften der Physikalisch‑Ökonomischen Gesellschaft zu Königsberg 16 (1876) 23‑24.  Sketches Euler's work, listing the seven bridges.  Says that a recent railway bridge, of 1865, connecting regions  B  and  C  on Euler's diagram, can be considered within the walkable region.  He shows there are  48 x 2 x 4  = 384  possible paths -- the  48  are the lists of regions visited starting with  A;  the  2  corresponds to reversing these lists;  the  4 (= 2 x 2)  corresponds to taking each of the two pairs of bridges connecting the same regions in either order,  He lists the  48  sequences of regions which start at  A.  I wrote a program to compute Euler paths and I tested it on this situation.  I find that Saalschütz has omitted two cases, leading to four sequences or  16  paths starting at  A  or  32  paths considering both directions.  That is, his  48  should be  52  and his  384  should be  416.

Kamp.  Op. cit. in 5.B.  1877.  Pp. 322‑327 show several unicursal problems.

No. 8 is the five‑brick pattern as in Clausen.

No. 10 is two overlapping squares.

No. 11 is a diagram from which one must remove some lines to leave an Eulerian figure.

C. Hierholzer.  Ueber die Möglichkeit, einen Linienzug ohne Wiederholung und ohne Unterbrechung zu umfahren.  Math. Annalen 6 (1873) 30‑32.  (English is in BLW, 11‑12.)

G. Tarry.  Géométrie de situation: Nombre de manières distinctes de parcourir en une seule course toutes les allées d'un labyrinthe rentrant, en ne passant qu'une seule fois par chacune des allées.  Comptes Rendus Assoc. Franç. Avance. Sci. 15, part 2 (1886) 49‑53 & Plates I & II.  General technique for the number of Euler circuits.

Lucas. RM2.  1883.  Le jeu de dominos -- Dispositions rectilignes, pp. 63‑77  &  Note 1:  Sur le jeu de dominos, p. 229.

                    RM4.  1894.  La géométrie des réseaux et le problème des dominos, pp. 123‑151. 

                    Cites Reiss's work and says (in RM4) that it has been confirmed by Jolivald.  The note in RM2 is expanded in RM4 to explain the connection between dominoes and  K2n+1.  There are obviously 2 Euler circuits on  K3.  He sketches Tarry's method and uses it to compute that  K5  has  88  Euler circuits and  K7  has  1299 76320.  [This gives  28 42582 11840  domino rings for the double-6 set.]  He says Tarry has found that  K9  has  911 52005 70212 35200.

Tom Tit, vol. 3.  1893.  Le rectangle et ses diagonales, pp. 155-156.  = K, no. 16: The rectangle and its diagonals, pp. 46‑48.  = R&A, The secret of the rectangle, p. 100.  Trick solutions by folding the paper and making an arc on the back.

Hoffmann.  1893.  Chap. X, no. 9: Single‑stroke figures, pp. 338 & 375 = Hoffmann-Hordern, pp. 230-231.  Three figures, including the double crescent 'Seal of Mahomet'.  Answer states Euler's condition.

Dudeney.  The shipman's puzzle.  London Mag. 9 (No. 49) (Aug 1902) 88‑89  &  9 (No. 50) (Sep 1902) 219 (= CP, prob. 18, pp. 40‑41 & 173).  Number of Euler circuits on  K5.

Benson.  1904.  A geometrical problem, p. 255.  Seal of Mahomet.

William F. White.  A Scrap‑Book of Elementary Mathematics.  Open Court, 1908.  [The 4th ed., 1942, is identical in content and pagination, omitting only the Frontispiece and the publisher's catalogue.]  Bridges and isles, figure tracing, unicursal signatures, labyrinths, pp. 170‑179.  On p. 174, he gives the five‑brick puzzle, asking for a route along its edges.

Dudeney.  Perplexities: No. 147: An old three‑line puzzle.  Strand Magazine 46 (Jul 1913) 110  &  (Aug 1913) 221.  c= AM, prob. 239: A juvenile puzzle, pp. 68‑69 & 197.  Five‑brick form to be drawn or rubbed out on a board in three strokes.  Either way requires doing two lines at once, either by folding the paper as you draw or using two fingers to rub out two lines at once.  "I believe Houdin, the conjurer, was fond of showing this to his child friends, but it was invented before his time -- perhaps in the Stone Age."

Loyd.  Problem of the bridges.  Cyclopedia, 1914, pp. 155 & 359‑360.  = MPSL1, prob. 28, pp. 26‑27 & 130‑131.  Eight bridges.  Asks for number of routes.

Loyd.  Puzzle of the letter carrier's route.  Cyclopedia, 1914, pp 243 & 372.  Asks for a circuit on a  3 x 4  array with a minimal length of repeated path.

Dudeney.  AM.  1917. 

Prob. 242: The tube inspector's puzzle, pp. 69 & 198.  Minimal route on a  3 x 4  array.

Prob. 261: The monk and the bridges, pp. 75-76 & 202-203.  River with one island.  Four bridges from island, two to each side of the river, and another bridge over the river.  How many Euler paths from a given side of the river to the other?  Answer:  16.

Collins.  Book of Puzzles.  1927.  The fly on the octahedron, pp. 105-108.  Asserts there are  1488  Euler circuits on the edges of an octahedron.  He counts the reverse as a separate circuit.

Harry Houdini [pseud. of Ehrich Weiss]  Houdini's Book of Magic.  1927 (??NYS); Pinnacle Books, NY, 1976, p. 19: Can you draw this?  Take a square inscribed in a circle and draw both diagonals.  "The idea is to draw the figure without taking your pencil off the paper and without retracing or crossing a line.  There is a trick to it, but it can be done.  The trick in drawing the figure is to fold the paper once and draw a straight line between the folded halves; then, not removing your pencil, unfold the paper.  You will find that you have drawn two straight lines with one stroke.  The rest is simple."  This perplexed me for some time, but I believe the idea is that holding the pencil between the two parts of the folded sheet and moving the pencil parallel to the fold, one can draw a line, parallel to the fold, on each part.

Loyd Jr.  SLAHP.  1928.  Pp. 7‑8.  Discusses what he calls the "Five‑brick puzzle", the common pattern of five rectangles in a rectangle.  He says that the object was to draw the lines in four strokes -- which is easily done -- but that it was commonly misprinted as three strokes, which he managed to do by folding the paper.  He says "a similar puzzle ... some ten or fifteen years ago" asked for a path crossing each of the 16 walls once, which is also impossible.

The Bile Beans Puzzle Book.  1933. 

No. 32.  Draw the triangular array of three on an edge without crossing.

No. 36.  Draw the five-brick pattern in three lines.  Folds paper and draws two lines at once.

R. Ripley.  Believe It Or Not!  Book 2.  (Simon & Schuster, 1931);  Pocket Books, NY, 1948, pp. 70‑71.  = Omnibus Believe It Or Not! Stanley Paul, London, nd [c1935?], p. 270.  Gives the five‑brick problem of drawing a path crossing each wall once, with the trick solution having the path going along a wall.  Asserts "This unicursal problem was solved thus by the great Euler himself." and cites the Euler paper above!!

Meyer.  Big Fun Book.  1940. 

Tryangle, pp. 98 & 731.  Triangle subdivided into triangles, with three small triangles along each edge.  Draw an Euler circuit without crossings.

Cutting the walls, pp. 637 & 794.  Five-brick problem.  Solution has line crossing through a vertex.

Ern Shaw.  The Pocket Brains Trust - No. 2.  W. H. Allen, London, nd but inscribed 1944.  Prob. 29: Five bricks teaser, pp. 10 & 39.

Leeming.  1946.  Chap. 6, prob. 2: Through the walls, pp. 70 & 184.  Five‑brick puzzle, with trick solution having the path go through an intersection.

John Paul Adams.  We Dare You to Solve This!.  Op. cit. in 5.C.  1957?  Prob. 49: In just one line, pp. 30 & 48-49.  Five-brick puzzle, with answer having the path going along a wall, as in Ripley.  Asserts Euler invented this solution.

Gibson.  Op. cit. in 4.A.1.a.  1963.  Pp. 70 & 75: The "impossible" diagram.  Same as Tom Tit.

Gardner.  SA (Apr 1964) = 6th Book, chap. 10.  Says Carroll knew that a planar Eulerian graph could be drawn without crossings.  Gives a method of O'Beirne for doing this -- two colour the regions and then make a path which separates the colours into simply connected regions.

Ripley's Puzzles and Games.  1966.  P. 39.  Euler paths on the 'envelope', i.e. a rectangle with its diagonals drawn and an extra connection between the top corners, looking like an unfolded envelope.  Asserts the envelope has 50 solutions, but it is not clear if the central crossing is a further vertex.  I did this by hand but did not get 50, so I wrote a program to count Euler paths.  If the central crossing is not a vertex, then I find 44 paths from one of the odd vertices to the other, and of course 44 going the other way -- and I had found this number by hand.  However, if the central crossing is a vertex, then my hand solution omitted some cases and the computer found 120 paths from one odd vertex to the other. 

                    Pp. 40-43 give many problems of drawing non-crossing Euler paths or circuits.

W. Wynne Willson.  How to abolish cross‑roads.  MTg 42 (Spring 1968) 56‑59.  Euler circuit of a planar graph can be made without crossings.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  Tempo (Grosset & Dunlap), NY, 1973 (& 1978?? -- both dates are given -- I'm presuming the 1978 is a 2nd ptg or a reissue under a different imprint??).  One line/no crossing, pp. 85-86.  Non-crossing Euler circuits on the triangular array of side 3 and non-crossing Euler paths on the 'envelope' -- cf under Ripley's, above.  Asserts the envelope has 50 solutions.  I adapted the program mentioned above to count the number of non-crossing Euler paths -- one must rearrange the first case as a planar graph -- and there are  16  in the first case and  26  in the second case.  Taking the reversals doubles these numbers so it is possible that the Scotts meant the second case and missed one path and its reversal.

David Singmaster, proposer;  Jerrold W. Grossman & E. M. Reingold, solvers.  Problem E2897 -- An Eulerian circuit with no crossings.  AMM 88:7 (Aug 1981) 537-538  &  90:4 (Apr 1983) 287-288.  A planar Eulerian graph can be drawn with no crossings.  Solution cites some previous work.

Anon. [probably Will Shortz  ??check with Shortz].  The impossible file.  No. 2: In just one line.  Games (Apr 1986) 34 & 64  & (Jul 1986) 64.  Five brick pattern -- draw a line crossing each wall once.  Says it appeared in a 1921 newspaper [perhaps by Loyd Jr??].  Gives the 1921 solution where the path crosses a corner, hence two walls at once.  Also gives a solution with the path going along a wall.  In the July issue, Mark Kantrowitz gives a solution by folding over a corner and also a solution on a torus.

Marcia Ascher.  Graphs in cultures: A study in ethnomathematics.  HM 15 (1988) 201‑227.  Discusses the history of Eulerian circuits and non-crossing versions and then exposits many forms of the idea in many cultures.

Marcia Ascher.  Ethnomathematics.  Op. cit. in 4.B.10.  1991.  Chapter Two: Tracing graphs in the sand, pp. 30-65.  Sketches the history of Eulerian graphs with some interesting references -- ??NYS.  Describes graph tracing in three cultures: the Bushoong and the Tshokwe of central Africa and the Malekula of Vanuatu (ex-New Hebrides).  Extensive references to the ethnographic literature.

 

          5.E.1. MAZES

 

          This section is mainly concerned with the theory.  The history of mazes is sketched first, with references to more detailed sources.  There is even a journal, Caerdroia (53 Thundersley Grove, Thundersley, Essex, SS7 3EB, England), devoted to mazes and labyrinths, mostly concentrating on the history.  It is an annual, began in 1980 and issue 31 appeared in 2000.

          Mazes are considered under Euler Circuits, since the method of Euler Circuits is often used to find an algorithm.  However, some mazes are better treated as Hamiltonian Circuits -- see 5.F.2.

          A maze can be considered as a graph formed by the nodes and paths -- the path graph.  For the usual planar maze, one can also look at the graph formed by the walls -- the wall graph, which is a kind of dual to the path graph.  In later mazes, the walls do not form a connected whole, and an isolated part of the wall appears as a region or 'face' in the path graph.  Such isolated bits of walling are sometimes called islands, but they are the same as the components of the wall graph, with the outer wall being one component, so the number of components is one more than the number of islands.  The 'hand-on-wall' method will solve a maze if and only if the goals are adjacent to walls in the component of the outer wall. 

          A 'ring maze' is a plate with holes and raised areas with an open ring which must be removed by moving it from hole to hole.  I have put these in  11.K.5  as they are a kind of mechanical or topological puzzle, though there are versions with a simple two legged spacer.

 

                    HISTORICAL SOURCES

 

W. H. Matthews.  Mazes & Labyrinths:  A General Account of Their History and Developments.  Longmans, Green and Co., London, 1922.  = Mazes and Labyrinths:  Their History and Development.  Dover, 1970.  (21 pages of references.)  [For more about the book and the author, see:  Zeta Estes; My Father, W. H. Matthews; Caerdroia (1990) 6-8.]

Walter Shepherd.  For Amazement Only.  Penguin, 1942;  Let's go amazing, pp. 5-12.  Revised as:  Mazes and Labyrinths -- A Book of Puzzles.  Dover, 1961;  Let's go a‑mazing, pp. v‑xi.  (Only a few minor changes are made in the text.)  Sketch of the history.

Sven Bergling invented the rolling ball labyrinth puzzle/game and they began being produced in 1946.  [Kenneth Wells; Wooden Puzzles and Games; David & Charles, Newton Abbot, 1983, p. 114.]

Walter Shepherd.  Big Book of Mazes and Labyrinths.  Dover, 1973, More amazement, pp. vii-x.  Extends the historical sketch in his previous book, arguing that mazes with multiple choices perhaps derive from Iron Age hill forts whose entrances were designed to confuse an enemy.

Janet Bord.  Mazes and Labyrinths of the World.  Latimer, London, 1976.  (Extensively illustrated.)

Nigel Pennick.  Mazes and Labyrinths.  Robert Hale, London, 1990.

Adrian Fisher [& Georg Gerster (photographer)].  The Art of the Maze.  Weidenfeld and Nicolson, London, 1990.  (Also as:  Labyrinth; Solving the Riddle of the Maze; Harmony (Crown Publishers), NY, 1990.)  Origins and History occupies pp. 11-56, but he also describes many recent developments and innovations.  He has convenient tables of early examples.

Adrian Fisher & Diana Kingham.  Mazes.  Shire Album 264.  Shire, Aylesbury, 1991.

Adrian Fisher & Jeff Saward.  The British Maze Guide.  Minotaur Designs, St. Alban's, 1991.

 

 

                    ┌────────────────────────────────────────────┐

                                                               

                      ┌──────────────────────────────────────┐ 

                                                             

                        ┌────────────────────────────────┐   

                                                           

                          ┌──────────────────────────┐     

                                                         

                            ┌────────────────────┐       

                                                       

                              ┌──────────────┐         

                                                     

                                ┌────────┐           

                                                   

                                  ┌──┐             

                                                 

                                   J              

                            └─────┘    └─────┘        

                                                       

                          └───────────┼───────────┘       

                                                         

                        └───────────┐    ┌───────────┘     

                                                         

                      └────────────        ────────────┘   

                                                           

                    └─────────────────┤    └─────────────────┘ 

                                                              

                                        ├───────────────────────┘

                                       Ý

                                       

 

                    HISTORICAL SKETCH

 

          Up to about the 16C, all mazes were unicursal, i.e. with no decision points.  The word labyrinth is sometimes used to distinguish unicursal mazes from others, but this distinction is not made consistently.  Until about 1000, all mazes were of the classical 'Cretan' seven-ring type shown above.  (However, see Shepherd's point in his 1973 book, above.)  The oldest examples are rock carvings, the earliest being perhaps that in the Tomba del Labirinto at Luzzanas, Sardinia, c‑2000 [Fisher, pp. 12, 25, 26, with photo on p. 12].  (In fact, Luzzanas is a local name for an uninhabited area of fields, so does not appear on any ordinary map.  It is near Benetutti.  See my A Mathematical Gazetteer or Mazing in Sardinia (Caerdroia 30 (1999) 17-21).  Jeff Saward writes that current archaeological feeling is that the maze is Roman, though the cave is probably c-2000.)  On pottery, there are labyrinths on fragments, c‑1300, from Tell Rif'at, Syria [the first photos of this appeared in [Caerdroia 30 (2000) 54-55]), and on tablets, c‑1200, from Pylos.  Fisher [p. 26] lists the early examples.  Staffen Lundán; The labyrinth in the Mediterranean; Caerdroia 27 (1996) 28-54, catalogues all known 'Cretan' labyrinths from prehistory to the end of antiquity, c250, excluding the Roman 'spoked' form.  All these probably had some mystical significance about the difficulty of reaching a goal, often with substantial mythology -- e.g. Theseus in the Labyrinth or, later, the Route to Jerusalem. 

          Roman mosaics were unicursal but essentially used the Cretan form four times over in the four corners.  Lundán, above, calls these 'spoked'.  Most of the extant examples are 2C‑4C, but some BC examples are known -- the earliest seems to be c-110 at Selinunte, Sicily.  Fisher [pp. 36-37] lists all surviving examples.  Saward says the earliest Roman example is at Pompeii, so £ 79.

          In the medieval period, the Christians developed a quite different unicursal maze.  See Fisher [pp. 60-67] for detailed comparison of this form with the Roman and Cretan forms.  The earliest large Christian example is the Chemin de Jerusalem of 1235 on the floor of Chartres Cathedral.  Fisher [pp. 41 & 48] lists early and later Christian examples.

          The legendary Rosamund's Bower was located in Woodstock Park, Oxfordshire, and its purported site is marked by a well and fountain.  It was some sort of maze to conceal Rosamund Clifford, the mistress of Henry II (1133‑1189), from the Queen, Eleanor of Aquitaine.  Legend says that about 1176, Eleanor managed to solve the maze and confronted Rosamund with the choice of a dagger or poison -- she drank the poison and Henry never smiled again.  [Fisher, p. 105].  Historically, Henry had imprisoned Eleanor for fomenting rebellion by her sons and Rosamund was his acknowledged mistress.  Rosamund probably spent her last days at a nunnery in Godstow, near Oxford.  The legend of the bower dates from the 14C and her murder is a later addition [Collins, Book of Puzzles, 1927, p. 121.]  In the 19C, many puzzle collections had a maze called Rosamund's Bower.

          The earliest record of a hedge maze is of one destroyed in a siege of Paris in 1431.

          Non-unicursal mazes and islands in the wall graph start to appear in the late 16C.  Matthews [p. 96] says that: "A simple "interrupted-circle" type of labyrinth was adopted as a heraldic device by Gonzalo Perez, a Spanish ecclesiastic ... and published ... in 1566 ..." in his translation of the Odyssey.  Matthews doesn't show this, but he then [pp. 96-97] describes and illustrates a simple maze used as a device by Bois-dofin de Laval, Archbishop of Embrum.  He copies it from Claude Paradin; Devises Héroiques et Emblèmes of the early 17C.  It has four entrances and possibly three goals, with walls having 8 components, two being part of the outer wall.  The central goals is accessible from two of the entrances, but the two minor goals are each accessible from just one of the other entrances.  Presumably this sort of thing is what Matthews meant as an "interrupted circle".

          However, Saward has found a mid 15C anonymous English poem, The Assembly of Ladies, which describes the efforts of a group of ladies to reach the centre of a maze, which, as he observes, implies there must be some choices involved.

          [Matthews, p. 114] has three examples from a book by Androuet du Cerceau; Les Plus Excellents Bastiments de France of 1576.  Fig. 82 was in the gardens at Charleval and has four entrances, only one of which goes to the central goal.  There are four minor goals.  The N entrance connects to the NE and SE goals, with several dead ends.  The E entrance is a dead end.  The S entrance goes to the SW goal.  The W entrance goes to the central goal, but the NW goal is on an island, though 'left-hand-on-wall' goes past it.  Figs. 83 and 84 are essentially identical and seem to be corruptions of unicursal examples so that most of the maze is bypassed.  In fig. 84, one has to walk around to the back of the maze to find the correct entrance to get to the central goal, which is an interesting idea.  A small internal change in both cases and moving the entrances converts them to a standard unicursal pattern.

          Matthews' Chap. XIII [pp. 100-109] is on floral mazes and reproduces some from Jan Vredeman De Vries; Hortorum Viridariorumque Formae; Antwerp, 1583.  Fig. 74 is one of these and has two components and a short dead-end, but the 'hand-on-wall' rule solves it.  Fig. 73 is another of De Vries's, but it is not all shown.  It appears to have two entrances and there is certainly a decision point by the far gate, but one route goes to the apparent exit at the bottom of the page.  There is a small dead end near the central goal.  Fig. 78 shows a maze from a 17C manuscript book in the Harley Manuscripts at the BL, identified on p. 224 as BM Harl. 5308 (71, a, 12).  This has two components with the central goal in the inner component, so the 'hand-on-wall' rule fails, but it brings you within sight of the centre and Matthews describes it as unicursal!  Fig. 79 is from Adam Islip; The Orchard and the Garden, compiled from continental sources and published in 1602.  It has 5 components, but four of these are small enclosures which could be considered as minor goals, especially if they had seats in them.  The 'hand-on-wall' rule gets to the central goal.  There is a lengthy dead end which goes to two of the inner islands.  Fig. 80 is from a Dutch book: J. Commelyn; Nederlantze Hesperides of 1676.  It has two components, a central goal and four minor goals.  The 'hand-on-wall' gets you to the centre and passes two minor goals.  One minor goal is on a dead end so 'left-hand-on-wall' gets to it, but 'right-hand-on-wall' does not.  The fourth minor goal is on the island.

          At Versailles, c1675, André Le Nôtre built a Garden Maze, but the objective was to visit, in correct order, 40 fountains based on Aesop's Fables.  Each node of the maze had at least one fountain.  Some fountains were not at path junctions, but one can consider these as nodes of degree two.  This is an early example of a Hamiltonian problem, except that one fountain was located at the end of a short dead end.  [Fisher, pp. 49, 79, 130 & 144-145, with contemporary map on p. 144.  Fisher says there are 39 fountains, and the map has 40.  Close examination shows that the map counts two statues at the entrance but omits to count a fountain between numbers 37 and 38.  Matthews, pp. 117‑121, says it was built by J. Hardouin-Mansart and his map has 39 fountains.]  It has a main entrance and exit but there is another exit, so the perimeter wall already has three components, and there are 14 other components.  Sadly, it was destroyed in 1775.

          Several other mazes, of increasing complexity, occur in the second half of the 17C [Matthews, figs. 93-109, opp. p. 120 - p. 127].  Several of these could be from 20C maze books.  Fig. 94, designed for Chantilly by Le Nôtre, is surprisingly modern in that there are eight paths spiralling to the centre.  The entrance path takes you directly to the centre, so the real problem is getting back out!  One of the mazes presently at Longleat has this same feature.

          The Hampton Court Maze, planted c1690, is the oldest extant hedge maze and one of the earliest puzzle mazes.  ([Christopher Turner; Hampton Court, Richmond and Kew Step by Step; (As part of: Outer London Step by Step, Faber, 1986); Revised and published in sections, Faber, 1987, p. 16] says the present shape was laid out in 1714, replacing an earlier circular shape, but I haven't seen this stated elsewhere.)  Matthews [p. 128] says it probably replaced an older maze.  It has dead ends and one island, i.e. the graph has two components, though the 'hand on wall' rule will solve it.

          The second Earl Stanhope (1714-1786) is believed to be the first to design mazes with the goal (at the centre) surrounded by an island, so that the 'hand on wall' rule will not solve it.  It has seven components and only a few short dead ends..  The fourth Earl planted one of these at Chevening, Kent, in c1820 and it is extant though not open to the public.  [Fisher, p. 71, with photo on p. 72 and diagram on p. 73.]  However, investigation in Matthews revealed the earlier examples above.  Further Bernhard Wiezorke (below at 2001) has found a hedge maze in Germany, dating from c1730, which is not solved by the 'hand on wall' rule.  This maze has 12 components.

          In 1973, Stuart Landsborough, an Englishman settled at Wanaka, South Island, New Zealand, began building his Great Maze.  This was the first of the board mazes designed by Landsborough which were immensely popular in Japan.  Over 200 were built in 1984-1987, with 20 designed by Landsborough.  Many of these were three dimensional -- see below.  About 60 have been demolished since then.  [Fisher, pp. 78‑79 & 118-121 has 6 colour photos, pp. 156-157 lists Landsborough's designs.]

          If Minos' labyrinth ever really existed, it may have been three dimensional and there may have been garden examples with overbridges, but I don't know of any evidence for such early three dimensional mazes.  Lewis Carroll drew mazes which had paths that crossed over others making a simple three dimensional maze, in his Mischmasch of c1860, see below.  John Fisher [The Magic of Lewis Carroll; (Nelson, 1973), Penguin, 1975, pp. 19-20] gives this and another example.  Are there earlier examples?  Boothroyd & Conway, 1959, seems to be the earliest cubical maze.  Much more complex versions were developed by Larry Evans from about 1970 and published in a series of books, starting with 3-Dimensional Mazes (Troubador Press, San Francisco, 1976).  His 3‑Dimensional Maze Art (Troubador, 1980) sketches some general history of the maze and describes his development of pictures of three dimensional mazes.  The first actual three dimensional maze seems to be Greg Bright's 1978 maze at Longleat House, Warminster.  [Fisher, pp. 74, 76, 94-95 & 152-153, with colour photos on pp. 94-95.]  Since then, Greg Bright, Adrian Fisher, Randoll Coate, Stuart Landsborough and others have made many innovations.  Bright seems to have originated the use of colour in mazes c1980 and Fisher has extensively developed the idea.  [Fisher, pp. 73-79.]

 

Abu‘l-Rayhan Al‑Biruni (= ’Abû-alraihân [the  h  should have an underdot] Muhammad ibn ’Ahmad [the  h  should have an underdot] Albêrûnî).  India.  c1030.  Chapter XXX.  IN: Al‑Beruni's India, trans. by E. C. Sachau, 2 vols., London, 1888,  vol. 1, pp. 306-307 (= p. 158 of the Arabic ed., ??NYS).  In describing a story from the fifth and sixth books of the Ramayana, he says that the demon Ravana made a labyrinthine fortress, which in Muslim countries "is called Yâvana-koti, which has been frequently explained as Rome."  He then gives "the plan of the labyrinthine fortress", which is the classical Cretan seven-ring form.  Sachau's notes do not indicate whether this plan is actually in the Ramayana, which dates from perhaps -300.

Pliny.  Natural History.  c77.  Book 36, chap. 19.  This gives a brief description of boys playing on a pavement where a thousand steps are contained in a small space.  This has generally been interpreted as referring to a maze, but it is obviously pretty vague.  See: Michael Behrend; Julian and Troy names; Caerdroia 27 (1996) 18-22, esp. note 5 on p. 22.

Pacioli.  De Viribus.  c1500.  Part II: Cap. (C)XVII.  Do(cumento). de saper fare illa berinto con diligentia secondo Vergilio, f. 223v  = Peirani 307-308.  A sheet (or page) of the MS has been lost.  Cites Vergil, Æneid, part six, for the story of Pasiphæ and the Minotaur, but the rest is then lost.

Sebastiano Serlio.  Architettura, 5 books, 1537-1547.  The separate books had several editions before they were first published together in 1584.  The material of interest is in Book IV which shows two unicursal mazes for gardens.  I have seen the following.

                    Tutte l'Opere d'Architetture et Prospetiva, ....  Giacomo de'Franceschi, Venice, 1619;  facsimile by Gregg Press, Ridgewood, New Jersey, 1964.  F. 199r shows the designs and f. 197v has some text, partly illegible in my photocopy.  [Cf Caerdroia 30 (1999) 15.]

                    Sebastiano Serlio on Architecture  Volume One  Books I-V of 'Tutte l'Opere d'Architettura et Prospetiva'.  Translated and edited by Vaughan Hart and Peter Hicks.  Yale Univ. Press, New Haven, 1996.  P. 388 shows the designs and p. 389 has the text, saying these 'are for the compartition of gardens'.  The sidenotes state that these pages are ff. LXXVr and LXXIIIIr of the 3rd ed. of 1544  and  ff. 198v-199r and 197v-198r of the 1618/19 ed.

William Shakespeare.  A Midsummer Night's Dream.  c1610.  Act II, scene I, lines 98-100:  "The nine men's morris is fill'd up with mud,  And the quaint mazes in the wanton green  For lack of tread are undistinguishable."  Fiske 126 opines that the latter two lines may indicate that the board was made in the turf, though he admits that they may refer just to dancers' tracks, but to me it clearly refers to turf mazes.

John Cooke.  Greene's Tu Quoque; or the Cittie Gallant; a Play of Much Humour.  1614.  ??NYS -- quoted by Matthews, p. 135.  A challenge to a duel is given by Spendall to Staines. 

                    Staines.   I accept it ; the meeting place? 

                    Spendall.   Beyond the maze in Tuttle.

          This refers to a maze in Tothill Fields, close to Westminster Abbey.

Lewis Carroll.  Untitled maze.  In: Mischmasch, the last of his youthful MS magazines, with entries from 1855 to 1862.  Transcribed version in:  The Rectory Umbrella and Mischmasch; Cassell, 1932; Dover, 1971; p. 165 of the Dover ed.  John Fisher [The Magic of Lewis Carroll; (Nelson, 1973), Penguin, 1975, pp. 19-20] gives this and another example.  Cf Carroll-Wakeling, prob. 35: An amazing maze, pp. 46-47 & 75 and Carroll-Gardner, pp. 80-81 for the Mischmasch example.  I don't find the other example elsewhere, but it was for Georgina "Ina" Watson, so probably c1870.

Mittenzwey.  1880.  Prob. 281, pp. 50 & 100;  1895?: 310, pp. 53-54 & 102;  1917: 310, pp. 49 & 97.  The garden of a French place has a maze with 31 points to see.  Find a path past all of them with no repeated edges and no crossings.  The pattern is clearly based on the Versailles maze of c1675 mentioned in the Historical Sketch above, but I don't recall the additional feature of no crossings occurring before.

C. Wiener.  Ueber eine Aufgabe aus der Geometria situs.  Math. Annalen 6 (1873) 29‑30.  An algorithm for solving a maze.  BLW asserts this is very complicated, but it doesn't look too bad.

M. Trémaux.  Algorithm.  Described in Lucas, RM1, 1891, pp. 47‑51.  ??check 1882 ed.  BLW assert Lucas' description is faulty.  Also described in MRE,  1st ed., 1892, pp. 130‑131;  3rd ed., 1896, pp. 155-156;  4th ed., 1905, pp. 175-176 is vague;  5th-10th ed., 1911‑1922, 183;  11th ed., 1939, pp. 255‑256 (taken from Lucas);  (12th ed. describes Tarry's algorithm instead) and in Dudeney, AM, p. 135 (= Mazes, and how to thread them, Strand Mag. 37 (No. 220) (Apr 1909) 442‑448, esp. 446‑447).

G. Tarry.  Le problème des labyrinthes.  Nouv. Annales de Math. (3) 4 (1895) 187‑190.  ??NYR

Collins.  Book of Puzzles.  1927.  How to thread any maze, pp. 122-124.  Discusses right hand rule and its failure, then Trémaux's method.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 2.  5 x 5 x 5  cubical maze.  Get from a corner to an antipodal corner in a minimal number of steps.

Anneke Treep.  Mazes... How to get out!  (part I).  CFF 37 (Jun 1995) 18-21.  Based on her MSc thesis at Univ. of Twente.  Notes that there has been very little systematic study.  Surveys the algorithms of Tarry, Trémaux, Rosenstiehl.  Rosenstiehl is greedy on new edges, Trémaux is greedy on new nodes and Trémaux is a hybrid of these.  ??-oops-check.  Studies probabilities of various routes and the expected traversal time.  When the maze graph is a tree, the methods are equivalent and the expected traversal time is the number of edges.

Bernhard Wiezorke.  Puzzles und Brainteasers.  OR News, Ausgabe 13 (Nov 2001) 52-54.  This reports his discovery of a hedge maze in Germany -- the first he knew of.  It is in Altjessnitz, near Dessau in Sachsen-Anhalt.  (My atlas doesn't show such a place, but Jessnitz is about 10km south of Dessau.)  This maze dates from 1720 and has 12 components, with the goal completely separated from the outside so that the 'hand on wall' rule does not solve it.  Torsten Silke later told Wiezorke of two other hedge mazes in Germany.  One, in Probststeierhagen, Schleswig-Holstein, about 12km NE of Kiel, is in the grounds of the restaurant Zum Irrgarten (At the Labyrinth) and is an early 20C copy of the Altjessnitz example.  The other, in Kleinwelka, Sachsen, about 50km NE of Dresden, was made in 1992 and is private.  Though it has 17 components, the 'hand on wall' method will solve it.  He gives plans of both mazes.  He discusses the Seven Bridges of Königsberg, giving a B&W print of the 1641 plan of the city mentioned at the beginning of Section 5.E -- he has sent me a colour version of it.  He also describes Tremaux's solution method.

 

          5.E.2. MEMORY WHEELS  =  CHAIN CODES

 

          These are cycles of  2n  0s  and  1s  such that each  n‑tuple of  0s  and  1s  appears just once.  They are sometimes called De Bruijn sequences, but they have now been traced back to the late 19C.  An example for  n = 3  is  00010111.

 

Émile Baudot.  1884.  Used the code for  25  in telegraphy.  ??NYS -- mentioned by Stein.

A. de Rivière, proposer;  C. Flye Sainte-Marie, solver.  Question no. 58.  L'Intermédiare des Mathématiciens 1 (1894)  19-20 & 107-110.  ??NYS -- described in Ralston and Fredricksen (but he gives no. 48 at one point).  Deals with the general problem of a cycle of  kn  symbols such that every  n‑tuple of the  k  basic symbols occurs just once.  Gives the graphical method and shows that such cycles always exist and there are  k!g(n)/ kn  of them,  where g(n) = kn‑1.  This work was unknown to the following authors until about 1975.

N. G. de Bruijn.  A combinatorial problem.  Nederl. Akad. Wetensch. Proc. 49 (1946) 758‑764.  ??NYS -- described in Ralston and Fredricksen.  Gives the graphical method for finding examples and finds there are  2f(n) solutions,  where  f(n) = 2n-1 - n.

I. J. Good.  Normal recurring decimals.  J. London Math. Soc. 21 (1946) 167-169.  ??NYS -- described in Ralston and Fredricksen.  Shows there are solutions but doesn't get the number.

R. L. Goodstein.  Note 2590:  A permutation problem.  MG 40 (No. 331) (Feb 1956) 46‑47.  Obtains a kind of recurrence for consecutive  n‑tuples.

Sherman K. Stein.  Mathematics:  The Man‑made Universe.  Freeman, 1963.  Chap. 9: Memory wheels.  c= The mathematician as explorer, SA (May 1961) 149‑158.  Surveys the topic.  Cites the c1000 Sanskrit word:  yamátárájabhánasalagám  used as the mnemonic for  01110100(01)  giving all triples of short and long beats in Sanskrit poetry and music.  Describes the many reinventions, including Baudot (1882), ??NYS, and the work of Good (1946), ??NYS, and de Bruijn  (1946), ??NYS.  15 references.

R. L. Goodstein.  A generalized permutation problem.  MG 54 (No. 389) (Oct 1970) 266‑267.  Extends his 1956 note to find a cycle of  an  symbols such that the  n‑tuples are distinct.

Anthony Ralston.  De Bruijn sequences -- A model example of the interaction of discrete mathematics and computer science.  MM 55 (1982) 131‑143 & cover.  Deals with the general problem of cycles of  kn  symbols such that every  n‑tuple of the  k  basic symbols occurs just once.  Discusses the history and various proofs and algorithms which show that such cycles always exist.  27 references.

Harold Fredricksen.  A survey of full length nonlinear shift register cycle algorithms.  SIAM Review 24:2 (Apr 1982) 195-221.  Mostly about their properties and their generation, but includes a discussion of the door lock connection, a mention of using the  23  case as a switch for three lights, and gives a good history.  The door lock connection is that certain push button door locks will open when a four digit code is entered, but they open if the last four buttons pressed are the correct code, so using a chain code reduces the number of button pushes required by a burglar to  1/4  of the number required if he tries all four digit combinations.  58 references.

At G4G2, 1996, Persi Diaconis spoke about applications of the chain code in magic and mentioned uses in repeated measurement designs, random number generators, robot location, door locks, DNA comparison. 

                    They were first used in card tricks by Charles T. Jordan in 1910.  Diaconis' example had a deck of cards which were cut and then five consecutive cards were dealt to five people in a row.  He then said he would determine what cards they had, but first he needed some help so he asked those with red cards to step forward.  The position of the red cards gives the location of the five cards in a cycle of  32  (which was the size of the deck)!  Further, there are simple recurrences for the sequence so it is fairly easy to determine the location.  One can code the binary quintuples to give the suit and value of the first card and then use the succeeding quintuples for the succeeding cards.

                    Long versions of the chain code are printed on factory floors so that a robot can read it and locate itself.

In Jan 2000, I discussed the Sanskrit chain code with a Sanskrit scholar, Dominik Wujastyk, who said that there is no known Sanskrit source for it.  He has asked numerous pandits who did not know of it and he said there is is a forthcoming paper on it, but that it did not locate any Sanskrit source.

 

          5.E.2.a.        PANTACTIC SQUARES

 

          Haubrich's 1995-1996 surveys, op. cit. in 5.H.4, include this.

 

B. Astle.  Pantactic squares.  MG 49 (No. 368) (May 1965) 144‑152.  This is a two‑dimensional version of the memory wheel.  Take a  5 x 5  array of cells marked  0  or  1  (or Black or White).  There are  16  ways to take a  2 x 2  subarray from the  5 x 5  array.  If these give all  16  2 x 2  binary patterns, the array is called pantactic.  The author shows a number of properties and some types of such squares.

C. J. Bouwkamp, P. Janssen & A. Koene.  Note on pantactic squares.  MG 54 (No. 390) (Dec 1970) 348‑351.  They find 800 such squares, forming 50 classes of 16 forms.

[Surprisingly, neither paper considers a  4 x 4  array viewed toroidally, which is the natural generalization of the memory wheel.  Precisely two of the fifty classes, namely nos. 25 & 41, give such a solution and these are the same pattern on the torus.  One can also look at the  4 x 4  subarrays of a  131 x 131  or a  128 x 128  array, etc., as well as 3 and higher dimensional arrays.  I submitted the question of the existence and numbers of these as a problem for CM, but it was considered too technical.]

Ivan Moscovich.  US Patent 3,677,549 -- Board Game Apparatus.  Applied: 14 Jun 1971;  patented: 18 Jul 1972.  Front page, 1p diagrams, 2pp text.  Reproduced in Haubrich, About ..., 1996, op. cit. in 5.H.4.  2pp + 2pp diagrams.  This uses the  16  2 x 2  binary patterns as game pieces.  He allows the pieces to be rotated, scoring different values according to the orientation.  No mention of reversing pieces or of the use of the pieces as a puzzle.

John Humphries.  Review of Q-Bits.  G&P 54 (Nov 1976) 28.  This is Moscovich's game idea, produced by Orda.  Though he mentions changing the rules to having non-matching, there is no mention of two-sidedness.

Pieter van Delft & Jack Botermans.  Creative Puzzles of the World.  (As:  Puzzels uit de hele wereld; Spectrum Hobby, 1978);  Harry N. Abrams, NY, 1978.  The colormatch square, p. 165.  See Haubrich,1994, for description.

Jacques Haubrich.  Pantactic patterns and puzzles.  CFF 34 (Oct 1994) 19-21.  Notes the toroidal property just mentioned.  Says Bouwkamp had the idea of making the 16 basic squares in coloured card and using them as a MacMahon-type puzzle, with the pieces double-sided and such that when one side had MacMahon matching, the other side had non-matching.  There are two different bijections between matching patterns and non-matching patterns, so there are also 800 solutions in 50 classes for the non-matching problem.  Bouwkamp's puzzle appeared in van Delft & Botermans, though they did not know about and hence did not mention the double-sidedness.  [In an email of 22 Aug 2000, Haubrich says he believes Bouwkamp did tell van Delft and Botermans about this, but somehow it did not get into their book.]  The idea was copied by two manufacturers (Set Squares by Peter Pan Playthings and Regev Magnetics) who did not understand Bouwkamp's ideas -- i.e. they permitted pieces to rotate.  Describes Verbakel's puzzle of 5.H.2.

Jacques Haubrich.  Letter: Pantactic Puzzles = Q-Bits.  CFF 37 (Jun 1995) 4.  Says that Ivan Moscovich has responded that he invented the version called "Q-Bits" in 1960-1964, having the same tiles as Bouwkamp's (but only one-sided [clarified by Haubrich in above mentioned email]).  His US Patent 3,677,549 (see above) is for a game version of he idea.  The version produced by Orda Ltd. was reviewed in G&P 54 (Nov 1976) (above).  So it seems clear that Moscovich had the idea of the pieces before Bouwkamp's version was published, but Moscovich's application was to use them in a game where the orientations could be varied.

 

          5.F.    HAMILTONIAN CIRCUITS

 

          For queen's, bishop's and rook's tours, see 6.AK.

          A tour is a closed path or circuit. 

          A path has end points and is sometimes called an open tour.

 

          5.F.1. KNIGHT'S TOURS AND PATHS

 

                    GENERAL REFERENCES

 

Antonius van der Linde.  Geschichte und Literatur des Schachspiels.  (2 vols., Springer, Berlin, 1874);  one vol. reprint, Olms, Zürich, 1981.  [There are two other van der Linde books:  Quellenstudien zur Geschichte des Schachspiels, Berlin, 1881, ??NYS;  and  Das Erste Jartausend [sic] der Schachlitteratur  (850‑1880), (Berlin, 1880); reprinted with some notes and corrections, Caissa Limited Editions, Delaware, 1979, which is basically a bibliography of little use here.]

Baron Tassilo von Heydebrand und von der Lasa.  Zur Geschichte und Literatur des Schachspiels.  Forschungen.  Leipzig, 1897.  ??NYS.

Ahrens.  MUS I.  1910.  Pp. 319-398.

Harold James Ruthven Murray.  A History of Chess.  OUP, 1913;  reprinted by Benjamin Press, Northampton, Massachusetts, nd [c1986].  This has many references to the problem, which are detailed below.

Reinhard Wieber.  Das Schachspiel in der arabischen Literatur von den Anfängen bis zur zweiten Hälfte des 16.Jahrhunderts.  Verlag für Orientkunde Dr. H. Vorndran, Walldorf‑Hessen, 1972.

George P. Jelliss. 

Special Issue:  Notes on the Knight's Tour.  Chessics 22 (Summer 1985) 61‑72.

Further notes on the knight's tour.  Chessics 25 (Spring 1986) 106‑107.

Notes on Chessics 22 continued.  Chessics 29 & 30 (1987) 160.

           This is a progress report on his forthcoming book on the knight's tour.  I will record some of his comments at the appropriate points below.  He also studies the  3 x n  board extensively.

 

          Two problems with knights on a  3 x 3  board are generally treated here, but cf 5.R.6.

          The 4 knights problem has two W and two B knights at the corners (same colours at adjacent corners) and the problem is to exchange them in 16 moves.  The graph of knight's connections is an 8-cycle with the pieces at alternate nodes.  [Putting same colours at opposite corners allows a solution in 8 moves.]

          The 7 knights problem is to place 7 knights on a  3 x 3  board in the 4 corners and 3 of the sides so each is a knight's move from the previously placed one.  This is equivalent to the octagram puzzle of 5.R.6.

          4 knights problem -- see:  at‑Tilimsâni, 1446;  Civis Bononiae, c1475; 

          7 knights problem -- see:  King's Library MS.13, A.xviii, c1275;  "Bonus Socius", c1275;  at‑Tilimsâni, 1446; 

 

 

Al‑Adli (c840) and as‑Suli (c880‑946) are the first two great Arabic chess players.  Although none of their works survive, they are referred to by many later writers who claim to have used their material.

 

 

Rudraţa: Kāvyālaʼnkāra [NOTE:  ţ  denotes a  t  with an underdot and  ʼn  denotes an  n  with an overdot.].  c900.  ??NYS -- described in Murray 53‑55, from an 1896 paper by Jacobi, ??NYS.  The poet speaks of verses which have the shapes of "wheel, sword, club, bow, spear, trident, and plough, which are to be read according to the chessboard squares of the chariot [= rook], horse [= knight], elephant [c= bishop], &c."  According to Jacobi, the poet placed syllables in the cells of a half chessboard so that it reads the same straight across as when following a piece's path.  With help from the commentator Nami, of 1069, the rook's and knight's path's are reconstructed, and are given on Murray 54.  Both are readily extended to full board paths, but not tours.  The elephant's path is confused.

Kitâb ash‑shatranj mimma’l‑lafahu’l‑‘Adli waş‑Şûlî wa ghair‑huma [Book of the Chess; extracts from the works of al‑'Adlî, aş‑Şûlî and others].  [NOTE:  ş,  Ş  denote  s,  S  with underdot.]  Copied by Abû Ishâq [the  h  should have an underdot] Ibrâhîm ibn al‑Mubârak ibn ‘Alî al‑Mudhahhab al‑Baghdâdî.  Murray 171‑172 says it is MS ‘Abd‑al‑Hamid [the  H  should have an underdot] I, no. 560, of 1140, and denotes it AH.  Wieber 12‑15 says it is now MS Lala Ismail Efendi 560, dates it July‑August 1141, and denotes it L.  Both cite van der Linde, Quellenstudien, no. xviii, p. 331+, ??NYS.  The author is unknown.  This MS was discovered in 1880.  Catalogues in Istanbul listed it as Risâla fi’sh‑shaţranj by Abû’l‑‘Abbâs Ahmad [the  h  should have an underdot] al‑‘Adlî.  It is sometimes attributed to al‑Lajlâj who wrote one short section of this book.  Murray, van der Linde and Wieber (p. 41) cite another version:  MS Khedivial Lib., Cairo, Mustafa Pasha, no. 8201, copied c1370, which Murray denotes as C and Wieber lists as unseen.

                    Murray 336 gives two distinct tours: AH91 & AH92.  The solution of AH91 is a numbered diagram, but AH92 is 'solved' four times by acrostic poems, where the initial letters of the lines give the tour in an algebraic notation.  Wieber 479‑480 gives 2 tours from ff. 74a‑75b: L74a = AH91 and L74b = reflection of AH92.  [Since the 'solutions' of AH92 are poetic, it is not unreasonable to consider the reflection as different.]  Also AH94 = L75b is a knight/bishop tour, where moves of the two types alternate.  These tours may be due to as‑Suli.  AH196 is a knight/queen tour.

Arabic MS Atif Efendi 2234 (formerly Vefa (‘Atîq Efendî) 2234), Eyyub, Istanbul.  Copied by Muhammad [the  h  should have an underdot] ibn Hawâ (or Rahwâr -- the MS is obscure) ibn ‘Othmân al‑Mu’addib in 1221.  Murray 174‑175 describes it as mostly taken from the above book and denotes it V.  A tour is shown on p. 336 as V93 = AH92.  Wieber 20‑24 denotes it A.  On p. 479, he shows the tour from f. 68b which is the same as L74b, the reflection of AH92.

King's Library MS.13, A.xviii, British Museum, in French,  c1275.  Described in van der Linde I 305‑306.  Described and transcribed in Murray 579‑582 & 588‑600, where it is denoted as K.  Van der Linde discusses the knight's path on I 295, with diagram no. 244 on p. 245.  Murray 589 gives the text and a numbered diagram of a knight's path as K1.  The path splits into two half board paths:  a1 to d1 and e3 to h1, so the first half and the whole are corner to corner.  The first half is also shown as diagram K2 with the half board covered with pieces and the path described by taking of pieces.  K3 is the 7 knights problem

"Bonus Socius" [perhaps Nicolas de Nicolaï].  This is the common name of a collection of chess problems, assembled c1275, which was copied and translated many times.  See Murray 618‑642 for about 11 MSS.  Some of these are given below.  Fiske 104 & 110‑111 discusses some MSS of this collection.

                    MS Lat. 10286, Nat. Lib., Paris.  c1350.  Van der Linde I 293‑295 describes this but gives the number as 10287 (formerly 7390).  Murray 621 describes it and denotes it PL.  Van der Linde describes a half board knight's path, with a diagram no. 243 shown on p. 245.  The description indicates a gap in the path which can only be filled in one way.  This is a path from a8 to h8 which cannot be extended to the full board.  Murray 641 says that PL275 is the same as problems in two similar MSS and as CB244, diagrammed on p. 674.  However, this is not the same as van der Linde's no. 243, though cells 1‑19 and 31‑32 are the same in both paths, so this is also an a8 to h8 path which does not extend to a full board.

                    Murray 620 mentions a path in a late Italian MS version of c1530 (Florence, Nat. Lib. XIX.7.51, which he denotes It) which may be the MS described by van der Linde I 284 as no. 4 and the half board path described on I 295 with diagram no. 245 on I 245.  Fiske 210-211 describes this and says von der Lasa 163-165 (??NYS) describes it as early 16C, but Murray does not mention von der Lasa.  Fiske says it contains a tour on f. 28b, which von der Lasa claims is "das älteste beispiel eines vollkommenen rösselsprunges", but Murray does not detail the problems so I cannot compare these citations.  Fiske also says it also contains the 7 knights problem.

Dresden MS 0/59, in French, c1400.  Murray describes this on pp. 607‑613 and denotes it D.  On p. 609, Murray describes D57 which asks for a knight's path on a  4 x 4  board.  No solution is given -- indeed this is impossible, cf Persian MS 211 in the RAS.  Ibid. is D62 which asks for a half board tour, but no answer is provided.

Persian MS 211 in Royal Asiatic Society.  Early 15C.  ??NYS.

                    Extensively described as MS 250 bequeathed by Major David Price in:  N. Bland; On the Persian game of chess; J. Royal Asiatic Soc. 13 (1852) 1‑70.  He dates it as 'at least 500 years old' and doesn't mention the knight's tour.

                    Described, as MS No. 260, and partially translated in Duncan Forbes; The History of Chess; Wm. H. Allen, London, 1860.  Forbes says Bland's description is "very detailed but unsatisfactory".  On p. 82 is the end of the translation of the preface:  '"Finally I will show you how to move a Knight from any individual square on the board, so that he may cover each of the remaining squares in as many moves and finally come to rest on that square whence he started.  I will also show how the same thing may be done by limiting yourself only to one half, or even to one quarter (1) of the board." -- Here the preface abruptly terminates, the following leaf being lost.'  Forbes's footnote (1) correctly doubts that a knight's tour (or even a knight's path) is possible on the  4 x 4  board.

                    Murray 177 cites it as MS no. 211 and denotes it RAS.  He says that it has been suggested that this MS may be the work of ‘Alâ'addîn Tabrîzî = ‘Alî ash‑Shatranjî, late 14C, described on Murray 171.  Murray mentions the knight's tour passage on p. 335.  This may be in van der Linde, ??NX.  Wieber 45 mentions the MS.

Abû Zakarîyâ Yahya [the  h  should have an underdot] ibn Ibrâhîm al‑Hakîm[the  H  should have an underdot].  Nuzhat al‑arbâb al‑‘aqûl fî’sh‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot] al‑manqûl (The delight of the intelligent, a description of chess).  Arabic MS 766, John Rylands Library, Manchester.

                    Bland, loc. cit., pp. 27‑28, describes this as no. 146 of Dr. Lee's catalogue and no. 76 of the new catalogue.  Forbes, loc. cit., says that Dr. Lee had loaned his two MSS to someone who had not yet returned them, so Forbes copies Bland's descriptions (on pp. 27‑31) as his Appendix C, with some clarifying notes.  (The other of Dr. Lee's MSS is described below.)  Van der Linde I 107ff (??NX) seems to copy Bland & Forbes.

                    Murray 175‑176 describes it as Arab. 59 at John Rylands Library and denotes it H.  He says it was Bland who had borrowed the MSS from Dr. Lee and Murray traces their route to Dr. Lee and to Manchester.  Murray says it is late 15C, is based on al‑Adli and as‑Suli and he also describes a later version, denoted Z, late 18C.  Wieber 32‑35 cites it as MS 766(86) at John Rylands, dates it 1430 and denotes it Y1.

                    Murray 336 gives three paths.  H73 = H75 are the same tour, but with different keys, one poetic as in Rudraţa [NOTE:  ţ  denotes a  t  with an underdot.], one numeric.  H74 is a path attributed to Ali Mani with similar poetic solution.  Wieber 480 shows two diagrams.  Y1‑39a, Y1‑39b, Y1‑41b are the same tour as H73, but with different descriptions, the latter two being attributed to al‑Adli.  Y1‑39a (second diagram) = H74 is attributed to ‘Ali ibn Mani.

Shihâbaddîn Abû’l‑‘Abbâs Ahmad [the  h  should have an underdot] ibn Yahya [the  h  should have an underdot] ibn Abî Hajala [the  H  should have an underdot] at‑Tilimsâni alH‑anbalî [the  H  should have an underdot].  Kitâb ’anmûdhaj al‑qitâl fi la‘b ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot] (Book of the examples of warfare in the game of chess).  Copied by Muhammed ibn ‘Ali ibn Muhammed al‑Arzagî in 1446.

                    Bland, loc. cit., pp. 28‑31, describes this as the second of Dr. Lee's MSS, old no. 147, new no. 77.  Forbes copies this and adds notes.  Van der Linde I 105‑107 seems to copy from Bland and Forbes.  Murray 176‑177 says the author died in 1375, so this might be c1370.  He says it is Dr. Lee's on 175‑176, that it is MS Arab. 93 at the John Rylands Library and denotes it Man.  Wieber 29‑32 cites it as MS 767(59) at the Rylands Library and denotes it H.  On p. 481, he shows a half‑board path which cannot be extended to the full board.

                    This MS also gives the 4 knights and 7 knights problems.  Murray 337, 673 (CB236) & 690 and Wieber 481 show these problems.

Risâlahĭ Shatranj.  Persian poem of unknown date and authorship.  A copy was sent to Bland by Dr. Sprenger of Delhi.  See Bland, loc. cit., pp. 43‑44.  [Bland uses á for â.]  Bland says it has the problem of the knight's tour or path.  [I think this is the poem mentioned on Murray 182-183 and hence on Wieber 42.]

Şifat mal ‘ûb al‑faras fî gamî abyât aš‑šaţranğ [NOTE:  Ş,  ţ.  denote  S,  t  with underdot.] MS Gotha 10, Teil 6; ar. 366; Stz. Hal. 408.  Date unknown.  Wieber 37 & 480 describes this and gives a path from h8 to e4 which occurs on ff. 70 & 68.

Civis Bononiae [Citizen of Bologna].  Like Bonus Socius, this is a collection of chess problems, from c1475, which exists in several MSS and printings.  All are in Latin, from Italy, and give essentially the same 288 problems.  See Murray 643‑703 for description of about 10 texts and transcription of the problems.  Many of the texts are not in van der Linde.  Murray 643 cites MS Lasa, in the library of Baron von der Lasa, c1475, as the most accurate and complete of the texts.  Two other well known versions are described below.

                    Paulo Guarino (di Forli) (= Paulus Guarinus).  No real title, but the end has 'Explicit liber de partitis scacorum' with the writer's name and the date 4 Jan 1512.  This MS was in the Franz Collection and is now (1913) in the John G. White Collection in Cleveland, Ohio.  This version only contains 76 problems.  Van der Linde I 295‑297 describes the MS and on p. 294 he describes a half board path and says Guarino's 74 is a reflection of his no. 243.  Murray 645 describes the MS but doesn't list the individual problems.  He implies that CB244, on p. 674, is the tour that appears in all of the Civis Bononiae texts, but this is not the same as van der Linde's no. 243.  CB236, pp. 673 & 690, is the 4 knights problem, which is Guarino's 42 [according to Lucas, RM4, p. 207], but I don't have a copy of van der Linde's no. 215 to check this, ??NX.

                    Anon. Sensuit Jeux Partis des eschez: composez nouvellement Pour recreer tous nobles cueurs et pour eviter oysivete a ceulx qui ont voulente: desir et affection de le scavoir et apprendre et est appelle ce Livre le jeu des princes et damoiselles.  Published by Denis Janot, Paris, c1535, 12 ff.  ??NYS.  (This is the item described by von der Lasa as 'bei Janot gedrucktes Quartbändchen' (MUS #195).)  This a late text of 21 problems, mostly taken from Civis Bononiae.  Only one copy is known, now (1913) in Vienna.  See van der Linde I 306‑307 and Murray 707‑708 which identify no. 18 as van der Linde's no. 243 and with CB244, as with the Guarino work.  I can't tell but van der Linde may identify no. 11 as the 4 knights problem (??NX).

                    Murray 730 gives another half board path, C92, of c1500 which goes from a8 to g5.  Murray 732 notes that a small rearrangement makes it extendable to the whole board.

Horatio Gianutio della Mantia.  Libro nel quale si tratta della Maniera di giuocar' à Scacchi, Con alcuni sottilissimi Partiti.  Antonio de' Bianchi, Torino, 1597.  ??NYS.  Gives half board tours which can be assembled into to a full tour.  (Not in the English translation:   The Works of Gianutio and Gustavus Selenus, on the game of Chess, Translated and arranged by J. H. Sarratt; J. Ebers, London, 1817, vol. 1. -- though the copy I saw didn't say vol. 1.  Van der Linde, Erste Jartausend ... says there are two volumes.)

Bhaţţa Nīlakaņţha.  [NOTE:  ţ,  ņ  denote  t,  n  with underdot.]  Bhagavantabhāskara.  17C.  End of 5th book.  ??NYS, described by Murray 63‑66.  The author gives three tours, in the poetic form of Rudraţa [NOTE:  ţ  denotes a  t  with an underdot.], which are the same tour starting at different points.  The tour has 180 degree rotational symmetry.

Ozanam.  1725.  Prob. 52, 1725: 260‑269.  Gives solutions due to Pierre Rémond de Montmort, Abraham de Moivre, Jean‑Jacques d'Ortous de Mairan (1678-1771).  Surprisingly, these are all distinct and different from the earlier examples.  Ozanam says he had the problem and the solution from de Mairan in 1722.  Says the de Moivre is the simplest.  Kraitchik, Math. des Jeux, op. cit. in 4.A.2, p. 359, dates the de Montmort as 1708 and the de Moivre as 1722, but gives no source for these.  Montmort died in 1719.  Ozanam died in 1717 and this edition was edited by Grandin.  Van der Linde and Ahrens say they can find no trace of these solutions prior to Ozanam (1725).  See Ozanam-Montucla, 1778.

                    Ball, MRE, 1st ed., 1892, p. 139, says the earliest examples he knows are the De Montmort & De Moivre of the late 17C, but he only cites them from Ozanam-Hutton, 1803,  &  Ozanam-Riddle, 1840.  In the 5th ed., 1911, p. 123, he adds that "They were sent by their authors to Brook Taylor who seems to have previously suggested the problem."  He gives no reference for the connection to Taylor and I have not seen it mentioned elsewhere.  This note is never changed and may be the source of the common misconception that knight's tours originated c1700!

Les Amusemens.  1749.  Prob. 181, p. 354.  Gives de Moivre's tour.  Says one can imagine other methods, but this is the simplest and most interesting.

L. Euler.  Letter to C. Goldbach, 26 Apr 1757.  In:  P.‑H. Fuss, ed.; Correspondance Mathématique et Physique de Quelques Célèbres Géomètres du XVIIIème Siècle; (Acad. Imp. des Sciences, St. Pétersbourg, 1843)  = Johnson Reprint, NY, 1968, vol. 1, pp. 654‑655.  Gives a 180o symmetric tour.

L. Euler.  Solution d'une question curieuse qui ne paroit soumise à aucune analyse.  (Mém. de l'Académie des Sciences de Berlin, 15 (1759 (1766)), 310‑337.)  = Opera Omnia (1) 7 (1923) 26‑56.  (= Comm. Arithm. Coll., 1849, vol. 1, pp. 337‑355.)  Produces many solutions;  studies 180o symmetry, two halves, and other size boards.

[Petronio dalla Volpe].  Corsa del Cavallo per tutt'i scacchi dello scacchiere.  Lelio della Volpe, Bologna, 1766.  12pp, of which 2 and 12 are blanks.  [Lelio della Volpe is sometimes given as the author, but he died c1749 and was succeeded by his son Petronio.]  Photographed and printed by Dario Uri from the example in the Libreria Comunale Archiginnasio di Bologna, no. 17 CAPS XVI 13.  The booklet is briefly described in:  Adriano Chicco; Note bibliografiche su gli studi di matematica applicata agli scacchi, publicati in Italia; Atti del Convegno Nazionale sui Giochi Creative, Siena, 11‑14 Jun 1981, ed. by Roberto Magari; Tipografia Senese for GIOCREA (Società Italiana Giochi Creativi), 1981; p. 155. 

                    The Introduction by the publisher cites Ozanam as the originator of this 'most ingenious' idea and says he gives examples due to Montmort, Moivre and Mairan.  He also says this material has 'come to hand' but doesn't give any source, so it is generally thought he was the author.  He gives ten paths, starting from each of the 10 essentially distinct cells.  He then gives the three cited paths from Ozanam.  He then gives six tours.  Each path is given as a numbered board and a line diagram of the path, which led Chicco to say there were 38 paths.  The line drawing of the first tour is also reproduced on the cover/title page.

Ozanam-Montucla.  1778.  Prob. 23, 1778: 178-182;  1803: 177-180;  1814: 155-157.  Prob. 22, 1840: 80‑81.  Drops the reference to de Mairan as the source of the problem and adds a fourth tour due to "M. de W***, capitaine au régiment de Kinski".  All of these have a misprint of 22 for 42 in the right hand column of De Moivre's solution.

H. C. von Warnsdorff.  Des Rösselsprunges einfachste und allgemeinste Lösung.  Th. G. Fr. Varnhagenschen Buchhandlung, Schmalkalden, 1823, 68pp.  ??NYS -- details from Walker.  Rule to make the next move to the cell with the fewest remaining neighbours.  Lucas, L'Arithmétique Amusante, p. 241, gives the place of publication as Berlin.

Boy's Own Book.  Not in 1828.  1828-2: 318 states a knight's tour can be made.

George Walker.  The Art of Chess-Play: A New Treatise on the Game of Chess.  (1832, 80pp.  2nd ed., Sherwood & Co, London, 1833, 160pp.  3rd ed., Sherwood & Co., London, 1841, 300pp.  All ??NYS -- details from 4th ed.)  4th ed., Sherwood, Gilbert & Piper, London, 1846, 375pp.  Chap. V -- section: On the knight, p. 37.  "The problem respecting the Knight's covering each square of the board consecutively, has attracted, in all ages, the attention of the first mathematicians."  States Warnsdorff's rule, without credit, but gives the book in his bibliography on p. 375, and asserts the rule will always give a tour.  No diagram.

Family Friend 2 (1850) 88 & 119, with note on 209.  Practical Puzzle -- No. III.  Find a knight's path.  Gives one answer.  Note says it has been studied since 'an early period' and cites Hutton, who copies some from Montucla, an article by Walker in Frasers Magazine (??NYS) which gives Warnsdorff's rule and an article by Roget in Philosophical Magazine (??NYS) which shows one can start and end on any two squares of opposite colours.  Describes using a pegged board and a string to make pretty patterns.

Boy's Own Book.  Moving the knight over all the squares alternately.  1855: 511-512;  1868: 573;  1881 (NY): 346-347.  1855 says the problem interested Euler, Ozanam, De Montmart [sic], De Moivre, De Majron [sic] and then gives Warnsdorff's rule, citing George Walker's 'Treatise on Chess' for it -- presumably 'A New Treatise', London, 1832, with 2nd ed., 1833  &  3rd ed., 1841, ??NYS.  Walker also wrote On Moving the Knight, London, 1840, ??NYS.  1868 drops all the names, but the NY ed. of 1881 is the same as the 1855.  Gives a circuit due to Euler.

Magician's Own Book.  1857.  Art. 46: Moving the knight over all the squares alternately, pp. 283-287.  Identical to Boy's Own Book, 1855, but adds Another Method.  = Book of 500 Puzzles; 1859, art. 46, pp. 97-101.  = Boy's Own Conjuring Book, 1860, prob. 45, pp. 246‑251. 

Landells.  Boy's Own Toy-Maker.  1858.  Moving the knight over all the squares alternately, p. 143.  This is the Another Method of Magician's Own Book, 1857.  Cf Illustrated Boy's Own Treasury, 1860.

Illustrated Boy's Own Treasury.  1860.  Prob. 47: Practical chess puzzle, pp. 404 & 443.  Knight's tour.  This is the Another Method of Magician's Own Book.

C. F. de Jaenisch.  Traité des Applications de l'Analyse Mathématiques au Jeu des Échecs.  3 vols., no publisher, Saint-Pétersbourg. 1862-1863.  Vol. 1: Livre I: Section III: De la marche du cavalier, pp. 186-259 & Plate III.  Vol. 2: Livre II: Problème du Cavalier, pp. 1-296 & 31 plates (some parts ??NYS).  Vol. 3: Addition au Livre II, pp. 239-243 (This Addition ??NYS).  This contains a vast amount of miscellaneous material and I have not yet read it carefully.  ??NYR

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 323, pp. 153-154 & 393: Rösselsprung-Aufgaben.  Three arrays of syllables and one must find a poetic riddle by following a knight's tour.  Arrays are  8 x 8,  8 x 8,  6 x 4.

C. Flye Sainte‑Marie.  Bull. Soc. Math. de France (1876) 144‑150.  ??NYS -- described by Jelliss.  Shows there is no tour on a  4 x n  board and describes what a path must look like.

Mittenzwey.  1880.  Prob. 222-223, pp. 40 & 91;  1895?: 247-248, pp. 44 & 93;  1917: 247‑248, pp. 40-41 & 89.  First is a knight's path.  Second is a board with word fragments and one has to make a poem, which uses the same path as in the first problem.

Paul de Hijo [= Abbé Jolivald].  Le Problème du Cavalier des Échecs.  Metz, 1882.  ??NYS -- described by Jelliss and quoted by Lucas.  Jelliss notes the BL copy of de Hijo was destroyed in the war, but he has since told me there are copies in The Hague and Nijmegen.  First determination of the five  6 x 6  tours with 4-fold rotational symmetry, the 150 ways to cover the  8 x 8  with two circuits of length 32 giving a pattern with 2‑fold rotational symmetry, the 378 ways giving reflectional symmetry in a median, the 140 ways with four circuits giving 4-fold rotational symmetry and the 301 ways giving symmetry in both medians (quoted in Lucas, L'Arithmétique Amusante, pp. 238-241).

Lucas.  Nouveaux jeux scientifiques ..., 1889, op. cit. in 4.B.3.  (Described on p. 302, figure on p. 301.)  'La Fasioulette' is an  8 x 8  board with 64 links of length  Ö5  to form knight's tours.

Knight's move puzzles.  The Boy's Own Paper 11 (Nos. 557  &  558) (14  &  21 Sep 1889) 799  &  814.  Four Shakespearean quotations concealed as knight's tours on a  8 x 8  board.  Beginnings not indicated!

Hoffmann.  1893.  Chap. X, no. 6: The knight's tour, pp. 335-336 & 367-373 = Hoffmann‑Hordern, pp. 225-229.  Gives knight's paths due to Euler and Du Malabare, a knight's tour due to Monneron, and four other unattributed tours.  Gives Warnsdorff's rule, citing Walker's A New Treatise on Chess, 1832.

Ahrens.  Mathematische Spiele.  Encyklopadie article, op. cit. in 3.B.  1904, pp. 1080‑1093.  Pp. 1084‑1086 gives many references to 19C work, including estimates of the number of tours and results on 'semi‑magic tours'.

C. Planck.  Chess Amateur (Dec 1908) 83.  ??NYS -- described by Jelliss. Shows there are 1728 paths on the  5 x 5  board.  Jelliss notes that this counts each path in both directions and there are only 112 inequivalent tours.

Ahrens.  1910.  MUS I 325.  Use of knight's tours as a secret code.

Dudeney.  AM.  1917.  Prob. 339: The four knight's tours, pp. 103 & 229.  Quadrisect the board into four congruent pieces such that there is a knight's tour on the piece.  Jelliss asserts that the solution is unique and says this may be what Persian MS 260 (i.e. 211) intended.  He notes that the four tours can be joined to give a tour with four fold rotational symmetry.

W. H. Cozens.  Cyclically symmetric knight's tours.  MG 24 (No. 262) (Dec 1940) 315‑323.  Finds symmetric tours on various odd‑shaped boards.

H. J. R. Murray.  The Knight's Tour.  ??NYS.  MS of 1942 described by G. P. Jelliss, G&PJ 2 (No. 17) (Oct 1999) 315.  Observes that a knight can move from the  (0, 0)  cell to the  (2, 1)  and  (1, 2)  cella and that the angle between these lines is the smaller angle of a  3, 4, 5  triangle.  One can see this by extending the lines to  (8, 4)  and  (5, 10)  and seeing these points form a  3, 4, 5  triangle with  (0, 0).

W. H. Cozens.  Note 2761:  On note 2592.  MG 42 (No. 340) (May 1958) 124‑125.  Note 2592 tried to find the cyclically symmetric tours on the  6 x 6  board and found 4.  Cozens notes two are reflections of the other two and that three such tours were omitted.  He found all these in his 1940 paper.

R. C. Read.  Constructing open knight's tours blindfold!  Eureka 22 (Oct 1959) 5-9.  Describes how to construct easily a tour between given cells of opposite colours, correcting a method of Roget described by Ball (MRE 11th ed, p. 181).  Says he can do it blindfold.

W. H. Cozens.  Note 2884:  On note 2592.  MG 44 (No. 348) (May 1960) 117.  Estimates there are 200,000 cyclically symmetric tours on the  10 x 10  board.

Roger F. Wheeler.  Note 3059:  The KNIGHT's tour on  42  and other boards.  MG 47 (No. 360) (May 1963) 136‑141.  KNIGHT means a knight on a toroidal board.  He finds 2688 tours of 19 types on the  42  toroid.  (Cf Tylor, 1982??)

J. J. Duby.  Un algorithme graphique trouvant tous les circuits Hamiltoniens d'un graphe.  Etude No. 8, IBM France, Paris, 22 Oct 1964.  [In English with French title and summary.]  Finds there are 9862 knight's tours on the  6 x 6  board, where the tours all start at a fixed corner and then go to a fixed one of the two cells reachable from the corner.  He also finds 75,000 tours on the  8 x 8  board which have the same first 35 moves.  He believes there may be over a million tours.

Karl Fabel.  Wanderungen von Schachfiguren.  IN:  Eero Bonsdorff, Karl Fabel & Olavi Riihimaa; Schach und Zahl; Walter Rau Verlag, Düsseldorf, 1966, pp. 40-50.  On p. 50, he says that there are  122,802,512  tours where the knight does two joined half-board paths.  He also says there are upper bounds, determined by several authors, and he gives  1.5 x 1026  as an example.

Gardner.  SA (Oct 1967) = Magic Show, chap. 14.  Surveys results of which boards have tours or paths.

D. J. W. Stone.  On the Knight's Tour Problem and Its Solution by Graph‑Theoretic and Other Methods.  M.Sc. Thesis, Dept. of Computing Science, Univ. of Glasgow, Jan. 1969.  Confirms Duby's 9862 tours on the  6 x 6  board.

David Singmaster.  Enumerating unlabelled Hamiltonian circuits.  International Series on Numerical Mathematics, No. 29.  Birkhäuser, Basel, 1975, pp. 117‑130.  Discusses the work of Duby and Stone and gives an estimate, which Stone endorses, that there are  1023±3  tours on the  8 x 8  board.

C. M. B. Tylor.  2‑by‑2 tours.  Chessics 14 (Jul‑Dec 1982) 14.  Says there are 17 knight's tours on a  2 x 2  torus and gives them.  Doesn't mention Wheeler, 1963.

Robert Cannon & Stan Dolan.  The knight's tour.  MG 70 (No. 452) (Jun 1986) 91‑100.  A rectangular board is tourable if it has a knight's path between any two cells of opposite colours.  They prove that  m x n  is tourable if and only if  mn is even  and  m ³ 6, n ³ 6.  They also prove that  m x n  has a knight's tour if and only if  mn is even  and [(m ³ 5, n ³ 5) or (m = 3, n ³ 10)]  and that when mn is even,  m x n  has a knight's path if and only if  m ³ 3, n ³ 3,  except for the  3 x 6  and  4 x 4  boards.  (These later results are well known -- see Gardner.  The authors only cite Ball's MRE.)

George Jelliss.  Figured tours.  MS 25:1 (1992/93) 16-20.  Exposition of paths and tours where certain stages of the path form an interesting geometric figure.  E.g. Euler's first paper has a path on the  5 x 5  such that the points on one diagonal are in arithmetic progression:  1, 7, 13, 19, 25.

Martin Loebbing & Ingo Wegener.  The number of knight's tours equals  33,439,123,484,294 ‑-- Counting with binary decision diagrams.  Electronic Journal of Combinatorics 3 (1996) article R5.  A somewhat vague description of a method for counting knight's tours -- they speak of directed knight's tours, but it is not clear if they have properly accounted for the symmetries of a tour or of the board.  Several people immediately pointed out that the number is incorrect because it has to be divisible by four.  Two comments have appeared, ibid.  On 15 May 1996, the authors admitted this and said they would redo the problem, but they have submitted no further comment as of Jan 2001.  On 18 Feb 1997, Brendan McKay announced that he had done the computation another way and found  13,267,364,410,532. 

                    In view of the difference between this and my 1975 estimate of  1023±3  tours, it might be worth explaining my reasoning.  In 1964, Duby found  75,000  tours with the same first 35 moves.  The average valence for a knight on an  8 x 8  board is 5.25, but one cannot exit from a cell in the same direction as one entered, so we might estimate the number of ways that the first 35 moves can be made as  4.2535 = 9.9 x 1021.  Multiplying by  75,000  then gives  7.4 x 1026.  I think I assumed that some of the first moves had already been made, e.g. we only allow one move from the starting cell, and factored by 8 for the symmetries of the square, to get  2.2 x 1025.  I can't find my original calculations, and I find the estimate  1025  in later papers, so I suppose I tried to reduce the effect of the  4.2535  some more.  In retrospect, I had no knowledge of how many of these had already been tried.  If about half of all moves from a cell had already been tried before any circuit was found, then the estimate would be more like  2.2534 x 75,000  =  7.1 x 1016.  If we divide the given number of circuits by  75,000  and take the 34th root, we get an average valence of 1.78 remaining, far less than I would have guessed.

                    I am grateful to Don Knuth for this reference.  Neither he nor I expected to ever see this number calculated!

 

          5.F.2. OTHER HAMILTONIAN CIRCUITS

 

          For circuits on the n‑cube, see also 5.F.4 and 7.M.1,2,3.

          For circuits on the chessboard, see also 6.AK.

 

Le Nôtre.  Le Labyrinte de Versailles, c1675.  This was a hedge or garden maze, but the objective was to visit, in correct order, 40 fountains based on Aesop's Fables.  Each node of the maze had at least one fountain.  Some fountains were not at path junctions, but one can consider these as nodes of degree two.  This is an early example of a Hamiltonian problem, except that one fountain was located at the end of a short dead end.  [Fisher, op. cit. in 5.E.1, pp. 49, 79, 130 & 144-145, with contemporary diagram on p. 144.  He says there are 39 fountains, but the diagram has 40.] 

 

T. P. Kirkman.  On the partitions of the R‑pyramid, being the first class of R‑gonous X‑edra.  Philos. Trans. Roy. Soc. 148 (1858) 145‑161.

W. R. Hamilton.  The Icosian Game.  4pp instructions for the board game.  J. Jaques and Son, London, 1859.  (Reproduced in BLW, pp. 32-35, with frontispiece photo of the board at the Royal Irish Academy.)

For a long time, the only known example of the game, produced by Jaques, was at the Royal Irish Academy in Dublin.  This example is inscribed on the back as a present from Hamilton to his friend, J. T. Graves.  It is complete, with pegs and instructions.  None of the obvious museums have an example.  Diligent searching in the antique trade failed to turn up an example in twenty years, but in Feb 1996, James Dalgety found and acquired an example of the board -- sadly the pegs and instructions were lacking.  Dalgety obtained another board in 1998, again without the pegs and instructions, but in 1999 he obtained another example, with the pegs.

Mittenzwey.  1880.  Prob. 281, pp. 50 & 100;  1895?: 310, pp. 53-54 & 102;  1917: 310, pp. 49 & 97.  The garden of a French palace has a maze with 31 points to see.  Find a path past all of them with no repeated edges and no crossings.  The pattern is clearly based on the Versailles maze of c1675 mentioned above, but I don't recall the additional feature of no crossings occurring before.

T. P. Kirkman.  Solution of problem 6610, proposed by himself in verse.  Math. Quest. Educ. Times 35 (1881) 112‑116.  On p. 115, he says Hamilton told him, upon occasion of Hamilton presenting him 'with his handsomest copy of the puzzle', that Hamilton got the idea for the Icosian Game from p. 160 of Kirkman's 1858 article,

Lucas.  RM2, 1883, pp. 208‑210.  First? mention of the solid version.  The 2nd ed., 1893, has a footnote referring to Kirkman, 1858.

John Jaques & Son.  The Traveller's Dodecahedron;  or,  A Voyage Round the World.  A New Puzzle.  "This amusing puzzle, exercising considerable skill in its solution, forms a popular illustration of Sir William Hamilton's Icosian Game.  A wood dodecahedron with the base pentagon stretched so that when it sits on the base, all vertices are visible.  With ivory? pegs at the vertices, a handle that screws into the base, a string with rings at the ends and one page of instructions, all in a box.  No date.  The only known example was obtained by James Dalgety in 2002.

Pearson.  1907.  Part III, no. 60: The open door, pp. 60 & 130.  Prisoner in one corner of an  8 x 8  array is allowed to exit from from the other corner provided he visits every cell once.  This requires him to enter and leave a cell by the same door.

Ahrens.  Mathematische Spiele.  2nd ed., Teubner, Leipzig, 1911.  P. 44, note, says that a Dodekaederspiel is available from Firma Paul Joschkowitz -- Magdeburg for  .65  mark.  This is not in the 1st ed. of 1907 and the whole Chapter is dropped in the 3rd ed. of 1916 and the later editions.

Anonymous.  The problems drive.  Eureka 12 (Oct 1949) 7-8 & 15.  No. 3.  How many Hamiltonian circuits are there on a cube, starting from a given point?  Reflections and reversals count as different tours.  Answer is 12, but this assumes also that rotations are different.  See Singmaster, 1975, for careful definitions of how to count.  There are 96 labelled circuits, of which 12 start at a given vertex.  But if one takes all the 48 symmetries of the cube as equivalences (six of which fix the given vertex), there are just 2 circuits from a given starting point.  However, these are actually the same circuit started at different points.  Presumably Kirkman and Hamilton knew of this.

C. W. Ceram.  Gods, Graves and Scholars.  Knopf, New York, 1956, pp. 26-29.  2nd ed., Gollancz, London, 1971, pp. 24-25.  Roman knobbed dodecahedra -- an ancient solid version??

R. E. Ingram.  Appendix 2: The Icosian Calculus.  In:  The Mathematical Papers of Sir William Rowan Hamilton.  Vol. III: Algebra.  Ed. by H. Halberstam & R. E. Ingram.  CUP, 1967, pp. 645‑647.  [Halberstam told me that this Appendix is due to Ingram.]  Discusses the method and asserts that the tetrahedron, cube and dodecahedron have only one unlabelled circuit, the octahedron has two and the icosahedron has 17.

David Singmaster.  Hamiltonian circuits on the regular polyhedra.  Notices Amer. Math. Soc. 20 (1973) A‑476, no. 73T‑A199.  Confirms Ingram's results and gives the number of labelled circuits.

David Singmaster.  Op. cit. in 5.F.1.  1975.  Carefully defines labelled and unlabelled circuits.  Discusses results on regular polyhedra in 3 and higher dimensions.

David Singmaster.  Hamiltonian circuits on the  n‑dimensional octahedron.  J. Combinatorial Theory (B) 18 (1975) 1‑4.  Obtains an explicit formula for the number of labelled circuits on the  n‑dimensional octahedron and shows it is  @ (2n)!/e.  Gives numbers for  n £ 8.  In unpublished work, it is shown that the number of unlabelled circuits is asymptotic to the above divided by  n!2n×4n.

Angus Lavery.  The Puzzle Box.  G&P 2 (May 1994) 34-35.  Alternative solitaire, p. 34.  Asks for a knight's tour on the 33-hole solitaire board.  Says he hasn't been able to do it and offers a prize for a solution.  In Solutions, G&P 3 (Jun 1994) 44, he says it cannot be done and the proof will be given in a future issue, but I never saw it.

 

          5.F.3.           KNIGHT'S TOURS IN HIGHER DIMENSIONS

 

A.‑T. Vandermonde.  Remarques sur les problèmes de situation.  Hist. de l'Acad. des Sci. avec les Mémoires (Paris) (1771 (1774)) Mémoires: pp. 566‑574 & Plates I & II.  ??NYS.  First? mention of  cubical problem.  (Not given in BLW excerpt.)

F. Maack.  Mitt. über Raumschak.  1909, No. 2, p. 31.  ??NYS -- cited by Gibbins, below.  Knight's tour on  4 x 4 x 4  board.

Dudeney.  AM.  1917.  Prob. 340: The cubic knight's tour, pp. 103 & 229.  Says Vandermonde asked for a tour on the faces of a  8 x 8 x 8  cube.  He gives it as a problem with a solution.

N. M. Gibbins.  Chess in three and four dimensions.  MG 28 (No. 279) (1944) 46‑50.  Gives knight's tour on  3 x 3 x 4  board -- an unpublished result due to E. Hubar‑Stockar of Geneva.  This is the smallest 3‑D board with a tour.  Gives Maack's tour on  4 x 4 x 4  board.

Ian Stewart.  Solid knight's tours.  JRM 4:1 (Jan 1971) 1.  Cites Dudeney.  Gives a tour through the entire  8 x 8 x 8  cube by stacking 8 knight's paths.

T. W. Marlow.  Closed knight tour of a  4 x 4 x 4  board.  Chessics 29 & 30 (1987) 162.  Inspired by Stewart.

 

          5.F.4.           OTHER CIRCUITS IN AND ON A CUBE

 

          The number of Hamiltonian Circuits on the  n-dimensional cube is the same as the number of Gray codes (see 7.M.3) and has been the subject of considerable research.  I will not try to cover this in detail.

 

D. W. Crowe.  The  n‑dimensional cube and the Tower of Hanoi.  AMM 63:1 (Jan 1956) 29‑30.

E. N. Gilbert.  Gray codes and paths on the  n-cube.  Bell System Technical Journal 37 (1958) 815-826.  Shows there are 9 inequivalent circuits on the 4-cube and 1 on the  n-cube for n = 1, 2, 3.  The latter cases are sufficiently easy that they may have been known before this.

Allen F. Dreyer.  US Patent 3,222,072 -- Block Puzzle.  Filed: 11 Jun 1965;  patented: 7 Dec 1965.  4pp + 2pp diagrams.  27 cubes on an elastic.  The holes are straight or diagonal so that three consecutive cubes are either in a line or form a right angle.  A solution is a Hamiltonian path through the 27 cells.  Such puzzles were made in Germany and I was given one about 1980 (see Singmaster and Haubrich & Bordewijk below).  Dreyer gives two forms.

Gardner.  The binary Gray code.  SA (Aug 1972)  c= Knotted, chap. 2.  Notes that the number of circuits on the  n-cube,  n > 4,  is not known.  SA (Apr 1973) reports that three (or four) groups had found the number of circuits on the 4-cube -- this material is included in the Addendum in Knotted, chap. 2, but none of the groups ever seem to have published their results elsewhere.  Unfortunately, none of these found the number of inequivalent circuits since they failed to take all the equivalences into account -- e.g. for  n = 1, 2, 3, 4, 5,  their enumerations give:  2, 8, 96, 43008, 5 80189 28640  for the numbers of labelled circuits.  Gardner's Addendum describes some further work including some statistical work which estimates the number on the 6-cube is about  2.4 x 1025.

David Singmaster.  A cubical path puzzle.  Written in 1980 and submitted to JRM, but never published.  For the  3 x 3 x 3  problem, the number,  S,  of straight through pieces (ignoring the ends) satisfies  2 £ S £ 11.

Mel A. Scott.  Computer output, Jun 1986, 66pp.  Determines there are 3599 circuits through the  3 x 3 x 3  cube such that the resulting string of 27 cubes can be made into a cube in just one way.  But cf the next article which gives a different number??

Jacques Haubrich & Nanco Bordewijk.  Cube chains.  CFF 34 (Oct 1994) 12‑15.  Erratum, CFF 35 (Dec 1994) 29.  Says Dreyer is the first known reference to the idea and that they were sold 'from about 1970'  Reproduces the first page of diagrams from Dreyer's patent.  Says his first version has a unique solution, but the second has 38 solutions.  They have redone previous work and get new numbers.  First, they consider all possible strings of 27 cubes with at most three in a line (i.e. with at most a single 'straight' piece between two 'bend' pieces and they find there are  98,515  of these.  Only  11,487  of these can be folded into a  3 x 3 x 3  cube.  Of these,  3654  can be folded up in only one way.  The chain with the most solutions had 142 different solutions.  They refer to Mel Scott's tables and indicate that the results correspond -- perhaps I miscounted Scott's solutions??

 

          5.G.   CONNECTION PROBLEMS

 

          5.G.1.          GAS, WATER AND ELECTRICITY

 

Dudeney.  Problem 146 -- Water, gas, and electricity.  Strand Mag. 46 (No. 271) (Jul 1913) 110  &  (No. 272) (Aug 1913) 221 (c= AM, prob. 251, pp. 73  &  200‑201).  Earlier version is slightly more interesting, saying the problem 'that I have called "Water, Gas, and Electricity" ... is as old as the hills'.  Gives trick solution with pipe under one house.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 96: The "three houses" problem, pp. 89-90 & 114.  "Were all the houses connected up with all three supplies or not?"  Answer is no -- one connection cannot be made.

Loyd, Jr.  SLAHP.  1928.  The three houses and three wells, pp. 6 & 87‑88.  "A puzzle ... which I first brought out in 1900 ..."  The drawing is much less polished than Dudeney's.  Trick solution with a pipe under one house, a bit differently laid out than Dudeney.

The Bile Beans Puzzle Book.  1933.  No. 46: Water, gas & electric light.  Trick solution almost identical to Dudeney.

Philip Franklin.  The four color problem.  In:  Galois Lectures; Scripta Mathematica Library No. 5; Scripta Mathematica, Yeshiva College, NY, 1941, pp. 49-85.  On p. 74, he refers to the graph as "the basis of a familiar puzzle, to join each of three houses with each of three wells (or in a modern version to a gas, water, and electricity plant)".

Leeming.  1946.  Chap. 6, prob. 4: Water, gas and electricity, pp. 71 & 185.  Dudeney's trick solution.

H. ApSimon.  Note 2312:  All modern conveniences.  MG 36 (No. 318) (Dec 1952) 287‑288.  Given  m  houses and  n  utilities, the maximum number of non‑crossing connections is  2(m+n‑2)  and this occurs when all the resulting regions are 4‑sided.  He extends to  p‑partite graphs in general and a special case.

John Paul Adams.  We Dare You to Solve This!  Op. cit. in 5.C.  1957?  Prob. 50: Another enduring favorite appears below, pp. 30 & 49.  Electricity, gas, water.  Dudeney's trick solution.

Young World.  c1960.  P. 4: Crossed lines.  Electricity, TV and public address lines.  Trick solution with a line passing under a house.

T. H. O'Beirne.  For boys, men and heroes.  New Scientist 12 (No. 266) (21 Dec 1961) 751‑753.  Shows you can join 4 utilities to 4 houses on a torus without crossing.

 

          5.H.   COLOURED SQUARES AND CUBES, ETC.

 

          5.H.1.          INSTANT INSANITY  =  THE TANTALIZER

          Note.  Often the diagrams do not show all sides of the pieces so I cannot tell if one version is the same as another.

 

Frederick A. Schossow.  US Patent 646,463 -- Puzzle.  Applied: 19 May 1899;  patented: 3 Apr 1900.  1p + 1p diagrams.  Described in S&B, p. 38, which also says it is described in O'Beirne, but I don't find it there??  Four cubes with suit patterns.  The net of each cube is shown.  The fourth cube has three clubs.

George Duncan Moffat.  UK Patent 9810 -- Improvements in or relating to Puzzle-apparatus.  Applied: 28 May 1900;  accepted: 30 Jun 1900.  2pp + 1p diagrams.  For a six cube version with  "letters  R,  K,  B,  W,  F  and  B-P,  the initials of the names of General Officers of the South African Field Force."

Joseph Meek.  UK Patent 2775 -- Improved Puzzle Game.  Applied: 5 Feb 1909;  complete specification: 16 Jun 1909;  accepted: 3 Feb 1910.  2pp + 1p diagrams.  A four cube version with suit patterns.  His discussion seems to describe the pieces drawn by Schossow.

Slocum.  Compendium.  Shows:  The Great Four Ace Puzzle (Gamage's, 1913);  Allies Flag Puzzle (Gamage's, c1915);  Katzenjammer Puzzle (Johnson Smith, 1919).

Edwin F. Silkman.  US Patent 2,024,541 -- Puzzle.  Applied: 9 Sep 1932;  patented: 17 Dec 1935.  2pp + 1 p diagrams.  Four cubes marked with suits.  The net of each cube is shown.  The third cube has three hearts.  This is just a relabelling of Schossow's pattern, though two cubes have to be reflected which makes no difference to the solution process.

E. M. Wyatt.  The bewitching cubes.  Puzzles in Wood.  (Bruce Publishing, Co., Milwaukee, 1928)  = Woodcraft Supply Corp., Woburn, Mass., 1980, p. 13.  A six cube, six way version.

Abraham.  1933.  Prob. 303 -- The four cubes, p. 141 (100).  4 cube version "sold ... in 1932".

A. S. Filipiak.  Four ace cube puzzle.  100 Puzzles, How do Make and How to Solve Them.  A. S. Barnes, NY, (1942)  = Mathematical Puzzles, and Other Brain Twisters; A. S. Barnes, NY, 1966;  Bell, NY, 1978;  p. 108.

Leeming.  1946.  Chap. 10, prob. 9: The six cube puzzle, pp. 128‑129 & 212.  Identical to Wyatt.

F. de Carteblanche [pseud. of Cedric A. B. Smith].  The coloured cubes problem.  Eureka 9 (1947) 9‑11.  General graphical solution method, now the standard method.

T. H. O'Beirne.  Note 2736:  Coloured cubes: A new "Tantalizer".  MG 41 (No. 338) (Dec 1957) 292-293.  Cites Carteblanche, but says the current version is different.  Gives a nicer version.

T. H. O'Beirne.  Note 2787:  Coloured cubes: a correction to Note 2736.  MG 42 (No. 342) (Dec 1958) 284.  Finds more solutions than he had previously stated.

Norman T. Gridgeman.  The 23 colored cubes.  MM 44:5 (Nov 1971) 243-252.  The  23  colored cubes are the equivalence classes of ways of coloring the faces with  1  to  6  colors.  He cites and describes some later methods for attacking Instant Insanity problems.

Jozsef Bognár.  UK Patent Application 2,076,663 A -- Spatial Logical Puzzle.  Filed 28 May 1981;  published 9 Dec 1981.  Cover page + 8pp + 3pp diagrams.  Not clear if the patent was ever granted.  Describes Bognár's Planets, which is a four piece instant insanity where the pieces are spherical and held in a plastic tube.  This was called Bolygok in Hungarian and there is a reference to an earlier Hungarian patent.  Also describes his version with eight pieces held at the corners of a plastic cube.

 

          5.H.2.          MACMAHON PIECES

 

          Haubrich's 1995-1996 surveys, op. cit. in 5.H.4, include MacMahon puzzles as one class.

          I have just added the Carroll result that there are 30 six-coloured cubes, but this must be older??

 

Frank H. Richards.  US Patent 331,652 -- Domino.  Applied: 13 Jun 1885;  patented: 1 Dec 1885.  2pp + 2pp diagrams.  Cited by Gardner in Magic Show, but with date 1895.  Reproduced in Haubrich, About ..., 1996, op. cit. in 5.H.4.  For triangular matching games, specifically showing the MacMahon 5-coloured triangles, but considering reflections as equivalences, so he has 35 pieces.  [One of the colours is blank and hence Gardner said it was a 4-colouring.]

Carroll-Wakeling.  c1890?  Prob. 15: Painting cubes, pp. 18-19 & 67.  This is one of the problems on undated sheets of paper that Carroll sent to Bartholomew Price.  How many ways can one six-colour a cube?  Wakeling gives a solution, but this apparently is not on Carroll's MS.

Percy Alexander MacMahon  &  Julian Robert John Jocelyn.  UK Patent 3927  A.D. 1892 -- Appliances to be used in Playing a New Class of Games.  Applied: 29 Feb 1892;  Complete Specification Left: 28 Nov 1892;  Accepted: 28 Jan 1893.  5pp + 2pp diagrams.  Reproduced in Haubrich, About ..., 1996, op. cit. in 5.H.4.  Describes the 24 triangles with four types of edge and mentions other numbers of edge types.  Describes various games and puzzles.

Percy Alexander MacMahon  &  Julian Robert John Jocelyn.  UK Patent 8275  A.D. 1892 -- Appliances for New Games of Puzzles.  Applied: 2 May 1892;  Complete Specification Left: 31 Jan 1893;  Accepted: 4 Mar 1893.  2pp.  27 cubes with three colours, opposite faces having the same colour.  Similar sets of 8, 27, 64, etc. cubes.  Various matching games suggested.  Using six colours and all six on each cube gives 30 cubes -- the MacMahon Cubes.  Gives a complex matching problem of making two  2 x 2 x 2  cubes.  Paul Garcia (email of 15 Nov 2002) commented:  "8275 describes 2 different sets of blocks, using either three colours or six colours.  The three colour blocks form a set of 27 that can be assembled into a large cube with single coloured faces and internal contact faces matching.  For the six colour cubes, the puzzle suggested is to pick out two associated cubes, and find the sixteen cubes that can be assembled to make a copy of each.  Not quite Mayblox, although using the same colouring system."

James Dalgety.  R. Journet & Company  A Brief History of the Company & its Puzzles.  Published by the author, North Barrow, Somerset, 1989.  On p. 13, he says Mayblox was patented in 1892.  In an email on 12 Nov 2002, he cited  UK Patent 8275. 

Anon.  Report:  "Mathematical Society, February 9".  Nature 47 (No. 1217) (23 Feb 1893) 406.  Report of MacMahon's talk:  The group of thirty cubes composed by six differently coloured squares.

See:  Au Bon Marché, 1907, in 5.P.2, for a puzzle of hexagons with matching edges.

Manson.  1911.  Likoh, pp. 171-172.  MacMahon's 24 four-coloured isosceles right triangles, attributed to MacMahon.

"Toymaker".  The Cubes of Mahomet Puzzle.  Work, No. 1447 (9 Dec 1916) 168.  8 six-coloured cubes to be assembled into a cube with singly-coloured faces and internal faces to have matching colours.

P. A. MacMahon.  New Mathematical Pastimes.  CUP, 1921.  The whole book deals with variations of the problem and calculates the numbers of pieces of various types.  In particular, he describes the  24  4-coloured triangles, the  24  3-coloured squares, the MacMahon cubes, some right-triangular and hexagonal sets and various subsets of these.  With  n  colours, there are  n(n2+2)/3  triangles,  n(n+1)(n2‑n+2)/4  squares and n(n+1)(n4-n3+n2+2)/6  hexagons.  [If one allows reflectional equivalence, one gets  n(n+1)(n+2)/6  triangles,  n(n+1)(n2+n+2)/8  squares  and  n(n+1)(n4-n3+4n2+2)/12  hexagons.  Problem -- is there an easy proof that the number of triangles is  BC(n+2, 3)?]  On p. 44, he says that Col. Julian R. Jocelyn told him some years ago that one could duplicate any cube with 8 other cubes such that the internal faces matched.

Slocum.  Compendium.  Shows Mayblox made by R. Journet from Will Goldston's 1928 catalogue.

F. Winter.  Das Spiel der 30 bunten Würfel  MacMahon's Problem.  Teubner, Leipzig, 1934, 128pp.  ??NYR.

Clifford Montrose.  Games to play by Yourself.  Universal Publications, London, nd [1930s?].  The coloured squares, pp. 78-80.  Makes 16 squares with four-coloured edges, using five colours, but there is no pattern to the choice.  Uses them to make a  4 x 4  array with matching edges, but seems to require the orientations to be fixed.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 6.  There are twelve ways to colour the edges of a pentagon, when rotations and reflections are considered as equivalences.  Can you colour the edges of a dodecahedron so each of these pentagonal colourings occurs once?  [If one uses tiles, one has to have reversible tiles.]  Solution says there are three distinct solutions and describes them by describing contacts between 10 pentagons forming a ring around the equator.

Richard K. Guy.  Some mathematical recreations I  &  II.  Nabla [= Bull. Malayan Math. Soc.] 7 (Oct  &  Dec 1960) 97-106  &  144-153.  Pp. 101-104 discusses MacMahon triangles, squares and hexagons.

T. H. O'Beirne.  Puzzles and paradoxes 5: MacMahon's three-colour set of squares.  New Scientist 9 (No. 220) (2 Feb 1961) 288-289.  Finds 18 of the 20 possible monochrome border patterns.

Gardner.  SA (Mar 1961) = New MD, Chap. 16.  MacMahon's 3-coloured squares and his cubes.  Addendum in New MD cites Feldman, below.

Gary Feldman.  Documentation of the MacMahon Squares Problem.  Stanford Artificial Intelligence Project Memo No. 12, Stanford Computation Center, 16 Jan 1964.  ??NYS  Finds  12,261  solutions for the  6 x 4  rectangle with monochrome border -- but see Philpott, 1982, for  13,328  solutions!!

Gardner.  SA (Oct 1968) = Magic Show, Chap. 16.  MacMahon's four-coloured triangles and numerous variants.

Wade E. Philpott.  MacMahon's three-color squares.  JRM 2:2 (1969) 67-78.  Surveys the topic and repeats Feldman's result.

N. T. Gridgeman, loc. cit. in 5.H.1, 1971, covers some ideas on the MacMahon cubes.

J. J. M. Verbakel.  Digitale tegels (Digital tiles).  Niet piekeren maar puzzelen (name of a puzzle column).  Trouw (a Dutch newspaper) (1 Feb 1975).  ??NYS -- described by Jacques Haubrich; Pantactic patterns and puzzles; CFF 34 (Oct 1994) 19-21.  There are 16 ways to 2‑colour the edges of a square if one does not allow them to rotate.  Assemble these into a  4 x 4  square with matching edges.  There are  2,765,440  solutions in  172,840  classes of 16.  One can add further constraints to yield fewer solutions -- e.g. assume the  4 x 4  square is on a torus and make all internal lines have a single colour.

Gardner.  Puzzling over a problem‑solving matrix, cubes of many colours and three‑dimensional dominoes.  SA 239:3 (Sep 1978) 20‑30 & 242  c= Fractal, chap. 11.  Good review of MacMahon (photo) and his coloured cubes.  Bibliography cites recent work on Mayblox, etc.

Wade E. Philpott.  Instructions for Multimatch.  Kadon Enterprises, Pasadena, Maryland, 1982.  Multimatch is just the 24 MacMahon 3-coloured squares.  This surveys the history, citing several articles ??NYS, up to the determination of the  13,328  solutions for the  6 x 4  rectangle with monochrome border, by Hilario Fernández Long (1977) and John W. Harris (1978).

Torsten Sillke.  Three  3 x 3  matching puzzles.  CFF 34 (Oct 1994) 22-23.  He has wanted an interesting 9 element subset of the MacMahon pieces and finds that of the 24 MacMahon 3-coloured squares, just 9 of them contain all three colours.  He considers both the corner and the edge versions.  The editor notes that a  3 x 3  puzzle has  36 x 32/2 = 576  possible edge contacts and that the number of these which match is a measure of the difficulty of the puzzle, with most  3 x 3  puzzles having  60  to  80  matches.  The corner version of Sillke's puzzle has  78  matches and one solution.  The edge version has  189  matches and many solutions, hence Sillke proposes various further conditions.

 

          5.H.3.          PATH FORMING PUZZLES

 

          Here we have a set of pieces and one has to join them so that some path is formed.  This is often due to a chain or a snake, etc.  New section.  Again, Haubrich's 1995-1996 surveys, op. cit. in 5.H.4, include this as one class.

 

 

Hoffmann.  1893.  Chap. III, No. 18: The endless chain, pp. 99-100 & 131 = Hoffmann‑Hordern, pp. 91-92, with photo.  18  pieces, some with parts of a chain, to make into an  8 x 8  array with the chain going through  34  of the cells.  All the pieces are rectangles of width one.  Photo shows The Endless Chain, by The Reason Manufacturing Co., 1880-1895.  Hordern Collection, p. 62, shows the same and La Chaine sans fin, 1880-1905.

Loyd.  Cyclopedia.  1914.  Sam Loyd's endless chain puzzle, pp. 280 & 377.  Chain through all 64 cells of a chessboard, cut into 13 pieces.  The chessboard dissection is of type:  13: 02213 131.

Hummerston.  Fun, Mirth & Mystery.  1924.  The dissected serpent, p. 131.  Same pieces as Hoffmann, and almost the same pattern.

Collins.  Book of Puzzles.  1927.  The dissected snake puzzle, pp. 126-127.  17 pieces forming an  8 x 8  square.  All the piece are rectangular pieces of width one except for one L‑hexomino -- if this were cut into straight tetromino and domino, the pieces would be identical to Hoffmann.  The pattern is identical to Hummerston.

See Haubrich in 5.H.4.

 

          5.H.4.          OTHER AND GENERAL

 

          These all have coloured edges unless specified.  See S&B, p. 36, for examples.

 

Edwin L[ajette] Thurston.  US Patent 487,797 -- Puzzle.  Applied: 30 Sep 1890;  patented: 13 Dec 1892.  3pp + 3pp diagrams.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  4 x 4  puzzles with 6-coloured corners or edges, but assuming no colour is repeated on a piece -- indeed he uses the  15 = BC(6,2)  ways of choosing 4 out of 6 colours once only and then has a sixteenth with the same colours as another, but in different order.  Also a star-shaped puzzle of six parallelograms.

Edwin L. Thurston.  US Patent 487,798 -- Puzzle.  Applied: 30 Sep 1890;  patented: 13 Dec 1892.  2pp + 1p diagrams.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  As far as I can see, this is the same as the  4 x 4  puzzle with 6‑coloured edges given above, but he seems to be emphasising the 15 pieces.

Edwin L. Thurston.  US Patent 490,689 -- Puzzle.  Applied: 30 Sep 1890;  patented: 31 Jun 1893.  2pp + 1p diagrams.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  The patent is for  3 x 3  puzzles with 4‑coloured corners or edges, but with pieces having no repeated colours and in a fixed orientation.  He selects some 8 of these pieces for reasons not made clear and mentions moving them "after the  manner of the old 13, 14, 15 puzzle."  S&B, p. 36, describes the Calumet Puzzle, Calumet Baking Powder Co., Chicago, which is a  3 x 3  head to tail puzzle, claimed to be covered by this patent.

Le Berger Malin.  France, c1900.  3 x 3  head to tail puzzle, but the edges are numbered and the matching edges must add to 10.  ??NYS -- described by K. Takizawa, N. Takashima & N. Yoshigahara; Vess Puzzle and Its Family -- A Compendium of  3 by 3  Card Puzzles; published by the authors, Tokyo, 1983.  Slocum has this in two different boxes and dates it to c1900 -- I had c1915 previously.  Haubrich has one version, Produced by GB&O N.K. Atlas.

Angus K. Rankin.  US Patent 1,006,878 -- Puzzle.  Applied: 3 Feb 1911;  patented: 24 Oct 1911.  2pp + 1p diagrams.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  Described in S&B, p. 36.  Grandpa's Wonder Puzzle.  3 x 3  square puzzle.  Each piece has corners coloured, using four colours, and the colours meeting at a corner must differ.  The patent doesn't show the advertiser's name -- Grandpa's Wonder Soap -- but is otherwise identical to S&B's photo.

Daily Mail World Record Net Sale puzzle.  1920‑1921.  Instructions and picture of the pieces.  Letter from Whitehouse to me describing its invention.  19  6-coloured hexagons without repeated colours.  Daily Mail articles as follows.  There may be others that I missed and sometimes the page number is a bit unclear.  Note that 5 Dec was a Sunday.

  9 Nov 1920, p. 5.  "Daily Mail" puzzle.  To be issued on 7 Dec.

13 Nov 1920, p. 4.  Hexagon mystery.

17 Nov 1920, p. 5.  New mystery puzzle.  Asserts the inventor does not know the solution -- i.e. the solution has been locked up in a safe.

20 Nov 1920, p. 4.  What is it?

23 Nov 1920, p. 5.  Fascinating puzzle.  The most fascinating puzzle since "Pigs in Clover".

25 Nov 1920, p. 5.  Can you do it?

29 Nov 1920, p. 5.  £250 puzzle.

  1 Dec 1920, p. 4.  Mystery puzzle clues.

  2 Dec 1920, p. 5.  £250 puzzle race.

  3 Dec 1920, p. 5.  The puzzle.

  4 Dec 1920, p. 4.  The puzzle.  Amplifies on the inventor not knowing the solution -- after the idea was approved, a new pattern was created by someone else and locked up.

  6 Dec 1920, unnumbered back page.  Photo with caption:  £250 for solving this.

  7 Dec 1920, p. 7.  "Daily Mail" Puzzle.  Released today.  £100 for getting the locked up solution.  £100 for the first alternative solution and £50 for the next alternative solution.  "It is believed that more than one solution is possible."

  8 Dec 1920, p. 5.  "Daily Mail" puzzle.

  9 Dec 1920, p. 5.  Can you do it?

10 Dec 1920, p. 4.  It can be done.

13 Dec 1920, p. 9.  Most popular pastime.  "More than 500,000 Daily Mail Puzzles have been sold."

15 Dec 1920, p. 4.  Puzzle king & the 19 hexagons.  Dudeney says he does not think it can be solved "except by trial."

16 Dec 1920, p. 4.  Tantalising 19 hexagons.

16 Dec 1920, unnumbered back page.  Banner at top has:  "The Daily Mail" puzzle.  Middle of page has a cartoon of sailors trying to solve it.

17 Dec 1920, p. 5?  The Xmas game.

18 Dec 1920, p. 7.  Puzzle Xmas 'card'.

20 Dec 1920, p. 7.  Hexagon fun.

22 Dec 1920, p. 3.  3,000,000 fascinated.  It is assumed that about 5 people try each example and so this indicates that nearly 600,000 have been sold.

23 Dec 1920, p. 3.  Too many cooks.

23 Dec 1920, unnumbered back page.  Cartoon:  The hexagonal dawn!

28 Dec 1920, p. 3?  Puzzled millions.  "On Christmas Eve the sales exceeded 600,000 ...."

29 Dec 1920, p. 3?  "I will do it." 

30 Dec 1920, p. 8.  Puzzle fun.

  3 Jan 1921, p. 3.  The Daily Mail Puzzle.  C. Lewis, aged 21, a postal clerk solved it within two hours of purchase and submitted his solution on 7 Dec.  Hundreds of identical solutions were submitted, but no alternative solutions have yet appeared.  There are two pairs of identical pieces:  1 & 12,  4 & 10.

  3 Jan 1921, p. 10 = unnumbered back page.  Hexagon Puzzle Solved, with photo of C. Lewis and diagram of solution.

10 Jan 1921, p. 4.  Hexagon puzzle.  Since no alternative hexagonal solutions were received, the other £150 is awarded to those who submitted the most ingenious other solution -- this was judged to be a butterfly shape, submitted by 11 persons, who shared the £150.

Horace Hydes & Francis Reginald Beaman Whitehouse.  UK Patent 173,588 -- Improvements in Dominoes.  Applied: 29 Sep 1920;  complete application: 29 Jun 1921;  accepted: 29 Dec 1921.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  3pp + 1p diagrams.  This is the patent for the above puzzle, corresponding to provisional patent  27599/20  on the package.  The illustration shows a solved puzzle based on  'A stitch in time saves nine'.

George Henry Haswell.  US Patent 1,558,165 -- Puzzle.  Applied: 3 Jul 1924;  patented: 11 Sep 1925.  Reproduced in Haubrich, About ..., 1996, op. cit. below.  2pp + 1p diagrams.  For edge-matching hexagons with further internal markings which have to be aligned.  [E.g. one could draw a diagonal and require all diagonals to be vertical -- this greatly simplifies the puzzle!]  If one numbers the vertices  1, 2, ..., 6,  he gives an example formed by drawing the diagonals  13, 15, 42, 46  which produces six triangles along the edges and an internal rhombus.

C. Dudley Langford.  Note 2829:  Dominoes numbered in the corners.  MG 43 (No. 344) (May 1959) 120‑122.  Considers triangles, squares and hexagons with numbers at the corners.  There are the same number of pieces as with numbers on the edges, but corner numbering gives many more kinds of edges.  E.g. with four numbers, there are 24 triangles, but these have 16 edge patterns instead of 4.  The editor (R. L. Goodstein) tells Langford that he has made cubical dominoes "presumably with faces numbered".  Langford suggests cubes with numbers at the corners.  [I find 23 cubes with two corner numbers and 333 with three corner numbers.  ??check]

Piet Hein.  US Patent 4,005,868 -- Puzzle.  Applied: 23 Jun 1975;  patented: 1 Feb 1977.  Front page + 8pp diagrams + 5pp text.  Basically non-matching puzzles using marks at the corners of faces of the regular polyhedra.  He devises boards so the problems can be treated as planar.

Kiyoshi Takizawa;  Naoaki Takashima  &  Nob. Yoshigahara.  Vess Puzzle and Its Family -- A Compendium of  3  by  3  Card Puzzles.  Published by the authors, Tokyo, Japan, 1983.  Studies  32  types (in  48  versions) of  3 x 3  'head to tail' matching puzzles and  4  related types (in  4  versions).  All solutions are shown and most puzzles are illustrated with colour photographs of one solution.  (Haubrich counts 51 versions -- check??)

Melford D. Clark.  US Patent 4,410,180 -- Puzzle.  Applied: 16 Nov 1981;  patented: 18 Oct 1983.  Reproduced in Haubrich, About ..., 1996, op. cit. in 5.H.4.  2pp + 2pp diagrams.  Corner matching squares, but with the pieces marked  1, 2, ...,  so that the pieces marked  1 form a  1 x 1  square, the pieces marked  2  allow this to be extended to a  2 x 2  square, etc.  There are  n2 - (n-1)2  pieces marked  n.

Jacques Haubrich.  Compendium of Card Matching Puzzles.  Printed by the author, Aeneaslaan 21, NL-5631 LA Eindhoven, Netherlands, 1995.  2 vol., 325pp. describing over  1050  puzzles.  He classifies them by the nine most common matching rules:  Heads and Tails;  Edge Matching (i.e. MacMahon);  Path Matching;  Corner Matching;  Corner Dismatching;  Jig-Saw-Like;  Continuous Path;  Edge Dismatching;  Hybrid.  He does not include Jig-Saw-Like puzzles here.  Using the number of cards and their shape, then the matching rules, he has 136 types.  31  different numbers of cards occur:  4, 6-16, 18-21, 23-25, 28, 30, 36, 40, 45, 48, 56, 64, 70, 80, 85, 100.  There is an index of 961 puzzle names.  He says Hoffmann is the earliest published example.  He notes that most path puzzles have a global criterion that the result have a single circuit which slightly removes them from his matching criterion and he does not treat them as thoroughly.  He has developed computer programs to solve each type of puzzle and has checked them all.

Jacques Haubrich.  About, Beyond and Behind Card Matching Puzzles.  [= Vol. 3 of above].  Ibid, Apr 1996, 87pp.  This is a general discussion of the different kinds of puzzles, how to solve them and their history, reproducing ten patents and two obituaries.

 

          5.I.     LATIN SQUARES AND EULER SQUARES

 

          This topic ties in with certain tournament problems but I have not covered them.  See also Hoffmann and Loughlin & Flood in 5.A.2 for examples of two orthogonal 3 x 3 Latin squares.  The derangement problems in 5.K.2 give Latin rectangles.

 

Ahrens-1 & Ahrens-2.  Opp. cit. in 7.N.  1917 & 1922.  Ahrens-1 discusses and cites early examples of Latin squares, going back to medieval Islam (c1200), where they were used on amulets.  Ahrens-2 particularly discusses work of al‑Buni -- see below. 

(Ahmed [the  h  should have an underdot] ibn ‘Alî ibn Jûsuf) el‑Bûni, (Abû'l‑‘Abbâs, el‑Qoresî.)  = Abu‑l‘Abbas al‑Buni.  (??= Muhyi'l‑Dîn Abû’l-‘Abbâs al‑Bûnî  -- can't relocate my source of this form.)  Sams al‑ma‘ârif  = Shams al‑ma‘ârif al‑kubrâ  = Šams al-ma‘ārif.  c1200.  ??NYS.  Ahrens-1 describes this briefly and incorrectly.  He expands and corrects this work in Ahrens-2.  See 7.N for more details.  Ahrens notes that a  4 x 4  magic square can be based on the pattern of two orthogonal Latin squares of order  4,  and Al-Buni's work indicates knowledge of such a pattern, exemplified by the square

           8, 11, 14,  1;    13,  2,  7, 12;     3, 16,  9,  6;    10,  5,  4, 15   considered  (mod 4).  He also has Latin squares of order  4  using letters from a name of God.  He goes on to show  7  Latin squares of order  7,  using the same  7  letters each time -- though four are corrupted.  (Throughout, the Latin squares also have 'Latin' diagonals, i.e. the diagonals contain all the values.)  These are arranged so each has a different letter in the first place.  It is conjectured that these are associated with the days of the week or the planets.

Tagliente.  Libro de Abaco.  (1515).  1541.  F. 18v.  7 x 7  Latin square with entries  1, 13, 2, 14, 3, 10, 4  cyclically shifted forward -- i.e. the second row starts  13, 2, ....  This is an elaborate plate which notes that the sum of each file is 47 and has a motto: Sola Virtu la Fama Volla, but I could find no text or other reason for its appearance!

 

Inscription on memorial to Hannibal Bassett, d. 1708, in Meneage parish church, St. Mawgan, Cornwall.  I first heard of this from Chris Abbess, who reported it in some newsletter in c1993.  However, [Peter Haining; The Graveyard Wit; Frank Graham, Newcastle, 1973, p. 133] cites this as being at Cunwallow, near Helstone, Cornwall.  [W. H. Howe; Everybody's Book of Epitaphs Being for the Most Part What the Living Think of the Dead; Saxon & Co., London, nd [c1895] (facsimile by Pryor Publications, Whitstable, 1995); p. 173] says it is in Gunwallow Churchyard.  Spelling and punctuation vary a bit.  The following gives a detailed account.

Alfred Hayman Cummings.  The Churches and Antiquities of Cury & Gunwalloe, in the Lizard District, including Local Traditions.  E. Marlborough & Co., London & Truro, 1875, pp. 130-131.  ??NX.  "It has been said that there once existed ... the curious epitaph --" and gives a considerable rearrangement of the inscription below.  He continues "But this is in all probability a mistake, as repeated search has been made for it, not only by the writer, but by a former Vicar of Gunwalloe, and it could nowhere be found, while there is a plate with an inscription in the church at Mawgan, the next parish, which might be very easily the one referred to."  He gives the following inscription, saying it is to Hannibal Basset, d. 1708-9.  Chris Weeks was kind enough to actually go to the church of St. Winwaloe, Gunwalloe, where he found nothing, and to St. Mawgan in Meneage, a few miles away.  Chris Weeks sent pictures of Gunwallowe -- the church is close to the cliff edge and it looks like there could once have been more churchyard on the other side of the church where the cliff has fallen away.  In the church at St. Mawgan is the brass plate with 'the Acrostic Brass Inscription', but it is not clearly associated with a grave and I wonder if it may have been moved from Gunwallowe when a grave was eroded by the sea.  It is on the left of the arch by the pulpit.  I reproduce Chris Weeks' copy of the text.  He has sent a photograph, but it was dark and the photo is not very clear, but one can make out the Latin square part.

 

                              Hanniball Baòòet here Inter'd doth lye

                              Who dying lives to all Eternitye

                              hee departed this life the 17th of Ian

                              1709/8 in the 22th year of his age ~

                                        A lover of learning

 

                              Shall       wee        all           dye

                              Wee       shall        dye         all

                              all           dye         shall        wee

                              dye         all           wee        shall

 

                    The  òò  are old style long esses.  The superscript  th  is actually over the numeral.  The  9  is over the  8  in the year and there is no stroke.  This is because it was before England adopted the Gregorian calendar and so the year began on 25 Mar and was a year behind the continent between 1 Jan and 25 Mar.  Correspondence of the time commonly would show  1708/9  at this time, and I have used this form for typographic convenience, but with the  9  over the  8  as on the tomb.

                    A word game book points out that this inscription is also palindromic!!

Richard Breen.  Funny Endings.  Penny Publishing, UK, 1999, p. 35.  Gives the following form:  Shall we all die?  /  We shall die all.  /  All die shall we?  /  Die all we shall  and notes that it is a word palindrome and says it comes from Gunwallam [sic], near Helstone.

 

Joseph Sauveur.  Construction générale des quarrés magiques.  Mémoires de l'Académie Royale des Sciences 1710(1711) 92‑138.  ??NYS -- described in Cammann‑4, p. 297, (see 7.N for details of Cammann) which says Sauveur invented Latin squares and describes some of his work.

Ozanam.  1725.  1725:  vol. IV, prob. 29, p. 434 & fig. 35, plate 10 (12).  Two  4 x 4  orthogonal squares, using  A, K, Q, J  of the 4 suits, but it looks like: 

                    J¨, A©, K§, Qª;  Qª, K§, A¨, J©;  A§, Jª, Q©, K¨;  K©, Q¨, J§, Aª;  but the  §  and ª  look very similar.  From later versions of the same diagram, it is clear that the first row should have its  §  and ª  reversed.   Note the diagonals also contain all four ranks and suits.  (I have a reference for this to the 1723 edition.)

Minguet.  1733.  Pp. 146-148 (1864: 142-143; not noticed in other editions).  Two  4 x 4  orthogonal squares, using  A, K, Q, J  (= As, Rey, Caballo (knight), Sota (knave)) of the 4 suits, but the Spanish suits, in descending order, are: Espadas, Bastos, Oros, Copas.  The result is described but not drawn, as: 

                    RO, AE, CC, SB;  SC, CB, AO, RE;  AB, RC, SE, CO;  CE, SO, RB, AC; 

          which would translate into the more usual cards as: 

                    K¨, A§, Qª, J©;  J§, Q©, A¨, Kª;  A©, K§, Jª, Q¨;  Qª, J¨, K©, A§.

          However, I'm not sure of the order of the Caballo and Sota; if they were reversed, which would interchange Q and J in the latter pattern, then both Ozanam and Minguet would have the property that each row is a cyclic shift or reversal of  A, K, Q, J.

Alberti.  1747.  Art. 29, p. 203 (108) & fig. 36, plate IX, opp. p. 204 (108).  Two  4 x 4  orthogonal squares, figure simplified from the correct form of Ozanam, 1725.

L. Euler.  Recherches sur une nouvelle espèce de Quarrés Magiques.  (Verhandelingen uitgegeven door het zeeuwsch Genootschap der Wetenschappen te Vlissingen (= Flessingue) 9 (1782) 85‑239.)  = Opera Omnia (1) 7 (1923) 291‑392.  (= Comm. Arithm. 2 (1849) 302‑361.)

Manuel des Sorciers.  1825.  Pp. 78-79, art. 39.  ??NX  Correct form of Ozanam.

The Secret Out.  1859.  How to Arrange the Twelve Picture Cards and the four Aces of a Pack in four Rows, so that there will be in Neither Row two Cards of the same Value nor two of the same Suit, whether counted Horizontally or Perpendicularly, pp. 90-92.  Two  4 x 4  orthogonal Latin squares, not the same as in Ozanam.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. XI, 1884: 200‑202.  Two  4 x 4  orthogonal squares.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles, No. XVI, pp. 17-18.  Similar to Ozanam.

Hoffmann.  1893.  Chap. X, no. 14: Another card puzzle, pp. 342 & 378-379 = Hoffmann‑Hordern, pp. 234 & 236.  Two orthogonal Latin squares, but the diagonals do not contain all the suits and ranks.

          Aª, J©, Q¨, K§;   J¨, A§, Kª, Q©;   Q§, K¨, A©, Jª;   K©, Qª, J§, A¨.

G. Tarry.  Le probleme de 36 officiers.  Comptes Rendus de l'Association Française pour l'Avancement de Science Naturel 1 (1900) 122‑123  &  2 (1901) 170‑203.  ??NYS

Dudeney.  Problem 521.  Weekly Dispatch (1 Nov, 15 Nov, 1903) both p. 10.

H. A. Thurston.  Latin squares.  Eureka 9 (Apr 1947) 19-21.  Survey of current knowledge.

T. G. Room.  Note 2569:  A new type of magic square.  MG 39 (No. 330) (Dec 1955) 307.  Introduces 'Room Squares'.  Take the  2n(2n‑1)/2  combinations from  2n  symbols and insert them in a  2n‑1 x 2n‑1  grid so that each row and column contains all  2n  symbols.  There are  n  entries and  n‑1  blanks in each row and column.  There is an easy solution for  n = 1.  n = 2  and  n = 3  are impossible.  Gives a solution for  n = 4.  This is a design for a round‑robin tournament with the additional constraint of  2n‑1  sites such that each player plays once at each site.

Parker shows there are two orthogonal Latin squares of order 10 in 1959.

R. C. Bose  &  S. S. Shrikande.  On the falsity of Euler's conjecture about the nonexistence of two orthogonal Latin squares of order  4t+2.  Proc. Nat. Acad. Sci. (USA) 45: 5 (1959) 734‑737.

Gardner.  SA (Nov 1959)  c= New MD, chap. 14.  Describes Bose & Shrikande's work.  SA cover shows a  10 x 10  counterexample in colour.  Kara Lynn and David Klarner actually made a quilt of this, thereby producing a counterpane counterexample!  They told me that the hardest part of the task was finding ten sufficiently contrasting colours of material.

H. Howard Frisinger.  Note:  The solution of a famous two-centuries-old problem:  the Leonhard Euler-Latin square conjecture.  HM 8 (1981) 56-60.  Good survey of the history.

Jacques Bouteloup.  Carrés Magiques, Carrés Latins et Eulériens.  Éditions du Choix, Bréançon, 1991.  Nice systematic survey of this field, analysing many classic methods.  An Eulerian square is essentially two orthogonal Latin squares.

 

          5.I.1.  EIGHT QUEENS PROBLEM

 

          See MUS I 210-284.  S&B 37 shows examples.  See also 5.Z.  See also 6.T for examples where no three are in a row.

 

Ahrens.  Mathematische Spiele.  Encyklopadie article, op. cit. in 3.B.  1904.  Pp. 1082‑1084 discusses history and results for the  n  queens problem, with many references.

Paul J. Campbell.  Gauss and the eight queens problem.  HM 4:4 (Nov 1977) 397‑404.  Detailed history.  Demonstrates that Gauss did not obtain a complete solution and traces how this misconception originated and spread.

 

"Schachfreund" (Max Bezzel).  Berliner Schachzeitung 3 (Sep 1848) 363. ??NYS

Solutions.  Ibid. 4 (Jan 1849) 40.  ??NYS  (Ahrens says this only gives two solutions.  A. C. White says two or three.  Jaenisch says a total of  5  solutions were published here and in 1854.)

Franz Nauck.  Eine in das Gebiet der Mathematik fallende Aufgabe von Herrn Dr. Nauck in Schleusingen.  Illustrirte Zeitung (Leipzig) 14 (No. 361) (1 Jun 1850) 352.  Reposes problem.  [The papers do not give a first name or initial.  The only Nauck in the first six volumes of Poggendorff is Ernst Friedrich (1819-1875), a geologist.  Ahrens gives no initial.  Campbell gives Franz.]

Franz Nauck.  Briefwechseln mit Allen für Alle.  Illustrirte Zeitung (Leipzig) 15 (No. 377) (21 Sep 1850) 182.  Complete solution.

Editorial comments: Briefwechsel.  Illustrirte Zeitung (Leipzig) 15 (No. 378) (28 Sep 1850) 207.  Thanks 6 correspondents for the complete solution and says Nauck reports that a blind person has also found all  92  solutions.

Gauss read the Illustrirte Zeitung and worked on the problem, corresponding with his friend Schumacher starting on 1 Sep 1850.  Campbell discusses the content of the letters, which were published in:  C. A. F. Peters, ed; Briefwechsel zwischen C. F. Gauss und H. C. Schumacher; vol. 6, Altona, 1865, ??NYS.  John Brillhart writes that there is some material in Gauss' Werke, vol. XII: Varia kleine Notizen verschiednen Inhalts ... 5, pp. 19-28,  ??NYS -- not cited by Campbell.

F. J. E. Lionnet.  Question 251.  Nouvelles Annales de Mathématiques 11 (1852) 114‑115.  Reposes problem and gives an abstract version.

Giusto Bellavitis.  Terza rivista di alcuni articoli dei Comptes Rendus dell'Accademia delle Scienze di Francia e di alcuni questioni des Nouvelles Annales des mathématiques.  Atti dell'I. R. Istituto Veneto di Scienze, Lettere ed Arti (3) 6 [= vol. 19] (1860/61) 376-392 & 411‑436 (as part of Adunanza del Giorno 17 Marzo 1861 on pp. 347‑436).  The material of interest is: Q. 251.  Disposizione sullo scacchiere di otto regine, on pp. 434‑435.  Gives the  12  essentially different solutions.  Lucas (1895) says Bellavitis was the first to find all solutions, but see above.  However this may be the first appearance of the  12  essentially different solutions.

C. F. de Jaenisch.  Op. cit. in 5.F.1.  1862.  Vol. 1, pp. 122-135.  Gives the 12 basic solutions and shows they produce 92.  Notes that in every solution, 4 queens are on white squares and 4 are on black. 

A. C. Cretaine.  Études sur le Problème de la Marche du Cavalier au Jeu des Échecs et Solution du Problème des Huit Dames.  A. Cretaine, Paris, 1865.  ??NYS -- cited by Lucas (1895).  Shows it is possible to solve the eight queens problem after placing one queen arbitrarily.

G. Bellavitis.  Algebra N. 72 Lionnet.  Atti dell'Istituto Veneto (3) 15 (1869/70) 844‑845.

Siegmund Günther.  Zur mathematische Theorie des Schachbretts.  Grunert's Archiv der Mathematik und Physik 56 (1874) 281-292.  ??NYS.  Sketches history of the problem -- see Campbell.  He gives a theoretical, but not very practical, approach via determinants which he carries out for  4 x 4  and  5 x 5.

J. W. L. Glaisher.  On the problem of the eight queens.  Philosophical Magazine (4) 48 (1874) 457-467.  Gives a sketch of Günther's history which creates several errors, in particular attributing the solution to Gauss -- see Campbell, who suggests Glaisher could not read German well.  (However, in 1921 & 1923, Glaisher published two long articles involving the history of 15-16C German mathematics, showing great familiarity with the language.)  Simplifies and extends Günther's approach and does  6 x 6,  7 x 7,  8 x 8  boards.

Lucas.  RM2, 1883.  Note V:  Additions du Tome premier.  Pp. 238-240.  Gives the solutions on the  9 x 9  board, due to P. H. Schoute, in a series of articles titled Wiskundige Verpoozingen in Eigen Haard.  Gives the solutions on the  10 x 10  board, found by M. Delannoy.

S&B, p. 37, show an 1886 puzzle version of the six queens problem.

A. Pein.  Aufstellung von  n  Königinnen auf einem Schachbrett von  n2  Feldern.  Leipzig.  ??NYS -- cited by Ball, MRE, 4th ed., 1905 as giving the 92 inequivalent solutions on the  10 x 10.

Ball.  MRE, 1st ed., 1892.  The eight queens problem, pp. 85-88.  Cites Günther and Glaisher and repeats the historical errors.  Sketches Günther's approach, but only cites Glaisher's extension of it.  He gives the numbers of solutions and of inequivalent solutions up through  10 x 10  -- see Dudeney below for these numbers, but the two values in  ( )  are not given by Dudeney.  He states results for the  9 x 9  and  10 x 10,  citing Lucas.  Says that a  6 x 6  version "is sold in the streets of London for a penny".

Hoffmann.  1893. 

     Chap. VI, pp. 272‑273 & 286 = Hoffmann-Hordern, pp. 187-189, with photo.

No. 24: No two in a row.  Eight queens.  Photo on p. 188 shows Jeu des Sentinelles, by Watilliaux, dated 1874-1895.

No. 25: The "Simple" Puzzle.  Nine queens.  Says a version was sold by Messrs. Feltham, with a notched board but the pieces were allowed to move over the gaps, so it was really a  9 x 9  board.

     Chap. X, No. 18: The Treasure at Medinet, pp. 343‑344 & 381 = Hoffmann-Hordern, pp. 237-239.  This is a solution of the eight queens problem, cut into four quadrants and jumbled.  The goal is to reconstruct the solution.  Photo on p. 239 shows Jeu des Manifestants, with box. 

                    Hordern Collection, p. 94, and S&B, p. 37, show a version of this with same box, but which divides the board into eight  2 x 4  rectangles.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 1: The famous Italian pin puzzle.  6  queens puzzle.  No solution.

Lucas.  L'Arithmétique Amusante.  1895.  Note IV:  Section I: Les huit dames, pp. 210-220.  Asserts Bellavitis was the first to find all solutions.  Discusses symmetries and shows the 12 basic solutions.  Correctly describes Jaenisch as obscure.  Gives an easy solution of Cretaine's problem which can be remembered as a trick.  Shows there are six solutions which can be superimposed with no overlap, i.e. six solutions using disjoint sets of cells.

C. D. Locock, conductor.  Chess Column.  Knowledge 19  (Jan 1896) 23-24;  (Feb 1896) 47‑48;  (May 1896) 119;  (Jul 1896) 167-168.  This series begins by saying most players know there is a solution, "but, possibly, some may be surprised to learn that there are ninety-two ways of performing the feat, ...."  He then enumerates them.  Second article studies various properties of the solutions, particularly looking for examples where one solution shifts to produce another one.  Third article notes some readers' comments.  Fourth article is a long communication from W. J. Ashdown about the number of distinct solutions, which he gets as  24  rather than the usual  12.

T. B. Sprague.  Proc. Edinburgh Math. Soc. 17 (1898-9) 43-68.  ??NYS -- cited by Ball, MRE, 4th ed., 1905, as giving the 341 inequivalent solutions on the  11 x 11.

Benson.  1904.  Pins and dots puzzle, p. 253.  6 queens problem, one solution.

Ball.  MRE, 4th ed., 1905.  The eight queens problem, pp. 114-120.  Corrects some history by citing MUS, 1st ed., 1901.  Gives one instance of Glaisher's method -- going from  4 x 4  to  5 x 5  and its results going up to  8 x 8.  Says the 92 inequivalent solutions on the  10 x 10  were given by Pein and the 341 inequivalent solutions on the  11 x 11  were given by Sprague.  The 5th ed., pp. 113-119 calls it "One of the classical problems connected with a chess-board" and adds examples of solutions up to  21 x 21  due to Mr. Derington.

Pearson.  1907.  Part III, no. 59: Stray dots, pp. 59 & 130.  Same as Hoffmann's Treasure at Medinet.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  The eight provinces, pp. 14-15 & 65.  Same as Hoffmann's Treasure at Medinet.

A. C. White.  Sam Loyd and His Chess Problems.  1913.  Op. cit. in 1.  P. 101 says Loyd discovered that all solutions have a piece at  d1  or equivalent.

Williams.  Home Entertainments.  1914.  A draughtboard puzzle, p. 115.  "Arrange eight men on a draughtboard in such a way that no two are upon the same line in any direction."  This is not well stated!!  Gives one solution:  52468317  and says  "Work out other solutions for yourself."

Dudeney.  AM.  1917.  The guarded chessboard, pp. 95‑96.  Gives the number of ways of placing n queens and the number of inequivalent ways.  The values in  ( )  are given by Ball, but not by Dudeney.

 

            n                          4      5    6      7      8            9          10      11            12            13

          ways                       2    10    4    40    92      (352)      (724)                        -               -

inequivalent ways               1      2    1      6    12          46          92     341      (1766)      (1346)

 

Ball.  MRE, 9th ed., 1920.  The eight queens problem, pp. 113-119.  Omits references to Pein and Sprague and adds the number of inequivalent solutions for the  12 x 12  and  13 x 13.

Blyth.  Match-Stick Magic.  1921.  No pairs allowed, p. 74.  6 queens problem.

Hummerston.  Fun, Mirth & Mystery.  1924.  No two in a line, p. 48.  Chessboard.  Place 'so that no two are upon the same line in any direction along straight or diagonal lines?'  Gives one solution:  47531682,  'but there are hundreds of other ways'.  You can let someone place the first piece.

Rohrbough.  Puzzle Craft.  1932.  Houdini Puzzle, p. 17.  6 x 6  case.

Rohrbough.  Brain Resters and Testers.  c1935.  Houdini Puzzle, p. 25.  6 x 6  problem.  "-- From New York World some years ago, credited to Harry Houdini."  I have never seen this attribution elsewhere.

Pál Révész.  Mathematik auf dem Schachbrett.  In:  Endre Hódi, ed. Mathematisches Mosaik.  (As:  Matematikai Érdekességek; Gondolat, Budapest, 1969.)  Translated by Günther Eisenreich.  Urania‑Verlag, Leipzig, 1977.  Pp. 20‑27.  On p. 24, he says that all solutions have 4 queens on white and 4 on black.  He says that one can place at most 5 non‑attacking queens on one colour.

Doubleday - 2.  1971.  Too easy?, pp. 97-98.  The two solutions on the  4 x 4  board are disjoint.

Dean S. Clark & Oved Shisha.  Proof without words: Inductive construction of an infinite chessboard with maximal placement of nonattacking queens.  MM 61:2 (1988) 98.  Consider a  5 x 5  board with queens in cells  (1,1), (2,4), (3,2), (4,5), (5,3).  5 such boards can be similarly placed within a  25 x 25  board viewed as a  5 x 5  array of  5 x 5  boards and this has no queens attacking.  Repeating the inflationary process gives a solution on the board of edge  53,  then the board of edge  54,  ....  They cite their paper:  Invulnerable queens on an infinite chessboard; Annals of the NY Acad. of Sci.: Third Intern. Conf. on Comb. Math.; to appear.  ??NYS.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Squares before your eyes, pp. 21 & 106.  Asks for solutions of the eight queens problem with no piece on either main diagonal.  Two of the 12 basic solutions have this, but one of these is the symmetric case, so there are 12 solutions of this problem.

Donald E. Knuth.  Dancing links.  25pp preprint of a talk given at Oxford in Sep 1999, sent by the author.  See the discussion in 6.F.  He finds the following numbers of solutions for placing  n  queens,  n = 1, 2, ..., 18.

          1,  0,  0,  2,  10,    4,  40,  92,  352,  724,    2680,  14200,  73712,  3 65596,  22 79184,  147 72512,  958 15104,  6660 90624.

 

          5.I.2.  COLOURING CHESSBOARD WITH NO REPEATS IN A LINE

 

          New section.  I know there is a general result that an  n x n  board can be  n‑coloured if  n  satisfies some condition like  º 1 or 5 (mod 6),  but I don't recall any other old examples of the problem.

 

Dudeney.  Problem 50: A problem in mosaics.  Tit‑Bits 32 (11 Sep 1897) 439  &  33 (2 Oct 1897) 3.  An  8 x 8  board with two adjacent corners omitted can be 8‑coloured with no two in a row, column or diagonal.  = Anon. & Dudeney; A chat with the Puzzle King; The Captain 2 (Dec? 1899) 314-320;  2:6 (Mar 1900) 598-599  &  3:1 (Apr 1900) 89.

Dudeney.  AM.  1917.  Prob. 302: A problem in mosaics, pp. 90 & 215-216.  The solution to the previous problem is given and then it is asked to relay the tiles so that the omitted squares are the  (3,3)  and  (3,6)  cells.

Hummerston.  Fun, Mirth & Mystery.  1924.  Q.E.D. -- The office boy problem, Puzzle no. 30, pp. 82 & 176.  Wants to mark the cells of a  4 x 4  board with no two the same in any 'straight line ..., either horizontally, vertically, or diagonally.'  His answer is:  ABCD,  CDEA,  EABC,  BCDA,  which has no two the same on any short diagonal.  The problem uses coins of values:  A, B, C, D, E  =  12, 30, 120, 24, 6  and the object is to maximize the total value of the arrangement.  In fact, there are only two ways to 5-colour the board and they are mirror images.  Four colours are used three times and one is used four times -- setting the value 120 on the latter cells gives the maximum value of 696.

 

          5.J.    SQUARED SQUARES, ETC.

 

          NOTE.  Perfect means no two squares are the same size.  Compound means there is a squared subrectangle.  Simple means not compound.

 

Dudeney.  Puzzling Times at Solvamhall Castle: Lady Isabel's casket.  London Mag. 7 (No. 42) (Jan 1902) 584  &  8 (No. 43) (Feb 1902) 56.  = CP, prob. 40, pp. 67 & 191‑193.  Square into  12  unequal squares and a rectangle.

Max Dehn.  Über die Zerlegung von Rechtecken in Rechtecke.  Math. Annalen 57 (1903) 314‑332.  Long and technical.  No examples.  Shows sides must be parallel and commensurable.

Loyd.  The patch quilt puzzle.  Cyclopedia, 1914, pp. 39 & 344.  = MPSL1, prob. 76, pp. 73 & 147‑148.  c= SLAHP: Building a patchquilt, pp. 30 & 92.  13 x 13  into  11  squares, not simple nor perfect.  (Gardner, in 536, says this appeared in Loyd's "Our Puzzle Magazine", issue 1 (1907), ??NYS.)

Loyd.  The darktown patch quilt party.  Cyclopedia, 1914, pp. 65 & 347.  12 x 12  into  11  squares, not simple nor perfect, in two ways.

P. J. Federico.  Squaring rectangles and squares -- A historical review with annotated bibliography.  In:  Graph Theory and Related Topics; ed. by J. A. Bondy & U. S. R. Murty; Academic Press, NY, 1979, pp. 173‑196.  Pp. 189‑190 give the background to Moroń's work.  Moroń later found the first example of Sprague but did not publish it.

Z. Moroń.  O rozkładach prostokątów na kwadraty (In Polish) (On the dissection of a rectangle into squares).  Przegląd Matematyczno‑Fizyczny (Warsaw) 3 (1925) 152‑153.  Decomposes rectangles into 9 and 10 unequal squares.  (Translation provided by A. Mąkowski, 1p.  Translation also available from M. Goldberg, ??NYS.)

M. Kraitchik.  La Mathématique des Jeux, 1930, op. cit. in 4.A.2, p. 272.  Gives Loyd's "Patch quilt puzzle" solution and Lusin's opinion that there is no perfect solution.

A. Schoenflies.  Einführung in der analytische Geometrie der Ebene und des Raumes.  2nd ed., revised and extended by M. Dehn, Springer, Berlin, 1931.  Appendix VI: Ungelöste Probleme der Analytischen Geometrie, pp. 402‑411.  Same results as in Dehn's 1903 paper.

Michio Abe.  On the problem to cover simply and without gap the inside of a square with a finite number of squares which are all different from one another (in Japanese).  Proc. Phys.‑Math. Soc. Japan 4 (1931) 359‑366.  ??NYS

Michio Abe.  Same title (in English).  Ibid. (3) 14 (1932) 385‑387.  Gives  191 x 195  rectangle into  11  squares.  Shows there are squared rectangles arbitrarily close to squares.

Alfred Stöhr.  Über Zerlegung von Rechtecken in inkongruente Quadrate.  Schr. Math. Inst. und Inst. angew. Math. Univ. Berlin 4:5 (1939), Teubner, Leipzig, pp. 119‑140.  ??NYR.  (This was his dissertation at the Univ. of Berlin.)

S. Chowla.  The division of a rectangle into unequal squares.  Math. Student 7 (1939) 69‑70.  Reconstructs Moroń's  9  square decomposition.

Minutes of the 203rd Meeting of the Trinity Mathematical Society (Cambridge) (13 Mar 1939).  Minute Books, vol. III, pp. 244‑246.  Minutes of A. Stone's lecture:  "Squaring the Square".  Announces Brooks's example with  39  elements, side  4639,  but containing a perfect subrectangle.

Minutes of the 204th Meeting of the Trinity Mathematical Society (Cambridge) (24 Apr 1939).  Minute Books, vol. III, p. 248.  Announcement by C. A. B. Smith that Tutte had found a perfect squared square with no perfect subrectangle.

R. Sprague.  Recreation in Mathematics.  Op. cit. in 4.A.1.  1963.  The expanded foreword of the English edition adds comments on Dudeney's "Lady Isabel's Casket", which led to the following paper.

R. Sprague.  Beispiel einer Zerlegung des Quadrats in lauter verschiedene Quadrate.  Math. Zeitschr. 45 (1939) 607‑608.  First perfect squared square --  55  elements, side  4205.

R. Sprague.  Zur Abschätzung der Mindestzahl inkongruenter Quadrate, die ein gegebenes Rechteck ausfüllen.  Math. Zeitschrift 46 (1940) 460‑471.  Tutte's 1979 commentary says this shows every rectangle with commensurable sides can be dissected into unequal squares.

A. H. Stone, proposer;  M. Goldberg & W. T. Tutte, solvers.  Problem E401.  AMM 47:1 (Jan 1940) 48  &  AMM 47:8 (Oct 1940) 570‑572.  Perfect squared square --  28  elements, side  1015.

R. L. Brooks,  C. A. B. Smith,  A. H. Stone  &  W. T. Tutte.  The dissection of rectangles into squares.  Duke Math. J. 7 (1940) 312‑340.  = Selected Papers of W. T. Tutte; Charles Babbage Research Institute, St. Pierre, Manitoba, 1979; pp. 10-38, with commentary by Tutte on pp. 1-9.  Tutte's 1979 commentary says Smith was perplexed by the solution of Dudeney's "Lady Isabel's Casket" -- see also his 1958 article.

A. H. Stone, proposer;  Michael Goldberg, solver.  Problem E476.  AMM 48 (1941) 405 ??NYS  &  49 (1942) 198-199.  An isosceles right triangle can be dissected into  6  similar figures, all of different sizes.  Editorial notes say that Douglas and Starke found a different solution and that one can replace  6  by any larger number, but it is not known if  6  is the least such.  Stone asks if there is any solution where the smaller triangles have no common sides.

M. Kraitchik.  Mathematical Recreations, op. cit. in 4.A.2, 1943.  P. 198.  Shows the compound perfect squared square with  26  elements and side  608  from Brooks, et al.

C. J. Bouwkamp.  On the construction of simple perfect squared squares.  Konink. Neder. Akad. van Wetensch. Proc. 50 (1947) 72-78  =  Indag. Math. 9 (1947) 57-63.  This criticised the method of Brooks, Smith, Stone & Tutte, but was later retracted.

Brooks, Smith, Stone & Tutte.  A simple perfect square.  Konink. Neder. Akad. van Wetensch. Proc. 50 (1947) 1300‑1301.  = Selected Papers of W. T. Tutte; Charles Babbage Research Institute, St. Pierre, Manitoba, 1979; pp. 99-100, with commentary by Tutte on p. 98.  Bouwkamp had published several notes and was unable to make the authors' 1940 method work.  Here they clarify the situation and give an example.  One writer said they give details of Sprague's first example, but the example is not described as being the same as in Sprague.

W. T. Tutte.  The dissection of equilateral triangles into equilateral triangles.  Proc. Camb. Phil Soc. 44 (1948) 464‑482.  = Selected Papers of W. T. Tutte; Charles Babbage Research Institute, St. Pierre, Manitoba, 1979; pp. 106-125, with commentary by Tutte on pp. 101-105. 

T. H. Willcocks, proposer and solver.  Problem 7795.  Fairy Chess Review 7:1 (Aug 1948) 97 & 106 (misnumberings for 5 & 14).  Refers to prob. 7523 -- ??NYS.  Finds compound perfect squares of orders  27, 27, 28 and 24.

T. H. Willcocks.  A note on some perfect squares.  Canadian J. Math. 3 (1951) 304‑308.  Describes the result in Fairy Chess Review prob. 7795.

T. H. Willcocks.  Fairy Chess Review (Feb & Jun 1951).  Prob. 8972.  ??NYS -- cited and described by G. P. Jelliss; Prob. 44 -- A double squaring, G&PJ 2 (No. 17) (Oct 1999) 318-319.  Squares of edges  3, 5, 9, 11, 14, 19, 20, 24, 31, 33, 36, 39, 42  can be formed into a  75 x 112  rectangle in two different ways.  {These are reproduced, without attribution, as Fig. 21, p. 33 of Joseph S. Madachy; Madachy's Mathematical Recreations; Dover, 1979 (this is a corrected reprint of Mathematics on Vacation, 1966, ??NYS).  The 1979 ed. has an errata slip inserted for p. 33 as the description of Fig. 21 was omitted in the text, but the erratum doesn't cite a source for the result.}  The G&PJ problem then poses a new problem from Willcocks involving  21  squares to be made into a rectangle in two different ways -- it is not clear if these have to be the same shape.

M. Goldberg.  The squaring of developable surfaces.  SM 18 (1952) 17‑24.  Squares cylinder, Möbius strip, cone.

W. T. Tutte.  Squaring the square.  Guest column for SA (Nov 1958).  c= Gardner's 2nd Book, pp. 186‑209.  The latter  =  Selected Papers of W. T. Tutte; Charles Babbage Research Institute, St. Pierre, Manitoba, 1979; pp. 244-266, with a note by Tutte on p. 244, but the references have been omitted.  Historical account -- cites Dudeney as the original inspiration of Smith.

R. L. Hutchings  &  J. D. Blake.  Problems drive 1962.  Eureka 25 (Oct 1962) 20-21 & 34-35.  Prob. G.  Assemble squares of sides  2, 5, 7, 9, 16, 25, 28, 33, 36  into a rectangle.  The rectangle is  69 x 61  and is not either of Moroń's examples. 

W. T. Tutte.  The quest of the perfect square.  AMM 72:2, part II (Feb 1965) 29-35.  = Selected Papers of W. T. Tutte; Charles Babbage Research Institute, St. Pierre, Manitoba, 1979; pp. 432-438, with brief commentary by Tutte on p. 431.  General survey, updating his 1958 survey.

Blanche Descartes [pseud. of Cedric A. B. Smith].  Division of a square into rectangles.  Eureka 34 (1971) 31-35.  Surveys some history and Stone's dissection of an isosceles right triangle into  6  others of different sizes (see above).  Tutte has a dissection of an equilateral triangle into  15  equilateral triangles -- but some of the pieces must have the same area so we consider up and down pointing triangles as + and - areas and then all the areas are different.  Author then considers dissecting a square into incongruent but equiareal rectangles.  He finds it can be done in  n  pieces for any  n ³ 7.

A. J. W. Duijvestijn.  Simple perfect squared square of lowest order.  J. Combinatorial Thy. B 25 (1978) 240‑243.  Finds a perfect square of minimal order  21.

A. J. W. Duijvestin, P. J. Federico & P. Leeuw.  Compound perfect squares.  AMM 89 (1982) 15‑32.  Shows Willcocks' example has the smallest order for a compound perfect square and is the only example of its order,  24.

 

          5.J.1. MRS PERKINS'S QUILT

 

This is the problem of cutting a square into smaller squares.

 

Loyd.  Cyclopedia, 1914, pp. 248 & 372,  307 & 380.  Cut  3 x 3  into 6 squares:  2 x 2  and  5  1 x 1.

Dudeney.  AM.  1917.  Prob. 173: Mrs Perkins's quilt, pp. 47 & 180.  Same as Loyd's "Patch quilt puzzle" in 5.J.

Dudeney.  PCP.  1932.  Prob. 117: Square of Squares, pp. 53 & 148‑149.  = 536, prob. 343, pp. 120 & 324‑325.  c= "Mrs Perkins's quilt".

N. J. Fine & I. Niven, proposers;  F. Herzog, solver.  Problem E724 -- Admissible Numbers.  AMM 53 (1946) 271  &  54 (1947) 41‑42.  Cubical version.

J. H. Conway.  Mrs Perkins's quilt.  Proc. Camb. Phil. Soc. 60 (1964) 363‑368.

G. B. Trustrum.  Mrs Perkins's quilt.  Ibid. 61 (1965) 7‑11.

Ripley's Puzzles and Games.  1966.  Pp. 16-17, item 7.  "Can you divide a square into 6 perfect squares?"  Answer as in Loyd.

Nick Lord.  Note 72.11:  Subdividing hypercubes.  MG 72 (No. 459) (Mar 1988) 47‑48.  Gives an upper bound for impossible numbers in  d  dimensions.

David Tall.  To prove or not to prove.  Mathematics Review 1:3 (Jan 1991) 29-32.  Tall regularly uses the question as an exercise in problem solving.  About ten years earlier, a 14 year old girl pointed out that the problem doesn't clearly rule out rejoining pieces.  E.g. by cutting along the diagonals and rejoining, one can make two squares.

 

          5.J.2. CUBING THE CUBE

 

S. Chowla.  Problem 1779.  Math. Student 7 (1939) 80.  (Solution given in Brooks, et al., Duke Math. J., op. cit. in 5.J, section 10.4, but they give no reference to a solution in Math. Student.)

 

          5.J.3. TILING A SQUARE OF SIDE 70 WITH SQUARES OF SIDES  1, 2, ..., 24

 

J. R. Bitner.  Use of Macros in Backtrack Programming.  M.Sc. Thesis, ref. UIUCDCS‑R‑74‑687, Univ. of Illinois, Urbana‑Champaign, 1974, ??NYS.  Shows such a tiling is impossible.

 

          5.K.   DERANGEMENTS

 

          Let  D(n) = the number of derangements of  n  things,  i.e. permutations leaving no point fixed.

 

Eberhard Knobloch.  Euler and the history of a problem in probability theory.  Gaņita-Bhāratī [NOTE:  ņ  denotes an  n  with an underdot] (Bull. Ind. Soc. Hist. Math.) 6 (1984) 1‑12.  Discusses the history, noting that many 19C authors were unaware of Euler's work.  There is some ambiguity in his descriptions due to early confusion of  n  as the number of cards and  n  as the number of the card on which a match first occurs.  Describes numerous others who worked on the problem up to about 1900:  De Moivre, Waring, Lambert, Laplace, Cantor, etc.

 

Pierre Rémond de Montmort.  Essai d'analyse sur les jeux de hazards.  (1708);  Seconde edition revue & augmentee de plusieurs lettres,  (Quillau, Paris, 1713 (reprinted by Chelsea, NY, 1980));  2nd issue, Jombert & Quillau, 1714.  Problèmes divers sur le jeu du trieze, pp. 54‑64.  In the original game, one has a deck of 52 cards and counts  1, 2, ..., 13  as one turns over the cards.  If a card of rank  i  occurs at the  i-th count, then the player wins.  In general, one simplifies by assuming there are  n  distinct cards numbered  1, ..., n  and one counts  1, ..., n.  One can ask for the probability of winning at some time and of winning at the  k-th draw.  In 1708, Montmort already gives tables of the number of permutations of  n  cards such that one wins on the  k-th draw, for  n = 1, ..., 6.  He gives various recurrences and the series expression for the probability and (more or less) finds its limit.  In the 2nd ed., he gives a proof of the series expression, due to Nicholas Bernoulli, and John Bernoulli says he has found it also.  Nicholas' solution covers the general case with repeated cards.  [See:  F. N. David; Games, Gods and Gambling; Griffin, London, 1962, pp. 144‑146 & 157.]  (Comtet and David say it is in the 1708 ed.  I have seen it on pp. 54-64 of an edition which is uncertain, but probably 1708, ??NX.  Knobloch cites 1713, pp. 130-143, but adds that Montmort gave the results without proofs in the 1708 ed. and includes several letters from and to John I and Nicholas I Bernoulli in the 1713 ed., pp. 290-324, and mentions the problem in his Preface -- ??NYS.)

Abraham de Moivre.  The Doctrine of Chances: or, A Method of Calculating the Probability of Events in Play.  W. Pearson for the Author, London, 1718.  Prob. XXV, pp. 59-63.  (= 2nd ed, H. Woodfall for the Author, London, 1738.  Prob. XXXIV, pp. 95-98.)  States and demonstrates the formula for finding the probability of  p  items to be correct and  q  items to be incorrect out of  n  items.  One of his examples is the probability of six items being deranged being  53/144.

L. Euler.  Calcul de la probabilité dans le jeu de rencontre.  Mémoires de l'Académie des Sciences de Berlin (7) (1751(1753)) 255‑270.  = Opera Omnia (1) 7 (1923) 11‑25.  Obtains the series for the probability and notes it approaches  1/e.

L. Euler.  Fragmenta ex Adversariis Mathematicis Deprompta.  MS of 1750‑1755.  Pp. 287‑288: Problema de permutationibus.  First published in Opera Omnia (1) 7 (1923) 542‑545.  Obtains alternating series for  D(n).

Ozanam-Montucla.  1778.  Prob. 5, 1778: 125-126;  1803: 123-124;  1814: 108-109;  1840: omitted.  Describes Jeu du Treize, where a person takes a whole deck and turns up the cards, counting  1, 2, ..., 13  as he goes.  He wins if a card of rank  i  appears at the  i‑th count.  Montucla's description is brief and indicates there are several variations of the game.  Hutton gives a lengthier description of one version.  Cites Montmort for the probability of winning as  .632..

L. Euler.  Solutio quaestionis curiosae ex doctrina combinationum.  (Mem. Acad. Sci. St. Pétersbourg 3 (1809/10(1811)) 57‑64.)  = Opera Omnia (1) 7 (1923) 435‑440.  (This was presented to the Acad. on 18 Oct 1779.)  Shows  D(n) = (n‑1) [D(n‑1) + D(n‑2)]  and  D(n) = nD(n‑1) + (‑1)n.

Ball.  MRE.  1st ed., 1892.  Pp. 106-107: The mousetrap and Treize.  In the first, one puts out  n  cards in a circle and counts out.  If the count  k  occurs on the  k-th card, the card is removed and one starts again.  Says Cayley and Steen have studied this.  It looks a bit like a derangement question. 

Bill Severn.  Packs of Fun.  101 Unusual Things to Do with Playing Cards and To Know about Them.  David McKay, NY, 1967.  P. 24: Games for One: Up and down.  Using a deck of 52 cards, count through  1, 2, ..., 13  four times.  You lose if a card of rank  i  appears when you count  i, i.e. you win if the cards are a generalized derangement.  Though a natural extension of the problem, I can't recall seeing it treated, perhaps because it seems to get very messy.  However, a quick investigation reveals that the probability of such a generalized derangement should approach  e-4.

Brian R. Stonebridge.  Derangements of a multiset.  Bull. Inst. Math. Appl. 28:3 (Mar 1992) 47-49.  Gets a reasonable extension to multisets, i.e. sets with repeated elements.

 

          5.K.1.          DERANGED BOXES OF  A,  B  AND  A & B

 

          Three boxes contain  A  or  B  or  A & B,  but they have been shifted about so each is in one of the other boxes.  You can look at one item from one box to determine what is in all of them.  This is just added and is certainly older than the examples below.

 

Simon Dresner.  Science World Book of Brain Teasers.  1962.  Op. cit. in 5.B.1.  Prob. 84: Marble garble, pp. 40 & 110.  Black and white marbles.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.  No. 26: Mexican jumping beans, pp. 40-41 & 96.  Red and black beans in matchboxes.  The problem continues with a Bertrand box paradox -- see 8.H.1.

Doubleday - 3.  1972.  Open the box, pp. 147-148.  Black and white marbles.

 

          5.K.2.          OTHER LOGIC PUZZLES BASED ON DERANGEMENTS

 

          These typically involve a butcher, a baker and a brewer whose surnames are Butcher, Baker and Brewer, but no one has the profession of his name.  I generally only state the beginning of the problem.

          New section -- there must be older examples.  Gardner, in an article:  My ten favorite brainteasers in Games (collected in Games Big Book of Games, 1984, pp. 130-131) says this is one of his favorite problems.  ??locate

          I now see these lead to Latin rectangles, cf Section 5.I.

 

R. Turner, proposer: The sons of the dons; Eureka 2 (May 1939) 9-10.  K. Tweedie, solver: On the problem of the sons of the dons.  Eureka 4 (May 1940) 21-23.  Six dons, in analysis, geometry, algebra, dynamics, physics and astronomy, each have a son who studies one of these subjects, but none studies the subject of his father.  Several further restrictions, e.g., there are no two students who each study the subject of the other's father.

M. Adams.  Puzzle Book.  1939.  Prob. B.91: Easter bonnet, pp. 80 & 107.  Women named Green, Black, Brown and White with 4 colours of hats and 4 colours of dresses, but name, hat and dress are always distinct.

J. B. Parker.  Round the table.  Eureka  5 (Jan 1941) 20-21  &  6 (May 1941) 11.  Seven men, whose names are colours, with ties, socks and cars, being coloured with three of the names of other men and all colours used for each item, sitting at a table with eight places. 

Anonymous.  The umbrella problem.  Eureka  9 (Apr 1947) 22  &  10 (Mar 1948) 25.  Six men 'of negligible honesty' each go away with another's umbrella. 

Jonathan Always.  Puzzles to Puzzle You.  Tandem, London, 1965.  No. 30: Something about ties, pp. 16 & 74-75.  Black, Green and Brown are wearing ties, but none has the colour of his name, remarked the green tie wearer to Mr. Black.

David Singmaster.  The deranged secretary.  If a secretary puts  n  letters all in wrong envelopes, how many envelopes must one open before one knows what is each of the unopened envelopes?

Problem proposal and solution 71.B.  MG 71 (No. 455) (Mar 1987) 65  &  71 (No. 457) (Oct 1987) 238-239.

Open question.  The Weekend Telegraph (11 Jun 1988) XV  &  (18 Jun 1988) XV. 

 

          5.K.3.          CAYLEY'S MOUSETRAP

 

          This is a solitaire (= patience) game developed by Cayley, based on Treize.  Take a deck of cards, numbered  1, 2, ..., n,  and shuffle them.  Count through them.  If a card does not match its count, put it on bottom and continue.  If it matches, set it aside and start counting again from  1.  One wins if all cards are set aside.  In this case, pick up the deck and start a new game.

 

T. W. O. Richards, proposer;  Richard I. Hess, solver.  Prob. 1828.  CM 19 (1993) 78  &  20 (1994) 77-78.  Asks whether there is any arrangement which allows three or more consecutive wins.  No theoretical solution.  Searching finds one solution for  n = 6  and  n = 8  and  8  solutions for  n = 9.

 

          5.L.    MÉNAGE PROBLEM

 

          How many ways can  n  couples be seated, alternating sexes, with no couples adjacent?

 

A. Cayley.  On a problem of arrangements.  Proc. Roy. Soc. Edin. 9 (1878) 338‑342.  Problem raised by Tait.  Uses inclusion/exclusion to get a closed sum.

T. Muir.  On Professor Tait's problem of arrangements.  Ibid., 382‑387.  Uses determinants to get a simple  n‑term recurrence.

A. Cayley.  Note on Mr. Muir's solution of a problem of arrangement.  Ibid., 388‑391.  Uses generating function to simplify to a usable form.

T. Muir.  Additional note on a problem of arrangement.  Ibid., 11 (1882) 187‑190.  Obtains Laisant's 2nd order and 4th order recurrences.

É. Lucas.  Théorie des Nombres.  Gauthier‑Villars, Paris, 1891;  reprinted by Blanchard, Paris, 1958.  Section 123, example II, p. 215  &  Note III, pp. 491‑495.  Lucas appears not to have known of the work of Cayley and Muir.  He describes Laisant's results.  The 2nd order, non‑homogeneous recurrence, on pp. 494‑495, is attributed to Moreau.

C. Laisant.  Sur deux problèmes de permutations.  Bull. Soc. Math. de France 19 (1890‑91) 105‑108.  General approach to problems of restricted occupancy.  His work yields a 2nd order non-homogeneous recurrence and homogeneous 3rd and 4th order recurrences.  He cites Lucas, but says Moreau's work is unpublished.

H. M. Taylor.  A problem on arrangements.  Messenger of Math. 32 (1903) 60‑63.  Gets almost to Muir & Laisant's 4th order recurrence.

J. Touchard.  Sur un problème de permutations.  C. R. Acad. Sci. Paris 198 (1934) 631‑633.  Solution in terms of a complicated integral.  States the explicit summation.

I. Kaplansky.  Solution of the "problème des ménages".  Bull. Amer. Math. Soc. 49 (1943) 784‑785.  Obtains the now usual explicit summation.

I. Kaplansky & J. Riordan.  The problème de ménages.  SM 12 (1946) 113‑124.  Gives the history and a uniform approach.

J. Touchard.  Permutations discordant with two given permutations.  SM 19 (1953) 109‑119.  Says he prepared a 65pp MS developing the results announced in 1934 and rediscovered in Kaplansky and in Kaplansky & Riordan.  Proves Kaplansky's lemma on selections by finding the generating functions which involve Chebyshev polynomials.  Obtains the explicit summation, as done by Kaplansky.  Extends to more general problems.

M. Wyman & L. Moser.  On the 'problème des ménages'.  Canadian J. Math. 10 (1958) 468‑480.  Analytic study.  Updates the history -- 26 references.  Gives table of values for  n = 0 (1) 65.

Jacques Dutka.  On the 'Problème des ménages'.  Math. Intell. 8:3 (1986) 18‑25 & 33.  Thorough survey & history -- 25 references.

Kenneth P. Bogart & Peter G. Doyle.  Non‑sexist solution of the ménage problem.  AMM 93 (1986) 514‑518.  14 references.

 

 

 

          5.M.   SIX PEOPLE AT A PARTY  --  RAMSEY THEORY

 

 

 

          In a group of six people, there is a triple who all know each other or there is a triple who are all strangers.  I.e., the Ramsey number  R(3,3) = 6.  I will not go into the more complex aspects of this -- see Graham & Spencer for a survey.

 

P. Erdös & G. Szekeres.  A combinatorial problem in geometry.  Compositio Math. 2 (1935) 463‑470.  [= Paul Erdös; The Art of Counting; Ed. by Joel Spencer, MIT Press, 1973, pp. 5‑12.]  They prove that if  n ³ BC(a+b-2, a-1),  then any two‑colouring of  Ka  contains a monochromatic  Ka  or  Kb.

William Lowell Putnam Examination, 1953, part I, problem 2.  In:  L. E. Bush; The William Lowell Putnam Mathematical Competition; AMM 60 (1953) 539-542.  Reprinted in:  A. M. Gleason, R. E. Greenwood & L. M. Kelly; The William Lowell Putnam Mathematical Competition Problems and Solutions -- 1938‑1964; MAA, 1980; pp. 38 & 365‑366.  The classic six people at a party problem.

R. E. Greenwood & A. M. Gleason.  Combinatorial relations and chromatic graphs.  Canadian J. Math. 7 (1955) 1-7.  Considers  n = n(a,b,...)  such that a two colouring of  Kn  contains a  Ka  of the first colour or a  Kb  of the second colour or ....  Thus  n(3,3) = 6.  They find the bound and many other results of Erdös & Szekeres.

C. W. Bostwick, proposer;  John Rainwater & J. D. Baum, solvers.  Problem E1321 -- A gathering of six people.  AMM 65 (1958) 446  &  66 (1959) 141‑142.

Gamow & Stern.  1958.  Diagonal strings.  Pp. 93‑95.

G. J. Simmons.  The Game of Sim.  JRM 2 (1969) 66.

M. Gardner.  SA (Jan 1973) c= Knotted, chap. 9.  Exposits Sim.  Reports Simmons' result that it is second person (determined after his 1969 article above).  The Addendum in Knotted reports that several people have shown that Sim on five points is a draw.  Numerous references.

Ronald L. Graham & Joel H. Spencer.  Ramsey theory.  SA 263:1 (Jul 1990) 80‑85.  Popular survey of Ramsey theory beginning from Ramsey and Erdös & Szekeres.

 

          5.N.   JEEP OR EXPLORER'S PROBLEM

 

          See Ball for some general discussion and notation.

 

Alcuin.  9C.  Prob. 52: Propositio de homine patrefamilias.  Wants to get 90 measures over a distance of 30 leagues.  He is trying to get the most to the other side, so this is different than the 20C versions.  Solution is confusing, but Folkerts rectifies a misprint and this makes it less confusing.  Alcuin's camels only eat when loaded!!  (Or else they perish when their carrying is done??)  The camel take a load to a point 20 leagues away and leaves 10 there, then returns.  This results in getting 20 to the destination. 

                    The optimum solution is for the camel to make two return trips and a single trip to  10  leucas, so he will have consumed  30  measures and he has  60  measures to carry on.  He now makes one return and a single trip of another  15  leucas, so he will have consumed another  30  measures, leaving  30  to carry on the last  5  leucas, so he reaches home with  25  measures.

Pacioli.  De Viribus.  c1500.  Probs. 49‑52.  Agostini only describes Prob. 49 in some detail.

Ff. 94r - 95v.  XLIX. (Capitolo) de doi aportare pome ch' piu navanza (Of two ways to transport as many apples as possible).  = Peirani 134‑135.  One has 90 apples to transport 30 miles from Borgo [San Sepolcro] to Perosia [Perugia], but one eats one apple per mile and one can carry at most 30 apples.  He carries 30 apples 20 miles and leaves 10 there and returns, without eating on the return trip!  (So this  = Alcuin.)  Pacioli continues and gives the optimum solution!

F. 95v.  L. C(apitolo). de .3. navi per .30. gabelle  90. mesure (Of three ships holding 90 measures, passing 30 customs points).  Each ship has to pay one measure at each customs point.  Mathematically the same as the previous. 

F. 96r.  LI. C(apitolo). de portar .100. perle .10. miglia lontano  10. per volta et ogni miglio lascia 1a (To carry 100 pearls 10 miles, 10 at a time, leaving one every mile).  = Peirani 136-137.  Takes them 2 miles in ten trips, giving 80 there.  Then takes them to the destination in 8 trips, getting 16 to the destination.

Ff. 96v - 97r.  LII. C(apitolo). el medesimo con piu avanzo per altro modo (The same with more carried by another method).  Continues the previous problem and takes them 5 miles in ten trips, giving 50 there.  Then takes them to the destination in 5 trips, getting 25 to the destination. 

           [This is optimal for a single stop -- if one makes the stop at distance  a,  then one gets  a(10-a)  to the destination.  One can make more stops, but this is restricted by the fact that pearls cannot be divided.  Assuming that the amount of pearls accumulated at each depot is a multiple of ten, one can get 28 to the destination by using depots at  2 and 7  or  5 and 7.  One can get 27 to the destination with depots at  4 and 9  or  5 and 9.  These are all the ways one can put in two depots with integral multiples of 10 at each depot and none of these can be extended to three such depots.  If the material being transported was a continuous material like grain, then I think the optimal method is to first move 1 mile to get 90 there, then move another  10/9  to get 80 there, then another  10/8  to get 70 there, ..., continuing until we get 40 at  8.4563...,  and then make four trips to the destination.  This gets  33.8254  to the destination.  Is this the best method??]

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 57, ff. EE.vi.v - EE.vii.v (pp. 152‑153).  Complicated problem involving carrying food and material up the Tower of Babel!  Tower is assumed 36 miles high and seems to require 15625 porters.

Mittenzwey.  1880.  Prob. 135, pp. 28-29;  1895?: 153, p. 32;  1917: 153, pp. 29.  If eight porters can carry eight full loads from  A  to  B  in an hour, how long will it take four porters?  The obvious answer is two hours, but he observes that the porters have to return from  B  to  A  and it will take three hours.  [Probably a little less as they should return in less time than they go.]

Pearson.  1907.  Part II, pp. 139 & 216.  Two explorers who can carry food for 12 days.  (No depots, i.e. form  A  of Ball, below.)

Loyd.  A dash for the South Pole.  Ladies' Home Journal (15 Dec 1910).  ??NYS -- source?? -- WS??

Ball.  MRE, 5th ed., 1911.  Exploration problems, pp. 23‑24.  He distinguishes two forms of the problem, with  n  explorers who can carry food for  d  days.

                    A.  Without depots, they can get one man  nd/(n+1)  days into the desert and back.

                    B.  With depots permitted, they can get a man  d/2 (1/1 + 1/2 + ... + 1/n)  into the desert and back.  This is the more common form.

Dudeney.  Problem 744: Exploring the desert.  Strand Mag. (1925).  ??NX. (??= MP 49)

Dudeney.  MP.  1926.  Prob. 49: Exploring the desert, pp. 21 & 111 (= 536, prob. 76, pp. 22 & 240).  A version of Ball's form  A,  with  n = 9,  d = 10,  but replacing days by stages of length 40 miles.

Abraham.  1933.  Prob. 34 -- The explorers, pp. 13 & 25 (9‑10 & 112).  4 explorers, each carrying food for 5 days.  Mentions general case.  This is Ball's form  A.

Haldeman-Julius.  1937.  No. 10: The four explorers, pp. 4 & 21.  Ball's form A, with  n = 4, d = 5.

Olaf Helmer.  Problem in logistics: The Jeep problem.  Project Rand Report RA‑15015 (1 Dec 1946) 7pp.

N. J. Fine.  The jeep problem.  AMM 54 (1947) 24‑31.

C. G. Phipps.  The jeep problem: a more general solution.  AMM 54 (1947) 458‑462.

G. G. Alway.  Note 2707:  Crossing the desert.  MG 41 (No. 337) (1957) 209.  If a jeep can carry enough fuel to get halfway across, how much fuel is needed to get across?  For a desert of width  2,  this leads to the series  1 + 1/3 + 1/5 + 1/7 + ....  See Lehmann and Pyle below.

G. C. S[hephard, ed.]  The problems drive.  Eureka 11 (Jan 1949) 10-11 & 30.  No. 2.  Four explorers, starting from a supply base.  Each can carry food for 100 miles and goes 25 miles per day.  Two men do the returning to base and bringing out more supplies.  the third man does ferrying to the fourth man.  How far can the fourth man get into the desert and return?  Answer is 100 miles.  Ball's form B would give  104 1/6.

Gamow & Stern.  1958.  Refueling.  Pp. 114‑115.

Pyle, I. C.  The explorer's problem.  Eureka 21 (Oct 1958) 5-7.  Considers a lorry whose load of fuel takes it a distance which we assume as the unit.  What is the widest desert one can cross?  And how do you do it?  This is similar to Alway, above.  He starts at the far side and sees you have to have a load at distance 1 from the far side, then two loads at distance  1 + 1/3  from the far side,  then three loads at  1 + 1/3 + 1/5,  ....  This diverges, so any width can be crossed.  Does examples with given widths of 2, 3 and 4 units.  Editor notes that he is not convinced the method is optimal.

Martin Gardner, SA (May  &  June 1959) c= 2nd Book, chap. 14, prob. 1.  (The book gives extensive references which were not in SA.)

R. L. Goodstein.  Letter: Explorer's problem.  Eureka 22 (Oct 1959) 23.  Says Alway shows that Pyle's method is optimal.  Editor notes Gardner's article and that Eureka was cited in the solutions in Jun.

David Gale.  The jeep once more or jeeper by the dozen  &  Correction to "The Jeep once more or jeeper by the dozen".  AMM 77:5 (May 1970) 493-501  &  78:6 (Jun-Jul 1971) 644-645.  Gives an elaborate approach via a formula of Banach for path lengths in one dimension.  This formally proves that the various methods used are actually optimal and that a continuous string of depots cannot help, etc.  Notes that the cost for a round trip is only slightly more than for a one-way trip -- but the Correction points out that this is wrong and indeed the round trip is nearly four times as expensive as a one-way trip.  Considers sending several jeeps.  Says he hasn't been able to do the round trip problem when there is fuel on both sides of the desert.  Comments on use of dynamic programming, noting that R. E. Bellman [Dynamic Programming; Princeton Univ. Press, 1955, p. 103, ex. 54-55] gives the problem as exercises without solution and that he cannot see how to do it!

Birtwistle.  Math. Puzzles & Perplexities.  1971.

The expedition, pp. 124-125, 183 & 194.  Ball's form A, first with  n = 5,  d = 6,  then in general.

Second expedition, pp. 125-126.  Ball's form B, done in general.

Third expedition, pp. 126, 183-184 & 194.  Three men want to cross a 180 mile wide desert.  They can travel 20 miles per day and can carry food for six days, which can be stored at depots.  Minimize the total distance travelled.  Solution seems erroneous to me.

A. K. Austin.  Jeep trips and card stacks.  MTg 58 (1972) 24‑25.  There are  n  flags located at distances  a1,  a1 + a2,  a1 + a2 + a3,  ....  Jeep has to begin at the origin, go to the first flag, return to the origin, go to the second flag, return, ....  He can unload and load fuel at the flags.  Can he do this with  F  fuel?  Author shows this is equivalent to successfully stacking cards over a cliff with successive overhangs being  a1,  a2,  a3,  ....

Doubleday - 3.  1972.  Traveller's Tale, pp. 63-64.  d = 8  and we want one man to get across the desert of width  12.  How many porters, who return to base, are needed?  The solution implies that no depots are used.  Reasoning as in Ball's case  A,  we see that  n  men can support one man crossing a desert of width  2nd/(n+1).  If depots are permitted, this is essentially the jeep problem and  n  men can support a man getting across a desert of width  d [1 + 1/3 + 1/5 + ... + 1/(2n-1)]

Johannes Lehmann.  Kurzweil durch Mathe.  Urania Verlag, Leipzig, 1980.  No. 13, pp. 27 & 129.  d = 4  and we want to get a man across a desert of width  6.  Similar to Doubleday - 3.

Pierre Berloquin.  [Le Jardin du Sphinx.  Dunod, Paris, 1981.]  Translated by Charles Scribner Jr as:  The Garden of the Sphinx.  Scribner's, NY, 1985.

          Prob. 1: Water in the desert, pp. 3 & 85.

          Prob. 40: Less water in the desert, pp. 26 & 111.

          Prob. 80: Beyond thirst, pp. 48 & 140.

          Prob. 141: The barrier of thirst, pp. 79 & 181.

          Prob. 150: No holds barred, pp. 82 & 150.

          In all of these,  d = 5  and we want to get a man across a desert of width  4,  and sometimes back, which is slightly different than the problem of getting to the maximum distance and back.

                    Prob. 1 is Ball's form  A,  with  n = 4  men, using  20  days' water.

                    Prob. 40 is Ball's form  B,  but using only whole day trips, using  14  days' water.

                    Prob. 80 is Ball's form  B,  optimized for width  4, using  11½  days' water.

                    Prob. 141 uses depots and bearers who don't return, as in Alcuin??  You can get one man, who is the only one to return, a distance  d (1/2 + 1/3 + ... + 1/(n+1))  into the desert this way.  He gives the optimum form for width  4,  using    days' water.

                    Prob. 150 is like prob. 141, except that no one returns!  You can get him  d (1 + 1/2 + ... + 1/n)  into the desert this way.  The optimum here uses  4  days' water.

D. R. Westbrook.  Note 74.7:  The desert fox, a variation of the jeep problem.  MG 74 (No. 467) (1990) 49‑50.  A more complex version, posed by A. K. Dewdney in SA (Jan 1987), is solved here.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Round-trip, pp. 96-97 & 140.  Plane wants to circle the earth, but can only carry fuel to go half-way.  Other planes can accompany and transfer fuel, but must return to base.

Dylan Gow.  Flyaway.  MS 25:3 (1992/3) 84-86.  Considers the standard problem without return as in Alway, Pyle and Lehmann -- but finds a non-optimal solution.

Wolfram Hinderer.  Optimal crossing of a desert.  MS 26:4 (1993/4) 100-102.  Finds optimal solutions for Gow's problem and for the case with return -- i.e. Ball's  B.  Also considers use of extra jeeps that do not return, i.e. Berloquin's 141 & 150.  Notes that extra jeeps that must return to base do not change the distance that one jeep can reach.  [But it changes the time required.]

Harold Boas.  Letter:  Crossing deserts.  MS 26:4 (1993/4) 122.  Notes the problem has a long history and cites Fine, Phipps, Gale (and correction), Alway.

David Singmaster.  Letter:  Crossing deserts.  MS 27:3 (1994/5) 63.  Points out that the history is far older and sketches the history given above.

Günter Rote & Guochuan Zhang.  Optimal Logistics for Expeditions: The Jeep Problem with Complete Refilling.  Karl-Franzens-Universität Graz & Technische Universität Graz.  Bericht 71 (24 Jun 1996).  This deals with a variant.  "We have  n  cans of fuel on the edge of a desert and a jeep with an empty tank whose capacity is just one can.  The jeep can carry one can in addition to the fuel in its tank.  Moreover, when a can is opened, the fuel must immediately be filled into the jeep's tank.  The goal is to find the farthest point in the desert which the jeep can reach by consuming the  n  cans of fuel.  Derick Wood [1984] treated this problem similarly to the classical problem and gave the first solution.  Ute and Wilfried Brauer [1989] presented a new strategy and got a better solution than Wood's.  They also conjectured that their solution was optimal for infinitely many values of  n.  We give an algorithm which produces a better solution than Brauers' for all  n > 6, and we use a linear programming formulation to derive an upper bound which shows that our solution is optimal."  14 references, several not given above.

 

5.O.   TAIT'S COUNTER PUZZLE:  BBBBWWWW  TO  WBWBWBWB

 

          See S&B 125.

          The rules are that one can move two counters as an ordered pair, e.g. from  BBBBWWWW  to  BBB..WWWBW,  but not to  BBB..WWWWB  -- except in Lucas (1895) and AM prob. 237, where such reversal must be done.  Also, moving to  BBB..WWW.BW  is sometimes explicitly prohibited, but it is not always clear just where one can move to.  It is also not always specified where the blank spaces are at the beginning and end positions. 

          Gardner, 1961, requires that the two counters must be  BW  or  WB.

          Barbeau, 1995, notes that moving to  BWBWBWBW  is a different problem, requiring an extra move.  I had not noticed this difference before -- indeed I previously had it the wrong way round in the heading of this section.  I must check to see if this occurs earlier.  See Achugbue & Chin, 1979-80, for this version.

 

Genjun Nakane (= Hōjiku Nakane).  Kanja‑otogi‑soshi (Book of amusing problems for the entertainment of thinkers).  1743.  ??NYS.  (See:  T. Hayashi; Tait's problem with counters in the Japanese mathematics; Bibl. Mathem. (3) 6 (1905) 323, for this and other Japanese references of 1844 and 1879, ??NYS.)

P. G. Tait.  Listing's Topologie.  Philosophical Mag. (Ser. 5) 17 (No. 103) (Jan 1884) 30‑46 & plate opp. p. 80.  Section 12, pp. 39‑40.  He says he recently saw it being played on a train.

George Hope Verney (= Lloyd‑Verney).  Chess eccentricities.  Longmans, 1885.  P. 193: The pawn puzzle.  ??NX  With  4 & 4.

Lucas.  Amusements par les jetons.  La Nature 15 (1887, 2nd sem.) 10-11.  ??NYS -- cited by Ahrens, title obtained from Harkin.  Probably c= the material in RM3, below.

Ball.  MRE, 1st ed., 1892, pp. 48‑49.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. XIV: The eight-card puzzle, pp. 14-15.  Uses cards:  BRBRBRBR  and asks to bring the colours together, explicitly requiring the moved cards to be placed in contact with the unmoved cards.

Hoffmann.  1893.  Chap. VI, pp. 270‑271 & 284‑285 = Hoffmann-Hordern, pp. 184-186, with photo.

No. 19: The "Four and Four" puzzle.  Photo on p. 184 shows a version named Monkey Puzzle advertising Brooke's Soap to go from  BBBBBWWWW..  to  ..WBWBWBWB .

No. 20: The "Five and Five" puzzle.

No. 21: The "Six and Six" puzzle.

Lucas.  RM3.  1893.  Amusements par les jetons, pp. 145‑151.  He gives Delannoy's general solution for  n  of each colour in  n  moves.  Remarks that one can reverse the moved pair.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 11: The Egyptian disc puzzle.  4 & 4.  "Two discs adjoining each other to be moved at a time; no gaps to be left in the line."  -- this seems to prevent one from making any moves at all!!  No solution.

Lucas.  L'Arithmétique Amusante.  1895.  Pp. 84-108.

Prob. XXI - XXIV and Méthode générale, pp. 84-97.  Gives solution for  4, 5, 6, 7  and the general solution for  n & n  in  n  moves due to Delannoy.

Rouges et noires, avec interversion, prob. XXV - XXVIII and Méthode générale, pp. 97‑108.  Interversion means that the two pieces being moved are reversed or turned over, e.g. from  BBBBWWWW  to  BBB..WWWWB,  but not to  BBB..WWWBW.  Gives solutions for  4, 5, 6, 7, 8  pairs and in general in  n  moves, but he ends with a gap, e.g.  ....BB..BB  and it takes an extra move to close up the gap.

Ball.  MRE, 3rd ed., 1896, pp. 65‑66.  Cites Delannoy's solution as being in La Nature (Jun 1887) 10.  ??NYS.

Ahrens.  MUS I.  1910.  Pp. 14-15 & 19-25.  Cites Tait and gives Delannoy's general solution, from Lucas.

Ball.  MRE, 5th ed., 1911, pp. 75-77.  Adds a citation to Hayashi, but incorrectly gives the date as 1896.

Loyd.  Cyclopedia.  1914.  After dinner tricks, pp. 41 & 344.  4 & 4.

Williams.  Home Entertainments.  1914.  The eight counters puzzle, pp. 116-117.  Standard version, but with black and white reversed, in four moves.  Says the moved counters must be placed in line with and touching the others.

Dudeney.  AM.  1917. 

Prob. 236: The hat puzzle, pp. 67 & 196-197.  BWBWBWBWBW..  to have the  Bs  and  Ws  together and two blanks at an end.  Uses 5 moves to get to  ..WWWWWBBBBB.

Prob. 237: Boys and girls, pp. 67-68 & 197.  ..BWBWBWBW  to have the  Bs  and  Ws  together with two blanks at an end, but pairs must be reversed as they are moved.  Solution in 5 moves to  WWWWBBBB...  = Putnam, no. 2.  Cf Lucas, 1895.

Blyth.  Match-Stick Magic.  1921.  Transferring in twos, pp. 80-81.  WBWBWBWB..  to  ..BBBBWWWW  in four moves.

King.  Best 100.  1927.  No. 66, pp. 27 & 55.  = Foulsham's, no. 9, pp. 9 & 13.  BWBWBWBW..  to  ..WWWWBBBB,  specifically prescribed.

Rohrbough.  Brain Resters and Testers.  c1935.  Alternate in Four Moves, p. 4.  ..BBBBWWWW  to  WBWBWBWB.. , but he doesn't specify the blanks, showing all stages as closed up to 8 spaces, except the first two stages have a gap in the middle.

McKay.  At Home Tonight.  1940. 

Prob. 43: Arranging counters, pp. 73 & 87-88.  RBRBRB....  to  ....BBBRRR  in three moves.  Sketches general solution.

Prob. 45: Triplets, pp. 74 & 88.  YRBYRBYRB..  to  BBBYYYRRR..  in 5 moves.

McKay.  Party Night.  1940.  Heads and tails again, p. 151.  RBRBR..  to  ..BBRRR  in three moves.  RBRBRB..  to ..BBBRRR  in four moves.  RBRBRBRB..  to  ..BBBBRRRR  in four moves.  Notes that the first move takes coins 2 & 3 to the end and thereafter one is always filling the spaces just vacated.

Gardner.  SA (Jun & Jul 1961)  =  New MD, chap. 19, no. 1: Collating the coins.  BWBWB  to  BBBWW,  moving pairs of  BW  or  WB  only, but the final position may be shifted.  Gardner thanks H. S. Percival for the idea.  Solution in 4 moves, using gaps and with the solution shifted by six spaces to the right.  Thanks to Heinrich Hemme for this reference.

Joseph S. Madachy.  Mathematics on Vacation.  (Scribners, NY, 1966, ??NYS);  c= Madachy's Mathematical Recreations.  Dover, 1979.  Prob. 3: Nine-coin move, pp. 115 & 128-129 (where the solution is headed Eight-coin move).  This uses three types of coin, which I will denote by  B, R, W.  BRWBRWBRW ® WWWRRRBBB  by moving two adjacent unlike coins at a time and not placing the two coins away from the rest.  Eight move solution leaves the coins in the same places, but uses two extra cells at each end.  From the discussion of Bergerson's problem, see below, it is clear that the earlier book omitted the word unlike and had a nine move solution, which has been replaced by Bergerson's eight move solution.

Yeong‑Wen Hwang.  An interlacing transformation problem.  AMM 67 (1967) 974‑976.  Shows the problem with  2n  pieces,  n > 2,  can be solved in  n  moves and this is minimal.

Doubleday - 1.  1969.  Prob. 70: Oranges and lemons, pp. 86 & 170.  = Doubleday - 4, pp. 95‑96.  BWBWBWBWBW..  considered as a cycle.  There are two solutions in five moves:  to  ..WWWWWBBBBB, which never uses the cycle;  and to:  BBWWWWWW..BBB.

Howard W. Bergerson, proposer;  Editorial discussion;  D. Dobrev, further solver;  R. H. Jones, further solver.  JRM  2:2 (Apr 1969) 97;  3:1 (Jan 1970) 47-48;  3:4 (Oct 1970) 233-234;  6:2 (Spring 1973) 158.  Gives Madachy's 1966 problem and says there is a shorter solution.  The editor points out that Madachy's book and Bergerson have omitted unlike.  Bergerson has an eight move solution of the intended problem, using two extra cells at each end, and Leigh James gives a six move solution of the stated problem, also using two extra cells at each end.  Dobrev gives solutions in six and five steps, using only two extra cells at the right.  Jones notes that the problem does not state that the coins have to be adjacent and produces a four move solution of the stated problem, going from  ....BRWBRWBRW....  to  WWW..R..R..R..BBB.

Jan M. Gombert.  Coin strings.  MM 42:5 (Nov 1969) 244-247.  Notes that  BWBWB...... ® ......BBBWW  can be done in four moves.  In general,  BWB...BWB,  with  n  Ws  and  n+1  Bs  alternating can be transformed to  BB...BWW...W  in  n2  moves and this is minimal.  This requires shifting the whole string  n(n+1)  to the right and a move can go to places separated from the rest of the pieces.  By symmetry,  ......BWBWB ® WWBBB......  in the same number of moves.

Doubleday - 2.  1971.  Two by two, pp. 107-108.  ..BWBWBWBW  to  WWWWBBBB...  He doesn't specify where the extra spaces are, but says the first two must move to the end of the row, then two more into the space, and so on.  The solution always has two moving into an internal space after the first move.

Wayne A. Wickelgren.  How to Solve Problems.  Freeman, 1974.  Checker-rearrangement problem, pp. 144‑146.  BWBWB  to  BBBWW  by moving two adjacent checkers, of different colours, at a time.  Solves in four moves, but the pattern moves six places to the left.

Putnam.  Puzzle Fun.  1978.

No. 1: Nickles [sic] & dimes, pp. 1 & 25.  Usual version with 8 coins.  Solution has blanks at the opposite end to where they began.

No. 2: Nickles [sic] & dimes variation, pp. 1 & 25.  Same, except the order of each pair must be reversed as it moves.  Solution in five moves with blanks at opposite end to where they started.  = AM 237.  Cf Lucas, 1895.

James O. Achugbue  &  Francis Y. Chin.  Some new results on a shuffling problem.  JRM 12:2 (1979-80) 125-129.  They demonstrate that any pattern of  n & n  occupying  2n  consecutive cells can be transformed into any other pattern in the same cells, using only two extra cells at the right, except for the case  n = 3  where  10  cells are used.  They then find an optimal solution for  BB...BW...WW ® BWBW...BW  in  n+1  moves using two extra cells.  They seem to leave open the question of whether the number of moves could be shortened by using more cells.

Walter Gibson.  Big Book of Magic for All Ages.  Kaye & Ward, Kingswood, Surrey, 1982. 

Six cents at a time, p. 117.  Uses pennies and nickels.  .....PNPNP  to  NNPPP.....  in four moves.

Tricky turnover, p. 137.  HTHTHT  to  HHHTTT  in two moves.  This requires turning over one of the two coins on each move.

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Pp. 117, 119 & 123-126.  He asks to move  BBBWWW  to  WBWBWB  and to  BWBWBW  and notes that the latter takes an extra move.  He sketches the general solutions. 

 

5.P.    GENERAL MOVING PIECE PUZZLES

 

          See also under 5.A.

         

5.P.1. SHUNTING PUZZLES

 

          See Hordern, op. cit. in 5.A, pp. 167‑177, for a survey of these puzzles.  The Chifu‑Chemulpo (or Russo‑Jap Railway) Puzzle of 1903 is actually not of this type since all the pieces can move by themselves -- Hordern, pp. 124‑125 & plate VIII.

          See S&B 124‑125.

          A 'spur' is a dead‑end line.  A 'side‑line' is a line or siding joined to another at both ends.

 

Mittenzwey.  1880.  Prob. 219-221, pp. 39-40 & 91;  1895?: 244-246, pp. 43-44 & 93;  1917: 244-246, pp. 40 & 89.  First two have a canal too narrow to permit boats to pass, with a 'bight', or widening, big enough to hold one boat while another passes.  First problem has two boats meeting one boat; second problem has two boats meeting two boats.  The third problem has a single track railway with a side-line big enough to hold an engine and 16 wagons on the side-line or on the main line between the switches.  Two trains consisting of an engine and 20 wagons meet.

Lucas.  RM2, 1883, pp. 131‑133.  Passing with a spur and with a side‑line.

Alexander Henry Reed.  UK Patent 15,051 -- Improvements in Puzzles.  Complete specification: 8 Dec 1885.  4pp + 1p diagrams.  Reverse a train using a small turntable on the line.  This has forms with one line and with two crossing lines.  One object is to spell 'Humpty Dumptie'.  He also has a circular line with three turntables (equivalent to the recent Top-Spin Puzzle of F. Lammertinck).

Pryse Protheroe.  US Patent 332,211 -- Puzzle.  Applied: 18 Sep 1885;  patented: 8 Dec 1885.  3pp + 1p diagrams.  Described in Hordern, p. 167.  Identical to the Reed patent above!  Both Reed and Protheroe are described as residents of suburban London.  The Reed patent says it was communicated from abroad by an Israel J. Merritt Jr of New York and it doesn't assert that Reed is the inventor, so perhaps Reed and Merritt were agents for Protheroe.

Jeffrey & Son (Syracuse, NY).  Great Railroad Puzzle.  Postcard puzzle produced in 1888.  ??NYS.  Described in Hordern, pp. 175‑176.  Passing with a turntable that holds two wagons.

Arthur G. Farwell.  US Patent 437,186 -- Toy or Puzzle.  Applied: 20 May 1889;  patented: 30 Sep 1890.  1p + 1p diagrams.  Described in  Hordern, pp. 167‑169.  Great Northern Puzzle.  This requires interchanging two cars on the legs of a 'delta' switch which is too short to allow the engine through, but will let the cars through.  Hordern lists 6 later patents on the same basic idea.

Ball.  MRE, 1st ed., 1892, pp. 43‑44.  Great Northern Puzzle "which I bought some eight or nine years ago."  (Hordern, p. 167, erroneously attributes this quote to Ahrens.)

Loyd.  Problem 28: A railway puzzle.  Tit‑Bits 32 (10 Apr  &  1 May 1897) 23  &  79.  Engine and 3 cars need to pass 4 cars by means of a 'delta' switch whose branches and tail hold only one car.  Solution with 28 reversals.

Loyd.  Problem 31: The turn‑table puzzle.  Tit‑Bits 32 (1  &  22 May 1897) 79  &  135.  Reverse an engine and 9 cars with an 8 track turntable whose lines hold 3 cars.  The turntable is a double curved connection which connects, e.g. track 1 to tracks 4 or 6.

E. Fourrey.  Récréations Arithmétiques.  Op. cit. in 4.A.1.  1899.  Art. 239: Problèmes de Chemin de fer, pp. 184-189.

I.  Three parallel tracks with two switched crossing tracks.  Train of 21 wagons on the first track must leave wagons 9 & 12 on third track.

II. Delta shape with a turntable at the point of the delta, which can only hold the wagons and not the engine, so this is isomorphic to Farwell.

III. This is a more complex railway problem involving timetables on a circular line.

J. W. B.  Shunting!  c1900.  ??NYS.  Described in Hordern, pp. 176‑177 & plate XII.  Reversing a train with a turntable that holds three wagons.

Orril L. Hubbard.  US Patent 753,266 -- Puzzle.  Applied: 21 Apr 1902;  patented: 1 Mar 1904.  3pp + 1p diagrams.  Great Railroad Puzzle, described in Hordern, pp. 175‑176.  Improved version of the Jeffrey & Son puzzle of 1888.  Engine & 2 cars to pass engine & 3 cars, using a turntable that holds two cars, preserving order of each train.

Harry Lionel Hook  &  George Frederick White.  UK Patent 26,645 -- An Improved Puzzle or Game.  Applied: 3 Dec 1902;  accepted: 11 Jun 1903.  2pp + 1p diagrams.  This is very cryptic, but appears to be a kind of sliding piece Puzzle using turntables.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:1 (May 1903) 50-51  &  10:2 (Jun 1903) 140-141  & 10:4 (Aug 1903) 336-337.  A railway puzzle.  One north-south line with a spur heading north which is holding 7 trucks, but cannot hold the engine as well, so the engine is on the main line heading south.  An engine pulling seven trucks arrives from the north and wants to get past.  First solution uses 17 stages; second uses 12 stages.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:5 (Sep 1903) 426-427  &  10:6 (Oct 1903) 530‑531.  A shunting problem.  Same as Fourrey - II, hence isomorphic to Farwell.  Solution in 17 stages.

Celluloid Starch Puzzle.  c1905.  Described in Hordern, pp. 169‑170.  Cars on the three parts of a 'delta' switch with an engine approaching.  Reverse the engine, leaving all cars on their original places.  More complexly, suppose the tail of the 'delta' only holds one car or the engine.

Livingston B. Pennell.  US Patent 783,589 -- Game Apparatus.  Applied: 20 Mar 1902;  patented: 28 Feb 1905.  3pp + 1p diagrams.  Described in Hordern, p. 173.  Passing with a side line -- engine & 3 cars to pass engine & 3 cars using a siding which already contains 3 cars, without couplings, so these three can only be pushed.  Also the engines can move at most three cars at a time.

William Rich & Harry Pritchard.  UK Patent 7647 -- Railway Game and Puzzle.  Applied: 11 Apr 1905;  complete specification: 11 Oct 1905;  accepted: 14 Dec 1905.  2pp + 1p diagrams.  Main line with two short and two long spurs.

Ball.  MRE, 4th ed., 1905, pp. 61-63, adds a problem with a side-line, "on sale in the streets in 1905“.  The 5th ed., 1911, pp. 69-71 & 82, adds the name "Chifu-Chemulpo Puzzle" and that the minimum number of moves is 26, in more than one way.  P. 82 gives solutions of both problems.

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  Great Northern Puzzle.  He says the "Railway puzzle" was very popular "about twenty years ago".

Ahrens.  MUS I.  1910.  Pp. 3-4.  Great Northern.  Says it is apparently modern and cites Fourrey for other examples.

Anon.  Prob. 6.  Hobbies 32  (No. 814) (20 May 1911) 145  &  (No. 817) (10 Jun 1911) 208.  Great Northern Puzzle.  Solution asks if readers know any other railway puzzles.

Loyd.  The switch problem  &  Primitive railroading problem.  Cyclopedia, 1914, pp. 167 & 361;  89 & 350 (= MPSL2, prob. 24, pp. 18‑19;  MPSL1, prob. 95, pp. 92 & 155).  Passing with a 'delta' switch  &  passing with a spur.  The first is like Tit-Bits Problem 28, but the engine and 3 cars have to pass 5 cars.  Solution in 32 moves.  See Hordern, pp. 170‑171.

Hummerston.  Fun, Mirth & Mystery.  1924.  The Chinese railways, pp. 103 & 188.  Imagine a line of positions:  ABCEHGJLMN  with single positions  D, I, F, K attached to positions  C, H, G, L.  You have eight engines at  ABCD  and  KLMN  and the object is to exchange them, preserving the order.  He does it in  18  moves, where a move can be of any length.

King.  Best 100.  1927.  No. 14, pp. 12 & 41.  Side‑line with a bridge over it too low for the engine.  Must interchange two wagons on the side‑line which are on opposite sides of the bridge.

B. M. Fairbanks.  Railroad switching problems.  IN:  S. Loyd Jr., ed.; Tricks and Puzzles; op. cit. in 5.D.1 under Chapin; 1927.  P. 85 & Answers p. 7.  Three realistic problems with several spurs and sidelines.

Loyd Jr.  SLAHP.  1928.  Switching cars, pp. 54 & 106.  Great Northern puzzle.  See Hordern, pp. 168‑169.

Doubleday - 2.  1971.  Traffic jam, pp. 85-86.  Version with cars in a narrow lane and a lay-by.  Two cars going each way.  Though the lay-by is three cars wide and just over a car long, he restricts its use so that it acts like it is two cars wide.

 

          5.P.2. TAQUIN

 

          Jacques Haubrich has kindly enlightened me that 'taquin' simply means 'teaser'.  So these items should be re-categorised.

 

Lucas.  RM3.  1893.  3ème Récréation -- Le jeu du caméléon et le jeu des jonctions de points, pp. 89‑103.  Pp. 91‑97 -- Le taquin de neuf cases avec un seul port.  I thought that taquin was the French generic term for such puzzles, but I find no other usage than that below, except in referring to the 15 Puzzle -- see references to taquin in 5.A.

Au Bon Marché (the Paris department store).  Catalogue of 1907, p. 13.  Reproduced in Mary Hillier; Automata and Mechanical Toys; An Illustrated History; Jupiter Books, London, 1976, p. 179.  This shows  Le Taquin Japonais  Jeu de Patience  Casse-tete.  This comprises 16 hexagonal pieces, looking like a corner view of a die, so each has three rhombic parts containing a pattern of pips.  They are to be placed as the corners of four interlocked hexagons with the numbers on adjacent rhombi matching.

 

5.Q.   NUMBER OF REGIONS DETERMINED BY  N  LINES OR PLANES

 

Mittenzwey.  1880.  Prob. 200, pp. 37 & 89;  1895?: 225, pp. 41 & 91;  1917: 225, pp. 38 & 88.  Family of 4 adults and 4 children.  With three cuts, divide a cake so the adults and the children get equal pieces.  He makes two perpendicular diametrical cuts and then a circular cut around the middle.  He seems to mean the adults get equal pieces and the children get equal pieces, not necessarily the same.  But if the circular cut is at  Ö2/2  of the radius, then the areas are all equal.  Not clear where this should go -- also entered in 5.T.

Jakob Steiner.  Einige Gesetze über die Theilung der Ebene und des Raumes.  (J. reine u. angew. Math. 1 (1826) 349‑364)  = Gesam. Werke, 1881, vol. 1, pp. 77‑94.  Says the plane problem has been raised before, even in a Pestalozzi school book, but believes he is first to consider 3‑space.  Considers division by lines and circles (planes and spheres) and allows parallel families, but no three coincident.

Richard A. Proctor.  Some puzzles;  Knowledge 9 (Aug 1886) 305-306  &  Three puzzles;  Knowledge 9 (Sep 1886) 336-337.  "3.  A man marks 6 straight lines on a field in such a way as to enclose 10 spaces.  How does he manage this?"  Solution begins:  "III.  To inclose ten spaces by six ropes fastened to nine pegs."  Take  (0,0), (1,0), ..., (n,0), (0,n), ..., (0,1),  as  2n+1  points, using  n+2  ropes from  (0,0) to (n,0) and to (0,n)  and from  (i,0) to (0,n+1-i)  to enclose  n(n+1)/2  areas. 

Richard A. Proctor.  Our puzzles.  Knowledge 10  (Nov 1886) 9  &  (Dec 1886) 39-40.  Describes several ways of solving previous problem and asks for a symmetric version.

G. Chrystal.  Algebra -- An Elementary Text-Book.  Vol. 2, A. & C. Black, Edinburgh, 1889.  [Note -- the 1889 version of vol. 1 is a 2nd ed.]  Chap. 23, Exercises IV, p. 34.  Several similar problems and the following.

No. 7 -- find number of interior and of exterior intersections of the diagonals of a convex  n-gon.

No. 8 --  n  points in general position in space, draw planes through every three and find number of lines and of points of intersection.

L. Schläfli.  Theorie der vielfachen Kontinuität.  Neue Denkschriften der allgemeinen schweizerischen Gesellschaft für die Naturwissenschaften 38:IV, Zürich, 1901, 239 pp.  = Ges. Math. Abh., Birkhäuser, Basel, 1950‑1956, vol. 1, pp. 167‑392.  (Pp. 388‑392 are a Nachwort by J. J. Burckhardt.)  Material of interest is Art. 16: Über die Zahl der Teile, ..., pp. 209‑212.  Obtains formula for  k  hyperplanes in  n  space.

Loyd, Dudeney, Pearson & Loyd Jr. give various puzzles based on this topic.

Howard D. Grossman.  Plane- and space-dissection.  SM 11 (1945) 189-190.  Notes Schläfli's result and observes that the number of regions determined by  k+1  hyperspheres in  n  space is twice the number of regions determined by  k  hyperplanes and gives a two to one correspondence for the case  n = 2.

Leo Moser, solver.  MM 26 (Mar 1953) 226.  ??NYS.  Given in:  Charles W. Trigg; Mathematical Quickies; (McGraw‑Hill, NY, 1967);  corrected ed., Dover, 1985.  Quickie 32: Triangles in a circle, pp. 11 & 90‑91.  N  points on a circle with all diagonals drawn.  Assume no three diagonals are concurrent.  How many triangles are formed whose vertices are internal intersections?

Timothy Murphy.  The dissection of a circle by chords.  MG 56 (No. 396) (May 1972) 113‑115  +  Correction (No. 397) (Oct 1972) 235‑236.  N  points on a circle, in a plane or on a sphere;  or  N  lines in a plane or on a sphere, all simply done, using Euler's formula.

Rowan Barnes-Murphy.  Monstrous Mysteries.  Piccolo, 1982.  Slicing cakes, pp. 33 & 61.  Cut a circular cake into  12  equal pieces with  4  cuts.  [From this, we see that  N  full cuts can yield either  2N  or  4(N-1)  equal pieces.  Further, if we make  k  circular cuts producing  k+1  regions of equal area and then make  N-k  diametric cuts equally spaced, we get  2(k+1)(N-k)  pieces of the same size.]

Looking at this problem, I see that one can obtain any number of pieces from  N+1  up through the maximum.

 

5.Q.1.      NUMBER OF INTERSECTIONS DETERMINED BY  N  LINES

 

Chrystal.  Text Book of Algebra.  2nd ed., vol. 2, 1889, p. 34, ex. 7.  See above.

Loyd Jr.  SLAHP.  1928.  When drummers meet, pp. 74 & 115.  Six straight railroads can meet in 15 points.

Paul Erdös, proposer;  Norbert Kaufman & R. H. Koch and Arthur Rosenthal, solvers.  Problem E750.  AMM 53 (1946) 591  &  54 (1947) 344.  The first solution is given in Trigg, op. cit. in 5.Q, Quickie 191: Intersections of diagonals, pp. 53 & 166‑167.  In a convex  n‑gon, how many intersections of diagonals are there?  This counts a triple intersection as three ordinary (i.e. double) intersections or assumes no three diagonals are concurrent.  Editorial notes add some extra results and cite Chrystal.

 

          5.R.   JUMPING PIECE GAMES

 

          See also 5.O.  Some of these are puzzles, but some are games and are described in the standard works on games -- see the beginning of 4.B.

 

          5.R.1. PEG SOLITAIRE

 

          See MUS I 182-210.

 

Ahrens, MUS I 182‑183, gives legend associating this with American Indians.  Bergholt, below, and Beasley, below, find this legend in the 1799 Encyclopédie Méthodique: Dictionnaire des Jeux Mathématique (??*), ??NYS.  Ahrens also cites some early 19C material which has not been located.  Bergholt says some maintain the game comes from China.

 

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De Ludo dicto Ufuba wa Hulana, p. 233.  This has a  5 x 5  board with each side having  12  men, but the description is extremely brief.  It seems to have two players, but this may simply refer to the two types of piece.  I'm not clear whether it's played like solitaire (with the jumped pieces being removed) or like frogs & toads.  I would be grateful if someone could read the Latin carefully.  The name of the puzzle is clearly Arabic and Hyde cites an Arabic source, Hanzoanitas (not further identified on the pages I have) -- I would be grateful to anyone who can track down and translate Arabic sources.

G. W. Leibniz.  Le Jeu du Solitaire.  Unpublished MS LH XXXV 3 A 10 f. 1-2, of c1678.  Transcribed in:  S. de Mora-Charles; Quelques jeux de hazard selon Leibniz; HM 19 (1992) 125-157.  Text is on pp. 152-154.  37 hole board.  Says the Germans call it 'Die Melancholy' and that it is now the mode at the French court.

Claude‑Auguste Berey.  Engraving:  Madame la Princesse de Soubize jouant au Jeu de Solitaire.  1697(?).  Beasley (below) discovered and added this while his book was in proof.  It shows the 37‑hole French board.  Reproduced in:  Pieter van Delft & Jack Botermans; Creative Puzzles of the World; op. cit. in 5.E.2.a, p. 170.

G. W. Leibniz.  Jeu des Productions.  Unpublished MS LH XXXV 8,30 f. 4, of 1698.  Transcribed in:  de Mora-Charles, loc. cit. above.  Text is on pp. 154-155.  37 hole board.  Considers the game in reverse.

Trouvain.  Engraving:  Dame de Qualité Jouant au Solitaire.  1698(?).

Claude‑Auguste Berey.  Engraving:  Nouveau Jeu de Solitaire.  Undated, but Berey was active c1690‑c1730.  Reproduced in:  R. C. Bell; The Board Game Book; Marshall Cavendish, London, 1979, pp. 54‑55  and in: Jasia Reichardt, ed.; Play Orbit [catalogue of an exhibition at the ICA, London, and elsewhere in 1969-1970]; Studio International, 1969, p. 38.  Beasley's additional notes point out that this engraving is well known, but he had not realised its date until the earlier Berey engraving was discovered.  This engraving includes the legend associating the game with the American indians -- "son origine vient de l'amerique ou les Peuples vont seuls à la chasse, et au retour plantent leurs flèches en des trous de leur cases, ce qui donna idée a un françois de composer ce jeu ...."  Reichardt says the original is in the Bibliothèque Nationale.

The three engravings above are reproduced in:  Henri d'Allemagne; Musée rétrospectif de la classe 100, Jeux, à l'exposition universelle international de 1900 à Paris, Tome II, pp. 152‑158.  D'Allemagne says the originals are in the Bibliothèque Nationale, Paris.  He (and de Mora-Charles) also cites Rémond de Montmort, 2nd ed., 1713 -- see below.

G. W. Leibniz.  Annotatio de quibusdam Ludis; inprimis de Ludo quodam Sinico, differentiaque Scachici et Latrunculorum & novo genere Ludi Navalis.  Misc. Berolinensia (= Misc. Soc. Reg., Berlin) 1 (1710) 24.  Last para. on p. 24 relates to solitaire.  (English translation on p. xii of Beasley, below.)

Pierre Rémond de Montmort.  Essai d'analyse sur les jeux de hazards.  (1708);  Seconde edition revue & augmentee de plusieurs lettres,  (Quillau, Paris, 1713 (reprinted by Chelsea, NY, 1980));  2nd issue, Jombert & Quillau, 1714.  Avertissement (to the 2nd ed.), xli-xl.  "J'ai trouvé dans le premier volume de l'Academie Royale de Berlin, ...; il propose ensuite des Problèmes sur un jeu qui a été à la mode en France il y a douze ou quinze ans, qui se nomme Le Solitaire."

Edward Hordern's collection has a wooden 37 hole board on the back of which is inscribed "Invented by Lord Derwentwater when Imprisoned in the Tower".  The writing is old, at least 19C, possibly earlier.  However the Encyclopedia Britannica article on Derwentwater and the DNB article on Radcliffe, James, shows that the relevant Lord was most likely to have been James Radcliffe (1689-1716), the 3rd Earl from 1705, who joined the Stuart rising in 1715, was captured at Preston, was imprisoned in the Tower and was beheaded on 24 Feb 1716, so the implied date of invention is 1715 or 1716.  The third Earl became a figure of romance and many stories and books appeared about him, so the invention of solitaire could well have been attributed to him. 

                    Though the title was attainted and hence legally extinct, it was claimed by relatives.  Both James's brother Charles (1693‑1746), the claimed 5th Earl from 1731, and Charles's son James Bartholomew (1725-1786), the claimed 6th Earl from 1746, spent time in prison for their Stuart sympathies.  Charles escaped from Newgate Prison after the 1715 rising, but both were captured on their way to the 1745 rising and taken to the Tower where Charles was beheaded.  If either of these is the Lord Derwentwater referred to, then the date must be 1745 or 1746.  A guide book to Northumberland, where the family lived at Dyvelston (or Dilston) Castle, near Hexham, asserts the last Derwentwater was executed in 1745, while the [Blue Guide] says the last was executed for his part in the 1715 uprising.

                    In any case, the claim seems unlikely.

G. W. Leibniz.  Letter to de Montmort (17 Jan 1716).  In:  C. J. Gerhardt, ed.; Die Philosophischen Schriften von Gottfried Wilhelm Leibniz; (Berlin, 1887)  = Olms, Hildesheim, 1960; Vol. 3, pp. 667‑669.  Relevant passage is on pp. 668‑669.  (Poinsot, op. cit. in 5.E, p. 17, quotes this as letter VIII in Leibn. Opera philologica.)

J. C. Wiegleb.  Unterricht in der natürlichen Magie.  Nicolai, Berlin & Stettin, 1779.  Anhang von dreyen Solitärspielen, pp. 413‑416, ??NYS -- cited by Beasley.  First known diagram of the 33‑hole board.

Catel.  Kunst-Cabinet.  1790.  Das Grillenspiel (Solitaire), p. 50 & fig. 167 on plate VI.  33 hole board.  (Das Schaaf- und Wolfspiel, p. 52 & fig. 169 on plate VI, is a game on the 33-hole board.)

Bestelmeier.  1801.  Item 511: Ein Solitair, oder Nonnenspiel.  33 hole board.

Strutt.  Op. cit. in 4.B.1.  The Solitary Game.  (1801: Book IV, p. 238.  ??NYS -- cited by Beasley -- may be actually 1791??)  1833: Book IV, chap. II, art. XV, p. 319.  c= Strutt-Cox, p. 259.  Beasley says this is the first attribution to a prisoner in the Bastille.  The description is vague: "fifty or sixty" holes and "a certain number of pegs".  Strutt-Cox adds a note that "The game of Solitaire, reimported from France, ..., came again into Fashion in England in the late" 1850s and early 1860s.

Ada Lovelace.  Letter of 16 Feb 1840 to Charles Babbage.  BM MSS 37191, f. 331.  ??NYS -- reproduced in Teri Perl; Math Equals; Addison-Wesley, Menlo Park, California, 1978, pp. 109-110.  Discusses the 37 hole board and wonders if there is a mathematical formula for it.

M. Reiss.  Beiträge zur Theorie des Solitär‑Spiels.  J. reine angew. Math. 54 (1857) 376‑379.

St. v. Kosiński & Louis Wolfsberg.  German Patent 42919 -- Geduldspiel.  Patented: 25 Sep 1877.  1p + 1p diagrams.  33 hole version.

The Sociable.  1858.  The game of solitaire, pp. 282-284.  37 hole board.  "It is supposed to have been invented in America, by a Frenchman, to beguile the wearisomeness attendant upon forest life, and for the amusement of the Indians, who pass much of their time alone at the chase, ...."

Anonymous.  Enquire Within upon Everything.  66th ed., 862nd thousand, Houlston and Sons, London, 1883, HB.  Section 135: Solitaire, p. 49.  Mentions a 37 hole board but shows a 33 hole board.  This material presumably goes back some time before this edition.  It later shows Fox and Geese on the 33 hole board. 

Hoffmann.  1893.  Chap. X, no. 11: Solitaire problems, pp. 339-340 & 376-377 = Hoffmann‑Hordern, pp. 232-233, with photo on p. 235.  Three problems.  Photo on p. 235 shows a 33-hole board in a square frame, 1820-1840, and a 37-hole board with a holding handle, 1840-1890.

Ernest Bergholt.  Complete Handbook to the Game of Solitaire on the English Board of Thirty-three Holes.  Routledge, London, nd [Preface dated Nov 1920] -- facsimile produced by Naoaki Takashima, 1993.  This is the best general survey of the game prior to Beasley.

King.  Best 100.  1927.  No. 68, pp. 28 & 55.  = Foulsham's no. 24, pp. 9 & 13.  3 x 3  array of men in the middle of a  5 x 5  board.  Men can jump diagonally as well as orthogonally.  Object is to leave one man in the centre.

Rohrbough.  Puzzle Craft.  1932.  Note on Solitaire  &  French Solitaire, pp. 14-15 (= pp. 6-7 of 1940s?).  33 hole board, despite being called French.

B. M. Stewart.  Solitaire on a checkerboard.  AMM 48 (1941) 228-233.  This surveys the history and then considers the game on the 32 cell board comprising the squares of one colour on a chessboard.  He tilts this by  45o  to get a board with 7 rows, having  2, 4, 6, 8, 6, 4, 2  cells in each row.  He shows that each beginning-ending problem which is permitted by the parity rules is actually solvable, but he gives examples to show this need not happen on other boards.

Gardner.  SA (Jun 1962).  Much amended as: Unexpected, chap. 11, citing results of Beasley, Conway, et al.  Cites Leibniz and mentions Bastille story.

J. D. Beasley.  Some notes on solitaire.  Eureka 25 (Oct 1962) 13-18.  No history of the game.

Jeanine Cabrera & René Houot.  Traité Pratique du Solitaire.  Librairie Saint‑Germain, Paris, 1977.  On p. 2, they give the story that it was invented by a prisoner in the Bastille, late 18C, and they even give the name of the reputed inventor:  "Comte"(?) Pellisson.  They say that a Paul Pellisson‑Fontanier was in the Bastille in 1661‑1666 and was a man of some note, but history records no connection between him and the game.

The Diagram Group.  Baffle Puzzles -- 3: Practical Puzzles.  Sphere, 1983.  No. 12.  On the 33-hole board place  16  markers:  1 in row 2;  3 in row 3;  5 in row 4;  7 in row 5;  making a triangle centred on the mid-line.  Can you remove all the men, except for one in the central square?  Gives a solution in 15 jumps. 

J. D. Beasley.  The Ins and Outs of Peg Solitaire.  OUP, 1985.  History, pp. 3‑7;  Selected Bibliography, pp. 253‑261.  PLUS  Additional notes, from the author, 1p, Aug 1985.  57 references and 5 patents, including everything known before 1850.

Franco Agostini & Nicola Alberto De Carlo.  Intelligence Games.  (As:  Giochi della Intelligenza; Mondadori, Milan, 1985.)  Simon & Schuster, NY, 1987.  This gives the legend of the nobleman in the Bastille.  Then says that "it would appear that a very similar game" is mentioned by Ovid "and again, it was widely played in ancient China -- hence its still frequent alternative name, "Chinese checkers.""  I have included this as an excellent example of how unreferenced statements are made in popular literature.  I have never seen either of these latter statements made elsewhere.  The connection with Ovid is pretty tenuous -- he mentions a game involving three in a row and otherwise is pretty cryptic and I haven't seen anyone else claiming Ovid is referring to a solitaire game -- cf 4.B.5.  The connection with Chinese checkers is so far off that I wonder if there is a translation problem -- i.e. does the Italian name refer to some game other than what is known as Chinese checkers in English??

Nob Yoshigahara.  Puzzlart.  Tokyo, 1992.  Coin solitaire, pp. 5 & 90.  Four problems on a  4 x 4  board.

Marc Wellens, et al.  Speelgoed Museum Vlaanderen -- Musée du Jouet Flandre -- Spielzeug Museum Flandern -- Flanders Toy Museum.  Speelgoedmuseum Mechelen, Belgium, 1996, p. 90 (in English), asserts  'It was invented by the French nobleman Palissen, who had been imprisoned in the Bastille by Louis XIV' in the early 18C.

 

          5.R.1.a.        TRIANGULAR VERSION

 

          The triangular version of the game has only recently been investigated.  The triangular board is generally numbered as below.

 

                          1

                       2     3

                    4     5    6

                  7    8    9   10

              11  12  13  14  15

 

Herbert M. Smith.  US Patent 462,170 -- Puzzle.  Filed: 13 Mar 1891;  issued: 27 Oct 1891.  2pp + 1p diagrams.  A board based on a triangular lattice.

Rohrbough.  Puzzle Craft.  1932.  Triangle Puzzle, p. 5 (= p. 6 in 1940s?).  Remove peg 13 and leave last peg in hole 13.

Maxey Brooke.  (Fun for the Money, Scribner's, 1963);  reprinted as:  Coin Games and Puzzles; Dover, 1973.  All the following are on the 15 hole board.

Prob. 1: Triangular jump, pp. 10-11 & 75.  Remove one man and jump to leave one man on the board.  Says Wesley Edwards asserts there are just six solutions.  He removes the middle man of an edge and leaves the last man there.

Prob. 2: Triangular jump, Ltd., pp. 12-13 & 75.  Removes some of the possible jumps.

Prob. 3: Headless triangle, pp. 14 & 75.  Remove a corner man and leave last man there.

M. Gardner.  SA (Feb 1966)  c= Carnival, 1975, chap. 2.  Says a 15 hole version has been on sale as Ke Puzzle Game by S. S. Adams for some years.  Addendum cites Brooke and Hentzel and says much unpublished work has been done.

Irvin Roy Hentzel.  Triangular puzzle peg.  JRM 6:4 (1973) 280-283.  Gives basic theory for the triangular version.  Cites Gardner.

[Henry] Joseph & Lenore Scott.  Quiz Bizz.  Puzzles for Everyone -- Vol. 6.  Ace Books (Charter Communications), NY, 1975.  Pennies for your thoughts, pp. 179-182.  Remove a coin and solve.  Hint says to remove the coin at  13  and that you should be able to have the last coin at  13.  The solution has this property.

Alan G. Henney & Dagmar R. Henney.  Computer oriented solutions.  CM 4:8 (1978) 212‑216.  Considers the 'Canadian I. Q. Problem', which is the 15 hole board, but they also permit such jumps as  1 to 13,  removing  5.  They find solutions from each initial removal by random trial and error on a computer.

Putnam.  Puzzle Fun.  1978.  No. 15: Jumping coins, pp. 5 & 28.  15 hole version, remove peg  1  and leave last man there.

Benjamin L. Schwarz & Hayo Ahlburg.  Triangular peg solitaire -- A new result.  JRM 16:2 (1983-84) 97-101.  General study of the 15 hole board showing that starting and ending with  5  is impossible.

J. D. Beasley.  The Ins and Outs of Peg Solitaire.  Op. cit. above, 1985.  Pp. 229-232 discusses the triangular version, citing Smith, Gardner and Hentzel, saying that little has been published on it.

Irvin Roy Hentzel & Robert Roy Hentzel.  Triangular puzzle peg.  JRM 18:4 (1985-86) 253‑256.  Develops theory.

John Duncan & Donald Hayes.  Triangular solitaire.  JRM 23:1 (1991) 26-37.  Extended analysis.  Studies army advancement problem.

William A. Miller.  Triangular peg solitaire on a microcomputer.  JRM 23:2 (1991) 109-115  &  24:1 (1992) 11.  Summarises and extends previous work.  On the 10 hole triangular board, the classic problem has essentially a unique solution -- the removed man must be an edge man (e.g. 2) and the last man must be on the adjacent edge and a neighbour of the starting hole (i.e. 3 if one starts with 2).  On the 15 hole board, the removed man can be anywhere and there are many solutions in each case.

                    Remove man from hole:                       1                  2                  4                  5

                    Number of solutions:                     29760           14880           85258             1550

          Considers the 'tree' formed by the first four rows and hole 13.

 

          5.R.1.b.       OTHER SHAPES

 

          New section.  See also King and Stewart in 5.R.1 for some forms based on a square board.

                                                                           A

                                                                     B  C D  E

                                                                        F   G

                                                                           H

                                                                       I       J

 

Putnam.  Puzzle Fun.  1978.  No. 53: Checker star, pp. 10 & 34.  Use the 10 points of a pentagram, as above, and leave one of the inner points empty.  Reduce to one man.  [Parity shows the one man must be at an outer point and any outer point can be achieved.  If one leaves an outer point empty, then the last man must be on an inner point and any of these can be achieved.]

Hummerston.  Fun, Mirth & Mystery.  1924. 

Perplexity, pp. 22-23.  Using the octagram board shown in 5.A, place 15 markers on it, leaving cell 16 empty.  It is possible to remove all but one man.  [I can't see how to apply parity to this board.]

Solplex, p. 25.  In playing his Perplexity, specify where you will leave the last man?

Leap frog, Puzzle no. 22, pp. 64 & 175.  Take a  4 x 3  board with the long edge extended by one more cell at the upper left and lower right.  Put white counters on the  4 x 3  area, put a black counter in one of the extra cells and leave the other extra cell empty.  Remove all but the black man.  Counting multiple jumps of the same man as a single move, he does it in eight moves, getting the black man back to its starting point.

 

          5.R.2. FROGS AND TOADS:  BBB_WWW  TO  WWW_BBB

 

          In the simplest version, one has  n  black men at the left and  n  white men at the right of a strip of  2n+1  cells, e.g.  BBB_WWW.  One can slide a piece forward (i.e. blacks go left and whites go right) into an adjacent place or one can jump forward over one man of the other colour into an empty place.  The object is to reverse the colours, i.e. to get  WWW_BBB.  S&B 121 & 125, shows versions. 

          One finds that the solution never has a man moving backward nor a man jumping another man of the same colour.  Some authors have considered relaxing these restrictions, particularly if one has more blank spaces, when these unusual moves permit shorter solutions.  Perhaps the most general form of the one-dimensional problem would be the following.  Suppose we have  m  men at the left of the board,  n  men at the right and  b  blank spaces in the middle.  The usual case has  b = 1,  but when  b > 1,  the kinds of move permitted do change the number of moves in a minimal solution.  First, considering slides, can a piece slide backward?  Can a piece slide more than one space?  If so, is there a maximum distance,  s,  that it is allowed to slide?  (The usual problem has  s = 1.)  Of course  £ b.  Second, considering jumps, can a piece jump backward?  Can a piece jump over a piece (or pieces) of its own colour and/or a blank space (or spaces) and/or a mixture of these?  If so, is there a maximum number of pieces,  p,  that it can jump over?  (The usual problem has  p = 1.)  It is not hard to construct simple examples with  s > 1  such that shorter solutions exist when unusual moves are permitted.  Are there situations where one can show that backward moves are not needed?

          The game is sometimes played on a 2-dimensional board, where one colour can move down or right and the other can move up or left.  See:  Hyde ??;  Lucas (1883);  Ball;  Hoffmann  and 5.R.3.  Chinese checkers is a later variation of this same idea.  On these more complex boards, one is usually allowed to make multiple jumps and the object is usually to minimize the number of moves to accomplish the interchange of pieces.

          There is a trick version to convert full and empty glasses:  FFFEEE  to  FEFEFE  in one move, which is done by pouring.  I've just noted this in a 1992 book and I'll look for earlier examples.

 

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De Ludo dicto Ufuba wa Hulana, p. 233.  This has a  5 x 5  board with each side having  12  men, but the description is extremely brief.  It seems to have two players, but this may simply refer to the two types of piece.  I'm not clear whether it's played like solitaire (with the jumped pieces being removed) or like frogs & toads.  I would be grateful if someone could read the Latin carefully.  The name of the puzzle is clearly Arabic and Hyde cites an Arabic source, Hanzoanitas (not further identified on the pages I have) -- I would be grateful to anyone who can track down and translate Arabic sources.

American Agriculturist (Jun 1867).  Spanish Puzzle.  ??NYR -- copy sent by Will Shortz.

Anonymous.  Every Little Boy's Book  A Complete Cyclopædia of in and outdoor games with and without toys, domestic pets, conjuring, shows, riddles, etc.  With two hundred and fifty illustrations.  Routledge, London, nd.  HPL gives c1850, but the material is clearly derived from Every Boy's Book, whose first edition was 1856.  But the text considered here is not in the 1856 ed of Every Boy's Book (with J. G. Wood as unnamed editor), nor in the 8th ed of 1868 (published for Christmas 1867), which was the first seriously revised edition, with Edmund Routledge as editor, nor in the 13th ed. of 1878.  So this material is hard to date, though in 4.A.1, I've guessed this book may be c1868.

                    P. 12: Frogs and toads.  "A new and fascinating game of skill for two players; played on a leather board with twelve reptiles; the toads crawling, and the frogs hopping, according to certain laws laid down in the rules.  The game occupies but a few minutes, but in playing it there is scarcely any limit to the skill that can be exhibited, thus forming a lasting amusement.  (Published by Jaques, Hatton Garden.)"  This does not sound like our puzzle, but perhaps it is related.  Unfortunately Jaques' records were destroyed in WW2, so it is unlikely they can shed any light on what the game was.  Does anyone know what it was?

Hanky Panky.  1872.  Checker puzzle, p. 124.  Three and three, with solution.

Mittenzwey.  1880.  Prob. 239, pp. 44 & 94;  1895?: 267-268, pp. 48 & 96;  1917:  267-268, pp. 44 & 91‑92.  Problem with  3 & 3  brown and white horses in stalls.  1895? adds a version with  4 & 4.

Bazemore Bros. (Chattanooga, Tennessee).  The Great "13" Puzzle! Copyright No. 1033 ‑ O ‑ 1883.  Hammond & Jones Printers.  Advertising puzzle consisting of two  3 and 3  versions arranged in an  X  pattern.

Lucas.  RM2.  1883. 

Pp. 141‑143.  Finds number of moves for  n and n.

Pp. 144‑145.  Considers game on  5 x 5,  7 x 7,  ...,  boards and gives number of moves.

Edward Hordern's collection has an example called  Sphinxes and Pyramids  from the 1880s.

Sophus Tromholt.  Streichholzspiele.  (1889;  5th ed, 1892.)  Revised from 14th ed. of 1909 by R. Thiele; Zentralantiquariat der DDR, Leipzig, 1986.  Prob. 11, 41, 81 are the game for  4 & 4,  2 & 2,  3 & 3.

Ball.  MRE, 1st ed., 1892, pp. 49‑51.  3 & 3  case, citing Lucas, with generalization to  n & n;  7 x 7  board, citing Lucas, with generalization to  2n+1 x 2n+1.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  German counter puzzle, p. 112.  3 & 3 case.

Hoffmann.  1893.  Chap. VI, pp. 269‑270 & 282‑284 = Hoffmann-Hordern, pp. 182-185, with photo.

No. 17: The "Right and Left" puzzle.  Three and three.  Photo on p. 184 shows: a cartoon from Punch (18 Dec 1880): The Irish Frog Puzzle -- with a Deal of Croaking; and an example of a handsome carved board with square pieces with black and white frogs on the tops, registered 1880.  Hordern Collection, p. 77, shows the latter board and two further versions: Combat Sino-Japonais (1894‑1895) and Anglais & Boers (1899-1902).

No. 18.  Extends to a  7 x 5  board.

Puzzles with draughtsmen.  The Boy's Own Paper 17 or 18?? (1894??) 751.  3 and 3.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XXXV: Le bal des crapauds et des grenouilles, pp. 117-124.  Does  2 and 2,  3 and 3,  4 and 4  and the general case of  n and n,  showing it can be done in  n(n+2)  moves  --  n2  jumps and  2n  steps.  The general solution is attributed to M. Van den Berg.  M. Schoute notes that each move should make as little change as possible from the previous with respect to the two aspects of changing type of piece and changing type of move.

Clark.  Mental Nuts.  1904, no. 72;  1916, no. 62.  A good study.  3 and 3.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  Doola's Game, pp. 42-43 & 61-62.  3 and 3.

Anon.  Prob. 47: The monkey's dilemma.  Hobbies 30  (No. 762) (21 May 1910) 168 & 182  &  (No. 765) (11 Jun 1910) 228.  Basically  3 & 3,  but there are eight posts for crossing a river, with the monkeys on  1,2,3  and  6,7,8.  The monkeys can jump onto the bank and we want the monkeys to all get to the bank they are headed for, so this is not the same as  BBB..WWW  to  WWW..BBB.  The solution doesn't spell out all the steps, so it's not clear what the minimum number of moves is -- could we have a monkey jumping another of the same colour?

Ahrens.  MUS I.  1910.  Pp. 17-19.  Basically repeats some of Lucas's work from 1883 & 1895.

Williams.  Home Entertainments.  1914.  The cross-over puzzle, pp. 119-120.  3 and 3  with red and white counters.  Doesn't say how many moves are required.

Dudeney.  AM.  1917.  Prob. 216: The educated frogs, pp. 59-60 & 194.  _WWWBBB  to  BBBWWW_  with frogs able to jump either way over one or two men of either colour.  Solution in 10 jumps.

Ball.  MRE, 9th ed., 1920, pp. 77-79, considers the  m & n  case, giving the number of steps in the solution.

Blyth.  Match-Stick Magic.  1921.  Matchstick circle transfer, pp. 81‑82.  3 and 3  in 15 moves.

Hummerston.  Fun, Mirth & Mystery.  1924.  The frolicsome frogs, Puzzle no. 2, pp. 17 & 172.  Two  3 & 3  problems with the boards crossing at the centre cell.  He notes that the easiest solution is to solve the boards one at a a time.  He says: "It is not good play to jump a counter over another of the same colour."

Lynn Rohrbough, ed.  Socializers.  Handy Series, Kit G, Cooperative Recreation Service, Delaware, Ohio, 1925.  Six Frogs, p. 5.  Dudeney's 1917 problem done in 11 moves.

Botermans et al.  The World of Games.  Op. cit. in 4.B.5.  1989.  P. 235 describes this as The Sphinx Puzzle, "very popular around the turn of the century, particularly in the United States and France" and they show an example of the period labelled The Sphinx and Pyramid Puzzle -- An Egyptian Novelty.

Haldeman-Julius.  1937.  No. 162: Checker problem, pp. 18 & 29.  3 & 3.

See Harbin in 5.R.4 for a 1963 example.

Doubleday - 1.  1969. 

Prob. 77: Square dance, pp. 93 & 171.  = Doubleday - 5, pp. 103-104.  Start with  _WWWBBB.  He says they must change places, with a piece able to move into the vacant space by sliding (either way) or by jumping one or two pieces of any colour.  Asks for a solution in 10 moves.  His solution gets to  BBBWWW_,  which does not seem to be 'changing places' to me.

Prob. 79: All change, pp. 95 & 171.  = Doubleday - 5, pp. 105-106.             BB_WW

Start with the pattern at the right and change the whites and                        BB_WW

blacks in 10 moves, where a piece can slide one place into an

adjacent vacant square or jump one or two pieces into a vacant square.  However, the solution simply does each row separately.

Katharina Zechlin.  (Dekorative Spiele zum Selbermachen; Verlag Frech,                              WWWWW

          Stuttgart-Botnang, 1973.)  Translated as:  Making Games in Wood  Games                  BWWWW

          you can build yourself.  Sterling, 1975, pp. 24-27: The chess knight game.                     BBOWW

          5 x 5  board with 12 knights of each of two colours, arranged as at the right.                   BBBBW

          The object is to reverse them by knight's moves.  Says it can be done within                    BBBBB

          50 moves and 'is almost impossible to do it in less than 45'.

Wickelgren.  How to Solve Problems.  Op. cit. in 5.O.  1974.  Discrimination reversal problem, pp. 78‑81.  _WWWBBB  to  BBBWWW  with the extra place not specified in the goal, with pieces allowed to move into the vacant space by sliding or by hopping over one or two pieces.  Gets to  BBBWW_W  in 9 moves.  [I find it takes 10 moves to get to  BBBWWW_ .]

Joe Celko.  Jumping frogs and the Dutch national flag.  Abacus 4:1 (Fall 1986) 71-73.  Same as Wickelgren.  Celko attributes this to Dudeney.  Gives a solution to  BBBWWW_  in 10 moves and asks for results for higher numbers.

Johnston Anderson.  Seeing induction at work.  MG 75 (No. 474) (Dec 1991) 406-414.  Example 2: Frogs, pp. 408-411.  Careful proof that  BB...BB_WW...WW  to  WW...WW_BB...BB,  with  n  counters of each colour, requires  n2 + 2n  moves.

 

          5.R.3. FORE AND AFT  --  3 BY 3  SQUARES MEETING AT A CORNER

 

          This is Frogs and Toads on part of the  5 x 5  board consisting of two  3 x 3  subarrays at diagonally opposite corners.  They overlap in the central square.  One square has 8 black men and the other has 8 white men, with the centre left vacant.

 

Ball.  MRE, 1st ed., 1892, pp. 51‑52.  51 move solution.  In the third ed., 1896, pp. 69‑70, he says he believes he was the first to publish the puzzle but "that it has been since widely distributed in connexion with an advertisement and probably now is well known".  He gives a 48 move solution.

Hoffmann.  1893.  Chap. VI, no. 26: The "English Sixteen" puzzle, pp. 273‑274 & 287 = Hoffmann-Hordern, pp. 188-189, with photo.  Mentions that it is produced by Messrs Heywood, as below.  Solution in 52 moves, which he believes is minimal.  Hordern notes that the minimum is 46.  Photo on p. 188 of the Heywood version, see next entry.

John Heywood, Manchester, produced a version called 'The English Sixteen Puzzle', undated, but by 1893 as Hoffmann cites it.  Photo in Hoffmann-Hordern, p. 188, dated 1880‑1895.

Charles A. Emerson.  US Patent 522,250 -- Puzzle.  Applied: 3 Nov 1893;  patented: 3 July 1894.  2pp + 1p diagrams.  The Fore and Aft Puzzle.  Says it can be done in  48, 49, 50, 51 or 52  moves.

Dudeney.  Problem 66: The sixteen puzzle.  Tit‑Bits 33 (1 Jan  &  5 Feb 1898) 257  &  355.  "It was produced, I believe, in America, many years ago, and has since been issued over here in the form of an advertisement by a prominent commercial house."  Solution in 46 moves.  He says published solutions assert the minimum number of moves is  53, 52 or 50.  The 46 move solution is given in Ball, MRE, 5th ed., 1911, 79‑80.

Ball.  MRE, 5th ed., 1911, pp. 79-80.  Drops his historical claims and includes a 46 move solution due to Dudeney.

Loyd.  Fore and aft puzzle.  Cyclopedia, 1914, pp. 108 & 353 (solution misprinted, but claimed to be 47 moves in contrast to 52 move solutions 'in the puzzle books'.)  (c= MPSL1, prob. 4, pp. 3‑4 & 121 (only referring to Dudeney's 46 move solution)).

Loyd Jr.  SLAHP.  1928.  A joke on granddad, pp. 29 & 93.  Says 'our granddaddies, who used to play this puzzle game 75 years ago, when it was universally popular.  The old‑time books explain how the solution is accomplished in 52 moves, "the shortest possible method."'  He then asks for and gives a 46 move solution.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 24, pp. 17 & 132: "General post".  Gives a solution which takes 46 moves, but gives no discussion of it.

Rohrbough.  Puzzle Craft.  1932.  Migration (or Fore and Aft), p. 12 (= p. 15 of 1940s?).  Says it was popular 75 years ago and it has recently been shown that it can be done in 46 moves, then gives a solution which stops at 42 moves!

M. Gardner.  SA (Sep 1959) = 2nd Book, pp. 210‑219.  Discusses the puzzle.  On pp. 218‑219, he gives Dudeney's 46 move solution and says 48 different solutions and several proofs that 46 is minimal were sent to him.

Uwe Schult.  Das Seemanns‑Spiel: Mathematisch erledigt.  Reported in Das Mathematische Kabinett column, Bild der Wissenschaft 19:11 (Nov 1982) 181-184.  (A version is given in Neues aus dem Mathematischen Kabinett, ed. by Thiagar Devendran, Hugendubel, Munich, 1985, pp. 102‑103.)  There are  218,790  possible patterns of the pieces.  Reversing black and white takes  46  moves and there are  1026  different halfway positions that can occur in a  46  move solution.  There are two patterns which require  47  moves, namely, after reversing black and white, put one of the far corner pieces in the centre.

Nob Yoshigahara, postcard to me on 18 Aug 1994, announces he has found the worst solution -- in 58 moves.

 

          5.R.4. REVERSING FROGS AND TOADS:  _12...n  TO  _n...21 , ETC.

 

          A piece can slide into the empty cell or jump another piece into the empty cell.

 

Dudeney.  AM.  1917. 

Prob. 214: The six frogs, pp. 59 & 193.  Case of  n = 6,  solved in  21  moves, which he says is minimal.  In general, the minimal solution takes  n(n+1)/2  moves, including  n  steps, when  n  is even  and  (n2+3n-8)/2  moves, including  2n-4  steps, when  n  is odd.  "This complete general solution is published here for the first time."

Prob. 215: The grasshopper puzzle, pp. 59 & 193-194.  Problem for a circular arrangement.  Example has  n = 12.  Says he invented it in 1900.  Solvable in  44  moves.  General solution is complex -- he says that for  n > 4,  it can be done in  (n2+4n-16)/4  moves when  n  is even and in  (n2+6n-31)/4  moves when  n  is odd.

Rohrbough.  Puzzle Craft.  1932.  The Reversible Frogs, p. 22 (= The Jumping Frogs, pp. 20‑21 of 1940s?).  n = 8, citing Dudeney, AM.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  Hopover, p. 89.  First gives  3 and 3  Frogs and Toads, then asks for complete reversal from  123_456  to  654_321.

[Henry] Joseph and Lenore Scott.  Master Mind Pencil Puzzles.  Tempo Books (Grosset & Dunlap), NY, 1973 (& 1978?? -- both dates are given -- I'm presuming the 1978 is a 2nd ptg or a reissue under a different imprint??).  Reverse the numbers, pp. 117‑118.  Give the problem for  n = 6  and a solution in  21  moves.  For  n  even, the method gives a solution in  n(n+1)/2,  it is not shown that this is optimal, nor is a general method given for odd  n.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  13-hour clock, pp. 43-44.  Case  n = 12  considered in a circle can be done in  44  moves.

Joe Celko.  Jumping frogs and the Dutch national flag.  Abacus 4:1 (Fall 1986) 71-73.  Cites Dudeney and gives the results.

Jim Howson.  The Computer Weekly Book of Puzzlers.  Computer Weekly Publications, Sutton, Surrey, 1988, unpaginated.  [The material comes from his column which started in 1966, so an item may go back to then.]  Prob. 54 -- same as the Scotts in Master Mind Pencil Puzzles.

 

          5.R.5. FOX AND GEESE, ETC.

 

          There are a number of similar games on different boards -- too many to describe completely here, so I will generally just cite extensive descriptions.  See any of the main books on games mentioned at the beginning of 4.B, such as Bell or Falkener.  The key feature is that one side has more, but weaker, pieces.  These are sometimes called hunt games.  The standard Fox and Geese is played on a 33 hole Solitaire board, with diagonal moves allowed.  I have recently acquired but not yet read Murray's History of Board Games other than Chess which should have lots of material.

 

Gretti's Saga, late 12C.  Mention of Fox and Geese.  Also in Edward IV's accounts.  ??NYS -- cited by Botermans et al, below.

Shackerley (or Schackerley or Shakerley) Marmion.  A Fine Companion (a play).  1633.  IN: The Dramatic Works of Shackerley Marmion; William Paterson, Edinburgh  &  H. Sotheran & Co., London,  1875.  II, v, pp. 140-141.  "..., let him sit in the shop ..., and play at fox and geese with the foreman, ....."  Earliest English occurrence of fox-and-geese.  Quoted by OED and cited by Fiske, below.

Richard Lovelace.  To His Honoured Friend On His Game of Chesse-Play  or  To Dr. F. B. on his Book of Chesse.  1656?, published in his Posthume Poems, 1659.  Lines 1-4.  My edition of Lovelace notes that F. B. was Francis Beale, author of 'Royall Game of Chesse Play,' 1656.  Lovelace died in 1658.

                    Sir, now unravell'd is the Golden Fleece,

                    Men that could onely fool at Fox and Geese,

                    Are new-made Polititians by the Book,

                    And can both judge and conquer with a look.

Henry Brooke.  Fool of Quality.  [A novel.]  1766-1768.  Vol. I, p. 367.  ??NYS -- quoted by Fiske, below.  "Can you play at no kind of game, Master Harry?"  "A little at fox‑and‑geese, madam."

Catel.  Kunst-Cabinet.  1790. 

Das Fuchs- und Hühnerspiel, pp. 51-52 & fig. 168 on plate VI.  11 chickens against one fox on a  4 x 4  board with all diagonals drawn, giving  16 + 9  playing points.

Das Schaaf- und Wolfspiel, p. 52 & fig. 169 on plate VI, is the same game on the 33-hole solitaire board with 11 sheep and one wolf, no diagonals

Bestelmeier.  1801.

Item 83: Das Schaaf- und Wolfspiel.  Same diagram and game as Catel, p. 52.

Item 833: Ein Belagerungspiel.  33 hole board with a fortress on one arm, with diagonals drawn.

Strutt.  Op. cit. in 4.B.1.  Fox and Geese.  1833: Book IV, chap. II, art. XIV, pp. 318‑319.  = Strutt-Cox, p. 258 & plate opp. p. 246.  Fig. 107 (= plate opp. p. 246) shows the 33 hole board with its diagonals drawn.

Gomme.  Op. cit. in 4.B.1.  I 141‑142 refers to Strutt and Micklethwaite.

Illustrated Boy's Own Treasury.  1860.  Fox and Geese, pp. 406‑407.  33 hole Solitaire board with diagonals drawn.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 320, p. 152: Fuchs und Gänse.  Shows 33 hole solitaire board with diagonals drawn.

Stewart Culin.  Chinese Games with Dice and Dominoes.  From the Report of the U. S. National Museum for 1893, pp. 489‑537.  Pp. 874-877 describes:  the Japanese game of  Juroku Musashi  (Sixteen Soldiers)  with 16 men versus a general;  the Chinese game of  Shap luk kon tséung kwan  (The sixteen pursue the commander);  another Chinese game of  Yeung luk sz' kon tséung kwan  with 27 men against a commander (described by Hyde -- ?? I didn't see this);  the Malayan game of  Dam Hariman  (Tiger Game),  identical to the Hindu game of  Mogol Putt'han  (= Mogul Pathan (Mogul against Pathan)),  similar to a Peruvian game of Solitario and the Mexican game of Coyote;  the Siamese game of  Sua ghin gnua  (Tiger and Oxen)  and the similar Burmese game of  Lay gwet kyah,  with three big tigers versus 11 or 12 little tigers;  the Samoan game of  Moo;  the Hawaiian game of Konane;  a similar Madagascarian game;  the Hindu game of  Pulijudam  (Tiger Game) with three tigers versus 15 lambs.

Fiske.  Op. cit. in 4.B.1.  1905.  Fox-and-Geese, pp. 146-156 & 359, discusses the history of the game, especially as to whether it is identical to the old Norse game of Hnefatafl.  On p. 359, he says that John of Salisbury (c1150) used 'vulpes' as the name of a game, but there is no indication of what it was.  He says "the fox-and-geese board, in comparatively modern times, has begun to be used for games more or less different in their nature, especially for one called in England solitaire and in France "English solitaire", and for another, known in Spain and Italy as asalto (assalto), in French as assaut, in Danish as belejringsspel."  He then surveys the various sources that he treated under Mérelles -- see 4.B.1 and 4.B.5 for details.  He is not sure that Brunet is really describing the game in the Alfonso MS (op. cit. in 4.B.5 and below).  He cites an 1855 Italian usage as Jeu de Renard or Giuoco della Volpe.  In Come Posso Divertirmi? (Milan, 1901, pp. 231-233), it is said that the game is usually played with 17 geese rather than 13 -- Fiske notes that this assertion is of "some historical value, if it be true."  Moulidars calls it Marelle Quintuple, quotes Maison des Jeux Académiques (Paris, 1668) for a story that it was invented by the Lydians and gives the game with 13 or 17 geese.  Asalto has 2 men against 24.  Fiske quotes Shackley Marmion, above, for the oldest English occurrence of fox-and-geese and then Henry Brooke, above.  Fiske follows with German, Swedish and Icelandic (with 13 geese) references.

H. Parker.  Ancient Ceylon.  Op. cit. in 4.B.1, 1909.  Pp. 580-583 & 585 describe four forms of The Leopards Game, with one tiger against seven leopards, three leopards against 15 dogs, two leopards against 24 cattle and one leopard against six cattle on a  12 x 12  board.  The first two are played on a triangular board. 

Robert Kanigel.  The Man Who Knew Infinity.  A Life of the Genius Ramanujan.  (Scribner's, NY, 1991);  Abacus (Little, Brown & Co. (UK)), London, 1992.  Pp. 18 & 377:  Ramanujan and his mother used to play the game with three tigers and fifteen goats on a kind of triangular board.

The Spanish Treatise on Chess-Play written by order of King Alfonso the Sage in the year 1283.  [= Libro de Acedrex, Dados e Tablas of Alfonso El Sabio, generally known as the Alfonso MS.]  MS in Royal Library of the Escorial (j.T.6. fol).  Complete reproduction in 194 Phototypic Plates.  2 vols.,  Karl W. Hirsemann, Leipzig, 1913.  See 4.B.5 for more details of this work.  See below.

Botermans et al.  The World of Games.  Op. cit. in 4.B.5.  1989.  P. 147 says De Cercar La Liebre (Catch the Hare) occurs in the Alfonso MS and is the earliest example of a hunt game in European literature, but undoubtedly derived from an Arabic game of the Alquerque type -- I didn't see this when I briefly looked at the facsimile -- ??NYS.  They say Murray has noted that hunt games are popular in Asia, but not in Africa, leading to the conjecture that they originated in Asia.  They describe it on a  5 x 5  array of points with verticals and horizontals and some diagonals drawn, with one hare against 12 hunters. 

                    Botermans et al. continue on pp. 148-155 to describe the following.

Shap Luk Kon Tseung Kwan (Sixteen Pursue the General) played on a  5 x 5  board like Catch the Hare with an extra triangle on one side and capturing by interception.

Yeung Luk Sz'Kon Tseung Kwan, seen in Nanking by Hyde and described by him in 1694, somewhat similar to the above, but with 26 rebels against a general.  (??NYS)

Fox and Geese, mentioned in Gretti's Saga of late 12C and in Edward IV's accounts.  They give a version called Lupo e Pecore from a 16C Venetian book, using a Solitaire board extended by three points on each arm, giving 45 points.  They give a 1664 engraving showing Le Jeu du Roi which they say is a rather complex form of fox and geese, but looks like a four-handed game on a cross-shaped board with  7 x 5  arms on a  7 x 7  central square and 4 groups of  7 x 4  men.

Leopard games, from Southeast Asia, with a kind of triangular board.  Len Choa, from Thailand, has a tiger against six leopards.  Hat Diviyan Keliya, from Sri Lanka, has a tiger against seven leopards.

Tiger games, also from Southeast Asia, are similar to leopard games, but use an extended Alquerque board (as in Catch the Hare).  Rimau (Tiger), from Malaysia, has 24 men versus a tiger and Rimau-Rimau (Tigers) is a version with two tigers versus 22 men.

Murray.  1913.

P. 347 cites a 1901 Indian book for 2 lions against 32 goats on a chessboard.

P. 371 cites a Soyat (North Asia) example (19C?) of Bouge‑Shodra (Boar's Chess) with 2 boars against 24 calves on a chessboard.

Pp. 569 & 616‑617 cite the Alfonso MS of 1283 for 'De cercar la liebre', played on a  5 x 5  board with  10, 11 or 12  men against a hare.

P. 585 shows Cott. 6 (c1275) of 8 pawns against a king on a chessboard.

Pp. 587 & 590 give Cott. 11 = K6: Le Guy de Alfins with king and 4 bishops against a king on a chessboard.

Pp. 589-590 shows K4 = CB249: Le Guy de Dames and No. 5 = K5: Le Guy de Damoyselles, which have 16 pawns against a king on a chessboard.

P. 617 discusses Fox and Geese, with 13, 15 or 17 geese against a fox on the solitaire board.  Edward IV, c1470, bought "two foxis and 46 hounds".  Murray says more elaborate forms exist and refers to Hyde and Fiske (see 4.B.1 and 5.F.1 for more on these), ??NYS.

Pp. 675 & 692 show CB258: Partitum regis Francorum with king and four pawns against king on the chessboard.  It says the first side wins.

P. 758 describes a 16C Venetian board (then) at South Kensington (V&A??) with the Solitaire board for Fox and Geese and an enlarged board for Fox and Geese.

P. 857 mentions Fox and Geese in Iceland.

Family Friend 2 (1850) 59.  Fox and geese.  4 geese against 1 fox on a chess board.

The Sociable.  1858.  Fox and geese, p. 281.  17 geese against a fox on the solitaire board.  Four men versus a king on the draughts board, saying the first side wins even allowing the king to be placed anywhere against the men who start on one side.

Stewart Culin.  Korean Games, op. cit. in 4.B.5, 1895.  Pp. 76-77 describes some games of this type, in particular a Japanese game called Yasasukari Musashi with 16 soldiers versus a general on a  5 x 5  board, taken from a 1714 (or 1712) Japanese book:  Wa Kan san sai dzu e "Japanese, Chinese, Three Powers picture collection", published in Osaka.

Anonymous.  Enquire Within upon Everything.  66th ed., 862nd thousand, Houlston and Sons, London, 1883, HB.  Section 2593: Fox and Geese, p. 364.  33 hole Solitaire board with 17 geese against a fox.  4 geese against a fox on the chessboard.  Says the geese should win in both cases.

Slocum.  Compendium.  Shows Solitaire and Solitaire & Tactic Board from Gamage's 1913 catalogue.  Like Bestelmeier's 833, but without diagonals.

Bell & Cornelius.  Board Games Round the World.  Op. cit. in 4.B.1.  1988.  Games involving unequal forces, pp. 43-52.  Discusses the following.

The Maharajah and the Sepoys.  1 against 16 on a chessboard.

Fox and Geese.  Cites an Icelandic work of c1300 (probably Gretti's Saga?).  1 against 13 or 17 on a Solitaire board.

Lambs and Tigers, from India.  3 against 15.

Cows and Leopards, from SE Asia.  2 against 24.

Vultures and Crows, also called Kaooa, from India.  1 against 7 on a pentagram board.

The New Military Game of German Tactics, c1870.  2 against 24 on a Solitaire Board with a fortress, as in Bestelmeier.

Yuri I. Averbakh.  Board games and real events.  IN: Alexander J. de Voogt, ed.; New Approaches to Board Games Research: Asian Origins and Future Perspectives; International Institute for Asian Studies, Leiden, 1995; pp. 17-23.  Notes that Murray believes hunt games evolved from war games, but he feels the opposite is true.  He describes a Nepalese game of Baghachal with four tigers versus 20 goats -- this is Murray's 5.6.22.  He corrects some of Murray's assertions about Boar Chess and describes other Tuvinian hunt games: Bull's Chess and Calves' Chess, probably borrowed from the Mongols.  The latter has a three-in-a-row pattern and he wonders if there is some connection with morris or noughts and crosses (which he says is "played everywhere").  He mentions Cercar la Liebre from the Alfonso MS.  Fox and Geese type games are mentioned in the Icelandic sagas as 'the fox game'.  He describes several forms.

 

          5.R.6. OCTAGRAM PUZZLE

 

          One has an octagram and seven men.  One has to place a man on a vacant point and then slide him to an adjacent vacant point, then do the same with the next man, ...,  so as to cover seven of the points.  The diagram is just an 8-cycle and is the same as the knight's connections on the  3 x 3  board, so the octagram puzzle is equivalent to the 7 knights problem mentioned in 5.F.1.  Further, the 4 knights problem of 5.F.1 has the same 8-cycle, with men at alternate points of it.

          Versions with different numbers of points.

  5 points:  Rohrbough.

  7 points:  Mittenzwey; Meyer.

  9 points:  Dudeney.

10 points:  Bell & Cornelius; Hoffmann; Cohen; Williams; Toymaker; Rohrbough; Putnam.

13 points:  Berkeley & Rowland.

 

Bell & Cornelius.  Board Games Round the World.  Op. cit. in 4.B.1.  1988.  Pentalpha, p. 15.  Says that a pentagram board occurs at Kurna, Egypt, c-1400 and that the solitaire game of Pentalpha is played in Crete.  This has 9 men to be placed on the vertices and the intersections of the pentagram.  Each man must be placed on a vacant point, then slid ahead two positions along one straight line.  The intermediate point may be occupied, but the ending point must be unoccupied.  Unfortunately we don't know if the Egyptian board was used for this game.

Pacioli.  De Viribus.  c1500. 

Ff. 112r - 113v.  .C(apitolo). LXVIII. D(e). cita ch' a .8. porti ch' cosa convi(e)ne arepararli (Chap. 68.  Of a city with 8 gates which admits of division ??).  = Peirani 158-160.  Octagram puzzle with a complex story about a city with 8 gates and 7 disputing factions to be placed at the gates.

F. IVv.  = Peirani 8.  The Index gives the above as Problem 83.  Problem 82: De .8. donne ch' sonno aun ballo et de .7. giovini quali con loro sa con pagnano (Of 8 ladies who are at a ball and of 7 youths who accompany them).

Schwenter.  1636.  Part 2, exercise 36, pp. 149-150.  Octagram.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 4, pp. 224-225.  Octagram, taken from Schwenter.

Les Amusemens.  1749.  P. xxxiii. 

Catel.  Kunst-Cabinet.  1790.  Das Achteck, pp. 12-13 & fig. 36 on plate II.  The rules are not clearly described.

Bestelmeier.  1801.  Item 290: Das Achteckspiel.  Text copies part of Catel.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  Ff. 131-133 are an analysis of the heptagram puzzle.

Rational Recreations.  1824.  Feat 34, pp. 161-164.  Octagram.

Endless Amusement II.  1826?  Prob. 28, pp. 203-204.  = New Sphinx, c1840, pp. 137-138.

Nuts to Crack IV (1835), no. 194 -- part of a long section called Tricks upon Travellers.

Family Friend (Dec 1858) 359.  Practical puzzles -- 1.  I don't have the answer.

The Boy's Own Magazine 3 (1857) 159 & 192.  Puzzle of the points.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles No. 6, pp. 396 & 436.

The Secret Out.  1859.  To Place Seven Counters upon an Eight-Pointed Star, pp. 373-374.

J. J. Cohen, New York.  Star puzzle.  Advertising card for Star Soap, Schultz & Co., Zanesville, Ohio, Copyright May 1887.  Reproduced in: Bert Hochberg; As advertised  Puzzles from the collection of Will Shortz; Games Magazine 17:1 (No. 113) 10-13, on p. 11.  Identical to pentalpha - see Bell & Cornelius above.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. IX: The reversi puzzle, pp. 8-10.  Version with 13 cards in a circle and one can move ahead by any number of steps.  If there are  x  cards and one moves ahead  s  steps, then  x  and  s  must have no common factor.

Hoffmann.  1893.  

     Chap. VI, pp. 267-268 & 280-281 = Hoffmann-Hordern, pp. 180-181, with photos.

No. 13.  No name.  Basic octagram puzzle.  Photo on p. 181 shows: The Seven Puzzles, by W. & T. Darton, dated 1806-1811; a Tunbridge ware version dated 1825-1840; and Jeu de Zig‑Zag, by M. D., Paris, 1891-1900.

No. 14.  The "Okto" puzzle, pp. 268 & 281.  Here the counters and points are coloured.  Photo on p. 181 of The "Okto" Puzzle by McGaw, Stevenson & Orr, Ltd. for John Stewart, dated 1880-1895.

     Chap. X, no. 8: Crossette, pp. 337 & 374-375 = Hoffmann-Hordern, pp. 229-230, with photo.  10 counters in a circle.  Start anywhere and move ahead three.  Photo on p. 230 shows The Mystic Seven, a seven counter version, by the Lord Roberts Workshops, 1914-1920.

Mittenzwey.  1895?  Prob. 329, pp. 58 & 106;  1917: 329, pp. 52 & 101.  Heptagram.

Dudeney.  Problem 58: A wreath puzzle.  Tit‑Bits 33 (6  &  27 Nov 1897) 99  &  153.  Complex nonagram puzzle involving moves in either direction and producing the original word again.

Clark.  Mental Nuts.  1897, no. 54;  1904, no. 80;  1916, no. 69.  A little puzzle.  Usual octagram.

Benson.  1904.  The eight points puzzle, pp. 250‑251.  c= Hoffmann, no. 13.

Slocum.  Compendium.  Shows the "Octo" Star Puzzle from Gamage's 1913 catalogue.

Williams.  Home Entertainments.  1914.

Crossette, pp. 115-116.  Ten points, advancing three places.

Eight points puzzle, pp. 120-121.  Usual octagram.

"Toymaker".  Top in Hole Puzzle.  Work (23 Dec 1916) 200.  10 holes and one has to move to the third position and reverse the top in that hole.

Blyth.  Match-Stick Magic.  1921.  Crossing the points, pp. 83-84.

Hummerston.  Fun, Mirth & Mystery.  1924. 

The sacred seven, Puzzle no. 5, pp. 26 & 173.  Octagram puzzle on the outer points of the diagram shown in 5.A.

The four rabbits, Puzzle no. 6, pp. 26 & 173.  Using the octagram shown in 5.A, put black counters on locations  1 and 2  and white counters on  7 and 8.  The object is to interchange the colours.  This is like the 4 knights problem except the corresponding 8-cycle has men at positions  1, 2, 5, 6.  He counts a sequence of steps by the same man as a move and hence solves it in 6 moves (comprising 16 steps).

Will Blyth.  Money Magic.  C. Arthur Pearson, London, 1926.  Turning the tails, pp. 66-69.  8 coins in a circle, tails up.  Count from a tail four ahead and reverse that coin.  Get 7 heads up.  Counting four ahead means that if you start at 1, you count  1, 2, 3, 4  and reverse 4.

King.  Best 100.  1927.  No. 64, pp. 26-27 & 54.

Rohrbough.  Puzzle Craft.  1932. 

Count 4, p. 6.  10 points on a circle, moving ahead 3.  (= Rohrbough; Brain Resters and Testers; c1935, p. 21.)

Star Puzzle, p. 8 (= p. 10 of 1940s?).  Consider the pentagram with its internal vertices.  First puzzle is Pentalpha.  Second is to place a counter and move ahead three positions.  The object is to get four counters on the points, which is the same as the pentagram puzzle, moving one position.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 18: Odd man out, pp. 27 & 184.  Version with 7 positions in a circle and 6 men where one must place a man and then move him three places ahead.

Putnam.  Puzzle Fun.  1978.  No. 63: Ten card turnover, pp. 11 & 35.  Ten face down cards in a circle.  Mark a card, count ahead three and turnover.

 

          5.R.7. PASSING OVER COUNTERS

 

          The usual version is to have 8 counters in a row which must be converted to 4 piles of two, but each move must pass a counter over two others.  Martin Gardner pointed out to me that the problem for  10, 12, 14, ...  counters is easily reduced to that for 8.  The problem is impossible for  2, 4, 6.  There are many later appearances of the problem than given here.  In describing solutions,  4/1  means move the 4th piece on top of the 1st piece.

          There are trick solutions where a counter moves to a vacated space or even lands between two spaces.  See:  Mittenzwey; Haldeman-Julius; Hemme.

          Berkeley & Rowland give a problem where each move must pass a counter over two piles.  This makes the problem easier and it is solvable for any even number of counters  ³ 6,  but it gives more solutions.  See: Berkeley & Rowland; Wood; Indoor Tricks & Games; Putnam; Doubleday - 1. 

          One could also permit passing over one pile, which is solvable for any even number  ³ 4.

          Mittenzwey, Double Five Puzzle, Hummerston, and Singmaster & Abbott deal with the problem in a circle and with piles to be left in specific locations.

          Mittenzwey, Lucas and Putnam consider making piles of three by passing over 3, etc.

 

Kanchusen.  Wakoku Chiekurabe.  1727.  Pp. 38-39.  Jukkoku-futatsu-koshi (Ten stones jumping over two).  Ten counters, one solution.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has "The question of the shillings passing at each time over two or a certain number   8 is the least number   Any number being given and any law of transit  Dr Roget"  The layout suggests that Roget had posed the general version.  Adjacent is a diagram with a row of 10 counters and the first move 1 to 4 shown, but with some unclear later moves.

Endless Amusement II.  1826?  Prob. 10, p. 195.  10 halfpence.  One solution: 4/1 7/3 5/9 2/6 8/10.  = New Sphinx, c1840, pp. 135-135.

Nuts to Crack II (1833), no. 122.  10 counters, identical to Endless Amusement II.

Nuts to Crack V (1836), no. 68.  Trick of the eight sovereigns.  Usual form.

Young Man's Book.  1839.  P. 234.  Ingenious Problem.  10 halfpence.  Identical to Endless Amusement II.

Family Friend 2 (1850) 178 & 209.  Practical Puzzle, No. VI.  Usual form with eight counters or coins.  One solution.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 2, p. 176 (1868: 187).  Passing over coins.  Gives two symmetric solutions.

Magician's Own Book.  1857.  Prob. 34: The counter puzzle, pp. 277 & 300.  Identical to Book of 500 Puzzles, prob. 34.

The Sociable.  1858.  Prob. 16: Problem of money, pp. 291-292 & 308.  Start with 10 half‑dimes, says to pass over one, but solution has passing over two.  One solution.  = Book of 500 Puzzles, 1859, prob. 16, pp. 9-10 & 26.

Book of 500 Puzzles.  1859.

Prob. 16: Problem of money, pp. 9-10 & 26.  As in The Sociable.

Prob. 34: The counter puzzle, pp. 91 & 114.  Eight counters, two solutions given.  Identical to Magician's Own Book.

The Secret Out.  1859.  The Crowning Puzzle, p. 386.  'Crowning' is here derived from the idea of crowning in draughts or checkers.  One solution:  4/1 6/9 8/3 2/5 7/10.

Boy's Own Conjuring Book.  1860.

Prob. 33: The counter puzzle, pp. 240 & 264.  Identical to Magician's Own Book, prob. 34.

The puzzling halfpence, p. 342.  Almost identical to The Sociable, prob. 16, with half-dimes replaced by halfpence.

Illustrated Boy's Own Treasury.  1860.  Prob. 17, pp. 398 & 438.  Same as prob. 34 in Magician's Own Book but only gives one solution.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 593, part 6, pp. 299-411: Sechs Knacknüsse.  10 counters, one solution.

Hanky Panky.  1872.  Counter puzzle, p. 132.  Gives two solutions for 8 counters and one for 10 counters.

Kamp.  Op. cit. in 5.B.  1877.  No. 12, p. 325.

Mittenzwey.  1880.  Prob. 235-238, pp. 43-44 & 93-94;  1895?: 262-266, pp. 47-48 & 95-96;  1917: 262-266, pp. 43-44 & 91.

235 (262).  Usual problem, with 10 counters.  Two solutions.

--- (263).  Added in 1895?  Same with 8 counters.  Two solutions.

236 (264).  12 numbered counters in a circle.  Pass over two to leave six piles of two on the first six positions.  Solution is misprinted in all editions!

237 (265).  12 counters in a circle.  Pass over three to leave six piles of two, except the last move goes over six.  The solution allows landing counters on vacated locations!

238 (266).  15 counters in a row.  Pass over 3 to leave five piles of three.  The solution allows landing a counter on a vacated location and landing a counter between two locations!!

Lucas.  RM2.  1883.  Les huit pions, pp. 139-140.  Solves for  8, 10, 12, ...  counters.  Says Delannoy has generalized to the problem of  mp  counters to be formed into  m  piles  (m ³ 4)  of  p  by passing over  p  counters. 

                    [More generally, using one of Berkeley & Rowland's variations (see below), one can ask when the following problem is solvable:  form a line of  n = kp  counters into  k  piles of  p  by passing over  q  [counters or piles].  Does  q  have to be  £ p?]

Double Five Puzzle.  c1890.  ??NYS -- described by Slocum from his example.  10  counters in a circle, but the final piles must alternate with gaps, e.g. the final piles are at the even positions.  This is also solvable for  4, 8, 12, 16, ...,  and I conjectured it was only solvable for  4n  or  10  counters -- it is easy to see there is no solution for  2  or  6  counters and my computer gave no solutions for  14  or  18.  For  4, 8, 10 or 12  counters, one can also leave the final piles in consecutive locations, but there is no such solution for  6, 14, 16 or 18  counters.  See Singmaster & Abbott, 1992/93, for the resolution of these conjectures.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles.

No. VII: The halma puzzle, pp. 6-7.  Arrange the first ten cards of a suit in a row so that passing over two cards leaves five piles whose cards total 11 and are in the odd places.  Arrangement is  7,6,3,4,5,2,1,8,9,10.  Move  2 to 9,  4 to 7,  8 to 3,  6 to 5,  10 to 1.

No. VIII: Another version, p. 7.  With the cards in order and passing over two piles, leave five piles of two.  But this is so easy, he adds that one wants to leave as low a total as possible on the tops of the piles.  He moves  7 to 10,  6 to 3,  4 to 9,  1 to 5,  2 to 8,  leaving a total of 20. 

           [However, this is not minimal -- there are six solutions leaving 18 exposed, e.g.  1 to 4,  3 to 6,  7 to 10,  5 to 9,  2 to 8.  For 6 cards, the minimum is 6, achieved once;  for 8 cards, the minimum is 11, achieved 3 times;  for 12 cards, the minimum is 27, achieved 10 times.  For the more usual case of passing over two cards, the minimum for 8 cards is 15, achieved twice;  for 10 cards, the minimum is 22, achieved 4 times, e.g. by  7 to 10,  5 to 2,  3 to 8,  1 to 4,  6 to 9;  for 12 cards, the minimum is 31, achieved 6 times;  for 14 cards, the minimum is 42, achieved 8 times.  For passing over one pile, the minimum for 4 cards is 3, achieved once;  the minimum for 6 cards is 7, achieved twice;  the minimum for 8 cards is 13, achieved 3 times;  for 10 cards, the minimum is 21, achieved 4 times;  the minimum for 12 cards is 31, achieved 5 times.  Maxima are obtained by taking mirror images of the minimal solutions.]

Puzzles with draughtsmen.  The Boy's Own Paper 17 or 18?? (1894??) 751.  8 men, passing over two men each time.  Notes that it can be extended to any even number of counters.

Clark.  Mental Nuts.  1897, no. 63: Toothpicks;  1904, no. 83  &  1916, no. 70:  Place 8 toothpicks in a row.  One solution.

Parlour Games for Everybody.  John Leng, Dundee & London, nd [1903 -- BLC], p. 32: The five pairs.  10 counter version, one solution.

Wehman.  New Book of 200 Puzzles.  1908.  P. 15: The counter puzzle  and  Problem of money.  8 and 10 counter versions, the latter using pennies.  Two and one solutions.

Ahrens.  MUS I.  1910.  Pp. 15-17.  Essentially repeats Lucas.

Manson.  1911.  Decimal game, pp. 253-254.  Ten rings on pegs.  "Children are frequently seen playing the game out of doors with pebbles or other convenient articles."

Blyth.  Match-Stick Magic.  1921.  Straights and crosses, pp. 85-87.  10 matchsticks, one must pass over two of them.  Two solutions, both starting with 4 to 1.

Hummerston.  Fun, Mirth & Mystery.  1924. 

The pairing puzzle, Puzzle no. 8, pp. 27 & 173.  Essential 8 counters in a circle, with four in a row being white, the other four being black.  Moving only the whites, and passing over two, form four piles of two.

Pairing the pennies, Puzzle no. 39, pp. 102 & 178.  Ten pennies, one solution.

Will Blyth.  Money Magic.  C. Arthur Pearson, London, 1926.  Marrying the coins, pp. 113‑115.  Ten coins or eight coins, passing over 2.  Gives two solutions for 10, not noting that the case of 10 is immediately reduced to 8.  Says there are several solutions for 8 and gives two.

Wood.  Oddities.  1927.  Prob. 45: Fish in the basket, pp. 39-40.  12 fish in baskets in a circle.  Move a fish over two baskets, continuing moving in the same direction, to get get two fish in each of six baskets, in the fewest number of circuits.

Rudin.  1936.  No. 121, pp. 43 & 103.  10 matches.  Two solutions.

J. C. Cannell.  Modern Conjuring for Amateurs.  C. Arthur Pearson, London, nd [1930s?].  Straights and crosses, pp. 105-106.  As in Blyth, 1926.

Indoor Tricks and Games.  Success Publishing, London, nd [1930s??]. 

How to pair the pennies, p. 4.  8 pennies, one solution.

The ten rings. p. 4.  10 rings, passing over two piles, one solution.

Haldeman-Julius.  1937. 

No. 91: Jumping pennies, pp. 11 & 25.  Six pennies to be formed into two piles of three by jumping over three pennies each time.  Solution has a trick move.  Jump 1 to 5, 6 to 3 and 2 to 1/5, which gives the position:  6/3  4  2/1/5.  He then says: "No. 4 jumps over 5, 1 and 2 -- then jumps back over 5, 1 and 2 and lands upon 3 and 6, ...."  Since the rules are not clear about where a jumping piece can land, the trick move can be viewed as a legitimate jump to the vacant 6 position, then a legitimate move over the 2/1/5 pile and the now vacant 4 position onto the 6/3 pile.  If the pennies are considered as a cycle, this trick is not needed. 

No. 148: Half dimes, pp. 16 & 143.  10 half dimes, passing over one dime (i.e. two counters).

Sullivan.  Unusual.  1947.  Prob. 39: On the line.  Ten pennies.

Doubleday - 1.  1969.  Prob. 75: Money moves, pp. 91 & 170.  Ten pennies.  Jump over two piles.  Says there are several solutions and gives one, which sometimes jumps over three or four pennies.

Putnam.  Puzzle Fun.  1978.

No. 26: Pile up the coins, pp. 7 & 31.  12 in a row.  Make four piles of three, passing over three coins each time.

No. 27: Pile 'em up again, pp. 7 & 31.  16 in a row.  Make four piles of four, passing over four or fewer each time.

No. 60: Coin assembly, pp. 11 & 35.  Ten in a row, passing over two each time.

No. 61: Alternative coin assembly, pp. 11 & 35.  Ten in a row, passing over two piles each time.

David Singmaster, proposer;  H. L. Abbott, solver.  Problem 1767.  CM 18:7 (1992) 207  &  19:6 (1993).  Solves the general version of the Double Five Puzzle, which the proposer had not solved.  One can leave the counters on even numbered locations if and only if the number of counters is a multiple of 4 or a multiple of 10.  One can leave the counters in consecutive locations if and only if the number of counters is  4, 8, 10 or 12.

Heinrich Hemme.  Email of 25 Feb 1999.  Points out that the rules in the usual version should say that the counter must land on a pile of a single coin.  This would eliminate the trick solutions given by Mittenzwey and Haldeman-Julius.  Hemme says that without this rule, the problem is easy and can be solved for 4 and 6 counters!

 

          5.S.    CHAIN CUTTING AND REJOINING

 

          The basic problem is to minimise the cost or effort of reforming a chain from some fragments.

 

Loyd.  Problem 25: A brace of puzzles -- No. 25: The chain puzzle.  Tit‑Bits 31 (27 Mar 1897)  &  32 (17 Apr 1897) 41.  13 lengths:  5, 6, 7, 7, 7, 7, 8, 8, 8, 8, 8, 9, 12.  (Not in the Cyclopedia.)

Loyd.  Problem 42: The blacksmith puzzle.  Tit‑Bits 32 (10  &  31 Jul  &  21 Aug 1897) 273,  327  &  385.  Complex problem involving 10 pieces of lengths from 3 to 23 to be joined.

Clark.  Mental Nuts.  1897, no. 7  &  1904, no. 14: The chain question;  1916, no. 59: The chain puzzle.  5 pieces of 3 links to make into a single length.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 9:4 (Feb 1903) 390-391  & 9:5 (Mar 1903) 490-491.  The five chains.  5 pieces of 3 links to make into a single length.

Pearson.  1907.  Part II, no. 67, pp. 128 & 205.  5 pieces of 3 links to make into a single length.

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  He attributes such puzzles to Loyd (Tit‑Bits prob. 25) and gives that problem.

Cecil H. Bullivant.  Home Fun.  T. C. & E. C. Jack, London, 1910.  Part VI, Chap IV, No. 9: The broken chain, pp. 518 & 522.  5 3‑link pieces into an open chain.

Loyd.  The missing link.  Cyclopedia, 1914, pp. 222 (no solution) (c= MPSL2, prob. 25, pp. 19 & 129).  6 5‑link pieces into a loop.

Loyd.  The necklace puzzle.  Cyclopedia, 1914, pp. 48 & 345 (= MPSL1, prob. 47, pp. 45‑46 & 138).  12 pieces, with large and small links which must alternate.

D. E. Smith.  Number Stories.  1919.  Pp. 119 & 143‑144.  5 pieces of 3 links to make into one length.

Hummerston.  Fun, Mirth & Mystery.  1924.  Q.E.D. -- The broken chain, Puzzle no. 38, pp. 99 & 178.  Pieces of lengths  2, 2, 3, 3, 4, 4, 6  to make into a closed loop.

Ackermann.  1925.  Pp. 85‑86.  Identical to the Loyd example cited by Dudeney.

Dudeney.  MP.  1926.  Prob. 212: A chain puzzle, pp. 96 & 181 (= 536, prob. 513, pp. 211‑212 & 408).  13 pieces, with large and small links which must alternate.

King.  Best 100.  1927.  No. 7, pp. 9 & 40.  5 pieces of three links to make into one length.

William P. Keasby.  The Big Trick and Puzzle Book.  Whitman Publishing, Racine, Wisconsin, 1929.  A linking problem, pp. 161 & 207.  6 pieces comprising  2, 4, 4, 5, 5, 6  links to be made into one length.

 

          5.S.1. USING CHAIN LINKS TO PAY FOR A ROOM

 

          The landlord agrees to accept one link per day and the owner wants to minimise the number of links he has to cut.  The solution depends on whether the chain is closed in a cycle or open at the ends.  Some weighing problems in 7.L.2.c and 7.L.3 are phrased in terms of making daily payment, but these are like having the chain already in pieces.  See the Fibonacci in 7.L.2.c.

          New section.  I recall that there are older versions.

 

Rupert T. Gould.  The Stargazer Talks.  Geoffrey Bles, London, 1944.  A Few Puzzles -- write up of a BBC talk on 10 Jan 1939, pp. 106-113.  63  link chain with three cuts.  On p. 106, he says he believes it is quite modern -- he first heard it in 1935.  On p. 113, he adds a postscript that he now believes it first appeared in John O'London's Weekly (16 Mar 1935) ??NYS.

Anonymous.  Problems drive, 1958.  Eureka 21 (Oct 1958) 14-16 & 30.  No. 3.  Man has closed chain of 182 links and wants to stay 182 days.  What is the minimum number of links to be opened?

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Pp. 13-16.  Begins with seven link open-ended bracelet.  Then how big a bracelet can be dealt with using only two cuts?  Gets  23.  Then does general case, getting  n + (n+1)(2n+1 - 1).

Angela Fox Dunn.  Second Book of Mathematical Bafflers.  Dover, 1983.  Selected from Litton's Problematical Recreations, which appeared in 1959‑1971.  Prob. 26, pp. 28 & 176.  23  link case.

Howson.  Op. cit. in 5.R.4.  1988.  Prob. 30.  Says a  23  link chain need only be cut twice, giving lengths  1, 1, 3, 6, 12,  which make all values up to  23.  Asks for three cuts in a  63  link chain and the maximum length chain one can deal with in  n  cuts.

 

          5.T.    DIVIDING A CAKE FAIRLY

 

Mittenzwey.  1880.  Prob. 200, pp. 37 & 89;  1895?: 225, pp. 41 & 91;  1917: 225, pp. 38 & 88.  Family of 4 adults and 4 children.  With three cuts, divide a cake so the adults and the children get equal pieces.  He makes two perpendicular diametrical cuts and then a circular cut around the middle.  He seems to mean the adults get equal pieces and the children get equal pieces, not necessarily the same.  But if the circular cut is at  Ö2/2  of the radius, then the areas are all equal.  Not clear where this should go -- also entered in 5.Q.

B. Knaster.  Sur le problème du partage pragmatique de H. Steinhaus.  Annales de la Société Polonaise de Mathématique 19 (1946) 228‑230.  Says Steinhaus proposed the problem in a 1944 letter to Knaster.  Outlines the Banach & Knaster method of one cutting  1/n  and each being allowed to diminish it -- last diminisher takes the piece.  Also shows that if the valuations are different, then everyone can get  > 1/n  in his measure.  Gives Banach's abstract formulations.

H. Steinhaus.  Remarques sur le partage pragmatique.  Ibid., 230‑231.  Says the problem isn't solved for irrational people and that Banach & Knaster's method can form a game.

H. Steinhaus.  The problem of fair division.  Econometrica 16:1 (Jan 1948) 101‑104.  This is a report of a paper given on 17 Sep.  Gives Banach & Knaster's method.

H. Steinhaus.  Sur la division pragmatique.  (With English summary) Econometrica 17 (Supplement) (1949) 315‑319.  Gives Banach & Knaster's method.

Max Black.  Critical Thinking.  Prentice‑Hall, Englewood Cliffs, (1946, ??NYS), 2nd ed., 1952.  Prob. 12, pp. 12 & 432.  Raises the question but only suggests combining two persons.

 

          5.U.   PIGEONHOLE RECREATIONS

 

van Etten.  1624.  Prob. 89, part II, pp. 131‑132 (not in English editions).  Two men have same number of hairs.  Also:  birds & feathers,  fish & scales,  trees & leaves, flowers or fruit,  pages & words -- if there are more pages than words on any page.

E. Fourrey.  Op. cit. in 4.A.1, 1899, section 213: Le nombre de cheveux, p. 165.  Two Frenchmen have the same number of hairs.  "Cette question fut posée et expliquée par Nicole, un des auteurs de la Logique de Port‑Royal, à la duchesse de Longueville."  [This would be c1660.]

The same story is given in a review by T. A. A. Broadbent in MG 25 (No. 264) (May 1941) 128.  He refers to MG 11 (Dec 1922) 193, ??NYS.  This might be the item reproduced as MG 32 (No. 300) (Jul 1948) 159.

The question whether two trees in a large forest have the same number of leaves is said to have been posed to Emmanuel Kant (1724-1804) when he was a boy.  [W. Lietzmann; Riesen und Zwerge im Zahlbereich; 4th ed., Teubner, Leipzig, 1951, pp. 23-24.]  Lietzmann says that an oak has about two million leaves and a pine has about ten million needles.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles, no. 9, pp. 2-3 & 53.  Two people in the world have the same number of hairs on their head.

Manuel des Sorciers.  1825.  Pp. 84-85.  ??NX  Two men have the same number of hairs, etc.

Gustave Peter Lejeune Dirichlet.  Recherches sur les formes quadratiques à coefficients et à indéterminées complexes.  (J. reine u. angew. Math. (24 (1842) 291‑371)  = Math. Werke, (1889‑1897), reprinted by Chelsea, 1969, vol. I, pp. 533‑618.  On pp. 579‑580, he uses the principle to find good rational approximations.  He doesn't give it a name.  In later works he called it the "Schubfach Prinzip".

Illustrated Boy's Own Treasury.  1860.  Arithmetical and Geometrical Problems, No. 34, pp. 430 & 434.  Hairs on head.

Pearson.  1907.  Part II, no. 51, pp. 123 & 201.  "If the population of Bristol exceeds by two hundred and thirty‑seven the number of hairs on the head of any one of its inhabitants, how many of them at least, if none are bald, must have the same number of hairs on their heads?"  Solution says 474!

Dudeney.  The Paradox Party.  Strand Mag. 38 (No. 228) (Dec 1909) 670‑676 (= AM, pp. 137‑141).  Two people have same number of hairs.

Ahrens.  A&N, 1918, p. 94.  Two Berliners have same number of hairs.

Abraham.  1933.  Prob. 43 -- The library, pp. 16 & 25 (12 & 113).  All books have different numbers of words and there are more books than words in the largest book.  (My copy of the 1933 ed. is a presentation copy inscribed 'For the Athenaeum Library No 43  p 16  R M Abraham  Sept 19th 1933'.)

Perelman.  FMP.  c1935?  Socks and gloves.  Pp. 277 & 283‑284.  = FFF, 1957: prob. 25, pp. 41 & 43;  1977, prob. 27, pp. 53‑54 & 56.  = MCBF, prob. 27, pp. 51 & 54.  Picking socks and gloves to get pairs from 10 pairs of brown and 10 pairs of black socks and gloves. 

P. Erdös & G. Szekeres.  Op. cit. in 5.M.  1935.  Any permutation of the first  n2 ‑ 1  integers contains an increasing or a decreasing subsequence of length  > n.

P. Erdös, proposer;  M. Wachsberger & E. Weiszfeld, M. Charosh, solvers.  Problem 3739.  AMM 42 (1935) 396  &  44 (1937) 120.  n+1  integers from first  2n  have one dividing another.

H. Phillips.  Question Time.  Dent, London, 1937.  Prob. 13: Marbles, pp. 7 & 179.  12 black,  8 red  &  6 white  marbles -- choose enough to get three of the same colour.

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  Pp. 148‑149, prob. 6.  Blind maid bringing stockings from a drawer of white and black stockings.

I am surprised that the context of picking items does not occur before Perelman, Phillips and Home Book.

Sullivan.  Unusual.  1943.  Prob. 18: In a dark room.  Picking shoes and socks to get pairs.

H. Phillips.  News Chronicle "Quiz" No. 3: Natural History.  News Chronicle, London, 1946.  Pp. 22 & 43.  12  blue,  9  red and  6  green marbles in a bag.  Choose enough to have three of one colour and two of another colour.

H. Phillips.  News Chronicle "Quiz" No. 4: Current Affairs.  News Chronicle, London, 1946.  Pp. 17 & 40.  6  yellow,  5  blue and  2  red marbles in a bag.  Choose enough to have three of the same colour.

L. Moser, proposer;  D. J. Newman, solver.  Problem 4300 -- The identity as a product of successive elements.  AMM 55 (1948) 369  &  57 (1950) 47.  n  elements from a group of order  n  have a a subinterval with product  = 1.

Doubleday - 2.  1971.

In the dark, pp. 145-146.  How many socks do you have to pick from a drawer of white and black socks to get two pairs (possibly different)?

Lucky dip, pp. 147-148.  How many socks do you have to pick from a drawer of with many white and black socks to get nine pairs (possibly different)?  Gives the general answer  2n+1  for  n  pairs.  [Many means that the drawer contains more than  n  pairs.]

Doubleday - 3.  1972.  In the dark, pp. 35-36.  Four sweaters and  5, 12, 4, 9  socks of the same colours as the sweaters.  Lights go out.  He can only find two of the sweaters.  How many socks must he bring down into the light to be sure of having a pair matching one of the sweaters?

 

          5.V.   THINK‑A‑DOT, ETC.

 

          I managed to acquire one of these without instructions or packaging some years ago.  Michael Keller provided an example complete with instructions and packaging.  I have recently seen Dockhorn's article on variations of the idea.  This is related to Binary Recreations, 7.M.

          The device was produced by E.S.R., Inc.  The box or instructions give an address of 34 Label St., Montclair, New Jersey, 07042, USA, but the company has long been closed.  In Feb 2000, Jim McArdle wrote that he believed that this became the well known Edmund Scientific Co. (101 East Gloucester Pike, Barrington, New Jersey, 08007, USA; tel: 609‑547 3488; email: scientifics@edsci.com; web: http://www.edmundscientific.com).  But he later wrote that investigation of the manuals of DifiComp, one of their other products, reveals that there appears to be no connection.  E.S.R. = Education Science Research.  The inventors of DigiComp, as listed in the patent for it, are: Irving J. Lieberman, William H, Duerig and Charles D. Hogan, all of Montclair, and they were the founders of the company.  The DigiComp manuals say Think‑A‑Dot was later invented by John Weisbacker.  There is a website devoted to DigiComp which contains this material and/or pointers to related sites and has a DigiComp emulator: http://members.aol.com/digicomp1/DigiComp.html .  www.yahoo.com  has a Yahoo club called Friends of DigiComp.  There is another website with the DigiComp manual: http://galena.tj.edu.inter.net/digicomp/ .

 

E.S.R.  Instructions, 8pp, nd -- but box says ©1965.  No patent number anywhere but leaflet says the name Think-A-Dot is trademarked.

E.S.R., Inc. Corporation.  US trademark registration no. 822,770.  Filed: 8 Dec 1965;  registered: 24 Jan 1967.  First used 23 Aug 1965.  Expired.  The US Patent and Trademark Office website entry says the owner is the company and gives no information about the inventor(s).  The name has been registered for a computer game on 23 Jul 2002.

Benjamin L. Schwartz.  Mathematical theory of Think‑A‑Dot.  MM 40:4 (Sep 1967) 187‑193.  Shows there are two classes of patterns and that one can transform any pattern into any other pattern in the same class in at most 15 drops.

Ray Hemmings.  Apparatus Review:  Think‑a‑Dot.  MTg 40 (1967) 45.

Sidney Kravitz.  Additional mathematical theory of Think‑A‑Dot.  JRM 1:4 (Oct 1968) 247‑250.  Considers problems of making ball emerge from one side and of viewing only the back of the game.

Owen Storer.  A think about Think‑a‑dot.  MTg 45 (Winter 1968) 50‑55.  Gives an exercise to show that any possible transformation can be achieved in at most 9 drops.

T. H. O'Beirne.  Letter:  Think‑a‑dot.  MTg 48 (Autumn 1969) 13.  Proves Steiner's (Storer?? - check) assertion about 9 drops and gives an optimal algorithm.

John A. Beidler.  Think-A-Dot revisited.  MM 46:3 (May 1973) 128-136.  Answers a question of Schwarz by use of automata theory.  Characterizes all minimal sequences.  Suggests some generalized versions of the puzzle.

Hans Dockhorn.  Bob's binary boxes.  CFF 32 (Aug 1993) 4-6.  Bob Kootstra makes boxes with the same sort of T-shaped switch present in Think-A-Dot, but with just one entrance.  One switch with two exits is the simplest case.  Kootstra makes a box with three switches and four exits along the bottom, and the successive balls come out of the exits in cyclic sequence.  Using a reset connection between switches, he also makes a two switch, three exit, box.

Boob Kootstra.  Box seven.  CFF 32 (Aug 1993) 7.  Says he has managed to design and make boxes with  5, 6, 7, 8  exits, again with successive balls coming out the exits in cyclic order, but he cannot see any general method nor a way to obtain solutions with a minimal number of movable parts (switches and reset levers).  Further his design for 7 exits is awkward and the design of an optimal box for seven is posed as a contest problem. 

 

          5.W.  MAKING THREE PIECES OF TOAST

 

          This involves an old‑fashioned toaster which does one side of two pieces at a time.  An alternative version is frying steaks or hamburgers on a grill which holds two objects, assuming each side has to be cooked the same length of time.  The problem is probably older than these examples.

 

Sullivan.  Unusual.  1943.  Prob. 7: For the busy housewife.

J. E. Littlewood.  A Mathematician's Miscellany.  Op. cit. in 5.C.  1953.  P. 4 (26).  Mentions problem and solution.

Simon Dresner.  Science World Book of Brain Teasers.  1962.  Op. cit. in 5.B.1.  Prob. 40: Minute toast, pp. 18 & 93.

D. St. P. Barnard.  50 Daily Telegraph Brain‑Twisters.  1985.  Op. cit. in 4.A.4.  Prob. 5: Well done, pp. 16, 80, 103‑104.  Grilling three steaks on a grill which only holds two.  He complicates the problem in two ways:  a)  each side takes a minute to season before cooking;  b)  the steaks want to be cooked  4, 3, 2  minutes per side.

Edward Sitarski.  When do we eat?  CM 27:2 (Mar 2001) 133-135.  Hamburgers which require time  T  per side.  After showing that three hamburgers take  3T,  he asks how long it will take to cook  H  hamburgers.  Easily shows that it can be done in  HT,  except for  H = 1,  which takes  2T.  Then remarks that this is an easy version of a scheduling problem -- in reality, the hamburgers would have different numbers of sides, there would be several grills and each hamburger would have different parts requiring different grills, but in a particular order!

 

          5.W.1.         BOILING EGGS

 

          New section.  These are essentially parodies of the Cistern Problem, 7.H.

 

McKay.  Party Night.  1940.  No. 28, p. 182.  "An egg takes    minutes to boil.  How long should 12 eggs take?"

Jonathan Always.  Puzzles to Puzzle You.  Op. cit. in 5.K.2.  1965.  No. 88: A boiling problem, pp. 29 & 82.  "If it takes    minutes to boil 2 eggs, how long will it take to boil 4 eggs?"

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 161, the editor mentions "If a girl on a hilltop can see two miles, how far would two girls be able to see?"

 

          5.X.   COUNTING FIGURES IN A PATTERN

 

          New section -- there must be older examples.  There are two forms of such problems depending on whether one must use the lattice lines or just the lattice points.

          For counting several shapes, see:  Young World (c1960); Gooding (1994) in 5.X.1.

 

          5.X.1.          COUNTING TRIANGLES

 

          Counting triangles in a pattern is always fraught with difficulties, so I have written a program to do this, but I haven't checked all the examples here.

 

Pearson.  1907.  Part II.

No. 74: A triangle of triangles, p. 74.  Triangular array with four on a side, but with all the altitudes also drawn.  Gets  653  triangles of various shapes.

No. 75: Pharaoh's seal, pp. 75 & 174.  Isosceles right triangles in a square pattern with some diagonals.

Anon.  Prob. 76.  Hobbies 31  (No. 791) (10 Dec 1910) 256  &  (No. 794) (31 Dec 1910) 318.  Make as many triangles as possible with six matches.  From the solution, it seems that the tetrahedron was expected with four triangles, but many submitted the figure of a triangle with its altitudes drawn, but only one solver noted that this figure contains 16 triangles!  However, if the altitudes are displaced to give an interior triangle, I find 17 triangles!!

Loyd.  Cyclopedia.  1914.  King Solomon's seal, pp. 284 & 378.  = MPSL2, No. 142, pp. 100 & 165  c= SLAHP: Various triangles, pp. 25 & 91.  How many triangles in the triangular pattern with 4 on a side?  Loyd Sr. has this embedded in a larger triangle.

Collins.  Book of Puzzles.  1927.  The swarm of triangles, pp. 97-98.  Same as Pearson No. 74.  He says there are  653  triangles and that starting with 5 on a side gives  1196  and  10,000  on a side gives  6,992,965,420,382.  When I gave August's problem in the Weekend Telegraph, F. R. Gill wrote that this puzzle with 5 on a side was given out as a competition problem by a furniture shop in north Lancashire in the late 1930s, with a three piece suite as a prize for the first correct solution.

Evelyn August.  The Black-Out Book.  Harrap, London, 1939.  The eternal triangle, pp. 64 & 213.  Take a triangle,  ABC,  with midpoints  a, b, c,  opposite  A, B, C.  Take a point  d  between  a  and  B.  Draw  Aa, ab, bc, ca, bd, cd.  How many triangles?  Answer is given as  24,  but I (and my program) find  27  and others have confirmed this.

Anon.  Test your eyes.  Mathematical Pie 7 (Oct 1952) 51.  Reproduced in:  Bernard Atkin, ed.; Slices of Mathematical Pie; Math. Assoc., Leicester, 1991, pp. 15 & 71 (not paginated - I count the TP as p. 1).  Triangular pattern with  2  triangles on a side, with the three altitudes drawn.  Answer is 47 'obtained by systematic counting'.  This is correct.  Cf Hancox, 1978.

W. Leslie Prout.  Think Again.  Frederick Warne & Co., London, 1958.  How many triangles, pp. 43 & 130.  Take a pentagon and draw the pentagram inside it.  In the interior pentagon, draw another pentagram.  How many triangles are there?  Answer is  85.

Young World.  c1960.  P. 57: One for Pythagoras.  Consider a L-tromino.  Draw all the midlines to form  12  unit squares.  Or take a  4 x 4  square array and remove a  2 x 2  array from a corner.  Now draw the two main diagonals of the  4 x 4  square - except half of one diagonal would be outside our figure.  How many triangles and how many squares are present?  Gives correct answers of  26  &  17.

J. Halsall.  An interesting series.  MG 46 (No. 355) (Feb 1962) 55‑56.  Larsen (below) says he seems to be the first to count the triangles in the triangular pattern with  n  on a side, but he does not give any proof.

Although there are few references before this point, the puzzle idea was pretty well known and occurs regularly.  E.g. in the children's puzzle books of Norman Pulsford which start c1965, he gives various irregular patterns and asks for the number of triangles or squares.

J. E. Brider.  A mathematical adventure.  MTg (1966) 17‑21.  Correct derivation for the number of triangles in a triangle.  This seems to be the first paper after Halsall but is not in Larsen.

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966.  Prob. 2/12, pp. 23 & 75.  Consider an isosceles right triangle with legs along the axes from  (0,0)  to  (4,0)  and  (0,4).  Draw the horizontals and verticals through the integer lattice points, except that the lines through  (1,1)  only go from the legs to this point and stop.  Draw the diagonals through even-integral lattice points, e.g. from  (2,0)  to  (0,2).  How many triangles.  Says he found  27,  but his secretary then found  29.  I find  29.

Ripley's Puzzles and Games.  1966.  Pp. 72-73 have several problems of counting triangles.

Item 3.  Consider a Star of David with the diameters of its inner hexagon drawn.  How many triangles are in it?  Answer: 20, which I agree with.

Item 4.  Consider a  3 x 3  array of squares with their diagonals drawn.  How many triangles are there?  Answer: 150, however, there are only 124.

Item 5.  Consider five squares, with their midlines and diagonals drawn, formed into a Greek cross.  How many triangles are there?  Answer: 104, but there are 120.

Doubleday - 2.  1971.  Count down, pp. 127-128.  How many triangles in the pentagram (i.e. a pentagon with all its diagonals)?  He says  35.

Gyles Brandreth.  Brandreth's Bedroom Book.  Eyre Methuen, London, 1973.  Triangular, pp. 27 & 63.  Count triangles in an irregular pattern.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  An unusual star, pp. 49-50.  Consider a pentagram and draw lines from each star point through the centre to the opposite crossing point.  How many triangles?  They say 110.

[Henry] Joseph and Lenore Scott.  Master Mind Pencil Puzzles.  1973.  Op. cit. in 5.R.4. 

Diamonds are forever, pp. 35-36.  Hexagon with Star of David inside and another Star of David in the centre of that one.  How many triangles?  Answer is 76.

Count the triangles, pp. 55-56.  Ordinary Greek cross of five squares, with all the diagonals and midlines of the five squares drawn.  How many triangles>  Answer is 104.

C. P. Chalmers.  Note 3353:  More triangles.  MG 58 (No. 403) (Mar 1974) 52‑54.  How many triangles are determined by  N  points lying on  M  lines?  (Not in Larsen.)

Nicola Davies.  The 2nd Target Book of Fun and Games.  Target (Universal-Tandem), London, 1974.  Squares and triangles, pp. 18 & 119.  Consider a chessboard of  4 x 4  cells.  Draw all the diagonals, except the two main ones.  How many squares and how many triangles?

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 136: The triangles, pp. 85 & 133.  How many triangles in a Star of David made of 12 equilateral triangles?

Michael Holt.  Figure It Out -- Book Two.  Granada, London, 1978.  Prob. 67, unpaginated.  How many triangles in a Star of David made of 12 equilateral triangles?

Putnam.  Puzzle Fun.  1978.  No. 91: Counting triangles, pp. 12 & 37.  Same as Doubleday - 2.

D. J. Hancox, D. J.  Number Puzzles  For all The Family.  Stanley Thornes, London, 1978. 

Puzzle 8, pp. 2 & 47.  Draw a line with five points on it, say  A, B, C, D, E,  making four segments.  Connect all these points to a point  F  on one side of the line and to a point  G  on the other side of the line, with  FCG  collinear.  How many triangles are there?  Answer is 24, which is correct.

Puzzle 53, pp. 24 & 54.  Same as Anon.; Test Your Eyes, 1952.  Answer is 36, but there are 47.

The Diagram Group.  The Family Book of Puzzles.  The Leisure Circle Ltd., Wembley, Middlesex, 1984. 

Problem 40, with Solution at the back of the book.  Same as Doubleday - 2.

Problem 116, with Solution at the back of the book.  Count the triangles in a 'butterfly' pattern.

Sue Macy.  Mad Math.  The Best of DynaMath Puzzles.  Scholastic, 1987.  (Taken from Scholastic's DynaMath magazine.)  Shape Up, pp. 5 & 56. 

                    Take a triangle, trisect one edge and join the points of trisection to the opposite vertex.  How many triangles?  [More generally, if one has  n  points on a line and joins them all to a vertex, there are  1 + 2 + ... + n-1  = n(n-1)/2  triangles.]

                    Take a triangle, join up the midpoints of the edges, giving four smaller triangles, and draw one altitude of the original triangle.  How many triangles?

1980 Celebration of Chinese New Year Contest Problem No. 5;  solution by Leroy F. Meyers.  CM 17 (1991) 2  &  18 (1992) 272-273.  n x n  array of squares with all diagonals drawn.  Find the number of isosceles right triangles.  [Has this also been done in half the diagram?  That is, how many isosceles right triangles are in the isosceles right triangle with legs going from  (0,0)  to  (n,0)  and  (0,n)  with all verticals, horizontals and diagonals through integral points drawn?]

Mogens Esrom Larsen.  The eternal triangle -- a history of a counting problem.  Preprint, 1988.  Surveys the history from Halsall on.  The problem was proposed at least five times from 1962 and solved at least ten times.  I have sent him the earlier references.

Marjorie Newman.  The Christmas Puzzle Book.  Hippo (Scholastic Publications), London, 1990.  Star time, pp. 26 & 117.  Consider a Star of David formed from  12  triangles, but each of the six inner triangles is subdivided into  4  triangles.  How many triangles in this pattern?  Answer is 'at least  50'.  I find  58.

Erick Gooding.  Polygon counting.  Mathematical Pie No. 131 (Spring 1994) 1038  &  Notes, pp. 1-2.  Consider the pentagram, i.e. the pentagon with its diagonals drawn.  How many triangles, quadrilaterals and pentagons are there?  Gets  35, 25, 92,  with some uncertainty whether the last number is correct.

When F. R. Gill (See Pearson and Collins above) mentioned the problem of counting the triangles in the figure with all the altitudes drawn, I decided to try to count them myself for the figure with  N  intervals on each side.  The theoretical counting soon gets really messy and I adapted my program for counting triangles in a figure (developed to verify the number found for August's problem).  However, the number of points involved soon got larger than my simple Basic could handle and I rewrote the program for this special case, getting the answers of 653 and 1196 and continuing to N = 22.  I expected the answers to be like those for the simpler triangle counting problem so that there would be separate polynomials for the odd and even cases, or perhaps for different cases (mod 3 or 4 or 6 or 12 or ??).  However, no such pattern appeared for moduli 2, 3, 4 and I did not get enough data to check modulus 6 or higher.  I communicated this to Torsten Sillke and Mogens Esrom Larsen.  Sillke has replied with a detailed answer showing that the relevant modulus is 60!  I haven't checked through his work yet to see if this is an empirical result or he has done the theoretical counting.

Heather Dickson, Heather, ed.  Mind-Bending Challenging Optical Puzzles.  Lagoon Books, London, 1999, pp. 40 & 91.  Gives the version  m = n = 4  of the following.  I have seen other versions of this elsewhere, but I found the general solution on 4 Jul 2001 and am submitting it as a problem to AMM. 

                    Consider a triangle  ABC.  Subdivide the side  AB  into  m  parts by inserting  m‑1  additional points.  Connect these points to  C.  Subdivide the side  AC  into  n  parts by  inserting  n-1  additional points and connect them to  B.  How many triangles are in this pattern?  The number is  [m2n + mn2]/2.  When  m = n,  we get  n3,  but I cannot see any simple geometric interpretation for this.

 

          5.X.2.          COUNTING RECTANGLES OR SQUARES

 

          I have just seen M. Adams.  There are probably earlier examples of these types of problems.

 

Anon.  Prob. 63.  Hobbies  30 (No. 778) (10 Sep 1910) 488  &  31 (No. 781) (1 Oct 1910) 2.  How many rectangles on a  4 x 4  chessboard?  Solution says  100,  which is correct, but then says they are of  17  different types -- I can only get  16  types.

Blyth.  Match-Stick Magic.  1921.  Counting the squares, p. 47.  Count the squares on a  4 x 4  chessboard made of matches with an extra unit square around the central point.  The extra unit square gives 5 additional squares beyond the usual  1 + 4 + 9 + 16.

King.  Best 100.  1927.  No. 9, pp. 10 & 40.  = Foulsham's, no. 5, pp. 6 & 10.  4 x 4  board with some diagonals yielding one extra square.

Loyd Jr.  SLAHP.  1928.  How many rectangles?, pp. 80 & 117.  Asks for the number of squares and rectangles on a  4 x 4  board (i.e. a  5 x 5  lattice of points).  Says answers are  1 + 4 + 9 + 16  and  (1 + 2 + 3 + 4)2  and that these generalise to any size of board.

M. Adams.  Puzzles That Everyone Can Do.  1931.                                                      o o

Prob. 37, pp. 22 & 134: 20 counter problem.  Given the pattern of                     o o

20 counters at the tight, 'how many perfect squares are                               o o o o o o

contained in the figure.'  This means having their vertices                o o o o o o

at counters.  There are surprisingly more than I expected.                     o o

Taking the basic spacing as one, one can have squares of                      o o

edge  1,  Ö2,  Ö5,  Ö8,  Ö13,  giving 21 squares in all.

           He then asks how many counters need to be removed in order to destroy all the squares?  He gives a solution deleting six counters.

Prob. 217, pp. 83 & 162: Match squares.  He gives 10 matches making a row of three equal squares and asks you to add 14 matches to form 14 squares.  The answer is to make a  3 x 3  array of squares and count all of the squares in it.

J. C. Cannell.  Modern Conjuring for Amateurs.  C. Arthur Pearson, London, nd [1930s?].  Counting the squares, pp. 84-85.  As in Blyth.

Indoor Tricks and Games.  Success Publishing, London, nd [1930s??].  Square puzzle, p. 62.  Start with a square and draw its diagonals and midlines.  Join the midpoints of the sides to form a second level square inscribed in the first level original square.  Repeat this until the 9th level.  How many squares are there?  Given answer is  16,  but in my copy someone has crossed this out and written  45,  which seems correct to me.

Meyer.  Big Fun Book.  1940.  No. 9, pp. 162 & 752.  Draw four equidistant horizontal lines and then four equidistant verticals.  How many squares are formed?  This gives a  3 x 3  array of squares, but he counts all sizes of squares, getting  9 + 4 + 1 = 14.  (Also in 7.AU.)

Foulsham's New Party Book.  Foulsham, London, nd [1950s?].  P. 103: How many squares?  4 x 4  board with some extra diagonals giving one extra square.

Although there are few references before this point, the puzzle idea was pretty well known and occurs regularly in the children's puzzle books of Norman Pulsford which start c1965.  He gives various irregular patterns and asks for the number of triangles or squares.

Jonathan Always.  Puzzles to Puzzle You.  Op. cit. in 5.K.2.  1965.  No. 140: A surprising answer, pp. 43 & 90.  4 x 4  chessboard with four corner cells deleted.  How many rectangles are there?

Anon.  Puzzle page: Strictly for squares.  MTg 30 (1965) 48  &  31 (1965) 39  &  32 (1965) 39.  How many squares on a chessboard?  First solution gets  S(8) = 1 + 4 + 9 + ... + 64 = 204.  Second solution observes that there are skew squares if one thinks of the board as a lattice of points and this gives  S(1) + S(2) + ... + S(8) = 540  squares.

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966. 

Prob. 2/11, pp. 22 & 74.  4 x 4  array of squares bordered on two sides by bricks  1 x 2, 1 x 3, 2 x 1, 2 x 1.  Count the squares and the rectangles.  Gets 35 and 90.

Prob. 2/14, pp. 23 & 75.  Pattern of squares making the shape of a person -- how many squares in it?

Ripley's Puzzles and Games.  1966.  Pp. 72-73 have several problems of counting squares.

Item 4.  Consider a  3 x 3  array of squares with their diagonals drawn.  The solution says this has 30 squares.  I get 31, but perhaps they weren't counting the whole figure.  I have computed the total number of squares for an  n x n  array and get  (2n3 + n2)/2  squares for  n  even and  (2n3 + n2 -1)/2  squares for  n  odd.

Unnumbered item at lower right of p. 73.  4 x 4  array of squares with their diagonals drawn, except that the four corner squares have only one diagonal -- the one not pointing to an opposite corner -- and this reduces the number of squares by eight, agreeing with the given answer of  64.

Doubleday - 2.  1971.  Bed of nails, pp. 129-130.  20 points in the form of a Greek cross with double-length arms (so that the axes are five times the width of the central square, or the shape is a 9-omino).  How many squares can be located on these points?  He finds 21.

W. Antony Broomhead.  Note 3315:  Two unsolved problems.  MG 55 (No. 394) (Dec 1971) 438.  Find the number of squares on an  n x n  array of dots, i.e. the second problem in MTg (1965) above, and another problem.

W. Antony Broomhead.  Note 3328:  Squares in a square lattice.  MG 56 (No. 396) (May 1972) 129.  Finds there are  n2(n2 ‑ 1)/12  squares and gives a proof due to John Dawes.  Editorial note says the problem appears in:  M. T. L. Bizley; Probability: An Intermediate Textbook; CUP, 1957, ??NYS.  A. J. Finch asks the question for cubes.

Gyles Brandreth.  Brandreth's Bedroom Book.  Eyre Methuen, London, 1973.  Squares, pp. 26 & 63.  Same as Briggs.

Nicola Davies.  The 2nd Target Book of Fun and Games.  1974.  See entry in 5.X.1.

Putnam.  Puzzle Fun.  1978. 

No. 107: Square the coins, pp. 17 & 40.  20 points in the form of a Greek cross made from five  2 x 2  arrays of points.  How many squares -- including skew ones?  Gets 21.

No. 108: Unsquaring the coins, pp. 17 & 40.  How many points must be removed from the previous pattern in order to leave no squares?  Gets 6.

 

          5.X.3.          COUNTING HEXAGONS

 

M. Adams.  Puzzle Book.  1939.  Prob. C.157: Making hexagons, pp. 163 & 190.  The hexagon on the triangular lattice which is two units along each edge contains 8 hexagons.  [It is known that the hexagon of side  n  contains  n3  hexagons.  I recently discovered this but have found that it is known, though I don't know who discovered it first.]

The Diagram Group.  The Family Book of Puzzles.  The Leisure Circle Ltd., Wembley, Middlesex, 1984.  Problem 32, with Solution at the back of the book.  Count the hexagons in the hexagon of side three on the triangular lattice.  They get 27.

 

          5.X.4.          COUNTING CIRCLES

 

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966.  Prob. 2/13, pp. 23 & 75.  Pattern with hexagonal symmetry and lots of overlapping circles, some incomplete.

 

          5.Y.   NUMBER OF ROUTES IN A LATTICE

 

          The common earlier form was to have the route spell a word or phrase from the centre to the boundaries of a diamond.  I will call this a word diamond.  Sometimes the phrase is a palindrome and one reads to the centre and then back to the edge.  See Dudeney, CP, for analysis of the most common cases.  I have seen such problems on the surface of a  3 x 3 x 3  cube.  The problems of counting Euler or Hamiltonian paths are related questions, but dealt with under 5.E and 5.F.

          New section -- in view of the complexity of the examples below, there must be older, easier, versions, but I have only found the few listed below.  The first entry gives some ancient lattices, but there is no indication that the number of paths was sought in ancient times.

 

Roger Millington.  The Strange World of the Crossword.  M. & J. Hobbs, Walton‑on‑Thames, UK, 1974.  (This seems to have been retitled: Crossword Puzzles: Their History and Cult for a US ed from Nelson, NY.)

                    On pp. 38-39 & 162, he gives the cabalistic triangle shown below and says it is thought to have been constructed from the opening letters of the Hebrew words  Ab (Father),  Ben (Son),  Ruach Acadash (Holy Spirit).  He then asks how many ways one can read  ABRACADABRA  in it, though there is no indication that the ancients did this.  His answer is  1024  which is correct.

 

                                        A B R A C A D A B R A

                                         A B R A C A D A B R

                                          A B R A C A D A B

                                           A B R A C A D A

                                            A B R A C A D

                                             A B R A C A

                                             A B R A C

                                               A B R A

                                                A B R

                                                 A B

                                                  A

 

                    On pp. 39-40 he describes and illustrates an inscription on the Stele of Moschion from Egypt, c300.  This is a  39 x 39  square with a Greek text from the middle to the corner, e.g. like the example in the following entry.  The text reads:  ΟΥIΡIΔIΜΟΥΧIΩΝΥΓIΑΥΘΕIΥΤΟΝΠΟΔΑIΑΤΠΕIΑIΥ  which means:  Moschion to Osiris, for the treatment which cured his foot.  Millington does not ask for the number of ways to read the inscription, which is  4 BC(38,19) = 14 13810 55200.

 

Curiosities for the Ingenious selected from The most authentic Treasures of                           E D C D E

          Nature, Science and Art, Biography, History, and General Literature.                          D C B C D

          (1821); 2nd ed., Thomas Boys, London, 1822.  Remarkable epitaph,                          C B A B C

          p. 97.  Word diamond extended to a square, based on 'Silo Princeps Fecit',                  D C B C D

          with the  ts  at the corners.  An example based on 'ABCDEF' is shown                        E D C D E

          at the right.  Says this occurs on the tomb of a prince named Silo at the

          entrance of the church of San Salvador in Oviedo, Spain.  Says the epitaph can be read in 270 ways.  I find there are  4 BC(16, 8)  = 51490  ways.

In the churchyard of St. Mary's, Monmouth, is the gravestone of John Rennie, died 31 May 1832, aged 33 years.  This has the inscription shown below.  Further down the stone it gives his son's name as James Rennie.  Apparently an  N  has been dropped to get a message with an odd number of letters.  I have good photos.  Nothing asks for the number of ways of reading the inscription.  I get  4 BC(16,9)  =  45760  ways.

 

                                                            eineRnhoJsJohnRenie

                                                            ineRnhoJsesJohnReni

                                                            neRnhoJseiesJohnRen

                                                             eRnhoJseiliesJohnRe

                                                             RnhoJseileliesJohnR

                                                              nhoJseilereliesJohn

                                                               hoJseilerereliesJoh

                                                               oJseilereHereliesJo

                                                               hoJseilerereliesJoh

                                                              nhoJseilereliesJohn

                                                             RnhoJseileliesJohnR

                                                             eRnhoJseiliesJohnRe

                                                            neRnhoJseiesJohnRen

                                                            ineRnhoJsesJohnReni

                                                            eineRnhoJsJohnRenie

 

Nuts to Crack I (1832), no. 200.  The example from Curiosities for the Ingenious with 'SiloPrincepsFecit', but no indication of what is wanted -- perhaps it is just an amusing picture.

W. Staniforth.  Letter.  Knowledge 16 (Apr 1893) 74-75.  Considers            1   2   3   4   5   6

          "figure squares" as at the right.  "In how many different ways may       2   3   4   5   6   7

          the figures in the square be read from  1  to  11  consecutively?"          3   4   5   6   7   8

          He computes the answers for the  n x n  case for the first few               4   5   6   7   8   9

          cases and finds a recurrence.  "Has such a series of numbers any           5   6   7   8   9 10

          mathematical designation?"  The editor notes that he doesn't                 6   7   8   9 10 11

          know.

J. J. Alexander.  Letter.  Knowledge 16 (May 1893) 89.  Says Staniforth's numbers are the sums of the squares of the binomial coefficients  BC(n, k),  the formula for which is  BC(2n, n).  Editor say he has received more than one note pointing this out and cites a paper on such figure squares by T. B. Sprague in the Transactions of the Royal Society of Edinburgh -- ??NYS, no more details provided.

Loyd.  Problem 12: The temperance puzzle.  Tit‑Bits 31 (2  &  23 Jan 1897) 251  &  307.  Red rum & murder.  = Cyclopedia, 1914, The little brown jug, pp. 122 & 355.  c= MPSL2, no. 61, pp. 44 & 141.  Word diamond based on 'red rum & murder', i.e. the central line is  redrum&murder.  He allows a diagonal move from an  E  back to an inner  R  and this gives  372  paths from centre to edge, making  3722 = 138,384  in total.

Dudeney.  Problem 57: The commercial traveller's puzzle.  Tit‑Bits 33 (30 Oct  &  20 Nov 1897) 82  &  140.  Number of routes down and right on a  10 x 12  board.  Gives a general solution for any board.

Dudeney.  A batch of puzzles.  The Royal Magazine 1:3 (Jan 1899) 269-274  &  1:4 (Feb 1899) 368-372.  A "Reviver" puzzle.  Complicated pattern based on 'reviver'.  544 solutions.

Dudeney.  Puzzling times at Solvamhall Castle.  London Magazine 7 (No. 42) (Jan 1902) 580‑584  &  8 (No. 43) (Feb 1902) 53-56.  The amulet.  'Abracadabra' in a triangle with   A  at top,  two  B's  below,  three  R's  below that,  etc.  Answer:  1024.  = CP, 1907, No. 38, pp. 64-65 & 190.  CF Millington at beginning of this section.

Dudeney.  CP.  1907. 

Prob. 30: The puzzle of the canon's yeoman, pp. 55-56 & 181-182.  Word diamond based on 'was it a rat I saw'.  Answer is  63504  ways.  Solution observes that for a diamond of side  n+1,  with no diagonal moves, the number of routes from the centre to an edge is  4(2n-1)  and the number of ways to spell the phrase is this number squared.  Analyses four types with the following central lines:  A ‑ 'yoboy';  B - 'level';  C - 'noonoon';  D ‑ 'levelevel'. 

           In  A, one wants to spell 'boy', so there are  4(2n-1) solutions. 

           In  B, one wants to spell 'level' and there are  [4(2n-1)]2  solutions. 

           In  C, one wants to spell 'noon' and there are  8(2n-1)  solutions. 

           In  D, one wants to spell 'level' and there are complications as one can start and finish at the edge.  He obtains a general formula for the number of ways.  Cf Loyd, 1914.

Prob. 38: The amulet, pp. 64-65 & 190.  See:  Dudeney, 1902.

Pearson.  1907.  Part II: A magic cocoon, p. 147.  Word diamond based on 'cocoon', so the central line is  noocococoon.  Because one can start at the non‑central  Cs,  and can go in as well as out, I get  948  paths.  He says  756.

Loyd.  Cyclopedia.  1914.  Alice in Wonderland, pp. 164 & 360.  = MPSL1, no. 109, pp. 107 & 161‑162.  Word diamond based on 'was it a cat I saw'.  Cf Dudeney, 1907.

Dudeney.  AM.  1917.

Prob. 256: The diamond puzzle, pp. 74 & 202.  Word diamond based on 'dnomaidiamond'.  This is type  A  of his discussion in CP and he states the general formula.  252  solutions.

Prob. 257: The deified puzzle, pp. 74-75 & 202.  Word diamond based on 'deifiedeified'.  This is type  D  in CP and has  1992  solutions.  He says 'madamadam' gives  400  and 'nunun' gives  64,  while 'noonoon' gives  56.

Prob. 258: The voter's puzzle, pp. 75 & 202.  Word diamond built on 'rise to vote sir'.  Cites CP, no. 30, for the result,  63504,  and the general formula.

Prob. 259: Hannah's puzzle, pp. 75 & 202.  6 x 6  word square based on 'Hannah' with  Hs  on the outside,  As  adjacent to the  Hs  and four  Ns  in the middle.  Diagonal moves allowed.  3468  ways.

Wood.  Oddities.  1927.  Prob. 44: The amulet problem, p. 39.  Like the original  ABRACADABRA  triangle, but with the letters in reverse order.

Collins.  Book of Puzzles.  1927.  The magic cocoon puzzle, pp. 169-170.  As in Pearson.

Loyd Jr.  SLAHP.  1928.  A strolling pedagogue, pp. 38 & 97.  Number of routes to opposite corner of a  5 x 5  array of points.

D. F. Lawden.  On the solution of linear difference equations.  MG 36 (No. 317) (Sep 1952) 193-196.  Develops use of integral transforms and applies it to find that the number of king's paths going down or right or down‑right from  (0, 0)  to  (n, n)  is  Pn(3)  where Pn(x)  is the Legendre polynomial. 

Leo Moser.  King paths on a chessboard.  MG 39 (No. 327) (Feb 1955) 54.  Cites Lawden and gives a simpler proof of his result  Pn(3).

Anon.  Puzzle Page: Check this.  MTg 36 (1964) 61  &  27 (1964) 65.  Find the number of king's routes from corner to corner when he can only move right, down or right‑down.  Gets  48,639  routes on  8 x 8  board.

Ripley's Puzzles and Games.  1966.  P. 32.  Word diamond laid out differently so                         A A A

          one has to read from one side to the opposite side.  Rotating by 45o, one gets                        B B 

          the pattern at the right for edge three.  One wants the number of ways to                            C C C

          read  ABCDEF.  In general, when the first line of  As  has  n  positions,                                D D 

          the total number of ways to reach the first row is  n.  For each successive                            E E E

          row, the total number is alternately twice the number for the previous row less                       F F 

          twice the end term of that row or just twice the the number for the previous

          row.  In our example with  n = 3,  the number of ways to reach the second row is  4 = 2x3 - 2x1.  The number of ways to reach the third row is  8 = 2x4.  The number of ways to reach the fourth row is  12 = 2x8 - 2x2,  then we get  24 = 2 x 12;  36 = 2x24 ‑ 2x6.  It happens that the first  n  end terms are the central binomial coefficients  BC(2k,k),  so this is easy to calculate.  I find the total number of routes, for  n = 2, 3, ..., 7,  is  4,  18,  232,  1300,  6744,  33320,  the last being the desired and given answer for the given problem.

Pál Révész.  Op. cit. in 5.I.1.  1969.  On p. 27, he gives the number of routes for a king moving forward on a chessboard and a man moving forward on a draughtsboard.

Putnam.  Puzzle Fun.  1978.  No. 8: Level - level, pp. 3 & 26.  Form a wheel of 16 points labelled  LEVELEVELEVELEVE.  Place 4  Es  inside, joined to two consecutive  Vs  and the intervening  L.  Then place a  V  in the middle, joined to these four  Es.  How many ways to spell LEVEL?  He gets 80, which seems right.

 

          5.Z.    CHESSBOARD PLACING PROBLEMS

 

          See MUS I 285-318, some parts of the previous chapter and the Appendix in II 351-360.  See also 5.I.1, 6.T.

          There are three kinds of domination problems. 

                    In strong domination, a piece dominates the square it is on.

                    In weak domination, it does not, hence more pieces may be needed to dominate the board. 

                    Non‑attacking domination is strong domination with no piece attacking another.  Graph theorists say the pieces are independent.  This also may require more pieces than strong domination, but it may require more or fewer pieces than weak domination. 

          The words 'guarded' or 'protected' are used for weak domination, but 'unguarded' or 'unprotected' may mean either strong or non‑attacking domination.

          Though these results seem like they must be old, the ideas seem to have originated with the eight queens problem, c1850, (cf 5.I.1) and to have been first really been attacked in the late 19C.  There are many variations on these problems, e.g. see Ball, and I will not attempt to be complete on the later variations.  In recent years, this has become a popular subject in graph theory, where the domination number,  γ(G),  is the size of the smallest strongly dominating set on the graph  G  and the independent domination number,  i(G),  is the size of the smallest non-attacking (= independent) dominating set.

          Mario Velucchi has a web site devoted to the non-dominating queens problem and related sites for similar problems.  See:  http://www.bigfoot.com/~velucchi/papers.html  and  http://www.bigfoot.com/~velucchi/biblio.html.

 

Ball.  MRE, 3rd ed., 1896, pp. 109-110: Other problems with queens.  Says:  "Captain Turton has called my attention to two other problems of a somewhat analogous character, neither of which, as far as I know, has been hitherto published, ...."  These ask for ways to place queens so as to attack as few or as many cells as possible -- see 5.Z.2.

Ball.  MRE, 4th ed., 1905, pp. 119-120: Other problems with queens;  Extension to other chess pieces.  Repeats above quote, but replaces 'hitherto published' by 'published elsewhere', extends the previous text and adds the new section.

Ball.  MRE, 5th ed., 1911.  Maximum pieces problem;  Minimum pieces problem, pp. 119‑122.  [6th ed., 1914 adds that Dudeney has written on these problems in The Weekly Dispatch, but this is dropped in the 11th ed. of 1939.]  Considerably generalizes the problems.  On the  8 x 8  board, the maximum number of non-attacking  kings is 16,  queens is 8,  bishops is 14 [6th ed., 1914, adds there are 256 solutions],  knights is 32 with 2 solutions  and  rooks is 8 with  88  solutions [sic, but changed to  8! in the 6th ed.].  The minimum number of pieces to strongly dominate the board is  9 kings,  5 queens with 91 inequivalent solutions [the 91 is omitted in the 6th ed., since it is stated later],  8 bishops,  12 knights,  8 rooks.  The minimum number of pieces to weakly dominate the board is  5 queens,  10 bishops,  14 knights,  8 rooks.

Dudeney.  AM.  1917.  The guarded chessboard, pp. 95‑96.  Discusses different ways pieces can weakly or non‑attackingly dominate  n x n  boards.

G. P. Jelliss.  Multiple unguard arrangements.  Chessics 13 (Jan/Jun 1982) 8‑9.  One can have  16 kings,  8 queens,  14 bishops,  32 knights  or  8 rooks  non‑attackingly placed on a  8 x 8  board.  He considers mixtures of pieces -- e.g. one can have 10 kings and 4 queens non‑attacking.  He tries to maximize the product of the numbers of each type in a mixture -- e.g. scoring 40 for the example.

 

          5.Z.1. KINGS

 

Ball.  MRE, 4th ed., 1905.  Other problems with queens;  Extension to other chess pieces, pp. 119-120.  Says problems have been proposed for  k  kings on an  n x n,  citing L'Inter. des math. 8 (1901) 140, ??NYS.

Gilbert Obermair.  Denkspiele auf dem Schachbrett.  Hugendubel, Munich, 1984.  Prob. 27, pp. 29 & 58.  9  kings strongly, and  12  kings weakly, dominate an  8 x 8  board.

 

          5.Z.2. QUEENS

 

          Here the graph is denoted  Qn,  but I will denote  γ(Qn)  by  γ(n)  and  i(Qn)  by  i(n).

 

Murray.  Pp. 674 & 691.  CB249 (c1475) shows  16  queens weakly dominating an  8 x 8  board, but the context is unclear to me.

de Jaenisch.  Op. cit. in 5.F.1.  Vol. 3, 1863.  Appendice, pp. 244-271.  Most of this is due to "un de nos anciens amis, Mr de R***".  Finds and describes the 91 ways of placing 5 queens so as to non-attackingly dominate the  8 x 8  board.  Then considers the  n x n  board for  n = 2, ..., 7  with strong and non-attacking domination.  Up through 5, he gives the number of pieces being attacked in each solution which allows one to determine the weak solutions.  For  n < 6,  he gets the answers in the table below, but for  n = 6,  he gets  21  non-attacking solutions instead of  17?.

Ball.  MRE, 3rd ed., 1896.  Other problems with queens, pp. 109-110.  "Captain [W. H.] Turton has called my attention to two other problems of a somewhat analogous character, neither of which, as far as I know, has been hitherto published, or solved otherwise than empirically."  The first is to place 8 queens so as to strongly dominate the fewest squares.  The minimum he can find is 53.  (Cf Gardner, 1999.)  The second is to place  m  queens,  m £ 5,  so as to strongly dominate as many cells as possible.  With 4 queens, the most he can find is 62.

Dudeney.  Problem 54: The hat‑peg puzzle.  Tit‑Bits 33 (9  &  30 Oct 1897) 21  &  82.  Problem involves several examples of strong domination by 5 queens on an  8 x 8  board leading to a non‑attacking domination.  He says there are just 728 such.  This  = 8 x 91.  = Anon. & Dudeney; A chat with the Puzzle King; The Captain 2 (Dec? 1899) 314-320; 2:6 (Mar 1900) 598-599  &  3:1 (Apr 1900) 89.  = AM; 1917; pp. 93-94 & 221.

Ball.  MRE, 4th ed., 1905, loc. cit. in 5.Z.1.  Extends 3rd ed. by asking for the minimum number of queens to strongly dominate a whole  n x n  board.  Says there seem to be 91 ways of having 5 non-attacking queens on the  8 x 8,  citing L'Inter. des Math. 8 (1901) 88, ??NYS. 

Ball.  MRE, 5th ed., 1911, loc. cit. in 5.Z.  On pp. 120-122, he considers queens and states the minimum numbers of queens required to strongly dominate the board and the numbers of inequivalent solutions for  2 x 2,  3 x 3,  ...,  7 x 7,  citing the article cited in the 4th ed. and Jaenisch, 1862, without a volume number.  For  n = 7,  he gives the same unique solution for strongly dominating as for non-attacking dominating.  [In the 6th ed., this is corrected and he says it is a solution.]  He says Jaenisch also posed the question of the minimum number of non-attacking queens to dominate the board and gives the numbers and the number of inequivalent ways for the  4 x 4,  ..,  8 x 8,  except that he follows Jaenisch in stating that there are 21 solutions on the  6 x 6.  [This is changed to 17 in the 6th ed.]

Dudeney.  AM.  1917.  Loc. cit. in 5.Z.  He uses 'protected' for 'weakly', but he seems to copy the values for 'strongly' from Jaenisch or Ball.  His 'not protected' seems to mean 'non-attacking'.  However,  some values are different and I consequently am very uncertain as to the correct values?? 

Pál Révész.  Op. cit. in 5.I.1.  1969.  On pp. 24‑25, he shows 5 queens are sufficient to strongly dominate the board and says this is minimal.

 

Below, min. denotes the minimum number of queens to dominate and no. is the number of inequivalent ways to do so.

 

                              STRONG      WEAK    NON-ATTACKING

          n                     min.   no.     min.   no.     min.    no.

 

          1                      1       1         0       0         1       1

          2                      1       1         2       2         1       1

          3                      1       1         2       5         1       1

          4                      2       3         2       3         3       2

          5                      3     37         3     15         3       2

          6                      3       1         4     ³2         4    17?

          7                      4                  4      5?        4       1

          8                      5  ³150        5    ³41         5    91

 

Rodolfo Marcelo Kurchan, proposer;  Henry Ibstedt & proposer, solver.  Prob. 1738 -- Queens in space.  JRM 21:3 (1989) 220  &  22:3 (1989) 237.  How many queens are needed to strongly dominate an  n x n x n  cubical board?  For  n = 3, 4, ..., 9,  the best known numbers are:  1, 4, 6, 8, 14, 20, 24.  The solution is not clear if these are minimal, but it seems to imply this.

Martin Gardner.  Chess queens and maximum unattacked cells.  Math Horizons (Nov 1999).  Reprinted in Workout, chap. 34.  Considers the problem of Turton described in Ball, 3rd ed, above:  place  8  queens on an  8 x 8  board so as to strongly dominate the fewest squares.  That is, leave the maximum number of unattacked squares.  More generally, place  k  queens on an  n x n  board to leave the maximum number of unattacked squares.  He describes a simple problem by Dudeney (AM, prob. 316) and recent work on the general problem.  He cites Velucchi, cf below, who provides the following table of maximum numbers of unattacked cells and number of solutions for the maximum.  I'm not sure if some of these are still only conjectured.

 

   n         1      2      3      4      5      6      7      8      9    10    11    12    13    14    15    16    17

Max        0      0      0      1      3      5      7    11    18    22    30    36    47    56    72    82    97

Sols        0      0      0    25      1      3    38      7      1      1      2      7      1      4      3      1

 

Mario Velucchi has a web site devoted to the non-dominating queens problem and related sites for similar problems.  See:  http://www.bigfoot.com/~velucchi/papers.html  and  http://www.bigfoot.com/~velucchi/biblio.html.

A. P. Burger & C. M. Mynhardt.  Symmetry and domination in queens graphs.  Bull. Inst. Combinatorics Appl. 29 (May 2000) 11-24.  Extends results to  n = 30, 45, 69, 77.  Summarizes the field, with 14 references, several being earlier surveys.  The table below gives all known values.  It will be seen that the case  n = 4k + 1  seems easiest to deal with.  The values separated by strokes,  /,  indicate cases where the value is one of the two given values, but it is not known which.

 

  n              4        5        6        7        8        9      10      11      12      13      14      15      16      17

γ(n)            2        3        3        4        5        5        5        5        6        7     7/8     8/9     8/9        9

i(n)             3        3        4        4        5        5        5        5        7        7        8        9        9        9

 

  n             18      19      21      25      29      30      31      33      37      41      45      49      53      57

γ(n)            9      10      11      13      15      15      16      17      19      21      23      25      27      29

i(n)                               11      13                                    17                          23         

 

  n             61      69      77

γ(n)           31      35      39

 

          5.Z.3. BISHOPS

 

Dudeney.  AM.  1917.  Prob. 299: Bishops in convocation, pp. 89 & 215.  There are  2n  ways to place  2n‑2  bishops non‑attackingly on an  n x n  board.  At loc. cit. in 5.Z, he says that for  n = 2, ..., 8,  there are  1, 2, 3, 6, 10, 20, 36  inequivalent placings.

Pál Révész.  Op. cit. in 5.I.1.  1969.  On pp. 25‑26, he shows the maximum number of non‑attacking bishops on one colour is 7 and there are 16 ways to place them.

Obermair.  Op. cit. in 5.Z.1.  1984.  Prob. 17, pp. 23 & 50.  8 bishops strongly, and 10 bishops weakly, dominate the  8 x 8  board.

 

          5.Z.4. KNIGHTS

 

Ball.  MRE, 4th ed., 1905.  Loc. cit. in 5.Z.1.  Says questions as to the maximum number of non-attacking knights and minimum number to strongly dominate have been considered, citing L'Inter. des math. 3 (1896) 58,  4 (1897) 15-17  &  254, 5 (1898) 87 [5th ed. adds 230‑231], ??NYS.

Dudeney.  AM.  1917.  Loc. cit. in 5.Z.  Notes that if  n  is odd, one can have  (n2+1)/2  non‑attacking knights in one way, while if  n  is even, one can have  n2/2  in two equivalent ways.

Irving Newman, proposer;  Robert Patenaude, Ralph Greenberg and Irving Newman, solvers.  Problem E1585 -- Nonattacking knights on a chessboard.  AMM 70 (1963) 438  &  71 (1964) 210-211.  Three easy proofs that the maximum number of non-attacking knights is 32.  Editorial note cites Dudeney, AM, and Ball, MRE, 1926, p. 171 -- but the material is on p. 171 only in the 11th ed., 1939.

Gardner.  SA (Oct 1967,  Nov 1967  &  Jan 1968) c= Magic Show, chap. 14.  Gives Dudeney's results for the  8 x 8.  Golomb has noted that Greenberg's solution of E1585 via a knight's tour proves that there are only two solutions.  For the  k x k  board,  k = 3, 4, ..., 10,  the minimal number of knights to strongly dominate is:  4, 4, 5, 8, 10, 12, 14, 16.  He says the table may continue:  21, 24, 28, 32, 37.  Gives numerous examples.

Obermair.  Op. cit. in 5.Z.1.  1984.  Prob. 16, pp. 21 & 47.  14 knights are necessary for weak domination of the 8x8 board.

E. O. Hare & S. T. Hedetniemi.  A linear algorithm for computing the knight's domination number of a  k x n  chessboard.  Technical report 87‑May‑1, Dept. of Computer Science, Clemson University.  1987??  Pp. 1‑2 gives the history from 1896 and Table 2 on p. 13 gives their optimal results for strong domination on  k x n  boards,  £ k £ 9,  £ n £ 12  and also for  k = n = 10.  For the  k x k  board,  k = 3, ..., 10,  they confirm the results in Gardner.

Anderson H. Jackson & Roy P. Pargas.  Solutions to the  N x N  knight's cover problem.  JRM 23:4 (1991) 255-267.  Finds number of knights to strongly dominate by a heuristic method, which finds all solutions up through  N = 10.  Improves the value given by Gardner for  N = 15  to 36 and finds solutions for  N = 16, ..., 20  with  42, 48, 54, 60, 65  knights.

 

          5.Z.5. ROOKS

 

É. Lucas.  Théorie des Nombres.  Gauthier‑Villars, Paris, 1891;  reprinted by Blanchard, Paris, 1958.  Section 128, pp. 220‑223.  Determines the number of inequivalent placings of  n  nonattacking rooks on an  n x n  board in general and gives values for  n £ 12.  For  n = 1, ..., 8,  there are  1, 1, 2, 7, 23, 115, 694, 5282  inequivalent ways.

Dudeney.  AM.  1917.  Loc. cit. at 5.Z.  Notes there are  n!  ways to place  n  non‑attacking rooks and asks how many of these are inequivalent.  Gives values for  n = 1, ..., 5.  AM prob. 296, pp. 88 & 214, is the case  n = 4.

D. F. Holt.  Rooks inviolate.  MG 58 (No. 404) (Jun 1974) 131‑134.  Uses Burnside's lemma to determine the number of inequivalent solutions in general, getting Lucas' result in a more modern form.

 

          5.Z.6. MIXTURES

 

Ball.  MRE, 5th ed., 1911.  Loc. cit. in 5.Z.  P. 122:  "There are endless similar questions in which combinations of pieces are involved."  4 queens and king or queen or bishop or knight or rook or pawn can strongly dominate  8 x 8.

King.  Best 100.  1927.

          No. 77, pp. 30 & 57.  4 queens and a rook strongly dominate  8 x 8.

          No. 78, pp. 30 & 57.  4 queens and a bishop strongly dominate  8 x 8.

 

5.AA.   CARD SHUFFLING

 

          New section.  I have been meaning to add this sometime, but I have just come across an expository article, so I am now starting.  The mathematics of this gets quite formidable.  See 5.AD for a somewhat related topic.

          A faro, weave, dovetail or perfect (riffle) shuffle starts by cutting the deck in half and then interleaving the two halves.  When the deck has an even number of cards, there are two ways this can happen -- the original top card can remain on top (an out shuffle) or it can become the second card of the shuffled deck (an in shuffle).  E.g. if our deck is  123456,  then the out shuffle yields  142536  and the in shuffle yields  415263.  Note that removing the first and last cards converts an out shuffle on  2n  cards to an in shuffle on  2n-2  cards.  When the deck has an odd number of cards, say  2n+1,  we cut above or below the middle card and shuffle so the top of the larger pile is on top, i.e. the larger pile straddles the smaller.  If the cut is below the middle card, we have piles of  n+1  and  n  and the top card remains on top, while cutting above the middle card leaves the bottom card on bottom.  Removing the top or bottom card leaves an in shuffle on  2n  cards.

          Monge's shuffle takes the first card and then alternates the next cards over and under the resulting pile, so  12345678  becomes  86421357.

          At G4G2, 1996, Max Maven gave a talk on some magic tricks based on card shuffling and gave a short outline of the history.  The following is an attempt to summarise his material.  The faro shuffle, done by inserting part of the deck endwise into the other part, but not done perfectly, began to be used in the early 18C and a case of cheating using this is recorded in 1726.  The riffle shuffle, which is the common American shuffle, depends on mass produced cards of good quality and began to be used in the mid 19C.  However, magicians did not become aware of the possibilities of the perfect shuffle until the mid 20C, despite the early work of Stanyans C. O. Williams and Charles T. Jordan in the 1910s.

 

Hooper.  Rational Recreations.  Op. cit. in 4.A.1.  1774.  Vol. 1, pp. 78-85: Of the combinations of the cards.  This describes a shuffle, where one takes the top two cards, then puts the next two cards on top, then the next three cards underneath, then the next two on top, then the next three underneath.  For ten cards  1234567890,  it produces  8934125670,  a permutation of order 7.  Tables of the first few repetitions are given for  10, 24, 27 and 32  cards, having orders  7, 30, 30, 156.

The Secret Out.  1859.  Permutation table, pp. 394-395 (UK: 128-129).  Describes Hooper's shuffle for ten cards.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. XV, 1884: 214-222.  Discusses Monge's shuffle and its period.

John Nevil Maskelyne.  Sharps and Flats.  1894.  ??NYS -- cited by Gardner in the Addendum of Carnival.  "One of the earliest mentions".  Called the "faro dealer's shuffle".

Ahrens.  MUS I.  1910.  Ein Kartenkunststück Monges, pp. 152-145.  Expresses the general form of Monge's shuffle and finds its order for  n = 1, 2, ..., 10.  Mentions the general question of finding the order of a shuffle.

Charles T. Jordan.  Thirty Card Mysteries.  The author, Penngrove, California, 1919 (??NYS), 2nd ed., 1920 (?? I have copy of part of this).  Cited by Gardner in the Addendum to Carnival.  First magician to apply the shuffle, but it was not until late 1950s that magicians began to seriously use and study it.  The part I have (pp. 7-10) just describes the idea, without showing how to perform it.  The text clearly continues to some applications of the idea.  This material was reprinted in The Bat (1948-1949).

Frederick Charles Boon.  Shuffling a pack of cards and the theory of numbers.  MG 15 (1930) 17-20.  Considers the Out shuffle and sees that it relates to the order of  2 (mod 2n+1) and gives some number theoretic observations on this.  Also considers odd decks.

J. V. Uspensky & M. A. Heaslet.  Elementary Number Theory.  McGraw-Hill, NY, 1939.  Chap. VIII: Appendix: On card shuffling, pp. 244-248.  Shows that an In shuffle of a deck of  2n  cards takes the card in position  i  to position  2i (mod 2n+1), so the order of the permutation is the exponent or order of  2 (mod 2n+1), which is  52  when  n = 26.  [Though not discussed, this shows that the order of the Out shuffle is the order of  2 (mod 2n-1), which is only  8  for  n = 26.  And the order of a shuffle of  2n+1  cards is the order of  2 (mod 2n+1).]  Monge's shuffle is more complex, but leads to congruences (mod 4n+1) and has order equal to the smallest exponent  e  such that  2e º ±1 (mod 4n+1),  which is  12  for  n = 26.

T. H. R. Skyrme.  A shuffling problem.  Eureka 7 (Mar 1942) 17-18.  Describes Monge's shuffle with the second card going under or over the first.  Observes that in the under shuffle for an even number of cards, the last card remains fixed, while the over shuffle for an odd number of cards also leaves the last card fixed.  By appropriate choice, one always has the  n-th card becoming the first.  Finds the order of the shuffle essentially as in Uspensky & Heaslet.  Makes some further observations.

N. S. Mendelsohn, proposer and solver.  Problem E792 -- Shuffling cards.  AMM 54 (1947) 545 ??NYS  &  55 (1948) 430-431.  Shows the period of the out shuffle is at most  2n-2.  Editorial notes cite Uspensky & Heaslet and MG 15 (1930) 17-20 ??NYS.

Charles T. Jordan.  Trailing the dovetail shuffle to its lair.  The Bat (Nov, Dec 1948;  Jan, Feb, Mar, 1949).  ??NYS -- cited by Gardner.  I have  No. 59 (Nov 1948) cover & 431-432, which reprints some of the material from his book.

Paul B. Johnson.  Congruences and card shuffling.  AMM 63 (1956) 718-719.  ??NYS -- cited by Gardner.

Alexander Elmsley.  Work in Progress.  Ibidem 11 (Sep 1957) 222.  He had previously coined the terms 'in' and 'out' and represented them by  I  and  O.  He discovers and shows that to put the top card into the  k‑th position, one writes  k-1  in binary and reads off the sequence of  1s  and  0s,  from the most significant bit, as  I  and  O  shuffles.  He asks but does not solve the question of how to move the  k-th card to the top -- see Bonfeld and Morris.

Alexander Elmsley.  The mathematics of the weave shuffle,  The Pentagram  11:9 (Jun 1957) 70‑71;  11:10 (Jul 1957) 77-79;  11:11 (Aug 1957) 85;  12 (May 1958) 62.  ??NYR -- cited by Gardner in the bibliography of Carnival, but he doesn't give the Ibidem reference in the bibliography, so there may be some confusion here??  Morris only cites Pentagram.

Solomon W. Golomb.  Permutations by cutting and shuffling.  SIAM Review 3 (1961) 293‑297.  ??NYS -- cited by Gardner.  Shows that cuts and the two shuffles generate all permutations of an even deck.  However, for an odd deck of  n  cards, the two kinds of shuffles can be intermixed and this only changes the cyclic order of the result.  Since cutting also only changes the cyclic order, the number of possible permutations is  n  times the order of the shuffle.

Gardner.  SA (Oct 1966) = Carnival, chap. 10.  Defines the in and out shuffles as above and gives the relation to the order of 2.  Notes that it is easier to do the inverse operations, which consist of extracting every other card.  Describes Elmsley's method.  Addendum says no easy method is known to determine shuffles to bring the  k‑th card to the top.

Murray Bonfeld.  A solution to Elmsley's problem.  Genii 37 (May 1973) 195-196.  Solves Elmsley's 1957 problem by use of an asymmetric in-shuffle where the top part of the deck has 25 cards, so the first top card becomes second and the last two cards remain in place.  (If one ignores the bottom two cards this is an in-shuffle of a 50 card deck.)

S. Brent Morris.  The basic mathematics of the faro shuffle.  Pi Mu Epsilon Journal 6 (1975) 86-92.  Obtains basic results, getting up to Elmsley's work.  His reference to Gardner gives the wrong year.

Israel N. Herstein & Irving Kaplansky.  Matters Mathematical.  1974;  slightly revised 2nd ed., Chelsea, NY, 1978.  Chap. 3, section 4: The interlacing shuffle, pp. 118-121.  Studies the permutation of the in shuffle, getting same results as Uspensky & Heaslet.

S. Brent Morris.  Faro shuffling and card placement.  JRM 8:1 (1975) 1-7.  Shows how to do the faro shuffle.  Gives Elmsley's and Bonfeld's results.

Persi Diaconis, Ronald L. Graham & William M. Kantor.  The mathematics of perfect shuffles.  Adv. Appl. Math. 4 (1983) 175-196.  ??NYS.

Steve Medvedoff & Kent Morrison.  Groups of perfect shuffles.  MM 60:1 (1987) 3-14.  Several further references to check.

Walter Scott.  Mathematics of card sharping.  M500 125 (Dec 1991) 1-7.  Sketches Elmsley's results.  States a peculiar method for computing the order of  2 (mod 2n+1)  based on adding translates of the binary expansion of  2n+1  until one obtains a binary number of all  1s.  The number of ones is the order  a  and the method is thus producing the smallest  a  such that  2a-1  is a multiple of  2n+1.

John H. Conway & Richard K. Guy.  The Book of Numbers.  Copernicus (Springer-Verlag), NY, 1996.  Pp. 163-165 gives a brief discussion of perfect shuffles and Monge's shuffle.

 

          5.AB.           FOLDING A STRIP OF STAMPS

 

É. Lucas.  Théorie des Nombres.  Gauthier‑Villars, Paris, 1891;  reprinted by Blanchard, Paris, 1958.  P. 120. 

Exemple II -- La bande de timbres-poste. -- De combien de manières peut-on replier, sur un seul, une bande de  p  timbres-poste?

Exemple III -- La feuille de timbres-poste. -- De combien de manières peut-on replir, sur un seul, une feuille rectangulaire de  pq  timbres-poste?

"Nous ne connaissons aucune solution de ces deux problèmes difficiles proposés par M. Em. Lemoine."

M. A. Sainte-Laguë.  Les Réseaux (ou Graphes).  Mémorial des Sciences Mathématiques, fasc. XVIII.  Gauthier-Villars, Paris, 1926.  Section 62: Problème des timbres-poste, pp. 39‑41.  Gets some basic results and finds the numbers for a strip of  n,  n = 1, 2, ..., 10  as:  1, 2, 6, 16, 50, 144, 448, 7472, 17676, 41600.

Jacques Devisme.  Contribution a l'étude du problème des timbres-poste.  Comptes-Rendus du Deuxième Congrès International de Récréation Mathématique, Paris, 1937.  Librairie du "Sphinx", Bruxelles, 1937, pp. 55-56.  Cites Lucas (but in the wrong book!) and Sainte-Laguë.  Studies the number of different forms of the result, getting numbers:  1, 2, 3, 8, 18, 44, 115, 294, 783.

 

5.AC.      PROPERTIES OF THE SEVEN BAR DIGITAL DISPLAY

                                                                                                                            ┌─┐       2

          The seven bar display, in the form of a figure  8,  as at the right, is                  │ │       1   3

now the standard form for displaying digits on calculators, clocks, etc.                       ├─┤       4

This lends itself to numerous problems of a combinatorial/numerical                          │ │       5   7

          New Section.                     └─┘                                                                    6

            For reference, we number the seven bars in the reverse-S pattern

shown.  We can then refer to a pattern by its binary  7-tuple or its decimal equivalent.  E.g. the number one is displayed by having bars  3  and  7  on, which gives a binary pattern  1000100  corresponding to decimal  68.  NOTE that there is some ambiguity with the 6 / 9.  Most versions use the upper / lower bar for these, i.e.  1101111 / 1111011,  but the bar is sometimes omitted, giving  1001111 / 1111001.  I will assume the first case unless specified.

          I have been interested in these for some time for several reasons.  First, my wife has such a clock on her side of the bed and she often has a glass of water in front of it, causing patterns to be reversed.  At other times the clock has been on the floor upside down, causing a different reversal of patterns.  Second, segments often fail or get stuck on and I have tried to analyse which would be the worst segment to fail or get stuck.  As an example, the clock in my previous car went from  16:59 to 15:00.  Third, I have analysed which segment(s) in a clock are used most/least often.

 

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 35: New numbers, pp. 26-27 & 83.  Asks for the number of new digits one can make, subject to their being connected and full height.  Says it is difficult to determine when these are distinct -- e.g. calculators differ as to the form of their 6s and 9s -- so he is not sure how to count, but he gives 22 examples.  I find there are 55 connected, full-height patterns.

Gordon Alabaster, proposer  &  Robert Hill, solver.  Problem 134.3 -- Clock watching.  M500 134 (Aug 1993) 17  &  135 (Oct 1993) 14-15.  Proposer notes that one segment of the units digit of the seconds on his station clock was stuck on, but that the sequence of symbols produced were all proper digits.  Which segment was stuck?  Asks if there are answers for 2, ..., 6  segments stuck on.  Solver gives systematic tables and discusses problems of how to determine which segment(s) are stuck and whether one can deduce the correct time when the stuck segments are known.

Martin Watson.  Email to NOBNET, 17 Apr 2000 08:17:32 PDT [NOBNET 2334].  Observes that the 10 digits have a total of 49 segments and asks if they can be placed on a  4 x 5  square grid.  He calls these forms 'digigrams'.  He had been unable to find a solution but Leonard Campbell has found 5 distinct solutions, though they do no differ greatly.  He has the pieces and some discussion on his website: http://martnal.tripod.com/puzzles.html .  Dario Uri [22 Apr 2000 14:44:35 +0200] found two extra solutions, but Rick Eason [22 Apr 2000 09:37: -0400] also found these, but points out that these have an error due to misreading the lattice which gives the two bars of the 1 being parallel instead of end to end.  Eason's program also found the 5 solutions.

 

5.AD.          STACKING A DECK TO PRODUCE A SPECIAL EFFECT

 

          New section.  This refers to the process of arranging a deck of cards or a stack of coins so that dealing it by some rule produces a special effect.  In many cases, this is just inverting the permutation given by the rule and the Josephus problem (7.B) is a special case.  Other cases involve spelling out the names of cards, etc.

 

Will Blyth.  Money Magic.  C. Arthur Pearson, London, 1926.  Alternate heads, pp. 61-63.  Stack of eight coins.  Place one on the table and the next on the bottom of the stack.  The sequence of placed coins is to alternate heads and tails.  How do you arrange the stack?  Answer is  HHTHHTTT.  This is the same process as counting out by  2s -- see 7.B.

Doubleday - 2.  1971.  Heads and tails, pp. 105-106.  Same as Blyth, but with six coins and solution  HTTTHH.

 

          5.AE.           REVERSING CUPS

 

          New section.  There are several versions of this and they usually involve parity.  The basic move is to reverse two of the cups.  The classic problem seems to be to start with  UDU  and produce  DDD  in three moves.  A trick version is to demonstrate this several times to someone and then leave him to start from  DUD.  Another easy problem is to leave three cups as they were after three moves.  This is equivalent to a  3 x 3  array with an even number in each row and column -- see 6.AO.2.  These problems must be much older than I have, but the following are the only examples I have yet noted.

 

Anonymous.  Social Entertainer and Tricks (thus on spine, but running title inside is New Book of Tricks).  Apparently a compilation with advertisements for Johnson Smith (Detroit, Michigan) products, c1890?.  P. 38a: Bottoms up.  Given  UDU,  produce  DDD  in three moves.

Young World.  c1960.  P. 39: Water switch.  Full and empty glasses:  FFFEEE.  Make them alternately full and empty in one move.

Putnam.  Puzzle Fun.  1978.

No. 3: Tea for three, pp. 1 & 25.  Cups given as  UDU.  Produce  DDD  in three moves.

No. 16: Glass alignment, pp. 5 & 28.  Six cups arranged  UUUDDD.  Produce an alternating row.  He gets  UDUDUD  in three moves.  I can get  DUDUDU  in four moves.

 

          5.AF. SPOTTING DICE

 

          New section.  In the early 1980s, I asked Richard Guy what was the 'standard' configuration for a die and later asked Ray Bathke if he used a standard pattern.  Assuming opposite sides add to seven there are two handednesses.  But also the spot pattern of the two, three and six has two orientations, giving 16 different patterns of die.  Ray said that when he furnished dice with games, some customers had sent them back because they weren't the same.  Within about three years, I had obtained examples of all sixteen patterns!  Indeed, I often found several patterns in a single batch from one manufacturer.  Ray Bathke also pointed out that the small dice that come from the oriental games have the two arranged either horizontally or vertically rather than diagonally, giving another 16 patterns.  I have only obtained five of these, but with both handednesses included.  I used this idea in one of my Brain Twisters, cf below.

          Since the 2, 3 and 6 faces all meet at a corner, one has just to describe this corner.  The 2, 3, 6 can be clockwise around the corner or anti-clockwise.  Note that  236  is clockwise if and only if  132  is clockwise.  The position of the 2 and 3 can be described by saying whether the pattern points toward or away from the corner.  If we place the 2 upward, then 6 will be a vertical face and we can describe it by saying whether the lines of three spots are vertical or horizontal.  Guy told me a system for describing a die, but it's not in Winning Ways and I've forgotten it, so I'll invent my own.

          We write the sequence  236  if  236  is arranged clockwise at the  236  corner and we write  263  otherwise.  When looked at cornerwise, with the  2  on top, the pattern of the  2  may appear vertical or horizontal.  We write  2  when it is vertical and  2  when it is horizontal.  (For oriental dice, the  2  will appear on a diagonal and can be indicated by  2  or  2.  If we now rotate the cube to bring the  3  on top, its pattern will appear either vertical or horizontal and we write  3  or  3.  Putting the  2  back on top, the  6  face will be upright and the lines of three spots will be either vertical or horizontal, which we denote by  6  or  6.

 

David Singmaster.  Dicing around.  Weekend Telegraph (16 Dec 1989).  = Games & Puzzles No. 15 (Jun 1995) 22-23  &  16 (Jul 1995) 43-44.  How many dice are there?  Describes the normal 16 and mentions the other 16.

Ian Stewart.  The lore and lure of dice.  SA (Nov 1997) ??.  He asserts that the standard pattern has  132  going clockwise at a corner, except that the Japanese use the mirror-image version in playing mah-jongg.  His picture has both 2 and 3 toward the 236 corner and the 6 being vertical, i.e. in pattern 236.  He discusses crooked dice of various sorts and that the only way to make all values from 1 to 12 equally likely is to have 123456 on one die and 000666 on the other.

Ricky Jay.  The story of dice.  The New Yorker (11 Dec 2000) 90-95. 

 

          5.AG. RUBIK'S CUBE AND SIMILAR PUZZLES

 

          I have previously avoided this as being too recent to be covered in a historical work, but it is now old enough that it needs to be covered, and there are some older references.  Much of the history is given in my Notes on Rubik's Cube and my Cubic Circular.  Jaap Scherphuis has sent me a file of puzzle patents and several dozen of them could be entered here, but I will only enter older or novel items.  Scherphuis's file has about a dozen patents for the  4 x 4 x 4  and  5 x 5 x 5  cubes!  See Section 5.A for predecessors of the idea.  However, this Section will mostly deal with puzzles where pieces are permuted without having any empty places, so these are generally permutation puzzle.

 

          5.AG.1.       RUBIK'S CUBE

          New section.  Much to be added.

 

Richard E. Korf.  Finding optimal solutions to Rubik's Cube using pattern databases.  Proc. Nat. Conf. on Artificial Intelligence (AAAI-97), Providence, Rhode Island, Jul 1997, pp. 700-705.  Studies heuristic methods of finding optimal solutions of the Cube.  Claims to be the first to find optimal solutions for random positions of the Cube -- but I think others such as Kociemba and Reid were doing it up to a decade earlier.  For ten random examples, he found optimal solutions took 16 moves in one case, 17 moves in three cases, 18 moves in six cases, from which he asserts the median optimal solution length seems to be 18.  He uses the idea of axial moves and obtains the lower bound of 18 for God's Algorithm, as done in my Notes in 1980.  Cites various earlier work in the field, but only one reference to the Cube literature.

Richard E. Korf  &  Ariel Felner.  Disjoint pattern database heuristics.  Artificial Intelligence 134 (2002) 9-22.  Discusses heuristic methods of solving the Fifteen Puzzle, Rubik's Cube, etc.  Asserts the median optimal solution length for the Cube is only 18.  Seems to say one of the problems in the earlier paper took a couple of weeks running time, but improved methods of Kociemba and Reid can find optimal solutions in about an hour. 

 

          5.AG.2.       HUNGARIAN RINGS, ETC.

 

          New section.  Much to be added.

 

William Churchill.  US Patent 507,215 -- Puzzle.  Applied: 28 May 1891;  patented: 24 Oct 1893.  1p + 1p diagrams.  Two rings of 22 balls, intersecting six spaces apart.

Hiester Azarus Bowers.  US Patent 636,109 -- Puzzle.  Filed: 16 Aug 1899;  patented 31 Oct 1899.  2pp + 1p diagrams.  4 rotating discs which overlap in simple lenses.

Ivan Moscovich.  US Patent 4,509,756 -- Puzzle with Elements Transferable Between Closed‑loop Paths.  Filed: 18 Dec 1981;  patented: 9 Apr 1985.  Cover page + 3pp + 2pp diagrams.  Two rings of 18 balls, each stretched to have two straight sections with semicircular ends.  The rings cross in four places, at the ends of the straight sections, so adjacent crossing points are separated by two balls.  I'm not sure this was ever produced.  Mentions three circular rings version, but there each pair of rings only overlaps in two places so this is a direct generalization of the Hungarian Rings.

David Singmaster.  Hungarian Rings groups.  Bull. Inst. Math. Appl. 20:9/10 (Sep/Oct 1984) 137-139.  [The results were stated in Cubic Circular 5 & 6 (Autumn & Winter 1982) 9‑10.]  An article by Philippe Paclet [Des anneaux et des groupes; Jeux et Stratégie 16 (Aug/Sep 1982) 30-32] claimed that all puzzles of two rings have groups either the symmetric or the alternating group on the number of balls.  This article shows this is false and determines the group in all cases.  If we have rings of size  m, n  and the intersections are distances  a, b  apart on the two rings.  Then the group,  G(m, n, a, b)  is the symmetric group on  m+n-2  if  mn  is even and is the alternating group if  mn  is odd; except that  G(4, 4, 1, 1)  is the exceptional group described in R. M. Wilson's 1974 paper: Graph puzzles, homotopy and the alternating group -- cited in Section 5.A under The Fifteen Puzzle -- and is also the group generated by two adjacent faces on the Rubik Cube acting on the six corners on those faces; and except that  G(2a, 2b, a, b)  keeps antipodal pairs at antipodes and hence is a subgroup of the wreath product  Z2 wr Sa+b‑1, with three cases depending on the parities of  a  and  b.

Bala Ravikumar.  The Missing Link and the Top-Spin.  Report TR94-228, Department of Computer Science and Statistics, University of Rhode Island, Jan 1994.  Top-Spin has a cycle of 20 pieces and a small turntable which permits inverting a section of four pieces.  After developing the group theory and doing the Fifteen Puzzle and the Missing Link, he shows the state space of Top-Spin is  S20. 

 

6.       GEOMETRIC RECREATIONS

 

          6.A.   PI

 

          This is too big a topic to cover completely.  The first items should be consulted for older material and the general history.  Then I include material of particular interest.  See also 6.BL which has some formulae which are used to compute  π.  I have compiled a separate file on the history of  π.

 

Augustus De Morgan.  A Budget of Paradoxes.  (1872);  2nd ed., edited by D. E. Smith, (1915), Books for Libraries Press, Freeport, NY, 1967.

J. W. Wrench Jr.  The evolution of extended decimal approximations to  π.  MTr 53 (Dec 1960) 644‑650.  Good survey with 55 references, including original sources.

Petr Beckmann.  A History of  π.  The Golem Press, Boulder, Colorado, (1970);  2nd ed., 1971.

Lam Lay-Yong & Ang Tian-Se.  Circle measurements in ancient China.  HM 13 (1986) 325‑340.  Good survey of the calculation of  π  in China.

Dario Castellanos.  The ubiquitous  π.  MM 61 (1988) 67-98  &  148-163.  Good survey of methods of computing  π. 

Joel Chan.  As easy as pi.  Math Horizons 1 (Winter 1993) 18-19.  Outlines some recent work on calculating π and gives several of the formulae used.

David Singmaster.  A history of π.  M500 168 (Jun 1999) 1-16.  A chronology.  (Thanks to Tony Forbes and Eddie Kent for carefully proofreading and amending my file.)

 

Aristophanes.  The Birds.  ‑414.  Lines 1001‑1005.  In:  SIHGM I 308‑309.  Refers to 'circle-squarers', possibly referring to the geometer/astronomer Meton.

 

E. J. Goodwin.  Quadrature of the circle.  AMM 1 (1894) 246‑247.

House Bill No. 246, Indiana Legislature, 1897.  "A bill for an act introducing a new mathematical truth ..."  In Edington's paper (below), p. 207, and in several of the newspaper reports.

(Indianapolis) Journal (19 Jan 1897) 3.  Mentions the Bill in the list of bills introduced.

Die Quadratur des Zirkels.  Täglicher Telegraph (Indianapolis) (20 Jan 1897) ??.  Surveys attempts since -2000 and notes that Lindemann and Weierstrass have shown that the problem is impossible, like perpetual motion.

A man of 'genius'.  (Indianapolis) Sun (6 Feb 1897) ??.  An interview with Goodwin, who says: "The astronomers have all been wrong.  There's about 40,000,000 square miles on the surface of this earth that isn't here."  He says his results are revelations and gives several rules for the circle and the sphere.

Mathematical Bill passed.  (Indianapolis) Journal (6 Feb 1897) 5.  "This is the strangest bill that has ever passed an Indiana Assembly."  Gives whole text of the Bill.

Dr. Goodwin's theaorem (sic)  Resolution adopted by the House of Representatives.  (Indianapolis) News (6 Feb 1897) 4.  Gives whole text of the Bill. 

The Mathematical Bill  Fun-making in the Senate yesterday afternoon -- other action.  (Indianapolis) News (13 Feb 1897) 11.  "The Senators made bad puns about it, ...."  The Bill was indefinitely postponed.

House Bills in the Senate.  (Indianapolis) Sentinel (13 Feb 1897) 2.  Reports the Bill was killed.

(No heading??)  (Indianapolis) Journal (13 Feb 1897) 3, col. 4.  "... indefinitely postponed, as not being a subject fit for legislation."

Squaring the circle.  (Indianapolis) Sunday Journal (21 Feb 1897) 9.  Says Goodwin has solved all three classical impossible problems.  Says  π = 3.2, using the fact that  Ö2 = 10/7,  giving diagrams and a number of rules. 

My thanks to Underwood Dudley for locating and copying the above newspaper items.

C. A. Waldo.  What might have been.  Proc. Indiana Acad. Science 26 (1916) 445‑446.

W. E. Edington.  House Bill No. 246, Indiana State Legislature, 1897.  Ibid. 45 (1935) 206‑210.

A. T. Hallerberg.  House Bill No. 246 revisited.  Ibid. 84 (1975) 374‑399.

Manuel H. Greenblatt.  The 'legal' value of pi, and some related mathematical anomalies.  American Scientist 53 (Dec 1965) 427A‑434A.  On p. 427A he tries to interpret the bill and obtains three different values for  π.

David Singmaster.  The legal values of pi.  Math. Intell. 7:2 (1985) 69‑72.  Analyses Goodwin's article, Bill and other assertions to find 23 interpretable statements giving 9 different values of  π !

Underwood Dudley.  Mathematical Cranks.  MAA, 1992.  Legislating pi, pp. 192-197.

 

C. T. Heisel.  The Circle Squared Beyond Refutation.  Published by the author, 657 Bolivar Rd., Cleveland, Ohio, 1st ed., 1931, printed by S. J. Monck, Cleveland;  2nd ed., 1934, printed by Lezius‑Hiles Co., Cleveland, ??NX  +  Supplement: "Fundamental Truth", 1936, ??NX,  distributed by the author from 2142 Euclid Ave., Cleveland.  This is probably the most ambitious publication of a circle-squarer -- Heisel distributed copies all around the world.

 

Underwood Dudley.  πt:  1832-1879.  MM 35 (1962) 153-154.  He plots 45 values of  π  as a function of time over the period 1832-1879 and finds the least-squares straight line which fits the data, finding that  πt = 3.14281 + .0000056060 t,  for  t  measured in years AD.  Deduces that the Biblical value of 3 was a good approximation for the time and that Creation must have occurred when  πt = 0,  which was in  -560,615.

Underwood Dudley.  πt.  JRM 9 (1976-77) 178 & 180.  Extends his previous work to 50 values of  π  over 1826-1885, obtaining  πt = 4.59183 - .000773 t.  The fact that  πt is decreasing is worrying -- when  πt = 1,  all circles will collapse into straight lines and this will certainly be the end of the world, which is expected in 4646 on 9 Aug at 20:55:33 -- though this is only the expected time and there is considerable variation in this prediction.  [Actually, I get that this should be on 11 Aug.  However, it seems to me that circles will collapse once  πt = 2,  as then the circumference corresponds to going back and forth along the diameter.  This will occur when  t = 3352.949547,  i.e. in 3352, on 13 Dec at 14:01:54 -- much earlier than Dudley's prediction, so start getting ready now!]

 

          6.B.   STRAIGHT LINE LINKAGES

 

          See Yates for a good survey of the field.

 

James Watt.  UK Patent 1432 -- Certain New Improvements upon Fire and Steam Engines, and upon Machines worked or moved by the same.  Granted: 28 Apr 1784;  complete specification: 24 Aug 1784.  14pp + 1 plate.  Pp. 4-6 & Figures 7‑12 describe Watt's parallel motion.  Yates, below, p. 170 quotes one of Watt's letters:  "... though I am not over anxious after fame, yet I am more proud of the parallel motion than of any other invention I have ever made."

P. F. Sarrus.  Note sur la transformation des mouvements rectilignes alternatifs, en mouvements circulaires; et reciproquement.  C. R. Acad. Sci. Paris 36 (1853) 1036‑1038.  6 plate linkage.  The name should be Sarrus, but it is printed Sarrut on this and the following paper.

Poncelet.  Rapport sur une transformation nouvelle des mouvements rectilignes alternatifs en mouvements circulaires et reciproquement, par Sarrut.  Ibid., 36 (1853) 1125‑1127.

A. Peaucellier.  Lettre au rédacteur.  Nouvelles Annales de Math. (2) 3 (1864) 414‑415.  Poses the problem.

A. Mannheim.  Proces‑Verbaux des sceances des 20 et 27 Juillet 1867.  Bull. Soc. Philomathique de Paris (1867) 124‑126.  ??NYS.  Reports Peaucellier's invention.

Lippman Lipkin.  Fortschritte der Physik (1871) 40‑??  ??NYS

L. Lipkin.  Über eine genaue Gelenk‑Geradführung.  Bull. Acad. St. Pétersbourg [=? Akad. Nauk, St. Petersburg, Bull.] 16 (1871) 57‑60.  ??NYS

L. Lipkin.  Dispositif articulé pour la transformation rigoureuse du mouvement circulaire en mouvement rectiligne.  Revue Univers. des Mines et de la Métallurgie de Liége 30:4 (1871) 149‑150.  ??NYS.  (Now spelled Liège.)

A. Peaucellier.  Note sur un balancier articulé a mouvement rectiligne.  Journal de Physique 2 (1873) 388‑390.  (Partial English translation in Smith, Source Book, vol. 2, pp. 324‑325.) Says he communicated it to Soc. Philomath. in 1867 and that Lipkin has since also found it.  There is also an article in Nouv. Annales de Math. (2) 12 (1873) 71‑78 (or 73?), ??NYS.

E. Lemoine.  Note sur le losange articulé du Commandant du Génie Peaucellier, destiné a remplacer le parallélogramme de Watt.  J. de Physique 2 (1873) 130‑134.  Confirms that Mannheim presented Peaucellier's cell to Soc. Philomath. on 20 Jul 1867.  Develops the inversive geometry of the cell.

[J. J. Sylvester.]  Report of the Annual General Meeting of the London Math. Soc. on 13 Nov 1873.  Proc. London Math. Soc. 5 (1873) 4 & 141.  On p. 4 is:  "Mr. Sylvester then gave a description of a new instrument for converting circular into general rectilinear motion, and into motion in conics and higher plane curves, and was warmly applauded at the close of his address."  On p. 141 is an appendix saying that Sylvester spoke "On recent discoveries in mechanical conversion of motion" to a Friday Evening's Discourse at the Royal Institution on 23 Jan 1874.  It refers to a paper 20 pages long but is not clear if or where it was published.

H. Hart.  On certain conversions of motion.  Cambridge Messenger of Mathematics 4 (1874) 82‑88 and 116‑120 & Plate I.  Hart's 5 bar linkage.  Obtains some higher curves.

A. B. Kempe.  On some new linkages.  Messenger of Mathematics 4 (1875) 121‑124 & Plate I.  Kempe's linkages for reciprocating linear motion.

H. Hart.  On two models of parallel motions.  Proc. Camb. Phil. Soc. 3 (1876‑1880) 315‑318.  Hart's parallelogram (a 5 bar linkage) and a 6 bar one.

V. Liguine.  Liste des travaux sur les systèms articulés.  Bull. d. Sci. Math. 18 (or (2) 7) (1883) 145‑160.  ??NYS ‑ cited by Kanayama.  Archibald; Outline of the History of Mathematics, p. 99, says Linguine is entirely included in Kanayama.

Gardner D. Hiscox.  Mechanical Appliances  Mechanical Movements  and  Novelties of Construction.  A second volume to accompany his previous Mechanical Movements, Powers and Devices.  Norman W. Henley Publishing Co, NY, (1904), 2nd ed., 1910.  This is filled with many types of mechanisms.  Pp. 245-247 show five straight-line linkages and some related mechanisms.

(R. Kanayama).  (Bibliography on linkages.  Text in Japanese, but references in roman type.)  Tôhoku Math. J. 37 (1933) 294‑319.

R. C. Archibald.  Bibliography of the theory of linkages.  SM 2 (1933‑34) 293‑294.  Supplement to Kanayama.

Robert C. Yates.  Geometrical Tools.  (As:  Tools; Baton Rouge, 1941);  revised ed., Educational Publishers, St. Louis, 1949.  Pp. 82-101 & 168-191.  Gets up to outlining Kempe's proof that any algebraic curve can be drawn by a linkage.

R. H. Macmillan.  The freedom of linkages.  MG 34 (No. 307) (Feb 1960) 26‑37.  Good survey of the general theory of linkages.

Michael Goldberg.  Classroom Note 312:  A six‑plate linkage in three dimensions.  MG 58 (No. 406) (Dec 1974) 287‑289.

 

          6.C.   CURVES OF CONSTANT WIDTH

 

          Such curves play an essential role in some ways to drill a square hole, etc.

 

L. Euler.  Introductio in Analysin Infinitorum.  Bousquet, Lausanne, 1748.  Vol. 2, chap. XV, esp. § 355, p. 190 & Tab. XVII, fig. 71.  = Introduction to the Analysis of the Infinite; trans. by John D. Blanton; Springer, NY, 1988-1990; Book II, chap. XV: Concerning curves with one or several diameters, pp. 212-225, esp. § 355, p. 221 & fig. 71, p. 481.  This doesn't refer to constant width, but fig. 71 looks very like a Reuleaux triangle.

L. Euler.  De curvis triangularibus.  (Acta Acad. Petropol. 2 (1778(1781)) 3‑30)  = Opera Omnia (1) 28 (1955) 298‑321.  Discusses triangular versions.

M. E. Barbier.  Note sur le problème de l'aiguille et jeu du joint couvert.  J. Math. pures appl. (2) 5 (1860) 273‑286.  Mentions that  perimeter = π * width.

F. Reuleaux.  Theoretische Kinematik; Vieweg, Braunschweig, 1875.  Translated:  The Kinematics of Machinery.  Macmillan, 1876;  Dover, 1964.  Pp. 129‑147.

Gardner D. Hiscox.  Mechanical Appliances  Mechanical Movements  and  Novelties of Construction.  A second volume to accompany his previous Mechanical Movements, Powers and Devices.  Norman W. Henley Publishing Co, NY, (1904), 2nd ed., 1910. 

Item 642: Turning a square by circular motion, p. 247.  Plain face, with four pins forming a centred square, is turned by the lathe.  A triangular follower is against the face, so it is moved in and out as a pin moves against it.  This motion is conveyed by levers to the tool which moves in and out against the work which is driven by the same lathe.

Item 681: Geometrical boring and routing chuck, pp. 257-258.  Shows it can make rectangles, triangles, stars, etc.  No explanation of how it works.

Item 903A: Auger for boring square holes, pp. 353-354.  Uses two parallel rotating cutting wheels.

H. J. Watts.  US Patents 1,241,175‑7 -- Floating tool‑chuck;  Drill or boring member;  Floating tool‑chuck.  Applied: 30 Nov 1915;  1 Nov 1916;  22 Nov 1916;  all patented: 25 Sep 1917.  2 + 1,  2 + 1,  4 + 1  pp + pp diagrams.  Devices for drilling square holes based on the Reuleaux triangle.

T. Bonnesen & W. Fenchel.  Theorie der konvexen Körper.  Berlin, 1934;  reprinted by Chelsea, 1971.  Chap. 15: Körper konstanter Breite, pp. 127‑141.  Surveys such curves with references to the source material.

G. D. Chakerian & H. Groemer.  Convex bodies of constant width.  In:  Convexity and Its Applications; ed. by Peter M. Gruber & Jörg M. Wills; Birkhäuser, Boston, 1983.  Pp. 49‑96.  (??NYS -- cited in MM 60:3 (1987) 139.)  Bibliography of some 250 items since 1930.

 

          6.D.   FLEXAGONS

 

          These were discovered by Arthur H. Stone, an English graduate student at Princeton in 1939.  American paper was a bit wider than English and would not fit into his notebooks, so he trimmed the edge off and had a pile of long paper strips which he played with and discovered the basic flexagon.  Fellow graduate students Richard P. Feynman, Bryant Tuckerman and John W. Tukey joined in the investigation and developed a considerable theory.  One of their fathers was a patent attorney and they planned to patent the idea and began to draw up an application, but the exigencies of the 1940s led to its being put aside, though knowledge of it spread as mathematical folklore.  E.g. Tuckerman's father, Louis B. Tuckerman, lectured on it at the Westinghouse Science Talent Search in the mid 1950s.

          S&B, pp. 148‑149, show several versions.  Most square versions (tetraflexagons or magic books) don't fold very far and are really just extended versions of the Jacob's Ladder -- see 11.L

 

Martin Gardner.  Cherchez la Femme [magic trick].  Montandon Magic Co., Tulsa, Okla., 1946.  Reproduced in:  Martin Gardner Presents; Richard Kaufman and Alan Greenberg, 1993, pp. 361-363.  [In:  Martin Gardner Presents, p. 404, this is attributed to Gardner, but Gardner told me that Roger Montandon had the copyright -- ??  I have learned a little more about Gardner's early life -- he supported himself by inventing and selling magic tricks about this time, so it may be that Gardner devised the idea and sold it to Montandon.].  A hexatetraflexagon.

"Willane".  Willane's Wizardry.  Academy of Recorded Crafts, Arts and Sciences, Croydon, 1947.  A trick book, pp. 42-43.  Same hexatetraflexagon.

Sidney Melmore.  A single‑sided doubly collapsible tessellation.  MG 31 (No. 294) (1947) 106.  Forms a Möbius strip of three triangles and three rhombi.  He sees it has two distinct forms, but doesn't see the flexing property!!

Margaret Joseph.  Hexahexaflexagrams.  MTr 44 (Apr 1951) 247‑248.  No history.

William R. Ransom.  A six‑sided hexagon.  SSM 52 (1952) 94.  Shows how to number the 6 faces.  No history.

F. G. Maunsell.  Note 2449:  The flexagon and the hexahexaflexagram.  MG 38 (No. 325) (Sep 1954) 213‑214.  States that Joseph is first article in the field and that this is first description of the flexagon.  Gives inventors' names, but with Tulsey for Tukey.

R. E. Rogers & Leonard L. D'Andrea.  US Patent 2,883,195 -- Changeable Amusement Devices and the Like.  Applied: 11 Feb 1955;  patented: 21 Apr 1959.  2pp + 1p correction + 2pp diagrams.  Clearly shows the 9 and 18 triangle cases and notes that one can trim the triangles into hexagons so the resulting object looks like six small hexagons in a ring.

M. Gardner.  Hexa‑hexa‑flexagon  and  Cherchez la femme.  Hugard's MAGIC Monthly 13:9 (Feb 1956) 391.  Reproduced in his:  Encyclopedia of Impromptu Magic; Magic Inc., Chicago, 1978, pp. 439-442.  Describes hexahexa and the hexatetra of Gardner/Montandon & Willane.

M. Gardner.  SA (Dec 1956) = 1st Book, chap. 1.  His first article in SA!!

Joan Crampin.  Note 2672:  On note 2449.  MG 41 (No. 335) (Feb 1957) 55‑56.  Extends to a general case having  9n  triangles of  3n  colours.

C. O. Oakley & R. J. Wisner.  Flexagons.  AMM 64:3 (Mar 1957) 143‑154.

Donovan A. Johnson.  Paper Folding for the Mathematics Class.  NCTM, 1957, section 61, pp. 24-25: Hexaflexagons.  Describes the simplest case, citing Joseph.

Roger F. Wheeler.  The flexagon family.  MG 42 (No. 339) (Feb 1958) 1‑6.  Improved methods of folding and colouring.

M. Gardner.  SA (May 1958) = 2nd Book, chap. 2.  Tetraflexagons and flexatube.

P. B. Chapman.  Square flexagons.  MG 45 (1961) 192‑194.  Tetraflexagons.

Anthony S. Conrad  &  Daniel K. Hartline.  Flexagons.  TR 62-11, RIAS, (7212 Bellona Avenue, Baltimore 12, Maryland,) 1962, 376pp.  This began as a Science Fair project in 1956 and was then expanded into a long report.  The authors were students of Harold V. McIntosh who kindly sent me one of the remaining copies in 1996.  They discover how to make any chain of polygons into a flexagon, provided certain relations among angles are satisfied.  The bibliography includes almost all the preceding items and adds the references to the Rogers & D'Andrea patent, some other patents (??NYS) and a number of ephemeral items:  Conrad produced an earlier RIAS report, TR 60-24, in 1960;  Allan Phillips wrote a mimeographed paper on hexaflexagons;  McIntosh wrote an unpublished paper on flexagons;  Mike Schlesinger wrote an unpublished paper on Tuckerman tree theory. 

Sidney H. Scott.  How to construct hexaflexagons.  RMM 12 (Dec 1962) 43‑49.

William R. Ransom.  Protean shapes with flexagons.  RMM 13 (Feb 1963) 35‑37.  Describes 3‑D shapes that can be formed.  c= Madachy, below.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  The magic book, pp. 124-125.  As in Gardner's Cherchez la Femme and Willane.

Pamela Liebeck.  The construction of flexagons.  MG 48 (No. 366) (Dec 1964) 397‑402.

Joseph S. Madachy.  Mathematics on Vacation.  Op. cit. in 5.O, (1966), 1979.  Other flexagon diversions, pp. 76‑81.  Describes 3‑D shapes that one can form.  Based on Ransom, RMM 13.

Lorraine Mottershead.  Investigations in Mathematics.  Blackwell, Oxford, 1985.  Pp. 66-75.  Describes various tetra- and hexa-flexagons.

Douglas A. Engel.  Hexaflexagon + HFG = slipagon!  JRM 25:3 (1993) 161-166.  Describes his slipagons, which are linked flexagons.

Robert E. Neale (154 Prospect Parkway, Burlington, Vermont, 05401, USA).  Self-designing tetraflexagons.  12pp document received in 1996 describing several ways of making tetraflexagons without having to tape or paste.  He starts with a creased square sheet, then makes some internal tears or cuts and then folds things through to miraculously obtain a flexagon!  A slightly rearranged version appeared in: Elwyn R. Berlekamp  &  Tom Rodgers, eds.;  The Mathemagician and Pied Puzzler  A Collection in Tribute to Martin Gardner;  A. K. Peters, Natick, Massachusetts, 1999, pp. 117-126.

Jose R. Matos.  US Patent 5,735,520 -- Fold-Through Picture Puzzle.  Applied: 7 Feb 1997;  patented: 7 Apr 1998.  Front page + 6pp diagrams + 13pp text.  Robert Byrnes sent an example of the puzzle.  This is a square in thin plastic, 100mm on an edge.  Imagine a  2 x 2  array of squares with their diagonals drawn.  Fold along all the diagonals and between the squares.  This gives an array of 16 isosceles right triangles.  Now cut from the centres of the four squares to the centre of the whole array.  This produces an  X  cut in the middle.  This object can now be folded through itself in various ways to produce a double thickness square of half the area with various logos.  The example is 100mm along the edge of the large square and has four logos advertising Beanoland (at Chessington, 3 versions) and Strip Cheese.  The patent is assigned to Lulirama International, but Byrnes says it has not been a commercial success as it is too complicated.  The patent cites 19 earlier patents, back to 1881, and discusses the history of such puzzles.  It also says the puzzle can form three dimensional objects.

 

          6.E.    FLEXATUBE

 

          This is the square cylindrical tube that can be inverted by folding.  It was also invented by Arthur H. Stone, c1939, cf 6.D.

 

J. Leech.  A deformation puzzle.  MG 39 (No. 330) (Dec 1955) 307.  Doesn't know source.  Says there are three solutions.

M. Gardner.  Flexa-tube puzzle.  Ibidem 7 (Sep 1956) 129.  Cites the inventors of the flexagons and the articles of Maunsell and Leech (but he doesn't have its details).  (I have a note that this came with attached sample, but the copy I have doesn't indicate such.)

T. S. Ransom.  Flexa-tube solution.  Ibidem 9 (Mar 1957) 174.

M. Gardner.  SA (May 1958) = 2nd Book, chap. 2.  Says Stone invented it and shows Ransom's solution.

H. Steinhaus.  Mathematical Snapshots.  Not in the Stechert, NY, 1938, ed. nor the OUP, NY, 1950 ed.  OUP, NY:  1960: pp. 189‑193 & 326;  1969 (1983): pp. 177-181 & 303.  Erroneous attribution to the Dowkers.  Shows a different solution than Ransom's.

John Fisher.  John Fisher's Magic Book.  Muller, London, 1968.  Homage to Houdini, pp. 152‑155.  Detailed diagrams of the solution, but no history.

Highland Games (2 Harpers Court, Dingwall, Ross-Shire, IV15 9HT) makes a version called Table Teaser, made in a strip with end pieces magnetic.  Pieces are coloured so to produce several folding and inverting problems other than the usual one.  Bought in 1995.

 

          6.F.    POLYOMINOES, ETC.

 

          See S&B, pp. 15‑18.  See 6.F.1, 6.F.3, 6.F.4 & 6.F.5 for early occurrences of polyominoes.  See Lammertink, 1996 & 1997 for many examples in two and three dimensions.

          NOTATION.  Each of the types of puzzle considered has a basic unit and pieces are formed from a number of these units joined edge to edge.  The notation  N: n1, n2, ....  denotes a puzzle with  N  pieces, of which  ni  pieces consist of  i  basic units.  If  ni are single digit numbers, the intervening commas and spaces will be omitted, but the digits will be grouped by fives, e.g.  15: 00382 11.

 

Polyiamonds:  Scrutchin;  John Bull;  Daily Sketch;  Daily Mirror;  B. T.s Zig-Zag;  Daily Mail;  Miller (1960);  Guy (1960);  Reeve & Tyrrell;  O'Beirne (2 & 9 Nov 1961);  Gardner (Dec 1964 & Jul 1965);  Torbijn;  Meeus;  Gardner (Aug 1975);  Guy (1996, 1999);  Knuth,

Polycubes:  Rawlings (1939);  Editor (1948);  Niemann (1948);  French (1948);  Editor (1948);  Niemann (1948);  Gardner (1958);  Besley (1962);  Gardner (1972)

Solid Pentominoes:  Nixon (1948);  Niemann (1948);  Gardner (1958);  Miller (1960);  Bouwkamp (1967, 1969, 1978);  Nelson (2002); 

Cylindrical Pentominoes:  Yoshigahara (1992);

Polyaboloes:  Hooper (1774);  Book of 500 Puzzles (1859);  J. M. Lester (1919);  O'Beirne (21 Dec 61  &  18 Jan 62)

Polyhexes:  Gardner (1967);  Te Riele & Winter

Polysticks:  Benjamin;  Barwell;  General Symmetrics;  Wiezorke & Haubrich;  Knuth;  Jelliss;

Polyrhombs or Rhombiominoes:  Lancaster (1918);  Jones (1992).

Polylambdas:  Roothart.

Polyspheres -- see Section 6.AZ.

 

                    GENERAL REFERENCES

 

G. P. Jelliss.  Special Issue on Chessboard Dissections.  Chessics 28 (Winter 1986) 137‑152.  Discusses many problems and early work in Fairy Chess Review.

Branko Grünbaum & Geoffrey C. Shephard.  Tilings and Patterns.  Freeman, 1987.  Section 9.4: Polyiamonds, polyominoes and polyhexes, pp. 497-511.  Good outline of the field with a number of references otherwise unknown.

Michael Keller.  A polyform timeline.  World Game Review 9 (Dec 1989) 4-5.  This outlines the history of polyominoes and other polyshapes.  Keller and others refer to polyaboloes as polytans.

Rodolfo Marcelo Kurchan (Parana 960 5 "A", 1017 Buenos Aires, Argentina).  Puzzle Fun, starting with No. 1 (Oct 1994).  This is a magazine entirely devoted to polyomino and other polyform puzzles.  Many of the classic problems are extended in many ways here.  In No. 6 (Aug 1995) he presents a labelling of the 12 hexiamonds by the letters  A, C, H, I, J, M, O, P, S, V, X, Y,  which he obtained from Anton Hanegraaf.  I have never seen this before.

 

Hooper.  Rational Recreations.  Op. cit. in 4.A.1.  1774.  Vol. 1, recreation 23, pp. 64-66.  Considers figures formed of isosceles right triangles.  He has eight of these, coloured with eight colours, and uses some of them to form "chequers or regular four-sided figures, different either in form or colour".

Book of 500 Puzzles.  1859.  Triangular problem, pp. 74-75.  Identical to Hooper, dropping the last sentence.

Dudeney.  CP.  1907.  Prob. 74: The broken chessboard, pp. 119‑121 & 220‑221.  The 12 pentominoes and a  2 x 2.

A. Aubry.  Prob. 3224.  Interméd. Math 14 (1907) 122-124.  ??NYS -- cited by Grünbaum & Shephard who say Aubry has something of the idea or the term polyominoes.

G. Quijano.  Prob. 3430.  Interméd. Math 15 (1908) 195.  ??NYS -- cited by Grünbaum & Shephard, who say he first asked for the number of  n‑ominoes.

Thomas Scrutchin.  US Patent 895,114 -- Puzzle.  Applied: 20 Feb 1908;  patented: 4 Aug 1908.  2pp + 1p diagrams.  Mentioned in S&B, p. 18.  A polyiamond puzzle -- triangle of side 8, hence with 64 triangles, apparently cut into 10 pieces (my copy is rather faint -- replace??).

Thomas W. Lancaster.  US Patent 1,264,944 -- Puzzle.  Filed: 7 May 1917;  patented: 7 May 1918.  2pp + 1p diagrams.  For a general polyrhomb puzzle making a rhombus.  His diagram shows an  11 x 11  rhombus filled with  19  pieces formed from  4  to  10  rhombuses.

John Milner Lester.  US Patent 1,290,761 -- Game Apparatus.  Filed: 6 Feb 1918;  patented: 7 Jan 1919.  2pp + 3pp diagrams.  Fairly general assembly puzzle claims.  He specifically illustrates a polyomino puzzle and a polyabolo puzzle.  The first has a Greek cross of edge  3  (hence containing  45  unit cells) to be filled with polyominoes --  11: 01154.  The second has an  8-pointed star formed by superimposing two  4 x 4  squares.  This has area  20  and hence contains  40  isosceles right triangles of edge  1, which is the basic unit of this type of puzzle.  There are  11: 0128  pieces.

Blyth.  Match-Stick Magic.  1921.  Spots and squares, pp. 68-73.  He uses matchsticks broken in thirds, so it is easier to describe with units of one-third.  6 units,  4 doubles and  2 triples.  Some of the pieces have black bands or spots.  Object is to form polyomino shapes without pieces crossing, but every intersection must have a black spot.  19  polyomino shapes are given to construct, including  7  of the pentominoes, though some of the shapes are only connected at corners.

"John Bull" Star of Fortune Prize Puzzle.  1922.  This is a puzzle with  20  pieces, coloured red on one side, containing  6  through  13  triangles to be assembled into a star of David with  4  triangles along each edge (hence  12 x 16 = 192  triangles).  Made by Chad Valley.  Prize of £250 for a red star matching the key solution deposited at a bank;  £150 for solution closest to the key;  £100 for a solution with  10  red and  10  grey pieces, or as nearly as possible.  Closing date of competition is 27 Dec 1922.  Puzzle made by Chad Valley Co. as a promotional item for John Bull magazine, published by Odhams Press.  A copy is in the toy shop of the Buckleys Shop Museum, Battle, East Sussex, to whom I am indebted for the chance to examine the puzzle and a photocopy of the puzzle, box and solution.

Daily Sketch Jig-Saw Puzzle.  By Chad Valley.  Card polyiamonds.  39: 0,0,1,5,6, 12,9,6,  with a path printed on one side, to assemble into a shape of  16  rows of  15  with four corners removed and so the printed sides form a continuous circuit.  In box with shaped bottom.  Instructions on inside cover and loose sheet to submit solution.  No dates given, but appears to be 1920s, though it is somewhat similar to the Daily Mail Crown Puzzle of 1953 -- cf below -- so it might be much later.

Daily Mirror Zig-Zag £500 Prize Puzzle.  By Chad Valley.  Card polyiamonds.  29: 0,0,1,1,4,  5,6,3,1,4,  1,2,1.  One-sided pieces to fit into frame in card box.  Three pieces are duplicated and one is triplicated.  Solution and claim instructions appeared in Daily Mirror (17 Jan 1930) 1-2.  See: Tom Tyler & Felicity Whiteley; Chad Valley Promotional Jig-Saw Puzzles; Magic Fairy Publishing, Petersfield, Hampshire, 1990, p. 55. 

B. T.s Zig-Zag.  B.T. is a Copenhagen newspaper.  Polyiamond puzzle.  33: 0,0,1,2,5,  6,7,2,2,4,  1,1,1,  Some repetitions, so I only see 20 different shapes.  To be fit into an irregular frame.  Solution given on 23 Nov 1931, pp. 1-2.  (I have a photocopy of the form to fill in; an undated set of rules, apparently from the paper, saying the solutions must be received by 21 Nov; and the pages giving the solution; provided by Jan de Geus.)

Herbert D. Benjamin.  Problem 1597: A big cutting-out design -- and a prize offer.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:9 (Dec 1934) 92.  Finds the  35  hexominoes and asks if they form a  14 x 15  rectangle.  Cites Dudeney (Tribune (20 Dec 1906));  Loyd (OPM (Apr-Jul 1908)) (see 6.F.1);  Dudeney (CP, no. 74) (see above)  and some other chessboard dissections.  Jelliss says this is the first dissection problem in this journal.

F. Kadner.  Solution 1597.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:10 (Feb 1935) 104-105.  Shows the  35  hexominoes cannot tile a rectangle by two arguments, both essentially based on two colouring.  Gives some other results and some problems are given as 1679-1681 -- ??NYS.

William E. Lester.  Correction to 1597.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:11 (Apr 1935) 121.  Corrects an error in Kadner.  Finds a number of near-solutions.  Editor says Kadner insists the editor should take credit for the two-colouring form of the previous proof.

Frans Hansson, proposer & solver?.  Problem 1844.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:12 (Jun 1935) 128  &  2:13 (Aug 1935) 135.  Finds both  3 x 20  pentomino rectangles.

W. E. Lester & B. Zastrow, proposers.  Problem 1923.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:13 (Aug 1935) 138.  Take an  8 x 8  board and remove its corners.  Fill this with the  12  pentominoes.

H. D. Benjamin, proposer.  Problem 1924.  Problemist Fairy Chess Supplement (later called Fairy Chess Review) 2:13 (Aug 1935) 138.  Dissect an  8 x 8  into the  12  pentominoes and the I-tetromino.  Need solution -- ??NYS.

Thomas Rayner Dawson & William E. Lester.  A notation for dissection problems.  Fairy Chess Review 3:5 (Apr 1937) 46‑47.  Gives all  n‑ominoes up to  n = 6.  Describes the row at a time notation.  Shows the pentominoes and a  2 x 2  cover the chessboard with the  2 x 2  in any position.  Asserts there are  108  7‑ominoes and  368  8‑ominoes -- citing F. Douglas & W. E. L[ester] for the hexominoes and J. Niemann for the heptominoes.

H. D. Benjamin, proposer.  Problem 3228.  Fairy Chess Review 3:12 (Jun 1938) 129.  Dissect a  5 x 5  into the five tetrominoes and a pentomino so that the pentomino touches all the tetrominoes along an edge.  Asserts the solution is unique.  Refers to problems 3026‑3030 -- ??NYS.

H. D. Benjamin, proposer.  Problem 3229.  Fairy Chess Review 3:12 (Jun 1938) 129.  Dissect an  8 x 8  into the  12  pentominoes and a tetromino so that all pieces touch the edge of the board.  Asserts only one tetromino works.

T. R. D[awson], proposer.  Problems 3230-1.  Fairy Chess Review 3:12 (Jun 1938) 129.  Extends prob. 3229 to ask for solutions with  12  pieces on the edge, using two other tetrominoes.  Thinks it cannot be done with the remaining two tetrominoes.

Editorial note:  The colossal count.  Fairy Chess Review 3:12 (Jun 1938) 131.  Describes progress on enumerating  8-ominoes (four people get  368  but Niemann gets  369)  and  9‑ominoes (numbers vary from  1237  to  1285).  All workers are classifying them by the size of the smallest containing rectangle.

W. H. Rawlings, proposer.  Problem 3930.  Fairy Chess Review 4:3 (Nov 1939) 28.  How many pentacubes are there?  Ibid. 4:4 (Feb 1940) 75,  reports that both  25  or  26  are claimed, but the editor has only seen  24.  Ibid. 4:5 (Apr 1940) 85, reports that R. J. F[rench] has clearly shown there are  23  -- but this considers reflections as equal -- cf the 1948 editorial note.

R. J. French, proposer and solver.  Problem 4149.  Fairy Chess Review 4:3 (Nov 1939) 43  &  4:6 (Jun 1940) 93.  Asks for arrangement of the pentominoes with the largest hole and gives one with  127  squares in the hole.  (See:  G. P. Jelliss; Comment on Problem 1277; JRM 22:1 (1990) 69.  This reviews various earlier solutions and comments on Problem 1277.)

J. Niemann.  Item 4154: "The colossal count".  Fairy Chess Review 4:3 (Nov 1939) 44-45.  Announces that there are  369  8-ominoes,  1285  9-ominoes and  4654  10-ominoes, but Keller and Jelliss note that he missed a  10‑omino which was not corrected until 1966.

H. D. Benjamin.  Unpublished notes.  ??NYS -- cited and briefly described in G. P. Jelliss; Prob. 48 -- Aztec tetrasticks; G&PJ 2 (No. 17) (Oct 1999) 320.  Jelliss says Benjamin studied polysticks, which he called 'lattice dissections' around 1946-1948 and that some results by him and T. R. Dawson were entered in W. Stead's notebooks but nothing is known to have been published.  For orders  1, 2, 3, 4,  there are  1, 2, 5, 16  polysticks.  Benjamin formed these into a  6 x 6  lattice square.  Jelliss then mentions Barwell's rediscovery of them and goes on to a new problem -- see Knuth, 1999.

D. Nixon, proposer and solver.  Problem 7560.  Fairy Chess Review 6:16 (Feb 1948) 12  &  6:17 (Apr 1948) 131.  Constructs  3 x 4 x 5  from solid pentominoes.

Editorial discussion: Space dissection.  Fairy Chess Review 6:18 (Jun 1948) 141-142.  Says that several people have verified the  23  pentacubes but that  6  of them have mirror images, making  29  if these are considered distinct.  Says F. Hansson has found  77  6‑cubes (these exclude mirror images and the  35  solid  6-ominoes).  Gives many problems using  n-cubes and/or solid polyominoes, which he calls flat  n-cubes -- some are corrected in 7:2 (Oct 1948) 16 (erroneously printed as 108).

J. Niemann.  The dissection count.  Item 7803.  Fairy Chess Review 7:1 (Aug 1948) 8 (erroneously printed as 100).  Reports on counting  n‑cubes.  Gets the following.

                  n =                             4                    5                    6                    7

          flat pieces                            5                  12                  35                 108

          non-flats                              2                  11                  77                 499

                    TOTAL                    7                  23                 112                 607

          mirror images                      1                    6                  55                 416

              GRAND TOTAL             8                  29                 167               1023

R. J. French.  Space dissections.  Fairy Chess Review 7:2 (Oct 1948) 16 (erroneously printed as 108).  French writes that he and A. W. Baillie have corrected the number of  6-cubes to  35 + 77 + 54 = 166.  Baillie notes that every  6-cube lies in two layers -- i.e. has some width  £ 2  -- and asks for the result for  n‑cubes as prob. 7879.  [I suspect the answer is that  n £ 3k  implies that an  n-cube has some width  £ k.]  Editor adds some corrections to the discussion in 6:18.

Editorial note.  Fairy Chess Review 7:3 (Dec 1948) 23.  Niemann and Hansson confirm the number  166  given in 7:2.

Daily Mail Crown Puzzle.  Made by Chad Valley Co.  1953.  26 pieces, coloured on one side, to be fit into a crown shape.  11 are border pieces and easily placed.  The other 15 are polyiamonds:  15: 00112 24012 11.  Prize of £100 for solution plus best slogan, entries due on 8 Jun 1953.

S. W. Golomb.  Checkerboards and polyominoes.  AMM 61 (1954) 675‑682.  Mostly concerned with covering the  8 x 8  board with copies of polyominoes.  Shows one covering with the  12  pentominoes and the square tetromino.  Mentions that the idea can be extended to hexagons.  S&B, p. 18, and Gardner (Dec 1964) say he mentions triangles, but he doesn't.

Walter S. Stead.  Dissection.  Fairy Chess Review 9:1 (Dec 1954) 2‑4.  Gives many pentomino and hexomino patterns -- e.g. one of each pattern of  8 x 8  with a  2 x 2  square deleted.  "The possibilities of the  12  fives are not infinite but they will provide years of amusement."  Includes  3 x 20,  4 x 15,  5 x 12  and  6 x 10  rectangles.  No reference to Golomb.  In 1955, Stead uses the  108  heptominoes to make a  28 x 28  square with a symmetric hole of size  28  in the centre -- first printed as cover of Chessics 28 (1986).

Jules Pestieau.  US Patent 2,900,190 -- Scientific Puzzle.  Filed: 2 Jul 1956;  patented: 18 Aug 1959.  2pp + 1p diagrams.  For the  12  pentominoes!  Diagram shows the  6 x 10  solution with two  5 x 6  rectangles and shows the two-piece non-symmetric equivalence of the  N  and  F  pieces.  Pieces have markings on one side which may be used -- i.e. pieces may not be turned over.  Mentions possibility of using  n-ominoes.

Gardner.  SA (Dec 1957) = 1st Book, chap. 13.  Exposits Golomb and Stead.  Gives number of  n-ominoes for  n = 1, ..., 7.  1st Book describes Scott's work.  Says a pentomino set called 'Hexed' was marketed in 1957.  (John Brillhart gave me and my housemates an example in 1960 -- it took us two weeks to find our first solution.)

Dana Scott.  Programming a Combinatorial Puzzle.  Technical Report No. 1, Dept. of Elec. Eng., Princeton Univ., 1958, 20pp.  Uses MANIAC to find  65  solutions for pentominoes on an  8 x 8  board with square  2 x 2  in the centre.  Notes that the  3 x 20  pentomino rectangle has just two solutions.  In 1999, Knuth notes that the total number of solutions with the  2 x 2  being anywhere does not seem to have ever been published and he finds  16146.

M. Gardner.  SA (Sep 1958) c= 2nd Book, chap. 6.  First general mention of solid pentominoes, pentacubes, tetracubes.  In the Addendum in 2nd Book, he says Theodore Katsanis of Seattle suggested the eight tetracubes and the  29  pentacubes in a letter to Gardner on 23 Sep 1957.  He also says that Julia Robinson and Charles W. Stephenson both suggested the solid pentominoes.

C. Dudley Langford.  Note 2793:  A conundrum for form VI.  MG 42 (No. 342) (Dec 1958) 287.  4  each of the  L,  N,  and  T (= Y)  tetrominoes make a  7 x 7  square with the centre missing.  Also nine pieces make a  6 x 6  square but this requires an even number of  Ts.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 8.  Use the pentominoes to make two  5 x 5  squares at the same time.  Solution just says there are several ways to do so.

J. C. P. Miller.  Pentominoes.  Eureka 23 (Oct 1960) 13‑16.  Gives the Haselgroves' number of  2339  solutions for the  6 x 10  and says there are  2  solutions for the  3 x 20.  Says Lehmer suggests assembling  12  solid pentominoes into a  3 x 4 x 5  and van der Poel suggests assembling the  12  hexiamonds into a rhombus.

C. B. & Jenifer Haselgrove.  A computer program for pentominoes.  Ibid., 16‑18.  Outlines program which found the  2339  solutions for the  6 x 10.  It is usually said that they also found all solutions of the  3 x 20,  4 x 15  and  5 x 12,  but I don't see it mentioned here and in JRM 7:3 (1974) 257, it is reported that Jenifer (Haselgrove) Leech stated that only the  6 x 10  and  3 x 20  were done  in 1960, but that she did the  5 x 12  and  4 x 15  with a new program in c1966.  See Fairbairn, c1962, and Meeus, 1973.

Richard K. Guy.  Some mathematical recreations I  &  II.  Nabla [= Bull. Malayan Math. Soc.] 7 (Oct  &  Dec 1960) 97-106  &  144-153.  Considers handed polyominoes, i.e. polyominoes when reflections are not considered equivalent.  Notes that neither the  5  plain nor the  7  handed tetrominoes can form a rectangle.  The  10  chequered handed tetrominoes form  4 x 10  and  5 x 8  rectangles and he has several solutions of each.  There is no  2 x 20  rectangle.  Discusses MacMahon pieces -- cf 5.H.2 -- and polyiamonds.  He uses the word 'hexiamond', but not 'polyiamond' -- in an email of 8 Apt 2000, Guy says that O'Beirne invented all the terms.  He considers making a 'hexagon' from the 19 hexiamonds.  Part II considers solid problems and uses the term 'solid pentominoes'.

Solomon W. Golomb.  The general theory of polyominoes: part 2 -- Patterns and polyominoes.  RMM 5 (Oct 1961) 3-14.  ??NYR.

J. E. Reeve & J. A. Tyrrell.  Maestro puzzles.  MG 45 (No. 353) (Oct 1961) 97‑99.  Discusses hexiamond puzzles, using the  12  reversible pieces.  [The puzzle was marketed under the name 'Maestro' in the UK.]

T. H. O'Beirne.  Pell's equation in two popular problems.  New Scientist 12 (No. 258) (26 Oct 1961) 260‑261.

T. H. O'Beirne.  Pentominoes and hexiamonds.  New Scientist 12 (No. 259) (2 Nov 1961) 316‑317.  This is the first use of the word 'polyiamond'.  He considers the  19  one‑sided pieces.  He says he devised the pieces and R. K. Guy has already published many solutions in Nabla.  He asks for the number of ways the 18 one-sided pentominoes can fill a  9 x 10.  In 1999, Knuth found this would take several months.

T. H. O'Beirne.  Some hexiamond solutions:  and an introduction to a set of  25  remarkable points.  New Scientist 12 (No. 260) (9 Nov 1961) 378‑379.

Maurice J. Povah.  Letter.  MG 45 (No. 354) (Dec 1961) 342.  States Scott's result of  65  and the Haselgroves' result of  2339  (computed at Manchester).  Says he has over  7000  solutions for the  8 x 8  board using a  2 x 2.

T. H. O'Beirne.  For boys, men and heroes.  New Scientist 12 (No. 266) (21 Dec 1961) 751‑752.

T. H. O'Beirne.  Some tetrabolic difficulties.  New Scientist 13 (No. 270) (18 Jan 1962) 158‑159.  These two columns are the first mention of tetraboloes, so named by S. J. Collins.

R. A. Fairbairn.  Pentomino Problems: The  6 x 10,  5 x 12,  4 x 15,  and  3 x 20  Rectangles -- The Complete Drawings.  Unpublished MS, undated, but c1962, based on the Haselgroves' work of 1960.  ??NYS -- cited by various authors, e.g.  Madachy (1969), Torbijn (1969), Meeus (1973).  Madachy says Fairbairn is from Willowdale, Ontario, and takes some examples from his drawings.  However, the dating is at variance with Jenifer Haselgrove's 1973 statement - cf Haselgrove, 1960.  Perhaps this MS is somewhat later??  Does anyone know where this MS is now?  Cf Meeus, 1973.

Serena Sutton Besley.  US Patent 3,065,970 -- Three Dimensional Puzzle.  Filed: 6 Jul 1960;  issued: 27 Nov 1962.  2pp + 4pp diagrams.  For the  29  pentacubes, with one piece duplicated giving a set of  30.  Klarner had already considered omitting the  1 x 1 x 5  and found that he could make two separate  2 x 5 x 7s.  Besley says the following can be made:  5 x 5 x 6,  3 x 5 x 10,  2 x 5 x 15,  2 x 3 x 25;  3 x 5 x 6,  3 x 3 x 10,  2 x 5 x 9,  2 x 3 x 15;  3 x 4 x 5,  2 x 5 x 6,  2 x 3 x 10  (where the latter three are made with the  12  solid pentominoes and the previous four are made with the  18  non-planar pentacubes) but detailed solutions are only given for the  5 x 5 x 6,  3 x 5 x 6,  3 x 4 x 5.  Mentions possibility of  n-cubes. 

M. Gardner.  Polyiamonds.  SA (Dec 1964) = 6th Book, chap. 18.  Exposits basic ideas and results for the 12 double sided hexiamonds.  Poses several problems which are answered by readers.  The six-pointed star using 8 pieces has a unique solution.  John G. Fletcher and Jenifer (Haselgrove) Leech both showed the  3 x 12  rhombus is impossible.  Fletcher found the  3 x 11  rhombus has  24  solutions, all omitting the 'bat'.  Leech found  155  solutions for the  6 x 6  rhombus and  74  solutions for the  4 x 9.  Mentions there are  160  9-iamonds, one with a hole. 

John G. Fletcher.  A program to solve the pentomino problem by the recursive use of macros.  Comm. ACM 8 (1965) 621-623.  ??NYS -- described by Knuth in 1999 who says that Fletcher found the 2339 solutions for the  6 x 10  in 10 minutes on an IBM 7094 and that the program remains the fastest known method for problems of placing the 12 pentominoes.

M. Gardner.  Op art.  SA (Jul 1965) = 6th Book, chap. 24.  Shows the 24 heptiamonds and discusses which will tile the plane.

Solomon W. Golomb.  Tiling with polyominoes.  J. Combinatorial Theory 1 (1966) 280-296.  ??NYS.  Extended by his 1970 paper.

T. R. Parkin.  1966.  ??NYS -- cited by Keller.  Finds  4655  10-ominoes.

M. Gardner.  SA (Jun 1967) = Magic Show, chap. 11.  First mention of polyhexes.

C. J. Bouwkamp.  Catalogue of Solutions of the Rectangular  3 x 4 x 5  Solid Pentomino Problem.  Dept. of Math., Technische Hogeschool Eindhoven, July 1967, reprinted 1981, 310pp.

C. J. Bouwkamp.  Packing a rectangular box with the twelve solid pentominoes.  J. Combinatorial Thy. 7 (1969) 278‑280.  He gives the numbers of solutions for rectangles as 'known'.

               2 x 3 x 10   can be packed in        12  ways, which are given.

               2 x 5 x   6   can be packed in      264  ways.

               3 x 4 x   5   can be packed in    3940  ways.  (See his 1967 report.)

T. R. Parkin,  L. J. Lander  &  D. R. Parkin.  Polyomino enumeration results.  Paper presented at the SIAM Fall Meeting, Santa Barbara, 1 Dec 1967.  ??NYS -- described by Madachy, 1969.  Gives numbers of  n-ominoes, with and without holes, up to  n = 15,  done two independent ways.

Joseph S. Madachy.  Pentominoes -- Some solved and unsolved problems.  JRM 2:3 (Jul 1969) 181-188.  Gives the numbers of Parkin, Lander & Parkin.  Shows various examples where a rectangle splits into two congruent halves.  Discusses various other problems, including Bouwkamp's  3 x 4 x 5  solid pentomino problem.  Bouwkamp reports that the final total of 3940 was completed on 16 Mar 1967 after about three years work using three different computers, but that a colleague's program would now do the whole search in about three hours.

P. J. Torbijn.  Polyiamonds.  JRM 2:4 (Oct 1969) 216-227.  Uses the double sided hexiamonds and heptiamonds.  A few years before, he found, by hand, that there are  156  ways to cover the  6 x 6  rhombus with the 12 hexiamonds and  74  ways for the  4 x 9,  but could find no way to cover the  3 x 12.  The previous year, John G. Fletcher confirmed these results with a computer and he displays all of these -- but this contradicts Gardner (Dec 64) -- ??  He gives several other problems and results, including using the 24 heptiamonds to form  7 x 12,  6 x 14,  4 x 21  and  3 x 28  rhombuses.

Solomon W. Golomb.  Tlling with sets of polyominoes.  J. Combinatorial Theory 9 (1970) 60‑71.  ??NYS.  Extends his 1966 paper.  Asks which heptominoes tile rectangles and says there are two undecided cases -- cf Marlow, 1985.  Gardner (Aug 75) says Golomb shows that the problem of determining whether a given finite set of polyominoes will tile the plane is undecidable.

C. J. Boukamp & D. A. Klarner.  Packing a box with  Y-pentacubes.  JRM 3:1 (1970) 10-26.  Substantial discussion of packings with  Y‑pentominoes and  Y-pentacubes.  Smallest boxes are  5 x 10  and  2 x 5 x 6  and  3 x 4 x 5.

Fred Lunnon.  Counting polyominoes.  IN:  Computers in Number Theory, ed. by A. O. L. Atkin & B. J. Birch; Academic Press, 1971, pp. 347-372.  He gets up through 18‑ominoes, but the larger ones can have included holes.  The numbers for  n = 1, 2, ...,  are as follows:  1, 1, 2, 5, 12,   35, 108, 369, 1285, 4655,   17073, 63600, 238591, 901971, 3426576,   13079255, 50107911, 192622052.  These values have been quoted numerous times.

Fred Lunnon.  Counting hexagonal and triangular polyominoes.  IN:  Graph Theory and Computing, ed. by R. C. Read; Academic Press, 1972, pp. 87-100.  ??NYS -- cited by Grünbaum & Shephard. 

M. Gardner.  SA (Sep 1972).  c= Knotted, chap. 3.  Says the  8  tetracubes were made by E. S. Lowe Co. in Hong Kong and marketed as "Wit's End".  Says an MIT group found  1390  solutions for the  2 x 4 x 4  box packed with tetracubes.  He reports that several people found that there are  1023  heptacubes -- but see Niemann, 1948, above.  Klarner reports that the heptacubes fill a  2 x 6 x 83.

Jean Meeus.  Some polyomino and polyamond [sic] problems.  JRM 6:3 (1973) 215-220.  (Corrections in 7:3 (1974) 257.)  Considers ways to pack a  5 x n  rectangle with some  n  pentominoes.  A. Mank found the number of ways for  n = 2, 3, ..., 11  as follows, and the number for  n = 12  was already known: 

                    0,   7,  50,   107,   541,   1387,   3377,   5865,   6814,   4103,   1010.

          Says he drew out all the solutions for the area 60 rectangles in 1972 (cf Fairbairn, c1962).  Finds that  520  of the  6 x 10  rectangles can be divided into two congruent halves, sometimes in two different ways.  For  5 x 12,  there are  380;  for  4 x 15,  there are 94.  Gives some hexomino rectangles by either deleting a piece or duplicating one, and an 'almost  11 x 19'.  Says there are  46  solutions to the  3 x 30  with the 18 one-sided pentominoes and attributes this to Mrs (Haselgrove) Leech, but the correction indicates this was found by A. Mank.

Jenifer Haselgrove.  Packing a square with Y-pentominoes.  JRM 7:3 (1974) 229.  She finds and shows a way to pack 45 Y-pentominoes into a  15 x 15,  but is unsure if there are more solutions.  In 1999, Knuth found  212  solutions.  She also reports the impossibility of using the Y-pentominoes to fill various other rectangles.

S. W. Golomb.  Trademark for 'PENTOMINOES'.  US trademark 1,008,964 issued 15 Apr 1975;  published 21 Jan 1975 as SN 435,448.  (First use:  November 1953.)  [These appear in the Official Gazette of the United States Patent Office (later Patent and Trademark Office) in the Trademarks section.]

M. Gardner.  Tiling with polyominoes, polyiamonds and polyhexes.  SA (Aug 75) (with slightly different title) = Time Travel, chap. 14.  Gives a tiling criterion of Conway.  Describes Golomb's 1966 & 1970 results. 

C. J. Bouwkamp.  Catalogue of solutions of the rectangular  2 x 5 x 6  solid pentomino problem.  Proc. Koninklijke Nederlandse Akad. van Wetenschappen A81:2 (1978) 177‑186.  Presents the  264  solutions which were first found in Sep 1967.

H. Redelmeier.  Discrete Math. 36 (1981) 191‑203.  ??NYS -- described by Jelliss.  Obtains number of n‑ominoes for  n £ 24.

Karl Scherer.  Problem 1045: Heptomino tessellations.  JRM 14:1 (1981‑82) 64.                              XX 

          Says he has found that the heptomino at the right fills a  26 x 42  rectangle.                     XXXXX

          See Dahlke below.

David Ellard.  Poly-iamond enumeration.  MG 66 (No. 438) (Dec 1982) 310‑314.  For  n = 1, ..., 12,  he gets  1, 1, 1, 3, 4,   12, 24, 66, 160, 448,   1186, 3342  n-iamonds.  One of the  8-iamonds has a hole and there are many later cases with holes.

Anon.  31: Polyominoes.  QARCH 1:8 (June 1984) 11‑13.  [This is an occasional publication of The Archimedeans, the student maths society at Cambridge.]  Good survey of counting and asymptotics for the numbers of polyominoes, up to  n = 24,  polycubes, etc.  10 references.

T. W. Marlow.  Grid dissections.  Chessics 23 (Autumn, 1985) 78‑79.

                               X                                          XX

          Shows  XXXXX  fills a  23 x 24  and   XXXXX   fills a  19 x 28.

Herman J. J. te Riele & D. T. Winter.  The tetrahexes puzzle.  CWI Newsletter [Amsterdam] 10 (Mar 1986) 33‑39.  Says there are:  7  tetrahexes,  22  pentahexes,  82  hexahexes,  333  heptahexes,  1448  octahexes.  Studies patterns of  28  hexagons.  Shows the triangle cannot be constructed from the  7  tetrahexes and gives  48  symmetric patterns that can be made.

Karl A. Dahlke.  Science News 132:20 (14 Nov 1987) 310.  (??NYS -- cited in JRM 21:3 and

                                                                         XX

          22:1 and by Marlow below.)  Shows  XXXXX  fills a  21 x 26  rectangle.

                    The results of Scherer and Dahlke are printed in JRM 21:3 (1989) 221‑223 and Dahlke's solution is given by Marlow below.

Karl A. Dahlke.  J. Combinatorial Theory A51 (1989) 127‑128.  ??NYS -- cited in JRM 22:1.  Announces a  19 x 28  solution for the above heptomino problem, but the earlier  21 x 26  solution is printed by error.  The  19 x 28  solution is printed in JRM 22:1 (1990) 68‑69.

Tom Marlow.  Grid dissections.  G&PJ 12 (Sep/Dec 1989) 185.  Prints Dahlke's result.

Brian R. Barwell.  Polysticks.  JRM 22:3 (1990) 165-175.  Polysticks are formed of unit lengths on the square lattice.  There are:  1, 2, 5, 16, 55  polysticks formed with  1, 2, 3, 4, 5  unit lengths.  He forms  5 x 5  squares with one  4-stick omitted, but he permits pieces to cross.  He doesn't consider the triangular or hexagonal cases.  See also Blyth, 1921, for a related puzzle.  Cf Benjamin, above, and Wiezorke & Haubrich, below.

General Symmetrics (Douglas Engel) produced a version of polysticks, ©1991, with  4  3‑sticks and  3  4-sticks to make a  3 x 3  square array with no crossing of pieces.

Nob Yoshigahara.  Puzzlart.  Tokyo, 1992.  Ip-pineapple (pineapple delight), pp. 78-81.  Imagine a cylindrical solution of the  6 x 10  pentomino rectangle and wrap it around a cylinder, giving each cell a depth of one along the radius.  Hence each cell is part of an annulus.  He reduces the dimensions along the short side to make the cells look like tenths of a slice of pineapple.  Nob constructed and example for Toyo Glass's puzzle series and it was later found to have a unique solution.

Kate Jones, proposer;  P. J. Torbijn, Jacques Haubrich, solvers.  Problem 1961 -- Rhombiominoes.  JRM 24:2 (1992) 144-146  &  25:3 (1993) 223‑225.  A rhombiomino or polyrhomb is a polyomino formed using rhombi instead of squares.  There are  20  pentarhombs.  Fit them into a  10 x 10  rhombus.  Various other questions.  Haubrich found many solutions.  See Lancaster, 1918.

Bernard Wiezorke & Jacques Haubrich.  Dr. Dragon's polycons.  CFF 33 (Feb 1994) 6-7.  Polycons (for connections) are the same as the polysticks described by Barwell in 1990, above.  Authors describe a Taiwanese version on sale in late 1993, using  10  of the  4‑sticks suitably shortened so they fit into the grooves of a  4 x 4  board -- so crossings are not permitted.  (An  n x n  board has  n+1  lines of  n  edges in each direction.)  They fit  15  of the  4-sticks onto a  5 x 5  board and determine all solutions.

                    CFF 35 (Dec 1994) 4 gives a number of responses to the article.  Brain Barwell wrote that he devised them as a student at Oxford, c1970, but did not publish until 1990.  He expected someone to say it had been done before, but no one has done so.  He also considered using the triangular and hexagonal lattice.  He had just completed a program to consider fitting  15  of the 4-sticks onto a  5 x 5  board and found over  180,000  solutions, with slightly under half having no crossings, confirming the results of Wiezorke & Haubrich.

                    Dario Uri also wrote that he had invented the idea in 1984 and called them polilati (polyedges).  Giovanni Ravesi wrote about them in Contromossa (Nov 1984) 23 -- a defunct magazine.

Chris Roothart.  Polylambdas.  CFF 34 (Oct 1994) 26-28.  A lambda is a  30o-60o-90o  triangle.  These may be joined along corresponding legs, but not along hypotenuses.  For  n = 1, 2, 3, 4, 5,  there are  1, 4, 4, 11, 12  n-lambdas.  He gives some problems using various sets of these pieces.

Richard Guy.  Letters of 29 May and 13 Jun 1996.  He is interested in using the  19  one-sided hexiamonds.  Hexagonal rings of hexagons contain  1, 6, 12  hexagons, so the hexagon with three hexagons on a side has 19 hexagons.  If these hexagons are considered to comprise six equilateral triangles, we have a board with  19 x 6  triangles.  O'Beirne asked for the number of ways to fill this board with the one-sided hexiamonds.  Guy has collected over  4200  solutions.  A program by Marc Paulhus found  907  solutions in eight hours, from which it initially estimated that there are about  30,000  solutions.  The second letter gives the final results -- there are  124,518  solutions.  This is modulo the  12  symmetries of the hexagon.  In 1999, Knuth found  124,519  and Paulhus has rerun his program and found this number.

Ferdinand Lammertink.  Polyshapes.  Parts 1 and 2.  The author, Hengelo, Netherlands, 1996 & 1997.  Part 1 deals with two dimensional puzzles.  Good survey of the standard polyform shapes and many others.

Hilarie Korman.  Pentominoes: A first player win.  IN: Games of No Chance; ed. by Richard Nowakowski; CUP, 1997??, ??NYS - described in  William Hartston; What mathematicians get up to; The Independent Long Weekend (29 Mar 1997) 2.  This studies the game proposed by Golomb -- players alternately place one of the pentominoes on the chess board, aligned with the squares and not overlapping the previous pieces, with the last one able to play being the winner.  She used a Sun IPC Sparcstation for five days, examining about  22 x 109  positions to show the game is a first player win.

Nob Yoshigahara found in 1994 that the smallest box which can be packed with W-pentacubes is  5 x 6 x 6.  In 1997, Yoshya (Wolf) Shindo found that one can pack the  6 x 10 x 10  with Z-pentacubes, but it is not known if this is the smallest such box.  These were the last unsolved problems as to whether a box could be packed with a planar pentacube (= solid pentomino).

Marcel Gillen  &  Georges Philippe.  Twinform  462 Puzzles in one.  Solutions for Gillen's puzzle exchange at IPP17, 1997, 32pp + covers.  Take 6 of the pentominoes and place them in a  7 x 5  rectangle, then place the other six to make the same shape on top of the first shape.  There are  462  (= BC(12,6)/2)  possible puzzles and all of them have solutions.  Taking  F, T, U, W, X, Z  for the first layer, there is just one solution; all other cases have multiple solutions, totalling  22,873  solutions, but only one solution for each case is given here.)

Richard K. Guy.  O'Beirne's hexiamond.  In:  The Mathemagican and Pied Puzzler; ed. by Elwyn Berlekamp & Tom Rodgers, A. K. Peters, Natick, Massachusetts, 1999, pp. 85‑96.  He relates that O'Beirne discovered the 19 one-sided hexiamonds in c1959 and found they would fill a hexagonal shape in Nov 1959 and in Jan 1960 he found a solution with the hexagonal piece in the centre.  He gives Paulhus's results (see Guy's letters of 1996), broken down in various ways.  He gives the number of double-sided (i.e. one can turn them over) and single-sided  n-iamonds for  n = 1, ..., 7.  Cf Ellard, 1982, for many more values for the double-sided case.

                         n             1   2   3   4   5     6     7

                    double 1   1   1   3   4   12   24

                    single            1   1   1   4   6   19   44

          In 1963, Conway and Mike Guy considered looking for 'symmetric' solutions for filling the hexagonal shape with the 19 one-sided hexiamonds.  A number of these are described.

Donald E. Knuth.  Dancing links.  25pp preprint of a talk given at Oxford in Sep 1999, sent by the author.  Available as:  http://www-cs-faculty.stanford.edu/~knuth/preprints.html .  In this he introduces a new technique for backtrack programming which runs faster (although it takes more storage) and is fairly easy to adapt to different problems.  In this approach, there is a symmetry between pieces and cells.  He applies it to several polyshape problems, obtaining new, or at least unknown, results.  He extends Scott's 1958 results to get  16146  ways to pack the  8 x 8  with the 12 pentominoes and the  2 x 2.  He describes Fletcher's 1965 work.  He extends Haselgrove's 1974 work and finds 212 ways to fit 15 Y-pentominoes in a  15 x 15.  Describes Torbijn's 1969 work and Paulhus' 1996 work on hexiamonds, correcting the latter's number to  124,519.  He then looks for the most symmetric solutions for filling the hexagonal shape with the 19 one-sided hexiamonds, in the sense discussed by Guy (1999).  He then considers the 18 one-sided pentominoes (cf Meeus (1973)) and tries the  9 x 10,  but finds it would take a few months on his computer (a 500 MHz Pentium III), so he's abandoned it for now.  He then considers polysticks, citing an actual puzzle version that I've not seen.  He adapts his program to them.  He considers the 'welded tetrasticks' which have internal junction points.  There are six of these and ten if they are taken as one-sided.  The ten can be placed in a  4 x 4 grid.  There are  15  unwelded, one-sided, tetrasticks, but they do not form a square, nor indeed any nice shape.  He considers all  25  one-sided tetrasticks and asks if they can be fit into what he calls an Aztec Diamond, which is the shape looking like a square tilted 45o on the square lattice.  The rows contain  1, 3, 5, 7, 9, 7, 5, 3, 1  cells.  He thinks an exhaustive search is beyond present computing power.

G. P. Jelliss.  Prob. 48 -- Aztec tetrasticks.  G&PJ 2 (No. 17) (Oct 1999) 320.  Jelliss first discusses Benjamin's work on polysticks (see at 1946-1948 above) and Barwell's rediscovery of them (see above).  He then describes Knuth's Dancing Links and gives the Aztec Diamond problem.  Jelliss has managed to get all but one of the polysticks into the shape, but feels it is impossible to get them all in.

Harry L. Nelson.  Solid pentomino storage, Question and answer.  1p HO at G4G5, 2002.  1: Can one put all the solid pentominoes into a cube of edge  4.5?  What is the smallest cube into which they can all be placed?  He gives 2 solutions to 1 and a solution due to Wei-Hwa Huang for a cube of edge 4.405889..., which is conjectured to be minimal.  In fact, one edge of the packing is actually 4, so the volume is less than  (4.405889...)3.  This leads me to ask what is the smallest volume of a cuboid, with edges less than 5, that contains all the solid pentominoes.  In Summer 2002, Harry gave me a set of solid pentominoes in a box with a list of various rectangles and boxes to fit them into:  3 x 22;  3 x 21;  3 x 20;  4 x 16;  4 x 15;  5 x 13;  5 x 12;  6 x 11;  6 x 10;  7 x 9;  8 x 8;   2 x 4 x 8;  2 x 5 x 7;  2 x 5 x 6;  2 x 6 x 6;  3 x 4 x 6;  3 x 4 x 5;  3 x 5 x 5;  4 x 4 x 5;  the given box: 4.4 x 4.4 x 4.9.)

 

          6.F.1.           OTHER CHESSBOARD DISSECTIONS

 

          See S&B, pp. 12‑14.  See also 6.F.5 for dissections of uncoloured boards.

 

Jerry Slocum.  Compendium of Checkerboard Puzzles.  Published by the author, 1983.  Outlines the history and shows all manufactured versions known then to him:  33 types in 61 versions.  The first number in Slocum's numbers is the number of pieces. 

Jerry Slocum  &  Jacques Haubrich.  Compendium of Checkerboard Puzzles.  2nd ed., published by Slocum, 1993.  90 types in 161 versions, with a table of which pieces are in which puzzles, making it much easier to see if a given puzzle is in the list or not.  This gives many more pictures of the puzzle boxes and also gives the number of solutions for each puzzle and sometimes prints all of them.  The Slocum numbers are revised in the 2nd ed. and I use the 2nd ed. numbers below.  (There was a 3rd ed. in 1997, with new numbering of 217 types in 376 versions.  NYR.  Haubrich is working on an extended version with Les Barton providing information.)

 

Henry Luers.  US Patent 231,963 -- Game Apparatus or Sectional Checker Board.  Applied: 7 Aug 1880;  patented: 7 Sep 1880.  1p + 1p diagrams.  15: 01329.  Slocum 15.5.1.  Manufactured as: Sectional Checker Board Puzzle, by Selchow & Righter.  Colour photo of the puzzle box cover is on the front cover of the 1st ed. of Slocum's booklet.  B&W photo is on p. 14 of S&B.

??  UK patent application 16,810.  1892.  Not granted, so never published.  I have spoken to the UK Patent Office and they say the paperwork for ungranted applications is destroyed after about three to five years.  (Edward Hordern's collection has an example with this number on it, by Feltham & Co.  In the 2nd ed., the cover is reproduced and it looks like the number may be 16,310, but that number is for a locomotive vehicle.)  14: 00149.  Slocum 14.20.1.  Manufactured as: The Chequers Puzzle, by Feltham & Co.

Hoffmann.  1893.  Chap. III, no. 16: The chequers puzzle, pp. 97‑98 & 129‑130 = Hoffmann‑Hordern, pp. 88-89, with photos.  14: 00149.  Slocum 14.20.1.  Says it is made by Messrs. Feltham, who state it has over 50 solutions.  He gives two solutions.  Photo on p. 89 of a example by Feltham & Co., dated 1880-1895. 

                    At the end of the solution, he says Jacques & Son are producing a series of three "Peel" puzzles, which have coloured squares which have to be arranged so the same colour is not repeated in any row or column.  Photo on p. 89 shows an example,  9: 023,  with the trominoes all being L-trominoes.  This makes a  5 x 5  square, but the colours have almost faded into indistinguishability. 

Montgomery Ward & Co.  Catalog No 57, Spring & Summer, 1895.  Facsimile by Dover, 1969, ??NX.  P. 237 describes item 25470:  "The "Wonder" Puzzle.  The object is to place 18 pieces of 81 squares together, so as to form a square, with the colors running alternately.  It can be done in several different ways." 

Dudeney.  Problem 517 -- Make a chessboard.  Weekly Dispatch (4  &  18 Oct 1903), both p. 10.  8: 00010 12111 001.  Slocum 8.3.1.

Benson.  1904.  The chequers puzzle, pp. 202‑203.  As in Hoffmann, with only one solution.

Dudeney.  The Tribune (20  &  24 Dec 1906) both p. 1.  ??NX  Dissecting a chessboard.  Dissect into maximum number of different pieces.  Gets 18: 2,1,4,10,0, 0,0,1.  Slocum 18.1, citing later(?) Loyd versions.

Loyd.  Sam Loyd's Puzzle Magazine (Apr-Jul 1908) -- ??NYS, reproduced in:  A. C. White; Sam Loyd and His Chess Problems; 1913, op. cit. in 1; no. 58, p. 52.  = Cyclopedia, 1914, pp. 221 & 368, 250 & 373.  = MPSL2, prob. 71, pp. 51 & 145.  = SLAHP: Dissecting the chessboard, pp. 19 & 87.  Cut into maximum number of different pieces -- as in Dudeney, 1906.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  The rug, pp. 7-13 & 65.  14: 00149.  Not in Slocum.

Loyd.  A battle royal.  Cyclopedia, 1914, pp. 97 & 351 (= MPSL1, prob. 51, pp. 49 & 139).  Same as Dudeney's prob. 517 of 1903.

Dudeney.  AM. 1917.  Prob. 293: The Chinese chessboard, pp. 87 & 213‑214.  Same as Loyd, p. 221.

Western Puzzle Works, 1926 Catalogue.  No. 79: "Checker Board Puzzle, in 16 pieces", but the picture only shows 14 pieces.  14: 00149.  Picture doesn't show any colours, but assuming the standard colouring of a chess board, this is the same as Slocum 14.15.

John Edward Fransen.  US Patent 1,752,248 -- Educational Puzzle.  Applied: 19 Apr 1929;  patented: 25 Mar 1930.  1p + 1p diagrams.  'Cut thy life.'  11: 10101 43001.  Slocum 11.3.1.

Emil Huber-Stockar.  Patience de l'echiquier.  Comptes-Rendus du Premier Congrès International de Récréation Mathématique, Bruxelles, 1935.  Sphinx, Bruxelles, 1935, pp. 93-94.  15: 01329.  Slocum 15.5.  Says there must certainly be more than 1000 solutions.

Emil Huber-Stockar.  L'echiquier du diable.  Comptes-Rendus du Deuxième Congrès International de Récréation Mathématique, Paris, 1937.  Librairie du "Sphinx", Bruxelles, 1937, pp. 64-68.  Discusses how one solution can lead to many others by partial symmetries.  Shows several solutions containing about 40 altogether.  Note at end says he has now got  5275  solutions.  This article is reproduced in Sphinx 8 (1938) 36-41, but without the extra pages of diagrams.  At the end, a note says he has  5330  solutions.  Ibid., pp. 75-76 says he has got 5362 solutions and ibid. 91-92 says he has  5365.  By use of Bayes' theorem on the frequency of new solutions, he estimates  c5500  solutions.  Haubrich has found 6013.  Huber-Stocker intended to produce a book of solutions, but he died in May 1939 [Sphinx 9 (1939) 97].

F. Hansson.  Sam Loyd's 18-piece dissection -- Art. 48 & probs. 4152‑4153.  Fairy Chess Review 4:3 (Nov 1939) 44.  Cites Loyd's Puzzles Magazine.  Asserts there are many millions of solutions!  He determines the number of chequered handed  n-ominoes for  n = 1, 2, ..., 8  is  2, 1, 4, 10, 36, 110, 392, 1371.  The first 17 pieces total 56 squares.  Considers 8 ways to dissect the board into 18 different pieces.  Problems ask for the number of ways to choose the pieces in each of these ways and for symmetrical solutions.  Solution in 4:6 (Jun 1940) 93-94 (??NX of p. 94) says there are a total of  3,309,579  ways to make the choices.

C. Dudley Langford.  Note 2864:  A chess‑board puzzle.  MG 43 (No. 345) (Oct 1959) 200.  15: 01248.  Not in Slocum.  Two diagrams followed by the following text.  "The pieces shown in the diagrams can be arranged to form a square with either side uppermost.  If the squares of the underlying grid are coloured black and white alternately, with each white square on the back of a black square, then there is at least one more way of arranging them as a chess-board by turning some of the pieces over."  I thought this meant that the pieces were double-sided with the underside having the colours being the reverse of the top and the two diagrams were two solutions for this set of pieces.  Jacques Haubrich has noted that the text is confusing and that the second diagram is NOT using the set of double-sided pieces which are implied by the first diagram.  We are not sure if the phrasing is saying there are two different sets of pieces and hence two problems or if we are misinterpreting the description of the colouring.

B. D. Josephson.  EDSAC to the rescue.  Eureka 24 (Oct 1961) 10‑12 & 32.  Uses the EDSAC computer to find two solutions of a 12 piece chessboard dissection.  12: 00025 41.  Slocum 12.9.

Leonard J. Gordon.  Broken chessboards with unique solutions.  G&PJ 10 (1989) 152‑153.  Shows Dudeney's problem has four solutions.  Finds other colourings which give only one solution.  Notes some equivalences in Slocum.

 

          6.F.2.           COVERING DELETED CHESSBOARD WITH DOMINOES

 

          See also 6.U.2.

          There is nothing on this in Murray. 

 

Pál Révész.  Op. cit. in 5.I.1.  1969.  On p. 22, he says this problem comes from John [von] Neumann, but gives no details.

Max Black.  Critical Thinking, op. cit. in 5.T.  1946 ed., pp. 142 & 394, ??NYS.  2nd ed., 1952, pp. 157 & 433.  He simply gives it as a problem, with no indication that he invented it.

H. D. Grossman.  Fun with lattice points: 14 -- A chessboard puzzle.  SM 14 (1948) 160.  (The problem is described with 'his clever solution' from M. Black, Critical Thinking, pp. 142 & 394.)

S. Golomb.  1954.  Op. cit. in 6.F.

M. Gardner.  The mutilated chessboard.  SA (Feb 1957) = 1st Book, pp. 24 & 28.

Gamow & Stern.  1958.  Domino game.  Pp. 87‑90.

Robert S. Raven, proposer;  Walter P. Targoff, solver.  Problem 85 -- Deleted checkerboard.  In:  L. A. Graham; Ingenious Mathematical Problems and Methods; Dover, 1959, pp. 52 & 227.

R. E. Gomory.  (Solution for deletion of any two squares of opposite colour.)  In:  M. Gardner, SA (Nov 1962) = Unexpected, pp. 186‑187.  Solution based on a rook's tour.  (I don't know if this was ever published elsewhere.)

Michael Holt.  What is the New Maths?  Anthony Blond, London, 1967.  Pp. 68 & 97.  Gives the  4 x 4  case as a problem, but doesn't mention that it works on other boards.  (I include this as I haven't seen earlier examples in the educational literature.)

David Singmaster.  Covering deleted chessboards with dominoes.  MM 48 (1975) 59‑66.  Optimum extension to  n‑dimensions.  For an  n-dimensional board, each dimension must be  ³ 2.  If the board has an even number of cells, then one can delete any  n-1  white cells and any  n-1  black cells and still cover the board with dominoes (i.e.  2 x 1 x 1 x ... x 1  blocks).  If the board has an odd number of cells, then let the corner cells be coloured black.  One can then delete any  n  black cells and any  n-1  white cells and still cover the board with dominoes.

I-Ping Chu & Richard Johnsonbaugh.  Tiling deficient boards with trominoes.  MM 59:1 (1986) 34-40.  (3,n) = 1  and  n ¹ 5  imply that an  n x n  board with one cell deleted can be covered with  L  trominoes.  Some  5 x 5  boards with one cell deleted can be tiled, but not all can.

 

          6.F.3.           DISSECTING A CROSS INTO  Zs  AND  Ls

 

          The  L  pieces are not always drawn carefully, and in some cases the unit pieces are not all square.  I have enlarged and measured those which are not clear and approximated them as  n-ominoes.

 

Minguet.  1733.  Pp. 119-121 (1755: 85-86; 1822: 138-139; 1864: 116-117).  The problem has two parts.  The first is a cross into 5 pieces: L-tetromino, 2 Z-pentominoes, L‑hexomino, Z-hexomino.  The two hexominoes are like the corresponding pentominoes lengthened by one unit.  Similar to Les Amusemens, but one  Z  is longer and one  L  is shorter.  The diagram shows  8  L  and  Z  shaped pieces formed from squares, but it is not clear what the second part of the problem is doing -- either a piece or a label is erroneous or missing.  Says one can make different figures with the pieces.

Les Amusemens.  1749.  P. xxxi.  Cross into 3  Z  pentominoes and 2  L  pieces.  Like Minguet, but the  Ls  are much lengthened and are approximately a L-heptomino and an L‑octomino.

Catel.  Kunst-Cabinet.  1790.  Das mathematische Kreuz, p. 10 & fig. 27 on plate I.  As in Les Amusemens, but the Ls are approximately a 9-omino and a 10-omino.

Bestelmeier.  1801.  Item 274 -- Das mathematische Kreuz.  Cross into 6 pieces, but the picture has an erroneous extra line.  It should be the reversal of the picture in Catel.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has the dissection of the cross into 3  Z  pentominoes and two  L  pieces.  I don't have a copy of this, but my sketch looks like the Ls are a tetromino and a pentomino, or possibly a pentomino and a hexomino.

Manuel des Sorciers.  1825.  Pp. 204-205, art. 21.  ??NX.  Dissect a cross into three  Zs  and two  Ls.  My notes don't indicate the size of the Ls.

Boy's Own Book.  1843 (Paris): 435 & 440, no. 3.  As in Les Amusemens, but with the Ls apparently intended to be a pentomino and a hexomino.  = Boy's Treasury, 1844, pp. 424-425 & 428.  = de Savigny, 1846, pp. 353 & 357, no. 2, except the solution has been redrawn with some slight changes and so the proportions are less clear.

Family Friend 3 (1850) 330 & 351.  Practical puzzle, No. XXI.  As in Les Amusemens.

Magician's Own Book.  1857.  Prob. 31: Another cross puzzle, pp. 276 & 299.  As in Les Amusemens.

Landells.  Boy's Own Toy-Maker.  1858.  P. 152.  As in Les Amusemens.

Book of 500 Puzzles.  1859.  Prob. 31: Another cross puzzle, pp. 90 & 113.  As in Les Amusemens.  = Magician's Own Book.

Indoor & Outdoor.  c1859.  Part II, p. 127, prob. 5: The puzzle of the cross.  As in Les Amusemens.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 24, pp. 399 & 439.  Identical to Magician's Own Book.

Boy's Own Conjuring book.  1860.  Prob. 30: Another cross puzzle, pp. 239 & 263.  = Magician's Own Book, 1857.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 584-2, pp. 286 & 404.  4  Z  pentominoes to make a (Greek) cross.  (Also entered in 6.F.5.)

Prob. 584-8, pp. 287 & 405.  3  Z  pentominoes,  L  tetromino and  L  pentomino to make a Greek cross.  Despite specifically asking for a Greek cross, the answer is a standard Latin cross with  height : width = 4 : 3.

Mittenzwey.  1880.  Prob. 173-174, pp. 33 & 85;  1895?: 198-199, pp. 38 & 87;  1917: 198‑199, pp. 35 & 84.  The first is 3  Z  pentominoes,  L  tetromino and  L  pentomino to make a cross.  The second is 4  Z  pentominoes to make a (Greek) cross.  (Also entered in 6.F.5.)

Cassell's.  1881.  P. 93: The magic cross.  = Manson, 1911, p. 139.  Same pattern as Les Amusemens, but one end of the Zs is decidedly longer than the other and the middle 'square' of the Zs is decidedly not square.  The Ls are approximately a pentomino and a heptomino,  But the middle 'square' of the Zs is almost a domino and that makes the Zs into heptominoes, with the Ls being a hexomino and a nonomino.

S&B, p. 20, shows a 7 piece cross dissection, Jeu de La Croix, into 3  Zs,  2  Ls  and 2 straights, from c1890.  The Zs are pentominoes, with the centre 'square' lengthened a bit.  The Ls appear to be a heptomino and an octomino and the straights appear to be a hexomino and a tetromino.  Cf Hoffmann-Hordern for a version without the straight pieces.

Handy Book for Boys and Girls.  Showing How to Build and Construct All Kinds of Useful Things of Life.  Worthington, NY, 1892.  Pp. 320-321: The cross puzzle.  As in Cassell's.

Hoffmann.  1893.  Chap. III, no. 29: Another cross puzzle, pp. 103 & 136 = Hoffmann‑Hordern, pp. 100-101, with photo.  States that the two Ls are the same shape, but the solution is as in Les Amusemens, with the Ls approximately a hexomino and a heptomino.  Hordern has corrected the problem statement.  Photo on p. 100 shows an ivory version, dated 1850-1900, of the same proportions.  Hordern Collection, p. 65, shows two wood versions, La Croix Brisée and Jeu de la Croix, dated 1880-1905, both with Ls being approximately a heptomino and an octomino.

Benson.  1904.  The Latin cross puzzle, p. 200.  As in Hoffmann, but the solution is longer, as in Les Amusemens.

Wehman.  New Book of 200 Puzzles.  1908.  Another cross puzzle, p. 32.  As in Les Amusemens, with the Ls being a pentomino and a hexomino.

S. Szabo.  US Patent 1,263,960 -- Puzzle.  Filed: 20 Oct 1917;  patented: 23 Apr 1918.  1p + 1p diagrams.  As in Les Amusemens, with even longer  Ls, approximately a 10-omino and an 11-omino.

 

          6.F.4.           QUADRISECT AN  L‑TROMINO, ETC.

 

          See also 6.AW.1 & 4.

          Mittenzwey and Collins quadrisect a hollow square obtained by removing a  2 x 2  from the centre of a  4 x 4.

          Bile Beans quadrisects a  5 x 5  after deleting corners and centre.

 

Minguet.  1733.  Pp. 114-115 (1755: 80; 1822: 133-134; 1864: 111-112).  Quadrisect  L‑tromino.

Alberti.  1747.  Art. 30: Modo di dividere uno squadro di carta e di legno in quattro squadri equali, p. ?? (131) & fig. 56, plate XVI, opp. p. 130.

Les Amusemens.  1749.  P. xxx.  L-tromino ("gnomon") into 4 congruent pieces.

Vyse.  Tutor's Guide.  1771?  Prob. 9, 1793: p. 305, 1799: p. 317 & Key p. 358.  Refers to the land as a parallelogram though it is drawn rectangular.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has an entry  "8½ a  Prob of figure"  followed by the  L‑tromino.  8½ b is the same with a mitre and there are other dissection problems adjacent -- see 6.F.3, 6.AQ, 6.AW.1, 6.AY, so it seems clear that he knew this problem.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles, no. 3, pp. 23 & 83 & plate I, fig. 2.

Manuel des Sorciers.  1825.  Pp. 203-204, art. 20.  ??NX.  Quadrisect L-tromino.

Family Friend 2 (1850) 118 & 149.  Practical Puzzle -- No. IV.  Quadrisect L-tromino of land with four trees.

Family Friend 3 (1850) 150 & 181.  Practical puzzle, No. XV.  15/16  of a square with 10 trees to be divided equally.  One tree is placed very close to another, cf Magician's Own Book and Hoffmann, below.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 8, p. 179 (1868: 190).  Land in the shape of an  L-tromino to be cut into four congruent parts, each with a cherry tree.

Magician's Own Book.  1857.

Prob. 3: The divided garden, pp. 267 & 292.  15/16  of a square to be divided into five (congruent) parts, each with two trees.  The missing  1/16  is in the middle.  One tree is placed very close to another, cf Family Friend 3, above, and Hoffmann below.

Prob. 22: Puzzle of the four tenants, pp. 273 & 296.  Same as Parlour Pastime, but with apple trees.  (= Illustrated Boy's Own Treasury, 1860, No. 10, pp. 397 & 437.) 

Prob. 28: Puzzle of the two fathers, pp. 275-276 & 298.  Each father wants to divide  3/4  of a square.  One has  L‑tromino, other has the mitre shape.  See 6.AW.1.

Landells.  Boy's Own Toy-Maker.  1858. 

P. 144.  = Magician's Own Book, prob. 3.

Pp. 148-149.  = Magician's Own Book, prob. 27.

Book of 500 Puzzles.  1859.

Prob. 3: The divided garden, pp. 81 & 106.  Identical to Magician's Own Book.

Prob. 22: Puzzle of the four tenants, pp. 87 & 110.  Identical to Magician's Own Book.

Prob. 28: Puzzle of the two fathers, pp. 89-90 & 112.  Identical to Magician's Own Book.  See also 6.AW.1.

Charades, Enigmas, and Riddles.  1860: prob. 28, pp. 59 & 63;  1862: prob. 29, pp. 135 & 141;  1865: prob. 573, pp. 107 & 154.  Quadrisect  L-tromino, attributed to Sir F. Thesiger.

Boy's Own Conjuring book.  1860.

Prob. 3: The divided garden, pp. 229 & 255.  Identical to Magician's Own Book.

Prob. 21: Puzzle of the four tenants, pp. 235 & 260.  Identical to Magician's Own Book.

Prob. 27: Puzzle of the two fathers, pp. 237‑238 & 262.  Identical to Magician's Own Book.

Illustrated Boy's Own Treasury.  1860.  Prob. 21, pp. 399 & 439.  15/16  of a square to be divided into five (congruent) parts, each with two trees.  c= Magician's Own Book, prob. 3.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 175, p. 88.  L-tromino into four congruent pieces, each with two trees.  The problem is given in terms of the original square to be divided into five parts, where the father gets a quarter of the whole in the form of a square and the four sons get congruent pieces.

Hanky Panky.  1872.  The divided orchards, p. 130.  L‑tromino into 4 congruent pieces, each with two trees.

Boy's Own Book.  The divided garden.  1868: 675.  = Magician's Own Book, prob. 3.

Mittenzwey.  1880. 

Prob. 192, pp. 36 & 89;  1895?: 217, pp. 40 & 91;  1917: 217, pp. 37 & 87.  Cut  1 x 1  out of the centre of a  4 x 4.  Divide the rest into five parts of equal area with four being congruent.  He cuts a  2 x 2  out of the centre, which has a  1 x 1  hole in it, then divides the rest into four L-trominoes.

Prob. 213, pp. 38 & 90;  1895?: 238, pp. 42 & 92;  1917: 238, pp. 39 & 88.  Usual quadrisection of an L-tromino.

Prob. 214, pp. 38 & 90;  1895?: 239, pp. 42 & 92;  1917: 239, pp. 39 & 88.  Square garden with mother receiving 1/4 and the rest being divided into four congruent parts.

Cassell's.  1881.  P. 90: The divided farm.  = Manson, 1911, pp. 136-137.  = Magician's Own Book, prob. 3.

Lemon.  1890.

The divided garden, no. 259, pp. 38 & 107.  = Magician's Own Book, prob. 3.

Geometrical puzzle, no. 413, pp. 55 & 113 (= Sphinx, no. 556, pp. 76 & 116).  Quadrisect  L-tromino.

Hoffmann.  1893.  Chap. X, no. 41: The divided farm, pp. 352‑353 & 391 = Hoffmann‑Hordern, p. 250.  = Magician's Own Book, prob. 3.  [One of the trees is invisible in the original problem, but Hoffmann-Hordern has added it, in a more symmetric pattern than in Magician's Own Book.]

Loyd.  Origin of a famous puzzle -- No. 18: An ancient puzzle.  Tit‑Bits 31 (13 Feb  &  6 Mar 1897) 363  &  419.  Nearly 50 years ago he was told of the quadrisection of  3/4  of a square, but drew the mitre shape instead of the  L‑tromino.  See 6.AW.1.

Clark.  Mental Nuts.  1897, no. 73;  1904, no. 31.  Dividing the land.  Quadrisect an  L‑tromino.  1904 also has the mitre -- see 6.AW.1.

Benson.  1904.  The farmer's puzzle, p. 196.  Quadrisect an  L‑tromino.

Wehman.  New Book of 200 Puzzles.  1908.

The divided garden, p. 17.  = Magician's Own Book, prob. 3

Puzzle of the two fathers, p. 43.  = Magician's Own Book, prob. 28.

Puzzle of the four tenants, p. 46.  = Magician's Own Book, prob. 22.

Dudeney.  Some much‑discussed puzzles.  Op. cit. in 2.  1908.  Land in shape of an  L‑tromino to be quadrisected.  He says this is supposed to have been invented by Lord Chelmsford (Sir F. Thesiger), who died in 1878 -- see Charades, Enigmas, and Riddles (1860).  But cf Les Amusemens.

M. Adams.  Indoor Games.  1912.  The clever farmer, pp. 23‑25.  Dissect  L‑tromino into four congruent pieces.

Blyth.  Match-Stick Magic.  1921.  Dividing the inheritance, pp. 20-21.  Usual quadrisection of  L-tromino set out with matchsticks.

Collins.  Book of Puzzles.  1927.  The surveyor's puzzle, pp. 2-3.  Quadrisect  3/4  of a square, except the deleted  1/4  is in the centre, so we are quadrisecting a hollow square -- cf Mittenzwey,

The Bile Beans Puzzle Book.  1933. 

No. 22: Paper squares.  Quadrisect a P-pentomino into P-pentominoes.  One solution given, I find another.  Are there more?  How about quadrisecting into congruent pentominoes?  Which pentominoes can be quadrisected into four copies of themself?

No. 41: Five lines.  Consider a  5 x 5  square and delete the corners and centre.  Quadrisect into congruent pentominoes.  One solution given.  I find three more.  Are there more?  One can extend this to consider quadrisecting the  5 x 5  with just the centre removed into congruent hexominoes.  I find seven ways.

Depew.  Cokesbury Game Book.  1939.  A plot of ground, p. 227.  3/4 of                                   XX     

          a square to be quadrisected, but the shape is as shown at the right.                                     XXX  

                                                                                                                                             X   XX

                                                                                                                                              XXXX

Ripley's Puzzles and Games.  1966.  Pp. 18 & 19, item 8.  Divide an L-tromino into eight congruent pieces.

F. Göbel.  Problem 1771: The L‑shape dissection problem.  JRM 22:1 (1990) 64‑65.  The  L‑tromino can be dissected into  2, 3, or 4  congruent parts.  Can it be divided into 5 congruent parts?

Rowan Barnes-Murphy.  Monstrous Mysteries.  Piccolo, 1982.  Apple-eating monsters, pp. 40 & 63.  Trisect into equal parts, the shape consisting of a  2 x 4  rectangle with a  1 x 1  square attached to one of the central squares of the long side.  [Actually, this can be done with the square attached to any of the squares, though if it is attached to the end of the long side, the resulting pieces are straight trominoes.]

 

          6.F.5.           OTHER DISSECTIONS INTO POLYOMINOES

 

Catel.  Kunst-Cabinet.  1790.

Das Zakk- und Hakenspiel, p. 10 & fig. 11 on plate 1.  4  Z‑pentominoes and 4  L‑tetrominoes make a  6 x 6  square.

Die zwolf Winkelhaken, p. 11 & fig. 26 on plate 1.  8  L‑pentominoes and 4  L‑hexominoes make a  8 x 8  square.

Bestelmeier.  1801.  Item 61 -- Das Zakken und Hakkenspiel.  As in Catel, p. 10, but not as regularly drawn.  Text copies some of Catel.

Manuel des Sorciers.  1825.  Pp. 203-204, art. 20.  ??NX  Use four L-trominoes to make a  3 x 4  rectangle or a  4 x 4  square with four corners deleted.

Family Friend 3 (1850) 90 & 121.  Practical puzzle -- No. XIII.  4 x 4  square, with 12 trees in the corners, centres of sides and four at the centre of the square, to be divided into 4 congruent parts each with 3 trees.  Solution uses  4  L-tetrominoes.  The same problem is repeated as Puzzle 17 -- Twelve-hole puzzle in (1855) 339 with solution in (1856) 28.

Magician's Own Book.  1857.  Prob. 14: The square and circle puzzle, pp. 270 & 295.  Same as Family Friend.  = Book of 500 Puzzles, 1859, prob. 14, pp. 84 & 109.  = Boy's Own Conjuring book, 1860, prob. 13, pp. 231-232 & 257.  c= Illustrated Boy's Own Treasury, 1860, prob. 8, pp. 396 & 437.  c= Hanky Panky, 1872, A square of four pieces, p. 117.

Landells.  Boy's Own Toy-Maker.  1858.  Pp. 146-147.  Identical to Family Friend.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 584-2, pp. 286 & 404.  4  Z‑pentominoes to make a Greek cross.  (Also entered in 6.F.3.)

Prob. 584-3, pp. 286 & 404.  4  L-tetrominoes to make a square.

Prob. 584-5, pp. 286 & 404.  8  L‑pentominoes and 4  L‑hexominoes make a  8 x 8  square.  Same as Catel, but diagram is inverted.

Prob. 584-7, pp. 287 & 405.  4  Z‑pentominoes and 4  L‑tetrominoes make a  6 x 6  square.  Same as Catel, but diagram is inverted.

Mittenzwey.  1880. 

Prob. 174, pp. 33 & 85;  1895?: 199, pp. 38 & 87;  1917: 199, pp. 35 & 84.  4  Z  pentominoes to make a (Greek) cross.  (Also entered in 6.F.5.)

Prob. 186, pp. 35 & 88;  1895?: 211, pp. 40 & 90;  1917: 211, pp. 36 & 87.  4 x 4  square into 4  L-tetrominoes. 

Prob. 187, pp. 35 & 88;  1895?: 212, pp. 40 & 90;  1917: 212, pp. 36 & 87.  6 x 6  square into 4  Z‑pentominoes and 4  L‑tetrominoes, as in Catel, p. 10.

Prob. 215, pp. 38 & 90;  1895?: 240, pp. 42 & 92;  1917: 240, pp. 39 & 88.  Square garden with 12 trees quadrisected into four L-tetrominoes. 

S&B, p. 20, shows a 7 piece cross dissection into  3  Zs,  2  Ls  and  2  straights, from c1890.

Hoffmann.  1893.  Chap. X, no. 37: The orchard puzzle, pp. 350 & 390  = Hoffmann-Hordern, pp. 247, with photo.  Same as Family Friend 3.  Photo on p. 247 shows St. Nicholas Puzzle Card, © 1892 in the USA.

Tom Tit, vol 3.  1893.  Les quatre  Z  et des quatre  L,  pp. 181-182.  = K, No. 27: The four  Z's  and the four  L's,  pp. 70‑71.  = R&A, Squaring the  L's  and  Z's,  p. 102.  6 x 6  square as in Catel, p. 10.

Sphinx.  1895.  The Maltese cross, no. 181, pp. 28 & 103.  Make a Maltese cross (actually a Greek cross of five equal squares) from 4  P-pentominoes.  Also:  quadrisect a  P‑pentomino.

Wehman.  New Book of 200 Puzzles.  1908.  The square and circle puzzle, p. 5.  = Family Friend.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  The Zoltan's orchard, pp. 24-28 & 64.  = Family Friend.

Anon.  Prob. 84.  Hobbies 31 (No. 799) (4 Feb 1911) 443.  Use at least one each of:  domino;  L-tetromino;  P and X pentominoes  to make the smallest possible square   Due to ending of this puzzle series, no solution ever appeared.  I find numerous solutions for  5 x 5,  6 x 6,  8 x 8,  of which the first is easily seen to be the smallest possibility.

A. Neely Hall.  Carpentry & Mechanics for Boys.  Lothrop, Lee & Shepard, Boston, nd [1918].  The square puzzle, pp. 20‑21.  7 x 7  square cut into 1 straight tromino, 1  L‑tetromino and 7  L‑hexominoes.

Collins.  Book of Puzzles.  1927.  The surveyor's puzzle, pp. 2-3.  Quadrisect  3/4  of a square, except the deleted  1/4  is in the centre, so we are quadrisecting a hollow square.

Arthur Mee's Children's Encyclopedia 'Wonder Box'.  The Children's Encyclopedia appeared in 1908, but versions continued until the 1950s.  This looks like 1930s??  4  Z‑pentominoes and 4  L‑tetrominoes make a  6 x 6  square and a  4 x 9  rectangle.

W. Leslie Prout.  Think Again.  Frederick Warne & Co., London, 1958.  All square, pp. 42 & 129.  Make a  6 x 6  square from the staircase hexomino, 2  Y-pentominoes, an  N‑tetromino, an  L-tetromino  and 3  T-tetrominoes.  None of the pieces is turned over in the solution, though this restriction is not stated.

 

          6.G.   SOMA CUBE

 

Piet Hein invented the Soma Cube in 1936.  (S&B, pp. 40‑41.)  ??Is there any patent??

M. Gardner.  SA (Sep 1958) = 2nd Book, Chap 6.

Richard K. Guy.  Loc. cit. in 5.H.2, 1960.  Pp. 150-151 discusses cubical solutions --  234  found so far.  He proposes the 'bath' shape -- a  5 x 3 x 2  cuboid with a  3 x 1 x 1  hole in the top layer.  In a 1985 letter, he said that O'Beirne had introduced the Soma to him and his family. in 1959 and they found 234 solutions before Mike Guy went to Cambridge -- see below.

P. Hein, et al.  Soma booklet.  Parker Bros., 1969, 56pp.  Asserts there are  240  simple solutions and  1,105,920  total solutions, found by J. H. Conway & M. J. T. Guy with a a computer (but cf Gardner, below) and by several others.  [There seem to be several versions of this booklet, of various sizes.]

Thomas V. Atwater, ed.  Soma Addict.  4 issues, 1970‑1971, produced by Parker Brothers.  (Gardner, below, says only three issues appeared.)  ??NYS -- can anyone provide a set or photocopies??

M. Gardner.  SA (Sep 1972) c= Knotted, chap. 3.  States there are  240  solutions for the cube, obtained by many programs, but first found by J. H. Conway & M. J. T. Guy in 1962, who did not use a computer, but did it by hand "one wet afternoon".  Richard Guy's 1985 letter notes that Mike Guy had a copy of the Guy family's 234 solutions with him.

SOMAP  ??NYS -- ??details.  (Schaaf III 52)

Winning Ways, 1982, II, 802‑803 gives the SOMAP.

Jon Brunvall et al.  The computerized Soma Cube.  Comp. & Maths. with Appl. 12B:1/2 (1986  [Special issues  1/2  &  3/4  were separately printed as:  I. Hargittai, ed.; Symmetry -- Unifying Human Understanding; Pergamon, 1986.] 113‑121.  They cite Gardner's 2nd Book which says the number of solutions is unknown and they use a computer to find them.

 

          6.G.1.          OTHER CUBE DISSECTIONS

 

          See also 6.N, 6.U.2, 6.AY.1 and 6.BJ.  The predecessors of these puzzles seem to be the binomial and trinomial cubes showing  (a+b)3  and  (a+b+c)3.  I have an example of the latter from the late 19C.  Here I will consider only cuts parallel to the cube faces -- cubes with cuts at angles to the faces are in 6.BJ.  Most of the problems here involve several types of piece -- see 6.U.2 for packing with one kind of piece.

 

Catel.  Kunst-Cabinet.  1790.  Der algebraische Würfel, p. 6 & fig 50 on plate II.  Shows a binomial cube:  (a + b)3  =  a3 + 3a2b + 3ab2 + b3.

Bestelmeier.  1801.  Item 309 is a binomial cube, as in Catel.  "Ein zerschnittener Würfel, mit welchem die Entstehung eines Cubus, dessen Seiten in  2  ungleiche Theile  a + b  getheilet ist, gezeigt ist."

Hoffmann.  1893.  Chap. III, no. 39: The diabolical cube, pp. 108 & 142 = Hoffmann-Hordern, pp. 108-109, with photos.  6: 0, 1, 1, 1, 1, 1, 1,  i.e. six pieces of volumes  2, 3, 4, 5, 6, 7.  Photos on p. 108 shows Cube Diabolique and its box, by Watilliaux, dated 1874-1895.

J. G.‑Mikusiński.  French patent.  ??NYS -- cited by Steinhaus.

H. Steinhaus.  Mikusiński's Cube.  Mathematical Snapshots.  Not in Stechert, 1938, ed.  OUP, NY:  1950: pp. 140‑142 & 263;  1960, pp. 179‑181 & 326;  1969 (1983): pp. 168-169 & 303.

John Conway.  In an email of 7 Apr 2000, he says he developed the dissection of the  3 x 3 x 3  into  3  1 x 1 x 1  and  6  1 x 2 x 2  in c1960 and then adapted it to the  5 x 5 x 5  into  3  1 x 1 x 3,  1  2 x 2 x 2,  1  1 x 2 x 2  and  13  1 x 2 x 4  and the  5 x 5 x 5  into  3  1 x 1 x 3  and  29  1 x 2 x 2.  He says his first publication of it was in Winning Ways, 1982 (cf below).

Jan Slothouber  &  William Graatsma.  Cubics.  Octopus Press, Deventer, Holland, 1970.  ??NYS.  3 x 3 x 3  into  3  1 x 1 x 1  and  6  1 x 2 x 2.  [Jan de Geus has sent a photocopy of some of this but it does not cover this topic.]

M. Gardner.  SA (Sep 1972) c= Knotted, chap. 3.  Discusses Hoffmann's Diabolical Cube and Mikusiński's cube.  Says he has  8  solutions for the first and that there are just  2  for the second.  The Addendum reports that Wade E. Philpott showed there are just  13  solutions of the Diabolical Cube.  Conway has confirmed this.  Gardner briefly describes the solutions.  Gardner also shows the Lesk Cube, designed by Lesk Kokay (Mathematical Digest [New Zealand] 58 (1978) ??NYS), which has at least  3  solutions.

D. A. Klarner.  Brick‑packing puzzles.  JRM 6 (1973) 112‑117.  Discusses 3 x 3 x 3  into  3  1 x 1 x 1  and  6  1 x 2 x 2  attributed to Slothouber‑Graatsma;  Conway's  5 x 5 x 5  into  3  1 x 1 x 3  and  29  1 x 2 x 2;  Conway's  5 x 5 x 5  into  3  1 x 1 x 3,  1  2 x 2 x 2,  1  1 x 2 x 2  and  13  1 x 2 x 4.  Because of the attribution to Slothouber & Graatsma and not knowing the date of Conway's work, I had generally attributed the  3 x 3 x 3  puzzle to them and Stewart Coffin followed this in his book.  However, it now seems that it really is Conway's invention and I must apologize for misleading people.

Leisure Dynamics, the US distributor of Impuzzables, a series of  6  3 x 3 x 3  cube dissections identified by colours, writes that they were invented by Robert Beck, Custom Concepts Inc., Minneapolis.  However, the Addendum to Gardner, above, says they were designed by Gerard D'Arcey.

Winning Ways.  1982.  Vol. 2, pp. 736-737 & 801.  Gives the  3 x 3 x 3  into  3  1 x 1 x 1  and  6  1 x 2 x 2  and the  5 x 5 x 5  into  3  1 x 1 x 3,  1  2 x 2 x 2,  1  1 x 2 x 2  and  13  1 x 2 x 4,  which is called Blocks-in-a-Box.  No mention of the other  5 x 5 x 5.  Mentions Foregger & Mather, cf in 6.U.2.

Michael Keller.  Polycube update.  World Game Review 4 (Feb 1985) 13.  Reports results of computer searches for solutions.  Hoffmann's Diabolical Cube has  13;  Mikusinski's Cube has  2;  Soma Cube has  240;  Impuzzables:  White -- 1;  Red -- 1;  Green -- 16;  Blue -- 8;  Orange -- 30;  Yellow -- 1142.

Michael Keller.  Polyform update.  World Game Review 7 (Oct 1987) 10‑13.  Says that Nob Yoshigahara has solved a problem posed by O'Beirne:  How many ways can  9  L‑trominoes make a cube?  Answer is  111.  Gardner, Knotted, chap. 3, mentioned this.  Says there are solutions with  n  L‑trominoes and  9‑n  straight trominoes for  n ¹ 1 and there are  4  solutions for  n = 0.  Says the Lesk Cube has  4  solutions.  Says Naef's Gemini Puzzle was designed by Toshiaki Betsumiya.  It consists of the  10  ways to join two  1 x 2 x 2  blocks.

H. J. M. van Grol.  Rik's Cube Kit -- Solid Block Puzzles.  Analysis of all  3 x 3 x 3  unit solid block puzzles with non‑planar  4‑unit and  5‑unit shapes.  Published by the author, The Hague, 1989, 16pp.  There are  3  non‑planar tetracubes and  17  non‑planar pentacubes.  A  3 x 3 x 3  cube will require the  3  non‑planar tetracubes and  3  of the non‑planar pentacubes -- assuming no repeated pieces.  He finds  190  subsets which can form cubes, in  1  to  10  different ways.

Nob Yoshigahara.  (Title in Japanese: (Puzzle in Wood)).  H. Tokuda, Sowa Shuppan, Japan, 1987.  Pp. 68-69 is a  3^3  designed by Nob --  6: 01005.

 

          6.G.2.          DISSECTION OF  63  INTO  33,  43  AND  53,  ETC.

 

H. W. Richmond.  Note 1672:  A geometrical problem.  MG 27 (No. 275) (Jul 1943) 142.  AND  Note 1704:  Solution of a geometrical problem (Note 1672).  MG 28 (No. 278) (Feb 1944) 31‑32.  Poses the problem of making such a dissection, then gives a solution in 12 pieces:  three  1 x 3 x 3;  4 x 4 x 4;  four 1 x 5 x 5;  1 x 4 x 4;  two  1 x 1 x 2  and a  V‑pentacube.

Anon. [= John Leech, according to Gardner, below].  Two dissection problems, no. 2.  Eureka  13 (Oct 1950) 6  &  14 (Oct 1951) 23.  Asks for such a dissection using at most 10 pieces.  Gives an 8 piece solution due to R. F. Wheeler.  [Cundy & Rollett; Mathematical Models; 2nd ed., pp. 203‑205, say Eureka is the first appearance they know of this problem.  See Gardner, below, for the identity of Leech.]

Richard K. Guy.  Loc. cit. in 5.H.2, 1960.  Mentions the 8 piece solution.

J. H. Cadwell.  Some dissection problems involving sums of cubes.  MG 48 (No. 366) (Dec 1964) 391‑396.  Notes an error in Cundy & Rollett's account of the Eureka problem.  Finds examples for  123 + 13 = 103 + 93  with 9 pieces  and  93 = 83 + 63 + 13  with 9 pieces.

J. H. Cadwell.  Note 3278:  A three‑way dissection based on Ramanujan's number.  MG 54 (No. 390) (Dec 1970) 385‑387.  7 x 13 x 19  to  103 + 93  and  123 + 13  using 12 pieces.

M. Gardner.  SA (Oct 1973) c= Knotted, chap. 16.  He says that the problem was posed by John Leech.  He gives Wheeler's initials as E. H. ??  He says that J. H. Thewlis found a simpler 8‑piece solution, further simplified by T. H. O'Beirne, which keeps the  4 x 4 x 4  cube intact.  This is shown in Gardner.  Gardner also shows an 8‑piece solution which keeps the  5 x 5 x 5  intact, due to E. J. Duffy, 1970.  O'Beirne showed that an 8‑piece dissection into blocks is impossible and found a 9‑block solution in 1971, also shown in Gardner.

Harry Lindgren.  Geometric Dissections.  Van Nostrand, Princeton, 1984.  Section 24.1, pp. 118‑120 gives Wheeler's solution and admires it.

Richard K. Guy, proposer;  editors & Charles H. Jepson [should be Jepsen], partial solvers.  Problem 1122.  CM 12 (1987) 50  &  13 (1987) 197‑198.  Asks for such dissections under various conditions, of which (b) is the form given in Eureka.  Eight pieces is minimal in one case and seems minimal in two other cases.  Eleven pieces is best known for the first case, where the pieces must be blocks, but this appears to be the problem solved by O'Beirne in 1971, reported in Gardner, above.

Charles H. Jepsen.  Additional comment on Problem 1122.  CM 14 (1988) 204‑206.  Gives a ten piece solution of the first case.

Chris Pile.  Cube dissection.  M500 134 (Aug 1993) 2-3.  He feels the  1 x 1 x 2  piece occurring in Cundy & Rollett is too small and he provides another solution with 8 pieces, the smallest of which contains 8 unit cubes.  Asks how uniform the piece sizes can be.

 

          6.G.3.          DISSECTION OF A DIE INTO NINE  1 x 1 x 3

 

Hoffmann.  1893.  Chap. III, no. 17: The "Spots" puzzle, pp. 98‑99 & 130‑131 = Hoffmann‑Hordern, pp. 90-91, with photo.  Says it is made by Wolff & Son.  Photo on p. 91 shows an example made by E. Wolff & Son, London.

Benson.  1904.  The spots puzzle, pp. 203‑204.  As in Hoffmann.

Collins.  Book of Puzzles.  1927.  Pp. 131‑134: The dissected die puzzle.  The solution is different than Hoffmann's.

Rohrbough.  Puzzle Craft.  1932.  P. 21 shows a dissected die, but with no text.  The picture is the same as in Hoffmann's solution.

Slocum.  Compendium.  Shows Diabolical Dice from Johnson Smith catalogue, 1935.

Harold Cataquet.  The Spots puzzle revisited.  CFF 33 (Feb 1994) 20-21.  Brief discussion of two versions.

David Singmaster.  Comment on the "Spots" puzzle.  29 Sep 1994, 2pp.  Letter in response to the above.  I note that there is no standard pattern for a die other than the opposite sides adding to seven.  There are  23 = 8  ways to orient the spots forming  2, 3, and 6.  There are two handednesses, so there are 16 dice altogether.  (This was pointed out to me perhaps 10 years before by Richard Guy and Ray Bathke.  I have since collected examples of all 16 dice.)  However, Ray Bathke showed me Oriental dice with the two spots of the 2 placed horizontal or vertically rather than diagonally, giving another 16 dice (I have 5 types), making 32 dice in all.  A die can be dissected into  9  1 x 1 x 3  pieces in 6 ways if the layers have to alternate in direction, or in 21 ways in general.  I then pose a number of questions about such dissections.

 

          6.G.4.          USE OF OTHER POLYHEDRAL PIECES

 

S&B.  1986.  P. 42 shows Stewart Coffin's 'Pyramid Puzzle' using pieces made from truncated octahedra and his 'Setting Hen' using pieces made from rhombic dodecahedra.  Coffin probably devised these in the 1960s -- perhaps his book has some details of the origins of these ideas.  ??check.

Mark Owen & Matthew Richards.  A song of six splats.  Eureka 47 (1987) 53‑58.  There are six ways to join three truncated octahedra.  For reasons unknown, these are called 3‑splats.  They give various shapes which can and which cannot be constructed from the six 3‑splats.

 

          6.H.   PICK'S THEOREM

 

Georg Pick.  Geometrisches zur Zahlenlehre.  Sitzungsberichte des deutschen naturwissenschaftlich‑medicinischen Vereines für Böhmen "Lotos" in Prag (NS) 19 (1899) 311‑319.  Pp. 311‑314 gives the proof, for an oblique lattice.  Pp. 318‑319 gives the extension to multiply connected and separated regions.  Rest relates to number theory.  [I have made a translation of the material on Pick's Theorem.]

Charles Howard Hinton.  The Fourth Dimension.  Swan Sonnenschein & Co., London, 1906.  Metageometry, pp. 46-60.  [This material is in Speculations on the Fourth Dimension, ed. by R. v. B. Rucker; Dover, 1980, pp. 130-141.  Rucker says the book was published in 1904, so my copy may be a reprint??]  In the beginning of this section, he draws quadrilateral shapes on the square lattice and determines the area by counting points, but he counts  I + E/2 + C/4,  which works for quadrilaterals but is not valid in general.

H. Steinhaus.  O mierzeniu pól płaskich.  Przegląd Matematyczno‑Fizyczny 2 (1924) 24‑29.  Gives a version of Pick's theorem, but doesn't cite Pick.  (My thanks to A. Mąkowski for an English summary of this.)

H. Steinhaus.  Mathematical Snapshots.  Stechert, NY, 1938, pp. 16-17 & 132.  OUP, NY:  1950: pp. 76‑77 & 260 (note 77);  1960: pp. 99‑100 & 324 (note 95);  1969 (1983): pp. 96‑97 & 301 (note 107).  In 1938 he simply notes the theorem and gives one example.  In 1950, he outlines Pick's argument.  He  refers to Pick's paper, but in "Ztschr. d. Vereins 'Lotos' in Prag".  Steinhaus also cites his own paper, above.

J. F. Reeve.  On the volume of lattice polyhedra.  Proc. London Math. Soc. 7 (1957) 378‑395.  Deals with the failure of the obvious form of Pick's theorem in 3‑D and finds a valid generalization.

Ivan Niven & H. S. Zuckerman.  Lattice points and polygonal area.  AMM 74 (1967) 1195‑1200.  Straightforward proof.  Mention failure for tetrahedra.

D. W. De Temple & J. M. Robertson.  The equivalence of Euler's and Pick's theorems.  MTr 67 (1974) 222‑226.  ??NYS.

W. W. Funkenbusch.  From Euler's formula to Pick's formula using an edge theorem.  AMM 81 (1974) 647‑648.  Easy proof though it could be easier.

R. W. Gaskell, M. S. Klamkin & P. Watson.  Triangulations and Pick's theorem.  MM 49 (1976) 35‑37.  A bit roundabout.

Richard A. Gibbs.  Pick iff Euler.  MM 49 (1976) 158.  Cites DeTemple & Robertson and observes that both Pick and Euler can be proven from a result on triangulations.

John Reay.  Areas of hex-lattice polygons, with short sides.  Abstracts Amer. Math. Soc. 8:2 (1987) 174, #832-51-55.  Gives a formula for the area in terms of the boundary and interior points and the characteristic of the boundary, but it is an open question to determine when this formula gives the actual area.

 

6.I.     SYLVESTER'S PROBLEM OF COLLINEAR POINTS

 

          If a set of non‑collinear points in the plane is such that the line through any two points of the set contains a third point of the set, then the set is infinite.

 

J. J. Sylvester.  Question 11851.  The Educational Times 46 (NS, No. 383) (1 Mar 1893) 156.

H. J. Woodall & editorial comment.  Solution to Question 11851.  Ibid. (No. 385) (1 May 1893) 231.  A very spurious solution.

(The above two items appear together in Math. Quest. with their Sol. Educ. Times 59 (1893) 98‑99.)

E. Melchior.  Über Vielseite der projecktiven Ebene.  Deutsche Math. 5 (1940) 461‑475.  Solution, but in a dual form.

P. Erdös, proposer;  R. Steinberg, solver & editorial comment giving solution of T. Grünwald (later = T. Gallai).  Problem 4065.  AMM 50 (1943) 65  &  51 (1944) 169‑171.

L. M. Kelly.  (Solution.)  In:  H. S. M. Coxeter; A problem of collinear points; AMM 55 (1948) 26‑28.  Kelly's solution is on p. 28.

G. A. Dirac.  Note 2271:  On a property of circles.  MG 36 (No. 315) (Feb 1952) 53‑54.  Replace 'line' by 'circle' in the problem.  He shows this is true by inversion.  He asks for an independent proof of the result, even for the case when  two, three  are replaced by  three, four.

D. W. Lang.  Note 2577:  The dual of a well‑known theorem.  MG 39 (No. 330) (Dec 1955) 314.  Proves the dual easily.

H. S. M. Coxeter.  Introduction to Geometry.  Wiley, 1961.  Section 4.7: Sylvester's problem of collinear points, pp. 65-66.  Sketches history and gives Kelly's proof.

W. O. J. Moser.  Sylvester's problem, generalizations and relatives.  In his:  Research Problems in Discrete Geometry 1981, McGill University, Montreal, 1981.  Section 27, pp. 27‑1 -- 27‑14.  Survey with 73 references.  (This problem is not in Part 1 of the 1984 ed. nor in the 1986 ed.)

 

6.J.    FOUR BUGS AND OTHER PURSUIT PROBLEMS

 

          The general problem becomes too technical to remain recreational, so I will not try to be exhaustive here.

 

Arthur Bernhart. 

Curves of pursuit.  SM 20 (1954) 125‑141.

Curves of pursuit -- II.  SM 23 (1957) 49‑65.

Polygons of pursuit.  SM 24 (1959) 23‑50. 

Curves of general pursuit.  SM 24 (1959) 189‑206.

Extensive history and analysis.  First article covers one dimensional pursuit, then two dimensional linear pursuit.  Second article deals with circular pursuit.  Third article is the 'four bugs' problem -- analysis of equilateral triangle, square, scalene triangle, general polygon, Brocard points, etc.  Last article includes such variants as variable speed, the tractrix, miscellaneous curves, etc.

 

Mr. Ash, proposer; editorial note saying there is no solver.  Ladies' Diary, 1748-47  =  T. Leybourn, II: 15-17, quest. 310, with 'Solution by ΦIΛΟΠΟΝΟΣ, taken from Turner's Exercises, where this question was afterwards proposed and answered ...'  A fly is constrained to move on the periphery of a circle.  Spider starts 30o away from the fly, but walks across the circle, always aiming at the fly.  If she catches the fly 180o from her starting point, find the ratio of their speeds.  ΦIΛΟΠΟΝΟΣ solves the more general problem of finding the curve when the spider starts anywhere.

Carlile.  Collection.  1793.  Prob. CV, p. 62.  A dog and a duck are in a circular pond of radius 40 and they swim at the same speed.  The duck is at the edge and swims around the circumference.  The dog starts at the centre and always swims toward the duck, so the dog and the duck are always on a radius.  How far does the dog swim in catching the duck?  He simply gives the result as  20π.  Letting  R  be the radius of the pond and  V  be the common speed, I find the radius of the dog,  r,  is given by  r = R sin Vt/R.  Since the angle,  θ,  of both the duck and the dog is given by  θ = Vt/R,  the polar equation of the dog's path is  r = R sin θ  and the path is a semicircle whose diameter is the appropriate radius perpendicular to the radius to the duck's initial position.

Cambridge Math. Tripos examination, 5 Jan 1871, 9 to 12.  Problem 16, set by R. K. Miller.  Three bugs in general position, but with velocities adjusted to make paths similar and keep the triangle similar to the original.

Lucas.  (Problem of three dogs.)  Nouvelle Correspondance Mathématique 3 (1877) 175‑176.  ??NYS -- English in Arc., AMM 28 (1921) 184‑185 & Bernhart.

H. Brocard.  (Solution of Lucas' problem.)  Nouv. Corr. Math. 3 (1877) 280.  ??NYS -- English in Bernhart.

Pearson.  1907.  Part II, no. 66: A duck hunt, pp. 66 & 172.  Duck swims around edge of pond;  spaniel starts for it from the centre at the same speed.

A. S. Hathaway, proposer and solver.  Problem 2801.  AMM 27 (1920) 31  &  28 (1921) 93‑97.  Pursuit of a prey moving on a circle.  Morley's and other solutions fail to deal with the case when the velocities are equal.  Hathaway resolves this and shows the prey is then not caught.

F. V. Morley.  A curve of pursuit.  AMM 28 (1921) 54-61.  Graphical solution of Hathaway's problem.

R. C. Archibald [Arc.] & H. P. Manning.  Remarks and historical notes on problems 19 [1894], 160 [1902], 273 [1909] & 2801 [1920].  AMM 28 (1921) 91-93.

W. W. Rouse Ball.  Problems -- Notes: 17: Curves of pursuit.  AMM 28 (1921) 278‑279.

A. H. Wilson.  Note 19: A curve of pursuit.  AMM 28 (1921) 327.

Editor's note to Prob. 2 (proposed by T. A. Bickerstaff), National Mathematics Magazine (1937/38) 417 cites Morley and Archibald and adds that some authors credit the problem to Leonardo da Vinci -- e.g. MG (1930-31) 436 -- ??NYS

Nelson F. Beeler & Franklyn M. Branley.  Experiments in Optical Illusion.  Ill. by Fred H. Lyon.  Crowell, 1951, An illusion doodle, pp. 68-71, describes the pattern formed by four bugs starting at the corners of a square, drawing the lines of sight at (approximately) regular intervals.  Putting several of the squares together, usually with alternating directions of motion, gives a pleasant pattern which is now fairly common.  They call this 'Huddy's Doodle', but give no source.

J. E. Littlewood.  A Mathematician's Miscellany.  Op. cit. in 5.C.  1953.  'Lion and man', pp. 135‑136 (114‑117).  The 1986 ed. adds three diagrams and revises the text somewhat.  I quote from it.  "A lion and a man in a closed circular arena have equal maximum speeds.  What tactics should the lion employ to be sure of his meal?"  This was "invented by R. Rado in the late thirties" and "swept the country 25 years later".  [The 1953 ed., says Rado didn't publish it.]  The correct solution "was discovered by Professor A. S. Besicovitch in 1952".  [The 1953 ed. says "This has just been discovered ...; here is the first (and only) version in print."]

C. C. Puckette.  The curve of pursuit.  MG 37 (No. 322) (Dec 1953) 256‑260.  Gives the history from Bouguer in 1732.  Solves a variant of the problem.

R. H. Macmillan.  Curves of pursuit.  MG 40 (No. 331) (Feb 1956) 1‑4.  Fighter pursuing bomber flying in a straight line.  Discusses firing lead and acceleration problems.

Gamow & Stern.  1958.  Homing missiles.  Pp. 112‑114.

Howard D. Grossman, proposer;  unspecified solver.  Problem 66 -- The walk around.  In:  L. A. Graham; Ingenious Mathematical Problems and Methods; Dover, 1959, pp. 40 & 203‑205.  Four bugs -- asserts Grossman originated the problem.

I. J. Good.  Pursuit curves and mathematical art.  MG 43 (No. 343) (Feb 1959) 34‑35.  Draws tangent to the pursuit curves in an equilateral triangle and constructs various patterns with them.  Says a similar but much simpler pattern was given by G. B. Robison; Doodles; AMM 61 (1954) 381-386, but Robison's doodles are not related to pursuit curves, though they may have inspired Good to use the pursuit curves.

J. Charles Clapham.  Playful mice.  RMM 10 (Aug 1962) 6‑7.  Easy derivation of the distance travelled for  n  bugs at corners of a regular  n‑gon.  [I don't see this result in Bernhart.]

C. G. Paradine.  Note 3108:  Pursuit curves.  MG 48 (No. 366) (Dec 1964) 437‑439.  Says Good makes an error in Note 3079.  He shows the length of the pursuit curve in the equilateral triangle is    of the side and describes the curve as an equiangular spiral.  Gives a simple proof that the length of the pursuit curve in the regular  n‑gon is the side divided by  (1 ‑ cos 2π/n).

M. S. Klamkin & D. J. Newman.  Cyclic pursuit or "The three bugs problem".  AMM 78 (1971) 631‑639.  General treatment.  Cites Bernhart's four SM papers and some of the history therein.

P. K. Arvind.  A symmetrical pursuit problem on the sphere and the hyperbolic plane.  MG 78 (No. 481) (Mar 1994) 30-36.  Treats the  n  bugs problems on the surfaces named.

Barry Lewis.  A mathematical pursuit.  M500 170 (Oct 1999) 1-8.  Starts with equilateral triangular case, giving QBASIC programs to draw the curves as well as explicit solutions.  Then considers regular  n-gons.  Then considers simple pursuit, one beast pursuing another while the other moves along some given path.  Considers the path as a straight line or a circle.  For the circle, he asserts that the analytic solution was not determined until 1926, but gives no reference.

 

6.K.   DUDENEY'S SQUARE TO TRIANGLE DISSECTION

 

Dudeney.  Weekly Dispatch (6 Apr, 20 Apr, 4 May, 1902) all p. 13.

Dudeney.  The haberdasher's puzzle.  London Mag. 11 (No. 64) (Nov 1903) 441 & 443.  (Issue with solution not found.)

Dudeney.  Daily Mail (1  &  8 Feb 1905) both p. 7.

Dudeney.  CP.  1907.  Prob. 25: The haberdasher's puzzle, pp. 49‑50 & 178‑180.

Western Puzzle Works, 1926 Catalogue.  No. 1712 -- unnamed, but shows both the square and the triangle.  Apparently a four piece puzzle.

M. Adams.  Puzzle Book.  1939.  Prob. C.153: Squaring a triangle, pp. 162 & 189.  Asserts that Dudeney's method works for any triangle, but his example is close to equilateral and I recall that this has been studied and only certain shapes will work??

Robert C. Yates.  Geometrical Tools.  (As:  Tools; Baton Rouge, 1941);  revised ed., Educational Publishers, St. Louis, 1949.  Pp. 40-41.  Extends to dissecting a quadrilateral to a specified triangle and gives a number of related problems.

 

          6.L.    CROSSED LADDERS

 

          Two ladders are placed across a street, each reaching from the base of the house on one side to the house on the other side. 

          The simple problem gives the heights  a,  b  that the ladders reach on the walls.  If the height of the crossing is  c,  we easily get  1/c = 1/a + 1/b.  NOTATION -- this problem will be denoted by  (a, b).

          The more common and more complex problem is where the ladders have lengths  a  and  b,  the height of their crossing is  c  and one wants the width  d  of the street.  If the heights of the ladder ends are  x,  y,  this situation gives  x2 ‑ y2 = a2 ‑ b2  and  1/x + 1/y = 1/c  which leads to a quartic and there seems to be no simple solution.  NOTATION -- this will be denoted  (a, b, c).

 

Mahavira.  850.  Chap. VII, v. 180-183, pp. 243-244.  Gives the simple version with a modification -- each ladder reaches from the top of a pillar beyond the foot of the other pillar.  The ladder from the top of pillar  Y  (of height  y)  extends by  m  beyond the foot of pillar  X  and the ladder from the top of pillar  X  (of height  x)  reaches  n  beyond the foot of pillar  Y.  The pillars are  d  apart.  Similar triangles then yield:  (d+m+n)/c  =  (d+n)/x + (d+m)/y  and one can compute the various distances along the ground.  He first does problems with  m = n = 0,  which are the simple version of the problem, but since  d  is given, he also asks for the distances on the ground.

                    v. 181.  (16, 16)  with  d = 16.

                    v. 182.  (36, 20)  with  d = 12.

                    v. 183.  x, y, d, m, n  =  12, 15, 4, 1, 4.

Bhaskara II.  Lilavati.  1150.  Chap. VI, v. 160.  In Colebrooke, pp. 68‑69.  (10, 15).  (= Bijaganita, ibid., chap. IV, v. 127, pp. 205‑206.)

Fibonacci.  1202.  Pp. 397‑398 (S: 543-544) looks like a crossed ladders problem but is a simple right triangle problem.

Pacioli.  Summa.  1494.  Part II.

F. 56r, prob. 48.  (4, 6).

F. 60r, prob. 64.  (10, 15).

Hutton.  A Course of Mathematics.  1798?  Prob. VIII,  1833: 430;  1857: 508.  A ladder  40  long in a roadway can reach  33  up one side and, from the same point, can reach  21  up the other side.  How wide is the street?  This is actually a simple right triangle problem.

Victor Katz reports that Hutton's problem, with values  60; 37, 23  appears in a notebook of Benjamin Banneker (1731-1806).

Loyd.  Problem 48: A jubilee problem.  Tit‑Bits 32 (21 Aug,  11  &  25 Sep 1897) 385,  439  &  475.  Given heights of the ladder ends above ground and the width of the street, find the height of the intersection.  However one wall is tilted and the drawing has it covered in decoration so one may interpret the tilt in the wrong way.

Jno. A. Hodge, proposer;  G. B. M. Zerr, solver.  Problem 131.  SSM 8 (1908) 786  &  9 (1909) 174‑175.  (100, 80, 10).

W. V. N. Garretson, proposer;  H. S. Uhler, solver.  Problem 2836.  AMM 27 (1920)  &  29 (1922) 181.  (40, 25, 15).

C. C. Camp, proposer;  W. J. Patterson & O. Dunkel, solvers.  Problem 3173.  AMM 33 (1926) 104  &  34 (1927) 50‑51.  General solution.

Morris Savage, proposer;  W. E. Batzler, solver.  Problem 1194.  SSM 31 (1931) 1000  &  32 (1932) 212.  (100, 80, 10).

S. A. Anderson, proposer;  Simon Vatriquant, solver.  Problem E210.  AMM 43 (1936) 242  &  642‑643.  General solution in integers. 

C. R. Green, proposer;  C. W. Trigg, solver.  Problem 1498.  SSM 37 (1937) 484  &  860‑861.  (40, 30, 15).  Trigg cites Vatriquant for smallest integral case.

A. A. Bennett, proposer;  W. E. Buker, solver.  Problem E433.  AMM 47 (1940) 487  &  48 (1941) 268‑269.  General solution in integers using four parameters.

J. S. Cromelin, proposer;  Murray Barbour, solver.  Problem E616 -- The three ladders.  AMM 51 (1944) 231  &  592.  Ladders of length  60  &  77  from one side.  A ladder from the other side crosses them at heights  17  &  19.  How long is the third ladder and how wide is the street?

Geoffrey Mott-Smith.  Mathematical Puzzles for Beginners and Enthusiasts.  (Blakiston, 1946);  revised ed., Dover, 1954.  Prob. 103: The extension ladder, pp. 58-59 & 176‑178.  Complex problem with three ladders.

Arthur Labbe, proposer;  various solvers.  Problem 25 -- The two ladders.  Sep 1947 [date given in Graham's second book, cited at 1961].  In:  L. A. Graham; Ingenious Mathematical Problems and Methods; Dover, 1959, pp. 18 & 116‑118.  (20, 30, 8).

M. Y. Woodbridge, proposer and solver.  Problem 2116.  SSM 48 (1948) 749  &  49 (1949) 244‑245.  (60, 40, 15).  Asks for a trigonometric solution.  Trigg provides a list of early references.

Robert C. Yates.  The ladder problem.  SSM 51 (1951) 400‑401.  Gives a graphical solution using hyperbolas.

G. A. Clarkson.  Note 2522:  The ladder problem.  MG 39 (No. 328) (May 1955) 147‑148.  (20, 30, 10).  Let  A = Ö(a2 ‑ b2)  and set  x = A sec t,  y = A tan t.  Then  cos t + cot t = A  and he gets a trigonometrical solution.  Another method leads to factoring the quartic in terms of a constant  k  whose square satisfies a cubic.

L. A. Graham.  The Surprise Attack in Mathematical Problems.  Dover, 1968.  Problem 6: Searchlight on crossed ladders, pp. 16-18.  Says they reposed Labbe's Sep 1947 problem in Jun 1961.  Solution by William M. Dennis which is the same trigonometric method as Clarkson.

H. E. Tester.  Note 3036:  The ladder problem.  A solution in integers.  MG 46 (No. 358) (Dec 1962) 313‑314.  A four parameter, incomplete, solution.  He finds the example  (119, 70, 30).

A. Sutcliffe.  Complete solution of the ladder problem in integers.  MG 47 (No. 360) (May 1963) 133‑136.  Three parameter solution.  First few examples are:  (119, 70, 30);  (116, 100, 35);  (105, 87, 35).  Simpler than Anderson and Bennett/Buker.

Alan Sutcliffe, proposer;  Gerald J. Janusz, solver.  Problem 5323 -- Integral solutions of the ladder problem.  AMM 72 (1965) 914  &  73 (1966) 1125-1127.  Can the distance  f  between the tops of the ladders be integral?  (80342, 74226, 18837)  has  x = 44758,  y = 32526,  d = 66720,  f = 67832.  This is not known to be the smallest example.

Anon.  A window cleaner's problem.  Mathematical Pie 51 (May 1967) 399.  From a point in the road, a ladder can reach  30  ft up on one side and  40  ft up on the other side.  If the two ladder positions are at right angles, how wide is the road?

J. W. Gubby.  Note 60.3:  Two chestnuts (re-roasted).  MG 60 (No. 411) (Mar 1976) 64-65.  1.  Given heights of ladders as  a, b,  what is the height  c  of their intersection?  Solution:  1/c = 1/a + 1/b  or  c = ab/(a+b).  2.  The usual ladder problem -- he finds a quartic.

J. Jabłkowski.  Note 61:11:  The ladder problem solved by construction.  MG 61 (No. 416) (Jun 1977) 138.  Gives a 'neusis' construction.  Cites Gubby.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 83, A second ladder problem, pp. 58-59 & 115-118.  (15, 20, 6).  Uses  xy  as a variable to simplify the quartic for numerical solution and eventually gets  11.61.

See:  Gardner, Circus, p. 266 & Schaaf for more references.  ??follow up.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  The tangled ladders, pp. 43-44 & 116.  (30, 20, 10).  Gives answer  12.311857...  with no explanation.

 

6.L.1.        LADDER OVER BOX

 

 

          A ladder of length  L  is placed to just clear a box of width  w  and height  h  at the base of a wall.  How high does the ladder reach?  Denote this by  (w, h, L).  Letting  x  be the horizontal distance of the foot and  y  be the vertical distance of the top of the ladder, measured from the foot of the wall, we get  x2 + y2 = L2  and  (x‑w)(y‑h) = wh,  which gives a quartic in general.  But if  w = h,  then use of  x + y  as a variable reduces the quartic to a quadratic.  In this case, the idea is old -- see e.g. Simpson.

          The question of determining shortest ladder which can fit over a box of width  w  and height  h  is the same as determining the longest ladder which will pass from a corridor of width  w  into another corridor of width  h.  See Huntington below and section 6.AG.

 

 

Simpson.  Algebra.  1745.  Section XVIII, prob. XV, p. 250 (1790: prob. XIX, pp. 272-273).  "The Side of the inscribed Square  BEDF,  and the Hypotenuse  AC  of a right-angled Triangle  ABC  being given; to determine the other two Sides of the Triangle  AB  and  BC."  Solves "by considering  x + y  as one Quantity".

Pearson.  1907.  Part II, no. 102: Clearing the wall, p. 103.  For  (15, 12, 52),  the ladder reaches 48.

D. John Baylis.  The box and ladder problem.  MTg 54 (1971) 24.  (2, 2, 10).  Finds the quartic which he solves by symmetry.  Editorial note in MTg 57 (1971) 13 says several people wrote to say that use of similar triangles avoids the quartic.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  The ladder and the box problem, pp. 44-45.  = Birtwistle; Calculator Puzzle Book; 1978; Prob. 53: A ladder problem, pp. 37 & 96‑98.  (3, 3, 10).  Solves by using  x + y - 6  as a variable.

Monte Zerger.  The "ladder problem".  MM 60:4 (1987) 239‑242.  (4, 4, 16).  Gives a trigonometric solution and a solution via two quadratics.

Oliver D. Anderson.  Letter.  MM 61:1 (1988) 63.  In response to Zerger's article, he gives a simpler derivation.

Tom Heyes.  The old box and ladder problem -- revisited.  MiS 19:2 (Mar 1990) 42‑43.  Uses a graphic calculator to find roots graphically and then by iteration.

A. A. Huntington.  More on ladders.  M500 145 (Jul 1995) 2-5.  Does usual problem, getting a quartic.  Then finds the shortest ladder.  [This turns out to be the same as the longest ladder one can get around a corner from corridors of widths  w  and  h,  so this problem is connected to 6.AG.]

David Singmaster.  Integral solutions of the ladder over box problem.  In preparation.  Easily constructs all the primitive integral examples from primitive Pythagorean triples.  E.g. for the case of a square box, i.e.  w = h,  if  X, Y, Z  is a primitive Pythagorean triple, then the corresponding primitive solution has  w = h = XY,  x = X (X + Y),  y = Y (X + Y),  L = Z (X + Y),  and remarkably,  x - h = X2,  y - w = Y2.

 

          6.M.   SPIDER & FLY PROBLEMS

 

          These involve finding the shortest distance over the surface of a cube or cylinder.  I've just added the cylindrical form -- see Dudeney (1926), Perelman and Singmaster.  The shortest route from a corner of a cube or cuboid to a diagonally opposite corner must date back several centuries, but I haven't seen any version before 1937!  I don't know if other shapes have been done -- the regular (and other) polyhedra and the cone could be considered.

          Two-dimensional problems are in 10.U.

 

Loyd.  The Inquirer (May 1900).  Gives the Cyclopedia problem.  ??NYS -- stated in a letter from Shortz.

Dudeney.  Problem 501 -- The spider and the fly.  Weekly Dispatch (14  &  28 Jun 1903) both p. 16.  4 side version.

Dudeney.  Breakfast table problems, No. 320 -- The spider and the fly.  Daily Mail (18  &  21 Jan 1905) both p. 7.  Same as the above problem.

Dudeney.  Master of the breakfast table problem.  Daily Mail (1  &  8 Feb 1905) both p. 7.  Interview with Dudeney in which he gives the 5 side version. 

Ball.  MRE, 4th ed., 1905, p. 66.  Gives the 5 side version, citing the Daily Mail of 1 Feb 1905.  He says he heard a similar problem in 1903 -- presumably Dudeney's first version.  In the 5th ed., 1911, p. 73, he attributes the problem to Dudeney.

Dudeney.  CP.  1907.  Prob. 75: The spider and the fly, pp. 121‑122 & 221‑222.  5 side version with discussion of various generalizations.

Dudeney.  The world's best problems.  1908.  Op. cit. in 2.  P. 786 gives the five side version.

Sidney J. Miller.  Some novel picture puzzles -- No. 6.  Strand Mag. 41 (No. 243) (Mar 1911) 372  &  41 (No. 244) (Apr 1911) 506.  Contest between two snails.  Better method uses four sides, similar to Dudeney's version, but with different numbers.

Loyd.  The electrical problem.  Cyclopedia, 1914, pp. 219 & 368 (= MPSL2, prob. 149, pp. 106 & 169  = SLAHP: Wiring the hall, pp. 72 & 114).  Same as Dudeney's first, four side, version.  (In MPSL2, Gardner says Loyd has simplified Dudeney's 5 side problem.  More likely(?) Loyd had only seen Dudeney's earlier 4 side problem.)

Dudeney.  MP.  1926.  Prob. 162: The fly and the honey, pp. 67 & 157.  (= 536, prob. 325, pp. 112 & 313.)  Cylindrical problem.

Perelman.  FFF.  1934.  The way of the fly.  1957: Prob. 68, pp. 111‑112 & 117‑118;  1979: Prob. 72, pp. 136 & 142‑144.  MCBF: Prob. 72, pp. 134 & 141-142.  Cylindrical form, but with different numbers and arrangement than Dudeney's MP problem.

Haldeman-Julius.  1937.  No. 34: The louse problem, pp. 6 & 22.  Room  40 x 20 x 10  with louse at a corner wanting to go to a diagonally opposite corner.  Problem sent in by J. R. Reed of Emmett, Idaho.  Answer is  50!

M. Kraitchik.  Mathematical Recreations, 1943, op. cit. in 4.A.2, chap. 1, prob. 7, pp. 17‑21.  Room with 8 equal routes from spider to fly.  (Not in his Math. des Jeux.)

Sullivan.  Unusual.  1943.  Prob. 10: Why not fly?  Find shortest route from a corner of a cube to the diagonally opposite corner.

William R. Ransom.  One Hundred Mathematical Curiosities.  J. Weston Walch, Portland, Maine, 1955.  The spider problem, pp. 144‑146.  There are three types of path, covering  3, 4 and 5  sides.  He determines their relative sizes as functions of the room dimensions.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

Round the cone, pp. 144 & 195.  What is the shortest distance from a point  P  around a cone and back to  P?  Answer is "An ellipse", which doesn't seem to answer the question.  If the cone has height  H,  radius  R  and  P  is  l  from the apex,  then the slant height  L  is  Ö(R2 + H2),  the angle of the opened out cone is  θ = 2πR/L  and the required distance is  2l sin θ/2.

Spider circuit, pp. 144 & 198.  Spider is at the midpoint of an edge of a cube.  He wants to walk on each of the faces and return.  What is his shortest route?  Answer is "A regular hexagon.  (This may be demonstrated by putting a rubber band around a cube.)"

David Singmaster.  The spider spied her.  Problem used as:  More than one way to catch a fly, The Weekend Telegraph (2 Apr 1984).  Spider inside a glass tube, open at both ends, goes directly toward a fly on the outside.  When are there two equally short paths?  Can there be more than two shortest routes?

Yoshiyuki Kotani has posed the following general and difficult problem.  On an  a x b x c  cuboid, which two points are furthest apart, as measured by an ant on the surface?  Dick Hess has done some work on this, but I believe that even the case of square cross-section is not fully resolved.

 

6.N.   DISSECTION OF A  1 x 1 x 2  BLOCK TO A CUBE

 

W. F. Cheney, Jr., proposer;  W. R. Ransom; A. H. Wheeler, solvers.  Problem E4.  AMM 39 (1932) 489;  40 (1933) 113-114  &  42 (1934) 509-510.  Ransom finds a solution in  8  pieces;  Wheeler in  7.

Harry Lindgren.  Geometric Dissections.  Van Nostrand, Princeton, 1964.  Section 24.2, p. 120 gives a variant of Wheeler's solution.

Michael Goldberg.  A duplication of the cube by dissection and a hinged linkage.  MG 50 (No. 373) (Oct 1966) 304‑305.  Shows that a hinged version exists with 10 pieces.  Hanegraaf, below, notes that there are actually 12 pieces here.

Anton Hanegraaf.  The Delian Altar Dissection.  Polyhedral Dissections, Elst, Netherlands, 1989.  Surveys the problem, gives a 6 piece solution and a 7 piece hinged solution.

 

          6.O.   PASSING A CUBE THROUGH AN EQUAL OR SMALLER CUBE. PRINCE RUPERT'S PROBLEM

 

          The projection of a unit cube along a space diagonal is a regular hexagon of side  Ö2/Ö3.  The largest square inscribable in this hexagon has edge  Ö6 - Ö2 = 1.03527618.  By passing the larger cube on a slant to the space diagonal, one can get the larger cube having edge  3Ö2/4 = 1.06066172.

          There are two early attributions of this.  Wallis attributes it to Prince Rupert, but Hennessy says Philip Ronayne of Cork invented it.  I have discovered a possible connection.  Prince Rupert of the Rhine (1619-1682), nephew of Charles I, was a major military figure of his time, becoming commander-in-chief of Charles I's armies in the 1640s.  In 1648-1649, he was admiral of the King's fleet and was blockaded with 16 ships in Kinsale Harbor for 20 months.  Kinsale is about 20km south of Cork.

          Ronayne wrote an Algebra, of which only a second edition of 1727 is in the BL.  Schrek has investigated the family histories and says Ronayne lived in the early 18C.  This would seem to make him too young to have met Rupert.  Perhaps Rupert invented the problem while in Kinsale and this was conveyed to Ronayne some years later.  Does anyone know the dates of Ronayne or of the 1st ed (Schrek only located the BL example of the 2nd ed)?  I cannot find anything on him in Wallis, May, Poggendorff, DNB, but Google has turned up a reference to a 1917 history of the family which Schrek cites, but I have not yet tried to find this.

          Hennessy's article says a little about Daniel Voster and details are in Wallis's .  His father, Elias (1682 ‑ >1728) wrote an Arithmetic, of which Wallis lists 30 editions.  The BL lists one as late as 1829.  The son, Daniel (1705 ‑ >1760) was a schoolmaster and instrument maker who edited later versions of his father's arithmetic.  The 1750 History of Cork quoted by Hennessy says the author had seen the cubes with Daniel.  Hennessy conjectures that his example was made specially, perhaps under the direction of a mathematician.  It seems likely that Daniel knew Ronayne and made this example for him.

 

John Wallis.  Perforatio cubi, alterum ipsi aequalem recipiens.  (De Algebra Tractatus; 1685; Chap. 109)  = Opera Mathematica, vol. II, Oxford, 1693, pp. 470‑471, ??NYS.  Cites Rupert as the source of the equal cube version.  (Latin and English in Schrek.)  Scriba, below, found an errata slip in Wallis's copy of his Algebra in the Bodleian.  This corrects the calculations, but was published in the Opera, p. 695.

Ozanam‑Montucla.  1778.  Percer un cube d'une ouverture, par laquelle peut passer un autre cube égal au premier.  Prob. 30 & fig. 53, plate 7, 1778: 319-320;  1803: 315-316;  1814: 268-269.  Prob. 29, 1840: 137.  Equal cubes with diagonal movement.

J. H. van Swinden.  Grondbeginsels der Meetkunde.  2nd ed., Amsterdam, 1816, pp. 512‑513, ??NYS.  German edition by C. F. A. Jacobi, as:  Elemente der Geometrie, Friedrich Frommann, Jena, 1834, pp. 394-395.  Cites Rupert and Wallis and gives a simple construction, saying Nieuwland has found the largest cube which can pass through a cube.

Peter Nieuwland.  (Finding of maximum cube which passes through another).  In:  van Swinden, op. cit., pp. 608‑610;  van Swinden‑Jacobi, op. cit. above, pp. 542-544, gives Nieuwland's proof.

Cundy and Rollett, p. 158, give references to Zacharias (see below) and to Cantor, but Cantor only cites Hennessy.

H. Hennessy.  Ronayne's cubes.  Phil. Mag. (5) 39 (Jan‑Jun 1895) 183‑187.  Quotes, from Gibson's 'History of Cork', a passage taken from Smith's 'History of Cork', 1st ed., 1750, vol. 1, p. 172, saying that Philip Ronayne had invented this and that a Daniel Voster had made an example, which may be the example owned by Hennessy.  He gives no reference to Rupert.  He finds the dimensions.

F. Koch & I. Reisacher.  Die Aufgabe, einen Würfel durch einen andern durchzuschieben.  Archiv Math. Physik (3) 10 (1906) 335‑336.  Brief solution of Nieuwland's problem.

M. Zacharias.  Elementargeometrie und elementare nicht-Euklidische Geometrie in synthetischer Behandlung.  Encyklopädie der Mathematischen Wissenschaften.  Band III, Teil 1, 2te Hälfte.  Teubner, Leipzig, 1914-1931.  Abt. 28: Maxima und Minima.  Die isoperimetrische Aufgabe.  Pp. 1133-1134.  Attributes it to Prince Rupert, following van Swinden.  Cites Wallis & Ronayne, via Cantor, and Nieuwland, via van Swinden.

U. Graf.  Die Durchbohrung eines Würfels mit einem Würfel.  Zeitschrift math. naturwiss. Unterricht 72 (1941) 117.  Nice photos of a model made at the Technische Hochschule Danzig.  Larger and better versions of the same photos can be found in: W. Lietzmann & U. Graf; Mathematik in Erziehung und Unterricht; Quelle & Meyer, Leipzig, 1941, vol. 2, plate 3, opp. p. 168, but I can't find any associated text for it.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 12: Curios [sic] cubes, p. 14.  First says it can be done with equal cubes and then a larger can pass through a smaller.  Claims that the larger cube can be about  1.1,  but this is due to an error -- he thinks the hexagon has the same diameter as the cube itself.

H. D. Grossman, proposer;  C. S. Ogilvy & F. Bagemihl, solvers.  Problem E888 -- Passing a cube through a cube of same size.  AMM 56 (1949) 632 ??NYS  &  57 (1950) 339.  Only considers cubes of the same size, though Bagemihl's solution permits a slightly larger cube.  No references.

D. J. E. Schrek.  Prince Rupert's problem and its extension by Pieter Nieuwland.  SM 16 (1950) 73‑80 & 261‑267.  Historical survey, discussing Rupert, Wallis, Ronayne, van Swinden & Nieuwland.  Says Ronayne is early 18C.

M. Gardner.  SA (Nov 1966) = Carnival, pp. 41‑54.  The largest square inscribable in a cube is the cross section of the maximal hole through which another cube can pass.

Christoph J. Scriba.  Das Problem des Prinzen Ruprecht von der Pfalz.  Praxis der Mathematik 10 (1968) 241-246.  ??NYS -- described by Scriba in an email to HM Mailing List, 20 Aug 1999.  Describes the correction to Wallis's work and considers the problem for the tetrahedron and octahedron.

 

          6.P.    GEOMETRICAL VANISHING

 

Gardner.  MM&M.  1956.  Chap. 7 & 8: Geometrical Vanishing -- Parts I & II, pp. 114‑155.  Best extensive discussion of the subject and its history.

Gardner.  SA (Jan 1963) c= Magic Numbers, chap. 3.  Discusses application to making an extra bill and Magic Numbers adds citations to several examples of people trying it and going to jail.

Gardner.  Advertising premiums to beguile the mind: classics by Sam Loyd, master puzzle‑poser.  SA (Nov 1971) = Wheels, Chap. 12.  Discusses several forms.

S&B, p. 144, shows several versions.

 

          6.P.1.           PARADOXICAL DISSECTIONS OF THE CHESSBOARD BASED ON FIBONACCI NUMBERS

 

          Area 63 version:  AWGL, Dexter, Escott, White, Loyd, Ahrens, Loyd Jr., Ransom. 

 

(W. Leybourn.  Pleasure with Profit.  1694.  ??  I cannot recall the source of this reference and think it may be an error.  I have examined the book and find nothing relevant in it.)

Loyd.  Cyclopedia, 1914, pp. 288 & 378.  8 x 8  to  5 x 13  and to an area of  63.  Asserts Loyd presented the first of these in 1858.  Cf Loyd Jr, below.

O. Schlömilch.  Ein geometrisches Paradoxon.  Z. Math. Phys., 13 (1868) 162.  8 x 8  to  5 x 13.  (This article is only signed Schl.  Weaver, below, says this is Schlömilch, and this seems right as he was a co‑editor at the time.  Coxeter (SM 19 (1953) 135‑143) says it is V. Schlegel, apparently confusing it with the article below.)  Doesn't give any explanation, leaving it as a student exercise.

F. J. Riecke.  Op. cit. in 4.A.1.  Vol. 3, 1873.  Art. 16: Ein geometrisches Paradoxon.  Quotes Schlömilch and explains the paradox.

G. H. Darwin.  Messenger of Mathematics 6 (1877) 87.  8 x 8  to  5 x 13  and generalizations.

V. Schlegel.  Verallgemeinerung eines geometrischen Paradoxons.  Z. Math. Phys. 24 (1879) 123‑128 & Plate I.  8 x 8  to  5 x 13  and generalizations.

Mittenzwey.  1880.  Prob. 299, pp. 54 & 105;  1895?: 332, pp. 58 & 106-107;  1917: 332, pp. 53 & 101.  8 x 8  to  5 x 13.  Clear explanation.

The Boy's Own Paper.  No. 109, vol. III (12 Feb 1881) 327.  A puzzle.  8 x 8  to  5 x 13  without answer.

Richard A. Proctor.  Some puzzles.  Knowledge 9 (Aug 1886) 305-306.  "We suppose all the readers ... know this old puzzle."  Describes and explains  8 x 8  to  5 x 13.  Gives a different method of cutting so that each rectangle has half the error -- several typographical errors.

Richard A. Proctor.  The sixty-four sixty-five puzzle.  Knowledge 9 (Oct 1886) 360-361.  Corrects the above and explains it in more detail.

Will Shortz has a puzzle trade card with the  8 x 8  to  5 x 13  version, c1889.

Ball.  MRE, 1st ed., 1892, pp. 34‑36.  8 x 8  to  5 x 13  and generalizations.  Cites Darwin and describes the examples in Ozanam-Hutton (see Ozanam-Montucla in 6.P.2).  In the 5th ed., 1911, p. 53, he changes the Darwin reference to Schlömilch.  In the 7th ed., 1917, he only cites the Ozanam-Hutton examples.

Clark.  Mental Nuts.  1897, no. 33;  1904, no. 41;  1916, no. 43.  Four peculiar drawings.  8 x 8  to  5 x 13.

Carroll-Collingwood.  1899.  Pp. 316-317 (Collins: 231 and/or 232 (lacking in my copy))  = Carroll-Wakeling II, prob. 7: A geometrical paradox, pp. 12 & 7.  8 x 8  to  5 x 13.  Carroll may have stated this as early as 1888.  Wakeling says the papers among which this was found on Carroll's death are now in the Parrish Collection at Princeton University and suggests Schlömilch as the earliest version.

AWGL (Paris).  L'Echiquier Fantastique.  c1900.  Wooden puzzle of  8 x 8  to  5 x 13  and to area  63.  ??NYS -- described in S&B, p. 144.

Walter Dexter.  Some postcard puzzles.  Boy's Own Paper (14 Dec 1901) 174‑175.  8 x 8  to  5 x 13  and to area  63.

C. A. Laisant.  Initiation Mathématique.  Georg, Geneva  &  Hachette, Paris, 1906.  Chap. 63: Un paradoxe:  64 = 65,  pp. 150-152.

Wm. F. White.  In the mazes of mathematics.  A series of perplexing puzzles.  III.  Geometric puzzles.  The Open Court 21 (1907) 241‑244.  Shows  8 x 8  to  5 x 13  and a two‑piece  11 x 13  to area  145.

E. B. Escott.  Geometric puzzles.  The Open Court 21 (1907) 502‑505.  Shows  8 x 8  to area  63  and discusses the connection with Fibonacci numbers.

William F. White.  Op. cit. in 5.E.  1908.  Geometric puzzles, pp. 109‑117.  Partly based on above two articles.  Gives  8 x 8  to  5 x 13  and to area  63.  Gives an extension which turns  12 x 12  into  8 x 18  and into area  144,  but turns  23 x 23  into  16 x 33  and into area  145.  Shows a puzzle of Loyd:  three‑piece  8 x 8  into  7 x 9.

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  5 x 5  into four pieces that make a  3 x 8.

M. Adams.  Indoor Games.  1912.  Is  64  equal to  65?  Pp. 345-346 with fig. on p. 344.

Loyd.  Cyclopedia.  1914.  See entry at 1858.

W. Ahrens.  Mathematische Spiele.  Teubner, Leipzig.  3rd ed., 1916, pp. 94‑95 & 111‑112.  The 4th ed., 1919, and 5th ed., 1927, are identical with the 3rd ed., but on different pages:  pp. 101‑102  &  pp. 118‑119.  Art. X.  65 = 64 = 63  gives  8 x 8  to  5 x 13  and to area  63.  The area  63  case does not appear in the 2nd ed., 1911, which has Art. V.  64 = 65,  pp. 107 & 118‑119 and this material is not in the 1st ed. of 1907.

Tom Tit??  In Knott, 1918, but I can't find it in Tom Tit.  No. 3: The square and the rectangle:  64 = 65!,  pp. 15-16.  Clearly explained.

Hummerston.  Fun, Mirth & Mystery.  1924.  A puzzling paradox, pp. 44 & 185.  Usual  8 x 8  to  5 x 13,  but he erases the chessboard lines except for the cells the cuts pass through, so one way has 16 cells, the other way has 17 cells.  Reasonable explanation.

Collins.  Book of Puzzles.  1927.  A paradoxical puzzle, pp. 4-5.  8 x 8  to  5 x 13.  Shades the unit cells that the lines pass through and sees that one way has 16 cells, the other way has 17 cells, but gives only a vague explanation.

Loyd Jr.  SLAHP.  1928.  A paradoxical puzzle, pp. 19‑20 & 90.  Gives  8 x 8  to  5 x 13.  "I have discovered a companion piece ..." and gives the  8 x 8  to area  63  version.  But cf AWGL, Dexter, etc. above.

W. Weaver.  Lewis Carroll and a geometrical paradox.  AMM 45 (1938) 234‑236.  Describes unpublished work in which Carroll obtained (in some way) the generalizations of the  8 x 8  to  5 x 13  in about 1890‑1893.  Weaver fills in the elementary missing arguments.

W. R. Ransom, proposer;  H. W. Eves, solver.  Problem E468.  AMM 48 (1941) 266  &  49 (1942) 122‑123.  Generalization of the  8 x 8  to area  63  version.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 23: Summat for nowt?, pp. 27-28.  8 x 8  to  5 x 13,  clearly drawn.

Warren Weaver.  Lewis Carroll: Mathematician.  Op. cit. in 1.  1956.  Brief mention of  8 x 8  to  5 x 13.  John B. Irwin's letter gives generalizations to other consecutive triples of Fibonacci numbers (though he doesn't call them that).  Weaver's response cites his 1938 article, above.

 

          6.P.2. OTHER TYPES

 

          In several early examples, the authors appear unaware that area has vanished!

 

Pacioli.  De Viribus.  c1500.  Ff. 189v - 191r.  Part 2.  LXXIX. Do(cumento). un tetragono saper lo longare con restregnerlo elargarlo con scortarlo (a tetragon knows lengthening and contraction, enlarging with shortening ??)  = Peirani 250-252.  Convert a  4 x 24  rectangle to a  3 x 32  using one cut into two pieces.  Pacioli's

          description is cryptic but seems to have two cuts, making                                                 d       c

          three pieces.  There is a diagram at the bottom of f. 190v, badly                                  k         f  e

          redrawn on Peirani 458.  Below this is a inserted note which Peirani

          252 simply mentions as difficult to read, but can make sense.  The                                   g          

          points are as laid out at the right.  abcd  is the original  4 x 24                                       h  a    o  b

          rectangle.  g  is one unit up from  a  and  e  is one unit down from  c.

          Cut from  c  to  g  and from  e  parallel to the base, meeting  cg  at  f.  Then move  cdg  to  fkh  and move  fec  to  hag.  Careful rereading of Pacioli seems to show he is using a trick!  He cuts from  e  to  f  to  g.  then turns over the upper piece and slides it along so that he can continue his cut from  g  to  h,  which is where  f  to  c  is now.  This gives three pieces from a single cut!  Pacioli clearly notes that the area is conserved.

                    Although not really in this topic, I have put it here as it seems to be a predecessor of this topic and of 6.AY.

Sebastiano Serlio.  Libro Primo d'Architettura.  1545.  This is the first part of his Architettura, 5 books, 1537-1547, first published together in 1584.  I have seen the following editions.

                    With French translation by Jehan Martin, no publisher shown, Paris, 1545, f. 22.r.  ??NX

                    1559.  F. 15.v.

                    Francesco Senese & Zuane(?) Krugher, Venice, 1566, f. 16.r.  ??NX

                    Jacomo de'Franceschi, Venice, 1619, f. 16.r.

                    Translated into Dutch by Pieter Coecke van Aelst as:  Den eerstē vijfsten boeck van architecturē; Amsterdam, 1606.  This was translated into English as:  The Five Books of Architecture; Simon Stafford, London, 1611  = Dover, 1982.  The first Booke, f. 12v.

                    3 x 10  board is cut on a diagonal and slid to form a  4 x 7  table with  3 x 1  left over, but he doesn't actually put the two leftover pieces together nor notice the area change!

Pietro Cataneo.  L'Architettura di Pietro Cataneo Senese.  Aldus, Venice, 1567.  ??NX.  Libro Settimo.

P. 164, prop. XXVIIII: Come si possa accresciere una stravagante larghezza.  Gives a correct version of Serlio's process.

P. 165, prop. XXX: Falsa solutione del Serlio.  Cites p. xxii of Serlio.  Carefully explains the error in Serlio and says his method is "insolubile, & mal pensata".

Schwenter.  1636.  Part 15, ex. 14, p. 541: Mit einem länglichten schmahlen Brett / für  ein bräites Fenster einen Laden zu machen.  Cites Gualtherus Rivius, Architectur.  Discusses Serlio's dissection as a way of making a  4 x 7  from a  3 x 10  but doesn't notice the area change.

Gaspar Schott.  Magia Universalis.  Joh. Martin Schönwetter, Bamberg, Vol. 3, 1677.  Pp. 704-708 describes Serlio's error in detail, citing Serlio.  ??NX of plates.

I have a vague reference to the 1723 ed. of Ozanam, but I have not seen it in the 1725 ed. -- this may be an error for the 1778 ed. below.

Minguet.  1755.  Pp. not noted -- ??check (1822: 145-146; 1864: 127-128).  Same as Hooper.  Not in 1733 ed.

Vyse.  Tutor's Guide.  1771?  Prob. 8, 1793: p. 304, 1799: p. 317 & Key p. 358.  Lady has a table  27  square and a board  12 x 48.  She cuts the board into two  12 x 24  rectangles and cuts each rectangle along a diagonal.  By placing the diagonals of these pieces on the sides of her table, she makes a table  36  square.  Note that  362 = 1296  and  272 + 12 x 48  =  1305.  Vyse is clearly unaware that area has been created.  By dividing all lengths by  3,  one gets a version where one unit of area is lost.  Note that  4, 8, 9  is almost a Pythagorean triple.

William Hooper.  Rational Recreations.  1774.  Op. cit. in 4.A.1.  Vol. 4, pp. 286‑287: Recreation CVI -- The geometric money.  3 x 10  cut into four pieces which make a  2 x 6  and a  4 x 5.  (The diagram is shown in Gardner, MM&M, pp. 131‑132.)  (I recently saw that an edition erroneously has a  3 x 6  instead of a  2 x 6  rectangle.  This must be the 1st ed. of 1774, as it is correct in my 2nd ed. of 1782.)

Ozanam-Montucla.  1778.  Transposition de laquelle semble résulter que le tout peut être égal à la partie.  Prob. 21 & fig. 127, plate 16, 1778: 302-303 & 363;  1803: 298-299 & 361;  1814: 256 & 306;  1840: omitted.   3 x 11  to  2 x 7  and  4 x 5.  Remarks that M. Ligier probably made some such mistake in showing  172 = 2 x 122  and this is discussed further on the later page.

E. C. Guyot.  Nouvelles Récréations Physiques et Mathématiques.  Nouvelle éd.  La Librairie, Rue S. André‑des‑Arcs[sic], Paris, Year 7 [1799].  Vol. 2, Deuxième récréation: Or géométrique -- construction, pp. 41‑42 & plate 6, opp. p. 37.  Same as Hooper.

Manuel des Sorciers.  1825.  Pp. 202-203, art. 19.  ??NX  Same as Hooper.

The Boy's Own Book.  The geometrical money.  1828: 413;  1828-2: 419;  1829 (US): 212;  1855: 566‑567;  1868: 669.  Same as Hooper.

Magician's Own Book.  1857.  Deceptive vision, pp. 258-259.  Same as Hooper.  = Book of 500 Puzzles, 1859, pp. 72-73.

Illustrated Boy's Own Treasury.  1860.  Optics: Deceptive vision, p. 445.  Same as Hooper.  Identical to Book of 500 Puzzles.

Wemple & Company (New York).  The Magic Egg Puzzle.  ©1880.  S&B, p. 144.  Advertising card, the size of a small postcard, but with ads for Rogers Peet on the back.  Starts with 9 eggs.  Cut into four rectangles and reassemble to make 6, 7, 8, 10, 11, 12  eggs.

R. March & Co. (St. James's Walk, Clerkenwell).   'The Magical Egg Puzzle', nd [c1890].  (I have a photocopy.)  Four rectangles which produce  6, 7, ..., 12  eggs.  Identical to the Wemple version, but with Wemple's name removed.  I only have a photocopy of the front of this and I don't know what's on the back. I also have a photocopy of the instructions.

Loyd.  US Patent 563,778 -- Transformation Picture.  Applied: 11 Mar 1896;  patented: 14 Jul 1896.  1p + 1p diagrams.  Simple rotating version using 8 to 7 objects.

Loyd.  Get Off the Earth.  Puzzle notices in the Brooklyn Daily Eagle (26 Apr ‑ 3 May 1896), printing individual Chinamen.  Presenting all of these at an office of the newspaper gets you an example of the puzzle.  Loyd ran discussions on it in his Sunday columns until 3 Jan 1897 and he also sold many versions as advertising promotions.  S&B, p. 144, shows several versions.

Loyd.  Problem 17: Ye castle donjon.  Tit‑Bits 31 (6  &  27 Feb  &  6  &  20 Mar 1897) 343,  401,  419  &  455.  = Cyclopedia, 1914, The architect's puzzle, pp. 241 & 372.  5 x 25  to area 124.

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  Discusses and shows Get Off the Earth.

Ball.  MRE, 4th ed., 1905, pp. 50-51: Turton's seventy-seven puzzle.  Additional section describing Captain Turton's  7 x 11  to  7 x 11  with one projecting square, using bevelled cuts.  This is dropped from the 7th ed., 1917.

William F. White.  1907 & 1908.  See entries in 6.P.1.

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  Gives "Get Off the Earth" on p. 785.

Loyd.  Teddy and the Lions.  Gardner, MM&M, p. 123, says he has seen only one example, made as a promotional item for the Eden Musee in Manhattan.  This has a round disc, but two sets of figures -- 7 natives and 7 lions which become 6 natives and 8 lions.

Dudeney.  A chessboard fallacy.  The Paradox Party.  Strand Mag. 38 (No. 228) (Dec 1909) 676 (= AM, prob. 413, pp. 141 & 247).  (There is a solution in Strand Mag. 39 (No. 229) (Jan 1910) ??NYS.)  8 x 8  into 3 pieces which make a  9 x 7.

Fun's Great Baseball Puzzle.  Will Shortz gave this out at IPP10, 1989, as a colour photocopy, 433 x 280 mm (approx. A3).  ©1912 by the Press Publishing Co (The New York World).  I don't know if Fun was their Sunday colour comic section or what.  One has to cut it diagonally and slide one part along to change from 8 to 9 boys.

Loyd.  The gold brick puzzle.  Cyclopedia, 1914, pp. 32 & 342 (= MPSL1, prob. 24, pp. 22 & 129).  24 x 24  to  23 x 25.

Loyd.  Cyclopedia.  1914.  "Get off the earth", p. 323.  Says over 10 million were sold.  Offers prizes for best answers received in 1909.

Loyd Jr.  SLAHP.  1928.  "Get off the Earth" puzzle, pp. 5‑6.  Says 'My "Missing Chinaman Puzzle"' of 1896.  Gives a simple and clear explanation.

John Barnard.  The Handy Boy's Book.  Ward, Lock & Co., London, nd [c1930?].  Some interesting optical illusions, pp. 310-311.  Shows a card with 11 matches and a diagonal cut so that sliding it one place makes 10 matches.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 24: A chessboard fallacy, pp. 28-29.  8 x 8  cut with a diagonal of a  8 x 7  region, then pieces slid and a triangle cut off and moved to the other end to make a  9 x 7.  Clear illustration.

Mel Stover.  From 1951, he devised a number of variations of both Get off the Earth (perhaps the best is his Vanishing Leprechaun) and of Teddy and the Lions (6 men and 4 glasses of beer become 5 men and 5 glasses).  I have examples of some of these from Stover and I have looked at his notebooks, which are now with Mark Setteducati.  See  Gardner, MM&M, pp. 125-128.

Gardner.  SA (May 1961)  c= NMD, chap. 11.  Mentioned in Workout, chap. 27.  Describes his adaptation of a principle of Paul Curry to produce The Disappearing Square puzzle, where 16 or 17 pieces seem to make the same square.  The central part of the 17 piece version consists of five equal squares in the form of a Greek cross.  The central part of the 16 piece version has four of the squares in the shape of a square.  This has since been produced in several places.

Ripley's Puzzles and Games.  1966.  P. 60.  Asserts that when you cut a  2½ x 4½  board into six right triangles with legs  1½ and 2½,  then they assemble into an equilateral triangle of edge 5.  This has an area loss of about 4%.

John Fisher.  John Fisher's Magic Book.  Muller, London, 1968.

Financial Wizardry, pp. 18-19.  7 x 8  region with  £  signs marking the area.  A line cuts off a triangle of width 7 and height 2 at the top.  The rest of the area is divided by a vertical into strips of widths 4 and 3, with a small rectangle 3 by 1 cut from the bottom of the width 3 strip.  When the strips are exchanged, one unit of area is lost and one  £  sign has vanished.

Try-Angle, pp. 126-127.  This is one of Curry's triangles -- see Gardner, MM&M, p. 147.

Alco-Frolic!, pp. 148-149.  This is a form of Stover's  6 & 4  to  5 & 5  version.

D. E. Knuth.  Disappearances.  In:  The Mathematical Gardner; ed. by David Klarner; Prindle, Weber & Schmidt/Wadsworth, 1981.  P. 264.  An eight line poem which rearranges to a seven line poem.

Dean Clark.  A centennial tribute to Sam Loyd.  CMJ 23:5 (Nov 1992) 402‑404.  Gives an easy circular version with  11 & 12  astronauts around the earth and a  15 & 16  face version with three pieces, a bit like the Vanishing Leprechaun.

 

          6.Q.   KNOTTING A STRIP TO MAKE A REGULAR PENTAGON

 

Urbano d'Aviso.  Trattato della Sfera e Pratiche per Uso di Essa.  Col modeo di fare la figura celeste, opera cavata dalli manoscritti del. P. Bonaventura Cavalieri.  Rome, 1682.  ??NYS cited by Lucas (1895) and Fourrey.

Dictionary of Representative Crests.  Nihon Seishi Monshō Sōran (A Comprehensive Survey of Names and Crests in Japan),  Special issue of  Rekishi Dokuhon (Readings in History),  Shin Jinbutsu Oraisha, Tokyo, 1989, pp. 271-484.  Photocopies of relevant pages kindly sent by Takao Hayashi. 

                    Crests 3504 and 3506 clearly show a strip knotted to make a pentagon.  3507 has two such knots and 3508 has five.  I don't know the dates, but most of these crests are several centuries old.

Lucas.  RM2, 1883, pp. 202‑203.

Tom Tit.

Vol. 2, 1892.  L'Étoile à cinq branches, pp. 153-154.  = K, no. 5: The pentagon and the five pointed star, pp. 20‑21.  He adds that folding over the free end and holding the knot up to the light shows the pentagram.

Vol. 3, 1893.  Construire d'un coup de poing un hexagone régulier, pp. 159-161.  = K, no. 17: To construct a hexagon by finger pressure, pp. 49‑51.  Pressing an appropriate size Möbius strip flat gives a regular hexagon.

Vol. 3, 1893.  Les sept pentagones, pp. 165-166.  = K, no. 19: The seven pentagons, pp. 54‑55.  By tying five pentagons in a strip, one gets a larger pentagon with a pentagonal hole in the middle.

Somerville Gibney.  So simple!  The hexagon, the enlarged ring, and the handcuffs.  The Boy's Own Paper 20 (No. 1012) (4 Jun 1898) 573-574.  As in Tom Tit, vol. 3, pp. 159-161. 

Lucas.  L'Arithmétique Amusante.  1895.  Note IV:  Section II: Les Jeux de Ruban, Nos. 1 & 2:  Le nœud de cravate  &  Le nœud marin, pp. 220-222.  Cites d'Aviso and says he does both the pentagonal and hexagonal knots, but Lucas only shows the pentagonal one. 

E. Fourrey.  Procédés Originaux de Constructions Géométriques.  Vuibert, Paris, 1924.  Pp. 113 & 135‑139.  Cites Lucas and cites d'Aviso as Traité de la Sphère and says he gives the pentagonal and hexagonal knots.  Fourrey shows and describes both, also giving the pictures on his title page.

F. V. Morley.  A note on knots.  AMM 31 (1924) 237-239.  Cites Knott's translation of Tom Tit.  Says the process generalizes to  (2n+3)‑gons by using  n  loops.  Gets even-gons by using two strips.  Discusses using twisted strips.

Robert C. Yates.  Geometrical Tools.  (As:  Tools; Baton Rouge, 1941);  revised ed., Educational Publishers, St. Louis, 1949.  Pp. 64-65 gives square (a bit trivial), pentagon, hexagon, heptagon and octagon.  Even case need two strips.

Donovan A. Johnson.  Paper Folding for the Mathematics Class.  NCTM, 1957, pp. 16-17: Polygons constructed by tying paper knots.  Shows how to tie square, pentagon, hexagon, heptagon and octagon.

James K. Brunton.  Polygonal knots.  MG 45 (No. 354) (Dec 1961) 299‑302.  All regular  n‑gons,  n > 4,  can be obtained, except  n = 6  which needs two strips.  Discusses which can be made without central holes.

Marius Cleyet-Michaud.  Le Nombre d'Or.  Presses Universitaires de France, Paris, 1973.  On pp. 47-48, he calls this the 'golden knot' (Le "nœud doré") and describes how to make it.

 

          6.R.   GEOMETRIC FALLACIES

 

          General surveys of such fallacies can be found in the following.  See also:  6.P, 10.A.1.

          These fallacies are actually quite profound as the first two point out some major gaps in Euclid's axioms -- the idea of a point being inside a triangle really requires notions of order of points on a line and even the idea of continuity, i.e. the idea of real numbers.

 

Ball.  MRE.  1st ed., 1892, pp. 31‑34, two examples, discussed below.  3rd ed., 1896, pp. 39‑46  =  4th ed., 1905, pp. 41-48, seven examples.  5th ed., 1911, pp. 44-52  =  11th ed., 1939, pp. 76-84, nine example.

Walther Lietzmann.  Wo steckt der Fehler?  Teubner, Stuttgart, (1950), 3rd ed., 1953.  (Strens/Guy has 3rd ed., 1963(?).)  (There are  2nd ed, 1952??;  5th ed, 1969;  6th ed, 1972.  MG 54 (1970) 182 says the 5th ed appears to be unchanged from the 3rd ed.)  Chap. B: V, pp. 87-99 has 18 examples. 

                    (An earlier version of the book, by Lietzmann & Trier, appeared in 1913, with 2nd ed. in 1917.  The 3rd ed. of 1923 was divided into two books:  Wo Steckt der Fehler?  and  Trugschlüsse.  There was a 4th ed. in 1937.  The relevant material would be in Trugschlüsse, but I have not seen any of the relevant books, though E. P. Northrop cites Lietzmann, 1923, three times -- ??NYS.)

E. P. Northrop.  Riddles in Mathematics.  1944.  Chap. 6,  1944: 97-116, 232-236 & 249-250;  1945: 91-109, 215-219 & 230-231;  1961: 98-115, 216-219 & 229.  Cites Ball, Lietzmann (1923), and some other individual items.

V. M. Bradis,  V. L. Minkovskii  &  A. K. Kharcheva.  Lapses in Mathematical Reasoning.  (As: [Oshibki v Matematicheskikh Rassuzhdeniyakh], 2nd ed, Uchpedgiz, Moscow, 1959.)  Translated by J. J. Schorr-Kon, ed. by E. A. Maxwell.  Pergamon & Macmillan, NY, 1963.  Chap. IV, pp. 123-176.  20 examples plus six discussions of other examples.

Edwin Arthur Maxwell.  Fallacies in Mathematics.  CUP, (1959), 3rd ptg., 1969.  Chaps. II-V, pp. 13-36, are on geometric fallacies.

Ya. S. Dubnov.  Mistakes in Geometric Proofs.  (2nd ed., Moscow?, 1955).  Translated by Alfred K. Henn & Olga A. Titelbaum.  Heath, 1963.  Chap 1-2, pp. 5-33.  10 examples.

А. Г. Конфорович. [A. G. Konforovich].  (Математичні Софізми і Парадокси [Matematichnī Sofīzmi ī Paradoksi] (In Ukrainian).  Радянська Школа [Radyans'ka Shkola], Kiev, 1983.)  Translated into German by Galina & Holger Stephan as: Konforowitsch, Andrej Grigorjewitsch; Logischen Katastrophen auf der Spur – Mathematische Sophismen und Paradoxa; Fachbuchverlag, Leipzig, 1990.  Chap. 4: Geometrie, pp. 102-189 has 68 examples, ranging from the type considered here up through fractals and pathological curves.

S. L. Tabachnikov.  Errors in geometrical proofs.  Quantum 9:2 (Nov/Dec 1998) 37-39 & 49.  Shows:  every triangle is isosceles (6.R.1);  the sum of the angles of a triangle is 180o without use of the parallel postulate;  a rectangle inscribed in a square is a square;  certain approaching lines never meet (6.R.3);  all circles have the same circumference (cf Aristotle's Wheel Paradox in 10.A.1);  the circumference of a wheel is twice its radius;  the area of a sphere of radius  R  is  π2R2.

 

          6.R.1. EVERY TRIANGLE IS ISOSCELES

 

This is sometimes claimed to have been in Euclid's lost Pseudaria (Fallacies).

Ball.  MRE, 1st ed., 1892, pp. 33‑34.  On p. 32, Ball refers to Euclid's lost Fallacies and presents this fallacy and the one in 6.R.2:  "I do not know whether either of them has been published previously."  In the 3rd ed., 1896, pp. 42-43, he adds the heading:  To prove that every triangle is isosceles.  In the 5th ed., 1911, p. 45, he adds a note that he believes these two were first published in his 1st ed. and notes that Carroll was fascinated by them and they appear in The Lewis Carroll Picture Book (= Carroll-Collingwood) -- see below.

Mathesis (1893).  ??NYS.  [Cited by Fourrey, Curiosities Geometriques, p. 145.  Possibly Mathesis (2) 3 (Oct 1893) 224, cited by Ball in MRE, 3rd ed, 1896, pp. 44-45, cf in Section 6.R.4.]

Carroll-Collingwood.  1899.  Pp. 264-265 (Collins: 190-191).  = Carroll-Wakeling II, prob. 27: Every triangle has a pair of equal sides!, pp. 43 & 27.   Every triangle is isosceles.  Carroll may have stated this as early as 1888.  Wakeling's solution just suggests making an accurate drawing.  Carroll-Gardner, p. 65, mentions this and says it was not original with Carroll.

Ahrens.  Mathematische Spiele.  Teubner.  Alle Dreiecke sind gleichschenklige.  2nd ed., 1911, chap. X, art. VI, pp. 108 & 119‑120.  3rd ed., 1916, chap. IX, art. IX, pp. 92-93 & 109-111.  4th ed., 1919  &  5th ed., 1927, chap IX, art. IX, pp. 99‑101 & 116‑118.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  Call Mr. Euclid -- No. 15: To prove all triangles are equilateral, pp. 16-17.  Clear exposition of the fallacy.

See Read in 6.R.4 for a different proof of this fallacy.

 

          6.R.2. A RIGHT ANGLE IS OBTUSE

 

Ball.  MRE, 1st ed., 1892, pp. 32‑33.  See 6.R.1.  In the 3rd ed., 1896, pp. 40-41, he adds the heading:  To prove that a right angle is equal to an angle which is greater than a right angle.

Mittenzwey.  1895?.  Prob. 331, pp. 58 & 106;  1917: 331, pp. 53 & 101.

Carroll-Collingwood.  1899.  Pp. 266‑267 (Collins 191-192).  An obtuse angle is sometimes equal to a right angle.  Carroll-Gardner, p. 65, mentions this and says it was not original with Carroll.

H. E. Licks.  1917.  Op. cit. in 5.A.  Art. 82, p. 56.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  Call Mr. Euclid -- No. 16: To prove one right angle greater than another right angle, pp. 18-19.  "Here again, if you take the trouble to draw an accurate diagram, you will find that the "construction" used for the alleged proof is impossible."

E. A. Maxwell.  Note 2121:  That every angle is a right angle.  MG 34 (No. 307) (Feb 1950) 56‑57.  Detailed demonstration of the error.

 

          6.R.3. LINES APPROACHING BUT NOT MEETING

 

Proclus.  5C.  A Commentary on the First Book of Euclid's Elements.  Translated by Glenn R. Morrow.  Princeton Univ. Press, 1970.  Pp. 289-291.  Gives the argument and tries to refute it.

van Etten/Henrion/Mydorge.  1630.  Part 2, prob. 7: Mener une ligne laquelle aura inclination à une autre ligne, & ne concurrera jamais contre l'Axiome des paralelles, pp. 13‑14.

Schwenter.  1636.  To be added.

Ozanam-Montucla.  1778.  Paradoxe géométrique des lignes ....  Prob. 70 & fig. 116-117, plate 13, 1778: 405-407;  1803: 411-413;  1814: 348-350.  Prob. 69, 1840: 180-181.  Notes that these arguments really produce a hyperbola and a conchoid.  Hutton adds that a great many other examples might be found.

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 209-211 & 239;  1945: 195‑197 & 222;  1961: 197‑198 & 222.  Gives the 'proof' and its fallacy, with a footnote on p. 253 (1945: 234; 1961: 233) saying the argument "has been attributed to Proclus."

Jeremy Gray.  Ideas of Space.  OUP, 1979.  Pp. 37-39 discusses Proclus' arguments in the context of attempts to prove the parallel postulate.

 

          6.R.4. OTHERS

 

Ball.  MRE, 3rd ed, 1896, pp. 44-45.  To prove that, if two opposite sides of a quadrilateral are equal, the other two sides must be parallel.  Cites Mathesis (2) 3 (Oct 1893) 224 -- ??NYS

Cecil B. Read.  Mathematical fallacies  &  More mathematical fallacies.  SSM 33 (1933) 575‑589  &  977-983.  There are two perpendiculars from a point to a line.  Part of a line is equal to the whole line.  Every triangle is isosceles (uses trigonometry).  Angle trisection (uses a marked straightedge).

P. Halsey.  Class Room Note 40:  The ambiguous case.  MG 43 (No. 345) (Oct 1959) 204‑205.  Quadrilateral  ABCD  with  angle A = angle C  and  AB = CD.  Is this a parallelogram?

 

          6.S.    TANGRAMS, ET AL.

 

          GENERAL HISTORIES.

 

Hoffmann.  1893.  Chap III, pp. 74‑90, 96-97, 111-124 & 128 = Hoffmann-Hordern, pp. 62‑79 & 86-87 with several photos.  Describes Tangrams and Richter puzzles at some length.  Lots of photos in Hordern.  Photos on pp. 67, 71, 75, 87 show Richter's: Anchor (1890‑1900, = Tangram), Tormentor (1898), Pythagoras (1892), Cross Puzzle (1892), Circular Puzzle (1891), Star Puzzle (1899), Caricature (1890-1900, = Tangram) and four non-Richter Tangrams in Tunbridge ware, ivory, mother-of-pearl and tortoise shell.  Hordern Collection, pp. 45-57 & 60, (photos on pp. 46, 49, 50, 52, 54, 56, 60) shows different Richter versions of Tormentor (1880-1900), Pythagoras (1880-1900), Circular Puzzle (1880-1900), Star Puzzle (1880‑1900) and has a wood non-Richter version instead of the ivory version in the last photo.

Ronald C. Read.  Tangrams -- 330 Puzzles.  Dover, 1965.  The Introduction, pp. 1-6, is a sketch of the history.  Will Shortz says this is the first serious attempt to counteract the mythology created by Loyd and passed on by Dudeney.  Read cannot get back before the early 1800s and notes that most of the Loyd myth is historically unreasonable.  However, Read does not pursue the early 1800s history in depth and I consider van der Waals to be the first really serious attempt at a history of the subject.

Peter van Note.  Introduction.  IN: Sam Loyd; The Eighth Book of Tan; (Loyd & Co., 1903); Dover, 1968, pp. v-viii.  Brief debunking of the Loyd myth.

Jan van der Waals.  History  &  Bibliography.  In:  Joost Elffers; Tangram; (1973), Penguin, 1976.  Pp. 9‑27  &  29‑31.  Says the Chinese term "ch'i ch'ae" dates from the Chu era (‑740/‑330), but the earliest known Chinese book is 1813.  The History reproduces many pages from early works.  The Bibliography cites 8 versions of 4 Chinese books (with locations!) from 1813 to 1826 and 18 Western books from 1805 to c1850.  The 1805, and several other references, now seem to be errors.

S&B.  1986.  Pp. 22‑33 discusses loculus of Archimedes, Chie no Ita, Tangrams and Richter puzzles.

Alberto Milano.  Due giochi di società dell'inizio dell'800.  Rassegna di Studi e di Notizie 23 (1999) 131-177.  [This is a publication by four museums in the Castello Sforzesco, Milan:  Raccolta delle Stampe Achille Bertarelli; Archivio Fotografico; Raccolte d'Arte Applicata; Museo degli Strumenti Musicali.  Photocopy from Jerry Slocum.]  This surveys early books on tangrams, some related puzzles and the game of bell and hammer, with many reproductions of TPs and problems.

Jerry Slocum.  The Tangram Book.  (With Jack Botermans, Dieter Gebhardt, Monica Ma, Xiaohe Ma, Harold Raizer, Dic Sonneveld and Carla van Spluntern.)  ©2001 (but the first publisher collapsed), Sterling, 2003.  This is the long awaited definitive history of the subject!  It will take me sometime to digest and summarize this, but a brief inspection shows that much of the material below needs revision!

 

          Recent research by Jerry Slocum, backed up by The Admired Chinese Puzzle, indicates that the introduction of tangrams into Europe was done by a person or persons in Lord Amherst's 1815-1817 embassy to China, which visited Napoleon on St. Helena on its return voyage.  If so, then the conjectural dating of several items below needs to be amended.  I have amended my discussion accordingly and marked such dates with ??.  Although watermarking of paper with the correct date was a legal requirement at the time, paper might have been stored for some time before it was printed on, so watermark dates only give a lower bound for the date of printing.  I have seen several further items dated 1817, but it is conceivable that some material may have been sent back to Europe or the US a few years earlier -- cf Lee.

 

          On 2 Nov 2003, I did the following brief summary of Slocum's work in a letter to an editor.  I've made a few corrections and added a citation to the following literature.

          Tangrams.  The history of this has now been definitely established in Jerry Slocum's new book: The Tangram Book;  ©2001 (but the first publisher collapsed), Sterling, 2003.  This history has been extremely difficult to unravel because Sam Loyd deliberately obfuscated it in 1903, claiming the puzzle went back to 2000 BC, because the only previous attempt at a history had many errors, and because much of the material doesn't survive, or only a few examples survive.  The history covers a wide range in both time and location, as evidenced by the presence of seven co-authors from several countries.

          Briefly, the puzzle, in the standard form, dates from about 1800, in China.  It is attributed to Yang-cho-chü-shih, but this is a pseudonym, meaning 'dim-witted recluse', and no copies of his work are known.  The oldest known example of the game is one dated 1802 in a museum near Philadelphia -- see Lee, below.  The oldest known book on the puzzle had a preface by Sang-hsia-ko [guest under the mulberry tree] dated June 1813 and a postscript by Pi-wu-chü-shih dated July 1813.  This is only known from a Japanese facsimile of it made in 1839.  This book was republished, with a book of solutions, in two editions in 1815 -- one with about four problems per page, the other with about eleven.  The latter version was the ancestor of many 19C books, both in China and the west.  Another 2 volume version appeared later in 1815.  Sang-hsia-ko explicitly says "The origin of the Tangram lies within the Pythagorean theorem". 

          In 1816, several ships brought copies of the eleven problems per page books to the US, England and Europe.  The first western publication of the puzzle is in early 1817 when J. Leuchars of 47 Piccadilly registered a copyright and advertised sets for sale.  But the craze was really set off by the publication of The Fashionable Chinese Puzzle and its Key by John and Edward Wallis and John Wallis Jr in March 1817.  This included a poem with a note that the game was "the favourite amusement of Ex-Emperor Napoleon".  This went through many printings, with some (possibly the first) versions having nicely coloured illustrations.  By the end of the year, there were many other books, including examples in France, Italy and the USA. 

          Dic Sonneveld, one of the co-authors of Slocum's book, managed to locate the tangram and books that had belonged to Napoleon in the Château de Malmaison, outside Paris, but there is no evidence that Napoleon spent much time playing with it.  St. Helena was a regular stop for ships in the China trade.  Napoleon is recorded as having bought a chess set from one ship and several notables are recorded as having presented Napoleon with gifts of Chinese objects.  A diplomatic letter of Jan 1817 records sending an example of the game from St. Helena to Prince Metternich, but this example has not been traced.

          The first American book was Chinese Philosophical and Mathematical Trangram by James Coxe, appearing in Philadelphia in August 1817.  The word 'trangram' meaning 'an odd, intricately contrived thing' according to Johnson's Dictionary, was essentially obsolete by 1817, but was still in some use in the US.  The earliest known use of the word 'tangram' is in Thomas Hill's Geometrical Puzzles for the Young, Boston, 1848.  One suspects that he was influenced by Coxe's book, but he may have known that 'T'ang' is the Cantonese word for 'Chinese'.  Hill later became President of Harvard University and was an active promoter and inventor of games for classroom use.  In 1864, the word was in Webster's Dictionary.

          However, the above is the story of the seven-piece tangram that we know today.  There is a long background to this, dating back to the 3rd century BC, when Archimedes wrote a letter to Eratosthenes describing a fourteen piece puzzle, known as the Stomachion or Loculus of Archimedes.  The few surviving texts are not very clear and there are two interpretations -- in one the standard arrangement of the pieces is a square and in the other it is a rectangle twice as wide as high.  There are six (at least) references to the puzzle in the classical world, the last being in the 6th century.  The puzzle was used to make a monstrous elephant, a brutal boar, a ship, a sword, etc., etc.  The puzzle then disappears, and no form of it appears in the Arabic world, which has always surprised me, given the Arabic interest in patterns.

          Further, several eastern predecessors of the tangrams are known.  The earliest is a Japanese version of 1742 by Ganriken (or Granreiken) which has seven pieces, attributed (as were many things) to Sei Shonagon, a 10th century courtesan famous for her ingenuity.  By the end of the 18th century, three other dissection/arrangement puzzles appeared in Japan, with 15, 19 and 19 pieces, including some semi-circles.  An 1804 print by Utamaro shows courtesans playing with some version of the puzzle -- only two copies of this print have been located.

          But the basic puzzle idea has its roots in Chinese approaches to the Theorem of Pythagoras and similar geometric proofs by dissection and rearrangement which date back to the 3rd century (and perhaps earlier).  But the tangram did not develop directly from these ideas.  From the 12th century, there was a Chinese tradition of making "Banquet Tables" in the form of several pieces that could be arranged in several ways.  The first known Chinese book on furniture, by Huang Po-ssu in 1194, describes a Banquet Table formed of seven rectangular pieces: two long, two medium and three short.  In 1617, Ko Shan described 'Butterfly Wing" tables with 13 pieces, including isosceles right triangles, right trapeziums and isosceles trapeziums.  In 1856, a Chinese scholar noted the resemblance of these tables with the tangram and a modern Chinese historian of mathematics has observed that half of the butterfly arrangement can be easily transformed into the tangrams.  No examples of these tables have survived, but tables (and serving dishes) in the tangram pattern exist and are probably still being made in China.

 

          SPECIFIC ITEMS

 

Kanchusen.  Wakoku Chiekurabe.  1727.  Pp. 9 & 28-29: a simple dissection puzzle with 8 pieces.  The problem appears to consist of a mitre comprising  ¾  of a unit square;  4 isosceles right triangles of hypotenuse 1  and  3 isosceles right triangles of side  ½,  but the solution shows that all the triangles are the same size, say having hypotenuse  1,  and the mitre shape is actually formed from a rectangle of size  1 x Ö2.

"Ganriken" [pseud., possibly of Fan Chu Sen].  Sei Shōnagon Chie-no-Ita (The Ingenious Pieces by Sei Shōnagon.) (In Japanese).  Kyoto Shobo, Aug 1742, 18pp, 42 problems and solutions.  Reproduced in a booklet, ed. by Kazuo Hanasaki, Tokyo, 1984, as pp. 19‑36.  Also reproduced in a booklet, transcribed into modern Japanese, with English pattern names and an English abstract, by Shigeo Takagi, 1989.  This uses a set of seven pieces different than the Tangram.  S&B, p. 22, shows these pieces.  Sei Shōnagon (c965-c1010) was a famous courtier, author of The Pillow Book and renowned for her intelligence.  The Introduction is signed Ganriken.  S&B say this is probably Fan Chu Sen, but Takagi says the author's real name is unknown.

Utamaro.  Interior of an Edo house, from the picture‑book:  The Edo Sparrows (or Chattering Guide), 1786.  Reproduced in B&W in:  J. Hillier; Utamaro -- Colour Prints and Paintings; Phaidon Press, Oxford, (1961), 2nd ed., 1979, p. 27, fig. 15.  I found this while hunting for the next item.  This shows two women contemplating some pieces but it is hard to tell if it is a tangram‑type puzzle, or if perhaps they are cakes.  Hiroko and Mike Dean tell me that they are indeed cooking cakes.

Utamaro.  Woodcut.  1792.  Shows two courtesans working on a tangram puzzle.  Van der Waals dated this as 1780, but Slocum has finally located it, though he has only been able to find two copies of it!  The courtesans are clearly doing a tangram-like puzzle with 12(?) pieces -- the pieces are a bit piled up and one must note that one of the courtesans is holding a piece.  They are looking at a sheet with 10 problem figures on it.

Early 19C books from China -- ??NYS -- cited by Needham, p. 111.

Jean Gordon Lee.  Philadelphians and the China Trade  1784-1844.  Philadelphia Museum of Art, 1984, pp. 122-124.  (Photocopy from Jerry Slocum.)  P. 124, item 102, is an ivory tangram in a cardboard box, inscribed on the bottom of the box: F. Waln April 4th 1802.  Robert Waln was a noted trader with China and this may have been a present for his third son Francis (1799‑1822).  This item is in the Ryerss Museum, a city museum in Philadelphia in the country house called Burholme which was built by one of Robert Waln's sons-in-law.

A New Invented Chinese Puzzle, Consisting of Seven Pieces of Ivory or Wood, Viz. 5 Triangles, 1 Rhomboid, & 1 Square, which may be so placed as to form the Figures represented in the plate.  Paine & Simpson, Boro'.  Undated, but the paper is watermarked 1806.  This consists of two 'volumes' of 8 pages each, comprising 159 problems with no solutions.  At the end are bound in a few more pages with additional problems drawn in -- these are direct copies of plates 21, 26, 22, 24, and 28 (with two repeats from plate 22) of The New and Fashionable Chinese Puzzle, 1817.  Bound in plain covers.  This is in Edward Hordern's collection and he provided a photocopy.  Dalgety also has a copy.

Ch'i Ch'iao t'u ho‑pi (= Qiqiao tu hebi) (Harmoniously combined book of tangram problems OR Seven clever pieces).  1813.  (Bibliothek Leiden 6891, with an 1815 edition at British Library 15257 d 13.)  van der Waals says it has 323 examples.  The 1813 seems to be the earliest Chinese tangram book of problems, with the 1815 being the solutions.  Slocum says there was a solution book in 1815 and that the problem book had a preface by  Sang‑hsia K'o  (= Sang-xia-ke),  which was repeated in the solution book with the same date.  Milano mentions this, citing Read and van der Waals/Elffers, and says an example is on the BL.  A version of this appears to have been the book given to Napoleon and to have started the tangram craze in Europe.  I have now received a photocopy from Peter Rasmussen & Wei Zhang which is copied from van der Waals' copy from BL 15257 d 13.  It has a cover, 6 preliminary pages and 28 plates with 318 problems.  The pages are larger than the photocopies of 1813/1815 versions in the BL that Slocum gave me, which have 334 problems on 86 pages, but I see these are from 15257 d 5 and 14.  I have a version of the smaller page format from c1820s which has 334 problems on 84pp, apparently lacking its first sheet.  The problems are not numbered, but given Chinese names.  They are identical to those appearing in Wallis's Fashionable Chinese Puzzle, below, except the pages are in different order, two pages are inverted, Wallis replaces Chinese names by western numbers and draws the figures a bit more accurately.  Wallis skips one number and adds four new problems to get 323 problems - van der Waals seems to have taken 323 from Wallis.

Shichi‑kou‑zu Gappeki [The Collection of Seven‑Piece Clever Figures].  Hobunkoku Publishing, Tokyo, 1881.  This is a Japanese translation of an 1813 Chinese book "recognized as the earliest of existing Tangram book", apparently the previous item.  [The book says 1803, but Jerry Slocum reports this is an error for 1813!]  Reprinted, with English annotations by Y. Katagiri, from N. Takashima's copy, 1989.  129 problems (but he counts 128 because he omits one after no. 124), all included in my version of the previous item, no solutions.

 

Anonymous.  A Grand Eastern Puzzle.  C. Davenporte & Co.  Registered on 24 Feb 1817, hence the second oldest English (and European?) tangram book [Slocum, p. 71.]  It is identical to Ch'i Ch'iao t'u ho‑pi, 1815, above, except that plates 25 and 27 have been interchanged.  It appears to be made by using Chinese pages and putting a board cover on it.  On the front cover is the only English text:

 

                                                                           A

 

                                                           Grand Eastern Puzzle

                                                                      ----------

 

                                        THE following Chineze Puzzle is recommended

                              to the Nobility, Gentry, and others, being superior to

                              any hitherto invented for the Amusement of the Juvenile

                              World, to whom it will afford unceasing recreation and

                              information; being formed on Geometrical principles, it

                              may not be considered as trifling to those of mature

                              years, exciting interest, because difficult and instructive,

                              imperceptibly leading the mind on to invention and per-

                              severence. -- The Puzzle consists of five triangles, a

                              square, and a rhomboid, which may be placed in upwards

                              of THREE HUNDRED and THIRTY Characters, greatly re-

                              sembling MEN, BEASTS, BIRDS, BOATS, BOTTLES, GLASS-

                              ES, URNS, &c.  The whole being the unwearied exertion

                              of many years study and application of one of the Lite-

                              rati of China, and is now offered to the Public for their

                              patronage and support.

 

                                                ENTERED AT STATIONERS HALL

                                                                          ----

                                                            Published and sold by

                                                        C. DAVENPORTE and Co.

                                            No. 20, Grafton Street, East Euston Square.

 

The Fashionable Chinese Puzzle.  Published by J. & E. Wallis, 42, Skinner Street and J. Wallis Junr, Marine Library, Sidmouth, nd [Mar 1817].  Photocopy from Jerry Slocum.  This has an illustrated cover, apparently a slip pasted onto the physical cover.  This shows a Chinese gentleman holding a scroll with the title.  There is a pagoda in the background, a bird hovering over the scroll and a small person in the foreground examining the scroll.  Slocum's copy has paper watermarked 1816.

PLUS 

A Key to the New and Fashionable Chinese Puzzle, Published by J. and E. Wallis, 42, Skinner Street, London, Wherein is explained the method of forming every Figure contained in That Pleasing Amusement.  Nd [Mar 1817].  Photocopy from the Bodleian Library, Oxford, catalogue number Jessel e.1176.  TP seems to made by pasting in the cover slip and has been bound in as a left hand page.  ALSO a photocopy from Jerry Slocum.  In the latter copy, the apparent TP appears to be a paste down on the cover.  The latter copy does not have the Stanzas mentioned below.  Slocum's copy has paper watermarked 1815; I didn't check this at the Bodleian.

                    NOTE.  This is quite a different book than The New and Fashionable Chinese Puzzle published by Goodrich in New York, 1817.

                    Bound in at the beginning of the Fashionable Chinese Puzzle and the Bodleian copy of the Key is:  Stanzas, Addressed to Messrs. Wallis, on the Ingenious Chinese Puzzle, Sold by them at the Juvenile Repository, 42, Skinner Street.  In the Key, this is on different paper than the rest of the booklet.  The Stanzas has a footnote referring to the ex-Emperor Napoleon as being in a debilitated state.  (Napoleon died in 1821, which probably led to the Bodleian catalogue's date of c1820 for the entire booklet - but see below.  Then follow 28 plates with 323 numbered figures (but number 204 is skipped), solved in the Key.  In the Bodleian copy of the Key, these are printed on stiff paper, on one side of each sheet, but arranged as facing pairs, like Chinese booklets.

                    [Philip A. H. Brown; London Publishers and Printers  c. 1800-1870; British Library, 1982, p. 212] says the Wallis firm is only known to have published under the imprint J. & E. Wallis during 1813 and Ruth Wallis showed me another source giving 1813?-1814.  This led me to believe that the booklets originally appeared in 1813 or 1814, but that later issues or some owner inserted the c1820 sheet of Stanzas, which was later bound in and led the Bodleian to date the whole booklet as c1820.  Ruth Wallis showed me a source that states that John Wallis (Jun.) set up independently of his father at 186 Strand in 1806 and later moved to Sidmouth.  Finding when he moved to Sidmouth might help date this publication more precisely, but it may be a later reissue.  However, Slocum has now found the book advertised in the London Times in Mar 1817 and says this is the earliest Western publication of tangrams, based on the 1813/1815 Chinese work.  Wallis also produced a second book of problems of his own invention and some copies seem to be coloured.

                    In AM, p. 43, Dudeney says he acquired the copy of  The Fashionable Chinese Puzzle  which had belonged to Lewis Carroll.  He says it was "Published by J. and E. Wallis, 42 Skinner Street, and J. Wallis, Jun., Marine Library, Sidmouth" and quotes the Napoleon footnote, so this copy had the Stanzas included.  This copy is not in the Strens Collection at Calgary which has some of Dudeney's papers. 

                    Van der Waals cites two other works titled  The Fashionable Chinese Puzzle.  An 1818 edition from A. T. Goodridge [sic], NY, is in the American Antiquarian Society Library (see below) and another, with no details given, is in the New York Public Library.  Could the latter be the Carroll/Dudeney copy?

                    Toole Stott 823 is a copy with the same title and imprint as the Carroll/Dudeney copy, but he dates it c1840.  This version is in two parts.  Part I has 1 leaf text + 26 col. plates -- it seems clear that col. means coloured, a feature that is not mentioned in any other description of this book -- perhaps these were hand-coloured by an owner.  Unfortunately, he doesn't give the number of puzzles.  I wonder if the last two plates are missing from this??  Part II has 1 leaf text + 32 col. plates, giving 252 additional figures.  The only copy cited was in the library of J. B. Findlay -- I have recently bought a copy of the Findlay sale catalogue, ??NYR.

                    Toole Stott 1309 is listed with the title: Stanzas, ....  J. & F. [sic] Wallis ... and Marine Library, Sidmouth, nd [c1815].  This has 1 leaf text and 28 plates of puzzles, so it appears that the Stanzas have been bound in and the original cover title slip is lost or was not recognised by Toole Stott.  The date of c1815 is clearly derived from the Napoleon footnote but 1817 would have been more reasonable, though this may be a later reissue.  Again only one copy is cited, in the library of Leslie Robert Cole.

                    Plates 1-28 are identical to plates 1-28 of The Admired Chinese Puzzle, but in different order.  The presence of the Chinese text in The Admired Chinese Puzzle made me think the Wallis version was later than it.

                    Comparison of the Bodleian booklet with the first 27 plates of Giuoco Cinese, 1818?, reveals strong similarities.  5 plates are essentially identical, 17 plates are identical except for one, two or three changes and 3 plates are about 50% identical.  I find that 264 of the 322 figures in the Wallis booklet occur in  Giuoco Cinese,  which is about 82%.  However, even when the plates are essentially identical, there are often small changes in the drawings or the layout.

                    Some of the plates were copied by hand into the Hordern Collection's copy of  A New Invented Chinese Puzzle, c1806??.

 

The Admired Chinese Puzzle  A New & Correct Edition From the Genuine Chinese Copy.  C. Taylor, Chester, nd [1817].  Paper is clearly watermarked 1812, but the Prologue refers to the book being brought from China by someone in Lord Amherst's embassy to China, which took place in 1815-1817 and which visited Napoleon on St. Helena on its return.  Slocum dates this to after 17 Aug 1817, when Amherst's mission returned to England and this seems to be the second western book on tangrams.  Not in Christopher, Hall, Heyl or Toole Stott -- Slocum says there is only one copy known in England!  It originally had a cover with an illustration of two Chinese, titled  The Chinese Puzzle,  and one of the men holds a scroll saying  To amuse and instruct.  The Chinese text gives the title  Ch'i ch'iao t'u ho pi (Harmoniously combined book of tangram problems).  I have a photocopy of the cover from Slocum.  Prologue facing TP; TP; two pp in Chinese, printed upside down, showing the pieces;  32pp of plates numbered at the upper left (sometimes with reversed numbers), with problems labelled in Chinese, but most of the characters are upside down!  The plates are printed with two facing plates alternating with two facing blank pages.  Plate 1 has 12 problems, with solution lines lightly indicated.  Plates 2 - 28 contain 310 problems.  Plates 29-32 contain 18 additional "caricature Designs" probably intended to be artistic versions of some of the abstract tangram figures.  The Prologue shows faint guide lines for the lettering, but these appear to be printed, so perhaps it was a quickly done copperplate.  The text of the Prologue is as follows.

                                        This ingenious geometrical Puzzle was introduced into this Kingdom from China.

                    The following sheets are a correct Copy from the Chinese Publication, brought to England by a Gentleman of high Rank in the suit [sic] of Lord Amherst's late Embassy.  To which are added caricature Designs as an illustration, every figure being emblematical of some Being or Article known to the Chinese.

          The plates are identical to the plates in The Fashionable Chinese Puzzle above, but in different order and plate 4 is inverted and this version is clearly upside down.

 

Sy Hall.  A New Chinese Puzzle,  The Above Consists of Seven Pieces of Ivory or Wood, viz. 5 Triangles, 1 Rhomboid, and 1 Square, which will form the 292 Characters, contained in this Book; Observing the Seven pieces must be used to form each Character.  NB.  This Edition has been corrected in all its angles, with great care and attention.  Engraved by Sy Hall, 14 Bury Street, Bloomsbury.  31 plates with 292 problems.  Slocum, the Hordern Collection and BL have copies.  I have a photocopy from a version from Slocum which has no date but is watermarked 1815.  Slocum's recent book [The Tangram Book, pp. 74-75] shows a version of the book with the publisher's name as James Izzard and a date of 1817.  Sy probably is an abbreviation of Sydney (or possibly Stanley?). 

                    (The BL copy is watermarked  IVY MILL 1815  and is bound with a large folding Plate 2 by Hall, which has 83 tinted examples with solution lines drawn in (by hand??), possibly one of four sheets giving all the problems in the book.  However there is no relationship between the Plate and the book -- problems are randomly placed and often drawn in different orientation.  I have a photocopy of the plate on two A3 sheets and a copy of a different plate with 72 problems, watermarked  J. Green  1816.)

A New Chinese Puzzle.  Third Edition: Universally allowed to be the most correct that has been published.  1817.  Dalgety has a copy.

A New Chinese Puzzle Consisting of Seven Pieces of Ivory or Wood, The Whole of which must be used, and will form each of the CHARACTERS.  J. Buckland, 23 Brook Street, Holborn, London.  Paper watermarked 1816.  (Dalgety has a copy, ??NYS.)

Miss D. Lowry.  A Key to the Only Correct Chinese Puzzle Which has Yet Been Published, with above a Hundred New Figures.  No. 1.  Drawn and engraved by Miss Lowry.  Printed by J. Barfield, London, 1817.  The initial D. is given on the next page.  Edward Hordern's collection has a copy.

W. Williams.  New Mathematical Demonstrations of Euclid, rendered clear and familiar to the minds of youth, with no other mathematical instruments than the triangular pieces commonly called the Chinese Puzzle.  Invented by Mr. W. Williams, High Beech Collegiate School, Essex.  Published by the author, London, 1817.  [Seen at BL.]

Enigmes Chinoises.  Grossin, Paris, 1817.  ??NYS -- described and partly reproduced in Milano.  Frontispiece facing the TP shows an oriental holding a banner which has the pieces and a few problems on it.  This is a small book, with five or six figures per page.  The figures seem to be copied from the Fashionable Chinese Puzzle, but some figures are not in that work.  Milano says this is cited as the first French usage of the term 'tangram', but this does not appear in Milano's photos and it is generally considered that Loyd introduced the word in the 1850s.  Milano's phrasing might be interpreted as saying this is the first French work on tangrams.

Chinesische-Raethsel.  Produced by Daniel Sprenger with designs by Matthaeus Loder, Vienna, c1818.  ??NYS -- mentioned by Milano.

Chinesisches Rätsel.  Enigmes chinoises.  Heinrich Friedrich Muller (or Mueller), Vienna, c1810??.  ??NYS (van der Waals).  This is probably a German edition of the above and should be dated 1817 or 1818.  However, Milano mentions a box in the Historisches Museum der Stadt Wien, labelled  Grosse Chinesische Raethsel,  produced by Mueller and dated 1815-1820. 

Passe-temps Mathématique, ou Récréation à l'ile Sainte-Hélène.  Ce jeu qui occupé à qu'on prétend, les loisirs du fameux exilé à St.-Hélène.  Briquet, Geneva, 1817.  21pp.  [Copy advertised by Interlibrum, Vaduz, in 1990.]

The New and Fashionable Chinese Puzzle.  A. T. Goodrich & Co., New York, 1817.  TP, 1p of Stanzas (seems like there should be a second page??), 32pp with 346 problems.  Slocum has a copy.

[Key] to the Chinese Philosophical Amusements.  A. T. Goodrich & Co., New York, 1817.  TP, 2pp of stanzas (the second page has the Napoleon footnote and a comment which indicates it is identical to the material in the problem book), Index to the Key to the Chinese Puzzle, 80pp of solutions as black shapes with white spacing.  Slocum has a copy.

                    NOTE.  This is quite a different book than The Fashionable Chinese Puzzle published in London by Wallis in 1817.

                    Slocum writes: "Although the Goodrich problem book has the same title as the British book by Wallis and Goodrich has the "Stanzas" poem (except for the first 2 paragraphs which he deleted) the problem books have completely different layouts and Goodrich's solution book largely copies Chinese books."

Il Nuovo e Dilettevole Giuoco Chinese.  Bardi, Florence, 1817.  ??NYS -- mentioned by Milano.

Buonapartes Geliefkoosste Vermaack op St. Helena, op Chineesch Raadsel.  1er Rotterdam by J. Harcke.  Prijs 1 - 4 ??.  2e Druck te(?) Rotterdam.  Ter Steendrukkery van F. Scheffers & Co.  Nanco Bordewijk has recently acquired this and Slocum has said it is a translation of one of the English items in c1818.  I have just a copy of the cover, and it uses many fancy letters which I don't guarantee to have read correctly.

Recueil des plus jolis Jeux de Sociéte,  dans lequel on trouve les gravures d'un grand nombre d'énigmes chinoises, et l'explication de ce nouveau jeu.  Chez Audot, Librairie, Paris, 1818.  Pp. 158-162: Le jeu des énigmes chinoises.  This is a short introduction, saying that the English merchants in Japan have sent it back to their compatriots and it has come from England to France.  This is followed by 11 plates.  The first three are numbered.  The first shows the pieces formed into a rectangle.  The others have 99 problems, with 7 shown solved (all six of those on plate 2 and one (the square) on the 10th plate.)

Das grosse chinesische Rätselspiel für die elegante Welt.  Magazin für Industrie (Leipzig) (1818).  ??NYS (van der Waals).  Jerry Slocum informs me that 'Magazin' here denotes a store, not a periodical, and that this is actually a game version with a packet of 50 cards of problems, occurring in several languages, from 1818.  I have acquired a set of the cards which lacks one card (no. 17), in a card box with labels in French and Dutch pasted on.  One side has:  Nouvelles / ENIGMES / Chinoises / en Figures et en Paysages  with a dancing Chinaman below.  The other side has:  Chineesch / Raadselspel, / voor / de Geleerde Waereld / in / 50 Beelaachlige / Figuren.  with two birds below.  Both labels are printed in red, with the dancing Chinaman having some black lines.  The cards are  82 x 55 mm and are beautifully printed with coloured pictures of architectonic, anthropomorphic and zoomorphic designs in appropriate backgrounds.  The first card has four shapes, three of which show the solution with dotted lines.  All other cards have just one problem shape.  The reverses have a simple design.  Slocum says the only complete set he has seen is in the British Library.  I have scanned the cards and the labels.

Gioco cinese chiamato il rompicapo.  Milan, 1818.  ??NYS (van der Waals).  Fratelli Bettali, Milan, nd, of which Dalgety has a copy.

Al Gioco Cinese Chiamato Il Rompicapo Appendice di Figure Rappresentanti ... Preceduta da un Discorso sul Rompicapo e sulla Cina intitolato Passatempo Preliminare scritto dall'Autore Firenze All'Insegna dell'Ancora 1818.  64pp + covers.  The cover or TP has an almond shape with the seven shapes inside.  Pp. 3-43 are text -- the Passatempo Preliminare and an errata page.  12 plates.  The first is headed Alfabeto in fancy Gothic.  Plates 1-3 give the alphabet (J and W are omitted).  Plate 4 has the positive digits.  Plates 5-12 have facing pages giving the names of the figures (rather orientalized) and contain 100 problems.  Hence a total of 133 problems, no solutions.  The Hordern collection has a copy and I have a photocopy from it.  This has some similarities to Giuoco Cinese.  Described and partly reproduced in Milano.

Al Gioco Cinese chiamato il Rompicapo Appendice.  Pietro & Giuseppe Vallardi, Milan, 1818.  Possibly another printing of the item above.  ??NYS -- described in Milano, who reproduces plates 1 & 2, which are identical to the above item, but with a simpler heading.  Milano says the plates are identical to those in the above item.

Nuove e Dilettevole Giuoco Chinese.  Milano presso li Frat. Bettalli Cont. del Cappello N. 4031.  Dalgety has a copy.  It is described and two pages are reproduced in Milano from an example in the Raccolta Bertarelli.  Milano dates it as 1818.  Cover illustration is the same as The Fashionable Chinese Puzzle, with the text changed.  But it is followed by some more text:  Questa ingegnosa invenzione è fondata sopra principi Geometrici, e consiste in 7 pezzi cioè 5. triangoli, un quadrato ed un paralellogrammo i quali possono essere combinati in modo da formare piu di 300 figure curiose.  The second photo shows a double page identical to pp. 3-4 of The Fashionable Chinese Puzzle, except that the page number on p. 4 was omitted in printing and has been written in.  (Quaritch's catalogue 646 (1947) item 698 lists this as  Nuovo e dilettevole Giuoco Chinese,  from Milan, [1820?])

Nuove e Dilettevole Giuoco Chinese.  Bologna Stamperia in pietra di Bertinazzi e Compag.  ??NYS -- described and partly reproduced in Milano from an example in the Raccolta Bertarelli.  Identical to the above item except that it is produced lithographically, the text under the cover illustration has been redrawn, the page borders, the page numbers and the figure numbers are a little different.  Milano's note 5 says the dating of this is very controversial.  Apparently the publisher changed name in 1813, and one author claims the book must be 1810.  Milano opts for 1813? but feels this is not consistent with the above item.  From Slocum's work and the examples above, it seems clear it must be 1818? 

Supplemento al nuovo giuoco cinese.  Fratelli Bettalli, Milan, 1818.  ??NYS -- described in Milano, who says it has six plates and the same letters and digits as Al Gioco Cinese Chiamato Il Rompicapo Appendice.

Giuoco Cinese  Ossia  Raccolta di 364. Figure Geometrica [last letter is blurred] formate con un Quadrato diviso in 7. pezzi, colli quali si ponno formare infinite Figure diversi, come Vuomini[sic], Bestie, Ucelli[sic], Case, Cocchi, Barche, Urne, Vasi, ed altre suppelletili domestiche: Aggiuntovi l'Alfabeto, e li Numeri Arabi, ed altre nuove Figure.  Agapito Franzetti alle Convertite, Rome, nd [but 1818 is written in by hand].  Copy at the Warburg Institute, shelf mark FMH 4050.  TP & 30 plates.  It has alternate openings blank, apparently to allow you to draw in your solutions, as an owner has done in a few cases.  The first plate shows the solutions with dotted lines, otherwise there are no solutions.  There is no other text than on the TP, except for a florid heading  Alfabeto  on plate XXVIII.  The diagrams have no numbers or names.  The upper part of the TP is a plate of three men, intended to be Orientals, in a tent?  The one on the left is standing and cutting a card marked with the pieces.  The man on the right is sitting at a low table and playing with the pieces.  He is seated on a box labelled  ROMPI CAPO.  A third man is seated behind the table and watching the other seated man.  On the ground are a ruler, dividers and right angle.  The Warburg does not know who put the date 1818 in the book, but the book has a purchase note showing it was bought in 1913.  James Dalgety has the only other copy known.  Sotheby's told him that Franzetti was most active about 1790, but Slocum finds Sotheby's is no longer very definite about this.  I thought it possible that a page was missing at the beginning which gave a different form of the title, but Dalgety's copy is identical to this one.  Mario Velucchi says it is not listed in a catalogue of Italian books published in 1800-1900.  The letters and numbers are quite different to those shown in Elffers and the other early works that I have seen, but there are great similarities to The New and Fashionable Chinese Puzzle, 1817 (check which??), and some similarities to Al Gioco Cinese above.  I haven't counted the figures to verify the 364.  Mentioned in Milano, based on the copy I sent to Dario Uri.

Jeu du Casse Tete Russe.  1817?  ??NYS -- described and partly reproduced in Milano from an example in the Raccolta Bertarelli but which has only four cards.  Here the figures are given anthropomorphic or architectonic shapes.  There are four cards on one coloured sheet and each card has a circle of three figures at the top with three more figures along the bottom.  Each card has the name of the game at the top of the circle and "les secrets des Chinois dévoliés" and "casse tête russe" inside and outside the bottom of the circle.  The figures are quite different than in the following item.

Nuovo Giuoco Russo.  Milano presso li Frat. Bettalli Cont. del Cappello.  [Frat. is an abbreviation of Fratelli (Brothers) and Cont. is an abbreviation of Contrade (road).]  Box, without pieces, but with 16 cards of problems (one being examples) and instruction sheet (or leaflet).  ??NYS - described by Milano with reproductions of the box cover and four of the cards.  This example is in the Raccolta Bertarelli.  Box shows a Turkish(?) man handing a box to another.  On the first card is given the title and publisher in French: Le Casse-Tête Russe  Milan, chez les Fr. Bettalli, Rue du Chapeau.  The instruction sheet says that the Giuoco Chinese has had such success in the principal cities of Europe that a Parisian publisher has conceived another game called the Casse Tête Russe and that the Brothers Bettalli have hurried to produce it.  Each card has four problems where the figures are greatly elaborated into architectonic forms, very like those in Metamorfosi, below.  Undated, but Milano first gives 1815‑1820, and feels this is closely related to Metamorfosi and similar items, so he concludes that it is 1818 or 1819, and this seems to be as correct as present knowledge permits.  The figures are quite different than in the French version above.

Metamorfosi del Giuoco detto l'Enimma Chinese.  Firenze  1818  Presso Gius. Landi Libraio sul Canto di Via de Servi.  Frontispiece shows an angel drawing a pattern on a board which has the seven pieces at the top.  The board leans against a plinth with the solution for making a square shown on it.  Under the drawing is A. G. inv.  Milano reproduces this plate.  One page of introduction, headed  Idea  della Metamorfosi Imaginata dell'Enimma Chinese.  100 shapes, some solved, then with elegant architectonic drawings in the same shapes, signed Gherardesce inv: et inc:  Milano identifies the artist as Alessandro Gherardesca (1779-1852), a Pisan architect.  See S&B, pp. 24‑25.

Grand Jeu du Casse Tête Français en X. Pieces.  ??NYS -- described and partly reproduced in Milano, who says it comes from Paris and dates it 1818?  The figures are anthropomorphic and are most similar to those in Jeu du Casse Tete Russe. 

Grande Giuocho del Rompicapo Francese.  Milano presso Pietro e Giuseppe Vallardi Contrada di S. Margherita No 401(? my copy is small and faint).  ??NYS -- described and partly reproduced in Milano, who dates it as 1818-1820.  Identical problems as in the previous item, but the figures have been redrawn rather than copied exactly.

Ch'i Ch'iao pan.  c1820.  (Bibliothek Leiden 6891; Antiquariat Israel, Amsterdam.)  ??NYS (van der Waals).

Le Veritable Casse‑tete, ou Enigmes chinoises.  Canu Graveur, Paris, c1820.  BL.  ??NYS (van der Waals).

L'unico vero Enimma Chinese  Tradotto dall'originale, pubblicato a Londra, da J. Barfield.  Florence, [1820?].  [Listed in Quaritch's catalogue 646 (1947) item 699.)

A tangram appears in Pirnaisches Wochenblatt of 16 Dec 1820.  ??NYS -- described in Slocum, p. 60.

Ch'i Ch'iao ch'u pien ho‑pi.  After 1820.  (Bibliothek Leiden 6891.)  ??NYS (van der Waals).  476 examples.

Nouveau Casse‑Tête Français.  c1820 (according to van der Waals).  Reproduced in van der Waals, but it's not clear how the pages are assembled.  Milano dates it a c1815 and indicates it is 16 cards, but van der Waals looks like it may have been a booklet of 16 pp with TP, example page and end page.  The 16 pp have 80 problems. 

                    Jerry Slocum has sent 2 large pages with 58 figurative shapes which are clearly the same pictures.  The instructions are essentially the same, but are followed by rules for a Jeu de Patience on the second page and there is a  6 x 6  table of words on the first page headed "Morales trouvées dans les ruines de la célébres Ville de Persépolis ..." which one has to assemble into moral proverbs.  It looks like these are copies of folding plates in some book of games. 

Chinese Puzzle  Georgina.  A. & S. Josh Myers, & Co  144, Leadenhall Street, London.  Ganton Litho.  81 examples on 8 plates with elegant TP.  Pages are one-sided sheets, sewn in the middle, but some are upside down.  Seen at BL (1578/4938).

Bestelmeier, 1823.  Item 1278: Chinese Squares.  It is not in the 1812 catalogue. 

Slocum.  Compendium.  Shows the above Bestelmeier entry.

Anonymous.  Ch'i ch'iao t'u ho pi (Harmoniously combined book of Tangram problems)  and  Ch'i ch'iao t'u chieh (Tangram solutions).  Two volumes of tangrams and solutions with no title page, Chinese labels of the puzzles, in Chinese format (i.e. printed as long sheets on thin paper, accordion folded and stitched with ribbon.  Nd [c1820s??], stiff card covers with flyleaves of a different paper, undoubtedly added later.  84 pages in each volume, containing 334 problems and solutions.  With ownership stamp of a cartouche enclosing EWSHING, probably a Mr. E. W. Shing.  Slocum says this is a c1820s reprint of the earliest Chinese tangram book which appeared in 1813 & 1815.  This version omits the TP and opening text.  I have a photocopy of the opening material from Slocum.  The original problem book had a preface by  Sang‑hsia K'o,  which was repeated in the solution book with the same date.  Includes all the problems of Shichi-kou-zu Gappeki, qv. 

New Series of Ch'i ch'iau puzzles.  Printed by Lou Chen‑wan, Ch'uen Liang, January 1826.  ??NYS.  (Copy at Dept. of Oriental Studies, Durham Univ., cited in R. C. Bell; Tangram Teasers.)

Neues chinesisches Rätselspiel für Kinder, in 24 bildlichen und alphabetischen Darstellungen.  Friese, Pirna.  Van der Waals, copying Santi, gives c1805, but Slocum, p. 60, reports that it first appears in Pirnaisches Wochenblatt of 19 Dec 1829, though there is another tangram in the issue of 16 Dec 1820.  ??NYS.

Child.  Girl's Own Book.  1833: 85;  1839: 72;  1842: 156.  "Chinese Puzzles -- These consist of pieces of wood in the form of squares, triangles, &c.  The object is to arrange them so as to form various mathematical figures."

Anon.  Edo Chiekata (How to Learn It??) (In Japanese).  Jan 1837, 19pp, 306 problems.  (Unclear if this uses the Tangram pieces.)  Reprinted in the same booklet as Sei Shōnagon, on pp. 37‑55.

A Grand Eastern Puzzle.  C. Davenport & Co., London.  Nd.  ??NYS (van der Waals).  (Dalgety has a copy and gives C. Davenporte (??SP) and Co., No. 20, Grafton Street, East Euston Square.  Chinese pages dated 1813 in European binding with label bearing the above information.)

Augustus De Morgan.  On the foundations of algebra, No. 1.  Transactions of the Cambridge Philosophical Society 7 (1842) 287-300.  ??NX.  On pp. 289, he says "the well-known toy called the Chinese Puzzle, in which a prescribed number of forms are given, and a large number of different arrangements, of which the outlines only are drawn, are to be produced."

Crambrook.  1843.  P. 4, no. 4: Chinese Puzzle.  Chinese Books, thirteen numbers.  Though not illustrated, this seems likely to be the Tangrams -- ??

Boy's Own Book.  1843 (Paris): 439. 

No. 19: The Chinese Puzzle.  Instructions give five shapes and say to make one copy of some and two copies of the others.  As written, this has two medium sized triangles instead of two large ones, though it is intended to be the tangrams.  11  problem shapes given, no answers.  Most of the shapes occur in earlier tangram collections, particularly in  A New Invented Chinese Puzzle.  "The puzzle may be purchased, ..., at Mr. Wallis's, Skinner street, Snow hill, where numerous books, containing figures for this ingenious toy may also be obtained."  = Boy's Treasury, 1844, pp. 426-427, no. 16.  It is also reproduced, complete with the error, but without the reference to Wallis, as:  de Savigny, 1846, pp. 355-356, no. 14: Le casse-tête chinois;  Magician's Own Book, 1857, prob. 49, pp. 289-290;  Landells, Boy's Own Toy-Maker, 1858, pp. 139-140;  Book of 500 Puzzles, 1859, pp. 103‑104;  Boy's Own Conjuring Book, 1860, pp. 251-252;  Wehman, New Book of 200 Puzzles, 1908, pp. 34‑35.

No. 20: The Circassian puzzle.  "This is decidedly the most interesting puzzle ever invented; it is on the same principle, but composed of many more pieces than the Chinese puzzle, and may consequently be arranged in more intricate figures.  ..."  No pieces or problems are shown.  In the next problem, it says: "This and the Circassian puzzle are published by Mr. Wallis, Skinner-street, Snow-hill."  = Boy's Treasury, 1844, p. 427, no. 17.  = de Savigny, 1846, p. 356, no. 15: Le problème circassien, but the next problem omits the reference to Wallis. 

Although I haven't recorded a Circassian puzzle yet -- cf in 6.S.2 -- I have just seen that the puzzle succeeding The Chinese Puzzle in Wehman, New Book of 200 Puzzles, 1908, pp. 35-36, is called The Puzzle of Fourteen which might be the Circassian puzzle.  Taking a convenient size, this has two equilateral triangles of edge 1 and four each of the following:  a  30o-60o-90o  triangle with edges  2, 1, Ö3;  a parallelogram with angles  60o and 120o  with edges 1 and 2;  a trapezium with base angles  60o and 60o,  with lower and upper base edges 2 and 1, height Ö3/4 and slant edges 1/2 and Ö3/2.  All 14 pieces make a rectangle  2Ö3 by 4.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 584-11, pp. 288 & 405: Chinesisches Verwandlungsspiel.  Make a square with the tangram pieces.  Shows just five of the pieces, but correctly states which two to make two copies of.

Prob. 584-16, pp. 289 & 406.  Make an isosceles right triangle with the tangram pieces.

Prob. 584-18/25, pp. 289-291 & 407: Hieroglyphenspiele.  Form various figures from various sets of pieces, mostly tangrams, but the given shapes have bits of writing on them so the assembled figure gives a word.  Only one of the shapes is as in Boy's Own Book.

Prob. 588, pp. 298 & 410: Etliche Knackmandeln.  Another tangram problem like the preceding, not equal to any in Boy's Own Book.

Adams & Co., Boston.  Advertisement in The Holiday Journal of Parlor Plays and Pastimes, Fall 1868.  Details?? -- photocopy sent by Slocum.  P. 6: Chinese Puzzle.  The celebrated Puzzle with which a hundred or more symmetrical forms can be made, with book showing the designs.  Though not illustrated, this seems likely to be the Tangrams -- ??

Mittenzwey.  1880.  Prob. 243-252, pp. 45 & 95-96;  1895?: 272-281, pp. 49 & 97-98;  1917: 272-282, pp. 45 & 92-93.  Make a funnel, kitchen knife, hammer, hat with brim being horizontal or hanging down or turned up, church, saw, dovecote, hatchet, square, two equal squares.

J. Murray (editor of the OED).  Two letters to H. E. Dudeney (9 Jun 1910  &  1 Oct 1910).  The first inquires about the word 'tangram', following on Dudeney's mention of it in his "World's best puzzles" (op. cit. in 2).  The second says that 'tan' has no Chinese origin;  is apparently mid 19C, probably of American origin;  and the word 'tangram' first appears in Webster's Dictionary of 1864.  Dudeney, AM, 1917, p. 44, excerpts these letters.

F. T. Wang & C.‑S. Hsiung.  A theorem on the tangram.  AMM 49 (1942) 596‑599.  They determine the 20 convex regions which 16 isosceles right triangles can form and hence the 13 ones which the Tangram pieces can form.

Mitsumasa Anno.  Anno's Math Games.  (Translation of: Hajimete deau sugaku no ehon; Fufkuinkan Shoten, Tokyo, 1982.)  Philomel Books, NY, 1987.  Pp. 38-43 & 95-96 show a simplified  5-piece tangram-like puzzle which I have not seen before.  The pieces are: a square of side  1;  three isosceles right triangles of side  1;  a right trapezium with bases  1  and  2,  altitude  1  and slant side  Ö2.  The trapezium can be viewed as putting together the square with a triangle.  19 problems are set, with solutions at the back.

James Dalgety.  Latest news on oldest puzzles.  Lecture to Second Meeting on the History of Recreational Mathematics, 1 Jun 1996.  10pp.  In 1998, he extracted the two sections on tangrams and added a list of tangram books in his collection as: The origins of Tangram; © 1996/98; 10pp.  (He lists about 30 books, eight up to 1850.)  In 1993, he was buying tangrams in Hong Kong and asked what they called it.  He thought they said 'tangram' but a slower repetition came out 'ta hau ban' and they wrote down the characters and said it translates as 'seven lucky tiles'.  He has since found the characters in 19C Chinese tangram books.  It is quite possible that Sam Loyd (qv under Murray, above) was told this name and wrote down 'tangram', perhaps adjusted a bit after thinking up Tan as the inventor.

At the International Congress on Mathematical Education, Seville, 1996, the Mathematical Association gave out  The  3, 4, 5  Tangram,  a cut card tangram, but in a  6 x 8  rectangular shape, so that the medium sized triangle was a  3-4-5 triangle.  I modified this in Nov 1999, by stretching along a diagonal to form a rhombus with angles double the angles of a 3-4-5 triangle, so that four of the triangles are similar to 3-4-5 triangles.  Making the small triangles be actually 3-4-5, all edges are integral.  I made up 35 problems with these pieces.  I later saw that Hans Wiezorke has mentioned this dissection in CFF, but with no problems.  I distributed this as my present at G4G4, 2000.

 

          6.S.1. LOCULUS OF ARCHIMEDES

 

See S&B 22.  I recall there is some dispute as to whether the basic diagram should be a square or a double square.

E. J. Dijksterhuis.  Archimedes.  Munksgaard, Copenhagen, 1956;  reprinted by Princeton Univ. Press, 1987.  Pp. 408‑412 is the best discussion of this topic and supplies most of the classical references.

 

Archimedes.  Letter to Eratosthenes, c-250?.  Greek palimpsest, c975, on MS no. 355, from the Cloister of Saint Sabba (= Mar Saba), Jerusalem, then at Metochion of the Holy Sepulchre, Constantinopole.  [This MS disappeared in the confusion in Asia Minor in the 1920s but reappeared in 1998 when it was auctioned by Christie's in New York for c2M$.  Hopefully, modern technology will allow a facsimile and an improved transcription in the near future.]  Described by J. L. Heiberg (& H. G. Zeuthen); Eine neue Schrift des Archimedes; Bibliotheca Math. (3) 7 (1906‑1907) 321‑322.  Heiberg describes the MS, but only mentions the loculus.  The text is in Heiberg's edition of Archimedes; Opera; 2nd ed., Teubner, Leipzig, 1913, vol. II, pp. 415‑424, where it has been restored using the Suter MSS below.  Heath only mentions the problem in passing.  Heiberg quotes Marius Victorinus, Atilius Fortunatianus and cites Ausonius and Ennodius. 

H. Suter.  Der Loculus Archimedius oder das Syntemachion des Archimedes.  Zeitschr. für Math. u. Phys. 44 (1899) Supplement  = AGM 9 (1899) 491‑499.  This is a collation from two 17C Arabic MSS which describe the construction of the loculus.  They are different than the above MS.  The German is included in Archimedes Opera II, 2nd ed., 1913, pp. 420‑424.

Dijksterhuis discusses both of the above and says that they are insufficient to determine what was intended.  The Greek seems to indicate that Archimedes was studying the mathematics of a known puzzle.  The Arabic shows the construction by cutting a square, but the rest of the text doesn't say much.

 

Lucretius.  De Rerum Natura.  c‑70.  ii, 778‑783.  Quoted and discussed in H. J. Rose; Lucretius ii. 778‑83; Classical Review (NS) 6 (1956) 6‑7.  Brief reference to assembling pieces into a square or rectangle.

Decimus Magnus Ausonius.  c370.  Works.  Edited & translated by H. G. Evelyn White.  Loeb Classical Library, ??date.  Vol. I, Book XVII: Cento Nuptialis (A Nuptial Cento), pp. 370-393 (particularly the Preface, pp. 374-375) and Appendix, pp. 395-397.  Refers to 14 little pieces of bone which form a monstrous elephant, a brutal boar, etc.  The Appendix gives the construction from the Arabic version, via Heiberg, and forms the monstrous elephant.

Marius Victorinus.  4C.  VI, p. 100 in the edition of Keil, ??NYS.  Given in Archimedes Opera II, 2nd ed., 1913, p. 417.  Calls it 'loculus Archimedes' and says it had 14 pieces which make a ship, sword, etc.

Ennodius.  Carmina: De ostomachio eburneo.  c500.  In:  Magni Felicis Ennodii Opera; ed. by F. Vogel, p. 340.  In:  Monumenta Germaniae Historica, VII (1885) 249.  ??NYS.  Refers to ivory pieces to be assembled.

Atilius Fortunatianus.  6C.  ??NYS  Given in Archimedes Opera II, p. 417.  Same comment as for Marius Victorinus.

E. Fourrey.  Curiositiés Géométriques.  (1st ed., Vuibert & Nony, Paris, 1907);  4th ed., Vuibert, Paris, 1938.  Pp. 106‑109.  Cites Suter, Ausonius, Marius Victorinus, Atilius Fortunatianus.

Collins.  Book of Puzzles.  1927.  The loculus of Archimedes, pp. 7-11.  Pieces made from a double square.

 

          6.S.2. OTHER SETS OF PIECES

 

          See Hoffmann & S&B, cited at the beginning of 6.S, for general surveys.

          See Bailey in 6.AS.1 for an 1858 puzzle with 10 pieces and The Sociable and Book of 500 Puzzles, prob. 10, in 6.AS.1 for an 11 piece puzzle. 

          There are many versions of this idea available and some are occasionally given in JRM.

          The Richter Anchor Stone puzzles and building blocks were inspired by Friedrich Froebel (or Fröbel) (1782‑1852), the educational innovator.  He was the inventor of Kindergartens, advocated children's play blocks and inspired both the Richter Anchor Stone Puzzles and Milton Bradley.  The stone material was invented by Otto Lilienthal (1848‑1896) (possibly with his brother Gustav) better known as an aviation pioneer -- they sold the patent and their machines to F. Adolph Richter for 1000 marks.  The material might better be described as a kind of fine brick which could be precisely moulded.  Richter improved the stone and began production at Rudolstadt, Thüringen, in 1882; the plant closed in 1964.  Anchor was the company's trademark.  He made at least 36 puzzles and perhaps a dozen sets of building blocks which were popular with children, architects, engineers, etc.  The Deutsches Museum in Munich has a whole room devoted to various types of building blocks and materials, including the Anchor blocks.  The Speelgoed Museum 'Op Stelten' (Sint Vincentiusstraat 86, NL-4902 (or 4901) GL Oosterhout, Noord-Brabant, The Netherlands; tel: 0262 452 825; fax: 0262 452 413) has a room of Richter blocks and some puzzles.  There was an Anker Museum in the Netherlands (Stichting Ankerhaus (= Anker Museum); Opaalstraat 2‑4 (or Postf. 1061), NL-2400 BB Alphen aan den Rijn, The Netherlands; tel: 01720‑41188) which produced replacement parts for Anker stone puzzles.  Modern facsimiles of the building sets are being produced at Rudolstadt.

 

In 1996 I noticed the ceiling of the room to the south of the Salon of the Ambassadors in the Alcazar of Seville.  This 15C? ceiling was built by workmen influenced by the Moorish tradition and has 112 square wooden panels in a wide variety of rectilineal patterns.  One panel has some diagonal lines and looks like it could be used as a 10 piece tangram-like puzzle.  Consider a  4 x 4  square.  Draw both diagonal lines, then at two adjacent corners, draw two lines making a unit square at these corners.  At the other two corners draw one of these two lines, namely the one perpendicular to their common side.  This gives six isosceles right triangles of edge  1;  two pentagons with three right angles and sides  1, 2, 1, Ö2, Ö2;  two quadrilaterals with two right angles and sides  2, 1, Ö2, 2Ö2.  Since geometric patterns and panelling are common features of Arabic art, I wonder if there are any instances of such patterns being used for a tangram-like puzzle?

Grand Jeu du Casse Tête Français en X. Pieces.  ??NYS -- described and partly reproduced in Milano, who says it comes from Paris and dates it 1818?  The figures are anthropomorphic and are most similar to those in Jeu du Casse Tete Russe. 

Grande Giuocho del Rompicapo Francese.  Milano presso Pietro e Giuseppe Vallardi Contrada di S. Margherita No 401(? my copy is small and faint).  ??NYS -- described and partly reproduced in Milano, who dates it as 1818-1820.  Identical problems as in the previous item, but the figures have been redrawn rather than copied exactly.

Allizeau.  Les Métamorphoses  ou  Amusemens Géometriques  Dédiée aux Amateurs  Par Allizeau.  A Paris chex Allizeau  Quai Malaquais, No 15.  ??NYS -- described and partly reproduced in Milano.  This uses 15 pieces and the problems tend to be architectural forms, like towers.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles, nos. 20-27, pp. 27-29 & 88-89 & plate II, figs. 15-22.  This is a set of 20 pieces of 8 shapes used to make a square, a right triangle, three squares, etc.

Crambrook.  1843.  P. 4, no. 1: Pythagorean Puzzle, with Book.  Though not illustrated, this is probably(??) the puzzle described in Hoffmann, below, which was a Richter Anchor puzzle No. 12 of the same name and is still occasionally seen.  See S&B 28.

Edward Hordern's collection has a Circassian Puzzle, c1870, with many pieces, but I didn't record the shapes -- cf Boy's Own Book, 1843 (Paris), in section 6.S.

Mittenzwey.  1880. 

Prob. 177-179, pp. 34 & 86;  1895?: 202-204, pp. 38-39 & 88;  1917: 202-204, pp. 35 & 84-85.  Consider the ten piece version of dissecting 5 squares to one (6.AS.1).  Use the pieces to make: 

           a squat octagon, a house gable-end, a church (no solution), etc.;

           two dissimilar rectangles;

           three dissimilar parallelograms, two dissimilar trapezoids.  Solution says one can make many other shapes with these pieces, e.g. a trapezoid with parallel sides in the proportion  9 : 11.

Prob. 181-184, pp. 34-35 & 87-88;  1895?: 206-209, pp. 39 & 89-90;  1917: 206-209, pp. 36 & 85-86.  Take six equilateral triangles of edge 2.  Cut an equilateral triangle of edge 1 from the corner of each of them, giving  12  pieces.  Make a hexagon in eight different ways [there are many more -- how many??] and three tangram-like shapes.

Prob. 195-196, pp. 36 & 89;  1895?: 220-221, pp. 41 & 91;  1917: 220-221, pp. 37 & 87.  Use four isosceles right triangles, say of leg 1, to make a square, a  1 x 4  rectangle and an isosceles right triangle.

Nicholas Mason.  US Patent 232,140 - Geometrical Puzzle-Block.  Applied: 13 May 1880; patented 14 Sep 1880.  1p plus 2pp diagrams.  Five squares, six units square, each cut into four pieces in the same way.  Start at the midpoint of a side and cut to an opposite corner.  (This is the same cut used to produce the ten piece 'Five Squares to One' puzzle.)  Cut again in the triangle just formed, from the same midpoint to a point one unit from the right angle corner of the piece just made.  This gives a right triangle of sides  3, 1, Ö10  and a triangle of sides  5, Ö10, 3Ö45.  Cut again from the same midpoint across the trapezoidal piece made by the first cut, to a point five units from the corner previously cut to.  This gives a triangle of sides  5, 3Ö5, 2Ö10  and a right trapezoid with sides  2,Ö10, 1, 6, 3.  This was produced as Hill's American Geometrical Prize Puzzle in England ("Price, One Shilling.") in 1882.  Harold Raizer produced a facsimile version, with facsimile box label and instructions for IPP22.  The instructions have 20 problems to solve and the solutions have to be submitted by 1 May 1882.

Hoffmann.  1893.  Chap. III, no. 3: The Pythagoras Puzzle, pp. 83-85 & 117-118 = Hoffmann‑Hordern, pp. 69-72.  This has 7 pieces and is quite like the Tangram -- see comment under Crambrook.  Photo on p. 71, with different version in Hordern Collection, p. 50.

C. Dudley Langford.  Note 1538:  Tangrams and incommensurables.  MG 25 (No. 266) (Oct 1941) 233‑235.  Gives alternate dissections of the square and some hexagonal dissections.

C. Dudley Langford.  Note 2861:  A curious dissection of the square.  MG 43 (No. 345) (Oct 1959) 198.  There are 5 triangles whose angles are multiples of  π/8 = 22½o.  He uses these to make a square.

See items at the end of 6.S.

 

          6.T.    NO THREE IN A LINE PROBLEM

 

          See also section 6.AO.2.

 

Loyd.  Problem 14: A crow puzzle.  Tit‑Bits 31 (16 Jan  &  6 Feb 1897) 287  &  343.  = Cyclopedia, 1914, Crows in the corn, pp. 110 & 353.  = MPSL1, prob. 114, pp. 113 & 163‑164.  8 queens with no two attacking and no three in any line.

Dudeney.  The Tribune (7 Nov 1906) 1.  ??NX.  = AM, prob. 317, pp. 94 & 222.  Asks for a solution with two men in the centre 2 x 2 square.

Loyd.  Sam Loyd's Puzzle Magazine, January 1908.  ??NYS.  (Given in A. C. White; Sam Loyd and His Chess Problems; 1913, op. cit. in 1; p. 100, where it is described as the only solution with 2 pieces in the 4 central squares.)

Ahrens, MUS I 227, 1910, says he first had this in a letter from E. B. Escott dated  1 Apr 1909.  (W. Moser, below, refers this to the 1st ed., 1900, but this must be due to his not having seen it.)

C. H. Bullivant.  Home Fun, 1910, op. cit. in 5.S.  Part VI, Chap. IV: No. 2: Another draught puzzle, pp. 515 & 520.  The problem says "no three men shall be in a line, either horizontally or perpendicularly".  The solution says "no three are in a line in any direction" and the diagram shows this is indeed true.

Loyd.  Picket posts.  Cyclopedia, 1914, pp. 105 & 352.  = MPSL2, prob. 48, pp. 34 & 136.  2 pieces initially placed in the 4 central squares.

Blyth.  Match-Stick Magic.  1921.  Matchstick board game, p. 73.  6 x 6  version phrased as putting "only two in any one line: horizontal, perpendicular, or diagonal."  However, his symmetric solution has three in a row on lines of slope 2.

King.  Best 100.  1927.  No. 69, pp. 28 & 55.  Problem on the  6 x 6  board -- gives a symmetric solution.  Says "there are two coins on every row" including "diagonally across it", but he has three in a row on lines of slope 2. 

Loyd Jr.  SLAHP.  1928.  Checkers in rows, pp. 40 & 98.  Different solution than in Cyclopedia.

M. Adams.  Puzzle Book.  1939.  Prob. C.83: Stars in their courses, pp. 144 & 181.  Same solution as King, but he says "two stars in each vertical row, two in each horizontal row, and two in each of the the two diagonals ....  There must not be more than two stars in the same straight line", but he has three in a row on lines of slope 2.

W. O. J. Moser & J. Pach.  No‑three‑in‑line problem.  In:  100 Research Problems in Discrete Geometry 1986; McGill Univ., 1986.  Problem 23, pp. 23.1 -- 23.4.  Survey with 25 references.  Solutions are known on the  n x n  board for  n £ 16  and for even  n £ 26.  Solutions with the symmetries of the square are only known for  n = 2, 4, 10.

 

          6.U.   TILING

 

          6.U.1. PENROSE PIECES

 

R. Penrose.  The role of aesthetics in pure and applied mathematical research.  Bull. Inst. Math. Appl. 10 (1974) 266‑272.

M. Gardner.  SA (Jan 1977).  Extensively rewritten as Penrose Tiles, Chaps. 1 & 2.

R. Penrose.  Pentaplexity.  Eureka 39 (1978) 16‑22.  =  Math. Intell. 2 (1979) 32‑37.

D. Shechtman, I. Blech, D. Gratias & J. W. Cohn.  Metallic phase with long‑range orientational order and no translational symmetry.  Physical Rev. Letters 53:20 (12 Nov 1984) 1951‑1953.  Describes discovery of 'quasicrystals' having the symmetry of a Penrose‑like tiling with icosahedra.

David R. Nelson.  Quasicrystals.  SA 255:2 (Aug 1986) 32‑41 & 112.  Exposits the discovery of quasicrystals.  First form is now called 'Shechtmanite'.

Kimberly-Clark Corporation has taken out two patents on the use of the Penrose pattern for quilted toilet paper as the non-repetition prevents the tissue from 'nesting' on the roll.  In Apr 1997, Penrose issued a writ against Kimberly Clark Ltd. asserting his copyright on the pattern and demanding damages, etc.

John Kay.  Top prof goes potty at loo roll 'rip-off'.  The Sun (11 Apr 1997) 7.

Patrick McGowan.  It could end in tears as maths boffin sues Kleenex over design.  The Evening Standard (11 Apr 1997) 5.

Kleenex art that ended in tears.  The Independent (12 Apr 1997) 2.

For a knight on the tiles.  Independent on Sunday (13 Apr 1997) 24.  Says they exclusively reported Penrose's discovery of the toilet paper on sale in Dec 1996.

D. Trull.  Toilet paper plagiarism.  Parascope, 1997 -- available on www.noveltynet.org/content/paranormal/www.parascope.com/arti...

 

          6.U.2. PACKING BRICKS IN BOXES

 

          In two dimensions, it is not hard to show that  a x b  packs  A x B  if and only if  a  divides either  A or B;  b  divides either  A or B;  A and B  are both linear combinations of  a and b.  E.g.  2 x 3  bricks pack a  5 x 6  box.

          See also 6.G.1.

 

Anon.  Prob. 52.  Hobbies 30  (No. 767) (25 Jun 1910) 268 & 283  &  (No. 770) (16 Jul 1910) 328.  Use at least one of each of  5 x 7,  5 x 10,  6 x 10  to make the smallest possible square.  Solution says to use  4, 4, 1,  but doesn't show how.  There are lots of ways to make the assembly.

Manuel H. Greenblatt (  -1972, see JRM 6:1 (Winter 1973) 69).  Mathematical Entertainments.  Crowell, NY, 1965.  Construction of a cube, pp. 80‑81.  Can  1 x 2 x 4  fill  6 x 6 x 6?  He asserts this was invented by R. Milburn of Tufts Univ.

N. G. de Bruijn.  Filling boxes with bricks.  AMM 76 (1969) 37‑40.  If  a1 x ... x an  fills  A1 x ... x An  and  b  divides  k  of the  ai,  then  b  divides at least  k  of the  Ai.  He previously presented the results, in Hungarian, as problems in Mat. Lapok 12, pp. 110‑112, prob. 109 and 13, pp. 314‑317, prob. 119.  ??NYS.

D. A. Klarner.  Brick‑packing puzzles.  JRM 6 (1973) 112‑117.  General survey.  In this he mentions a result that I gave him -- that  2 x 3 x 7  fills a  8 x 11 x 21,  but that the box cannot be divided into two packable boxes.  However, I gave him the case  1 x 3 x 4  in  5 x 5 x 12  which is the smallest example of this type.  Tom Lensch makes fine examples of these packing puzzles.

T. H. Foregger, proposer;  Michael Mather, solver.  Problem E2524 -- A brick packing problem.  AMM  82:3 (Mar 1975) 300  &  83:9 (Nov 1976) 741-742.  Pack  41  1 x 2 x 4  bricks in a  7 x 7 x 7  box.  One cannot get  42  such bricks into the box.

 

          6.V.   SILHOUETTE AND VIEWING PUZZLES

 

          Viewing problems must be common among draughtsmen and engineers, but I haven't seen many examples.  I'd be pleased to see further examples.

 

2 silhouettes.

Circle  &  triangle  --  van Etten,  Ozanam,  Guyot,  Magician's Own Book (UK version)

Circle  &  square  --  van Etten

Circle  &  rhombus  --  van Etten,  Ozanam

Rectangle with inner rectangle  &  rectangle with notch  --  Diagram Group.

3 silhouettes.

Circle,  circle,  circle  --  Madachy

Circle,  cross,  square  --  Shortz collection (c1884),  Wyatt,  Perelman

Circle,  oval,  rectangle  --  van Etten,  Ozanam,  Guyot,  Magician's Own Book (UK version)

Circle,  oval,  square  --  van Etten,  Tradescant,  Ozanam,  Ozanam‑Montucla,  Badcock,  Jackson,  Rational Recreations,  Endless  Amusement II,  Young Man's Book

Circle,  rhombus,  rectangle  --  Ozanam,  Alberti

Circle,  square,  triangle  --  Catel,  Bestelmeier,  Jackson,  Boy's Own Book,  Crambrook,  Family Friend,  Magician's Own Book,  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Illustrated Boy's Own Treasury,  Riecke,  Elliott,  Mittenzwey,  Tom Tit,  Handy Book,  Hoffmann,  Williams,  Wyatt,  Perelman,  Madachy.  But see Note below.

Square,  tee,  triangle  --  Perelman

4 silhouettes.

          Circle,  square,  triangle,  rectangle with curved ends  --  Williams

2 views.

          Antilog,  Ripley's,  Diagram Group;

3 views.

          Madachy,  Ranucci, 

 

          For the classic Circle, Square, Triangle, version, the triangle cannot be not equilateral.  Consider a circle, rectangle, triangle version.  If  D  is the diameter of the circle and  H  is the height of the plug, then the rectangle has dimensions  D x H  and the triangle has base  D  and side  S,  so  S = Ö(H2 + D2/4).  Making the rectangle a square, i.e.  H = D,  makes  S = DÖ5/2,  while making the triangle equilateral, i.e.  S = D,  makes  H = DÖ3/2.

 

van Etten.  1624. 

Prob. 22 (misnumbered 15 in 1626) (Prob. 20), pp. 19‑20 & figs. opp. p. 16 (pp. 35‑36): 2 silhouettes --  one circular, the other triangular, rhomboidal or square.  (English ed. omits last case.)  The 1630 Examen says the author could have done better and suggests:  isosceles triangle, several scalene triangles, oval or circle, which he says can be done with an elliptically cut cone and a scalene cone.  I am not sure I believe these.  It seems that the authors are allowing the object to fill the hole and to pass through the hole moving at an angle to the board rather than perpendicularly as usually understood.  In the English edition the Examination is combined with that of the next problem. 

Prob. 23 (21), pp. 20‑21 & figs. opp. p. 16 (pp. 37‑38): 3 silhouettes -- circle, oval and square or rectangle.  The 1630 Examen suggests:  square, circle, several parallelograms and several ellipses, which he says can be done with an elliptic cylinder of height equal to the major diameter of the base.  The English Examination says "a solid colume ... cut Ecliptick-wise" -- ??

John II Tradescant (1608-1662).  Musæum Tradescantianum: Or, A Collection of Rarities Preserved at South-Lambeth neer London By John Tradescant.  Nathaniel Brooke, London, 1656.  [Facsimile reprint, omitting the Garden List, Old Ashmolean Reprints I, edited by R. T. Gunther, on the occasion of the opening of the Old Ashmolean Museum as what has now become the Museum of the History of Science, Oxford.  OUP, 1925.]  John I & II Tradescant were gardeners to nobility and then royalty and used their connections to request naval captains to bring back new plants, curiosities and "Any thing that Is strang".  These were accumulated at his house and garden in south Lambeth, becoming known as Tradescant's Ark, eventually being acquired by Elias Ashmole and becoming the foundation of the Ashmolean Museum in Oxford.  This catalogue was prepared by Elias Ashmole and his friend Thomas Wharton, but they are not named anywhere in the book.  It was the world's first museum catalogue.

                    P. 37, last entry: "A Hollow cut in wood, that will fit a round, square and ovall figure."

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  He says  square, circle and triangle  is in a book in front of him dated 1674.  I suspect this must be the 1674 English edition of van Etten, but I don't find the problem in the English editions I have examined.  Perhaps Dudeney just meant that the idea was given in the 1674 book, though he is specifically referring to the square, circle, triangle version.

Ozanam.  1725.  Vol. II, prob. 58 & 59, pp. 455‑458 & plate 25* (53 (note there is a second plate with the same number)).  Circle and triangle;  circle and rhombus;  circle, oval, rectangle;  circle, oval, square.  Figures are very like van Etten.  See Ozanam-Montucla, 1778.

Ozanam.  1725.  Vol. IV.  No text, but shown as an unnumbered figure on plate 15 (17).  3 silhouettes:  circle, rhombus, rectangle.

Simpson.  Algebra.  1745.  Section XVIII, prob. XXIX, pp. 279-281.  (1790: prob. XXXVII, pp. 306-307.  Computes the volume of an ungula obtained by cutting a cone with a plane.  Cf Riecke, 1867.

Alberti.  1747.  No text, but shown as an unnumbered figure on plate XIIII, opp. p. 218 (112), copied from Ozanam, 1725, vol IV.  3 silhouettes:  circle, rhombus, rectangle.

Ozanam-Montucla.  1778.  Faire passer un même corps par un trou quarré, rond & elliptique.  Prob. 46, 1778: 347-348;  1803: 345-346;  1814: 293.  Prob. 45, 1840: 149-150.  Circle, ellipse, square.

Catel.  Kunst-Cabinet.  1790.  Die mathematischen Löcher, p. 16 & fig. 42 on plate II.  Circle, square, triangle.

E. C. Guyot.  Nouvelles Récréations Physiques et Mathématiques.  Op. cit. in  6.P.2.  1799.  Vol. 2, Quatrième récréation, p. 45 & figs. 1‑4, plate 7, opp. p. 45.  2 silhouettes:  circle & triangle;  3 silhouettes:  circle, oval, rectangle.

Bestelmeier. 

1801.  Item 536: Die 3 mathematischen Löcher.  (See also the picture of Item 275, but that text is for another item.)  Square, triangle and circle.

1807.  Item 1126: Tricks includes the  square, triangle and circle.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  P. 14, no. 23: How to make a Peg that will exactly fit three different kinds of Holes.  "Let one of the holes be circular, the other square, and the third an oval; ...."  Solution is a cylinder whose height equals its diameter.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles.

No. 16, pp. 26 & 86.  Circle, square, triangle,  with discussion of the dimensions:  "a wedge, except that its base must be a circle".

No. 29, pp. 30 & 89-90.  Circle, oval, square.

Rational Recreations.  1824.  Feat 19, p. 66.  Circle, oval, square.

Endless Amusement II.  1826?  P. 62:  "To make a Peg that will exactly fit three different kinds of Holes."  Circle, oval, square.  c= Badcock.

The Boy's Own Book.  The triple accommodation.  1828: 419;  1828-2: 424;  1829 (US): 215;  1855: 570;  1868: 677.  Circle, square and triangle.

Young Man's Book.  1839.  Pp. 294-295.  Circle, oval, square.  Identical to Badcock.

Crambrook.  1843.  P. 5, no. 16: The Mathematical Paradox -- the Circle, Triangle, and Square.  Check??

Family Friend 3 (1850) 60 & 91.  Practical puzzle -- No. XII.  Circle, square, triangle.  This is repeated as Puzzle 16 -- Cylinder puzzle in (1855) 339 with solution in (1856) 28.

Magician's Own Book.  1857.  Prob. 21: The cylinder puzzle, pp. 273 & 296.  Circle, square, triangle.  = Book of 500 Puzzles, 1859, prob. 21, pp. 87 & 110.  = Boy's Own Conjuring Book, 1860, prob. 20, pp. 235 & 260.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 42, pp. 403 & 442.  Identical to Magician's Own Book, with diagram inverted.

F. J. P. Riecke.  Op. cit. in 4.A.1, vol. 1, 1867.  Art. 33: Die Ungula, pp. 58‑61.  Take a cylinder with equal height and diameter.  A cut from the diameter of one base which just touches the other base cuts off a piece called an ungula (Latin for claw).  He computes the volume as  4r3/3.  He then makes the symmetric cut to produce the circle, square, triangle shape, which thus has volume  (2π ‑ 8/3) r3.  Says he has seen such a shape and a board with the three holes as a child's toy.  Cf Simpson, 1745.

Magician's Own Book (UK version).  1871.  The round peg in the square hole:  To pass a cylinder through three different holes, yet to fill them entirely, pp. 49-50.  Circle, oval, rectangle;  circle & (isosceles) triangle.

Alfred Elliott.  Within‑Doors.  A Book of Games and Pastimes for the Drawing Room.  Nelson, 1872.  [Toole Stott 251.  Toole Stott 1030 is a 1873 ed.]  No. 4: The cylinder puzzle, pp. 27‑28 & 30‑31.  Circle, square, triangle.

Mittenzwey.  1880.  Prob. 257, pp. 46 & 97;  1895?: 286, pp. 50 & 99-100;  1917: 286, pp. 45 & 94-95.  Circle, square, triangle.

Will Shortz has a puzzle trade card with the circle, cross, square problem, c1884.

Tom Tit, vol. 2.  1892.  La cheville universelle, pp. 161-162.  = K, no. 28: The universal plug, pp. 72‑73.  = R&A, A versatile peg, p. 106.  Circle, square, triangle.

Handy Book for Boys and Girls.  Op. cit. in 6.F.3.  1892.  Pp. 238-242: Captain S's peg puzzle.  Circle, square, triangle.

Hoffmann.  1893.  Chap. X, no. 20: One peg to fit three holes, pp. 344 & 381‑382 = Hoffmann-Hordern, pp. 238-239, with photo.  Circle, square, triangle.  Photo on p. 239 shows two examples: one simply a wood board and pieces; the other labelled The Holes and Peg Puzzle, from Clark's Cabinet of Puzzles, 1880-1900, but this seems to be just a card box with the holes.

Williams.  Home Entertainments.  1914.  The plug puzzle, pp. 103-104.  Circle, square, triangle and rectangle with curved ends.  This is the only example of this four-fold form that I have seen.  Nice drawing of a board with the plug shown in each hole, except the curve on the sloping faces is not always drawn down to the bottom.

E. M. Wyatt.  Puzzles in Wood, 1928, op. cit. in 5.H.1. 

The "cross" plug puzzle, p. 17.  Square, circle and cross.

The "wedge" plug puzzle, p. 18.  Square, circle and triangle.

Perelman.  FMP.  c1935?  One plug for three holes;  Further "plug" puzzles, pp. 339‑340 & 346.  6 simple versions;  3 harder versions:  square, triangle, circle;  circle, square, cross;  triangle, square, tee.  The three harder versions are also in FFF, 1957: probs. 69-71, pp. 112 & 118-119;  1979: probs. 73‑75, pp. 137 & 144  = MCBF: probs. 73-75, pp. 134-135 & 142-143.

Anonymous [Antilog].  An elevation puzzle.  Eureka 19 (Mar 1957) 11 & 19.  Front and top views are a square with a square inside it.  What is the side view?  Gives two solutions.

Anonymous.  An elevation puzzle.  Eureka 21 (Oct 1958) 7 & 29.  Front is the lower half of a circle.  Plan (= top view) is a circle.  What is the side view?  Solution is a  V  shape, but it ought to be the other way up!  Nowadays, one can buy potato crisps (= potato chips) in this shape.

Joseph S. Madachy.  3‑D in 2‑D.  RMM 2 (Apr 1961) 51‑53  &  3 (Jun 1961) 47.  Discusses 3 view and 3 silhouette problems.

                    3 circular silhouettes, but not a sphere.

                    Square, circle, triangle.

Ernest R. Ranucci.  Non‑unique orthographic projections.  RMM 14 (Jan‑Feb 1964) 50.  3 views such that there are 10 different objects with these views.

Ripley's Puzzles and Games.  1966.  Pp. 18-19, item 1.  Same problem as Antilog, 1957.  Gives one solution.

Cedric A. B. Smith.  Simple projections.  MG 62 (No. 419) (Mar 1978) 19-25.  This is about how different projections affect one's recognition of what an object is.  He starts with an example with two views and the isometric projection which is very difficult to interpret.  He gives three other views, each of which is easily interpreted.

The Diagram Group.  The Family Book of Puzzles.  The Leisure Circle Ltd., Wembley, Middlesex, 1984.  Problem 114, with Solution at the back of the book.  Front view is a rectangle with an interior rectangle.  Side view is a rectangle with a rectangular notch on front side.  Solution is a short cylinder with a straight notch in it.  This is a fairly classic problem for engineers but I haven't seen it in print elsewhere.

Marek Penszko.  Polish your wits -- 3: Loop the loop.  Games 11:2 (Feb/Mar 1987) 28 & 58.  Draw lines on a glass cube to produce three given projections.  Problem asks for all three projections to be the same.

 

          6.W.  BURR PUZZLES

 

          When assembled, a burr looks like three sticks crossing orthogonally, forming a 'star' with six points at the vertices of an octahedron.  Slocum says Wyatt [Puzzles in Wood, 1928, op. cit. in 5.H.1] is the first to use the word 'burr'.  Collins, Book of Puzzles, 1927, p. 135, calls them "Cluster, Parisian or Gordian Knot Puzzles" and states: "it is believed that they were first made in Paris, if, indeed, they were not invented there."  Since about 1990, there has been considerable development in new types of burr which use plates or boards rather than sticks, or whose central volume is subdivided more (cf in 6.W.1).

          See S&B, pp. 62‑85.

          See also 6.BJ.

 

          6.W.1.         THREE PIECE BURR

 

          Most of these have three pieces which are rectangular in cross-section (1 x 3 x 5) with slots of the same size and some of the pieces have notches from the slot to the outside.  When one piece is pushed, it slides, revealing its notch.  When placed properly, this allows a second piece to slide off and out. 

          In the 1990s, a more elaborate type of three piece burr appeared.  These have three  3 x 3 x 5  pieces which intersect in a central  3 x 3 x 3  region.  Within this region, some of the unit cubes are not present, which allows sliding of the pieces.  Some versions of the puzzle permit twisting of pieces though this usually requires a bit of rounding of edges and the actual examples tend to break, so these are not as acceptable.

 

Crambrook.  1843.  P. 5, no. 4:  Puzzling Cross  3 pieces.  This seems likely to be a three piece burr, but perhaps is in 6.W.3 -- ??  It is followed by  "Maltese Cross  6 pieces".

Edward Hordern's collection has examples in ivory from 1850-1900.

Hoffmann.  1893.  Chap. III, no. 35: The cross‑keys or three‑piece puzzle, pp. 106 & 139 = Hoffmann-Hordern, pp. 104-105, with photo.  One piece has an extra small notch which does not appear in other versions where the dimensions are better chosen.  I have recently acquired an example which appears identical to the illustrations but does not have the extra notch - this came from a Jaques puzzle box, c1900, and Dalgety has several examples of such boxes with the solution, where the puzzle is named The Cross Keys Puzzle (cf discussion at the beginning of Section 11).  The photo on p. 105 is an assembled version, with verbal instructions, by Jaques & Son, 1880-1895 (but Jaques was producing them up to at least c1910).  Hordern Collection, p. 67, shows Le Noeud Mystérieux, 1880‑1905, with a pictorial solution and this does not have the extra notch.

Benson.  1904.  The cross keys puzzle, pp. 205‑206.

Pearson.  1907.  Part III, no. 56: The cross‑keys, pp. 56 & 127‑128.

Anon.  A puzzle in wood.  Hobbies 31 (No. 795) (7 Jan 1911) 345.  Three piece burr with small extra notch as in Hoffmann.

Anon.  Woodwork Joints.  Evans, London, (1918), 2nd ed., 1919.  [I have also seen a 4th ed., 1925, which is identical to the 2nd ed., except for advertising pages at the end.]  A mortising puzzle, pp. 197‑199.

Collins.  Book of Puzzles.  1927.  Pp. 136-137: The cross‑keys puzzle.

E. M. Wyatt.  Three piece cross.  Puzzles in Wood, 1928, op. cit. in 5.H.1, pp. 24‑25.

Arthur Mee's Children's Encyclopedia 'Wonder Box'.  The Children's Encyclopedia appeared in 1908, but versions continued until the 1950s.  This looks like 1930s??  3-Piece Mortise with thin pieces.

A. S. Filipiak.  Burr puzzle.  Mathematical Puzzles, 1942, op. cit. in 5.H.1, p. 101.

Dic Sonneveld seems to be the first to begin designing three piece burrs of the more elaborate style, perhaps about 1985.  Trevor Wood has made several examples for sale.

Bill Cutler.  Email announcement to NOBNET on 27 Jan 1999.  He has begun analysing the newer style of three piece burr, excluding twist moves.  His first stage has examined cases where the centre cube of the central region is occupied and the piece this central cube belongs to has no symmetry.  He finds  202 x 109  assemblies (I'm not sure if this is an exact figure) and there are  33  level-8 examples (i.e. where it takes  8  moves to remove the first piece);  6674  level-7 examples;  73362  level-6 examples.  He thinks this is about  70%  of the total and it is already about six times the number of cases considered for the six piece burr (see 6.W.2).

Bill Cutler.  Christmas letter of 4 Dec 1999.  Says he has completed the above analysis and found  25 x 1010  possibilities, which took 225 days on a workstation.  The most elaborate examples require 8 moves to get a piece out and there are 80 of these.  He used one for his IPP19 puzzle.  He has a website with many of his results on burrs, etc.:  www.billcutlerpuzzles.com .

 

          6.W.2.         SIX PIECE BURR  =  CHINESE CROSS

 

          The usual form of these has six sticks, 2 x 2 x 6 (or 8), which have various notches in them.  In the 1990s, new forms were introduced, using plates or boards.  One version makes an open frame shape, something like a  3 x 3 x 3  chessboard.  In the other, 1 x 4 x 6 boards are paired side by side and the result looks like a classic six-piece burr with the end rectangle divided lengthwise rather than crosswise.  See also 6.W.7.

 

Jurgis Baltrušaitis.  Anamorphoses  ou magie artificielle des effets merveilleux.  Olivier Perrin Éditeur, Paris, 1969.  On pp. 110-116 & 184 is a discussion of a 1698 engraving "L'Académie des Sciences et des Beaux Arts" by Sébastien Leclerc (or Le Clerc).  In the right foreground is an object looking like a six piece burr.  James Dalgety discusses this in his Latest news on oldest puzzles; Lecture to Second Meeting on the History of Recreational Mathematics, 1 Jun 1996.  This image also exists in a large painted version (950 x 480 mm) which is more precise and more legible in many details, so it is supposed that the engraving was done in conjunction with the painting.  Though it was normal for a notable painting to be turned into an engraving, the opposite sometimes happened and Leclerc was a famous engraver.  The painter is unknown.  The divisions between the pairs of pieces of the 'burr' are pretty clear in the engraving, but two of them are not visible in the painting.  The 'burr' is also not quite correctly drawn, but all in all, it seems pretty convincing.  James Dalgety was the first to discover this picture and he has a copy of the engraving, but has not been able to locate the painting, though it was in the Bernard Monnier Collection exhibited at the Musée des Arts Decoratifs in Paris in 1975/76.

Camille Frémontier.  Sébastien Leclerc and the British Encyclopeaedists.  Sphæra [Newsletter of the Museum of the History of Science, Oxford] 6 (Aut 1997) 6-7.  Discusses the Leclerc engraving which was used as the frontispiece to several encyclopedias, the earliest being Chambers Cyclopaedia of 1728.

Minguet.  1733.  Pp. 103-105 (1755: 51-52; 1822: 122-124; 1864: 103-104).  Pieces diagrammed.  One plain key piece.

Catel.  Kunst-Cabinet.  1790.  Die kleine Teufelsklaue, p. 10 & fig. 16 on plate I.  Figure shows it assembled and fails to draw one of the divisions between pieces.  Description says it is 6 pieces, 2 inches long, from plum wood and costs 3 groschen (worth about an English penny of the time).  (See also pp. 9-10, fig. 20 on plate I for Die grosse Teufelsklaue -- the 'squirrelcage'.)

Bestelmeier.  1801.  Item 147: Die kleine Teufelsklaue.  (Note -- there is another item 147 on the next plate.)  Only shows it assembled.  Brief text may be copying part of Catel.  See also the picture for item 1099 which looks like a six‑piece burr included in a set of puzzles.  (See also Item 142: Die grosse Teufelsklaue.)

Edward Hordern's collection has examples, called The Oak of Old England, from c1840.

Crambrook.  1843.  P. 5, no. 5:  Maltese Cross   6 [pieces], three sorts.  Not clear if these might be here or in 6.W.4 or 6.W.5 -- ??

Magician's Own Book.  1857.  Prob. 1: The Chinese cross, pp. 266-267 & 291.  One plain key piece.  Not the same as in Minguét.

Landells.  Boy's Own Toy-Maker.  1858.  Pp. 137-139.  Identical to Magician's Own Book.

Book of 500 Puzzles.  1859.  1: The Chinese cross, pp. 80-81 & 105.  Identical to Magician's Own Book.

A. F. Bogesen (1792‑1876).  In the Danish Technical Museum, Helsingør (= Elsinore) are a number of wooden puzzles made by him, including a 6 piece burr, a 12 piece burr, an Imperial Scale? and a complex (trick??) joint.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 23: The Chinese Cross, pp. 399 & 439.  Identical to Magician's Own Book, except one diagram in the solution omits two labels.

Boy's Own Conjuring Book.  1860.  Prob. 1: The Chinese cross, pp. 228 & 254.  Identical to Magician's Own Book.

Hoffmann.  1893.  Chap. III, no. 36: The nut (or six‑piece) puzzle, pp. 106 & 139‑140 = Hoffmann-Hordern, pp. 104-106.  Different pieces than in Minguét and Magician's Own Book.

Dudeney.  Prob. 473 -- Chinese cross.  Weekly Dispatch (23 Nov  &  7 Dec 1902), both p. 13.  "There is considerable variety in the manner of cutting out the pieces, and though the puzzle has been given in some of the old books, I have purposely presented it in a form that has not, I believe, been published."

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  "... the "Chinese Cross," a puzzle of undoubted Oriental origin that was formerly brought from China by travellers as a curiosity, but for a long time has had a steady sale in this country."

Wehman.  New Book of 200 Puzzles.  1908.  The Chinese cross, pp. 40-41.  = Magician's Own Book.

Dudeney.  The world's best puzzles.  1908.  Op. cit. in 2.  P. 779 shows a '"Chinese Cross" which ... is of great antiquity.'

Oscar W. Brown.  US Patent 1,225,760 -- Puzzle.  Applied: 27 Jun 1916;  patented: 15 May 1917.  3pp + 1p diagrams.  Coffin says this is the earliest US patent, with several others following soon after.

Anon.  Woodwork Joints, 1918, op. cit. in 6.W.1.  Eastern joint puzzle, pp. 196‑197: Two versions using different pieces.  Six‑piece joint puzzle, pp. 199‑200.  Another version.

Western Puzzle Works, 1926 Catalogue.  No. 86: 6 piece Wood Block.  Several other possible versions -- see 6.W.7.

E. M. Wyatt.  Six‑piece burr.  Puzzles in Wood, 1928, op. cit. in 5.H.1, pp. 27‑28.  Describes 17 versions from 13 types of piece.

A. S. Filipiak.  Mathematical Puzzles, 1942, op. cit. in 5.H.1, pp. 79‑87.  73 versions from 38 types of piece.

William H. [Bill] Cutler.  The six‑piece burr.  JRM 10 (1977‑78) 241‑250.  Complete, computer assisted, analysis, with help from T. H. O'Beirne and A. C. Cross.  Pieces are considered as 'notchable' if they can be made by a sequence of notches, which are produced by two saw cuts and then chiselling out the space between them.  Otherwise viewed, notches are what could be produced by a wide cutter or router.  There are  25  of these which can occur in solutions.  (In 1994, he states that there are a total of  59  notchable pieces and diagrams all of them.)  One can also have more general pieces with 'right-angle notches' which would require four chisel cuts -- e.g. to cut a single  1 x 1 x 1  piece out of a  2 x 2 x 8  rod.  Alternatively, one can glue cubes into notches.  There are  369  which can occur in solutions.  (In 1994, he states that there are  837  pieces which produce  2225  different oriented pieces, and he lists them all.)  He only considers solid solutions -- i.e. ones where there are no internal holes.  He finds and lists the  314  'notchable' solutions.  There are  119,979  general solutions.

C. Arthur Cross.  The Chinese Cross.  Pentangle, Over Wallop, Hants., UK, 1979.  Brief description of the solutions in the general case, as found by Cutler and Cross.

S&B, p. 83, describes holey burrs.

W. H. [Bill] Cutler.  Christmas letter, 1987.  Sketches results of his (and other's) search for holey burrs with notchable pieces.

Bill Cutler.  Holey 6‑Piece Burr!  Published by the author, Palatine, Illinois.  (1986);  with addendum, 1988, 48pp.  He is now permitting internal holes.  Describes holey burrs with notchable pieces, particularly those with multiple moves to release the first piece. 

Bill Cutler.  A Computer Analysis of All 6-Piece Burrs.  Published by the author, ibid., 1994.  86pp.  Sketches complete history of the project.  (I have included a few details in the description of his 1977/78 article, above.)  In 1987, he computed all the notchable holey solutions, using about 2 months of PC AT time, finding  13,354,991  assemblies giving  7.4 million solutions.  Two of these were level 10 -- i.e. they require 10 moves to remove the first piece (or pieces), but the highest level occurring for a unique solution was 5.  After that he started on the general holey burrs and estimated it would take 400 years of PC AT time -- running at 8 MHz.  After some development, the actual time used was about 62.5 PC AT years, but a lot of this was done on by Harry L. Nelson during idle time on the Crays at Lawrence Livermore Laboratories, and faster PCs became available, so the whole project only took about 2½ years, being completed in Aug 1990 and finding  35,657,131,235  assemblies.  He hasn't checked if all assemblies come apart fully, but he estimates there are 5.75 billion solutions.  He estimates the project used 45 times the computing power used in the proof of the Four Color Theorem and that the project would only take two weeks on the eight RS6000 workstations he now supervises.  Some 70,000 high-level solutions were specifically saved and can be obtained on disc from him.  The highest level found was 12 and the highest level for a unique solution was 10.  See 6.W.1 for a continuation of this work.  He has a website with many of his results on burrs, etc.:  www.billcutlerpuzzles.com .

Bill Cutler & Frans de Vreugd.  Information leaflet accompanying their separate IPP22 puzzles, 2002.  In 2001, they did an analysis of six-board burrs, of the type where the boards are paired side by side.  There are 4096 possible such boards, but only 219 usable boards occur.  They looked at all combinations of six of these and found 14,563,061,989 assemblies.  Of these, the highest level found was 13.

 

          6.W.3.         THREE PIECE BURR WITH IDENTICAL PIECES

 

          See S&B, p. 66.

 

Crambrook.  1843.  P. 5, no. 4:  Puzzling Cross  3 pieces.  This seems likely to be a three piece burr, but perhaps is in 6.W.1 -- ??  It is followed by  "Maltese Cross  6 pieces".

Wilhelm Segerblom.  Trick wood joining.  SA (1 Apr 1899) 196.

 

          6.W.4.         DIAGONAL SIX PIECE BURR  =  TRICK STAR

 

          This version often looks like a stellated rhombic dodecahedron.  It has two basic forms, one with a key piece;  the other with all pieces identical, which assembles as two groups of three.

          See S&B, p. 78.

 

Crambrook.  1843.  P. 5, no. 5:  Maltese Cross   6 [pieces], three sorts.  Not clear if these belong here or in 6.W.2 or 6.W.5 -- ??

The Youth's Companion.  1875.  [Mail order catalogue.]  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 19.  Star Puzzle.  The picture does not show which form it is.  Slocum's Compendium also shows this. 

Samuel P. Chandler.  US Patent 393,816 -- Puzzle.  Applied: 9 Mar 1888;  patented: 23 Apr 1888.  1p + 1p diagrams.  Coffin says this is the earliest version, but it is more complex than usual, with  12  pieces, and has a key piece.

John S. Pinnell.  US Patent 774,197 -- Puzzle.  Applied: 9 Oct 1902;  patented: 8 Nov 1904.  2pp + 2pp diagrams.  Coffin notes that this extends the idea to  102  pieces!

William E. Hoy.  US Patent 766,444 -- Puzzle‑Ball.  Applied: 16 Oct 1902;  patented: 2 Aug 1904.  2pp + 2pp diagrams.  Spherical version with a key piece.

George R. Ford.  US Patent 779,121 -- Puzzle.  Applied: 16 May 1904;  patented: 3 Jan 1905.  1p + 1p diagrams.  With square rods, all identical.  He shows assembly by inserting a last piece rather than joining two groups of three.

Anon.  Simple wood puzzle.  Hobbies 31 (No. 786) (5 Nov 1910) 127.  With key piece.

E. M. Wyatt.  Woodwork puzzles.  Industrial Arts Magazine 12 (1923) 326‑327.  Version with a key piece and square rods.

Collins.  Book of Puzzles.  1927.  The bonbon or nut puzzle, pp. 137-139. 

Iffland Frères (Lausanne).  Swiss Patent 245,402 -- Zusammensetzspiel.  Received: 19 Nov 1945;  granted: 15 Nov 1946;  published: 1 Jul 1947.  2pp + 1p diagrams.  Stellated rhombic dodecahedral version with a key piece.  (Coffin says this is the first to use this shape, although Slocum has a version c1875.)

 

          6.W.5.         SIX PIECE BURR WITH IDENTICAL PIECES

 

          One form has six identical pieces and all move outward or inward together.  Another form with flat notched pieces has one piece with an extra notch or an extended notch which allows it to fit in last, either by sliding or twisting, but this is not initially obvious.  This form is sometimes made with equal pieces so that it can only be assembled by force, perhaps after steaming, and it then makes an unopenable money box.  This might be considered under 11.M.

 

Edward Hordern's collection has a version with one piece a little smaller than the rest from c1800.

Crambrook.  1843.  P. 5, no. 5:  Maltese Cross   6 [pieces], three sorts.  Not clear if these belong here or in 6.W.2 or 6.W.4 -- ??

C. Baudenbecher catalogue, c1850s.  Op. cit. in 6.W.7.  This has an example of the six equal flat pieces making an unopenable(?) money box.

F. Chasemore.  Some mechanical puzzles.  In:  Hutchison; op. cit. in 5.A; 1891, chap. 70, part 1, pp. 571‑572.  Item 5: The puzzle box, p. 572.  Six U pieces make a uniformly expanding cubical box.

Hoffmann.  1893.  Chap. III, no.33: The bonbon nut puzzle, pp. 104 & 138 = Hoffmann‑Hordern, pp. 102-103, with photo.  One piece has an extra notch to simplify the assembly.  Photo on p. 103 shows an example, almost certainly by Jaques & Son, 1860-1895.

Burnett Fallow.  How to make a puzzle money-box.  The Boy's Own Paper 15 (No. 755) (1 Jul 1893) 638.  Equal flat notched pieces forced together to make an unopenable box.

Burnett Fallow.  How to make a puzzle picture-frame.  The Boy's Own Paper 16 (No. 815) (25 Aug 1894) 749.  Each corner has the same basic forced construction as used in the puzzle money-box.

Benson.  1904.  The bonbon nut puzzle, p. 204.

Bartl.  c1920.  Several versions on p. 306.

Western Puzzle Works, 1926 Catalogue.  Last page shows 20 Chinese Wood Block Puzzles, High Grade.  Some of these are of the present type.

Collins.  Book of Puzzles.  1927.  The bonbon or nut puzzle, pp. 137-139.  As in Hoffmann.

Iona & Robert Opie and Brian Alderson.  Treasures of Childhood.  Pavilion (Michael Joseph), London, 1989.  P. 158 shows a "cluster puzzle which Professor Hoffman [sic] names the 'Nut (or Six‑piece) Puzzle', but which is usually called 'The Maltese Puzzle'." 

 

          6.W.6.         ALTEKRUSE PUZZLE

 

William Altekruse.  US Patent 430,502 -- Block-Puzzle.  Applied: 3 Apr 1890;  patented: 17 Jun 1890.  1p + 1p diagrams.  Described in S&B, p. 72.  The standard version has 12 pieces, but variations discovered by Coffin have  14, 36 & 38  pieces.

Western Puzzle Works, 1926 Catalogue.  No. 112: 12 piece Wood Block.  Possibly Altekruse.

 

          6.W.7.  OTHER BURRS

 

          See also 6.BJ for other 3D dissections.  I have avoided repeating items, so 6.BJ should also be consulted if you are reading this section.

 

Catel.  Kunst-Cabinet.  1790.  Die grosse Teufelsklaue, pp. 9-10 & fig. 20 on plate I.  24 piece 'squirrel cage'.  Cost 16 groschen.

Bestelmeier.  1801.  Item 142: Die grosse Teufelsklaue.  The 'squirrelcage', identical to Catel, with same drawing, but reversed.  Text may be copying some of Catel.

C. Baudenbecher, toy manufacturer in Nuremberg.  Sample book or catalogue from c1850s.  Baudenbecher was taken over by J. W. Spear & Sons in 1919 and the catalogue is now in the Spear's Game Archive, Ware, Hertfordshire.  It comprises folio and double folio sheets with finely painted illustrations of the firm's products.  One whole folio page shows about 20 types of wooden interlocking puzzles, including most of the types mentioned elsewhere in this section and in 6.W.5 and 6.BJ.  Until I get a picture, I can't be more specific.

The Youth's Companion.  1875.  [Mail order catalogue.]  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 19.  Shows a 'woodchuck' type puzzle, called White Wood Block Puzzle, from The Youth's Companion, 1875.  I can't see how many pieces it has:  12 or 18??  Slocum's Compendium also shows this. 

Slocum.  Compendium.  Shows: "Mystery", Magic "Champion Puzzle" and "Puzzle of Puzzles" from Bland's Catalogue, c1890.

The first looks like a 6 piece burr with circular segments added to make it look like a ball.  So it may be a 6 piece burr in disguise.  See also Hoffmann, Chap. III, no. 38, pp. 107‑108 & 141‑142 = Hoffmann-Hordern, pp. 106-108  = Benson, p. 205.

The second is a six piece puzzle, but the pieces are flattish and it may be of the type described in 6.W.5. 

The third is complex, with perhaps 18 pieces.

Bartl.  c1920.  Several versions on pp. 306-307, including some that are in 6.W.5 and some 'Chinese block puzzles'.

Western Puzzle Works, 1926 Catalogue.  Shows a number of burrs and similar puzzles.

No. 86: 6 piece Wood Block.

No. 112: 12 piece Wood Block.  Possibly Altekruse.

No. 212: 11 piece Wood Block

The last page shows 20 Chinese Wood Block Puzzles, High Grade.  Some of these are burrs.

Collins.  Book of Puzzles.  1927.  Other cluster puzzles, pp. 139-142.  Describes and illustrates:  The cluster;  The cluster of clusters;  The gun cluster;  The point cluster;  The flat cluster;  The cluster (or secret) table;  The barrel;  The Ball;  The football.  All of these have a key piece.

Jan van de Craats.  Das unmögliche Escher-puzzle.  (Taken from:  De onmogelijke Escher-puzzle; Pythagoras (Amsterdam) (1988).)  Alpha 6 (or:  Mathematik Lehren / Heft 55 -- ??) (1992) 12-13.  Two Penrose tribars made into an impossible 5-piece burr.

 

          6.X.   ROTATING RINGS OF POLYHEDRA

 

          Generally, these have edge to edge joints.  'Jacob's ladder' joints are used by Engel -- see 11.L for other forms of this joint.

 

I am told these may appear in Fedorov (??NYS).

Max Brückner.  Vielecke und Vielfläche.  Teubner, Leipzig, 1900.  Section 162, pp. 215‑216 and Tafel VIII, fig. 4.  Describes rings of  2n  tetrahedra joined edge to edge, called stephanoids of the second order.  The figure shows the case  n = 5.

Paul Schatz.  UK Patent 406,680 -- Improvements in or relating to Boxes or Containers.  Convention date (Germany): 10 Dec 1931;  application date (in UK): 19 Jul 1932;  accepted: 19 Feb 1934.  6pp + 6pp diagrams.  Six and four piece rings of prisms which fold into a box.

Paul Schatz.  UK Patent 411,125 -- Improvements in Linkwork comprising Jointed Rods or the like.  Convention Date (Germany): 31 Aug 1931;  application Date (in UK): 31 Aug 1932;  accepted: 31 May 1934.  3p + 6pp diagrams.  Rotating rings of six tetrahedra and linkwork versions of the same idea, similar to Flowerday's Hexyflex.

Ralph M. Stalker.  US Patent 1,997,022 -- Advertising Medium or Toy.  Applied: 27 Apr 1933;  patented: 9 Apr 1935.  3pp + 2pp diagrams.  "... a plurality of tetrahedron members or bodies flexibly connected together."  Shows six tetrahedra in a ring and an unfolded pattern for such objects.  Shows a linear form with 14 tetrahedra of decreasing sizes.

Sidney Melmore.  A single‑sided doubly collapsible tessellation.  MG 31 (No. 294) (1947) 106.  Forms a Möbius strip of three triangles and three rhombi, which is basically a flexagon (cf 6.D).  He sees it has two distinct forms, but doesn't see the flexing property!!  He describes how to extend these hexagons into a tessellation which has some resemblance to other items in this section.

Alexander M. Shemet.  US Patent 2,688,820 -- Changeable Display Amusement Device.  Applied: 25 Jul 1950;  patented: 14 Sep 1954.  2pp + 2pp diagrams.  Basically a rotating ring of six tetrahedra, but says 'at least six'.  Gives an unfolded version or net for making it and a mechanism for flexing it continually.  Cites Stalker.

Wallace G. Walker invented his "IsoAxis" ® in 1958 while a student at Cranbrook Academy of Art, Michigan.  This is approximately a ring of ten tetrahedra.  He obtained a US Patent for it in 1967 -- see below.  In 1973(?) he sent an example to Doris Schattschneider who soon realised that the basic idea was a ring of tetrahedra and that Escher tessellations could be adapted to it.  They developed the idea into "M. C. Escher Kaleidocycles", published by Ballantine in 1977 and reprinted several times since.

Douglas Engel.  Flexahedrons.  RMM 11 (Oct 1962) 3‑5.  These have 'Jacob's ladder' hinges, not edge‑to‑edge hinges.  He says he invented these in Fall, 1961.  He formed rings of  4, 6, 7, 8  tetrahedra and used a diagonal joining to make rings of 4 and 6 cubes.

Wallace G. Walker.  US Patent 3,302,321 -- Foldable Structure.  Filed: 16 Aug 1963;  issued: 7 Feb 1967.  2pp + 6pp diagrams.

Joseph S. Madachy.  Mathematics on Vacation.  Op. cit. in 5.O, (1966), 1979.  Solid Flexagons, pp. 81‑84.  Based on Engel, but only gives the ring of 6 tetrahedra.

D. Engel.  Flexing rings of regular tetrahedra.  Pentagon 26 (Spring 1967) 106‑108.  ??NYS -- cited in Schaaf II 89 -- write Engel.

Paul Bethell.  More Mathematical Puzzles.  Encyclopædia Britannica International, London, 1967.  The magic ring, pp. 12-13.  Gives diagram for a ten-tetrahedra ring, all tetrahedra being regular.

Jan Slothouber  &  William Graatsma.  Cubics.  Octopus Press, Deventer, Holland, 1970.  ??NYS.  Presents versions of the flexing cubes and the 'Shinsei Mystery'.  [Jan de Geus has sent a photocopy of some of this but it does not cover this topic.]

Jan Slothouber.  Flexicubes -- reversible cubic shapes.  JRM 6 (1973) 39‑46.  As above.

Frederick George Flowerday.  US Patent 3,916,559 -- Vortex Linkages.  Filed: 12 Aug 1974 (23 Aug 1973 in UK);  issued: 4 Nov 1975.  Abstract + 2pp + 3pp diagrams.  Mostly shows his Hexyflex, essentially a six piece ring of tetrahedra, but with just four edges of each tetrahedron present.  He also shows his Octyflex which has eight pieces.  Text refers to any even number  ³ 6.

Naoki Yoshimoto.  Two stars in a cube (= Shinsei Mystery).  Described in Japanese in:  Itsuo Sakane; A Museum of Fun; Asahi Shimbun, Tokyo, 1977, pp. 208‑210.  Shown and pictured as Exhibit V‑1 with date 1972 in:  The Expanding Visual World -- A Museum of Fun; Exhibition Catalogue, Asahi Shimbun, Tokyo, 1979, pp. 102 & 170‑171.  (In Japanese).  ??get translated??

Lorraine Mottershead.  Investigations in Mathematics.  Blackwell, Oxford, 1985.  Pp. 63-66.  Describes Walkers IsoAxis and rotating rings of six and eight tetrahedra.

 

          6.Y.   ROPE ROUND THE EARTH

 

          The first few examples illustrate what must be the origin of the idea in more straightforward situations.

 

Lucca 1754.  c1330.  F. 8r, pp. 31‑32.  This mentions the fact that a circumference increases by  44/7  times the increase in the radius.

Muscarello.  1478. 

Ff. 932-93v, p. 220.  A circular garden has outer circumference  150  and the wall is    thick.  What is the inner circumference?  Takes  π  as  22/7.

F. 95r, p. 222.  The internal circumference of a tower is  20  and its wall is  3  thick.  What is the outer circumference?  Again takes  π  as  22/7.

Pacioli.  Summa.  1494.  Part II, f. 55r, prob. 33.  Florence is 5 miles around the inside.  The wall is    braccia wide and the ditch is  14  braccia wide -- how far is it around the outside?  Several other similar problems.

William Whiston.  Edition of Euclid, 1702.  Book 3, Prop. 37, Schol. (3.).  ??NYS -- cited by "A Lover" and Jackson, below.

"A Lover of the Mathematics."  A Mathematical Miscellany in Four Parts.  2nd ed., S. Fuller, Dublin, 1735.  The First Part is:  An Essay towards the Probable Solution of the Forty five Surprising PARADOXES, in GORDON's Geography, so the following must have appeared in Gordon.  Part I, no. 73, p. 56.  "'Tis certainly Matter of Fact, that three certain Travellers went a Journey, in which, Tho' their Heads travelled full twelve Yards more than their Feet, yet they all return'd alive, with their Heads on."

Carlile.  Collection.  1793.  Prob. XXV, p. 17.  Two men travel, one upright, the other standing on his head.  Who "sails farthest"?  Basically he compares the distance travelled by the head and the feet of the first man.  He notes that this argument also applies to a horse working a mill by walking in a circle; the outside of the horse travels about six times the thickness of the horse further than the inside on each turn.

Jackson.  Rational Amusement.  1821.  Geographical Paradoxes, no. 54, pp. 46 & 115-116.  "It is a matter of fact, that three certain travellers went on a journey, in which their heads travelled full twelve yards more than their feet; and yet, they all returned alive with their heads on."  Solution says this is discussed in Whiston's Euclid, Book 3, Prop. 37, Schol. (3.).  [This first appeared in 1702.]

K. S. Viwanatha Sastri.  Reminiscences of my esteemed tutor.  In:  P. K. Srinivasan, ed.; Ramanujan Memorial Volumes:  1: Ramanujan -- Letters and Reminiscences;  2: Ramanujan -- An Inspiration;  Muthialpet High School, Number Friends Society, Old Boys' Committee, Madras, 1968.  Vol. 1, pp. 89-93.  On p. 93, he relates that this was a favourite problem of his tutor, Srinivasan Ramanujan.  Though not clearly dated, this seems likely to be c1908-1910, but may have been up to 1914.  "Suppose we prepare a belt round the equator of the earth, the belt being    feet longer, and if we put the belt round the earth, how high will it stand?  The belt will stand  1  foot high, a substantial height."

Dudeney.  The paradox party.  Strand Mag. 38 (No. 228) (Dec 1909) 673‑674 (= AM, p. 139).

Anon.  Prob. 58.  Hobbies 30  (No. 773) (6 Aug 1910) 405  &  (No. 776) (27 Aug 1910) 448.  Double track circular railway, five miles long.  Move all rails outward one foot.  How much more material is needed?  Solution notes the answer is independent of the length.

Ludwig Wittgenstein was fascinated by the problem and used to pose it to students.  Most students felt that adding a yard to the rope would raise it from the earth by a negligible amount -- which it is, in relation to the size of the earth, but not in relation to the yard.  See:  John Lenihan;  Science in Focus;  Blackie, 1975, p. 39.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927.  Prob. 5, p. 87 & Answers p. 7.  A yard is added to a band around the earth.  Can you raise it 5 inches?  Answer notes the size of the earth is immaterial.

Collins.  Book of Puzzles.  1927.  The globetrotter's puzzle, pp. 68‑69.  If you walk around the equator, how much farther does your head go?

Abraham.  1933.  Prob. 33 -- A ring round the earth, pp. 12 & 24 (9 & 112).

Perelman.  FMP.  c1935??  Along the equator, pp. 342 & 349.  Same as Collins.

Sullivan.  Unusual.  1943.

Prob. 20: A global readjustment.  Take a wire around the earth and insert an extra 40 ft into it -- how high up will it be?

Prob. 23: Getting ahead.  If you walk around the earth, how much further does your head go than your feet?

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  Things are seldom what they seem -- No. 42a, 43, 44, pp. 50-51.  42a and 43 ask how much the radius increases for a yard gain of circumference.  No. 44 asks if we add a yard to a rope around the earth and then tauten it by pulling outward at one point, how far will that point be above the earth's surface?

Richard I Hess.  Puzzles from Around the World.  The author, 1997.  (This is a collection of 117 puzzles which he published in Logigram, the newsletter of Logicon, in 1984-1994, drawn from many sources.  With solutions.)  Prob. 28.  Consider a building 125 ft wide and a rubber band stretched around the earth.  If the rubber band has to stretch an extra 10 cm to fit over the building, how tall is the building?  He takes the earth's radius as  20,902,851 ft.  He gets three trigonometric equations and uses iteration to obtain  85.763515... ft.

Erwin Brecher  &  Mike Gerrard.  Challenging Science Puzzles.  Sterling, 1997.  [Reprinted by Goodwill Publishing House, New Delhi, India, nd [bought in early 2000]].  Pp. 38-39 & 77.  The M25 is a large ring road around London.  A man commutes from the south to the north and finds the distance is the same if goes by the east or the west, so he normally goes to the east in the morning and to the west in the evening.  Recalling that the English drive on the left, he realised that his right wheels were on the outside in both journeys and he worried that they were wear out sooner.  So he changed and drove both ways by the east.  But he then worried whether the wear on the tires was the same since the evening trip was on the outer lanes of the Motorway.

 

          6.Z.    LANGLEY'S ADVENTITIOUS ANGLES

 

          Let ABC be an isosceles triangle with  Ð B = Ð C = 80o.  Draw  BD  and  CE,  making angles  50o  and  60o  with the base.  Then  Ð CED = 20o.

 

JRM 15 (1982‑83) 150 cites Math. Quest. Educ. Times 17 (1910) 75.  ??NYS

Peterhouse and Sidney Entrance Scholarship Examination.  Jan 1916.  ??NYS.

E. M. Langley.  Note 644:  A Problem.  MG 11 (No. 160) (Oct 1922) 173.

Thirteen solvers, including Langley.  Solutions to Note 644.  MG 11 (No. 164) (May 1923) 321‑323.

Gerrit Bol.  Beantwoording van prijsvraag No. 17.  Nieuw Archief voor Wiskunde (2) 18 (1936) 14‑66.  ??NYS.  Coxeter (CM 3 (1977) 40) and Rigby (below) describe this.  The prize question was to completely determine the concurrent diagonals of regular polygons.  The  18‑gon is the key to Langley's problem.  However Bol's work was not geometrical.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Find the angle, pp. 86-87.  Short solution using law of sines and other simple trigonometric relations.

Colin Tripp.  Adventitious angles.  MG 59 (No. 408) (Jun 1975) 98‑106.  Studies when РCED  can be determined and all angles are an integral number of degrees.  Computer search indicates that there are at most  53  cases.

CM 3 (1977) 12  gives 1939 & 1950 reappearances of the problem and a 1974 variation.

D. A. Q. [Douglas A. Quadling].  The adventitious angles problem: a progress report.  MG 61 (No. 415) (Mar 1977) 55-58.  Reports on a number of contributions resolving the cases which Tripp could not prove.  All the work is complicated trigonometry -- no further cases have been demonstrated geometrically.

CM 4 (1978) 52‑53 gives more references.

D. A. Q. [Douglas A. Quadling].  Last words on adventitious angles.  MG 62 (No. 421) (Oct 1978) 174-183.  Reviews the history, reports on geometric proofs for all cases and various generalizations.

J[ohn]. F. Rigby.  Adventitious quadrangles: a geometrical approach.  MG 62 (No. 421) (Oct 1978) 183-191.  Gives geometrical proofs for almost all cases.  Cites Bol and a long paper of his own to appear in Geom. Dedicata (??NYS).  He drops the condition that  ABC  be isosceles.  His adventitious quadrangles correspond to Bol's triple intersections of diagonals of a regular  n-gon.

MS 27:3 (1994/5) 65  has two straightforward letters on the problem, which was mentioned in ibid. 27:1 (1994/5) 7.  One letter cites 1938 and 1955 appearances.  P. 66 gives another solution of the problem.  See next item.

Douglas Quadling.  Letter: Langley's adventitious angles.  MS 27:3 (1994/5) 65‑66.  He was editor of MG when Tripp's article appeared.  He gives some history of the problem and some life of Langley (d. 1933).  Edward Langley was a teacher at Bedford Modern School and the founding editor of the MG in 1894-1895.  E. T. Bell was a student of Langley's and contributed an obituary in the MG (Oct 1933) saying that Langley was the finest expositor he ever heard -- ??NYS.  Langley also had botanical interests and a blackberry variety is named for him.

 

          6.AA.          NETS OF POLYHEDRA

 

Albrecht Dürer.  Underweysung der messung mit dem zirckel uň [NOTE:  ň  denotes an  n  with an overbar.] richtscheyt, in Linien ebnen unnd gantzen corporen.  Nürnberg, 1525, revised 1538.  Facsimile of the 1525 edition by Verlag Dr. Alfons Uhl, Nördlingen, 1983.  German facsimile with English translation of the 1525 edition, with notes about the 1538 edition:  The  Painter's Manual; trans. by Walter L. Strauss; Abaris Books, NY, 1977.  Figures 29‑43 (erroneously printed 34) (pp. 316-347 in The Painter's Manual, Dürer's 1525 ff. M-iii-v - N-v-r) show nets and pictures of the regular polyhedra, an approximate sphere (16 sectors by 8 zones), truncated tetrahedron, truncated cube, cubo-octahedron, truncated octahedron, rhombi‑cubo-octahedron, snub cube, great rhombi-cubo-octahedron, truncated cubo‑octahedron (having a pattern of four triangles replacing each triangle of the cubo‑octahedron -- not an Archimedean solid) and an elongated hexagonal bipyramid (not even regular faced).  (See 6.AT.3 for more details.)  (Panofsky's biography of Dürer asserts that Dürer invented the concept of a net -- this is excerpted in The World of Mathematics I 618‑619.)  In the revised version of 1538, figure 43 is replaced by the icosi-dodecahedron and great rhombi-cubo-octahedron (figures 43 & 43a, pp. 414‑419 of The Painter's Manual) to make  9  of the Archimedean polyhedra.

Albrecht Dürer.  Elementorum Geometricorum (?) -- the copy of this that I saw at the Turner Collection, Keele, has the title page missing, but Elementorum Geometricorum is the heading of the first text page and appears to be the book's title.  This is a Latin translation of Unterweysung der Messung ....  Christianus Wechelus, Paris, 1532.  This has the same figures as the 1525 edition, but also has page numbers.  Liber quartus, fig. 29-43, pp. 145-158 shows the same material as in the 1525 edition. 

Cardan.  De Rerum Varietate.  1557, ??NYS  = Opera Omnia, vol. III, pp. 246-247.  Liber XIII.  Corpora, qua regularia diei solent, quomodo in plano formentur.  Shows nets of the regular solids, except the two halves of the dodecahedron have been separated to fit into one column of the text.

Barbaro, Daniele.  La Practica della Perspectiva.  Camillo & Rutilio Borgominieri, Venice, (1569);  facsimile by Arnaldo Forni, 1980, HB.  [The facsimile's TP doesn't have the publication details, but they are given in the colophon.  Various catalogues say there are several versions with dates on the TP and colophon varying independently between 1568 and 1569.  A version has both dates being 1568, so this is presumed to be the first appearance.  Another version has an undated title in an elaborate border and this facsimile must be from that version.]  Pp. 45-104 give nets and drawings of the regular polyhedra and 11 of the 13 Archimedean polyhedra -- he omits the two snub solids.

E. Welper.  Elementa geometrica, in usum geometriae studiosorum ex variis Authoribus collecta.  J. Reppius, Strassburg, 1620.  ??NYS -- cited, with an illustration of the nets of the octahedron, icosahedron and dodecahedron, in Lange & Springer Katalog 163 -- Mathematik & Informatik, Oct 1994, item 1350 & illustration on back cover, but the entry gives Trassburg.

Athanasius Kircher.  Ars Magna, Lucis et Umbrae.  Rome, 1646.  ??NX.  Has net of a rhombi-cuboctahedron.

Pike.  Arithmetic.  1788.  Pp. 458-459.  "As the figures of some of these bodies would give but a confused idea of them, I have omitted them; but the following figures, cut out in pasteboard, and the lines cut half through, will fold up into the several bodies."  Gives the regular polyhedra.

Dudeney.  MP.  1926.  Prob. 146: The cardboard box, pp. 58 & 149 (= 536, prob. 316, pp. 109 & 310).  All  11  nets of a cube.

Perelman.  FMP.  c1935?  To develop a cube, pp. 179 & 182‑183.  Asserts there are  10  nets and draws them, but two "can be turned upside down and this will add two more ...."  One shape is missing.  Of the two marked as reversible, one is symmetric, hence equal to its reverse, but the other isn't.

C. Hope.  The nets of the regular star‑faced and star‑pointed polyhedra.  MG 35 (1951) 8‑11.  Rather technical.

H. Steinhaus.  One Hundred Problems in Elementary Mathematics.  (As:  Sto Zadań, PWN -- Polish Scientific Publishers, Warsaw, 1958.)  Pergamon Press, 1963.  With a Foreword by M. Gardner; Basic Books, NY, 1964.  Problem 34: Diagrams of the cube, pp. 20 & 95‑96.  (Gives all  11  nets.)  Gardner (pp. 5‑6) refers to Dudeney and suggests the four dimensional version of the problem should be easy.

M. Gardner.  SA (Nov 1966) c= Carnival, pp. 41‑54.  Discusses the nets of the cube and the Answers show all  11  of them.  He asks what shapes these  11  hexominoes will form -- they cannot form any rectangles.  He poses the four dimensional problem;  the Addendum says he got several answers, no two agreeing.

Charles J. Cooke.  Nets of the regular polyhedra.  MTg 40 (Aut 1967) 48‑52.  Erroneously finds  13  nets of the octahedron.

Joyce E. Harris.  Nets of the regular polyhedra.  MTg 41 (Winter 1967) 29.  Corrects Cooke's number to  11.

A. Sanders  &  D. V. Smith.  Nets of the octahedron and the cube. MTg 42 (Spring 1968) 60‑63.  Finds  11  nets for the octahedron and shows a duality with the cube.

Peter Turney.  Unfolding the tesseract.  JRM 17 (1984‑85) 1‑16.  Finds  261  nets of the  4‑cube.  (I don't believe this has ever been confirmed.)

Peter Light  &  David Singmaster.  The nets of the regular polyhedra.  Presented at New York Acad. Sci. Graph Theory Day X, 213 Nov 1985.  In Notes from New York Graph Theory Day X, 23 Nov 1985; ed. by J. W. Kennedy & L. V. Quintas; New York Acad. Sci., 1986, p. 26.  Based on Light's BSc project in 1984-1984 under my supervision.  Shows there are  43,380  nets for the dodecahedron and icosahedron.  I may organize this into a paper, but several others have since verified the result.

 

          6.AB.           SELF‑RISING POLYHEDRA

 

H. Steinhaus.  Mathematical Snapshots.  Stechert, NY, 1938.  (= Kalejdoskop Matematyczny.  Książnica‑Atlas, Lwów and Warsaw, 1938, ??NX.)  Pp. 74-75 describes the dodecahedron and says to see the model in the pocket at the end, but makes no special observation of the self-rising property.  Described in detail with photographs in OUP, NY, eds:  1950: pp. 161-164;  1960: pp. 209‑212;  1969 (1983): pp. 196-198.

Donovan A. Johnson.  Paper Folding for the Mathematics Class.  NCTM, 1957, p. 29, section 66: Pop-up dodecahedron.

M. Kac.  Hugo Steinhaus -- a reminiscence and a tribute.  AMM 81 (1974) 572‑581.  Material is on pp. 580‑581, with picture on p. 581.

A pop‑up octahedron was used by Waddington's as an advertising insert in a trade journal at the London Toy Fair about 1981.  Pop-up cubes have also been used.

 

          6.AC.           CONWAY'S LIFE

 

          There is now a web page devoted to Life run by Bob Wainwright -- address is:

http://members.aol.com/life1ine/life/lifepage.htm [sic!].

 

M. Gardner.  Solitaire game of "Life".  SA (Oct 1970).  On cellular automata, self‑reproduction, the Garden‑of‑Eden and the game of "Life".  SA (Feb 1971).  c= Wheels, chap. 20-22.  In the Oct 1970 issue, Conway offered a $50 prize for a configuration which became infinitely large -- Bill Gosper found the glider gun a month later.  At G4G2, 1996, Bob Wainwright showed a picture of Gosper's telegram to Gardner on 4 Nov 1970 giving the coordinates of the glider gun.  I wasn't clear if Wainwright has this or Gardner still has it.

Robert T. Wainwright, ed. (12 Longue Vue Avenue, New Rochelle, NY, 10804, USA).  Lifeline (a newsletter on Life), 11 issues, Mar 1971 -- Sep 1973.  ??NYR.

John Barry.  The game of Life: is it just a game?  Sunday Times (London) (13 Jun 1971).  ??NYS -- cited by Gardner.

Anon.  The game of Life.  Time (21 Jan 1974).  ??NYS -- cited by Gardner.

Carter Bays.  The Game of Three‑dimensional Life.  Dept. of Computer Science, Univ. of South Carolina, Columbia, South Carolina, 29208, USA, 1986.  48pp.

A. K. Dewdney.  The game Life acquires some successors in three dimensions.  SA 256:2 (Feb 1987) 8‑13.  Describes Bays' work.

Bays has started a quarterly 3‑D Life Newsletter, but I have only seen one (or two?) issues.  ??get??

Alan Parr.  It's Life -- but not as we know it.  MiS 21:3 (May 1992) 12-15.  Life on a hexagonal lattice.

 

          6.AD.          ISOPERIMETRIC PROBLEMS

 

          There is quite a bit of classical history which I have not yet entered.  Magician's Own Book notes there is a connection between the Dido version of the problem and Cutting a card so one can pass through it, Section 6.BA.  There are several relatively modern surveys of the subject from a mathematical viewpoint -- I will cite a few of them.

 

Virgil.  Aeneid.  ‑19.  Book 1, lines 360‑370.  (p. 38 of the Penguin edition, translated by W. F. Jackson Knight, 1956.)  Dido came to a spot in Tunisia and the local chiefs promised her as much land as she could enclose in the hide of a bull.  She cut it into a long strip and used it to cut off a peninsula and founded Carthage.  This story was later adapted to other city foundations.  John Timbs; Curiosities of History; With New Lights; David Bogue, London, 1857, devotes a section to Artifice of the thong in founding cities, pp. 49-50, relating that in 1100, Hengist, the first Saxon King of Kent, similarly purchased a site called Castle of the Thong and gives references to Indian, Persian and American versions of the story as well as several other English versions.

Pappus.  c290.  Synagoge [Collection].  Book V, Preface, para. 1‑3, on the sagacity of bees.  Greek and English in SIHGM II 588‑593.  A different, abridged, English version is in HGM II 389‑390.

The Friday Night Book  (A Jewish Miscellany).  Soncino Press, London, 1933.  Mathematical Problems in the Talmud: Arithmetical Problems, no. 2, pp. 135-136.  A Roman Emperor demanded the Jews pay him a tax of as much wheat as would cover a space  40 x 40  cubits.  Rabbi Huna suggested that they request to pay in two instalments of  20 x 20  and the Emperor granted this.  [The Talmud was compiled in the period -300 to 500.  This source says he is one of the few mathematicians mentioned in the Talmud, but gives no dates and he is not mentioned in the EB.  From the text, the problem would seem to be sometime in the 1-5 C.]

The 5C Saxon mercenary, Hengist or Hengest, is said to have requested from Vortigern: "as much land as can be encircled by a thong".  He "then took the hide of a bull and cut it into a single leather thong.  With this thong he marked out a certain precipitous site, which he had chosen with the greatest possible cunning."  This is reported by Geoffrey of Monmouth in the 12C and this is quoted by the editor in:  The Exeter Book Riddles;  8-10C (the book was owned by Leofric, first Bishop of Exeter, who mentioned it in his will of 1072);  Translated and edited by Kevin Crossley-Holland;  (As: The Exeter Riddle Book, Folio Society, 1978, Penguin, 1979);  Revised ed., Penguin, 1993; pp. 101-102.

Lucca 1754.  c1330.  Ff. 8r‑8v, pp. 31‑33.  Several problems, e.g. a city  1 by 24  has perimeter  50  while a city  8 by 8  has perimeter  32  but is  8/3  as large;  stitching two sacks together gives a sack  4  times as big.

Calandri.  Arimethrica.  1491.  F. 97v.  Joining sacks which hold  9  and  16  yields a sack which holds  49!!

Pacioli.  Summa.  1494.  Part II, ff. 55r-55v.  Several problems, e.g. a cord of length 4 encloses 100 ducats worth, how much does a cord of length 10 enclose?  Also stitching bags together.

Buteo.  Logistica.  1559.  Prob. 86, pp. 298-299.  If 9 pieces of wood are bundled up by    feet of cord, how much cord is needed to bundle up 4 pieces?  5 pieces?

Pitiscus.  Trigonometria.  Revised ed., 1600, p. 223.  ??NYS -- described in: Nobuo Miura; The applications of trigonometry in Pitiscus: a preliminary essay; Historia Scientarum 30 (1986) 63-78.  A square of side  4  and triangle of sides  5, 5, 3  have the same perimeter but different areas.  Presumably he was warning people not to be cheated in this way.

J. Kepler.  The Six‑Cornered Snowflake, op. cit. in 6.AT.3.  1611.  Pp. 6‑11 (8‑19).  Discusses hexagons and rhombic interfaces, but only says "the hexagon is the roomiest" (p. 11 (18‑19)).

van Etten.  1624.  Prob. 90 (87).  Pp. 136‑138 (214‑218).  Compares fields  6 x 6  and  9 x 3.  Compares 4 sacks of diameter 1 with 1 sack of diameter 4.  Compares 2 water pipes of diameter 1 with 1 water pipe of diameter 2.

Ozanam.  1725.

Question 1, 1725: 327.  Question 3, 1778: 328;  1803: 325;  1814: 276;  1840: 141.  String twice as long contains four times as much asparagus.

Question 2, 1725: 328.  If a cord of length 10 encloses 200, how much does a cord of length 8 enclose?

Question 3, 1725: 328.  Sack 5 high by 4 across versus 4 sacks 5 high by 1 across.  c=  Q. 2, 1778: 328;  1803: 324;  1814: 276;  1840: 140-141, which has sack 4 high by 6 around versus two sacks 4 high by 3 around.

Question 4, 1725: 328‑329.  How much water does a pipe of twice the diameter deliver?

Les Amusemens.  1749. 

Prob. 211, p. 376.  String twice as long contains four times as much asparagus.

Prob. 212, p. 377.  Determine length of string which contains twice as much asparagus.

Prob. 223-226, pp. 386-389.  Various problems involving changing shape with the same perimeter.  Notes the area can be infinitely small.

Ozanam‑Montucla.  1778.

Question 1, 1778: 327;  1803: 323-324;  1814: 275-276;  1840: 140.  Square versus oblong field of the same circumference.

Prob. 35, 1778: 329-333;  1803: 326-330;  1814: 277-280;  1840: 141-143.  Les alvéoles des abeilles (On the form in which bees construct their combs).

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles.

No. 30, pp. 30 & 90.  Square field versus oblong (rectangular?) field of the same perimeter.

No. 31, pp. 30 & 90-91.  String twice as long contains four times as much asparagus.

Magician's Own Book (UK version).  1871.  To cut a card for one to jump through, p. 124, says:  "The adventurer of old, who, inducing the aborigines to give him as much land as a bull's hide would cover, and made it into one strip by which acres were enclosed, had probably played at this game in his youth."  See 6.BA.

M. Zacharias.  Elementargeometrie und elementare nicht-Euklidische Geometrie in synthetischer Behandlung.  Encyklopädie der Mathematischen Wissenschaften.  Band III, Teil 1, 2te Hälfte.  Teubner, Leipzig, 1914-1931.  Abt. 28: Maxima und Minima.  Die isoperimetrische Aufgabe.  Pp. 1118-1128.  General survey, from Zenodorus (-1C) and Pappus onward.

 

          6.AD.1.       LARGEST PARCEL ONE CAN POST

 

          New section.  I have just added the problem of packing a fishing rod as the diagonal of a box.  Are there older examples?

 

Richard A. Proctor.  Greatest content with parcels' post.  Knowledge 3 (3 Aug 1883) 76.  Height + girth  £  6 ft.  States that a cylinder is well known to be the best solution.  Either for a cylinder or a box, the optimum has  height = 2,  girth = 4,  with optimum volumes  2  and  8/π = 2.54... ft3.

R. F. Davis.  Letter:  Girth and the parcel post.  Knowledge 3 (17 Aug 1883) 109-110, item 897.  Independent discussion of the problem, noting that  length  £  3½ ft  is specified, though this doesn't affect the maximum volume problem.

H. F.  Letter:  Parcel post problem.  Knowledge 3 (24 Aug 1883) 126, item 905.  Suppose 'length' means "the maximum distance in a straight line between any two points on its surface".  By this he means the diameter of the solid.  Then the optimum shape is the intersection of a right circular cylinder with a sphere, the axis of the cylinder passing through the centre of the sphere, and this has the 'length' being the diameter of the sphere and the maximum volume is then  2⅓ ft3.

Algernon Bray.  Letter:  Greatest content of a parcel which can be sent by post.  Knowledge 3 (7 Sep 1883) 159, item 923.  Says the problem is easily solved without calculus.  However, for the box, he says "it is plain that the bulk of half the parcel will be greatest when [its] dimensions are equal".

Pearson.  1907.  Part II, no. 20: Parcel post limitations, pp. 118 & 195.  Length  £  3½ ft;  length + girth  £  6 ft.  Solution is a cylinder.

M. Adams.  Puzzle Book.  1939.  Prob. B.86: Packing a parcel, pp. 79 & 107.  Same as Pearson, but first asks for the largest box, then the largest parcel.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 18, pp. 27 & 89.  Ship a rifle about    yards long when the post office does not permit any dimension to be more than  1  yard.

T. J. Fletcher.  Doing without calculus.  MG 55 (No. 391) (Feb 1971) 4‑17.  Example 5, pp. 8‑9.  He says only that  length + girth  £  6 ft.  However, the optimal box has length 2, so the maximal length restriction is not critical.

I have looked at the current parcel post regulations and they say  length £ 1.5m  and  length + girth £ 3m,  for which the largest box is  1 x ½ x ½,  with volume  1/4 m3.  The largest cylinder has length  1  and radius  1/π  with volume  1/π m3.

I have also considered the simple question of a person posting a fishing rod longer than the maximal length by putting it diagonally in a box.  The longest rod occurs at a boundary maximum, at  3/2 x 3/4 x 0  or  3/2 x 0 x 3/4,  so one can post a rod of length  3Ö5/4  =  1.677... m, which is about  12%  longer than  1.5m.  In this problem, the use of a cylinder actually does worse!

 

          6.AE. 6"  HOLE THROUGH SPHERE LEAVES CONSTANT VOLUME

 

Hamnet Holditch.  Geometrical theorem.  Quarterly J. of Pure and Applied Math. 2 (1858) ??NYS, described by Broman.  If a chord of a closed curve, of constant length  a+b,  be divided into two parts of lengths  a, b  respectively, the difference between the areas of the closed curve, and of the locus of the dividing point as the chord moves around the curve, will be  πab.  When the closed curve is a circle and  a = b,  then this is the two dimensional version given by Jones, below.  A letter from Broman says he has found Holditch's theorem cited in 1888, 1906, 1975 and 1976.

Richard Guy (letter of 27 Feb 1985) recalls this problem from his schooldays, which would be late 1920s-early 1930s, and thought it should occur in calculus texts of that time, but could not find it in Lamb or Caunt.

Samuel I. Jones.  Mathematical Nuts.  1932.  P. 86.  ??NYS.  Cited by Gardner, (SA, Nov 1957) = 1st Book, chap. 12, prob. 7.  Gardner says Jones, p. 93, also gives the two dimensional version:  If the longest line that can be drawn in an annulus is  6"  long, what is the area of the annulus?

L. Lines.  Solid Geometry.  Macmillan, London, 1935;  Dover, 1965.  P. 101, Example 8W3:  "A napkin ring is in the form of a sphere pierced by a cylindrical hole.  Prove that its volume is the same as that of a sphere with diameter equal to the length of the hole."  Solution is given, but there is no indication that it is new or recent.

L. A. Graham.  Ingenious Mathematical Problems and Methods.  Dover, 1959.  Prob. 34: Hole in a sphere, pp. 23 & 145‑147.  [The material in this book appeared in Graham's company magazine from about 1940, but no dates are provided in the book.  (??can date be found out.)]

M. H. Greenblatt.  Mathematical Entertainments, op. cit. in 6.U.2, 1965.  Volume of a modified bowling ball, pp. 104‑105.

C. W. Trigg.  Op. cit. in 5.Q.  1967.  Quickie 217: Hole in sphere, pp. 59 & 178‑179.  Gives an argument based on surface tension to see that the ring surface remains spherical as the hole changes radius.  Problem has a 10" hole.

Andrew Jarvis.  Note 3235:  A boring problem.  MG 53 (No. 385) (Oct 1969) 298‑299.  He calls it "a standard problem" and says it is usually solved with a triple integral (??!!).  He gives the standard proof using Cavalieri's principle.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

Tangential chord, pp. 71-73.  10"  chord in an annulus.  What is the area of the annulus?  Does traditionally and then by letting inner radius be zero.

The hole in the sphere, pp. 87-88 & 177-178.  Bore a hole through a sphere so the remaining piece has half the volume of the sphere.  The radius of the hole is approx.  .61  of the radius of the sphere.

Another hole, pp. 89, 178 & 192.  6"  hole cut out of sphere.  What is the volume of the remainder?  Refers to the tangential chord problem.

Arne Broman.  Holditch's theorem: An introductory problem.  Lecture at ICM, Helsinki, Aug 1978.  Broman then sent out copies of his lecture notes and a supplementary letter on 30 Aug 1978.  He discusses Holditch's proof (see above) and more careful modern versions of it.  His letter gives some other citations.

 

          6.AF.           WHAT COLOUR WAS THE BEAR?

 

          A hunter goes 100 mi south, 100 mi east and 100 mi north and finds himself where he started.  He then shoots a bear -- what colour was the bear?

          Square versions:  Perelman;  Klamkin, Breault & Schwarz;  Kakinuma, Barwell & Collins;  Singmaster.

          I include other polar problems here.  See also 10.K for related geographical problems.

 

"A Lover of the Mathematics."  A Mathematical Miscellany in Four Parts.  2nd ed., S. Fuller, Dublin, 1735.  The First Part is:  An Essay towards the Probable Solution of the Forty five Surprising PARADOXES, in GORDON's Geography, so the following must have appeared in Gordon.  Part I, no. 10, p. 9.  "There is a particular Place of the Earth where the Winds (tho' frequently veering round the Compas) do always blow from the North Point."

Philip Breslaw (attrib.).  Breslaw's Last Legacy; or the Magical Companion: containing all that is Curious, Pleasing, Entertaining and Comical; selected From the most celebrated Masters of Deception: As well with Slight of Hand, As with Mathematical Inventions.  Wherein is displayed The Mode and Manner of deceiving the Eye; as practised by those celebrated Masters of Mirthful Deceptions.  Including the various Exhibitions of those wonderful Artists, Breslaw, Sieur, Comus, Jonas, &c.  Also the Interpretation of Dreams, Signification of Moles, Palmestry, &c.  The whole forming A Book of real Knowledge in the Art of Conjuration.  (T. Moore, London, 1784, 120pp.)  With an accurate Description of the Method how to make The Air Balloon, and inject the Inflammable Air.  (2nd ed., T. Moore, London, 1784, 132pp;  5th ed., W. Lane, London, 1791, 132pp.)  A New Edition, with great Additions and Improvements.  (W. Lane, London, 1795, 144pp.)  Facsimile from the copy in the Byron Walker Collection, with added Introduction, etc., Stevens Magic Emporium, Wichita, Kansas, 1997.  [This was first published in 1784, after Breslaw's death, so it is unlikely that he had anything to do with the book.  There were versions in  1784, 1791, 1792, 1793, 1794, 1795, 1800, 1806, c1809, c1810, 1811, 1824.  Hall, BCB 39-43, 46-51.  Toole Stott 120-131, 966‑967.  Heyl 35-41.  This book went through many variations of subtitle and contents -- the above is the largest version.].  I will cite the date as  1784?.

                    Geographical Paradoxes.

Paradox I, p. 35.  Where is it noon every half hour?  Answer: At the North Pole in Summer, when the sun is due south all day long, so it is noon every moment!

Paradox II, p. 36.  Where can the sun and the full moon rise at the same time in the same direction?  Answer: "Under the North Pole, the sun and the full moon, both decreasing in south declination, may rise in the equinoxial points at the same time; and under the North Pole, there is no other point of compass but south."  I think this means at the North Pole at the equinox.

Carlile.  Collection.  1793.  Prob. CXVI, p. 69.  Where does the wind always blow from the north?

Jackson.  Rational Amusement.  1821.  Geographical Paradoxes.

No. 7, pp. 36 & 103.  Where do all winds blow from the north?

No. 8, pp. 36 & 110.  Two places  100  miles apart, and the travelling directions are to go  50  miles north and  50  miles south.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:3 (Jul 1903) 246-247.  A safe catch.  Airship starts at the North Pole, goes south for seven days, then west for seven days.  Which way must it go to get back to its starting point?  No solution given.

Pearson.  1907.

Part II, no. 21: By the compass, pp. 18 & 190.  Start at North Pole and go  20  miles southwest.  What direction gets back to the Pole the quickest?  Answer notes that it is hard to go southwest from the Pole!

Part II, no. 15: Ask "Where's the north?" -- Pope, pp. 117 & 194.  Start  1200  miles from the North Pole and go  20  mph due north by the compass.  How long will it take to get to the Pole?  Answer is that you never get there -- you get to the North Magnetic Pole.

Ackermann.  1925.  P. 116.  Man at North Pole goes  20  miles south and  30  miles west.  How far, and in what direction, is he from the Pole?

Richard Guy (letter of 27 Feb 1985) recalls this problem (I think he is referring to the 'What colour was the bear' version) from his schooldays in the 1920s. 

H. Phillips.  Week‑End.  1932.  Prob. 8, pp. 12 & 188.  = his Playtime Omnibus, 1933, prob. 10: Popoff, pp. 54 & 237.  House with four sides facing south.

H. Phillips.  The Playtime Omnibus.  Faber & Faber, London, 1933.  Section XVI, prob. 11: Polar conundrum, pp. 51 & 234.  Start at the North Pole, go  40  miles South, then  30  miles West.  How far are you from the Pole.  Answer:  "Forty miles.  (NOT thirty, as one is tempted to suggest.)"  Thirty appears to be a slip for fifty??

Perelman.  FFF.  1934.  1957: prob. 6, pp. 14-15 & 19-20: A dirigible's flight;  1979: prob. 7, pp. 18-19 & 25-27: A helicopter's flight.  MCBF: prob. 7, pp. 18-19 & 25-26: A helicopter's flight.  Dirigible/helicopter starts at Leningrad and goes  500km  N,  500km  E,  500km  S,  500km  W.  Where does it land?  Cf Klamkin et seq., below.

Phillips.  Brush.  1936.  Prob. A.1: A stroll at the pole, pp. 1 & 73.  Eskimo living at North Pole goes 3 mi south and 4 mi east.  How far is he from home?

Haldeman-Julius.  1937.  No. 51: North Pole problem, pp. 8 & 23.  Airplane starts at North Pole, goes 30 miles south, then 40 miles west.  How far is he from the Pole?

J. R. Evans.  The Junior Week‑End Book.  Gollancz, London, 1939.  Prob. 9, pp. 262 & 268.  House with four sides facing south.

Leopold.  At Ease!  1943.  A helluva question!, pp. 10 & 196.  Hunter goes  10  mi south,  10  mi west, shoots a bear and drags it  10  mi back to his starting point.  What colour was the bear?  Says the only geographic answer is the North Pole.

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 5-6;  1945: 5-6;  1961: 15‑16.  He starts with the house which faces south on all sides.  Then he has a hunter that sees a bear 100 yards east.  The hunter runs 100 yards north and shoots south at the bear -- what colour ....  He then gives the three‑sided walk version, but doesn't specify the solution.

E. J. Moulton.  A speed test question; a problem in geography.  AMM 51 (1944) 216 & 220.  Discusses all solutions of the three-sided walk problem.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 50: A fine outlook, pp. 54-55.  House facing south on all sides used by an artist painting bears!

Leeming.  1946.  Chap. 3, prob. 32: What color was the bear?, pp. 33 & 160.  Man walks  10  miles south, then  10  miles west, where he shoots a bear.  He drags it  10  miles north to his base.  What color ....  He gives only one solution.

Darwin A. Hindman.  Handbook of Indoor Games & Contests.  (Prentice‑Hall, 1955);  Nicholas Kaye, London, 1957.  Chap. 16, prob. 4: The bear hunter, pp. 256 & 261.  Hunter surprises bear.  Hunter runs  200  yards north, bear runs  200  yards east, hunter fires south at bear.  What colour ....

Murray S. Klamkin, proposer;  D. A. Breault & Benjamin L. Schwarz, solvers.  Problem 369.  MM 32 (1958/59) 220  &  33 (1959/60) 110  &  226‑228.  Explorer goes  100  miles north, then east, then south, then west, and is back at his starting point.  Breault gives only the obvious solution.  Schwartz gives all solutions, but not explicitly.  Cf Perelman, 1934.

Benjamin L. Schwartz.  What color was the bear?.  MM 34 (1960) 1-4.  ??NYS -- described by Gardner, SA (May 1966) = Carnival, chap. 17.  Considers the problem where the hunter looks south and sees a bear  100  yards away.  The bear goes  100  yards east and the hunter shoots it by aiming due south.  This gives two extra types of solution.

Ripley's Puzzles and Games.  1966.  Pp. 18, item 5.  50 mi N,  1000 mi W,  10 mi S  to return to your starting point.   Answer only gives the South Pole, ignoring the infinitely many cases near the North Pole.  Looking at this made me realise that when the sideways distance is larger than the circumference of the parallel at that distance from the pole, then there are other solutions that start near the pole.  Here there are three solutions where one starts at distances  109.2, 29.6 or 3.05  miles from the South Pole, circling it 1, 2 or 3 times.

Yasuo Kakinuma, proposer;  Brian Barwell and Craig H. Collins, solvers.  Problem 1212 -- Variation of the polar bear problem.  JRM 15:3 (1982‑83) 222  &  16:3 (1983-84) 226‑228.  Square problem going one mile south, east, north, west.  Barwell gets the explicit quadratic equation, but then approximates its solutions.  Collins assumes the earth is flat near the pole.

David Singmaster.  Bear hunting problems.  Submitted to MM, 1986.  Finds explicit solutions for the general version of Perelman/Klamkin's problem.  [In fact, I was ignorant of (or had long forgotten) the above when I remembered and solved the problem.  My thanks to an editor (Paul Bateman ??check) for referring me to Klamkin.  The Kakinuma et al then turned up also.]  Analysis of the solutions leads to some variations, including the following.

David Singmaster.  Home is the hunter.  Man heads north, goes ten miles, has lunch, heads north, goes ten miles and finds himself where he started.

                    Used as:  Explorer's problem by Keith Devlin in his Micromaths Column; The Guardian (18 Jun 1987) 16  &  (2 Jul 1987) 16. 

                    Used by me as one of:  Spring term puzzles; South Bank Polytechnic Computer Services Department Newsletter (Spring 1989) unpaged [p. 15].

                    Used by Will Shortz in his National Public Radio program 6? Jan 1991. 

                    Used as:  A walk on the wild side, Games 15:2 (No. 104) (Aug 1991) 57 & 40.

                    Used as:  The hunting game, Focus 3 (Feb 1993) 77 & 98.

                    Used in my Puzzle Box column, G&P, No. 11 (Feb 1995) 19  &  No. 12 (Mar 1995) 41.

Bob Stanton.  The explorers.  Games Magazine 17:1 (No. 113) (Feb 1993) 61 & 43.  Two explorers set out and go  500  miles in each direction.  Madge goes N, W, S, E, while Ellen goes E, S, W, N.  At the end, they meet at the same point.  However, this is not at their starting point.  How come?  and how far are they from their starting point, and in what direction?  They are not near either pole.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995.  Chap. 11, prob. 9: What color was that bear? (A lesson in non-Euclidean geometry), pp. 97 & 185-191.  Camper walks south 2 km, then west 5 km, then north 2 km; how far is he from his starting point?  Solution analyses this and related problems, finding that the distance  x  satisfies  0 £ x £ 7.183,  noting that there are many minimal cases near the south pole and if one is between them, one gets a local maximum, so one has to determine one's position very carefully. 

David Singmaster.  Symmetry saves the solution.  IN: Alfred S. Posamentier & Wolfgang Schulz, eds.; The Art of Problem Solving: A Resource for the Mathematics Teacher; Corwin Press, NY, 1996, pp. 273-286.  Sketches the explicit solution to Klamkin's problem as an example of the use of symmetric variables to obtain a solution.

Anonymous.  Brainteaser B163 -- Shady matters.  Quantum 6:3 (Jan/Feb 1996) 15 & 48.  Is there anywhere on earth where one's shadow has the same length all day long?

 

          6.AG. MOVING AROUND A CORNER

 

          There are several versions of this.  The simplest is moving a ladder or board around a corner -- here the problem is two-dimensional and the ladder is thin enough to be considered as a line.  There are slight variations -- the corner can be at a  T  or  +  junction;  the widths of the corridors may differ;  the angle may not be a right angle;  etc.  If the object being moved is thicker -- e.g. a table -- then the problem gets harder.  If one can use the third dimension, it gets even harder.

 

H. E. Licks.  Op. cit. in 5.A, 1917.  Art. 110, p. 89.  Stick going into a circular shaft in the ceiling.  Gets  [h2/3 + d2/3)]3/2  for maximum length, where  h  is the height of the room and  d  is the diameter of the shaft.  "A simple way to solve a problem which has proved a stumbling block to many."

Abraham.  1933.  Prob. 82 -- Another ladder, pp. 37 & 45 (23 & 117).  Ladder to go from one street to another, of different widths.

E. H. Johnson, proposer;  W. B. Carver, solver.  Problem E436.  AMM 47 (1940) 569  &  48 (1941) 271‑273.  Table going through a doorway.  Obtains  6th order equation.

J. S. Madachy.  Turning corners.  RMM 5 (Oct 1961) 37,  6 (Dec 1961) 61  &  8 (Apr 1962) 56.  In 5, he asks for the greatest length of board which can be moved around a corner, assuming both corridors have the same width, that the board is thick and that vertical movement is allowed.  In 6, he gives a numerical answer for his original values and asserts the maximal length for planar movement, with corridors of width  w  and plank of thickness  t,  is  2 (wÖ2 ‑ t).  In vol. 8, he says no two solutions have been the same.

L. Moser, proposer;  M. Goldberg and J. Sebastian, solvers.  Problem 66‑11 -- Moving furniture through a hallway.  SIAM Review 8 (1966) 381‑382  &  11 (1969) 75‑78  &  12 (1970) 582‑586.  "What is the largest area region which can be moved through a "hallway" of width one (see Fig. 1)?"  The figure shows that he wants to move around a rectangular corner joining two hallways of width one.  Sebastian (1970) studies the problem for moving an arc.

J. M. Hammersley.  On the enfeeblement of mathematical skills ....  Bull. Inst. Math. Appl. 4 (1968) 66‑85.  Appendix IV -- Problems, pp. 83‑85, prob. 8, p. 84.  Two corridors of width  1  at a corner.  Show the largest object one can move around it has area  < 2 Ö2 and that there is an object of area  ³  π/2 + 2/π  =  2.2074. 

                    Partial solution by T. A. Westwell, ibid. 5 (1969) 80, with editorial comment thereon on pp. 80‑81.

T. J. Fletcher.  Easy ways of going round the bend.  MG 57 (No. 399) (Feb 1973) 16‑22.  Gives five methods for the ladder problem with corridors of different widths.

Neal R. Wagner.  The sofa problem.  AMM 83 (1976) 188‑189.  "What is the region of largest area which can be moved around a right‑angled corner in a corridor of width one?"  Survey.

R. K. Guy.  Monthly research problems, 1969‑77.  AMM 84 (1977) 807‑815.  P. 811 reports improvements on the sofa problem.

J. S. Madachy  &  R. R. Rowe.  Problem 242 -- Turning table.  JRM 9 (1976‑77) 219‑221.

G. P. Henderson, proposer;  M. Goldberg, solver;  M. S. Klamkin, commentator.  Problem 427.  CM 5 (1979) 77  &  6 (1979) 31‑32 & 49‑50.  Easily finds maximal area of a rectangle going around a corner.

Research news:  Conway's sofa problem.  Mathematics Review 1:4 (Mar 1991) 5-8 & 32.  Reports on Joseph Gerver's almost complete resolution of the problem in 1990.  Says Conway asked the problem in the 1960s and that L. Moser is the first to publish it.  Says a group at a convexity conference in Copenhagen improved Hammersley's results to  2.2164.  Gerver's analysis gives an object made up of 18 segments with area  2.2195.  The analysis depends on some unproven general assumptions which seem reasonable and is certainly the unique optimum solution given those assumptions.

A. A. Huntington.  More on ladders.  M500 145 (Jul 1995) 2-5.  Does usual problem, getting a quartic.  The finds the shortest ladder.  [This turns out to be the same as the longest ladder one can get around a corner from corridors of widths  w  and  h,  so 6.AG is related to 6.L.]

 

          6.AH. TETHERED GOAT

 

          A goat is grazing in a circular field and is tethered to a post on the edge.  He can reach half of the field.  How long is the rope?  There are numerous variations obtained by modifying the shape of the field or having buildings within it.  In recent years, there has been study of the form where the goat is tethered to a point on a circular silo in a large field -- how much area can he graze?

 

Upnorensis, proposer; Mr. Heath, solver.  Ladies Diary, 1748-49  =  T. Leybourn, II: 6-7, quest. 302.  [I have a reference to p. 41 of the Ladies' Diary.]  Circular pond enclosed by a circular railing of circumference 160 yards.  Horse is tethered to a post of the railing by a rope 160 yards long.  How much area can he graze?

Dudeney.  Problem 67: Two rural puzzles -- No. 67: One acre and a cow.  Tit‑Bits 33 (5 Feb  &  5 Mar 1898) 355  &  432.  Circular field opening onto a small rectangular paddock with cow tethered to the gate post so that she can graze over one acre.  By skilful choice of sizes, he avoids the usual transcendental equation.

Arc. [R. A. Archibald].  Involutes of a circle and a pasturage problem.  AMM 28 (1921) 328‑329.  Cites Ladies Diary and it appears that it deals with a horse outside a circle.

J. Pedoe.  Note 1477:  An old problem.  MG 24 (No. 261) (Oct 1940) 286-287.  Finds the relevant area by integrating in polar coordinates centred on the post.

A. J. Booth.  Note 1561:  On Note 1477.  MG 25 (No. 267) (Dec 1941) 309‑310.  Goat tethered to a point on the perimeter of a circle which can graze over  ½, ⅓, ¼  of the area.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.

No. 8: "Don't fence me in", pp. 87.  Equilateral triangular field of area 120.  Three goats tethered to the corners with ropes of length equal to the altitude.  Consider an area where  n  goats graze as contributing  1/n  to each goat.  What area does each goat graze over?

No. 53: Around the silo, pp. 71 & 112-113.  Goat tethered to the outside of a silo of diameter 20 by a rope of length  10π,  i.e. he can just get to the other side of the silo.  How big an area can he graze?  The curve is a semicircle together with two involutes of a circle, so the solution uses some calculus.

Marshall Fraser.  A tale of two goats.  MM 55 (1982) 221‑227.  Gives examples back to 1894.

Marshall Fraser.  Letter:  More, old goats.  MM 56 (1983) 123.  Cites Arc[hibald].

Bull, 1998, below, says this problem has been discussed by the Internet newsgroup  sci.math  some years previously.

Michael E. Hoffman.  The bull and the silo: An application of curvature.  AMM 105:1 (Jan 1998) ??NYS -- cited by Bull.  Bull is tethered by a rope of length  L  to a circular silo of radius  R.  If  L £ πR,  then the grazeable area is  L3/3R + πL2/2.  This paper considers the problem for general shapes. 

John Bull.  The bull and the silo.  M500 163 (Aug 1998) 1-3.  Improves Hoffman's solution for the circular silo by avoiding polar coordinates and using a more appropriate variable, namely the angle between the taut rope and the axis of symmetry. 

Keith Drever.  Solution 186.5 -- Horse.  M550 188 (Oct 2002) 12.  A horse is tethered to a point on the perimeter of a circular field of radius 1.  He can graze over all but  1/π  of the area.  How long is the rope?  This turns out to make the problem almost trivial -- the rope is  Ö2  long and the angle subtended at the tether is  π/2.

 

          6.AI.            TRICK JOINTS

 

          S&B, pp. 146‑147, show several types.

          These are often made in two contrasting woods and appear to be physically impossible.  They will come apart if one moves them in the right direction.  A few have extra complications.  The simplest version is a square cylinder with dovetail joints on each face -- called common square version below.  There are also cases where one thinks it should come apart, but the wood has been bent or forced and no longer comes apart -- see also 6.W.5.

 

See Bogesen in 6.W.2 for a possible early example.

Johannes Cornelus Wilhelmus Pauwels.  UK Patent 15,307 -- Improved Means of Joining or Fastening Pieces of Wood or other Material together, Applicable also as a Toy.  Applied: 9 Nov 1887;  complete specification: 9 Aug 1888;  accepted: 26 Oct 1888.  2pp + 1p diagrams.  It says Pauwels is a civil engineer of The Hague.  Common square version.

Tom Tit, vol. 2.  1892.  Assemblage paradoxal, pp. 231-232.  = K, no. 155: The paradoxical coupling, pp. 353‑354.  Common square version with instructions for making it by cutting the corners off a larger square.

Emery Leverett Williams.  The double dovetail and blind mortise.  SA (25 Apr 1896) 267.  The first is a trick T‑joint.

T. Moore.  A puzzle joint and how to make it.  The Woodworker 1:8 (May 1902) 172.  S&B, p. 147, say this is the earliest reference to the common square version -- but see Pauwels, above.  "... the foregoing joint will doubtless be well-known to our professional readers.  There are probably many amateur woodworkers to whom it will be a novelty."

Hasluck, Paul N.  The Handyman's Book.  Cassell, 1903;  facsimile by Senate (Tiger Books), Twickenham, London, 1998.  Pp. 220‑223 shows various joints.  Dovetail halved joint with two bevels, p. 222 & figs. 703-705 of pp. 221-222.  "... of but little practical value, but interesting as a puzzle joint."

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  Shows the common square version "given to me some ten years ago, but I cannot say who first invented it."  He previously published it in a newspaper.  ??look in Weekly Dispatch.

Samuel Hicks.  Kinks for Handy Men: The dovetail puzzle.  Hobbies 31 (No. 790) (3 Dec 1910) 248-249.  Usual square dovetail, but he suggests to glue it together!

Dudeney.  AM.  1917.  Prob. 424: The dovetailed block, pp. 145 & 249.  Shows the common square version -- "... given to me some years ago, but I cannot say who first invented it."  He previously published it in a newspaper.  ??as above

Anon.  Woodwork Joints, 1918, op. cit. in 6.W.1.  A curious dovetail joint, pp. 193, 195.  Common square version.  Dovetail puzzle joint, pp. 194‑195.  A singly mortised T‑joint, with an unmortised second piece.

E. M. Wyatt.  Woodwork puzzles.  Industrial‑Arts Magazine 12 (1923) 326‑327.  Doubly dovetailed tongue and mortise T‑joint called 'The double (?) dovetail'.

Sherman M. Turrill.  A double dovetail joint.  Industrial‑Arts Magazine 13 (1924) 282‑283.  A double dovetail right angle joint, but it leaves sloping gaps on the inside which are filled with blocks.

Collins.  Book of Puzzles.  1927.  Pp. 134‑135: The dovetail puzzle.  Common square version.

E. M. Wyatt.  Puzzles in Wood, 1928, op. cit. in 5.H.1.

The double (?) dovetail, pp. 44‑45.  Doubly dovetailed tongue and mortise T‑joint.

The "impossible" dovetail joint, p. 46.  Common square version.

Double‑lock dovetail joint, pp. 47‑49.  Less acceptable tricks for a corner joint.

Two‑way fanned half‑lap joint, pp. 49‑50.  Corner joint.

A. B. Cutler.  Industrial Arts and Vocational Education (Jan 1930).  ??NYS.  Wyatt, below, cites this for a triple dovetail, but I could not not find it in vols. 1‑40.

R. M. Abraham.  Prob. 225 -- Dovetail Puzzle.  Winter Nights Entertainments.  Constable, London, 1932, p. 131.  (= Easy‑to‑do Entertainments and Diversions with coins, cards, string, paper and matches; Dover, 1961, p. 225.)  Common square version.

Abraham.  1933.  Prob. 304 -- Hexagon dovetail;  Prob. 306 -- The triangular dovetail, pp. 142‑143 (100 & 102).

Bernard E. Jones, ed.  The Practical Woodworker.  Waverley Book Co., London, nd [1940s?].  Vol. 1: Lap and secret dovetail joints, pp. 281‑287.  This covers various secret joints -- i.e. ones with concealed laps or dovetails.  Pp. 286-287 has a subsection: Puzzle dovetail joints.  Common square version is shown as fig. 28.  A pentagonal analogue is shown as fig. 29, but it uses splitting and regluing to produce a result which cannot be taken apart.

E. M. Wyatt.  Wonders in Wood.  Bruce Publishing Co., Milwaukee, 1946.

Double‑double dovetail joint, pp. 26‑27.  Requires some bending.

Triple dovetail puzzle, pp. 28‑29.  Uses curved piece with gravity lock.

S&B, p. 146, reproduces the above Wyatt and shows a 1948 example.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  Dovetail deceptions, p. 64.  Common square version and a tapered T joint.

Allan Boardman.  Up and Down Double Dovetail.  Shown on p. 147 of S&B.  Square version with alternate dovetails in opposite directions.  This is impossible!

I have a set of examples which belonged to Tom O'Beirne.  There is a common square version and a similar hexagonal version.  There is an equilateral triangle version which requires a twist.  There is a right triangle version which has to be moved along a space diagonal!  [One can adapt the twisting method to  n-gons!]

Dick Schnacke (Mountain Craft Shop, American Ridge Road, Route 1, New Martinsville, West Virginia, 26155, USA) makes a variant of the common square version which has two dovetails on each face.  I bought one in 1994.

 

          6.AJ.  GEOMETRIC ILLUSIONS

 

          There are a great many illusions.  This will only give some general studies and some specific sources, though the sources of many illusions are unknown.

 

An exhibition by Al Seckel says there are impossible geometric patterns in a mosaic floor in the Roman villa at Fishbourne, c75, but it is not clear if this was intentional.

Anonymous 15C French illustrator of Giovanni Boccaccio, De Claris Mulieribus, MS Royal 16 Gv in the British Library.  F. 54v: Collecting cocoons and weaving silk.  ??NYS -- reproduced in: The Medieval Woman  An Illuminated Book of Postcards, HarperCollins, 1991.  This shows a loom(?) frame with uprights at each corner and the crosspieces joining the tops of the end uprights as though front and rear are reversed compared to the ground.

Seckel, 2002a, below, p. 25 (= 2002b, p. 175), says Leonardo da Vinci created the first anamorphic picture, c1500.

Giuseppe Arcimboldo (1537-1593).  One of his paintings shows a bowl of vegetables, but when turned over, it is a portrait.  Seckel, 2000, below, fig. 109, pp. 120 & 122 (= 2002b, fig, 107, pp. 118 & 120), noting that this is the first known invertible picture, but see next entry.

Topsy turvy coin, mid 16C.  Seckel, 2002a, fig. 65, p. 80 (omitted in 2002b), shows an example which shows the Pope, but turns around to show the Devil.  Inscription around edge reads:  CORVI MALUM OVUM MALII.

Robert Smith.  A Compleat System of Opticks in Four Books.  Cambridge, 1738.  He includes a picture of a distant windmill for which one cannot tell whether the sails are in front or behind the mill, apparently the first publication of this visual ambiguity.  ??NYS -- cited by: Nicholas J. Wade; Visual Allusions  Pictures of Perception; Lawrence Erlbaum Associates, Hove, East Sussex, 1990, pp. 17 & 25, with a similar picture.

L. A. Necker.  LXI. Observations on some remarkable optical phœnomena seen in Switzerland; and on an optical phœnomenon which occurs on viewing a figure of a crystal or geometrical solid.  Phil. Mag. (3) 1:5 (Nov 1832) 329-337.  This is a letter from Necker, written on 24 May 1832.  On pp. 336-337, Necker describes the visual reversing figure known as the Necker cube which he discovered in drawing rhomboid crystals.  This is also quoted in Ernst; The Eye Beguiled, pp. 23-24].  Richard L. Gregory [Mind in Science; Weidenfeld and Nicolson, London, 1981, pp. 385 & 594] and Ernst say that this was the first ambiguous figure to be described.

See Thompson, 1882, in 6.AJ.2, for illusions caused by rotations.

F. C. Müller-Lyer.  Optische Urtheilstusehungen.  Arch. Physiol. Suppl. 2 (1889) 263-270.  Cited by Gregory in The Intelligent Eye.  Many versions of the illusion.  But cf below.

Wehman.  New Book of 200 Puzzles.  1908.  The cube puzzle, p. 37.  A 'baby blocks' pattern of cubes, which appears to show six cubes piled in a corner one way and seven cubes the other way.  I don't recall seeing this kind of puzzle in earlier sources, though this pattern of rhombuses is common on cathedral floors dating back to the Byzantine era or earlier.

James Fraser.  British Journal of Psychology (Jan 1908).  Introduces his 'The Unit of Direction Illusion' in many forms.  ??NYS -- cited in his popular article in Strand Mag., see below.  Seckel, 2000, below, has several versions.  On p. 44, note to p. 9 (= 2002b, p. 44, note to p. 9), he says Fraser created a series of these illusions in 1906.

H. E. Carter.  A clever illusion.  Curiosities section, Strand Mag. 378 (No. 219) (Mar 1909) 359.  An example of Fraser's illusion with no indication of its source.

James Fraser.  A new illusion.  What is its scientific explanation?  Strand Mag. 38 (No. 224) (Aug 1909) 218-221.  Refers to the Mar issue and says he introduced the illusion in the above article and that the editors have asked him for a popular article on it.  16 illustrations of various forms of his illusion.

Lietzmann, Walther  &  Trier, Viggo.  Wo steckt der Fehler?  3rd ed., Teubner, 1923.  [The Vorwort says that Trier was coauthor of the 1st ed, 1913, and contributed most of the Schülerfehler (students' mistakes).  He died in 1916 and Lietzmann extended the work in a 2nd ed of 1917 and split it into Trugschlüsse and this 3rd ed.  There was a 4th ed., 1937.  See Lietzmann for a later version combining both parts.]  II. Täuschungen der Anschauung, pp. 7-13.

Lietzmann, Walther.  Wo steckt der Fehler?  3rd ed., Teubner, Stuttgart, (1950), 1953.  (Strens/Guy has 3rd ed., 1963.)  (See: Lietzmann & Trier.  There are 2nd ed, 1952??;  5th ed, 1969;  6th ed, 1972.  Math. Gaz. 54 (1970) 182 says the 5th ed appears to be unchanged from the 3rd ed.)  II. Täuschungen der Anschauung, pp. 15-25.  A considerable extension of the 1923 ed.

Williams.  Home Entertainments.  1914.  Colour discs for the gramophone, pp. 207-212.  Discusses several effects produced by spirals and eccentric circles on discs when rotated.

Gerald H. Fisher.  The Frameworks for Perceptual Localization. Report of MOD Research Project70/GEN/9617, Department of Psychology, University of Newcastle upon Tyne, 1968.  Good collection of examples, with perhaps the best set of impossible figures.

                    Pp. 42‑47 -- reversible perspectives.

                    Pp. 56‑65 -- impossible and ambiguous figures.

                    Appendix 6, p.190 -- 18 reversible figures.

                    Appendix 7, pp. 191‑192 -- 12 reversible silhouettes.

                    Appendix 8, p. 193 -- 12 impossible figures.

                    Appendix 14, pp. 202‑203 -- 72 geometrical illusions.

Harvey Long.  "It's All In How You Look At It".  Harvey Long & Associates, Seattle, 1972.  48pp collection of examples with a few references.

Bruno Ernst [pseud. of J. A. F. Rijk].  (Avonturen met Onmogelijke Figuren; Aramith Uitgevers, Holland, 1985.)  Translated as:  Adventures with Impossible Figures.  Tarquin, Norfolk, 1986.  Describes tribar and many variations of it, impossible staircase, two‑pronged trident.  Pp. 76‑77 reproduces an Annunciation of 14C in the Grote Kerk, Breda, with an impossible perspective.  P. 78 reproduces Print XIV of Giovanni Battista Piranesi's "Carceri de Invenzione", 1745, with an impossible 4‑bar.

Diego Uribe.  Catalogo de impossibilidades.  Cacumen (Madrid) 4 (No. 37) (Feb 1986) 9‑13.  Good summary of impossible figures.  15 references to recent work.

Bruno Ernst.  Escher's impossible figure prints in a new context.  In:  H. S. M. Coxeter, et al., eds.; M. C. Escher -- Art and Science; North‑Holland (Elsevier), Amsterdam, 1986, pp. 124‑134.  Pp. 128‑129 discusses the Breda Annunciation, saying it is 15C and quoting a 1912 comment by an art historian on it.  There is a colour reproduction on p. 394.  P. 130 shows and discusses briefly Bruegel's "The Magpie on the Gallows", 1568.  Pp. 130‑131 discusses and illustrates the Piranesi.

Bruno Ernst.  (Het Begoochelde Oog, 1986?.)  Translated by Karen Williams as:  The Eye Beguiled.  Benedikt Taschen Verlag, Köln, 1992.  Much expanded version of his previous book, with numerous new pictures and models by new artists in the field.  Chapter 6: Origins and history, pp. 68-93, discusses and quotes almost everything known.  P. 68 shows a miniature of the Madonna and Child from the Pericope of Henry II, compiled by 1025, now in the Bayersche Staatsbibliothek, Munich, which is similar in form to the Breda Annunciation (stated to be 15C).  (However, Seckel, 1997, below, reproduces it as 2© and says it is c1250.)  P. 69 notes that Escher invented the impossible cube used in his Belvedere.  P. 82 is a colour reproduction of Duchamp's 1916-1917 'Apolinère Enameled' - see 6.AJ.2.  Pp. 83-84 shows and discusses Piranesi.  Pp. 84-85 show and discuss Hogarth's 'False Perspective' of 1754.  Reproduction and brief mention of Brueghel (= Bruegel) on p. 85.  Discussion of the Breda Annunciation on pp. 85-86.  Pp. 87-88 show and discuss a 14C Byzantine Annunciation in the National Museum, Ochrid.  Pp. 88-89 show and discuss Scott Kim's impossible four-dimensional tribar. 

J. Richard Block  &  Harold E. Yuker.  Can You Believe Your Eyes?  Brunner/Mazel, NY, 1992.  Excellent survey of the field of illusions, classified into 17 major types -- e.g. ambiguous figures, unstable figures, ..., two eyes are better than one.  They give as much information as they can about the origins.  They give detailed sources for the following -- originals ??NYS.  These are also available as two decks of playing cards.

W. E. Hill.  My wife and my mother-in-law.  Puck, (6 Nov 1915) 11.  [However, Julian Rothenstein & Mel Gooding; The Paradox Box; Redstone Press, London, 1993; include a reproduction of a German visiting card of 1888 with a version of this illusion.  The English caption by James Dalgety is:  My Wife and my Mother-in-law.  Cf Seckel, 1997, below.]  Ernst, just above, cites Hill and says he was a cartoonist, but gives no source.  Long, above, asserts it was designed by E. G. Boring, an American psychologist.

G. H. Fisher.  Mother, father and daughter.  Amer. J. Psychology 81 (1968) 274-277.

G. Kanisza.  Subjective contours.  SA 234:4 (Apr 1976) 48-52.  (Kanisza triangles.)

Al Seckel, 1997.  Illusions in Art.  Two decks of playing cards in case with notes.  Deck 1 -- Classics.  Works from Roman times to the middle of the 20th Century.  Deck 2 -- Contemporary.  Works from the second half of the 20th Century.  Y&B Associates, Hempstead, NY, 1997.  This gives further details on some of the classic illusions -- some of this is entered above and in 6.AU and some is given below.

10¨: Rabbit/Duck.  Devised by Joseph (but notes say Robert) Jastrow, c1900.  Seckel, 2000, below, p. 159 (= 2002b, p. 156), says Joseph Jastrow, c1900.

10§: My Wife and My Mother-in-Law, anonymous, 1888.  However, in an exhibition, Seckel's text implies the 1888 German card doesn't have a title and the title first occurs on an 1890 US card.  Seckel, 2000, below, p. 122 (= 2002b, p. 120), says Boring took it from a popular 19C puzzle trading card.

Al Seckel, 2000.  The Art of Optical Illusions.  Carlton, 2000.  144 well reproduced illusions with brief notes.  All figures except 69-70 are included in Seckel, 2002b.

J. Richard Block.  Seeing Double  Over 200 Mind-Bending Illusions.  Routledge, 2002.  Update of Block & Yuker, 1992. 

Edgar Rubin.  Rubin's Vase.  1921.  This is the illusion where there appears to be a vase, but the outsides appear to be two face profiles.  [Pp. 8-11.]  But Seckel, 2000, above, p. 122 (= 2002b, p. 120), says Rubin's inspiration was a 19C puzzle card.

My wife and my mother-in-law.  P. 17 says Hill's version may derive from a late 1880s advertising postcard for Phenyo-Caffein (Worcester, Massachusetts), labelled 'My Girl & Her Mother', reproduced on p. 17.

P. 18 has G. H. Fisher's 1968 triple image, labelled 'Mother, Father and Daughter-in-Law'.

P. 44 says that Rabbit/Duck was devised by Joseph Jastrow in 1888.

Al Seckel, 2002a.  More Optical Illusions.  Carlton, 2002.  137 well reproduced illusions with brief notes, different than in Seckel, 2000, above.  All figures except 65-66, 86-87, 95-95, 137 are included in Seckel, 2002b, but with different figure and page numbers.

Al Seckel, 2002b.  The Fantastic World of Optical Illusions.  Carlton, 2002.  This is essentially a combination of Seckel, 2000, and Seckel, 2002a, both listed above.  The Introduction is revised.  Figures 69-70 of the first book and 65-66, 86-87, 94-95, 137 of the second book are omitted.  The remaining figures are then numbered consecutively.  The page of Further Reading in the first book is put at the end of this combined book.

 

Here I make some notes about origins of other illusions, but I have fewer details on these.

 

The Müller-Lyer Illusion --  <->  vs  >---<  was proposed by Zollner in 1859 and described by Johannes Peter Müller (1801-1858) & Lyer in 1889.  This seems to be a confusion, as the 1889 article is by F. C. Müller-Lyer, cf above.  Lietzmann & Trier, p. 7, date it as 1887.

The Bisection Illusion -- with a vertical segment bisecting a horizontal segment, but above it -- was described by Albert Oppel (1831-1865) and Wilhelm Wundt (1832-1920) in 1865.

Zollner's Illusion -- parallel lines crossed by short lines at  45o,  alternately in opposite directions -- was noticed by Johann K. F. Zollner (1834-1882) on a piece of fabric with a similar design.

Hering's Illusion -- with parallel lines crossed by numerous lines through a point between the lines -- was invented by Ewals Hering (1834-1918) in 1860.

 

          6.AJ.1          TWO PRONGED TRIDENT

 

          I have invented this name as it is more descriptive than any I have seen.  The object or a version of it is variously called:  Devil's Fork;  Three Stick Clevis;  Widgit; Blivit; Impossible Columnade;  Trichometric Indicator Support;  Triple Encabulator for Tuned Manifold;  Hole Location Gage;  Poiyut;  Triple-Pronged Fork with only Two Branches;  Old Roman Pitchfork.

 

Oscar Reutersvård.  Letters quoted in Ernst, 1992, pp. 69-70, says he developed an equivalent type of object, which he calls impossible meanders, in the 1930s.

R. L. Gregory says this is due to a MIT draftsman (= draughtsman) about 1950??

California Technical Industries.  Advertisement.  Aviation Week and Space Technology 80:12 (23 Mar 1964) 5.  Standard form.  (I wrote them but my letter was returned 'insufficient address'.)

Hole location gage.  Analog Science Fact • Science Fiction 73:4 (Jun 1964) 27.  Classic Two pronged trident, with some measurements given.  Editorial note says the item was 'sent anonymously for some reason' and offers the contributor $10 or a two year subscription if he identifies himself.  (Thanks to Peter McMullen for the Analog items, but he doesn't recall the contributor ever being named.)

Edward G. Robles, Jr.  Letter (Brass Tacks column).  Analog Science Fact • Science Fiction 74:4 (Dec 1964) 4.  Says the Jun 1964 object is a "three-hole two slot BLIVIT" and was developed at JPL (Jet Propulsion Laboratory, Pasadena) and published in their Goddard News.  He provides a six-hole five-slot BLIVIT, but as the Editor comments, it 'lacks the classic simple elegance of the Original.'  However, a letter of inquiry to JPL resulted in an email revealing that Goddard News is not their publication, but comes from the Goddard Space Flight Center.  I have had a response from Goddard, ??NYR.

D. H. Schuster.  A new ambiguous figure: a three‑stick clevis.  Amer. J. Psychol. 77 (1964) 673.  Cites Calif. Tech. Ind. ad.  [Ernst, 1992, pp. 80-81 reproduces this article.]

Mad Magazine.  No. 93 (Mar 1965).  (I don't have a copy of this -- has anyone got one for sale?)  Cover by Norman Poiyut (?) shows the figure and it is called a poiyut.  Miniature reproduction in:  Maria Reidelbach; Completely Mad -- A History of the Comic Book and Magazine; Little, Brown & Co., Boston, 1991, p. 82.  Shows a standard version.  Al Seckel says they thought it was an original idea and they apologised in the next issue -- to all of the following!  I now have the relevant issue, No. 95 (Jun 1965) and p. 2 has 15 letters citing earlier appearances in Engineering Digest, The Airman (official journal of the U.S. Airforce), Analog, Astounding Science Fact -- Science Fiction (Jun 1964, see above), The Red Rag (engineering journal at the University of British Columbia), Society of Automotive Engineers Journal (designed by by Gregory Flynn Jr. of General Motors as Triple Encabulator Tuned Manifold),  Popular Mechanics, Popular Science (Jul 1964), Road & Track (Jun 1964).  Other letters say it was circulating at:  the Engineering Graphics Lab of the University of Minnesota at Duluth;  the Nevada Test Site;  Eastman Kodak (used to check resolution);  Industrial Camera Co. of Oakland California (on their letterhead).  Two letters give an impossible crate and an impossible rectangular frame (sort of a Penrose rectangle).

Sergio Aragones.  A Mad look at winter sports.  Mad Magazine (?? 1965);  reprinted in:  Mad Power;  Signet, NY, 1970, pp. 120‑129.  P. 124 shows a standard version.

Bob Clark, illustrator.  A Mad look at signs of the times.  Loc. cit. under Aragones, pp. 167‑188.  P. 186 shows standard version.

Reveille (a UK weekly magazine) (10 Jun 1965).  ??NYS -- cited by Briggs, below -- standard version.

Don Mackey.  Optical illusion.  Skywriter (magazine of North American Aviation) (18 Feb 1966).  ??NYS -- cited by Conrad G. Mueller et al.; Light and Vision; Time-Life Books Pocket Edition, Time-Life International, Netherlands, 1969, pp. 171 & 190.  Standard version with nuts on the ends.

Heinz Von Foerster.  From stimulus to symbol: The economy of biological computation.  IN:  Sign Image Symbol;  ed. Gyorgy Kepes;  Studio Vista, London, 1966, pp. 42-60.  On p. 55, he shows the "Triple-pronged fork with only two branches" and on p. 54, he notes that although each portion is correct, it is impossible overall, but he gives no indication of its history or that it is at all new.

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966.  Pp. 17-18 shows the unnamed trident in a version from Adcock & Shipley (Sales) Ltd., machine tool makers in Leicester.  Cites Reveille, above.  Standard versions.

Harold Baldwin.  Building better blivets.  The Worm Runner's Digest 9:2 (1967) 104‑106.  Discusses relation between numbers of slots and of prongs.  Draws a three slot version and 2 and 4 way versions.

Charlie Rice.  Challenge!  Op. cit. in 5.C.  1968.  P. 10 shows a six prong, four slot version, called the "Old Roman Pitchfork".

Roger Hayward.  Blivets; research and development.  The Worm Runner's Digest 10 (Dec 1968) 89‑92.  Several fine developments, including two interlaced frames and his monumental version.  Cites Baldwin.

M. Gardner.  SA (May 1970) = Circus, pp. 3‑15.  Says this became known in 1964 and cites Mad & Hayward, but not Schuster.

D. Uribe, op. cit. above, gives several variations.

 

          6.AJ.2.         TRIBAR AND IMPOSSIBLE STAIRCASE

 

Silvanus P. Thompson.  Optical illusions of motion.  Brain 3 (1882) 289-298.  Hexagon of non‑overlapping circles. 

Thomas Foster.  Illusions of motion and strobic circles.  Knowledge 1 (17 Mar 1882) 421-423.  Says Thompson exhibited these illusions at the British Association meeting in 1877.

Pearson.  1907.  Part II, no. 3: Whirling wheels, p. 3.  Gives Thompson's form, but the wheels are overlapping, which makes it look a bit like an ancestor of the tribar.

Marcel Duchamp (1887-1968).  Apolinère Enameled.  A 'rectified readymade' of 1916-1917 which turned a bedframe in an advertisement for Sapolin Enamel into an impossible figure somewhat like a Penrose Triangle and a square version thereof.  A version is in the Philadelphia Museum of Art and is reproduced and discussed in Ernst; The Eye Beguiled, p. 82.  (Duchamp's 'readymades' were frequently reproduced by himself and others, so there may be other versions of this.)

Oscar Reutersvård.  Omöjliga Figure [Impossible Figures -- In Swedish].  Edited by Paul Gabriel.  Doxa, Lund, (1982);  2nd ed., 1984.  This seems to be the first publication of his work, but he has been exhibiting since about 1960 and some of the exhibitions seem to have had catalogues.  P. 9 shows and discusses his Opus 1 from 1934, which is an impossible tribar made from cubes.  (Reproduced in Ernst, 1992, p. 69 as a drawing signed and dated 1934.  Ernst quotes Reutersvård's correspondence which describes his invention of the form while doodling in Latin class as a schoolboy.  A school friend who knew of his work showed him the Penroses' article in 1958 -- at that time he had drawn about 100 impossible objects -- by 1986, he had extended this to some 2500!)  He has numerous variations on the tribar and the two‑pronged trident.  An exhibition by Al Seckel says Reutersvård had produced some impossible staircases, e.g. 'Visualized Impossible Bach Scale', in 1936-1937, but didn't go far with it until returning to the idea in 1953.

Oscar Reutersvård.  Swedish postage stamps for 25, 50, 75 kr.  1982, based on his patterns from the 1930s.  The 25 kr. has the tribar pattern of cubes which he first drew in 1934.  (Also the 60 kr.??)

L. S. & R. Penrose.  Impossible objects: A special type of visual illusion.  British Journal of Psychology 49 (1958) 31‑33.  Presents tribar and staircase.  Photo of model staircase, which Lionel Penrose had made in 1955.  [Ernst, 1992, pp. 71-73, quotes conversation with Penrose about his invention of the Tribar and reproduces this article.  Penrose, like the rest of us, only learned about Reutersvård's work in the 1980s.]

Anon.(?)  Don't believe it.  Daily Telegraph (24 Mar 1958) ?? (clipping found in an old book).  "Three pages of the latest issue of the British Journal of Psychology are devoted to "Impossible Objects.""  Shows both the tribar and the staircase.

M. C. Escher.  Lithograph:  Belvedere.  1958.

L. S. & R. Penrose.  Christmas Puzzles.  New Scientist (25 Dec 1958) 1580‑1581 & 1597.  Prob. 2: Staircase for lazy people.

M. C. Escher.  Lithograph:  Ascending and Descending.  1960.

M. C. Escher.  Lithograph:  Waterfall.  1961.

Oscar Reutersvård, in 1961, produced a triangular version of the impossible staircase, called 'Triangular Fortress without Highest Level'.

Joseph Kuykendall.  Letter.  Mad Magazine 95 (Jun 1965) 2.  An impossible frame, a kind of Penrose rectangle.

S. W. Draper.  The Penrose triangle and a family of related figures.  Perception 7 (1978) 283‑296.  ??NYS -- cited and reproduced in Block, 2002, p. 48.  A Penrose rectangle.

Uribe, op. cit. above, gives several variations, including a perspective tribar and Draper's rectangle.

Jan van de Craats.  Das unmögliche Escher-puzzle.  (Taken from:  De onmogelijke Escher-puzzle; Pythagoras (Amsterdam) (1988).)  Alpha 6 (or:  Mathematik Lehren / Heft 55 -- ??) (1992) 12-13.  Two Penrose tribars made into an impossible 5-piece burr.

 

          6.AJ.3.         CAFÉ WALL ILLUSION

 

          This is the illusion seen in alternatingly coloured staggered brickwork where the lines of bricks distinctly seem tilted.  I suspect it must be apparent in brickwork going back to Roman times.

 

The illusion is apparent in the polychrome brick work on the side wall inside Keble College Chapel, Oxford, by William Butterfield, completed in 1876 [thanks to Deborah Singmaster for observing this].

Lietzmann & Trier, op. cit. at 6.AJ, 1923.  Pp. 12-13 has a striking version of this, described as a 'Flechtbogen der Kleinen'.  I can't quite translate this -- Flecht is something interwoven but Bogen could be a ribbon or an arch or a bower, etc.  They say it is reproduced from an original by Elsner.  See Lietzmann, 1953.

Ogden's Optical Illusions.  Cigarette card of 1927.  No. 5.  Original ??NYS -- reproduced in:  Julian Rothenstein & Mel Gooding; The Paradox Box; Redstone Press, London, 1993  AND in their:  The Playful Eye; Redstone Press, London, 1999, p. 56.  Vertical version of this illusion.

B. K. Gentil.  Die optische Täuschung von Fraser.  Zeitschr. f. math. u. naturw. Unterr. 66 (1935) 170 ff.  ??NYS -- cited by Lietzmann.

Nelson F. Beeler & Franklyn M. Branley.  Experiments in Optical Illusion.  Ill. by Fred H. Lyon.  Crowell, 1951, p. 42, fig. 39, is a good example of the illusion.

Lietzmann, op. cit. at 6.AJ, 1953.  P. 23 is the same as above, but adds a citation to Gentil, listed above.

Leonard de Vries.  The Third Book of Experiments.  © 1965, probably for a Dutch edition.  Translated by Joost van de Woestijne.  John Murray, 1965; Carousel, 1974.  Illusion 10, pp. 58-59, has a clear picture and a brief discussion.

Richard L. Gregory & Priscilla Heard.  Border locking and the café wall illusion.  Perception 8 (1979) 365‑380.  ??NYS -- described by Walker, below.  [I have photos of the actual café wall in Bristol.]

Jearl Walker.  The Amateur Scientist: The café‑wall illusion, in which rows of tiles tilt that should not tilt at all.  SA 259:5 (Nov 1988) 100‑103.  Good summary and illustrations.

 

          6.AJ.4.         STEREOGRAMS

 

          New section, due to reading Glass's assertion as to the inventor, who is different than other names that I have seen.

 

Don Glass, ed.  How Can You Tell if a Spider is Dead? and More Moments of Science.  Indiana Univ Press, Bloomington, Indiana, 1996.  Now you see it, now you don't, pp. 131-132.  Asserts that Christopher Tyler, of the Smith-Kettlewell Eye Research Institute, San Francisco, is the inventor of stereograms.

 

          6.AJ.5.         IMPOSSIBLE CRATE.

 

          This is like a Necker Cube where all the edges are drawn as wooden slats in an impossible configuration. 

 

Escher.  Man with Cuboid, which is essentially a detail from Belvedere, both 1958, are apparently the first examples of this impossible object.

Chuck Mathias.  Letter  Mad Magazine 95 (Jun 1965) 2.  Gives an impossible crate.

Jerry Andrus developed his actual model in 1981 and it appeared on the cover of Omni in 1981.  But Al Seckel's exhibition says the first physical example was The Feemish Crate, due to C. F. Cochran.

Seckel, 2002a, figs. 27 A&B, pp. 36-37 (= 2002b, figs. 169 A&B, pp. 186-187), shows and discusses Andrus' crate from two viewpoints.

 

          6.AK. POLYGONAL PATH COVERING  N x N LATTICE OF POINTS, QUEEN'S TOURS, ETC.

 

          For magic circuits, see 7.N.4.

3x3 problem:  Loyd (1907),  Pearson,  Anon.,  Bullivant,  Goldston,  Loyd (1914),  Blyth,  Abraham,  Hedges,  Evans,  Doubleday - 1,  Piggins & Eley

4x4 problem:  King,  Abraham,  M. Adams,  Evans,  Depew,  Meyer,  Ripley's, 

Queen's tours:  Loyd (1867, 1897, 1914),  Loyd Jr.

Bishop's tours:  Dudeney (1932),  Doubleday - 2,  Obermair

Rook's tours:  Loyd (1878),  Proctor,  Loyd (1897),  Bullivant,  Loyd (1914),  Filipiak,  Hartswick,  Barwell,  Gardner,  Peters,  Obermair

Other versions:  Prout,  Doubleday - 1

Trick solutions:  Fixx,  Adams,  Piggins,  Piggins & Eley

          Thanks to Heinrich Hemme for pointing out Fixx, which led to adding most of the material on trick solutions.

 

Loyd.  ??Le Sphinx (Mar 1867 -- but the Supplement to Sam Loyd and His Chess Problems corrects this to 15 Nov 1866).  = Chess Strategy, Elizabeth, NJ, 1878, no. or p. 336(??).  = A. C. White; Sam Loyd and His Chess Problems; 1913, op. cit. in 1; no. 40, pp. 42‑43.  Queen's circuit on  8 x 8  in  14  segments.  (I.e. closed circuit, not leaving board, using queen's moves.)  No. 41 & 42 of White give other solutions.  White quotes Loyd from Chess Strategy, which indicates that Loyd invented this problem.  Tit‑Bits No. 31 & SLAHP: Touring the chessboard, pp. 19 & 89, give No. 41.

Loyd.  Chess Strategy, 1878, op. cit. above, no. or p. 337 (??)  (= White, 1913, op. cit. above, no. 43, pp. 42‑43.)  Rook's circuit on  8 x 8  in  16  segments.  (I.e. closed circuit, not leaving board, using rook's moves, and without crossings.)

Richard A. Proctor.  Gossip column.  Knowledge 10  (Dec 1886) 43  &  (Feb 1887) 92.  6 x 6  array of cells.  Prisoner in one corner can exit from the opposite corner if he passes "once, and once only, through all the  36  cells."  "... take the prisoner into either of the cells adjoining his own, and back into his own, ....  This puzzle is rather a sell, ...."  Letter and response [in Gossip column, Knowledge 10 (Mar 1887) 115-116] about the impossibility of any normal solution.

Loyd.  Problem 15: The gaoler's problem.  Tit‑Bits 31 (23 Jan  &  13 Feb 1897) 307  &  363.  Rook's circuit on  8 x 8  in  16  segments, but beginning and ending on a central square.  Cf The postman's puzzle in the Cyclopedia, 1914.

Loyd.  Problem 16: The captive maiden.  Tit‑Bits 31 (30 Jan  &  20 Feb 1897) 325  &  381.  Rook's tour in minimal number of moves from a corner to the diagonally opposite corner, entering each cell once.  Because of parity, this is technically impossible, so the first two moves are into an adjacent cell and then back to the first cell, so that the first cell has now been entered.

Loyd.  Problem 20: Hearts and darts.  Tit‑Bits 31 (20 Feb,  13  &  20 Mar 1897) 381,  437,  455.  Queen's tour on  8 x 8,  starting in a corner, permitting crossings, but with no segment going through a square where the path turns.  Solution in  14  segments.  This is No. 41 in White -- see the first Loyd entry above.

Ball.  MRE, 4th ed., 1905, p. 197.  At the end of his section on knight's tours, he states that there are many similar problems for other kinds of pieces.

Loyd.  In G. G. Bain, op. cit. in 1, 1907.  He gives the  3 x 3  lattice in four lines as the Columbus Egg Puzzle.

Pearson.  1907.  Part I, no. 36: A charming puzzle, pp. 36 & 152‑153.  3 x 3  lattice in  4  lines.

Loyd.  Sam Loyd's Puzzle Magazine (Apr 1908) -- ??NYS, reproduced in:  A. C. White; Sam Loyd and His Chess Problems; 1913, op. cit. in 1; no. 56, p. 52.  = Problem 26: A brace of puzzles -- No. 26: A study in naval warfare; Tit‑Bits 31 (27 Mar 1897) 475  &  32 (24 Apr 1897) 59.  = Cyclopedia, 1914, Going into action, pp. 189 & 364.  = MPSL1, prob. 46, pp. 44 & 138.  = SLAHP: Bombs to drop, pp. 86 & 119.  Circuit on  8 x 8  in  14  segments, but with two lines of slope  2.  In White, p. 43, Loyd says an ordinary queen's tour can be started "from any of the squares except the twenty which can be represented by  d1, d3 and d4."  This problem starts at  d1.  However I think White must have mistakenly put down twenty for twelve??

Anon.  Prob. 67.  Hobbies 31  (No. 782) (8 Oct 1910) 39  &  (No. 785) (29 Oct 1910) 94.  3 x 3  lattice in  4  lines "brought under my notice some time back".

C. H. Bullivant.  Home Fun, 1910, op. cit. in 5.S.  Part VI, Chap. IV.

No. 1: The travelling draught‑man, pp. 515 & 520.  Rook's circuit on  8 x 8  in 16 segments, different than Loyd's.

No. 3: Joining the rings.  3 x 3  in 4 segments.

Will Goldston.  More Tricks and Puzzles without Mechanical Apparatus.  The Magician Ltd., London, nd [1910?].  (BMC lists Routledge & Dutton eds. of 1910.)  (There is a 2nd ed., published by Will Goldston, nd [1919].)  The nine‑dot puzzle, pp. 127‑128 (pp. 90‑91 in 2nd ed.).

Loyd.  Cyclopedia, 1914, pp. 301 & 380.  = MPSL2, prob. 133 -- Solve Christopher's egg tricks, pp. 93 & 163 (with comment by Gardner).  c= SLAHP: Milkman's route, pp. 34 & 96.  3 x 3  case.

Loyd.  Cyclopedia, 1914, pp. 293 & 379.  Queen's circuit on  7 x 7  in  12  segments.

Loyd.  The postman's puzzle.  Cyclopedia, 1914, pp. 298 & 379.  Rook's circuit on  8 x 8  array of points, with one point a bit out of line, starting and ending at a central square, in  16  segments.  P. 379 also shows another  8 x 8  circuit, but with a slope  2  line.  See also pp. 21 & 341 and SLAHP, pp. 85 & 118, for two more examples.

Loyd.  Switchboard problem.  Cyclopedia, 1914, pp. 255 & 373.  (c= MPSL2, prob. 145, pp. 102 & 167.)  Rook's tour with minimum turning.

Blyth.  Match-Stick Magic.  1921.  Four-way game, pp. 77-78.  3 x 3  in  4  segments.

King.  Best 100.  1927.  No. 16, pp. 12 & 43.  4 x 4  in  6  segments, not closed, but easily can be closed.

Loyd Jr.  SLAHP. 1928.  Dropping the mail, pp. 67 & 111.  4 x 4  queen's tour in  6  segments.

Collins.  Book of Puzzles.  1927.  The star group puzzle, pp. 95-96.  3 x 3  in  4  segments.

Dudeney.  PCP.  1932.  Prob. 264: The fly's tour, pp. 82 & 169.  = 536, prob. 422, pp. 159 & 368.  Bishop's path, with repeated cells, going from corner to corner in  17  segments.

Abraham.  1933.  Probs. 101, 102, 103, pp. 49 & 66 (30 & 118).  3 x 3,  4 x 4  and  6 x 6  cases.

The Bile Beans Puzzle Book.  1933.  No. 4: The puzzled milkman.  3 x 3  array in four lines.

Sid G. Hedges.  More Indoor and Community Games.  Methuen, London, 1937.  Nine spot, p. 110.  3 x 3.  "Of course it can be done, but it is not easy."  No solution given.

M. Adams.  Puzzle Book.  1939.  Prob. C.64: Six strokes, pp. 140 & 178.  4 x 4  array in  6  segments which form a closed path, though the closure was not asked for.

J. R. Evans.  The Junior Week‑End Book.  Op. cit. in 6.AF.  1939.  Probs. 30 & 31, pp. 264 & 270.  3 x 3  &  4 x 4  cases in  4 & 6  segments, neither closed nor staying within the array.

Depew.  Cokesbury Game Book.  1939.  Drawing, p. 220.  4 x 4  in  6  segments, not closed, not staying within the array.

Meyer.  Big Fun Book.  1940.  Right on the dot, pp. 99 & 732.  4 x 4  in  6  segments.

A. S. Filipiak.  Mathematical Puzzles, 1942, op. cit. in 5.H.1, pp. 50‑51.  Same as Bullivant, but opens the circuit to make a 15 segment path.

M. S. Klamkin, proposer and solver;  John L. Selfridge, further solver.  Problem E1123 -- Polygonal path covering a square lattice.  AMM 61 (1954) 423  &  62 (1955) 124 & 443.  Shows  N x N  can be done in  2N‑2  segments.  Selfridge shows this is minimal.

W. Leslie Prout.  Think Again.  Frederick Warne & Co., London, 1958.  Joining the stars, pp. 41 & 129.  5 x 5  array of points.  Using a line of four segments, pass through  17  points.  This is a bit like the  3 x 3  problem in that one must go outside the array.

R. E. Miller & J. L. Selfridge.  Maximal paths on rectangular boards.  IBM J. Research and Development 4:5 (Nov 1960) 479-486.  They study rook's paths where a cell is deemed visited if the rook changes direction there.  They find maximal such paths in all cases.

Ripley's Puzzles and Games.  1966.  Pp. 72-73, item 2.  4 x 4  cases with closed solution symmetric both horizontally and vertically.

F. Gregory Hartswick.  In:  H. A. Ripley & F. Gregory Hartswick, Detectograms and Other Puzzles, Scholastic Book Services, NY, 1969.  Prob. 4, pp. 42‑43 & 82.  Asks for  8 x 8  rook's circuit with minimal turning and having a turn at a central cell.  Solution gives two such with  16  segments and asserts there are no others.

Doubleday - 1.  1969.  Prob. 60: Test case, pp. 75 & 167.  = Doubleday - 4, pp. 83-84.  Two  3 x 3  arrays joined at a corner, looking like the Fore and Aft board (cf 5.R.3), to be covered in a minimum number of segments.  He does it in seven segments by joining two  3 x 3  solutions.

Brian R. Barwell.  Arrows and circuits.  JRM 2 (1969) 196‑204.  Introduces idea of maximal length rook's tours.  Shows the maximal length on a  4 x 4  board is  38  and finds there are  3  solutions.  Considers also the  1 x n  board.

Solomon W. Golomb & John L. Selfridge.  Unicursal polygonal paths and other graphs on point lattices.  Pi Mu Epsilon J. 5 (1970) 107‑117.  Surveys problem.  Generalizes Selfridge's 1955 proof to  M x N  for which  MIN(2M, M+N‑2)  segments occur in a minimal circuit.

Doubleday - 2.  1971.  Path finder, pp. 95-96.  Bishop's corner to corner path, same as Dudeney, 1932.

James F. Fixx.  More Games for the Superintelligent.  (Doubleday, 1972);  Muller, (1977), 1981.  6. Variation on a variation, pp. 31 & 87.  Trick solution in three lines, assuming points of finite size.

M. Gardner.  SA (May 1973) c= Knotted, chap. 6.  Prob. 1: Find rook's tours of maximum length on the  4 x 4  board.  Cites Barwell.  Knotted also cites Peters, below.

Edward N. Peters.  Rooks roaming round regular rectangles.  JRM 6 (1973) 169‑173.  Finds maximum length on  1 x N  board is  N2/2  for  N  even;  (N‑1)2/2 + N‑1  for  N  odd, and believes he has counted such tours.  He finds tours on the  N x N  board whose length is a formula that reduces to  4 BC(N+1, 3) ‑ 2[(N‑1)/2].  I am a bit unsure if he has shown that this is maximal.

James L. Adams.  Conceptual Blockbusting.  Freeman, 1974, pp. 16-22.  3rd ed., (A-W, 1986), Penguin, 1987, pp. 24-33.  Trick solution of  3 x 3  case in three lines, assuming points of finite size, which he says was submitted anonymously when he and Bob McKim used the puzzle on an ad for a talk on problem-solving at Stanford.  Also describes a version using paperfolding to get all nine points into a line.  The material is considerably expanded in the 3rd ed. and adds several new versions.  From the references in Piggins and Eley, it seems that these all appeared in the 2nd ed of 1979 -- ??NYS.

                    Cut out the  3 x 1  parts and tape them into a straight line.

                    Take the paper and roll it to a cylinder and then draw a slanting line on the cylinder which goes through all nine, largish, points.

                    Cut out bits with each point on and skewer the lot with a pencil.

                    Place the paper on the earth and draw a line around the earth to go through all nine points.  One has to assume the points have some size.

                    Wodge the paper, with large dots, into a ball and stick a pencil through it.  Open up to see if you have won -- if not, try again!

                    Use a very fat line, i.e. as thick as the spacing between the edges of the array.

David J. Piggins.  Pathological solutions to a popular puzzle.  JRM 8:2 (1975-76) 128-129.  Gives two trick solutions.

                    Three parallel lines, since they meet at infinity.

                    Put the figure on the earth and use a slanting line around the earth.  This works in the limit, but otherwise requires points of finite size, a detail that he doesn't mention. 

                    No references for these versions.

David J. Piggins & Arthur D. Eley.  Minimal path length for covering polygonal lattices:  A review.  JRM 14:4 (1981‑82) 279‑283.  Mostly devoted to various trick solutions of the  3 x 3  case.  They cite Piggins' solution with three parallel lines.  They say that Gardner sent them the trick solution in 1973 and then cite Adams, 1979.  They give solutions using points of different sizes, getting both three and two segment solutions and mention a two segment version that depends on the direction of view.  They then give the solution on a sphere, citing Adams, 1979, and Piggins.  They give several further versions using paper folding, including putting the surface onto a twisted triangular prism joined at the ends to make the surfaces into a Möbius strip -- Zeeman calls this a umbilical bracelet or a Möbius bar. 

Obermair.  Op. cit. in 5.Z.1.  1984.

Prob. 19, pp. 23 & 50.  Bishop's path on  8 x 8  in  17  segments, as in Dudeney, PCP, 1932.

Prob. 41, p. 72.  Rook's path with maximal number of segments, which is  57.  [For the  2 x 2,  3 x 3,  4 x 4  boards, I get the maximum numbers are  3, 6, 13.]

Nob Yoshigahara.  Puzzlart.  Tokyo, 1992.  Section: The wisdom of Solomon, pp. 40-47, abridged from an article by Solomon W. Golomb in Johns Hopkins Magazine (Oct 1984).  Classic  3 x 3  problem.  For the  4 x 4 case:  1) find four closed paths;  2), says there are about 30 solutions and gives 19 beyond the previous 4.  Find the unique 5‑segment closed path on the  3 x 4.  Gives 3 solutions on  5 x 5.  10-segment solution on  6 x 6  which stays on the board.  Loyd's 1867? Queen's circuit.  Queen's circuit on  7 x 7,  attributed to Dudeney, though my earliest entry is Loyd, 1914 -- ??CHECK.

 

          6.AL.           STEINER‑LEHMUS THEOREM

 

          This has such an extensive history that I will give only a few items.

 

C. L. Lehmus first posed the problem to Jacob Steiner in 1840.

Rougevin published the first proof in 1842.  ??NYS.

Jacob Steiner.  Elementare Lösung einer Aufgabe über das ebene und sphärische Dreieck.  J. reine angew. Math. 28 (1844) 375‑379 & Tafel III.  Says Lehmus sent it to him in 1840 asking for a purely geometric proof.  Here he gives proofs for the plane and the sphere and also considers external bisectors.

Theodor Lange.  Nachtrag zu dem Aufsatze in Thl. XIII, Nr. XXXIII.  Archiv der Math. und Physik 15 (1850) 221‑226.  Discusses the problem and gives a solution by Steiner and two by C. L. Lehmus.  Steiner also considers the external bisectors.

N. J. Chignell.  Note 1031: A difficult converse.  MG 16 (No. 219) (Jul 1932) 200-202.  [The author's name is omitted in the article but appears on the cover.]  'Three fairly simple proofs', due to:  M. J. Newell;  J. Travers, improving J. H. Doughty, based on material in Lady's and Gentleman's Diary (1859) 87-88 & (1860) 84-86;  Wm. Mason, found by Doughty, in Lady's and Gentleman's Diary (1860) 86.

H. S. M. Coxeter.  Introduction to Geometry.  Wiley, 1961.  Section 1.5, ex. 4, p. 16.  An easy proof is posed as a problem with adequate hints in four lines.

M. Gardner.  SA (Apr 1961) = New MD, chap. 17.  Review of Coxeter's book, saying his brief proof came as a pleasant shock.

G. Gilbert & D. MacDonnell.  The Steiner‑Lehmus theorem.  AMM 70 (1963) 79‑80.  This is the best of the proofs sent to Gardner in response to his review of Coxeter.  A later source says this turned out to be identical to Lehmus' original proof!

Léo Sauvé.  The Steiner‑Lehmus theorem.  CM 2:2 (Feb 1976) 19‑24.  Discusses history and gives 22 references, some of which refer to 60 proofs.

Charles W. Trigg.  A bibliography of the Steiner‑Lehmus theorem.  CM 2:9 (Nov 1976) 191‑193.  36 references beyond Sauvé's.

David C. Kay.  Nearly the last comment on the Steiner‑Lehmus theorem.  CM 3:6 (1977) 148‑149.  Observes that a version of the proof works in all three classical geometries at once and gives its history.

 

          6.AM.          MORLEY'S THEOREM

 

          This also has an extensive history and I give only a few items.

 

T. Delahaye and H. Lez.  Problem no. 1655 (Morley's triangle).  Mathesis (3) 8 (1908) 138‑139.  ??NYS.

E. J. Ebden, proposer;  M. Satyanarayana, solver.  Problem no. 16381 (Morley's theorem).  The Educational Times (NS) 61 (1 Feb 1908) 81  &  (1 Jul 1908) 307‑308  = Math. Quest. and Solutions from "The Educational Times" (NS) 15 (1909) 23.  Asks for various related triangles formed using interior and exterior trisectors to be shown equilateral.  Solution is essentially trigonometric.  No mention of Morley.

Frank Morley.  On the intersections of the trisectors of the angles of a triangle.  (From a letter directed to Prof. T. Hayashi.)  J. Math. Assoc. of Japan for Secondary Education 6 (Dec 1924) 260‑262.  (= CM 3:10 (Dec 1977) 273‑275.

Frank Morley.  Letter to Gino Loria.  22 Aug 1934.  Reproduced in:  Gino Loria; Triangles équilatéraux dérivés d'un triangle quelconque.  MG 23 (No. 256) (Oct 1939) 364‑372.  Morley says he discovered the theorem in c1904 and cites the letter to Hayashi.  Loria mentions other early work and gives several generalizations.

H. F. Baker.  Note 1476:  A theorem due to Professor F. Morley.  MG 24 (No. 261) (Oct 1940) 284‑286.  Easy proof and reference to other proofs.  He cites a related result of Steiner.

Anonymous [R. P.]  Morley's trisector theorem.  Eureka 16 (Oct 1953) 6-7.  Short proof, working backward from the equilateral triangle.

Dan Pedoe.  Notes on Morley's proof of his theorem on angle trisectors.  CM 3:10 (Dec 1977) 276‑279.  "... very tentative ... first steps towards the elucidation of his work."

C. O. Oakley & Charles W. Trigg.  A list of references to the Morley theorem.  CM 3:10 (Dec 1977) 281‑290  &  4 (1978) 132.  169 items.

André Viricel (with Jacques Bouteloup).  Le Théorème de Morley.  L'Association pour le Développement de la Culture Scientifique, Amiens, 1993.  [This publisher or this book was apparently taken over by Blanchard as Blanchard was selling copies with his label pasted over the previous publisher's name in Dec 1994.]  A substantial book (180pp) on all aspects of the theorem.  The bibliography is extremely cryptic, but says it is abridged from Mathesis (1949) 175  ??NYS.  The most recent item cited is 1970.

 

          6.AN.          VOLUME OF THE INTERSECTION OF TWO CYLINDERS

 

Archimedes.  The Method:  Preface, 2.  In:  T. L. Heath; The Works of Archimedes, with a supplement "The Method of Archimedes"; (originally two works, CUP, 1897 & 1912)  = Dover, 1953.  Supplement, p. 12, states the result.  The proof is lost, but pp. 48‑51 gives a reconstruction of the proof by Zeuthen.

Liu Hui.  Jiu Zhang Suan Chu Zhu (Commentary on the Nine Chapters of the Mathematical Art).  263.  ??NYS -- described in Li & Du, pp. 73‑74 & 85.  He shows that the ratio of the volume of the sphere to the volume of Archimedes' solid, called mou he fang gai (two square umbrellas), is  π/4,  but he cannot determine either volume.

Zu Geng.  c500.  Lost, but described in:  Li Chunfeng; annotation to Jiu Zhang (= Chiu Chang Suan Ching) made c656.  ??NYS.  Described on  pp. 86‑87 of:  Wu Wenchun; The out‑in complementary principle; IN:  Ancient China's Technology and Science; compiled by the Institute of the History of Natural Sciences, Chinese Academy of Sciences; Foreign Languages Press, Beijing, 1983, pp. 66‑89.  [This is a revision and translation of parts of:  Achievements in Science and Technology in Ancient China [in Chinese]; China Youth Publishing House, Beijing(?), 1978.]

                    He considers the shape, called fanggai, within the natural circumscribed cube and shows that, in each octant, the part of the cube outside the fanggai has cross section of area  h2  at distance  h  from the centre.  This is equivalent to a tetrahedron, whose volume had been determined by Liu, so the excluded volume is    of the cube.

                    Li & Du, pp. 85‑87, and say the result may have been found c480 by Zu Geng's father, Zu Chongzhi.

Lam Lay-Yong & Shen Kangsheng.  The Chinese concept of Cavalieri's Principle and its applications.  HM 12 (1985) 219-228.  Discusses the work of Liu and Zu. 

Shiraishi Chōchū.  Shamei Sampu.  1826.  ??NYS -- described in Smith & Mikami, pp. 233-236.  "Find the volume cut from a cylinder by another cylinder that intersects is orthogonally and touches a point on the surface".  I'm not quite sure what the last phrase indicates.  The book gives a number of similar problems of finding volumes of intersections.

P. R. Rider, proposer;  N. B. Moore, solver.  Problem 3587.  AMM 40 (1933) 52 (??NX) &  612.  Gives the standard proof by cross sections, then considers the case of unequal cylinders where the solution involves complete elliptic integrals of the first and second kinds.  References to solution and similar problem in textbooks.

Leo Moser, solver;  J. M. Butchart, extender.  MM 25 (May 1952) 290  &  26 (Sep 1952) 54.  ??NX.  Reproduced in Trigg, op. cit. in 5.Q: Quickie 15, pp. 6 & 82‑83.  Moser gives the classic proof that  V = 16r3/3.  Butchart points out that this also shows that the shape has surface area  16r2.

 

          6.AO. CONFIGURATION PROBLEMS

 

          NOTATION:  (a, b, c)  denotes the configuration of  a  points in  b  rows of  c  each.  The index below covers articles other than the surveys of Burr et al. and Gardner.

 

(  5,   2, 3):             Sylvester

(  6,   3, 3):             Mittenzwey

(  7,   6, 3):             Criton

(  9,   8, 3):             Sylvester;  Carroll;  Criton

(  9,   9, 3):             Carroll;  Bridges;  Criton

(  9, 10, 3):              Jackson;  Family Friend;  Parlour Pastime;  Magician's Own Book;  The Sociable;  Book of 500 Puzzles;  Charades etc.;  Boy's Own Conjuring Book;  Hanky Panky;  Carroll;  Crompton;  Berkeley & Rowland;  Hoffmann;  Dudeney (1908);  Wehman;  Williams;  Loyd Jr;  Blyth;  Rudin;  Young World;  Brooke;  Putnam;  Criton

(10,   5, 4):              The Sociable;  Book of 500 Puzzles;  Carroll;  Hoffmann;  Dudeney (1908);  Wehman;  Williams;  Dudeney (1917);  Blyth;  King;  Rudin;  Young World;  Hutchings  & Blake;  Putnam

(10, 10, 3):              Sylvester

(11, 11, 3):              The Sociable;  Book of 500 Puzzles; Wehman

(11, 12, 3):              Hoffmann;  Williams;  Young World

(11, 13, 3):              Prout

(11, 16, 3):              Wilkinson -- in Dudeney (1908 & 1917);  Macmillan

(12,   4, 5) -- Trick version of a hollow  3 x 3  square with doubled corners, as in 7.Q:  Family Friend (1858);  Secret Out;  Illustrated Boy's Own Treasury; 

(12,   6, 4):              Endless Amusement II;  The Sociable;  Book of 500 Puzzles;  Boy's Own Book;  Cassell's;  Hoffmann;  Wehman;  Rudin;  Criton

(12,   7, 4) -- Trick version of a  3 x 3  square with doubled diagonal:  Secret Out;  Hoffmann (1876);  Mittenzwey;  Hoffmann (1893), no. 8

(12,   7, 4):              Dudeney (1917);  Putnam

(12, 19, 3):              Macmillan

(13,   9, 4):              Criton

(13, 12, 3):              Criton

(13, 18, 3):              Sylvester

(13, 22, 3):              Criton

(15, 15, 3):              Jackson

(15, 16, 3):              The Sociable;  Book of 500 Puzzles;  H. D. Northrop;  Wehman

(15, 23, 3):              Jackson

(15, 26, 3):              Woolhouse

(16, 10, 4):              The Sociable;  Book of 500 Puzzles;  Hoffmann;  Wehman

(16, 12, 4):              Criton

(16, 15, 4):              Dudeney (1899, 1902, 1908);  Brooke;  Putnam; Criton

(17, 24, 3):              Jackson

(17, 28, 3):              Endless Amusement II;  Pearson

(17, 32, 3):              Sylvester

(17,  7, 5):               Ripley's

(18, 18, 4):              Macmillan

(19, 19, 4):              Criton

(19,   9, 5):              Endless Amusement II;  The Sociable;  Book of 500 Puzzles;  Proctor;  Hoffmann;  Clark;  Wehman;  Ripley;  Rudin;  Putnam;  Criton

(19, 10, 5):              Proctor

(20, 12, 5):              trick method: Doubleday - 3

(20, 18, 4):              Loyd Jr

(20, 21, 4):              Criton

(21,   9, 5):              Magician's Own Book;  Book of 500 Puzzles;  Boy's Own Conjuring Book;  Blyth;  Depew

(21, 10, 5):              Mittenzwey

(21, 11, 5):              Putnam

(21, 12, 5):              Dudeney (1917);  Criton

(21, 30, 3):              Secret Out;  Hoffmann

(21, 50, 3):              Sylvester

(22, 15, 5):              Macmillan

(22, 20, 4):              Dudeney (1899)

(22, 21, 4):              Dudeney (1917);  Putnam

(24, 28, 3):              Jackson;  Parlour Pastime

(24, 28, 4):              Jackson;  Héraud;  Benson;  Macmillan

(24, 28, 5):              Jackson

(25, 12, 5):              Endless Amusement II;  Young Man's Book;  Proctor;  Criton

(25, 18, 5):              Bridges

(25, 30, 4):              Macmillan

(25, 72, 3):              Sylvester

(26, 21, 5):              Macmillan

(27,   9, 6):              The Sociable;  Book of 500 Puzzles;  Hoffmann;  Wehman

(27, 10, 6):              The Sociable;  Book of 500 Puzzles;  Wehman

(27, 15, 5):              Jackson

(29, 98, 3):              Sylvester

(30, 12, 7):              Criton

(30, 22, 5):              Criton

(30, 26, 5):              Macmillan

(31,   6, 6) -- with 7 circles of 6:  The Sociable;  Book of 500 Puzzles;  Magician's Own Book (UK version);  Wehman

(31, 15, 5):              Proctor

(36, 55, 4):              Macmillan

(37, 18, 5):              Proctor

(37, 20, 5):              The Sociable;  Book of 500 Puzzles;  Illustrated Boy's Own Treasury;  Hanky Panky; Wehman

(49, 16, 7):              Criton

 

Trick versions -- with doubled counters:  Family Friend (1858);  Secret Out;  Illustrated Boy's Own Treasury;  Hoffmann (1876);  Mittenzwey;  Hoffmann (1893), nos. 8 & 9;  Pearson;  Home Book ....; Doubleday - 3.  These could also be considered as in 7.Q.2 or 7.Q.

A different type of configuration problem is considered by Shepherd, 1947.

 

Jackson.  Rational Amusement.  1821.  Trees Planted in Rows, nos. 1-10, pp. 33-34 & 99-100 and plate IV, figs. 1-9.  [Brooke and others say this is the earliest statement of such problems.]

1.  (9, 10, 3).  Quoted in Burr, below.

           "Your aid I want, nine trees to plant

                     In rows just half a score;

           And let there be in each row three.

                     Solve this: I ask no more."

2.  (n, n, 3),  He does the case  n = 15.

3.  (15, 23, 3).

4.  (17, 24, 3).

5.  (24, 24, 3)  with a pond in the middle.

6.  (24, 28, 4).

7.  (27, 15, 5)

8.  (25, 28, c)  with  c = 3, 4, 5.

9.  (90, 10, 10)  with equal spacing -- decagon with 10 trees on each side.

10. Leads to drawing square lattice in perspective with two vanishing points, so the diagonals of the resulting parallelograms are perpendicular.

Endless Amusement II.  1826? 

Prob. 13, p. 197.  (19, 9, 5).  = New Sphinx, c1840, p. 135.

Prob. 14, p. 197.  (12, 6, 4).  = New Sphinx, c1840, p. 135.

Prob. 26, p. 202.  (25, 12, 5).  Answer is a  5 x 5  square array.

Ingenious artists, how may I dispose

Of five-and-twenty trees, in just twelve rows;

That every row five lofty trees may grace,

Explain the scheme -- the trees completely place.

Prob. 35, p. 212.  (17, 28, 3).  [This is the problem that is replaced in the 1837 ed.]

Young Man's Book.  1839.  P. 239.  Identical to Endless Amusement II.

Crambrook.  1843.  P. 5, no. 15: The Puzzle of the Steward and his Trees.  This may be a configuration problem -- ??

Boy's Own Book.  1843 (Paris): 438 & 442, no. 15: "Is it possible to place twelve pieces of money in six rows, so as to have four in each row?"  I. e.  (12, 6, 5).  = Boy's Treasury, 1844, pp. 426 & 429, no. 13.  = de Savigny, 1846, pp. 355 & 358, no. 11.

Family Friend 1 (1849) 148 & 177.  Family Pastime -- Practical Puzzles -- 1. The puzzle of the stars.  (9, 10, 3). 

                    Friends of the Family Friend, pray show

                    How you nine stars would so bestow

                    Ten rows to form -- in each row three --

                    Tell me, ye wits, how this can be?

                                                                Robina.

          Answer has

                    Good-tempered Friends! here nine stars see:

                    Ten rows there are, in each row three!

W. S. B. Woolhouse.  Problem 39.  The Mathematician 1 (1855) 272.  Solution:  ibid. 2 (1856) 278‑280.  ??NYS -- cited in Burr, et al., below, who say he does  (15, 26, 3).

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles.

No. 1, p. 176 (1868: 187).  (9, 10, 3). 

Ingenious artist pray disclose,

How I nine trees can so dispose,

That these ten rows shall formed be,

And every row consist of three?

No. 12, p. 182 (1868: 192-193).  (24, 28, 3),  but with a central pond breaking 4 rows of 6 into 8 rows of 3.

Magician's Own Book.  1857.

Prob. 33: The puzzle of the stars, pp. 277 & 300.  (9, 10, 3),

Friends one and all, I pray you show

How you nine stars would so bestow,

Ten rows to form -- in each row three --

Tell me, ye wits, how this can be?

Prob. 41: The tree puzzle, pp. 279 & 301.  (21, 9, 5),  unequally spaced on each row.  Identical to Book of 500 Puzzles, prob. 41.

The Sociable.  1858.  = Book of 500 Puzzles, 1859, with same problem numbers, but page numbers decreased by 282.

Prob. 3: The practicable orchard, pp. 286 & 302.  (16, 10, 4).

Prob. 8: The florist's puzzle, pp. 289 & 303-304.  (31, 6, 6)  with 7 circles of 6.

Prob. 9: The farmer's puzzle, pp. 289 & 304.  (11, 11, 3).

Prob. 12: The geometrical orchard,  p. 291 & 306.  (27, 9, 6).

Prob. 17: The apple-tree puzzle, pp. 292 & 308.  (10, 5, 4).

Prob. 22: The peach orchard puzzle, pp. 294 & 309.  (27, 10, 6).

Prob. 26: The gardener's puzzle, pp. 295 & 311.  (12, 6, 4)  two ways.

Prob. 27: The circle puzzle, pp. 295 & 311.  (37, 20, 5)  equally spaced along each row.

Prob. 29: The tree puzzle, pp. 296 & 312.  (15, 16, 3)  with some bigger rows.  Solution is a  3 x 4  array with three extra trees halfway between the points of the middle line of four.

Prob. 32: The tulip puzzle, pp. 296 & 314.  (19, 9, 5).

Prob. 36: The plum tree puzzle, pp. 297 & 315.  (9, 10, 3).

Family Friend (Dec 1858) 359.  Practical puzzles -- 2.  "Make a square with twelve counters, having five on each side."  (12, 4, 5).  I haven't got the answer, but presumably it is the trick version of a hollow square with doubled corners, as in 7.Q.  See Secret Out, 1859 & Illustrated Boy's Own Treasury, 1860.

Book of 500 Puzzles.  1859.  Prob.  3, 9, 12, 17, 22, 26, 27, 29, 32, 36  are identical to those in The Sociable, with page numbers decreased by 282.

Prob. 33: The puzzle of the stars, pp. 91 & 114.  (9, 10, 3),  identical to Magician's Own Book, prob. 33.

Prob. 41: The tree puzzle, pp. 93 & 115.  (21, 9, 5),  identical to Magician's Own Book, prob. 41.  See Illustrated Boy's Own Treasury.

The Secret Out.  1859. 

To place twelve Cards in such a manner that you can count Four in every direction, p. 90.  (12, 7, 4)  trick of a  3 x 3  array with doubling along a diagonal.  'Every direction' must refer to just the rows and columns, but one diagonal also works.

The magical arrangement, pp. 381-382  =  The square of counters, (UK) p. 9.  (12, 4, 5) -- trick version.  Same as Family Friend & Illustrated Boy's Own Treasury, prob. 13.

The Sphynx, pp. 385-386.  (21, 30, 3).  = Hoffmann, no. 15.

Charades, Enigmas, and Riddles.  1860: prob. 13, pp. 58 & 61;  1862: prob. 13, pp. 133 & 139;  1865: prob. 557, pp. 105 & 152.  (9, 10, 3).  (The 1862 and 1865 have slightly different typography.)

                    Sir Isaac Newton's Puzzle (versified).

                              Ingenious Artist, pray disclose

                              How I, nine Trees may so dispose,

                              That just Ten Rows shall planted be,

                              And every Row contain just Three.

Boy's Own Conjuring Book.  1860.

Prob. 40: The tree puzzle, pp. 242 & 266.  (21, 9, 5),  identical to Magician's Own Book, prob. 41.

Prob. 42: The puzzle of the stars, pp. 243 & 267.  (9, 10, 3),  identical to Magician's Own Book, prob. 33, with commas omitted.

Illustrated Boy's Own Treasury.  1860.

Prob. 2, pp. 395 & 436.  (37, 20, 5),  equally spaced on each row, identical to The Sociable, prob. 27.

Prob. 13, pp. 397 & 438.  "Make a square with twelve counters, having five on each side."  (12, 4, 5).  Trick version of a hollow square with doubled corners.  Presumably identical to Family Friend, 1858.  Same as Secret Out.

J. J. Sylvester.  Problem 2473.  Math. Quest. from the Educ. Times 8 (1867) 106‑107.  ??NYS -- Burr, et al. say he gives  (10, 10, 3),  (81, 800, 3)  and  (a, (a‑1)2/8, 3).

Magician's Own Book (UK version).  1871.  The solution to The florist's puzzle (The Sociable, prob. 8) is given at the bottom of p. 284, apparently to fill out the page as there is no relevant text anywhere. 

Hanky Panky.  1872. 

To place nine cards in ten rows of three each, p. 291.  I.e.  (9, 10, 3).

Diagram with no text, p. 128.  (37, 20, 5),  equally spaced on each line as in The Sociable, prob. 27.

Hoffmann.  Modern Magic.  (George Routledge, London, 1876);  reprinted by Dover, 1978.  To place twelve cards in rows, in such a manner that they will count four in every direction, p. 58.  Trick version of a  3 x 3  square with extras on a diagonal, giving a form of  (12, 7, 4).  Same as Secret Out.

Lewis Carroll.  MS of 1876.  ??NYS -- described in:  David Shulman; The Lewis Carroll problem; SM 6 (1939) 238-240. 

Given two rows of five dots, move four to make 5 rows of 4.  Shulman describes this case, following Dudeney, AM, 1917, then observes that since Dudeney is using coins, there are further solutions by putting a coin on top of another.  He refers to Hoffmann and Loyd.  The same problem is in Carroll-Wakeling, prob. 1: Cakes in a row, pp. 1-2 & 63, but undated and the answer mentions the possibility of stacking the counters.

(9, 10, 3).  Shulman quotes from Robert T. Philip; Family Pastime; London, 1852, p. 30, ??NYS, but this must refer to the item in Family Friend, which was edited by Robert Kemp Philp.  BMC indicates Family Pastime which may be another periodical.  Shulman then cites Jackson and Dudeney.  Carroll-Wakeling, prob. 2: More cakes in a row, pp. 3 & 63, gives the problems  (9, 8, 3),  (9, 9, 3),  (9, 10, 3),  undated.

Mittenzwey.  1880. 

Prob. 151, pp. 31 & 83;  1895?: 174, pp. 36 & 85;  1917: 174, pp. 33 & 82.  (6, 3, 3)  in three ways.

Prob. 152, pp. 31 & 83;  1895?: 175, pp. 36 & 85;  1917: 175, pp. 33 & 82.  Arrange 16 pennies as a  3 x 3  square so each row and column has four in it.  Solution shows a  3 x 3  square with extras on the diagonal -- but this only uses 12 pennies!  So this the trick version of  (12, 7, 4)  as in Secret Out & Hoffmann (1876).

Prob. 153, pp. 31 & 83;  1895?: 176, pp. 36 & 85;  1917: 176, pp. 33 & 82.  (21, 10, 5).

Cassell's.  1881.  P. 92: The six rows puzzle.  = Manson, 1911, p. 146.

J. J. Sylvester.  Problem 2572.  Math. Quest. from the Educ. Times 45 (1886) 127‑128.  ??NYS -- cited in Burr, below.  Obtains good  examples of  (a, b, 3)  for each  a.  In most cases, this is still the best known.

Richard A. Proctor.  Some puzzles;  Knowledge 9 (Aug 1886) 305-306  &  Three puzzles;  Knowledge 9 (Sep 1886) 336-337.  (19, 9, 5).  Generalises to  (6n+1, 3n, 5).

Richard A. Proctor.  Our puzzles.  Knowledge 10  (Nov 1886) 9  &  (Dec 1886) 39-40.  Gives several solutions of  (19, 9, 5)  and asks for  (19, 10, 5).  Gossip column, (Feb 1887) 92, gives another solution

William Crompton.  The odd half-hour.  The Boy's Own Paper 13 (No. 657) (15 Aug 1891) 731-732.  Sir Isaac Newton's puzzle (versified).  (9, 10, 3).

                              Ingenious artist pray disclose

                              How I nine trees may so dispose

                              That just ten rows shall planted be

                              And every row contain just three.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. IV, p. 3.  (9, 10, 3).

Hoffmann.  1893.  Chap. VI, pp. 265‑268 & 275‑281 = Hoffmann-Hordern, pp. 174-182, with photo.

No. 1: (11, 12, 3).

No. 2: (9, 10, 3).

No. 3: (27, 9, 6).

No. 4: (10, 5, 4).

No. 5: (12, 6, 4).  Photo on p. 177 shows L'Embarras du Brigadier, by Mauclair-Dacier, 1891‑1900, which has a board with a  7 x 6  array of holes and 12 pegs.  The horizontal spacing seems closer than the vertical spacing.

No. 6: (19, 9, 5).

No. 7: (16, 10, 4).

No. 8: (12, 7, 4)  --  Trick version of a  3 x 3  square with extras on a diagonal as in Secret Out, Hoffmann (1876) & Mittenzwey.

No. 9: 9 red + 9 white,  form  10 + 8  lines of 3 each.  Puts a red and a white point at the same place, so this is a trick version.

No. 11: (10, 8, 4)  --  counts in 8 'directions', so he counts each line twice!

No. 12: (13, 12, 5)  --  with double counting as in no. 11.

No. 15: (21, 30, 3)  --  but points must lie on a given figure, which is the same as in The Secret Out.

Clark.  Mental Nuts.  1897, no. 19: The apple orchard;  1904, no. 91: The lovers' grove.  (19, 9, 5).  1897 just has "Place an orchard of nineteen trees so as to have nine rows of five trees each."  1904 gives a poem.

                    I am required to plant a grove

                    To please the lady whom I love.

                    This simple grove to be composed

                    Of nineteen trees in nine straight rows;

                    Five trees in each row I must place,

                    Or I shall never see her face.

          Cf Ripley, below.

Dudeney.  A batch of puzzles.  Royal Magazine 1:3 (Jan 1899)  &  1:4 (Feb 1899) 368-372.  (22, 20, 4)  with trees at lattice points of a  7 x 10  lattice.  Compare with AM, prob. 212.

Anon. & Dudeney.  A chat with the Puzzle King.  The Captain 2 (Dec? 1899) 314-320  &  2:6 (Mar 1900) 598-599  &  3:1 (Apr 1900) 89.  (16, 15, 4).  Cf 1902.

Dudeney.  "The Captain" puzzle corner.  The Captain 3:2 (May 1900) 179.  This gives a solution of a problem called Joubert's guns, but I haven't seen the proposal.  (10, 5, 4)  but wants the maximum number of castles to be inside the walls joining the castles.  Manages to get two inside.  = Dudeney; The puzzle realm; Cassell's Magazine ?? (May 1908) 713-716; no. 6: The king and the castles.  = AM, 1917, prob. 206: The king and the castles, pp. 56 & 189. 

H. D. Northrop.  Popular Pastimes.  1901.  No. 11: The tree puzzle, pp. 68 & 73.  = The Sociable, no. 29.

Dudeney.  The ploughman's puzzle.  In:  The Canterbury Puzzles, London Magazine 9 (No. 49) (Aug 1902) 88‑92  &  (No. 50) (Sep 1902) 219.  = CP; 1907; no. 21, pp. 43‑44 & 175‑176.  (16, 15, 4).  Cf 1899.

A. Héraud.  Jeux et Récréations Scientifiques -- Chimie, Histoire Naturelle, Mathématiques.  Baillière et Fils, Paris, 1903.  P. 307: Un paradoxe mathématique.  (24, 28, 4).  I haven't checked for this problem in the 1884 ed.

Pearson.  1907. 

Part I, no. 77: Lines on an old sampler, pp. 77 & 167.  (17, 28, 3).

Part II, no. 83: For the children, pp. 83 & 177.  Trick version of  (12, 4, 5),  as in Family Friend (1858).

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  He says  (9, 10, 3)  "is attributed to Sir Isaac Newton, but the earliest collection of such puzzles is, I believe, in a rare little book that I possess -- published in 1821."  [This must refer to Jackson.]  Says Rev. Mr. Wilkinson gave  (11, 16, 3)  "some quarter of a century ago" and that he, Dudeney, published  (16, 15, 4)  in 1897 (cf under 1902 above).  He leaves these as problems but doesn't give their solutions in the next issue.

Wehman.  New Book of 200 Puzzles.  1908.

P. 4: The practicable orchard.  (16, 10, 4).  = The Sociable, prob. 3.

P. 7: The puzzle of the stars.  (9, 10, 3).  = Magician's Own Book, prob. 33.

P. 8: The apple-tree puzzle.  (10, 5, 4).  = The Sociable, prob. 17.

P. 8: The peach orchard puzzle.  (27, 10, 6).  = The Sociable, prob. 22.

P. 8: The plum tree puzzle.  (9, 10, 3).  = The Sociable, prob. 36.

P. 12: The farmer's puzzle.  (11, 11, 3).  = The Sociable, prob. 9.

P. 19: The gardener's puzzle.  (12, 6, 4)  two ways.  = The Sociable, prob. 26.

P. 26: The circle puzzle.  (37, 20, 5)  equally spaced along each row.  = The Sociable, prob. 27.

P. 30: The tree puzzle.  (15, 16, 3)  with some bigger rows.  = The Sociable, prob. 29.

P. 31: The geometrical orchard.  (27, 9, 6).  = The Sociable, prob. 12.

P. 31: The tulip puzzle.  (19, 9, 5).  = The Sociable, prob. 32.

P. 41: The florist's puzzle.  (31, 6, 6)  with seven circles of six.  = The Sociable, prob. 8.

J. K. Benson, ed.  The Pearson Puzzle Book.  C. Arthur Pearson, London, nd [c1910, not in BMC or NUC].  [This is almost identical with the puzzle section of Benson, but has 13 pages of different material.]  A symmetrical plantation, p. 99.  (24, 28, 4).

Williams.  Home Entertainments.  1914.  Competitions with counters, p. 115.  (11, 12, 3);  (9, 10, 3);  (10, 5, 4).

Dudeney.  AM.  1917.  Points and lines problems, pp. 56-58 & 189-193.

Prob. 206: The king and the castles.  See The Captain, 1900.

Prob. 207: Cherries and plums.  Two  (10, 5, 4)  patterns among 55 of the points of an  8 x 8  array.

Prob. 208: A plantation puzzle.  (10, 5, 4)  among 45 of the points of a  7 x 7  array.

Prob. 209: The twenty-one trees.  (21, 12, 5).

Prob. 210: The ten coins.  Two rows of five.  Move four to make  (10, 5, 4).  Cf Carroll, 1876.  Shows there are 2400 ways to do this.  He shows that there are six basic solutions of the  (10, 5, 4)  which he calls:  star, dart, compasses, funnel, scissors, nail and he describes the smallest arrays on which they can fit.

Prob. 211: The twelve mince-pies.  12 points at the vertices and intersections of a Star of David.  Move four to make  (12, 7, 4).

Prob. 212: The Burmese plantation.  (22, x, 4)  among the points of a  7 x 7  array.  Finds  x = 21.  Cf 1899.

Prob. 213: Turks and Russians, pp. 58 & 191‑193.  Complicated problem leading to  (11, 16, 3)  -- cites his Afridi problem in Tit-Bits and attributes the pattern to Wilkinson 'some twenty years ago', cf 1908.

Blyth.  Match-Stick Magic.  1921.

Four in line, p. 48.  (10, 5, 4).

Three in line, p. 77.  (9, 10, 3).

Five-line game, pp. 78-79.  (21, 9, 5).

King.  Best 100.  1927.  No. 62, pp. 26 & 54.  = Foulsham's no. 21, pp. 9 & 13.  (10, 5, 4).

Loyd Jr.  SLAHP.  1928.  Points and lines puzzle, pp. 20 & 90.  Says Newton proposed  (9, 10, 3).  Asks for  (20, 18, 4)  on a  7 x 7  array.

R. Ripley.  Believe It or Not!  Book 2.  Op. cit. in 5.E, 1931.  The planter's puzzle, p. 197, asks for  (19, 9, 5)  but no solution is given.  See Clark, above, for a better version of the verse.

          "I am constrained to plant a grove

          For a lady that I love.

          This ample grove is too composed;

          Nineteen trees in nine straight rows.

          Five trees in each row I must place,

          Or I shall never see her face."

Rudin.  1936.  Nos. 105-108, pp. 39 & 99-100.

No. 105: (9, 10, 3).

No. 106: (10, 5, 4)  -- two solutions.

No. 107: (12, 6, 4)  -- two solutions.

No. 108: (19, 9, 5).

Depew.  Cokesbury Game Book.  1939.  The orange grower, p. 221.  (21, 9, 5).

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  P. 147, prob. 1 & 2.  Place six coins in an  L  or a cross and make two rows of four, i.e.  (6, 2, 4),  which is done by the simple trick of putting a coin on the intersection.

R. H. Macmillan.  Letter:  An old problem.  MG 30 (No. 289) (May 1946) 109.  Says he believes Newton and Sylvester studied this.  Says he has examples of  (11, 16, 3), (12, 19, 3),  (18, 18, 4),  (24, 28, 4),  (25, 30, 4),  (36, 55, 4),  (22, 15, 5),  (26, 21, 5),  (30, 26, 5).

G. C. Shephard.  A problem in orchards.  Eureka 9 (Apr 1947) 11-14.  Given  k  points in  n‑dimensions, the general problem is to draw  N(k, n)  hyperplanes to produce  k  regions, each containing one point.  The most common example is  k = 7,  n = 2,  N = 3.  [See Section 5.Q for determining  k  as a function of  n  and  N.]  The author investigates the question of determining the possible locations of the seventh point given six points.  He gives a construction of a set  T  such that being in  T  is necessary and sufficient for three such lines to exist.

J. Bridges.  Potter's orchard.  Eureka  11 (Jan 1949) 30  &  12 (Oct 1949) 17.  Start with an orchard  (9, 9, 3).  Add 16 trees to make  (25, 18, 5).  The nine trees are three points in a triangle, with the three midpoints of the sides and the three points halfway between these.  Six of the new trees are one third of the way along the sides of the original triangle; another six are one third of the way along the lines joining the midpoints of the original triangle; one point is the centre of the original triangle and the last three are easily seen.

W. Leslie Prout.  Think Again.  Frederick Warne & Co., London, 1958.  Thirteen rows of three, pp. 45 & 132.  (11, 13, 3).

Young World.  c1960.  Pp. 10-11.

Three coin lines.  (9, 10, 3).

Five coin lines.  (10, 5, 4).

Eleven coin trick.  (11, 12, 3).

Maxey Brooke.  Dots and lines.  RMM 6 (Dec 1961) 51‑55.  Cites Jackson and Dudeney.  Says Sylvester showed that  n  points can be arranged in at least  (n‑1)(n‑2)/6  rows of three.  Shows  (9, 10, 3)  and  (16, 15, 4).

R. L. Hutchings  &  J. D. Blake.  Problems drive 1962.  Eureka 25 (Oct 1962) 20-21 & 34-35.  Prob. F.  (10, 5, 4)  with points in the centres of cells of a chess board.  Actually only needs a  7 x 7  board.

Ripley's Puzzles and Games.  1966.  Pp. 18-19, item 4.  (17, 7, 5).

Doubleday - 3.  1972.  Count down, pp. 125-126.  Start with a  4 x 4  array of coins.  Add four coins so that each row, column and diagonal has the same number.  Solution doubles the coins in the  1, 3, 4, 2  positions in the rows.

S. A. Burr, B. Grünbaum & N. J. A. Sloane.  The orchard problem.  Geometria Dedicata 2 (1974) 397‑424.  Establishes good examples of  (a, b, 3)  slightly improving on Sylvester, and establishes some special better examples.  Gives upper bounds for  b  in (a, b, 3). Sketches history and tabulates best values and upper bounds for  b  in (a, b, 3), for  a = 1 (1) 32. 

                    The following have the maximal possible value of  b  for given  a  and  c.

          (3, 1, 3);  (4, 1, 3);  (5, 2, 3);  (6, 4, 3);  (7, 6, 3);  (8, 7, 3);  (9, 10, 3);  (10, 12, 3);  (11, 16, 3);  (12, 19, 3);  (16, 37, 3).

                    The following have the largest known value of  b  for the given  a  and  c. 

          (13, 22, 3);  (14, 26, 3);  (15, 31, 3);  (17, 40, 3);  (18, 46, 3);  (19, 52, 3);  (20, 57, 3);  (21, 64, 3);  (22, 70, 3);  (23, 77, 3);  (24, 85, 3);  (25, 92, 3);  (26, 100, 3);  (27, 109, 3);  (28, 117, 3);  (29, 126, 3);  (30, 136, 3);  (31, 145, 3);  (32, 155, 3).

M. Gardner.  SA (Aug 1976).  Surveys these problems, based on Burr, Grünbaum & Sloane.  He gives results for  c = 4. 

                    The following have the maximal possible value of  b  for the given  a  and  c.

          (4, 1, 4);  (5, 1, 4);  (6, 1, 4);  (7, 2, 4);  (8, 2, 4);  (9, 3, 4);  (10, 5, 4);  (11, 6, 4);  (12, 7, 4).

                    The following have the largest known value of  b  for the given  a  and  c.

          (13, 9, 4);  (14, 10, 4);  (15, 12, 4);  (16, 15, 4);  (17, 15, 4);  (18, 18, 4);  (19, 19, 4);  (20, 20, 4).

Putnam.  Puzzle Fun.  1978.  Nos. 17-23: Bingo arrangements, pp. 6 & 29-30.  (21, 11, 5),  (16, 15, 4),  (19, 9, 5),  (9, 10, 3),  (12, 7, 4),  (22, 21, 4),  (10, 5, 4).

S. A. Burr.  Planting trees.  In:  The Mathematical Gardner; ed. by David Klarner; Prindle, Weber & Schmidt/Wadsworth, 1981.  Pp. 90‑99.  Pleasant survey of the 1974 paper by Burr, et al.

Michel Criton.  Des points et des Lignes.  Jouer Jeux Mathématiques 3 (Jul/Sep 1991) 6-9.  Survey, with a graph showing  c  at  (a, b).  Observes that some solutions have points which are not at intersections of lines and proposes a more restrictive kind of arrangement of  b  lines whose intersections give  a  points with  c points on each line.  He denotes these with square brackets which I write as  [a, b, c].  Pictures of  (7, 6, 3),  [9, 8, 3],  (9, 9, 3),  (12, 6, 4),  [13, 9, 4],  (13, 12, 3),  (13, 22, 3), (16, 12, 4),  (19, 19, 4),  (19, 19, 5),  (20, 21, 4),  [21, 12, 5],  (25, 12, 5),  (30, 12, 7),  (30, 22, 5),  (49, 16, 7)  and mentions of  (9, 10, 3),  (16, 15, 4),

 

          6.AO.1.       PLACE FOUR POINTS EQUIDISTANTLY  =  MAKE FOUR TRIANGLES WITH SIX MATCHSTICKS

 

          I am adding the problem of making three squares with nine matchsticks here a it uses the same thought process -- see Mittenzwey and see the extended discussion at Anon., 1910.

 

Pacioli.  De Viribus.  c1500.  Ff. 191r - 192r.  LXXX. Do(cumento). commo non e possibile piu ch' tre ponti o ver tondi spere tocarse in un piano tutti (how it is not possible for more than three points or discs or spheres to all touch in a plane).  = Peirani 252-253.  Says you can only get three discs touching in the plane, but you can get a fourth so they are all touching by making a pyramid.

Endless Amusement II.  1826?  Prob. 21, p. 200.  "To place 4 poles in the ground, precisely at an equal distance from each other."  Uses a pyramidal mound of earth.

Young Man's Book.  1839.  P. 235.  Identical to Endless Amusement II.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 6, p. 178 (1868: 189).  Plant four trees at equal distances from each other.

Frank Bellew.  The Art of Amusing.  1866.  Op. cit. in 5.E.  1866: pp. 97-98 & 105-106;  1870: pp. 93‑94 & 101‑102.

Mittenzwey.  1880. 

Prob. 161, pp. 32 & 84;  1895?: 184, pp. 37 & 86;  1917: 184, pp. 34 & 83.  Use six sticks to make four congruent triangles.  Solution is a rectangle (should be a square) with its diagonals, but then two of the sticks have to be longer than the others.

Prob. 163, pp. 32 & 84;  1895?: 186 & 194, pp. 37 & 86-87;  1917: 186 & 194, pp. 34 & 83-84.  Use six equally long sticks to make four congruent triangles -- solution is a tetrahedron.  The two problems in the 1895? are differently phrased, but identical in content, while the first solution is a picture and the second is a description.

Prob. 171, pp. 33 & 85;  1895?: 195, pp. 38 & 87;  1917: 195, pp. 34 & 84.  Use nine equal sticks to make three squares.  Solution is three faces of a cube.

F. Chasemore.  Loc. cit. in 6.W.5.  1891.  Item 3: The triangle puzzle, p. 572.

Hoffmann.  1893. 

Chap. VII, no. 15, pp. 290 & 298 = Hoffmann-Hordern, pp. 195.  Four matches.

Chap. X, no. 19: The four wine glasses, pp. 344 & 381 = Hoffmann-Hordern, pp. 238‑239, with photo on p. 239 of a version by Jaques & Son, 1870-1900.  I usually solve the second version by setting one glass on top of the other three, but here he wants the centre of the feet of the glasses to be equally spaced and he turns one glass over and places it in the centre of the other three, appropriately spaced.

Loyd.  Problem 34: War‑ships at anchor.  Tit‑Bits 32 (22 May  &  12 Jun 1897) 135  &  193.  Place four warships equidistantly so that if one is attacked, the others can come to assist it.  Solution is a tetrahedron of points on the earth's oceans.

Parlour Games for Everybody.  John Leng, Dundee & London, nd [1903 -- BLC], p. 30.  "With 6 matches form 4 triangles of equal size."

Pearson.  1907.  Part III, no. 77: Three squares, p. 77.  Make three squares with nine matches.  Solution is a triangular prism!

Anon.  Prob. 66.  Hobbies 31  (No. 781) (1 Oct 1910) 2  &  (No. 784) (22 Oct 1910) 68.  Use nine matches to make three squares.  "... the only possible solution" is to make two adjacent squares with seven matches, then bisect each square to produce a third square which overlaps the other two.

                    I re-invented this problem in Apr 1999 and posted it on NOBNET on 19 Apr 1999.  Solution (1) is the idea I had when I made up the puzzle, but various friends gave more examples and then I found solution (3).

                    (1).  Arrange the nine matches to form the following.

                                    _

                              |_|  |_|

                                 |    |

 

          Then  4  is a square,  9  is a square and  49  is a square. 

                    (2).  Use the matches to form a triangular prism.  One may object that this also makes two triangles.

                    (3).  Make three squares forming three faces of a cube, all meeting at one corner.  Cf Mittenzwey 171.

                    (4).  Make two adjacent squares with seven of the matches.  Now bisect each of the squares with a match parallel to the common edge of the squares.  This produces a row of four adjacent half-squares as below.  The middle two form a new square.  Here one may object that the squares are overlapping.

                     ─── ─── 

                    │ │ │ │ │

                     ─── ─── 

                    (5).  Use the matches to make the figures  0,  1  and  4.

                    One can use the matches to make squares whose edge is half the match length, but one only needs eight matches to make three squares. 

                    There are other solutions which use the fact that matches have squared off ends and have square cross-section, but these properties do not hold for paper matches torn from a matchbook or for other equivalent objects like toothpicks and hence I don't consider them quite reasonable.

Anon.  Prob. 76.  Hobbies 31  (No. 791) (10 Dec 1910) 256  &  (No. 794) (31 Dec 1910) 318.  Make as many triangles as possible with six matches.  From the solution, it seems that the tetrahedron was expected with four triangles, but many submitted the figure of a triangle with its altitudes drawn, but only one solver noted that this figure contains 16 triangles!  However, if the altitudes are displaced to give an interior triangle, I find 17 triangles!!

Williams.  Home Entertainments.  1914.  Tricks with matches: To form four triangles with six matches, p. 106.

Blyth.  Match-Stick Magic.  1921.  Four triangle puzzle, p. 23.  Make four triangles with six matchsticks.

King.  Best 100.  1927.  No. 59, pp. 24 & 53.  = Foulsham's no. 20, pp. 8 & 12.  Use six matches to make four triangles.

 

          6.AO.2.       PLACE AN EVEN NUMBER ON EACH LINE

 

          See also section 6.T.

          Sometimes the diagonals are considered, but it is not always clear what is intended. 

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 564-31, pp. 254 & 396.  From a  6 x 6  array, remove  6  to leave an even number in each row.  (The German 'Reihe' can be interpreted as row or column or both.)  If we consider this in the first quadrant with coordinates going from 1 to 6, the removed points are:  (1,2), (1,3), (2,1), (2,2), (6,1), (6,3).  The use of the sixth column is peculiar and has the effect of making both diagonals odd, while the more usual use of the third column would make both diagonals even.

Prob. 583-5, pp. 285 & 403: Von folgenden 36 Punkten sechs zu streichen.  As above, but each file ('Zeile') in 'all four directions' has four or six points.  Deletes:  (1,1), (1,2), (2,2), (2,3), (6,1), (6,3)  which makes one diagonal even and one odd.

Mittenzwey.  1880.  Prob. 154, pp. 31 & 83;  1895?: 177, pp. 36 & 85;  1917: 177, pp. 33 & 82.  Given a  4 x 4  array, remove 6 to leave an even number in each row and column.  Solution removes a  2 x 3  rectangle from a corner.  [This fails -- it leaves two rows and a diagonal with an odd number.  One can use the idea mentioned for Leske 564-31 to get a solution with both diagonals also being even.]

Hoffmann.  1893.  Chap. VI, pp. 271-272 & 285 = Hoffmann-Hordern, pp. 186-187.

No. 22: The thirty‑six puzzle.  Place 30 counters on a  6 x 6  board so each horizontal and each vertical line has an even number.  Solution places the six blanks in a  3 x 3  corner in the obvious way.  This also makes the diagonals have even numbers. 

No. 23: The "Five to Four" puzzle.  Place 20 counters on a  5 x 5  board subject to the above conditions.  Solution puts blanks on the diagonal.  This also makes the diagonals have even number.

Dudeney.  The puzzle realm.  Cassell's Magazine ?? (May 1908) 713-716.  The crack shots.  10 pieces in a  4 x 4  array making the maximal number of even lines -- counting diagonals and short diagonals -- with an additional complication that pieces are hanging on vertical strings.  The picture is used in AM, prob. 270.

Loyd.  Cyclopedia.  1914.  The jolly friar's puzzle, pp. 307 & 380.  (= MPSL2, no. 155, pp. 109 & 172.  = SLAHP: A shifty little problem, pp. 64 & 110.)  10 men on a  4 x 4  board -- make a maximal number of even rows, including diagonals and short diagonals.  This is a simplification of Dudeney, 1908.

King.  Best 100.  1927.  No. 72, pp. 29 & 56.  As in Hoffmann's No. 22, but specifically asks for even diagonals as well.

The Bile Beans Puzzle Book.  1933.  No. 19: Thirty-six coins.  As in Hoffmann's No. 22, but specifically asks for even diagonals as well.

Rudin.  1936.  No. 151, pp. 53-54 & 111.  Place 12 counters on a  6 x 6  board with two in each 'row, column and diagonal'.  Reading the positions in each row, the solution is:  16, 34, 25, 25, 34, 16.  Some of the short diagonals and some of the broken diagonals are empty, so he presumably isn't including these, or he meant to ask for each of these to have an even number of at most two.

M. Adams.  Puzzle Book.  1939.  Prob. C.179: Even stars, pp. 169 & 193.  Same as Loyd.

Doubleday - 1.  1969.  Prob. 61: Milky Way, pp. 76 & 167.  = Doubleday - 5, pp. 85-86.  6 x 6  array with two opposite corners already filled.  Add ten more counters so that no row, column or diagonal has more than two counters in it.  Reading the positions in each row, the solution is:  13, 35, 12, 67, 24, 46.  Some short diagonals are empty or have one counter and some broken diagonals have one or four counters, so he seems to be ignoring them.  Hence this is the same problem as Rudin, but with a less satisfactory solution.

Obermair.  Op. cit. in 5.Z.1.  1984.  Prob. 37, pp. 38 & 68.  52 men on an  8 x 8  board with all rows, columns and diagonals (both long and short) having an even number.

 

          6.AP  .         DISSECTIONS OF A TETRAHEDRON

 

          6.AP.1.        TWO PIECES

 

Richard A. Proctor.  Our puzzles;  Knowledge 10 (Feb 1887) 83  &  Solutions of puzzles;  Knowledge 10 (Mar 1887) 108-109.  "Puzzle XIX.  Show how to cut a regular tetrahedron (equilateral triangular pyramid) so that the face cut shall be a square: also show how to plug a square hole with a tetrahedron."  Solution shows the cut clearly.

Edward T. Johnson.  US Patent 2,216,915 -- Puzzle.  Applied: 26 Apr 1939;  patented: 8 Oct 1940.  2pp + 1p diagrams.  Described in S&B, p. 46.

E. M. Wyatt.  Wonders in Wood.  Op. cit. in 6.AI.  1946.  Pp. 9 & 11: the tetrahedron or triangular pyramid.  P. 9 is reproduced in S&B, p. 46.

Donovan A. Johnson.  Paper Folding for the Mathematics Class.  NCTM, 1957, p. 26, section 62: Pyramid puzzle.  Gives instructions for making the pieces from paper.

Claude Birtwistle.  Editor's footnote.  MTg 21 (Winter 1962) 32.  "The following interesting puzzle was given to us recently."

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Bisected tetrahedron, pp. 157-158.  Gives the net so one can make a drawing, cut it out and fold it up to make one piece. 

 

          6.AP.2.        FOUR PIECES

 

          These dissections usually also work with a tetrahedron of spheres and hence these are related to ball pyramid puzzles, 6.AZ.

          The first version I had in mind dissects each of the two pieces of 6.AP.1 giving four congruent rhombic pyramids.  Alternatively, imagine a tetrahedron bisected by two of its midplanes, where a midplane goes halfway between a pair of opposite edges.  This puzzle has been available in various versions since at least the 1970s, including one from Stokes Publishing Co., 1292 Reamwood Avenue, Sunnyvale, California, 94089, USA., but I have no idea of the original source.  The same pieces are part of a more complex dissection of a cube, PolyPackPuzzle, which was produced by Stokes in 1996.  (I bought mine from Key Curriculum Press.)

          In 1997, Bill Ritchie, of Binary Arts, sent a quadrisection of the tetrahedron that they are producing.  Each piece is a hexahedron.  The easiest way to describe it is to consider the tetrahedron as a pile of spheres with four on an edge and hence  20  altogether.  Consider a planar triangle of six of these spheres with three on an edge and remove one vertex sphere to produce a trapezium (or trapezoid) shape.  Four of these assemble to make the tetrahedron.  Writing this has made me realise that Ray Bathke has made and sold these  5-sphere  pieces as Pyramid 4 for a few years.  However, the solid pieces used by Binary Arts are distinctly more deceptive.

          Len Gordon produced another quadrisection of the  20  sphere tetrahedron                        0       0

using the planar shape at the right.  This was c1980??                                                               0  0  0  

 

David Singmaster.  Sums of squares and pyramidal numbers.  MG 66 (No. 436) (Jun 1982) 100-104.  Consider a tetrahedron of spheres with  2n  on an edge.  The quadrisection described above gives four pyramids whose layers are the squares  1, 4, ..., n2.  Hence  four times the sum of the first  n  squares is the tetrahedral number for  2n,  i.e. 4 [1 + 4 + ... + n2]  =  BC(n+2, 3).  

 

          6.AQ. DISSECTIONS OF A CROSS,  T  OR  H

 

          The usual dissection of a cross has two diagonal cuts at  45o  to the sides and passing through two of the reflex corners of the cross and yielding five pieces.  The central piece is six-sided, looking like a rectangle with its ends pushed in and being symmetric.  Depending on the relative lengths of the arms, head and upright of the cross, the other pieces may be isosceles right triangles or right trapeziums.  Removing the head of the cross gives the usual dissection of the  T  into four pieces -- then the central piece is five-sided.  Sometimes the central piece is split in halves.  Occasionally the angle of the cuts is different than  45o.  Dissections of an  H  have the same basic idea of using cuts at  45o  -- the result can be a bit like two  Ts  with overlapping stems and the number of pieces depends on the relative size and positioning of the crossbar of the  H  -- see: Rohrbough.

 

S&B, pp. 20‑21, show several versions.  They say that crosses date from early 19C.  They show a 6‑piece Druid's Cross, by Edwards & Sons, London, c1855.  They show several  T‑puzzles -- they say the first is an 1903 advertisement for White Rose Ceylon Tea, NY -- but see 1898 below.  They also show some  H‑puzzles.

 

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has a cross cut into 5 pieces in the usual way.

Endless Amusement II.  1826?  Prob. 30, p. 207.  Usual five piece cross.  The three small pieces are equal.  = New Sphinx, c1840, pp. 139-140.

Crambrook.  1843.  P. 4.

No. 10: Five pieces to form a Cross.

No. 11: The new dissected Cross.

Without pictures, I cannot tell what dissections are used??

Boy's Own Book.  1843 (Paris): 435 & 440, no. 2.  Usual five piece cross, very similar to Endless Amusement.  One has to make three pieces of fig. 2.  = Boy's Treasury, 1844, pp. 424 & 428.  = de Savigny, 1846, pp. 353 & 357, no. 1.

Family Friend 2 (1850) 58 & 89.  Practical Puzzle -- No. II.  = Illustrated Boy's Own Treasury, 1860, No. 32, pp. 401 & 440.  Usual five piece cross to "form that which, viewed mentally, comforts the afflicted."  Three pieces of fig. 1.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 7, p. 178-179 (1868: 189).  Five piece dissection of a cross, but the statement of the problem doesn't say which piece to make multiple copies of.

Magician's Own Book.  1857.  Prob. 17: The cross puzzle, pp. 272 & 295.  Usual 5 piece cross, essentially identical to Family Friend, except this says to "form a cross."  = Book of 500 Puzzles, 1859, prob. 17, pp. 86 & 109.  = Boy's Own Conjuring Book, 1860, prob. 16, pp. 234 & 258. 

Charades, Enigmas, and Riddles.  1860: prob. 33, pp. 60 & 66;  1862: prob. 33, pp. 136 & 143;  1865: prob. 577, pp. 108 & 156.  Usual five piece cross, showing all five pieces.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 584-12, pp. 288 & 406: Ein Kreuz.  Begins as the usual five piece cross, but the central piece is then bisected into two mitres and the base has two bits cut off to give an eight piece puzzle.

Frank Bellew.  The Art of Amusing.  1866.  Op. cit. in 5.E.  1866: pp. 239-240;  1870: pp. 236‑238.  Usual five piece cross.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 1: The cross puzzle, pp. 27 & 30.  Usual five piece cross, but instructions say to cut three copies of the wrong piece.

Mittenzwey.  1880.  Prob. 188, pp. 35 & 88;  1895?: 213, pp. 40 & 91;  1917: 213, pp. 37 & 87.  This is supposed to be a 10 piece dissection of a cross obtained by further dissecting the usual five pieces.  However, pieces 3 & 4 are drawn as trapezoids in the problem and triangles in the solution and piece 2 in the solution is half the size given in the problem.  Further, pieces 1 & 2 appear equilateral in the problem, but are isosceles right triangles in the solution.  One could modify this to get a 9 piece version where eight of the pieces are right trapezoids -- four having edges  1, 1, 2, Ö2  and four having edges  Ö2, Ö2, 2Ö2, 2,  but the arms would be twice as long as they are wide.  Or one can make the second four pieces be  Ö2, Ö2, 2  isosceles right triangles.  In either case, the ninth piece would be a rectangle.

Lemon.  1890.  A card board puzzle, no. 33, pp. 8 & 98.  Usual five piece cross.

Hoffmann.  1893.  Chap. III, no. 12: The Latin cross puzzle, pp. 93 & 126 = Hoffmann‑Hordern, pp. 82-83, with photo.  As in Indoor & Outdoor.  Photo on p. 83 shows two versions: one in metal by Jaques & Sons, 1870-1895; the other in ivory, 1850-1900.  Hordern Collection, p. 59, shows a Druid's Cross Puzzle.

Lash, Inc. -- Clifton, N.J. -- Chicago, Ill. -- Anaheim, Calif.  T  Puzzle.  Copyright Sept. 1898.  4‑piece  T  puzzle to be cut out from a paper card, but the angle of the cuts is about  35o  instead of  45o  which makes it less symmetric and less confusing than the more common version.  The resulting  T  is somewhat wider than usual, being about  16%  wider than it is tall.  It advertises:  Lash's Bitters  The Original Tonic Laxative.  Photocopy sent by Slocum.

Benson.  1904.  The cross puzzle, pp. 191‑192.  Usual 5 piece version.

Wehman.  New Book of 200 Puzzles.  1908.  The cross puzzle, p. 17.  Usual 5 piece version.

A. Neely Hall.  Op. cit. in 6.F.5.  1918.  The T‑puzzle, pp. 19‑20.  "A famous old puzzle ...."  Usual 4‑piece version, but with long arms.

Western Puzzle Works, 1926 Catalogue.  No. 1394: Four pieces to form Letter  T.  The notched piece is less symmetric than usual.

Collins.  Book of Puzzles.  1927.  The crusader's cross puzzle, pp. 1-2.  The three small pieces are equal.

Arthur Mee's Children's Encyclopedia 'Wonder Box'.  The Children's Encyclopedia appeared in 1908, but versions continued until the 1950s.  This looks like 1930s??  Usual 5 piece cross.

A. F. Starkey.  The  T  puzzle.  Industrial Arts and Vocational Education 37 (1938) 442.  "An interesting novelty ...."

Rohrbough.  Puzzle Craft.  1932.  The "H" Puzzle, p. 23.  Very square  H  -- consider a  3 x 3  board with the top and bottom middle cells removed.  Make a cut along the main diagonal and two shorter cuts parallel to this to produce four congruent isosceles right triangles and two odd pentagons. 

See Rohrbough in 6.AS.1 for a very different  T  puzzle.

 

          6.AR. QUADRISECTED SQUARE PUZZLE

 

          This is usually done by two perpendicular cuts through the centre.  A dissection proof of the Theorem of Pythagoras described by Henry Perigal (Messenger of Mathematics 2 (1873) 104) uses the same shapes -- cf 6.AS.2. 

          The pieces make a number of other different shapes.

 

Crambrook.  1843.  P. 4, no. 17: Four pieces to form a Square.  This might be the dissection being considered here??

A. Héraud.  Jeux et Récréations Scientifiques -- Chimie, Histoire Naturelle, Mathématiques.  (1884);  Baillière, Paris, 1903.  Pp. 303‑304: Casse‑tête.  Uses two cuts which are perpendicular but are not through the centre.  He claims there are  120  ways to try to assemble it, but his mathematics is shaky -- he adds the numbers of ways at each stage rather than multiplying!  Also, as Strens notes in the margin of his copy (now at Calgary), if the crossing is off-centre, then many of the edges have different lengths and the number of ways to try is really only one.  Actually, I'm not at all sure what the number of ways to try is -- Héraud seems to assume one tries each orientation of each piece, but some intelligence sees that a piece can only fit one way beside another.

Handy Book for Boys and Girls.  Op. cit. in 6.F.3.  1892.  P. 14: The divided square puzzle.  Crossing is off-centre.

Tom Tit, vol 3.  1893.  Carré casse-tête, pp. 179-180.  = K, no. 26: Puzzle squares, pp. 68‑69.  = R&A, Puzzling squares, p. 99.  Not illustrated, but described:  cut a square into four parts by two perpendicular cuts, not necessarily through the centre.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 77: Pattern making, pp. 69-70 & 109.  Make five other shapes.

M. Adams.  Puzzle Book.  1939.  Prob. C.12: The broken square, pp. 125 & 173.  As above, but notes that the pieces also make a square with a square hole.

 

          6.AS. DISSECTION OF SQUARES INTO A SQUARE

 

Lorraine Mottershead.  Investigations in Mathematics.  Blackwell, Oxford, 1985.  P. 102 asserts that dissections of squares to various hexagons and heptagons were known c1800 while square to rectangle dissections were known to Montucla -- though she illustrates the latter with examples like 6.Y, she must mean 6.AS.5.

 

          6.AS.1.        TWENTY 1, 2, Ö5 TRIANGLES MAKE A SQUARE OR FIVE EQUAL SQUARES TO A SQUARE

 

          The basic puzzle has been varied in many ways by joining up the 20 triangles into various shapes, but I haven't attempted to consider all the modern variants.  A common form is a square with a skew  #  in it, with each line joining a corner to the midpoint of an opposite side, giving the 9 piece version.  This has four of the squares having a triangle cut off.  For symmetry, it is common to cut off a triangle from the fifth square, giving 10 pieces, though the assembly into one square doesn't need this.  See Les Amusemens for details.

          Cf Mason in 6.S.2 for a similar puzzle with twenty pieces.

          If the dividing lines are moved a bit toward the middle and the central square is bisected, we get a 10 piece puzzle, having two groups of four equal pieces and a group of two equal pieces, called the Japan square puzzle.  I have recently noted the connection of this puzzle with this section, so there may be other examples which I have not previously paid attention to -- see:  Magician's Own Book,  Book of 500 Puzzles,  Boy's Own Conjuring Book,  Illustrated Boy's Own Treasury,  Landells,  Hanky Panky,  Wehman.

 

Les Amusemens.  1749.  P. xxxii.  Consider five  2 x 2  squares.  Make a cut from a corner to the midpoint of an opposite side on each square.  This yields five  1, 2, Ö5  triangles and five pieces comprising three such triangles.  The problem says to make a square from five equal squares.  So this is the 10 piece version.

Minguet.  1755.  Pp. not noted -- ??check (1822: 145-146; 1864: 127-128).  Not in 1733 ed.  10  piece version.  Also a  15  piece version where triangles are cut off diagonally opposite corners of each small square leaving parallelogram pieces as in Guyot.

Vyse.  Tutor's Guide.  1771?  Prob. 6, 1793: p. 304, 1799: p. 317 & Key p. 357.  2 x 10  board to be cut into five pieces to make into a square.  Cut into a  2 x 2  square and four  2, 4, 2Ö5  triangles.

Ozanam‑Montucla.  1778.  Avec cinq quarrés égaux, en former un seul.  Prob. 18 & fig. 123, plate 15, 1778: 297;  1803: 292-293;  1814: 249-250;  1840: 127.  9 piece version.  Remarks that any number of squares can be made into a square -- see 6.AS.5.

Catel.  Kunst-Cabinet.  1790.

Das mathematische Viereck, pp. 10-11 & fig. 15 on plate I.  10 piece version with solution shown.  Notes these make five squares.

Das grosse mathematische Viereck, p. 11 & fig. 14 on plate I.  Cut the larger pieces to give five more  1, 2, Ö5  triangles and five  Ö5, Ö5, 2  triangles.  Again notes these make five squares.

Guyot.  Op. cit. in 6.P.2.  1799.  Vol. 2: première récréation: Cinq quarrés éqaux étant sonnés, en former un seul quarré, pp. 40‑41 & plate 6, opp. p. 37.  10 piece version.  Suggests cutting another triangle off each square to give 10 triangles and 5 parallelograms.

Bestelmeier.  1801.  Item 629: Die 5 geometrisch zerschnittenen Quadrate, um aus 5 ein einziges Quadrat zu machen.  As in Les Amusemens.  S&B say this is the first appearance of the puzzle.  Only shown in a box with one small square visible.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles.

No. 8, pp. 25 & 84 & plate I, fig. 5, no. 1.  = Vyse.

No. 10, pp. 25 & 84-85 & plate I, fig. 7, no. 1.  Five squares to one.  Nine piece version.

Rational Recreations.  1824.  Feat 35, pp. 164-166.  Usual 20 piece form.

Manuel des Sorciers.  1825.  Pp. 201-202, art. 18.  ??NX  Five squares to one -- usual 10 piece form and 15 piece form as in Guyot.

Endless Amusement II.  1826? 

[1837 only]  Prob. 35, p. 212.  20 triangles to form a square.  = New Sphinx, c1840, p. 141, with problem title: Dissected square.

Prob. 37, p. 215.  10 piece version.  = New Sphinx, c1840, p. 141.

Boy's Own Book.  The square of triangles.  1828: 426;  1828-2: 430;  1829 (US): 222;  1855: 576;  1868: 676.  Uses 20 triangles cut from a square of wood.  Cf 1843 (Paris) edition, below.  c= de Savigny, 1846, p. 272: Division d'un carré en vingt triangles.

Nuts to Crack IV (1835), no. 195.  20 triangles -- part of a long section: Tricks upon Travellers.  The problem is used as a wager and the smart-alec gets it wrong.

The Riddler.  1835.  The square of triangles, p. 8.  Identical to Boy's Own Book, but without illustration, some consequent changing of the text, and omitting the last comment.

Crambrook.  1843.  P. 4.

No. 7: Egyptian Puzzle.  Probably the 10 piece version as in Les Amusemens.  See S&B below, late 19C.  Check??

No. 23: Twenty Triangles to form a Square.  Check??

Boy's Own Book.  1843 (Paris): 436 & 441, no. 5: "Cut twenty triangles out of ten square pieces of wood; mix them together, and request a person to make an exact square with them."  As stated, this is impossible; it should be as in Boy's Own Book, 1828 etc., qv.  = Boy's Treasury, 1844, pp. 425 & 429.  = de Savigny, 1846, pp. 353 & 357, no. 4.  Also copied, with the error, in:  Magician's Own Book, 1857, prob. 29: The triangle puzzle;  Book of 500 Puzzles, 1859, prob. 29: The triangle puzzle;  Boy's Own Conjuring Book, 1860, prob. 28: The triangle puzzle.  c= Hanky Panky, 1872, p. 122.

Magician's Own Book.  1857. 

How to make five squares into a large one without any waste of stuff, p. 258.  9 piece version.

Prob. 29: The triangle puzzle, pp. 276 & 298.  Identical to Boy's Own Book, 1843 (Paris).

Prob. 35: The Japan square puzzle, pp. 277 & 300.  Make two parallel cuts and then two perpendicular to the first two so that a square is formed in the centre.  This gives a 9 piece puzzle, but here the central square is cut by a vertical through its centre to give a 10 piece puzzle.  = Landells, Boy's Own Toy-Maker, 1858, pp. 145-146. 

Charles Bailey (manufacturer in Manchester, Massachusetts).  1858.  An Ingenious Puzzle for the Amusement of Children ....  The 10 pieces of Les Amusemens, with 19 shapes to make, a la tangrams.  Sent by Jerry Slocum -- it is not clear if there were actual pieces with the printed material.

The Sociable.  1858. 

Prob. 10: The protean puzzle, pp. 289 & 305-306.  Cut a  5 x 1  into 11 pieces to form eight shapes, e.g. a Greek cross.  It is easier to describe the pieces if we start with a  10 x 2.  Then three squares are cut off.  One is halved into two  1 x 2  rectangles.  Two squares have two  1, 2, Ö5  triangles cut off leaving triangles of sides  2, Ö5, Ö5.  The remaining double square is almost divided into halves each with a  1, 2, Ö5  triangle cut off, but these two triangles remain connected along their sides of size  1,  thus giving a  4, Ö5, Ö5  triangle and two trapeziums of sides  2, 2, 1, Ö5.  = Book of 500 Puzzles, 1859, prob. 10, pp. 7 & 23-24.

Prob. 42: The mechanic's puzzle, pp. 298 & 317.  Cut a  10 x 2  in five pieces to make a square, as in Vyse.  = Book of 500 Puzzles, 1859, prob. 16, pp. 16 & 35.

Book of 500 Puzzles.  1859.

Prob. 10: The protean puzzle, pp. 7 & 23-24.  As in The Sociable.

Prob. 42: The mechanic's puzzle, pp. 16 & 35.  As in The Sociable.

How to make five squares into a large one without any waste of stuff, p. 72.  Identical to Magician's Own Book.

Prob. 29: The triangle puzzle, pp. 90 & 113.  Identical to Boy's Own Book, 1843 (Paris).

Prob. 35: The Japan square puzzle, pp. 91 & 114. 

Indoor & Outdoor.  c1859.  Part II, prob. 11: The mechanic's puzzle, pp. 130-131.  Identical to The Sociable.

Boy's Own Conjuring Book.  1860.

Prob. 28: The triangle puzzle, pp. 238 & 262.  Identical to Boy's Own Book, 1843 (Paris) and Magician's Own Book.

Prob. 34: The Japan square puzzle, pp. 240 & 264.  Identical to Magician's Own Book.

Illustrated Boy's Own Treasury.  1860.

Prob. 9, pp. 396 & 437.  [The Japan square puzzle.]  Almost identical to Magician's Own Book.

Optics: How to make five squares into a large one without any waste of stuff, p. 445.  Identical to Book of 500 Puzzles, p. 72.

Vinot.  1860.  Art. LXXV: Avec cinq carrés égaux, en faire un seul, p. 90.  Nine piece version.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 174, pp. 87-88.  Nine piece version.

Prob. 584-6, pp. 287 & 405.  Ten piece version of five squares to one.

Hanky Panky.  1872.

The puzzle of five pieces, p. 118.  9 piece version.

Another [square] of four triangles and a square, p. 120.  10 x 2  into five pieces to make a square.

[Another square] of ten pieces, pp. 121-122.  Same as the Japan square puzzles in Magician's Own Book.

[Another square] of twenty triangles, p. 122.  Similar to Boy's Own Book, 1843 (Paris), but with no diagram and less text, making it quite cryptic.

Mittenzwey.  1880.  Prob. 175, pp. 33-34 & 85;  1895?: 200, pp. 38 & 87;  1917: 200, pp. 35 & 84.  10 pieces as in Les Amusemens.  See in 6.AS.2 and 6.S.2 for the use of these pieces to make other shapes.

See Mason, 1880, in 6.S.2 for a similar, but different, 20 piece puzzle.

S&B, pp. 11 & 19, show a 10 piece version called 'Egyptian Puzzle', late 19C?

Lucas.  RM2.  1883.  Les vingt triangles, pp. 128‑129.  Notes that they also make five squares in the form of a cross.

Tom Tit, vol. 2.  1892.  Diviser un carré en cinq carrés égaux, pp. 147‑148.  = K, no. 2: To divide a square into five equal squares, pp. 12-14.  = R&A, Five easy pieces, p. 105.  Uses 9 pieces, but mentions use of 10 pieces.

Hoffmann.  1893. 

Chap. III, no. 21: The five squares, pp. 100 & 132‑133 = Hoffmann-Hordern, p. 94, with photo.  9 piece version, as in Magician's Own Book.  Photo on p. 94 shows: an ivory version, 1850-1900;, and a wood version, named Egyptian Puzzle, by C. N. Mackie, 1860-1890; both with boxes.

Chap. III, no. 24: The twenty triangles, pp. 101 & 134 = Hoffmann-Hordern, pp. 96-97, with photo.  As in Boy's Own Book.  Photo on p. 97 shows The Twenty Triangle Puzzle, with box, by Jaques & Son, 1870-1895.  Hordern Collection, p. 64, shows Apollonius, with box, by W. X., Paris, 1880-1900, in a solution very different to the usual one.

Chap. III, no. 30: The carpenter's puzzle -- no. 1, pp. 103 & 136‑137 = Hoffmann‑Hordern, p. 101.  Cut a  5 x 1  board into five pieces to make a square.

Chap. X, no. 25: The divided square, pp. 346 & 384 = Hoffmann-Hordern, p. 242.  9 piece puzzle as a dissection of a square which forms 5 equal squares.  He places the five squares together as a  2 x 2  with an adjacent  1 x 1,  but he doesn't see the connection with 6.AS.2.

Montgomery Ward & Co.  Catalog No 57, Spring & Summer, 1895.  Facsimile by Dover, 1969, ??NX.  P. 237 shows the Mystic Square, as item 25463, which is the standard 10 piece version.

Benson.  1904.

The carpenter's puzzle (No. 2), p. 191.  = Hoffmann, p. 103.

The five‑square puzzle, pp. 196‑197.  = Hoffmann, p. 100.

The triangle puzzle, p. 198.  = Hoffmann, p. 101.

Wehman.  New Book of 200 Puzzles.  1908.

P. 3: The triangle puzzle.  20 pieces.  = Boy's Own Book, omitting the adjuration to use wood and smooth the edges

P. 12: The protean puzzle.  c= The Sociable, prob. 10, with the instructions somewhat clarified.

P. 14: The Japan square puzzle.  c= Magician's Own Book.

P. 19: To make five squares into a large one.  10 piece version.

P. 27: The mechanic's puzzle.  = The Sociable, prob. 42.

J. K. Benson, ed.  The Pearson Puzzle Book.  C. Arthur Pearson, London, nd [c1910, not in BMC or NUC].  [This is almost identical with the puzzle section of Benson, but has 13 pages of different material.]  Juggling geometry, pp. 97-98.  Five triangles, which should be viewed as  2, 4, 2Ö5.  Cut one from the midpoint of the hypotenuse to the midpoint of the long leg and assemble into a square, so this becomes a six-piece or five-piece version as in Vyse, etc.

I have seen a 10 piece French example, called Jeu du Carré, dated 1900‑1920.

I have seen a 9(?) piece English example, dated early 20C, called The Euclid Puzzle.

Dudeney.  Perplexities column, no. 109: A cutting-out puzzle.  Strand Magazine  45 (No. 265) (Jan 1913) 113  &  (No. 266) (Feb 1913) 238.  c= AM, prob. 153 -- A cutting-out puzzle, pp. 37 & 172.  Cut a  5 x 1  to make a square.  He shows a solution in five pieces and asks for a solution in four pieces.  AM states the generalized form given in 6.AS.5.

Rohrbough.  Puzzle Craft.  1932.  Square "T", p. 23 (= The "T" Puzzle, p. 23 of 1940s?).  1 x 1  square and two  1 x 2  rectangles cut diagonally can be formed into a square or into a  T.

Gibson.  Op. cit. in 4.A.1.a.  1963.  Pp. 71 & 76: Square away.  A five piece puzzle, approximately that formed by drawing parallel lines from two diagonally opposite corners to the midpoints of opposite sides and then cutting a square from the middle of the central strip.  As drawn, the lines meet the opposite sides a bit further along than the midpoints.

Ripley's Puzzles and Games.  1966. 

Pp. 58-59, item 6.  Five right triangles to a square.  Though not specified, the triangles have sides proportional to  1, 2, Ö5.  Solution is as in Benson.

Pp. 60-61, item 5.  Start with the large square, which is  2Ö5  on a side.  Imagine the 9 piece puzzle where one line goes from the upper left corner to the midpoint of the right side.  Number the outer pieces clockwise from the upper left, so that pieces  1, 3, 5, 7  are the small triangles and  2, 4, 6, 8  are the trapezia.  The pieces of this puzzle are as follows:  combine pieces  1, 8, 7  into a  Ö5, 2Ö5, 5  triangle;  combine pieces  2, 3, 4  into an irregular hexagon;  take separate pieces 5 and 6 and the central square.  These five pieces form:  a square; a Greek cross; a  5 x 4  rectangle;  a triangle of sides  2Ö5, 4Ö5, 10;  etc.  I think there are earlier versions of this, e.g. in Loyd, but I have just observed the connection with this section.

 

          6.AS.1.a.     GREEK CROSS TO A SQUARE

 

          Note that a proper Greek cross is formed from five equal squares.

 

Lucas.  RM2.  1883.  Loc. cit. in 6.AS.1.  Uses 20 triangles.

Lemon.  1890.  The Maltese cross squared, no. 369, pp. 51 & 111.  Cut a Maltese cross (really a Greek cross) by two cuts into four pieces that make a square.

Hoffmann.  1893.  Chap. III, no. 13: The Greek cross puzzle, pp. 94 & 126 = Hoffmann‑Hordern, pp. 82 & 84.  Has four pieces made by two cuts.

Loyd.  Tit‑Bits 31 (10,  17  &  31 Oct 1896) 25,  39  &  75.  = Cyclopedia, 1914, p. 14.  Four pieces as in Hoffmann.

Loyd.  Problem 23: A new "square and cross" puzzle.  Tit‑Bits 31 (13 Mar 1897) 437  &  32 (3 Apr 1897) 3.  = Cyclopedia, 1914, pp. 58, 270 & 376.  Four congruent pieces.

Loyd.  Problem 27: The swastika problem.  Tit‑Bits 32 (3  &  24 Apr 1897) 3  &  59.  = Cyclopedia, 1914, p. 58.  Quadrisect square to make two equal Greek crosses.

Loyd.  Problem 30: The Easter problem.  Tit‑Bits 32 (24 Apr  &  15 May 1897) 59  &  117.  Dissect square into five pieces to make two unequal Greek crosses.

Dudeney.  Problem 56: Two new cross puzzles.  Tit‑Bits 33 (23 Oct  &  13 Nov 1897) 59  &  119.  Dissect a half square (formed by cutting a square either vertically or diagonally) to a Greek cross.  Solutions in 3 and 4 pieces.  [The first case  = Loyd, Cyclopedia, 1914, Easter 1903, pp. 46 & 345.]

Benson.  1904.  The Greek cross puzzle, p. 197.  = Hoffmann, p. 94.

Dudeney.  Cutting-out paper puzzles.  Cassell's Magazine ?? (Dec 1909) 187-191 & 233-235. 

States that the dissection with four pieces in two cuts is relatively 'recent'.  c= AM, 1917, p. 29, which dates this to 'the middle of the nineteenth century'.

Fold a Greek cross so that one cut gives four congruent pieces which form a square.  = AM, 1917, prob. 145, pp. 35 & 169.

M. Adams.  Indoor Games.  1912.  The Greek cross, p. 349 with figs. on p. 347.

 

          6.AS.1.b.     OTHER GREEK CROSS DISSECTIONS

 

          See also 6.F.3 and 6.F.5.

 

Dudeney.  A batch of puzzles.  Royal Magazine 1:3 (Jan 1899)  &  1:4 (Feb 1899) 368-372.  Squares and cross puzzle.  = AM, 1917, p. 34.  Dissect a Greek cross into five pieces which make two squares, one three times the edge of the other.  If the squares in the Greek cross have edge  Ö2,  then the cross has area  10  and the two squares have areas  1  and  9.  The dissection arise by joining the midpoints of the edges of the central square of the cross and extending these lines in one direction symmetrically.

Dudeney.  AM.  1917.  Greek cross puzzles, pp. 28-35.  This discusses a number of examples and gives a few problems.

Collins.  Book of Puzzles.  1927.  The Greek cross puzzle, pp. 98-100.  Take a Greek cross whose squares have side 2, so the cross has area 10.  Take another cross of area 5 and place it inside the large cross.  If this is done centrally and the small one turned to meet the edges of the large one, there are four congruent heptagonal pieces surrounding the small one which make another Greek cross of area 5.

Eric Kenneway.  More Magic Toys, Tricks and Illusions.  Beaver Books (Arrow (Hutchinson)), London, 1985.  On pp. 56-58, he considers a Greek cross cut by two pairs of parallel lines into nine pieces which would make five squares.  The lines join an outer corner to the midpoint of an opposite segment.  This produces a tilted square in the centre.  By pairing the other pieces, he gets four identical pieces which make a square and a Greek cross in a square.

 

          6.AS.2.        TWO (ADJACENT) SQUARES TO A SQUARE

 

          The smaller square often has half the edge of the larger, which connects this with 6.AS.1, but this is not essential.  The two squares are usually viewed as one piece, i.e. a  P‑pentomino.  These items are dissection proofs of the Theorem of Pythagoras -- see Yates (op. cit. in 6.B, pp. 38-39) for some other examples of this point.

          See Ripley's for a similar example, but the  2 x 2  square has a  Ö2, Ö2, 2  triangle attached to an edge.

          Another version has squares of area 1 and 8.  The area 8 square is  cut into four pieces which combine with the area 1 square to make an area 9 square.  I call this the 4 - 5 piece square.

 

Walther Karl Julius Lietzmann (1880-1959).  Der Pythagoreische Lehrsatz.  Teubner, (1911, 2nd ed., 1917), 6th ed., 1951.  [There was a 7th ed, 1953.]  Pp. 23-24 gives the standard dissection proof for the Theorem of Pythagoras.  The squares are adjacent and if considered as one piece, the dissection has three pieces.  He says it was known to Indian mathematicians at the end of the 9C as the Bride's Chair (Stuhl der Braut).  (I always thought this name referred to the figure of the Euclid I, 47 -- ??)

Thabit ibn Qurra (= Thābit ibn Qurra).  c875.  Gives the standard dissection proof for the Theorem of Pythagoras.  The squares are adjacent and if considered as one piece, the dissection has three pieces.  [Q. Mushtaq & A. L. Tan; Mathematics: The Islamic Legacy; Noor Publishing House, Farashkhan, Delhi, 1993, pp. 71-72] give this and cite Lietzmann.  Greg Frederickson [email of 18 Oct 1996] cites  Aydin Sayili; Thabit ibn Qurra's generalization of the Pythagorean theorem; Isis 51 (1960) 35-37.

Abu'l 'Abbas al-Fadhl ibn Hatim al-Narizi (or Annairizi).  (d. c922.)  Ed. by Maximilian Curtze, from a translation by Gherardo of Cremona, as: Anaritii In decem libros priores Elementorum Euclidis Commentarii,  IN: Euclidis Opera Omnia; Supplementum; Teubner, Leipzig, 1899.  ??NYS -- information supplied by Greg Frederickson.

Johann Christophorus Sturm.  Mathesis Enumerata, 1695, ??NYS.  Translated by J. Rogers? as:  Mathesis Enumerata: or, the Elements of the Mathematicks;  Robert Knaplock, London, 1700, ??NYS -- information provided by Greg Frederickson, email of 14 Jul 1995.  Fig. 29 shows it clearly and he attributes it to Frans van Schooten (the Younger, who was the more important one), but this source hasn't been traced yet.

Les Amusemens.  1749.  Prob. 216, p. 381 & fig. 97 on plate 8: Réduire les deux quarrés en un seul.  Usual dissection of two adjacent squares, attributed to 'Sturmius', a German mathematician, i.e the previous entry.

Ozanam‑Montucla.  1778.  Diverses démonstrations de la quarante-septieme du premier livre d'Euclide, ..., version 2.  Fig. 27, plate 4.  1778: 288;  1803: 284;  1814: 241-243;  1840: 123-124.  This is a version of the proof that  (a + b)2 = c2 + 4(ab/2),  but the diagram includes extra lines which produce the standard dissection of two adjacent triangles.

Crambrook.  1843.  P. 4, no. 19: One Square to form two Squares -- ??

E. S. Loomis.  The Pythagorean Proposition.  2nd ed., 1940;  reprinted by NCTM, 1968.  On pp. 194‑195, he describes the usual dissection by two cuts as Geometric Proof 165 and gives examples back to 1849, Schlömilch.

Family Friend 2 (1850) 298 & 353.  Practical Puzzle -- No. X.  = Illustrated Boy's Own Treasury, 1860, Prob. 11, pp. 397 & 437.  The larger square has twice the edge of the smaller and is shown divided into four, so this is clearly related to 6.AS.1, though the shape is considered as one piece, i.e. a  P-pentomino, to be cut into three parts to make a square.

Magician's Own Book.  1857.  To form a square, p. 261.  = Book of 500 Puzzles, 1859, p. 75.  An abbreviated version of Family Friend.  Refers to dotted lines in the figure which are drawn solid.

Charades, Enigmas, and Riddles.  1860: prob. 31, pp. 60 & 65;  1862: prob. 32, pp. 136 & 142;  1865: prob. 576, pp. 108 & 155.  Dissect a  P-pentomino into three parts which make a square.  Usual solution.

Peter Parley, the Younger.  Amusements of Science.  Peter Parley's Annual for 1866, pp. 139‑155. 

Pp. 143-144: "To form two squares of unequal size into one square, equal to both the original squares."  Usual method, with five pieces.  On pp. 146-148, he discusses the Theorem of Pythagoras and shows the dissection gives a proof of it.

P. 144: "To make two smaller squares out of one larger."  Cuts the larger square along both diagonals and assembles the pieces into two squares.

Hanky Panky.  1872.  To form a square, pp. 116-117.  Very similar to Magician's Own Book.

Henry Perigal.  Messenger of Mathematics 2 (1873) 104.  ??NYS -- described in Loomis, op. cit. above, pp. 104-105 & 214, where some earlier possible occurrences are mentioned.  He gives a dissection proof of the theorem of Pythagoras using the shapes that occur in the quadrisection of the square -- Section 6.AR.  For sides  a < b,  perpendicular cuts through the centre are made in the square of side  b  so they meet the sides at distance  (b-a)/2  from a corner.  These pieces then fit around the square of side  a  to make a square of side  c. 

                    I invented a hinged version of this, in the 1980s?, which is described in: Greg N. Frederickson; Hinged Dissections: Swinging & Twisting; CUP, 2002, pp. 33-34.  I am shown demonstrating this on Frederickson's website: www.cs.purdue.edu/homes/gnf/book2/Booknews2/singm.html . 

                    I have seen the assembly of these four pieces and the square of edge  a  into the square on the hypotenuse in a photo of the Tomb of Ezekiel in the village of Al-Kifil, near Hillah, Iraq.

Mittenzwey.  1880. 

Prob. 176, pp. 34 & 85;  1895?: 201, pp. 38 & 88;  1917: 201, pp. 35 & 84.  Use the 10 pieces of 6.AS.1, as in Les Amusemens, to make squares of edge 1 and edge 2.

Prob. 180, pp. 34 & 86;  1895?: 205, pp. 39 & 89;  1917: 205, pp. 35-36 & 85.  Cut a  2 x 2  and a  4 x 4  into five pieces which make a square.  Both the problem and the solutions are inaccurately drawn.  The smaller square has a  1, 2, Ö5  cut off, as for 6.AS.1.  The larger square has the same cut off at the lower left and a  2, 4, 2Ö5  cut off at the lower right -- these two touch at the midpoint of the bottom edge -- leaving a quadrilateral with edges  4, 2Ö5, Ö5, 3  and two right angles.  This is a variant of the standard five piece method.

Alf. A. Langley.  Letter:  Three-square puzzle.  Knowledge 1 (9 Dec 1881) 116, item 97.  Cuts two squares into five pieces which form a single square.

Alexander J. Ellis.  Letter:  The three-square puzzle.  Knowledge 1 (23 Dec 1881) 166, item 146.  Usual dissection of two adjacent squares, considered as one piece, into three parts by two cuts, which gives Langley's five pieces if the two squares are divided.  Suppose the two squares are on a single piece of paper and are  ABCD  and  DEFG,  with  E  on side  CD  of the larger square  ABCD.  He notes that if one folds the paper so that  B  and  F  coincide, then the fold line meets the line  ADG  at the point  H  such that the desired cuts are  BH  and  HF.

R. A. Proctor.  Letter or editorial reply:  Three square puzzle.  Knowledge 1 (30 Dec 1881) 184, item 152.  Says there have been many replies, cites Todhunter's Euclid, p. 266 and notes the pieces can be obtained by flipping the large square over and seeing how it cuts the two smaller ones.

R. A. Proctor.  Our mathematical column:  Notes on Euclid's first book.  Knowledge 5 (2 May 1884) 318.  "The following problem, forming a well-known "puzzle" exhibits an interesting proof of the 47th proposition."  Gives the usual three piece form, as in Ellis.

B. Brodie.  Letter:  Superposition.  Knowledge 5 (30 May 1884) 399, item 1273.  Response to the above, giving the five piece version, as in Langley.

Hoffmann.  1893.  Chap. III, no. 11: The two squares, pp. 93 & 125‑126 = Hoffmann-Hordern, pp. 82-83, with photo.  Smaller square has half the edge.  The squares are viewed as a single piece.  Photo on p. 83 shows The Five Squares Puzzle in paper with box, by Jaques & Son, 1870‑1895, and an ivory version, with box, 1850-1900.

Loyd.  Tit‑Bits 31 (3,  10  &  31 Oct 1896) 3,  25  &  75.  General two cut version.

Herr Meyer.  Puzzles.  The Boy's Own Paper 19 (No. 937) (26 Dec 1896) 206  &  (No. 948) (13 Mar 1897) 383.  As in Hoffmann.

Benson.  1904.  The two‑square puzzle, pp. 192‑193.

Pearson.  1907.  Part II, no. 108: Still a square, pp. 108 & 182.  Smaller square has half the edge.

Loyd.  Cyclopedia.  1914.  Pythagoras' classical problem, pp. 101 & 352.  c= SLAHP, pp. 15‑16 & 88.  The adjacent squares are viewed as one piece of wood to be cut.  Uses two cuts, three pieces.

Williams.  Home Entertainments.  1914.  Square puzzle, p. 118.  P-pentomino to be cut into three pieces to make a square.  No solution given.

A. W. Siddons.  Note 1020: Perigal's dissection for the Theorem of Pythagoras.  MG 16 (No. 217) (Feb 1932) 44.  Here he notes that the two cutting lines of Perigal's 1873 dissection do not have to go through the centre, but this gives dissections with more pieces.  He shows examples with six and seven pieces.  These cannot be hinged.

The Bile Beans Puzzle Book.  1933. 

No. 37: No waste.  Consider a square of side 2 extended by an isosceles triangle of hypotenuse 2.  Convert to a square using two cuts.

No. 39: Square building.  P-pentomino to square in two cuts.

Slocum.  Compendium.  Shows 4 - 5 piece square from Johnson Smith catalogue, 1935.

M. Adams.  Puzzle Book.  1939.  Prob. C.120: One table from two, pp. 154 & 185.  3 x 3  and  4 x 4  tiled squares to be made into a  5 x 5  but only cutting along the grid lines.  Solves with each table cut into two pieces.  (I think there are earlier examples of this -- I have just added this variant.)

Ripley's Puzzles and Games.  1966.  Pp. 58-59.

Item 5.  Two joined adjacent squares to a square, using two cuts and three pieces.

Item 6.  Consider a  2 x 2  square with a  Ö2, Ö2, 2  triangle attached to an edge.  Two cuts and three pieces to make a square.

 

          6.AS.2.a.     TWO EQUAL SQUARES TO A SQUARE

 

          Further subdivision of the pieces gives us 6.AS.4.

 

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles, no. 13, pp. 25 & 85-86.  Cut two equal squares each into two pieces to make a square.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 172, p. 87.  Cut one square into pieces to make two equal squares.  Cuts along the diagonals.

Mittenzwey.  1880.  Prob. 241, pp. 44 & 94;  1895?: 270, pp. 48 & 96;  1917: 270, pp. 44 & 92.  As is Leske.

 

          6.AS.3.        THREE EQUAL SQUARES TO A SQUARE

 

Crambrook.  1843.  P. 4, no. 21: One [Square to form] three [Squares] -- ??

"Student".  Proposal [A pretty geometrical problem].  Knowledge 1 (13 Jan 1882) 229, item 184.  Dissect an  L-tromino into a square.  Says there are  25  solutions -- editor says there are many more.

Editor.  A pretty geometrical problem.  Knowledge 1 (3 Mar 1882) 380.  Says only the proposer has given a correct solution, which cuts off one square, then cuts the remaining double square into three parts, so the solution has four pieces.  Says there are several other ways with four pieces and infinitely many with five pieces.

Hoffmann.  1893.  Chap. III, no. 23: The dissected square, pp. 101 & 134 = Hoffmann‑Hordern, pp. 96-97, with photo.  Cuts three squares identically into three pieces to form one square.  Photo on p. 97 shows The Dissected Square, with box, by Jaques & Son, 1870-1895.  Hordern Collection, p. 63, shows Arabian Puzzle, with box and some problem shapes to make, 1870‑1890.

Loyd.  Problem 3: The three squares puzzle.  Tit‑Bits 31 (17 Oct,  7  &  14 Nov 1896) 39,  97  &  112.  Quadrisect  3 x 1  rectangle to a square.  Sphinx (i.e. Dudeney) notes it also can be done with three pieces.

M. Adams.  Indoor Games.  1912.  The divided square, p. 349 with figs. on pp. 346-347.  3 squares, 4 cuts, 7 pieces.

Loyd.  Cyclopedia.  1914.  Pp. 14 & 341.  = SLAHP: Three in one, pp. 44 & 100.  Viewed as a  3 x 1  rectangle, solution uses 2 cuts, 3 pieces.  Viewed as 3 squares, there are 3 cuts, 6 pieces.

Johannes Lehmann.  Kurzweil durch Mathe.  Urania Verlag, Leipzig, 1980.  No. 6, pp. 61 & 160.  Claims the problem is posed by Abu'l-Wefa, late 10C, though other problems in this section are not strictly as posed by the historic figures cited.  Two of the squares to be divided into 8 parts so all nine parts make a square.  The solution has the general form of the quadrisection of the square of side  Ö2  folded around to surround a square of side 1 (as in Perigal's(?) dissection proof of the Theorem of Pythagoras), thus forming the square of side  Ö3.  The four quadrisection pieces are cut into two triangles of sides:  1, Ö3/2, (1+Ö2)/2  and  1, Ö3/2, (Ö2-1)/2.  Two of each shape assemble into a square of side 1 which can be viewed as having a diagonal cut and then cuts from the other corners to the diagonal, cutting off  (Ö2-1)/2  on the diagonal.

 

          6.AS.3.a.     THREE EQUAL 'SQUARES' TO A HEXAGON

 

Catel.  Kunst-Cabinet.  Vol. 2, 1793.  Das Parallelogramm, pp. 14-15 & fig. 249 a,b,c,d on plate XII.  This shows three squares, each dissected the same way into 4 pieces which will make a hexagon or two equal equilateral triangles.  Consider a hexagon and connect three alternate vertices to the centre.  Join up the same vertices and drop perpendiculars from the centre to three of the sides of the hexagon.  However, close examination shows that the squares have dimensions  3/2  by  Ö3.  The figure of the three adjacent squares has the divisions between them hard to make out.

Bestelmeier.  1801.  Item 292/293 -- Das Parallelogram.  Almost identical to Catel, except the diagrams are reversed, and worse, several of the lines are missing.  Mathematical part of the text is identical.

 

          6.AS.4.        EIGHT EQUAL SQUARES TO A SQUARE

 

          Divide four of the squares in half diagonally.

 

Magician's Own Book.  1857.  Prob. 8: The accommodating square, pp. 269 & 293.  c= Landells, Boy's Own Toy-Maker, 1858, p. 144.  = Book of 500 Puzzles, 1859, prob. 8, pp. 83 & 107.  = Boy's Own Conjuring Book, 1860, prob. 7, pp. 230 & 256.  = Illustrated Boy's Own Treasury, 1860, no. 24, pp. 399 & 439.

Hanky Panky.  1872.  [Another square] of four squares and eight triangles, p. 120.

Cassell's.  1881.  Pp. 92-93: The accommodating square.  = Manson, 1911, p. 131.

Handy Book for Boys and Girls.  Op. cit. in 6.F.3.  1892.  Pp. 321-322: Square puzzle.

Hoffmann.  1893.  Chap. III, no. 20: Eight squares in one, pp. 100 & 132 = Hoffmann‑Hordern, p. 94.

Wehman.  New Book of 200 Puzzles.  1908.  The accommodating square, p. 13.  c= Magician's Own Book.

 

          6.AS.5.        RECTANGLE TO A SQUARE OR OTHER RECTANGLE

 

          New section.  See comment at 6.AS.  The dissection of a  5 x 1  into five pieces which make a square is explicitly covered in 6.AS.1, and the other cases above can be viewed as dissections of  2 x 1,  3 x 1  and  8 x 1.  There must be older examples of the general case??

 

Ozanam‑Montucla.  1778.

Avec cinq quarrés égaux, en former un seul.  Prob. 18 & fig. 123, plate 15, 1778: 297;  1803: 292-293;  1814: 249-250;  1840: 127.  9 pieces.  Remarks that any number of squares can be made into a square.

Prob. 19 & fig. 124-126, plate 15 & 16, 1778: 297-301;  1803: 293-296;  1814: 250‑253;  1840: 127-129.  Dissect a rectangle to a square.

Prob. 20 & fig. 125-126, plate 15 & 16, 1778: 301-302;  1803: 297;  1814: 253;  1840: 129.  Dissect a square into  4, 5, 6, etc. parts which form a rectangle.

"Mogul".  Proposal [A pretty geometrical problem].  Knowledge 1 (13 Jan 1882) 229, item 184.  Dissect a rectangle into a square.  Editor's comment in (3 Mar 1882) 380 says only the proposer has given a correct solution but it will be held over.

"Mogul".  Mogul's Problem.  Knowledge 1 (31 Mar 1882) 483.  Gives a general construction, noting that if the ratio of length to width is  £ 2,  then it takes two cuts;  if the ratio is in the interval  (2, 5],  it takes three cuts;  if the ratio is in  (5, 10],  it takes four cuts;  if the ratio is in (10, 17], it takes five cuts.  In general if the ratio is in  (n2+1, (n+1)2+1],  it takes  n+2  cuts.

Richard A. Proctor.  Our puzzles;  Knowledge 10  (Nov 1886) 9  &  (Dec 1886) 39-40  &  Solution of puzzles;  Knowledge 10 (Jan 1887) 60-61.  "Puzzle XII.  Given a rectangular carpet of any shape and size to divide it with the fewest possible cuts so as to fit a rectangular floor of equal size but of any shape."  He says this was previously given and solved by "Mogul".  Solution notes that this is not the problem posed by "Mogul" and that the shape of the second rectangle is assumed as given.  He distinguishes between the cases where the actual second rectangular area is given and where only its shape is given.  Gives some solutions, remarking that more cuts may be needed if either rectangle is very long.  Poses similar problems for a parallelogram.

Tom Tit, vol. 3.  1893.  Rectangle changé en carré. en deux coups de ciseaux, pp. 175-176.  = K, no. 24: By two cuts to change a rectangle into a square, pp. 64-65.  Consider a square  ABCD  of side one.  If you draw  AA'  at angle  α  to  AB  and then drop  BE  perpendicular to  AA', the resulting three pieces make a rectangle of size  sin α  by  csc α,  where  α  must be  ³ 450,  so the rectangle cannot be more than twice as long as it is wide.  If one starts with such a rectangle  ABCD,  where  AB  is the length, then one draws  AA'  so that  DA'  is the geometric mean of  AB  and  AB - AD.  Dropping  CE  perpendicular to  AA'  gives the second cut.

Dudeney.  Perplexities column, no. 109: A cutting-out puzzle.  Strand Magazine  45 (No. 265) (Jan 1913) 113  &  (No. 266) (Feb 1913) 238.  c= AM, prob. 153 -- A cutting-out puzzle, pp. 37 & 172.  Cut a  5 x 1  into four pieces to make a square.  AM states the generalized form:  if  length/breadth  is in  [(n+1)2, n2),  then it can be done with  n+2  pieces, of which  n-1  are rectangles of the same breadth but having the desired length.  The cases  1 x (n+1)2 are exceptional in that one of pieces vanishes, so only  n+1  pieces are needed.  He doesn't describe this fully and I think one can change the interval above to  ((n+1)2, n2].

Anonymous.  Two dissection problems, no. 1.  Eureka  13 (Oct 1950) 6  &  14 (Oct 1951) 23.  An  n-step  is formed by  n  lines of unit squares of lengths  1, 2, ..., n,  with all lines aligned at one end.  Hence a  1-step  is a unit square, a  2-step is an L-tromino and an  n‑step is what is left when an  (n-1)-step is removed from a corner of an  n x n  square.  Show any  n-step  can be cut into four pieces to make a square, with three pieces in one case.  Cut parallel to a long side at distance  (n+1)/2  from it.  The small piece can be rotated  180o about a corner to make an  n x (n+1)/2  rectangle.  Dudeney's method cuts this into three pieces which make a square, and the cuts do not cut the small part, so we can do this with a total of four pieces.  When  n = 8,  the rectangles is  8 x 9/2,  which is similar to  16 x 9  which can be cut into two pieces by a staircase cut, so the problem can be done with a total of three pieces.  A little calculation shows this is the only case where  n x (n+1)/2  is similar to  k2 x (k-1)2.

Lorraine Mottershead.  Investigations in Mathematics.  Blackwell, Oxford, 1985.  P. 105.  Dissect a  2 x 5  rectangle into four pieces that make a square.

 

          6.AT. POLYHEDRA AND TESSELLATIONS

 

          These have been extensively studied, so I give only the major works.  See 6.AA for nets of polyhedra.

 

          6.AT.1.        REGULAR POLYHEDRA

 

Gwen White.  Antique Toys and Their Background.  Batsford, 1971.  (Reprinted by Chancellor Press, London, nd [c1989].)  P. 9 has a sketch of "Ball of stone, Scotland", which seems to be tetrahedral and she says: "... one of the earliest toys known is a stone ball.  Perhaps it is not a plaything, no one knows why it was made, but it is a convenient size to hold in the hand."

Dorothy N. Marshall.  Carved stone balls.  Proc. Soc. of Antiquaries of Scotland 108 (1976-7) 40-72.  Survey of the Scottish neolithic carved stone balls.  Lists 387 examples in 36 museums and private collections, mostly of 70mm diameter and mostly from eastern Scotland.  Unfortunately Marshall is not interested in the geometry and doesn't clearly describe the patterns -- she describes balls with  3, 4, 5, 6, 7, 8, 9, 10 - 55  and  70 - 160 knobs, but emphasises the decorative styles.  From the figures, there are clearly tetrahedral, cubical, dodecahedral(?) and cubo-octahedral shapes.  Many are in the National Museum of Antiquities of Scotland (= Royal Museum, see below), but the catalogue uses a number of unexplained abbreviations of collections.

Royal Museum of Scotland, Queen Street, Edinburgh, has several dozen balls on display, showing cubical, tetrahedral, octahedral and dodecahedral symmetry, and one in the form of the dual of the pentagonal prism.  [This museum has now moved to a new building beside its other site in Chambers Street and has been renamed the Museum of Scotland.  When I visited in 1999, I was dismayed to find that only three of the carved stone balls were on display, in a dimly lit case and some distance behind the glass so that it was difficult to see them.  Admittedly, the most famous example, the tetrahedral example with elaborate celtic decorative spirals, NMA AS10 from Glasshill, Towie, Aberdeenshire, is on display -- photo in [Jenni Calder; Museum of Scotland; NMS Publishing, 1998, p. 21].  They are on Level 0 in the section called In Touch with the Gods.]

Ashmolean Museum, Oxford, has six balls on display in case 13a of the John Evans Room.  One is tetrahedral, three are cubical, one is dodecahedral and one is unclear. 

Keith Critchlow.  Time Stands Still -- New Light on Megalithic Science.  Gordon Fraser, London, 1979.  Chap. 7: Platonic spheres -- a millennium before Plato, pp. 131‑149.  He discusses and depicts Neolithic Scottish stones carved into rounded polyhedral shapes.  All the regular polyhedra and the cubo‑octahedron occur.  He is a bit vague on locations -- a map shows about 50 discovery sites and he indicates that some of these stones are in the Ashmolean Museum, Dundee City Museum and 'in Edinburgh'.  Likewise, the dating is not clear -- he only says 'Neolithic' -- and there seem to be no references.

D. V. Clark.  Symbols of Power at the Time of Stonehenge.  National Museum of Antiquities, Edinburgh, 1985.  Pp. 56-62 & 171.  ??NYS -- cited by the Christie's Catalogue, below.

Robert Dixon.  Mathographics.  Blackwell, 1987, fig. 5.1B, p. 130, is a good photo of the Towie example.

Anna Ritchie.  Scotland BC.  HMSO, Edinburgh, for Scottish Development Department -- Historic Buildings and Monuments, 1988. 

P. 8 has a colour photo of a neolithic cubical ball from the Dark Age fort of Dunadd, Argyll.

P. 14 has a colour photo of a cubical and a tetrahedral ball from Skara Brae, Orkney Islands, c-2800.

Simant Bostock of Glastonbury has made a facsimile of the Towie example, casts of which are available from Glastonbury Film Office, 3 Market Place, Glastonbury, Somerset, BA6 9HD; tel: 01458-830228.  You can also contact him at 24 Northload Street, Glastonbury, Somerset, BA6 9JJ; tel: 01458-833267 and he has a mail order catalogue.  Since he worked from photographs, there are some slight differences from the original, and the facsimile is slightly larger.

Three examples of tetrahedral stone balls were in Christie's South Kensington antiquities sale of 12 Apr 2000, lots 124 and 125 (2 balls), p. 62, with colour photo of item 124 and the better example in lot 124 on p. 63.  (Thanks to Christine Insley Green for a copy of the catalogue.)  The descriptive text says 'their exact use is unclear'.  Cites Clark, above.

 

The British Museum has icosahedral dice from Egypt, dated -700/-500.

Moritz Cantor.  Vorlesungen über Geschichte der Mathematik.  Vol. I, 4th ed., 1906, pp. 174‑176.  He feels all the regular solids were known to Pythagoras, with the tetrahedron, cube and octahedron having been known long before.  Says to see various notices by Count Leopold Hugo in CR 77 for a bronze dodecahedron, a work by Conze on a Celtic bronze example and the paper of Lindemann, below, for a north Italian example.  However, he says the dates of these are not determined and I think these are now all dated to later Roman times -- see below.  He also notes that moderately regular dodecahedra and icosahedra occur in mineral deposits on Elba and in the Alps and wonders if Pythagoras could have known of these.

Thomas L. Heath.  Note about Scholium 1 of Book XIII of Euclid.  The Thirteen Books of Euclid's Elements; trans. & ed. by Thomas L. Heath; (1908?); 2nd ed., (1926); Dover, 3 vols., 1956, vol. 3, p. 438.  "And it appears that dodecahedra have been found, of bronze or other material, which may belong to periods earlier than Pythagoras' time by some centuries (for references see Cantor's Geschichte der Mathematik I3, pp. 175-6)." 

HGM I 160 cites Hugo and Lindemann, dating the Monte Loffa example as -1000/-500.

HGM I 162 discusses the Scholium, giving it as:  "the five so-called Platonic figures, which, however, do not belong to Plato, three of the five being due to the Pythagoreans, namely the cube, the pyramid, and the dodecahedron, while the octahedron and icosahedron are due to Theaetetus".  He cites Heiberg's Euclid, vol. v., p. 654.

Thomas, SIHGM I 223 says "A number of objects of dodecahedral form have survived from pre-Pythagorean days."  But he gives no details or references.  Cf Heath's note to Euclid, above.

Plato (‑427/‑347).  Timaeus.  c-350.  Page references are based on the 1578 edition of Plato which has been used for all later references: pp. 54-56.  I use the version in: Edith Hamilton and Huntington Cairns, eds; The Collected Dialogues of Plato including the Letters; Bollingen Series LXXI, Pantheon Books (Random House), (1961), corrected 2nd ptg, 1963, which is the translation of Benjamin Jowett in his The Dialogues of Plato, OUP, (1872), 4th revised ed, 1953.  Discusses the regular polyhedra, describing the construction of the tetrahedron, octahedron, icosahedron and cube from triangles since he views the equilateral triangle as made from six 30-60-90 triangles and the square made from eight 45-45-90 triangles.  "There was yet a fifth combination which God used in the delineation of the universe with figures of animals."  He then relates the first four to the elements:  tetrahedron -- fire;  octahedron -- air;  icosahedron -- water;  cube -- earth.  [These associations are believed to derive from the Pythagoreans.]  However, these associations contribute very little to the rest of the dialogue.  The incidental appearance of the dodecahedron lends support to the belief that it was discovered or became known after the initial relation between regular polyhedra and the elements had been established and had to be added in some ad hoc manner.  [It is believed that the later Pythagoreans related it to the universe as a whole.]

Scholium 1 of Book XIII of Euclid.  Discussed in: The Thirteen Books of Euclid's Elements; op. cit. above, vol. 3, p. 438.  Heath's discussion of the Scholia in vol. 1, pp. 64-74, indicates this may be c600.  The Scholium asserts that only the tetrahedron, cube and dodecahedron were known to the Pythagoreans and that the other two were due to Theaetetus.  Heath thinks the Pythagoreans had all five solids (cf his note to IV.10, vol. 2, pp. 97-100) and the Scholium is taken from Geminus, who may have been influenced by the fact that Theaetetus was the first to write about all five solids and hence the first to write much about the latter two polyhedra.

Stefano de'Stefani.  Intorno un dodecaedro quasi regolare di pietra a facce pentagonali scolpite con cifre scoperto nelle antichissime capanne di pietra del Monte Loffa.  Atti del Reale Istituto Veneto di Scienze e Lettere, (Ser. 6) 4 (1885) 1437-1459 + plate 18.  Separately reprinted by G. Antonelli, Venezia, 1886, which has pp. 1-25 and Tavola 18.  This describes perhaps the oldest known reasonably regular dodecahedron, in the Museo Civico di Storia Naturale, Palazzo Pompei, Largo Porta Vittoria 9, Verona, Italy, in the central case of Sala XIX.  This is discussed by Herz-Fischler [op. cit. below, p. 61], I have been to see it and the Director, Dr. Alessandra Aspes, has kindly sent me a slide and a photocopy of this article.

                    The dodecahedron was discovered in 1886 at Monte Loffa, NE of Verona, and has been dated as far back as -10C, but is currently considered to be -3C or -2C [Herz‑Fischler, p. 61].  Dr. Aspes said the site was inhabited by tribes who had retreated into the mountains when the Romans came to the area, c-3C.  These tribes were friendly with the Romans and were assimilated over a few centuries, so it is not possible to know if this object belongs to the pre-Roman culture or was due to Roman influence.  She dates it as ‑4C/‑1C.  The stone apparently was cut with a bronze saw and these existed before the Roman incursion (stated in Lindemann, below).  It is clearly not perfectly regular -- some of the face angles appear to be 90o and some edges are clearly much shorter than others.  But it also seems clear that it is an attempt at a regular dodecahedron -- the faces are quite flat.  Its faces are marked with holes and lines, but their meaning and the function of the object are unknown.  de'Stefani conjectures it is a kind of die.  Lindemann notes that the symbols are not Etruscan nor Greek, but eventually gets to an interpretation of them, which seems not too fanciful, using the values:  3,  6,  9,  10,  12,  15,  16,  20,  21,  24,  60,  300.  (Are there any  ancient Greek models of the regular polyhedra?)  But see also the carved stone balls above. 

 

About 90 examples of a Roman dodecahedron have been found at Roman sites, north of the Alps, from Britain to the Balkans, dating about 200-400.  These are bronze and hollow, but also each face has a hole in it, almost always circular, and each corner has a knob at it, making it look like it could be used for Hamilton's Icosian game!  The shape is quite precise.

The Society of Antiquaries, London, has the largest extant specimen, dug up on the north side of the Church of St. Mary, Carmarthen, in 1768 and presented to the Society about 1780.  The edge length is 2 1/5 in (= 56mm) and, unusually, has plain faces -- almost all examples have some incised decoration on the faces.  [Rupert Bruce-Mitford; The Society of Antiquaries of London  Notes on Its History and Possessions; The Society, 1951, p. 75, with photograph as pl. XXIV (b) on p. 74.

The Gallo-Romeins Museum (Kielenstraat 15, B-3700 Tongeren, Belgium; Tel: 12-233914) has an example which is the subject of an exhibition and they have produced facsimiles for sale.  "The precise significance and exact use of this object have never been explained and remain a great mystery."  Luc de Smet says the bronze facsimile is slightly smaller than the original and that the Museum also sells a tin and a bronzed version of the original size. 

Other examples are in the Newcastle University Museum (only about half present) and the Hunt Museum, Limerick.

C. W. Ceram.  Gods, Graves and Scholars.  Knopf, New York, 1956, pp. 26-29.  2nd ed., Gollancz, London, 1971, pp. 24-25.  In the first edition, he illustrated this as an example of the mysterious objects which archaeologists turn up and said that it had been described as a toy, a die, a model for teaching measurement of cylinders, a candleholder.  His picture shows one opening as being like a key-hole.  In the second edition, he added that he had over a hundred suggestions as to what it was for and thinks the most probable answer is that it was a musical instrument.

Jacques Haubrich has recently sent an example of a hollow cubical stone object with different size holes in the faces, apparently currently made in India, sold as a candleholder.

 

See Thomas, SIHGM I 216‑225, for brief references by Philolaus, Aëtius, Plato, Iamblichus.

Euclid.  Elements.  c-300.  Book 13, props. 13‑18 and following text.  (The Thirteen Books of Euclid's Elements, edited by Sir Thomas L. Heath.  2nd ed., (CUP, 1925??),  Dover, vol. 3, pp. 467‑511.)  Constructs the 5 regular polyhedra in a sphere, compares them.  In Prop. 18, he continues "I say next that no other figure, besides the said five figures, can be constructed which is contained by equilateral and equiangular figures equal to one another."

Leonardo Pisano, called Fibonacci (c1170->1240).  La Practica Geometriae.  1221.  As:  La  Practica Geometriae  di  Leonardo Pisano  secondo la lezione  del Codice Urbinate no. 202  della Biblioteca Vaticana.  In: Scritti di Leonardo Pisano; vol. II, ed. and pub. by B. Boncompagni; Tipografia delle Scienze Matematiche e Fisiche, Rome, 1862, pp. 1‑224.  On p. 159, he says there are many polyhedra and mentions there are ones with 8, 12 and 20 faces which Euclid constructs in a sphere in his book XIIII.  On pp. 161-162, he describes division in mean and extreme ratio and the construction of the regular pentagon in a circle, then says you can construct, in a sphere, a solid with 20 equilateral triangular faces or with 12 pentagonal faces.  After some discussion, he says you can also construct solids with  4, 6, 8, 12, 20  faces, in a sphere.  Division in mean and extreme ratio and the construction of the icosahedron are later covered in detail on pp. 196-202.  His only drawings of solids are of cubes and pyramids.

Drawings of all the regular polyhedra are included in works, cited in 6.AA, 6.AT.2 and 6.AT.3, by della Francesca (c1480 & c1487), Pacioli (1494), Pacioli & da Vinci (1498), Dürer (1525), Jamnitzer (1568), and Kepler (1619).

F. Lindemann.  Zur Geschichte der Polyeder und der Zahlzeichen.  Sitzungsber. der math.‑phys. Classe k. b. Akademie der Wissenschaften zu München 26 (1896) 625‑758 & plates I-IX.  Discusses and illustrates many ancient polyhedra.  Unfortunately, most of these are undated and/or without provenance.  He generally dates them as  -7C/5C.

                    A bronze rhombic triacontahedron, which he dates as first half of the first millennium AD.

                    Roman knobbed dodecahedra, which he describes as Celtic, going back to the La Tène period (Bronze Age) -- these are now dated to late Roman times.  He lists 26 examples listed from the works of Conze and Hugo (cf Cantor, above). 

                    A dodecahedral die;  an irregular rhombi-cubo-octahedral die;  a bronze dodecahedral die (having two 1s, three 2s, two 3s, one 4 and four 5s).

                    The Verona dodecahedron (from de'Stefani), which he dates as -1000/-500.

                    An enamelled icosahedron in Turin with Greek letters on the faces.

                    An octagonal bipyramid (elongated) from Meclo, South Tyrol, marked with a form of Roman numerals in a somewhat irregular order.  It is dated to before the Barbarian migrations.

                    Three bronze cubo-octahedra.

                    He then does a long analysis of north-Italian culture and its relations to other cultures and of their number symbols, eventually obtaining an interpretation of the symbols on the Monte Loffa dodecahedron, which he then justifies further with Pythagorean number relations.

Roger Herz-Fischler.  A Mathematical History of Division in Extreme and Mean Ratio.  Wilfrid Laurier University Press, Waterloo, Ontario, 1987.  Corrected and extended as: A Mathematical History of the Golden Number; Dover, 1998.  P. 61 discusses the history of the dodecahedron and refers to the best articles on the history of polyhedra.  Discusses the Verona dodecahedron, see above.

Judith V. Field.  Kepler's Geometrical Cosmology.  Athlone Press, London, 1988.  This gives a good survey of the work of Kepler and his predecessors.  In particular, Appendix 4: Kepler and the rhombic solids, pp. 201-219, is most informative.  Kepler described most of his ideas several times and this book describes all of them and the relationships among the various versions.

The regular polyhedra in four dimensions were described by Ludwig Schläfli, c1850, but this was not recognised and in the 1880s, several authors rediscovered them.

H. S. M. Coxeter.  Regular skew polyhedra.  Proc. London Math. Soc. (2) 43 (1937) 33-62.  ??NYS -- cited and discussed by Gott, qv.

J. R. Gott III.  Pseudopolyhedrons.  AMM 74:5 (May 1967) 497-504.  Regular polyhedra have their sum of face angles at a vertex being less than 360o and approximate to surfaces of constant positive curvature, while tessellations, with angle sum equal to 360o, correspond to surfaces of zero curvature.  The pseudopolyhedra have angle sum greater than 360o and approximate to surfaces of negative curvature.  There are seven regular pseudopolyhedra.  Each is a periodic structure.  He subsequently discovered that J. F. Petrie and Coxeter had discovered three of these in 1926 and had shown that they were the only examples satisfying an additional condition that arrangement of polygons at any vertex have rotational symmetry, and hence that the dihedral angles between adjacent faces are all equal.  Coxeter later refers to these structures as regular honeycombs.  Some of Gott's examples have some dihedral angles of 180o.  Two of these consist of two planes, with a regular replacement of pieces in the planes by pieces joining the two planes.  The other five examples go to infinity in all directions and divide space into two congruent parts.  He makes some remarks about extending this to general and Archimedean pseudopolyhedra.

 

          6.AT.2         STAR AND STELLATED POLYHEDRA

 

I have heard it stated that Kircher was the first to draw star polygons.

Paolo Uccello (1397‑1475).  Mosaic square on the floor at the door of San Pietro in San Marco, Venice.  1425‑1430.  (This doorway is not labelled on the maps that I have seen -- it is the inner doorway corresponding to the outer doorway second from the left, i.e. between Porta di Sant'Alipio and Porta di San Clemente, which are often labelled.)  This seems to show the small stellated dodecahedron  {5/2, 5}.  This mosaic has only recently (1955 & 1957) been attributed to Uccello, so it can only be found in more recent books on him.  See, e.g., Ennio Flaiano & Lucia Tongiorgi Tomase; L'Opera Completa di Paolo Uccello; Rizzoli, Milan, 1971 (and several translations).  The mosaic is item 5.A:  Rombo con elementi geometrici in the Catalogo delle Opere, with description and a small B&W picture on p. 85.  [Bokowski & Wills, below, give the date 1420.]

                    Coxeter [Elem. der Math. 44 (1989) 25‑36] says it "is evidently intended to be a picture of this star polyhedron."

                    However, J. V. Field tells me that the shape is not truly the small stellated dodecahedron, but just a 'spiky' dodecahedron.  She has examined the mosaic and the 'lines' of the pentagrams are not straight.  [The above cited photo is too small to confirm this.]  She says it appears to be a direct copy of a drawing in Daniele Barbaro; La Practica della Perspettiva; Venice, 1568, 1569, see below, and is most unlikely to be by Uccello.  See Field, Appendix 4, for a discussion of early stellations.

                    In 1998, I examined the mosaic and my photos of it and decided that the 'lines' are pretty straight, to the degree of error that a mason could work, and some are dead straight, so I agree with Coxeter that it is intended to be the small stellated dodecahedron.  I now have a postcard of this.  However, I have recently seen a poster of a different mosaic of the same shape which is distinctly irregular, so the different opinions may be based on seeing different mosaics! 

                    Both mosaics are viewed directly onto a pentagonal pyramid, but the pyramids are distinctly too short in the poster version.  The only spiky dodecahedron in Barbaro is on p. 111, fig. 52, and this is viewed looking at a common edge of two of the pyramids and the pyramids are distinctly too tall, so this is unlikely to be the source of the mosaics.  The 'elevated dodecahedron' in Pacioli & da Vinci, plate XXXI, f. CVI-v, has short pyramids and looks quite like the second mosaic, but it is viewed slightly at an angle so the image does not have rotational symmetry.  If anyone is in Venice, perhaps they could check whether there are two (or more?) mosaics and get pictures of them. 

Luca Pacioli & Leonardo da Vinci.  Untitled MS of 1498, beginning:  Tavola dela presente opera e utilissimo compendio detto dela divina proportione dele mathematici discipline e lecto  -- generally called  De divina proportione.  Ill. by Leonardo da Vinci.  See the entry in 6.AT.3 for fuller details of the facsimiles and details about which plates are in which of the editions.

                    Discussed by Mackinnon (see in 6.AT.3 below) and Field, pp. 214-215.  Clearly shows the stella octangula in one of the superb illustrations of Leonardo, described as a raised or elevated octahedron (plates XVIIII & XX).  Field, p. 214, gives the illustration.  None of the other raised shapes is a star, but the raised icosahedron is close to a star shape. 

Barbaro, Daniele (1514-1570).  La Practica della Perspettiva.  Camillo & Rutilio Borgominieri, Venice, 1568, 2nd ptg, 1569.  (Facsimile from a 1569 copy, Arnaldo Forni, Milan, 1980.  The facsimile's TP doesn't have the publication details, but they are given in the colophon.  Various catalogues say there are several versions with dates on the TP and colophon varying independently between 1568 and 1569.  A version has both dates being 1568, so this is presumed to be the first appearance.  Another version has an undated title in an elaborate border and this facsimile must be from that version.)  P. 111 has a dodecahedron with pyramids on each face, close to, but clearly not the stellated dodecahedron.  P. 112 has an icosahedron with pyramids on each face, again close to, but clearly not the stellated icosahedron.  I would have expected a reasonably accurate drawing, but in both drawings, several of the triples of segments which should lie on a single straight line clearly do not.  P. 113 shows an icosi-dodecahedron with pyramids on the triangular faces.  If the pyramids extended the edges of the pentagons, this would produce the dodecahedron!  But here the pyramids distinctly point much further out and the overall perspective seems wrong.  [Honeyman, no. 207, observing that some blocks come from the 1566 edition of Serlio which was dedicated to Barbaro.]

Wentzel Jamnitzer (or Jamitzer).  Perspectiva Corporum Regularum.  With 50 copper plates by Jost Amman.  (Nürnberg, 1568.)  Facsimile by Akademische Druck- u. Verlagsanstalt, Graz, 1973.  [Facsimiles or reprints have also been issued by Alain Brieux, Paris, 1964 and Verlag Biermann und Boukes, Frankfurt, 1972.]

                    This includes 164 drawings of polyhedra in various elaborations, ranging from the 5 regular solids through various stellations and truncations, various skeletal versions, pseudo-spherical shapes and even rings.  Some polyhedra are shown in different views on different pages.  Nameable objects, sometimes part of larger drawings, include:  tetrahedron, cubo-octahedron, truncated tetrahedron, stella octangula, octahedron, cube, truncated octahedron, rhombi-cuboctahedron, compound of a cube and an octahedron (not quite correct), great rhombi-cuboctahedron, icosahedron, great dodecahedron, dodecahedron, icosi-dodecahedron, rhombi-icosi-dodecahedron, truncated cube, and skeletal versions of:  stella octangula, octahedron, cube, icosahedron, dodecahedron, icosi-dodecahedron.  There are probably some uniform polyhedra, but I haven't tried to identify them, and some of the truncated and stellated objects might be nameable with some effort.

J. Kepler.  Letter to Herwart von Hohenberg.  6 Aug 1599.  In:  Johannes Kepler Gesammelte Werke, ed. by M. Caspar, Beck, Munich, 1938.  Vol 14, p. 21, letter 130, line 457.  ??NYS.  Cited by Field, op. cit. below.  Refers to (small??) stellated dodecahedron.

J. Kepler.  Letter to Maestlin (= Mästlin).  29 Aug 1599.  Ibid.  Vol. 14, p. 43, letter 132, lines 142-145.  ??NYS.  Cited by Field, below, and in [Kepler's Geometrical Cosmology; Athlone Press, London, 1988, p. 202].  Refers to (small??) stellated dodecahedron.

J. Kepler.  Harmonices Mundi.  Godfrey Tampach, Linz, Austria, 1619;  facsimile:  (Editions) Culture et Civilization, Brussels, 1968 (but my copy is missing three plates!)  [Editions probably should have É, but my only text which uses the word Editions is a leaflet in English.]  = Joannis Kepleri Astronomi Opera Omnia; ed. Ch. Frisch, Heyder & Zimmer, Frankfurt & Erlangen, 1864, vol. 5.  = Johannes Kepler Gesammelte Werke; ed. by M. Caspar, Beck, Munich, 1938, vol. 6, ??NYS. Book II.  Translated by J. V. Field; Kepler's star polyhedra; Vistas in Astronomy 23 (1979) 109‑141.

                    Prop. XXVI, p. 60 & figs. Ss & Tt on p. 53.  Describes both stellated dodecahedra,  {5/2, 5}  and  {5/2, 3}.  This is often cited as the source of the stella octangula, but the translation is referring to an 'eared cube' with six octagram faces and the stella octangula is clearly shown by Pacioli & da Vinci and by Jamnitzer.

Louis Poinsot.  Mémoire sur les polygones et les polyèdres.  J. de L'École Polytechnique 4 (1810) 16‑48 & plate opp. p. 48.  Art. 33‑40, pp. 39‑42, describe all the regular star polyhedra.  He doesn't mention Kepler here, but does a few pages later when discussing Archimedean polyhedra.

A. L. Cauchy.  Recherches sur les polyèdres.  J. de L'École Polytechnique 16 (1813) 68‑86.  ??NYS.  Shows there are no more regular star polyhedra and this also shows there are no more stellations of the dodecahedron.

H. S. M. Coxeter, P. Du Val, H. T. Flather & J. F. Petrie.  The Fifty-Nine Icosahedra.  Univ. of Toronto Press, 1938;  with new Preface by Du Val, Springer, 1982.  Shows that there are just 59 stellations of the icosahedron.  They cite earlier workers:  M. Brückner (1900) found 12;  A. H. Wheeler (1924) found 22.

Dorman Luke.  Stellations of the rhombic dodecahedron.  MG 41 (No. 337) (Oct 1957) 189‑194.  With a note by H. M. Cundy which says that the first stellation is well known (see 6.W.4) and that the second and third are in Brückner's Vielecke und Vielfläche, but that the new combinations shown here complete the stellations in the sense of Coxeter et al.

J. D. Ede.  Rhombic triacontahedra.  MG 42 (No. 340) (May 1958) 98‑100.  Discusses Coxeter et al. and says the main process generates 8 solids for the icosahedron.  He finds that the main process gives 13 for the rhombic triacontahedron, but makes no attempt to find the analogues of Coxeter et al.'s 59.

 

          6.AT.3.        ARCHIMEDEAN POLYHEDRA

 

Archimedes discovered the Archimedean solids, -3C, but his work is lost.  Heron quotes some of it and Pappus summarises it.  See HGM II 98-101.

Hero of Alexandria (c150).  Definitiones.  IN: Heronis Alexandrini  Opera quae supersunt omnia;  Vol. IV, Heronis Definitiones Cum Variis Collectionibus  Heronis Quae Feruntur Geometrica, ed. by J. L. Heiberg, Teubner, 1912, pp. 64-67.  Heath, HGM I 294-295 has a translation, but it doesn't give the complete text which seems open to two interpretations.  The German goes:  Archimedes aber sagt, es gebe in ganzen dreizehn Körper, die in einer Kugel eingeschreiben werden können, indem er ausser den genannten fünf noch acht hinzufügt; von diesen habe auch Platon das Tessareskaidekaeder gekannt, dies aber sei ein zweifaches, das eine aus acht Dreiecken und sechs Quadraten zusammengesetzt, aus Erde und Luft, welches auch einige von den Alten gekannt hätten, das andere umgekehrt aus acht Quadraten und sechs Dreiecken, welches schwieriger zu sein scheint.  My translation:  But Archimedes said, there are in total 13 bodies, which can be inscribed in a sphere, as he added eight beyond the named five [regular solids, which he had just defined]; but of these Plato knew the 14-hedron, however this is a double, one is composed of eight triangles and six squares, from Earth and Air, which some of the ancient also knew, the other conversely [is composed] of eight squares and six triangles, which seems to be more difficult.

                    Note that Hero has got the numbers wrong - Archimedes found 13 more than the 5 regular solids.  Secondly, the 'more difficult' solid does not exist!  Heath notes this and suggests that either the truncated cube or the truncated octahedron was intended.  The question of interpretation arises at the first semicolon -- is this continuing the statement of Archimedes or is Hero commenting on Archimedes' results?  Heath seems to say Archimedes is making the attribution to Plato, but see below.  MacKinnon, below, seems to be say this is being made by Hero.  Heath's discussion on HGM II 100 says "We have seen that, according to Heron, two of the semi-regular [i.e. Archimedean] solids had already been discovered by Plato" undoubtedly using the method of truncation.  However, I don't see that Heron is saying that Plato discovered the cubo-octahedron and the other solid, only that he knew it.  Mackinnon says "Plato is said by Heron to have discovered the cuboctahedron by making a model of it from a net."  But I don't see that Heron says this.

Pappus.  Collection.  c290.  Vol. 19.  In:  SIHGM II 194‑199.  Describes the 13 Archimedean solids.  "..., but also the solids, thirteen in number, which were discovered by Archimedes and are contained by equilateral and equiangular, but not similar, polygons."  He then describes each one.  Pappus' work has survived in a single MS (Vat. gr. 218) of the 10C in the Vatican and was not copied until 1550, but see Mackinnon, pp. 175-177, on whether it had been seen by Piero.  For the history of this MS, see also: Noel M. Swerdlow; The recovery of the exact sciences of antiquity: mathematics, astronomy, geography; IN: Anthony Grafton, ed.;  Rome Reborn  The Vatican Library and Renaissance Culture  Catalog of an exhibition at the Library of Congress, Washington, D.C., Jan. 6 ‑ Apr. 30, 1993; Library of Congress, Washington  &  Yale University Press, New Haven & London; in association with the Biblioteca Apostolica Vaticana; 1993; pp. 137-139.  [This exhibition is on-line at  www.ibiblio.org/expo/vatican.exhibit/vatican.exhibit.html.]

R. Ripley.  Believe It Or Not.  18th series, Pocket Books, NY, 1971.  P. 116 asserts the Romans used dice in the shape of cubo‑octahedra.

The British Museum, Room 72, Case 9, has two Roman cubo-octahedral dice on display.

F. Lindemann, op. cit. in 6.AT.1, 1896, describes and illustrates an antique rhombic triacontahedron, possibly a die, possibly from the middle of the Byzantine era.

Nick Mackinnon.  The portrait of Fra Luca Pacioli.  MG 77 (No. 479) (Jul 1993) plates 1-4 & pp. 129-219.  Discusses the various early authors, but has mistakes.

della Francesca.  Trattato.  c1480.  Ff. 105r - 117v (224-250) treats solid bodies, discussing all the regular polyhedra, with figures, though Arrighi gives only a projection of the octahedron.  Discusses and gives good diagrams of the truncated tetrahedron and cubo‑octahedron, apparently the first drawings of any Archimedean polyhedra.  Jayawardene refers to the cubo-octahedron as a truncated cube.

                    Davis notes that Pacioli's Summa, Part II, ff. 68v - 73v, prob. 1-56, are essentially identical to della Francesca's Trattato, ff. 105r - 120r.

Piero della Francesca.  Libellus de Quinque Corporibus Regularibus.  c1487 [Davis, p. 44, dates it to 1482-1492].  Piero would have written this in Italian and it is believed to have been translated into Latin by Matteo da Borgo [Davis, p. 54], who improved the style.  First post-classical discussion of the Archimedean polyhedra, but it was not published until an Italian translation (probably by Pacioli) was printed in Pacioli & da Vinci, qv, in 1509, as: Libellus in tres partiales tractatus divisus quae corpori regularium e depēdentiū actine perscrutatiōis ..., ff. 1‑27.  A Latin version was discovered by J. Dennistoun, c1850, and rediscovered by Max Jordan, 1880, in the Urbino manuscripts in the Vatican -- MS Vat.Urb.lat. 632; the Duke of Urbino was a patron of Piero and in the MS, Piero asks that it be placed by his De Prospectiva Pingendi in the Duke's library.  This was published by Girolamo Mancini in: L'opera "De corporibus regularibus" di Pietro Franceschi detto della Francesca usurpata da Fra Luca Pacioli, Memorie della Reale Accademia dei Lincei, Classe di Scienze Morali, Storiche e Filologiche (5) 14:7B (1915)  441-580 & 8 plates, also separately published by Tipografia della Reale Accademia dei Lincei, Rome, 1916.  Davis identifies  139  problems in this, of which  85  (= 61%)  are taken from the Trattato.  There is debate as to how much of this work is due to Piero and how much to Pacioli.  The Latin text differs a bit from the Italian.  See the works of Taylor and Davis in Section 1 under Pacioli and the discussion on della Francesca's Trattato and Pacioli's Summa in the common references. 

                    He describes a sphere divided into  6  zones and  12  sectors.  Mackinnon says Piero describes seven of the Archimedean polyhedra, but without pictures, namely:  cuboctahedron, truncated tetrahedron, truncated cube, truncated octahedron, truncated dodecahedron, truncated icosahedron, rhombi-cuboctahedron.  Field, op. cit. in 6.AT.1, p. 107, says Piero gives six of the Archimedean polyhedra.  In recent lectures Field has given a table showing which Archimedean polyhedra appear in Piero, Pacioli, Dürer and Barbaro and this lists just the first six of the above as being in Piero.  I find just the five truncated regular polyhedra -- see above for the cubo-octahedron -- and there is an excellent picture of the truncated tetrahedron on f. 22v of the printed version.  Mancini gives different diagrams than in the 1509 printed version, including clear pictures of the truncated icosahedron and the truncated dodecahedron.  della Francesca clearly has the general idea of truncation.  An internet biographical piece, apparently by, or taken from, J. V. Field, (http://www‑history.mcs.st‑andrews.ac.uk/history/Mathematicians.Francesca.html), shows that the counting is confused by the presence of the cubo-octahedron in the Trattato but not in the Libellus.  So della Francesca rediscovered six Archimedean polyhedra, but only five appear in the Libellus.  The work of Pappus was not known at this time. 

Pacioli.  Summa.  1494. 

f. 4.  Brief descriptions of the cubo-octahedron, truncated tetrahedron, icosidodecahedron, truncated icosahedron.  No drawings.

Part II, ff. 68v - 72r, sections 2 (unlabelled) - 35.  Discussion and some crude drawings of the regular polyhedra, the truncated tetrahedron and the cubo-octahedron.  Mackinnon says these are the first printed illustrations of any Archimedean polyhedra.  Davis notes that Part II, ff. 68v - 73v, prob. 1-56, are essentially identical to della Francesca's Trattato, ff. 105r - 120r.

Jacopo de'Barbari or Leonardo da Vinci.  Portrait of Fra Luca Pacioli.  1495.  In the Museo Nazionale di Capodimonte, Naples.  The upper left shows a glass rhombi-cuboctahedron half filled with water.  Discussed by Mackinnon, with colour reproduction on the cover.  Colour reproduction in Pacioli, Summa, 1994 reprint supplement.

Luca Pacioli & Leonardo da Vinci.  Untitled MS of 1498, beginning:  Tavola dela presente opera e utilissimo compendio detto dela divina proportione dele mathematici discipline e lecto  -- generally called  De divina proportione.  Three copies of this MS were made.  One is in the Civic Library of Geneva, one is the Biblioteca Ambrosiana in Milan and the third is lost.  Three modern versions of this exist.

                    Transcription published as Fontes Ambrosiani XXXI, Bibliothecae Ambrosianae, Milan, 1956.  This was sponsored by Mediobanca as a private edition.  There is a copy at University College London.

                    Colour facsimile of the Milan copy, Silvana Editoriale, Milano, (1982), 2nd ptg, 1986.  With a separate booklet giving bibliographical details and an Introduzione di Augusto Marinoni, 20pp + covers.  The booklet indicates this is Fontes Ambrosiani LXXII.

                    Printed version: [De] Divina proportione  Opera a tutti glingegni perspicaci e curiosi necessaria  Ove ciascun studioso di Philosophia: Prospectiva Pictura Sculptura: Architectura: Musica: e altre Mathematice: suavissima: sottile: e admirabile doctrina consequira: e delectarassi: cōvarie questione de secretissima scientia.  Ill. by Leonardo da Vinci.  Paganino de Paganini, Brescia, 1509.  Facsimile in series Fontes Ambrosiana, no. XXXI, Milan, 1956;  also by Editrice Dominioni, Maslianico (Como), 1967.  (On f. 23r, the date of completion of the original part is printed as 1497, but both MSS have 1498.) 

                              The printed version was assembled from three codices dating from 1497‑1498 and contains the above MS with several additional items.  However, the diagrams in the text are simplified and the plates are in a different order.  The MS has 60 coloured plates, double sided; the printed version has 59 B&W plates, single sided.  There are errors of pagination and plate numbering in both versions.  On f. 3 of the printed version is a list of plates and one sees that plate LXI should be numbered LVIIII and that plates LX, LXI were omitted and were to have been a hexagonal pyramid in solid and framework views (the framework view is in the MS, but the solid pyramid is not).

                              NOTE.  Simon Finch's Catalogue 48, item 4, describes the copy that was in the Honeyman Collection and says it has 59 printed plates of geometric figures and is unique in having two contemporary additional MS plates showing the hexagonal pyramid (numbered LX and LXI), which are given in the list of plates, but which do not appear in any other known copy.  It seems that these figures were overlooked in printing and that the owner of the Honeyman copy decided to make his own versions, or, more likely, got someone to make versions in the original style.  There is a framework hexagonal pyramid in the MS, and this makes it seem likely that these figures had been prepared and were omitted in printing -- indeed the Honeyman leaves could be the overlooked drawings.  That leaves the question of whether there was a solid hexagonal pyramid in the MS?

                    Most pictures come in pairs -- a solid figure and then a framework figure.  There are the five regular polyhedra, the following six Archimedean polyhedra: truncated tetrahedron, cubo-octahedron, truncated octahedron, truncated icosahedron, icosi-dodecahedron, rhombi-cubo-octahedron and also the stella octangula.  There are raised or elevated versions of the tetrahedron, cube, cubo-octahedron, icosahedron, dodecahedron, icosi-dodecahedron, rhombi-cubo-octahedron.  Also triangular, square, pentagonal and hexagonal prisms and tall triangular, square and pentagonal pyramids.  Also a triangular pyramid not quite regular and a sphere divided into 12 sectors and 6 zones.  There are also a solid sphere, a solid cylinder, a solid cone and a framework hexagonal pyramid (the last is not in the printed version).  Mackinnon says they give the same seven Archimedean polyhedra as Piero, but Piero gives five or six and Pacioli & da Vinci gives six, with only four common polyhedra.  Pacioli & da Vinci assert that the rhombi-cuboctahedron arises by truncating a cuboctahedron, but this is not exactly correct.

                    Part of the printed version is Libellus in tres partiales tractatus divisus quae corpori regularium e depēdentiū actine perscrutatiōis ..., which is an Italian translation (probably by Luca Pacioli) of Piero della Francesca's  Libellus de quinque corporibus regularibus.  There is debate as to whether this was actually written by Pacioli or whether Pacioli plagiarized it and whether it actually appeared in the 1509 printing or was added to a later reprinting, etc.

                    Davis [p. 65] says the drawings were made from models prepared by Da Vinci.  Davis [p. 74] cites Summa, Part II, f. 68v, and she quotes part of it on pp. 100-101.  This is also referred to by MacKinnon [p. 170] and Taylor [p. 344], neither giving details and no two of the three agreeing on what the passage means.  I have not been able to make complete sense of the passage, but it seems clearly to say that in Apr 1489, Pacioli presented models of at least the regular solids to the Duke of Urbino at the palace of Pacioli's protector [Cardinal Giuliano della Rovere, later Pope Julius II] in Rome.  He then says many other dependants [= variations] of the regular solids can be made, and models were made for Pietro Valletari, Bishop of Carpentras.  There is no reference to the number of models, nor their material, nor to a set being given to the Cardinal, nor whether the Cardinal was present when the models were given to the Duke.  Due to a missing right parenthesis,  ), the sense of one statement involving 'his own hands' could mean either that Pacioli presented the models to the Duke's own hands or that Pacioli had decorated the models himself.  I doubt whether there were many other solids at this time, otherwise he would have mentioned them in the Summa -- the Summa only describes four of the Archimedean solids and only two of them are in Part II.  I suggest that he didn't start developing the other shapes until about 1494, or later, in 1496 when he went to Milan and met Leonardo.

                    However, on f. 28v of De Divina Proportione, Pacioli is clearer and says that he arranged, coloured and decorated with his own hands 60 models in Milan and two other sets for Galeazzo Sanseverino in Milan and for Piero Soderini in Florence.  This refers to his time in Milan, which was 1496-1499, though the Soderini set might have been made after Pacioli and da Vinci moved to Florence.

Albrecht Dürer.  Underweysung der messung ....  1525 & 1538.  Op. cit. in 6.AA.  Figures 29‑43 (erroneously printed 34) (pp. 316-347 in The Painter's Manual, Dürer's 1525 ff. M-iii-v - N-v-r) show a net of each of the regular polyhedra, an approximate sphere (16 sectors and 8 zones), truncated tetrahedron, truncated cube, cubo-octahedron, truncated octahedron, rhombi-cubo-octahedron, snub cube, great rhombi-cubo-octahedron, polyhedron of six dodecagons and thirty-two triangles (having a pattern of four triangles replacing each triangle of the cubo-octahedron, so a sort of truncated cubo-octahedron -- not an Archimedean solid and not correctly drawn) and an elongated hexagonal bipyramid (not even regular faced).  This gives  7  of the  13 non-regular Archimedean polyhedra.  Mackinnon says Figures 29‑41 show a net of each of the regular polyhedra and the same seven Archimedean ones as given by Pacioli & da Vinci, but they give six and there are only four common ones.  In the revised version of 1538, figure 43 is replaced by the truncated icosahedron and icosi-dodecahedron (figures 43 & 43a, pp. 414‑419 in The Painter's Manual), giving  9  of the  13  non-regular Archimedean solids.  P. 457 shows the remaining four Archimedean cases from an 1892 edition. 

Albrecht Dürer.  Elementorum Geometricorum (?).  1534. Op. cit. in 6.AA.  Liber quartus, fig. 29-43, pp. 145-158 shows the same material as in the 1525 edition. 

See Barbaro, 1568, in 6.AT.2, pp. 45-104 for drawings and nets of 11 of the 13 Archimedean solids - he omits the two snub solids.

See Jamnitzer, 1568, in 6.AT.2 for drawings of eight of the 13 Archimedean solids.

J. Kepler.  Letter to Maestlin (= Mästlin).  22 Nov 1599.  In:  Johannes Kepler Gesammelte Werke, ed. by M. Caspar, Beck, Munich, 1938.  Vol 14, p. 87, letter 142, lines 21-22.  ??NYS.  Described by Field, p. 202.  Describes both rhombic solids.

J. Kepler.  Strena seu De Nive Sexangula [A New Year's Gift or The Six‑Cornered Snowflake].  Godfrey Tampach, Frankfurt am Main, 1611.  (Reprinted in: Johannes Kepler Gesammelte Werke; ed. by M. Caspar, Beck, Munich, 1938, vol. 4, ??NYS.)  Reprinted, with translation by C. Hardie and discussion by B. J. Mason & L. L. Whyte, OUP, 1966.  I will cite the pages from Kepler (and then the OUP pages).  P. 7 (10‑11).  Mentions 'the fourteen Archimedean solids' [sic!].  Describes the rhombic dodecahedron and mentions the rhombic triacontahedron.  The translator erroneously adds that the angles of the rhombi of the dodecahedron are 6Oo and 120o.  Kepler adds that the rhombic dodecahedron fills space.  Kepler's discussion is thorough and gives no references, so he seems to feel it was his own discovery.

J. Kepler.  Harmonices Mundi, 1619.  Book II, opp. cit. above.  Prop. XXVII, p. 61.  Proves that there are just two rhombic 'semi‑regular' solids, the rhombic dodecahedron and the rhombic triacontahedron, though the cube and the 'baby blocks' tessellation can also be considered as limiting cases.  He illustrates both polyhedra.  Def. XIII, p. 50 & plate (missing in facsimile).  Mentions prisms and antiprisms.  Prop. XXVIII, pp. 61‑65.  Finds the 13 Archimedean solids and illustrates them -- the first complete set -- but he does not formally show existence.

J. Kepler.  Epitome Astronomiae Copernicanae.  Linz, 1618-1621.  Book IV, 1620.  P. 464.  = Johannes Kepler Gesammelte Werke; ed. by M. Caspar, Beck, Munich, 1938, vol. 7, p. 272.  Shows both rhombic solids.

The following table shows which Archimedean polyhedra appear in the various early books from della Francesca to Kepler.

 

dF      = della Francesca, c1487, folio

          T indicates the object appears in the Trattato of c1480.

P        = Pacioli, Summa, Part II, 1494, folio.

P&dV = Pacioli & da Vinci, 1498, plate number of the solid version; the framework version is

          the next plate

D       = Dürer, 1525, plate

          + indicates the object is added in the 1538 edition.

B        = Barbaro, 1568, page

J         = Jamnitzer, 1568, plate

K       = Kepler, 1619, figure

                        DF       P                          P&dV           D        B          J           K

 

Truncated tetrahedron                       T,22v          4v,II-69v          3        35        56              A3        2

Truncated cube                                     22r                                            36        61              G2        1

Truncated dodecahedron                       21r                                                        76                          3

Truncated octahedron                           21v                                17        38        68         B2,B4        5

Great rhombi-cubo-octahedron                                                                41        88              B6        6

Great rhombi-icosidodecahedron                                                                        100                          7

Truncated icosahedron (football!)          20v                    4v        23          +        81                          4

Cubo-octahedron                                     T          4v,II-69v          9        37        58         B6,F1        8

Icosidodecahedron                                                         4v        29                    71         D4,F6        9

Rhombi-cubo-octahedron                                                            35        39        64              B6      10

Rhombi-icosidodecahedron                                                                                 94              D4      11

Snub cube                                                                                             40                                    12

Snub dodecahedron                                                                                                                       13

 

          NUMBER                                 5(6)                      4          6      7(9)        11                8      13

 

Richard Buckminster Fuller.  Centre spread card version of his Dymaxion World map on the cubo-octahedron.  Life (15 Mar 1943).  Reproduced in colour, with extended discussion, in:  Joachim Krausse & Claude Lichtenstein, eds;  Your Private Sky  R. Buckminster Fuller  The Art of Design Science [book accompanying a travelling exhibition in 2000];  Lars Müller Publishers, Bade, Switzerland, 1999, pp. 250-275.  (This quotes a Life article on 1 Mar 1943 and a Fuller article, Fluid Geography, of 1944 -- ??NYS.  It also reproduces a 1952 colour example of the icosahedral version.)

Richard Buckminster Fuller.  US Patent 2,393,676 -- Cartography.  Filed: 25 Feb 1944;  granted: 29 Jan 1946.  3pp + 5pp diagrams.  His world map on the cubo-octahedron.  It was later put on the icosahedron.  One page is reproduced in:  William Blackwell; Geometry in Architecture; Key Curriculum Press, Berkeley, 1984, p. 157.

J. H. Conway.  Four-dimensional Archimedean polytopes.  Proc. Colloq. Convexity, Copenhagen, 1965 (1967) 38-39.  ??NYS -- cited by Guy, CMJ 13:5 (1982) 290-299.

 

          6.AT.4.        UNIFORM POLYHEDRA

 

H. S. M. Coxeter, M. S. Longuet‑Higgins & J. C. P. Miller.  Uniform polyhedra.  Philos. Trans. Roy. Soc. 246A (1954) 401‑450.  They sketch earlier work and present 53 uniform polyhedra, beyond the 5 Platonic, 13 Archimedean and 4 Kepler‑Poinsot polyhedra and the prisms and anti‑prisms.  Three of these uniform polyhedra are actually infinite families.  "... it is the authors' belief that the enumeration is complete, although a rigorous proof has still to be given."

S. P. Sopov.  Proof of completeness of the list of uniform polyhedra.  Ukrain. Geometr. Sb. 8 (1970) 139-156.  ??NYS -- cited in Skilling, 1976.

J. S. Skilling.  The complete set of uniform polyhedra.  Philos. Trans. Roy. Soc. London Ser. A 278 (1975) 111‑135.  Demonstrates that the 1954 list of Coxeter, et al., is complete.  If one permits more than two faces to meet at an edge, there is one further polyhedron -- the great disnub dirhombidodecahedron.

J. S. Skilling.  Uniform compounds of uniform polyhedra.  Math. Proc. Camb. Philos. Soc. 79 (1976) 447-468.  ??NYS -- I am told it determines that there are 75 uniform compounds and also cites Sopov.

 

          6.AT.5.        REGULAR‑FACED POLYHEDRA

 

O. Rausenberger.  Konvexe pseudoreguläre Polyeder.  Zeitschr. für math. und naturwiss. Unterricht 46 (1915) 135‑142.  Finds the eight convex deltahedra.

H. Freudenthal & B. L. van der Waerden.  Over een bewering van Euclides [On an assertion of Euclid] [in Dutch].  Simon Stevin 25 (1946/47) 115‑121.  ??NYS.  Finds the eight convex deltahedra -- ignorant of Rausenberger's work.

H. Martyn Cundy.  "Deltahedra".  MG 36 (No. 318) (Dec 1952) 263‑266.  Suggests the name "deltahedra".  Exposits the work of Freudenthal and van der Waerden, but is ignorant of Rausenberger.  Considers non‑convex cases with two types of vertex and finds only 17 of them.  Considers the duals of Brückner's trigonal polyhedra.

Norman W. Johnson.  Convex polyhedra with regular faces.  Canad. J. Math. 18 (1966) 169‑200.  (Possibly identical with an identically titled set of lecture notes at Carleton College, 1961, ??NYS.)  Lists 92 such polyhedra beyond the 5 regular and 13 Archimedean polyhedra and the prisms and antiprisms.

Viktor A. Zalgaller.  Convex polyhedra with regular faces [in Russian].  Seminars in Mathematics, V. A. Steklov Mathematical Institute, Leningrad, vol. 2 (1967).  ??NYS.  English translation:  Consultants Bureau, NY, 1969, 95pp.  Gives details of computer calculations which show that Johnson's list is complete.  Defines a notion of simplicity and shows that the simple regular‑faced polyhedra are the prisms, the antiprisms (excepting the octahedron) and 28 others.  Names all the polyhedra and gives drawings of the simple ones.

 

          6.AT.6.        TESSELLATIONS

 

Albrecht Dürer.  Underweysung der messung ....  1525 & 1538.  Op. cit. in 6.AA.  Figures 22‑27 (pp. 156‑169 in The Painter's Manual, Dürer's 1525 ff. E‑vi‑v ‑ F-iii-v) show:  the three regular tessellations;  the quasi-regular one,  3636,  and some of its dual;  several irregular ones, including some partial tessellations with pentagons;  and the truncated square lattice,  482.  In the revised version of 1538, he adds some tilings by rhombuses (figures 23a & 24, pp. 410‑411 in The Painter's Manual).

Albrecht Dürer.  Elementorum Geometricorum (?).  Op. cit. in 6.AA, 1532.  Book II, fig. 22‑27, pp. 62-67, is the material from the 1525 version.

J. Kepler.  Letter to Herwart von Hohenberg.  6 Aug 1599.  Op. cit. in 6.AT.2.  Field, p. 105, says Kepler discusses tessellations here and this is the earliest of his writings to do so.

J. Kepler.  Harmonices Mundi.  1619.  Book II.  Opp. cit. above.  Prop. XVIII, p. 51 & plate (missing in my facsimile).  Shows there are only three regular plane tessellations and mentions the dual of 3636, which is the 'baby blocks' tessellation.  Prop. XIX‑XX, pp. 51‑56 & four plates (three missing in my facsimile).  Finds the 8 further Archimedean tessellations and 7 of the 10 further ways to fill 360 degrees with corners of regular polygons.  He misses  3,7,42;  3,8,24;  3,9,18  despite computing, e.g., that a triangle and a heptagon would leave a gap of  40/21  of a right angle.  Field, p. 109, notes that Kepler doesn't clearly have all vertices the same in some pictures -- e.g. he has both  3366  and  3636  patterns in his figure R.

Koloman Moser.  Ver Sacrum.  1902.  This Viennese art nouveau drawing is considered to be the first tessellation using life-like figures.  It has trout and the pattern has symmetry (or wallpaper) group  pg  and isohedral type  IH2.

Branko Grünbaum & G. C. Shephard.  Tilings and Patterns.  Freeman, 1986.  I haven't examined this thoroughly yet, but it clearly is the definitive work and describes everything known to date.

 

          6.AT.6.a.     TESSELLATING WITH CONGRUENT FIGURES

 

          This is a popular topic which I have just added.  Gardner's article and addendum in Time Travel gives most recent results, so I will just give just some highlights.  The facts that any triangle and any quadrilateral will tile the plane must be very old, perhaps Greek, but I have no early references.  Generally, I will consider convex polygons and most items only deal with the plane.

 

David Hilbert.  Mathematische Probleme.  Göttinger Nachrichten (= Nachrichten der K. Gesellschaft der Wissenschaften zu Göttingen, Math.‑phys. Klasse) 3 (1900) 253‑297.  This has been reprinted and translated many times, e.g. in the following.

                    R. Bellman, ed.  A Collection of Modern Mathematical Classics -- Analysis.  Dover, 1961.  Pp. 248‑292 [in German].

                    Translated by M. W. Newson.  Bull. Amer. Math. Soc. 8 (1902) 437‑479.  Reprinted in:  F. E. Browder, ed.  Mathematical Developments Arising from Hilbert Problems.  Proc. Symp. Pure Math. 28 (1976) 1‑34.

                    Problem 18: Aufbau des Raumes aus kongruenten Polyedern [Building up of space from congruent polyhedra].  "The question arises: Whether polyhedra also exist which do not appear as fundamental regions of groups of motions, by means of which nevertheless by a suitable juxtaposition of congruent copies a complete filling up of space is possible."  Hilbert also asks two other questions in this problem.

                    The problem is discussed by John Milnor in his contribution to the Symposium, but he only shows non‑convex  8‑  &  10‑gons which fill the plane.

K. Reinhardt.  Über die Zerlegung der Ebene in Polygone.  Dissertation der Naturwiss. Fakultät, Univ. Frankfurt/Main, Borna, 1918.  ??NYS -- cited by Kershner.  Finds the three types of hexagons and the first five types of pentagons which fill the plane.

Max Black.  Reported in: J. F. O'Donovan; Clear thinking; Eureka 1 (Jan 1939) 15 & 20.  Problem 2: which quadrilaterals can tile the plane?  Answer: all!

R. B. Kershner.  On paving the plane.  AMM 75:8 (Oct 1968) 839‑844.  Says the problem was posed by Hilbert.  Gives exhaustive lists of hexagons and a list of pentagons which he claimed to be exhaustive.  Cites previous works which had claimed to be exhaustive, but he has found three new types of pentagon.

J. A. Dunn.  Tessellations with pentagons.  MG 55 (No. 394) (Dec 1971) 366‑369.  Finds several types and asks if there are more.

M. M. Risueño, P. Nsanda Eba & Editorial comment by Douglas A. Quadling.  Letters:  Tessellations with pentagons.  MG 56 (No. 398) (Dec 1972) 332‑335.  Risueño's letter replies to Dunn by citing Kershner.  Eba constructs a re‑entrant pentagon.  [This is not cited by Gardner.]

Gardner.  On tessellating the plane with convex polygon tiles.  SA (Jul 1975).  Much extended in Time Travel, chap. 13.

Ivan Niven.  Convex polygons that cannot tile the plane.  AMM 85 (1978) 785-792.  n‑gons, with  n > 6,  cannot tile the plane.

Doris Schattschneider.  In praise of amateurs.  In:  The Mathematical Gardner; ed. by David A. Klarner; Wadsworth, Belmont, California, 1981, pp. 140‑166 & colour plates I‑V between 166 & 167.  Surveys history after Kershner, describing contributions of James & Rice.

Gardner.  On tessellating the plane with convex polygon tiles.  [Originally:  SA (Jul 1975).]  Much extended in Time Travel, 1988, chap. 13.  The original article generated a number of responses giving new pentagonal tilings, making 14 types in all.  Good survey of the recent literature.

 

          6.AT.7.        PLAITING OF POLYHEDRA

 

          New section.

 

John Gorham.  A System for the Construction of Crystal Models on the Type of an Ordinary Plait: Exemplified by the Forms Belonging to the Six Axial Systems in Crystallography.  E. & F. N. Spon, London, 1888.  Gorham's Preface says he developed the idea and demonstrated it to the Royal Society some 40 years earlier.

A. R. Pargeter.  Plaited polyhedra.  MG 43 (No. 344) (May 1959) 88‑101.  Cites and quotes Gorham.  Extends to plaiting dodecahedron, icosahedron and some archimedean, dual and stellated examples.

J. Brunton.  The plaited dodecahedron.  MG 44 (No. 347) (Feb 1960) 12‑14.  With comment by Pargeter.  Obtains a 3‑plait which almost completes the dodecahedron.

 

          6.AT.8.        DÜRER'S OCTAHEDRON

 

          New section -- I know of other articles claiming to 'solve' the problem.

 

Albrecht Dürer.  Melencolia I.  1514.  Two impressions are in the British Museum.  In the back left is an octahedron whose exact shape is the subject of this section.  It looks like a cube truncated at two opposite corners, but the angles do not quite look like 90o.

Albrecht Dürer.  Dresden Sketchbook.  Facsimile as The Human Figure, the complete Dresden Sketchbook; Dover, NY, 1972.  ??NYS -- cited by Sharp.  This has a sketch of the solid with hidden lines indicated, so the combinatorial shape is definitely known and is a hexahedron of six equal faces, truncated at two opposite corners. 

E. Schröder.  Dürer Kunst und Geometrie.  Birkhäuser, Basel, 1980.  ??NYS -- cited by Sharp and MacGillavry.

Caroline H. MacGillavry.  The polyhedron in A. Dürer's Melencolia I  An over 450 years old puzzle solved?  Koninklijke Nederlandse Akademie van Wetenschappen Proc. B 84:3 (28 Sep 1981) 287-294.  The rhombohedral angle, i.e. the angle between edges at the truncated top and bottom vertices of the rhombohedron, was estimated as 72o by Grodzinski.  She determines it is 79o ± 1o.  She then built and photographed such a polyhedron and then computed its projection, both of which seem identical to Dürer's picture.  Crystallographers believe Dürer was drawing an actual crystal, with a form of calcite having rhombohedral angle of 76o being the closest known shape, though it is not known to have been studied in Dürer's time, so others have suggested fluorite, though fluorite has two standard forms, neither of this form, but Dürer's 'hybrid' artistic version could have been derived from them.

Terence Lynch.  The geometric body in Dürer's engraving Melencolia I.  J. Warburg and Courtauld Institutes 45 (1982) 226-232 & plate a on p. 37?.  Lots of references to earlier work.  Notes that perspective was not sufficiently advanced for Dürer to construct a general drawing of such an object.  After many trials, he observes that a parallel projection of the solid fits onto a  4 x 4  grid -- like the magic square in the picture -- and that symmetry then permits the construction with straight edge and compass (which are both shown in the picture).  This shows that the original faces are rhombuses whose diameters are in the ratio  2 : Ö3.  And the dihedral angle between the triangular faces and the cut off rhombuses is 30o  Further, the actual drawing can then be made by one of the simplified perspective techniques known to Dürer.  However, Dürer has taken a little bit off the top and bottom of the figure and this distortion has misled many previous workers. 

John Sharp.  Dürer's melancholy octahedron.  MiS (Sep 1994) 18-20.  Asserts that the shape was first determined by Schröder in 1980 and verified by Lynch. 

 

          6.AT.9.        OTHER POLYHEDRA

 

          New section.

 

Stuart Robertson.  The twenty-two cuboids.  Mathematics Review 1:5 (May 1991) 18-21.  This considers polyhedra with six quadrilateral faces and determines what symmetries are possible -- there are 22 different symmetry groups.

 

          6.AU. THREE RABBITS, DEAD DOGS AND TRICK MULES

 

          See S&B, p. 34. 

          Loyd's Trick Mules has two mules and two riders which can only be placed correctly by combining each front with the other rear.

          Earlier forms showed two dead dogs which were brought to life by adding four lines. The resulting picture is a pattern, generally called 'Two heads, four children' and can be traced back to medieval Persian, Oriental and European forms.

          The three rabbits problem is: "Draw three rabbits, so that each shall appear to have two ears, while, in fact, they have only three ears between them."  Until about 1996, I only knew this from the 1857 Magician's Own Book and the many books which copied from it.  Someone at a conference at Oxford in 1996 mentioned that the pattern occurs in a stained glass window at Long Melford, Suffolk.  Correspondence revealed that the glass is possibly 15C and the pattern was apparently brought from Devon about that time.  More specifically, it comes from the east side of Dartmoor and inquiries there have turned up numerous examples as roof bosses from 13‑16C.  Totally serendipitously, I was reading a guide book to Germany in 1997 and discovered the pattern occurs in stonework, possibly 16C, at Paderborn, Germany.  A letter led to receiving a copy of Schneider's article (see below) which described the pattern occurring at Dunhuang, c600.  I am indebted to Miss Y. Yasumara, the Art Librarian at the School of Oriental and African Studies, for directing me to several works on Dunhuang.  However, I have not examined all these works in detail (the largest is five large volumes), so I may not have found all the examples of this pattern.  Miss Yasumara also directed me to Roderick Whitfield, of the School of Oriental and African Studies, who tells me there is no other example of this pattern in Chinese art, and to Susan Whitfield (no relation), head of the International Dunhuang Project at the British Library.  However, Greeves (see his articles, below) has found other examples of the pattern in Europe, Iran and Tibet and found that modern carpets with the pattern are being made in China.  A student of his recently went to Dunhuang and the locals told her that the pattern came from 'the West', meaning India, which opens up a whole new culture to examine.

          In 1997, I visited the Dartmoor area, seeing several examples and finding a reference (Hambling, below) to Tom Greeves' article.  In 2000, I again visited the area, seeing more churches and meeting Tom Greeves and his associate Sue Andrew.  Later in 2000, I visited Paderborn.  Later in 2000, I showed this material to Wei Zhang and Peter Rasmussen, leading collectors of Chinese puzzles and they have begun to investigate the Chinese material much more thoroughly than I have done -- see below.  In 2001, I went again to the area.

          I have read that rabbits were introduced to England in 1176 by the Normans and became common in the 13C, though I believe they weren't really wild for some time after that.  E.g. [J. A. R. Pimlott; Recreations; Studio Vista, 1968, p. 18] says: "The rabbit was not known in Britain until the thirteenth century and did not become plentiful until the fifteenth", citing Elspeth M. Veale; The English Fur Trade in the Later Middle Ages; 1966.  I have read that hares were introduced between -500 and +500, but I have just seen a mention that bones of a hare found in Ireland have been carbon dated to ‑26,000.  So it is probable that the animals in many of the images are hares, though some of the images are distinctly more rabbit-like than hare-like. 

          The material in this section has grown so much that it is now divided into seven subsections:  China;  Other Asia;  Paderborn;  Medieval Europe;  Modern Versions of the Three Rabbits Puzzle;  Dead Dogs;  Trick Mules. 

          I have about a dozen letters and emails which have not yet been processed.

          Rabbits going clockwise: Dunhuang (14 caves -- all except 407 & 420); Goepper; St. Petersburg; Iranian tray; Paderborn; Münster; Bestiary; Lyon; Throwleigh; Valentine; Clyst Honiton; Hasloch am Main; Michelstadt; Collins; Greeves letterhead; Urumqi;

          Rabbits going anti-clockwise: Dunhuang (2 caves -- 407 & 420); Corbigny; Lombard's Gloss; North Bovey; Long Crendon; Chester; Widecombe; Long Melford; Chagford; Best Cellars, Chagford; Tavistock; Broadclyst; Sampford Courtenay; Spreyton; Paignton; South Tawton; Valentine; Schwäbisch Hall; Baltrušaitis Fig. 97; Child; Magician's Own Book et al; Warren Inn; Best Cellars, Chagford; Newman; Lydford; Trinity Construction Services;

          The choice of clockwise versus anticlockwise seems to be random!  Except the Chinese clearly preferred clockwise.

          FOUR RABBITS versions.  Baltrušaitis;  Wilson; Goepper.  Bestiary; Andrew/Lombard; Lyon; Hamann-MacLean;

          MORE FIGURES.  Chichester Cathedral.  Boxgrove Priory.

 

                    CHINA

 

Jurgis Baltrušaitis.  Le Moyen Age Fantastique  Antiquités et exotismes dans l'art gothique.  (A. Colin, Paris, 1955, 299 pp.)  Revised, Flammarion, Paris, 1981, 281 pp.  Thanks to Peter Rasmussen for telling me about this.  Supposedly an English edition was published in 1998, but I have found an entry in the Warburg catalogue for The Fantastic in the Middle Ages, published by Boydell & Brewer, Woodbridge, Suffolk, 2000, marked 'order cancelled' -- so the publication seems to have never happened.  This is a major source used for When Silk Was Gold, below.  Pp. 132‑139 of the 1981 edition have many examples of three and four rabbits, four boys, etc.  He gives a small illustration (Fig. 96B) from Dunhuang (Touen-houang) (6C-10C), the oldest example he knows, and many others.  See other sections for more details.

[Huang Zu'an ?? -- Schneider, below, gives this author, but there is no mention of an author in the entire issue.]  Dunhuang -- Pearl of the Silk Road.  China Pictorial (1980:3) 10-23 with colour photo on p. 22.  9th article in a series on the Silk Road.  Colour photo of the three rabbits pattern with caption: "A ceiling design.  The three rabbits with three ears and the apsarases seem to be whirling.  Cave 407.  Sui Dynasty."  The Sui dynasty was from 581 to 618, so we can date this as c600.  The image is rather small, but the three hares can be made out.  There is no discussion of the pattern in the article. 

Chang Shuhong  &  Li Chegxian.  The Flying Devis of Dunhuang.  China Travel and Tourism Press, Beijing, 1980.  Unpaginated.  In the Preface, we find the following.  "What is particularly novel is the full-grown lotus flower painted in the centre of the canopy design on the ceiling of Cave 407.  In the middle of the flower there are three rabbits running one after the other in a circle.  For the three rabbits only three ears are painted, each of them borrowing one ear from another.  This is an ingenious conception of the master painter."  From this, it seems that this pattern is uncommon.  The best picture of the pattern that I have located is in this book, in the section on the Sui period.  I have now acquired a copy with its dust jacket and find a painting of the pattern is on the back of the dust jacket -- this is the painting also given in Li Kai et al, below, p, 27.  [Incidentally, a devi or apsaras is a kind of Buddhist angel.  The art of Dunhuang is quite lovely.]

The Dunhuang Institute for Cultural Relics.  The Mogao Grottos of Dunhuang.  5 vols. + Supplement.  Heibonsha Ltd., Tokyo, 1980-1982.  (In Japanese, with all the captions given in English at the end of the Supplementary volume.  Fortunately the plate and cave numbers are in western numerals.  A Chinese edition was planned.)  Vol. 2, plate 94, is a double-page spread of the ceiling of Cave 407, with the page division running right through the middle of the rabbits pattern!  Vol. 2, plate 95, is a half-page plate of the ceiling of Cave 406, and shows the rabbits pattern, but it seems rather faded.  The English captions simply say "Ornamental ceiling decoration".

R. Whitfield  &  A. Farrer.  Caves of the Thousand Buddhas.  British Museum, 1990, esp. pp. 12 & 16.  Though cited by Greeves, these pages only have general material on Dunhuang and the book does not mention any of the relevant caves.

Duan Wenjie.  Dunhuang Art  Through the Eyes of Duan Wenjie.  Indira Gandhi National Centre for the Arts, Abhinav Publications, New Delhi, 1994.  This gives much more detail about the caves.  Pp. 400-401 describes Caves 406-407.  Peter Rasmussen examined the book in detail and found it mentions 12 other caves with the pattern.  Peter has found that the book is accessible on-line at  www.ignca.nic.in/ks_19.htm.  This provides the facility to download a font which will display diacritical marks and this is worth doing before you start to browse. 

                    In late 2002, Peter Rasmussen and Wei Zhang were able to inspect 10 caves that they hadn't seen before.  Peter sent notes of their impressions of 12 caves on 24 Nov 2002 and I will add some of Peter's comments and additions in [ ] and marked PR.  This will make the following list the basic list of all sixteen of the caves.

 

                    Cave No. 127.  Late Tang (renovated in Five Dynasties and Qing).  "The ceiling exhibits lotus and three rabbits (joining as one) in the centre."  [PR: tan on turquoise, going clockwise.]

                    Cave No. 139.  Late Tang.  "The ceiling shows the three rabbits (joining as one) and lotus designs in the centre."  [PR: this is a small cave off the entry to Cave 138.  Tan on light green, going clockwise, in excellent condition.  "Rabbits beautifully drawn in pen-like detail, showing toes, eyes (with eyeballs!), all four legs, tail, nose, mouth, outline of thigh muscle, and hair on stomach, breast, legs and top of head."  Peter says this is by far the finest of the images; he is applying to get it photographed.]

                    Cave No. 144.  Middle and Late Tang (renovated during the Five Dynasties and Qing).  "The centre of the ceiling shows the three rabbits (joining as one) and floral designs."  [PR: white on aqua green, going clockwise.]

                    Cave No. 145.  Late Tang (renovated during the Five Dynasties and Song).  "The ceiling of the niche on the west wall shows lotus and three rabbits (joining as one), chess‑board and floral patterns."  [Zhang & Rasmussen's letter of 1 Jun 2001 says the rabbits are going clockwise.]

                    Cave No. 147.  Late Tang.  "Main Hall: The ceiling shows three rabbits (joining as one) and lotus designs in the centre".  [PR: tan or turquoise green, going clockwise.  Paint on ears has peeled off.]

                    Cave No. 200.  Middle Tang.  "The ceiling shows three rabbits (joining as one) and round petalled lotus designs in the centre".  [PR: white on turquoise green, going clockwise.]

                    Cave No. 205.  Early and High Tang (renovated during Middle Tang and the Five Dynasties).  Main Hall: "The ceiling has the three rabbits (joining as one) designs drawn in Early Tang".  [Zhang & Rasmussen's letter of 1 Jun 2001 says the rabbits are going clockwise.]

                    Cave No. 237.  Middle Tang (renovated in Western Xia and Qing).  Main Hall: "The ceiling shows the three rabbits (joining as one) and round petalled lotus designs in the centre."  [Zhang & Rasmussen's letter of 1 Jun 2001 says the rabbits are going clockwise.]

                    [PR: Cave No. 305.  Sui (renovated during the Five Dynasties and Qing).  "Rabbit paint is gone (only white silhouette remains on faint rusty red background)."  See: Decorative Patterns in the Dunhuang Art; Li Kai et al; Zhang & Rasmussen's letter of 1 Jun 2001, which says the rabbits are going clockwise]

                    Cave No. 358.  Middle Tang (renovated during the Five Dynasties, Western Xia and Qing).  "Main Hall: The caisson ceiling shows the three rabbits (joining as one) and round petalled lotus in the centre."  [PR: white on turquoise, going clockwise, faded.]

                    Cave No. 383.  Sui (renovated during Song, Western Xia and Qing Dynasties).  "Main Hall: The centre of the caisson ceiling shows the three rabbits (joining as one) and lotus flower designs".  [PR: brown with white outlines, going clockwise, fair condition.]

                    Cave No. 397.  Sui and Early Tang (renovated during the Five Dynasties and Qing).  "Main Hall: The caisson ceiling shows the three rabbits (joining as one) and lotus in the centre".  [PR: white on peeled-off aqua green, going clockwise. Poor to fair condition.]

                    Cave No. 406.  Sui (renovated in Song and Qing).  "The centre of the caisson ceiling shows four designs of a set of three rabbits (joining as one) and lotus".  I don't quite understand his phrasing -- there is a picture of a pattern of three rabbits in the centre of a lotus, as in Cave 407, but perhaps there are other patterns which are not reproduced??  [PR: white on tan, going clockwise, faded, fair condition.  Peter says nothing to clarify Duan's text.]

                    Cave No. 407.  Sui (renovated in Song and Qing).  "Main Hall: The caisson ceiling is covered with the three rabbits, lotus designs and flying figures drawn in Sui."  [PR: Black with white outline on turquoise green background, going anticlockwise, good to excellent condition.]

                    [PR: Cave No. 420.  Sui.  Zhang & Rasmussen's letter of 1 Jun 2001 says the rabbits are going anticlockwise.]

                    Cave No. 468.  Middle Tang (renovated during the Five Dynasties).  "Main Hall: The centre of caisson ceiling has three rabbits (joining as one) and a lotus design".  [PR: white on turquoise, going clockwise.]

 

                    CHRONOLOGY  Tang is 618-907, but there is a Later Tang (923-936).

          Sui (581-618, so c600):  305, 383, 406, 407, 420.

          Sui and Early Tang (c620):  397.

          Early/Mid Tang (7-8C):  205.

          Mid Tang (8C):  200, 237, 358, 468.

          Mid/Late Tang (8-9C):  144.

          Late Tang (9-10C): 127, 139, 145, 147.

 

Roderick Whitfield.  Dunhuang  Caves of the Singing Sands.  (Revision of a Japanese book by NHK, 1992.)  Textile & Art Publications, London, 1995.  On pp. 59 & 238, plates 66 & 361-362, are pictures of the roof of cave 420 which may be showing a three rabbit pattern, but it is too faded and too small to really be sure.

Decorative Patterns in the Dunhuang Art.  1996.  This has no English text except for the book title.  The dust jacket has a colour picture of a painting of the ceiling of cave 407.  Otherwise it gives only black and white drawings of patterns.  P. 31 is an introduction to a section and has an unidentified three rabbits pattern, probably from cave 407.  P. 32 is cave 305.  P. 33 is cave 397.

Li Kai, chief designer;  Zhao Le Nin  &  Luo Ke Hua, eds.  The Selections of Copied Art Works of Dun Huang Sunk Panel.  1997.  This presents various artists' paintings of the ceiling panels. 

                    Plate 21 is: Sui Dynasty "three rabbits and lotus flower" sunk panel (cave number 305).

                    Plate 26 is: Sui Dynasty "three rabbit and flying Apsarase" sunk panel (cave number 407) copied by Duan Wen Jie.

                    Plate 27 is: Sui Dynasty "three rabbits and flying Apsarase" sunk panel (cave number 407) copied by Guo Shi Qing and Chang Sha Na.  This is reproduced on the cover of the book, but it has been reversed in printing in both places.  The difference in coloration and even in the outlines show the difficulty of seeing what is present in a rather faded and damaged cave where the early artists had only daylight.  Neither of these is the same as the cover of the previous item!  See also Shuhong & Chegxian above for another version.

                    Plate 32 is: Sui Dynasty sunk panel (the cave number is not recorded).

                    Plate 26 is: Tang Dynasty sunk panel (cave number 205).

Wei Zhang  &  Peter Rasmussen.  Letter of 1 Jun 2001 reporting on their research and visit to Dunhuang.  They have made good contacts there and were given a lengthy special tour.  They have now found the three rabbits pattern occurs in 16 caves: 127, 139, 144, 145, 147, 200, 205, 237, 305, 358, 383, 397, 406, 407, 420 and 468, all in the Mogao caves and all as central ceiling panels.  They were able to see 145 (late Tang, 848-906), 205 (Early Tang, 618-704), 237 (mid Tang, 781-847), 305 (Sui, 581-618), 407 (Sui, 581‑618), 420 (Sui, 581-618).  In the first four, the rabbits are going clockwise, in the last two, they are going anticlockwise.  They also visited the Western Thousand Buddha Caves, but there was no sign of the three rabbits there.  However, none of the researchers there has investigated the three rabbits pattern.  Unfortunately, photography is not allowed, but they sent the previous two books and two reproductions from an otherwise unidentified book: Dunhuang of China (2000). 

                    P. 20 is: Hall With Inverted Funnel Shaped Ceiling Shape of Cave 305 (Sui Dynasty) and gives a good impression of the shape of these caves -- this one is roughly cubical with several statues, apparently life-size, on a central plinth, with a decorated ceiling with the Three Rabbits in the centre.  The central part of the ceiling usually is a panel which is sunk into the ceiling, i.e. higher than the rest of the ceiling, but some have two steps and some are more rounded.  [Greeves (2001)] suggests this is to represent a cloth canopy.

                    Pp. 40-41 is: Pattern (ceiling) Cave 407 (Sui Dynasty), showing the whole ceiling and the tops of the walls.

 

                    OTHER ASIA

 

Anna Filigenzi, of the Istituto Universitario Orientale di Napoli and the Italian Archaeological Mission in Pakistan reports and has sent an image of a three rabbits plaque found at Bir-kot-ghwandai, Swat, Pakistan and dating from 9C-11C.  The publication is: P. Callieri et al.; Bir-kot-Ghwandai 1990-92  A Preliminary Report; Supplemento n. 73 of Annali dell'Istituto Universitario Orientale de Napoli 52:4 (1992) 45 -- ??NYS.   [emails of 30 Oct & 10 Nov 2001.] 

In the Museum für Islamische Kunst, Berlin, is a glass medallion with the Three Hares pattern, attributed to Afghanistan.  It is 52mm in diameter.  The museum purchased it from a dealer and there is no record of its origin.  Peter Rasmussen & Wei Zhang were shown it in mid 2002 by Dr. Jens Kröger, the Curator of Islamic Art at the museum.  [Email from Rasmussen, 17 Aug 2002.]  There is a description of it in Kröger's book and in Stefano Carboni; Glass from Islamic Lands; Thames & Hudson, 2001, pp. 272-280.  This also describes a fragmentary glass piece with four rabbits in the al-Sabah Collection in Kuwait, supposed to come from Ghazni, Afghanistan.  [Email from Rasmussen, 20 Aug 2002.]

Eva Wilson.  Islamic Patterns (British Museum Pattern Books).  British Museum Publications, 1988.  Plate 42, bottom picture, is a four rabbits pattern.  The notes on p. 19 say:  Engraved design in the centre of a brass plate.  Diameter 7cm.  Iran, 12th century.  British Museum, London (1956 7‑26.12).

[Greeves (2000)] reports a light blue glass seal from Afghanistan, c1200.

In the Hermitage Museum, St. Petersburg, is an oriental silver flask of 12/13C with the three rabbits on the base.  Discussed and illustrated in Hamann-MacLean, below.  A small illustration is in Baltrušaitis (Fig. 96D), cf under China, above, who says it comes from near Perm and has a Kufic inscription on it.  Cf next entry.

Vladislav P. Darkevich.  Khudozhestvennyi metall Vostoka VII‑XIII.  Nauka, Moskva, 1976, 195 pp.  ??NYS – described by Peter Rasmussen [email of 8 Jan 2002] and photocopy of pl. 34 sent by him.  See pp. 16, 17, 115 and Таблуца 34: Серебряные чаша у флакон 5 - 8: Селянино Озеро (No. 19) [Tablutsa 34: Serebryanye chasha u flakon: 5 - 8: Selyanino Ozero (Plate 34: Silver basin and flagon: 5 - 8: Salt(?) Lake)].  This is the Hermitage Museum silver flask.  Fig. 5 is an overall view; fig. 6 is the bottom, which is rounded, rather like part of a sphere and has the three rabbits pattern; figs. 7 & 8 are the Kufic inscriptions.

[Greeves (2000)] reports a fine metal tray from Iran, c1200, in the Keir Collection and gives a fine photo of it.  He notes that Sassanian culture is believed to have spread outward and could have been the source for the Himalayan and Chinese versions as well as the western versions.  Sue Andrew tells me that similar trays are in the Victoria and Albert Museum, London.

[Greeves (2000)] reports that the pattern has been found in Nepalese temples and in c1200 wall paintings at the temple complex of Alchi in Ladakh, Jammu & Kashmir.  I can't find Alchi on maps, but a guide book says it is on the Indus about 1½ hours drive from Leh.

Roger Goepper.  Alchi: Buddhas, Göttinnen, Mandalas: Wandmalerei in einem Himalaya‑Kloster.  DuMont Buchverlag, Köln, 1982, 110 pp.  ??NYS -- photocopies sent by Peter Rasmussen.  Another version appeared in 1996 -- ??NYS, described by Rasmussen [email of 21 Dec 2001], see below -- but the earlier book has different pictures.  Rasmussen says this is about Sumtsek, the Buddhist temple at Alchi, Ladakh, presumably the temple mentioned by Greeves, above.  Goepper dates the temple as c1200, but another author suggests 11C or 12C [Lionel Fournier; The Buddhist Paradise: The Murals at Alchi; 1982].  On p. 127 of the 1996 book is a photo of a 4.63 m high statue of Maitreya (the Buddha) wearing a dhoti on which Rasmussen finds no less than 50 (fifty!) examples of the three hares pattern.  A close-up on p. 128 shows three complete examples.  Goepper refers to these examples as 'three or four deer-like animals, the three long ears being shared ...', but Rasmussen could not see any examples with four animals and the text only refers to three ears.  However, plate 7 of the 1982 book shows the same two roundels as on page 128 of the 1996 Alchi, but this illustration shows FOUR rabbits (sorry, hares) or deer or bulls (take your pick) in the upper left corner above Maya that were cropped out of the 1996 illustration, so Goepper's reference on page 126 of his 1996 book to "three or four deer‑like animals" was correct.  In a note on p. 278 of the 1996 book, Goepper says the occurrence of the same pattern at Paderborn is 'hardly anything more than a coincidence'.  Rasmussen finally notes that Goepper said the photographer, Jaroslav Poncar, and his group took about 3000 transparencies, a 'virtually complete documentation of the Alchi murals', but only 300 occur in the 1996 book.  I have seen a B&W copy of Plate 7 of the 1982 edition.  This has two images, each of a central character in a roundel surrounded by four frames in the form of a Greek cross with extra squares at the corners.  (The colour image from the 1993 book, see below, shows these are adjacent, indeed overlapping, images.)  These frames contain patterns of three and, in one case, four animals, but the identity of the animals is not clear.  The four-fold pattern seems to be rabbits sharing ears, but others seem to be horses (or bulls) sharing ears or deer sharing antlers or possibly bulls sharing horns.  All the animals are going clockwise.

                    Sue Andrew has been in contact with Goepper & Poncar.  Goepper said "just about a week ago I hinted at the 'intercultural' character of this strange motif during my lectures at Cologne University".

Pratapaditya Pal.  A Buddhist Paradise: The Murals of Alchi Western Himalayas.  Ravi Kumar for Visual Dharma Publications Ltd., Hong Kong, 1982, 67 pp.  ??NYS -- described by Peter Rasmussen [email of 4 Jan 2002].  Plate S9 shows Maitreya's full dhoti, S10 and S11 are closeups of the individual legs, and S12 and S13 are closeups of details.  Lionel Fournier's photography is very poor in comparison to Poncar's.  On pp. 51‑52, Pal states "Interspersed with the rondels [sic] are little cruciform blocks adorned with leaping bulls, whose exact function is not clear, but which remind one of similar though more naturalistic bulls on the ceilings of Ajanta."  So Goepper's deer are Pal's BULLS!!!  But the interesting part of the sentence is the reference to the ceilings of Ajanta.  It's not clear whether Pal means the Ajanta bulls are leaping in threes, but this statement reminded us that Terese Bartholomew of the San Francisco Asian Art Museum told us she thought she remembered the three‑animal motif being somewhere in the Ajanta caves.  In an email of 8 Jan, Rasmussen reports that he has gone through all the Ajanta books at UC Berkeley and could only find an image of four deer sharing a single head -- details are in the Dead Dogs section.

Roger Goepper.  The 'Great Stupa' at Alchi.  Artibus Asiae 53:1/2 (1993) 111-143.  ??NYS -- photocopy sent by Peter Rasmussen.  Figure 9: Deities in outer triangles of the ceiling.  Photo by Poncar.  This shows a number of three and four beast roundels forming a decorative band going out of the picture.  as with the other Alchi images, it is hard to tell whether the beasts are rabbits, hares, bulls, etc., but they are clearly sharing ears here.  all the beasts are going clockwise.

James C. Y. Watt  &  Anne E. Wardwell.  When Silk Was Gold  Central Asian and Chinese Textiles.  The Metropolitan Museum of Art in Cooperation with The Cleveland Museum of Art, dist. by Abrams, nd [mid 1990s?, after 1993].  Section 45: Cloth of gold with rabbit wheels, p. 158, with a colour plate opposite (no number on my copy).  This shows a square array of circles with four rabbits in each circle.  The direction of the rabbits alternates from one row of circles to the next.  It is from the 'Eastern Iranian world, second quarter to mid-13th century', i.e. c1240.  The cloth has a green-gold colour, while the patterns are outlined with red silk, giving red-gold lines.  A footnote says to consult Roes 1936-37, pp. 85‑105 for the history of the motif of the animal wheel.  Other notes cite Dunhuang, metalwork in Khurasan (1150-1225), the dhoti of the Maitreya at Sumtsek Temple in Alchi and the ceiling paintings of the Great Stupa in Alchi.  The section goes on to discuss the 'two heads, four boys' motif - see under Dead Dogs, below.

                    Sue Andrew, via Peter Rasmussen [email of 26 Dec 2001], reported finding this, but only gives Wardwell as author - perhaps she wrote this part, but this is not indicated on the photocopies sent by Rasmussen.  Peter Rasmussen [email of 4 Jan 2002] says Wardwell's main source was Baltrušaitis, under China, above.  However, Baltrušaitis doesn't mention several of the cited areas.

[Greeves (2001)] reports a pre-Islamic Mongol coin from north Iran, dated 1281, with the three hares on one side.

Jurgis Baltrušaitis.  Le Moyen Age Fantastique,....  Op. cit. under China, above.  Pp. 132‑139 of the 1981 edition have many examples of three and four rabbits, four boys, etc.  After discussing Dunhuang, he says the motif was taken on by Islam and cites the Petersburg cup.  He says there is a Mogul (school of Akbar) miniature with the pattern. 

Roger Goepper.  Alchi  Ladakh's Hidden Buddhist Sanctuary.  Serindia Publications, London, nd [1993?].  Colour image sent by Peter Rasmussen showing the two overlapping images of the dhoti of Maitreya shown in his 1982 book, above.  However, this image is more centralised and hence shows only one of the three animals patterns completely, with parts of several others.

In Cairo, the Museum of Islamic art has a fragment from the bottom of a bowl with a three rabbits pattern using three colours!  It is their item 6939/1, coming from Egypt or Syria in the 12-13C.  In 2001-2002, it was a featured item at the exhibition:  L'Orient de Saladin  L'Art des Ayyoubides  at the Institut du Monde Arabe, Paris.  It is reproduced in the catalogue:  Éric Delpont et al.  L'Orient de Saladin  L'Art des Ayyoubides.  [Catalogue for] Exposition présentée à l'institut du monde arabe, Paris  du 23 octobre 2001 au 10 mars 2002.  Institut du monde arabe / Éditions Gallimard, Paris, 2001.  Item 111, p. 123, is 'Tesson aux trois lièvres'.  A black on yellow version was adopted as the logo of the exhibition and hence appeared on many other items associated with the exhibition and on the advertisements for it.

 

                    PADERBORN

 

In the cloister (Kreuzgang) of the Cathedral (Dom) of Paderborn, Nordrhein-Westfalen, Germany, is the "Three Hares Window" (Dreihasenfenster), with hares instead of rabbits.  This faces the outside, i.e. into the central garden of the cloister.  I learned of this from the Michelin Green Guide - Germany (Michelin et Cie, Clermont-Ferrand, 1993, p. 229) and wrote a letter of enquiry.  A response from Dr. Heribert Schmitz in the Archbishopric states that the present form of the cloister dates from the early 16C and this is the date given on a postcard he included (and in a local guidebook).  ([Greeves (2000); Greeves (2001)] says probably 15C.)  He also included a guide to the Cathedral, a poster and a copy of the parish magazine with the article by Schneider, see below.  On 12 Jul 2000, I was able to visit Paderborn, see the window, meet Dr. Schmitz and obtain much more material.  The image is actually carved stone tracery in the arch over one of the triple windows of the cloister, and is about 3 ft (1 m) across.  The central stone image is supported only by the three rear feet of the hares which are on a circular rim -- the intermediate spaces are filled with leaded glass.  The local guidebook refers to the mason as 'crafty', perhaps implying that he saved having to carve three more ears.  Several of the photos show the bodies of the hares supported by metal rods, but there are presently no rods.  Later inquiry revealed that the original version is now in the Cathedral Museum and the version in the cloister is a recent copy.  The Cathedral guidebook refers to the 'well known' window and says the symbol is an old land-mark of the city and the poster describes it as a famous landmark to be studied and developed in a workshop for children.  One guide book shows three people dressed as hares who are a regular feature of parties and celebrations.  The Cathedral guidebook says the 'motif is also to be found in other buildings, but elsewhere is mostly smaller and less conspicuous', but no references are given and Dr. Schmitz's letter says that he knows of no other examples than Long Melford and the article by Schneider.  Greeves, below, notes that St. Boniface, the Apostle of the Germans, came from Crediton, Devon, some 10 mi east of Dartmoor!  Further, he consecrated a bishop at Paderborn.

Hans Schneider.  Symbolik des Hasenfensters in Nordwestchina entdeckt.  Die Warte (Heimatzeitschrift für die Kreise Paderborn und Höxter) 32 (Dec 1981) 9.  This was kindly provided by Dr. Schmitz of the Archbishopric of Paderborn.  First Schneider gives various interpretations of the symbolism of the three hares pattern:  old German fertility symbols from the myths of the gods;  the Easter rabbit as a symbol of the eternal power of nature;  a symbol of the Trinity.  In recent years, it has been connected to the patron saint of Paderborn, St. Liborius, by viewing his name as Leporius, which means 'hare man'.  But Schneider has discovered the article of Huang and gives a B&W reproduction of the picture.  Schneider notes that Paderborn had connections with the Islamic world -- e.g. Achmed el Taruschi and a delegation came to Paderborn in 970, [and we know Charlemagne's court had contacts with Constantinople, Córdoba and Baghdad].  Hence it is possible that the Chinese symbol could have been transmitted to Paderborn [and elsewhere].

A small illustration is in Baltrušaitis (Fig. 96D), under China, above. 

On 12 Jul 2000 I was able to visit Paderborn and meet Dr. Schmitz.  The way to the Dreihasenfenster is clearly signposted in the Cathedral and we found images of it elsewhere in the town, and the local guidebook mentions further locations, e.g. the Drei Hasen restaurant at 55 Königstrasse.  I got an English version of the Cathedral guide and a children's guide to the Cathedral.  I obtained two more postcards featuring the window and several multi-image postcards of Paderborn with the window as one image.  I also bought a stained glass roundel of the pattern, 225mm (8 3/4") across.  The city information office has the pattern on many of their guide books and I also got stickers, transfers, etc. 

                    I had met Michael Freude from Münster and he had recalled there was an example in Münster and that there was a children's rhyme about it, though he could not remember it, nor could his family.  We stayed with Michael Freude and Hanno Hentrich in Münster and they had located the example in Münster, which is a roof boss in the southwest corner of the south transept of the Dom (Cathedral) (St. Paul's), over the organ.  It is very high and I was unable to get a good picture of it.  [Greeves (2000)] notes that it is stunning, but he told me that he also had been unable to get a good picture.  Since the Dom was much restored after the War, I thought it might be a post-war addition, but Hentrich checked in a Münster history and photos showed this part of the Dom had survived.  [Greeves (2000)] says it is early 16C.  I told Dr. Schmitz of this example when I visited him and he did not know of it.  Freude & Hentrich thought it likely that the Paderborn example was a replacement after the War.

                    [Greeves (2000)] says there is an example in the cloisters of a former monastery at Hardehausen, S of Paderborn (but I can't locate this on my maps).

Theodore Fockele & Ewald Regniet.  Domführer für Jungen und Madchen.  Metropolitan-Kapitel, Paderborn, (1982), 6th ptg, 1999.  On p. 21 is a brief description and a drawing with caption being the rhyme:  "Der Hasen und der Löffel drei,  /  und doch hat jeder Hase zwei."  [The hares and ears are three,  /  and yet each hare has two[, you see].]  This rhyme also occurs in the local guidebook and on one of the available postcards.

Verkehrsverein Paderborn [Paderborn Tourist Information].  Paderborn  A short guide to the old city.  Paderborn, 1998, p. 13 and back cover.  This says the window is early 16C and gives an English version of the rhyme:  "Count the ears.  There are but three.  But still each hare has two, you see?" and I have now inserted 'you see' into my translation above.  The pattern is printed on the outside covers of this booklet

One of the postcards available in Paderborn has a longer poem. 

 

                    Viribus Auribusque Unitis (Mit vereinten Kräften und Ohren)

 

                    Jedweder Hase hat zwei Ohren.

                    Und hier ging jedem eins verloren.

                    Das Soll ist sechs, das Ist nur drei.

                    Und Schein und Sein sind zweierlei.

                    Was führt der Steinmetz wohl im Schilde?

                    Welch ein Gedanke liegt im Bilde?

                    Die Ohren sitzen an der Stirne,

                    Gehörtes fliess in drei Gehirne.

                    Drittselbst wird hier somit bedacht,

                    was Sorgen oder Freude macht.

                    Vereint geht manches leichtes eben

                    im Hasen- wie in Menschenleben.

                    Und überdies ist, was ihr seht,

                    'ne Spielart von der Trinität.

 

                    [With united powers and ears     Every hare has two ears.  And here each has lost one.  There should be six, there are only three.  And appearance and being are different.  How can the stone mason make an emblem?  What thought is in the picture?  The ears sit on the forehead, which flow into three heads.  A third itself is here thus considered, which makes fear or joy.  United, many things go easily even in the life of hares as in the life of men.  And moreover this is, as you see, a playful image of the Trinity.]

Annemarie Schimmel.  The Mystery of Numbers.  (As: Das Mysterium der Zahl; Eugen Diederichs Verlag, Munich, 1984.  Based on: Franz Carl Endres (1878-1954); Das Mysterium der Zahl, last edition of 1951.  It's not clear when Schimmel's work was done -- the Preface is dated Sep 1991 and her © is dated 1993, so perhaps 1984 refers to the last printing of the original Endres book??)  OUP, 1993, p. 63 has a drawing of the Paderborn three hares, but with no indication of the puzzle aspect.  "Hares, symbols of the tri-unity that is always awake, seeing and hearing everything.  Their ears form a triangle.".

 

                    MEDIEVAL EUROPE

 

          There are 17 churches with 28 roof bosses of the Three Rabbits in Devon.  The dating of these is not very exact and is not always given in the church guides.  I have now discovered Cave's 1948 book which mentions many of these.  See also Jenkins' book of 1999.

 

Carl Schuster  &  Edmund Carpenter.  Patterns That Connect  Social Symbolism in Ancient & Tribal Art.  Abrams, NY, 1996.  Pp. 158-159, fig. 453, is a 12C European wind chart with a central demon face with four mouths, but it has four pairs of eyes.

A stone boss supposed to come from a church demolished c1200 is built into a cellar of a house in Corbigny, Nièvre, Bourgogne, about 50km NE of Nevers.  [Greeves (2000) with colour photo.]  [Greeves (2001)] changes the name to Corbenay, but this is not in my French atlas.]

[Greeves (2000); Greeves (2001)] says the pattern occurs on a bell from the early 13C in the great abbey church at Kloster Haina.  I can't locate this on my maps, but there is a Haina, Hessen, about 40km SW of Kassel.

Bestiary.  MS Bodley 764, 1220-1250.  ??NYS.  Translated by Richard Barber as: Bestiary  Being an English version of the Bodleian Library, Oxford M.S. Bodley 764  with all the original miniatures reproduced in facsimile; Folio Society, 1992;  The Boydell Press, Woodbridge, Suffolk, 1993;  PB, 1999.  Pp. 66-67, the entry for hare is preceded by a 'four rabbits' pattern.  There were many medieval versions of the bestiary and the BL MS Harley 4751 is very similar to this.  However, Barber [p. 13] notes that the entry for the hare is not in earlier texts and rarely reappears in later texts.  Thanks to Sue Andrew for this reference and a copy of the photo she had done from the original MS. 

Sue Andrew also showed me an illuminated initial Q, also with four rabbits, from a c1285 MS, a French copy of Peter Lombard's Gloss on the Psalter, Bodleian MS Auct.D.2.8. f. 115r, commentary on Psalm 51.

A roof boss in the south choir aisle of Chichester Cathedral, dated to the first half of the 13C, shows six 'Green Men' sharing eyes.  The Green Men have foliage coming out of their mouths.  My thanks to Marianna Clark for noticing this and sending me an example of the colour postcard of it which is labelled: Six heads with six eyes between them.  Colin Clark, the Chief Guide to the Cathedral, told me of the example in Boxgrove Priory.  See Cave, 1930 & 1948.

A roof boss in Boxgrove Priory is roughly contemporary with that in Chichester Cathedral, but shows eight faces sharing eyes.  Photo 300 in Cave, who includes it on a page of Foliate Heads.  On p. 184, he says alternate heads have a stem from the mouth, but this is very small, leading one to wonder if the foliage has been broken off.  [Jenkins, pp. 686-687] says: "The second boss from the altar is so crafted that each of eight faces comes complete with two eyes, yet there are only eight eyes in all."  See Cave, 1948.

A roof boss in the Chapter House of a former Benedictine Abbey, now the sacristy of the church of Saints Peter and Paul, Wissembourg, Bas-Rhin, Alsace, is dated to c1300.  Nearby bosses include a Green Man, one of the common figures in Devon bosses.  [Greeves (2000); Greeves (2001).]

On the right side of the southern west doorway of St. John's Cathedral (Primatiale St-Jean) in Lyon, there is a quatrefoil panel with four rabbits, from about 1315.  The rabbits are going clockwise.  Discussed and illustrated in Hamann-MacLean, below, and in Baltrušaitis (Fig. 96A), cf under China, above.  I photographed this in Lyon and my picture is much better than that in Hamann-MacLean, possibly because his photograph was taken before the facade was cleaned in 1982.  However, there is no postcard of the pattern and I could not find a picture in any of the material available -- but see the next item.

                    John Winterbottom & Diana Hall have sent a photo of a similar four rabbits pattern, but going anti-clockwise, at Bord du Forêt, near Lyon.  Here the rabbits are leaping upward, so the central square of ears is smaller and tilted by about 30o.

N. Reveyron.  Primatial Church of Saint-John-the-Baptist, Cathedral of Lyon.  Translated by Valérie Thollon & Diana Sarran.  Association Lyon Cathédrale, Lyon, nd [c2000].  Pp. 22-29 describe the western doorways, saying the sculptures were made during 1308‑1332.  The South doorway is described on pp. 28-29, and the third section, on p. 28, headed Bestiary, includes: "The four hares, arranged like a swastika."  This is the only time I have seen the four rabbits pattern compared to a swastika and I don't think there is any similarity!

In the Church of St. John the Baptist, North Bovey, Devon, there is a carved wood roof boss of the Tinners' Hares which possibly dates to the 13C.  The leaflet guide to the Church says the pattern was an emblem of the tin-miners of the 14C and is thought by some to refer to the Trinity.  The symbol is used by a number of local firms as a logo -- though I didn't see any on my visit in 1997.  Thanks to Harry James, Churchwarden, for his letter of 25 Apr 1997, his drawing of the pattern and a copy of the guide.

In the Church of St. Mary the Virgin, Long Crendon, Buckinghamshire, about halfway between Aylesbury and Oxford, is a mid-14C tile showing the Three Rabbits pattern.  About a third of it is missing.  It is just at the altar step, which has preserved it somewhat from wear.  The church guide says the tile was made in nearby Penn and is unique.  This is an encaustic tile with the rabbits in yellow on a pale orange background.  (This church is generally locked; try telephoning the Vicar on 01844-208363 or the Churchwarden on 01844-208665 if you want to get in.  My thanks to Avril Neal, the Churchwarden, for letting me in.)  The pattern is reproduced in B&W in  Carol Belanger Grafton; Old English Tile Designs; Dover, 1985 [selected from Haberly, cf below], p. 91, but there is no text or indication of the source of it.  My thanks to Sue Andrew for finding the tile and the Grafton reproduction.  She has persuaded a potter to make facsimiles of the tile.  The potter is Diana Hall, Anne's Cottage, Wimborne St. Giles, Dorset, BH21 5NG; tel: 01725‑517475.  I have now purchased some of these and they are very fine.

                    Loyd Haberly.  Mediaeval English Pavingtiles.  Blackwell for Shakespeare Head Press, Oxford, 1937.  Shows the complete pattern in red and white on the TP.  Shows it in B&W as fig. CXIII on p. 168, saying it occurs in Long Crendon Church and Notley.  "The design is also found elsewhere in glass and carved on stone.  Some say it symbolizes the Trinity, others the Trivium, or three Liberal Arts of Grammar, Rhetoric and Logic.  The conies, these say, represent the scholars, a feeble folk, who have an ear for each of the three Arts.  One writer thinks the donor of the this design was therefore a man of academic distinction."  Thanks to John Winterbottom and Diana Hall for copies of this material.

                    [Greeves (2000)] discusses this and says it is the earliest known British example, presumably because the North Bovey example is not precisely dated.  [Greeves (2001)] thinks this is roughly contemporary with the Chester tile, see below, and another tile in Anglesey.

In the Grosvenor Museum, Chester, is a floor tile from the Cathedral with an inlaid design of the Three Rabbits pattern.  Sue Andrew has found that this is shown in Jane A. Wight; Mediaeval Floor Tiles; John Baker, London, 1975 and she has kindly sent a photocopy.  On p. 48 is fig. 15: "Trinity Rabbits: Narrow inlaid design of linked rabbits, symbolising the Holy Trinity, in Chester Cathedral.  (About 5⅜ inches square.)"  These are just outline rabbits, looking much as though Matisse had drawn them.  The outer edges of the ears are curved to produce a circle, so the usual delta shape is here very curved.  On pp. 12‑13, Wight discusses the pattern.

                              Some signs are ambiguous, hovering between religion and magic, like the three rabbits or hares linked by their shared ears, that may act as a symbol of the Holy Trinity.  In The Leaping Hare by George Ewart Evans and David Thomson (1972) it is pointed out that these are correctly hares, 'joined in a kind of animated Catherine-wheel' and 'another instance of a pre-Christian symbol being adopted by the church'.  (On roof bosses in the Dartmoor churches, financed by money from the stanneries or mines, the creatures appear as the craft-badge of the tin-miners.)  This 'Holy Trinity' is found on tiles in Chester Cathedral and in Buckinghamshire.

          It is clear that Wight and Evans & Thomson have no knowledge of the puzzle aspect of the pattern.  The connection with the tin miners is now known to be a modern myth -- see Greeves, below.  Hares certainly have pre-Christian associations, but I don't know of any pre-Christian example of the Three Rabbits pattern except at Dunhuang and possibly St. Petersburg, see above.  [Greeves (2000)] mentions this example and gives more non-Christian examples.  [Greeves (2001)] suggests this is early 14C, roughly contemporary with the Long Crendon example and gives a photo of the actual tile on p. 62 -- this is slightly damaged in the middle, but does not have the circularity I noted above which is thus an artist's liberty.  Greeves, Andrew & Chapman recently visited Chester, but the only tiles shown to them were from excavations in the Cathedral nave in the 1990s, which makes us wonder if the tile described by Wight is still somewhere in the Cathedral??

Sue Andrew (Nov 2002) reports that Diana Hall has learned of a 14C tile from the Stadion's Prebendary Court in Constance, now in the Zurich Schweizerisches Landesmuseum.

In St. Pancras Church, Widecombe, Devon, is a roof boss of the three rabbits pattern, probably from the late 14C.  The guide book to the church says it is "a symbol of the Trinity connected with tin-mining."  This is shown on a postcard by Judges of Hastings, card number c11933X, where it is called the Hunt of Venus.  A separate guide to the roof bosses also calls it the Hunt of Venus and suggests the tinners took the imagery from either the mines being like rabbit burrows, or from Venus, the goddess of Cyprus, the island which produced the copper that the tin was combined with.  Thanks to the Rector of the Church, Derek Newport, for the material.  [Jenkins, 1999, p. 142] says: "This rare symbol of the Trinity is formed of three animal heads sharing just three ears."

At The Great Church of the Holy Trinity, Long Melford, Suffolk, there is an example of the three rabbits pattern in the 15C(?) stained glass.  Christopher Sansbury, the Rector, wrote on 3 Jun 1996 that the motif is common on the east side of Dartmoor and that it may have been brought to Suffolk by the Martyn family c1500.  He says it is old glass, older than the church, which was built in 1484, but doesn't specify a date for it, nor does the commercial postcard (Jarrold & Sons, Norwich, no. CKLMC 6).  The pattern is considered to be an emblem of the Trinity.  In a later latter, he cited Chagford, North Bovey and Widecombe as churches with the pattern near Dartmoor and cites Greeves' article.  Greeves, below, says the pattern seems to be about 5.5cm in diameter, but I wonder if he is measuring a picture as when I visited the Church, it seemed to be perhaps 5" or 6".  Sansbury wrote that it was less than a foot across.  Greeves also says that the Cornish merchant family of Martin came to Long Melford c1490, so the glass is more likely to be 16C.  [Jenkins, 1999, pp. 658-659.]

The Church of St. Michael the Archangel in Chagford, Devon, has a roof boss in the chancel depicting the Tinners' Rabbits, from the 15C.  My photo is not very clear.  The Rector, P. Louis Baycock, states that there are one or two other bosses in the wooden ceilings, but they are dark and were obscured by the lighting so I was unable to locate them.  On a later visit, I managed to see one at the top of the aisle vault a bit to the left of the entrance door -- one needs to shield one's eyes from the light bulb near by.  There is a kneeler in the church with the pattern and it figures at the lower right of a large modern embroidery 'Chagford through the ages' hanging at the back of the church (postcard and explanatory leaflet available).  The leaflet says that a 'rabbit' was a tool used in tin-mining.  On a later visit, I found also a pew seat cover with four examples of the pattern on it.  James Dalgety tells me that there was a gift shop called Tinners' Rabbits in Chagford some years ago.  [Mary Gray; Devon's Churches; James Pike Ltd, St. Ives, 1974, pp. 7 & 15] only mentions the pattern at Chagford, describing it as 'the old tinners' design'. 

                    The local wine shop, called Best Cellars, has the pattern on its window.  On my 2001 visit, the owner said the shop had previously been the Tinners' Rabbits.  The Newsagent's in the main square sells a Chagford plate featuring the Tinners Rabbits and with the usual mythology on the back.  There is a new building, Stannary Place, with a modern crest of three rabbits over the doorway, on New Street going away from the Church. 

The 15C house of Cardinal Jouffroy at Luxeuil-les-Bains, Haute-Saône, Franche-Comté, about 40km NW of Belfort, has the pattern carved under a balcony.  [Greeves (2000); Greeves (2001).]

The church of St. Andrew, Sampford Courtenay, Devon, has two roof bosses in the Three Rabbits pattern.  A letter from D. P. Miles says these date from c1450.  Cf [Jenkins, 1999, pp. 137-138.]

St. Michael's Church, Spreyton, Devon, has a roof boss of the Three Rabbits pattern.  The sign at the entrance to the churchyard is a large and beautiful version of the pattern painted by Helen Powlesland, the nicest I have seen.  Thanks to Rev. John Withers for information.  Mrs. Powlesland informs me that the roof boss is 15C.  [Greeves (2000); Greeves (2001)] says the roof is dated 1451.  [Greeves (2001)] has a photo.  Spreyton has produced a Millennium cup showing the school and an inscription surrounded by small squares containing the three rabbits pattern in white on blue and blue on white -- sent by Helen Powlesland.  In 2003, they produced a cup for the Church with one side having the three rabbits pattern -- kindly sent by Helen Powlesland.  Cave, p. 211, says there are two bosses, one in the chancel and one in the nave, and that there is an inscription on the roof of the chancel dated c1450.

The Chapel of Saints Cyr and Julitta (now called St. Anne and St. Catherine), Cotehele House, Cornwall, was built in 1485-1489 and has an example of the pattern as a roof boss -- see [Greeves (2000); Greeves (2001)].  The boss is on the midline, one in from the E end of the Chapel.  A chandelier hangs from it.  It is unpainted and so dark that one really needs binoculars to make out the pattern.  One can get a better view from a squint off the South Room on the first floor.

In Scarborough, North Yorkshire, there is a c1350 building in Sandgate (or Sandside) on the seafront of South Bay called the King Richard III Inn because he is reputed to have stayed there.  On the ceiling of one of the upper rooms is a 'Three Rabbits' pattern, but this is in the landlord and landlady's rooms and the landlady was unwilling to let me see it.  Inquiry to the Scarborough Museums and Gallery Officer elicited a photo held by the Planning Department in which the pattern can just be discerned (though I can't see which way it is going) and the information that it is in 16C plasterwork apparently done by Italian workers.  [Greeves (2000)] describes this and notes that Richard III's wife's family (the Nevills) owned the manor of North Bovey in the early 16C.  Greeves, Andrew & Chapman have been to see and photograph it -- the pattern has been painted red.

[Greeves (2000)] reports two occurrences of plaster ceilings with the pattern in Devon.  A 16C example at Treasbeare Farm, near Clyst Honiton, and a mid-17C(?) example at Upcott Barton, Cheriton Fitzpaine.  [Greeves (2001)] says the Clyst Honiton example is 17C and gives a fine photo of it.

Baltrušaitis.  Op. cit. below.  Fig. 97 is a 1576 Dutch engraving of three rabbits.

In Throwleigh, Devon, a few miles from Chagford, there is a roof boss of the Tinners' Rabbits in the 16C north aisle of the Church of St. Mary the Virgin.  (Thanks to the Rector of Chagford and Throwleigh, P. Louis Baycock, for directing me to this site.)  Photo in Sale, below, p. 63.

[Greeves (2000)] has a picture of the example in the parish church, Tavistock, Devon, but only dated as medieval.

[Greeves (2000)] reports there is a fine stone lintel from Charmois-L'Orgueilleux (I can't locate this) in the Musée d'Art Ancien et Contemporain in Epinal, Vosges, Lorraine, dated as 16C, but Greeves thinks the carving may be more primitive than 16C.  He also reports two medieval stone roof bosses at Ingwiller, Bas-Rhin, Alsace, about 35km NW of Strasbourg,  and  Xertigny, Vosges, Lorraine, about 15km S of Epinal.  Greeves says there are other French examples: in the chapel of the Hotel de Cluny (= Cluny Museum), Paris ([Greeves (2001)] says this is 15C; I have a good photo);  at Vienne, Isère, Rhône-Alpes;  at St. Bonnet le Chateau. Loire, Rhône-Alpes;  and that potters at Soufflenheim, Bas-Rhin, Alsace, use the pattern in current production.

In 1997, an old trunk and crates in the Statens Museum for Kunst (National Gallery), Copenhagen were opened and found to contain over 20,000 prints which had been stored during re-organization in the 1830s.  One of these is a three rabbits engraving, very similar to the 1576 Dutch engraving shown in Baltrušaitis, below.  Here the rabbits are going clockwise.  This material is the basis of an exhibition continuing until 16 Feb 2003.  This picture is being used to advertise the exhibition and appears in Copenhagen This Week for Jan 2003 with a short English text, on  www.ctw.dk/Sider/Articles.html .  The museum's site is  www.smk.dk , but I cannot see an English catalogue available.  Information and photo from Diana Hall and John Winterbottom -- John's son saw it in Copenhagen on a stopover at the airport.

The Church of St. John the Baptist, Broadclyst, Devon, has nine roof bosses of the Three Rabbits pattern.  (Greeves (1991) erroneously has eight.)  They were made in 1833 but are said to be careful plaster copies of the medieval examples, but my sources give no date for the originals.  There is one three rabbits boss in the central aisle -- Chris Chapman thinks some of the bosses in the central aisle may be originals, but this one has been recently painted and doesn't look old to me.

[Greeves (2000); Greeves (2001)] mentions two early 17C plaster ceilings with the pattern at Burg Breuberg in the Odenwald (I can't locate this on my maps, but the Odenwald is where Hessen, Baden-Württemberg and Bayern meet)  and  at Seligenstadt, Hessen, about 20km SE of Frankfurt. 

 

Basil Valentine.  (He may be catalogued as Basilius Valentinus (or Valentis) and entered under  B  rather than  V.)  De Macrocosmo, oder von der grossen Heimlichkeit der Welt und ihrer Artzney, dem Menschen zugehorig.  c1600.  ??NYS -- reproduced and briefly discussed in Greeves.  This shows the Hunt of Venus, with three hares going clockwise, but with each hare pursued by an unconnected dog and Greeves notes that the dogs are an essential part of a hunt symbol.  Inside the triangle of ears is the astrological/alchemical symbol for Mercury, which is claimed to be similar to a symbol used for tin (or Jupiter) -- but I have now looked at a book on alchemy and there is no similarity; further both symbols are included in the surrounding text and they are clearly distinct.  The pattern is drawn inside a circle of text which contains the astrological symbols of all seven planets and hence is rather cryptic -- see Roob, below, for the text.  The top of the picture has a flaming heart pierced by an arrow and the legend VENUS.  So there is no real connection with tin, though this is the source cited by the first book on Dartmoor to mention the symbol, in 1856.  Greeves notes that the three ears produce a delta-shape and this has connotations of fertility, both as the Nile delta and as the female pubic triangle.

                    Valentine is a (semi-?) mythical character.  He was allegedly a Benedictine monk of the early 15C, but no trace of his writings occurs before the 17C.  (However, de Rola (below) asserts that Antonius Guainerius (d. 1440) praised Valentine and that Valentine himself says he was a Benedictine monk at the monastery of St. Peter in Erfurt.  But his real name is unknown and so he cannot be traced in the records of the monastery or at Erfurt.  De Rola quotes a 1675 report that Valentine was at St. Peter's in 1413.)  Legend says his works were discovered when a pillar in the Cathedral of Erfurt split open (a variant of a story often used to give works a spurious age) -- cf the next item.  Despite their uncertain origin, the works were well received and remained popular for about 200 years, with the pictures of his 'Die zwölff Schlüssel' (The Twelve Keys) being used to the present day.   He was probably an alias of Johann Thölde (fl. 1595-1625).  The work cited may occur in his Last Will and Testament (1st English ed of 1657) or in his Chymische Schriften (Gottfried Liebezeite, Hamburg, 1700).

Basilius Valentinus.  The Last Will and Testament of Basil Valentine, Monke of the Order of St. Bennet ... To which is added Two Treatises: ....  Never before Published in English.  Edward Brewster, London, 1671.  ??NYS -- seen in a bookdealer's catalogue, 2003.

Basilius Valentinus.  Letztes Testament / Darinnen die Geheime Bücher vom groffen Stein der uralten Weifen / und anderen verborgenen Geheimnüssen der Natur  Auss dem Original, so zu Erfurt in hohen Altar / unter einem Marmorsteineren Täflein gefunden worden / nachgeschrieben:  Und nunmehr auf vielfältiges Begehren / denen Filiis Doctrinæ zu gutem / neben angehengten XII. Schlüsseln / und in Kupffer gebrachten Figuren ie. deffen Innhalt nach der Vorrede zu sehen / zum vierdtenmahl ans Liecht gegehen / deme angehänget ein Tractätlein von der ALCHIMIE, Worinnen von derselben Usprung / Fortgang und besten Scriptoribus gehandelt / auff alle Einwürffe der Adversariorum geantwortet / und klar bewiesen wird / dass warhafftig durch die Alchimie der rechte Lapis Philosophorum als eine Universal Medicin Könne bereitet werden / von Georg Philips Nenter.  Johann Reinhold Dulssecker, Strasburg, 1712.  This has several parts with separate pagination.  About halfway through is a new book with TP starting:  Von dem Grossen Stein der Uhralten / Daran so viel tausend Meister anfangs der Welt hero gemacht haben.  Neben angehängten Tractätlein / derer Inhalt nach der Vorrede zu finden.  Den Filis doctrinæ zu gutem publicirt / und jetzo von neuen mit seinen zugehörigen Figuren in Kupffer and Leight gebracht.  Strassburg /  Im Jahr M. DCC. XI.  Part V. of this is: Von der grossen Heimligkeit der Welt / und ihrer Artzney / dem Mensche zugehörig -- from Greeves, it appears this is also called Macrocosmo.  On p. 140 is a smaller and simpler version of the three hares picture, with 'folio 222' on it.  This has the hares going anticlockwise, the only example of Basilius' picture with this feature that I have seen.  Further, the word 'recht' is missing from the text around the picture and has been written in.  Below the picture is a lengthy poem, starting "Ein Venus-Jagt ist angestallt".  I translate the first part as:  "A Hunt of Venus is prepared.  The hound catches, so the hare will not now grow old.  I say this truly that Mercury will protect well when Venus begins to rage, so there occur fearfully many hares.  Then Mars guards with your [sic, but must be  his] sword, so that Venus does not become a whore." 

 

                    Part II. of  Von dem Grossen Stein der Uhralten  is  Die zwölff Schlüssel ... and Der neundte Schlüssel is on p. 51, with 'folio 70' on it.  In the centre of the circular part of this is a pattern of three hearts with serpents growing out of them and biting the next heart.  Above this are a man and a woman, each nude and in a sitting position, with bottoms almost touching.  Baltrušaitis mentions this in his discussion of 'two heads, four bodies' pictures, but I don't feel this is really a picture of that type.  A different and less clear version of this picture in reproduced in the following.

                    Stanislas Klossowski de Rola.  The Golden Game  Alchemical Engravings of the Seventeenth Century.  Thames & Hudson, 1998.  This reproduces the entire Twelve Keys from:  Michael Maier; Tripus aureus; Lucas Jennis, Frankfurt, 1618, and gives some discussion.  P. 119 is the title page of the Valentine; p. 103 includes the Ninth Key; pp. 125-126 give explications of the Keys.

                    Roob (below), p. 678, from:  D. Stolcius v. Stolcenberg; Viridarium chymicum; Frankfurt, 1624. 

                    Adam McLean.  The Silent Language  The Symbols of Hermetic Philosophy  Exhibition in the Bibliotheca Philosophica Hermetica.  In de Pelikaan, Amsterdam, 1994, p. 47,  reproduced from:  Johann Grasshoff; Dyas chymica tripartita; Lucas Jennis, Frankfurt, 1625. 

                    Baltrušaitis, above, Fig. 103, but he cites:  B. Valentino;  Les Sept Clefs de Sagesse;  vers 1413;  via an 1891 book.

                    This three heart/serpent pattern is known as the Ouroboros or Ouroborus.  McLean explains that man has three hearts: physical, soul and spiritual.

 

Basilius Valentius.  Chymische Schriften.  Hamburg, 1717.  ??NYS -- reproduced in:  Alexander Roob; Alchemie & Mystik  Das Hermetische Museum; Taschen, Köln, et al., 1996, p. 676.  Same picture as in De Macrocosmo, with some descriptive text.  The circle of text is translated as: "Sol and Luna and Mars mit Jupiter jagen / Saturnus muss die Garne trage / Steltt Mercurius recht nach dem Wind / so wird Frau Venus Kind"  (Sun and Moon and Mars hunt with Jupiter / Saturn must carry the net (or decoy) / Place Mercury correctly according to the Wind / then Frau Venus captures a child)  and Roob says this is an alchemical description of the preparation of the 'Universal Medicine' from copper vitriol, which Basilius called the highest of all salts.  Externally it is green, but inside it is fiery red from its father Mars, an oily balsam.  He then gives lines 5-8 of the poem and says the hare is a known symbol for the fleetness of Mercury.

 

At Schwäbisch Hall, Baden-Württemberg, about 60km NE of Stuttgart, the pattern appears painted on the ceiling of a synagogue, from 1738/9 [Greeves (2000); Greeves (2001), with picture.]  Sol Golomb notes that the commandment against making graven images is interpreted as referring to solid images and excepts decorative patterns even if they use animals.

The town of Hasloch (am Main), Bayern, about 70km SE of Frankfurt, uses the pattern as its town crest (colour photo of an 1842 example in [Greeves (2000)]). 

C. J. P. Cave.  The roof bosses in Chichester Cathedral.  Sussex Archaeology 71 (1930) 1-9 & plate opp. p. 1.  Photo 11 is a B&W picture, but not as sharp or clear as the current postcard.  Discussion on p. 6 says the Boxgrove example is 'more boldly and better carved'.  See Cave, 1948.

Jurgis Baltrušaitis.  Le Moyen Age Fantastique,....  Op. cit. under China, above.  Pp. 132‑139 of the 1981 edition have many examples of three and four rabbits, four boys, etc.  Fig. 96 gives small rabbit illustrations from Dunhuang (10th c.), Islamic vase from the Hermitage Museum (12th‑13th c.), Lyon Cathedral (1310‑20), Paderborn (15th c.), and Fig 97 is a large Dutch engraving, Lièvres a Oreilles Communes, of 1576 in the BN, Paris, with rabbits going anti-clockwise.  He cites examples at Saint-Maurice de Vienne (15C); at the Hôtel de Cluny, at Saint-Benoit-le-Château (Loire) and says it is frequent in the east of France (Thiélouse, Xertigny), Switzerland (Abbey of Muototal), and in Germany (Munster, Paderborn).  He cites a 1928 book for the pattern being a symbol of the Trinity.  By the beginning of the 16C, he says it was used as a vignette by the printer Jacques Arnollet and as the sign of the Three Rabbits Inn (L'Hostellerie aux trois lapins).  He gives some lines from a poem of the time:

                    Tournez et retournez et nous tournerons aussi,

                    Afin qu'a chacun de vous nous donnions du plaisir.

                    Et lorsque nous aurez tournés faites compte de nos oreilles,

                    C'est là que, sans rien déguiser, vous trouverez une merveille.

          and says these or similar occur frequently in prints of the 16C and 17C, referring to the large Dutch engraving which has similar lines in Dutch and French around the border - but his reproduction (or the original) is truncated on the left.  He cites an 1879 history of inn signs and seems to say this inn is in Lyon and that the author had also seen a version with three deer.

                    See Dead Dogs for further material.

C. J. P. Cave.  Roof Bosses in Medieval Churches  An Aspect of Gothic Sculpture.  CUP, 1948.  This has several hundred B&W plates.  I have a reference to Cave's collection of photographs of roof-bosses and an index thereof at the Society of Antiquaries, so I looked up Cave at the Warburg and found this book.

Chapter VI: Beasts, birds and fish, pp. 69-75.  On p. 71, he says:  "There are a few rabbits or hares, it is difficult to tell which; the most curious are the three rabbits with only three ears between them, each rabbit sharing an ear with its neighbours; this device is found at Broadclyst, North Bovey (49), Chagford, Sampford Courtenay (182), Spreyton, South Tawton, Tavistock, and Widecombe-in-the-Moor, all villages on or not far from Dartmoor; at Selby there is a similar arrangement, but there is a fourth rabbit unconnected with the three.1

                  1  The three rabbits occur in stained glass at Long Melford.  The same motif occurs in Paderborn Cathedral in Germany, on a window known as the Hare Window; see Walter Hotz, Mittelalterliche Groteskplastic (Leipsig), p. 47 (text) and p. 68 (plate); this window is even mentioned in Baedeker's Guide to North Germany."

Plate 1.  Tickencote.  This has three heads in a circle, but there is no interconnection.

Plate 30.  Bristol Cathedral, north transept: Fish.  This has three fish arranged in a triangular pattern, each overlapping the next.

Plate 49.  North Bovey; rabbits.

Plate 182.  Sampford Courtenay.  Rabbits.

Plate 200.  Selby.  Triple face.  [This has three faces looking forward, a bit to the right and a bit to the left, sharing four eyes.  I have a photo of a similar keystone in Citta di Castello where the side faces are facing directly right and left.  This seems to be a development from the Roman double-sided heads of Janus.]

Plate 300.  Foliate Heads.  Boxgrove.

Appendix I  List of churches containing roof bosses.  Pp. 181-222.  He begins with a note: This list does not aim at completeness, ....  Names of places which I have not visited personally are in square brackets.]  He lists 208 sites, including:  Bovey (North); Boxgrove Priory; Broadclyst ('evidently modern copies of old designs', but he doesn't say how many there are); Chagford ('two bosses ..., one much restored'); Cheriton Bishop; Chichester Cathedral (cites his: The roof bosses in Chichester Cathedral, Sussex Archaeological Collections, vol. LXXI, 1940, above); Havant ('Two early thirteenth-century bosses in the chancel with affinities to those as Chichester and Boxgrove.' ??); Oxford, Christ Church ('four lion's bodies joined to one head'); Portsmouth Cathedral ('Two bosses in the north aisle of the quire in the Chichester-Boxgrove style.'  ??); Portsmouth, SS John and Nicholas (Garrison Chapel) ('The first and third bosses are early conventional foliage in the Chichester-Boxgrove style.'  ??); Sampford Courtenay (cites two examples, in the chancel and the nave); Selby Abbey (says many bosses are medieval, but says nothing about the Triple face (plate 200)); Spreyton (cites two examples, in the chancel and the nave); Tavistock; Tawton (South); Widecombe-in-the-Moor.  The nine three rabbits examples were known, but Chichester and Boxgrove are new and there are a number of possibles that seem worth investigating, but Plates 1 and 30, now added above, are the only previously unnoted pictures of marginal interest.

Richard Hamann-MacLean.  Künstlerlaunen im Mittelalter.  IN: Friedrich Möbius & Ernst Schubert, eds.; Skulptur des Mittelalters; Hermann Böhlaus Nachfolger, Weimar, 1987, pp. 385-452.  The material of interest is on pp. 400-403.  He discusses and illustrates:  the Paderborn example, saying it is 15C;  the silver flask in the Hermitage;  the four rabbit pattern at Lyon.  He notes that the last cannot be considered a symbol of the Trinity!  He mentions a cave painting in Chinese Turkestan, 10C?, which is probably the Dunhuang example.  He gives a number of references to earlier articles -- ??NYS.

Tom Greeves.  The Tinners' Rabbits -- chasing hares?.  Dartmoor Magazine 25 (Winter 1991) 4-6.  This was certainly the most informative discussion of the topic until his later articles.  Greeves is an archaeologist and an authority on the Dartmoor tin industry.  He has now been joined in this research by Sue Andrew and the photographer Chris Chapman, comprising the Three Hares Project.  Greeves says that it is claimed to be the emblem of the medieval tinners, and various connections between tinners and rabbits have been adduced, e.g. it is claimed that the pattern was the medieval alchemical symbol for tin.  It is also called the Hunt of Venus and/or an emblem of the Trinity.  However, the earliest reference to the pattern on Dartmoor is an 1856 description of Widecombe Church which only says that the roof was connected with the tinners and that the pattern had an alchemical connection.  Later guides to Dartmoor are still pretty vague, e.g. a 1956 writer connects the symbol with copper, not tin.  It is not until 1965 that the symbol is specifically called The Tinners' Rabbits.  See his 2000 article for more early references.

                    There is no particular Dartmoor mythology connected with rabbits, but there is much mythology of hares.  See the note at the beginning of this section about the introduction of rabbits to Britain.  Three unconnected rabbits do occur in some English crests.  Greeves reproduces and discusses the c1600 Valentine picture -- see above.

                    Since no list of occurrences of the pattern had ever been compiled, Greeves examined almost all the churches in the area and discovered roof bosses with the pattern in 12 churches -- see [Greeves (2000)] for a longer list).  These are all on the east side of Dartmoor or to the north, except Tavistock is on the west side and Broadclyst is some 20 mi further east.  Bridford, Iddesleigh, Sampford Courtenay and Spreyton have no significant tin-mining connections.  No examples are known from the much more important tin-mining area of Cornwall, but Greeves has since found an example at Cotehele, just over the border into Cornwall.

                    Greeves then discusses the examples at Long Melford and Paderborn, giving comments which are mentioned above.  He then briefly describes the Dunhuang example, citing Whitfield & Farrar, and then the St. Petersburg example.  He then notes some modern versions: a wooden teapot stand from Scandinavia and a Victorian(?) carving in Holy Trinity, Fareham, Hampshire.

                    Early Christianity took over many pagan symbols and the three hares or rabbits (like the Green Men) could have been adapted to represent the Trinity.

                    In a letter of 3 Jun 1997, Greeves says he has located further examples of the three rabbits in Cheriton Bishop and Paignton in Devon, Cotehele in Cornwall and in Wales, Scarborough (North Yorkshire), France, Germany, Switzerland, Bohemia and modern China, where the pattern is still woven into carpets.  See [Greeves (2000); Greeves (2001)] for more details of these.

                    See below for the continuation articles [Greeves (2000); Greeves (2001)].

Paul Hambling.  The Dartmoor Stannaries.  Orchard Publications, Newton Abbott, 1995, pp. 38-39.  This gives a short summary of Greeves' work.  He adds that a story is that the tinners adopted the rabbit as their emblem in allusion to their common underground mode of life.  Tinners are also said to have been responsible for some rabbit warrens, but there were lots of other warrens and they would have been too common to be specifically associated with the tinners.  He notes that the symbol of three intertwined fishes was a common Christian symbol.

Simon Jenkins.  England's Thousand Best Churches.  Allen Lane, 1999;  slightly revised, Penguin, 2000.  This mentions the Three Rabbits pattern in several churches. 

Boxgrove, pp. 686-687.  "The second boss from the altar is so crafted that each of eight faces comes complete with two eyes, yet there are only eight eyes in all."

Long Melford, pp. 658-659.  "... three rabbits sharing three ears, representing the Trinity."

Sampford Courtenay, pp. 136-137.

Widecombe in the Moor, pp. 142.  "This rare symbol of the Trinity is formed of three animal heads sharing just three ears."

Richard Sale.  Dartmoor  The Official National Park Guide.  Photos by Chris Chapman.  Pevensey Guides (David & Charles), 2000.  Pp. 63-64 says the symbol of 'three rabbits each with only one ear' was adopted by the tinners and may be an allusion to the Holy Trinity.  He also says the rabbits became a pest and myxomatosis was introduced in 1954 and rabbit warrening was abolished in 1956.  P. 63 has a photo of the boss at Throwleigh.

Tom Greeves.  Chasing three hares.  Dartmoor Magazine No. 61 (Winter 2000) 8-10.  P. 11 contains several advertisements for hotels and tours featuring the three rabbits motif.  This reports on information discovered since his previous article -- see above.  Many bits of information are incorporated above, citing this as Greeves (2000).

                    The earliest known connection of the pattern with the tinners is given in the Torquay Directory (25 Nov 1925), but the first popular usage seems to be Sylvia Sayers' The Outline of Dartmoor's Story (1925), p. 24.

                    He now has found 17 Devon churches with 28 examples of the pattern:  Ashreigney, Bridford, Broadclyst (9 examples from 1833 said to be careful copies of medieval bosses -- Greeves (1991) erroneously has 8), Chagford (2 examples), Cheriton Bishop, Iddesleigh, Ilsington, Kelly (2 examples, one a modern copy), Newton St. Cyres, North Bovey, Paignton, Sampford Courtenay (2 examples), South Tawton, Spreyton (2 examples), Tavistock, Throwleigh, Widecombe(-in-the-Moor).  Excepting the recent copies, there are 19 medieval (i.e. pre-1500) examples, all wooden roof bosses.  These stretch well beyond the Dartmoor area into mid and east Devon.  Many of these sites have no significant tin-mining connections.  No examples are known from the much more important tin-mining area of Cornwall, except an example at Cotehele, Cornwall, just over the Cornish border.  There are also examples, probably 15C, on the roof of the Lady Chapel in St. David's Cathedral, St. David's, Pembrokeshire,  and  on an arch of St. Aidan's church, near Llawhaden, Dyfed.

                    Greeves writes that his group are organising an exhibition on the pattern at the High Moorland Visitor Centre, Princetown, Devon, for several weeks from 22 Nov 2001.

Tom Greeves.  Three hares -- a Medieval Mongol Mystery.  Devon Today (Apr 2001) 58-63.  Many bits of information are incorporated above, citing this as Greeves (2001).  Notes that Easter is believed to derive from the festival of the pagan goddess Eastre, whose familiar spirit was the hare.  Reports a possibly 15C example at Corfe Mullen, Dorset.  Says the pattern was used in 18C Slovakian pottery.  Gives photos of examples at Paignton and South Tawton and of a modern carpet from Urumqi.

 

                    MODERN VERSIONS OF THE THREE RABBITS PUZZLE

 

Child.  Girl's Own Book.  Puzzle 10.  1833: 163;  1839: 143;  1842: 264;  1876: 221.  "Can you draw three rabbits, so that they will have but three ears between them; yet each will appear to have the two that belongs to it?"  (1839, 1842 and 1876 have  belong  instead of  belongs.)

Magician's Own Book.  1857.  Prob. 7: The three rabbits, pp. 269 & 293.  "Draw three rabbits, so that each shall appear to have two ears, while, in fact, they have only three ears between them."  The drawing is similar to, but reasonably different than that in Girl's Own Book.

Book of 500 Puzzles.  1859.  Prob. 7: The three rabbits, pp. 83 & 107.  Identical to Magician's Own Book.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 3, pp. 395 & 436.  Identical to Magician's Own Book, prob. 7.

Boy's Own Conjuring Book.  1860.  Prob. 6: The three rabbits, pp. 230 & 255.  Identical to Magician's Own Book.

The Drei Hasen hotel, Michelstadt, Hessen, about 40km NE of Heidelberg, uses 19C and later versions of the pattern.  [Greeves (2000); Greeves (2001) has a photo of a 19C stained glass window.]

Hanky Panky.  1872.  P. 87: The one-eared hares.  Very similar to Magician's Own Book.

Wehman.  New Book of 200 Puzzles.  1908.  The three rabbits, p. 21.  c= Magician's Own Book.

Collins.  Book of Puzzles.  1927.  The Manx rabbit puzzle, p. 153.  Says it was invented by a Manxman.  Shows three rabbits, each with two ears, and one has to assemble them to have just three ears.

The Warren House Inn is at one of the highest passes over Dartmoor, Devon, on the B3212 about halfway between Princetown and Moretonhampstead.  The pub sign shows the Three Rabbits and they sell a polo shirt with the pattern.  [Thanks to Tom Greeves for directing me to this.]

Marjorie Newman.  The Christmas Puzzle Book.  Hippo (Scholastic Publications, London, 1990.  Kangaroos' ears, pp. 69 & 126.  Like Magician's Own Book, prob. 7, but with kangaroos.

The Castle Inn, Lydford, Devon, has a fine stained glass window of the Three Rabbits by James Paterson (1915-1986).  ??NYS -- described and illustrated in colour in Greeves (2000).

Jan Misspent (??sp).  Design Sources for Symbolism.  Batsford, 1993, p. 18.  Shows the three rabbits, going anticlockwise, among other examples of three-fold rotational symmetry.  Sent by Diana Hall.

Tom Greeves (see above) uses a three rabbits logo as his letterhead.

Holy Trinity Church, Long Melford, Suffolk, uses a version of their stained glass as a letterhead.

Trinity Construction Services, London and Essex, uses a three rabbits logo as their letterhead.  A former director saw the pattern in Devon and liked it.

Laurie Brokenshire reports that the chaplain at HMS Raleigh, the naval training station near Plymouth, has a vestment with the three rabbits emblem.  He saw it many years ago in the area and thought it was an excellent symbol of the Trinity and had the vestment made, probably for use on Trinity Sunday.

Martin and Philip Webb run a company called Fine Stone Miniatures (www.finestoneminiatures.com) which makes miniatures of medieval beasties from cathedrals, etc.  They have recently introduced two versions of the three rabbits.

 

                    DEAD DOGS

 

G. Yazdani.  Ajanta  Monochrome Reproductions of the Ajanta Frescoes Based on Photography  Part I.  1930; reprinted by Swati Publications, Delhi, 1983.  Photocopies sent by Peter Rasmussen.  P. 3 has the following.  "A good example of the artistic fancy of the sculptors of Ajanta is the delineation of four deer on the capital of a column in this cave (Plate XLb).  They have been so carved that the one head serves for the body of any of the four.  The poses of the bodies are most graceful and absolutely realistic, showing close study of nature combined with high artistic skill."  He dates the cave to the end of the 5C.  Footnote 1 says the motif of the four deer also occurs in a cave at Ghatotkach, which have an inscription dating them to the end of the 5C.  Plate IV is a general view of Cave I, but I cannot recognise the image in it.  Plate XLb is "Four deer with a common head" but very unclear.

Narayan Sanyal.  Immortal Ajanta.  Hrishikes Barik, Calcutta, 1984.  Photocopies sent by Peter Rasmussen.  P. 18, fig. 4.1, is a Plan of Cave I with Exhibits.  P. 19 has a List of Exhibits in Cave I -- entry 17 is "Four deer with one head" and notes that it is shown in Yazdani.  On the plan, 17 points to the two middle columns on the right side.  P. 38 describes this: "On another [pillar-capital] there are four deer with a common head.  The local guide would often invite the attention of the tourist to this and similar freaks.  But the magnanimity of Ajanta lies not in such frivolities ...."

Benoy K. Behl, text and photographs.  The Ajanta Caves  Artistic Wonder of Ancient Buddhist India.  Abrams, NY, 1998, 255pp.  [= Benoy K. Behl, text and photographs.  The Ajanta Caves: Ancient Paintings of Buddhist India.  Thames and Hudson, London, 1998, 256pp.]  P. 18 has a colour photo of a relief  of four deer sharing a single head.  Peter Rasmussen sent a B&W copy, which is not good, and then an enlarged colour picture which does show the effect, but it is still not very good.  Behl says " Inside the main hall on the right-hand side one of the capitals is subtly carved to create the illusion of four recumbent and standing deer sharing a single head."  This is apparently the only example at Ajanta as other authors refer to it as 'artistic fancy', 'freaks', 'frivolities'.  The Ajanta Caves date from -2C to 6C.

Carl Schuster  &  Edmund Carpenter.  Patterns That Connect  Social Symbolism in Ancient & Tribal Art.  Abrams, NY, 1996.  P. 34, fig. 68, shows a painted pottery pattern from Panama with two heads and four legs, but rather more like Siamese twins than our 'two heads, four boys' pattern.

The Peterborough Psalter, Brussels.  c1310.  At the bottom of f. 48v, Psalm 68, is a somewhat crude example of 'two heads, four horses', where the two vertical horses have their front and back hooves touching so they are very curved while the horizontal horses have bellies on the ground.  B&W in:  Lucy Freeman Sandler; The Peterborough Psalter in Brussels and Other Fenland Manuscripts; Harvey Miller, London, 1974, plate 45, p. 28 (see p. 9 for the date of the MS).  Also in Baltrušaitis. below.

Anna Roes.  "Tierwirbel" in JPEK: Jahrbuch für prähistorische und ethnographische Kunst: Jahrgang 1936/37.  De Gruyter, Berlin, 1937, 182 pp.  ??NYS -- described by Peter Rasmussen [email of 8 Jan 2002].  Pp. 85‑105 gives a history of the animal wheel motif, with many illustrations, but none with three hares.

Jurgis Baltrušaitis.  Le Moyen Age Fantastique,....  Op. cit. under China, above.  Pp. 132‑139 of the 1981 edition have many examples of three and four rabbits, four boys, etc. 

                    He discusses the animal wheel motif where several animals share the same head.  This sort of image is rather more common than either the Three Rabbits or the Dead Dogs type of image and I have not tried to chronicle it.  But I will include a few early examples.  Baltrušaitis shows (Fig. 98) several examples of three fish with one head:  from Egypt (XVIII-XX dynasty, i.e. c‑1500);  by Villard de Honnecourt, c1235;  a pavement at Hérivaux, 13C;  an Arabised plate from Paterna, 13C-14C.  He cites examples in Italian and French ceramics. 

                    He then goes on to 'two heads, four animals' patterns, showing (Fig. 99) a version of 'two heads, four horses' by the Safavid artist Riza Abbasi, signed and dated 20 Oct 1616, though the authenticity of the signature and date have been disputed.  He notes that another version is in the Boston Museum of Fine Arts (cf below) and there are definite differences in the two versions. 

                    He then shows (Fig. 99) the example of 'two heads, four horses' in the Peterborough Psalter.  His note 160 cites J. van den Gheyn, op. cit., pl. xxi.  Considerable searching finds the citation as part of note 68: J. van den Gheyn; Le Psautier de Peterborough; Haarlem, nd, ??NYS.  There are a number of Peterborough Psalters, including one in the Society of Antiquaries in London, one now MS. 12 in the Fitzwilliam Museum, Cambridge, and the famous example in Brussels (which seems to be the one studied by van den Gheyn).

                    Baltrušaitis then shows (Fig. 100) an example of 'two heads, four men' from the Library Portal of the Cathedral of Rouen, 1290-1300.  This is a relief, nicely enclosed in a quatrefoil frame.  In his text, the author says the pattern also occurs on a misericord in the Cathedral and on the Palace of Justice in Rouen.  These are now the earliest known examples of this idea.  Baltrušaitis cites an expert on Iranian art who says the tradition is very old in Iran and could well have inspired these examples.  Baltrušaitis says the pattern also occurs:  in the Church at Rosny, Aube (15C);  on a cathedral stall at Vendôme;  at New College, Oxford. 

                    He then shows the Two Apes on Horseback (Fig. 101), apparently the form 1985n of Schreiber.  He says several engravings of the 15C show these musical apes with the movable centre and his note cites similar examples with 'amours' and men by Cornelius Reen (1560) and Adrien Herbert (1576) and very frequently in the 17C.  He then has a photo (Fig. 102) of a oriental scrollweight, and says it undoubtedly comes from the same Islamic source as the other compositions with interchangeable elements.  He then shows the Ninth Key of Basil Valentine, cf under Medieval Europe, above, but I don't feel it is a picture of the present type.  He then reproduces (Fig. 104) Reen's picture, Symbol d'un Amour Inconstant, from the BN, dated as c1561, with two cherubs.  And then he gives (Fig. 105) a 'two heads, four men' picture with a drinking man and a harlequin horn-player.  This is another 1576 Dutch engraving in the BN, in a circular frame with French and Dutch text.

 

Zwei Affen als Kunstreiter (Verwandlungsbild).  This has two apes on a horse with a small bit of paper on a pivot which shifts the mid-body connections so either head is connected to either legs.  Two versions described in W. L. Schreiber; Handbuch der Holz- und Metallschnitte des XV.Jahrhunderts.  Band IV: Holzschnitte, Nr. 1783-2047.  Verlag Karl W. Hiersemann, Leipzig, 1927.  P. 120.  Schreiber pasted up examples of all the woodcuts and metalcuts described in his Handbuch in four mammoth volumes.  This unique set was presented to the Warburg Institute, where they are RR 240 1-4.  The following plates are in plate vol. 3.  Schreiber describes coloured examples of each version, but he only has uncoloured examples in his plate volume.

                    1985m.  This was supposedly found in Ulm when a small church was demolished.  Schreiber says the painting points to a Swiss origin and guesses a date of 1460-1480.  His example shows the rotating piece in one of its positions.  He describes a coloured example in the Germanisches Museum, Nürnberg.

                              [This picture is used in: Jasia Reichardt, ed.; Play Orbit [catalogue of an exhibition at the ICA, London, and elsewhere in 1969‑1970]; Studio International, 1969, p. 45.  It is described as "Paper toy from an Ulm woodcut, 1470."  No further details are given in Reichardt and she tells me that it was found by a research assistant, probably at the V&A.  However, it seems likely it was found at the Warburg, though it is possible that the V&A also has an example of the same print.]

                              [Ray Bathke [email of 20 Aug 1998] says the Ulm woodcut, 1470, appears in: Karl Gröber; Children's Toys of Bygone Ages; Batsford, 1928, 1932.  ??NYS.]

                    1985n.  Basically the same picture, but elaborated and with much more decorative detail, clearly added to an earlier version (an extra column lacks a base), again probably from Switzerland, but Schreiber makes no estimate of a date.  His copy lacks the rotating piece and shows no indication of its existence.  He describes a coloured version at Zürich Zentralbibliothek.  Baltrušaitis, above, gives a version as Fig. 100.

Wolfgang Brückner.  Populäre Druckgraphik Europas  Deutschland  Vom 15. bis zum 20.Jahrhundert.  (As: Stampe Popolari Tedesche; Verlag Electra, Milan, 1969);  Verlag Georg D. W. Callwey, München, 1969.  Pp. 24-25 (Abb. 17 is on p. 25), 203.  Coloured example of 1985n from Zürich Zentralbibliothek, described as Swiss, 1460/80.  Cites Schreiber.  He says there is a replica at Nürnberg, but this must be confusing it with 1985m.  This example is slightly different than Schreiber's picture in that the circle where the rotating piece would rotate does show, but the print is pasted to another sheet and one back line of an ape has been drawn in on the backing sheet.  I have a slide.

Thomas Eser.  Schiefe Bilder  Die Zimmernsche Anamorphose und andere Augenspiele aus den Sammlungen des Germanischen Nationalmuseums [catalogue of an exhibition in 1998].  Germanisches Museum, Nürnberg, 1998, pp. 86-87.  The picture is a B&W version of a coloured example of 1985m, from the Museum Graphische Sammlung, H5690, Kapsel 8.  This is probably the same example described by Schreiber, but is described as a coloured woodcut, Swiss or Schwabian, 1460/70 and probably the oldest surviving example of a picture which the viewer can change.  He cites Schreiber and Brückner.

Barbara Maria Stafford &  Frances Terpak.  Catalog for the exhibition Devices of Wonder, at the Getty Museum, early 2002.  ??NYS - information sent by William Poundstone.  This shows (in colour) and discusses an oil painting of Hermes and Aphrodite, like the above apes image, with a rotating piece in the middle which covers one of the pairs of waists.  It dates from early 17C Bohemia, probably the school of Prague some time after Rudolf II's death in 1612.

James C. Y. Watt  &  Anne E. Wardwell.  When Silk Was Gold  Central Asian and Chinese Textiles.  The Metropolitan Museum of Art in Cooperation with The Cleveland Museum of Art, dist. by Abrams.  Section 45: Cloth of gold with rabbit wheels, p. 158.  After discussing the four rabbits textile (see above under Other Asia), this goes on to discuss the 'two heads, four boys' motif, with Figure 73 showing a Ming example in the Shanghai Museum dated to 16‑17 C.  These were often used as toggles.  This says that the pattern is called 'two boys make four images' and is a rebus for part of a famous line from the Yi-ching:  "The primal 'one' [taiji] begets the two opposites [yi], the two opposites beget the four elements [xiang], and the four elements beget the eight trigrams [gua]."  The word yi is a homonym for the word for 'boy' in some pronunciations and xiang means both 'element' and 'image'.

Rza (or Riza) Abbasi (1587‑1628).  Drawing:  Four Horses.  Drawing given in:  Kh. S. Mamedov; Crystallographic Patterns.  Comp. & Maths. with Appls. 12B:1/2 (1986) [= I. Hargittai, ed., Symmetry -- Unifying Human Understanding, as noted in 6.G.] 511‑529, esp. 525‑526.  Two heads and four bodies.  This seems to be an outline made from the original, probably by Mamedov?  See below for a possible original version.  Cf Baltrušaitis, above.

Early 17C Persian drawing:  Four Horses: Concentric Design.  Museum of Fine Arts, Boston.  Reproduced in:  Gyorgy Kepes; The New Landscape in Art and Science; Paul Theobold, Chicago, 1956, fig. 44 on p. 53 with caption on p. 52;  and in:  S&B, p. 34.  This picture and the drawing above differ in the position of the feet and other small details, so it is not clear if the above has been copied from this picture.  Seckel, 1997 (op. cit. in 6.AJ), reproduces it as 6© and says it is 18C.  Baltrušaitis (Fig. 99), above, has a similar, but different version.

Itsuo Sakane.  A Museum of Fun (in Japanese).  Asahi Shimbun, Tokyo, 1977.  Chap. 54-55, pp. 201-207.  Seven examples, but I haven't had the text translated.

Itsuo Sakane et al.  The Expanding Visual World -- A Museum of Fun.  Catalogue of a travelling exhibition with some texts in English.  Asahi Shimbun, Tokyo, 1979.  Section IV: Visual Games, no. 12-15, pp. 96-99, with short texts on p. 170.  Seven examples, mostly the same as in the previous book, including the following.  I give the page numbers of the previous book in ( ) when the picture is in both books.

IV-12, p. 96 (p. 205), both B&W.  Sadakage Gokotei.  Five Children, Ten Children.  Edo era (= 1603-1867).  Seckel, 1997 (op. cit. in 6.AJ), reproduces it in colour as 5© with the same data and Seckel, 2000 (op. cit. in 6.AJ), reproduces it in colour as fig. 44, p. 55 (= 2002a, fig, 44, p. 55), with no data.  However, the same picture is reproduced in colour in Julian Rothenstein & Mel Gooding; The Paradox Box; Redstone Press, London, 1993; with a caption by James Dalgety, saying that it is a painting by Yamamoto Hisabei, c1835, based on an earlier Chinese image, and giving the title as Ten Children with Five Heads.

Unnumbered, p. 97 (pp. 202-203, figs. 2, 3, 4), both B&W, but the titles have no English versions.  Though no credit is given, the top item is is the early 17C Persian drawing in the Museum of Fine Arts, Boston. 

           The second item seems to be from Renaissance Europe and can be called 'three heads, seven children'.  There are three heads and upper bodies alternating with three legs and lower bodies in a flattened hexagonal pattern so the the top head can also connect to the bottom legs giving a third arrangement of the pieces, though the other two children occur in the previous arrangements, so one gets a total of seven children rather than nine.  I have only recently discovered this is a European item and Bill Kalush has an example of a lead medallion with the same pattern which is dated to c1610 Prague -- I have a photo.  Edward Hordern's collection has a wooden box with this pattern on the cover, dated as 16C, but it looks later to me, though I have only seen photos.

           The third item is a version of the 'five heads, ten chldren' picture described above.  However, the references to this chapter mention Lietzmann and I have found it there, where it is stated to be a Japanese matchbox -- see below.

IV-14, p. 99 (colour) (p. 207, B&W).  Kuniyoshi Ichiyusai [= Utagawa Kuniyoshi].  Stop Yawning.  Late Edo era [mid 19C]. 

IV-15, p. 98 (colour).  Anon.  Four Heads, Twelve Horses.  Probably Persian.  This item belongs to Martin Gardner, having been left to him by M. C. Escher.  It seems to be a leather cushion cover, probably Persian.  Seckel, 1997 (op. cit. in 6.AJ), reproduces it as the 6ª and says it is 18C.  Seckel, 2000 (op. cit. in 6.AJ), reproduces it as fig. 88, pp. 99 & 122 (= 2002b, fig. 86, pp. 97 & 120) and says sometime in the 17C. 

Metal scrollweights of the 'two heads, four children' pattern have been made in China since at least the 17C.  I have three modern examples which Peter Hajek obtained for me in Hong Kong.  Edward Hordern's collection has a version from c1680 and a porcelain version, about six or eight inches across, among other examples -- I have the date somewhere.  James Dalgety also has an example of the porcelain version.

"Three Boys -- Nine Torsos".  Anonymous painting on silk from 1700-1710 in City Palace Museum, Jaipur, India.  Edward Hordern's collection has a modern replica.  This is similar to the Three Heads, Seven Children version mentioned above.  I have photos from Hordern's replica and a copy of his information sheet on it and another painting.  Reproduced from Hordern's example in: Rothenstein & Gooding, below, p. 16.

                    Mohammad Bagheri has sent me some souvenir material from the Museum, now named the Maharaja Sawai Man Singh II Museum, City Palace, Jaipur.  This shows a different version of the pattern, with the same basic geometry but very different figures and colours. 

Family Friend 1 (1849) 148 & 178.  Practical puzzles -- No. 3 -- Dead or alive?  "These dogs are dead you well may say:--   Add four lines more, they'll run away!"  Answer has  "See now the four lines.  "Tally-ho!"  We've touch'd the dogs, and away they go!"

Julian Rothenstein & Mel Gooding.  The Playful Eye.  Redstone Press, London, 1999.  They include a number of Japanese prints.  There are brief notes on p. 100.

P. 16.  "Three Boys -- Nine Torsos".  Photo from Hordern's facsimile, see above.

P. 20.  Three heads, six bodies.  Woodblock print, 1860s.

P. 20.  Five heads, ten bodies, similar to the version given by Sakane, above, but with different costumes.  Woodblock print, 1860s.

P. 21.  Two or three bodies, one head.  Woodblock print, c1855.  This has six groups of bodies where one head or hat can be associated with two or three of the bodies.  There is a set of four upper bodies which can be associated with one pair of legs.

P. 22.  Five heads, ten bodies, almost identical to the version given by Sakane, above, but in colour with  MADE IN JAPAN  on the bottom edge.  The notes say it is a Japanese matchbox label, c1900.

P. 22.  Six heads, twelve bodies.  Two rings of three heads, six bodies, with elephants.  Japanese matchbox label, c1900.

P. 24.  Five heads, ten bodies, similar to the version given by Sakane, above, but with women.  Woodblock print, c1850.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 5, p. 177 (1868: 188): Alive or Dead.  "These dogs are dead, perhaps you'll say;   Add four lines and then they'll run away."

Magician's Own Book.  1857.  Prob. 27: The dog puzzle, pp. 275 & 298.  "The dogs are, by placing two lines upon them, to be suddenly aroused to life and made to run.  Query, How and where should these lines be placed, and what should be the forms of them?"  S&B, p. 34.

Book of 500 Puzzles.  1859.  Prob. 27: The dog puzzle, pp. 89 & 112.  Identical to Magician's Own Book.

Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 33: Dead or Alive, pp. 401 & 441  "These dogs are dead you well may say:--   Add four lines more, they'll run away!"

Boy's Own Conjuring Book.  1860.  Prob. 26: The dog puzzle, pp. 237 & 261.  Identical to Magician's Own Book.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 583-7, p. 285.  Dead dogs.

Magician's Own Book (UK version).  1871.  The solution drawing is given at the bottom of p. 231, apparently to fill out the page as there is no relevant text anywhere.  The drawing is better than in the 1857 US book of the same name.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 6: The dog puzzle, pp. 28 & 31.  "By connecting the dogs with four lines only they will suddenly start into life, and commence running.  Where should the lines be placed?"  However, he omits to give a picture!

Hoffmann.  1893.  Chap. 10, no. 32: The two dogs, pp. 348 & 387 = Hoffmann-Hordern, pp. 244-245.  No poetry, but the solution notes that you have to view the dogs sideways.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 9:5 (Mar 1903) 490-491.  Trick donkeys.  "Here are two apparently very dead donkeys.  To bring them to life it is only necessary to fill in the dotted lines and then turn the page half way round."

Benson.  1904.  The dead dogs puzzle, pp. 256‑257.  Prose version.

Pearson.  1907.  Part III, no. 83: Rousing dead dogs -- A good old puzzle, p. 83 .  "These dogs are dead, we all should say;   Give them four strokes, they run away."

Wehman.  New Book of 200 Puzzles.  1908.  The dog puzzle, p. 22.  c= Magicians Own Book.

W. Lietzmann.  Lustiges und Merkwürdiges von Zahlen und Formen.  1922.  I can't find it in the 2nd ed. of 1923.  4th ed, F. Hirt, Breslau, 1930, p. 208, unnumbered figure, shows the 'five heads, ten children' pattern mentioned as the third item on p. 97 of the Sakane book above, labelled:  Japanese Matchbox   How many people, how many heads, how many legs, how many arms are in this picture?

                    The material is also in the 6th ed., 1943, p. 200, fig. 46;  7th & 10th eds., 1950 & 1969, p. 196, fig. 37.

Collins.  Book of Puzzles.  1927.  The dead dogs puzzle, p. 152.

A two bodied woman.  In Seckel, 2002a, op. cit. in 6.AJ, fig. 2, pp. 11 & 44 (= 2002b, fig. 144, pp. 161 & 194).  A real photo of Lady Bird Johnson greeting a woman friend which shows just one head on two embracing bodies.

 

                    TRICK MULES

 

Loyd.  P. T. Barnum's Trick Mules.  1871.  Loyd registered this in 1871 and sold it to Barnum shortly thereafter.  Barnum used it in his Advance Courier.  See S&B, p. 34, for an illustration of Barnum's version and two recent versions.  See SLAHP: Out for a gallop, pp. 65 & 110.  See Gardner, SA (Aug 1957)  c= 1st Book, chap. 9.  In a 1907 interview, it was stated that thousands of millions of copies of the puzzle had been printed, with Loyd taking orders for a million at a time!

Gaston Tissandier.  Jeux et Jouets du jeune age  Choix de récréations  amusantes & instructives.  Ill. by Albert Tissandier.  G. Masson, Paris, nd [c1890].  Le mulet rigolo, pp. 36-37, with elegant coloured plate.  No reference to its history.

Mel Stover.  1980s??  Trick zebras puzzle.  This has two identical cards with two zebras and two riders.  The instructions say to cut one card into three parts along the dotted lines and put the riders on the zebras.  However, one zebra is facing the opposite way to the usual case and it takes some time to realise how to solve the problem.

Seckel, 2000 (op. cit. in 6.AJ), gives a nice colour version as fig. 105, pp. 116 & 122 (= 2002b, fig. 103, pp. 114 & 120), but only says it is due to Loyd.

 

          6.AV.           CUTTING UP IN FEWEST CUTS

 

Mittenzwey.  1880.  Prob. 191, pp. 36 & 88-89;  1895?: 216, pp. 40 & 91;  1917: 216, pp. 37 & 87.  Cut a  2 x 4  into eight unit squares with three cuts.  First cuts into two squares, then overlays them and then cuts both ways.

Perelman.  FFF.  1934.

Sectioning a cube.  Not in the 1957 ed.  1979: prob. 122, pp. 170-171 & 182.  MCBF: prob. 122, pp. 171-172 & 186.  How many cuts to cut a cube into 27 cubes? 

More sectioning.  Not in the 1957 ed.  1979: prob. 123, pp. 171 & 182-183.  MCBF: prob. 123, pp. 172-173 & 186.  How many cuts to cut a chessboard into 64 squares?

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. k, pp. 39, 189 & 191.  Cut a cube into 27 cubelets.

 

          6.AW.          DIVISION INTO CONGRUENT PIECES

 

          Polyomino versions occur in 6.F.4.

          Quadrisecting a square is 6.AR.

          See also:  6.AS.1, 6.AT.6.a, 6.AY, 6.BG.

          For solid problems, see:  6.G.3, 6.G.4, 6.AP, 6.AZ?, 6.BC.

 

See: Charades, Enigmas, and Riddles, 1862, in 6.AW.2 for a quadrisection with pieces not congruent to the original.

 

          6.AW.1.       MITRE PUZZLE

 

          Take a square and cut from two corners to the centre to leave  ¾  of the square.  The problem is to quadrisect this into four congruent parts.

 

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has an entry  "8½ a  Prob of figure"  followed by the  L‑tromino.  8½ b  is the same with a mitre and there are other dissection problems adjacent -- see 6.F.3, 6.F.4, 6.AQ, 6.AY -- so it seems clear that he knew this problem.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles.

No. 15, pp. 26 & 86 & plate II, fig. 11.  2 squares, one double the size of the other, to be cut into four pieces to make a mitre.  Just cut each along the diagonal.

No. 18, pp. 27 & 87 & plate II, fig. 14.  Six equal squares to form a mitre.  Cut each diagonally.  [Actually you only need to cut three of the squares.]

Endless Amusement II.  1826?  Prob. 5, pp. 192-193.  Mitre puzzle -- says the pieces are not precisely equal.  = New Sphinx, c1840, pp. 131-132.

Magician's Own Book.  1857.

Prob. 12: The quarto puzzle, pp. 269 & 294.  Solution is a bit crudely drawn, but the parts are numbered to make it clear how they are combined.  = Illustrated Boy's Own Treasury, 1860, No. 41, pp. 403 & 442.

Prob. 28: Puzzle of the two fathers, pp. 275-276 & 298.  One father has  L‑tromino (see 6.F.4), the other has the mitre.  Solution carefully drawn and shaded.  c= Landells, Boy's Own Toy-Maker, 1858, pp. 148-149.

Book of 500 Puzzles.  1859. 

Prob. 12: The quarto puzzle, pp. 83 & 108.  Identical to Magician's Own Book.

Prob. 28: Puzzle of the two fathers, pp. 89-90 & 112.  Identical to Magician's Own Book.. 

Boy's Own Conjuring Book.  1860.

Prob. 11: The quarto puzzle, pp. 231 & 257.  Identical with Magician's Own Book.

Prob. 27: Puzzle of the two fathers, pp. 237‑238 & 262.  Identical to Magician's Own Book.

Hanky Panky.  1872.  The one‑quarterless square, p. 132

Hoffmann.  1893.  Chap. X, no. 29: The mitre puzzle, pp. 347 & 386 = Hoffmann-Hordern, pp. 243 & 247.  Photo on p. 247 shows Enoch Morgan's Sons Sapolio Color-Puzzle.  This says to arrange the blocks 'in four equal parts so that each part will be the same size color and shape.'  It appears that the blocks are isosceles right triangles with legs equal to a quarter of the original square.  There are 6 blue triangles, 6 yellow triangles and 12 red triangles.  I think there were originally 6 red and 6 orange, but the colors have faded, and I think one wants each of the four parts having one color.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 5: The mitre puzzle.  Similar to Hoffmann.  No solution.

Loyd.  Origin of a famous puzzle -- No. 19: The mitre puzzle.  Tit‑Bits 31 (13 Feb  &  6 Mar 1897) 363  &  419.  Nearly 50 years ago someone told him to quadrisect  ¾  of a square into congruent figures.  The  L‑tromino was intended, but young Loyd drew the mitre shape instead.  He says it took him nearly a year to solve it.  But see Dudeney's comments below.

Clark.  Mental Nuts.  1904, no. 31.  Dividing the land.  Quadrisect an  L‑tromino and a mitre.

Pearson.  1907.  Part II, no. 87: Loyd's mitre puzzle, pp. 87 & 178.

Dudeney.  The world's best puzzles.  Op. cit. in 2.  1908.  He says he has traced it back to 1835 (Loyd was born in 1841) and that "strictly speaking, it is impossible of solution, but I will give the answer that is always presented, and that seems to satisfy most people."  See also the solution to AM, prob. 150, discussed in 6.AY.  Can anyone say what the 1835 source might be -- a version of Endless Amusement??

Wehman.  New Book of 200 Puzzles.  1908.

P. 43: Puzzle of the two fathers.  c= Magician's Own Book, with cruder solution.

P. 47: The quarto puzzle.  c= Magician's Own Book, without the numbering of parts.

Loyd.  Cyclopedia.  1914.  A tailor's problem, pp. 311 & 381.  Quadrisect half of a mitre.  This has a solution with each piece similar to to the half mitre.

Loyd Jr.  SLAHP.  1928.  Wrangling heirs, pp. 35 & 96.  Divide mitre into 8 congruent parts -- uses the pattern of Loyd Sr.

Putnam.  Puzzle Fun.  1978.  No. 106: Divide the shape, pp. 16 & 39.  "Divide the given shape into four pieces, such that each and every piece is the same area."  This is much easier than the usual version.  Put two half-size mitres on the bottom edge and two trapeziums are left.

 

          6.AW.2.       REP‑TILES

 

          Here one is cutting a shape into congruent pieces similar to the original shape.  Section 5.J is a version with similar but non‑congruent pieces.

 

Charades, Enigmas, and Riddles.  1862: prob. 34, pp. 137 & 143;  1865: prob. 578, pp. 109 & 156.  Divide isosceles right triangle and a hatchet-shaped 10-omino into four congruent pieces.  In the first case, they are congruent to the original, but in the second case, the pieces are right trapeziums of height  1  and bases  2  and  2½.

Dudeney.  The puzzle realm.  Cassell's Magazine ?? (May 1908) 713-716.  No. 3: An easy dissection puzzle.  Quadrisect a right trapezium in form of square plus half a square.  = AM, 1917, prob. 146, pp. 35 & 170.

C. Dudley Langford.  Note 1464:  Uses of a geometric puzzle.  MG 24 (No. 260) (Jul 1940) 209‑211.  Quadrisects:  L‑tromino,  P‑pentomino, right trapezium (trapezoid),  L‑tetromino, isosceles trapezium, another right trapezium, two squares joined at a corner.  Gives some 9‑sections and asks several questions, including asking about 3‑D versions.

R. Sibson.  Note 1485:  Comments on Note 1464.  MG 24 (No. 262) (Dec 1940) 343.  Says some of Langford's 4‑sections also give 9‑sections.  Mentions some 3‑D versions and 144‑sections.

Howard D. Grossman.  Fun with lattice points.  13.  A geometric puzzle.  SM 14 (1948) 157‑159.  Cites Langford & Sibson.  Gives two 9-sections obtained from Langford's 4‑sections as asserted by Sibson.  Gives alternative 4- & 9-sections.  Gives a method of generating infinitely many examples on both square and triangular lattices.

Gardner.  SA (May 1963) = Unexpected, chap. 19.  Says Golomb started considering rep‑tiles in 1962 and wrote three private reports on them (??NYS).  Gardner describes the ideas in them.

Solomon W. Golomb.  Replicating figures on the plane.  MG 48 (No. 366) (Dec 1964) 403‑412.  Cites Langford and adds numerous examples.  Defines  rep‑k  and shows all  k  can occur.

Roy O. Davis.  Note 3151:  Replicating boots.  MG 50 (No. 372) (May 1966) 175.

Rochell Wilson Meyer.  Mutession: a new tiling relationship among planar polygons.  MTg 56 (1971) 24‑27.  A  and  B  mutually tessellate if each tiles an enlargement of the other.

 

          6.AW.3.       DIVIDING A SQUARE INTO CONGRUENT PARTS

 

          In the 1960s, a common trick was to give someone a number of quadrisection problems where the parts happen to be congruent to the original figure -- e.g. the quadrisection of the square or the  L-tromino.  Then ask her to divide a square into five congruent parts.  She usually tries to use square pieces in some way and takes a long time to find the obvious answer.  c1980, Des MacHale told me that it was a serious question as to whether there was any non-trivial dissection of a square into five or even three congruent pieces.  Sometime later, I found a solution -- slice the square into  6  equal strips and say part  A  consists of the  1st and 4th strips,  part  B  is the 2nd and 5th,  part  C  is the 3rd and 6th.  However this is not what was intended by the problem though it leads to other interesting questions.  Since then I have heard that the problem has been 'solved' negatively several times on the backs of envelopes at conferences, but no proof seems to have ever appeared.  Very little seems to be published on this, so I give what little I know.  Much of this applies to rectangles as well as squares.  QARCH is an occasional problem sheet issued by the Archimedeans, the Cambridge (UK) student mathematics society. 

 

Gardner, in an article:  My ten favorite brainteasers in Games (collected in Games Big Book of Games, 1984, pp. 130-131) says the dissection of the square into five congruent parts is one of his favorite problems.  ??locate

David Singmaster.  Problem 12.  QARCH III (Aug 1980) 3.  Asks if the trisection of the square is unique.

David Singmaster.  Response to Problem 12 and Problem 21.  QARCH V (Jan 1981) 2 & 4.  Gives the trisection by using six strips and unconnected parts.  In general, we can have an  n-section by cutting the square into  kn  strips and grouping them regularly.  For  n = 2,  k = 4,  there is an irregular dissection by using the parts as strips  1, 4, 6, 7  and  2, 3, 5, 8.  If  p  is an odd prime, are there any irregular  p-sections?

John Smith, communicated in a letter from Richard Taylor, editor of QARCH, nd, early 1981?  Smith found that if you slice a square into 9 strips, then the following parts are congruent, giving an irregular trisection.  1, 2, 6;   3, 7, 8;   4, 5, 9.

David Singmaster.  Divisive difficulties.  Nature 310 (No. 5977) (9 Aug 1984) 521  &  (No. 5979) (23 Aug 1984) 710.  Poses a series of problems, leading to the trisections of the cube.  No solutions were received.

Angus Lavery asserts that he can trisect the cube, considered as a  3 x 3 x 3  array of indivisible unit cubes.  I had sought for this and was unable to find such a trisection and had thought it impossible and I still haven't been able to do it, but Angus swears it can be done, with one piece being the mirror image of the other two.

George E. Martin.  Polyominoes -- A Guide to Puzzles and Problems in Tiling.  MAA Spectrum Series, MAA, 1991.  Pp. 29-30.  Fig. 3.9 shows a  5 x 9  rectangle divided into 15  L-trominoes.  Shrinking the length  9  to  5  gives a dissection of the square into 15 congruent pieces which are shrunken  L-trominoes.  Prob. 3.10 (very hard) asks for a rectangle to be dissected into an odd number of congruent pieces which are neither rectangles nor shrunken  L-trominoes.  He doesn't give an explicit answer, but on p. 76 there are several rectangles filled with an odd number of  L,  P  and  Y  pentominoes.  One might argue that the  L  and  P  have the shape of some sort of shrunken  L-tromino, but the  Y-pentomino is certainly not.  Prob. 3.11 (unsolved) asks if a rectangle can be dissected into three congruent pieces which are not rectangles.  Prob. 3.12 is a technical generalization of this and hence is also unsolved.

On 19 Jun 1996, I proposed the trisection of the square and Lavery's problem on NOBNET.  Michael Reid demonstrated that Lavery's problem has no solution and someone said Lavery had only conjectured it.  Reid also cited the following two proofs that the square trisection is impossible.

I. N. Stewart & A. Wormstein.  Polyominoes of order 3 do not exist.  J. Combinatorial Theory A 61 (1992) 130-136.  ??NYS -- Reid says they show that if a rectangle is dissected into three congruent polyominoes, then each is a rectangle.

S. J. Maltby.  Trisecting a rectangle.  J. Combinatorial Theory A 66 (1994) 40-52.  ??NYS -- Reid says he proves the result of Stewart & Wormstein without assuming the pieces are polyominoes.

Martin Gardner.  Six challenging dissection tasks.  Quantum (May/Jun 1994).  Reprinted, with postscript, in Workout, chap. 16.  Trisecting the square into congruent parts is his first problem.  Cites Stewart & Wormstein.  Then asks if one can have three similar parts, with just two, or none, congruent.  Then asks the same questions about the equilateral triangle.  All except the first question are solved, but the solutions for the fifth and sixth are believed to be unique.  In the postscript, Gardner says Rodolfo Kurchan and Andy Liu independently suggested the problems with four parts and cites Maltby. 

 

          6.AW.4.       DIVIDING AN  L-TROMINO INTO CONGRUENT PARTS

 

          See also 6.F.4.

 

F. Göbel.  Problem 1771:  The  L‑shape dissection problem.  JRM 22:1 (1990) 64‑65.  The  L‑tromino can be dissected into  2, 3, or 4  congruent parts.  Can it be divided into 5 congruent parts?

                    Editorial comment -- The  L-shaped dissection problem.  JRM 23:1 (1991) 69-70.  Refers to Gardner.

                    Comments and partial solution by Michael Beeler.  JRM 24:1 (1992) 64-69.

Martin Gardner.  Tiling the bent tromino with  n  congruent shapes.  JRM 22:3 (1990) 185‑191.

 

          6.AX. THE PACKER'S SECRET

 

          This requires placing  12  unit discs snugly into a circular dish of radius  1 + 2 Ö3 = 4.464.

 

Tissandier.  Récréations Scientifiques.  5th ed., 1888, Le secret d'un emballeur, pp. 227-229.  Not in the 2nd ed. of 1881 nor the 3rd ed. of 1883.  Illustration by Poyet.  He shows the solution and how to get the pieces into that pattern.  No dimensions given.  = Popular Scientific Recreations; [c1890]; Supplement: The packer's secret, pp. 855‑856.

Hoffmann.  1893.  Chap. X, no. 48: The packer's secret, pp. 356 & 394 = Hoffmann-Hordern, p. 255.  Says the problem is of French origin.  Gives dimensions    and  ¾,  giving a ratio of  14/3 = 4.667.  "The whole are now securely wedged together ...."  [I think this would be a bit loose.]

"Toymaker".  The Japanese Tray and Blocks Puzzle.  Work, No. 1447 (9 Dec 1916) 168.  Says to make the dish of radius  7  and the discs of radius  1½,  again giving a ratio of  14/3 = 4.667.  Makes "a firm immovable job ...."

 

          6.AY. DISSECT  3A x 2B  TO MAKE  2A x 3B,  ETC.

 

          This is done by a 'staircase' cut.  See 6.AS.

 

Pacioli.  De Viribus.  c1500.  Ff. 189v - 191r.  Part 2.  LXXIX. Do(cumento). un tetragono saper lo longare con restregnerlo elargarlo con scortarlo (a tetragon knows lengthening and contraction, enlarging with shortening ??)  = Peirani 250-252.  Convert a  4 x 24  rectangle to a  3 x 32  using one cut into two pieces.  Pacioli's

          description is cryptic but seems to have two cuts, making                                                d        c

          three pieces.  There is a diagram at the bottom of f. 190v, badly                                  k         f  e

          redrawn on Peirani 458.  Below this is a inserted note which Peirani

          252 simply mentions as difficult to read, but can make sense.  The                                   g          

          points are as laid out at the right.  abcd  is the original  4 x 24                                      h   a    o  b

          rectangle.  g  is one unit up from  a  and  e  is one unit down from  c.

          Cut from  c  to  g  and from  e  parallel to the base, meeting  cg  at  f.  Then move  cdg  to  fkh  and move  fec  to  hag.  Careful rereading of Pacioli now seems to show he is using a trick!  He cuts from  e  to  f  to  g.  then turns over the upper piece and slides it along so that he can continue his cut from  g  to  h,  which is where  f  to  c  is now.  This gives three pieces from a single cut!  Pacioli clearly notes that the area is conserved.

                    Although not really in this topic, I have put it here as it seems to be a predecessor of this topic and of 6.P.2.

Cardan.  De Rerum Varietate.  1557, ??NYS.  = Opera Omnia, vol. III, p. 248 (misprinted 348 and with running head Lib. XII in the 1663 ed.).  Liber XIII.  Shows  2A x 3B  to  3A x 2B  and half of  3A x 4B  to  4A x 3B  and discusses the general process.

Kanchusen.  Wakoku Chiekurabe.  1727.  Pp. 11-12 & 26‑27.  4A x 3B  to  3A x 4B,  with the latter being square.  Solution asserts that any size of paper can be made into a square: 'fold lengthwise into an even number and fold the width into an odd number' -- cf Loyd (1914) & Dudeney (1926) below.

Minguet.  1733.  Pp. 117-119 (1755: 81-82; 1822: 136-137; 1864: 114-115).  3 x 4  to  4 x 3.  Shows a straight tetromino along one side moved to a perpendicular side so both shapes are  4 x 4.

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4v has an entry  "8½ c  Prob of figure"  followed by a staircase piece.  F. 145-146 show two pieces formed into both rectangles.  There are other dissection problems adjacent on F. 4v -- see 6.F.3, 6.F.4, 6.AQ, 6.AW.1.

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles.

No. 7, pp. 24 & 83-84 & plate I, fig. 4.  9 x 16  to  12 x 12.

No. 12, pp. 25 & 85 & plate I, fig. 9.  4 x 9  to  6 x 6.

No. 14, pp. 26 & 86 & plate I, fig. 10.  10 x 20  to  13 1/3 x 15.

Endless Amusement II.  1826? 

Prob. 1, p. 188.  5A x 6B  to  6A x 5B  and to  4A x 7B  with two  A x B  projections.  The  6A x 5B  looks to be square.  = New Sphinx, c1840, p. 136.

Prob. 22, pp. 200-201.  16 x 9  to fill a  12 x 12  hole.  Does it by cutting in four pieces -- one  12 x 9  and three  4 x 3.  = New Sphinx, c1840, pp. 136-137.

Prob. 33, pp. 210-211.  Take a rectangle of proportion  2 : 3  and cut it into two pieces to make a square.  Uses cut from  4A x 5B  to  5A x 4B,  but if we make the rectangle  4 x 6,  this makes  A = 1, B = 6/5  and the 'square' is  5 x 24/5.  = New Sphinx, c1840, p. 140.

Nuts to Crack II (1833), no. 124.  9 x 16  to fill a  12 x 12  hole using four pieces.  = Endless Amusement II, prob. 22.

Young Man's Book.  1839.  Pp. 241-242.  Identical to Endless Amusement II, prob. 22.

Boy's Own Book.  1843 (Paris): 436 & 441, no. 6.  5A x 6B  to  6A x 5B  and to  4A x 7B  with two  A x B  projections.  The  6A x 5B  looks to be square.  = Boy's Treasury, 1844, pp. 425 & 429.  = de Savigny, 1846, pp. 353 & 357, no. 5.  Cf de Savigny, below.

de Savigny.  Livre des Écoliers.  1846.  P. 283: Faire d'une carte un carré.  View a playing card as a  5A x 4B  rectangle and make a staircase cut and shift to  4A x 5B,  which will be nearly square.  [When applied to a bridge card,  3.5 x 2.25 in, the result is  2.8 x 2.8125 in.]

Magician's Own Book.  1857.  Prob. 2: The parallelogram, pp. 267 & 291.  Identical to Boy's Own Book, 1843 (Paris).

The Sociable.  1858.  Prob. 19: The perplexed carpenter, pp. 292 & 308.  2 x 12  to  3 x 8.  = Book of 500 Puzzles, 1859, prob. 19, pp. 10 & 26.  = The Secret Out, 1859, p. 392.

Book of 500 Puzzles.  1859. 

Prob. 19: The perplexed carpenter, pp. 10 & 26.  As in The Sociable.

Prob. 2: The parallelogram, pp. 81 & 105.  Identical to Boy's Own Book, 1843 (Paris).

Charades, Enigmas, and Riddles.  1860: prob. 29, pp. 60 & 64;  1862: prob. 30, pp. 136 & 142;  1865: prob. 574, pp. 107 & 155.  16 x 9  to  12 x 12.  All the solutions have an extraneous line in one figure.

Boy's Own Conjuring Book.  1860.  Prob. 2: The parallelogram, pp. 229 & 254.  Identical to Boy's Own Book, 1843 (Paris).

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 584-4, pp. 286 & 404.  Looks like  3 x 4  to  2 x 6.  The rectangles are formed by trimming a quarter off a playing card.  The diagrams are not very precise, but it seems that the card is supposed to be twice as long as wide.  If we take the card as  4 x 8, then the problem is  3 x 8  to  4 x 6.

Hanky Panky.  1872.  The parallelogram, p. 107.  "A parallelogram, ..., may be cut into two pieces, by which two other figures can be formed."  Shows  5A x 4B  cut, but no other figures.

Mittenzwey.  1880.  Prob. 253 & 255, pp. 45-46 & 96-97;  1895?: 282 & 284, pp. 49 & 98-99;  1917: 282 & 284, pp. 45 & 93-94.  4 x 9  to  6 x 6.  16 x 9  to  12 x 12.

Cassell's.  1881.  The carpenter's puzzle, p. 89.  = Manson, 1911, p. 133.  3 x 8  board to cover  2 x 12  area.

Richard A. Proctor.  Some puzzles;  Knowledge 9 (Aug 1886) 305-306  &  Three puzzles;  Knowledge 9 (Sep 1886) 336-337.  Cut  4 x 3  to  3 x 4.  Discusses general method for  nA x (n+1)B  to  (n+1)A x nB  and notes that the shape can be oblique as well as rectangular.

Lemon.  1890.  Card board puzzle, no. 58, pp. 11‑12 & 99.  c= The parallelogram puzzle, no. 620, pp. 77 & 120  (= Sphinx, no. 706, pp. 92 & 121).  Same as Boy's Own Book, 1843 (Paris).  In the pictures,  A  seems to be equal to  B.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 136, no. 8.  9 x 15  to  12 x 12.  No solution.

Tom Tit, vol. 2.  1892.  Les figures superposables, pp. 149-150.  3 x 2  to  2 x 3.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Chinese Geometrical Puzzles No. 1, pp. 108 & 111.  Same as Boy's Own Book, 1843 (Paris).

Hoffmann.  1893.

Chap. III, no. 8: The extended square, pp. 91 & 124‑125 = Hoffmann-Hordern, p. 80.  As in Boy's Own Book, 1843 (Paris), but  A  is clearly not equal to  B.

Chap. III, no. 31: The carpenter's puzzle -- no. 2, pp. 103 & 137 = Hoffmann-Hordern, p. 101.  12 x 36  to  18 x 24.

Clark.  Mental Nuts.  1897, no. 27.  The leaking ship.  12 x 12  to  9 x 16.

Benson.  1904.

The extended square, p. 190.  5A x 6B  square, but the other two figures are  8A x 7B  and  8A x 5B  with two  A x B  projections.

The carpenter's puzzle (No. 1), pp. 190‑191.  = Hoffmann, p. 103.

Anon [possibly Dudeney??]  Breakfast Table Problems No. 331: A carpenter's dilemma.  Daily Mail (31 Jan  &  1 Feb 1905) both p. 7.  16 x 9  to  12 x 12.

Pearson.  1907.  Part II, no. 1: The carpenter's puzzle, pp. 1‑2 & 185‑186.  2 x 12  to  3 x 8.

Wehman.  New Book of 200 Puzzles.  1908.

P. 6: The perplexed carpenter.  2 x 12  to  3 x 8.  c= The Sociable.

P. 9: The carpenter's puzzle.  "A plank was to be cut in two: the carpenter cut it half through on each side, and found he had two feet still to cut.  How was it?"  This is very vague and can only be recognised as a version of our present problem because the solution looks like cutting a  2 x 6  to make a  3 x 4.

P. 16: The parallelogram.  Identical to Boy's Own Book, 1843 (Paris).

P. 17: Another parallelogram.  Takes a parallelogram formed of a square and a half a square and intends to form a square.  He cuts  5A x 4B  and makes  4A x 5B.  But for this to be a square, it must be  20 x 20  and then the original was  25 x 16,  which is not quite in the given shape.

M. Adams.  Indoor Games.  1912.  A zigzag puzzle, p. 349, with figs. on 348.  5A x 6B  square, but the other two figures are  5A x 4B  and  7A x 4B  with two  A x B  projections.

Loyd.  Cyclopedia.  1914.

The smart Alec puzzle, pp. 27 & 342.  (= MPSL1, prob. 93, pp. 90‑91 & 153‑154.)  Cut a mitre into pieces which can form a square.  He trims the corners and inserts them into the notch to produce a rectangle and then uses a staircase cut which he claims gives a square using only four pieces.  Gardner points out the error, as carefully explained by Dudeney, below.  Since Dudeney gives this correction 1n 1911, he must have seen it in an earlier Loyd publication, possibly OPM?

The carpenter's puzzle, pp. 51 & 345.  Claims any rectangle can be staircase cut to make a square.  Shows  9 x 4  to  6 x 6  and  25 x 16  to  20 x 20.  Cf Kanchusen (1727) and Dudeney (1926).

Dudeney.  Perplexities.  Strand Magazine 41 (No. 246) (Jun 1911) 746  &  42 (No. 247) (Jul 1911) 108.  No. 45: Dissecting a mitre.  "I have seen an attempt, published in America, ..."  Sketches Loyd's method and says it is wrong.  "At present no solution has been found in four pieces, and one in five has not apparently been published."

Dudeney.  AM.  1917.  Prob. 150: Dissecting a mitre, pp. 35‑36 & 170‑171.  He fully describes "an attempt, published in America", i.e. Loyd's method.  If the original square has side  84,  then Loyd's first step gives a  63 x 84  rectangle, but the staircase cut yields a  72 x 73½  rectangle, not a square.  Dudeney gives a 5 piece solution and says "At present no solution has been found in four pieces, and I do not believe one possible."

Dudeney.  MP.  1926.  Prob. 115: The carpenter's puzzle, pp. 43‑44 & 132‑133.  = 536, prob. 338, pp. 116‑117 & 320‑321.  Shows  9 x 16  to  12 x 12.  "But nobody has ever attempted to explain the general law of the thing.  As a consequence, the notion seems to have got abroad that the method will apply to any rectangle where the proportion of length to breadth is within reasonable limits.  This is not so, and I have had to expose some bad blunders in the case of published puzzles ...."  He discusses the general principle and shows that an  n‑step cut dissects  n2 x (n+1)2  to a square of side  n(n+1).  Gardner adds a note referring to AM, prob. 150.  Cf Kanchusen (1727) & Loyd (1914)

Stephen Leacock.  Model Memoirs and Other Sketches from Simple to Serious.  John Lane, The Bodley Head, 1939, p. 299.  Mentions  12 x 12  to  9 x 16.

Harry Lindgren.  Geometric Dissections.  Van Nostrand, 1964.  P. 28 discusses Loyd's mitre dissection problem and variations.  He also thinks a four piece solution is impossible.

 

          6.AY.1.       O'BEIRNE'S STEPS

 

           This is a cube dissected into 6 pieces which form 6 cuboids, each of which can be 'staircased' in two ways.  There is a 6-cycle through the cuboids, with relative sizes:  12 x 12 x 12,  8 x 12 x 18,  8 x 9 x 24,  12 x 6 x 24,  16 x 6 x 18,  16 x 9 x 12.  I have a fine example using six different woods, that had been made for Tom O'Beirne, from Mrs. O'Beirne.

 

Richard K. Guy.  Op. cit. in 5.H.2.  1960.  Pp. 151-152 describes O'Beirne's invention.

T. H. O'Beirne.  Puzzles and paradoxes -- 9: A six-block cycle for six step-cut pieces.  New Scientist 9 (No. 224) (2 Mar 1961) 560-561.

 

          6.AY.2.       SWISS FLAG PUZZLE

 

          This appears to be a  7 x 5  flag with a Greek cross                      X X X X X O O     X X X X X

of  5  cells removed from the middle as in the first figure                          X X X   O O O     X X X O O

at the right.  One has to cut it into two pieces to make a                               X X       O O     X X O O O

perfect square.  This is done by cutting along a 'staircase'                         X X X   O O O     X X X O O

as shown.  However, this seems to produce a  5 x 6  flag,                       X X O O O O O     X X O O O

not a square.  But there is usually a swindle -- the diagram                                                   O O O O O

is not drawn with the unit cells square, but instead the unit

cells are  6/5  as wide as they are tall.  Normally the reader would not recognize this and the diagrams are often rather imprecise.

 

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 5: The flag puzzle.  Starts with a square and asks to make a Swiss flag.  The square is actually  31.5 mm = 1¼ in  on an edge and the flag is  26 x 37 mm  =  1 x 1½ in  with the Greek cross formed from squares of edge  5 mm = ¼ in,  so the areas do not add up!  No solution.

Loyd.  Problem 4: The Swiss flag puzzle.  Tit-Bits 31  (31 Oct 1896) 75  &  (21 Nov 1896) 131.  c= Cyclopedia, pp. 250 & 373.  Swiss flag puzzle with the flag at an angle and a slight wiggle in the edges, so the solution requires an extra cut to make it square. 

Loyd.  Cyclopedia, 1914, pp. 14 & 341:  A Swiss puzzle -- part 2.  = MPSL2, no. 144, pp. 101 & 166.  = SLAHP, pp. 48 & 102: How was this flag made?  Starts with the Swiss flag which is  47 x 30 mm  and the Greek cross has cells  6.5 x 5.5 mm,  so this is approximately the correct shape to make a square, but the resulting square is not drawn.

 

          6.AZ.           BALL PYRAMID PUZZLES

 

          This section is largely based on Gordon's Notes, cf below.  See also 6.AP.2 for dissections of a tetrahedron in general.  This section is now expanding to consider all polysphere puzzles.

          See also S&B, p. 42, which mentions Hein and some other versions.

 

Piet Hein.  Pyramystery.  Made by Skjøde of Skjern, Denmark, 1970.  With leaflet saying it was "recently invented by Piet Hein....  Responding to numerous requests, the inventor has therefore obliged the many admirers of the puzzle by also inventing its history".  He then gives a story about Cheops.  Peter Hajek and Jerry Slocum have different examples!!

                    Hajek's example has four planar rectangular pieces:  1 x 4,  2 x 3,  3 x 2,  4 x 1  rectangles.  It is the same as Tut's Tomb -- see below.  It has a 4pp English leaflet marked  © Copyright Piet Hein 1970.

                    Slocum's example has  6  planar pieces:  4  3-spheres and  2  4-spheres.  The leaflet is 34pp (?? -- Slocum only sent me part of it) with 3pp of instructions in each of 9 languages and then 6pp of diagrams of planar and 3-D problems.  It is marked  © 1970 Aspila,  so perhaps this is a later development from the above??  The story part of the text is very similar to the above, but slightly longer.  The pieces make an order  4  tetrahedron or two order  3  tetrahedra or two order  4  triangles and one can also divide them into two groups of three pieces such that one group makes an order  3  tetrahedron, but the other does not.

                    Advertising leaflet for Pyramystery from Piet Hein International Information Center, ©1976, describes the puzzle as having six pieces.

Mag‑Nif Inc.  Tut's Tomb.  c1972.  Same as the first Pyramystery.

Akira Kuwagaki & Sadao Takenaka.  US Patent 3,837,652 -- Solid Puzzle.  Filed: 1 May 1973;  patented: 24 Sep 1974.  2pp + 4pp diagrams.  Four planar  3-spheres and a  2-sphere to make a square pyramid of edge  3.  11  planar  4-spheres to make an octahedron shape of edge  4.  Cites a 1936 Danish patent -- Hein ??NYS

Len Gordon.  Perplexing Pyramid.  1974.  Makes a edge  4  tetrahedron with  6  planar right-angled pieces:  domino;  straight and L trominoes;  I,  L,  Y  tetrominoes.

Patrick A. Roberts.  US Patent 3,945,645 -- Tangential Spheres Geometric Puzzle.  Filed: 28 Jun 1976;  patented: 29 Nov 1977.  3pp + 3pp diagrams.  8  4-spheres and a  3-sphere to make a tetrahedron of side  5.  5  of the  4-spheres are non-planar. 

Robert E. Kobres.  US Patent 4,060,247 -- Geometric Puzzle.  Filed: 28 Jun 1976;  patented: 29 Nov 1977.  1p + 2pp diagrams.   5  pieces which make a  4 x 5  rhomboid or a tetrahedron.  Two pieces have the form of a  2 x 3  rhombus;  two pieces are  2-spheres  and the last piece is the linear  4-sphere.

Len Gordon.  Some Notes of Ball‑Pyramid and Related Puzzles.  Revised version, 10 Jul 1986, 14pp.  Available from the author, 2737 N. Nordic Lane, Tucson, Arizona, 85716, USA.

Ming S. Cheng.  US Patent 4,988,103 -- Geometric Puzzle of Spheres.  Filed: 2 Oct 1989;  patented: 29 Jan 1991.  Front page, 5pp diagrams, 4pp text.  A short version is given in Wiezorke, 1996, p. 64.  7  planar  5‑spheres to make a tetrahedron;  a hexagon with sides  3, 4, 4, 3, 4, 4;  an equilateral triangle lacking one vertex.

Bernhard Wiezorke.  Puzzling with Polyspheres.  Published by the author (Lantzallee 18, D‑4000 Düsseldorf 30, Germany), Mar 1990, 10pp.

Bernhard Wiezorke.  Compendium of Polysphere Puzzles.  (1995);  Second Preliminary Edition, as above, Aug, 1996.  64pp, reproducing the short versions of the above patents.  Despite Wiezorke's searches, nothing earlier than Hein's 1970 puzzles has come to light.

Torsten Sillke & Bernhard Wiezorke.  Stacking identical polyspheres.  Part 1: Tetrahedra.  CFF 35 (Dec 1994) 11-17.  Studies packing of tetrahedra with identical polysphere pieces, with complete results for tetrahedra of edges  4 - 8  and polyspheres of  3, 4, 5  spheres.  Some of the impossibility results have only been done by computer, but others have been verified by a proof. 

 

          6.BA.           CUTTING A CARD SO ONE CAN PASS THROUGH IT

 

Ozanam.  1725.  1725: vol. IV, prob. 34, pp. 436‑437 & fig. 40, plate 12 (14).

Minguet.  1733.  Pp. 115-117.  (1755: 83-84; 1864: 112-113; not noted in 1822, but it's likely to be at p. 138.)  Similar to Ozanam, 1725.

Alberti.  1747.  Art. 34, p. 208-209 (110) & fig. 42, plate XI, opp. p. 210 (109).  Copied from Ozanam, 1725.

Family Friend 3 (1850) 210 & 241.  Practical puzzle -- No. XVII.  Shows a 3 inch by 5 inch card.  Repeated as Puzzle 15 -- The wonder puzzle in (1855) 339 with solution in (1856) 28.

Magician's Own Book.  1857.  Prob. 10: The cardboard puzzle, pp. 269 & 294.  Problem shows 3 inch by 5 inch card.  Answer calls it "the cut card puzzle".  c= Landells, Boy's Own Toy-Maker, 1858, p. 142.  = Book of 500 Puzzles, 1859, pp. 83 & 108.  = Boy's Own Conjuring Book, 1860, prob. 9, pp. 230 & 256.

Indoor & Outdoor.  c1859.  Part II, prob. 7: The cardboard puzzle, p. 129.  No diagram, so the solution is a bit cryptic.

The Secret Out.  1859.  How to Cut a Visiting Card for a Cat to Jump through it, p. 382.

Illustrated Boy's Own Treasury.  1860.  No. 27, pp. 400 & 440.  Identical to Magician's Own Book, but solution omits the sentence:  "A laurel leaf may be treated in the same manner."

Magician's Own Book (UK version).  1871.  To cut a card for one to jump through, p. 124.  He adds: "The adventurer of old, who, inducing the aborigines to give him as much land as a bull's hide would cover, and made it into one strip by which acres were enclosed, had probably played at this game in his youth."  See 6.AD.

Elliott.  Within-Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 2: The cardboard puzzle, pp. 27 & 30‑31.  No diagram, so the solution is a bit cryptic.

Lemon.  1890.  Cardboard puzzle, no. 140, pp. 23 & 102.  = Sphinx, no. 467, pp. 65 & 113.

J. B. Bartlett.  How to walk through a laurel leaf.  The Boy's Own Paper 12 (No. 587) (12 Apr 1890) 440.

Hoffmann.  1893.  Chap. X, no. 28: The cut playing‑card, pp. 346 & 385‑386 = Hoffmann‑Hordern, p. 243.

Benson.  1904.  The elastic cardboard puzzle, pp. 200‑201.

Dudeney.  Cutting-out paper puzzles.  Cassell's Magazine ?? (Dec 1909) 187-191 & 233-235.  With photo of Dudeney going through the card.

Collins.  Book of Puzzles.  1927.  Through a playing card, pp. 16-17.

 

          6.BB.           DOUBLING AN AREA WITHOUT CHANGING ITS HEIGHT OR WIDTH

 

          The area is usually a square, but other shapes are possible.  If one views it as a reduction, one can reduce the area to any fraction of the original!

 

The Sociable.  1858.  Prob. 41: The carpenter puzzled, pp. 298 & 316.  3 x 3  square of wood with holes in it forming a  4 x 4  array with the corner holes at the corners of the board.  Claims one can cut  1/4  of the board out of the centre without including any holes.  But this only gets  2/9  of the area -- double the central square.  = Book of 500 Puzzles, 1859, prob. 41, pp. 16 & 34.  = Secret Out, 1859, pp. 386-387.

Indoor & Outdoor.  c1859.  Part II: prob. 14: The carpenter puzzled, pp. 133‑134.  Almost identical with The Sociable.

Hanky Panky.  1872.  P. 226 shows the same diagram as the solution in The Sociable, but there is no problem or text.

Lewis Carroll.  Letter of 15 Mar 1873 to Helen Feilden.  = Carroll-Collingwood, pp. 212-215 (Collins 154-155), without solution.  Cf Carroll-Wakeling, prob. 28: The square window, pp. 36-37 & 72.  Halve the area of a square window.  Wakeling and Carroll-Gardner, p. 52, give the surname as Fielden, but it is Feilden in Carroll-Collingwood and in Cohen, who sketches her life.  Wakeling writes that Feilden is correct.

Mittenzwey.  1880.  Prob. 216-217, pp. 38-39 & 90;  1895?: 241-242, pp. 43 & 92;  1917: 241‑242, pp. 39 & 88.  Divide a rectangle or square into two pieces with the same height and width as the square.  Solution is to draw a diagonal.

Lemon.  1890.  A unique window, no. 444, pp. 58 & 114.  The philosopher's puzzle, no. 660, pp. 82 & 121.

Don Lemon.  Everybody's Scrap Book of Curious Facts.  Saxon, London, 1890.  P. 82 quotes an article from The New York World describing this as 'an excellent, if an old, puzzle'.

Hoffmann.  1893.  Chap. IX, no. 28: A curious window, pp. 319 & 327 = Hoffmann-Hordern, pp. 211-212.  Notes that either a diamond or a triangle in appropriate position can be so doubled.

Clark.  Mental Nuts.  1897, no. 40.  The building lot.  "Have a lot  50 x 100.  Want to build a house  50 x 100  and have the yard same size.  How?"  Solution shows  50 x 100  with a diagonal drawn.

Pearson.  1907.  Part II, no. 79: At a duck pond, pp. 79 & 176.  A square pond is to be doubled without disturbing the duckhouses at its corners.

Wehman.  New Book of 200 Puzzles.  1908.  The carpenter puzzled, p. 39.  = The Sociable.

Will Blyth.  Handkerchief Magic.  C. Arthur Pearson, London, 1922.  Doubling the allotment, pp. 23-24.

Hummerston.  Fun, Mirth & Mystery.  1924.  Some queer puzzles, Puzzle no. 76, part 6, pp. 164 & 183.  Solution notes that a window in the shape of a diamond or a right triangle or an isosceles triangle can be doubled in area without changing its width or height.

King.  Best 100.  1927.

No. 2, pp. 7‑8 & 38.  Same as Indoor & Outdoor, with the same error.

No. 4, pp. 8 & 39.  Halve a square window.  See Foulsham's.

Foulsham's Games and Puzzles Book.  W. Foulsham, London, nd [c1930].  No. 2, pp. 5 & 10.  Double a window without changing its height or width.  (This is one of the few cases where the problem is not quite identical to King.)

M. Adams.  Puzzle Book.  1939.  Prob. B.117: Enlarging the allotment, pp. 86 & 110.  Double a square allotment without disturbing the trees at the corners.

 

          6.BC.           HOFFMAN'S CUBE

 

          This consists of  27  blocks,  a x b x c,  to make into a cube  a+b+c  on a side.  It was first proposed by Dean Hoffman at a conference at Miami Univ. in 1978.  See S&B, p. 43.  The planar version, to use  4  rectangles  a x b  to make a square of side  a + b  is easy.  These constructions are proofs of the inequality of the arithmetic and geometric means.  Sometime in the early 1980s, I visited David Klarner in Binghamton and Dean Hoffman was present.  David kindly made me a set of the blocks and a three-sided corner to hold them.

 

D. G. Hoffman.  Packing problems and inequalities.  In:  The Mathematical Gardner, op. cit. in 6.AO, 1981.  Pp. 212‑225.  Includes photos of Carl Klarner assembling the first set of the blocks.  Asks if there are analogous packings in  n  dimensions.

Berlekamp, Conway & Guy.  Winning Ways.  1982.  Vol. 2, pp. 739‑740 & 804‑806.  Shows all 21 inequivalent solutions.

 

          6.BD.  BRIDGE A MOAT WITH PLANKS

 

          In the simplest case, one has a  3 x 3  moat with a  1 x 1  island in the centre.  One wants to get to the island using two planks of length  1  or a bit less than  1.  One plank is laid diagonally across the corner of the moat and the second plank is laid from the centre of the first plank to the corner of the island.  If the width of the moat is  D  and the planks have length  L,  then the method works if  3L/2 > DÖ2,  i.e.  L > 2Ö2 D/3 = .94281.. D.  One really should account for the width of the planks, but it is not clear just how much overlap is required for stability.  Depew is the only example I have seen to use boards of different lengths.  With more planks, one can reach across an arbitrarily large moat, but the number of planks needed gets very large.  In this situation, the case of a circular moat and island is a bit easier to solve.

          The Magician's Own Book (UK version) version is quite different and quite erroneous.

 

Magician's Own Book (UK version).  1871.  The puzzle bridge, p. 123.  Stream 15 or 16 feet across, but none of the available planks is more than 6 feet long.  He claims that one can use a four plank version of the three knives make a support problem (section 11.N) to make a bridge.  However the diagram of the solution clearly has the planks nearly as long as the width of the stream.  In theory, one could build such a bridge with planks slightly longer than half the width of the stream, but to get good angles (e.g. everything crossing at right angles or nearly so), one needs planks somewhat longer than  Ö2/2  of the width.  E.g. for a width of 16 ft,  12 ft planks would be adequate.

Mittenzwey.  1880.  Prob. 298, pp. 54 & 105;  1895?: 330, pp. 58 & 106;  1917: 330, pp. 52‑53 & 101.  4 m gap bridged with two boards of length    m.  He only gives a diagram.  In fact this doesn't work because the ratio of lengths is  15/16 = .9375

Lucas.  RM2.  1883.  Le fossé du champ carré.  Bridge the gap with two planks whose length is exactly  1.  Notes this works because  3/2 > Ö2.

Hoffmann.  1893.  Chap. VII, no. 9, pp. 289 & 295.  Matchstick version = Hoffmann-Hordern, p. 193.

Benson.  1904.  The moat puzzle, p. 246.  Same as Hoffmann, but the second plank is shown under the first!!

Dudeney.  CP.  1907.  No. 54: Bridging the ditch, pp. 83-85 & 204.  Eight  9'  planks to cross a 10' ditch where it makes a right angle.

Pearson.  1907.  Part I, no. 34: Across the moat, pp. 122 & 186.

Blyth.  Match-Stick Magic.  1921.  Boy Scouts' bridge, p. 21.  Ordinary version done with matchsticks.

J. C. Cannell.  Modern Conjuring for Amateurs.  C. Arthur Pearson, London, nd [1930s?].  Boy Scouts' bridge, pp. 68-69.  As in Blyth.

Depew.  Cokesbury Game Book.  1939.  Crossing the moat, pp. 225-226.  Square moat  20  feet wide to be crossed with boards of width  18  and  15  ft.  In fact this doesn't work -- one needs  L1 + L2/2 > DÖ2.

"Zodiastar".  Fun with Matches and Matchboxes.  Op. cit. in 4.B.3.  Late 1940s?  The bridge, pp. 66-67 & 83.  Matchstick version of the square moat & square island problem.

F. D. Burgoyne.  Note 3106:  An  n  plank problem.  MG 48 (No. 366) (Dec 1964) 434‑435.  The island is a point in the centre of a  2 x 2  lake.  Given  n  planks of length  s,  can you get to the island?  He denotes the minimal length as  s(n)  and computes  s(1) = 1, s(2) = 2 Ö2/3,  s(3) = .882858...  and says  s(¥) = Ö2/2,  [but I believe it is  0,  i.e. one can get across an arbitrarily large moat with a fixed length of plank].

Jonathan Always.  Puzzling You Again.  Tandem, London, 1969.

Prob. 10: A damsel in distress, pp. 15 & 70‑71.  Use two planks of length  L  to reach a point in the centre of a circular moat of radius  R.  He finds one needs  L2 ³ 4R2/5.

Prob. 11: Perseus to the rescue again, pp. 15‑16 & 71‑72.  Same with five planks.  The solution uses only four and needs  L2 ³ 2R2/3.

C. V. G.[?] Howe?  Mathematical Pie 75 (Summer 1975) 590  &  76 (Autumn 1975) 603.  How big a square hole can be covered with planks of unit length?  Answer says there is no limit, but the height of the pile increases with the side of the square.

Highlights for Children (Columbus, Ohio).  Hidden Picture Favorites and Other Fun.  1981.  Brain Buster 4, pp. 12 & 32: Plink, plank, kerplunk?  Two children arrive at a straight(!) stream  4m  wide with two planks  3m  long.  Solution: extend one plank about  1¼ m over the stream and one child stands on the land end.  The second child carries the other plank over the stream and extends it to the other side and crosses.  He then pulls the plank so it extends about  1¼ m  over the stream.  The first child now extends her plank out to rest on the second plank and crosses, pulling up her plank and taking it with her.  I theory this technique will work if  L > ⅔D,  but one needs some overlap space, and the children may not have the same weight.

Richard I. Hess.  Puzzles from Around the World.  The author, 1997.  (This is a collection of 117 puzzles which he published in Logigram, the newsletter of Logicon, in 1984-1994, drawn from many sources.  With solutions.)  Prob. 64.  Usual problem with  D = 10,  but he says the board have width  W = 1  and so one use the diagonal of the board in place of  L.  In my introduction, we saw that the standard version leads to  L2 ³ 8D2/9,  so Hess's version leads to  L2  ³ 8D2/9 - W2  and we can get across with slightly shorter planks, but we have to tread very carefully!

 

          6.BE. REVERSE A TRIANGULAR ARRAY OF TEN CIRCLES

 

          One has a triangle of ten coins with four on an edge.  Reverse its direction by moving only three coins.  New section -- I'm surprised not to have seen older examples.

 

Sid G. Hedges.  More Indoor and Community Games.  Methuen, London, 1937.  The triangle trick, p. 54.  Uses peas, buttons or nuts.

M. Adams.  Puzzle Book.  1939.  Prob. B.34: Pitching camp, pp. 66 & 103.  Array of tents.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  The General inspects the balloons, pp. 106 & 214.  Array of 10 barrage balloons.

Leopold.  At Ease!  1943.  1, 2, 3 -- shift!, pp. 20 & 198.  Thanks to Heinrich Hemme for the lead to this.

Joseph Leeming.  Games with Playing Cards Plus Tricks and Stunts.  Franklin Watts, 1949.  ??NYS -- but two abridged versions have appeared.

                    Games and Fun with Playing Cards.  Dover, NY, 1980.  This contains everything except the section on bridge.

                    Tricks and Stunts with Playing Cards Plus Games of Solitaire.  Gramercy Publishing, NY, nd [1960s?].  This includes all the tricks, stunts, puzzles and solitaire games.

                    25 Puzzles with Cards, 8th puzzle.  Tricky triangle.  Dover: pp. 154-155 & 172.  Gramercy: pp. 45-46 & 65.  Both have fig. 25 & 42. 

Young World.  c1960.  P. 7: fifteen coin problem.  Reverse a triangle with five on a side by moving five coins.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  Birds in flight, p. 34.  Says this problem is described by Gardner, but gives no specific source.

Maxey Brooke.  (Fun for the Money, Scribner's, 1963);  reprinted as:  Coin Games and Puzzles, Dover, 1973.  Prob. 4: Bottoms up, pp. 15 & 75.  On p. 6, he acknowledges Leopold as his source.  Thanks to Heinrich Hemme for this reference.

D. B. Eperson.  Triangular (old) pennies.  MG 54 (No. 387) (Feb 1970) 48‑49.  The number of pennies which must be moved to reverse a triangle with  n  on a side is  [T(n)/3],  where T(n)  is the  n‑th triangular number, which is the number in the array.

James Bidwell.  The ten‑coin triangle.  MTg 54 (1971) 21‑22.  How many coins must be moved to reverse the triangle with  n  on an edge?  His students find the same value as Eperson, but they weren't sure they had proved it.

Putnam.  Puzzle Fun.  1978.  No. 25: Triangular reverse, pp. 6 & 31.  Usual 10 coin triangle.

 

          6.BF. PYTHAGOREAN RECREATIONS

 

          6.L  might be considered as part of this section.  There are some examples of problems with ladders which look like crossed ladders, but are simple Pythagorean problems.

          See also 6.AS.2 for dissection proofs of the theorem of Pythagoras.  I will include here only some interesting ancient examples.  See Elisha Scott Loomis; The Pythagorean Proposition; 2nd ed., NCTM, 1940, for many proofs.

          Aryabhata I, v. 17, states the Theorem of Pythagoras and the related theorem that if  ABC  is a diameter of a circle and  LBM  is a chord perpendicular to it, then  LB2 = AB x BC; Bhaskara I's commentary applies the latter in several forms where modern algebra would make it more natural to use the former.  Brahmagupta, v. 41, states  LM2 = AB x BC.

          I had overlooked the examples in Mahavira -- thanks to Yvonne Dold for pointing them out.

 

Fibonacci.  1202.  Pp. 397‑398 (S: 543-544) looks like a crossed ladders problem but is a simple right triangle problem.

Vyse.  Tutor's Guide.  1771? 

Prob. 9, 1793: p. 178, 1799: p. 189 & Key p. 224.  A ladder  40  long in a roadway can reach  33  up one side and, from the same point, can reach  21  up the other side.  This is actually a simple right triangle problem.  There is a misprint of  9  for  6  in the answer.

Prob. 17 (in verse), 1793: 179, 1799: p. 190 & Key p. 228.  A variation of the Broken Bamboo problem, cf below, with  D = 30,  H - X = 63,  which is a simple right triangle problem.

Hutton.  A Course of Mathematics.  1798? 

Prob. VIII,  1833: 430;  1857: 508.  = Vyse, prob. 19.

Prob. IX,  1833: 430;  1857: 508.  = Vyse, prob. 17 with  D = 15,  H - X = 39.

 

          6.BF.1         THE BROKEN BAMBOO

 

          A bamboo (or tree) of height  H  breaks at height  X  from the ground so that the broken part reaches from the break to the ground at distance  D  from the foot of the bamboo.  In fact the quadratic terms drop out of the solution, leaving a linear problem.  This may be of Babylonian origin??  The hawk and rat problems of 6.BF.3 are geometrically the same problem viewed sideways.

          In all cases below,  H  and  D  are given and  X  is sought, so I will denote the problem by  (H, D).

          See Tropfke, p. 620.

 

Chiu Chang Suan Ching (Jiu Zhang Suan Shu).  c‑150?  Chap. IX, prob. 13, p. 96.  [English in Mikami, p. 23 and in Swetz & Kao, pp. 44‑45, and in HM 5 (1978) 260.]  (10, 3).

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 17, part 2.  Sanskrit is on pp. 97‑103; English version of the examples is on pp. 296-300.  The material of interest is examples 4 and 5.  In the set-up described under 6.BF.3, the bamboo is  BOC  which breaks at  O  and the point  C  reaches the ground at  L.

                    Ex. 4:  (18, 6).  Shukla notes this is used by Chaturveda.

                    Ex. 5:  (16, 8).

Chaturveda.  860.  Commentary to the Brahma‑sphuta‑siddhanta, chap. XII, section IV, v. 41, example 2.  In Colebrooke, p. 309.  Bamboo: (18, 6).

Mahavira.  850.  Chap. VII, v. 190-197, pp. 246-248.

v. 191.  (25, 5),  but the answer has  H - X  rather than  X.

v. 192.  (49, 21),  but the answer has  H - X  rather than  X.

v. 193.  (50, 20),  but with the problem reflected so the known leg is vertical rather than horizontal.

v. 196.  This modifies the problem by imagining two trees of heights  H  and  h,  separated by  D.  The first, taller, tree breaks at height  X  from the ground and leans over so its top reaches the top of the other tree.  If we subtract  h  from  X  and  H,  then  │X - h│  is the solution of the problem  (H - h, D).  Because the terms are squared, it doesn't matter whether  X  is bigger or smaller than  h.  He does the case  H, h, D  =  23, 5, 12.

Bhaskara II.  Lilavati.  1150.  Chap. VI, v. 147‑148.  In Colebrooke, pp. 64‑65.  (32, 16).

Bhaskara II.  Bijaganita.  1150.  Chap. IV, v. 124.  In Colebrooke, pp. 203‑204.  Same as Lilavati.

Needham, p. 28, is a nice Chinese illustration from 1261.

Gherardi.  Libro di ragioni.  1328.  Pp. 75‑76: Regolla di mesura.  (40, 14).

Pseudo-dell'Abbaco.  c1440.  No. 166, p. 138 with B&W reproduction on p. 139.  Tree by stream.  (60, 30).  I have a colour slide of this.

Muscarello.  1478.  F. 96v, pp. 224-225.  Tree by a stream.  (40, 30).

Calandri.  Aritmetica.  c1485.  Ff. 87v-88r, pp. 175-176.  Tree by a stream.  (60, 30).  = Pseudo-dell'Abbaco.

Calandri.  Arimethrica.  1491.  F. 98r.  Tree by a river.  (50, 30).  Nice woodcut picture.  Reproduced in Rara, 48.

Pacioli.  Summa.  1494.  Part II, f. 55r, prob. 31.  (30, 10).  Seems to say this very beautiful and subtle invention is due to Maestro Gratia.

Clark.  Mental Nuts.  1897, no. 78.  The tree and the storm.  (100, 30).  [I have included this as this problem is not so common in the 19C and 20C as in earlier times.]

N. L. Maiti.  Notes on the broken bamboo problem.  Gaņita-Bhāratī [NOTE:  ņ  denotes an  n  with an underdot] (Bull. Ind. Soc. Hist. Math.) 16 (1994) 25-36 -- ??NYS -- abstracted in BSHM Newsletter 29 (Summer 1995) 41, o/o.  Says the problem is not in Brahmagupta, though this has been regularly asserted since Biot made an error in 1839 (probably a confusion with Chaturveda -- see above).  He finds eight appearances in Indian works, from Bhaskara I (629) to Raghunath-raja (1597).

 

          6.BF.2.        SLIDING SPEAR  =  LEANING REED

 

          A spear (or ladder) of length  H  stands against a wall.  Its base moves out  B  from the wall, causing the top to slide down  D.  Hence  B2 + (H - D)2 = H2.

          The leaning reed has height  H.  It reaches  D  out of the water when it is straight up.  When it leans over, it is just submerged when it is  B  away from its upright position.  This is identical to the sliding spear turned upside down.

          In these problems, two of  H,  B,  D  are given and one wants the remaining value.  I will denote them, by e.g.  H, D = 30, 6.

          See Tropfke, p. 619 & 621.

 

BM 85196.  Late Old Babylonian tablet in the British Museum, c‑1800.  Transcribed, translated and commented on by O. Neugebauer; Mathematische Keilschrift-texte II; Springer, Berlin, 1935, pp. 43++.  Prob. 9 -- translation on pp. 47-48, commentary on p. 53.  Quoted in B. L. van der Waerden; Science Awakening; OUP, 1961, p. 76.  See also:  J. Friberg; HM 8 (1981) 307-308.  Sliding beam(?) with  H, D = 30, 6  and with  H, B = 30, 18.

BM 34568.  Seleucid period tablet in the British Museum, c‑300.  Transcribed, translated and commented on by O. Neugebauer; Mathematische Keilschrift-texte III, Springer, Berlin, 1937, pp. 14-22 & plate 1.  Prob. 12 -- translation on p. 18, commentary on p. 22.  Quoted in van der Waerden, pp. 76‑77.  See also:  J. Friberg, HM 8 (1981) 307-308.  Sliding reed or cane,  B, D = 9, 3.

Papyri Cairo J. E. 89127‑30  &  89137‑43.  c‑260.  Shown and translated in:  Richard A. Parker; Demotic Mathematical Papyri; Brown Univ. Press, Providence, 1972; pp. 1, 3‑4, 35‑40 & Plates 9‑10.

Prob. 24‑26:  H, B  =  10, 6;  14½, 10;  10, 8.

Prob. 27‑29:  H, D  =  10, 2;  14½,   4;  10, 4.

Prob. 30‑31:  B, D  =    6, 2;        10, 4.

Chiu Chang Suan Ching (Jiu Zhang Suan Shu).  c‑150?  Chap. IX.

Prob. 6, p. 92.  [English in Mikami, p. 22.]  Leaning reed,  B, D = 5, 1.

Prob. 7, p. 93.  Version with a rope hanging and then stretched giving  B, D = 8, 3.

Prob. 8, p. 93.  [English in Mikami, p. 22 and in Swetz & Kao, pp. 30‑32 and in HM 4 (1977) 274.]  Ladder, but with vertical and horizontal reversed.  B, D = 10, 1.  Mikami misprints the answer as  55  rather than  50.5.

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 17, part 2.  Sanskrit is on pp. 97‑103, with reproductions of original diagrams on pp. 101-102; English version of the examples is on pp. 296-300.  The material of interest is examples 6 and 7.  In the setup of 6.BF.3, the lotus is  OBA  and  LBM  is the water level.

                    Ex. 6:  B, D  =  24, 8.  Shukla notes this is used by Chaturveda.

                    Ex. 7:  B, D  =  48, 6.

Chaturveda.  860.  Commentary to the Brahma‑sphuta‑siddhanta, chap. XII, section IV, v. 41, example 3.  In Colebrooke, pp. 309‑310.  Leaning lotus:  B, D = 24, 8.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 124, no. 42.  ??NYS - cited by Tropfke 621.

Bhaskara II.  Lilavati.  1150.  Chap. VI, v. 152‑153.  In Colebrooke, pp. 66.  Leaning lotus:  B, D = 2,  ½.

Bhaskara II.  Bijaganita.  1150.  Chap. IV, v. 125.  In Colebrooke, p. 204.  Same as Lilavati.

Fibonacci.  1202.  P. 397 (S: 543).  Sliding ladder:  H, B = 20, 12.

Leonardo Fibonacci.  La Practica di Geometria.  Volgarizzata da Cristofano di Gherardo di Dino, cittadino pisano.  Dal Codice 2186 della Biblioteca Riccardiana di Firenze, 1448.  Ed. by Gino Arrighi, Domus Galilaeana, Pisa, 1966.  P. 37 and fig. 18.  Same as Fibonacci 1202.

Zhu Shijie.  Siyuan Yujian (Precious Mirror of the Four Elements).  1303.  ??NYS -- English given in Li & Du, p. 179.  Questions in verse, no. 1.  Two reeds,  14  apart, which reach    and  1  out of the water.  When they lean together, they just touch at the water surface.  The water is assumed to have the same depth at both reeds, which reduces the problem from two variables to one variable.

Gherardi?.  Liber habaci.  c1310.  Pp. 139‑140.  H, B = 20, 12

Gherardi.  Libro di ragioni.  1328.  Pp. 77‑78.  Ship's mast with  H, D = 131, 4.  B&W picture on p. 77, from f. 46v.

Columbia Algorism.  c1350.

No. 135, pp. 138‑139.  Sliding ladder:  H, B = 10, 6.

No. 140, pp. 149‑150.  Leaning tree:  H, B = 20, 10.

(I have colour slides of the illustrations to these problems.)

Pseudo-dell'Abbaco.  c1440.  No. 167, pp. 138-140, with B&W picture on p. 139.  Sliding spear:  H, D = 30, 4.  I have a colour slide of this.

della Francesca.  Trattato.  c1480.  F. 44r (108).  B, D = 6, 2.  English in Jayawardene.

Pacioli.  Summa.  1494.  Part II, ff. 54v-55v.

Prob. 25 (misprinted 52).  B, D = 6, 2.  = della Francesca.

Prob. 26.  H,  H - D  =  10, 8.

Prob. 27.  H, D = 10, 4.

Prob. 28.  H,  D  =  10,  B/3.

Prob. 29.  D + H = 12,  DH = 12.

Prob. 30.  H = D + 4,  DH = 12.

Prob. 41.  Tree of height  40  with a rope of length  50  tied to the top which reaches to the ground at a point  30  away.  Length  10  of the rope is pulled, causing the tree to lean.  How high is the top of the tree now?  We now have a triangle of sides  40, 40, 30  and want the altitude to the side of length  30.

   PART II.

F. 68r, prob. 98.  Tree of height  30  has a rope of length  50  tied to the top, so it reaches the ground  40  away.  How much rope has to be pulled to move the top of the tree to being  8  away from the vertical?  He neglects to give the value  50  in the problem statement and the ground distances of  8  and  32  are interchanged in the diagram.

van Etten.  1624.  Prob. 89 (86), part V (4), p. 135 (214).  Sliding ladder:  H, B = 10, 6.

Ozanam.  1694.  Prob. 42 & fig. 48, plate 10, 1696: 123-124;  1708: 129;  1725: 320-321 & plate 10 (11).  Ladder  25  long with foot  7  from wall.  Foot is pulled out  8  more -- how much does the top come down?

Vyse.  Tutor's Guide.  1771?  Prob. 16 (in verse), 1793: p. 179, 1799: p. 190 & Key p. 228.  H, B = 100, 10.

Mittenzwey.  1880.  Prob. 294, pp. 53-54 & 104;  1895?: 324, pp. 57 & 106;  1917: 324, pp. 52 & 100.  Leaning reed,  B, D  =  5, 1.  [I have included this and the next entry as this problem is not so common in the 19C and 20C as in earlier times.]

Clark.  Mental Nuts.  1904, no. 69;  1916, no. 95.  The boatman's puzzle.  Leaning pole,  B, D =  12, 6.  Find  H - D.

 

          6.BF.3.        WELL BETWEEN TWO TOWERS

 

          The towers have heights  A,  B  and are  D  apart.  A well or fountain is between them and equidistant from the tops of the towers.  I denote this by  (A, B, D).  Vogel, in his DSB article on Fibonacci, says the problem is Indian, and Dold pointed me to Mahavira.  Pseudo‑dell'Abbaco introduces the question of a sliding weight or pulley -- see Pseudo‑dell'Abbaco, Muscarello, Ozanam-Montucla, Tate, Palmaccio, Singmaster.

          I have just found that Bhaskara I gives several unusual variations on this.

          See Tropfke, p. 622.  See also 10.U.

 

          INDEX of  A, B, D  problems, with  A £ B.

 

    0          4          8            Chaturveda

    0          9        27            Bhaskara II

    0        12        24            Bhaskara I

    0        18        81            Bhaskara I

    5          6        12            Bhaskara I

  10        10        12            Bhaskara I

  13        15        14            Mahavira

  18        22        20            Mahavira

  20        24        22            Mahavira

  20        30        30            Gherardi?

  20        30        50            Perelman

  30        40        50            Fibonacci,  Muscarello,  Cardan

  30        50       100            Bartoli

  30        70       100            Tate

  40        50        30            Muscarello

  40        50        70            Pseudo-dell'Abbaco

  40        50       100            della Francesca

  40        60        50            Lucca 1754

  60        80       100            Calandri c1485

  70       100       150            Columbia Algorism,  Pacioli

  80        90       100            Calandri 1491

 

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 17, part 2.  Sanskrit is on pp. 97‑103; English version of the examples is on pp. 296-300.  The material of interest is examples 2 and 3.

                    These are 'hawk and rat problems'.  A hawk is sitting on a wall of height  A  and a rat is distance  D  from the base of the wall.  The rat tries to get to its hole, in the wall directly under the hawk.  The hawk swoops, at the same speed as the rat runs, and catches the rat when it hits the ground.  Hence this is the same as our two tower problem, but with  B = 0,  so I will denote this version by  (A, 0, D).  Bhaskara I attributes this type of problem to unspecified previous writers.  Shukla adds that later writers have it, including Chaturveda and Bhaskara II, qqv.

                    Ex. 2:  (12, 0, 24).

                    Ex. 3:  (18, 0, 81).  Bhaskara I explains the solution in detail and Shukla gives an English precis of it.  Let  ABOC  be the horizontal diameter of a circle and let  LBM  be a vertical chord.  LB  is our pole, with the hawk at  L,  and the rat is at  C  and wants to get to  B.  The point of capture is  O,  because  LO = OC.  From  LB2 = AB x BC,  we can determine  AB  and hence the other values.

                    Looking at Chaturveda (below), I now see that turning this sideways gives the same diagram as the broken bamboo problem -- the tree was  BC  and breaks at  O  to touch the ground at  L.  So the broken bamboo problem  (H, D)  is the same as the two towers or hawk and rat problem  (D, 0, H).

Bhaskara I.  629.  Ibid.  Examples 8 and 9 are 'crane and fish problems'.  A fish is at the NE corner of a rectangular pool and a crane is at the NW corner and they move at the same speeds.  The fish swims obliquely to the south side, but the crane has to walk along the edge of the pool.  The fish unfortunately gets to the side just as the crane reaches the same point and gets eaten.  This again like our two tower problem, but with one pigeon unable to fly, so it has to walk down the tower and across.  Because the pool is rectangular, the two values  A  and  B  are equal.

                    Ex. 8:  (6, 6, 12).

                    Ex. 9:  (10, 10, 12).  The meeting point is  3 3/11  from the SW corner.

Chaturveda.  860.  Commentary to the Brahma‑sphuta‑siddhanta, chap. XII, section IV, v. 41, example 4.  In Colebrooke, p. 310.  Cat and rat, where the cat behaves like the hawk of Bhaskara I:  (4, 0, 8).

Mahavira.  850.  Chap. VII, v. 201-208, pp. 249-251.

          He gives several problems, but he usually also asks for the equal distance from the top

                    of each tower to the fountain. 

v. 204.  Two pillars, with a rope between them which touches the ground but with equal lengths to the tops.  (13, 15, 14).

v. 206.  Two hills with mendicants who are able to fly along the hypotenuses.  (22, 18, 20)

v. 208.  Same context.  (20, 24, 22).

Bhaskara II.  Lilavati.  1150.  Chap. VI, v. 149‑150.  In Colebrooke, pp. 65‑66.  Peacock and snake version of the hawk and rat problem:  (9, 0, 27).

Fibonacci.  1202.  De duabus avibis [On two birds], pp. 331‑332 (S: 462-463).  (40, 30, 50).  He does the same problem differently on pp. 398‑399 (S: 544-545).

Gherardi?.  Liber habaci.  c1310.  P. 139.  (20, 30, 30).

Lucca 1754.  c1330.  F. 54v, pp. 120-121.  (60, 40, 50).

Columbia Algorism.  c1350.  No. 136, pp. 139‑140.  (70, 100, 150).

Bartoli.  Memoriale.  c1420.  Prob. 10, f. 76r (= Sesiano 138-139 & 148-149, with reproduction of the relevant part of f. 76r on p. 139).  (50, 30, 100).

Pseudo-dell'Abbaco.  c1440.

Prob. 80, p. 72, with picture on p. 71.  (50, 40, 70).

Prob. 158‑159, pp. 129‑133, with illustrations on pp. 130 & 132, deal with the related problem where a rope with a sliding weight hangs between two towers, and the diagram clearly shows the weight in the air, not reaching the ground, so that the resulting triangles are similar.  [I found it an interesting question to determine when the rope was long enough to reach the ground, and if not, how much above the ground the weight was -- see Muscarello, Ozanam-Montucla, Singmaster below.]

           Prob. 158 has  A, B, D = 40, 60, 40  and a rope of length  L = 110,  so the rope is more than long enough for the weight to reach the ground, but all he does is show that the two parts of the rope are  66  and  44,  which is a bit dubious as there is slack in the rope.  The diagram clearly shows the weight in the air.  I have a colour slide of this.

           Prob. 159 has  A, B = 40, 60  with  L = 120  such that the weight just touches the ground -- find the distances of the weight to the towers.

Prob. 160, p. 133.  (40, 30, 50).

Muscarello.  1478. 

Ff. 95r-95v, pp. 222-223.  A, B, D = 50, 40, 30.  Place a rope between the towers just long enough to touch the ground.

Ff. 95v-96r, pp. 223-224.  A, B, D = 30, 20, 40.  A rope of length  L = 60  with a sliding weight is stretched between them -- where does the weight settle?

F. 99r, pp. 227-228.  Fountain between towers for doves:  (40, 30, 50).

della Francesca.  Trattato.  c1480.  F. 22r (72).  (40, 50, 100).  English in Jayawardene.

Calandri.  Aritmetica.  c1485.  Ff. 89r-89v, pp. 178‑179.  (60, 80, 100).  (Tropfke, p. 599, shows the illustration in B&W.)

Calandri.  Arimethrica.  1491.  F. 100v.  Well between two towers.  (80, 90, 100).  Nice double size woodcut picture.

Pacioli.  Summa.  1494.  Part II.

F. 59v, prob. 62.  (70, 100, 150)  = Columbia Algorism.

Ff. 59v-60r, prob. 63.  A, B, D = 30, 40, 20  with rope of length  25  between the towers with a sliding lead weight on it.  How high is the weight from the ground. 

Ff. 61r-61v, prob. 66.  Three towers of heights  A, B, C = 125, 135, 125, with distances  AB, AC, BC = 150, 130, 140.  Find the point on the ground equidistant from the tops of the towers.

Cardan.  Practica Arithmetice.  1539.  Chap. 67.

Section 9, f. NN.vi.r (p. 197).  (40, 30, 50).

Section 10, ff. NN.vi.r - NN.vii.v (pp. 197-198).  Three towers of heights  A, B, C = 40, 30, 70, with distances  AB, AC, BC = 50, 60, 20.  Find the point on the ground equidistant from the tops of the towers.  Same idea as Pacioli, prob. 66.

Ozanam‑Montucla.  1778.  Vol. II, prob. 7 & fig. 5, plate 1.  1778: 11;  1803: 11-12;  1814: 9‑10;  1840: 199.  Rope between two towers with a pulley on it.  Locate the equilibrium position.  Uses reflection.

Carlile.  Collection.  1793.  Prob. XLV, p. 25.  Find the position for a ladder on the ground between two towers so that leaning it each way reaches the top of each tower.  (80, 91, 100).  He simply states how to do the calculation,  x = (D2 + A2 - B2)/2D  for the distance from the base of tower  B.  

T. Tate.  Algebra Made Easy.  Chiefly Intended for the Use of Schools.  New edition.  Longman, Brown, Green, and Longman, London, 1848.  P. 111.

No. 36.  A, B, D = 30, 70, 100.  Locate  P  such that the towers subtend the same angle, i.e. the two triangles are similar.  Clearly  P  divides  D  as  30  to  70.

No. 37.  Same data.  Locate  P  so the distance to the tops is the same.  This gives  70  to 30  easily because  A + B = D.

No. 38.  Same data.  Locate  P  so the difference of the squares of the distances is  400.  Answer is  68  from the base of the shorter tower.

Perelman.  MCBF.  1937.  Two birds by the riverside.  Prob. 136, pp. 224-225.  (30, 20, 50).  "A problem of an Arabic mathematician of the 11th century."

Richard J. Palmaccio.  Problems in Calculus and Analytic Geometry.  J. Weston Walch, Portland, Maine, 1977.  Maximum-Minimum Problems, No. 3, pp. 9 & 70-71.  Cable supported pulley device over a factory.  A, B, D  =  24, 27, 108  with cable of length  L = 117.  Find the lowest point.  He sets up the algebraic equations corresponding to the the two right triangles, assumes the distance from one post and the height above ground are implicit functions of the length of the cable from the pulley to the same post and differentiates both equations and sets equal to zero, but it takes half a page to get to the answer and he doesn't notice that the two triangles are similar at the lowest point.

David Singmaster, proposer;  Dag Jonsson & Hayo Ahlburg, solvers.  Problem 1748: [The two towers].  CM 18:5 (1992) 140  &  19:4 (1993) 125-127.  Based on Pseudo-dell'Abbaco 158-159, but the solution by reflection was later discovered to be essentially Ozanam‑Montucla.

David Singmaster.  Symmetry saves the solution.  IN:  Alfred S. Posamentier & Wolfgang Schulz, eds.; The Art of Problem Solving: A Resource for the Mathematics Teacher; Corwin Press, NY, 1996, pp. 273-286.  Gives the reflection solution.

Yvonne Dold-Samplonius.  Problem of the two towers.  IN: Itinera mathematica; ed. by R. Franci, P. Pagli & L. Toti Rigatelli.  Siena, 1996.  Pp. 45-69.  ??NYR.  The earliest example she has found is Mahavira and it was an email from her in about 1995 that directed me to Mahavira.

 

          6.BF.4.        RAIL BUCKLING.

 

          A railway rail of length  L  and ends fixed expands to length  L + ΔL.  Assuming the rail makes two hypotenuses, the middle rises by a height,  H,  satisfying  H2 = {(L+ΔL)/2}2 ‑ (L/2)2,  hence  @ Ö(LΔL/2).

          However, one might assume the rail buckled into an arc of a circle of radius  r.  If we let the angle of the arc be  2θ,  then we have to solve     = (L + ΔL)/2;  r sin θ = L/2.  Taking  sin θ @ θ - θ3/6,  we get  r2 @ (L + ΔL)3/ 24 ΔL.  We have  H = r (1 - cos θ) @2/2  and combining this with earlier equations leads to  H @ Ö{3(L+ΔL)ΔL/8}  which is about  Ö3 / 2 = .866...  as big as the estimate in the linear case.

 

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  P. 149, prob. 12.  L = 1 mile,  ΔL = 1 ft or 2 ft -- text is not clear.  "Answer: More than 54 ft."  However, in the linear case,  ΔL = 1 ft  gives  H = 51.38 ft  and  ΔL = 2 ft  gives  H = 72.67 ft,  while the exact answers in the circular case are  44.50 ft  and  62.95  ft.

Sullivan.  Unusual.  1943.  Prob. 15: Workin' on the railroad.  L = 1 mile,  ΔL = 2 ft.  Answer:  about  73  ft.

Robert Ripley.  Mammoth Believe It or Not.  Stanley Paul, London, 1956.  If a railroad rail a mile long is raised  200  feet in the centre, how much closer would it bring the two ends?  I.e.  L = 1 mile,  H = 200 ft.  Answer is:  "less than  6  inches".  I am unable to figure out what Ripley intended.

Jonathan Always.  More Puzzles to Puzzle You.  Tandem, London, 1967.  Gives the same question as Ripley with answer "approximately  15  feet".  The exact answer is  15.1733.. feet  or  15 feet 2.08 inches. 

David Singmaster, submitter.  Gleaning:  Diverging lines.  MG 69 (No. 448) (Jun 1985) 126.  Quotes from Ripley and Always.

David Singmaster.  Off the rails.  The Weekend Telegraph (18 Feb 1989) xxiii  &  (25 Feb 1989) xxiii.  Gives the Ripley and Always results and asks which is correct and whether the wrong one can be corrected -- cf Ripley above.

Phiip Cheung.  Bowed rail problem.  M500 161 (?? 1998) 9.  ??NYS.  Paul Terry, Martin S. Evans, Peter Fletcher, solvers and commentators.  M500 163 (Aug 1998)  10-11.  L = 1 mile,  ΔL = 1 ft.  Terry treats the bowed rail as circular and gets  H = 44.49845 ft.  Evans takes  L = 1 nautical mile of 6000 ft and gets almost exactly  H = 50 ft.  Fletcher says it took 15 people to lift a 60ft length of rail, so if someone lifted the 1 mile rail to insert the extra foot, it would need about 1320 people to do the lifting.

 

          6.BF.5.        TRAVELLING ON SIDES OF A RIGHT TRIANGLE.

 

          New section.  See also the Mittenzwey example in 10.A.6.

 

Brahmagupta.  Brahma‑sphuta‑siddhanta.  628.  Chap. XII, sect. IV, v. 39.  In Colebrooke, p. 308.  Rule for the problem illustrated by Chaturveda.

Mahavira.  850.  Chap. VII, v. 210-211, pp. 251-252.  A slower traveller goes due east at rate  v.  A faster traveller goes at rate  V  and starts going north.  After time  t,  he decides to meet the other traveller and turns so as to go directly to their meeting point.  How long,  T,  do they travel?  This gives us a right triangle with sides  vT,  Vt,  V(T-t)  leading to a quadratic in  T  whose constant term drops out, yielding  T = 2t V2/(V2-v2).  If we set  r = v/V,  then  T = 2t/(1-r2),  so we can determine  T  from  t  and  r  without actually knowing  V  or  v.  Indeed, if we let  ρ = d/D,  we get  2ρ = 1 - r2.   v, V, t  =  2, 3, 5.

Chaturveda.  860.  Commentary to the Brahma‑sphuta‑siddhanta, chap. XII, section IV, v. 39.  In Colebrooke, p. 308.  Two ascetics are at the top of a (vertical!) mountain of height  A.  One, being a wizard, ascends a distance  X  and then flies directly to a town which is distance  D  from the foot of the mountain.  The other walks straight down the mountain and to the town.  They travel at the same speeds and reach the town at the same time.  Example with  A, D = 12, 48.

Bhaskara II.  Lilavati.  1150.  Chap. VI, v. 154-155.  In Colebrooke, pp. 66-67.  Similar to Chaturveda.  Two apes on top of a tower of height  A  and they move to a point  D  away.  A, D = 100, 200.

Bhaskara II.  Bijaganita.  1150.  Chap. IV, v. 126.  In Colebrooke, pp. 204-205.  Same as Lilavati.

 

          6.BG. QUADRISECT A PAPER SQUARE WITH ONE CUT

 

          This involves careful folding.  One can also make  mn  rectangles with a single cut.

 

Mittenzwey.  1880.  Prob. 193 & 194, pp. 36 & 89;  1895?: 218 & 219, pp. 41 & 91;  1917: 218 & 219, pp. 37 & 87.  Make four squares.  Make four isosceles right triangles.

Walter Gibson.  Big Book of Magic for All Ages.  Kaye & Ward, Kingswood, Surrey, 1982.  Tick-tack-toe, pp. 68-69.  Take a  4 x 4  array and mark alternate squares with  Xs  and  Os.  By careful folding and cutting, one produces eight free squares and a connected lattice of the other eight squares, with the free squares being either all the  Os  or all the  Xs,  depending on how the final part of the cut is made.

David Singmaster.  Square cutting.  Used in my puzzle columns.

          Weekend Telegraph (18  &  25 Mar 1989) both p. xxiii.

          G&P, No. 16 (Jul 1995) 26.  (Publication ceased with No. 16.)

 

          6.BH. MOIRÉ PATTERNS

 

          I have not yet found any real history of this topic.  One popular book says moirés were first made in 15C China.  The OED has several entries for Moire and Moiré.  It originally refers to a type of cloth and may be a French adaptation of the English word mohair -- Pepys refers to 'greene-waterd moyre' and this is the earliest citation.  In the early 19C, the term began to be used for the 'watered' effect on cloth and metal.  At some point, the term was transferred to the optical phenomena, but the OED does not have this meaning.

 

Journal of Science and Arts 5 (1818) 368.  On the Moiré Metallique, or Fer blanc moiré.  ??NYS -- cited in OED as their first citation for the noun use of the term.

John Badcock.  Domestic Amusements, or Philosophical Recreations, ... Being a Sequel Volume to Philosophical Recreations, or Winter Amusements.  T. Hughes, London, nd [Preface dated Feb 1823].  [BCB 16-17; OCB, pp. 180 & 196.  Heyl 21.  Toole Stott 78‑80.  Wallis 34 BAD.  HPL [Badcock].  These give dates of 1823, 1825, 1828.]  Pp. 139-141, no. 169: Moiré Metal, or Crystallised Tin & no. 170: Moiré Watering, by other Methods.  "Quite new and splendid as this art is, ....  M. Baget, a Frenchman, however, claims the honour of a discovery of this process, attributing the same to accident, ...."  Cited in the OED as the first adjectival use of the term, though the previous entry seems to also have an adjectival usage.

Rational Recreations.  1824.  Experiment 16, p. 15: Metallic watering, or, fer blanc moire.  Says it is of Parisian invention and gives the method of applying sulphuric acid to tin.

Endless Amusement II.  1826?  Pp. 24-25: Application of the moiré métallique to tin-foil.  This deals with obtaining a moiré effect in tin-foil and is quite different than Badcock.

Young Man's Book.  1839.  Pp. 312-314.  Identical to Endless Amusement II.

Tom Tit, vol. 2.  1892.  Le papier-canevas et les figures changeantes, pp. 137-138.  Uses perforated card.

Hans Giger.  Moirés.  Comp. & Maths. with Appls. 12B:1/2 (1986) [= I. Hargittai, ed., Symmetry -- Unifying Human Understanding, as noted in 6.G.] 329‑361.  Giger says the technique of moiré fabrics derives from China and was first introduced into France in 1754 by the English manufacturer Badger (or Badjer).  He also says Lord Rayleigh was the first to study the phenomenon, but gives no references.

 

          6.BI.  VENN DIAGRAMS FOR  N  SETS

 

          New topic.  I think I have seen more papers on this and Anthony Edwards has recently sent several more papers.

 

Martin Gardner.  Logic diagrams.  IN:  Logic Machines and Diagrams; McGraw‑Hill, NY, 1958, pp. 28-59.  Slightly amended in the 2nd ed., Univ. of Chicago Press, 1982, and Harvester Press, Brighton, 1983, pp. 28-59.  This surveys the history of all types of diagrams.  John Venn [Symbolic Logic, 2nd ed., ??NYS] already gave Venn diagrams with 4 ovals and with 4 ovals and a disconnected set.  Gardner describes various binary diagrams from 1881 onward, but generalised Venn diagrams seem to first occur in 1909 and then in 1938-1939, before a surge of interest from 1959.  His references are much expanded in the 2nd ed. and he cites most of the following items.

John Venn.  On the diagrammatic and mechanical representation of propositions and reasonings.  London, Edinburgh and Dublin Philos. Mag. 10 (1880) 1-18.  ??NYS -- cited by Henderson.

John Venn.  Symbolic Logic.  2nd ed., Macmillan, 1894.  ??NYS.  Gardner, p. 105, reproduces a four ellipse diagram.

Lewis Carroll.  Symbolic Logic, Part I.  4th ed., Macmillan, 1897;  reprinted by Dover, 1958.  Appendix -- Addressed to Teachers, sections 5 - 7: Euler's method of diagrams; Venn's method of diagrams; My method of diagrams, pp. 173-179.  Describes Euler's simple approach and Venn's thorough approach.  Reproduces Venn's four-ellipse diagram and his diagram for five sets using four ellipses and a disconnected region.  He notes that Venn suggests using two five-set diagrams to deal with six sets and does not go further.  He then describes his own method, which easily does up to eight sets.  The diagram for four sets is the same as the common Karnaugh diagram used by electrical engineers.  For more than four sets, the regions become disconnected with the cells of the four-set case being subdivided, using a simple diagonal, then his 2-set, 3-set and 4-set diagrams within each cell of the 4-set case.  ?? -- is this in the 1st ed. -- ??NYS  date??

                    Carroll-Gardner, p. 61, says this is in the first ed. of 1896.

William E. Hocking.  Two extensions of the use of graphs in elementary logic.  University of California Publications in Philosophy 2:2 (1909) 31(-??).  ??NYS -- cited by Gardner who says Hocking uses nonconvex regions to get a solution for any  n.

Edmund C. Berkeley.  Boolean algebra and applications to insurance.  Record of the Amer. Inst. of Actuaries 26:2 (Oct 1937)  &  27:1 (Jun 1938).  Reprinted as a booklet by Berkeley and Associates, 1952.  ??NYS -- cited by Gardner.  Uses nonconvex sets.

Trenchard More Jr.  On the construction of Venn diagrams.  J. Symbolic Logic 24 (Dec 1959) 303-304.  ??NYS -- cited by Gardner.  Uses nonconvex sets.

David W. Henderson.  Venn diagrams for more than four classes.  AMM 70:4 (1963) 424-426.  Gives diagrams with  5  congruent irregular pentagons and with  5  congruent quadrilaterals.  Considers problem of finding diagrams that have  n-fold rotational symmetry and shows that then  n  must be a prime.  Says he has found an example for  n = 7,  but doesn't know if examples can be found for all prime  n.

Margaret E. Baron.  A note on the historical development of logic diagrams: Leibniz, Euler and Venn.  MG 53 (No. 384) (May 1969) 113‑125.  She notes Venn's solutions for  n = 4, 5.  She gives toothed rectangles for  n = 5, 6.

K. M. Caldwell.  Multiple‑set Venn diagrams.  MTg 53 (1970) 29.  Does  n = 4  with rectangles and then uses indents.

A. K. Austin, proposer;  Heiko Harborth, solver.  Problem E2314 -- Venn again.  AMM 78:8 (Oct 1971) 904  &  79:8 (Oct 1972) 907-908.  Shows that a diagram for  4  or more sets cannot be formed with translates of a convex set, using simple counting and Euler's formula.  (The case of circles is in Yaglom & Yaglom I, pp. 103-104.)  Editor gives a solution of G. A. Heuer with  4  congruent rectangles and more complex examples yielding disconnected subsets.

Lynette J. Bowles.  Logic diagrams for up to n classes.  MG 55 (No. 394) (Dec 1971) 370‑373.  Following Baron's note, she gives a binary tooth‑like structure with examples for  n = 7, 8.

Vern S. Poythress & Hugo S. Sun.  A method to construct convex connected Venn diagrams for any finite number of sets.  Pentagon (Spring 1972) 80-83.  ??NYS -- cited by Gardner.

S. N. Collings.  Further logic diagrams in various dimensions.  MG 56 (No. 398) (Dec 1972) 309‑310.  Extends Bowles.

Branko Grünbaum.  Venn diagrams and independent families of sets.  MM 48 (1975) 12‑22.  Considers general case.  Substantial survey of different ways to consider the problem.  References to earlier literature.  Shows one can use 5 identical ellipses, but one cannot use ellipses for  n > 5.

B. Grünbaum.  The construction of Venn diagrams.  CMJ 15 (1984) 238‑247.  ??NYS.

Allen J. Schwenk.  Venn diagram for five sets.  MM 57 (1984) 297.  Five ovals in a pentagram shape.

A. V. Boyd.  Letter:  Venn diagram of rectangles.  MM 58 (1985) 251.  Does  n = 5  with rectangles.

W. O. J. Moser & J. Pach.  Research Problems in Discrete Geometry.  Op. cit. in 6.T.  1986.  Prob. 27: On the extension of Venn diagrams.  Considers whether a diagram for  n  classes can be extended to one for  n+1  classes.

Mike Humphries.  Note 71.11:  Venn diagrams using convex sets.  MG 71 (No. 455) (Mar 1987) 59.  His fourth set is a square;  fifth is an octagon.

J. Chris Fisher, E. L. Koh & Branko Grünbaum.  Diagrams Venn and how.  MM 61 (1988) 36‑40.  General case done with zig‑zag lines.  References.

Anthony W. F. Edwards.  Venn diagrams for many sets.  New Scientist 121 (No. 1646) (7 Jan 1989) 51-56.  Discusses history, particularly Venn and Carroll, the four set version with ovals and Carroll's four set version where the third and fourth sets are rectangles.  Edwards' diagram starts with a square divided into quadrants, then a circle.  Fourth set is a two-tooth 'cogwheel' which he relates to a Hamiltonian circuit on the 3-cube.  The fifth set is a four-tooth cogwheel, etc.  The result is rather pretty.  Edwards notes that the circle in the  n  set diagram meets the  2n-1  subsets of the  n-1  sets other than that given by the circle, hence travelling around the circle gives a sequence of the subsets of  n-1  objects and this is the Gray code (though he attributes this to Elisha Gray, the 19C American telephone engineer -- cf 7.M.3).  The relationship with the  n-cube leads to a partial connection between Edwards' diagram and the lattice of subsets of a set of  n  things.

                    New Scientist (11 Feb 1989) 77 has:  Drawing the lines -- letters from Michael Lockwood -- describing a version with indented rectangles -- and from Anthony Edwards -- noting some errors in the article.

Ian Stewart.  Visions mathématiques: Les dentelures de l'esprit.  Pour la Science No. 138 (Apr 1989) 104-109.  c= Cogwheels of the mind, IN:  Ian Stewart; Another Fine Math You've Got Me Into; Freeman, NY, 1992, chap. 4, pp. 51-64.  Exposits Edwards' work with a little more detail about the connection with the Gray code.

A. W. F. Edwards  &  C. A. B. Smith.  New 3-set Venn diagram.  Nature 339 (25 May 1989) 263.  Notes connection with the family of cosine curves,  y = 2-n cos 2nx  on  [0, π]  and Gray codes  and  with the family of sine curves,  y = 2-n sin 2nx  on [0, 2π]  and ordinary binary codes.  Applying a similar phase shift to Edwards' diagram leads to diagrams where more than two set boundaries are allowed to meet at a point.

A. W. F. Edwards.  Venn diagrams for many sets.  Bull. Intern. Statistical Inst., 47th Session, Paris, 1989.  Contrib. Papers, Book 1, pp. 311-312.

A. W. F. Edwards.  To make a rotatable Edwards map of a Venn diagram.  4pp of instructions and cut-out figures.  The author, Gonville and Caius College, Cambridge, CB2 1TA, 21 Feb 1991.

A. W. F. Edwards.  Note 75.39:  How to iron a hypercube.  MG 75 (No. 474) (1991) 433-436.  Discusses his diagram and its connection with the  n-cube.

Anthony Edwards.  Rotatable Venn diagrams.  Mathematics Review 2:3 (Feb 1992) 19-21.  +  Letter:  Venn revisited.  Ibid. 3:2 (Nov 1992) 29.

 

          6.BJ.  3D DISSECTION PUZZLES

 

          This will cover a number of cases which are not very mathematical.  I will record just some early examples.  See also 6.G (esp. 6.G.1), 6.N, 6.W (esp. 6.W.7), 6.AP for special cases.  The predecessors of these puzzles seem to be the binomial and trinomial cubes showing  (a+b)3  and  (a+b+c)3,  which I have placed in 6.G.1.  Cube dissections with cuts at angles to the faces were common in the 19C Chinese puzzle chests, often in ivory.  I have only just started to notice these.  It is hard to distinguish items in this section from other burr puzzles, 6.W.7, and I have tried to avoid repetition, so one must also look at that section when looking at this section.

 

Catel.  Kunst-Cabinet.  1790.  Der Vexierwürfel, p. 11 & fig. 32 on plate II.  Figure shows that there are some cuts at angles to the faces, so this is not an ordinary cube dissection, but is more like the 19C Chinese dissected cubes.

C. Baudenbecher.  Sample book or catalogue from c1850s.  Op. cit. in 6.W.7.  One whole folio page shows about 20 types of wooden interlocking puzzles, including most of the types mentioned in this section and in 6.W.5 and 6.W.7.  Until I get a picture, I can't be more specific.

Slocum.  Compendium.  Shows:  Wonderful "Coffee Pot";  Magic "Apple";  Magic "Pear";  Extraordinary "Cube";  Magic "Tub"  from Mr. Bland's Illustrated Catalogue of Extraordinary and Superior Conjuring Tricks, etc.; Joseph Bland, London, c1890.  He shows further examples from 1915 onward.

Hoffmann.  1893.  Chap. II, pp. 107-108 & 141-142 = Hoffmann-Hordern, pp. 106-107, with photo.

No. 37: The Fairy Tea-Table.  Photo on p. 107 shows a German example, 1870-1895.

No. 38: The Mystery.  Photo on p. 107 shows a German example, 1870-1895, with instructions.

Western Puzzle Works, 1926 Catalogue.

No. 5075.  Unnamed -- Fairy Tea-Table.

Last page shows 20 Chinese Wood Block Puzzles, High Grade.  These are unnamed, but the shapes include various burr-like objects, cube, spheres, egg, barrel, tankard, pear and apple.

P. M. Grundy.  A three-dimensional jig-saw.  Eureka 7 (Mar 1942) 8-10.  Consider a  2 x 2 x 2  array of unit cubes.  He suggests removing and/or adding lumps on the interior faces to make a jig-saw.  He then considers lumps in the form of a isosceles right triangular prism with largest face being a unit square.  He finds there are 25 such pieces, subject to the conditions that there are most two removals, and that when there are two, they must be parallel.  He gives a graphical view of a  2 x 2 x 2  formed from such pieces which gives some necessary, but not sufficient, conditions for a set of eight such pieces to be able to make a cube.  One solution is shown.  [If one assumes there is just one removal and one addition, I find just four pieces, which form a  1 x 2 x 2  block.  One could view this as a 3-dimensional matching puzzle, where the internal faces have to match both like a head to tail matching, but also with correct orientation.  MacMahon thought of notching pieces as an alternative to colouring edges, but his pieces were two-dimensional.]

 

          6.BK  SUPERELLIPSE

 

          New section.

 

Gardner.  SA (Sep 1965) = Carnival, chap. 18.  Describes how the problem arose in the design of Sergel's Square, Stockholm, in 1959.  Addendum in Carnival gives the results given by Gridgeman, below.  Also says road engineers used such curves with  n = 2.2,  called '2.2 ellipses', from the 1930s for bridge arches.

N. T. Gridgeman.  Lamé ovals.  MG 54 (No. 387) (Feb 1970) 31‑37.  Lamé (c1818) seems to be the first to consider  (x/a)n + (y/b)n = 1.  Hein's design in Stockholm uses  a/b = 6/5  and  n = 5/2.  Gerald Robinson used  a/b = 9/7  and  n = 2.71828...,  which he determined by a survey asking people which shape they liked most.  Gridgeman studies curvature, area, perimeter, evolutes, etc.

 

          6.BL. TAN-1 ⅓ + TAN-1 ½  =  TAN-1 1, ETC.

 

          This problem is usually presented with three squares in a row with lines drawn from one corner to the opposite corners of the squares.  New section.  Similar formulae occur in finding series for  π.  See 6.A and my Chronology of π.

 

L. Euler.  Introductio in Analysin Infinitorum.  Bousquet, Lausanne, 1748.  Vol. 1, chap. VIII, esp. § 142 (??NYS).  = Introduction to the Analysis of the Infinite; trans. by John D. Blanton; Springer, NY, 1988-1990; Book I, chap. VIII: On transcendental quantities which arise from the circle, pp. 101-115, esp. § 142, pp. 114-115.  Developing series to calculate  π,  he considers angles  a, b  such that  a + b = π/4,  then examines the formula for  tan (a + b)  and says:  "If we let  tan a = ½,  then  tan b = ⅓ ....  In this way we calculate ...  π,  with much more ease than ... before."  Conway & Guy give some more details.

Carroll.  ?? -- see Lowry (1972) and Conway & Guy (1996).

Størmer.  1896.  See Conway & Guy.

Gardner.  SA (Feb 1970) = Circus, chap. 11, prob. 3.  Says he received the geometric problem from Lyber Katz who had been given it when he was in 4th grade in Moscow. 

C. W. Trigg.  A three‑square geometry problem.  JRM 4:2 (Apr 1971) 90‑99.  Quotes a letter from Katz, dating his 4th year as 1931-32.  Trigg sketches 54 proofs of the result, some of which generalize.

H. V. Lowry.  Note 3331:  Formula for π/4.  MG 56 (No. 397) (Oct 1972) 224-225.  tan‑1 1/a  =  tan-1 1/b + tan-1 1/c  implies  a(b+c)  =  bc - 1,  hence  (b‑a)(c‑a)  =  a2 + 1,  whence all integral solutions can be determined.  Conway & Guy say this was known to Lewis Carroll.

J. R. Goggins & G. B. Gordon.  Note 3346:  Formula for π/4 (see Note 3331, Oct 1972).  MG 57 (No. 400) (Jun 1973) 134.  Goggins gets  π/4  =  Σn=1 tan-1 1/F2n+1,  where  Fn  is the  n‑th Fibonacci number.  [I think this formula was found by Lehmer some years before??]  Gordon also mentions Eureka No. 35, p. 22, ??NYS, and finds recurrences giving  tan-1 1/pn + tan-1 1/qn  =  tan-1 1/rn.

Douglas A. Quadling.  Classroom note 304:  The story of the three squares (continued).  MG 58 (No. 405) (Oct 1974) 212‑215.  The problem was given in Classroom note 295 and many answers were received, including four proofs published by Roger North.  Quadling cites Trigg and determines which proofs are new.  Trigg writes that tan‑1 1/F2n+2 + tan-1 1/F2n+1  =  tan-1 1/F2n,  which is the basis of Goggins' formula.

Alan Fearnehough.  On formulas for  π  involving inverse tangent functions  and  Prob. 23.7.  MS 23:3 (1990/91) 65-67 & 95.  Gives four basic theorems about inverse tangents leading to many different formulae for  π/4.  The problem gives a series using inverse cotangents.

John H. Conway & Richard K. Guy.  The Book of Numbers.  Copernicus (Springer-Verlag), NY, 1996.  Pp. 241-248 discusses relationships among values of  tan-1 1/n  which they denote as  tn  and call Gregory's numbers.  Euler knew  t = t2 + t3,  t1 = 2 t3 + t7  and  t1 = 5 t7 + 2 t18 - 2 t57  and used them to compute  π  to 20 places in an hour.  They say Lewis Carroll noted that  tn = tn+c + tn+d  if and only if  cd = n2 + 1.  In 1896, Størmer related Gaussian integers to Gregory numbers and showed how to obtain a Gregory number as a sum of other Gregory numbers.  From this it follows that the only two-term expressions for  π/4  are  t2 + t3,  2 t2 - 4 t7,  2 t3 + t7  and  4 t5 - t239.  This is described in Conway & Guy, but they have a misprint of  8  for  18  at the bottom of p. 246.

 

          6.BM.          DISSECT CIRCLE INTO TWO HOLLOW OVALS

 

          Consider a circle of radius  2.  Cut it by two perpendicular diameters and by the circle of radius  1  about the centre.  Two of the outer pieces (quarters of the annulus) and two of the inner pieces (quadrants) make an oval shape, with a hollow in the middle.  The problem often refers to making two oval stools and the hollows are handholds!  After the references below, the problem appears in many later books.

 

Jackson.  Rational Amusement.  1821.  Geometrical Puzzles, no. 9, pp. 25 & 84 & plate I, fig. 6.  Mentions handholes.  Solution is well drawn.

Endless Amusement II.  1826?  Mentions handholes.  Solution is well drawn.

Crambrook.  1843.  P. 4, no. 6: A Circle to form two Ovals.  Check??

Magician's Own Book.  1857.  Prob. 36: The cabinet maker's puzzle, pp. 277 & 300.  Solution is a bit crudely drawn.  = Book of 500 Puzzles, 1859, pp. 91 & 114.  = Boy's Own Conjuring Book, 1860, prob. 35, pp. 240 & 265.

Family Friend (Dec 1858) 359.  Practical puzzle -- 4.  I don't have the answer.

The Secret Out.  1859.  The Oval Puzzle, pp. 380-381.  Asks to 'produce two perfect ovals.'  Solution is a bit crudely drawn, as in Magician's Own Book, but the text and numbering of pieces is different.

Illustrated Boy's Own Treasury.  1860.  Prob. 34, pp. 401 & 441.  Same crude solution as Magician's Own Book, but with different text, neglecting to state that the stools have handholes in their centres.

Magician's Own Book (UK version).  1871.  On p. 282, in the middle of an unrelated problem, is the solution diagram, very poorly drawn -- the pieces of the oval stools are shown as having curved edges almost as though they were circular arcs.  There is no associated text.

Hanky Panky.  1872.  The oval puzzle, p. 123.  Same crude solution as Magician's Own Book, but different text, mentioning handholes.

Mittenzwey.  1880.  Prob. 256, pp. 46 & 97;  1895?: 285, pp. 50 & 99;  1917: 285, pp. 45 & 94.  The stools are very poorly drawn, with distinct wiggles in what should be straight lines.

Hoffmann.  1893.  Chap. II, no. 32: The cabinet maker's puzzle, pp. 104 & 137‑138 = Hoffmann-Hordern, p. 102.  Mentions hand holes.  Well drawn solution.

Benson.  1904.  The cabinet‑maker's puzzle, p. 201.  Mentions hand holes.  Poor drawing.

 

          6.BN. ROUND PEG IN SQUARE HOLE OR VICE VERSA

 

Wang Tao‑K'un.  How to get on.  Late 16C.  Excerpted and translated in:  Herbert A. Giles; Gems of Chinese Literature; 2nd ed. (in two vols., Kelly & Walsh, 1923), in one vol., Dover, 1965, p. 226.  "... like square handles which you would thrust into the round sockets ..."

Sydney Smith.  Sketches of Moral Philosophy.  Lecture IX.  1824.  "If you choose to represent the various parts in life by holes upon a table, of different shapes, -- some circular, some triangular, some square, some oblong, -- and the persons acting these parts by bits of wood of similar shapes, we shall generally find that the triangular person has got into the square hole, the oblong into the triangular, and a square person has squeezed himself into the round hole.  The officer and the office, the doer and the thing done, seldom fit so exactly that we can say they were almost made for each other."  Quoted in:  John Bartlett; Familiar Quotations; 9th ed., Macmillan, London, 1902, p. 461 (without specifying the Lecture or date).  Irving Wallace; The Square Pegs; (Hutchinson, 1958); New English Library, 1968; p. 11, gives the above quote and says it was given in a lecture by Smith at the Royal Institution in 1824.  Bartlett gives a footnote reference:  The right man to fill the right place -- Layard: Speech, Jan. 15, 1855.  It is not clear to me whether Layard quoted Smith or simply expressed the same idea in prosaic terms .  Partially quoted, from 'we shall ...' in The Oxford Dictionary of Quotations; 2nd ed. revised, 1970, p. 505, item 24.  Similarly quoted in some other dictionaries of quotations. 

I have located other quotations from 1837, 1867 and 1901.

William A. Bagley.  Paradox Pie.  Vawser & Wiles, London, nd [BMC gives 1944].  No. 17: Misfits, p. 18.  "Which is the worst misfit, a square peg in a round hole or a round peg in a square hole?"  Shows the round peg fits better.  He notes that square holes are hard to make.

David Singmaster.  On round pegs in square holes and square pegs in round holes.  MM 37 (1964) 335‑337.  Reinvents the problem and considers it in  n  dimensions.  The round peg fits better for  n < 9.  John L. Kelley pointed out that there must be a dimension between  8  and  9  where the two fit equally well.  Herman P. Robinson kindly calculated this dimension for me in 1979, getting  8.13795....

David Singmaster.  Letter:  The problem of square pegs and round holes.  ILEA Contact [London] (12 Sep 1980) 12.  The two dimensional problem appears as a SMILE card which was attacked as 'daft' in an earlier letter.  Here I defend the problem and indicate some extensions -- e.g. a circle fits better in a regular  n‑gon than vice‑versa for all  n.

 

          6.BO. BUTTERFLY PROBLEM

 

          I have generally avoided classical geometry of this sort, but Bankoff's paper deserves inclusion.

 

Leon Bankoff.  The metamorphosis of the butterfly problem.  MM 60 (1987) 195‑210.  Includes historical survey of different proofs.  The name first appears in the title of the solution of Problem E571, AMM 51 (1944) 91 (??NYS).  The problem first occurs in The Gentlemen's Diary (1815) 39‑40 (??NYS).

 

          6.BP. EARLY MATCHSTICK PUZZLES

 

          There are too many matchstick puzzles to try to catalogue.  Some of them occur in other sections, e.g. 6.AO.1.  Here I only include a few very early other examples.  At first I thought these would date from mid to late 19C when matches first started to become available, but the earliest examples refer to slips of paper or wood.  The earliest mention of matches is in 1858.

 

Rational Recreations.  1824.  Exer. 23, p. 132.  Double a sheep pen by adding just two hurdles.  I have just realised this is a kind of matchstick puzzle and I suspect there are other early examples of this.  It is a little different than most matchstick puzzles in that one is usually not given an initial pattern, but must figure it out.  (A hurdle is a kind of panel woven from sticks or reeds used by shepherds to make temporary pens.)

Family Friend 2 (1850) 148 & 179.  Practical Puzzle -- No. V.  = Illustrated Boy's Own Treasury, 1860, Prob. 46, pp. 404 & 443.  "Cut seventeen slips of paper or wood of equal lengths, and place them on a table, to form six squares, as in the diagram.  ...."

Magician's Own Book.  1857.  Prob. 20: Three square puzzle, pp. 273 & 296.  (I had 87 & 110 ??)  Almost identical to Family Friend, with a few changes in wording and a different drawing, e.g. "Cut seventeen slips of cardboard of equal lengths, and place them on a table to form six squares, as in the diagram.   ...." 

The Sociable.  1858.  Prob. 2: The magic square, pp. 286 & 301.  "With seventeen pieces of wood (lucifer matches will answer the purpose, but be careful to remove the combustible ends, and see that they are all of the same length) make the following figure:  [a  2 x 3  array of squares]", then remove 5 matches to leave three squares. 

Book of 500 Puzzles.  1859. 

Prob. 2: The magic square, pp. 4 & 19.  As in The Sociable.

Prob. 20: Three square puzzle, pp. 87 & 110.  Identical to Magician's Own Book.

Boy's Own Conjuring Book.  1860.  Three-square puzzle, pp. 235 & 259.  Identical to Magician's Own Book.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 5: The three squares, pp. 28 & 31.  Almost identical to Magician's Own Book, prob. 20, with a slightly different diagram.

Mittenzwey.  1880.  Prob. 156-171, 202-212, 240, 242, pp. 32-33, 37-38, 44 & 83-85, 90, 94;  1895?: 179-196, 227-237, 269, 271, pp. 37-38, 41-42, 48 & 85-87, 92, 96;  1917: 179‑196, 227-237, 269, 271, pp. 33-34, 38-39, 44 & 82-84, 88, 92.  This is the first puzzle book to have lots of matchstick problems, though he doesn't yet use the name, calling them 'Hölzchen' (= sticks).  Several problems occur elsewhere, e.g. in 6.AO.1.  The last problem is the same as in Jackson.

Sophus Tromholt.  Streichholzspiele.  Otto Spamer, Leipzig, 1889;  5th ed., 1892;  14th ed., Leipzig, 1909;  slightly revised and with a Preface by Rüdiger Thiele, Zentralantiquariat der DDR, Leipzig, 1986;  Hugendubel, 1986.  [Christopher 1017 is Spamer, 1890.  C&B give 1890.  There is an edition by Ullstein, Frankfurt, 1990.]  This is the earliest book I know which is entirely devoted to matchstick puzzles.

Gaston Tissandier.  Jeux et Jouets du jeune age  Choix de récréations  amusantes & instructives.  Ill. by Albert Tissandier.  G. Masson, Paris, nd [c1890].  P. 40, no. 4-5: Le problème des allumettes.  Make five squares with nine matches.  Solution has four small squares and one large one.

Clark.  Mental Nuts.  1897, no. 56.  The toothpicks.  Use 12 toothpicks to make a  2 x 2  array of squares.  Move three picks to form three squares.

H. D. Northrop.  Popular Pastimes.  1901.  No. 1: The magic square, pp. 65 & 71.  = The Sociable.

 

          6.BQ. COVERING A DISC WITH DISCS

 

          The general problem is too complex to be considered recreational.  Here I will mainly deal with the carnival version where one tries to cover a circular spot with five discs.  In practice, this is usually rigged by stretching the cloth.

 

Eric H. Neville.  On the solution of numerical functional equations, illustrated by an account of a popular puzzle and of its solution.  Proc. London Math. Soc. (2) 14 (1915) 308-326.  Obtains several possible configurations, but says "actual trial is sufficient to convince" that one is clearly the best, namely the elongated pentagon with 2-fold symmetry.  This leads to four trigonometric equations in four unknown angles which theoretically could be solved, but are difficult to solve even numerically.  He develops a modification of Newton's method and applies it to the problem, obtaining the maximal ratio of spot radius to disc radius as  1.64091.  Described by Gardner and Singleton.

Ball.  MRE.  10th ed., 1922.  Pp. 253-255: The five disc problem.  Sketches Neville's results.

Will Blyth.  More Paper Magic.  C. Arthur Pearson, London, 1923.  Cover the spot, pp. 66-67.  "This old "fun of the fair" game has been the means of drawing many pennies from the pockets of frequenters of fairs."  Says the best approach is an elongated pentagon which has only 2-fold symmetry.

William Fitch Cheney Jr, proposer;  editorial comment.  Problem E14.  AMM 39 (1932) 606  &  42 (1935) 622.  Poses the problem.  Editor says no solution of this, or its equivalent, prob. 3574, was received, but cites Neville. 

J. C. Cannell.  Modern Conjuring for Amateurs.  C. Arthur Pearson, London, nd [1930s?].  Cover the spot, pp. 132-134.  Uses discs of diameter  1 5/8 in  to cover a circle of diameter  2 1/2 in.  This is a ratio of  20/13 = 1.538...,  which should be fairly easy to cover??  His first disc has its edge passing through the centre of the circle.  His covering pattern has bilateral symmetry, though the order of placing the last two discs seems backward.

Walter B. Gibson.  The Bunco Book.  (1946);  reprinted by Citadel Press (Lyle Stuart Inc.), Secaucus, New Jersey, 1986.  Spotting the spot, pp. 24-25.  Also repeated in summary form, with some extra observations in:  Open season on chumps, pp. 97-106, esp. pp. 102‑103.  The circles have diameter  5"  and the discs "are slightly more than three inches in diameter."  He assumes the covering works exactly when the discs have five-fold symmetry -- which implies the discs are 3.090... inches in diameter -- but that the operator stretches the cloth so the spot is unsymmetric and the player can hardly ever cover it -- though it still can be done if one plays a disc to cover the bulge.  "There is scarcely one chance in a hundred that the spectator will start correctly ...."  On p. 103, he adds that "in these progressive times" the bulge can be made in any direction and that shills are often employed to show that it can be done, though it is still difficult and the operator generally ignores small uncovered bits in the shills' play in order to make the game seem easy.

Martin Gardner.  SA (Apr 1959)??  c= 2nd Book, chap. 13.  Describes the five disc version as Spot-the-Spot.  Cites Neville.

Colin R. J. Singleton.  Letter:  A carnival game -- covering disks with smaller disks.  JRM 24:3 (1992) 185-186.  Responding to a comment in JRM 24:1, he points out that the optimum placing of five discs does not have pentagonal symmetry but only bilateral.  Five discs of radius 1 can then cover a disc of radius  1.642..,  rather than  1.618...,  which occurs when there is pentagonal symmetry.  He cites Gardner and E. H. Neville.  His  1.642  arises because Gardner had truncated the reciprocal ratio to three places.

 

          6.BR. WHAT IS A GENERAL TRIANGLE?

 

David & Geralda Singmaster, proposers;  Norman Miller, solver.  Problem E1705 -- Skewness of a triangle.  AMM 71:6 (1964) 680  &  72:6 (1965) 669.  Assume  a £ b £ c.  Define the skewness of the triangle as  S = max {a/b, b/c, c/a} x min {a/b, b/c, c/a}.  What triangles have maximum and minimum skewness?  Minimum is  S = 1  for any isosceles triangle.  Maximum occurs for the degenerate triangle with sides  1, φ, 1+φ,  where  1 + φ = φ2,  so  φ = (1 + Ö5)/2  is the golden mean.

Baruch Schwarz & Maxim Bruckheimer.  Let  ABC  be any triangle.  MTr 81 (Nov 1988) 640-642.  Assume  AB < AC < BC  and  ÐA < 90o.  Drawing  BC  and putting  A  above it leads to a small curvilinear triangular region where  A  can be.  Making  A  equidistant from the three boundaries leads to a triangle with sides proportional to  Ö33, 7, 8  and with angles  44.5o,  58.5o,  77o.  The sides are roughly in the proportion  6 : 7 : 8.

Gontran Ervynck.  Drawing a 'general' triangle.  Mathematics Review (Nov 1991).  ??NYS -- cited by Anon., below.  Notes that if we take an acute triangle with angles as different as possible, then we get the triangle with angles  45o,  60o,  75o.

Anon. [possibly the editor, Tom Butts].  What is a 'general' triangle?  Mathematical Log 37:3 (Oct 1993) 1 & 6.  Describes above two results and mentions Guy's article in 8.C.  Gives an argument which would show the probability of an acute triangle is  0.

Anon. [probably the editor, Arthur Dodd].  A very scalene triangle.  Plus 30 (Summer 1995) 18-19 & 23.  (Content says this is repeated from a 1987 issue -- ??NYS.)  Uses the same region as Schwarz & Bruckheimer, below.  Looks for a point as far away from the boundaries as possible and takes the point which gives the angles  45o,  60o,  75o.

In 1995?, I experimented with variations on the definition of skewness given in the first item above, but have not gotten much.  However, taking  a = 1,  we have  £ b £ c £ b + 1.  Plotting this in the  b, c  plane gives us a narrow strip extending to infinity.  For generality, it would seem that we want  c = b + ½,  but there is no other obvious condition to select a central point in this region.  As fairly random points, I have looked at the case where  c = b2,  which gives  b = (1 + Ö3)/2 = 1.366..,  c = 1.866..  --  this triangle has angles about  31.47o,  45.50o,  103.03o  --  and at the case where  b = 3/2,  which gives a triangle with sides proportional to  2, 3, 4  with angles about  28.96o,  46.57o,  104.46o.  In Mar 1996, I realised that the portion of the strip corresponding to an acute triangle tends to  0 !!

I have now (Mar 1996) realised that the situation is not very symmetric.  Taking  c = 1,  we have  £ a £ b £ 1 £ a + b  and plotting this in the  a, b  plane gives us a bounded triangle with vertices at  (0, 1),  (½, ½),  (1, 1).  There are various possible central points of this triangle.  The centroid is at  (1/2, 5/6),  giving a triangle with sides  1/2, 5/6, 1  which is similar to  3, 5, 6,  with angles  29.93o,  56.25o,  93.82o.  An alternative point in this region is the incentre, which is at  (½, ½{-1 + 2Ö2}),  giving a triangle similar to  1, -1 + 2Ö2, 2  with angles  29.85o,  65.53o,  84.62o.  The probability of an acute triangle in this context is  2 - π/2  =  .429.

 

          6.BS. FORM SIX COINS INTO A HEXAGON

 

                                        O O O                                O O

          Transform                 O O O       into                O     O       in three moves.

                                                                                  O O

New section -- there must be older versions.

 

Young World.  c1960.  P. 13: Ringing the change.  He starts with the mirror image of the first array.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  The ring of coins, p. 31.  Says it is described by Gardner, but gives no details.  Notes that if you show the trick to someone

            O O O        and then give the coins in the mirror image pattern shown at the left, he will

          O O O          not be able to do it.

 

Putnam.  Puzzle Fun.  1978.

No. 12: Create a space, pp. 4 & 27.  With four coins, create the pattern                                          OO  

                              on the right.  [Takes two moves from a rhombic starting pattern.]                   O    O

No. 13: Create a space again, pp. 4 & 38.  Standard hexagon problem.

 

          6.BT. PLACING OBJECTS IN CONTACT

 

          New section.  The objects involved are usually common objects such as coins or cigarettes, etc.  The standard recreation is to have them all touching one another.  However, the more basic question of how many spheres can touch a sphere goes back to Kepler and perhaps the Greeks.  Similar questions have been asked about cubes, etc.

 

Endless Amusement II.  1826?  Problem II, p. 189.  "Five shillings or sixpences may be so placed over each other, as to be all visible and all be in contact."  Two solutions.  The first has two coins on the table, then two coins on top moved far enough onto one of the lower coins that a vertical coin can touch both of them and the two lower coins at once.  The second solution has one coin with two coins on top and two slanted coins sitting on the bottom coin and touching both coins in the second layer and then touching each other up in the air.  [I have recently read an article analysing this last solution and showing that it doesn't work if the coin is too thick and that the US nickel is too thick.]  = New Sphinx, c1840, p. 131.

Will Baffel.  Easy Conjuring  without apparatus.  Routledge  &  Dutton,  nd, 4th ptg [c1910], pp. 103-104.  Six matches, each touching all others.  Make a  V  with two matches and place a third match in the notch to make a short arrow.  Lie one of these on top of another.

Will Blyth.  Money Magic.  C. Arthur Pearson, London, 1926.  Five in contact, pp. 98-101.  Same as Endless Amusement II.

Rohrbough.  Puzzle Craft.  1932.  Six Nails, p. 22 (= p. 22 of 1940??).  As in Baffel.

Meyer.  Big Fun Book.  1940.  Five coins, p. 543.  Same as the first version in Endless Amusement II.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 29, pp. 35 & 92.  Second of the forms given in Endless Amusement II.

Ripley's Puzzles and Games.  1966.  P. 36.  Six cigarettes, as in Baffel.

I recall this is in Gardner and that a solution with 6 cigarettes was improved to 7.

 

          6.BU. CONSTRUCTION OF  N-GONS

 

          New Section.  This is really a proper geometric topic, but there is some recreational interest in two aspects.

          A.       Attempts to construct regular  n-gons for impossible values of  n,  e.g.  n = 7,  either by ruler and compass or by origami or by introducing new instruments -- see 6.BV.

          B.       Attempts to construct possible cases, e.g.  n = 5,  by approximate methods.

Aspect A is closely related to the classic impossible problems of trisecting an angle and duplicating a cube and hence some of the material occurs in books on mathematical cranks -- see Dudley.  Further, there were serious attempts on both aspects from classic times onward. 

 

Abu’l-Jūd.  11C.  He "devised a geometrical method to divide the circle into nine equal parts."  [Seyyed Hossein Nasr; Islamic Science -- an Illustrated Study; World of Islam Festival Publishing Co., London?, 1976, p. 82.  Q. Mushtaq & A. L. Tan; Mathematics: The Islamic Legacy; Noor Publishing House, Farashkhan, Delhi, 1993, p. 70.]

Pacioli.  De Viribus.  c1500.  These problems are discussed by Mackinnon, op. cit. in 6.AT.3, pp. 167, 169, citing Agostini, p. 5.  Let  ln  be the side of a regular  n-gon inscribed in a unit circle.

Ff. 146r-147r, XXIII afare la 7a fia dicta nonangolo. cioe de .9. lati difficile (XXIII to make the 7th figure called nonagon, that is of 9 sides, difficult)  = Peirani 198‑199.  Asserts  l9 = (l3 + l6)/4.  Mackinnon computes this gives  .6830  instead of the correct  .6840.

Ff. 148r-148v, XXV. Documento della 9 fia recti detta undecagono (XXV. on the 9th rectilinear figure called undecagon)  = Peirani 200.  Asserts  l11 = φ (l3 + l6)/3,  where  φ  is the golden mean:  (1 - Ö5)/2.  Mackinnon computes this gives  .5628  instead of the correct  .5635.

F. 148v, XXVI. Do. de' .13. (XXVI. on the 13th)  = Peirani 200.  Asserts  l13 = (1 φ)·5/4.  Mackinnon computes this gives  .4775  instead of the correct  .4786.

Ff. 149r-149v, XXVIII. Documento del .17. angolo cioe fia de .17. lati (XXVIII on the 17-angle, that is the figure of 17 sides)  = (Peirani 201-202).  Peirani says some words are missing in the second sentence of the problem and Agostini says the text is too corrupt to be reconstructed.  MacKinnon suggests  l17 =  (l3 - l6)/2  which gives  .3660  instead of the correct  .3675.

Barbaro, Daniele.  La Practica della Perspectiva.  Camillo & Rutilio Borgominieri, Venice, (1569);  facsimile by Arnaldo Forni, 1980, HB.  [The facsimile's TP doesn't have the publication details, but they are given in the colophon.  Various catalogues say there are several versions with dates on the TP and colophon varying independently between 1568 and 1569.  A version has both dates being 1568, so this is presumed to be the first appearance.  Another version has an undated title in an elaborate border and this facsimile must be from that version.]  Pp. 26-27 includes discussion of constructing a regular heptagon, but it just seems to say to divide the circumference of a circle into seven equal parts -- ??

Christian Huygens.  Oeuvres Complètes.  Vol. 14, 1920, pp. 498-500: problem dated 1662, "To inscribe a regular heptagon in a circle."  ??NYS -- discussed by Archibald.

R. C. Archibald.  Notes (to Problems and Solutions section) 24:  Problems discussed by Huygens.  AMM 28 (1921) 468‑479 (+??).  The third of the problems discussed is the construction of the heptagon quoted above.  Archibald gives an extensive survey of the topic on pp. 470-479.  A relevant cubic equation was already found by an unknown Arab writer, c980, and occurs in Vieta and in Kepler's Harmonices Mundi, book I, Prop. 45, where Kepler doubts that the heptagon can be constructed with ruler and compass.

                    An approximate construction was already given by Heron of Alexandria and may be due to Archimedes -- this says the side of the regular heptagon is approximately half the side of the equilateral triangle inscribed in the same circle.  Jordanus Nemorarius (c1230) called this the Indian method.  Leonardo da Vinci claimed it was exact.  For the central angle, this approximation gives a result that is about  6.5'  too small.

                    Archibald then goes on to consider constructions which claim to work or be good approximations for all  n-gons.  The earliest seems to be due to Antoine de Ville (1628), revised by A. Bosse (1665).  In 1891, A. A. Robb noted that a linkage could be made to construct the heptagon and J. D. Everett (1894) gave a linkage for  n-gons.

Italo Ghersi; Matematica Dilettevoli e Curiosa; 2nd ed., Hoepli, 1921; pp. 425-430: Costruzioni approssimate. 

                    He says the following construction is given by Housel; Nouvelles Annales de Mathématiques 12 (1853) 77-?? with no indication of its source.  Ghersi says it also occurs in Catalan's Trattato di Geometria, p. 277, where it is attributed to Bion.  However, Ghersi says it is due to Rinaldini (probably Carlo Renaldini (1615-1698)).  Let  AOB  be a horizontal diameter of a circle of radius  1  and form the equilateral triangle  ABC  with  C  below the diameter.  Divide  AB  into  n  equal parts and draw the line through  C  and the point  4/n  in from  B.  Where this line hits the circle, say  P,  is claimed to be  1/n  of the way around the circumference from  B.  Ghersi obtains the coordinates of  P  and the angle  BOP  and computes a table of these values compared to the real values.  The method works for  n = 2, 3, 4, 6.  For  n = 17,  the error is  36'37".

                    On pp. 428-430, he discusses a method due to Bardin.  Take  AOB  as above and draw the perpendicular diameter  COD.  Divide the diameter into  n  equal parts and extend both diameters at one end by this amount to points  M, N.  Draw the line  MN  and let it meet the circle near  B  at a point  P.  Now the line joining  P  to the third division point in from  B  is claimed to be an edge of the regular  n-gon inscribed in the circle.  Ghersi computes this length, finding the method only works for  n ³ 5, and gives a table of values compared to the real values.  This is exact for  n = 6  and is substantially more accurate than Renaldini's method.  For  n = 17,  the error is 1'10.32".

The "New" School of Art  Geometry,  Thoroughly Remodelled so as to Satisfy all the Requirements of the Science and Art Department for  Science Subject I. Sections I. and II,  Practical Plane and Solid Geometry,   (Cover says:  Gill's New School of Art  Geometry  Science Subject I.)  George Gill and Sons, London, 1890.

Pp. 26-27, prob. 66 -- To describe any regular Polygon on a given straight line, AB.  He constructs the centre of a regular  n-gon with AB as one edge.  Taking the side  AB  as  1,  the height  hn  of the centre is given by   hn  =  (n‑4) Ö3/4  ‑  (n-6)/4,  while the correct answer is  ½ cot π/n.  For large  n,  the relative error approaches  14.99%.  He gives no indication that the method is only approximate and doesn't even work for  n = 5.

Pp. 74-75, prob. 188 -- To inscribe any Regular Polygon in a given circle.  He gives three methods.  The first is to do it by trial!  The second requires being able to construct the regular  2n-gon!  The third construction is Renaldini's, which he does indicate is approximate.

R. C. Archibald, proposer;  H. S. Uhler, solver.  Problem 2932.  AMM 28 (1921) 467 (??NX)  &  30 (1923) 146-147.  Archibald gives De Ville's construction and asks for the error.  Uhler gives values of the error for  n = 5, 6, ..., 20,  and the central angles are about  1o  too large, even for  n = 6,  though the error seems to be slowly decreasing.

T. R. Running.  An approximate construction of the side of a regular inscribed heptagon.  AMM 30 (1923) 195-197.  His central angle is  .000061"  too small.

W. R. Ransom, proposer;  E. P. Starke, solver.  Problem E6.  AMM 39 (1932) 547 (??NX)  &  40 (1933) 175-176.  Gives Dürer's method for the pentagon and asks if it is correct.  Starke shows the central angle is about  22'  too large.

C. A. Murray, proposer;  J. H. Cross, E. D. Schell, Elmer Latshaw, solvers.  Problem E697 -- Approximate construction of regular pentagon.  AMM 52 (1945) 578 (??NX)  &  53 (1946) 336-337.  Describes a method similar to that of de Ville - Bosse and asks if it works for a pentagon.  Latshaw considers the general case.  The formula is exact for  n = 3, 4, 6.  For  n = 5,  the central angle is  2.82'  too small.  For  n > 6,  the central angle is too large and the error is increasing with  n.

J. C. Oldroyd.  Approximate constructions for 7, 9, 11, 13-sided polygons.  Eureka 18 (Oct 1955) 20.  Gives fairly simple constructions which are accurate to a few seconds.

Marius Cleyet-Michaud.  Le Nombre d'Or.  Presses Universitaires de France, Paris, 1973.  Méthode dite d'Albert Dürer, pp. 45-47.  Describes Dürer's approximate method for the pentagon and says it fails by  22'.

Underwood Dudley.  Mathematical Cranks.  MAA Spectrum, 1992.  This book discusses many related problems, e.g. duplication of the cube, trisection of the angle.  The chapter:  Nonagons, Regular,  pp. 231‑234 notes that there seem to be few crank constructors of the heptagon but that a nonagoner exists -- Dudley does not identify him.  Actually he constructs  10o  with an error of about  .0001',  so he is an excellent approximater, but he claims his construction is exact.

Robert Geretschläger.  Euclidean constructions and the geometry of origami.  MM 68:5 (Dec 1995) 357-371.  ??NYS -- cited in next article, where he states that this shows that all cubic equations can be solved by origami methods.

Robert Geretschläger.  Folding the regular heptagon.  CM 23:2 (Mar 1997) 81-88.  Shows how to do it exactly, using the result of his previous paper.

Dirk Bouwens, proposer;  Alan Slomson  &  Mick Bromilow, independent solvers.  An early protractor.  M500  171 (Dec 1999) 18  &  173 (Apr 2000) 16-17.  Draw a semicircle on diameter  AOB.  Draw the perpendicular through  O  and extend it to  C  so that  BC = BA (the diameter) [this makes  c = OC = Ö3] .  If  P  divides  AO in the ratio  λ : 1-λ,  then draw  CP  to meet the semicircle at  D  and  OD  divides the arc  ADE  in approximately the same ratio.  He finds the exact value of the angle  AOD  and finds that the maximum error in the process is only  .637o  (when  λ @ .18).  Second author provides a graph of the error and says the maximum error is at about 18o.  [I get  .637375o  at  .181625o.]

Ken Greatrix.  A better protractor.  M500 175 (Aug 2000) 14-15.  Taking  c = 1.67721  in the previous construction gives a more accurate construction, with maximum error about  .32o  at about  13o.  [I get  .324020o  at  .130164o.]

 

          6.BV. GEOMETRIC CONSTRUCTIONS

 

          New Section.  This is really a proper geometric topic, but there is some recreational interest in it, so I will cite some general references.

 

Robert C. Yates.  Geometrical Tools.  (As:  Tools; Baton Rouge, 1941);  revised ed., Educational Publishers, St. Louis, 1949.  Pp. 82-101 & 168-191.  Excellent survey.  After considering use of straightedge and compasses, he considers:  compasses only;  folds and creases;  straightedge only;  straight line linkages;  straightedge with fixed figure (circle, square or parallelogram);  straightedge with restricted compasses (collapsible compass, rigid (or rusty) compass or rigid dividers);  parallel and angle rulers;  higher order devices (marked ruler, carpenter's square, tomahawk, compasses of Hermes, two right angle rulers, straightedge with compasses and fixed conic);  plane linkages in general.  Each section has numerous references.

 

          6.BW.          DISTANCES TO CORNERS OF A SQUARE

 

          New section.  If  ABCD  is a rectangle,  P is a point and  a, b, c, d  are the distances of  P  from the corners of the rectangle, then the basic relation  a2 + c2 = b2 + d2  is easily shown.  This leads to a number of problems.  A little research has found references back to 1896, but the idea might be considerably older.

 

AMM 3 (1896) 155.  ??NYS -- this is the earliest reference given by Trigg, cf below.

SSM 15 (1915) 632.  ??NYS -- cited by Trigg, below.

AMM 35 (1928) 94.  ??NYS -- cited by Trigg, below.

SSM 32 (1932) 788.  ??NYS -- cited by Trigg, below.

NMM 12 (1937) 141.  ??NYS -- cited by Trigg, below.

AMM 47 (1940) 396.  ??NYS -- cited by Trigg, below.

NMM 16 (1941) 106.  ??NYS -- cited by Trigg, below.

NMM 17 (1942) 39.  ??NYS -- cited by Trigg, below.

AMM 50 (1943) 392.  ??NYS -- cited by Trigg, below.

SSM 46 (1946) 89, 783.  ??NYS -- cited by Trigg, below.

SSM 50 (1950) 324.  ??NYS -- cited by Trigg, below.

SSM 59 (1959) 500.  ??NYS -- cited by Trigg, below.

"A. Polter Geist", proposer;  Joseph V. Michalowicz, Mannis Charosh, solvers, with historical note by Charles W. Trigg.  Problem 865 -- Locating the barn.  MM  46:2 (Mar 1973) 104  &  47:1 (Jan 1974) 56-59.  For a square with an interior point,  a, b, c  =  13, 8, 5.  How far is  P from the nearest side?  First solver determines the side,  s,  by applying the law of cosines to triangles  BPA  and  BPC  and using that angles  ABP  and BPC  are complementary.  This gives a fourth order equation which is a quadratic in  s2.  Second solver uses a more geometric approach to determine  s.  The distances to the sides are then easily determined.  Trigg gives 13 references to earlier versions of the problem -- see above.

Anonymous.  Puzzle number 35 -- Eccentric lighting.  Bull. Inst. Math. Appl.  14:4 (Apr 1978) 110  &  13:5/6 (May/Jun 1978) 155.  Light bulb in a room, with distances measured from the corners.  a, b, c  =  9, 6, 2.  Find  d.  Solution uses the theorem of Apollonius to obtain the basic equation.

David Singmaster, proposer and solver.  Puzzle number 40 -- In the beginning was the light.  Bull. Inst. Math. Appl.  14:11/12 (Nov/Dec 1978) 281  &  15:1 (Jan 1979) 28.  Assuming  P  is inside the rectangle, what are the conditions on  a, b, c  for there to be a rectangle with these distances?  When is the rectangle unique?  When can  P  be on a diagonal?  Solution first obtains the basic relation, which does not depend on  P  being in the rectangle.  Reordering the vertices if necessary, assume  b  is the greatest of the distances.  Then  a2 + c2 ³ b2  is necessary and sufficient for a rectangle to exist with these distances.  This is unique if and only if equality holds, when  P = D.  If the distances are all equal, then  P  is at the centre of the rectangle, which can have a range of sizes.  If the distances are not all equal, there is a unique rectangle having  P  on a diagonal and it is on the diagonal containing the largest and smallest distances.

Marion Walter.  Exploring a rectangle problem.  MM 54:3 (1981) 131-134.  Takes  P  inside the rectangle with  a, b, c  = 3, 4, 5.  Finds the basic relation, noting  P  can be anywhere, and determines  d.  Then observes that the basic relation holds even if  P  is not in the plane of  ABCD.  Ivan Niven pointed out that the problem extends to a rectangular box.  Mentions the possibility of using other metrics.

James S. Robertson.  Problem 1147 -- Re-exploring a rectangle problem.  MM  55:3 (May 1982) 177  &  56:3 (May 1983) 180-181.  With  P  inside the rectangle and  a, b, c  given, what is the largest rectangle that can occur?  Observes that  a2 + c2 > b2  is necessary and sufficient for a rectangle to exist with  P  interior to it.  He then gives a geometric argument which seems to have a gap in it and finds the maximal area is  ac + bd.

I. D. Berg,  R. L. Bishop  &  H. G. Diamond, proposers.  Problem E 3208.  AMM 94:5 (May 1987) 456-457.  Given  a, b, c, d  satisfying the basic relation, show that a rectangle containing  P  can have any area from zero up to some maximum value and determine this maximum.

Problem 168.2.  M500 168 ???  Seven solvers, M500 170 (Oct 1999) 15-20.  Given  P  inside a square and  a, b, c = 5, 3, 4,  find the side of the square.

 

7.       ARITHMETIC  &  NUMBER‑THEORETIC RECREATIONS

 

          7.A.   FIBONACCI NUMBERS

 

          We use the standard form:  F0 = 0,  F1 = 1,  Fn+1 = Fn + Fn-1,  with the auxiliary Lucas numbers being given by:  L0 = 2,  L1 = 1,  Ln+1 = Ln + Ln-1.

 

Parmanand Singh.  The so-called Fibonacci numbers in ancient and medieval India.  HM 12 (1985) 229-244.  In early Indian poetry, letters had weights of  1  or  2  and meters were classified both by the number of letters and by the weight.  Classifying by weight gives the number of sequences of  1s  and  2s  which add to the weight  n  and this is  Fn+1. 

                    Pińgala [NOTE:  ń  denotes  n  with an overdot.] (c-450) studied prosody and gives cryptic rules which have been interpreted as methods for generating the next set of sequences, either classified by number of letters or by weight and several later writers have given similar rules.  The generation implies  Fn+1 = Fn + Fn-1.  Virahāńka [NOTE:  ń  denotes  n  with an overdot.] (c7C) is slightly more explicit.  Gopāla (c1134) gives a commentary on Virahāńka [NOTE:  ń  denotes  n  with an overdot.] which explicitly gives the numbers as  3, 5, 8, 13, 21.  Hemacandra (c1150) states "Sum of the last and last but one numbers ... is ... next."  This is repeated by later authors. 

                    The Prākŗta [NOTE:  ŗ  denotes  r  with an underdot] Paińgala [NOTE:  ń  denotes  n  with an overdot.] (c1315) gives rules for finding the  k-th sequence of weight  n  and for finding the position of a particular sequence in the list of sequences of weight  n  and the positions of those sequences having a given number of  2s  (and hence a given number of letters).  It also gives the relation   Fn+1 = Σi BC(n-i,i).

                    Narayana Pandita (= Nārāyaņa Paņdita’s [NOTE:  ņ  denotes  n  with an overdot and the  d  should have an underdot.]) Gaņita[NOTE:  ņ  denotes  n  with an underdot.]  Kaumudī (1356) studies additive sequences in chap. 13, where each term is the sum of the last  q  terms.  He gives rules which are equivalent to finding the coefficients of  (1 + x + ... + xq-1)p  and relates to ordered partitions using  1, 2, ..., q.

Narayana Pandita (= Nārāyaņa Paņdita [NOTE:  ņ  denotes  n  with an overdot and the  d  should have an underdot.]).  Gaņita[NOTE:  ņ  denotes  n  with an underdot.]  Kaumudī (1356).  Part I, (p. 126 of the Sanskrit ed. by P. Dvivedi, Indian Press, Benares, 1942), ??NYS -- quoted by Kripa Shankar Shukla in the Introduction to his edition of: Narayana Pandita (= Nārāyaņa Paņdita); Bījagaņitāvatamsa [NOTE: the  ņ  denotes an  n  with an under dot and there should be a dot over the  m.]; Part I; Akhila Bharatiya Sanskrit Parishad, Lucknow, 1970, p. iv.  "A cow gives birth to a calf every year.  The calves become young and themselves begin giving birth to calves when they are three years old.  Tell me, O learned man, the number of progeny produced during twenty years by one cow."

 

                    WESTERN HISTORIES

 

H. S. M. Coxeter.  The golden section, phyllotaxis, and Wythoff's game.  SM 19 (1953) 135‑143.  Sketches history and interconnections. 

H. S. M. Coxeter.  Introduction to Geometry.  Wiley, 1961.  Chap. 11: The golden section and phyllotaxis, pp. 160-172.  Extends his 1953 material.

Maxey Brooke.  Fibonacci numbers:  Their history through 1900.  Fibonacci Quarterly 2:2 (1964) 149‑153.  Brief sketch, with lots of typographical errors.  Doesn't know of Bernoulli's work.

Leonard Curchin & Roger Herz-Fischler.  De quand date le premier rapprochement entre la suite de Fibonacci et la division en extrême et moyenne raison?  Centaurus 28 (1985) 129-138.  Discusses the history of the result that the ratio  Fn+1/Fn  approaches  φ.  Pacioli and Kepler, described below, seem to be the first to find this.

Roger Herz‑Fischler.  Letter to the Editor.  Fibonacci Quarterly 24:4 (1986) 382.

Roger Herz-Fischler.  A Mathematical History of Division in Extreme and Mean Ratio.  Wilfrid Laurier University Press, Waterloo, Ontario, 1987.  Retitled: A Mathematical History of the Golden Number, with new preface and corrections and additions, Dover, 1998.  Pp. 157-162 discuss early work relating the Fibonacci sequence to division in extreme and mean ratio.  15 pages of references.

Georg Markovsky.  Misconceptions about the Golden Ratio.  CMJ 23 (1992) 2-19.  This surveys many of the common misconceptions -- e.g. that  --  appears in the Great Pyramid, the Parthenon, Renaissance paintings and/or the human body and that the Golden Rectangle is the most pleasing -- with 59 references.  He also discusses the origin of the term 'golden section', sketching the results given in Herz-Fischler's book.

Thomas Koshy.  Fibonacci and Lucas Numbers with Applications.  Wiley-Interscience, Wiley, 2001.  Claims to be 'the first attempt to compile a definitive history and authoritative analysis' of the Fibonacci numbers, but the history is generally second-hand and marred with a substantial number of errors,  The mathematical work is extensive, covering many topics not organised before, and is better done, but there are more errors than one would like.

Ron Knott has a huge website on Fibonacci numbers and their applications, with material on many related topics, e.g. continued fractions, π, etc. with some history.  www.ee.surrey.ac.uk/personal/r.knott/fibonacci/fibnat.html .

 

Fibonacci.  1202.  Pp. 283‑284 (S: 404-405): Quot paria coniculorum in uno anno ex uno pario germinentur [How many pairs of rabbits are created by one pair in one year].  Rabbit problem -- the pair propagate in the first month so there are  Fn+2  pairs at the end of the  n-th month.  (English translation in:  Struik, Source Book, pp. 2‑3.)  I have colour slides of this from L.IV.20 & 21 and Conventi Soppresi, C. I. 2616.  This is on ff. 130r-130v of L.IV.20,  f. 225v of L.IV.21,  f. 124r of CS.C.I.2616.

Unknown early 16C annotator.  Marginal note to II.11 in Luca Pacioli's copy of his 1509 edition of Euclid.  Reproduced and discussed in Curchin & Herz-Fischler and discussed in Herz-Fischler's book, pp. 157-158.  II.11 involves division in mean and extreme ratio.  Uses  89, 144, 233  and that  1442 = 89 * 233 + 1.  Also refers to  5, 8, 13.

Gori.  Libro di arimetricha.  1571.  F. 73r (p.81).  Rabbit problem as in Fibonacci.

J. Kepler.  Letter of Oct 1597 to Mästlin.  ??NYS -- described in Herz-Fischler's book, p. 158.  This gives a construction for division in extreme and mean ration.  On the original, Mästlin has added his numerical calculations, getting  1/φ = .6180340,  which Herz-Fischler believes to be the first time anyone actually calculated this number.

J. Kepler.  Letter of 12 May 1608 to Joachim Tanckius.  ??NYS -- described in Herz-Fischler (1986), Curchin & Herz-Fischler and Herz-Fishler's book, pp. 160-161.  Shows that he knows that the ratio  Fn+1/Fn  approaches  φ  and that  Fn2 + (-1)n  =  Fn-1Fn+1.

J. Kepler.  The Six‑Cornered Snowflake.  Op. cit. in 6.AT.3.  1611.  P. 12 (20‑21).  Mentions golden section in polyhedra and that the ratio  Fn+1/Fn  approaches  φ.  See Herz-Fischler's book, p. 161.

Albert Girard, ed.  Les Œuvres Mathematiques de Simon Stevin de Bruges.  Elsevier, Leiden, 1634.  Pp. 169‑170, at the end of Stevin's edition of Diophantos (but I have seen other page references).  Notes the recurrence property of the Fibonacci numbers, starting with  0,  and asserts that the ratio  Fn+1/Fn  approaches the ratio of segments of a line cut in mean and extreme ratio, i.e.  φ,  though he doesn't even give its value -- but he says  13, 13, 21  'rather precisely constitutes an isosceles triangle having the angle of a pentagon'.  Herz-Fischler's book, p. 162, notes that Girard describes it as a new result and includes  0  as the starting point of the sequence.

Abraham de Moivre.  The Doctrine of Chances: or, A Method of Calculating the Probability of Events in Play.  W. Pearson for the Author, London, 1718.  Lemmas II & III, pp. 128‑134.  Describes how to find the generating function of a recurrence.  One of his illustrations is the Lucas numbers for which he gets: 

                    x + 3x2 + 4x3 + 7x4 + 11x5 + ...  =  (2x + x2)/(1 - x - x2).  However, he does not have the Fibonacci numbers and he does not use the generating function to determine the individual coefficients of the sequence.  In Lemma III, he describes how to find the recurrence of  p(n) an  where  p(n)  is a polynomial.  Koshy [p. 215] says De Moivre invented generating functions to solve the Fibonacci recurrence, which seems to be reading much more into De Moivre than De Moivre wrote.  The second edition is considerably revised, cf below.

Daniel Bernoulli.  Observationes de seriebus quae formantur ex additione vel substractione quacunque terminorum se mutuo consequentium, ubi praesertim earundem insignis usus pro inveniendis radicum omnium aequationum algebraicarum ostenditur.  Comm. Acad. Sci. Petropolitanae 3 (1728(1732)) 85‑100, ??NYS.  = Die Werke von Daniel Bernoulli, ed. by L. P. Bouckaert & B. L. van der Waerden, Birkhäuser, 1982, vol. 2, pp. 49+??.  Section 6, p. 52, gives the general solution of a linear recurrence when the roots of the auxiliary equation are distinct.  Section 7, pp. 52‑53, gives the 'Binet' formula for  Fn.  [Binet's presentation is so much less clear that I suggest the formula should be called the Bernoulli formula.]

Abraham de Moivre.  The Doctrine of Chances: or, A Method of Calculating the Probability of Events in Play.  2nd ed, H. Woodfall for the Author, London, 1738.  Of the Summation of recurring Series, pp. 193-206.  This is a much revised and extended version of the material, but he says it is just a summary, without demonstrations, as he has given the demonstrations in his Miscellanea Analytica of 1730 (??NYS).  Gives the generating functions for various recurrences and even for a finite number of terms.  Prop. VI is: In a recurring series, any term may be obtained whose place is assigned.  He assumes the roots of the auxiliary equation are real and distinct.  E.g., for a second order recurrence with distinct roots  m, p,  he says the general term is  Amn + Bpn  where he has given  A  and  B  in terms of the first two values of the recurrence.  He even gives the general solution for a fourth order recurrence and expresses  A, B, C, D  in terms of the first four values of the recurrence.  Describes how to take the even terms and the odd terms of a recurrence separately and how to deal with sum and product of recurrences. 

R. Simson.  An explication of an obscure passage in Albert Girard's commentary upon Simon Stevin's works.  Phil. Trans. Roy. Soc. 48 (1753) 368‑377.  Proves that  Fn2 + (‑1)n  =  Fn-1Fn+1.  This says that the triple  Fn-1, Fn, Fn+1  "nearly express the segments of a line cut in extreme and mean proportion, and the whole line;" from which he concludes that the ratio  Fn+1/Fn  does converge to  φ.  Herz-Fischler's book, p. 162, notes that his proof is essentially an induction.  (He also spells the author Simpson, but it is definitely Simson on the paper.)  [Koshy, p. 74, says  Fn2 + (‑1)n  =  Fn-1Fn+1  was discovered in 1680 by Giovanni Domenico Cassini, but he gives no reference and neither Poggendorff nor BDM help to determine what paper this might be.]

Ch. Bonnet.  Recherches sur l'usage des feuilles dans les plantes.  1754, pp. 164‑188.  Supposed to be about phyllotaxis but only shows some spirals without any numbers.  Refers to Calandrini.  Nice plates.

Master J. Paty (at the Mathematical Academy, Bristol), proposer; W. Spicer, solver.  Ladies' Diary, 1768-69  =  T. Leybourn, II: 293, quest. 584.  Cow calves at age two and every year thereafter.  How many offspring in 40 years?  Answer is:  0 + 1 + 1 + 2 + 3 + ... + F39.  We would give this as:  F41 - 1,  but he gets it as:  2F39 + F38 - 1.

Eadon.  Repository.  1794.  P. 389, no. 47.  Same as the previous problem.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions, no. 32, pp. 22 & 82.  Similar to Fibonacci, with a cow and going for 20 generations.

Martin Ohm.  Die reine Elementar-Mathematik.  2nd ed., Jonas Verlags-Buchhandlung, Berlin, 1835.  P. 194, footnote to Prop. 5.  ??NYS -- extensively discussed in Herz-Fischler's book, p. 168.  This is the oldest known usage of 'goldene Schnitt'.  It does not appear in the 1st ed. of 1826 and here occurs as:  "... nennt man wohl auch den goldenen Schnitt" (... one also appropriately calls [this] the golden section).  The word 'wohl' has many, rather vague, meanings, giving different senses to Ohm's phrase.  Herz-Fischler interprets it as 'habitually', which would tend to imply that Ohm and/or his colleagues had been using the term for some time.  I don't really see this meaning and interpreting 'wohl' as 'appropriately' would give no necessity for anyone else to know of the phrase before Ohm.  However the term is used in several other German books by 1847.   [Incidentally, this is not the Ohm of Ohm's Law, but his brother.] 

A. F. W. Schimper & A. Braun.  Flora.  1835.  Pp. 145 & 737.  ??NYS

J. Binet.  Mémoire sur l'integration des équations linéaires aux différences finies, d'un ordre quelconque, à coefficients variables.  (Extrait par l'auteur).  CR Acad. Sci. Paris 17 (1843) 559‑567.  States the Binet formula as an example of a general technique for solving recurrences of the form:  v(n+2) = v(n+1) + r(n)v(n),  but the general technique is not clearly described, nor is the linear case.

B. Peirce.  Mathematical investigation of the fractions which occur in phyllotaxis.  Proc. Amer. Assoc. Adv. Sci. 2 (1849) 444‑447.  Not very interesting.

Gustav Theodor Fechner.  Vorschule der Ästhetik.  Breitkopf & Härtel, Leipzig, 1876.  ??NYS.  Origin of the aesthetic experiments on golden rectangles.

Koshy [p. 5] says Lucas originally called  Fn  the 'série de Lamé', but introduced the name Fibonacci numbers in May 1876.  However, he doesn't give a reference.  There are several papers by Lucas which might be the desired paper.

                    Note sur le triangle arithmétique de Pascal et sur la série de Lamé.  Nouvelle Correspondence Mathématique 2 (1876) 70-75; which might be the desired paper.

                    L'arithmétique, la série de Lamé, le problème de Beha-Eddin, etc.  Nouvelles Annales de Mathématiques 15 (1876) 20pp. 

Édouard Lucas.  Théorie des fonctions numériques simplement périodiques.  AJM 1 (1878) 184-240 (Sections 1-23)  &  289-321 (Sections 24-30).  [There is a translation by Sidney Kravitz of the first part as:  The Theory of Simply Periodic Numerical Functions, edited by Douglas Lind, The Fibonacci Association, 1969.  Dickson I 400, says this consists of 7 previous papers in Nouv. Corresp. Math. in 1877-1878 with some corrections and additions.  Robert D. Carmichael; Annals of Math. (2) 15 (1913) 30-70, ??NX, gives corrections.]  The classic work which begins the modern study of recurrences.

Koshy, p. 273, says Adolf Zeising's Der goldene Schnitt of 1884 put forth the theory that "the golden ratio is the most artistically pleasing of all proportions ...."  But cf Fechner, 1876.

Pearson.  1907.  Part II, no. 63: A prolific cow, pp. 126 & 203.  Same as Fibonacci's rabbits, but wants the total after 16 generations.

Koshy, p. 242, asserts that Mark Barr, an American mathematician, introduced the symbol  φ  (from Phidias) for the Golden Ratio, (1 + Ö5)/2,  about 1900, but he gives no reference.

Coxeter, 1953, takes  τ  from the initial letter of  τoμη,  the Greek word for section, but I have no idea if this was used before him.

There is a magic trick where you ask someone to pick two numbers and extend them to a sequence of ten by adding the last two numbers each time.  You then ask him to add up the ten numbers and you tell him the answer, which is 11 times the seventh number.  In general, if the two starting numbers are  A  and  B,  the  n-th  term is  Fn-2A + Fn-1B  and the sum of the first  2n  terms is  F2nA + (F2n+1-1)B  =  Ln (FnA + Fn+1B),  but only the case  n = 5  is interesting!  I saw Johnny Ball do this in 1989 and I have found it in: Shari Lewis; Abracadabra!  Magic and Other Tricks; (World Almanac Publications, NY, 1984);  Puffin, 1985;  Sum trick!, p. 14, but it seems likely to be much older.

 

          7.B.   JOSEPHUS OR SURVIVOR PROBLEM

 

          See Tropfke 652.

          This is the problem of counting out every  k‑th from a circle of  n.  Early versions counted out half the group;  later authors and the Japanese are interested in the last man -- the survivor.  Euler (1775) seems to be the first to ask for the last man in general which we denote as  L(n, k).  Cardan, 1539, is the first to associate this process with Josephus.  Some later authors derive this from the Roman practice of decimation. 

          For last man versions, see the general entries and:  Michinori?,  Kenkō,  Cardan,  Coburg,  Bachet,  van Etten,  Yoshida,  Muramatsu,  Schnippel,  Ozanam (1696 & 1725),  Les Amusemens,  Fujita,  Euler,  Miyake,  Matuoka,  Boy's Own Book,  Nuts to Crack,  The Sociable,  Indoor & Outdoor,  Secret Out (UK),  Leske,  Le Vallois,  Hanky Panky,  Kemp,  Mittenzwey,  Gaidoz,  Ducret,  Lemoine,  Akar et al,  Lucas,  Schubert,  Busche,  Tait,  Ahrens,  Rudin,  MacFhraing,  Mendelsohn,  Barnard,  Zabell,  Richards,  Dean,  Richards, 

 

  2 to last, counted by 9s:  Boy's Own Book, 

  3 to last, counted by 9s:  Boy's Own Book, 

  4 to last, counted by 9s:  Boy's Own Book, 

  5 to last, counted by 9s:  Boy's Own Book, 

  6 to last, counted by 9s:  Boy's Own Book, 

  7 to last, counted by 9s:  Boy's Own Book, 

  9 to last, counted by 9s:  Boy's Own Book, 

10 to last, counted by 9s:  Boy's Own Book, 

11 to last, counted by 9s:  Boy's Own Book, 

12 to last, counted by 9s:  Boy's Own Book,  Secret Out (UK), 

12 to last, counting number unspecified:  Coburg, 

13 to last, counted by 2s:  Ducret,  Leeming, 

13 to last, counted by 9s:  Boy's Own Book,  Secret Out (UK),  Leske,  Rudin, 

14 to all!, counted by 6s:  Secret Out,  

14 to last, counted by 10s:  Mittenzwey, 

17 to last, counted by 3s:  Barnard, 

21 to last, counted by 5s:  Hyde, 

21 to last, counted by 7s:  Nuts to Crack,  The Sociable,  Indoor & Outdoor,  Hanky Panky,  H. D. Northrop, 

21 to last, counted by 8s:  Mittenzwey, 

21 to last, counted by 10s:  Hyde, 

24 to last, counted by 9s:  Kemp, 

28 to last, counted by 9s:  Kemp, 

30 to last, counted by 9s:  Schnippel, 

30 to last, counted by 10s:  see entries in next table for  15 & 15 counted by 10s

40 to last man, counted by 3s:  van Etten (erroneous), 

41 to last man, counted by 3s:  van Etten,  Ozanam (1725),  Vinot,  Ducret,  Lucas (1895), 

General case:  Euler,  Lemoine,  Akar et al.,  Schubert,  Busche,  Tait,  Ahrens,  MacFhraing,  Mendelsohn,  Robinson,  Jakóbczyk,  Herstein & Kaplansky,  Zabell,  Richards, 

 

          There are a few examples where one counts down to the last two persons -- see references to Josephus and:  Pacioli,  Muramatsu,  Mittenzwey,  Ducret,  Les Bourgeois Punis.

 

          Almost all the authors cited consider  15 & 15 counted by 9s,  so I will only index other versions.

 

  2 &   2 counted by  3s:  Ball (1911), 

  2 &   2 counted by  4s:  Ball (1911), 

  3 &   3 counted by  7s:  Ball (1911), 

  3 &   3 counted by  8s:  Ball (1911), 

  4 &   4 counted by  2s:  Leeming, 

  4 &   4 counted by  5s:  Ball (1911), 

  4 &   4 counted by  9s:  Ball (1911), 

  5 &   5 counted by  ??:  Dudeney (1905),  Pearson,  Ball (1911),  Ball (1920),  Shaw, 

  6 &   6 counted by  ??:  Dudeney (1900), 

  8 &   2 counted by  ??:  Les Bourgeois Punis, 

  8 &   8 counted by  8s:  Kanchusen, 

  8 &   8 counted by  ??:  Dudeney (1899), 

12 & 12 counted by  6s:  Harrison, 

15 & 15 counted by  3s:  Tartaglia,  Alberti, 

15 & 15 counted by  4s:  Tartaglia, 

15 & 15 counted by  5s:  Tartaglia, 

15 & 15 counted by  6s:  AR,  Codex lat. Monacensis 14908,  Tartaglia, 

15 & 15 counted by  7s:  Tartaglia,  Schnippel, 

15 & 15 counted by  8s:  Codex lat. Monacensis 14836,  AR,  Codex lat. Monacensis 14908,  Tartaglia,  Alberti, 

15 & 15 counted by 10s:  Michinori?,  Reimar von Zweiter,  AR,  Codex lat. Monacensis 14908,  Chuquet,  Tartaglia,  Buteo,  Hunt,  Yoshida,  Muramatsu,  Wingate/Kersey,  Schnippel,  Alberti,  Shinpen Kinko-ki,  Fujita,  Miyake,  Matuoka,  Sanpo Chie Bukuro,  Hoffmann,  Brandreth,  Benson,  Williams,  Collins,  Dean.  (Almost all of these actually continue to the last person.) 

15 & 15 counted by 11s:  Tartaglia,  Schnippel, 

15 & 15 counted by 12s:  AR,  Codex lat. Monacensis 14908,  Tartaglia, 

15 & 15 counted by other values, not specified -- ??check:  Codex lat. Monacensis 14836,  Meermanische Codex,  at‑Tilimsâni,  Bartoli,  Murray 643,  Chuquet,  Keasby, 

17 & 15 counted by 10s:  Schnippel, 

17 & 15 counted by 12s:  Mittenzwey, 

18 &  2 counted by 12s:  Pacioli,  Rudin, 

18 &  6 counted by  8s:  Manuel des Sorciers, 

18 & 18 counted by  9s:  Chuquet, 

24 & 24 counted by  9s:  Chuquet, 

30 &  2 counted by  7s:  Pacioli, 

30 &  2 counted by  9s:  Pacioli, 

30 &  6 counted by 10s:  Ducret,  30 & 10 counted by 12s:  Endless Amusement II,  Magician's Own Book,  The Sociable,  Boy's Own Conjuring Book,  Lucas (1895), 

30 & 30 counted by 12s:  Sarma, 

36 &  4 counted by 10s:  Jackson, 

n2-n+1  &  n-1  counted by  n:  Lucas (1894),  Cesarò,  Franel,  Akar, 

 

          Many authors provide a mnemonic for the case of  15 and 15 counted by 9s.  In this case, the longest group of the same type is five, so a common device is to encode the numbers  1, 2, 3, 4, 5  by the vowels  a, e, i, o, u  and then produce a phrase with the vowels in the correct order.  I will call this a vowel mnemonic.  The most popular form is:  Populeam virgam Mater Regina ferebat,  giving the numerical sequence:  4, 5, 2, 1, 3, 1, 1, 2, 2, 3, 1, 2, 2, 1.  The first group of 4 are good guys, followed by 5 bad guys, etc.  Below I list the mnemonics and where they occur, but I did not always record them in my notes below, so I must check a number of the sources again ?? -- the classification was inspired by seeing that Franci (op. cit. in 3.A) describes a vowel mnemonic in Benedetto da Firenze which I had overlooked.  Ahrens gives many more verse and vowel mnemonics -- to be added below.  Hyde gives an Arabic mnemonic due to al-Safadi using the first letters of the Arabic alphabet: a, b, gj, d, h.

 

Dahbagja Ababgja Baba  [=  Da h b a gj a  A ba b gj a  Ba b a]:  Hyde from al-Safadi.

Unspecified(?) verse mnemonic:  ibn Ezra

Populea  irgam  mater  regina  reserra:  Pacioli; 

Populea virga pacem regina ferebat:  Minguét

Populeam jirgam mater Regina ferebat:  Badcock

Populeam virgam mater Regina ferebat:  van Etten;  Hunt;  Schnippel;  Ozanam 1725;  Les Amusemens;  Hooper;  Jackson;  Manuel des Sorciers;  Boy's Own Book;  The Sociable;  Le Vallois;  Gaidoz;  Lucas

Populeam  virgam  mater  regina  reserrat:  Agostini's version of Pacioli; 

Populeam virgam Mater Regina tenebat:  Hyde from Wit's Interpreter;  Murphy;  Schnippel/Bolte

Mort tu ne failliras pas en me liurant le trespas:  van Etten

Mort, tu ne falliras pas   En me livrant au trépas:  Manuel des Sorciers; 

Mort, tu ne falliras pas.  En me livrant le trépas:  Schnippel/Bolte;  Ozanam 1725;  Les Amusemens;  The Sociable;  Le Vallois (without the first comma);  Ducret;  Lucas

On tu ne dai la pace ei la rendea:  Schnippel/Bolte

Gott schuf den Mann in Amalek, der (or den) Israel bezwang:  Schnippel

Gott schlug den Mann in Amalek, den Israel bezwang:  Schnippel/Bolte

So du etwan bist gfalln hart, Stehe widr, Gnade erwart:  Schnippel/Bolte

Non dum pena minas a te declina degeas:  Schnippel/Bolte

Nove la pinta dà e certi mantena:  Benedetto da Firenze

From member's aid and art, Never will fame depart:  Schnippel/Bolte

From numbers, aid and art / Never will fame depart:  Wingate/Kersey

From numbers, aid, and art, Never will fame depart:  Ingleby;  Jackson;  Rational Recreations

From number's aid and art, Never will fame depart:  Gaidoz

From numbers aid and art / Never will fame depart:  The Sociable

 

          I have only one example of a mnemonic for 15 & 15 counted by 10s.

Rex Paphicum Gente Bonadat Signa Serena:  Hunt

 

          See 5.AD for the general problem of stacking a deck to produce a desired effect.

 

Josephus.  De Bello Judaico.  c80.  Book III, chap. 8, sect. 7.  (Translated by Whiston or by Thackeray (Loeb Classical Library, Heinemann, London, 1927, vol. 2, pp. 685‑687.))  (Many later authors cite Hegesippus which is a later version of Josephus.)  This says that Josephus happened to survive  "by chance or God's providence".

H. St. J. Thackeray.  Josephus, the Man and the Historian.  Jewish Institute Press, NY, 1929, p. 14.  Comments on the Slavonic text, which says that Josephus  "counted the numbers with cunning and thereby misled them all"  but gives no indication how.

 

Ahrens.  MUS II.  1918.  Chap. XV: Das Josephsspiel, pp. 118-169.  This is the most extended and thorough discussion of this problem and its history.  I have used it as the basis of this section.  He gives a rather complex method, based on work of Busche, Schubert and Tait, for determining the last man, or any other man in the sequence of counting out, which I never worked through, but which is clearly explained under Richards (1999/9).

 

Gerard Murphy.  The puzzle of the thirty counters.  Béaloideas -- The Journal of the Folklore of Ireland Society XII (1942) 3‑28.  In this work and the material cited (mostly ??NYS), the problem of  15 and 15 counted by 9s  is shown to have the medieval name  Ludus Sancti Petri  = St. Peters‑Spiel  = St. Peter's Lake (lake being an Old English word for game [or Anglo-Saxon for 'to play'])  = Sankt Peter Lek or Sankt Päders Lek (in Swedish).  Murphy cites Schnippel & Bolte to assert that it was known to the Arabs in the 14C -- cf below.  He says the usual European version has Christians and Jews on a ship, with St. Peter present and suggesting the counting out process.  [I had forgotten that such versions occur in Ahrens, MUS II 130.]  However, Murphy was unable to consult MUS, so his background is not as complete as it might be.

                    Murphy demonstrates that the problem was recently well‑known in both Scots and Irish Gaelic in a form where a woman has to choose between two groups of warriors seated in a circle, with emotional reasons for her preference.  The solution is given in a vernacular mnemonic, using actual numbers as in early Latin forms, while later Latin and vernacular forms used vowel mnemonics.  He gives an Irish reconstruction, with English translation, based on several 18C MSS whose texts he estimates as 13C to 17C, probably 16C.  This is titled:  Goid Fhinn Agus Dubháin Anso (Here is the Thieving of Fionn and Dubhán).  One MS has the Latin subtitle:  Populeam virgam Mater Regina tenebat,  which is a common Latin vowel mnemonic.  One of Murphy's sources says this refers to the Queenly Mary appearing to the ship's captain and holding a poplar rod.

                    Murphy also gives an extended Irish story (3pp) built around the problem:  Ceann Dubhrann na Ndumhchann Bán (Ceann Dubhrann of the White Sandhills).  The Gaelic names  Fionn  and  Dubhán  are derived from  'fionn'  and  'dub'  meaning  'white'  and  'black'.  Murphy gives a contemporary Irish version on board a ship with a white captain and a black wife and a crew of 15 and 15, with half having to go overboard due to lack of food.  He sketches numerous other Irish and Scots version with varying combinations of details, but using essentially the same verse mnemonic.

                    Murphy cites a study by Manitius of a 9C MS (Bib. Nat. Paris, No. 13029) where the problem begins  "Quadam nocte niger dub nomine, candiduus alter" (One night a black man named Dub and another [named] White).  They have to choose between the blacks and the whites to keep watch.  Cf. Codex Einsidelensis No. 326 below.  The MS ends with the prose line  "These two Irish soldiers, one named 'Find' the other 'Dub', were engaged in hunting.  'Find' means "white", 'dub' "black"."  The 12C Rouen MS No. 1409 attributes the problem to a Clemens Scottus, which Murphy interprets as Clement the Irishman.  The 12C MS Bib. Nat. Paris No. 8091 attributes it to a Thomas Scottus.

                    Murphy concludes that the problem has an Irish origin, c800.  He gives what he believes to be the earliest Latin form, basically Bib. Nat. Paris No. 13029, and opines there must have been an Irish predecessor.

Codex Einsidelensis No. 326.  10C.  F. 88'.  Latin verse.  Published by Th. Mommsen, Handschriftliches.  Zur lateinischen Anthologie.  Rheinischen Museum für Philologie (NS) 9 (1854) 296‑301, with material of interest on pp. 298-299.  Latin given in:  M. Curtze, Bibliotheca Math. (2) 9 (1895) 34‑35.  Latin & German in MUS II 123‑125.  Begins:  "Quadam nocte niger dux nomine, candidus alter".  15 white  &  15 black soldiers, half to keep watch, counted off by 9.  The colours refer to clothing, not skin!

Codex lat. Monacensis 14836.  11C.  F. 80' gives rules for  15 and 15 counted by 9 (though this value is not specified)  and mentions counting by 8 and other values.  No mention of what is being counted.  Quoted and discussed by:  M. Curtze; Zur Geschichte der Josephspiels; Bibliotheca Math. (2) 8 (1894) 116  and in:  Die Handschrift No. 14836 der Königl. Hof‑ und Staats‑bibliothek zu München; AGM 7 (1895) 105 & 111‑112 (Supplement to  Zeitsch. für Math. und Physik 40 (1895)).

Codex Bernensis 704.  12C.  Published by:  Hermann Hage; Carmina medii aevi maximam partem inedita; Ex Bibliothecis Helveticus collecta; Bern, 1877; no. 85, pp. 145‑146.  ??NYS.   Latin in:  Curtze, op. cit. at Codex Einsidelensis, pp. 35‑36;  and in:  MUS II 127.  Jews  &  Christians.

Ahrens, MUS II 118‑147, gives many further references from 10‑13C.  Originals ??NYS.

Meermanische Codex, 10C.  Mentions counting by other values.

Leiden Miscellancodex, 12C

Basel Miscellancodex, 13C

Michinori Fujiwara (1106-1159).  This work is lost, but has been conjectured to contain a form of the problem -- see under Kenkō, c1331, and Yoshida, 1634.

Rabbi Abraham ben Ezra.  Ta'hbula (or Tachbûla), c1150.  ??NYS -- described in:  Moritz Steinschneider; Abraham ibn Ezra (Abraham Judaeus, Avenare); Zur Geschichte der mathematischen Wissenschaft im XII Jahrhundert;  Zeitschr. für Math. und Physik 25 (1880): Supp: AGM 3, Part II (1880).  The material is Art. 20, pp. 123‑124.  15 students and 15 good-for-nothings on a ship, counted by 9s.  This seems to be the first extant example on board a ship.  Verse mnemonic, which Steinschneider says is not original.  Steinschneider cites further sources.  Smith & Mikami, p. 84, say ben Ezra died in 1067 -- ??

Reinmar von Zweter.  Meisterlied:  "Ander driu, wie man juden und cristen ûz zelt".  13C.  (In MUS II 128.)  Jews and Christians on a ship, counts by 10.

Kenkō, also known as  Kenkō Yoshida  or  Urabe no Kaneyoshi  or  Yoshida no Kaneyoshi  (1283-1350).  Tsurezuregusa.  c1331.  Translated by Donald Keene as: Essays in Idleness  The Tsurezuregusa of Kenkō; Columbia Univ. Press, NY, 1967.  (Kenkō was the author's monastic name.  His lay name was  Urabe no Kaneyoshi.  He lived for a long time at Yoshida in Kyoto.) 

                    This book is one of the classics of Japanese literature, consisting of 243 essays, ranging from single sentences to several pages.  The most common themes of these relate to the impermanence of life and the vanity of man. 

                    In Japanese, the Josephus problem is called  Mamakodate  or  Mamako-date San  (or  Mama-ko tate no koto - cf Matŭ-oka, 1808)  or  Mamagodate  (Scheme to benefit the step-children  or  Stepchild disposition).  It is said to have been in the lost work of Michinori Fujiwara (1106‑1159), qv.  The word  Mamagodate  first occurs in essay 137 of Kenkō, pp. 115-121 in Keene's version (including a double-page illustration which doesn't depict the problem), whose beginning is characteristic of Kenkō's style:  "Are we to look at cherry blossoms only in full bloom, the moon only when it is cloudless?  To long for the moon while looking on the rain, to lower the blinds and be unaware of the passing of the spring -- these are even more deeply moving."  The passage of interest is toward the end, on p. 120 of Keene: "When you make a mamagodate1 with backgammon counters, at first you cannot tell which of the stones arranged before you will be taken away.  Your count then falls on a certain stone and you remove it.  The others seem to have escaped, but as you renew the count you will thin out the pieces one by one, until none is left.  Death is like that."  The footnote refers to counting 15 and 15 by 10s, so that 14 white stones are eliminated, then the counting is reversed and all the black stones are eliminated.  "The Japanese name mamagodate (stepchild disposition) derives from the story of a man with fifteen children by one wife and fifteen by another; his estate was disposed of by means of the game, one stepchild in the end inheriting all."  Kenkō's text clearly shows he was familiar with the process of counting to the last man and the use of the name indicates that he was familiar with the version mentioned in the footnote, though its earliest explicit appearance in Japan is in Yoshida, 1634, qv.  My thanks to Takao Hayashi for the reference to Keene.

Thomas Hyde.  Mandragorias seu Historia Shahiludii, ....  (= Vol. 1 of De Ludis Orientalibus, see  4.B.5 for vol. 2.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  Prolegomena curiosa.  The initial material and the Prolegomena are unpaged but the folios of the Prolegomena are marked (a), (a 1), ....  The material is on  (e 1).v - (e 2).v, which are pages 34-36 if one starts counting from the beginning of the Prolegomena.  Cited by Bland (loc. cit. in 5.F.1 under Persian MS 211, p. 31);  Ahrens (MUS II 136)  &  Murray 280.  Several citations are to  ii.23,  which may be to the 1767 reprint of Hyde's works.

                    Hyde asserts that the problem of the ship  with 15 Moslems and 15 Christians on a ship, counted by 9s, was given by al-Safadi (Şalâhaddîn aş‑Şafadî [NOTE:  Ş,  ş  denote  S,  s,  with an underdot and the  h  should have an underdot.]  = al-Sâphadi  = AlSáphadi) (d. 1363) in his  Lâmiyato ’l Agjam  (variously printed in the text).  This must be his  Sharh [the  h  should have an underdot] Lâmîyat al‑‘Ajam  of c1350.  Hyde gives an Arabic mnemonic using the first five letters of the Arabic alphabet, which he transliterates as:  Dahbagja Ababgja Baba  [=  Da h b a gj a  A ba b gj a  Ba b a].  He says the problem occurs in an English book called Wit's Interpreter (??NYS) (8oS.87.Art) where the mnemonic  Populeam virgam mater Regina tenebat  is given.  He then says that the problem is also described in 'Megjdium & Abulphedam' -- p. 43 of his main text identifies Abulpheda as a prince born in 672 AH -- ??

Shihâbaddîn Abû’l‑‘Abbâs Ahmad [the  h  should have an underdot] ibn Yahya [the  h  should have an underdot] ibn Abî Hajala [the  H  should have an underdot] at‑Tilimsâni alH‑anbalî [the  H  should have an underdot].  Kitâb ’anmûdhaj al‑qitâl fi la‘b ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot] (Book of the examples of warfare in the game of chess).  Copied by Muhammed ibn ‘Ali ibn Muhammed al‑Arzagî in 1446.

                    This is the second of Dr. Lee's MSS, described in 5.F.1, denoted Man. by Murray.  Murray 280 says "Man. 36‑45 relate to  as‑Safadî's  problem of the ship (see Hyde, ii.23)", described by Murray as 15 Christians and 15 Muslims counted by  n.  Bland has "the well-known problem of the Ship, first as described by Safadi, and then in other varieties.  (Hyde, p. 23.)"

Murray 620 says the problem is of Muslim origin and says it appears in the c1530 Italian version of the Bonus Socius collection which Murray denotes It.  (See 5.F.1 under Bonus Socius.)  Murray refers to  15 & 15 counted by 9s,  but it is not clear if this refers to this particular MS. 

Murray 622 cites  MS Sloan 3281 in the BM, 14C, as giving the Latin mnemonic solution.

Bartoli.  Memoriale.  c1420.  F. 100r (= Sesiano p. 135).  Il giuocho de' Cristiani contra Saracin.  15 Christians and 15 Saracens  -- the text ends in the middle of the statement of the problem.

AR.  c1450.  Prob. 80, pp. 52, 181‑182 & 229.  15 Christians and 15 Jews.  Gives only mnemonics for counting by  10, 9, 8, 6 or 12.

Codex lat. Monacensis 14908.  c1460.  F. 76 gives mnemonics for 15 Jews and 15 Christians counted by  6, 8, 9, 10, 12.  Quoted and discussed by Curtze, opp. cit. under Cod. lat. Mon. 14836, above.  [In the first paper, the codex number is misprinted as 14809.]

Benedetto da Firenze.  c1465.  Pp. 142‑143.  15 Christians and 15 Jews on a boat counted by 9s.  Vowel mnemonic:  Nove la pinta dà e certi mantena.  Diagrammatic picture on p. 143.

Murray 643 says the MS Lasa version, c1475, of the Civis Bononiae collection (described in 5.F.1 under Civis Bononiae) has  "16 diagrams of the 'ship' puzzle under different conditions".

Chuquet.  1484.  Prob. 146.  English in FHM 228-230, with reproduction of the original on p. 229.  15 Jews and 15 Christians on a ship, counting by 9s.  Says one can have 18 or 24 of each and can count by 10s, etc.  The reproduction on FHM 229 shows a circle marked out, with  Populeam virgam matre regina tenebat  written in the middle.  The commentary says this "problem is comparatively rare in fifteenth century texts", which doesn't seem like a fair assessment to me.

Calandri.  Aritmetica.  c1485.  Ff. 102v-103v, pp. 205‑207.  Coloured plate opp. p. 192 of the text volume.  (Tropfke 654 gives this in B&W.)  Franciscans and Camoldensians on a boat:  15 & 15 counted by 9s.

Pacioli.  De Viribus.  c1500.  Probs. 56‑60.

Ff. 99r - 102r.  LVI. (Capitolo) de giudei Chri'ani in diversi modi et regole. a farne quanti se vole etc (Of Jews and Christians in diverse methods and rules, to make as many as one wants, etc.).  = Peirani 140-143.  Does  2 & 30 by 9s -- there is a diagram for this in the margin of f. 100r, but it is not in the transcription and Peirani says another diagram is lacking.  Pacioli suggests counting the passengers on shore and doing the counting out with coins or pebbles in case one will need to know the arrangement in a hurry.  He also says one might count by  8s, 7s, 6s, 13s, etc., with any number of Christians and Jews.

Ff. 102r - 102v.  [Unnumbered.]  de .18. Giudei et .2. Chri'ani.  = Peirani 144.  2 & 18 by 7s.

F. 102v.  LVII. C(apitolo). de .30. Giudei et .2. contando per .7. ch' toca va in aqua (Of 30 Jews and 2 counting by 7 with the touched going in the water).  = Peirani 144. 

Ff. 102v - 103r.  LVIII. C(apitolo). de .15. Giudei et .15. Chri'ani per .9. in aqua (Of 15 Jews and 15 Christians by 9 in the water.)  = Peirani 144-145.  15 & 15 by 9s.  In order to remember the arrangement, he says to see the next section.

Ff. 103r.  LIX. C(apitolo). Quater quinque. duo. unus. tres unus. et unus. bis duo. ter unus. duo duobus un' (4, 5, 2, 1, 3, 1, 1, 2, 2, 3, 1, 2, 2, 1).  = Peirani 145.  Pattern for 15 & 15 counted by 9s.

Ff. 103r - 103v.  LX. C(apitolo). si da unaltro verso viz. Populea. irga. mater regina. reserra.  ['unaltro' is in the margin with a mark showing where it is to go.]  Vowel mnemonic for 15 & 15 counted by 9s, which he explains in detail.  Agostini says this is intended to be:  Populeam  virgam  mater  regina  reserrat  but both Pacioli's heading and his discussion have  Populea  irgam  mater  regina  reserra.  The Index has LVIII-LX under one heading which only refers to 'versi memorevili'.

Elias Levita der Deutsche.  Ha‑Harkabah.  Rome, 1518.  ??NYS.   Attributes to ben Ezra, c1150??.  Smith & Mikami, p. 84, say this seems to be the first printed version of the problem.

Cardan.  Practica Arithmetice.  1539.  Chap. LXI, section 18, ff. T.iiii.r ‑ T.v.r, but the material of interest is just a few sentences on f. T.iv.v (p. 113).  Very brief description of 15 white and 15 black as 'ludus Josephus', saying one can work out any numbers with some pebbles.  MUS says this is first to relate the problem to Josephus as the last man, but he doesn't give any numerical details.

Hans Sachs (1494-1576).  Meisterleid:  'Historia  Die XV Christen und XV Türcken, so auff dem meer furen'.  (MUS II 132‑133 gives text.)

Tartaglia.  General Trattato, 1556, art. 203, pp. 264v‑265r.  15 whites and 15 blacks (or Turks and Christians) counted out by  3, 4, ..., 12.  No reference to Josephus.

Buteo.  Logistica.  1559.  Prob. 89, pp. 303-304.  15 Christians and 15 Jews on a ship counted by 10s.  [Mentioned in H&S 52.]

Simon Jacob von Coburg.  Ein new und Wolgegründt Rechenbuch ....  1565 or 1612 (in quarto, not to be confused with octavo versions of 1565 and 1613 which do not contain the problem), f. 250v.  ??NYS -- described in MUS II 133-134.  12 drinkers deciding who shall pay the bill.  Ahrens doesn't specify the counting number.  Ahrens & Bolte (below) say this is the earliest example, after Cardan, of finding the last man.  Ahrens describes numerous later examples of this type from 1693 on.

Prévost.  Clever and Pleasant Inventions.  (1584), 1998.  Pp. 183-185.  This seems like a version of the Josephus problem but isn't.  Place ten counters in a circle and then ten on top of them.  Start anywhere and count off five and remove the top counter.  He says to count five again -- starting on the place where the counter was removed, so we now would say he is counting four -- and remove the top counter.  Continue in this way, counting the places where a top counter has been removed and you manage to remove all the top counters.  In fact this is impossible, but after removing five counters, you subtly start counting from the next position rather than where the top counter was removed!  Hence you remove the top counters in the order  5, 9, 3, 7, 1, 6, 10, 4, 8, 2.  Your audience will not observe this and hence cannot reproduce the effect.

                    The mathematical description is simpler if one counts by fours, removing  4, 8, 2, 6, 10, 5, 9, 3, 7, 1.  The first five values are the values of  4a (mod 10)  for  a = 1, 2, 3, 4, 5.  Because  GCD (4, 10) = 2,  this sequence repeats with period  5.  Your trick shifts from the even values to the odd values and then you can count out the five odd values.

Bachet.  Problemes.  1612.  Préface, 1624: A.5.v - A.7.r;  1884: 8-9  &  prob. XX, 1612: 103‑106;  prob. XXIII,  1624: 174-177;  1884: 118‑121.  Turks & Christians -- discusses Josephus as last man.

van Etten.  1624.  Prob. 7 (7), pp. 7‑9 (16‑19).  15 Turks and 15 Christians counted by 9s.  Mnemonics:  Populeam virgam mater Regina ferebat;  Mort tu ne failliras pas en me liurant le trespas.  Discusses other cases, Roman decimation and Josephus as  40  counted by 3s.  In the 1630 edition,  40  is changed to  41.  Henrion's Notte, pp. 9‑10, refers to Bachet's prob. 23 and mentions the correction of  40  to  41.

Hunt.  1631 (1651).  Pp. 266-269 (258-261).  15 Christians & 15 Turks counted by 9s; mnemonic:  Populeam virga mater regina ferebat.  Then does the same counting by 10s and gives the mnemonic:  Rex Paphicum Gente Bonadat Signa Serena.

Yoshida (Shichibei) Kōyū (= Mitsuyoshi Yoshida) (1598-1672).  Jinkō‑ki.  Additional problems in the 2nd ed., 1634.  Op. cit. in 5.D.1.  ??NYS.  Shimodaira (see the entry in 5.D.1) discusses the Josephus problem on pp. 12-14.  He gives some of the information on the Japanese names and on Michinori (1106-1159) and Kenkō, c1331, which is presented under them.  I have a transcription of (some of?) Yoshida into modern Japanese which includes this material as prob. 3 on pp. 66-67.

                    15 children (in black) and 15 stepchildren (in white) counted by 10s.  When  14  stepchildren are eliminated, the last stepchild says the arrangement was unfair and requests the counting to go the other way from him (so that he is number 1 in the counting).  His stepmother agrees and thereby eliminates all her own children.

                    This is discussed in Smith & Mikami, pp. 80-84.  They quote a slightly later version by  Seki Kōwa (1642-1708)  where the stepmother simply reverses the order due to overconfidence.  (On p. 121, they identify the source as  Sandatsu Kempu,  a MS of Kōwa.)  This is also discussed in MUS II 139‑140, where it says that the change in counting was an error on the stepmother's part.  Needham, p. 62, gives a picture from the 1634 ed. of Yoshida, but this is different than the picture in my modern transcription.  I have a photocopy from an 1801 ed.  Dean, 1997, gives the picture, discusses this and provides some additional details, citing the Heibonsha encyclopaedia for the version with the intelligent stepchild.  Dean, 1997, also gives an illustration from a 1767 version called Shinpen Jinko-ki, cf at 1767.

Ahmed el-Qalyubi (d. 1659).  Naouadir (or Nauadir), c1650?, published at Boulaq (a suburb of Cairo), 1892, hist. 176, p. 82.  ??NYS -- described in Basset (1886-1887 below) and MUS II 136.  15 Moslems and 15 infidels on a ship counted by 9s.

Muramatsu Kudayū Mosei.  Mantoku Jinkō‑ri.  1665.  ??NYS -- described in MUS II 139 and Smith & Mikami, pp. 80-84.  Smith & Mikami, p. 81, and Dean, 1997, give Muramatsu's schematic diagrams.  The top diagram is for the classic 15 and 15 counted by 10s.  The second has 32 people counted by 10s to the last two, though the first 15 are coloured black and the second 15 are coloured white, with the last two drawn as squares marked by dice patterns for 5 and 6.

Wingate/Kersey.  1678?.  Prob. 3, pp. 531-532.  15 & 15 counted by 9s or 10s or any other.  Christians and Turks.  From numbers, aid and art / Never will fame depart.  Discusses Josephus.

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see above for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De Ludo Charîgj seu Charîtch, pp. 225-226.  Says it is an Arabic game, i.e. Ludus Exeundi & Eliminadi.  The description is very vague, but it seems to involve counting out in a circle.  The diagram shows a circle of 21 and the text mentions counting by ten or by five.  No reference to any other version of the process.  ??need to read the Latin more carefully.

Emil Schnippel (& Johannes Bolte).  Das St. Peters-Spiel (with a Nachtrag by Bolte).  Zeitschrift für Volkskunde 39 (1929) 190-192  (&  192-194).  Schnippel describes the appearance of solutions of the  St. Peters-Spiel  = Sankt Päders Lek  = Saint Peter's Lake  on 17-18C rune calendars from Sweden, which mystified academics until identified by G. Stephens in 1866.  He gives the vowel-mnemonic:  Populeam virgam mater regina ferebat.  The rune marks are  X  for  Χριστιαvoι (Xristianoi)  and  I  for  ’Ioυδαîoι (Ioudaioi).  He gives the German vowel-mnemonic:  Gott schuf den Mann in Amalek, der (or den) Israel bezwang.  He cites other writers (??NYS) who describe a 1497 MS with 15 & 15 Christians and Jews counted by 10s,  and versions counted by 7s and 11s,  and a version with 17 & 15 Christians and Jews counted by 12s.  He cites:  a 1604 reference to Josephus but without specific numbers;  a 1703 version with 15 & 15 French and Germans;  and a 1782 version with 30 deserters, 15 to be pardoned. 

                    [Nigel Pennick; Mazes and Labyrinths; Robert Hale, London, 1990, p. 37, says that in Finland, stone labyrinths are sometimes called  "Pietarinleikki (St Peter's Game).  The latter name refers to a traditional numerical sequence which appears to be related to the lunar cycle.  It is known from rock carvings and ancient Scandinavian calendars and as an anti-semitic folk-tale."  Can anyone provide details of a connection to the lunar cycle or its appearance in rock carvings??]

                    Bolte's Nachtrag cites Gaidoz et al. (below at 1886-1887) and MUS and an article by himself in Euphorion 3 (1896) 351-362, ??NYS -- cited MUS II 132.  He sketches the history as given by Ahrens.  Mentions the Japanese versions and reproduces Matuoka's picture.  He adds three citations including a 1908 Indian version with 15 honest men and 15 thieves counted by 9s to the last man (??).  He gives vowel-mnemonics in Latin, French, German, English and Italian as follows.

                    Non dum pena minas a te declina degeas.

                    Populeam virgam mater regina ferebat.

                    Mort, tu ne falliras pas.  En me livrant le trépas.

                    So du etwan bist gfalln hart, Stehe widr, Gnade erwart.

                    Gott schlug den Mann in Amalek, den Israel bezwang.

                    From member's (sic) aid and art, Never will fame depart.

                    On tu ne dai la pace ei la rendea.

Ozanam.  Murphy, note 4, says the problem is not in the 1694 ed. -- but see below which could explain why Murphy didn't find it here.

Ozanam.  1696.  Preface to vol. 2 -- first and second of unnumbered pages, which are pp. 269‑270.  1708: Author's Preface -- second and third of unnumbered pp.   Discusses Josephus, citing Bachet.

Ozanam.  1725.  Prob. 45, 1725: 246‑250.  Prob. 17, 1778: 168-171;  1803: 168-171;  1814: 148-150.  Prob. 16, 1840: 76-77.  15 Turks and 15 Christians counted by 9s.  Gives two verse mnemonics:  Mort, tu ne failliras pas, En me livrant le trépas;  Populeam virgam mater Regina ferebat.  Discusses decimation.  Quotes Bachet on Josephus and asserts Hegesippus says Josephus used the method and suggests  41  counted by  3s  (however, Hegesippus doesn't say this!).

Kanchusen.  Wakoku Chiekurabe.  1727.  Pp. 8 & 35.  8 and 8 counted by 8s.  This is pointing out the remarkable fact that one can count out either set first by starting at different points, in different directions.  See:  Dudeney, 1899 & 1905;  Shaw, 1944?

Minguet.  1733.  Pp. 152-154 (1755: 110-111; 1822: 169-171; 1864: 146-148).  15 & 15  by 9s,  whites and blacks.  Populea virga pacem regina ferebat.

Alberti.  1747.  'Modo di disporre 30 cose ...', pp. 132‑134 (77‑78).  15 Christians and 15 Turks or Jews, counted by  3, 8, 9, 10.

Les Amusemens.  1749.  Prob. 16, p. 138: Tiré de Josephe l'Historien.  15 and 15 counted by 9s.  French and Latin mnemonics:  Mort tu ne failliras pas  En me livrant le trépas;  Populeam Virgam Mater Regina ferebat.

Shinpen Jinko-ki (New Edition of the Jonko-ki), more correctly entitled Sanpo Shinan Guruma (A Mathematical Compass).  1767.  BL ORB 30/3411.  ??NYS -- illustration reproduced in Dean, 1997.

Fujita Sadasuke.  Sandatsu Kaigi.  1774.  ??NYS -- cited in a draft version of Dean, 1997, as a Japanese commentary on the problem.

Hooper.  Rational Recreations.  Op. cit. in 4.A.1.  1774.  Vol. 1, recreation XIII, pp. 42-43.  30 deserters of whom 15 are to be punished, counted by 9s.  Populeam virgam mater regina ferebat.

L. Euler.  Observationes circa novum et singulare progressionum genus.  (Novi Comment. Acad. Sci. Petropol. 20 (1775 (1776)) 123‑139.)  = Opera Omnia (1) 7, (1923) 246‑261.  Gets the recurrence for the last man:  L(n) º L(n-1) + k (mod n).

Miyake Kenryū.  Shojutsu Sangaku Zuye.  1795.  ??NYS.  (Described in MUS II 142‑143.)  First(?) to modify Yoshida's problem (1634?) so that the last stepchild sees his imminent fate and asks for the count to restart with him.  Smith, History II 543 and Smith & Mikami, p. 82, give a poorish picture from this.  Dean, 1997, is a better picture.

Matuoka (= Matŭ-oka ??= Matsuoka Nōichi).  1808.  ??NYS -- translated by Le Vallois, with reproductions of the pictures, cf below.  Ahrens, MUS II 140‑142, discusses this, based on Le Vallois and reproduces the main picture from Le Vallois.  Gives Miyake's version.  Le Vallois gives the title as: Mama-ko tate no koto (Problème des beaux-fils (i.e. step-sons)).  There is a diagram showing the counting-out processes.

Ingleby.  Ingleby's Whole Art of Legerdemain, containing all the Tricks and Deceptions, (Never before published) As performed by the Emperor of Conjurors, at the Minor Theatre, with copious explanations; Also, several new and astonishing Philosophical and Mathematical Experiments, with Preliminary Observations, Including directions for practicing the Slight of Hand.  T. Hughes  &  C. Chaple,  London, nd [1815].  Trick L.  The Turks and Christians, pp. 104-106.  15 & 15 counted by 9s.  "This ingenious trick, which is scarcely known, ...."  "From numbers, aid, and art, /  Never will fame depart."

Sanpo Chie Bukuro (A Bag of Mathematical Wisdom).  1818.  BL ORB 30/3411.  ??NYS - illustration reproduced and discussed in Dean, 1997.  Here a man and a woman are studying a set of 29 black and white go stones and the text describes the problem and how to arrange the children.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 83-85, no. 130: Thirty soldiers having deserted, so to place them in a ring, that you may save any fifteen you please, and it shall seem the effect of chance.  15 & 15 by 9s.  Populeam jirgam mater Regina ferebat.  (jirgam  must be a misprint of  virgam.)  Says Josephus and 'thirty or forty of his soldiers' hid in a cave and Josephus arranged to be one of the last.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 16, pp. 4-5 & 54-56.  15 Turks and 15 Christians counted by 9s.  Solution gives:  From numbers, aid, and art, Never will fame depart  and  Populeam virgam mater regina ferebat.

No. 39, pp. 9-10 & 61-62.  Decimation of a troop of 40 to be counted by 10s -- where to place the four ringleaders so they will be the four to be shot.

Rational Recreations.  1824.  Feat 27, p. 103.  15 Turks and 15 Christians counted by 9s.  Gives: From numbers, aid, and art, / Never will fame depart.

Manuel des Sorciers.  1825.  ??NX  Pp. 79-80, art. 40.  15 & 15 by 9s.  Populeam virgam mater regina ferebat.  Mort, tu ne falliras pas   En me livrant au trépas.  Says one can also do  18 & 6 by 8s, etc.  Cf Gaidoz, below, col. 429.

Endless Amusement II.  1826?  P. 117 (misprinted 711 in 1826?): Predestination illustrated.  30 and 10 counted by 12s.

Boy's Own Book.

The slighted lady.  1828: 411‑412;  1828-2: 417-418;  1829 (US): 210-211;  1855: 565‑566;  1868: 670.  13 counted down to last person by 9s.  Before 1868, it gives the survivor for  2, 3, ..., 13,  counted out by 9s.

The partial reprieve.  1828: 417‑418;  1828-2: 422;  1855: 571;  1868: 672-673;  1881: 214.  30 criminals counted by 9s to eliminate 15.  Populeam virgam mater regina ferebat.

Nuts to Crack XIV (1845), no. 72.  21  counted by  7s  to the last man.

Magician's Own Book.  1857.

The fortunate ninth, pp. 221-222.  15 oranges and 15 apples, counted by 9s.  English mnemonics based on vowel coding.

Another decimation of fruit, p. 224-225.  30 apples and 10 oranges, counted by 12s in order to get the oranges first.

The Sociable.  1858.

Prob. 31: The puzzle of the Christians and the Turks, pp. 296 & 312-314.  From numbers aid and art  /  Never will fame depart.  Mort, tu ne faillras pas  /  en me livrant le trepas.  Populeam Virgam Mater regina ferebat.  Then considers counting out 10 from 40, counting by 12s.  Discusses Josephus, citing Hegesippus, and suggests counting by 3s.  = Book of 500 Puzzles, 1859, prob. 31, pp. 14 & 30-32.

Prob. 39: The landlord tricked, pp. 298 & 316.  21 counted by 7s to the last man.  = Book of 500 Puzzles, 1859, prob. 39, pp. 16 & 34.  = Wehman; New Book of 200 Puzzles; 1908, p. 51.

The Secret Out.  1859.  The Circle of Fourteen Cards, p. 87.  This appears to be counting out all 14 cards by 6s (it says by 7s, but it takes the counted out card as one for the next stage), but it's not clear what the object is.  This seems to be a corruption of an earlier version??

Indoor & Outdoor.  c1859.  Part II, prob. 19: The landlord tricked, p. 136.  Identical to The Sociable.

Boy's Own Conjuring Book.  1860.

The fortunate ninth, pp. 190‑191.  Identical to Magician's Own Book.

Another decimation of fruit, p. 194.  Identical to Magician's Own Book.

Vinot.  1860.  Art. XXVI: De l'historien Josèphe, pp. 55-56.  Gives the Josephus story and does it as counting from  41  by  3s  to the last man.

The Secret Out (UK).  c1860.  A delicate distribution, p. 12.  Count 13 by 9s to the last person (different context than Leske).  Mentions counting 12 by 9s.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-30, pp. 254 & 396: Aus 12 Dreizehn machen.  Count 13 by 9s to the last person.

M. le Capitaine Le Vallois.  Les Sciences exactes chez les Japonais.  With comments by Louis de Zélinski & M. Sédillot.  Congrès International des Orientalists (= International Congress of Orientalists).  Compte-Rendu de la première session, Paris, 1873.  Maisonneuve et Cie., Paris, 1874.  T. 1, pp. 289‑299, with comments on pp. 299-303.  The material of interest is on pp. 294-298.  Gives a translation of Matŭ-oka, 1808, and reproduces the pictures, cf above.  Discusses Bachet, Ozanam,  Mort tu ne failliras pas  En me livrant le trépas,  Populeam virgam mater Regina tenebat,  Josephus (saying Josephus arranged to be last).

Hanky Panky.  1872.  The landlord tricked, pp. 129-130.  Identical to The Sociable, prob. 39.

Kamp.  Op. cit. in 5.B.  1877.  No. 7, pp. 323‑324.  28 counted by 9s until one is left.  Footnote seems to refer to 24 counted by 9s.

Mittenzwey.  1880.  Prob. 282-285, pp. 50-52 & 100-101;  1895?: 311-314, pp. 54-56 & 102-103;  1917: 311‑314, pp. 49‑50 & 97‑98.

282 (311):  15 Negroes and 17 Europeans on a ship, counted by 12s.

283 (312):  7 students and a crafty Jew who wishes to make two of the students,  A  &  B,  pay the bill since they had been rude to him.  Initially he is not included.  Starting with  A  and counting  clockwise by 3s,  A  &  B  are left.  Starting with  B  and counting anti‑clockwise by 3s,  A  &  B  are left.  Then the Jew is included and starts with himself, counting anticlockwise by 3s and again  A  &  B  are left.

284 (313):  14 counted by 10s to the last man.

285 (314):  21 counted by 8s to the last man.

Cassell's.  1881.  P. 103: To reward the favourites, and show no favouritism.  = Manson, 1911, p. 256.  15 & 15 counted by 9s.

Henri Gaidoz, Israël Lévi & René Basset.  Le jeu de Saint‑Pierre -- Amusement arithmétique.  This is a series of five notes in Mélusine 3 (1886‑87).

Gaidoz.  Part I.  Col. 273‑274.  Gives classical version with St. Peter, 15 Christians & 15 Jews counted by 9s.  He then gives two versions from Ceylon.  One version is called  Yonmaruma -- The massacre of the Moors --  and has 15 Portuguese & 15 Moors with a Singhalese verse mnemonic.  The second version involves the Portuguese siege of Kandy in 1821, again  15 & 15  by 9s, but different versions have the Portuguese winning or losing.  These versions come from:  The Orientalist 2 (1885) 177, ??NYS.  The editor of The Orientalist added a version learned from an Irish soldier with the vowel‑mnemonic:  From number's aid and art, Never will fame depart.  Gaidoz says he cannot venture a source for the puzzle.

Gaidoz.  Part II.  Col. 307‑308.  Comments on correspondence generated by Part I which provided:  'Populeam virgam mater regina ferebat';  the version with the Virgin instead of St. Peter;  a version with negroes and whites and a negro captain;  a version with French and English;  references to Josephus, Bachet and Ozanam.

Lévi.  Part III.  Col. 332.  Says ibn Ezra's "Tahboula" (Stratagem), c1150, is devoted to this game.  Cites  Schwenter (1623);  Steinschneider's 1880 article discussed above at Ezra;  Steinschneider's Catalog librorum hebr. Biblioth. Bodleianae, col. 687 -- all ??NYS.  Previously Steinschneider opined the game derived from Jahia ibn al‑Batrik's Secret of Secrets (8C), but Lévi says that that is a different amusement involving 9.

Gaidoz.  Part IV.  Col. 429.  Cites:  Le Manuel des Sorciers, Paris, 2nd ed., 1802, p. 70 for a version with French and English.  ??NYS, but see the 1825 ed above.

Basset.  Part V.  Col. 528.  Describes el-Qalyubi, c1650? -- cf above.

Robert Harrison.  UK Patent 15,105 -- An Improved Puzzle or Game.  Applied: 25 Sep 1889;  accepted: 2 Nov 1889.  2pp + 1p diagrams.  12 whites and 12 blacks on a boat with a lifeboat that will hold 12, counted by 6s, called  The Captain's Dilemma.

É. Ducret.  Récréations Mathématiques.  Op. cit. in 4.A.1.  1892? 

Pp. 105-106: Une dame pas contente.  13 counted by 2s to last person.

Pp. 118-119: Stratagéme de Joséphe.  41 counted by 3s to last two, claimed to be the method used by Josephus.

Pp. 120-121: Les marauders punis.  15 & 15 counted by 9s.  Officers and soldiers to be executed.

Pp. 121-122: Les naufrages.  Same numbers, with Turks and Christians on a boat.  Mort, tu ne failliras pas, En me livrant au trépas.

P. 122: Les Élections perfectionnées.  36 counted by 10s -- want the first six chosen.

Hoffmann.  1893.  Chap. 4, pp. 156-157 & 210-211 = Hoffmann-Hordern, pp. 134-135, with photo.

No. 54: Tenth man out.  15  whites and  15  blacks on a ship, counted by  10s,  but first  15  get to go into the lifeboats.  Photo on p. 135 shows L'Equipage Decime, with box and instructions, by Watilliaux, 1874-1895.

No. 55: Ninth man out.  Same, counted by  9s.  Hoffmann cites Bachet and gives a Latin mnemonic.  Photo on p. 135 shows La Question des Boches, with box having instructions on base, 1914-1918.

É. Lucas.  Problem 32.  Intermed. des Math. 1 (1894) 9.  n2  persons, counted by  n  until  n‑1  are left.  "Problème dit de Caligula".

E. Cesarò.  Solution to 32.  Ibid., pp. 30‑31.

J. Franel.  Deuxième réponse [to Problem 32].  Ibid., p. 31.  Cites:  Busche, CR 103, pp. 118, ??NYS.

Adrien Akar.  Troisième réponse [to Problem 32].  Ibid., pp. 189‑190.

E. Lemoine.  Problem 330.  Ibid, pp. 184‑185.  Asks for last man of  n  counted by  p.

Adrien Akar;  H. Delannoy;  J. Franel;  C. Moreau.  Independent solvers of Lemoine's problem.  Ibid., 2 (1895) 120‑122  &  229‑230.  Akar refers to Josephus, Bachet, etc.  Moreau has the clearest form of the recurrence.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 7: The tenth man out.  Almost identical to Hoffmann, no. 54.  No solution.

Lucas.  L'Arithmétique Amusante.  1895.  Pp. 12-18.

Le stratagème de Josèphe, pp. 12-17.  Prob. VI.  15 Christians and 15 Turks, counted by 9s.  Vowel mnemonics:  Mort, tu ne falliras pas, En me livrant le trépas!;  Populeam virgam mater Regina ferebat.  Discusses and quotes Bachet's 1624 Préface which gives the Josephus story and the idea of counting  41  by  3s.

Prob. VII: Le procédé de Caligula, pp. 17-18.  6  and  30  counted by  10s  so as to count the  6  first.

H. Schubert.  Zwölf Geduldspiele.  1895.  P. 125.  ??NYS -- cited by Ahrens; Mathematische Spiele; Encyklopadie article, op. cit. in 3.B; 1904.

E. Busche.  Ueber die Schubert'sche Lösung eines Bachet'schen Problems.  Math. Annalen 47 (1896) 105‑112.

Clark.  Mental Nuts.  1897, no. 13;  1904, no. 22.  The ship's crew.  1897 has the usual  15  and  15  counted by  9s,  starting with the captain, involving whites and blacks on a ship and half being thrown overboard.  1904 has  14  whites and  15  blacks and the captain must discharge  15  at a port.  He joins the crew and starts counting from himself and wants to discharge the  15 blacks.

P. G. Tait.  On the generalization of Josephus' problem.  Proc. Roy. Soc. Edin. 22 (1898) 165‑168.  = Collected Scientific Papers, vol. II, pp. 432‑435.  Says the Josephus passage is "very obscure, ..., but it obviously suggests deliberate fraud of some kind on Josephus' part."  Develops a way of computing the last man.

Les Bourgeois Punis.  Puzzle from c1900, shown in S&B, p. 133.  8 and 2 counted by ?? to leave the 2 at the end.

Dudeney.  A batch of puzzles.  Royal Magazine 1:3 (Jan 1899)  &  1:4 (Feb 1899) 368-372.  The prisoners of Omdurman.  8 Europeans followed by 8 Abyssinians in a ring.  Start counting with the first European.  Determine the counting-out number to eliminate the Abyssinians in sequence.  Doing it in reverse sequence works for any multiple of  16, 15, 14, ..., 9.  The LCM is  720720.  But doing it in forward sequence can be done with  360361.  Since the pattern is symmetric in the two types of people, a change of initial position, but keeping the same direction, will count out the others first.

Dudeney.  "The Captain" puzzle corner.  The Captain 3:2 (May 1900) 97 & 179  &  3:4 (Jul 1900) 303.  The "blacks" and "whites" puzzle  or  The twelve schoolboys.  6 consecutive "blacks" and 6 consecutive "whites" in a circle.  What is the smallest number to count out by which will count out the "whites" first?  You can start anywhere and in any direction.  Answer is  322  and one starts counting on the fourth "white" in the direction of counting.

Sreeramula Rajeswara Sarma.  Mathematical literature in Telugu: An overview.  Sree Venkateswara University Oriental Journal 28 (1985) 77-90.  Telugu is one of the Dravidian languages of south India, spoken in the area north of Madras, and is the state language of Andhra Pradesh.  On pp. 83-87 & 90, he reports finding examples in Telugu in the notebooks of the schoolmaster Panakalu Rayudu (1883-1928) who was a collector of material from many sources.  Unfortunately there is no indication of where Rayudu obtained these examples and Sarma knows of no Indian versions.  He has  15  thieves and  15  brahmins counted by  9s,  then  30  thieves and  30  brahmins counted by  12s.  Solutions are given in some literary form.  The second problem is new to me.  In his notes, Sarma cites the German mnemonic  Gott schuf den Mann in Amalek, der (or den) Israel bezwang  given by Schnippel, and that he has learned that the problem occurs in the Peddabālaśikşa [NOTE:  ş  denotes an  s  with an underdot.], a work which is unknown to me.

H. D. Northrop.  Popular Pastimes.  1901.  No. 8: the landlord tricked, pp. 67-68 & 72.  = The Sociable, no. 39.

Ahrens.  Mathematische Spiele.  Encyklopadie article, op. cit. in 3.B.  1904.  Pp. 1088‑1089 discusses Schubert's work and its later developments.

Benson.  1904.  The black and white puzzle, pp. 225‑226.  As in Hoffmann, no. 54, but first 15 get thrown overboard.  Solution is linear rather than circular.

Dudeney.  Tit‑Bits (14 Oct  &  28 Oct 1905).  ??NYS -- described by Ball; MRE, 5th ed., 1911, pp. 25‑27.  5 and 5 arranged so that one method eliminates one group while another method eliminates the other group.  Determine the two starting points and counts.  Ball doesn't give these values, but seems to imply that both counts go in the same direction, and this is the case in the examples given below.  Ball asks if the starting points can ever be the same for two groups of  C?  He gives solutions for  C = 2  (counted by  4 & 3),  3 (counted by  7 & 8),  4 (counted by  9 & 5).  He believes this question is new.  Note on p. 27 gives the solution for  C = 5,  but with different starting points.  See MRE, 10th ed., 1920, for a general solution with the same starting points.  See:  Kanchusen, 1727;  Dudeney, 1899;  Shaw, 1944?

Pearson.  1907.  Part II, no. 62, pp. 126 & 203.  15 Christians, including St. Peter, who does the counting, and 15 Jews, counted by 9s.

Ball.  MRE, 5th ed.  1911.  See under Dudeney, 1905.

Loyd.  Cyclopedia.  1914.  Christians and Turks, pp. 198 & 365.  = MPSL2, prob. 42, pp. 30‑31 & 134.  Like Dudeney's 1905 version with a different arrangement of 5 and 5.

Williams.  Home Entertainments.  1914.  A decimation problem, pp. 122-124.  15 whites & 15 blacks counted by 10s.  Half have to go over because of shortage of provisions.  Simple circular picture with man counting in middle.

Ball.  MRE, 10th ed., 1920, pp. 26-27.  See under Dudeney, 1905, for the previous version.  Incorporates the solution for the case  C = 5  into the text and adds a general solution due to Swinden.

See:  Will Blyth; Money Magic; 1926 for a related problem.

Collins.  Book of Puzzles.  1927.  Sailors don't care puzzle, pp. 70-71.  15 whites & 15 blacks counted by  10s.  Captain has to throw half over because of shortage of provisions.  Diagram of 15 circles in a row above a picture with 15 circles in a row below, but normally numbered -- it would seem natural in this problem to have the lower row numbered backward to simulate a circle.

William P. Keasby.  The Big Trick and Puzzle Book.  Whitman Publishing, Racine, Wisconsin, 1929.  Counting out problem, pp. 71 & 95.  15 & 15  shown in a circle with the starting point and direction given.  Determine the counting number.

Rudin.  1936.  Nos. 102-103, pp. 37-38 & 98.

No. 102.  13 counted by 9s until last man.

No. 103.  17 and 15 counted by 12s to eliminate the 15 first.

Ern Shaw.  The Pocket Brains Trust -- No. 2.  Op. cit. in 5.E.  c1944.  Prob. 50: Poser with pennies.  Pattern of  5  Hs  and  5  Ts  given -- determine the count to count the  Hs  first, which turns out to be  11.  Though not mentioned, the pattern of  Hs  is equivalent to that for  Ts,  so one can count out the  Ts  first by starting at a different point in the opposite direction.  The pattern is the same as Dudeney (1905).

Robert Gibbings.  Lovely is the Lee.  Dent, London, 1945.  Pp. 111-114.  He says he was shown the puzzle by an old man on the Aran Islands.  Cites Murphy, but his version is different than anything in Murphy.  ??NYS -- information sent by Michael Behrend in an email of 12 Jun 2000.

Le Rob Alasdair MacFhraing (= Robert A. Rankin, who tells me that the Gaelic particle  'Le'  means  'by'  and is not part of the name).  Àireamh muinntir Fhinn is Dhubhain, agus sgeul Iosephuis is an dà fhichead Iudhaich (The numbering of Fionn's men and Dubhan's men, and the story of Josephus and the forty Jews) (in Scots Gaelic with English summary).  Proc. Roy. Irish Acad. Sect. A. 52:7 (1948) 87-93.  A more detailed description than in the Summary appears in Rankin's review in:  Math. Reviews 10, #509b (= A99-5 in:  Reviews in Number Theory).  I had assumed that this was a development from Murphy's article, but Rankin writes that he had not heard of Murphy's article until I wrote for a reprint of Rankin's article in 1991.  He gives a Scottish Gaelic version which is clearly a variant of those studied by Murphy.  He then studies the problem of determining the last man, citing Tait.  For counting out by 2s, the rule is simple.  [There is a story that the the number of mathematicians fluent in Scots Gaelic is so small and the author's name is so obscured that the journal sent the paper to Rankin to referee, not knowing he was actually the author.  The story continues that the referee made a number of suggestions for improvement which the author gratefully accepted.  However, Rankin told me that he was not the referee.  But he did review it for Math. Reviews!]

Joseph Leeming.  Games with Playing Cards Plus Tricks and Stunts.  Op. cit. in 6.BE, 1949.  ??NYS -- but two abridged versions have appeared which contain the material -- see 6.BE.

                    24 Stunts with Cards, 7th & 19th stunts.  A surprising card deal  &  All in order.  Dover: pp. 124  &  131.  Gramercy: pp. 11  &  18.  Both stunts involve dealing cards by putting one out face up, then the next is put at the bottom of the deck, then the next is dealt face up, ....  This process is the same as counting out by  2s.  The object is to produce the cards in a particular order.  The first has 8 cards and wants an alternation of face cards and non-face cards.  The second has the 13 cards of a suit and wants them produced in order.  This is the only example that I can recall of the use of the Josephus idea as a card trick, though other forms of counting out are common, e.g. by counting  1, 2, 3, ...,  or by spelling the words  one, two, three, ....

N. S. Mendelsohn, proposer;  Roger Lessard, solver.  Problem E 898 -- Discarding cards.  AMM 57 (1950)  34-35  &  488-489.  Basically counting out by  2s.  Determine the position of the last card discarded.  No mention of Josephus, though the editor asks what happens if every  r-th card is discarded and gets the recurrence  f(N) º r + f(N‑1) (mod N).

W. J. Robinson.  Note 2876:  The Josephus problem.  MG 44 (No. 347) (Feb 1960) 47‑52.  Analyses what sequences of persons can be removed by varying the count.  Applies to Dudeney's problem.

Barnard.  50 Observer Brain-Twisters.  1962.  Prob. 36: Circle of fate, pp. 41‑42, 64‑65 & 95.  Princess counts out from 17 suitors by 3s.  She sees that her favourite will be the next one out, so she reverses direction and then the favourite is the survivor.

F. Jakóbczyk.  On the generalized Josephus problem.  Glasgow Math. J. 14 (1973) 168-173.  Gives a method of determining when the  i-th man is removed and which is the  k-th to be removed.  Somewhat similar to Rankin's method.

D. Woodhouse  The extended Josephus problem.  Revista Matematica Hispano-Americana 33 (1973) 207-218.  Gets recurrences for the last person, but unnecessarily complicates the process by considering the starting point.  By combining the recurrences, he gets an  n‑fold iteration for the result, but this doesn't really clarify anything.  Only cites Josephus.

Israel N. Herstein & Irving Kaplansky.  Matters Mathematical.  1974;  slightly revised 2nd ed., Chelsea, NY, 1978.  Chap. 3, section 5: The Josephus permutation, pp. 121-128.  They study the permutation where  f(i) = number of  i-th man eliminated, but restrict to the case where one counts by  2s,  which has considerable structure.  Gives a substantial bibliography, mostly included here.

Sandy L. Zabell.  Letter [on the history of the Josephus problem].  Fibonacci Quarterly 14 (1976) 48 & 51.  Sketches the history.

I. M. Richards.  The Josephus problem.  MS 24 (1991/92) 97-104.  Studies the case of counting out by 3s.  Shows the 'Tait numbers', i.e.  n  such that  L(n) = 1 or 2,  are given by  [η(3/2)i + 1/3],  where  η = 1.216703...,  and obtains a formula for  L(n).  Presumably this could be extended to the general case??

Michael Dean.  Josephus and the Mamako-date san (Scheme to benefit the step-children).  International Netsuke Society Journal 17:2 (Summer 1997) 41-53.  There are inro boxes from late 17C Japan which have pictures of the 15 children and 15 stepchildren  problem.  These initially mystified the art historians, but eventually they discovered the Josephus problem and its Japanese forms, but only as far back as Bachet.  Dean gives a brief history for the benefit of art collectors, with references to a number of Japanese sources (some of which I have not seen) -- see above at 1767, 1795, 1818 -- and some photos of the inro boxes (including a fine late 17C example from the collection of Michael and Hiroko Dean) and other material. 

Ian M. Richards.  The Josephus Problem and Ahrens arrays.  MS 31:2 (1998/9) 30-33.  He has finally obtained a copy of Ahrens' work, but from the first edition, and states the result clearly.  For  n  persons, labelled  1, 2, ..., n,  counted out by  k,  if we want to locate the  e-th man counted out, form a sequence starting with  1 + k(n-e)  and then form each next term by multiplying a term by  k/(k-1)  and rounding the result up to an integer.  (I.e.  xn+1 = éxn * k/(k-1)ù.)  Then the position number of the  e-th person eliminated is the difference between  kn + 1  and the largest term in the sequence less than  kn + 1.  The sequence is giving the points where  L(n, k)  is zero in some sense.  Note that when  e = n,  so we are looking for the last person, then the sequence starts at  1,  which is because we start counting with the first person as one.  kn + 1  is the total amount counted in counting  n  people by  k,  For other values of  e,  the change of the starting point of the sequence compensates for the fact that one only counts  k(n-e) + 1  to eliminate the  e-th person.  Ahrens then examined the sequences obtained, with rational multipliers, and found some nice properties which Richards states.  Richards generalises to arbitrary multipliers and finds connections with Beatty sequences,  ëanû.

Ian M. Richards.  Towards an analytic solution of the Josephus problem.  Unpublished preprint sent to me on 21 Mar 1999, 12pp.  (Available from the author, 3 Empress Avenue, Penzance, Cornwall, TR18 2UQ.)  Gets formulae for the case  k = 4  which give the result with a maximum error of ±1.

David Singmaster.  Adjacent survivors in the Josephus Problem.  Nov 2003, 5pp, but may be extended.  This was inspired by the first example in Pacioli's De Viribus, which has 2 'good guys' and 30 'bad guys' arranged in a circle and every 9th person is thrown overboard.  I was struck by the fact that the two survivors were adjacent in the original circle as clearly marked in the marginal diagram.  Offhand it seems an unlikely result, but one soon observes that this remains true as the counting out takes place.  That is, if the two survivors in counting off  N  by  Ks  are adjacent, then this is also true for counting off  n  by  Ks  for  3 £ n < N.  This paper investigates the maximal  N  for which counting out by  Ks  leaves two last survivors who were originally adjacent.

 

          7.C.   EGYPTIAN FRACTIONS

 

          The basic problem is to represent a given fraction as a sum of fractions with unit numerators and distinct denominators, as done by the Egyptians.

          NOTE: Dating of early Egyptian documents is rather uncertain and sources can vary by several hundred years.  I will tend to use dates of Neugebauer and Parker, as given in Gillings.  This dates the composition of the Rhind Papyrus and the Moscow Papyrus as 13th Dynasty, c‑1785, but other sources say the Moscow Papyrus is several hundred years older and other sources date the composition of the Rhind Papyrus to the 12th Dynasty, c-1825.

 

Papyrus Rhind, composed c-1785 (or c-1825), copied c‑1650 (or c-1700).  A. B. Chace, ed.  (1927‑29);  c= NCTM, 1978.  Pp. 21‑22, 50‑51.

Moscow Mathematical Papyrus.  c-1785.  W. W. Struve, ed; Mathematischer Papyrus des Staatlichen Museums der Schönen Künste in Moskau; Quellen und Studien zur Geschichte der Mathematik, Abt. A: Quellen, Band 1; Springer, 1930.

Fibonacci.  1202.  Pp. 77‑83 (S: 119-126): ... de disgregatione partium in singulis partibus [... on the separation of fractions into unit fractions].  He clearly has the idea of taking the smallest  n  such that  1/n £ a/b,  but he doesn't prove that this gives a finite sequence.

J. J. Sylvester.  On a point in the theory of vulgar fractions.  Amer. J. Math. 3 (1880) 332‑335 & 388‑389.

M. N. Bleicher.  A new algorithm for the expansion of Egyptian fractions.  J. Number Theory 4 (1972) 342‑382.  The Introduction, pp. 342‑344, outlines the history.  Pp. 381‑382 give 41 references.

E. J. Barbeau.  Expressing one as a sum of distinct reciprocals.  CM 3:7 (1977) 178‑181.  Bibliography of 20 items.

Paul J. Campbell.  A "practical" approach to Egyptian fractions.  JRM 10 (1977-78) 81-86.  Discusses Fibonacci & Sylvester's methods, etc.  22 references.

Charles S. Rees.  Egyptian fractions.  Math. Chronicle 10 (1981) 13‑30.  Survey with 47 references.

R. J. Gillings.  Mathematics in the Time of the Pharaohs.  Dover, 1982.  He has a long discussion on the Egyptian approach to this topic, discussing and comparing the work in the various sources:  Reisner Papyri (c-2134);  Rhind Papyrus (c-1785);  Moscow Papyrus (c-1785);  Kahun Papyri (c-1785, but later than the previous two items);  Egyptian Mathematical Leather Roll (c-1647),  but he certainly devotes most space to the Rhind Papyrus and the Leather Roll.

 

          7.D.   THE FIRST DIGIT PROBLEM

 

S. Newcomb.  Note on the frequency of use of the different digits in natural numbers.  Amer. J. Math. 4 (1881) 39‑40.  Obtains the law by simply considering logarithms.

F. Benford.  The law of anomalous numbers.  Proc. Amer. Phil Soc. 78 (1938) 551‑572.

E. H. Neville.  Note 2540:  On even distribution of numbers.  MG 39 (No. 329) (Sep 1955) 224‑225.  Says the problem is not precisely defined.  (Not cited in Raimi.)

R. A. Fairthorne.  Note 2541:  On digital distribution.  Ibid., p. 225.  Cites earlier results (see Raimi) and says the law is "a consequence of the way we talk about [numbers]."  (Not cited in Raimi.)

R. A. Raimi.  The first digit problem.  AMM 83 (1976) 521‑538.  Extensive survey and references.

G. T. Q. Hoare & E. E. Wright.  Note 70.5:  The distribution of first significant digits.  MG 70 (No. 451) (Mar 1986) 34‑37.  Generates numbers as ratios of reals uniformly distributed on  (0, 1).  Finds explicit and surprisingly simple probabilities for initial digits of these numbers, which are reasonably close to Benford's probabilities.

Peter R. Turner.  The distribution of l.s.d. and its implications for computer design.  MG 71 (No. 455) (Mar 1987) 26‑31.  l.s.d. = leading significant digit.  Cites some recent articles.

 

          7.E.    MONKEY AND COCONUTS PROBLEMS

 

          Most of these problems are determinate.  Mahavira gives two indeterminate problems, but the next are in Ozanam, with the classic version of the problem first reappearing in Carroll, 1888; Ball, 1890; Clark, 1904; and Pearson, 1907, qv.

          NOTATION.  The classic coconuts problem has the following recurrence for the number of coconuts remaining: 

                    Ai+1  =  (n-1)/n [Ai - 1], 

i.e. each sailor removes  1  (given to the monkey) and  1/n  of the rest.  There are two common endings of the problem.

Ending 0  --  the  n‑th  man leaves a multiple of  n,  so the monkey doesn't get a final coconut.  See:  Mahavira: 131, 132;  Williams;  Moritz;  Meynell;  Leeming.

Ending 1  --  the  n‑th  man leaves one more than a multiple of  n,  so the monkey gets another coconut.  See:  Carroll-Wakeling; Ball; Clark; Pearson; Roray; Collins; Kraitchik; Phillips; Home Book; Leeming; Devi; Allen.

One can extend this to  Ending E  --  the  n-th  man leaves a number  º E (mod n).

Other indeterminate versions:  Ozanam; Dudeney; Weber (Dirac); Rudin.

For the solution with  -(n-1)  coconuts, see:  Roray; Weber (Dirac); Birkhoff & Mac Lane; Anonymous in Eureka; Gardner; Pedoe, Shima & Salvatore; Singmaster.

See  Morris (1988);  Singmaster (1993)  for the alternative division form where the pile is divided equally and the monkey takes one from the remainder, i.e. each sailor takes  1/n  of the pile and then the monkey then takes  1  from the remainder, so the recurrence is 

                    Ai+1  =  (n‑1)Ai/n - 1.  This is similar to the form of recurrence occurring in the determinate versions of the problem, where division takes place first and then some more is included.  Comparing this with the standard form, we see that the forms can be described by the number of coconuts  (mod n)  at each stage.  In the classic form, each  Ai º 1 (mod n),  and in Morris's form, each  Ai º 0 (mod n), so we can conveniently name these  Form 1  and  Form 0.  Unless specified, all examples have  Form 1.

It is easy to generalize to giving  c  coconuts to the monkey at each stage, in either Form, which I call Forms 1c and 0c, but only  Anonymous in Eureka; Kircher; Pedoe, Shima & Salvatore; Singmaster consider this.

Only Kircher considers giving variable amounts to the monkey and he even permits negative values,  e.g. if the monkey is adding coconuts to the pile!

Birkhoff & Mac Lane; Herwitz; Pedoe, Shima & Salvatore consider a variation where no ending is specified except that there is an integral number left after the  n-th division.  A discussion of this version has now been added to Singmaster.

Jackson gives a simple form with no monkey.  Edwards gives a form where the monkey only gets a coconut at the end.

See Tropfke 582.  See also 7.S.1.

 

Hermelink, op. cit. in 3.A, says there are Egyptian versions, presumably meaning some of the simpler determinate types of heap or 'aha' problems in the Rhind Papyrus.

Old Babylonian tablet YBC 4652.  c-1700?.  Transcribed, translated and commented on in:  O. Neugebauer & A. Sachs; Mathematical Cuneiform Texts; American Oriental Society and American Schools of Oriental Research, New Haven, 1945, pp. 100-103, plate 13 & photo on plate 39.  This has fragments of 22 simple problems, of which six can be restored.  The authors say the dating of the tablets discussed in the book is quite uncertain, only stating "they are to be dated to the centuries around 1700 B.C."

No. 7 is reconstructed as:  I found a stone, but did not weigh it; after I added one-seventh and added one-eleventh, I weighed it: 1 ma-na.  What was the original weight of the stone?  In modern notation, this is:  x + x/7  +  (x + x/7) / 11  =  1,  or simply:  x (8/7) (12/11) = 1  which is a simple 'aha' problem. 

No. 8 leads to  x - x/7  +  (x - x/7) / 11  =  1.

No. 9 leads to  x - x/7  +  (x - x/7) / 11  -  [x - x/7 + (x - x/7)/11] / 13  =  1.

No. 19 leads to  6x + 2  +  (6x + 2)·24/21  =  1.

No. 20 leads to  8x + 3  +  (8x + 3)·21/39  =  1.

No. 21 leads to  x - x/6  +  (x - x/6) / 24  =  1.

Old Babylonian tablet YBC 4669.  c-1700?.  Neugebauer and Sachs continue on p. 103 with a new analysis of this table which Neugebauer had previously treated in Mathematische Keilschrift-texte III, op. cit. in 6.BF.2, p. 27.  It leads to  (2/3) (2/3) x + 10 = x/2.

Chiu Chang Suan Ching (Jiu Zhang Suan Shu).  c‑150. 

Chap. VI, prob. 27, p. 69.  Man carrying rice through customs pays  1/3,  then  1/5,  then  1/7  and has  5  left. 

Chap. VI, prob. 28, pp. 69‑70.  Man pays  1/2, 1/3, 1/4, 1/5, 1/6,  making  1  paid out.

Chap. VII, prob. 20, pp. 79‑80.  Man  gains  30%  and sends home  14000;  then  gains  30%  and sends  13000;  then  30%  and  12000;  then  30%  and  11000;  then  30%  and  10000;  leaving  0.  Capital was  30468  84876/371293.  (English in Lam & Shen, HM 16 (1989) 113.)

Zhang Qiujian (= Chang Chhiu‑Chien  = Chang Ch'iu Chien  = Zhang Yo Chien).  Zhang Qiujian Suan Jing (= Chang Chhiu‑Chien Suan Ching) (Zhang Qiujian's Mathematical Manual).  468.  ??NYS.  Chap. II, no. 17.  Man  gains  40%  and withdraws  16000;  then gains  40%  and withdraws  17000;  then gains  40%  and withdraws  18000;  then gains  40%  and withdraws  19000;  then gains  40%  and withdraws  2000;  leaving  0.  Capital was  35326  5918/16807.  (English in Lam & Shen, HM 16 (1989) 117.)

Anania Schirakatzi (= Ananias of Shirak).  Arithmetical problems.  c640.  Translated by:  P. Sahak Kokian as:  Des Anania von Schirak arithmetische Aufgaben; Zeitschrift für d. deutschösterr. Gymnasien 69 (1919) 112-117.  Kokian cites several versions and editions of this Armenian MS as well as some studies on Ananias, but I haven't been able to determine just where Shirak was.  The title varies on the different MSS and Kokian heads the text with one version translated into German:  Des Anania Vardapet Schirakuni Frage und Auflösung [Questions and Solutions of the Priest Ananias of Shirak.]  There are 24 problems, mostly of the 'aha' or 'heap' type.  Only the numerical solutions are given -- no methods are given.  There are several confusing errors which may be misprints or may be errors in the MS, but Kokian says nothing about them.  One problem seems to have omitted an essential datum of the number of grains of barley in a 'kaith'.  I cannot reconcile one solution with its problem (see 7.H).

Prob. 11.  Spend  5/6  thrice, leaving  11.  Answer:  2376.

Prob. 13.  Spend  3/4  thrice, leaving  5.  Answer:  320.

Prob. 19.  (Double and give away  25)  thrice to leave zero.  Answer:  21⅞.  Kokian notes that this and prob. 22 are the earliest occurrences of fraction signs in Armenian.  Hermelink, op. cit. in 3.A, points out that here the doubling is done by God in response to prayer in churches -- then the Arabic world converts the churches to mosques, and then the West reverts to churches, while in the Renaissance, the doubling is by winning at gambling.  In fact, during the Renaissance, it often was by profit from trade.

Prob. 21.  Give away  1/2,  then  1/7,  then  1/8,  then  1/14,  then  1/13,  then  1/9,  then  1/16,  then  1/20,  leaving  570.  Answer:  2240.

Papyrus of Akhmim.  c7C.  Jules Baillet, ed.  Le Papyrus Mathématique d'Akhmîm.  Mémoires publiés par les membres de la Mission Archéologique Français au Caire, vol. IX, part 1, (1892) 1‑89.  Brief discussion of the following problems on pp. 58-59.

Prob. 13, p. 70.  Take  1/13th,  then  1/17th  of the rest, leaving  150.  Answer:  172 + 1/2 + 1/8 + 1/48 + 1/96.  (Also given in HGM II 544.  Kaye I 48, op. cit. under Bakhshali MS, discusses the Akhmim problem and says both Heath and Cantor give misleading references, but I don't see what he means.)

Prob. 17, p. 72.  Take  1/17th,  then  1/19th  of the rest, leaving  200.  Answer:  224 + 1/4 + 1/18.

Bakhshali MS.  c7C. 

                    In:  G. R. Kaye, The Bakhshāli manuscript.  J. Asiatic Soc. Bengal (2) 8:9 (Sep 1912) 349‑361.  P. 358: Sutra 25: example -- merchant pays customs of  1/3,  1/4,  1/5  and finds he has paid  24.  = Kaye III 205, f. 14r.

                    Hoernle, 1888, op. cit. under Bakhshali MS, p. 277 gives the above and the following.  Merchant gains  1/3,  1/4,  1/5,  1/6  and finds he has gained  40.  (Kaye III 205, f. 14r gives this in less detail and it is not clear if Hoernle's statement is what is intended.)  Merchant loses  1/3,  1/4,  1/5  for a total loss of  27.  (Kaye III 205, f. 14v says the remainder is  27  but gives the original amount as  45,  so he seems to have loss and remainder interchanged.).  Merchant loses  1/3,  1/4,  1/5  leaving  20 (can't find in Kaye III ??).

                    Kaye I 48, section 89, says there are 17 examples of this general form, some with the initial value given and the final result wanted, others with the final result given and the initial value wanted.  Gives the first example above and two others with the same rates and a payment of  280  (Kaye III 165, ff. 52r-52v) or a result of  2x - 32,  where  x  is the initial value (Kaye III 207, f. 15r).  Kaye III 204, f. 13v:  start with  60, lose  1/2,  gain  1/3,  lose  1/4,  gain  1/5.  Kaye III 208, f. 16r:  give  2/3,  then  2/5,  then  2/7,  then  2/9,  leaving  3.  How much was given?

                    See also Datta, op. cit. under Bakhshali MS, pp. 44 & 52‑53.  He says the Akhmim problems give the remainder, while the Bakhshali MS and Mahavira problems give the amount paid -- but above we have seen both kinds.  Datta, p. 46, says (Kaye III 184,) f. 70v has a badly damaged problem about a king who gives away  1/2,  1/3  and  1/4  of his money, making  65  given away.  Datta says that the king had only  60  to start!!  But if this is a problem of the type being treated here, then the fractions are applied to the amount left after the previous stage and the king would have  1/4  of his original amount left and he must have had  86 1/3  to start.

Ripley's Puzzles and Games.  1966.  P. 78 asserts that Premysl of Staditze won the kingdom of Bohemia by solving the following.  Give (half and one more) twice, then half and three more to leave zero.  Typically Ripley's gives no details.  The Encyclopædia Britannica says the origin of the Premysl dynasty is obscure, deriving from a plowman who married the Princess Libuse, but giving no date, though apparently by the 9C.  [Rob Humphreys; Prague  The Rough Guide; The Rough Guides, London, (1992), 3rd ed, 1998, p. 249] gives the legends of the founding of Prague.  The maiden queen Libuše, in the 7C or 8C, fell into a trance and told her followers to seek a ploughman with two oxen.  Such a man, named Přemysl (meaning ploughman) was found and produced the dynasty.  He makes no mention of the problem, nor does the Blue Guide for Prague.

Mahavira.  850.  Chap. III, v. 129-140, pp. 67-69 are simple problems of this general type, involving sums of numbers diminished by fractions -- I give just some examples.  Chap. VI,  v. 112, 114, 130, 131, 132, pp. 116 & 123‑125.

     Chap. III.

133.  x(3/4)(4/5)(5/6) + y(1/2)(5/6)(4/5) + z(3/5)(3/4)(5/6)  =  1/2.  This reduces to:  x/2 + y/3 + 3z/8  =  1/2  and he arbitrarily picks two of the values, getting:  1/3, 1/4, 2/3.

134.  x(1/2)(5/6)(4/5)(7/8)(6/7)  =  1/6.

     Chap. VI.

112.  Double and subtract  5,  triple and subtract  5,  ...,  quintuple and subtract  5,  leaving  0  (Datta & Singh I 234, note this gives  43/12  flowers and hence replace  5  by  60  to give  43). 

114.  Less regular problem, leaving  0.

130.  Gives a general technique.

131.  Two sons and mangoes --  (subtract  1  and halve)  twice,  leaving some even number  --  i.e.  Ending 0  with  2  men.  Cf Pearson, 1907.

132.  Man placing flowers in a temple  --  (subtract  1  and delete  ⅓)  thrice,  leaving some multiple of  3  --  i.e.  Ending 0  with  3  men.  Cf Pearson, 1907.

There are some simpler problems in Chap. IV, v. 29‑32, pp. 74‑75.

Chaturveda.  860.  There are some simple examples on pp. 282‑283 of Colebrooke.

Sridhara.  c900.  V. 74(i), ex. 97, pp. 59‑60 & 96.  Give away  1/2,  then  2/3,  then  3/4,  then  4/5,  leaving  3.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  ??NYS.

Pp. 177f. & 128, no. 28 & 45.  Tropfke 585 says these are business trips. 

P. 127, no. 44.  Hermelink, op. cit. in 3.A, says this is Fibonacci's seven gate problem of p. 278, with oranges instead of apples.  Tropfke 585 says it is a problem with porters at an orange orchard.

Abraham.  Liber augmenti et diminutionis.  Translated from Arabic in 12C (Tropfke 662 says early 14C).  Given in:  G. Libri; Histoire des Sciences Mathématiques en Italie; vol. 1, pp. 304-376, Paris, 1838.  ??NYR -- cited by Hermelink, op. cit. in 3.A. 

Bhaskara II.  Bijaganita.  1150.  Chap. 4, v. 114.  In Colebrooke, pp. 196‑197.  (Lose  10,  double,  lose 20)  thrice to triple original.

Fibonacci.  1202.  Pp. 258‑267, 278, 313‑318 & 329 (S: 372-383, 397-398, 439-445, 460‑461) gives many versions!  Chap. 12, part 6, pp. 258‑267 (S: 372-383): De viagiorum propositionibus, atque eorum similium [On problems of travellers and also similar problems] is devoted to such problems.

P. 258 (S: 372-373).  (Double & spend  12)  thrice to leave  0  or  9.  Answers:  10½,  11⅝.  H&S 60 gives English.

P. 259 (S: 374).  Start with  10½,  (double and spend  x)  thrice to leave  0.  H&S 60 gives English.

P. 259 (S: 374).  Same, starting with  11⅝  and leaving  9.

P. 259 (S: 374-375).  (Triple and spend  18)  four times to leave  0.  Answer:  8 8/9.

P. 260 (S: 375).  Start with  8 8/9,  (triple and spend  x)  four times to leave  0.

Pp. 260‑261 (S: 375-376).  (Triple and spend  18)  four times to leave  12,  or the original amount,  or original amount + 20.

P. 261 (S: 376-377).  Three voyages, making profits of  1/2,  1/4,  1/6  and spending  15  each time to leave final profit of  1/2.  Answer:  24 6/7.  Same, with initial amount  24 6/7,  find the common expenditure.  Same, with 'leave final profit  1/2'  replaced by 'leave  21'.

Pp. 261‑266 (S: 377-383).  Many variations.

Pp. 266‑267 (S: 383).  Start with  13,  (double and spend  14)  X  times to leave  0.  H&S 60 gives English.  He gets  X = 3¾  voyages, by linear interpolation between  3  and  4.  Exact answer is  log2 14 = 3.80735.

P. 278 (S: 397-398).  De illo qui intravit in viridario pro pomis collegendis [On him who went into the pleasure garden to collect apples].  Man collects apples in a garden with  7  gates.  (Subtract half and one more)  seven times to leave  1.  H&S 60 and Sanford 221 give English.  Answer:  382.

Pp. 313‑316 (S: 439-443).  Man starts with  100  and spends  1/10  twelve times.  This is not strictly of the type we are looking at, but it is notable that he computes  100 (.9)12  using a form of decimal fraction, getting  28.2429536481.  See 7.L for related problems.

Pp. 316‑318 (S: 443-445).  Exit from a city with  10  gates.  He pays  2/3  of his money and  2/3  more, then  1/i  of his money and  1/i  more for  i = 2, ..., 10,  leaving  1.

P. 329 (S: 460).  Same as on p. 258, done by false position.

P. 329 (S: 460-461).  Start with  12,  (double and spend  x)  thrice to leave  0.

Abbot Albert.  c1240.  Prob. 8, p. 334.  (Double and subtract  1)  thrice, leaving  0.

Chu Shih‑Chieh (= Zhu Shijie).  Ssu Yuan Yü Chien (= Siyuan Yujian) (Precious Mirror of the four Elements  = Jade Mirror of the Four Unknowns).  1303.  Questions in Verse, prob. 4.  ??NYS.  English in Li & Du,  p. 179.  (Double and drink  19)  four times to leave  0.

BR.  c1305.

No. 89, pp. 108‑109.  (Double and spend  35)  thrice leaving  0.

No. 119, pp. 134‑135.  (Double and spend  40)  thrice leaving  0.

Folkerts.  Aufgabensammlungen.  13-15C. 

(Double and give  a)  n  times to leave nothing.  17 sources.  Folkerts notes the solution is  a - a/2n  and the MSS give a general rule.

(Give half and one more )  n  times to leave  c.  12 sources, with the MSS giving a general rule.  Two sources where  half  is replaced by  a quarter.  One irregular example:  Lose half, gain 2; lose half, gain 4; lose half, gain 6; to leave 4.

Munich 14684, XXXIV.  6 sources.

Cites a number of other sources, almost all cited in this section (two items are NYR).

Gherardi?.  Liber habaci.  c1310.  Pp. 144‑145.  Three porters --  i‑th takes half plus  i,  leaving none.

Gherardi.  Libro di Ragioni.  1328. 

Pp. 47‑48.  Man gathering apples.  Four porters --  i‑th takes half plus  5 ‑ i,  leaving  1.

P. 100.  Man makes  12d  on his first trip.  He earns at the same rate on his second trip and then has  100d.  This leads to a quadratic and he finds the positive solution.  See Van Egmond, op. cit. in Common References, pp. 168, 177 & 185 for Italian, English and algebraic versions and some corrections.

Lucca 1754.  c1330. 

Ff. 26r‑27r, pp. 64‑65.  Multiply by  6/5  and spend  12,  multiply by  5/3  and spend  17,  double and spend  20,  leaving  0.

F. 59v, p. 135.  (Double and spend  12)  thrice to leave  3.

Paolo dell'Abbaco.  Trattato di Tutta l'Arte dell'Abacho.  1339.  The first work in the codex Plimpton 167 in the Plimpton collection, Columbia University, New York, is a c1445 copy.  ??NYS -- described in Rara, 435‑440 and Van Egmond's Catalog 254-255.  Van Egmond 365 lists 9 MSS of this work.  MS B 2433, Biblioteca Universitaria, Bologna, is a c1513 copy of just the problems of this work -- Dario Uri has kindly sent a copy of this, but it is somewhat blurry and often illegible; he has now sent a version on a CD which is clearer.  It is dated 1339.  See: Van Egmond's Catalog 67-68.

                    Rara 438 calls it the Dagomari Manuscript and reproduces a figure of a garden with three gates and guards.  Only the first line of the text of the problem is included, but the text is on ff. 25r-25v of B 2433.  (Halve and subtract  1)  thrice to leave  3.

Munich 14684.  14C.

Prob. V, p. 78.  (Double and subtract  2)  some times to leave  0  -- determines initial values for various numbers of times as  2(1 - 1/2n).  The text seems to also consider  (Double and subtract  5).

Prob. VI, p. 78.  (Halve and subtract  1)  thrice to leave  3.

Prob. XXXIV, p. 84.  (Double and subtract  100)  thrice, then  (double and subtract  50)  thrice, leaving  0.  Answer:  92 31/32,

Bartoli.  Memoriale.  c1420.  Prob. 9, f. 75v (= Sesiano pp. 138 & 148).  Man going into a garden to get apples.  Gives  3/4  plus  3  more;  2/3  plus  2  more;  1/2  plus  1  more;  to leave  1.

Provençale Arithmétique.  Written (or more likely copied) at Pamiers, c1430.  MS in Bibliothèque Nationale, Paris, fonds français, nouvelle acquisition 4140.  Previously in the collections of Colbert (no. 5194) and the King (no. 7937).  Partially transcribed/translated and annotated by Jacques Sesiano; Une Arithmétique médiévale en langue provençale; Centaurus 27 (1984) 26-75.  The problems are not numbered, so I will give the folios and the pages in Sesiano.  However the indications of the original folios have not come through on a few pages of my copy and I then only give Sesiano's page.

P. 58.  Man doubles his money and spends  1,  triples and spends  2,  quadruples and spends  2,  leaving him with  3.

F. 113v-114r, p. 60.  (Sell  1/2  and one (or  1/2 ??)  more)  three times to leave  3.  The author gives a general solution as starting with the final result, (adding the extra number and double)  three times to get the original number.

Pseudo-dell'Abbaco.  c1440. 

Prob. 47, p. 44 with plate on p.45.  (Halve and subtract  2)  thrice to leave  7.

Prob. 71, pp. 65‑67 with plate on p. 66.  (Lose    and  6  more)  thrice to leave  24.  (The illustrations are very different from that in Rara (see previous entry).  Rara does not show enough text to see if the numbers used are the same as here, though the wording is clearly different.)  I have a colour slide of this.

AR.  c1450.  No. 185 & 187, pp. 87‑88, 173‑174 & 220.

185  = Fibonacci, p. 258.

187.  Double and spend  6,  double and spend  12,  double and spend  15,  leaving the initial amount.

Muscarello.  1478. 

Ff. 78v-79r, pp. 194-196.  Lose  1/2  and  6 - i  more,  for  i = 1, 2, 3, 4, 5,  leaving  1.

Ff. 84r-85r, pp. 201-204.  Merchant starts with  79  and makes profits of  17%,  19%,  21%,  23%  at four fairs.

della Francesca.  Trattato.  c1480. 

F. 23r (73-74).  Gain  1/3 + 1/4  and  20  more.  Then spend  1/4 + 1/5  and  20 more  to leave  24.

F. 37v (97-98).  Double and spend 11, triple and spend 47, double and spend 34, double and spend 16 to leave 0.  English in Jayawardene.

F. 41v (104).  Identical to f. 23r.

Chuquet.  1484. 

Prob. 30.  (Double and subtract  12)  thrice, leaving  0.  English in FHM 206.

Prob. 31-33 are generalized versions.  E.g. Prob. 31 is  double and spend  5,  triple and spend  9,  quadruple and spend  12  to leave  8.

Prob. 95, English in FHM 219.  Merchant makes a profit of  1/3  and  i  more on his  i-th journey.  He makes as many journeys as he has money to start with.  When does he have  15?  This gives a messy equation:  (4/3)x  =  3 - 9/(x+12).  Chuquet uses some interpolation to estimate  X = Ö(50 16297/16384) - 4 13/128  [FHM misprints this]  =  3.03949414, but I get  3.045827298.  Chuquet says ordinary interpolation is not valid.

Calandri.  Arimethrica.  1491. 

F. 66v.  (Double and spend 2) thrice leaving  0.

F. 74r.  Double and then gain  50%  giving  1000.  Woodcut of merchant on horse.

Pacioli.  Summa.  1494.

F. 105v, prob. 20.  (Give half and one more)  thrice leaving  1.  (See also H&S 58.)

F. 105v, prob. 22.  (Double and spend  12)  thrice leaving  0.

F. 187r, prob. 8.  Start with  x  and double  x  times to get  30.  This gives us  x 2x = 30,  whose answer is  3.21988....  He interpolates both factors linearly on the third day, getting  (3+y)(8+8y) = 30,  so  3+y = 1 + Ö(19/4) = 3.17945.... 

He approaches the following problems similarly.

Ff. 187r-187v, prob. 9.  Start with  x  and make  25%  on each of  x  trips to make  40%  overall.  This gives  x (1.25)x = 1.4 x.

F. 187v, prob. 10.  Leads to  x (1.2)x = x2.

F. 187v, prob. 11.  Leads to  x (1.4)x = 6x.

F. 188r, prob. 13.  Start with  13,  (double and spend  14)  x  times to leave  0.  He observes that each iteration doubles the distance from  14,  so the problem leads to  2x = 14,  but again he has to interpolate on the third day. 

Calandri, Raccolta.  c1495.  Prob. 16, pp. 17‑18.  Merchant gains    of his money plus  i  on the  i-th trip.  After three trips he has  15.

Pacioli.  De Viribus.  c1500.  Ff. 120r - 120v, 111r - 111v (some pages are misbound here). C(apitolo). LXVII. un signore ch' manda un servo a coglier pome o ver rose in un giardino (A master who sends a servant to gather apples or roses in a garden).  = Peirani 156-158.  (Lose half and one more)  three times to leave  1.  Discusses the problem in general and also does  (Lose half and one more)  five times to leave  1;  (Lose half and one more)  three times to leave  3.

Blasius.  1513.  Ff. F.iii.r - F.iii.v: Decimaquinta regula.  Sack of money.  First man takes half and returns  100;  second takes half and returns  50;  third takes half and returns  25;  leaving  100  in the sack.

Johannes Köbel.  Rechenbiechlein auf den linien mit Rechenpfeningen.  Augsburg, 1514.  With several variant titles, Oppenheim, 1518;  Frankfort, 1531, 1537, 1564.  1564 ed., f. 89r, ??NYS.  Lose half, gain  100,  lose half, gain  50,  lose half, gain  25  to yield  100.  (H&S 58‑59 gives German and English.)

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 29, f. 66v.  Start with  100  and have to bribe ten guards with  1/10  each time.  Computes the exact residue, i.e.  100 x .910.  (H&S 59-60).

Tonstall.  De Arte Supputandi.  1522.

Quest. 43, p. 173.  Give half plus  i+1,  for  i = 1, 2, 3,  leaving  1.  (H&S 61 cites this to the 1529 ed., f. 103)

Quest. 44, pp. 173-174.  Give half and get back  2i  for  i = 1, 2, 3,  leaving  12.

P. 246.  (Double and spend  12)  thrice to leave  0.

Riese.  Die Coss.  1524.  Several examples -- no. 35, 53, 55, 56, 58, 59, 60, 61, 62.  I describe a few.

No. 35, p. 45.  Man stealing apples:  (give half and one more)  four times, leaving  1.

No. 53, pp. 47‑48.  (Double and spend  12)  thrice, leaving  0.

No. 55, p. 48.  (Double and spend  i)  for  i = 1, 2, 3,  leaving  10.

No. 58, p. 48.  x + (4x+1) + (3(4x+1)+3)  =  56.

No. 61, p. 48.  (Give half plus  2+2i  more)  for  i = 1, 2, 3,  leaving  0.

No. 62, p. 49.  (Give half less  2i)  for  i = 1, 2, 3,  leaving  12.

Tartaglia.  General Trattato.  1556.  Book 12, art. 34, p. 199v.  Book 16, art. 47 & 113-116, pp. 246r & 253v‑254r.  Book 17, art. 9 & 20, p. 268v & 271r.  Final remainder specified in each case.

12-34.  (Take half plus  i  more)  for  i = 1, 2, 3, 4,  leaving  1.  Cf 16-115.

16-47.  Take  1/2  and  1  more,  1/3  and  2  more,  1/4  and  4  more, leaving  26.

16-113.  (Double and subtract  20)  thrice, leaving  0  (H&S 61 gives Latin and English of this one and says it appears in van der Hoecke (1537), Stifel (quoting Cardan) (1544), Trenchant (1566) and Baker (1568) (but see below).

16-114.  (Halve and subtract  1)  thrice, leaving  1, 2, ....

16-115.  Halve and subtract  1,  then  2,  3,  4,  leaving  1.  Cf 12-34.

16-116.  Lose  1/2  and  3  more,  lose  2/3  and get back  10,  lose  3/4  and  6  more,  lose  4/5  and get back  16,  leaving  24.

17-9.  Double & spend  4,  double & spend  8,  leaving  24.

17-20.  Double and spend  18,  double and spend  24,  double and spend  36,  leaving  280.

Buteo.  Logistica.  1559. 

Prob. 6, pp. 334-335.  Lose  1/2  and  3  more,  lose  1/3  and  4  more,  lose  1/4  and  gain  1,  to leave  100.

Prob. 13, pp. 342-343.  Start with  X,  gain  40.  Make the same rate of profit twice again and then the second of these gains is  90.

Prob. 19, p. 347.  Double and spend  12,  triple and spend  15,  quadruple and spend  14,  leaving  12.

Prob. 20, pp. 347-348.  Gain  1/4  and spend  7,  gain  1/3  and spend  10,  lose  3/7  and spend  8,  leaving  0.

Prob. 21, pp. 348-350.  (Double and spend  10)  X  times to leave  0.  He makes an error at  X = 8  and deduces  X = 7.

Baker.  Well Spring of Sciences.  1562?

Prob. 7,  1580?: ff. 192r-193r;  1646: pp. 302‑304;  1670: pp. 344-345.  Lose half and gain  12,  lose half and gain  7,  lose half and gain  4,  leaving  20.

Prob. 8,  1580?: ff. 193r-193v;  1646: p. 304;  1670: p. 345.  (Double and spend  10)  thrice leaving  12.

Gori.  Libro di arimetricha.  1571.  F. 72r (pp. 77‑78).  (Lose half and one more)  four times to leave  3.

Book of Merry Riddles.  1629?  (Take half and half more) thrice, leaving one.

Wells.  1698.  No. 118, p. 209.  Soldiers take half of a flock of sheep and half a sheep more, thrice, leaving 20.

Ozanam.  1725.  Prob. 28, question 1, 1725: 211‑212.  (Give half the eggs and half an egg)  thrice.  He doesn't specify the remainder and says that  8n‑1  eggs will leave  n‑1  and that one can replace  8  by  2k  if one does the process  k  times.  Montucla replaces this by some determinate problems -- see below.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XV, pp. 86-87 (1790: prob. XXVII, pp. 88-89.  Shepherd loses  (half and  1/2  more)  thrice to leave  5  (or  1)  sheep.

Les Amusemens.  1749. 

Prob. 108, p. 249.  (Double and give away  6)  thrice to leave  0.

Prob. 111, pp. 252-253.  Double and spend  20,  triple and spend  27,  double and spend 19,  leaving  250.

Prob. 118, p. 260.  (Halve and give  1/2  more)  thrice to leave  0.

Walkingame.  Tutor's Assistant.  1751. 

1777: p. 82, prob. 7;  1860: p. 111, prob. 7.  Stealing apples.  Give half and get back  10,  give half and get back  4,  give half, get back  1,  leaving  24.

1777: pp. 174-175, prob. 86; 1860: p. 183, prob. 85.  Sheep fold robbed of half its sheep and half a sheep more, thrice, leaving  20.

Euler.  Algebra.  1770.  I.IV.III: Questions for practice, no. 9, p. 204. 

          (Spend  100, then gain  ⅓)  thrice to double original value.

Vyse.  Tutor's Guide.  1771? 

Prob. 18, 1793: p. 32; 1799: p. 36 & Key p. 29.  Sheep fold.  (Lose half and  ½  more) thrice, leaving  20.

Prob. 31, 1793: p. 57; 1799: p. 62 & Key p. 69.  (Gain  ⅓,  less  100)  for    years to yield  £3179 11s 8d.  Solution assumes the final quarter year is the same as  (gain  1/12,  less  25),  but it is not obvious how to determine an appropriate expression for the quarterly effect.  In general, repeating  ax - b  four times gives  a4x - b(a4-1)/(a-1)  and setting this equal to  4x/3 - 100  gives  a = 1.075..,  b = 22.371.  However, the  100  is the expenses of the merchant's family and he may not be able to reduce it in one quarter.

Prob. 33, 1793: pp. 57-58; 1799: pp. 62-63 & Key p. 70.  Lose  ½,  get back  10;  lose  ⅓, get back  2;  lose  ½,  get back 1;  leaving 12.

Prob. 2, 1793: p. 129; 1799: p. 137 & Key p. 179.  (Double and spend  6)  thrice, leaving  0.  Solution by double position.

Prob. 4, 1793: p. 129; 1799: p. 137 & Key p. 180.  Lose  ½,  gain  10;  lose  ½,  gain  4;  lose  ½,  gain  1;  yielding  18.  Solution by double position.

Dodson.  Math. Repository.  1775.

P. 10, quest. XXIV.  Double and spend 6; triple and spend 12;  quadruple and spend 18; leaving 30.

P. 47, quest. C.  Shepherd loses ¼ of his flock and ¼ of a sheep; then ⅓ of his flock and ⅓ of a sheep; then ½ of his flock and ½ of a sheep; and has 25 sheep left.

P. 48, quest CI.  Man (spends 50 and gains ⅓ on the remainder) thrice, yielding double his original amount.

P. 49, quest. CII.  Lose ¼, win 3, lose ⅓, win 2, lose 1/7, yielding 12.

Ozanam‑Montucla.  1778.

Prob. 15, part a, 1778: 207-208;  1803: 203.  Prob. 14, 1814: 175‑176;  1840: 91.  (Sell half the eggs and half an egg more)  thrice to leave  36.  This was one of the more popular forms of the puzzle after this time -- see: Jackson, Endless Amusement II, Nuts to Crack, Young Man's Book, Boy's Own Book, Magician's Own Book, Boy's Own Conjuring Book, Wehman, Collins, Sullivan.

Prob. 15, part b, 1778: 208-209;  1803: 203-204.  Prob. 15, 1814: 176-177;  1840: 91.  (Spend half and  1/2  more)  thrice leaving  0.  Gives the rule for the problem with more iterations.

Bonnycastle.  Algebra.  1782.  P. 86, no. 20 (1815: p. 107, prob. 30).  Lose  1/4  of what he has, win  3,  lose  1/3,  win  2,  lose  1/7,  leaving  12.

Eadon.  Repository.  1794. 

P. 296, no. 9.  Man loses  1/4,  then gains  3,  then loses  1/3,  then gains  2,  then loses  1/7,  then has  12.

P. 296, no. 10.  Man (spends  50  and then gains  1/3)  thrice to double his money.

P. 297, no. 14.  Man sends out  1/3  and  25  more of his men, leaving  1/2  and  100  more.

Bonnycastle.  Algebra.  10th ed., 1815.  P. 106, no. 19.  (Double and pay  1)  four times, leaving  0.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.

No. 10, pp. 16 & 73.  (Give one half)  seven times, leaving  1.

No. 13, pp. 17 & 74.  (Give half plus  i)  for  i = 1, 2, 3,  leaving  1.

No. 18, pp. 18 & 75.  (Spend half plus half a guinea)  four times, leaving  0.

No. 19, pp. 18-19 & 75-76.  (Sell half, receive  10);  (sell third, receive  2);  (sell half, receive  1);  leaving  12.

No. 23, pp. 19-20 & 77.  Same as Ozanam-Montucla, prob. 15a.

Endless Amusement II.  1826?  Prob. 27, pp. 202-203.  Identical to the 1803 English of Ozanam-Montucla, prob. 15a.  = New Sphinx, c1840, p. 138.

Nuts to Crack II (1833), no. 126.  Same as Ozanam-Montucla, prob. 15a.

Young Man's Book.  1839.  Pp. 234-235.  Identical to the 1803 English of Ozanam-Montucla, prob. 15a.

Boy's Own Book.  1843 (Paris): 344.  Same as Ozanam-Montucla, prob. 15a.  = Boy's Treasury, 1844, p. 301.  = de Savigny, 1846, p. 289: La paysanne et les œufs.

Magician's Own Book.  1857.

Countrywoman and eggs, pp. 238-239.  Almost identical to Boy's Own Book.

The old woman and her eggs, p. 240.  (Give half and half an egg more)  thrice, leaving  1.

The apple woman, p. 252.  Sell half, gain  10,  sell third, gain  2,  sell half, gain  1, have  12  remaining.  = Book of 500 Puzzles, 1859, p. 66.  = Illustrated Boy's Own Treasury, 1860, prob. 15, pp. 428 & 431-432.

Boy's Own Conjuring Book.  1860. 

Countrywoman and eggs, p. 205.  Almost identical to Boy's Own Book.

The old woman and her eggs, p. 212.  Identical to Magician's Own Book.

The apple woman, p. 223.  Identical to Magician's Own Book.

Vinot.  1860.  Art. LV: Les œufs, pp. 72-73.  (Sell half the eggs and half an egg more)  thrice to leave  0. 

Lewis Carroll.  Letter of 22 Jan 1878 to Jessie Sinclair.  = Carroll-Collingwood, pp. 205-207.  Cf Carroll-Wakeling, prob. 26, pp. 34 & 72.  "Tell Sally it's all very well to say she can do the two thieves and the five apples, ...."  Wakeling omits the number of apples since it is the answer to the problem he poses.  Cohen and Wakeling give a possible version of the problem, provided by Peter Heath, as:  steal half and half an apple more, then the second thief steals half of what the first thief stole and half an apple more, leaving none, for which the answer is  5.  Cohen says John Fisher [The Magic of Lewis Carroll; op. cit. in 1, p. 79] cites a similar problem from the notebooks of Samuel Taylor Coleridge, but this is:  (sell half and half an egg more) thrice, leaving three.  Another possibility, which seems much more likely to me, would be  (steal half and half an apple more) twice, leaving  5,  for which the answer is  23.  This is the more common form of the problem, whereas Heath's version takes  5  as the answer rather than the data.  In Carroll-Gardner, pp. 77‑78, Gardner gives a totally different explanation, saying this is an old magic trick and explaining it.

Mittenzwey.  1880. 

Prob. 103, pp. 21 & 73;  1895?: 120, pp. 25-26 & 75;  1917: 120, pp. 24 & 72-73.  Egg woman sells (half of her eggs and half an egg more) four times, leaving  1.  In 1917, the solution is expanded and he notes that starting with  2k - 1,  four stages bring you to  2k-4 - 1,  but he doesn't seem to understand how to solve the problem in general.

Prob. 120, pp. 24-25 & 76;  1895?: 138, pp. 28-29 & 79;  1917: 138, pp. 26 & 76‑77.  Three men successively taking  1/3  of a pile of potatoes.  Remainder is  24.  Observes that the second person is entitled to  3/8  of the remainder and the third person gets the rest [but this is unnecessary information].  How many potatoes were there?

William J. Milne.  The Inductive Algebra Embracing a Complete Course for Schools and Academies.  American Book Company, NY, 1881.  Pp. 138 & 332, no. 81.  (Double and lose one)  thrice to get triple original amount.

Hoffmann.  1893.  Chap. IV gives several deterministic examples:  nos. 27, 39, 46, 67, 76 (? see 7.E.1), 111, 112.

Lucas.  L'Arithmétique Amusante.  1895.  P. 184: Prob. XLII: La marchande d'Œufs. 

          (Sell  1/2  plus half an egg more)  n  times to leave  0.

Carroll-Wakeling.  1888  Prob. 17: Four brothers and a monkey, pp. 21 & 68.  This has a pile of nuts on a table and is Form 1, Ending 0.  Wakeling gives the solution  765  and says there are other solutions, citing  2813  and  5885,  i.e.  765 + 2048 k,  but the general solution is actually  765 + 1024 k.

                    This is one of the problems on undated sheets of paper that Carroll sent to Bartholomew Price.  Wakeling said he will look for a watermark on it, but the date is now pretty definite.  Wakeling says there is no other mention of the problem in Carroll's work, MSS or correspondence.  Carroll-Gardner, p. 53, mentions Carroll-Wakeling and cites his 1958 article.

                    On 28 May 2003, Wakeling kindly sent me copies of three items of the Carroll/Price material.  First is Carroll's solution of the problem, which is typewritten, 'probably using Dodgson's Hammond typewriter, purchased in 1888.'  This solution is grossly erroneous -- he only takes three stages and obtains the answer  61 + 64k.  Most importantly, Wakeling sent a note from John (later Sir John) Evans to Price, dated 15 Oct 1888, thanking Price for his solution of the problem and saying that his attempt had gotten to a value of  1789  (which is a correct solution!).  Evans then adds that he cannot make Price's solution work.  Price must have given  253,  but after the fourth brother, there remain  78  which is not divisible by four (nor is it one more than a multiple of four).  Evans than says that  509 (= 253 + 256)  and  765 (= 253 + 512)  also fail, 'I think'.  Wakeling also sent the statement, only, of the problem, in Evans' handwriting, headed  Four Brothers & the Family Monkey -- this differs from the version in Carroll-Wakeling. 

                    Though this is a moderately messy problem, it is depressing to discover that three competent mathematicians were unable to get the correct solution and failed to check the solutions that they had obtained!!  However, we now know that the problem was in circulation in 1888, and the fact that wrong answers were being obtained shows that the problem was new at that time.

W. W. Rouse Ball.  Elementary Algebra.  CUP, 1890 [the 2nd ed. of 1897 is apparently identical except for minor changes at the end of the Preface].  Prob. 12, p. 260 & 475.  Three Arab jugglers and their monkey, on their way to Mecca, buy a basket of dates.  Ending 1.  He gives no background to the problem nor any indication that it is novel.  The solution is just the numerical answer. 

The presence of two versions by c1890 would indicate the problem must have been known to at least a few other people.  The fact that both Carroll and Ball knew of the problem leads one to conjecture a possible mode of transmission.  After the appearance of Alice in 1865, Carroll's reputation was immense.  Ball's A Short Account of the History of Mathematics appeared in 1888 and was well-known in the English-speaking world and Ball was becoming known as an authority on the history of mathematics and on mathematical recreations.  Either or both might have been sent the problem from India (or anywhere in the world), perhaps by the translator of Mahavira when he first came on the problem.  However, I have examined the Ball material at Trinity College Cambridge and it is clear that he (or his heirs) disposed of his correspondence, so this conjecture cannot be verified.  We hope 'something will turn up' to elucidate this.  Something has turned up -- see the further material under Carroll-Wakeling -- but this does not determine the source of the problem.

Clark.  Mental Nuts.  See also at 1904.

1897, no. 11;  1904, no. 19;  1916, no. 17.  The man and his money.  (Spend  1/2  and  1/2  more)  four times to leave  0.

1897, no. 20;  1904, no. 92;  1916, no. 8.  The man and the stores.  (Double and pay  10)  thrice to leave  0.

1897, no. 29;  1904, no. 40.  The farmer and his horses.  (Pay  1,  then sell  1/2,  then pay  1  more)  four times to leave  1.

Dudeney.  Problem 522.  Weekly Dispatch (8  &  22 Nov 1903) both p. 10.  Multiple of  25  eggs.  Sell half and half an egg more until all gone.

Clark.  Mental Nuts.  1904, no. 99.  Three boys and basket of apples.  Coconuts -- Ending 1  with  3  people.  (This is not in the 1897 or 1916 eds.  This complicates the possible connection with Mahavira -- cf the discussion under Pearson, below.)

Pearson.  1907.  Part II.  Several determinate versions, and the following.

No. 29: The men, the monkey, and the mangoes, pp. 119 & 197.  Coconuts -- Ending 1  with  3  people.  Gives only one solution.

No. 94: One for the parrot, pp. 133‑134 & 210.  Coconuts -- Ending 1  with  4  boys, a bag of nuts and a parrot.  Gives only one solution.

The connection of these with Mahavira, 850, bemused me and I conjectured that Pearson might have heard of Mahavira's work, though the translation didn't appear until 1912.  Kaye's note (see under Mahavira in the Abbreviations) shows that an advance version of the translation was produced in 1908, which makes my conjecture much more likely.  The Frontispiece of Pearson's book shows him as a clergyman of about 40-50 years old, and another of his books describes him as MA of Balliol College, Oxford and Rector of Drayton Parslow, Buckinghamshire, with a stamp underneath giving Springfield Rectory, Chelmsford [Essex], so he might have been a missionary or had Indian contacts.  (Incidentally, the publisher Cyril Arthur Pearson was his son.)  HOWEVER, I have now seen Carroll, Ball and Clark and this makes the connections less clear.

Wehman.  New Book of 200 Puzzles.  1908.

P. 50: The sheepfold robbery.  (Lose  1/2  and  1/2  a sheep more)  thrice leaving  2.

P. 51: The maid and her apples.  c= Magician's Own Book, p. 252.

P. 57: The countrywoman and her eggs.  Same as Ozanam-Montucla, prob. 15a.

Nelson L. Roray, proposer;  A. M. Harding, Norman Anning and the proposer, solvers.  Problem 288.  SSM 12 (1912) 235  &  520‑521.  Coconuts -- Ending 1  with  3  men.  Anning shows that the solution is  ‑2 (mod 34),  but none of the solvers generalise to  n  men.

Loyd.  Newsboys puzzle.  Cyclopedia, 1914, pp. 116 & 354.  (= MPSL2, prob. 9, pp. 8 & 123.  = SLAHP: Family rivalry, pp. 51 & 103.)  Complex specification of one amount.

Loyd.  A study in hams.  Cyclopedia, 1914, pp. 268 & 375.  (= SLAHP: The Ham peddler, pp. 81 & 117.)  (Half plus half a ham more)  four times,  (half a ham plus half),  (half plus half a ham),  leaving  0.

R. L. Weber.  A Random Walk in Science.  Institute of Physics, London & Bristol, 1973.  P. 97 excerpts a Russian book on humour in physics which states that P. A. M. Dirac heard a version of the problem with three fishermen and a pile of fish, but only three divisions, at a mathematical congress while he was a student (at Cambridge?) and gave the solution,  -2.  In fact, he only came to Cambridge as a graduate student in 1923 and became a fellow in 1927, so that the story, if true and if it refers to his time at Cambridge, relates to the mid 1920s.

Ben Ames Williams.  Coconuts.  Saturday Evening Post (9 Oct 1926) 10,11,186,188.  Reprinted in:  Clifton Fadiman, ed.; The Mathematical Magpie; Simon & Schuster, NY, 1962, pp. 196-214.  Ending 0  with  5  men.

R. S. Underwood, proposer;  R. E. Moritz, solver.  Problem 3242.  AMM 34 (1927) 98 (??NX)  &  35 (1928) 47‑48.  General version of coconuts problem,  Ending 0  with  n  men.

Wood.  Oddities.  1927. 

Prob. 52: A goose problem -- not for geese to solve, pp. 43 & 44.  Sell a half and half a goose more;  sell a third and a third of a goose more;  sell a quarter and 3/4 of a goose more;  sell a fifth and a fifth of a goose more;  leaving 19.

Prob. 57: Eggs this time, p. 46.  Sell half and half an egg more;  sell a third and a third of an egg more;  sell a quarter and a quarter of an egg more;  sell a fifth and a fifth of an egg more;  leaving a multiple of 13.  Determine the least number of possible eggs.  Gives answer  719.  Complete answer is  719 (mod 780).

Collins.  Book of Puzzles.  1927.  The basket of eggs puzzle, p. 77.  Same as Ozanam-Montucla, prob. 15a.

Collins.  Fun with Figures.  1928.  The parrot talks, pp. 183-185.  Four boys and a parrot and a bag of nuts.  Ending 1  with  n = 4.  = Pearson 94.

Kraitchik.  La Mathématique des Jeux, 1930, op. cit. in 4.A.2.  P. 13.

Prob. 39.  Monkey and mangoes problem,  Ending 1  with  3  men.  = Pearson 29.  = MR, 1942, prob. 35, pp. 32-33.

Prob. 40.  (Take  ⅓)  thrice, leaving  8.  = MR, 1942, prob. 36, p. 32.

He asserts these are Hindu problems but gives no source.

Rudin.  1936.  No. 5, pp. 3 & 76.  Three men.  (Take away ⅓)  thrice, then divide in thirds.  He gives only the answer  81,  though any multiple of  81  works.

Hubert Phillips.  Question Time.  Op. cit. in 5.U.  1937.  Prob. 203: Adventure island, pp. 137 & 246.  Ending 1  with  5  men and Friday instead of a monkey.

Francis & Vera Meynell.  The Week‑End Book.  Nonesuch Press, 1924 and numerous printings and editions.  I have 8th printing, 2nd ed., Mar 1925, and a 5th(?) ed., in 2 vols., Penguin, 1938.  The earlier edition has some extra text surrounding the problems, but has only 8 of the 12 problems in the Penguin ed.  This problem is not in the 2nd ed.  5th?? ed., prob. seven, p. 408: Three men and a monkey.  Ending 0,  with  3  men.  No solution.

Joseph Bowden.  Special Topics in Theoretical Arithmetic.  Published by the author, Lancaster, Pennsylvania, 1936.  The problem of the dishonest men, the monkeys and the coconuts, pp. 203-212.  ??NYS - cited by Pedoe, Shima & Salvatore.

McKay.  At Home Tonight.  1940. 

Prob. 33: The niggers and the orchard, pp. 69 & 82.  Three men and apples.  Ordinary division with the extra thrown away and Ending 1.

Prob. 35: Dividing nuts, pp. 70 & 83.  Divide nuts among  5  girls with one left over.  One girl divides hers among the rest, with one left over.  Then another girl divides hers among the rest, with one left over.

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  Pp. 149‑150: The bag of peanuts.  Five Italians, a bag of peanuts and a monkey.  Coconuts -- Ending 1  with  5 people.  The given answer is  3121,  but it should be  15621.  3121  would be the answer for Ending 0 -- cf Leeming, 1946.

Garrett Birkhoff & Saunders Mac Lane.  A Survey of Modern Algebra.  Macmillan, (1941, ??NYS), revised, 1953.  Prob. 11, p. 26 (p. 28 in the 4th ed. of 1977).  Usual five men and a monkey form, but no ending is specified.  That is, nothing is stated about the number left over in the morning except that it is an integer.  No answer given, but there is a hint to try  -4  coconuts.  Somewhat surprisingly, the answer is the same as for the  Ending 0  problem -- see Pedoe, Shima & Salvatore.

Leeming.  1946.  Chap. 5, prob. 16, pp. 57‑58 & 177.  Five Italian organ grinders, their monkey and a pile of peanuts.  Coconuts -- Ending 1  with  5 people.  The given answer is  3121,  but it should be  15621.  3121  would be the answer for Ending 0 -- cf Home Book, 1941.

Sullivan.  Unusual.  1947.  Prob. 40: Another old problem.  Sell (half plus half an egg more)  thrice, leaving  36.  = Ozanam-Montucla, prob. 15a.

Paul S. Herwitz.  The theory of numbers.  SA 185:1 (Jul 1951) 52-55.  With letter and response in SA 185:3 (Sep 1951) 2 & 4.  Gives the Birkhoff & Mac Lane version with  n = 3  and solves it by explicitly computing the number remaining in terms of the initial number, obtaining one linear diophantine equation in two unknowns.  He states the equation for the case  m = 4,  but doesn't give the solution, and for the general case, though he doesn't sum the geometric progression that appears.  The letter requests the solution for the case  m = 5  and Herwitz outlines how to find the solution by the Euclidean algorithm, obtaining  3121.

Anon.??  Monkeys and coconuts.  Mathematics Teacher 54 (Dec 1951) 560-562.  ??NYS - cited by Pedoe, Shima & Salvatore.

Anonymous.  The problems drive.  Eureka 17 (Oct 1954) 8-9 & 16-17.  No. 2.  Three men and cigarettes guarded by a Boy Scout.  Two are given to the scout at each division and at the end, so this is Form 1c, Ending c, with  c = 2.  Solution observes that  -(n-1)c  is a fixed point and the solution is  -(n-1)c (mod nn+1),  as seen in Singmaster, but the solution here doesn't give any proof. 

Ron Edwards.  The cocoanut poker deal.  The Cardiste (Mar 1958) 5-6.  Uses the three person problem as the basis of a card trick.  He states the original problem in a novel form -- each hunter finds the pile evenly divisible by three, so the monkey doesn't get any coconuts until the morning division when he gets one.  But in the trick, Edwards uses the classic form, with a variation.  The  52  cards are dealt into  3  piles,  with one extra put in a discard pile.  The spectator places one of the end piles on the middle giving a new deck of  34.  The process is repeated twice more leaving  22,  then  14  cards.  These  14  are dealt into three piles, but now there are two extras which are discarded and then the five discards are found to be a royal flush!  (The Cardiste was a mimeographed magic magazine which ran from 1957 to 1959.  My thanks to Max Maven for remembering and finding this and sending a copy.)

M. Gardner.  SA (Apr 1958)  = 2nd Book, chap. 9.  Describes the Williams story of 1926 and says the Post received 2000 letters the first week after publication and the editor telegrammed:  "For the love of Mike, how many coconuts?  Hell popping around here."  Gardner says Williams modified the older problem of Ending 1  with  5  men.  He gives the solution of  -4,  but says he could not trace its origin.  His addendum cites Anning (1912) for this.

Roger B. Kircher.  The generalized coconut problem.  AMM 67:6 (Jun/Jul 1960) 516-519.  Generalizes by taking any number of sailors, any number of divisions and allowing the  i‑th division to discard a variable amount  Vi  before taking away  1/n  of the rest, even allowing negative  Vi,  e.g. if the monkey is adding coconuts to the pile!  Sadly, his basic recurrence equations (1) and (2) are misprinted.  He solves this by use of difference calculus.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. c, pp. 188-190.  Take    thrice, leaving  8.  How to divide the remainder to make things equal?

Philip Haber.  Peter Pauper's Puzzles & Posers.  Peter Pauper Press, Mount Vernon, NY, 1963.  Prob. 125, pp. 34 & 57.  Basket of pears divided among four people.  First gets  ¼  of the total plus  ¼  of a pear.  Second gets    and  ⅓.  Third gets  ½  and  ½.  Fourth gets remainder, which is half of what the first got.

Harold H. Hart.  Grab a Pencil No. 3.  Hart Publishing, NY, 1971.  The horse trader, pp. 41 & 118.  (Pay  1,  then half, then  1  more)  thrice, leaving  1.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

The Arabs, the monkey and the dates, pp. 50-52.  Four people, Ending 1.  Gets  1021.

Baling out, pp. 52, 169 & 189.  (Lose ⅓ of a load of bales of hay and ⅓ of a bale more) four times, leaving an integral number and no broken bales.  Solution is to start with  80  bales.

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 144: The mango thieves, pp. 89 & 135.  Ending 1  with three boys, but no monkey -- the boys each eat a mango when they steal    and there is an extra mango when they divide in the morning.

Michael Holt.  Figure It Out -- Book One.  Dragon (Granada), London, 1978.  Problem 14 (no page number) gives a Russian version involving a man who sells his soul to the Devil.  (Double and spend  8)  thrice to leave  0.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 27: Shipwreck, pp. 21 & 80.  Four sailors, the ship's cat and a box of biscuits, but only three of the sailors take a fourth, with one for the cat, before the final division into four, with none for the cat.  Answer is  61,  but this should be taken  (mod 256).

Dan Pedoe, Timothy Shima & Gali Salvatore.  Of coconuts and integrity.  CM 4:7 (Aug/Sep 1978) 182‑185.  General discussion of history and examples, based on Gardner (1958).  They generalise to consider giving  c  to the monkey and to having  m  divisions.  They first do Birkhoff & Mac Lane's version and  Ending 1,  effectively noting that the latter is the same as the former but with an extra division step.  I.e., the Birkhoff & Mac Lane problem has  m = n,  while  Ending 1  has  m = n+1.  Examination of the results shows that the Birkhoff & Mac Lane problem with odd  n  has the same answer as the  Ending 0  problem, but for even  n,  it corresponds to an  Ending 2  problem, which turns out to be one greater than the  Form, Ending = 0, 1 problem.  They separately solve the  Ending 0  problem, again with general  c.

Ben Hamilton.  Brainteasers and Mindbenders.  (1979);  Prentice‑Hall, Englewood Cliffs, NJ, 1981.  Problem for March 29, pp. 36 & 156.  i‑th customer buys  i+1  plus  1/(i+1)  of the rest.  How many customers can be served?

Scot Morris.  The Next Book of Omni Games.  Plume (New American Library), NY, 1988.  The monkey and the coconuts, pp. 30-31 & 182-183.  Sketches usual history.  He then notes that the usual process has the pile  º 1 (mod n),  so the monkey essentially gets one, then the pile is divided into  n  parts.  But one could alternatively have the pile  º 0 (mod n),  so the pile is divided into  n  parts, the sailor takes his part and then the monkey takes  1  from the remainder.  I.e. rather than removing one and  1/n  of the rest, each step removes  1/n  plus one more.  I have now termed these  Form 1  and  Form 0.  In this situation he doesn't allow the monkey to get one in the final division, i.e. he considers  Ending 0,  but  Ending 1  could be permitted, as studied by me below.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  It really takes the biscuit, pp. 50-52 & 119.  Four boys and their cat dividing biscuits.  Form 1, Ending 1  with  4  boys.

David Singmaster.  Coconuts.  The history and solutions of a classic diophantine problem.  Technical Report SBU-CISM-93-02, School of Computing, Information Systems and Mathematics, South Bank Univ., Dec 1993, 21pp, Feb 1994.  Submitted to Mathematics Review (Univ. of Warwick), 1993, but the journal closed before using it.  Revised in 1996 and again as the 3rd ed. of 11 Sep 1997, 21pp.  Extensive history, based on the material in this section.  Following Morris's comment, I consider both division Forms and both Endings, giving four basic problems rather than the three that he mentions.  In the 3rd ed., I changed the terminology to Form and Ending and have converted the material in this section to conform with this.  A little reflection shows that the solution for  Form 0, Ending 0  is one less than for  Form 1, Ending 1.  Actual calculation shows that one of the four cases has the same sequence of pile sizes as another, but shifted by one stage.  When  n  is odd,  Form 0, Ending 0  is the same as  Form 0, Ending 1,  but starting one stage earlier.  When  n  is even,  Form 1, Ending 1  is the same as  Form 1, Ending 0,  but starting one stage earlier.  Although these results are easily seen from the algebraic expressions, I cannot see any intuitive reason for these last equalities.

                    Having now seen Pedoe, Shima & Salvatore, I have added two supplementary pages discussing the Birkhoff & Mac Lane problem and relating it to the standard versions.

 

          7.E.1. VERSIONS WITH ALL GETTING THE SAME

 

          The  i‑th child gets some linear function of  i  applied to the remainder, but all wind up with the same amount.

          See Tropfke 586.

 

Fibonacci.  1202.

P. 279 (S: 399).  i-th gets  i  +  1/7  of rest.  (Sanford 219 gives the English;  H&S 61‑62 gives Latin & English.)

P. 279 (S: 399).  i-th gets  i  +  2/11  of the rest, but he doesn't ask for the number of children.

Pp. 279‑280 (S: 399-401).  i-th gets  (3i‑1)  +  6/31  of the rest.

Pp. 280‑281 (S: 401).  i-th gets  (2i+1)  +  5/19  of the rest.

Maximus Planudes.  Ψηφηφoρια κατ' Ivδoυσ η Λεγoμεvη Μεγαλη [Psephephoria kat' Indous e Legomene Megale (Arithmetic after the Indian method)].  c1300.  (Greek ed. by Gerhardt, Das Rechenbuch des Maximus Planudes, Halle, 1865, ??NYS [Allard, below, pp. 20-22, says this is not very good].  German trans. by H. Waeschke, Halle, 1878, ??NYS [See HGM II 549;  not mentioned by Allard].)  Greek ed., with French translation by A. Allard; Maxime Planude -- Le Grand Calcul selon des Indiens; Travaux de la Faculté de Philosophie et Lettres de l'Univ. Cath. de Louvain -- XXVII, Louvain‑la‑Neuve, 1981. 

                    i-th gets  i  +  1/7  of the rest, pp. 191‑194 & 233‑234.  On pp. 233‑234, Allard discusses the history of the problem, citing:  Fibonacci;  BR;  Parisinus supp. gr. 387  &  Scoriolensis Φ I 16.

BR.  c1305.  No. 84, pp. 102‑105.  i-th gets  i  +  1/7  of the rest.

Gherardi.  Libro di ragioni.  1328.  Pp. 37‑38.  i-th daughter gets  i  +  1/10  of the rest.

Lucca 1754.  c1330.  F. 82v, p. 199.  i‑th gets  i  +  1/10  of rest.  (This problem is not clearly expressed.)

Bartoli.  Memoriale.  c1420.  Prob. 9, f. 75v (= Sesiano pp. 138 & 147-148).  i-th gets  i  +  1/7  of the rest.

Pseudo-dell'Abbaco.  c1440. 

Prob. 168, p. 140.  i-th gets  1000 i  +  1/10  of the rest. 

Prob. 169, pp. 140‑141.  i-th gets  1/6  plus  10 i.

AR.  c1450.  No. 114, 115, 352, pp. 64‑65, 154, 173‑174 & 220.

114:  i-th gets  i  +  1/10  of rest.

115:  i-th gets  i  +  1/6  of rest.

352:  i-th gets  1/5  of remainder  +  i ‑ 1.

Muscarello.  1478.  F. 85v, pp. 204-205.  i-th gets  i  +  1/9  of the rest.

Chuquet.  1484.  Probs. 129‑141.  English of prob. 129 in FHM 224-225, with some description of the others.  i-th child gets  (ai + b)  plus  r  of the rest;  i-th child gets  r  of amount and  a + bi  more.  Many problems have non‑integral number of children and amounts received -- e.g. prob. 133 has  2 5/6  children receiving  6 2/3,  with the  5/6  getting  5 5/9.

HB.XI.22.  1488.   Pp. 44‑45 (= Rath 247).  i-th gets  i  +  1/9  of rest.

Calandri.  Arimethrica.  1491.  F. 65r.  i-th gets  1/10  +  1000 i

Calandri, Raccolta.  c1495.  Prob. 26, pp. 25‑26.  i-th gets  1000 i  +  1/10  of the rest.

Ghaligai.  Practica D'Arithmetica.  1521. 

Prob. 24, ff. 65v-66r.  i-th gets  1000 i  +  1/7  of rest.

Prob. 25, f. 66r.  i-th gets  1/7  +  1000 i.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 65, f. FF.ii.r (p. 155).  i-th child gets  1/7  of remainder  +  100 i

Tartaglia.  Quesiti, et Inventioni Diverse.  Venice, 1546.  Book 9, quest. 2, pp. 98r‑98v.  i-th child gets  i  +  1/8  of the rest.

Tartaglia.  General Trattato, 1556, art. 46, pp 245v‑246r.  i-th child gets  i  +  1/6  of the rest;  discusses same with  1/7  and  1/13.

Buteo.  Logistica.  1559.  Prob. 78, pp. 286-288.  i-th child gets  1/6  +  100 i.

Bachet.  Problemes.  1612.  Addl. prob. VII: Un homme venant à mourir ..., 1612: 149-154;  1624: 221-226;  1884: 158‑161.  i-th child gets  i  +  1/7  of the rest;  also  ai  +  1/n  and  1/n  +  ai.  Asserts some cases are impossible, contrary to Chuquet's approach.  Labosne has much revised the entire problem.

Ozanam.  1725.  Prob. 10, question 9, 1725: 67‑68.  Prob. 1, 1778: 185;  1803: 182-183;  1814: 159;  1840: 82.  i-th gets  10000 i  +  1/7  of the remainder.

Les Amusemens.  1749. 

Prob. 55, pp. 187-188.  i-th gets  1000 i  +  1/7  of the rest.

Prob. 177, p. 328.  i-th gets  1000 i  +  1/5  of the rest.

Euler.  Vollständige Anleitung zur Algebra.  (St. Petersburg, 1770) Part 2, sect. 1, chap. 3, art. 42.  (= Opera Omnia (1) 1 (1911), pp. 226‑228.  = Algebra; 1770; I.IV.III.604: Question 21, pp. 202‑203.)  i-th gets  100 i  +  1/10  of rest.

Hutton.  A Course of Mathematics.  1798?  Prob. 10,  1833: 214-217;  1857: 218-221.  Father with three sons leaves  ai + 1/n  of the remainder to the  i-th and this exhausts the fortune (but they do not get equal amounts except when  n = 4).  Finds algebraic expressions for the total and each portion, e.g. the total fortune is 

          (6n2 - 4n + 1)a/(n - 1)2.

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Art. 114, pp. 29-30.  First gets  2  + 1/6  of rest;  second gets  3  + 1/6  of rest;  they find they got the same amount.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 48, pp. 66-71 is the same as Hutton's 1798? problem.  Art. 49, pp. 71-73 then treats the usual problem of the same form, finding  n-1  children and a fortune of  a(n-1)2.

Fred Barlow.  Mental Prodigies.  Hutchinson, London, (1951), corrected reprint, 1952.  On pp. 38-41, he describes Henri Mondeaux (1826-1862), an illiterate who developed powers of mental calculation, and was then taught by a schoolmaster, M. Jacoby.  At some time, he was asked to solve the problem of people taking  100i  + 1/13  of the rest  and he found the answer in a few seconds by taking  12 x 100  as the amount of each person and  12 x 12 x 100  as the total amount.  I wonder if he knew this type of problem beforehand??

Vinot.  1860.  Art. XXXIX: Du testament, pp. 38-39.  i-th gets  1000 i  +  1/5  of the rest.

Hoffmann.  1893.  Chap. IV, no. 76: Another eccentric testator, pp. 166 & 221‑222 = Hoffmann-Hordern, p. 148.  First son gets  1/6  plus  240,  second son gets  1/5  of the remainder plus  288,  ...,  fifth gets  1/2  of the remainder plus  720  and all wind up with the same amount.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XL: Le Testament du nabab, pp. 144-147.  i‑th gets  i  +  1/7  of the rest.  Gives general solution of:  i-th gets  i  +  1/n  of the rest, using a rectangular layout of markers similar to ancient Indian multiplication tables and Lucas thinks the ancient Indians must have known this problem and solution.

 

          7.F.    ILLEGAL OPERATIONS GIVING CORRECT RESULT

 

          'Two digit' refers to illegal cancellation with two digit numbers, e.g. 16/64 = 1/4, etc.

 

A. Witting.  Ernst und Scherz im Gebiete der Zahlen.  Zeitschr. math. u. naturw. Unterricht 41 (1910) 45‑50.  P. 49 gives the rule of Ahrens, below, pp. 75‑76, for the case  k = 2.  He also gives three of the two digit examples:  26/65,  16/64,  19/95  --  omitting  49/98.

Ahrens.  A&N, 1918.

Pp. 73‑74 finds the two digit solutions and some with more digits.

Pp. 75‑76 studies  (a + m/n)1/k  =  a * (m/n)1/k  and finds that  m = a,  n = ak ‑ 1  works.

W. Lietzmann.  Lustiges und Merkwürdiges von Zahlen und Formen.  1922.

                    2nd ed., F. Hirt, Breslau, 1923.  Pp. 103‑104.  Gives  26/65  and  16/64  and the general rule of Ahrens, pp. 75‑76, citing Witting and Ahrens. 

                    4th ed, same publisher, 1930, p. 153, says there are more two digit examples and gives also  266/665,  etc.

                    The material is also in the 6th ed. (1943), but not in the 7th ed. (1950).

R. K. Morley, proposer;  Pincus Schub, solver.  Problem E24.  AMM 40 (1933) 111  &  425‑426.  2 digit versions.

G. [presumably the editor, Jekuthiel Ginsburg].  Curiosa 31 -- Another illegal operation.  SM 5 (1939) 176.  Cites Morley.  Refers to E. Nannei presenting several larger examples in 1935, some involving cancellation of several digits.

William R. Ransom.  Op. cit. in 6.M.  1955.  Freak cancellations, pp. 100‑102.  Finds the 2‑digit versions and give examples of several 3‑digit forms:  138/345 = 18/45,  163/326 = 1/2,  201/603 = 21/63.

B. L. Schwartz, proposer;  C. W. Trigg, solver.  Problem 434 -- Illegal cancellation.  MM 34 (1961) ??NYS  &  367‑368.  3 digit versions.

Anon.  Curiosa 122 -- A common illegal operation.  SM 12 (1946) 111.  (a2 ‑ b2)/(a ‑ b)  =  a + b.

Alan Wayne, proposer;  solution ??NYS.  Problem 3568.  SSM 75:2 (No. 660) (Feb 1975) 204.  (3/2)2 - (1/2)2  gives the right answer when the exponents are interpreted as multipliers.

Ben Hamilton.  Op. cit. in 7.E, 1979.  Problem for April 9, pp. 40 & 157‑158.  249/498  gives  24/48  correctly, which gives  2/8  wrongly.

R. P. Boas.  Anomalous cancellation.  In:  R. Honsberger, ed.; Mathematical Plums; MAA, 1979.  Chap. 6, pp. 113‑129.  Surveys the problem and studies the two digit case for other bases, e.g.  32/13 = 2/1  in base 4.  Cites the SM report, 1939.

R. P. Boas.  Generalizations of the  64/16  problem.  JRM 12 (1979‑80) 116‑118.  Summarises the above paper and poses problems.

 

          7.G.   INHERITANCE PROBLEMS

 

          7.G.1.          HALF + THIRD + NINTH, ETC.

 

          This is usually called 'The 17 camels', etc. 

          Early versions of this problem simply divided the amount proportionally to the given numbers, regardless of whether the numbers added to one or not.  I mention a few early examples of this below.  By about the 15C, people began objecting to such proportions, though Tartaglia (qv) and others see no difficulty with the older idea.  Sanford 218‑219 says Tartaglia was among the first to suggest the 18th camel -- but I have not found this in Tartaglia so far.  H&S 87 says that the use of the 18th camel is really a modern problem.  I haven't found it occurring until late 19C, when several authors claim it is centuries old and comes from the Arabic world or India or China!  See 1872 below.  Not everyone is happy with the problem, even to this day -- see Ashley, 1997 -- and many strange explanations have been given.

 

 

1/2, 1/3                                       Pseudo-dell'Abbaco;  Tagliente;  Buteo

1/3, 1/4                                       Jackson

  9,    8,    7                                 Papyrus of Akhmim

5/6, 7/12, 9/20                             Recorde

4/5, 3/4, 2/3                                Tartaglia 44

2/3, 1/6, 1/8                                Eperson

2/3, 1/2, 1/4                                Chuquet;  Apianus; 

1/2, 1/3, 1/4                                Bakhshali MS;  Lucca 1754;  AR 204;  Calandri;  Tagliente;  Riese;  Recorde;  Tartaglia 42;  Buteo;  Ozanam;  Les Amusemens;  Decremps;  Bullen;  Collins;  Always;

1/2, 1/3, 1/6                                AR, 286

1/2, 1/3, 1/9                                AR, 170, 286;  Hanky Panky;  Cassell's;  Proctor;  Cole;  Lemon;  Hoffmann;  Brandreth Puzzle Book;  Loyd;  H. D. Northrop;  Benson;  White;  Ball‑FitzPatrick;  Dudeney;  Kraitchik;  McKay;  Sullivan;  Doubleday - 1;  Ashley;

1/2, 1/4, 1/5                                Lemon;  Clark;  Ernst;  King;  Foulsham

1/2, 1/4, 1/6                                Clark;

1/2, 1/4, 1/8                                Bath

2/5, 1/3, 1/4                                AR, 202;  Wagner

1/3, 1/4, 1/5                                BR;  Riese;  Tartaglia 43;  Jackson

1/4, 1/5, 1/6                                Apianus;  W. Leybourn

2/3, 1/2, 1/3, 1/4                          Papyrus Rhind;  Pike;  D. Adams, 1835

1/2, 1/3, 1/4, 1/5                          Chaturveda;  Blasius

1/2, 1/3, 1/4, 1/6                          Mahavira

1/2, 1/3, 1/6, 1/19                        Parlour Pastime

1/3, 1/4, 1/5, 1/6                          Walton;  Simpson;  Dodson;  J. King

1/3, 1/4, 1/6, 1/8                          D. Adams, 1801; 

  6,    5,     4,    3,    2                  Mahavira

7/2, 5/2, 15/4, 25/4, 4                  Papyrus of Akhmim

1/2, 1/3, 1/4, 1/5, 1/6                   Tonstall

1/3, 1/4, 1/5, 1/6, 1/7                   Walkingame;  Vyse

1/3, 1/4, 1/6, 1/8, 1/9                   Meyer;  Haldeman-Julius;  Leeming

 

Papyrus Rhind, c‑1650, loc. cit. in 7.C.  Problem 63, p. 101 of vol. 1 (1927) (= p. 53 (1978)).  Divide  700  loaves in proportion  ⅔ : ½ : ⅓ : ¼.

Papyrus of Akhmim.  c7C.  Jules Baillet, ed.  Le Papyrus Mathématique d'Akhmîm.  Mémoires publiés par les membres de la Mission Archéologique Français au Caire, vol. IX, part 1, (1892) 1‑89.  Brief discussion of this type of problem on p. 56.  Probs.  3, 4, 10, 11, 28?, 47, 48, 49  are of this type.  I describe two examples.

Prob. 3, pp. 64-65.  Divide  1000  in proportion  3 + 1/2  :  2 + 1/2  :  3 + 1/2 + 1/4  :  6 + 1/4  :  4.

Prob. 11, pp. 68-69.  Divide  3 + 1/2 + 1/4  in proportion  7 : 8 : 9.

Bakhshali MS.  c7C.  See in 7.E, where a king gives away  ½ + ⅓ + ¼  of his money!

Mahavira.  850.  Chap. VI, v. 80, 86, pp. 110-111.  Divide  120  in proportion  1/2 : 1/3 : 1/4 : 1/6.  Divide  480  in proportion  2 : 3 : 4 : 5 : 6.

BR.  c1305.  No. 71, pp. 94‑95.  1/3 + 1/4 + 1/5.

Lucca 1754.  c1330.  F. 61r, p. 140.  Divide into  ½ + ⅓ + ¼.  He divides in proportion  6 : 4 : 3.

Pseudo-dell'Abbaco.  c1440.  Prob. 122, pp. 97‑98.  Divide into  ½ + ⅓.  He divides in the ratio  3 : 2.

AR.  c1450.  Probs. 170, 202-204, 207, 229-230, 286.  Pp. 81‑82, 94‑96, 106-107, 130, 160‑161, 166‑167, 211‑213.

170:  1/2 + 1/3 + 1/9.

202:  1/3 + 1/4 + 2/5.

203:  Divide 384 into  2/3 and 6 more,  3/5 and 8 more,  5/6 and 10 more,  7/8 and 6 more.  He takes a common denominator of  360  and finds  2/3  of it is  240  and then adds  6  to get  246.  Similarly, he gets  224,  310,  321  and then divides in the proportion  246 : 224 : 310 : 321.  This is actually indeterminate as it depends on the choice of common denominator.  Vogel says the problem is unclear and the solution is false and notes that dividing  387  instead of  384  would give an integral solution.  He cites a number of other occurrences of this problem -- cf. Widman below.

204:  ½ + ⅓ + ¼.

207:  Divide 100 into  (1/3 ‑ 1/4)  +  (1/4 ‑ 1/5)  +  (1/5 ‑ 1/6).

229:  Divide 20 into  1½, 2½, 1, 1, 1, 1.  Does as  3 : 5 : 2 : 2 : 2 : 2.

230:  Divide 20 into  1½ + ⅓, 2½ + ¼, 1, 1, 1, 1.  Does as  22 : 33 : 12 : 12 : 12 : 12.

286 discusses problems where one removes fractions and deals with the remainder.  Notes that  1/2 + 1/3 + 1/6  leaves nothing, but  1/2 + 1/3 + 1/9  leaves something.

Chuquet.  1484.  Triparty, part 1.  English in FHM 75.  "I wish to divide 100 into three parts of such proportion as are  1/2,  2/3,  1/4  ..."

Ulrich Wagner.  Untitled text known as "Das Bamberger Rechenbuch".  Heinrich Petzensteiner, Babenberg (= Bamburg), 1483.  Reproduced, with transcription and notes by Eberhard Schröder as:  Das Bamberger Rechenbuch von 1483.  Akademie‑Verlag, Berlin, DDR, 1988. 

Pp. 69‑70 & 200.  = AR, no.  202.

Pp. 71-72 & 201-202.  = AR, no. 203.

Calandri.  Aritmetica.  c1485.  Ff. 93v-94r, pp. 187‑188.  Same as Lucca 1754.

Johann Widman.  Behēde und hubsche Rechnung auff allen kauffmanschafft.  Conrad Kacheloffen, Leipzig, 1489.  ??NYS.  (Rara 36‑40.  This is extensively described by:  J. W. L. Glaisher in Messenger of Mathematics 51 (1921‑22) 1‑148, but he gives the title as:  Behēde und hubsche Rechenung ....)  Smith and Glaisher give  Widman,  but Knobloch (7.L.2.c) uses  Widmann  and  Behende und hubsche Rechenung ....

F. 195v (Glaisher 14-15 & 122).  = AR, no. 230.

Ff. 195v-196 (Glaisher 15 & 122).  = AR, no. 207.

F. 196v (Glaisher 18-19, 38, 45, 122, 130).  = AR, no. 203.

           Glaisher notes that Pacioli's Summa, (see below), gives a more natural determinate interpretation for similar problems.  In this example, this would first subtract  6 + 8 + 10 + 6  from  384,  leaving  354  which would be divided in the proportion  2/3 : 3/5 : 5/6 : 7/8.  He also notes the appearance of Widman's problem and solution in Huswirt (1501) ??NYS  and of problems similar to Widman and done in the same way, in Arithmetice Lilium (a book of c1510, ??NYS) (divide  100  into  1/2 + 3,  1/3 + 2,  1/5 + 4)  and Tonstall.  Rudolff's Kunstliche Rechnung of 1526, ??NYS, does (divide into 1/2 and 6,  1/3 and 4,  1/4 less 2) in Pacioli's manner.  Cf Apianus for a similar version.  Riese's Rechenung nach der Lenge (1550?, ??NYR) does (divide  124½  into  2/3 less 12,  1/4 and 10,  5/6 less 24,  3/8 and 6,  2/5 less 7) in Pacioli's way.

Pacioli.  Summa.  1494.  These give the more natural interpretation of this type of problem.

F. 150r, prob. 3.  Divide 100 as  1/2 plus 5;  1/3 less 4.  Subtracts 1 from 100 and divides the resulting 99 in the proportion  3 : 2. 

Ff. 150r-150v, prob. 4.  Divide 100 as  1/2 plus 3;  1/3 less 5.  Divides 102 as  3 : 2.

F. 150v, prob. 5.  Divide 100 as  1/2 less 4;  1/3 less 2.  Divides 106 as  3 : 2.

F. 150v, prob. 6.  Divide 30 as  1/2 plus 2;  1/3 plus 3.  Divides 25 as  3 : 2.

F. 150v, prob. 7.  Divide 10 as  1/2 less 3;  1/3 plus 4.  Divides 9 as  3 : 2.

F. 150v, prob. 8.  Divide 1046 as  1/2 less 2;  1/3 less 1;  1/4 plus 5.  No working or answer.

Blasius.  1513.  F. F.ii.r: Prime regula.  Man leaves  6000  to be divided  1/2 + 1/3 + 1/4 + 1/5.  There is an error in the calculation.

Tagliente.  Libro de Abaco.  (1515).  1541.

Prob. 94, part 2, ff. 48v-49r.  Divide  120  into  ½ + ⅓.

Prob. 95, ff. 48v-49v.  Divide  12  into  ½ + ⅓ + ¼.

Riese.  Rechnung.  1522. 

1544 ed. -- pp. 81‑82;  1574 ed. -- pp. 55r‑55v.  1/3 + 1/4 + 1/5.

1544 ed. -- pp. 98‑99;  1574 ed. -- p. 66r.  Three men take  1/2 + 1/3 + 1/4  of the profits, making  50  all told.  What was the profit?  Answer:  50 x 12/13.

Tonstall.  De Arte Supputandi.  1522. 

Quest. 17, pp. 147-149.  Divide into  1/2 + 1/3 + 1/4 + 1/5 + 1/6.  Uses  4350  as common denominator.

Quest. 18, pp. 149-150.  Same as Quest. 17, done with denominator  60.

Quest. 21, pp. 151-152.  Divide into parts:  1/3 + 1/4;  1/4 + 1/5;  1/5 + 1/6.

Quest. 22, pp. 153-154.  Divide  600  into:  2/3 plus 9;  3/5 plus 8;  5/6 plus 7;  7/8 plus 6.  See:  AR;  Widman;  Pacioli for discussion of this type of problem.  Takes common denominator of  120  and then divides as  89 : 80 : 107 : 111  which is not the way I read the problem.  I would divide  570  as  80 : 72 : 100 : 105,  as done by Pacioli.

Apianus.  Kauffmanss Rechnung.  1527. 

F. H.v.r.  Divide  1300  as  1/2 plus 8,  1/3 less 5,  1/4 less 12.  Does in Pacioli's manner, dividing  1309  in the proportion  12 : 8 : 6  and then amending by  +8, -5, -12.

F. H.v.r.  Divide  58  as  1/2 + 2/3 + 1/4.  Cf Chuquet.

F. H.vii.r.  Divide  40  as  1/4 + 1/5 + 1/6.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 91, second part, f. HH.i.v (p. 166).  First has    plus  7;  second has  ¼  plus  13;  third has  ½  minus  28.  How much was there?

Recorde.  Second Part.  1552.  Pages from 1668 ed.

Pp. 284-289:  A question of building;  An impossible question;  The former question of building now possible.  Divide 3000 into:  1/2 plus 6;  1/3 plus 12;  2/3 less 8;  1/4 plus 20.  He first says it is impossible, then rephrases it and solves by Pacioli's method.

P. 293: Another question of a testament.  Divide  7851 into  ½ + ⅓ + ¼.  Usual solution.

P. 294: Another like question.  Divide  450  into  1/2 + 1/3,  1/3 + 1/4,  1/4 + 1/5.  Usual solution.

Tartaglia.  General Trattato.  1556.  Book 12, art. 42-44, pp. 200r-201r.

Art. 42.  ½ + ⅓ + ¼.  Long discussion of an error of Luca Pacioli and others who assert that such problems are impossible or illegal.  Tartaglia says simply to divide in the proportion  6 : 4 : 3  and can't understand why others are making such a fuss.

Art. 43.  1/3 + 1/4 + 1/5.

Art. 44.  4/5 + 3/4 + 2/3.

(Sanford 218‑219 says Tartaglia was among the first to suggest the 18th camel -- but I see nothing of this here.  H&S 87 says that this is really a modern problem in that previously property was divided in proportion to fractions, regardless of whether they summed to unity.  Tartaglia's discussion of Pacioli and others makes it clear that people were starting to object to this at this time, but examples continue and I don't see the modern version occurring until late 19C.)

Buteo.  Logistica.  1559. 

Prob. 5, pp. 203-204.  Divide 77 into  ½ + ⅓ + ¼.

Prob. 74, pp. 283-284.  Divide 30 into  ½ + ⅓.  Discusses the solution.

Prob. 75, pp. 284-285.  Divide 15 as  1/2 plus 2;  1/4 + 3.  He divides 10 into  ½ + ⅓,  as in Pacioli.  See:  AR;  Widman;  Pacioli  for discussion of this type of problem.

Prob. 76, pp. 285-286.  Divide 24 as  1/3 less 7;  1/4 less 4.

Prob. 77, p. 285.  Divide 12 as  2/3 less 3;  1/6 plus 4.

Prob. 22, pp. 350-351.  Divide 60 as  1/4;  1/3 plus 4;  3/4 less 8.

Prob. 26, pp. 353-354.  Divide 30 as  1/2 plus 2;  1/3 less 3.

Prob. 28, p. 355.  Divide 224 as  1;  6/5 plus 4.

Izaak Walton.  The Compleat Angler.  (R. Marriott, London, 1653);  Everyman edition, Dent, London, 1906, et seq.  Chap. V -- The Fourth Day, pp. 101‑102.  The World's Classics, OUP, 1935, Chap. V, pp. 114-116.  Divide 20 into  1/3 + 1/4 + 1/5 + 1/6.  Leaves one left over.

W. Leybourn.  Pleasure with Profit.  1694.  Prob. 11, pp. 37-38.  £6000 divided  1/4 + 1/5 + 1/6.  He divides in the proportion  15 : 12 : 10.  Cf Apianus.

Ozanam.  1725.  Prob. 24, question 5, 1725: 179.  Divide  26000  into  ½ + ⅓ + ¼.  Takes in proportion  12 : 8 : 6.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XIV, pp. 85-86 (1790: prob. XXIV, pp. 86-87).  Divide 20 into  1/3 + 1/4 + 1/5 + 1/6,  done by proportion.

Les Amusemens.  1749.  Prob. 52, p. 184.  Divide 78 into  ½ + ⅓ + ¼,  by using proportions.

Walkingame.  Tutor's Assistant.  1751.  1777: p. 177, prob. 119;  1835: p. 180, prob. 58;  not in 1860.  Divide £500 into  1/3 + 1/4 + 1/5 + 1/6 + 1/7.  Gives exact answer as integer plus a fraction.  1835 reduces the fractions to lowest terms.

Vyse.  Tutor's Guide.  1771?  1793: p. 105; 1799: Prob. 6, p. 113 & key p. 151.  Same as Walkingame.  Solution gives answers rounded to farthings and never gives the exact fractions.

Dodson.  Math. Repository.  1775.  P. 32, quest. LXXVIII.  Divide 20s in proportion: 1/3, 1/4, 1/5, 1/6.

Pike.  Arithmetic.  1788. 

P. 335, no. 4.  "Being a little dipped, they agreed that A should pay  2/3,  B  1/2,  C  1/3, and  D  1/4."  = D. Adams, 1835.  Cf D. Adams, 1801.

P. 355, no. 39.  A, B, C do a job.  A and B do  3/11  of it,  A and C do  5/13,  B and C do  4/14.  (Also entered at 7.H.)

Henri Decremps.  Codicile de Jérôme Sharp, ....  Op. cit. in 4.A.1.  1788.  Avant-propos, pp. 18-19 mentions  ½ + ⅓ + ¼,  but there is no solution.

Samuel Bullen.  A New Compendium of Arithmetic ...  Printed for the author, London, 1789.  Chap. 38, prob. 3, p. 239.  Divide into  ½ + ⅓ + ¼,  phrased as  2A = 3B = 4C.

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  The 1795 has dividing  100  into  1/3 + 1/4 + 1/5 + 1/6.  Gives usual solution.

D. Adams.  Scholar's Arithmetic.  1801.  P. 206, no. 35.  Four men divide a purse of  $12  as  1/3 + 1/4 + 1/6 + 1/8.  Divides in the proportion:  8 : 6 : 4 : 3.  Cf his 1835 book.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions, no. 16, pp. 18 & 75.  Division in the proportion of  1/3, 1/4, and 1/5,  but last person dies.  Solution indicates this as standard practice.

D. Adams.  New Arithmetic.  1835.  P. 249, no. 132.  = Pike, no. 4.  Divides in the proportion:  8 : 6 : 4 : 3.  which is the same proportion as in his 1801 version.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 4, p. 173 (1868: 184).  Pay  20s  with only  19s  by dividing into  1/2 + 1/3 + 1/6 + 1/19.  "This, however, is only a payment upon paper."

Hanky Panky.  1872.  A Chinese puzzle, pp. 73-74.  17  elephants to be divided  1/2 + 1/3 + 1/9.

Cassell's.  1881.  P. 102: The clever lawyer.  = Manson, 1911, p. 255.  17  horses divided  1/2 + 1/3 + 1/9.  = Rohrbough; Puzzle Craft; 1932, p. 7. 

Richard A. Proctor.  Some puzzles.  Knowledge 9 (Aug 1886) 305-306.  "... the familiar puzzle [of] the farmer, ignorant of numbers, who left  17  horses to his three sons (or, equally well it may be, an Arab sheik who left  17  camels)".  Points out that if there were  35  camels, then the Cadi could also be left a camel.

E. W. Cole.  Cole's Fun Doctor.  The Funniest Book in the World.  Routledge, London  &  E. W. Cole, Melbourne,  nd [HPL gives 1886 and lists the author as  Arthur C. Cole].  P. 224: A Chinese puzzle.  17 elephants left by a Chinaman to be divided  1/2 + 1/3 + 1/9.  Says it is in the Galaxy for August, which might have been an Australian publication by Proctor, who had connections there.

Lemon.  1890.  The legacy, no. 652, pp. 81 & 121.  19  camels divided  1/2 + 1/4 + 1/5.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 135, no. 2.  17  horses:  1/2 + 1/3 + 1/9.  Dervish loans them his horse.

Hoffmann.  1893.  Chap. IV, no. 11: An unmanageable legacy, pp. 147 & 191‑192 = Hoffmann-Hordern, p. 119.  1/2 + 1/3 + 1/9.  Answer says "this expedient is frequently employed" in "the Mahomedan Law of Inheritance".

Mittenzwey.  1895?.  Prob. 164, pp. 34 & 82;  1917: 164, pp. 31-32 & 80.  17 camels,  1/2 + 1/3 + 1/9,  dervish loans them his camel.

Brandreth Puzzle Book.  Brandreth's Pills (The Porous Plaster Co., NY), nd [1895].  P. 3: An unmanageable legacy, with nice colour picture.  Identical to Hoffmann.  No solution.

Loyd.  Problem 37: A queer legacy.  Tit‑Bits 32 (5 Jun  &  3 Jul 1897) 173  &  258.  = Cyclopedia, 1914, The herd of camels, pp. 57 & 346.  17 horses in proportion  1/2 : 1/3 : 1/9.  Says the use of proportion makes the solution actually correct.

Clark.  Mental Nuts.  1897, no. 25;  1904, no. 5;  1916, no. 18.  The heirs and the sheep.  1897 has  1/2 + 1/4 + 1/6  of 15 sheep.  This seems to have been a miscopying of the question with 11 sheep.  He says to borrow a sheep and distribute  8, 4, 3,  returning one.  1904 & 1916 amend this to  1/2 + 1/4 + 1/5  of 19 sheep.

H. D. Northrop.  Popular Pastimes.  1901.  No. 18: The clever lawyer, pp. 69 & 74.  = Cassell's.

Benson.  1904.  The lawyer's puzzle, p. 225.  1/2 + 1/3 + 1/9.  There originally were 18 horses, but one died.

William F. White.  Op. cit. in 5.E.  1908.  Puzzle of the camels, p. 193.  17 camels divided  1/2 + 1/3 + 1/9.

Ball-FitzPatrick.  2nd ed., 1908‑1909.  Part 1, p. 111, footnote says the problem is Arabic.  The material is not in the 1st ed., nor in Ball, 5th ed.  A&N, pp. 84‑85, cites this but says it has been in German oral tradition for a long time.  He gives it with 17 horses.

E. Ernst.  Mathematische Unterhaltungen und Spielereien.  Vol. 2, Otto Maier, Ravensburg, 1912.  P. 15: Das geschente Weinfass.  Divide  19  in  1/2 + 1/4 + 1/5.

Dudeney.  MP.  1926.  Prob. 89: The seventeen horses, pp. 33-34 & 123-124.  = 536, prob. 172, pp. 54‑55 & 266‑267.  Discusses interpretation of proportion, as in Loyd, in detail.

King.  Best 100.  1927.  No. 21, pp. 14 & 43.  19 horses into  1/2 + 1/4 + 1/5.

Collins.  Book of Puzzles.  1927. 

The lady bookmaker's problem, pp. 72-73.  Because  1/2 + 1/3 + 1/4 = 13/12,  one can offer odds in a three horse race of:  even money,  2 to 1  and  3 to 1.

The sheik and his camels, pp. 77-78.  Usual form.  Cadi loans them his camel.

M. Kraitchik.  La Mathématique des Jeux, 1930, op. cit. in 4.A.2, chap. 1, prob. 47, p. 15.  17 sheep into  1/2 + 1/3 + 1/9.  Says it is of Hindu origin.

Anon.  Foulsham's New Fun Book.  W. Foulsham, London, nd [1930s?].  Pp. 85‑86: The farmer's horses.  Identical to King, 1927.

The Bile Beans Puzzle Book.  1933.  No. 26: A farmer's will.  19  horses divided 

          1/2 + 1/4 + 1/5.

McKay.  Party Night.  1940.  No. 35, p. 184.  "It is said that an Arab had 17 cattle." 

Sullivan.  Unusual.  1943.  Prob. 12: Will trouble.  17 horses into  1/2 + 1/3 + 1/9.

Jerome S. Meyer.  Fun for the Family.  (Greenberg Publishers, 1937);  Permabooks, NY, 1959.  No. 30: Think cow it is done, pp. 42-43 & 241.  Herd to be divided  1/3 + 1/4 + 1/6 + 1/8 + 1/9.  Neighbour loans two cows and everything divides up properly with two cows left over for the neighbour.  How many cows were there?

Haldeman-Julius.  1937.  No. 131: Cow problem, pp. 15 & 27.  Same as Meyer, asking what is wrong with the problem and answering that "The problem is coo-coo because all the fractions do not add up to unity."

Leeming.  1946.  Chap. 5, prob. 24: The herd of cattle, pp. 61 & 179‑180.  Same as Meyer.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 66: The vicar's garden, pp. 25 & 52.  7s divided  1/2 + 1/4 + 1/8  by adding an extra shilling.  Solution doesn't seem to understand this and claims there really should be  7/8 s  left over.

Doubleday - 1.  1969.  Prob. 25: Milk shake, pp. 36 & 159.  = Doubleday - 5, pp. 35-36.  17 cows divided  1/2 + 1/3 + 1/9.  He states the usual solution and then asks what is wrong with it.  His solution notes that the fractions add to  17/18  and then says 'So, in making his will, the farmer hadn't distributed his entire herd.'  This seems confused to me as the entire herd has been distributed to the sons.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 35: An odd bequest, pp. 22 & 71.  Divide 13 camels as  1/2 + 1/3 + 1/4.  Executor borrows one camel, so there are now  12  camels, then he gives the first son  6 = 12/2  and the second son  4 = 12/3,  leaving 2.  But the third son ought to have  3 = 12/4,  so the executor returns the borrowed camel and gives the third son his three!  The answer says this is an amusing way of arriving at the intended division in the proportion  6 : 4 : 3.  See Singmaster, 2000, for an extension.

D. B. Eperson.  Puzzles, pastimes and problems.  MiS 3:6 (Nov 1974) 12‑13 & 26‑27.  Prob. 6: The Shah's Rolls‑Royces.  Divide 23 Rolls‑Royces into  2/3 + 1/6 + 1/8.  The answer erroneously asserts this works for  n º ‑1 (mod 24).

David Singmaster.  A Middle Eastern muddle.  41  oil wells to be divided into  1/2 + 1/3 + 1/7.  But then I ask if there are values other than  2, 3, 7, 41  which produce such a puzzle problem.  There are  12  such quadruples.  I recall seeing this when I was a student but I haven't relocated it.  Appeared in my puzzle columns as follows.

                    "Well, well, well."  Brain Twister.  Weekend Telegraph (27 Feb 1988) xv (misprinted),  (5 Mar 1988) (corrected) xv  &  (12 Mar 1988) xv.

                    Reprinted, with no title, in:  The Daily Telegraph Braintwisters No. 1;  Pan Books, London, 1993;  with Barry R. Clarke, Rex Gooch and Angela Newing.  Prob. 25, pp. 27, 76 & 117.

                    "A Middle Eastern muddle."  The Puzzle Box.  Games & Puzzles 12 (Mar 1995) 18-19  &  13 (Apr 1995) 40.

John P. Ashley.  Arithmetickle.  Arithmetic Curiosities, Challenges, Games and Groaners for all Ages.  Keystone Agencies, Radnor, Ohio, 1997.  P. 8: Omar divides 17 horses among 3 sons.  The Answer says: "It was a great solution, but it was not correct mathematics.  The sum of the fractional parts: 1/2, 1/3 and 1/9 do not add up to 1 but to 17/18.  Therefore, each of the heirs got a bit more than the will intended."

David Singmaster.  The seventeen camels and the thirteen camels.  Draft paper written in Dec 2000.  This discusses the classic 17 camels problem and the 13 camels problem given by Always (1971) finding all solutions for two, three or four sons.

 

          7.G.2.          POSTHUMOUS TWINS, ETC.

 

          A man dies, leaving a pregnant wife and a will explaining how his estate is to be divided between the wife and a son or a daughter.  The wife produces twins, one boy and one girl.  How is the estate to be divided?

          The most common version has  son : wife  =  wife : daughter  =  2 : 1  given in the will and derives  son : wife : daughter = 4 : 2 : 1.  I will denote this as the "usual form".  Other proportions cited by Smith are:  4 : 3 : 2;   2 : 2 : 1;   9 : 6 : 4. 

          Alcuin  &  BR  proceed by dividing the estate in half first.

          See Brooks for some odd versions.

          See Tropfke 655.

 

Moritz Cantor.  Vorlesungen über Geschichte der Mathematik.  Vol. 1, 3rd ed., Teubner, Leipzig, 1907.  Pp. 561-563 sketches the history of this problem.  This is the basis of Smith's discussion below.  This asserts that the problem is based on Roman Lex Falcidia of  ‑40  which required that at least  ¼  of an estate should go to the legal heir.  He says it first appears in the works of Celsus and quotes Julianus.  He also cites Caecilius Africanus (c100) and Julius Paulus (3C).  Describes the common case and says that if the will was invalidated, then only the children would inherit.

D. E. Smith.  Op. cit. in 3.  Based on Cantor, cites Lex Falcidia, Celsus, Julianus, Africanus.  He cites 22 medieval references, including Vander Schuere and Recorde of those below.

See also:  Sanford 218‑219;  H&S 87‑88;  A&N 24‑26.

Juventius Celsus.  De istituzione uxoris et postumi et postumae.  c75.  ??NYS -- cited by Smith.  Julianus cites him.

Salvianus Julianus.  c140.  Lex 13 principio.  Digestorum lib. 28, title 2.  ??NYS -- quoted by Cantor and cited by Buteo & Smith.  Cantor says this reports a case, though the quoted text isn't very specific.  Usual form.  Julianus cites Celsus.

Caecilius Africanus.  c150.  Lex 47 §1.  Digestorum lib 28, title 5.  ??NYS -- cited by Cantor & Smith.  Cantor says this refers to a case.

Julius Paulus.  3C.  Lex 81 principio.  Digestorum lib. 28, title 5.  ??NYS -- cited by Cantor & Smith.  Cantor says this refers to a case.

Alcuin.  9C.  Prob. 35: Propositio de obitu cujusdam patrisfamilias.  Problem of posthumous twins.  Ratios are  3 : 1  for  son : mother  and  5 : 7  for  mother : daughter.  He takes half the estate and shares it  3 : 1  and then the other half is shared   5 : 7.  This gives  9 : 8 : 7.  Ahrens, A&N, p. 26, suggests  15 : 5 : 7,  which is the result of the usual Roman process.

BR.  c1305.  No. 91, pp. 110‑111.  Son : wife  =  wife : daughter  =  3 : 2.  Divides in halves, as in Alcuin, and divides each half as  3 : 2,  giving  son : wife : daughter = 3 : 5 : 2.

Gherardi?.  Liber habaci.  c1310.  P. 145.  Usual form.

Gherardi.  Libro di ragioni.  1328.  P. 37.  Son : wife  =  3 : 1;  wife : daughter  =  2 : 1.  Divides as  6 : 2 : 1.

Lucca 1754.  c1330. 

F. 60r, pp. 136‑137.  Posthumous triplets, 2 boys and a girl with usual ratios.  He divides in proportion  4 : 4 : 2 : 1  for  boy : boy : mother: girl.

F. 83r, pp. 200‑201.  Posthumous twins.  Usual form.

Pseudo-dell'Abbaco.  c1440.  Prob. 100, p. 85 with plate on p. 86.  Posthumous twins.  Usual form.  I have a colour slide of this.

AR.  c1450.  Prob. 209, pp. 97, 176, 223.  Man has son, wife and two daughters and gives the usual ratios, hence divides in the proportion  4 : 2 : 1 : 1.

Muscarello.  1478.  Ff. 75r-76r, pp. 189-191.  Posthumous twins.  Usual form.

Wagner.  Op. cit. in 7.G.1.  1483.  Pp. 73‑75 & 202‑203.  Usual will, but wife produces a son and two daughters.  Divides as in AR.

Chuquet.  1484.  Prob. 205.  English & discussion in FHM 205.  Usual form.  FHM say it "goes back to the Roman emperor and legislator Justinian" and quotes Recorde.

HB.XI.22.  1488.  P. 44 (Rath 247).  Posthumous twins.

Pacioli.  Summa.  1494. 

F. 158r, prob. 80.  Usual posthumous twins.  Then says that Nofrio Dini of Florence, a respectable merchant in Pisa, at the shop of Giuliano Salviati, told him about such a will on 16 Dec 1486.  After a bequest to the church, there was an estate of 800 to be divided.  If a son was born, the mother was to get 400;  if a daughter, the mother was to get 300.  Twins were born and he says to divide as  3 : 3 : 5.  Says one can deal similarly with similar problems.  Of the Biographical Sources listed in Section 1, Taylor, p. 149 & Fennell, p. 11, mention this problem.

F. 158v, prob. 82.  Selling a pregnant cow which bears twins.  Gives some rules which determine the relative values.

Blasius.  1513.  F. F.ii.v: Quarta regula.  Dying man with pregnant wife.  If she has a son, he gets  3/5  and the mother and the church get  1/5  each.  If she has a daughter, the daughter and the mother get  2/5  each and the church gets  1/5.  She has a son and a daughter.  He divides in proportion  3 : 2 : 2 : 1,  but gives no reason.  Offhand, I would think that  1/5  should go to the church -- since this is specified in either case -- and then the remaining  4/5  should be divided in the proportion  3 : 1 : 1,  giving overall proportions of  12 : 4 : 4 : 5.  He says one can similarly deal with two sons or two daughters or two sons and one daughter. 

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 100, ff. 50v-51r.  Usual form.

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 23, f. 65v.  Usual posthumous twins.

Riese.  Rechnung.  1522.  1544 ed. -- p. 80;  1574 ed. -- p. 54v.  Father leaves a widow, a son and two daughters.  Divides as in AR.

Tonstall.  De Arte Supputandi.  1522.  Quest. 16, pp. 146-147.  Usual form.  Then considers 3 sons and 2 daughters!

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 87, ff. GG.vi.v - GG.vii.r (p. 164).  Posthumous twins.  Son : wife  =  4 : 1;  wife : daughter  =  2 : 1;  divides as  8 : 2 : 1

Giovanni Sfortunati.  Nuovo lume.  Venice, 1545.  F. 58v.  ??NYS -- described by Franci, op. cit. in 3.A, p. 38.  Franci's discussion is about several extended versions, but it seems to indicate that Sfortunati deals with a hermaphrodite.

Recorde.  Second Part.  1552.  Smith, op. cit. in 3.A, p. 69, quotes 1558, fol. X 8 (??NYS).  1668, pp. 289-293: A question of a Testament.

                    Man with fortune of 3600 and a pregnant wife makes a will and dies.  If she has a son, the son get ½ and she gets ⅓; if she has a daughter, she gets ½ and the daughter gets ⅓.  "If some cunning lawyers had this matter in scanning, they would determine this testament to be quite voyde, and so the man to die intestate, because the testament was made unsufficient."  (The 1668 has identical wording, except it uses 'void' and 'insufficient'.)  He divides in the proportion  9 : 6 : 4.

Tartaglia.  General Trattato.  1556.  Book 12, art. 35-41, pp. 199v-200v.

Art. 35-40 give various ratios  S : M  =  son : mother  and  M : D  =  mother : daughter, then divides in the usual way to get the proportions  s : m : d  such that  s/m = S/M  and  m/d = M/D.  E.g. when  S : M = 2 : 1 = M : D,  then  s : m : d = 4 : 2 : 1.

Art. 35.  S : M = 2 : 1;   M : D = 2 : 1.

Art. 36.  S : M = 5 : 3;   M : D = 1 : 1.

Art. 37.  S : M = 2 : 1;   M : D = 1 : 1.

Art. 38.  S : M = 2 : 1;   M : D = 2 : 1,  in a different context than Art. 35.

Art. 39.  S : M = 2 : 1;   M : D = 3 : 1.

Art. 40.  S : M = 3 : 1;   M : D = 2 : 1.

Art. 41.  S : M = 2 : 1;   M : D = 2 : 1,  but quadruplets are produced -- two sons and two daughters.  He divides in proportion  4 : 4 : 2 : 1 : 1.

Buteo.  Logistica.  1559. 

Prob. 60, pp. 264-266.  Usual form.  Cites Julianus.

Prob. 12, pp. 341-342.  Selling a pregnant cow, where the value depends on the sex of the calf.  Cow + daughter is worth  40,  while  cow + son is worth 45.  This is insufficient to determine the relative values, but he then adds excessive information:  C = 3D = 2S.  The cow produces twins -- one son and one daughter.

Gori.  Libro di arimetricha.  1571.

Ff. 75r‑75v (pp. 83-84).  Usual form.

F. 75v (p. 84).  Posthumous quintuplets -- divides in same proportions, though there is some confusion in the text of the solution.

Jacob Vander Schuere.  Arithmetica, oft Reken‑const.  G. Kooman, Haarlem, 1600.  ??NYS.  [Smith, Rara, 421‑423.]  F. 98 is quoted in Smith, op. cit. in 3, p. 69, note 7.  Posthumous triplets:  boy, girl and hermaphrodite.  Divides in proportion  12 : 4 : 2 : 7 = son : wife : daughter : hermaphrodite.  Smith doesn't give the original ratios, but they were probably  son : wife = 3 : 1,  wife : daughter = 2 : 1.

Schott.  1674. 

Ænigma VII, pp. 559-560.  Usual form.

Ænigma X, p. 560.  Son : wife = wife : daughter = 2 : 1,  but he interprets this as the son getting  2/3  of the estate, the wife getting  2/3  of the rest with residue going to the daughter, leading to  son : wife : daughter = 6 : 2 : 1.

W. Leybourn.  Pleasure with Profit.  1694.  Prob. 15, pp. 39-40.  Posthumous triplets: boy, boy, girl.  Usual ratios.  Divides  4 : 4 : 2 : 1.

Ozanam.  1725.  Prob. 24, 1725: 179.  Prob. 4, 1778: 187-188;  1803: 185-185;  1814: 160‑161;  1840: 83.  1725 gives just posthumous triplets -- two girls and a boy.  He divides  4 : 2 : 1 : 1.  Montucla does usual form, then remarks that one could have posthumous triplets, e.g. two sons and a daughter, and that he thinks that the will would be declared legally void.

Les Amusemens.  1749.  Prob. 54, p. 186.  Usual case.

Vyse.  Tutor's Guide.  1771?  Prob. 19, 1793: pp. 156-157; 1799: p. 167 & Key p. 209.  Usual case.

Dodson.  Math. Repository.  1775.  P. 13, Quest. XXXIV.  Usual case.

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping, ....  New edition, corrected and enlarged by Alexander Ingram.  [c1780?]  G. & J. Ross, Edinburgh, 1804.  [BMC earliest entry is 7th ed., 1785, then 14th ed., 1815.]  Prob. 53, p. 137.  Usual form, but states that the mother thereby loses  2400£  compared to the case of just having a girl.  What would she have got if she had only had a son?  Answer is  2100£  which assumes the usual  4 : 2 : 1  division for the case of twins.

Vinot.  1860.  Art. XLI: Testament à interpréter, pp. 61-62.  First gives usual solution.  The says the problem is not serious because French legislation gives a solution.  Since the wife receives at least a third in either case mentioned by the husband, she must receive a third in any case.  The author then suggests the rest be divided equally among the children if more than one is born.

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Several examples in an unusual context.

1863 -- p. 128, no. 15;  1873 -- pp. 167-168, no. 5.  Man left  $26,000  to wife, son and daughter.  If the daughter dies before coming of age, the widow gets  ¼;  if the son dies before coming of age, the widow gets  ¾;  what happens if all live?  Unusually for this book, this problem has a remark which says the division should be in the proportion  son : wife : daughter  =  9 : 3 : 1.

1863 -- p. 128, no. 18;  1873 -- p. 168, no. 9.  Man with children abroad and wife at home.  If the son does not return, the widow gets  ⅔;  if the daughter does not return, the widow gets  ⅓;  both return and it is found that the son gets  $3000  more than the daughter.  What was the estate?

1863 -- p. 128, no. 19;  1873 -- p. 168, no. 10.  A, B, C  are thinking of buying a farm.  They agree that if  A and B  buy it, then  A  pays  2/5  and if  B and C  buy it, then  B  pays  2/5.  All three buy it together and  C  is found to pay  $500  more than  A.  What was the cost?

Susan Cunnington.  The Story of Arithmetic.  Swan Sonnenschein,, London, 1904.  Prob. 11, p. 212.  Usual form.  Asserts it is a Roman problem of +300, but gives no references.

Collins.  Fun with Figures.  1928.  Then he put in his other foot, pp. 236-237.  Usual form.  He adds:  A further complication -- triplets, two boys and a girl.  "The easiest way to find out is to let the lawyers decide it, and it is the one best bet that they will get it all."

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  Pp. 53-54: The judge's dilemma.  Ratios are  2 : 1  for son : mother  and  3 : 1  for mother : daughter.  Divides as  6 : 3 : 1.

 

          7.H.   DIVISION AND SHARING PROBLEMS  --  CISTERN PROBLEMS

 

          See Tropfke 578.

          The earliest sources in this group include what I call 'assembly problems'.  In these, there are several processes which constitute a unit of work.  The rates for the processes are given and one has to determine the number of units which can be done in a day (or how long some number of units will take).  See the Babylonian examples below and:  Chiu Chang Suan Ching;  Heron;  Metrodorus 134, 136;  Bakhshali MS;  BR 97, 98;  AR 57, 75;  Muscarello;  Borghi;  Riese;  Cardan;  Tartaglia;  Pike;  Treatise, 1850;  Chambers;  Bullen;  Pearson.  I am indebted to Eleanor Robson for the Old Babylonian examples.  She has also provided most of the references to the source material which I not yet seen.  The dating of these examples is generally pretty vague.

 

          Note.  Cistern problems with two pipes have the same form as meeting problems, cf. 10.A.

 

          NOTATION:  (a, b, c, ...)  means that different pipes, etc. can do the job in  a, b, c, ....  How long for all together?

          Negative values indicate outlets.

 

    a      b     ...

  30    20                                    Lucca 1754;  Wingate/Kersey;  Wells; 

  20    16                                    Tartaglia

  20    15                                    Benedetto da Firenze

  18    12                                    Calandri, c1485

  16    10                                    De Morgan, 1831

  15    12                                    Calandri, 1491;  Tagliente;  De Morgan, 1836; 

  14    12                                    Hutton-Rutherford

  13    10                                    Vyse;  Bonnycastle;  Colenso

  12      7                                    Pike

  10      8                                    Pseudo-dell'Abbaco

  10      7                                    Muscarello

  10      6                                    BR 99

  10   -15                                    D. Adams, 1835

    9   -12                                    Les Amusemens

    8      6                                    Calandri, c1485;  Hutton;  Mittenzwey

    7      5                                    De Morgan, 1836;  Sonnenschein & Nesbit

    6      4                                    Benedetto da Firenze;  Calandri, 1491;  Tagliente;  Gori

    5      4                                    Pseudo-dell'Abbaco

    4      3                                    Lucca 1754;  D. Adams, 1835;  Burnaby

    4      1                                    Heron;  BR 98

    4   -11                                    Calandri, 1491;  Tonstall

14/4   5/2                                    Lacroix

    3      2                                    Gherardi?

    3     -9                                    BR 70

    2     -3                                    Buteo

 

  60    30    20                            Meichsner

  55    45   -30                            Pike

  50    40   -25                            Hutton, c1780?;  Eadon;  Colenso

  27    15    12                            D. Adams, 1801

  24    12      8                            Buteo

  20    15    12                            Unger 520

  20    15    10                            Calandri, c1485

  18    12      6                            Calandri, c1485

  16    12    10                            Simpson

  16    12      8                            Calandri, c1485

  15    12    10                            Benedetto da Firenze;  Calandri, 1491 (twice); 

                                                            Dictator Roffensis;

  12    10      9                            Sonnenschein & Nesbit

  12    10      8                            Mittenzwey

  12    10      6                            Calandri, c1485

  12      9      6                            Borghi;  Ozanam

  12      8      4                            Metrodorus 131

  12      8   -10                            Unger 521

  10      9      8                            Milne

  10      8      4                            Pacioli

  10      5      4                            Pacioli;  Tartaglia

    9      7     -2                            Pike

    9      6     -4                            Mittenzwey

    8      6      4                            Calandri, 1491;  Tonstall

    8      6      3                            Pacioli

    7      5      6                            King

    7      5      4                            AR 70

    7      5      3                            Gherardi

    6      5      4                            Fibonacci

    6      4      3                            Chuquet

    6      4      2                            Metrodorus 135;  Ozanam-Montucla

    6      4     -4                            Sonnenschein & Nesbit

    6      3      1                            Faulhaber

    5      4      3                            Lucca 1754;  Gori;  Les Amusemens

    5      3      2                            Calandri, c1485

    4      3      2                            Gherardi?;  AR 98;  Wagner;  Faulhaber

    4      2      1                            Gori

    3   8/3  12/5                            Newton;  Dodson;  Eadon;  Colenso

    3      2      1                            Metrodorus 133;  Anania(?);  al-Karkhi;  BR 64;  AR 51, 97; 

          Calandri, 1491;  Blasius;  Tonstall;  Riese;  Vyse;  King

    3      1   2/5                            Metrodorus 132

5/3   1/2   1/3                            Chaturveda

    1   3/4   1/2                            Wingate/Kersey

    1   1/2   1/4                            AR 281;  Tonstall

1/2   1/3   1/4                            Columbia Alg.;  Pike

1/2 -10/7  -7/3                            Wingate/Kersey

 

  80    40    20    10                    D. Adams, 1801

  72    60    20    12                    Levi ben Gershon

  27    24      9      6                    Fibonacci

    6      8      9    12                    Recorde

    6      5      4      3                    Bartoli

    6      5      3      2                    Muscarello

    6      4      3      2                    W. Leybourn

    4      3      2      1                    Metrodorus 130;  Fibonacci;  Tonstall

    4      3      2   1/2                    Metrodorus 7;  BR 65;  van Etten;  Wingate/Kersey; 

    4      3      2   1/4                    Schott;  Ozanam

    1   1/2   1/3   1/6                    Bhaskara II

    1   1/2   1/4   1/5                    Chaturveda

1/2   1/3   1/4   1/5                    Mahavira

1/2   1/4   1/5   1/6                    Sridhara

     

    5      3   5/2      1   1/3            Chiu Chang Suan Ching

    4      3      2     -4     -6            BR 116

    3      2     -3     -4     -5            della Francesca

1/2   1/3   1/4   1/5   1/6            Columbia Alg.

1/2   1/3   1/5   1/7   1/9            BR 25

     

    6      5      4      3      2      1    Bartoli

    4      3      2     -3     -4     -5      della Francesca

    3      2      1  -3/4     -4     -5      Cardan

    3      2      1     -2     -3     -4      Pacioli

 

  12    10      8      6     -3     -4    -5       -6     Bullen;  Treatise, 1850

 

          General solution -- see:  Levi ben Gershon;  Wells;  Newton;  Simpson;  Dodson;  Bonnycastle;  Hutton;  Lacroix;  De Morgan;  Bourdon;  Young;  Mittenzwey;  Milne.

 

          The earliest forms derive joint rates from individual rates.  Deriving individual rates from joint rates seems to begin in the 14C.

          NOTATION:  (A, x) in B  means the first can do it in  A  and the first and second together can do it in  B.  How long would it take the second?  For such problems, see:  BR;  Gherardi;  Pseudo-dell'Abbaco;  AR;  Treviso Arith.;  Chuquet;  Calandri, 1491;  Tonstall;  Gemma Frisius;  Tartaglia;  Buteo;  Wingate/Kersey;  Wells;  Simpson;  Euler;  Vyse;  Dodson;  Ozanam‑Montucla;  Bonnycastle;  Pike;  Bullen;  Eadon;  Hutton, 1798?;  Bonnycastle, 1815;  Jackson;  Nuts to Crack;  D. Adams, 1835;  Family Friend;  Treatise, 1850;  Colenso;  Docharty;  Thomson;  Brooks.

 

(50, x) in  36        Gherardi

(48, x) in  24        Docharty

(36, x) in  30        Gherardi

(36, x) in  24        Docharty

(30, x) in  12        Dodson;  Bonnycastle;  Hutton, 1798?;  Nuts to Crack

(20, x) in  60        Silvester

(20, x) in  14        Gemma Frisius

(20, x) in  12        Wingate/Kersey;  Wells;  Euler;  Dodson;  Pike;  Bonnycastle, 1815;  Mittenzwey

(20, x) in    8        Treviso Arith.

(18, x) in  11        Vyse

(35/2, x)              in  40                Docharty (gives a negative x!)

(16, x) in  10        Treatise, 1850

(15, x) in  18        Thomson (gives a negative x!)

(15, x) in  10        Treatise, 1850

(13, x) in    9        AR 76

(13, x) in    8        Pike

(12, x) in    3        Family Friend, 1849

(10, x) in    7        Colenso

( 9, x) in    5        Pseudo-dell'Abbaco

( 8, x) in    5        Buteo;  Eadon

( 7, x) in    5        D. Adams, 1835

( 5, x) in  15/8     BR 67

( 3, x) in   4/3      BR 66

( 3, -x)                in   9/2              BR 69 (negative value!)

(-9, x) in   9/2      BR 68

 

(80, 60, x)           in  30                Tartaglia

(44, 32, x)           in  16                Eadon

(40, 30, x)           in  15                Calandri, 1491;  Tonstall;  Wingate/Kersey

(37, 23, x)           in  15                Pike

(34, 24, x)           in  12                Vyse

(17/2, 21/4, x)     in   6/5              Treatise, 1850

(8,   6, x)            in   3                 Brooks

(5/2, 9/4, x)         in    1                Treatise, 1850

 

          For the general solution of:  (x, y) in A,  (y, z) in B,  (x, z) in C,  see:  della Francesca;  Simpson;  Euler;  Ozanam-Montucla;  Bonnycastle;  Hutton;  De Morgan, 1836;  Colenso;  Singmaster.  For examples of this form, see also:  Muscarello;  Dodson;  D. Adams, 1835;  Docharty;  Todhunter;  Sonnenschein & Nesbit.  This is a form of the type III problem in Section 7.R.1, where the inverses of the variables are used.  Singmaster asks how to choose  A, B, C  so that  x, y, z  and the time for all three together are all integers -- the case with data  20, 15, 12  is by far the simplest example and none of the other examples have this property.

 

    A      B      C

  60      4   -40                            Colenso

  30    20    15                            AR 182

  20    15    12                            Docharty;  Todhunter;  Singmaster

  15    12    10                            della Francesca

  14    12  21/2                            Colenso

  10      9      8                            Simpson;  Euler;  Dodson;  Ozanam-Montucla;  Bonnycastle; 

                                                     Hutton;  Docharty;  Vinot;

    9      8      6                            Sonnenschein & Nesbit

    5      4      3                            Muscarello

    4      6      5                            D. Adams, 1835

 

          Vyse, Docharty and Thomson are the only examples I have seen with four people and you know how long it takes each set of three.  Fish has five workers and you know how long each four take.  If you use the reciprocals of the times, then these are like type III problems in 7.R.1.  That is, if  A, B, C, D  take  A, B, C, D  days, their rate of work is  a = 1/A per day, etc.  Then saying that  A, B, C  can do it in  d4  days becomes  a + b + c = 1/d4, etc.

 

          For problems where the combinations involve one tap or worker working only part of the time that the other does, see:  Fibonacci;  Gherardi;  Chuquet;  Cardan;  Buteo;  Pike;  Jackson;  Treatise, 1850;  Colenso;  Young;  Chambers;  Brooks;  André;  Sonnenschein & Nesbit.

          For problems like  (x, x/2, x/3) in 2,  see:  di Bartolo;  Buteo;  Todhunter.

          For problems like  (x, x‑5) in 12,  which lead to quadratic equations, see:  Di Bartolo;  Buteo;  Tate;  Todhunter;  Briggs & Bryan.

          Sonnenschein & Nesbit has a version where pumps can work at half or full power.

          I have included a few direct rate problems as comparisons -- these usually involve money -- see:  Bakhshali;  Chaturveda;  Pike;  Chambers.

          See Clairaut for the use of this context to discuss negative solutions.

 

          See Smith, op. cit. in 3.  See also 7.E  &  H&S 69‑71.

          5.W.1 can be viewed as parodies of this problem.

 

          COMPARISON of assembly and cistern problems.  Consider the cistern-type problem  (a1, a2 , ...).  In the unit of time, the pipes do  1/a1, 1/a2, ...  of the work, so all together they do  S = Σ 1/ai  per unit time and so the whole job takes time  1/S. 

          In an assembly-type problem, we can do  ai  units of process  i  per unit of time.  Hence it takes  1/ai  time to do one unit of process  i.  If each process has to be done the same number of times, then it takes  S = Σ 1/ai  time to do a unit of work and so  1/S  units can be done in a unit of time.  In the Babylonian problems, the unit of work may require varying amounts of the different units.  If the unit of work requires  bi  units of process  i,  then we take  S = Σ bi/ai.

          Hence the problems are mathematically the same, though the formulations are different.

 

YBC 7164.  Old Babylonian problem tablet at Yale, problems 6 & 7, c‑1700?  Transcribed, translated and commented on in Neugebauer & Sachs, op. cit. in 7.E, 1945, pp. 81-88 & plate 10 & photo plate 35.  On pp. 148-149, a linguistic analysis says it probably comes from Larsa, in southern Mesopotamia.

Problem 7.  A canal has to be cleaned to 3 kùš deep.  A man can clear 20 gín of silt from the top kùš in a day or he can clear 10 gín from the lower level in a day.  How much can he clear in a day?  Here  a1 = 20,  a2 = 10,  and we can take  b1 = 1,  b2 = 2,  because the lower level is twice as thick as the upper level.

Problem 6.  This is the same, but with depth    kùš divided into three levels with the rate of doing the bottom level from 3 to    deep being only    gín per day.  So we just add  a3 = 7½  and  b3 = 1½  to the previous problem.

BM 85196.  Late Old Babylonian tablet in the British Museum, prob. 16, c‑1700?.  Transcribed, translated and commented on by O. Neugebauer; Mathematische Keilschrift-texte II; Springer, Berlin, 1935, pp. 45+, 49, 56+ -- ??NX.  [See 6.BF.2 for another problem from this tablet.]  But Neugebauer was not able to make sense of it until he saw the above problems, so it is reconsidered in Neugebauer & Sachs, pp. 88‑90.  Robson says it is definitely from Sippar (middle Mesopotamia) and cites Thureau-Dangin; Revue d'Assyriologie 32 (1935) 1+ for another publication of the text, ??NYS.  This problem and those of YBC 7164 are more recently discussed by Marvin A. Powell; Evidence for agriculture and waterworks in Babylonian mathematical texts; Bulletin on Sumerian Agriculture 4 (1988) 161-172, ??NYS

                    Like problem 6 above, with each level of depth 1 kùš and rates of  20, 10,  6⅔  gín per day.

In Spring 1994, I mentioned the assembly problems from the Chiu Chang Suan Ching (see below) in a lecture at Oxford.  Eleanor Robson told me that such problems occur in Old Babylonian times and she sent me details, including the above references, and later provided more details and references.  She described four further examples, without specific dates, and the next four examples are simplified from her letter.  The simplifications are basically to avoid use of coefficients giving the number of bricks per unit of weight, etc.

Haddad 104, c-1770.  Tablet from Tell Haddad, near Baghdad, found in the destruction layer from when Hammurabi conquered the site -- usually dated at -1762.  The tablet is in Baghdad.  See:  Farouk al-Rawi & Michael Roaf; Ten Old Babylonian mathematical problems from Tell Haddad; Sumer 43 (1984) 175-218.

Prob. ix -- Making bricks.  One man can dig  1/3  sar of earth in a day, or he can mix  1/6  sar or he can mould  1/3  sar into bricks.  If 1 sar makes  1620  bricks, how many bricks can a team of three make in a day?  For one man, we get  S = 3 + 6 + 3,  so he can process  1/12  sar per day, or  135  bricks, so three men can make  405  bricks.

Prob. x -- Carrying earth to make bricks.  Same problem as the previous, but the earth must be carried 5 nindan from the digging site to the works.  The amount one man can carry in a day is given somewhat cryptically.  The simplest interpretation is that one man can carry  1/3  sar of earth over the 5 nindan in a day, but there still are three workers in the group.  Here we get  S = 3 + 3 + 6 + 3,  so one man can process  1/15  sar per day or  108  bricks and three men make  324  bricks.

YBC 4669.  This and the following tablet are in the same hand, but have no provenance.  See Neugebauer, vol. III, pp. 28-29 & plate 3, ??NYS.  Reverse, col. 3, lines 7-17, c-1800 -- Demolishing walls.  A man can knock down  1/15  sar of wall in  1/5  of a day and he can carry away  1/12  sar in a day.  How much wall can he demolish and carry away in a day;  and what part of the day is devoted to each task?  Here  a1 = (1/15)/(1/5) = 1/3,  so  S = 3 + 12  and he can do  1/15  sar per day.

YBC 4673.  See Neugebauer, vol. III, pp. 30 & 32 & plate 3, ??NYS.  Obverse, col. 2, lines 10-18, c-1800 -- Constructing a pile of bricks.  A man can carry  1/18  sar of earth (bricks??) in a day.  He can pile up 1 sar of bricks in  14 2/5  days.  If a sar makes  5184  bricks of this size, how many bricks can he carry and pile up in a day?

Chiu Chang Suan Ching (Jiu Zhang Suan Shu).  c‑150?  Chap. VI. 

Prob. 22, p. 67.  Man can do two processes at rates of  38  in  3  days and  76  in  2  days.  How many of both together can he do in one day?  Answer is given as  25½  but Vogel's note on the calculation shows  25⅓  was meant, and this is erroneous -- the correct answer is  9½.  The error arises from taking  38  and  76  as rates per day.

Prob. 23, p. 67.  Three processes at rates  50, 30, 15  per day, how many together in a day?  Correct answer,  8⅓,  is obtained.  (Arrow shafts, arrow feathering, arrow heading.)

Prob. 25, p. 68.  Three processes at rates  7, 3, 5  per day, how many together in a day?  Correct answer,  105/71,  is obtained.

Prob. 26, pp. 68‑69.  Cistern:  (1/3, 1, 5/2, 3, 5).  Correct answer.  Vogel says this is the first appearance of the problem.

Heron (attrib.).  c150.  Περι Μετρov (Peri Metron).  In:  J. L. Heiberg, ed.; Heronis Alexandrini Opera Quae Supersunt Omnia, vol. V; Teubner, Leipzig, 1914; reprinted 1976, pp. 176‑177.  Greek and German texts.

Problem 20: Μετρησισ χιστερvασ (Metresis xisternas) [Vermessung einer Zisterne].  (1,4)  in hours, but he computes  1 + 4 = 5  and then sets the cistern  = 12 ft and computes  12/5  as the number of hours.  See BR, c1305, prob. 98, for the explanation of this.

Problem 21: Αλλωσ η μετρεσισ (Allos e metresis) [Die Vermessung in anderer Weise].  + 1/7 ‑ 1/11,  how long to make  100?  He treats the    as  +  and gets the correct answer for that case, though Heiberg says the calculation is senseless.

Smith, History II 538, quotes from Bachet's Diophantos, implying a date of c275, citing the 1570 edition with Fermat's notes, but Smith's citation is to the part of Bachet taken from Metrodorus!  It is Art. 130 of Metrodorus.

                    Sanford 216 also cites Diophantos, but her discussion is based on Smith's AMM article (op. cit. in 3), which is the basis of the section in Smith's History containing Smith's quote.  The problem is nowhere in Heath's edition of Diophantos.

                    However, Tropfke 578 gives a reference to the Tannery edition of Diophantos, vol. 2, p. 46 -- ??NYS.

Metrodorus.  c510.  8 cistern‑type problems.

Art. 7, pp. 30‑31.  "I am a brazen lion."  (2, 3, 4, 1/2),  where 6 hours is counted as  1/2  day, i.e. a day has 12 hours.

Art. 130, pp. 96‑97.  (1, 2, 3, 4).

Art. 131, pp. 96‑97.  (4, 8, 12).

Art. 132, pp. 96‑97.  "This is Polyphemus, the brazen cyclops."  (3, 1, 2/5).

Art. 133, pp. 96‑99.  (1, 3, 2).

Art. 134, pp. 98‑99.  Three spinners can do  1, 4/3, 1/2  unit per day, how long for all three to do one unit?

Art. 135, pp. 98‑99.  "We three Loves" (or Cupids).  (2, 4, 6).

Art. 136, pp. 98‑101.  'Brickmakers.'  Three brickmakers can make  300, 200, 250  per day.  How long for all three to make 300?

Bakhshali MS.  c7C.  Kaye I 49‑52 discusses several types, e.g. first gives 5/2 dinars in 3/2 days;  next gives 7/2 in 4/3;  third gives 9/2 in 5/4;  how long for all three to give 500 dinars? (= Kaye III 192, ff. 21v-22r).  Kaye III 191 has three rates of  1/(1/3), 1/(1/2), 3/5  -- how long to give  100?  I 51 (= III 233-234, ff. 44v-44r) is an example with an income, some capital and three rates of expenditure.  On I 50 (= III 234-235, ff. 44r-43v) is an example with an income, some capital and seven rates of expenditure!

Anania Schirakatzi (= Ananias of Shirak).  Arithmetical problems.  c640.  Translated by:  P. Sahak Kokian as:  Des Anania von Schirak arithmetische Aufgaben; Zeitschrift für d. deutschösterr. Gymnasien 69 (1919) 112-117.  See 7.E for description.  Prob. 24 is a cistern with pipes:  (1, 2, 3),  but he gives the answer:  1/4 + 1/6 + 1/16 + 1/18,  which  =  77/144,  which is close to the correct answer of  6/11.  No working is shown and I am unable to see how  77/144  can arise, even allowing for a possible misprint.

H&S 70 says the cistern problem appears in Alcuin, 9C, but the only possible problem is a trivial problem (8: Propositio de cupa) which mentions a barrel.

Mahavira.  850.  Chap. VIII, v. 32, p. 266.  Cistern:  (1/2, 1/3, 1/4, 1/5).

Chaturveda.  860.  Commentary on Brahma‑sphuta‑siddhanta, chap. XII, sect. 1, art. 9.  In Colebrooke, p. 282. 

Cistern:  (1, 1/2, 1/4, 1/5).  (Datta & Singh, I, 234 and others cite this as Brahmagupta.) 

Bestowing alms:  (1/3, 1/2, 5/3).

Sridhara.  c900.  V. 69, ex. 91, pp. 55‑56 & 95.  Cistern:  (1/2, 1/4, 1/5, 1/6).

al‑Karkhi.  c1010.  Sect II, no. 15‑16, p. 83.  Cistern:  (1, 2, 3).  No. 16 asks how often the cistern will be filled in 5 days.

Bhaskara II.  Lilavati.  1150.  Chap. IV, sect. II, v. 95.  In Colebrooke, p. 42.  Cistern:  (1, 1/2, 1/3, 1/6).  (Datta & Singh, I, 234, erroneously say this is the same problem as Brahmagupta, i.e. Chaturveda.)

Fibonacci.  1202.

P. 182 (S: 279-280):  De Leone et leopardo et urso [On the lion and leopard and the bear].  Lion, leopard and bear eating a sheep:  (4, 5, 6).

P. 182 (S: 280):  De duabus navibus ... [On two ships ...] is in 10.A.

P. 183 (S: 281):  cistern problems  (1, 2, 3, 4)  &  (6, 9, 24, 27).

Pp. 183‑186 (S: 282-285) -- several problems with water butts having different size openings at different heights.  E.g., pp. 183‑184 has four openings at  1/4, 2/4, 3/4, 4/4  of the way down, which could drain the whole butt in  4, 8, 12, 16  days.  How long to drain a full butt with all holes open?  Answer:  7 267/2275  days.

BR.  c1305.

No. 25, pp. 44‑45.  Ship with 5 sails:  (1/2, 1/3, 1/5, 1/7, 1/9).  Vogel says this is the first example of the formulation of a ship with sails.

No. 64, pp. 88‑89.  Cistern,  (1, 2, 3).

No. 65, pp. 88‑91.  Cistern -- 'I am a noble lion',  (2, 3, 4, 1/2).  = Metrodorus 7.

No. 66, pp. 90‑91.  Cistern,  (3, x)  in  4/3.

No. 67, pp. 90‑93.  Cistern,  (5, x)  in  15/8.

No. 68, pp. 92‑93.  Cistern,  (x,  ‑9)  in  9/2.

No. 69, pp. 92‑93.  Cistern,  (3,  ‑x)  in  9/2.

No. 70, pp. 94‑95.  Cistern,  (3, ‑9).

No. 96, pp. 114‑117.  3 cisterns of volumes  30, 60, 120  with pipes that fill them in  6, 4, 3.  Using all three pipes, how long to fill all three cisterns?

No. 97, pp. 116‑117.  Cistern,  + 1/7 ‑ 1/11  to yield  100.  This is Heron's prob. 21 and is done in the same way -- as though it were  + 1/7 + 1/11,  -- but one MS is worded so this is the correct method, as noted by Vogel.

No. 98, pp. 118‑119.  Cistern,  (1, 4).  This is Heron's prob. 20, again noted by Vogel.  The text says to set the cistern equal to  12  and then divides by  5 = 1 + 4.  Vogel notes that this does not give the time, but  12/5  is the volume delivered by the smaller pipe.

No. 99, pp. 118‑119.  Cistern,  (6, 10).

No. 116, pp. 132‑133.  Cistern,  (2, 3, 4, ‑6, ‑4).

Gherardi?.  Liber habaci.  c1310  Pp. 143‑144: Compangnia et viaggio.  Baratti xvii.  Three workers,  (2, 3, 4).   Ship with two sails,  (2, 3).

Levi ben Gershon.  Maaseh Hoshev (= Ma‘aseh hosheb (the  h  should have an underdot)) (Work of the Computer), also known as Sefer ha mispar (Book of Number).  1321 or 1322.  ??NYS -- translation of the following by Shai Simonson, sent by David E. Kullman. 

                    "Question: A certain container has various holes in it, and one of the holes lets all the contents drain out in a given times.  And so on for each of the holes.  How much time will it take to empty the container when all the holes are opened?

                    First, calculate how much drains from each hole in an hour, add them all up, and note the ratio to the full container.  This ratio is equal to the ratio of one hour to the time it will take t empty the container when all holes are open."

                    He then does  (72, 60, 20, 12).

Gherardi.  Libro di ragioni.  1328.

Pp. 44‑45:  Volare una bocte.  (3,5),  but the  5  is half‑way down the barrel while the  3  is at the bottom.

P. 45:  Ship with three sails,  (3, 5, 7).

Pp. 56‑57:  Uno chavaleri che vuole far fare uno pallagio.  Three workers,  (50, B) in 36,  (50, B, C) in 30.

Lucca 1754.  c1330.  F. 59r, p. 134.

Ship with 3 or 2 sails:  (3, 4, 5),  (3, 4).

Two couriers meeting:  (20, 30).

Columbia Algorism.  c1350.

Prob. 66, pp. 87‑88.  Cask can be emptied in  1/2, 1/3, 1/4, 1/5, 1/6  of a day.  (See also Cowley 399.)

Prob. 141, p. 150.  Same with  1/2, 1/3, 1/4.

Giovanni di Bartolo.  Certi Chasi.  c1400.  Copied by Maestro Benedetto (da Firenze), in Cod. L.IV.21, Biblioteca degli Intronati di Siena, 1463.  Edited by M. Pancanti, Quaderni del Centro Studi della Matematica Medioevale, No. 3, Univ. di Siena, 1982.  Cf Van Egmond's Catalog 189-190 which doesn't mention this material.

Prob. 1, pp. 4‑5.  (x, x/2)  in  5.

Prob. 2‑8, pp. 5‑17 are more complex examples, often leading to quadratic equations, e.g.  (x, x‑5)  in  12.

Bartoli.  Memoriale.  c1420. 

Prob. 14, f. 76v (= Sesiano, pp. 140-142 & 149, with facsimile of the relevant part of f. 76v on p. 141.  Cask with four taps:  (3, 4, 5, 6).

Prob. 33, f. 79r (= Sesiano, pp. 146 & 150).  Six workers building a wall:  (1, 2, 3, 4, 5, 6).  He correctly finds the total rate is  137 [/ 60],  but then uses  36  instead of  60  -- "Now make  6  times  6,  which is  36,  because there are  6  workers."  Sesiano describes this as fantasy.

Pseudo-dell'Abbaco.  c1440.

Prob. 23, p. 32.  (4, 5).

Prob. 50, p. 49 with plate on p. 50.  Ship with two sails,  (8, 10).  I have a colour slide of this.

Prob. 62, p. 59.  (9, x)  in  5.

AR.  c1450.  Prob. 51, 57, 70, 75, 76, 97, 98, 182, 281.  Pp. 42, 44‑45, 48‑50, 58‑59, 85‑86, 128‑129, 157, 160‑161, 165‑166, 175, 211‑213, 221.

51:  cistern with three drains,  (1, 2, 3),  erroneously done -- see 97.

57:  three mills, but gives amounts each can do per day.

70:  three builders,  (7, 5, 4).

75:  three tailors, but gives amounts each can do per day.

76:  (13, B)  in  9.

97:  cistern with three drains,  (1, 2, 3),  = Metrodorus 133.

98:  ship with three sails,  (2, 3, 4).

182:  three scribes;  (A, B) in 20,  (A, C) in 30,  (B, C) in 15,  how long for each?

281:  barrel with three taps,  (1, ½, ¼).

Benedetto da Firenze.  c1465.

Pp. 90‑91:  ship with three sails  (10, 12, 15).

P. 91:  cistern  (4, 6).

P. 91:  two workers  (20, 5).

"The Treviso Arithmetic" = Larte de labbacho (there is no actual title).  Treviso, 1478.  Translated by David Eugene Smith, with historical commentary by Frank J. Swetz, as:  Capitalism and Arithmetic; Open Court, La Salle, Illinois, (1987), improved ed., 1989.  This is discussed in:  D. E. Smith; The first printed arithmetic (Treviso, 1478); Isis 6 (1924) 310‑331.  Facsimile edition, from the copy at the Diocese of Treviso, with commentary booklet by Giuliano Romano, (Editore Zoppelli, sponsored by Cassa di Risparmio della Marca Trevigian, Treviso, 1969);  updated ed., Libreria Canova, CalMaggiore 31, Treviso (tel: 0422-546253), 1995 [Swetz, p. 324, cites the 1969 ed.]  See:  www.calion.com/cultu/abbacho/abbacen.htm  for a description of the book and how to order it.  The facsimile has taken its title from the end of the opening sentence.  Romano's commentary calls it:  L'Arte dell'Abbacho.  The text nowhere gives a publisher's name.  Smith, Rara, pp. 3-7, says it was probably published by Manzolo or Manzolino, while Swetz, p. 26, specifies Michael Manzolo or Manazolus.  Romano says it was published by Gerardus de Lisa.  There was a copy in the Honeyman Collection, with title  Arte dell'Abbaco  and publisher [Gerardus de Lisa], who is described as the prototypographer at Treviso from 1471.  The entry says only ten copies are known -- the web page says nine.

                    F. 57r (pp. 162‑163 in Swetz).  Carpenters,  (20, x)  in  8.

Muscarello.  1478. 

F. 58r, p. 161-162.  (B, C)  in  3,  (A, C)  in  4,  (A, B)  in  5.  There are two copying errors in the MS answers.

F. 62v, pp. 169-170.  Four workers,  (2, 3, 5, 6).

Ff. 77r-77v, pp. 192-193.  Three mills can do  9, 8, 5  per day.  How long will it take them to do  6  and how much does each do?

F. 77v, p. 193.  Ship with two sails,  (7, 10).

F. 81r, p. 196.  Cask with three spouts which can let out  6, 7, 8  per hour.  How long will it take to empty a cask of  23?

della Francesca.  Trattato.  c1480. 

Ff. 15r-15v (61-62).  Basin with three inlets and three outlets,  (2, 3, 4, -3, -4, -5).  Plug the  +4  pipe, which gives  (2, 3, -3, -4, -5).  English in Jayawardene.

F. 127r (269).  Three workers:  A & B in 15;  A & C in 12;  B & C in 10.  English in Jayawardene.

Wagner.  Op. cit. in 7.G.1.  1483.  Regel von einem Fass, pp. 114 & 224.  Cask with three taps  (2, 3, 4).

Chuquet.  1484. 

Prob. 21.  English in FHM 204.  Cistern emptying,  (3, 4, 6).

Prob. 53.  English in FHM 209-210.  The first says:  "If you help me  8  days, I will build it in  20".  The second responds:  "If you help me  10  days, I can do it in  15".  How long for each alone?

Prob. 54.  Same as prob. 53 with parameters   5, 17;  6, 24.

Borghi.  Arithmetica.  1484. 

F. 106v (1509: ff. 91r-91v).  Three mills can grind  6, 9, 11  per day.  How long to do  100?

F. 109r (1509: ff. 91v-92r).  Ship with three sails,  (6, 9, 12).  (H&S 70 gives Latin and English.)

Calandri.  Aritmetica.  c1485.

F. 91r, p. 182.  Ship with 2 sails:  (12, 18).

F. 91v, p. 183.  Three men in prison:  (6, 12, 18).  (Tropfke 520 reproduces this in B&W.)

F. 93r, p. 186.  Emptying a cask:  (6, 8).

F. 95v, p. 191.  Ship with three sails:  (6, 10, 12).  (Coloured plate opp. p. 120 of the text volume.)

F. 96v, p. 193.  Emptying a cask:  (8, 12, 16).

F. 97v, p. 195.  Lion, wolf & fox eating a goat:  (2, 3, 5).  (Tropfke 581 reproduces this in B&W.)

Ff. 98v-99r, pp. 197‑198.  Furnace with 3 fires:  (10, 15, 20).

Johann Widman.  Op. cit. in 7.G.1.  1489.  (On pp. 131-132, Glaisher mentions the following.)  Ff. 136r‑138v:  Eyn fasz mit dreyen Czapfen;  Von der Mulen;  Leb, wolff, hunt;  Schiff.  (Cistern problem;  3 mills;  lion, wolf, dog eating a sheep;  ship with 3 sails.)

Calandri.  Arimethrica.  1491.

F. 68v.  Ship with two sails.  (12, 15).  Woodcut of ship with indeterminate number of sails.

F. 69r.  Cask with two taps.  (4, 6).  Woodcut of cask with two taps.

F. 70r.  Ship with three sails.  (10, 12, 15).  Same woodcut as on f. 68v.

F. 70r.  Cask with three taps.  (4, 6, 8).  Same woodcut as on f. 69r.

F. 70v.  Three masters build a house.  (10, 12, 15).  Woodcut of two builders.  (H&S 70 gives Italian and English and says it also occurs in the Treviso Arithmetic (1478) [but that has a very different type!], Pacioli, Cataneo, Tartaglia, Buteo (1559), Clavius, Tonstall.)

F. 70v.  Three masters doing a job.  (30, 40, x)  in 15. 

F. 71v.  Cistern.  (4, ‑11).  Woodcut of cistern.  (Rara, 48 is a reproduction.)

F. 72v.  Lion, leopard & wolf eating a sheep.  (1, 2, 3)  days.  Nice woodcut.  (H&S 70 gives Italian and English, says there is a remarkable picture and says it occurs in Fibonacci [again, there it occurs in a different form] and Cataneo.)

Pacioli.  Summa.  1494.  See also Buteo.

F. 99r, prob. 6.  Building a house,  (8, 10, 4).  Says one can have more builders and it is similar to dog, wolf & lion eating a sheep.

F. 99v, prob. 16.  Three mills,  (6, 8, 3)  days.

F. 99v, prob. 17(not printed).  Three mills,  (10, 5, 4)  days.

   PART II.

F. 66v. prob. 91.  Cask with four taps.  Volume above highest tap is  1/3  of the cask.  Volume between highest and second highest is  1/4;  volume between second and third highest is  1/5;  volume between third highest and lowest tap is the rest of the cask.  Each tap can empty the section just above it in  1, 2, 3, 4  days.  How long to empty with all taps?  He assumes the cask holds  60  so the rates  are  20, 15/2, 12/3, 13/4  per day.  Answer is  80/139 + 60/59 + 48/29 + 4,  but he gives the sum as  7 245235/2959139.  Clearly the denominator denotes  29·59·139 = 237829,  but the correct sum is  7 58901/237829  and I cannot see how his expression relates to the answer.  The answer is not  7 + 24/29 + 52/59 + 35/139,  nor any similar expression that I can think of. 

Ff. 66v-67r, prob. 92.  Basin has inlets which fill it in  1, 2, 3  hours and outlet which empty it in  2, 3, 4  hours, i.e.  (1, 2, 3, -2, -3, -4).  How long to fill?  He follows with remarks that all such problems can be done similarly.  Cf della Francesca.

Blasius.  1513.  F. F.iii.r: Decimatertia regula.  Three rivers can water a field in  (1, 2, 3)  days.  Gets  13 1/11  hours for all three -- so he is using  24  hour days.

Tagliente.  Libro de Abaco.  (1515).  1541. 

Prob. 117, f. 58r.  Ship with two sails --  (12, 15).

Prob. 119, f. 58v.  Cask with two taps --  (4, 6).

Tonstall.  De Arte Supputandi.  1522. 

Quest. 26, pp. 157-159.  Three mills can do at rates of  18, 13, 8  per day.  How long to do 24?

Quest. 27, pp. 159-161.  Cistern,  (1, 2, 3)  and  (4, 6, 8).

Quest. 28, pp. 161-162.  Cistern,  (1/4, 1/2, 1).

Quest. 29, pp. 162-163.  Cistern,  (4, -11).

Quest. 32, pp. 164-165.  Four architects building a house,  (1, 2, 3, 4)  years.  Says it is similar to a cistern problem.

Quest. 33, p. 166.  Three architects building a house,  (30, 40, x)  in  15.

Riese.  Die Coss.  1524.

No. 117, p. 55.  Cask with three taps,  (1, 2, 3).

No. 118, p. 56.  Three windmills can grind  20, 17, 15  per day.  How long to do  24?

Cardan.  Practica Arithmetice.  1539.

Chap. 47, ff. L.iii.r - L.iii.v (pp. 70-71).  Simple example --  5  mills grind  7, 5, 3, 2, 1  per hour, how long will they take to grind  500? 

Chap. 66, section 125, ff. kk.vi.r - kk.vi.v (pp. 180-181).  Cask with four taps located at levels  1/3,  1/3 + 1/4,  1/3 + 1/4 + 1/6,  1  from the top and which empty their respective portions in  4, 3, 2, 1  hours.  How long to empty the cask with all taps?

Chap. 66, section 126 (misprinted 123), ff. kk.vi.v - kk.vii.r (p. 181).  Cistern:  (1, 2, 3, ‑4, -5, -3/4).

Gemma Frisius.  Arithmetica.  1540.  (20, x)  in  14  -- man & wife drinking a cask of wine.  ??NYS -- Latin given in H&S, p. 71.

Recorde.  Second Part.  1552.  1668, pp. 329-330: A question of water, the eighth example.  (6, 8, 9, 12).

Tartaglia.  General Trattato, 1556, art. 74, p. 248v;  art. 176‑177, p. 261v;  art. 187‑188, pp. 262r‑262v.

Art. 74:  120  per  40  and  15 ½  per  6  to do  120.

Art. 176:  (16, 20).

Art. 177:  (60, 80, x)  in  30.

Art. 187:  1 per 8,  1 per 6  and  1 per 3  to do  25.

Art. 188:  (10, 5, 4).

Buteo.  Logistica.  1559. 

Prob. 6, pp. 205-206.  Three mighty drinkers drinking an amphora of wine in  (24, 12, 8)  hours.  Cites Pacioli.  (H&S 71)

Prob. 7, pp. 206-208.  Three architects build a house:  (x, x/2, x/3)  in  2  months.  Says Pacioli gives  (x, x+6, x+8)  in  2  and solves it wrongly.

Prob. 8, pp. 208-209.  Ship with two sails,  (8, x)  in  5.

Prob. 61, pp. 266-268.  Cask with three taps  1/4, 2/3, 1  of the way down which could empty the whole cask in  (6, 3, 3)  hours.

Prob. 62, pp. 268-269.  Cistern,  (+2, -3).

Gori.  Libro di arimetricha.  1571.

F. 74v (pp. 82‑83).  Cistern empties in  (4, 6)  hours.  Ship with three sails,  (3, 4, 5)  days.

F. 77v (p.83).  Lion, bear and wolf eating a sheep,  (4, 2, 1)  hours.

Johann Faulhaber.  Arithmetischer Wegweiser ....  Ulm, 1614.  ??NYS.  A 1708 ed. is quoted in Hugo Grosse; Historische Rechenbücher des 16. und 17. Jahrhunderts; (1901);  reprinted by Sändig, Wiesbaden, 1965, p. 120. 

No. 91, p. 228:  wolf, sheepdog and dog eating a sheep,  (1, 3, 6). 

No. 92, p. 229:  ship with three sails,  (2, 3, 4).

van Etten.  1624.  Prob. 83 (76): Du Lyon de Bronze posé sur une fontaine avec cette epigraphe, pp. 94‑95 (140).  (2, 3, 4, 1/2)  = Metrodorus, art. 7.

Georg Meichsner.  Arithmetica Historica.  Hieronymus Körnlein, Rotenburg/Tauber, 1625.  No. 68, p. 209.  ??NYS.  Quoted in Hugo Grosse, op. cit. under Faulhaber, above, p. 77.  Three men with devices to pump out flooded lands in Holland,  (60, 30, 20).

Schott.  1674.  Ex. 1, pp. 570-571.  Cistern:  (2, 3, 4, 1/4)  done several ways.  Cites Clavius for the lion fountain (Metrodorus 7).

Wingate/Kersey.  1678?. 

Quest. 4, pp. 476-477.  Workmen:  (20, 30).

Quest. 5, pp. 477-478.  "I am a brazen lion ..." in Latin.  (2, 3, 4, 1/2),  where 6 hours is counted as  1/2  day, i.e. a day has 12 hours.  = Metrodorus, Art. 7.

Quest. 6, pp. 478-479.  (1/2, -10/7, -7/3).

Quest. 7, pp. 479-480.  Dog, wolf, lion eating a sheep: (1, 3/4, 1/2),  but the lion has a head-start of  1/8  hour.

Quest. 12, p. 484.  (20, x)  in  12.  Man and wife drinking beer.

Quest. 13, pp. 484-485.  (30, 40, x)  in  15.  Carpenters building a house.

W. Leybourn.  Pleasure with Profit.  1694.  Prob. 14, p. 39.  Cistern emptying: (6, 4, 3, 2).

Wells.  1698. 

No. 105, p. 206.  (20, 30)  and  (a, b).

No. 106, p. 206.  (20, x)  in  12  and  (a, x)  in  c.

Isaac Newton.  Arithmetica Universalis, 1707.  ??NYS.  English version:  Universal Arithmetic, translated by Mr. Ralphson, with revisions and additions by Mr. Cunn, Colin Maclaurin, James Maguire and Theaker Wilder; W. Johnston, London, 1769.  (De Morgan, in Rara, 652‑653, says there were Latin editions of 1722, 1732, 1761 and Raphson's English editions of 1720 and 1728, ??NYS.)  Resolution of Arithmetical Questions, Problem VII, pp. 184‑185.  "The Forces of several Agents being given, to determine  x  the Time, wherein they will jointly perform a given Effect  d."  Gives general approach for three workers.  Example is  (3, 8/3, 12/5),  where the Force of the second is expressed as saying he can do the work "thrice in  8  weeks".

Ozanam.  1725.

Prob. 24, question 9, 1725: 180‑181.  Prob. 5, 1778: 188-189;  1803: 185-186;  1814: 161-162;  1840: 84.  Same as Metrodorus 7, except that a day is considered as  24  hours, so the problem is done as  (2, 3, 4, ¼).

Prob. 24, question 10, 1725: 181.  (6, 9, 12)  months to print a book.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790).

Prob. XI, pp. 83-84 (1790: prob. XXX, pp. 91‑92).  General problems,  (a, b)  and  (a, b, c).  Example:  (10, 12, 16).

Prob. XXXI, pp. 90-92 (1790: prob. XXXVI, pp. 96‑98).  General problem:  given  (x, y)  in  a,  (x, z)  in  b,  (y, z)  in  c, determine  x, y, z.  Example with  a, b, c   =  8, 9, 10.

Alexis-Claude Clairaut.  Elémens d'Algèbre.  1746.  Vol. I, art. LVI (my source quotes from the 6th ed. of 1801).  ??NX.  He uses the context of this type of problem to study the meaning of negative solutions.  A cistern of size  a  is filled by a source running for time  b  together with another source running for time  c.  Another reservoir of size  d  is filled by the sources in times  e  and  f.  Determine the rate of each source.

Les Amusemens.  1749. 

Prob. 173, p. 321.  Cistern:  (9, -12).

Prob. 174, pp. 322-323.  Reservoir with three nymphs:  (3, 4, 5).

"By his Holiness the Pope".  The Gentleman and Lady's Palladium (1750) 22.  Qn. 11.  (??NYS, cited by E. H. Neville; Gleaning 1259: On Gleaning 1146; MG 23 (No. 254)  (May 1939) 149.  "If a Cardinal can pray a soul out of purgatory ...."  See Welch, 1833, below.

Dictator Roffensis, proposer;  Steph. Hodges & Will. Smith, solvers.  Ladies' Diary, 1750-51  =  T. Leybourn, II: 45-46, quest. 334.  [??NX of p. 46.]  Three drinkers:  (10, 12, 15)  for 12 hour days,  how long together for 10 hour days.

Arthur Young.  Rural Oeconomy: or, Essays on the Practical Parts of Husbandry.  Dublin, 1770, p. 32.  ??NYS - described in: Keith Thomas; Children in early modern England; IN: Gillian Avery &  Julia Briggs; Children and Their Books  A Celebration of the Work of Iona and Peter Opie; OUP, (1989), PB ed, 1990, pp. 45-77, esp. pp. 66 & 76.  Proverb: one boy, one day's work; two boys, half a day's work; three boys, no work at all.

Euler.  Algebra.  1770.  I.IV.III: Questions for practice.

No. 14, pp. 204‑205.  Cistern,  (x, 20)  in  12.

No. 22, p. 205.  Same as the example in Simpson's prob. XXXI

Vyse.  Tutor's Guide.  1771? 

Prob. 61, 1793: p. 69; 1799: p. 74 & Key p. 100.  Two workers,  (10, 13).

Prob. 62, 1793: p. 69; 1799: pp. 74-75 & Key p. 100.  Boatbuilders,  (18, x) in 11.

Prob. 6, 1793: p. 128; 1799: p. 136 & Key p. 178.  Cistern emptying,  (1, 2, 3,).  Gives volume of cistern but never uses it.  = Metrodorus 133.

Prob. 15, 1793: p. 156; 1799: p. 167 & Key p. 208.  Builders,  (34, x, 24) in 12.

Prob. 5, 1793: p. 189; 1799: p. 201 & Key pp. 244-245.  Trenching a field.  A, B, C  can do in  12;  B, C, D  can do in  14;  C, D, A  can do in  15;  D, A, B  can do in  18.  How long for all four together and for each one singly?  Solution in decimals.

Dodson.  Math. Repository.  1775.

P. 22, Quest. LVIII.  Man and wife drinking beer.  (30, x) in 12.

P. 23, Quest LIX.  Cistern:  (20, x) in 12.

Pp. 52-53, Quest. CV.  Workers.  A & B in 8;  A & C in 9;  B & C in 10.

P. 56, Quest 56.  (3, 8/3, 12/5),  phrased as in Newton.  Does the problem in general, then applies to the data.

Ozanam-Montucla.  1778.

Question 5, 1778: 193-194;  1803: 190-191;  1814: 165-166;  1840: omitted.  Same as Metrodorus 135.

Prob. 22, 1778: 214;  1803: 209.  Prob. 21, 1814: 181;  1840: 94.  Same as the example in Simpson XXXI.

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?]  1804: prob. 70, pp. 139-140.  Cistern:  (40, 50, -25).

Bonnycastle.  Algebra.  1782.

Pp. 82-83, no. 10 (1815: p. 102, no. 9).  (10, 13).  = Vyse, prob. 61.

P. 83, no. 11 (1815: pp. 102-103, no. 10).  (a, b)  done in general.

P. 86, no. 25 (1815: p. 108, no. 34).  Two drinkers:  (30, x)  in  12  days.

P. 86, no. 26 (1815: p. 108, no. 36).  Same as the example in Simpson's prob. XXXI.

P. 86, no. 27.  "If three agents,  A,  B,  and  C,  can produce the effects  a,  b,  c,  in the times  e,  f,  g,  respectively;  in what time would they jointly produce the effect  d?"

Pike.  Arithmetic.  1788. 

P. 335, no. 8.  Cistern:  (1/2, 1/4, 1/3).

P. 350, no. 14.  Merchant gaining and losing, equivalent to  (7, 9, ‑2)  -- how long to empty a full tank?

P. 350, no. 19.  Two workers:  (7, 12).

Pp. 350-351, no. 20.  Boatbuilders:  (20, x)  in  12.

P. 351, no. 21.  Two workers:  (13, x)  in  8.

P. 351, no. 22.  Three workers:  (23, 37, x)  in  15.

P. 351, no. 23.  Cistern:  (55, 45, -30).

P. 351, no. 24.  Cistern of  73,  inflow of  7/5  and outflow of  20/17  both run for two hours, then the outflow is stopped.

P. 355, no. 39.  A,  B,  C  do a job.  A  and  B  do  3/11  of it,  A  and  C  do  5/13,  B  and  C  do  4/14.  (Also entered at 7.G.1.)

Bullen.  Op. cit. in 7.G.1.  1789.  Chap. 38. 

Prob. 34, p. 243.  Mother & two daughters spin  3  lb flax in  1  day;  mother can do it in    days;  elder daughter in    days;  how long does it take the younger daughter? 

Prob. 51, pp. 245‑246.  Cistern:  (6, 8, 10, 12, ‑6, ‑5, ‑4, ‑3)  -- how long to empty from full?

Eadon.  Repository.  1794. 

P. 78, no. 21.  If  3  men or  4  women can do a job in  68  days, how long will it take  2  men and  3  women?

P. 79, no. 26.  If  5  oxen or  7  colts can eat a close in  87  days, how long will it take  2  oxen and  3  colts?  Answer is  105,  which neglects growth of grass.

P. 195, no. 10.  (8, x)  in  5.

P. 195, no. 11.  (32, 44, x)  in  16.

P. 195, no. 12.  (3, 8/3, 12/5).  = Newton's example.

P. 367, no. 6.  (40, 50, -25).

John King, ed.  John King   1795   Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 100:  (1, 2, 3);  (7, 5, 6).

Hutton.  A Course of Mathematics.  1798? 

Prob. 8,  1833: 212-213;  1857: 216-217.  (6, 8),  then generalises to  (a, b, c, d).

Prob. 15,  1833: 220-221;  1857: 224-225..  A  and  B  can do in  a  days;  A  and  C  in  b  days;  B  and  C  in  c  days.  How long for each singly and all three together?

Prob. 39,  1833: 223;  1857: 227.  (x, 30)  in  12.

Prob. 40,  1833: 223;  1857: 227.  Problem 15 above with numerical values:  a, b, c  =  8, 9, 10.

D. Adams.  Scholar's Arithmetic.  1801.  P. 125, nos. 26 & 27.  (80, 40, 20, 10)  &  (27, 15, 12).

Bonnycastle.  Algebra.  10th ed., 1815.  P. 226, no. 6.  Cistern:  (20, x)  in  12  hours.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.

No. 4, pp. 15 & 71.  A  &  B  earn  40s  in  6  days;  A  &  C  earn  54s  in  9  days;  B  &  C  earn  80s  in  15  days.  What does each earn per day?

No. 11, pp. 16 & 73.  "A  in five hours a sum can count, Which  B  can in eleven;  How much more then is the amount They both can count in  7?"

No. 29, pp. 21 & 80-81.  Lion, wolf, dog eating a sheep:  (1/2, 3/4, 1),  but the lion begins  1/8  before the others.

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Section 14, ex. 1, pp. 25-27.  (5/2, 15/4)  and  (a, b)  in general.

Augustus De Morgan.  Arithmetic and Algebra.  (1831?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Arts. 3 & 112, pp. 1-2 & 28-29.  In Art. 3, he mentions the problem  (10, 16) as an example of algebraic formulation.  In Art. 112, he solves it and  (a, b).

Welch.  Improved American Arithmetic.  1833 ed.  This must be Oliver Welch's American Arithmetic, first published in 1812 and which went through at least eight editions up to 1847.  [Halwas 459-465, of which 1833 is 462.]  "If a Cardinal can pray a soul out of purgatory by himself in 1 hour, a bishop in 3 hours, a Priest in 5 hours, a Friar in 7 hours, in what time can they pray out 3 souls, all praying together?"  In the 1842 ed., this was changed to steam, water, wind and horse power.  ??NYS -- quoted in Gleaning 1146, MG 21 (No. 245) (Oct 1937), 258.  See above at 1750 for an earlier version.

Nuts to Crack II (1833), no. 129.  (30, x)  in  12.  Identical to Bonnycastle, 1782, no. 25.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 57, p. 85.  General solution for  (a/b, c/d, e/f).

D. Adams.  New Arithmetic.  1835. 

P. 243, no. 74.  Ship has a leak which will fill it in 10 and a pump which will empty it in 15, i.e.  (10, -15).

P. 247, no. 109.  Two workers.  (3, 4).

P. 247, no. 110.  Three workers.  A & B  can do in  4;  B & C  in  6;  A & C  in  5.

P. 247, no. 111.  Two workers.  (7, x)  in  5.

Augustus De Morgan.  On The Study and Difficulties of Mathematics.  First, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  P. 31.  (12, 15),  then does  (a, b).

Augustus De Morgan.  Examples of the Processes of Arithmetic and Algebra.  Third, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836. 

Prob. 16, p. 28.  Given  (5, 7),  how long will it take to do  ⅔?

P. 91.  A & B  in  c;  A & C  in  b;  B & C  in  a.

Unger.  Arithmetische Unterhaltungen.  1838. 

Pp. 135 & 258, no. 519.  A  can do  63  in  8  days;  B  can do  37  in  6  days  and  C  can do  25  in  3  days.  How long for all three to do  268½?

Pp. 135 & 258, no. 520.  Cistern  (20, 15, 12).

Pp. 135-136 & 258, no. 521.  Cistern  (12, 8, -10).

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 8,  1857: 81.  Two workers:  (12, 14).

T. Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.

P. 75, no. 11.  (x + 6, x + 3/2)  in  x.

P. 76, no. 12.  (x, x+5)  in  6.

P. 85, no. 8.  A  can reap a field in  a  days.  If assisted by  B  for  b  days, then  A  only has to work  c  days.

Family Friend 1 (1849) Answers to correspondents, pp. 4 & 6.  Questions requiring answers. No. 1.  (12, x)  in  3.

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  Third edition, revised, improved and enlarged.  Published by direction of the Commissioners of National Education in Ireland, Dublin, 1850.  Many examples, of which the following are the more interesting.

Pp. 200-201, no. 133.  (17/2, 21/4, x)  in  6/5.

P. 347, no. 8.  (15, x)  in  10.

P. 347, no. 11.  5  mills grind  7, 5, 4, 3, 1  bushels per hour.  How long for all five to grind  500  bushels?

P. 356, no. 10.  A,  B  &  C  can do in  10  days;  B  &  C  can do in  16  days;  how long for  A?  (This is equivalent to  (x, 16)  in  10.)

P. 358, no. 30.  A,  with  2  days' help from  B,  can do in  12  days.  B,  with  4  days' help from  A,  can do in  8  days.  How long for both together?

P. 358, no. 31.  A  and  B  can do in  8  12-hour days.  A  can do in  12  16-hour days.  How many  14-hour days would  B  need?

P. 359, no. 36.  (5/2, 9/4, x)  in  1.

P. 359, no. 39.  (6, 8, 10, 12, -6, -5, -4, -3)  -- how long to empty from full?

John William Colenso (1814-1883).  Arithmetic Designed for the Use of Schools ....  New edition.  Longman, Brown, Green, and Longmans, 1853.  ??NX - Wallis 246 COL.  I have 1857, which seems identical for pages 1 - 164, then adds a chapter on decimal coinage on pp. 165-171.  I also have 1871, which rearranges the material at the end and adds Notes and Examination-Papers -- the advertisement on p. v says the additional material was added by J. Hunter in 1864.  This gives a large number of variations of the problem which I have included here as representative of mid 19C texts. 

     Miscellaneous Examples, pp. 122‑136, with answers on pp. 161-163 (1871: 211‑213).

No. 21.  (10, 13).

No. 27.  (10, B)  in  7.

No. 31.  Cistern, (40,  50, ‑25).

No. 52.  (3, 8/3, 12/5),  where  B  is determined from  "B  can do thrice as much in  8  days".  I.e. Newton's example; see also Eadon.

No. 80.  If  5  oxen or  7  horses can eat the grass of a field in  87  days, how long will it take  2  oxen and  3  horses?  (The grass is not assumed to grow.)   = Eadon, No. 26.

No. 101.  If  3  men,  5  women or  8  children can do a job in  26½  hours, how long will it take  2  men,  3  women and  4  children?

     Examination--Paper VIII, (1864), 1871: pp. 170‑172, with answers on p. 214.

No. 1.  M  can do in  20  7‑hour days.  N  can do in  14  8‑hour days.  How many hours per day must they work together to do it in  10  days?

No. 2.  Cistern,  (20, 24, ‑30).  How full is it after 15?

No. 3.  Reapers,  (F, G) in  8¾,  with  3½ : F  =  5 : G.

No. 4.  (34, 38),  but second man stops  4  days before the end.

No. 5.  Cistern.  (A, B)  fills in  4;  (A, ‑C)  empties in  40;  (B, ‑C)  fills in  60.

No. 6.  4  men, working various parts of the time.

No. 7.  (A, B)  in  14,  (B, C)  in  10½,  (A, C)  in  12.

No. 8.  B = twice (A, C);  C = thrice (A, B);  (A, B, C)  in 5.

No. 9.  Three men working various parts of the time.

No. 10.  Cistern with 2 inlets and 2 outlets running various parts of the time.

John Radford Young (1799-1885).  Simple Arithmetic, Algebra, and the Elements of Euclid.  IN: The Mathematical Sciences, Orr's Circle of the Sciences series, Orr & Co., London, 1854.  [An apparently earlier version appeared as Circle of the Sciences, vol. (2 &) 3, with no TP or other details except two editorial remarks referring to Professor Young as author of the Arithmetic and Algebra sections.  I have vol. 3, which begins with the last sheet of Arithmetic and then covers the other material.  (I saw a vol. 8? on zoology? in the same bookshop.)  This version includes detailed solutions and some plates not connected with the text.  The 1854 text is identical except that the order of topics has been changed and there are some consequent changes to the text.]  This is a typical mid 19C text with a number of cistern problems, of which the more interesting are the following.

No. 16, p. 178.  A  can do a job in  10  days.  After he has worked for  4  days,  B  comes to assist and they finish it in  2  more days.  How long would  B  take by himself?

No. 4, p. 207.  A man and his wife can drink a barrel in  15  days.  After  6  days the man leaves and the woman finishes it in  30  days.  How long would it take her to drink the whole barrel by herself?

No. 10, p. 208.  Similar to the last, with two workers and numbers  16,  4,  36.

No. 12, p. 208.  General solution for  (a, b, c).

Gerardus Beekman Docharty.  A Practical and Commercial Arithmetic: ....  Harper & Brothers, NY, 1854.  Many examples on pp. 166-167, 242-243, 247, including Simpson's XXXI;  the same problem with values  12, 20, 15;  (35/2, x)  in  40  and the following.

P. 167, no. 36.  A,  B  and  C  can do in  24;  A  and  B  can do in  48;  A  and  C  can do in  36.  How long for each separately?

Pp. 247-248, no. 64.  = Vyse, prob. 5, but only asks how long for all together, and gives solution in fractions.

Vinot.  1860.  Art. LVII: Les trois Ouvriers, pp. 74-75.  Same as the example in Simpson XXXI.

James B. Thomson.  Higher Arithmetic; or the Science and Application of Numbers; ....  Designed for Advanced Classes in Schools and Academies.  120th ed., Ivison, Phinney & Co, New York, (and nine copublishers), 1862.  Lots of straightforward examples and the following.

Prob. 65, p. 397 & 422.  (15, x)  in  18.  Note that  x  has a negative value, i.e. is an outlet.  See also BR & Docharty.

Prob. 93, p. 398 & 422.  = Vyse, prob. 5, with solution in fractions.

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Lots of examples.  I mention only those of some novelty to illustrate mid/late 19C texts.

1863 -- p. 122;  1873 -- pp. 140-141;  no. 16.  (6, 8, x)  in  3.

1863 -- p. 122;  1873 -- p. 141;  no. 18.  (A, B, C)  in  4;  (A, B)  in  6;  (B, C)  in  9.

1863 - pp. 122-123;  1873 -- p. 141;  no. 19.  (A, B, C)  in  6;  (A, B)  in  8;  B  in  12.

1863 -- p. 123;  1873 - p. 141;  no. 26.  (A, B, C)  can do in  6;  (A, B)  can do in  9;  all three work for  2  days, then  C leaves -- how long for  A, B  to finish?

1863 - p. 155, no. 2;  1873 -- p. 174, no. 9.  (A, B, C)  can do in  20;  (A, B)  can do in  30;  (B, C)  can do in  40;  all three work for  5  days, then  B  leaves -- how long for  A, C  to finish?

Boy's Own Magazine 2:2 (No. 8) (Aug 1863) 183  &  2:4 (No. 11) (Nov 1863) 367.  (Reprinted as (Beeton's) Boy's Own Magazine 3:8 (Aug 1889) 351  &  3:10 (Oct 1889) 431.)  Mathematical question 87.  Complicated version, typical of its time.  Bacchus drinks from a cask for    of the time it would take Silenus to drink the whole cask.  Silenus then finishes it off and the total time is two hours longer than if they had drunk together.  But if they had drunk together, Bacchus would only have drunk half as much as he left for Silenus.

[Robert Chambers].  Arithmetic.  Theoretical and Practical.  New Edition.  Part of: Chambers's Educational Course -- edited by W. & R. Chambers.  William and Robert Chambers, London and Edinburgh, nd, [1870 written on fep].  [Though there is no author given, Wallis 242 CHA is the same item, attributed to Robert Chambers, with 1866 on the fly-leaf, so I will date this as 1866? --??check in BMC.]

P. 263, quest. 4.  Person doing business, equivalent to a cistern of  8000  with inlets of  1500  and  1000  per year and a drain of  3000  per year.  When is he broke?  This is not really a cistern problem since the rates are given rather than the times to fill or empty, but the format is sufficiently similar that I have included it here as an example of the more straightforward rate problems.  Also, the formulation with money is not common.

P. 266, quest. 44.  Cistern  (10, 8),  but the second pipe is not turned on until the cistern is half full.

Stoddard, John F.  The American Intellectual Arithmetic: Containing An Extensive Collection of Practical Questions on the General Principles of Arithmetic.  With Concise and Original Methods of Solution, Which Simplify Many of the Most Important Rules of Arithmetic.  Sheldon & Company, New York & Chicago, 1866.  P. 120, no. 30.  "If a wolf can eat a sheep in    of an hour, and a bear can eat it in  ¾  of an hour, how long would it take them together to eat what remained of a sheep after the wolf had been eating  ½  of an hour?"  Thanks to David E. Kullman for sending this.

Todhunter.  Algebra, 5th ed.  1870.  Many examples -- the less straightforward are the following.

Examples X, no. 33, pp. 87 & 577.  (C/3, 2C/3, C)  in 6.

Examples XXIV, no. 23, pp. 212 & 586.  (A, A-2) in 15/8.

Miscellaneous Examples, no. 48, pp. 548 & 604.  Simpson's XXXI with values  12, 15, 20.  Cf Docharty.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. V, 1884: 187-188.  Poetic and complicated form.  In  3/10  of the time that Silenus would take to drink the whole amphora, Bacchus drinks  1/4  of what he leaves for Silenus to finish.  But if they drank it all together, they would finish it in two hours less than the previous time.

Daniel W. Fish, ed.  The Progressive Higher Arithmetic, for Schools, Academies, and Mercantile Colleges.  Forming a Complete Treatise of Arithmetical Science, and its Commercial and Business Applications.  Ivison, Blakeman, Taylor & Co., NY, nd [but prefaces give: 1860; Improved Edition, 1875].  P. 419, no. 80.  Five persons building a house.  All but A can do in 14 days;  all but B can do in 19;  all but C can do in 12;  all but D can do in 15;  all but E can do in 13.  How long for all five and who is the fastest worker?  Answer:  11 4813/12137.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  Several straightforward problems and the following.  Prob. 99, p. 62.  Four companies of workers can do a job in  45, 9, 27, 36  days.  How long will it take  2/5  of the first company,  3/4  of the second company,  1/2 of the third company and  1/3  of the fourth company?

Fred Burnaby.  On Horseback Through Asia Minor.  Sampson Low, et al., London, 1877.  Vol. 1, pp. 208‑210.  One man can mow in  3  days,  the other in  4.  How long together?

Mittenzwey.  1880. 

Prob. 70-73, pp. 14 & 65-66;  1895?: 77-80, pp. 18-19 & 68;  1917: 77-80, pp. 17-18 & 64-65.  Cistern problems:  (6, 8);  (x, 20) in 12;  (9, 6, -4);  same with openings delayed by  2, 1, 3  hours.  His solutions indicate the general method.

Prob. 81, p. 15-16 & 66-67;  1895?: 89, pp. 20 & 69;  1917: 89, pp. 18-19 & 66.  Three drinkers,  (12, 10, 8).

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.

No. 106, pp. 142 & 332.  Asks for general solution of  (a, b)  and two specific cases.

No. 6, pp. 162 & 334.  (8, 9, 10).

No. 2, pp. 166 & 334.  General solution of  (a, b)  and one specific case.

William Briggs & George Hartley Bryan.  The Tutorial Algebra, based on the Algebra of Radhakrishnan -- Part II -- Advanced Course.  W. B. Clive, London, (1898), 1900.  Exercises X, prob. 6, pp. 123 & 579.  Two reapers,  (x, x‑5)  in  6.

A. Sonnenschein & H. A. Nesbit.  The New Science and Art of Arithmetic For the Use of Schools.  A. & C. Black, London, 1903.  This is a typical text of its time and has a number of variations on the basic problem.

Pp. 233-234 & 479, prob. 37.  Workers --  (10, 12, 9).  How long will it take to do    tasks -- how much does each person do?

Pp. 234 & 479, prob. 38.  Workers:  man, woman & boy can do in  (5, 8, 12).  Man works    days, then is joined by the woman for    days, then the remainder is left to the boy.  When does he finish and how much does each do?

Pp. 234 & 479, prob. 44.  Cistern of size  600  --  (5, 7).  How much does each pipe pass?

Pp. 234 & 479, prob. 45.  Workers --  (A, B)  in  6;  (A, C)  in  8;  (B, C)  in  9.

Pp. 235 & 479, prob. 54.  Cistern --  (4, 6, -4).

Pp. 288 & 484, prob. 42.  Two pumps and they can work at half or full speed.  If  (A, B/2)  in  5  and  (A/2, B)  in  4,  determine  A,  B  and  (A,B).

Pearson.  1907.  Part II, no. 162, pp. 145 & 223.  A  brings a pint every  3  minutes,  B  a quart every  5  minutes and  C  a gallon every  7  minutes.  How long to fill a  55  gallon drum and who finishes the job?

Stephen Leacock.  A, B, and C.  IN: Literary Lapses, (1910).  The book has been frequently reprinted and the piece has been widely anthologised.  It is pp. 237-245 in my 9th English ed.

Collins.  Fun with Figures.  1928.  According to Hoyle, not arithmetic, pp. 32-33.  "[I]f your father can build a chicken coop in  7  days and your Uncle George can build it in  9  days, how long ...."  "They'd never get it done;  they'd sit down and swap stories of rum runners, and bootleggers and hijackers."

C. Dudley Langford.  Note 1558:  A graphical method of solving problems on "Rate of Work" and similar problems.  MG 25 (No. 267) (Dec 1941) 304-307.  +  Note 2110:  Addition to Note 1558:  "Rate of Work" problems.  MG 34 (No. 307) (Feb 1950) 44.  Uses a graph to show  (a, b)  problems as meeting problems.  Also solves problems  (A, x)  in  B  and  (a, -b),  the latter appearing as an overtaking problem.  The Addition gives a clearer way of viewing  (a, ‑b)  problems as overtaking problems.

David Singmaster.  How long is a brick wall? (my title is: Three bricklayers).  Weekend Telegraph (30 May 1992) xxxii  &  (7 Jun 1992) xxx.  Al and Bill can build a wall in 12 days;  Al & Charlie in 15;  Bill & Charlie in 20.  How long does it take each individually and how long does it take all three together?  This is well known, but then I ask how can you determine integer data to make all the results come out integers?  Let  A,  B,  C  denote the amounts each can build in a day.  To make all the data and results come out as integers, we have to have all of  A,  B,  C,  A+B,  A+C,  B+C,  A+B+C  be fractions with unit numerators.  To combine them easily, imagine that all these fractions have been given a common denominator  d,  so we can consider  A = a/d,  B = b/d,  etc., and we want  a,  b,  c,  a+b,  a+c,  b+c,  a+b+c  to all divide  d.  We can achieve this easily by taking any three integers  a,  b,  c,  and letting  d  be the least common multiple of  a,  b,  c,  a+b,  a+c,  b+c,  a+b+c.  Taking  a, b, c = 3, 2, 1,  we find  d = 60  and the given problem is by far the simplest example with distinct rates  A,  B,  C. 

                    I feel this is based on my remembering the problem from somewhere, but the only previous use of this data is in Docharty, but he doesn't consider the diophantine problem, and none of the other data has this property.  I suspect this was an AMM or similar problem some years ago.

John Silvester recently asked me if I knew the following version, which he heard from John Reeve.  A man can pack his bag to go to a meeting in 20 minutes.  But if his wife helps him, it takes an hour.  How long would it take his wife on her own?  I.e.  (20, x) in 60.  The answer is -30 minutes!

 

          7.H.1.          WITH GROWTH  --  NEWTON'S CATTLE PROBLEM

 

          The example of Ray has led me to re-examine the relation between this topic and the cistern problems.  Ray's problem can be recast as follows.

                    There is a cistern with an input pipe and a number of equal outlet taps.  When  a  taps are turned on, the cistern empties in time  c;  when  d  taps are turned on, the cistern empties in time  f;  how long  [h]  will it take to empty when  x  taps are turned on?

Despite the similarity, we do not have the times for the individual inlets and outlets to fill or empty the cistern and so it is much easier to use rates rather than times.

 

Isaac Newton.  Arithmetica Universalis, 1707.  ??NYS.  English version:  Universal Arithmetic, translated by Mr. Ralphson, with revisions and additions by Mr. Cunn, Colin Maclaurin, James Maguire and Theaker Wilder; W. Johnston, London, 1769.  (De Morgan, in Rara, 652‑653, says there were Latin editions of 1722, 1732, 1761 and Raphson's English editions of 1720 and 1728, ??NYS.)  Resolution of Arithmetical Questions, Problem 11, pp. 189‑191.  (Sanford 165 quotes from the 1728 English edition and it is the same as the 1769.)  "If the Number of Oxen  a  eat up the Meadow  b  in the time  c;  and the Number of Oxen  d  eat up as good a Piece of Pasture  e  in the Time  f,  and the Grass grows uniformly;  to find how many Oxen  [x]  will eat up the like Pasture  g  in the Time  h."  Gives a general solution:  (gbdfh ‑ ecagh ‑ bdcgf + ecfga)/(befh - bceh)  and an example with  a, b, c;  d, e, f;  g, h  =  12, 3⅓, 4;  21, 10, 9;  24, 18.  One easily finds that the rate of grass growth per unit area per unit time is   G  =  (ace - bdf)/cf(bd - ae)   and the rate of grass consumption per ox per unit time is   E  =  be(c-f)/cf(bd - ae).   There are actually three unknowns since we also don't know the initial amount of grass per unit area,  G0.  We can either take proportions of these or we can adopt a unit of grass such that the initial amount of grass per unit area is  1.  In the second case, the basic equation   b (G0 + Gc)  =  aEc   becomes   b (1 + cG)  =  acE.

Walkingame.  Tutor's Assistant.  1751.  1777: p. 177, prob. 121;  1860: p. 185, prob. 114.  Identical to Newton.

Eadon.  Repository.  1794. 

Pp. 208-209, no. 6.  Same as Newton.  Gives a specific solution.  "This is deemed a curious and difficult question, it was first proposed by Sir Isaac Newton, in his Universal Arithmetic, and there universally solved by an algebraic process: but I have never seen a numerical solution independent of algebra, except this of my own; ...."  He then states the general solution as  [bdfg(h-c) + aceg(f-h)]/beh(f-c). 

Pp. 209-210, no. 7.  "If  3  oxen or  5  colts can eat up  4 1/5  acres of pasture in  7  weeks, and  5  oxen and  3  colts can eat up  9  acres of like pasture in  10  weeks, the grass gowing [sic] uniformly;  how many sheep will eat up  48  acres in  20  weeks, supposing  567  sheep to eat just as much as  6  oxen and  11  colts?"  Answer is  1736.  He introduces heifers, where a heifer is  1/5  of an ox or  1/3  of a colt, which brings the problem into Newton's form with values  a, b, c;  d, e, f;  g, h  =  15, 4 1/5, 7;  34, 9, 10;  48, 20,  which gives the answer in terms of heifers, which he converts to sheep.

Stoddard, John F.  Stoddard's Practical Arithmetic.  The Practical Arithmetic, Designed for the use of Schools and Academies; Embracing Every Variety of Practical Questions Appropriate to Written Arithmetic, with Original, Concise and Analytic Methods of Solution.  Sheldon & Company, NY, 1852.  Pp. 281-282, no. 23.  a, b, c;  d, e, f;  g, h  =  14, 2, 3;  16, 6, 9;  24, 6.  Thanks to David E. Kullman for sending this.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 6, p. 174 (1868: 185): Sir Isaac Newton's problem.  Gives the numerical values from Newton, but  3⅓  is often mis-set as  3½,  and when so done, is identical in all three editions.

Vinot.  1860.  Art. LI: Les boeufs de Newton, pp. 67-68.  a, b, c;  d, e, f;  g, h  =  3, 2, 2;  2, 2, 4;  6, 6.  Answer: 5.

A. Schuyler.  A Complete Algebra for Schools and Colleges.  Van Antwerp, Bragg & Co., Cincinnatti & NY.  Pp. 99-100, nos. 31-32, gives Newton's specific problem, then the general version.  Thanks to David E. Kullman for sending this.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  Same as Vinot, but no answer.

Joseph Ray, revised by J. M. Greenwood.  Ray's New Higher Arithmetic.  American Book Co., Cincinnati, 1880.  ??NYS -- quoted and discussed in:  David E. Kullman; Story problems with a flavor of the old Northwest; preprint of 8pp sent by the author in Apr 1999, p. 4.  "There is coal now on the dock, and coal is running on also, from a shoot [sic], at a uniform rate.  Six men can clear the dock in one hour, but 11 men can clear it in 20 minutes: how long would it take 4 men?  Ans: 5 hr."  This is like Newton's problem, but assuming  b = e = g,  and solving for  h  rather than  x. In a letter Kullman notes that this problem is not in Ray's original edition, called Higher Arithmetic, date not given, but Ray died in 1855.

G. H. Mapleton, proposer;  Charles Hammond, solver.  Arithmetical problem.  Knowledge 1 (30 Dec 1881) 191  &  (20 Jan 1882) 258, item 9.  a, b, c;  d, e, f;  g, h  =  12, 10, 16;  18, 10, 8;  40, 6.  Cites Newton, 1722 ed., p. 90.

Charles Pendlebury.  Arithmetic.  Bell, London, (1886), 30th corrected and expanded printing, 1924.  Section XLIV: Pasture with growing grass, pp. 336q-336s & answers, part II, p. xv.  (These pages were added after the 6th ed. of 1893, by the 10th ed. of 1897, possibly for the 10th ed?)  This section carefully works an example to answer different questions, then gives 12 problems.  In all cases, all the fields have the same size, which much simplifies things. 

The example problem has  a, b, c;  d, e, f;  g, h  =  70, b, 24;  30, b, 60;  b, 96. 

Prob. 12.  A cistern has a number of equal holes in its base and a pipe is adding water.  When 10 holes are open, the cistern will empty in 20 minutes.  When 8 are open, it empties in 35 minutes.  How long will it take with 12 holes open?  This is equivalent to  a, b, c;  d, e, f;  g, h  =  10, b, 20;  8, b, 35;  b, 12. 

I have read an article (in HM??) which described this as a common problem in 19C textbooks.

Walter Percy Workman.  The Tutorial Arithmetic.  University Tutorial Press, (1902);  2nd ed., 1902.  [There is a 3rd ed., (c1908), c1928, which contains a few more pages.]  Section IX [= Chap. XXXI in the 1928 ed.], examples CXLV, prob. 59‑60, pp. 430 & 544 [= 436 & 577 in the 1928 ed.]. 

Prob. 59.  a, b, c;  d, e, f;  g, h  =  15, b, 8;  9, b, 16;  b, 12  -- this has all the fields being the same.

Prob. 60 is more complex.  The field is divided into two equal halves.  50  oxen eat all of one half in  4  weeks.  Four oxen are slaughtered and the rest are put in the other half.  After  6  weeks,  3  oxen are slaughtered.  In another week, all the grass is eaten up and the oxen are sold.  After another week, the division is removed and  85  oxen are put in the field.  How long will it last them?

Wehman.  New Book of 200 Puzzles.  1908.  An ox problem, p. 55.  Gives Newton's problem with just the numerical values.

Loyd.  Cyclopedia, 1914, pp. 47 & 345.  = MPSL1, prob. 48, pp. 46 & 138‑139.  Cow, goat and goose.

Wood.  Oddities.  1927.  Prob. 71: Ox and grass, p. 55.  a, b, c;  d, e, f;  g, h  =  6, 10, 16;  18, 10, 8;  40, 6.   Gets  88,  but is confused about the growing of the grass -- Newton's formula gives  104.  Says Newton divides the oxen into those that eat the accumulated grass and those that eat the increase, but he doesn't apply this correctly.  Indeed, for this data, the grass grows at a negative rate!  Undoubtedly intended to be the data of Mapleton, 1881, for which the answer is  88.

Perelman.  MCBF.  1937.

Cows in the meadow.  Prob. 139, pp. 229-234.  Same as Pendlebury's example.

Newton's problem.  Prob. 140, pp. 234-235.  Same data as Newton.

A. I. Ostrovsky.  Oxen grazing in a field.  MG 50 (No. 371) (Feb 1966) 46‑48.  Quotes Newton and gives a graphical solution which converts this into an overtaking problem, where the grass starts growing  12  weeks before the cows are put in.

John Bull.  Grazing Oxen.  M500 165 (Dec 1998) 1-4.  He is unhappy with some of the limiting situations and proposes a different basic equation.  However, his unhappiness is really due to not understanding the basic equation properly.

 

          7.H.2.          DIVISION OF CASKS

 

          NOTATION:  (a, b, c)  among  n   means to divide  a  full,  b  half‑full and  c  empty casks among  n  people so that each has an equal amount of contents and of casks.

          Dividing  kn  casks containing  1, 2, ..., kn  among  k  people so each gets the same amount of contents and of casks  --  see:  Albert;  Munich 14684;  AR;  Günther (1887);  Singmaster (1998).

          See Tropfke 659.

 

Alcuin.  9C.

Prob. 12: Propositio de quodam patrefamilias et tribus filius ejus.  (10, 10, 10)  among  3.  This has  5  solutions -- he gives just 1:  0, 10, 0;  5, 0, 5;  5, 0, 5.

Prob. 51: Propositio de vino in vasculis a quodam patre distributo.  Divide  4  casks containing  10, 20, 30, 40  among  4  -- solution involves shifting contents, so this is not really a problem of the present type.

Abbot Albert.  c1240.  Prob. 3, p. 333.  Divide  9  casks containing  1, 2, ..., 9  among three.  He gives only one of the two solutions:  1, 5, 9;  2, 6, 7;  3, 4, 8.  See Singmaster, 1998, for a generalization.

BR.  c1305.  No. 40, pp. 58‑61.  300  ewes,  100  each with  1, 2, 3  lambs, to be divided among three sons so each son has the same number of ewes and lambs and no lamb is separated from its mother.  This is the same as  (100,100,100)  among  3.  He gives one solution:  0, 100, 0;  50, 0, 50;  50, 0, 50.  [There are  234  solutions!]

Munich 14684.  14C.  Prob. XXIV, f. 32r.  Same as Abbot Albert and with the same solution.

Folkerts.  Aufgabensammlungen.  13-15C.  10 sources of Abbot Albert's problem.  Also cites Albert, AR, Günther.

AR.  c1450.  Prob. 351, pp. 154, 182.  Same as Abbot Albert, with the same solution, but arranged in columns.  Vogel comments that this makes a 'half‑magic square' and cites Günther, 1887, as having already noted this.

Tartaglia.  General Trattato, 1556, art. 130‑131, p. 255v. 

Art. 130:  (7, 7, 7)  among  3.  Gives one of the two solutions:  3, 1, 3;  3, 1, 3;  1, 5, 1.

Art. 131:  (9, 9, 9)  among  3.  Gives one of the three solutions:  4, 1, 4;  3, 3, 3;  2, 5, 2.

Bachet.  Problemes.  1612.  Addl. prob. IX: Trois hommes ont à partager 21 tonneaux ...., 1612: 161‑164;  1624: 233-236;  1884: 168‑171.

(7, 7, 7)  among  3 -- gives all two solutions.  1612 cites Tartaglia. 

(9, 9, 9)  among  3 -- gives all three solutions.  1612 cites Tartaglia. 

(8, 8, 8)  among  4 -- Bachet erroneously does  (6, 12, 6)  among  4,  but the editor gives all four solutions of the original problem.

Labosne adds:  (5, 11, 8)  among  3  -- giving all three solutions.

          (Ahrens, A&N, 29, says that the 1st ed. also does  (5, 5, 5)  among  3.)

van Etten.  1624.  Prob. 89 (86), part IV, pp. 134‑135 (213).  (7, 7, 7)  among  3.  One solution:  3, 1, 3;  3, 1, 3;  1, 5, 1.

Hunt.  1651.  Pp. 284-285.  Of three men that bought wine.  (7, 7, 7)  among  3.  Two answers.

Ozanam.  1725.  Prob. 44, 1725: 242‑246.  Prob. 24, 1778: 182-184;  1803: 180-182;  1803: 158-159;  1814: 158-159;  1840: 81-82. 

(7, 7, 7)  among  3  -- gives both solutions.  Notes that this cannot be divided among  4  persons because  4  does not divide  21.

1725 has a very confusing attempt at  (8, 8, 8)  among  4,  which is done as though it were  (6, 12, 6)  among  4,  and he seems to think half‑empty is different than half‑full!!  1778 onwards just do  (8, 8, 8)  among  3,  giving  3  of the  4  solutions, omitting   4, 0, 4;   4, 0, 4;   0, 8, 0.

(9, 9, 9)  among  3  -- gives all  3  solutions.

Les Amusemens.  1749.  Prob. 15, p. 137: Les Tonneaux. 

(7, 7, 7)  among  3  -- gives both solutions. 

(11, 11, 11)  among  3  -- gives  2  of the  4  solutions.

Bestelmeier.  1801.  Item 717: Die sonderbare Weintheilung unter 3 Erben.  Says there are  21  casks, so presumably  (7, 7, 7)  among  3.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 11, pp. 3 & 53.  (7, 7, 7)  among  3.  Both solutions.

No. 51, pp. 11 & 66.  (8, 8, 8)  among  3.  Gets three of the four solutions, omitting  4, 0, 4;  4, 0, 4;  0, 8, 0.

Endless Amusement II.  1826?  Prob. 20, pp. 199-200.  (7, 7, 7)  -- two solutions.

Young Man's Book.  1839.  Pp. 239-240.  Identical to Endless Amusement II.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 211.  (7, 7, 7)  with both solutions.

Magician's Own Book.  1857.  The wine and the tables, p. 225.  (7, 7, 7),  (8, 8, 8),  (9, 9, 9)  among  3  -- gives two solutions for each.  = Boy's Own Conjuring Book, 1860, p. 195. 

Vinot.  1860.  Art. XL: Un partage curieux, pp. 59-60.  (7, 7, 7)  -- two solutions.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 593, pp. 299 & 410: Sechs Knacknüsse -- part 1.  (7, 7, 7)  -- one solution shown diagrammatically:    3, 1, 3;  3, 1, 3;  1, 5, 1.

Mittenzwey.  1880.  Prob. 104, pp. 21 & 73;  1895?: 121, pp. 26 & 75;  1917: 121, pp. 24 & 73.  (7, 7, 7)  among 3.  Gives three solutions, but one is a rearrangement of another.  Solution asks, if pouring is allowed, can all three get the same inheritance?  It says to pour two half barrels into two other half barrels, obtaining  (9, 3, 9).

Siegmund Günther.  Geschichte des mathematischen Unterrichts im deutschen Mittelalter bis zum Jahre 1525.  Monumenta Germaniae Paedagogica III.  1887.  Facsimile reprint by Sändig Reprint Verlag, Vaduz, Liechtenstein, 1969.  He discusses Abbot Albert and his problem on pp. 35-36, noting that the solution can be viewed as a set of lines in a magic square so that the perpendicular lines give a second solution, but that magic squares were then unknown in Europe.  He gives no other examples.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XV: Jeux de tonneau, pp. 50-51.  (7, 7, 7)  among  3.  Gives both solutions.  Notes that  half empty = half full,  so doubling gives us  empty = full!

Ahrens.  A&N.  1918.  Pp. 29‑33.  Gives all solutions for  (n, n, n)  among  3  for

          n  =  5 (1) 10;  (8, 8, 8)  among  4  &  6;  (11, 5, 8)  among  3; 

          (5, 11, 8)  among  3;  (4, 12, 8) among  3.

McKay.  Party Night.  1940.  No. 14, p. 178.  (7, 7, 7)  among three.  One solution.

M. Kraitchik.  Mathematical Recreations, 1943, op. cit. in 4.A.2, chap. 2, prob. 34, pp. 31‑32.  9  full,  9  three-quarter full,  9  half,  9  quarter full,  9  empty  among  5.

David Singmaster.  Triangles with integer sides and sharing barrels.  CMJ 21 (1990) 278‑285.  Shows the number of ways of sharing  (N, N, N)  among  3  is the same as the number of integer sided triangles of perimeter  N.  Shows this is the number of partitions of  N/2  or  (N‑3)/2  into  3  parts.  Finds necessary and sufficient conditions for sharing  (a, b, c)  among  k  people.

David Singmaster.  Fair division of the first  kn  integers into  k  parts.  Written in 1998.  This generalizes the problem of Abbot Albert and determines that there is a partition of the first  kn  integers into  k  sets of  n  values with each set having the same sum if and only if  n  is even  or  n > 1  and  k  is odd.  I call this a fair division of the first  kn  integers into  k  parts.  The number of such divisions is large, but might be worth examining.

 

          7.H.3.          SHARING UNEQUAL RESOURCES  --  PROBLEM OF THE PANDECTS

 

            NOTATION:  (a, b, ...; c)  means persons contribute  a, b, ...   which is shared equally among themselves and an extra person who pays  c.  See Clark for a variant formulation with the same result.

            See McKay; Party Night; 1940 in 7.H.5 for a related form.

 

            INDEX  of  (a, b, ...; c)  problems with  a £ b £ ....

 

  2            3          3            Kraitchik

  2            3          5            Fibonacci;  al-Qazwînî;  Bartoli;  Calandri

  2            3        35            McKay

  3            4          4            Kraitchik

  3            4          7            Kraitchik

  3            4        14            Gherardi

  3            5          4            Vinot

  3            5          5            Benedetto da Firenze

  3            5          8            D. Adams, 1801;  Jackson;  Badcock;  New Sphinx;  Magician's Own Book;  Bachet‑Labosne;  Mr. X;  Pearson;  Clark;  Kraitchik;  Rohrbough;  Sullivan

  3            5        10            AR

  3            5        80            Kraitchik

  4            5          4            Pseudo-dell'Abbaco

  4            6        10            Mittenzwey

  5            7        12            Kraitchik

  7            8        30            Kraitchik

10          14          6            Hummerston

31          50        40            Tagliente

90         120        70            Kraitchik

 

11          14        17        42            Kraitchik

 

  2            3          6          9          4            Kraitchik

 

Fibonacci.  1202.  P. 283 (S: 403-404).  (3, 2; 5).

Qazwini  = Zakariyâ ibn Muhammad ibn Mahmûd [the  h  should have an underdot] abû Yahya [the  h  should have an underdot] al-Qazwînî.  (= al-Kazwînî [the  K  should have an underdot]  = Zakariyyā' b. Muhammad b. Mahmūd [the  h  should have an underdot] Abū Yahya [the  h  should have an underdot] al-Kazwīnī [the  K  should have an underdot]  =  Zakarīyā ibn Muhammad [the  h  should have an underdot] al-Qazwīnī).   (Kitâb) ‘Ajâ’ib al‑Makhlûqât wa Gharâ’ib al-Mawjûdât (= Adjāyib al-Makhlūkāţ [NOTE:  ţ  denotes a  t  with an underdot and the second  k  should hve an underdot.] wa Ghārā'ib al-Mawdjūdāţ [NOTE:  ţ  denotes a  t  with an underdot.]  =  ‘Ajā’ib al‑makhlūqāt wa-gharā’ib al-mawjūdāt) ((The Book of the) Wonders of the Creation and Unique [Phenomena] of the Existence  = Prodigies of Things Created and Miraculous Aspects of Things Existing  = The Wonders of Creation and the Peculiarities of Existing Things  = The Cosmography).  c1260.  ??NYS -- the earliest dated copy, of 1458, and several others are in the Wellcome Institute; BL has a page from a 14C copy on display.  Part 8: On the arts;  chap. 9: On reckoning.  In:  J. Ruska; Kazwīnīstudien[the  K  should have an underdot];  Der Islam 4 (1913) 14‑66 & 236‑262.  German translation (Arabic omitted) of this problem on pp. 252‑253.  (3, 2; 5).  Story says one proposes a  3 : 2  split, but  4 : 1  is found to be correct.  [Qazwini also wrote a Geography, in two editions, and its titles are slightly similar to the above.  I previously had reference to the Arabic titles of the other book, but rereading of Ruska and reference to the DSB article shows the above is correct.]

Gherardi.  Libro di ragioni.  1328.  Pp. 40‑41: Chopagnia.  (3, 4; 14).

Bartoli.  Memoriale.  c1420.  Prob. 26, ff. 77v-78r (= Sesiano, pp. 144 & 149).  (2, 3; 5),  correctly solved.

Pseudo-dell'Abbaco.  c1440.  Prob. 94, p. 81 with plate on p. 82.  (5, 4; 5).  I have a colour slide of this.

AR.  c1450.  Prob. 212, p. 98.  (5, 3; 10)  correctly solved.  (Unusually, Vogel's notes, pp. 160‑161 & 211‑213, say nothing about this problem.)

Benedetto da Firenze.  c1465.  c1480.  P. 106.  Part of the text is lacking, but it must be  (3, 5; 5).

Calandri.  Arimethrica.  1491.  F. 63v.  (2, 3; 5).

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 130, ff. 60v-61r.  Philippo and Jacomo share lunch with Constanzo --  (50, 31; 40).

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 27, ff. 66r-66v.  Three men have  3, 2, 1  loaves of bread and other foods worth  8, 6, 4.  A fourth comes and shares with them, paying  9.  How much should each of the three get?  The total value of the food is  4 x 9 = 36,  so the bread is worth  36 - 8 - 6 - 4 = 18,  or  3 per loaf.  So the first should get  9,  the second  3  and the third owes them  2 !!

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 16, ff. CC.iv.v - CC.v.v (pp. 140-141).  Three men with bread, wine and fish in the amounts:  3, 4, 0;  0, 5, 6;  2, 0, 7;   which are considered of equal value.  A fourth man with one bread comes and they share the meal and the fourth man pays  5.

Tartaglia.  General Trattato.  1556.  Book 12, art. 33, pp. 199r-199v.  Three men have quails and bread.  The first has  6  quails and  2s  worth of bread;  second has  4  quails and  3s  worth of bread;  third has  2  quails and  5s  worth of bread.  They share with a fourth person who pays  8s.

Buteo.  Logistica.  1559. 

Prob. 3, pp. 201-202.  Four share their food.  First has  4  breads and  20  carrots;  second has  1  bread and  32p  of wine;  third has  7  breads and  8  carrots;  fourth has a cheese.  If all have equal value, what is the value of each item?

Prob. 4, pp. 202-203.  Three share their food with a fourth.  First has  2  breads and  7  nummos worth of fish;  second has  4  breads and  5  nummos worth of condiments;  third has  1  bread and  8  nummmos worth of wine.  Fourth pays  12  nummos for his share.

D. Adams.  Scholar's Arithmetic.  1801.  P. 210, no. 6.  (3, 5; 8).  Doesn't give any attempts at division, nor a solution.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles, no. 12, pp. 3-4 & 53-54.  (3, 5; 8).  Says it appears in an Arabian manuscript.  The man with five loaves divides  5, 3;  the other protests and divides  4, 4;  judge divides  7, 1.  Why?  = Magician's Own Book (UK version), 1871, Arithmetical paradox, pp. 28-29.

John Badcock.  Domestic Amusements, or Philosophical Recreations.  Op. cit. in 6.BH.  [1823].  Pp. 186-187, no. 255: Arithmetical paradox.  Same as Jackson, saying it appears in an Arabic manuscript.

The New Sphinx.  c1840.  (3, 5; 8).  Three Greeks.  Extra person divides his money  3, 5,  but the second was dissatisfied and had the matter referred to Solon, who gave the right division.

Magician's Own Book.  1857.  The three travellers, pp. 225-226.  (3, 5; 8).  = Boy's Own Conjuring Book, 1860, pp. 195‑196.

Vinot.  1860.  Art. LX: Chacun son écot, p. 77.  (3, 5; 4).

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Several ordinary examples and some unusual examples.

1863 -- p. 146, no. 17;  1873 -- p. 154, no. 15.  (6, 10; 16)  but the third person eats  4  more than each of the other two eat.

1863 -- p. 146, no. 19;  1873 -- p. 155, no. 17.  (5, 9; 24)  but  C  eats twice as much as  B,  who eats twice as much as  A.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. I: Deux Arabes allaient dîner: ..., 1884: 181.  (3, 5; 8).

Mittenzwey.  1880.  Prob. 61, pp. 12-13 & 64;  1895?: 67, pp. 17 & 66;  1917: 67, pp. 16 & 63.  (6, 4; 10).  One suggest dividing  6, 4,  the other suggests  5, 5.  Gives correct solution.

Hoffmann.  1893.  Chap IV, no. 74: The three Arabs, pp. 165 & 220 = Hoffmann-Hordern, p. 147.  (3, 5; 8).  First says to divide  3, 5;  second says  4, 4;  third says both are wrong.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:6 (Oct 1903) 530-531.  The stranger's dinner.  (3, 5; 8).  Refers to Arabians.  Similar to Jackson.

Pearson.  1907.  Part II, no. 138, pp. 141 & 218.  (3, 5; 8).

Ball‑FitzPatrick.  Footnote cited in 7.G.1.  1908.  Says the problem is Arabic, but gives no reason.

Clark.  Mental Nuts.  1916, no. 22.  Real estate.  A invests  $5000,  B invests  $3000.  They buy three houses of equal value.  They each take one and then sell the third for  $8000.  How do they divide the money?  Answer is  $7000  and  $1000.  His  1904, no. 25;  1916, no. 39  is a standard version of  (3, 5; 8)  with sandwiches.

Hummerston.  Fun, Mirth & Mystery.  1924.  A partnership problem, Puzzle no. 3, pp. 19 & 172.  (10, 14; 6),  then the second uses his receipts to buy more which are shared equally and his colleagues pay him -- how many more could he now buy?

Kraitchik.  La Mathématiques des Jeux.  Op. cit. in 4.A.2.  1930.  Chap. 1, pp. 7‑8: Problème des Pandects.  He gives several examples and says they come from Unterrichtsblätter für Mathematik und Naturwissenschaften 11, pp. 81‑85, ??NYS.

No. 22:  (3, 5; 8).

No. 23:  (2, 3; 3).

No. 24:  (11, 14, 17; 42).

No. 25:  (5, 7; 12).

No. 26:  (7, 8; 30).  Caius and Sempronius share  7  and  8  with Titus who paid them  14  and  16.  Sempronius protests and a judge divides it as  12  and  18.  "C'est probablement cette version qui a donné à ce problème le nom de celui de Pandectes."

No. 27:  (3, 5; 80).

No. 28:  (3, 4; 7)  -- cultivating fields.

No. 29:  (90, 120; 70)  -- digging a ditch.

No. 30:  (3, 4; 4).

No. 31:  (2, 3, 6, 9; 4)  -- heating a workshop.

Rohrbough.  Brain Resters and Testers.  c1935.  The Travelers' Dinner, pp. 23-24.  Arabs,  (3, 5,; 8).

McKay.  Party Night.  1940.  No. 25, p. 182.   A  &  B  give a party and invite  2  and  3  guests.  The party costs  35s  -- how do they divide the expense?  Initial reaction is in the ratio  2 : 3,  but it should be  3 : 4.  (Also entered in 7.H.5)

Kraitchik.  Mathematical Recreations.  Op. cit. in 4.A.2.  1943.  The problem of the Pandects, pp. 28‑29.  c= No. 26 of Math. des Jeux.

Sullivan.  Unusual.  1943.  Prob. 19: An Arab picnic.  (3, 5; 8).

 

          7.H.4.          EACH DOUBLES OTHERS' MONEY TO MAKE ALL EQUAL, ETC.

 

          See 7.R for related problems.

          See Tropfke 647‑648.

 

Diophantos.  Arithmetica.  c250.  Book I.

No. 22, p. 138.  "To find three numbers such that, if each give to the next following a given fraction of itself, in order, the results after each has given and taken may be equal."  Illustrates with fractions  1/3, 1/4, 1/5. 

No. 23, pp. 138‑139, is the same with four numbers, illustrated with fractions  1/3, 1/4, 1/5, 1/6.

Mahavira.  850.  Chap. VI, v. 259‑267, pp. 160‑162.

260:  3  men, each doubles the next to make all equal.

262:  same with  5  men.

263:  each  3/2's  the next.  Also each  5/3's  the next.

266:  each doubles others.  Also each  3/2's  others.

Fibonacci.  1202.  Pp. 287‑293 (S: 409-415) gives several versions.

Pp. 287‑288 (S: 409-410).  w  5/2's  x;  x  7/3's  y;  y  9/4's  z;  z  11/5's  x  and then all are equal.  Answer:  5862, 2858, 2760, 2380  and he notes that these can be multiplied by any value.

Pp. 288‑291 (S: 410-413).  Same operations, but the results are in the proportion  5 : 4 : 3 : 2.  Answer:  22875, 10000, 8355, 7280,  and he also divides these by 5.

Pp. 291‑292 (S: 413-414).  w  5/2's  all the others, etc. with the above ratios, then all are equal.  Answer:  8436, 3288, 1440, 696.

Pp. 292‑293 (S: 414-415).  Same as the previous, but with results in proportion  5 : 4 : 3 : 2.  Answer:  29706, 11568, 498, 2256. (He erroneously has  29826  for the first value.)

Fibonacci.  Flos.  c1225.  In Picutti, pp. 326-332, numbers VIII-X.

Pp. 243-244: De quatuor hominibus qui invenerunt bizantios.  First man doubles the second's money,  then the second man triples the third's,  ...,  to make all equal.  Answer:  89, 77, 47, 27  with total of  240.

Pp. 244-245: Above continued.  First doubles the others,  second triples the others,  ...,  to make all equal.  Answer:  241, 161, 61, 17  with total  480.

Pp. 246-247: Questio similis suprascripte de tribus hominibus.  First  5/2's  the others,  second  10/3's  the others,  third  17/4's  the others to make all equal.  Answer:  1554, 738, 258  with total  2550  and he then divides through by  6.

Lucca 1754.  c1330.  Ff. 61v‑62r, p. 142.  4 & 3 men.  Each doubles the others to make all equal.  In the 4 man case, he specifies the total money is 400.

Giovanni di Bartolo.  Op. cit. in 7.H.  c1400.  Prob. 9, pp. 17‑18.  Four men, each doubles the others' money and the product of the results is 1000.  He assumes, for no clear reason, that the original amounts are proportional to  8 : 4 : 2 : 1.

AR.  c1450.  Prob. 231, pp. 107‑108 & 169‑171.  Two men, each doubles the others' money and then both have  13½.

Muscarello.  1478.  Ff. 78r-78v, pp. 193-194.  Four men, each doubles the others' money, then all are equal.  Answer:  33, 17, 9, 5.

Chuquet.  1484.  Prob. 148.  3 people.  Mentioned in passing on FHM 230.

Calandri.  Aritmetica.  c1485.  Ff. 101r-102r, pp. 202‑204.  6 men.  Each doubles the others to make all equal.

Pacioli.  Summa.  1494. 

F. 105r, prob. 16.  First wins  1/2  of the second's;  second wins  1/3  of the third's;  third wins  1/4  of the first's to make all equal  100.  I get  (200, 400, 300)/3.  See Tonstall for corrections and intended interpretation.  Pacioli gives no working and just states answers that are printed differently in the two editions, but from Tonstall we see that they are intended to be:  55 5/9 (given as  44 4/9  and as  144 4/9),  111 1/9,  133 1/3.

F. 189r, prob. 6.  First gives  7/12  of his money to the second, who then gives  11/30  of his money to the first, when both are equal.  He gives no working and just one answer:  70/9,  658/95.  I find the general answer is  47x = 48y.

Tonstall.  De Arte Supputandi.  1522.  P. 245.  First wins  1/2  of the second's;  second wins  1/3  of the third's;  third wins  1/5  of the first's to make all equal  100.  This is a correct version of Pacioli.  He gives  (500, 1000, 1200)/9  and shows the calculation which implies all three calculations are done at once -- that is, the  1/5  of the first's money is based on what he had to start, not what he has after winning from the second.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 91, ff. GG.viii.v - HH.i.v (pp. 165‑166).  A situation somewhat similar to 7.P.7, where the first two take money leaving the third with  5.  Friend says the first is to give  10  and    of what he has left to the second and the second is then to give  7  and  ¼  of what he has left to the third to make their amounts proportional to  (3, 2, 1).  Answer is  x = 172,  y = 39  and the total sum is  216.

Tartaglia.  General Trattato, 1556, art. 11, 18, 37, 38, pp. 240v, 241v, 244v‑245r.  2 and 3 people versions.

Buteo.  Logistica.  1559.  Prob. 63, pp. 270-273.  Three players -- first wins  1/2  of second's,  second wins  1/3  of third's,  third wins  1/4  of what the first had originally,  and all wind up with  100.  That is, we have  x + y/2 ‑ x/4  =  y ‑ y/2 + z/3  =  z ‑ z/3 + x/4  =  100.

Bachet.  Problemes.  1612.  Addl. prob. VIII, 1612: 154-160;  1624: 226‑233;  1884: 162‑167.  3  people;  also with tripling.  Labosne adds the general case.

van Etten.  1624.  Prob. 57 (52), pp. 52‑53 (78).  3 people version used as a kind of divination.

Ozanam.  1694.  Prob. 26, 1696: 81;  1708: 72.  Prob. 47, 1725: 253.  Prob. 17, 1778: 209;  1803: 204.  Prob. 16, 1814: 177;  1840: 91.  Three person version, resulting in each having 8, used as a kind of divination.

Euler.  Algebra.  1770.  I.IV,IV.616: Question 4, pp. 211‑212.  Three players, all winding up with 24.

Hutton.  A Course of Mathematics.  1798?  Prob. 42,  1833: 223;  1857: 227.  Five players, all ending up with £32.

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Section 82, ex. 9, p. 123.  Three players, each doubles others, all ending with  120.

Thomas Grainger Hall.  The Elements of Algebra: Chiefly Intended for Schools, and the Junior Classes in Colleges.  Second Edition: Altered and Enlarged.  John W. Parker, London, 1846.  Pp. 130-131, ex. 16.  Each  3/2's  others, with total given as 162.  Cf Mahavira 266.

G. Ainsworth & J. Yeats.  A Treatise on the Elements of Algebra.  H. Ingram, London, 1854.  Exercise XXXVIII, pp. 81‑83 & 178.

No. 14: A  doubles  B,  B  doubles  A,  A  doubles  B,  to make both have  80.  How much originally?  Answer is  100, 50  instead of  110, 50.

No. 15: same as no. 14, with one more stage to make both equal  n.

No. 28: usual problem with four gamblers, ending with  64.

No. 29: usual problem with  7  baskets of apples, all ending with  128.

No. 30: usual problem with  n  persons, all ending with  a.  Solution is badly misprinted -- the  i-th should start with an amount  a(2n-i + 1)/n×2n.

Vinot.  1860.  Art. LXXI: Les trois joueurs, pp. 86-87.  Three person version, all ending up with 24.

Boy's [Own] Magazine 2:5 (No. 11) (Nov 1863) 459 [answer would be in Jan 1864, ??NYS].  (Reprinted as (Beeton's) Boy's Own Magazine 3:11 (Nov 1889) 479.   Mathematical question 137.  Three boys playing.  Each pays the winner half of what he has.  Each one wins once and then they have  30d, 60d, 120d.  How much did they have to start?  [In fact, they had the same amounts.  In general, if they start with  1, 2, 4,  then the amounts after the first, second, third wins are:  4, 1, 2;  2, 4, 1;  1, 2, 4.]

Todhunter.  Algebra, 5th ed.  1870.  Examples XIII, no. 26, pp. 104 & 578.  3  men, each doubles the others to make all equal to  16. 

Mittenzwey.  1880. 

Prob. 113, pp. 23 & 76;  1895?: 131, pp. 27 & 78;  1917: 131, pp. 25 & 75.  = Ainsworth & Yeats, no. 14, except he doesn't say what the final result is.  Solution is  110, 50.

Prob. 126, pp. 26 & 76;  1895?: 144, pp. 30 & 79;  1917: 144, pp. 27 & 77.  Four gamblers, each doubles the others, winding up with  64. 

1895?: Prob. 76, pp. 18 & 68;  1917: 76, pp. 17 & 64.  Three piles, each is used to double the next, making all have  8.  1895? just states the answer; 1917 sets up and solves the equations.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XLIII: Les jouers, pp. 184-185.  Three players winding up with 12 each.  Does general solution for  n  players winding up with  a  each.

Workman.  Op. cit. in 7.H.1.  1902.  Section IX (= Chap. XXXI in c1928 ed.), examples CXLV, prob. 45, pp. 428 & 544 (434 & 577 in c1928 ed.).  Version with five people.

Hermann Schubert.  Beispiel‑Sammlung zur Arithmetik und Algebra.  3rd ed., Göschen, Berlin, 1913.  Section 17, no. 107, pp. 66 & 140.  Three people.  Each gives away half his money to be equally shared by the others, and then they all have 8.  Solution:  4, 7, 13.

Loyd.  Cyclopedia.  1914.  Sam Loyd's Mystery Puzzle, pp. 226 & 369.  (= MPSL2, prob. 85, pp. 61 & 148.  = SLAHP: An initiation fee, pp. 63 & 109.)  3 people version, resulting in the first person having lost 100.

Collins.  Book of Puzzles.  1927.  The five gamblers puzzle, p. 75.  They all end up with $32.

 

          7.H.5.          SHARING COST OF STAIRS, ETC.

 

            See Tropfke 529.  The simplest form is given by Sridhara, BR, Pseudo-dell'Abbaco, Gori, von Schinnern.

 

Mahavira.  850.  Chap. VI, v. 226‑232, pp. 151‑153.

227.  Porter carrying  32  jack‑fruits over distance  1  will receive    of the fruits.  He breaks down at distance  ½.  How much is he due?  Rule says  x/32*1 = (7½ ‑ x)/(32 ‑ x)(1 ‑ ½),  i.e. the wages per fruit‑mile should be the same for both parts of the journey.  (Properly, this problem leads to an exponential.)

229.  Porter carrying  24  jack‑fruits for distance  5  will earn  9  of them.  Two porters share the work, the first earns  6  and the second earns  3.  How far did the first one carry the fruits?

232.  Twenty men are to carry a palanquin a distance of  2  for wages of  720.  But two men drop off after distance  ½;  three more drop off after another  ½  and five more drop off after half the remaining distance.  How much does each earn?  He says each quarter of the distance is worth  180  and divides this equally among the carriers for each quarter.

Sridhara.  c900. 

V. 67(ii), ex. 84‑85, pp. 53 & 95.  Porter carrying  200  palas of oil for  5  panas wages.  But the bottle leaks and only  20  palas remain at the end.  How much should he be paid?  Rule says to pay  (20 + 180/2)/200  of the wages.

V. 68, ex. 86‑90, pp. 53‑55 & 95.

  Ex. 86‑87.  Four men watch a dance for  ¼, ½, ¾  and all of a day.  The dancers' fee is  96.  How much should each pay?  He charges  24/4  per watcher for the first quarter,  24/3  per watcher for the second quarter,  ..., giving payments  6, 14, 26, 50.

  Ex. 88.  Ten men are to carry a palanquin a distance of  3  for wages  100.  Two men drop off after distance  1  and another three after a total distance of  2.  How much does each earn?  Divides as in Mahavira's 232.

  Ex. 89.  Five chanters perform  1, 2, 3, 4, 5  chants for a fee of  300.  How much does each earn?  Each chant earns  60,  divided among the chanters of it.

V. 70‑71, ex. 92‑94, pp. 56‑58 & 96.

  Ex. 92.  Porter carrying  24  jack‑fruits for distance  5  will earn  9.  What does he earn for carrying distance  2?

  Ex. 93‑94.  Porter carrying  24  jack‑fruits for distance  5  will earn  9  of them.  Two porters split the carrying, the first earns  4  and the second earns  5.  How far did they carry?  Solutions of these are based on the rule in Mahavira 227.

BR.  c1305.  No. 36, pp. 54‑57.  Divide oil among  12  lamps which are to burn  1, 2, ..., 12  hours.

Pseudo-dell'Abbaco.  c1440.  Prob. 92, pp. 78‑81.  House rented to  1  person the first month,  who shares with a  2nd the  2nd month,  who share with a  3rd the  3rd month,  ...,  who share with a  12th the  12th month.  How much does each pay?  He says many obtain  12/78,  11/78,  ...,  1/78,  but that it is more correct if the first pays  (1 + 1/2 + 1/3 + ... + 1/12) * 1/12,  the second pays  (1/2 + 1/3 + ... + 1/12) * 1/12,  ...,  the  12th pays  1/12 * 1/12.

Gori.  Libro di arimetricha.  1571.  F. 73r (p. 80).  Same as Prob. 92 of Pseudo-dell'Abbaco, but with only the first solution.

Bullen.  Op. cit. in 7.G.1.  1789.  Chap. 38, prob. 32, p. 243.  Four men hire a coach to go  130  miles (misprinted as  100).  After  40  miles, two more men join.  How much does each pay?

Clemens Rudolph Ritter von Schinnern.  Ein Dutzend mathematischer Betrachtungen.  Geistinger, Vienna, 1826, pp. 14‑16.  Discusses general problem of sharing a cost of  n  for lighting a  x  floor staircase.  Does the case  n = 48,  x = 4,  getting  3, 7, 13, 25,  which are the same proportions as Sridhara, ex. 86‑87.

Dana P. Colburn.  Arithmetic and Its Applications.  H, Cowperthwait & Co., Philadelphia, 1856.  Miscellaneous Examples No. 66, p. 365.  A and B hire a horse and carriage for $7 to go 42 miles and return.  After 12 miles, C joins them, and after 24 miles, D joins them.  How much should each pay?  No solution is given, but there is a Note after the question saying there are two common ways to do the allocation.  First is in proportion to the miles travelled, so  A : B : C : D = 42 : 42 : 30 : 18.  Second is to divide the cost of each section among the number of riders, which here gives  A : B : C : D = 29 : 29 : 17 : 9

Clark.  Mental Nuts.  1897, no. 95;  1904, no. 71;  1916, no. 14.  The livery team.  I hire a livery team for $4 to go to the next city, 12 miles away.  At the crossroads 6 miles away, I pick up a rider to the city who then rides back to the crossroads.  How much should he pay?  Answer is $1.

M. Adams.  Puzzle Book.  1939.  Prob. C.135: Answer quickly!, pp. 158 & 187.  Man hires car to go to theatre.  He picks up and drops off a friend who lives half way to the theatre.  How do they divide the fare?  Answer is  3 : 1.

Depew.  Cokesbury Game Book.  1939.  Passenger, p. 219.  Similar to M. Adams.

McKay.  At Home Tonight.  1940.  Prob. 13: Sharing the cost, pp. 65 & 79.  Similar to M. Adams.

McKay.  Party Night.  1940.  No. 25, p. 182.   A  &  B  give a party and invite  2  and  3  guests.  The party costs  35s  -- how do they divided the expense?  Initial reaction is in the ratio  2 : 3,  but it should be  3 : 4.  (Also entered in 7.H.3.)

Doubleday - 3.  1972.  Fair's fair, pp. 61-62.  Man hires a taxi to go to the city and pays in advance.  Halfway there, he picks up a friend.  Later they go back, with the friend dropped at the halfway point.  How should they share the fare?  Answer says the friend should pay one third, because the man 'had only hired the taxi to take him into town ... not for a round-journey.'  This may be introducing the following extra feature.  Normally the friend would pay  1/4  of the total cost.  But when the taxi has got halfway back, the fare shown will only be  3/4  of the total cost and the friend should pay  1/3  of that amount.  The problem does say that the same driver has been used and there was no charge for the waiting time.  But I wonder whether taxi-meters can be stopped and restarted in this way.  It would be more natural if they caught another taxi back.  Then at the halfway point, the fare shown would be  1/4  of the total cost and the friend show pay all of the amount shown!

 

          7.H.6.          SHARING A GRINDSTONE

 

            New section.  I have seen other examples.  A grindstone of radius  R  is to be shared between two (or  k)  buyers -- one grinding until his share is used.  An inner circle of radius  r  is unusable.  The first man should grind to radius  x  where  x2 - r2  =  (R2 - r2)/k   or  x2  =  [(k-1)R2 + r2]/k.  It is straightforward to adapt this to the case when the buyers contribute unequal amounts to the purchase price.

            Clark gives a problem of sawing through a tree which uses the fact that the area of a segment of a circle of radius  R  and segment height  H  is  R2cos‑1(R‑H)/R ‑ (R‑H)Ö(2RH‑H2).  Though well-known, this seems about on the border of what I consider to be recreational.

 

Anonymous proposer; solution lacking.  Ladies' Diary, 1709-10  =  T. Leybourn, I: 5-6, quest. 9.  [??NX of p. 6.]  Share a grindstone among seven people.  2R = 60". 

Carlile.  Collection.  1793.  Prob. XXIV, p. 16.  Three men buy a grindstone of radius  20  for  20s.  They pay  9s,  6s,  5s  respectively.  How much should each man get to grind?  He makes no allowance for wastage.

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 22,  1857: 558.  Share a grindstone among seven people.  2R = 60".  He takes  r = 0.

Dana P. Colburn.  Arithmetic and Its Applications.  H, Cowperthwait & Co., Philadelphia, 1856.  Miscellaneous Examples No. 43, p. 362.  Four men sharing a grindstone 4 ft in diameter.  No indication of inner wastage.  No solution given.

Clark.  Mental Nuts.  1904, no. 93. Cutting trees.  Three men cutting through a tree of diameter three.  One cuts in one from one side, the next cuts in one from the opposite side.  How much is left?  Answer is  41.64%,  which is correct.  [The obvious question is how far should the first two men get so that all three cut an equal area?  I find it should be  .7350679153972  of the radius.]

Collins.  Fun with Figures.  1928.  A grindstone dispute, p. 22.  2R = 5 ft 6 in;  2r = 18 in.

 

          7.H.7.          DIGGING PART OF A WELL.

 

            Recently separated from 7.H.5.

            See Tropfke 529.

            The 'digging a well' problem has a contract to dig a well  a  deep for payment  b,  but the digging stops at  c.  How much should be paid?  The value of  b  is not always given and then only the ratio of the part payment to the total payment is sought.

            NOTATION -- this problem is denoted  (a, b; c).

If the difficulty is proportional to depth, then integration yields that the payment should be proportional to  (a/c)2.  A common medieval approach is use the proportion  1 + ... + c  :  1 + ... + a.  We let  Ta  =  1 + ... + a,  so this proportion is  Tc : Ta.

            Benedetto da Firenze;  della Francesca;  Calandri, 1491,  are the only cases where  c > a.

            Della Francesca begins the inverse problem -- if the contract for depth  a  is worth  b  and work stops at  x  such that the value of the dug hole is  d,  what is  x?  Denote this situation as  (a, b; x)  worth  d.  See:  della Francesca;  Pacioli;  Calandri, Raccolta;  Buteo.

            Ozanam, Vyse and Jackson are the only ones to consider use of other arithmetic progressions. 

            Berloquin gives a simple argument that work is proportional to  a2/2.

 

In Neugebauer & Sachs, op. cit. in 7.E, the problems discussed on pp. 81-91 involve digging out ditches and the cost or difficulty of digging increases with the depth, but none of these are like the problem considered here, though Tropfke 529 notes a resemblance.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  ??NYS - quoted and discussed by Tropfke 529. 

P. 96, part III, No. 17.  Calculation of ditchwork.  (15, 30; 10).  Tropfke doesn't give a solution but says it is similar to the following.

P. 227, part VI, no. 61.  "Apportionment of wells and cisterns.  In order to dig a well, the earth must be lifted out.  For the first ell, the earth comes up one ell; for the second ell, the earth comes up two ells; for the third, three ells; etc. until the end.  For this calculation, one must use the series of of natural numbers.  So we take 1 as first term, to which 2 as second term gives 3, to which 3 as third term gives 6 as sum, etc.  In this way, we do each time, and that is the 'basţ' [NOTE:  ţ  denotes a  t  with an underdot.] (literally 'extension').  Example: the well depth is 10 ells; what is the 'basţ'?  1 + 2 = 3,  3 + 3 = 6,  ...,  45 + 10 = 55.  That is the 'basţ' of a well 10 ells deep."  So he divides in the standard ratio  T10/T15 = 55/120 = 1/3 + 1/8.  "We multiply that by 30; this yields  13 [+] 1/2 [+] 1/4.  That is the payment for a well of 10 ells, when the other well costs 30 dirhems."  Tropfke says the problem is reminiscent of a Babylonian one - cf above.

Qazwini  = al‑Qazwînî.  Loc. cit. in 7.H.3.  c1260.  P. 253.  German translation only -- man contracts to dig a well  10  ells deep for  10  dirhems.  He stops at  9  ells, so we have  (10, 10; 9).  Man asks for  9  dirhems, but an expert says only  8  and somewhat more.

BR.  c1305.  No. 22, pp. 40‑43.  Man contracts to dig  10 x 10 x 10  cistern but only does  5 x 5 x 5.  Text gives him  1/8  of the value.

Lucca 1754.  c1330.  F. 64v, p. 152.  Man digging a well,  (10, b; 8).  He divides in ratio   T8 : T10 = 36 : 55.

Pseudo-dell'Abbaco.  c1440.  Prob. 102, p. 87 with plate on p. 88.  Man contracts to dig a well 20 deep and stops at  14,  i.e.  (20, b; 14).  Author divides in ratio  T14 : T20 = 1 : 2.  But he says he doesn't think this is a correct method, though he doesn't know a better one.  I have a colour slide of this.

Benedetto da Firenze.  c1465.  Pp. 115‑116.  If a well  12  deep is worth  12,  how much is a well  14 deep worth?  This is  (12, 12; 14).  He takes values proportional to Td.

Muscarello.  1478.  Ff. 66v-67r, pp. 176-177.  Man to dig a hole but hits water and has to stop,  (10, b; 7).  Divides in the ratio   T7 : T10 = 28 : 55.

della Francesca.  Trattato.  c1480.  F. 52r (121-122).  Men agree to dig a well of depth  4  for  10,  but no water is found and they continue until the cost is  11  more.  I.e.  (4, 10; x)  worth 21.  Since  T4 = 10  and  T6 = 21,  they dig to  6.  English in Jayawardene.

Calandri.  Arimethrica.  1491.  F. 65v.  (12, 12; 16).

Pacioli.  Summa.  1494. 

F. 40v, prob. 8.  Dig a well,  (11, 11; 6).  Divides as  T6 : T11 = 7 : 22,  so the partial well is worth  7/2. 

F. 40v, prob. 9.  Dig a well,  (11, 11; x)  worth  7/2.

Part II, f. 55v, prob. 38.  Dig a well,  (10, 10; 6).  Divides as  T6 : T10 = 21 : 55.

Part II, f. 55v, prob. 39.  Dig a well,  (10, 10; x)  worth  4.  He notes  4 : 10 = 22 : 55,  so we want  n(n+1)/2 = 22,  which would give  n = (-1 + Ö177)/2 = 6.15207....  He interpolates as  6  days plus  1/7  of the sixth day, i.e.  n = 6.14286....

Calandri.  Raccolta.  c1495.  Prob. 38, pp. 33‑34.  If a well  24  deep is worth  24,  how deep a well is worth  40?  I.e.  (24, 24; x)  worth  40.  He takes values proportional to  Td.

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 110, ff. 55r-55v.  Dig a well,  (9, 24; 5).  Divides in ratio  T5 : T9 = 15 : 45.

Apianus.  Kauffmanss Rechnung.  1527.  F. D.vi.v.  Mason to build a tower  100  high  for  100.  He falls ill after  84.  Divides in ratio  T84 : T100.

Cardan.  Practica Arithmetice.  1539.  Chap. 66.

Section 8, ff. CC.i.r - CC.i.v (p. 138).  Men dig a well,  (34, 60; 20).  Divides in ratio  T20 : T34 = 6 : 17.

Section 10, ff. CC.i.v ‑ CC.ii.v (p. 138).  Man building a wall  5  high at prices  10, 20, 40, 80, 160  per unit of height.  He stops at height  2½.  Takes half a unit of height as having  Ö2  times the value of the previous half-unit, so the interval from  2  to  3  high worth  40  divides into two halves worth  x  and  xÖ2,  giving  x = 40/(1+Ö2).

Buteo.  Logistica.  1559. 

Prob. 35, pp. 238-240.  Dig a well,  (100, 50; 50).  Divides as  T50 : T100 = 1275 : 5050.  Also does  (200, 50; 100).  His depths are initially in cubits, but are converted to 'semipedes' --  120  cubits is  50  semipedes.

Prob. 36, pp. 240-241.  Dig a well,  (100, 50; x)  worth  28 22/101.

Ozanam.  1694, 1725.

Prob. 7, question 1, 1696: 30;  1708: 27.  Prob. 10, question 1, 1725: 60-61.  Prob. 2, 1778: 65;  1803: 67-68;  1814: 60;  1840: 32-33.  Man digging a well  20  feet deep, to receive  3, 5, 7, ...,  for each successive foot.  [This is not really in this section, but is included because later ed. use it as the basis of the next problem.]

Prob. 51, question 1, 1725: 256‑257.  Prob. 3, 1778: 66-67;  1803: 68-69;  1814: 61-62;  1840: 33.  Man digging a well,  (20, b; 12)  (1778 et seq. change  12  to  8).  1725 divides as  T12 : T20 = 78 : 210.  1778 notes that the difficulty of the work increases in arithmetic progression, but that there are many such progressions.  He then posits that the first unit is worth  1/4  when the agreed payment is  20  and this gives a difference of  30/11  for the arithmetic progression.

            If the cost per unit depth is an arithmetic progression:  A, A+D, ..., A+(a‑1)D  and δ = D/A  and  d  is the value of the partial well, then 

d/b = ((2A-D)c + Dc2)/((2A-D)a + Da2). 

Les Amusemens.  1749.  Prob. 44, p. 176.  Mason to dig a well,  (10, b; 4).  Divides as  T4 : T10 = 10 : 55.

Vyse.  Tutor's Guide.  1771?  Prob. 17, 1793: p. 136; 1799: pp. 144-145 & Key p. 188.  Same as Ozanam, prob. 7, but with depth  30.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions, no. 25, pp. 20 & 79.  Dig  20  yards for  £20.  Man falls sick after  8  yards.  "How much was then due to him, on a supposition that the labour increases in arithmetical proportion as the depth?"  I.e.  (20, 20; 8).  Solution notes that the data does not determine what the arithmetic progression is and chooses  5s  as cost of the first yard -- see Ozanam.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 182 & 263, no. 693.  Man digging a well  49  feet deep.  First foot costs  15,  but each successive foot costs  6  more than the previous.  Find cost of last foot and total cost.  So this is really an arithmetic progression problem, but I haven't seen others of those using this context.

Vinot.  1860.  Art. CVII: Problème du puits et du Maçon, pp. 126-127,  (20, 400; 10).  He assumes the cost of digging up a unit depth is  5  and that lifting the  i-th unit raises it from its centre of gravity, so is given by  A, 3A, 5A, ..., 39A,  where  A  has to be determined from the total cost.  He finds  A = 3/4  and  d = 125.

Pierre Berloquin.  The Garden of the Sphinx, op. cit. in 5.N.  1981.  Prob. 117: A bailout fee, pp. 66 & 166.  Man contracts to bail out a  20  yard deep well for  $400  and gives up at  10  yards.  Answer says value is proportional to the depth  d  times the average distance lifted, i.e.  d/2,  hence value is proportional to  d2/2  and this is the result that integration produces.

 

          7.I.     FOUR FOURS, ETC.

 

            Express an integer using four 4s, etc.  Cupidus Scientiae, 1881, seems to be the first to ask for solutions to a lot of the integers, rather than a few specific examples.  The next examples of the general form are Cunningham & Wiggins (1905), Pearson (1907), Ball (1911), Ball (1912).  Dawson (1916) is the first to ask for four  R's,  where  R  is indeterminate, e.g.  3  =  (R+R+R)/R.  I have included examples where a set of numbers and operations is given and one has to obtain a given value.  This overlaps a bit with 7.I.1, where the object is to find the maximum possible value, and with 7.AC.3-6, where one uses all nine or ten of the digits and I have included problems of inserting signs into  12...9  to make  100  in 7.AC.3.

 

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.

Problem 4:  "Let  12  be set down in four Figures and let each Figure be the same."

Problem 9.  "Says Jack to his brother Harry, I can place four threes in such manner that they shall just make  34;  can you do so too?"

Les Amusemens.  1749.  Pp. 52-54.  Several problems leading to:  7 7/7,  33 3/3,  55 5/5,  99 9/9,  77 77/77,  2222 2222/2222,  11 1/1, etc.  See the entry in 7.AN.

Vyse.  Tutor's Guide.  1771?  Prob. 1, 1793: p. 155; 1799: p. 165 & Key p. 206.  "Four Figures of nine may be so placed and disposed of as to denote and read for  100,  neither more nor less.  Pray how is that to be done?"

Pike.  Arithmetic.  1788.  P. 350, no. 16.  "Said Harry to Edmund, I can place four  1's  so that, when added, they shall make precisely  12;  Can you do so too?"

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 2,  pp.  1 & 51.  "It is required to express  100  by four  9's."

No. 4,  pp.  2 & 51.  Use three  2s  to make  ½,  1  and  2.

No. 5,  pp.  2 & 51.  Express  12  by four equal figures.

No. 17, pp.  5 & 56.  Express    by four  5s.  Answer is:  5.5 + 5/5.

No. 33, pp.  8 & 59.  Use four  2s  to make  1/8,  1/2,  2  and  8.

No. 40, pp. 10 & 62.  Use three  3s  to make  1/3,  1  and  3.

No. 42, pp. 10 & 62.  Use four  3s  to make  1/243,  1/27,  1/3,  3,  27  and  243.

No. 43, pp. 10 & 62.  Use five  3s  to make the same numbers as in no. 42.

No. 44, pp. 10 & 63.  Express  78  by six equal digits.

Endless Amusement II.  1826?  Prob. 23, p. 201.  "Put down four nines, so that they will make one hundred."

Child.  Girl's Own Book.  Arithmetical puzzles, no. 5.  1832: 170 & 179;  1833: 184 & 193;  1839: 164 & 173;  1842: 282 & 291;  1876: 231 & 244.  "Place four nines together, so as to make exactly one hundred.  In the same way, four may be made from three threes, three may be made from three twos, &c."  The 1833 solution is printed rather oddly as  199 9‑9, while the 1839 and 1842 solution is  99 9-9  and the 1876 solution is  99 9¸9. 

Nuts to Crack III (1834), no. 211.  "Write down four nines so as to make a hundred."

Family Friend 1 (1849) 150 & 178.  Problems, arithmetical puzzles, &c. -- 4.  "Put four fives in such a manner, that they shall make  6½.  --  D. F."  Answer is   5 5/5 + .5.   = The Sociable, 1858, Prob. 46: A dozen quibbles, part 12: pp. 300 & 318.  = Book of 500 Puzzles, 1859, prob. 46: part 12, pp. 18 & 36.

Boy's Own Book.  To place four figures of  9  in such a manner as to represent 100.  1855: 601.

Magician's Own Book.  1857.  Quaint questions, p. 253.  [No. 4] -- "Place three sixes together, so as to make seven."  [No. 6] -- "Place four fives so as to make six and a half."  [Boy's Own Conjuring Book, 1860, pp. 224-225, has the Quaint Questions, but omits these two questions!]

Book of 500 Puzzles.  1859.

Prob. 46: A dozen quibbles: part 12, pp. 18 & 36.  As in Family Friend.

Quaint questions, p. 67.  [Nos. 4 & 6] -- Identical to Magician's Own Book.

Charades, Enigmas, and Riddles.  1860: prob. 30, pp. 60 & 64;  1862: prob. 31, pp. 136 & 142;  1865: prob. 575, pp. 108 & 155.  "Write a Hundred with 4 nines."  (1862 & 1865 have slightly different typography.)

Illustrated Boy's Own Treasury.  1860. 

Prob. 1, pp. 427 & 431.  "Put down four nines, so that they shall make one hundred."

Prob. 26, pp. 429 & 433.  "Put four fives in such a manner, that they shall make  6½."

Prob. 38, pp. 430 & 434.  "It is required to place four  2's  in such a manner as to form four numbers in geometrical progression?"  Uses four  2s  to make each of  1/8,  1/2,  2,  8.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 564-15, pp. 253 & 395.  Write  100  with  4  nines.  Write  1000  with no zeroes -- answer: 999 9/9.

Prob. 564-18, pp. 253 & 395.  Write  100  with no zeroes, using  4, 6 or 8 figures.  Answers:  99 3/3;  99 44/44;  99 999/999.

Magician's Own Book (UK version).  1871.  Paradoxes [no. 3], p. 37.  "With four fives make  6½? -- 5 5/5 .5."  where the  5/5  is written as a 5 over a 5 with no fraction bar.  Cf Jackson and Magician's Own Book.

Hugh Rowley.  More Puniana; or, Thoughts Wise and Other-Why's.  Chatto & Windus, London, 1875.  P. 300.  "Write down one hundred with four nines."

Mittenzwey.  1880.  Prob. 1, pp. 1 & 58;  1895?: 1, pp. 7 & 62;  1917: 1, pp. 7 & 56.  Write  100  with six equal digits.

"Cupidus Scientiae" (possibly the editor, Richard A. Proctor).  Four fours, singular numerical relation.  Knowledge 1 (30 Dec 1881) 184, item 151.  A bit vague as to what operations are permitted, but wants four  4s  to make various values.  Says he has not been able to make  19.

H. Snell.  Singular property of number  4.  Knowledge 1 (6 Jan 1882) 209, item 178.  19 = 4! ‑ 4 - 4/4.  Editor says  4!  is not reasonable for the problem as posed.

Solutions from various contributors.  Four fours.  Knowledge 1 (13 Jan 1882) 229, item 184.  Numerous solutions for  1  through  20,  except  19.  Solutions for  19  are:  4/.4 + 4/.4;   4! - 4 - 4/4;   4/Ö.4 ‑ 4/4 ("manifestly erroneous");   (4 + 4 - .4)/.4;   (x + x ‑ .x)/.x  in general.  Four  3s  give same results as three  5s,  except for  17.

Albert Ellery Berg, ed.  Op. cit. in 4.B.1.  1883.  P. 373.  "Place three sixes together so as to make seven."

Lemon.  1890. 

Vagaries, no. 217(b), pp. 33 & 105.  Three  6s  to make 7. 

Arithmetical, no. 752, pp. 92 & 124.  = Sphinx, no. 600, pp. 81 & 118.  "Place four nines so as to make one hundred."

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. II: A curious addition sum, p. 2.  Mentions "writing down 100 with four nines" as  99 9/9.

(Sam Loyd.)  One hundred pounds for correct answer to a puzzle.  Tit-Bits (14 Oct 1893) 25.  "Find How to Arrange the Figures  · 4 · 5 · 6 · 7 · 8 · 9 · 0 ·  in an Arithmetical Sum which Adds up the Nearest to  82."  "Mr. Loyd is confident that no one will find it out."  Indeed, Loyd will be paid £100 only if no correct answer is received.

(Sam Loyd.)  Solution of Mr. Sam Loyd's one hundred pound puzzle.  Tit-Bits (18 Nov 1893) 111.  80·5 + ·97 + ·46  =  82.  (There are points over the  5, 9, 7, 4, 6,  but my printer may not print these clearly.)  Here the mid-line dot (·) is used for a decimal point.  Because of the number of correct solutions, ten extra names were drawn from them for additional £5 prizes.  "It seems that not a single person in the whole of America has sent the correct answer when a prize was offered there, but here we have received a very large number actually correct."  [See MRE for another solution.]

Report on the 82 puzzle appeared in 25 Nov and  letter from Loyd appeared about two weeks later - photocopies on order.

Hoffmann.  1893.  Chap. IV, no. 18: Another way to make a hundred, pp. 148 & 193 = Hoffmann-Hordern, p. 120.  Use six  9s  to make 100.

Ball.  MRE, 3rd ed., 1896.  P. 13.  "... a question which attracted some attention in London in October, 1893, ...."  [See Loyd above.]  He says that the problem is to make  82  with the seven digits  9, 8, 7, 6, 5, 4, 0  and gives one solution as  80·69 + ·74 + ·5  (with points over the  9, 4, 5  -- there should also be points over the  6  and  7).

H. D. Northrop.  Popular Pastimes.  1901.  No. 10: A dozen quibbles, no. 12, pp. 68 & 73.  Express    by four  5s.  Answer is:  5 5/5 . 5,  which seems pretty poor to me.  c= Jackson, no. 17.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:5 (Sep 1903) 426-427.  "How to arrange four 9's to make 100."

Ball.  MRE, 4th ed., 1905.  P. 14.  Repeats the material in the 3rd ed of 1896, again omitting two points, and adds further questions.  Use the  10  digits to total  1  -- a solution is  35/70 + 148/296  -- or to total  100  -- a solution is  50 + 49 + 1/2 + 38/76.  Use the  9  digits to make four numbers which total  100  -- a solution is  78 + 15 + 2Ö9 + 3Ö64.

A. Cunningham & T. Wiggins.  ??  Math. Quest. Educ. Times 7 (1905) 43‑46.  ??NYS -- cited in Dickson I 460, item 45d.  Expressions using four  9s  and four  4s.

Pearson.  1907. 

Part I, no. 43: The nimble nines, pp. 125 & 187.  Verse asking for three  9s  to make  16  --  solution is   96/6 !!

Part II: On all fours, p. 107.  Four fours in general, with a few examples.

Wehman.  New Book of 200 Puzzles.  1908.  P. 26.  = Magician's Own Book, no. 4.

Ball.  MRE, 5th ed., 1911.  Pp. 13-14.  Briefly restates the material in the 4th ed. as "questions which have been propounded in recent years.  ...   To the making of such questions of this kind there is no limit, but their solution involves little or no mathematical skill."

                    "Another traditional and easy recreation ....  I have never seen this recreation in print, but it seems to be an old and well-known question."  Deals just with four  4s  and says one can get up through 170.  G. N. Watson has pointed out that one can get further by using factorials and subfactorials.  (The subfactorial of  n  is  n¡ = n![1/0! ‑ 1/1! + 1/2! - 1/3! + ... ± 1/n!].)  The topic is not in earlier editions.

W. W. Rouse Ball.  Four fours.  Some arithmetical puzzles.  MG 6 (No. 98) (May 1912) 289‑290.  "An arithmetical amusement, said to have been first propounded in 1881, ...."  [This would seem to refer to Knowledge, above.]  Studies various forms of the problem.  Says it occurs in his MRE -- see above.  MRE 6th ed., 1914, p. 14, cites this article.

Ball.  MRE, 6th ed., 1914.  Pp. 13-14.  He now splits the material into three sections.

                    Empirical Problems.  Restates the material in the 5th ed. as "... numerous empirical problems, ..." and omits Loyd's problem.  "To the making of such questions there is no limit, but their solution involves little or no mathematical skill."

                    He then introduces the "Four Digits Problem".  "I suggest the following problem as being more interesting."  Using the digits  1, 2, ..., n,  express the integers from  1  up using four different digits and the operations of sum, product, positive integral power and base-10  notation (or also allowing iterated square roots and factorials).  With  n = 4,  he can get to  88  or to  264.  With  n = 5,  he can get to  231  or  790.  Using  0, 1, 2, 3,  he can get to  36  (or  40).

                    Under Four Fours Problem, he discusses what operations are permitted and says he can get to  112,  or to  877  if subfactorials are permitted (citing his MG article for this).  Mentions four 9s and four 3s problems.

Williams.  Home Entertainments.  1914.  The six 9's, p. 119.  "Express the number  100  by means of six  9's."

Thomas Rayner Dawson.  1916.  ??NYS.  Cited in:  G&PJ 3 (Jan 1988) 45  &  4 (Mar 1988) 61.  Asks for four  R's,  where  R  is indeterminate, e.g.  3  =  (R+R+R)/R.

Ball.  MRE, 7th ed., 1917.  Pp. 13-14.  The material of the first two sections is repeated, but under "Four Fours Problem", he discusses the operations in more detail.  With  +,  -,  x,  ¸,  brackets and base-10  notation, he can get to  22.  Allowing also finitely iterated square roots, he can get to  30.  Allowing also factorials, he can get to  112.  Allowing also integral indices expressible by  4s  and infinitely iterated square roots, he can get to  156.  Allowing also subfactorials, he can get to  877.  (In the 11th ed., 1939, pp. 15-16, two footnotes are added giving expressions for  22  in the first case and  99  in the third case.)  Gives some results for four  2s,  four  3s,  four  5s,  four  9s.  Mentions the general problem of  n  ds.

Smith.  Number Stories.  1919.  Pp. 112‑113 & 140‑141.  Use four  9s  to make  19,  2  and  20.

Ball.  MRE, 9th ed., 1920.  Pp. 13-14.  In the "Four Digits Problem", he considers  n = 4,  i.e. using  1, 2, 3, 4,  and discusses the operations in more detail.  Using sum, product, positive integral power and base-10  notation, he can get to  88.  Allowing also finitely iterated square roots and factorials, he can get to  264.  Allowing also negative integral indices, he can get to  276.  Allowing also fractional indices, he can get to  312.  He then mentions using  0, 1, 2, 3  or four of the five digits  1, ..., 5.

                    Under "Four Fours Problem", he repeats the material of the 7th ed., but adds some extra results so he has results for four  ds,  d = 1, 2, 3, 5, 6, 7, 8, 9.

Ball.  MRE, 10th ed., 1922.  Pp. 13-14.  In the "Four Digits Problem", he repeats the material of the 9th ed., but at the end he adds that using all of the five digits,  1, ..., 5,  he has gotten to  3832  or  4282,  depending on whether negative and fractional indices are excluded or allowed.

Hummerston.  Fun, Mirth & Mystery.  1924.  Some queer puzzles, Puzzle no. 76, part 2, pp. 164 & 183.  "Express 100 by using the same figure six times."

Dudeney.  MP.  1926.  Prob. 58: The two fours, pp. 23‑24 & 114.  = 536, prob. 109, pp. 34 & 248‑249, with extensive comments by Gardner.

King.  Best 100.  1927.

No. 44, pp. 20-21 & 48.  = Foulsham's no. 14, pp. 8 & 11.  "Can you put down four fifteens so that they come to  16,665?

No. 45, p. 21 & 49.  "Arrange the figures  1  to  7  so that they will amount to  100,  when added together."  Arrange four  9s  to make  100.  Gives two answers for the first part.

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.  P. 112.  Write 1000 with seven or five equal digits.

Perelman.  FFF.  1934.  1957: probs. 98, 100 & 102, pp. 137 & 143-144;  1979: probs. 101, 103 & 105, pp. 166-167 & 174-175.  = MCBF, probs. 101, 103 & 105, pp. 167 & 176‑178.

101: Two digits:  "What is the smallest integer that can be written with two digits?"  1/1 = 2/2 = ....  [Though I think  0/1  might be counted.]

103: Five 9's:  "Write 10 with five  9's.  Do it in at least two ways."

105: Four ways:  "Show four different ways of writing 100 with five identical digits."

Perelman.  MCBF.  1937.  Any number via three twos.  Prob. 202, pp. 398-399.  "A witty algebraic brain-teaser that amused the participants of a congress of physicists in Odessa."  n  =  ‑ log2 log2 Ön2,  where  Ön  means  n‑fold iterated square root.

Haldeman-Julius.  1937.  No. 16: Adding fives, pp. 5 & 21.  Use four 5s to make  6½.  Answer is:  5 + 5/5 + .5.

M. Adams.  Puzzle Book.  1939.  Prob. B.83: Figure juggling, part 3, pp. 78 & 107.  Use a digit  8  times to make  1000.  Answer uses  8s.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  Number, please!, pp. 20 & 210. 

Use the same odd figure five times to make  14. 

Four  9s  to make  100. 

Four  5s  to make  6½.

Depew.  Cokesbury Game Book.  1939. 

Three eights, p. 216.  Use three  8s  to make  7.

Twenty-four, p. 227.  Use a digit three times to make  24.  Answers:  33 - 3,   22 + 2.

McKay.  Party Night.  1940.  No. 9, p. 177.

(a)  Use three  9s  to make  10.  Answer:  9 9/9.

(b)  Use four  9s  to make  20.  Answer:  9 99/9.

(c)  Use three  9s  to make  100.  Answer:  99.9 (or 99.9 for clarity).

(d)  Use two  9s  to make  10.  Answer:  9/.9  or  9.9 (or 9.9).

Meyer.  Big Fun Book.  1940.  A half dozen equals 12, pp. 119 & 738.  Use six  1s  to make  12.  Answer:  11 + 11/11.

George S. Terry.  The Dozen System.  Longmans, Green & Co., NY, 1941.  ??NYS -- quoted in:  Underwood Dudley; Mathematical Cranks; MAA, 1992, p. 25.  What numbers can be expressed with four  4s  duodecimally?  About 5 dozen.  How many numbers can be expressed using each of the digits  1,  2,  3,  4  once only, again duodecimally?  About 9 dozen and nine.  Terry (or Dudley) also gives the results for decimal working as  22  and  88.

Sullivan.  Unusual.  1943.  Prob. 17: Five of a kind.  Write 100 with the same figure five times "and the usual mathematical symbols".  Says it can be done with  1s,  2s,  5s (two ways),  9s  and perhaps others.

J. A. Tierney, proposer;  Manhattan High School of Aviation Trades, D. H. Browne, H. W. Eves, solvers.  Problem E631 -- Two fours.  AMM 51 (1944) 403  &  52 (1945) 219.  Express 64 using two 4s.

Vern Hoggatt & Leo Moser, proposers and solvers.  Problem E861 -- A curious representation of integers.  AMM 56 (1945) 262  &  57 (1946) 35.  Represent any integer with  p  a's,  for any  p ³ 3  and any  ¹ 1.  Solution for  ±n  uses  log  to base  ÖÖ...Öa,  with  n  radicals.

S. Krutman.  Curiosa 138: The problem of the four  n's.  SM 13 (1947) 47.

Sullivan.  Unusual.  1947.  Prob. 28: A problem in arithmetic.  What is the smallest number of eights which make 1000?

G. C. S[hephard, ed.]  The problems drive.  Eureka 11 (Jan 1949) 10-11 & 30. 

No. 4.  Use  1, 2, 3,  once each to make  19.  Answer:  (2/.1) - [Ö3].  Ibid. 12 (Oct 1949) 17 gives a simpler answer:  Ö(1 + 3!!/2).

No. 7/  Use four 1s to express 7, 37, 71, 99.  Answers:  (1+1+1)! + 1;  111 x Ö.1  [.1  is  .111..., but may not show up clearly];  .1 x (Ö(1/.1)!! - 1  [same comment on  .1];  1/(.1 x .1) - 1.

Anonymous.  The problems drive.  Eureka 13 (Oct 1950) 11 & 20-21. 

No. 4: Start with  2  and use cubing and integral part of square root to form any positive integer.  m  cubings, followed by  n  roots gives  2^(3m/2n) = 2^(2ma-n),  where  a = log2 3.  Since  a  is irrational, we can choose  ma - n  so that  2^(2ma-n)  is arbitrarily close to  N + ½,  so the integer part of it is  N.

No. 6: Use four  4s  to approximate  π.  They get  3.14159862196...,  using a nine-fold root.

Anonymous.  The problems drive.  Eureka 17 (Oct 1954) 8-9 & 16-17.  No. 5.  Use four 4s to express  37; 57; 77; 97; 123.

D. G. King‑Hele.  Note 2509:  The four 4's problem.  MG 39 (No. 328) (May 1955) 135.  n  =  log[{log 4}/{log Ön 4}]/log Ö4  expresses any positive integer  n  in terms of three  4s.  A slight variation expresses  n  in terms of four  x's,  for any real  x ¹ 0, 1.  1  can be used by taking  x = .1.  He also expresses  n  in terms of  m  x's  for real  x ¹ 0, 1  with  m > 5  and also with  m = 5.

Anonymous.  Problems drive.  Eureka 18 (Oct 1955) 15-17 & 21.  No. 6.  Use four 4s to make  7; 17; 37; 3,628,800.

Anonymous.  Problems drive, 1958.  Eureka 21 (Oct 1958) 14-16 & 30.  No. 10.  Use  1, 2, 3, 4, 5,  in order to form  100;  3 1/7;  32769.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.

No. 12: One hundred every time, pp. 10 & 41.  Make an arrangement of  x's  which gives the result  100  for  x = 1, 2, ..., 9.  Answer:  xxx/x - xx/x.

No. 20: Form fours, pp. 12 & 42.  Eight  4s  to make  500.

No. 74: Signs wanted, pp. 28 & 54.  Insert signs  (+,  -,  x,  /)  into a row of four  x's  to make  10 - x,  for  x = 2, ..., 9.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 3.  Use three 1s to make the integers from one to twelve, using only arithmetic symbols.  (No trigonometric functions or integer parts allowed.)

Young World.  c1960.  P. 54.  Use five  9s  to make  1000.

B. D. Josephson  &  J. M. Boardman.  Problems drive 1961.  Eureka 24 (Oct 1961) 20-22 & 24.  Prob. K.  Use three 7s to express  1, ..., 11,  using only arithmetic symbols.

J. H. Conway  &  M. J. T. Guy.  π  in four 4's.  Eureka 25 (Oct 1962) 18‑19.  Cite Eureka 13 (1950).  Note that  π  =  Ö[(‑Ö4/4)!]4,  if non‑integral factorials are allowed.  Show that any real number can be arbitrarily well approximated using four  4s.

R. L. Hutchings  &  J. D. Blake.  Problems drive 1962.  Eureka 25 (Oct 1962) 20-21 & 34-35.  Prob. H.  Use four identical digits to represent 100 in as many ways as possible, but not using representations which are independent of the digit used, like  (5x5)/(.5x.5).  The give eight examples, using 9, 5, 5, 4, 4, 3, 3, 9, 1, and say there are more.

D. E. Knuth.  Representing numbers using only one  4.  MM 37 (1964) 308‑310.

Gardner.  SA (Jan 1964) adapted as Magic Numbers, chap. 5.  Cites Knowledge as the origin.  Magic Numbers gives numerous other citations.

Marjorie Bicknell & Verner E. Hoggatt.  64 ways to write 64 using four  4's.  RMM 14 (Jan‑Feb 1964) 13‑15.

Jerome S. Meyer.  Arithmetricks.  Scholastic Book Services, NY, 1965.  Juggling numbers, no. 3, pp. 83 & 88.

Make  100  with four  7s.   77/.77.

Make  20  with two  3s.   3!/.3.

Make  7  with four  2s.   (2/.2)/2 + 2.

Make  37  with six  6s.   6*6 + 66/66.

Ripley's Puzzles and Games.  1966.  Pp. 16-17, item 2.  Use  13  3s  to make  100.

Steven Everett.  Meanwhile back in the labyrinth.  Manifold 10 (Autumn 1971) 14‑16.  (= Seven Years of Manifold 1968‑1980; ed. by I. Stewart & J. Jaworski; Shiva Publishing, Cheshire, 1981, pp. 64‑65.)  n  =  Ö4 * log4 log4 Ön4,  where  log4  means log to the base 4 and  Ön  means  n‑fold iterated square root.  This is a variant of King‑Hele's form.  This article is written in a casual style and seems to indicate that this formula was devised by Niels Bohr.  He gets a form for  e,  but it uses infinitely many factorial signs!!!!...

                    Editor's note on p. 2 (not in the collection) gives an improvement due to Michael Gerzon, but it is unclear what is intended.  The note gives another method due to Professor Burgess using  sec tan-1 Öm  =  Ö(m+1)  and  1  =  Önn  which expresses  n  by one  1.

[Henry] Joseph and Lenore Scott.  Master Mind Pencil Puzzles.  1973.  Op. cit. in 5.R.4.  Numbers-numbers, part 3, pp. 109-110.  Use  13  3s  to make  100.  The give  33 + 33 + 33 + (3/3)3 + 3x3 + 3x3.  I found  33 + 33 + 33 + 33/3 - 3x3 - 3/3,  which seems simpler.

Ball.  MRE, 12th ed., 1974.  Pp. 15-17.  Under "Four fours problem", the material of the 9th ed. and the footnotes mentioned at 7th ed are repeated, but the bound for four  9s  is increased.

Bronnie Cunningham.  Funny Business.  An Amazing Collection of Odd and Curious Facts with Some Jokes and Puzzles Too.  Puffin, 1978.  Pp. 38 & 142.  Arrange three  9s  to make  20.  Answer:  (9 + 9)/.9.

Putnam.  Puzzle Fun.  1978.

Nos. 54-57: Ten is the number, pp. 10 & 35.  Express 10 using five  9s,  in four different ways.

Nos. 58-59:  3 + 3 + 3 = 30,  pp. 10 & 35.  Express 30 using three  3s,  in two different ways.

No. 97: Eight to one thousand, pp. 13 & 37.  Use a digit eight times to express 1000.

P. Grammer, I. McFiggans, N. Blacknell, T. Joyce, J. Anstey & A. Devonald.  Counting in fours.  MiS 9:4 (Sep 1980) 21-22.  Uses four 4s to express  1, ..., 50.  Says  51 - 100 will appear in next issue, but they didn't.

J. Bellhouse.  Four fours.  MiS 14:1 (Jan 1985) 15.  Says the promised table for  51 - 100  (see Sep 1980 above) had not appeared, so his students found their own.

Anne Williamson.  1985.  MiS 14:4 (Sep 1985) 7.  Use the four digits  1, 9, 8, 5  to express integers  1 - 100.  Unhappy with expressions for  24, 31, 65  which use !.

Ken Lister.  Letter.  MiS 15:2 (Mar 1986) 47.  Responding to Bellhouse (Jan 1985).  Corrects and improves some values, but says  71 and 73  have not been done.  Expresses  a/b,  for single digits  a, b,  by use of four  4s.

Angie Aurora.  Letter.  MiS 15:3 (May 1986) 48.  Improvements for Williamson's problem -- Sep 1985 above.

Joyce Harris.  Letter:  Four fours.  MiS 15:3 (May 1986) 48.  Responding to Lister (Mar 1986), gives expressions for  71 and 73.

Bob Wasyliw.  Letter:  Four 4's -- the ultimate solution.  MiS 15:5 (1986) 39.  Adapts Everett's 1971 method to include non‑positive integers.

Simon Gray & Colin Abell.  Letters:  Four fours again.  MiS 16:2 (1987) 47.  Gray notes that  4 = Ö4 * Ö4,  so that 'four 4s' is the same as 'at most four 4s'.  He gives  π  =  Ö4 * sin‑1(4/4)  and more complex forms.  Abell gives  π  =  Ö(‑4/4) * log (‑4/4)  [The first minus sign is ambiguous??] and  π  =  Ö(4*4) * Tan-1(4/4)  [The minus sign is wrong.]

Tim Sole.  The Ticket to Heaven and Other Superior Puzzles.  Penguin, 1988.

     Number play (ii) ‑ (iv), pp. 15, 29 & 178‑180.

(ii).  Using  +,  ‑,  *,  /,  .,  !  and brackets,  (,  ),  one can express  1 ‑ 48  with four  3s.  Allowing also  SQR,  one can express  1 ‑ 64  with four  3s.

(iii).  Using all 7 symbols in (ii) and brackets, one can express  1 ‑ 112  with four  4s.  Gives solutions with two and three  4s.

(iv).  If we also allow  Σ(n) [=  Tn,  the  n-th triangle number], he finds solutions with one, two or three  4s.  Using  log,  all integers can be expressed with three  4s.

     Three of the best -- (iii), pp. 17 & 32.  Some solutions with one  4,  using  !,  Ö,  [n] = INT(n)  and  Σ(n).

Tony Forbes.  Fours.  M500 116 (Nov 1989) 4‑5.  Says someone (possibly Marion Stubbs?) gave a simple variation of King‑Hele's and Everett's formulae to use exactly four  4s  to yield  n.  Forbes suggests using one  4  and the three operations:  !,  Ö  and  INT.  He already gets stuck at 12.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Take two, pp. 96 & 139.  Use five  2s  to express  1, 2, ..., 26,  particularly 17 and 26.

P. H. R.  Fawlty keys.  Mathematical Pie, 129 (Summer 1993) 1023  &  Notes, p. 1.  Calculator whose only keys are  5,  7,  +,  -,  x,  ¸  and  =.  Make numbers from 0 to 20.  Solution notes you can make any number by adding enough terms of the form  5 ¸ 5  and then gives short solutions for 1 through 20.

Clifford A. Pickover.  Phi in four  4's.  Theta (Crewe) 7:2 (Autumn 1993) 5-8.  In Sep 1991 he asked for good approximations to  φ  using four  4s,  either with as many symbols as you want or with each symbol used at most four times.  Says he was inspired by Conway & Guy's paper of 1962.  Brian Boutel produced  φ  =  (Ö4 + Ö{4!-4})/4.  Pickover then extended the question and various solvers got  φ  in five  5s,  seven  6s,  eight  8s,  nine  9s  and  2k-5  ks.

John Seldon.  Fours.  M500 136 (Jan 1994) 15‑16.  Answers Forbes' 1989 problem of expressing  1 ‑ 100  with one  4  and any number of factorials,  x!,  and integer square roots,  [Öx].

David Crawford and students.  1999  the end of an era.  MiS 28:4 (Sep 1999) 25.  Uses  1, 9, 9, 9  to make all integers from  1  to  100.  Notes that  2000, 2001, ...  are not going to be very useful for such puzzles!

Derek Ball.  Four 4s.  MTg 173 (Dec 2000) 18.  Says his fifth year teacher discovered the following for  n  in terms of four 4s:  n = logÖ4/4 log4 Ön 4,  where  Ön denotes  n-fold repeated square root.  Cf Perelman, 1937; King-Hele, 1955; Everett, 1971 -- Everett is very close to this and the others are not quite that close.

 

          7.I.1.  LARGEST NUMBER USING FOUR ONES, ETC.

 

Mittenzwey.  1880.  Prob. 142, pp. 30 & 79-80;  1895?: 162, pp. 34 & 82;  1917: 162, pp. 31 & 79-80.  Find largest number using four digits.  Gets  9^9^9^9  and tries to contemplate its size.  9^9  is given as  387,420,488 (last digit should be 9),  so  9^9^9  has  369,693,100  digits.

James Joyce.  Ulysses.  (Dijon, 1922);  Modern Library (Random House), NY, 1934, apparently printed 1946.  P. 684 (Gardner says the 1961 ed. has p. 699).  Bloom estimates that  9^(9^9)  would occupy "33 closely printed volumes of 1000 pages each", but he erroneously phrases the number as "the 9th power of the 9th power of 9", which is only  981.

King.  Best 100.  1927.  No. 35, pp. 19 & 46.  Largest number using two  4s.  Gives  44 = 256.

Perelman.  FFF.  1934.  Four 1's.  1957: prob. 103, pp. 137 & 144-145;  1979: prob. 106, pp. 167 & 175.  = MCBF, prob. 106, pp. 167 & 178  = MCBF, prob. 131, p. 217.  "What is the biggest number that can be written with four  1's?

Perelman.  MCBF.  1937.  Probs. 128-132, pp. 214-219.  Largest numbers with:  Three twos;  Three threes;  Three fours;  Four ones;  Four twos.

Sullivan.  Unusual.  1947.  Prob. 30: Not  999.  Largest number that can be written with three integers [sic!].  Answer:  9^(9^9).

G. C. S[hephard, ed.]  The problems drive.  Eureka 11 (Jan 1949) 10-11 & 30.  No. 5.  Find the largest numbers expressible using four 2s or four 4s, no symbols allowed.  Answers:  2^2^22;  4^4^4^4.

Leroy F. Meyers.  An integer construction problem.  AMM 66:7 (Aug/Sep 1959) 556-561.  This deals with Ball's "Four Digits Problem" (see MRE, 6th ed., 1914 in 7.I) and generalizations.  In particular, he shows that if one uses  1, 2, 3, 4,  with operations  +,  ‑,  x  and brackets, then one can obtain precisely the following:  1, 2, ..., 28, 30, 32, 36.  In general he obtains the largest integer expressible using a given multiset of integers (i.e. one is allowed a fixed number of repeats of a value) using the operations  +,  x  and brackets.  He also shows that allowing also  -,  for both negation and subtraction, does not increase the maximum obtainable value.  He conjectures that allowing also  ¸,  for both reciprocation and division, does not increase the maximum obtainable, but Meyers has written that a student once showed him a counterexample, but he cannot remember it.  He applies his general results to show that no other values are obtainable when using  1, 2, 3, 4.

Problematical Recreations 4.  Problem 1 and its answer, pp. 3 & 36.  (This is one of a series of booklets issued by Litton Industries, Beverly Hills, California, nd [c1963], based on the series of the same name in Aviation Week and Electronic News during 1959-1971.  Unfortunately, neither the date nor location nor author is given and the booklet is unpaginated.  The answer simply states the maximum value with no argument.)  Reproduced with a proper solution in:  Angela Dunn; Mathematical Bafflers; (McGraw-Hill, 1964, ??NYS);  revised and corrected 2nd ed., Dover, 1980, pp. 119 & 132  and with just the answer in:  James F. Hurley; Litton's Problematical Recreations; Van Nostrand Reinhold, NY, 1971, chap. 7, prob. 8, pp. 238 & 329.   "What is the largest number which can be obtained as the product of positive integers which add up to 100?"  (This type of problem must be much older than this??  Meyers writes that he first encountered such problems as an undergraduate in 1947.  If one looks at maximizing the LCM instead of the product of the terms, this is the problem of finding a permutation of  100  letters with maximum order.)

Sol Golomb.  Section 13.8  The minimization of the cost of a digital device (the juke-box problem).  IN:  Ben Nobel;  Applications of undergraduate mathematics in engineering.  MAA & Macmillan, 1967, pp. 284-286.  This considers the problem of Problematical Recreations in the inverse form.  We have  r  k-state devices which allow  kr  choices and the cost is proportional to  rk.  Minimize the ratio of cost to capacity or maximize the ratio of capacity to cost.  [Another way to express this is to ask which base is best for a computer to use?  I recall this formulation from when I was a student in the early 1960s.  The answer is  e,  but here only integer values are used.  One can generalise to: given a value, find the smallest sum of numbers whose product is the given value.]  The connection with juke-boxes is that they typically have two rows of  12  buttons and one has to press two buttons to make a selection of one from  144  records.  One can do much better with five rows of three buttons, but asking a customer to punch five buttons may be unreasonable, so perhaps three rows of five or six buttons might be best.

The Fortieth William Lowell Putnam Mathematical Competition, 1 Dec 1979.  Problem A-1.  Reproduced in:  Gerald L. Alexanderson, Leonard F. Klosinski & Loren C. Larson;  The William Lowell Putnam Mathematical Competition  Problems and Solutions: 1965‑1984;  pp. 33 & 109.  "Find positive integers  n  and  a1,  a2, ...,  an  such that  a1 + a2 + ... + an = 1979  and the product  a1a2...an  is as large as possible."

Cliff Pickover & Ken Shirriff.  The terrible twos problem.  Theta (Crewe) 6:2 (Autumn 1992) 3-7.  They study the problem of making numbers using just  +,  -,  x,  ^,  and  1s  (or  1s  and  2s).  For a given  n,  what is the least number of digits required?  They later permit concatenation, e.g.  11  or  12  is permitted.  They report results from various programs and mention some related problems.

Bryan Dye.  1, 2, 3, 4 -- four digits that dwarfed the universe.  Micromath 10:3 (Aut 1994) 12‑13.  Says a version appeared in SA a few years ago and is discussed in:  Clifford A. Pickover; Computers and the Imagination; Alan Sutton, 1991.  Dye's version is to make the largest number using  1,  2,  3,  4  once and the signs  +,  -,  x,  ¸,  ( ),  . (i.e. decimal point).  Exponentiation was not considered a sign and was permitted.  Pickover's version allowed only the signs  -,  ( ),  . (i.e. decimal point).  The largest value found actually fits Pickover's conditions:  .3^-(.2^-{.1^-4})  has  106990  digits.  The largest number using just exponentiation was  2^(3^41))  with  1019  digits.

 

          7.J.    SALARY PUZZLE

 

            It is better to get a rise of  5  every six months than a rise of  20  every year.  The interpretation of the first phrase is somewhat ambiguous -- see Mills (1993).  If the salary is  S  every six months, the usual interpretation of the first phrase is that the half-yearly payments are:  S,  S + 5,  S + 10,  S + 15,  S + 20,  S + 25,  ...,   while the second phrase gives payments of:  S,  S,  S + 10,  S + 10,  S + 20,  S + 20,  ...,   and the former gets  5  extra every year.

 

Ball.  MRE, 3rd ed., 1896. pp. 26‑27.  £20 per year versus £5 every half year.  He says this is a question "which I have often propounded in past years."  It is not in the 1st ed.

Workman.  Op. cit. in 7.H.1.  1902.  Section IX (= Chap. XXXI in c1928 ed.), examples CXLV, prob. 16, pp. 425 & 544 (431 & 577 in c1928 ed.).  Compares rise of £15 per year every 3 years with £5 every year.  This represents a precursor of the puzzle version.

A. Sonnenschein & H. A. Nesbit.  The New Science and Art of Arithmetic For the Use of Schools.  A. & C. Black, London, 1903.  Prob. 12 & 14, pp. 341 & 489.  "A youth entered an office at the age of 15 at a salary of £40 a year, with an annual rise of £12.  ..."  "... What total sum would he have received in 30 years? and what would he have received if the increase had been at the rate of £1 per month?"

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:4 (Aug 1903) 336-337.  Not too obvious.  Half-yearly rise of £5 versus £20 a year.  No solution given.

Susan Cunnington.  The Story of Arithmetic.  Swan Sonnenschein,, London, 1904.  Prob. 35, p. 217.  £5 raise each six months versus £10 raise each year.

Dudeney??  Breakfast Table Problems No. 330: Smith's salary.  Daily Mail (30  &  31 Jan 1905) both p. 7.  Raise of  10  per year versus    every six months.

Pearson.  1907.  Part II, no. 87, pp. 132 & 208.  As in Ball.

Loyd.  Salary puzzle.  Cyclopedia, 1914, pp. 312 & 381.  = MPSL1, prob. 84, pp. 81 & 150‑151.  = SLAHP: The stenographer's raise, pp. 60 & 108.  Raise of  100  per year versus  25  every half year.  Interpreting the raise of  25  as worth only  12.50  in a half‑year, this option loses, contrary to all other approaches.

Clark.  Mental Nuts.  1916, no. 6.  The two clerks.  $25 rise each six months versus $100 rise each year.

Dudeney.  AM.  1917.  Prob. 26: The junior clerk's puzzle, pp. 4 & 150.  Two clerks getting £50 per year with one getting a raise of £10 per year versus the other getting a raise of  £2 10s  every six months with a complicated further process of savings at different rates for five years.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 276, pp. 103 & 171: A problem in salaries.  £20 rise every six months versus £80 rise each year.

T. O'Conor Sloane.  Rapid Arithmetic.  Quick and Special Methods in Arithmetical Calculation Together with a Collection of Puzzles and Curiosities of Numbers.  Van Nostrand, 1922.  [Combined into: T. O'Conor Sloane, J. E. Thompson & H. E. Licks; Speed and Fun With Figures; Van Nostrand, 1939.]  The two clerks, p. 168.  Raise of $50 every six months versus $200 each year.

F. & V. Meynell.  The Week‑End Book, op. cit. in 7.E.  1924.  Prob. one, p. 274 (2nd ed.), pp. 406‑407 (5th? ed.).  Raise of 20 per year versus 5 every half year.

Peano.  Giochi.  1924.  Prob. 16, p. 5.  1000 per year with rise of 20 each year versus 500 each half-year with rise of 5 each half-year.

Wood.  Oddities.  1927.  Prob. 31: A matter of incomes, p. 31.  $1000 per year with $20 per year increase versus "$5 each half year increase".

Collins.  Fun with Figures.  1928.  Do figures really lie?, pp. 35-36.  $50 every six months versus $200 per year.

R. Ripley.  Believe It Or Not!  Book 2.  (Simon & Schuster, 1931);  Pocket Books, NY, 1948.  P. 123: Figure your raise in pay.  A raise of one every day is better than a raise of 35 every week.  (Assumes a six day week.)

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 20, pp. 16 & 189‑190.  Raise of 2000 per year versus 500 half‑yearly.

Phillips.  Brush.  1936.  Prob. G.1: The two clerks, pp. 20 & 87.  Raise of 200 annually versus 50 half‑yearly.

McKay.  At Home Tonight.  1940.  Prob. 4: A choice of rises, pp. 63 & 76.  £5 per half-year versus £15 per year.  Solution is unclear and seems to be wrong.  "£5 each six months is £5 in the first half-year and £10 in the second -- that is, £15 per year.  But the man who gets £5 per six months gets £5 in the first year, and of course he keeps this advantage year by year."  I get that the first case is ahead by  £5n  in the  n-th year.

Sullivan.  Unusual.  1943.  Prob. 5: Raising the raise question.  Raise of $20 per year versus $5 every half year.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 56: The interview, pp. 39 & 99-100.  £3050  yearly plus  £100  each year versus  £1500  half-yearly plus  £50  each half-year.

D. J. Hancox, D. J.  Number Puzzles  For all The Family.  Stanley Thornes, London, 1978.  Puzzle 11, pp. 4 & 48..  £1 rise per month every month versus £144 rise per year every year.  Says the first gives £66 extra in the first year and £210 extra in the second year, while the second gives no extra in the first years and £144 extra in the second year.  He then says: "Hence he would never get the £66 he would have received in the first year."  In fact, he loses £66 every year.

Stuart E. Mills.  Dollars and sense.  CMJ 24 (1993) 446-448.  Raise of $1000 per year versus $300 every half year.  Discusses various interpretations of the second phrase and gives some recent references.

                    Comments by myself and various others seem to have appeared (??NYS) as they are included in the collection of these columns:  Edward J. Barbeau; Mathematical Fallacies, Flaws, and Flimflam; Spectrum Series, MAA, 2000, pp. 11-14.  This gives references to various recent appearances of the problem in 1943, 1983, 1992.

John P. Ashley.  Arithmetickle.  Arithmetic Curiosities, Challenges, Games and Groaners for all Ages.  Keystone Agencies, Radnor, Ohio, 1997.  P. 80: The best job offer.  $20,000 per year plus a $500 raise every six months versus $20,000 per year plus a raise of $1000 each year.  He says the six-monthly payments of the first are  10,000,  10,500,  11,000, 11,500,  ...  and the payments for the second are  10,000,  10,000,  11,000,  11,000,  ....  But the second is a raise of $2,000 per year!

 

          7.K.   CONGRUENCES

 

The Friday Night Book  (A Jewish Miscellany).  Soncino Press, London, 1933.  Mathematical Problems in the Talmud: The Divisibility Test, p. 137.  Hebrew law requires fields to lie fallow every seventh year, and this is to hold for all fields at once!  Rabbi Huna gave the following rule.  Write the year as  y = 100a + b.  Form  2a + b.  Then the year is a Sabbatical year if  7  divides  2a + b.  No explanation is given in the Talmud, but we clearly have  y º 2a + b (mod 7).  [The Talmud was compiled in the period -300 to 500.  This source says Rabbi Huna is one of the few mathematicians mentioned in the Talmud, but gives no dates and he is not mentioned in the EB.  From the text of another problem attributed to him (cf in Section 6.AD), the problem would seem to be sometime in the 1-5 C.]

 

          7.K.1.          CASTING OUT NINES

 

            See Smith, History II 151‑154 for a detailed discussion.  He says it appears in al‑Khowarizmi and al‑Karkhi and that it is generally assumed to come from India, but his earliest Indian source is Lilavati, 1150.  G. R. Kaye;  References to Indian mathematics in certain Mediæval works;  J. Asiatic Society of Bengal (NS) 7:11 (Dec 1911) 801-816 notes the appearances in al-Khwârizmî, Avicenna and Maximus Planudes [Arithmetic after the Indian method; c1300; op. cit. in 7.E.1] but asserts it does not occur in early Indian sources -- but cf Aryabhata II, 950.

            Dickson I, chap. XII, pp. 337‑346, especially p. 337, gives a concise history.  He says al‑Karkhi was the first to use a  (mod 11)  check.

            See Tropfke, pp. 165-167.

            I have recently realised that certain puzzle problems should be listed here, but so far I have only noted Boy's Own Book, Boy's Own Book (Paris), Carroll, Peano, Parlour Games for Everyone -- there must be many more 19C and even 18C examples.  Basically these involve getting someone to produce a number divisible by nine and asking him to delete one digit and tell you the others -- you tell him the missing digit.  There are many of these and I probably won't try to record all of them, but subtracting a number from its reversal may be the forerunner of the  1089  puzzle of 7.AR.

 

St. Hippolytus.  Κατα Πασωv Αιρεσεωv Ελεγχoσ (??= Philosphumena) (= Refutatio Omnium Haeresium  = Refutation of all Heresies).  c200.  Part iv, c. 14.  ??NYS.  Discusses adding up digits corresponding to letters  (mod 9)  and mentions considering it  (mod 7).  (HGM I 115‑117)  See also:  Smith II 152,  Dickson I 337  and Saidan (below), p. 472.  Hippolytus doesn't use the method to check any arithmetic.  (St. Hippolytus may be the only antipope to be counted a saint!  A reference says he was Bishop of Portus and the MS was discovered at Mt. Athos in 1842.)

Iamblichus.  On Nicomachus's Introduction to Arithmetic.  c325.  ??NYS.  In:  SIHGM I 108‑109.  Special case.  (See also HGM I 114‑117.)

Muhammad [the  h  should have an underdot] ibn Mûsâ al‑Khwârizmî.  c820.  Untitled Latin MS of 13C known as Algorismus or Arithmetic, Cambridge Univ. Lib. MS Ii.6.5.  Facsimile ed., with transcription and commentary by Kurt Vogel as:  Mohammed ibn Musa Alchwarizmi's Algorismus, Das früheste Lehrbuch zum Rechnen mit indischen Ziffern; Otto Zeller Verlagsbuchhandlungen, Aalen, 1963.  English translation by John N. Crossley & Alan S. Henry as:  Thus spake al‑Khwārizmī: A translation of the text of Cambridge University Library Ms.Ii.vi.5; HM 17 (1990) 103‑131.  [Crossley & Henry name this author  Abū Ja‘far Muhammad [the  h  should have an underdot] ibn Mūsā al‑Khwārizmī,  but I have seen no other authority giving  Abū Ja‘far  -- several give  Abū ‘Abdallāh  and Rosen's translation The Algebra of Mohammed ben Musa specifically says our author "must therefore be distinguished from  Abu Jafar Mohammed ben Musa,  likewise a mathematician and astronomer, who flourished under the Caliph Al Motaded" (c900).  F. 108r  = Vogel p. 25  = Crossley & Henry p. 117.  Describes casting out  9s  in doubling and in multiplication.

Aryabhata II.  Mahâ‑siddhânta.  950.  Edited by M. S. Dvivedi, Braj Bhushan Das & Co., Benares, 1910.  English Introduction, pp. 21‑23;  Sanskrit text, p. 245.  Casting out  9s  for multiplication, division, squaring, cubing, and taking square and cube roots.  (Datta & Singh I 181 give the text in English.)

Abû al-Hassan [the  H  should have an underdot] Ahmad[the  h  should have an underdot] Ibn Ibrâhîm al-Uqlîdisî.  Kitâb al Fuşûl [NOTE:  ş  denotes an  s  with an underdot.] fî al‑Hisâb[the  H  should have an underdot]  al‑Hindî.  952/953.  MS 802, Yeni Cami, Istanbul.  Translated and annotated by A. S. Saidan as:  The Arithmetic of Al‑Uqlīdisī; Reidel, 1978.  Book II, chap. 13, pp. 153‑155 and Book III, chap. 7‑8, pp. 195‑201 deal with checking by casting out  9s,  which is given only briefly, apparently being well‑known.  He applies it to division and square roots.  The method is also mentioned in Book II, chap. 2.  On pp. 468‑472, Saidan discusses the appearance of various rules in early texts.  His earliest Indian example is Lilavati, 1150, but he gives no reference.

Kūshyār ibn Labbān  = Abū ăl-Hasan [the  H  should have an underdot] Kūšyār ibn Labbān ibn Bāšahri al-Ğīlī.  Kitāb fī usūl Hisāb al-Hind [Principles of Hindu Reckoning].  c1000.  Facsimile with translation by Martin Levey & Marvin Petruck.  Univ. of Wisconsin Press, Madison, 1965. 

First Book, Ninth Section: On arithmetic checks, f. 274a, pp. 70-71.  Brief description of casting out nines.

Second Book, Eleventh Section: On checks, ff. 280a-280b, pp. 94-97.  Brief description of casting out nines in base 60.

Introduction, pp. 32-33, discusses the above, noting that use of  9  in base  60  is unreasonable, but others also did it.  Says Sibţ [NOTE:  ţ  denotes  t  with an underdot.] al-Māridīnī (16C) used casting out  8s  and  7s  in base  60,  but that al-Kāshī (15C) used casting out  59s  in base  60.  Kūshyār does not state that the check proves the correctness of the result, though this was commonly believed, e.g. by Fibonacci and Sibţ [NOTE:  ţ  denotes  t  with an underdot.], though al-Kāshī clearly discusses the question.

Ibn Sina  = Avicenna.  Treatise on Arithmetic.  c1020.  ??NYS.  Complete rules for checking operations by casting out  9s,  attributed to the Hindus, (Smith, Isis 6 (1924) 319).  (See also Cammann -- 3 (cited in 7.N);  Datta & Singh, I, 184;  and Kaye, above, who cite F. Woepcke;  Mémoire sur la propagation des chiffres indiens;  J. Asiatique (6) 1 (1863) 27-529;  p. 502,  ??NYS.)  The DSB entry indicates that the material is in Ibn Sina's Al‑Shifâ' (The Healing) and there doesn't appear to be a translation.  Suter 89 mentions some Latin translations but I'm not clear whether they are this book or a related book.

                    Saidan's discussion says Woepcke (p. 550 [sic]) construes ibn Sina as saying that the method is Indian, but this is a contentious interpretation.  Kaye, above, says Woepcke is wrong.  Smith, History II 151, says the expression "has been variously interpreted".

Bhaskara II.  Lilivati.  1150.  Smith, History II 152, cites this in Taylor's edition, p. 7, but the method is not in Colebrooke and neither Dickson nor Datta & Singh cite it, so perhaps it is an addition in the text Taylor used??

Fibonacci.  1202.  Pp. 8-9, 20, 39, 45 (S: 24-26, 41, 67, 74) uses checks  (mod 7, 9 and 11).  On p. 8 (S: 24), he implies that if the 'proof' is right, then the calculation is correct -- see comments at Kūshyār above.

Maximus Planudes.  Ψηφηφoρια κατ' Ivδoυσ η Λεγoμεvη Μεγαλη (Psephephoria kat' Indous e Legomene Megale) (Arithmetic after the Indian method).  c1300.  (Greek ed. by Gerhardt, Das Rechenbuch des Maximus Planudes, Halle, 1865, ??NYS [Allard, below, pp. 20-22, says this is not very good].  German trans. by H. Waeschke, Halle, 1878, ??NYS [See HGM II 549;  not mentioned by Allard].)  Greek ed., with French translation by A. Allard; Maxime Planude -- Le Grand Calcul selon des Indiens; Travaux de la Faculté de Philosophie et Lettres de l'Univ. Cath. de Louvain -- XXVII, Louvain‑la‑Neuve, 1981.  Proofs by casting out 9s are given in the material on the operations of arithmetic.

Narayana Pandita (= Nārāyaņa Paņdita [NOTE:  ņ  denotes  n  with an overdot and the  d  should have an underdot.]).  Gaņita[NOTE:  ņ  denotes  n  with an underdot.]  Kaumudī (1356).  Edited by P. Dvivedi, Indian Press, Benares, 1942.  Introduction in English, p. xv, discusses the material.  Allows any modulus.  (English in Datta & Singh I 183.)

The Treviso Arithmetic = Larte de labbacho.  Op. cit. in 7.H.  1478.  F. 4v onward (p. 46 in Swetz) uses casting out  9s  as a check on many examples.  Swetz (p. 189) refers to Avicenna and the Hindus.  On f. 10v (Swetz p. 59), the anonymous author says that proving a subtraction by addition "is more rapid and also more certain than the proofs by  9s"  and he makes similar statements regarding multiplication and division.

                    On p. 323 of his Isis article, Smith says the author "gives a proof by casting out sevens".  This would be on or near f. 17v (Swetz p. 73).  I can find nothing of the sort -- the author has an example of multiplication by  7,  but he checks it by casting out  9s.

Borghi.  Arithmetica.  1484.  Ff. 8r-9r (1509: ff. 9r-9v).  Casting out  7s  and  9s.  This is applied over the next few sections, but I don't see any indication that casting out  9s  is not a certain test.  However he uses casting out  7s  more often than  9s  which may indicate that he was aware that  7s  is a more secure test than  9s.

Chuquet.  1484.  Triparty, part 1.  English in FHM 41-42.  "There are several kinds of proofs such as the proof by  9,  by  8,  by  7,  and so on by other individual figures down to  2,  ....  Of these only the proof by  9,  because it is easy to do, and the proof by  7,  because it is even more certain than that by  9  are treated here."  He then notes that these proofs are not always certain.

Pacioli.  Summa.  1494.  Ff. 20v-23v.  Discusses casting out  9s  and  7s  and notes that these tests are not sufficient.

Apianus.  Kauffmanss Rechnung.  1527.  Gives numerous examples of testing by  9s,  and also by  8s,  7s  and  6s,  in his sections on the four arithmetic operations and also under arithmetic progressions.

Recorde.  First Part.  1543.  Discusses the proof by nines in his chapters on:  addition, ff. D.i.r - D.iii.v (1668: 29-32: The proof of Addition);  subtraction, ff. F.iii.r - F.iiii.r (omitted in 1668);  multiplication, ff. G.vi.r - G.vi.v (1668: 70-72: Proof of Multiplication);  and division, ff. H.iii.v - H.iiii.v (1668: 82-84: Proof of Division).

Hutton.  A Course of Mathematics.  1798?  1833 & 1857: 6-12.  In his discussion of the basic arithmetic operations, we find on p. 7 under  To Prove Addition,  "Then, if the excess of 9's in this sum, ...., be equal to the excess of 9's in the total sum ..., the work is right."  A footnote explains the idea and is less clear as to the direction of implication being asserted: "it is plain that this last excess must be equal to the excess of 9's contained in the total sum".  The note concludes: "This rule was first given by Dr Wallis in his Arithmetic, published in the year 1657."  However, Hutton does not mention the rule under subtraction and under multiplication on pp. 10-11, he says the "remainders must be equal when the work is right."  All in all, it seems that he is surprisingly unclear for his time.

Boy's Own Book. 

The number nine: "To add a figure to any given Number, which shall render it divisible by Nine".  1828: 179-180;  1828-2: 235;  1829 (US) & 1881 (NY): 103;  1855 & 1859: 389;  1868: 429;  1880: 459.  Here the digit is actually added, but then he indicates that it can be inserted.  [Incidentally, this avoids the hazard of discovering the missing digit might be either 0 or 9.]  This section is extended and combined into Properties of certain numbers from 1868.  Cf 1843 (Paris): 342

The cancelled figure guessed.  1828: 177;  1828-2: 237;  1829 (US) & 1881 (NY): 105; 1843 (Paris): 346: A person striking a figure out of the sum of two given numbers, to tell what that figure was;  1855 & 1859: 392-393;  1868: 430-431;  1880: 460‑461.  Give a person several multiples of nine, then ask him to add two of them and strike out one digit from the total and tell you the sum of the other digits in the total.  No consideration of the case when the cancelled digit could be 0 or 9.  = Boy's Treasury, 1844, pp. 303-304.  = de Savigny, 1846, p. 346.

Boy's Own Book.  1843 (Paris): 342.  "To make any number divisible by nine, by adding a figure to it."  Only appends or inserts the necessary digit.  = Boy's Treasury, 1844, p. 299.

Lewis Carroll.  Diary entry for 8 Feb 1856,  In Carroll-Gardner, pp. 43-44.  Observes that a number minus its reverse is divisible by nine, so you can ask someone to delete a digit and show you the rest and you announce the deleted digit.  Gardner points out that one can subtract any permutation of the original digits.

Mittenzwey.  1880. 

Prob. 40-41, pp. 9 & 60-61;  1895?: 46-47, pp. 14 & 64;  1917: 46-47, pp. 13 & 58-59.  Deduce the figure deleted in  9x;  in  x  minus the sum of its digits.

1895?: prob. 118, pp. 74-75;  1917: 118, pp. 23 & 71-72.  Casting out  9s  as a method of checking multiplication, claiming that the verification shows the calculation is correct.

Parlour Games for Everybody.  John Leng, Dundee & London, nd [1903 -- BLC], p. 42: The expunged figure.  Have someone write a number, form the sum of its digits and subtract that from the given number.  Get him to strike out any figure and tell you the sum of the remaining figures.  Says that if the result is a multiple of nine, then a nine was struck out.

Peano.  Giochi.  1924.  Prob. 50, p. 13.  Take  x,  form  10x  and subtract  x.  Cancel a non-zero figure from the result and tell me the other figures.  I will tell what number you cancelled.

 

          7.L.    GEOMETRIC PROGRESSIONS

 

            See Tropfke 628.

            These also occur in 7.M, 7.S.1 and 10.A. 

            I am starting to include early problems which involve interpolation in a geometric series here -- these were normally solved by linear interpolation.  From about 1400, such problems arise in compound interest but I will omit most such problems.  See Chuquet here and  Chiu Chang Suan Ching  &  Cardan in 10.A.

 

H. V. Hilprecht.  Mathematical, Metrological and Chronological Tablets from the Temple Library of Nippur.  Univ. of Pennsylvania, Philadelphia, 1906.  Pp. 13, 28‑34, 62, 69, pl. 15, PL. IX, are about a tablet which has a geometric progression from c‑2300.  The progression is double:  an = 125 * 2n  and  604/an  for  n = 0, 1, ..., 7.  There is no summation.

Tablet K 90 of the British Museum.  A moon tablet deciphered by Hincks containing  5, 10, 20, 40, 80  followed by  80, 96, 112, 128, ..., 240.  Described in The Literary Gazette (5 Aug 1854) -- ??NYS.  Cited and described in:  Nicomachus of Gerasa: Introduction to Arithmetic; Translated by Martin Luther D'Ooge, with notes by Frank Egleston Robbins and Louis Charles Karpinski; Macmillan, London, 1926; p. 12.

Euclid.  IX: 35, 36.  This gives the general rule for the sum of a geometric progression.

The Friday Night Book  (A Jewish Miscellany).  Soncino Press, London, 1933.  Mathematical Problems in the Talmud, pp. 132-133.  The Talmud says that any one visiting a sick person takes away a sixtieth of his illness.  This led to the question of what happened if sixty people visited the person.  This was answered by saying that the visitor took away a sixtieth of the illness that the person had, i.e. the patient was left with  59/60  of his illness, so that  60  visits left him still with  (59/60)60  of his illness.  The text quoted in the source says this 'is still approximately one-quarter of the original illness', but it is  .36479.  The modern compiler adds that 'The Talmud does not indicate the method of working out the remainder after each visitor, and it is to be noted that although the summation of series was known to the Greeks, there is no mention of it anywhere in the Talmud.'  To me, this shows some confusion as I don't see that summation of series is needed!  [The Talmud was compiled in the period -300 to 500, but nothing in the source gives any more precise dating for this problem.]

Zhang Qiujian .  Zhang Qiujian Suan Jing.  Op. cit. in 7.E.  468,  ??NYS.  Mikami 42 gives:  "A horse, halving its speed every day, runs 700 miles in 7 days.  What are his daily journeys?" -- i.e.  x*(1 + 1/2 + ... + 1/64)  =  700.  Solved by adding up.

Mahavira.  850.

Chap. IV, v. 28, p. 74:  x ‑ x/2 ‑ x/4 ‑ ... ‑ x/256  =  32.

Chap. VI, v. 314, pp. 175‑176:  Let  ai+1 = r*ai + c.  He sums such terms.

Fibonacci.  1202.  Pp. 313‑316 (S: 439-443).  Man has  100  and gives away  1/10  of his wealth  12  times.  This has been described under 7.E.

Lucca 1754.  c1330.  F. 10v, pp. 36‑37.  Computes  240  &  2100  by repeated squaring.

Columbia Algorism.  c1350.  Prob. 63, pp. 84‑85.  Same as the Fibonacci, but he converts to pounds, shillings and pence!

Folkerts.  Aufgabensammlungen.  13-15C.  Four sources with progressions with ratio 7 and seven sources with ratio 12.

Chuquet.  1484.  He gives a number of such problems -- see also 7.E.

Prob. 96, English in FHM 219.  Cask of    drains so first barrel takes  1  hour,  second barrel takes  2  hours,  third barrel takes  4  hours,  ....  How long to empty?  I.e. he wants the sum of    terms of a geometric progression.  He gets the correct answer of  29.5 - 1  hours.

Prob. 97, English in FHM 219.  Man travels  1, 3, 9, ...  leagues per day.  How far has he travelled in    days?  He gets the correct answer of  (35.5-1)/2  days.

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 29, f. 66v.  Same as Fibonacci.  (H&S 59-60.)

Buteo.  Logistica.  1559.

Prob. 69, pp. 276-278.  1 + .9 + (.9)2 + ... + (.9)x  =  7.5.  The phrasing of the problem is unclear, but this is what he considers.  He interpolates linearly between  12  and  13,  getting  12.164705107,  while the exact answer is  (log .25/log .9) ‑ 1  =  12.15762696.

Prob. 84, pp. 294-296.  Relates  2i  and  5i  for  i = 1, ..., 7.  He determines  23.5  as  Ö128  which he estimates as  11⅓.  Similarly he estimates  53.5  as  279.

Jacques Chauvet Champenois.  Les Institutions de L'Arithmetique.  Hierosme de Marnef, Paris, 1578, p. 70.  ??NYS.  Problem of tailor and robe involving  4888  divided by  2  twenty times.  (French quoted in H&S 14‑15.)

van Etten.  1624.  Prob. 87 (84): Des Progressions & de la prodigieuse multiplication des animaux, des Plantes, des fruicts de l'or & de l'argent quand on va tousjours augmentant par certaine proportion,  pp. 111‑118 (177‑183).  Numerous examples including horseshoe problem and chessboard problem, with ratios 1000, 4, 50.  Henrion's Notte, p. 38, observes that there are many arithmetical errors which the reader can easily correct.  In part X: Multiplication des Hommes, he considers one of the children of Noah, says a generation takes 30 years and that, when augmented to the seventh, one family can easily produce 800,000 souls.  The 1674 English ed. has: "... if we take but one of the Children of Noah, and suppose that a new Generation of People begin at every 30 years, and that it be continued to the Seventh Generation, which is 200 years; ... then of one only Family there would be produced 111000 Souls, 305 to begin the World: ... which number springing onely from a simple production of one yearly ...."

W. Leybourn.  Pleasure with Profit.  1694.  Chap. VI, pp. 24-28: Of the Increase of Swine, Corn, Sheep, &c.  Examples with ratios 4, 40, 2, 1000, 2, mostly taken from van Etten.  Then art. VI: Of Men, discusses the repopulation of the world from Noah's children: "... if we take but one of the Children of Noah, and suppose that a New Generation of People begin at every 30 years, and that it be continued to the seventh Generation, which is 210 years; ... then, of one only family there would be produced 111305, that is, One hundred and eleven thousand, three hundred and five Souls to begin the World ....  ... such a number arising only from a simple production of only One yearly ...."  I cannot work out how 111305 arises -- the fact that he spells it out makes it unlikely to be a misprint.

Ozanam.  1694.  Prob. 8, 1696: 33-35;  1708: 29-32.  Prob. 11, 1725: 68-75.  Section II, 1778: 68-74;  1803: 70-76;  1814: ??NYS;  1840: 34-36.  A discussion of geometric progression and a mention of  1, 2, 4, ....  1778 et seq. also mention 1, 3, 9, ....

Ozanam.  1725.  Prob. 11, questions 6 & 7, 1725: 79‑82.  Prob. 3, parts 1-3, 1778: 80-82;  1803: 82-84;  1814: 72-75;  1840: 38-39.  Examples of population growth in Biblical and biological contexts.  In 1725, he has ratios of 2, 50, 3, 4, 1000, The examples vary a bit between 1725 and 1778.

Walkingame.  Tutor's Assistant.  1751.  The section Geometrical Progression gives several problems with powers of  2  and the following less common types. 

Prob. 5, 1777: p. 95;  1835: p. 103;  1860: p. 123.  Find  1 + 4 + 16 + ... + 411  farthings.

Prob. 8, 1777: p. 96;  1835: p. 104;  1860: p. 123.  Find  2 + 6 + 18 + ... + 2 x 321.  If these are pins, worth 100 to the farthing, what is the value?

Vyse.  Tutor's Guide.  1771?    The section Geometrical Progression, 1793: 35, pp. 138-143; 1799: XXXV, pp. 146-151 & Key pp. 190-192, gives several examples with doublings and triplings as well as examples with ratios of  3/2  and  10.  There is a major error in the solution of prob. 7, to find  2 + 6 + 18 + ... + 2 x 319.

Pike.  Arithmetic.  1788.  Pp. 237-239.  Numerous fairly standard examples, mostly doubling, but with examples of powers of  3  and of  10  and the following.  D. Adams, 1835, copies two examples, but not the following.

                    Pp. 239-240, no. 8.  One farthing placed at  6%  compound interest in year  0  is worth what after  1784  years?  And supposing a cubic inch of gold is worth  £53 2s 8d,  how much gold does this make?  This is very close to  2150  farthings and makes about  4 x 1014  solid gold spheres the size of the earth!

Eadon.  Repository.  1794.  P. 241, ex. 3.  Doubling  20  times from a farthing.

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 100.  10 + 102 + ... + 1011  grains of wheat, converted to bushels and value at  4s  per bushel.

(Beeton's) Boy's Own Magazine 3:6 (Jun 1889) 255  &  3:8 (Aug 1889) 351.  (This is undoubtedly reprinted from Boy's Own Magazine 1 (1863).)  Mathematical question 59.  Seller of  12  acres asks  1  farthing for the first acre,  4  for the second acre,  16 for the third acre, ....  Buyer offers  £100  for the first acre,  £150 for the second acre,  £200  for the third acre, ....  What is the difference in the prices asked and offered?  Also entered in 7.AF.

Lewis Carroll.  Sylvie and Bruno Concluded.  Macmillan, London, 1893.  Chap. 10, pp. 131‑132.  Discusses repeated doubling of a debt each year as a way of avoiding paying the debt -- "You see it's always worth while waiting another year, to get twice as much money!"  = Carroll-Wakeling II, prob. 5: A new way to pay old debts, pp. 9 & 66, where Wakeling adds some problems based on repeated doubling and gives the chessboard problem.

 

          7.L.1. 1 + 7 + 49 + ...  &  ST. IVES

 

          See Tropfke 629.

 

Papyrus Rhind, c‑1650, loc. cit. in 7.C.  Problem 79, p. 112 of vol. 1 (1927) (= p. 59 (1978)).  7 + 49 + 343 + 2401 + 16807.  (Sanford 210 and H&S 55 give Peet's English.)  Houses, cats, mice, ears of spelt, hekats.

L. Rodet.  Les prétendus problèmes d'algèbre du manuel du calculateur Égyptien (Papyrus Rhind).  J. Asiatique 18 (1881) 390‑459.  Appendice, pp. 450‑454.  Discusses this problem and its appearance in Fibonacci (below).

F. Cajori.  History of Mathematics.  2nd ed., Macmillan, 1919;  Chelsea, 1980.  P. 90 gives the legend that Buddha was once asked to compute  717.

Shakuntala Devi.  The Book of Numbers.  Orient Paperbacks (Vision Books), Delhi, 1984.  This gives more details of the Buddha story, saying it occurs in the Lalitavistara and Buddha finds the number of atoms (of which there are seven to a grain of dust) in a mile, obtaining a number of 'about 50 digits'.  Note:  758  =  1.04 x 1049.

Alcuin.  9C.  Prob. 41: Propositio de sode et scrofa.  This has sows which produce  7  piglets, but this results in a GP of ratio  8.

Fibonacci.  1202. 

Pp. 311‑312 (S: 438).  7 + ... + 117649.  Old women [The Latin is  vetule,  which is corrupt and the gender is not clear -- Sigler says  old men.], mules, sacks, loaves, knives, sheathes.  (English in:  N. L. Biggs; The roots of combinatorics; HM 6 (1979) 109‑136 (on p. 110) and Sanford 210.  I have slides of this from L.IV.20 & 21.  It is on f. 147r of L.IV.20 and f. 225r of L.IV.21.)  

P. 312 (S: 438-439).  100 + 10000 + ... + 108.  Branches, nests, eggs, birds.

Munich 14684.  14C.  Prob. XXXI, pp. 83‑84.  Refers to  7 + 49 + ... + 117649.

AR.  On p. 227, Vogel refers to an example in CLM 4390 which has not been published.

Peter van Halle.  MS. 3552 in Royal Library Brussels, beginning "Dit woort Arithmetica coomt uuter griexer spraeken ...."  1568.  F. 23v.  "There were 5 women and each woman had 5 bags but in each bag were 5 cats and each cat had 5 kittens question how many feet are there to jump with?"  Copy of original Dutch text and English translation provided by Marjolein Kool, who notes that van Halle only counts the feet on the kittens.

Josse Verniers.  MS. 684 in University Library of Ghent, beginning "Numeration heet tellen ende leert hoemen die ghetalen uutghespreken ende schrijven ...."  1584.  F. 7v.  "Item there is a house with 14 rooms and in each room are 14 beds and each bed lay 14 soldiers and each soldier has 14 pistols and in each pistol are 3 bullets Question when they fire how many men, pistols and bullets are there"  Copy of original Dutch text and English translation provided by Marjolein Kool.

Harley MS 7316, in the BM.  c1730.  ??NYS -- quoted in:  Iona & Peter Opie; The Oxford Dictionary of Nursery Rhymes; OUP, (1951);  2nd ed., 1952, No. 462, p. 377.  The Opies give the usual version with  7s,  but their notes quote Harley MS 7316 as:  "As I went to St. Ives I met Nine Wives And every Wife had nine Sacs And every Sac had nine Cats And every Cat had Nine Kittens."  The Opies' notes also cite Mother Goose's Quarto (Boston, USA, c1825), a German version with  9s  and a Pennsylvania Dutch version with  7s.

                    Halliwell, James Orchard.  Popular Rhymes & Nursery Tales of England.  John Russell Smith, London, 1849.  Variously reprinted -- my copy is Bodley Head, London, 1970.  P. 19 refers to "As I was going to St. Ives" in MS. Harl. 7316 of early 18C, but doesn't give any more details.

D. Adams.  Scholar's Arithmetic.  1801. 

                              As I was going to St. Ives,

                              I met seven wives,

                              Every wife had seven sacks,

                              Every sack had seven cats,

                              Every cat had seven kits,

                              Kits, cats, sacks and wives,

                              How many were going to St. Ives?

          No solution.

Child.  Girl's Own Book.  1842: Enigma 35, pp. 233-234;  1876: Enigma 27, pp. 196-197.  "As I was going to St. Ives,  I chanced to meet with nine old wives:  Each wife had nine sacks,  Each sack had nine cats,  Each cat had nine kits;  Kits, cats, sacks, and wives,  Tell me how many were going to St. Ives?"  Answer is  "Only myself.  As I met all the others, they of course were coming from St. Ives."  The 1876 has a few punctuation changes.

                    = Fireside Amusements.  1850: No. 48, pp. 114 & 181;  1890: No. 34, p. 102.  The 1850 solution is a little different: "Only myself.  As I was going to St. Ives, of course all the others were coming from it."  The 1890 solution differs a little more: "Only myself.  As I was going to St. Ives, all the others I met were coming from it."

Kamp.  Op. cit. in 5.B.  1877.  No. 20, p. 327.  12  women, each with  12  sticks, each with  12  strings, each with  12  bags, each with  12  boxes, each with  12  shillings.  How many shillings?

Mittenzwey.  1880.  Prob. 20, pp. 3 & 59;  1895?: 24, pp. 9 & 63;  1917: 24, pp. 9 & 57.  Man going to Stötteritz meets  9  old women, each with  9  sacks, each with  9  cats, each with  9  kittens.  How many were going to Stötteritz?  Answer is one.

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.  Pp. 105-106: Wieviel Füsse sind es?  Hunter going into the woods meets an old woman with a sack which has six cats, each of which has six young.  How many feet all together were going into the woods?  His answer is 'Only one', which has confused 'feet' with 'walker' -- this may be an obscure German usage, but I can't find it in my dictionaries.

Joseph Leeming.  Riddles, Riddles, Riddles.  Franklin Watts, 1953;  Fawcett Gold Medal, 1967.  P. 150, no. 11: "As I was going to St. Ives, I chanced to meet with nine old wives; ...."  [I don't recall any other contemporary examples using 9 as multiplier.]

Mary & Herbert Knapp.  One Potato, Two Potato ...  The Secret Education of American Children.  Norton, NY, 1976.  Pp. 107-108 gives a modern New York City version:  "There once was a man going to St. Ives Place.  He had seven wives;  each wife had seven sacks;  each sack had seven cats;  each cat had seven kits.  How many altogether were going to St. Ives Street?  One."  St. Ives has become attached to a location in New York!

Colin Gumbrell.  Puzzler's A to Z.  Puffin, 1989.  Pp. 9 & 119: As I was going ... 

          "As I was going to St Ives,   I met a man with seven wives;   And every wife had seven sons;   But they were not the only ones,   For every son had seven sisters!   Bewildered by so many misters   And by so many misses too,   I quickly cried: 'Bonjour!  Adieu!'   And hurried to another street,   Away from all their trampling feet.   Now, here's the point that puzzles me yet:   Just how many people had I met?"  Answer is 106, or 64 if there are just seven girls who are half-sisters to all the 49 sons.

About 2000, someone told me that the answer to the classic St. Ives riddle ought to me 'None' as it asks how many of the kits, cats, sacks, and wives were going to St. Ives.

Ed Pegg Jr, Marek Penszko and Michael Kleber circulated a new version in early 2001 on the Internet.  Tim Rowett has adapted one of the St Ives postcards with this new text.

                    As I was going to St Ives

                    I met a man with seven wives

                    Every wife had seven sacks

                    Every sack had seven cats

                    Every cat had seven kits

                    We traded bits

                    Each cat, sack, wife, and he

                    Took a kit.  The rest for me.

                    So now I have a kit supply.

                    How many kits did I just buy?

 

The St. Ives of the riddle is usually thought to be the one in Cornwall, but there are also St. Ives in Cambridgeshire (near Huntingdon) and in Dorset (near Ringwood) and a St. Ive in eastern Cornwall (near Liskeard).

Darrell Bates.  The Companion Guide to Devon and Cornwall.  Collins, 1976.  P. 301 says the Cornish St. Ives is named for a 6C Irish lady missionary named St. Ia who crossed the Irish Sea on a leaf.  John Dodgson [Home Town  What's behind the name; Drive Publications for the Automobile Association, Basingstoke, 1984; p. 40] agrees.

Gilbert H. Doble.  St. Ives  Its Patron Saint and its Church.  Cornish Parish Histories No. 4, James Lanham Ltd, St. Ives, 1939.  This says the lady was named Ya, Hya or Ia.  The 'v' was probably inserted due to the influence of the Breton St. Yves, with the first appearance of the form 'Ives' being in 1571.  The earliest reference to St. Hya is c1300 and says she was an Irish virgin of noble birth, who found her friends had departed for Cornwall.  As she prayed she saw a leaf in the water and touched, whereupon it grew big enough to support her and she was wafted to Cornwall, arriving before her friends.  Doble suggests that Ireland refers to Wales here.  The next mention is in 1478 and says she was the sister of St. Erth and St. Uny and was buried at St. Hy.  There is only one other old source, a mention in 1538.  There seems to be very little, if anything, known about this saint!

The Michelin Green Guide to Brittany (3rd ed., Michelin et Cie, Clermont‑Ferrand, 1995, pp. 178 & 237-238) describes the Breton St. Yves, which I had assumed to be the eponym of the Cornish St. Ives.  St. Yves (Yves Helori (1253‑1303)) was once parish priest at Louannec, Côtes-d'Armor, where a chasuble of his is preserved in the church.  His tomb is in the Cathedral of St. Tugdual in Tréguier, Côtes-d'Armor.  His head is in a reliquary in the Treasury.  He worked as a lawyer and is the patron saint of lawyers.  He was born in the nearby village of Minihy-Tréguier and his will is preserved in the Chapel there.  The Chapel cemetery contains a monument known as the tomb of St. Yves, but this is unlikely to contain him.  Attending his festival, known as a 'pardon', is locally known as 'going to St. Yves' -- !!  The Cornish and Breton stories may have influenced each other.

 

          7.L.2. 1 + 2 + 4 + ...

 

See Høyrup in 7.L.2.a for other early examples of doubling 30 times.

Chiu Chang Suan Ching.  c‑150?  Chap. III, prob. 4, pp. 28‑29.  Weaver weaves  a (1 + 2 + 4 + 8 + 16),  making  5  in all.  (English in Needham, pp. 137‑138.  Needham says this problem also occurs in Sun Tzu (presumably the work cited in 7.P.2, 4C), ??NYS.)

Alcuin.  9C.  Prob. 13: Propostio de rege et de ejus exercitu.  1 + 1 + 2 + 4 + ... + 229  =  230.  Calculations are suppressed in the Alcuin text, but given in the Bede.  Murray 167 wonders if there is any connection between this and the Chessboard Problem (7.L.2.a).

Bhaskara II.  Lilavati.  1150.  Chap. V, sect. II, v. 128.  In Colebrooke, pp. 55‑56.  2 + 4 + ... + 230.

W. Leybourn.  Pleasure with Profit.  1694.  See in 7.L.

Wells.  1698.  No. 103, p. 205.  Weekly salary doubles each week for a year:  1 + 2 + 4 + ... + 251.

Ozanam.  1725.  Prob. 11, question 5, 1725: 78‑79.  1 + 2 + ... + 231.

Walkingame.  Tutor's Assistant.  1751.  The section Geometrical Progression gives several problems with straightforward doublings -- see 7.L and 7.L.2.b for some more interesting examples.

Vyse.  Tutor's Guide.  1771?  Same note as for Walkingame.

Eadon.  Repository.  1794.  P. 241, ex. 3.  Doubling 20 times from a farthing.

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 115.  For  20  horses, is starting with a farthing and doubling up through the  19th  horse, with the  20th  free, more or less expensive than  £20  per horse? 

Boy's Own Book.

Curious calculation.  1868: 433.  1 + 2 + ... + 251  pins would weigh  628,292,358  tons and require  27,924  ships as large as the Great Eastern to carry them.

Arithmetical [sic] progression.  1868: 433.  1 + 2 + ... + 299  farthings.  Answer is wrong.

Ripley's Believe It or Not, 4th Series, 1957.  P. 15 asserts that the Count de Bouteville directed that his widow, age 20, should receive one gold piece during the first year of widowhood, the amount to be doubled each successive year she remained unmarried.  She survived 69 years without marrying!  Ripley says the Count 'never suspected the cumulative powers of arithmetical [sic!] progression'.

 

          7.L.2.a.        CHESSBOARD PROBLEM

 

          See Tropfke 630.  See 5.F.1 for more details of books on the history of chess.

 

Jens Høyrup.  Sub-scientific mathematics: Undercurrents and missing links in the mathematical technology of the Hellenistic and Roman world.  Preprint from Roskilde University, Institute of Communication Research, Educational Research and Theory of Science, 1990, Nr. 3.  (Written for: Aufsteig und Niedergang der römischen Welt, vol. II 37,3 [??].)  He discusses this type of problem, citing al-Uqlīdisī [Abû al-Hassan [the  H  should have an underdot] Ahmad[the  h  should have an underdot] Ibn Ibrâhîm al-Uqlîdisî.  Kitâb al Fuşûl [NOTE:  ş  denotes an  s  with an underdot.] fî al‑Hisâb[the  H  should have an underdot]  al‑Hindî..  952/953.  MS 802, Yeni Cami, Istanbul.  Translated and annotated by A. S. Saidan as:  The Arithmetic of Al‑Uqlīdisī; Reidel, 1978.  ??NYS.  P. 337] as saying: "this is a question many people ask.  Some ask about doubling one 30 times, and others ask about doubling it 64 times".  Høyrup says that doubling 30 times is found in Babylonia, Roman Egypt, Carolingian France, medieval Damascus and medieval India. 

                    On pp. 23-24, he describes the first two examples mentioned above and then mentions Alcuin and al-Uqlīdisī.  The last example is probably Bhaskara II.

                    A cuneiform tablet from Old Babylonian Mari [Denis Soubeyran; Textes mathématiques de Mari; Revue d'Assyriologie 78 (1984) 19-48.  ??NYS.  P. 30] has, in Høyrup's translation: "To one grain, one grain has been added:  Two grains on the first day;  Four grains on the second day;  ...."  this goes on to 30 days.  The larger amounts are not computed as numbers, but converted to larger units.  Old Babylonian is c-1700.

                    Papyrus Ifao 88 [B. Boyaval; Le P. Ifao 88: Problèmes de conversion monétaire; Zeitschrift für Papyrologie und Epigraphik 7 (1971) 165-168, Tafel VI, ??NYS] starts with 5 and doubles 30 times, again using larger units for the later stages.  Høyrup says this is a Greco-Egyptian papyrus 'probably to be dated to the Principate but perhaps as late as the fourth century' -- I am unable to determine what the Principate was.

Perelman.  FFF.  1934.  1957: prob. 52, pp. 74-80;  1979: prob. 55, pp. 92-98.  = MCBF, prob. 55, pp. 90-98.  This describes a Roman version where the general Terentius can take  1  coin the first day,  2  the second day,  4  the third day,  ..., until he can't carry any more, which occurs on the  18th  day.  A footnote says this is a translation "from a Latin manuscript in the keeping of a private library in England."  ??

Murray mentions the problem on pp. 51‑52, 155, 167, 182 and discusses it in detail in his Chapter XII: The Invention of Chess in Muslim Legend, pp. 207‑219.  He discusses various versions of the invention of chess, some of which include the doubling reward.  He describes the doubling legends in the following.

al‑Ya‘qûbî (c875).

al-Maş‘udi [NOTE:  ş  denotes an  s  with an underdot.] (943).

Firdawsî's Shâhnâma (1011).

Kitâb ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot.] [= AH] (1141) as AH f. 3b (= Abû Zakarîyâ [= H] f. 6a).

BM MS Arab. Add. 7515 (Rich) [= BM] (c1200?).

von Eschenbach (c1220).

BM Cotton Lib. MS Cleopatra, B.ix [= Cott.] (13C).

ibn Khallikan (1256).

Dante (1321).

Shihâbaddîn at‑Tilimsâni [= Man.] (c1370), which gives five versions.

Kajînâ [= Y] (16C?).

          Murray 218 mentions two treatises on the problem:

Al‑Missisî.  Tad‘îf buyût ash‑shaţranj [Note:  ţ  denotes a  t  with an underdot and the  d  should have a dot under it.].  9C or 10C.

Al‑Akfânî.  Tad‘îf ‘adad ruq‘a ash‑shaţranj [Note:  ţ  denotes a  t  with an underdot and the first  d  should have a dot under it.] .  c1340.

                    On p. 217, Murray gives 10 variant spellings of Sissa and feels that Bland's connection of the name with Xerxes is right.

                    On p. 218, he says the reward of corn would cover England to a depth of  38.4  feet.

Murray 218.  "This calculation is undoubtedly of Indian origin, the early Indian mathematicians being notoriously given to long‑winded calculations of the character."  He suggests the problem may be older than chess itself.

Al‑Ya‘qûbî.  Ta’rîkh.  c875.  Ed. by Houtsma, Leyden, 1883, i, 99‑105.  ??NYS.  Cited by Murray 208 & 212.  "Give me a gift in grains of corn upon the squares of the chessboard.  On the first square one grain (on the second two), on the third square double of that on the second, and continue in the same way until the last square." [Quoted from Murray 213.]

al-Maş‘udi [NOTE:  ş  denotes an  s  with an underdot.] (= Mas'udi  = Maçoudi).  Murûj adh‑dhahab [Meadows of gold].  943.  Translated by:  C. Barbier de Meynard & P. de Courteille as:  Les Prairies d'Or; Imprimerie Impériale, Paris, 1861.  Vol. 1, Chap. VII, pp. 159‑161.  "The Indians ascribe a mysterious interpretation to the doubling of the squares of the chessboard;  they establish a connexion between the First Cause which soars above the spheres and on which everything depends, and the sum of the square of its squares.  This number equals  18,446,744,073,709,551,615 ...."  [Quoted from Murray 210.  The French ed. has two typographical errors in the number.]  No mention of the Sessa legend.

Muhammad ibn Ahmed Abû’l-Rîhân (the  h  should have an underdot) el-Bîrûnî  (= al‑Bîrûnî  = al-Biruni).  Kitâb al‑âtâr al‑bâqîya ‘an al‑qurûn al‑halîya  (= al‑Âthâr al‑bâqiya min al‑qurûn al khâliya   = Athâr‑ul‑bákiya)  (The Chronology of Ancient Nations).  1000.  Arabic (and/or German??) ed. by E. Sachau, Leipzig, 1876 (or 1878??), pp. 138‑139.  ??NYS.  English translation by E. Sachau, William H. Allen & Co., London, 1879, pp. 134‑136.  An earlier version is:  E. Sachau; Algebraisches über das Schach bei Bîrûnî; Zeitschr. Deut. Morgenländischen Ges. 29 (1876) 148‑156, esp. 151‑155.  Wieber, pp. 113‑115, gives another version of the same text.  Computes  1 + 2 + 4 + ... + 263  as  264 ‑ 1  by repeated squaring.  Doesn't mention Sessa.  He shows the total is  2,305  mountains.  "But these are (numerical) notions that the earth does not contain."  Murray 218 gives: "which is more than the world contains." but I'm not sure if al-Biruni means the mountains or the numbers are more than earth can contain.

BM MS Arab. Add. 7515 (Rich).  Arabic MS with the spurious title  "Kitâb ash‑Shaţranj [NOTE:  ţ  denotes a  t  with an underdot.] al Başrî [NOTE:  ş  denotes an  s  with an underdot.]",  perhaps c1200.  Copied in 1257.  Described by Bland and Forbes, loc. cit. in 5.F.1 under Persian MS 211, and by Murray on p. 173.  Murray denotes it BM.

                    Bland, p. 26, says p. 6 of the MS gives the story of  Súsah ben Dáhir  and the reward.  Bland, p. 62, says the various forms of the name Sissah are corruptions of Xerxes.  Forbes, pp. 74‑76, does not mention the story or the reward.

                    Murray 217 says all the Arabic MSS include the reward problem as part of one of their stories of the invention of chess, but on pp. 173 & 211‑219, he doesn't mention the story in this MS specifically.  However, on p. 173, he notes that the spurious first page gives the calculation in 15C Arabic and again in Turkish.

Fibonacci.  1202.  Pp. 309‑310 (S: 435-437).  He induces the repeated squaring process and gets  264 ‑ 1.  He computes the equivalent number of shiploads of grain -- there is a typographic error in his result.

Wolfram von Eschenbach.  Willehalm.  c1220.  Ed. by Lachmann, p. 151, ??NYS -- quoted by Murray 755.  "Ir hers mich bevilte, der Zende ûz zwispilte ame schâchzabel ieslîch velt mit cardamôm."

Murray 755 gives several other medieval European references.

(Al-Kâdi Shemseddîn Ahmed) Ibn Khallikan.  Entry for:  Abû Bakr as‑Sûli.  In:  Kitab wafayât al‑a‘yân.  1256.  Translated by MacGuckin de Slane as:  Biographical Dictionary; (London, 1843‑1871;) corrected reprint, Paris, 1868.  Vol. III, p. 69‑73.  Sissah ibn Dâhir,  King Shihrâm  and the chessboard on pp. 69‑71.  An interpolation(?) mentions King Balhait.

BM Cotton Lib. MS Cleopatra, B.ix.  c1275.  Anglo‑French MS of c1275, described by Murray 579‑580, where it is denoted Cott.  No. 18, f. 10a, gives doubling.

Dante.  Divina Commedia: Paradiso XXVIII.92.  1321.  "Ed eran tante che'l numero loro Piu che'l doppiar degli scacchi s'imila."  [Quoted from Murray 755.]

Paolo dell'Abbaco.  Trattato di Tutta l'Arta dell'Abacho.  1339.  Op. cit. in 7.E.  B 2433 f. 21v has an  8 x 8  board with two columns filled in with powers of two, starting with  2.  Below he gives  264  and treats it as farthings and converts to danari, soldi, libri??, soldi d'oro, libri?? d'oro and then a further step that I cannot understand.  No mention of chess or a reward.

Thomas Hyde.  Mandragorias seu Historia Shahiludii, ....  (= Vol. 1 of De Ludis Orientalibus, see  4.B.5 for vol. 2.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  Prolegomena curiosa.  The initial material and the Prolegomena are unpaged but the folios of the Prolegomena are marked (a), (a 1), ....  The material is on  (d 1).r - (d 4).v, which are pages 25-32 if one starts counting from the beginning of the Prolegomena.          He mentions Wallis (see below) and arithmetic (sic!) progressions and says the story is given in  al-Safadi (Şalâhaddîn aş‑Şafadî [NOTE:  Ş,  ş  denote  S,  s,  with an underdot and the  h  should have an underdot.]  = al-Sâphadi  = AlSáphadi) (d. 1363) in his  Lâmiyato ’l Agjam  (variously printed in the text).  This must be his  Sharh [the  h  should have an underdot] Lâmîyat al‑‘Ajam  of c1350.  Hyde gives some Arabic text and a Latin translation.  Wallis gives the full Arabic text and translation.  This refers to Ibn Khallikan.  In his calculation, he uses various measures until he takes  239  grains as a granary, then  1024  granaries (=  249  grains) as a city, so the amount on the  64th square is  16384  (=  214)  cities, “but you know there are not so many cities in the whole world".  He then gives  264 - 1  correctly and converts it into cubic miles, but seems off by a factor of ten -- see Wallis, below, who gives details of the units and calculations involved, noting that al‑Safadi is finding the edge (= height) of a square pyramid of the volume of the pile of wheat.  Hyde then adds a fragment from a Persian MS,  Mu’gjizât,  which gives the story with drachmas instead of grains of wheat, but the calculations are partly illegible.  In his main text, pp. 31-52 are on the invention of the game and he gives various stories, but doesn't mention the reward.

Folkerts.  Aufgabensammlungen.  13-15C.  7 sources.

Shihâbaddîn Abû’l‑‘Abbâs Ahmad [the  h  should have an underdot] ibn Yahya [the  h  should have an underdot] ibn Abî Hajala [the  H  should have an underdot] at‑Tilimsâni alH‑anbalî [the  H  should have an underdot].  Kitâb ’anmûdhaj al‑qitâl fi la‘b ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot] (Book of the examples of warfare in the game of chess).  Copied by Muhammed ibn ‘Ali ibn Muhammed al‑Arzagî in 1446.

                    This is the second of Dr. Lee's MSS, described in 5.F.1, denoted Man. by Murray.  Described by Bland and Forbes, loc. cit. in 5.F.1 under Persian MS 211, and by Murray 175-177 (as Man) & 207‑219.  Gives five versions of the chessboard story.  The first is that of ibn Khallikan;  others come from  ar-Râghib's  Muhâdarât (the  h  and  d  should have underdots) al-Udabâ’;  from  Quţbaddîn [NOTE:  ţ  denotes  t  with an underdot.] Muhammad (the  h  should have an underdot) ibn ‘Abdalqâdir's  Durrat al-Mudî’a (the  d  should have an underdot)  and from  al-Akfânî.  One calculates in lunar years and another version calculates in miles!!

Columbia Algorism.  c1350.  Prob. 88, pp. 106‑107.  Chessboard.  Uses repeated squaring.  Copying error in the final value.

Persian MS 211.  Op. cit. in 5.F.1.  c1400.  Bland, loc. cit., p. 14, mentions "the well known story of the reward asked in grain".  Forbes' pp. 64‑66 is a translation of the episode of the Indian King Kaid's reward to Sassa.  On p. 65, Forbes mentions various interpretations of the total.

AR.  c1450.  Prob. 319, pp. 141, 180, 227.  Chessboard, with very vague story.

Benedetto da Firenze.  Trattato di Praticha d'Arismetrica.  Italian MS, c1464, Plimpton 189, Columbia University, New York.  ??NYS.  Chessboard.  (Rara, 464‑465; Van Egmond's Catalog 257-258.)

Pacioli.  Summa.  1494.  Ff. 43r-43v, prob. 28.  First mentions  1, 2, 6, 18, 54, ..., where each cell has double the previous total.  Then does usual chessboard problem, but with no story.  Computes by repeated squaring.  Converts to castles of grain.  Shows how to do  1 + 2 + 6 + 18 + ...  for  64  cells and computes the result.

Muhammad ibn ‘Omar Kajînâ.  Kitab al‑munjih fî‘ilm ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot and the second  h  should have an underdot.] (A book to lead to success in the knowledge of chess).  16C?  Translated into Persian by  Muhammad ibn Husâm ad‑Daula,  copied in 1612.  Described by Bland and Forbes and more correctly by Murray on p. 179, where it is identified as MS BM Add. 16856 and denoted Y, since it was a present from Col. Wm. Yule.

                    Bland, p. 20, mentions  Sísah ben Dáhir al Hindi  and the reward claimed in grain.  "The geometrical progression of the sixty four squares ... is computed here at full length, commencing with a Dirhem on the first square, and amounting to two thousand four hundred times the size of the whole globe in gold."

                    Forbes describes this on pp. 76‑77 and in the note on p. 65, where he computed the reward to make a cube of gold about 6 miles along an edge.  He says the above Persian value is wrong somewhere, but he hasn't been able to see the original.  [I can't tell if he means the Persian or the Arabic MS.  If a dirhem was the size of an English  2p  coin or an American quarter, the reward is about  2 x 104 km3,  compared to earth's volume of about 1012 km3.  The reward would make a cube about  27 km on an edge or about 17 miles on an edge.]  Murray doesn't refer to this MS specifically.

Ian Trenchant.  L'Arithmetique.  Lyons, 1566, 1571, 1578, ...   ??NYS.  1578 ed., p. 297.  1, 3, 9, 27.  (H&S 91 gives French and English and says similar appear in Vander Hoecke (1537), Gemma Frisius (1540) and Buteo (1556).)

Clavius.  ??NYS.   Computes number of shiploads of wheat required.  (H&S 56.)

van Etten.  1624.  Prob. 87, pp. 111‑118 (not in English editions).  Includes chessboard as part XI, on p. 117.  Henrion's Notte, p. 38, observes that there are many arithmetical errors in prob. 87 which the reader can easily correct.

John Wallis.  (Mathesis Universalis.  T. Robinson, Oxford, 1657.  Chap. 31.)  = Operum Mathematicoroum.  T. Robinson, Oxford, 1657.  Part 1, chap. 31: De progressione geometrica, pp. 266-285.  This includes the story of Sessa and the Chessboard in Arabic & Latin, taken from  al-Safadi,  c1350, giving much more text than Hyde (see above) does and explaining the units and the calculation, showing that  al-Safadi's  60 miles should be about 6 miles and this is the edge and height of a square pyramid of the same volume as the wheat.  He then computes all the powers of two up to the  63rd and adds them!  John Ayrton Paris [Philosophy in Sport made Science in Earnest; (Longman, Rees, Orme, Brown, and Green; London, 1827);, 8th ed., Murray, 1857, p. 515] says Wallis got 9 English miles for the height and edge.

Anonymous proposer; a Lady, solver, with Additional Solution.  Ladies' Diary, 1709-10  =  T. Leybourn, I: 3-4, quest. 6.  64 diamonds sold for  1 + 2 + 4 + ... + 263  grains of wheat.  Suppose a pint of wheat contains 10,000 grains, a bushel of wheat weighs half a hundredweight [a hundredweight is 112 lb], the value is 5s per bushel, a horse can carry 1000 lb and a ship can carry 100 tons, then how much is the payment worth and how many horses or ships would be needed to carry it?

Euler.  Algebra.  1770.  I.III.XI: Questions for practice, no. 3, p. 170.  Payment to Sessa, converted to value.

Ozanam‑Montucla.  1778.  Prob. 3, 1778: 76-78;  1803: 78-81;  1814: 70-72;  1840: 37‑38.  Problem wants the results of doublings, with no story.  Discussion gives the story of  Sessa,  taken from  Al‑Sephadi.  Gives various descriptions of the pile of grain, citing Wallis for one of these and says it would cover three times the area of France to a depth of one foot.

Eadon.  Repository.  1794.  Pp. 369-370, no. 11.  Indian merchant selling  64  diamonds to a Persian king for grains of wheat, in verse.  Supposing a pint holds  10000  grains and a bushel of  64  pints weighs  50  pounds, how many horse loads (of a thousand pounds each) does this make?  How many ships of  100  tons capacity?

Manuel des Sorciers.  1825.  P. 84.  ??NX  Chessboard.

The Boy's Own Book.  The sovereign and the sage.  1828: 182;  1828‑2: 238;  1829 (US): 106;  1855: 393;  1868: 431.  Uses 63 doublings for no reason.

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 26,  1857: 82.  Reward to Sessa for inventing chess.  Takes a pint as 7680 grains and 512 pints as worth  27/6  to value the reward at  6.45 x 1012 £.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 200.  King conferring reward on a general.  Computes number of seers, which contain 15,360 grains, and the value if 30 seers are worth one rupee.

Nuts to Crack XIV (1845), no. 73.  The sovereign and the sage.  Almost identical to Boy's Own Book.

Magician's Own Book.  1857.  The sovereign and the sage, pp. 242-243.  A simplified version of Ozanam-Montucla.  = Book of 500 Puzzles, 1859, pp. 56-57.  = Boy's Own Conjuring Book, 1860, p. 213.

Vinot.  1860.  Art. XVIII: Problème des échecs, pp. 36-37.  Uses grains of wheat and says there are  20,000  grains in a litre.  Says the reward would cover France to a depth of  1.6 m.  He gives the area of France as  9,223,372 km2.

[Chambers].  Arithmetic.  Op. cit. in 7.H.  1866?  P. 267, quest. 54.  Story of Sessa with grains of wheat.  Suppose  7680  grains make a pint and a quarter is worth  £1 7s 6d,  how much was the wheat worth?

James Cornwell & Joshua G. Fitch.  The Science of Arithmetic: ....  11th ed., Simpkin, Marshall, & Co., London, et al., 1867.  (The 1888 ed. is almost identical to this, so I suspect they are close to identical to the 2nd ed. of 1856.)  Exercises CXLIII, no. 6, pp. 299 & 371.  Chessboard problem with no story, assumes  7680  grains to a pint.

Mittenzwey.  1880.  Prob. 94, pp. 19 & 68;  1895?: 109, pp. 24 & 71;  1917: 109, pp. 22 & 68.  King Shehran rewarding Sessa Eba Daher, according to Asephad.

Cassell's.  1881.  P. 101: Sovereign and the sage.  Uses sage's and king's common age of 64, with no reference to chessboard.

Lucas.  L'Arithmétique Amusante.  1895.  Le grains du blé de Sessa, pp. 150-151.  Says it would take 8 times the surface of the earth to grow enough grain.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  The chess inventor's reward, pp. 112‑114.  Assumes  7489 29/35  grains to the pound, with  112  pounds to the cwt.,  20  cwt. to the ton and  1024  tons to the cargo, getting  1,073,741,824  cargoes, less one grain.  The number of grains is chosen so that a ton contains exactly  224  grains of rice and the answer is  230  cargoes less one grain.

 

          7.L.2.b.        HORSESHOE NAILS PROBLEM

 

          See Tropfke 632.

 

AR.  c1450.  Prob. 274, 317, 318, 353.  Pp. 125, 140, 154, 180, 227.

274:  32  horseshoe nails.

317:  16  cow nails.

318:  32  horseshoe nails.

353:  28  nails --  text is obscure.

Riese.  Rechenung nach der lenge ....  1525.  (Loc. cit. under Riese, Die Coss.)  Prob. 32, p. 20.  32  horseshoe nails.

Christoff Rudolff.  Künstliche rechnung mit der ziffern und mit den zal pfenninge.  Vienna, 1526;  Nürnberg, 1532, 1534, et seq.  F. N.viii.v.  ??NYS.  32  horseshoe nails.  (H&S 56 gives German.)

Apianus.  Kauffmanss Rechnung.  1527.  Ff. D.vi.r - D.vi.v.  32 horseshoe nails.

Anon.  Trattato d'Aritmetica, e del Misure.  MS, c1535, in Plimpton Collection, Columbia Univ.  ??NYS.  Horseshoe problem:  1 + 2 + 4 + ... + 223.  (Rara, 482‑484, with reproduction on p. 484.)

Recorde.  First Part.  1543.  Ff. L.ii.r - L.ii.v (1668: 141-142: A question of an Horse).  24  horseshoe nails.

Buteo.  Logistica.  1559.  Prob. 34, pp. 237-238.  24  horseshoe nails.  (H&S 56.)

van Etten.  1624.  Prob. 87, pp. 111‑118 (not in English editions).  Includes  24  horseshoe nails problem as part VII on p. 115.  Henrion's Notte, p. 38, observes that there are many arithmetical errors in prob. 87 which the reader can easily correct.

Wells.  1698.  No. 102, p. 205.  24  nails.

Ozanam.  1725.  Prob. 11, question 4, 1725: 77‑78 & 80.  Part of prob. 3, 1778: 79-80;  1803: 81;  1814: 72;  1840: 38.  24  nails -- first asks for the price of the  24th, then the total.

Dilworth.  Schoolmaster's Assistant.  1743.  P. 96, no. 1.  32  nails, starting with a farthing.

Walkingame.  Tutor's Assistant.  1751.  Geometrical Progression, prob. 6, 1777: p. 95;  1835: p. 103;  1860: p. 123.  32  nails,  one farthing for the first,  wants total, which he gives in £ s/d.

Mair.  1765?  P. 493, ex. III.  "What will a horse cost by tripling the 32 nails in his shoes with a farthing?"  I.e., 32 horseshoe nails, but with tripling! 

Euler.  Algebra.  1770.  I.III.XI.511, p. 166.  Horse to be sold for the value of  32  nails,  1  penny for the first, ....

Vyse.  Tutor's Guide.  1771?  Prob. 2, 1793: p. 140; 1799: p. 148 & Key p. 190.  36  horseshoe nails.  Want value of last one, starting  ¼, ½, 1, ....

Bullen.  Op. cit. in 7.G.1.  1789.  Chap. 31, prob. 1, p. 215.  Same as Walkingame.

Bonnycastle.  Algebra.  10th ed., 1815.  P. 80, no. 6.  Same as Walkingame.

Manuel des Sorciers.  1825.  P. 84.  ??NX  24 horseshoe nails.

The Boy's Own Book.  The horsedealer's bargain.  1828: 182;  1828-2: 238;  1829 (US): 106;  1843 (Paris): 346;  1855: 393‑394;  1868: 431-432.  Wants value of  24th nail, starting with a farthing.  = Boy's Treasury, 1844, p. 304.  = de Savigny, 1846, p. 292: Le marché aux chevaux.

Nuts to Crack XIV (1845), no. 74.  The horsedealer's bargain.  Almost identical to Boy's Own Book.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 199.  32  nails.

[Chambers].  Arithmetic.  Op. cit. in 7.H.  1866?  P. 268, quest. 64.  24  nails, starting with a farthing.  Finds total.

Mittenzwey.  1880.  Prob. 93, pp. 19 & 68;  1895?: 108, pp. 23-24 & 71;  1917: 109, pp. 21-22 & 68.  32 horseshoe nails, starting at 1 pf.

Cassell's.  1881.  P. 101: The horse‑dealer's bargain.  24  nails, unclear, but uses  223  farthings as the answer.

 

          7.L.2.c.        USE OF 1, 2, 4, ... AS WEIGHTS, ETC.

 

            See Tropfke 633.

            A special case of this is the use of such amounts to make regular unit payments, e.g. rent of one per day.  See:  Knobloch;  Fibonacci;  BR;  Widman;  Tartaglia;  Gori;  Les Amusemens.

 

Eberhard Knobloch.  Zur Überlieferungsgeschichte des Bachetschen Gewichtsproblems.  Sudhoffs Archiv 57 (1973) 142-151.  This describes the history of this topic and 7.L.3 from Fibonacci to Ozanam (1694).  He gives a table showing occurrences of:  powers of two,  powers of three,  weight problem,  payment problem.  I am not entirely clear what he means in the first three cases -- I would have two kinds of weight problem corresponding to the first two cases and perhaps some of his references in the first case are listed under 7.L.2.  However, the last case clearly corresponds to the problem of making a payment of one unit per day as in Fibonacci.  He lists this as occurring in Fibonacci,  BR,  Widmann  and  Tartaglia  and notes that Sanford, H&S 91, only noticed Fibonacci.  Knobloch notes that Ball's citations are not very good and that Ahrens' note about them does not go much deeper.  I have a number of references listed below which were not available to Knobloch.

 

Fibonacci.  1202.  P. 298 (S: 421).  Uses  5  ciphi of value  1, 2, 4, 8, 15  to pay a man at rate of  1  per day for  30  days.

BR.  c1305.  No. 93, pp. 112‑113.  Use of  1, 2, 4  as payments at rate of one per year for  7  years.

Widman.  Op. cit. in 7.G.1.  1489.  Ff. 138v-139r.  ??NYS -- Knobloch says he uses values of  1, 2, 4, 8, 16  to pay for  31  days.

Pacioli.  Summa.  1494.  Ff. 97v-98r, no. 35.  Use five cups to pay daily rent for 30 days.  Uses cups of weight  1, 2, 4, 8, 15.  In De Viribus, c1500, F. XIIIv, item 86 in the Indice for the third part is: De 5 tazze, diversi pesi ogni di paga l'oste (Of 5 cups of diverse weights to pay the landlord every day)  = Peirani 20,  but at the end Pacioli says this problem is in 'libro nostro', i.e. the Summa.  Cf Agostini, p. 6.

Tartaglia.  General Trattato, 1556, part 2, book 1, chap. 16, art. 32: Di una particolar proprieta della progression doppia geometrica, p. 17v.  Weights:  1, 2, 4, 8, ...  (See MUS I 89.).  Also does payments with  1, 2, 4, 8, 16, 29.  Knobloch also refers to art. 33-35 -- ??NYS -- and notes that the folios are misnumbered, but miscites 'doppia' as 'treppia' here.  This covers the powers of  3  also.

Buteo.  Logistica.  1559.  Prob. 91, pp. 309-312.  Use of  1, 2, 4, 8, 16, ...  as weights.  (Cited by Knobloch.)

Knobloch also cites  Ian Trenchant (1566),  Daniel Schwenter (1636),  Franz van Schooten (1657).

Gori.  Libro di arimetricha.  1571.  Ff. 71r‑71v (p. 76).  Use of cups weighing  1, 2, 4  to make all weights through  7,  to pay for days at one per day.

Bachet.  Problemes.  1612.  Addl. prob. V & V(bis), 1612: 143-146;  as one prob. V, 1624: 215-219;  1884: 154-156.  Mentions  1, 2, 4, 8, 16  and cites Tartaglia, art. 32 only.  This was omitted in the 1874 ed.  Knobloch cites 1612, pp. 127 & 143-146, but but p. 127 is Addl. prob. I, which is a Chinese Remainder problem?

van Etten/Henrion.  1630.  Notte to prob. 53, pp. 20‑21.  Refers to Bachet and compares with ternary weights.

Ozanam.  1694. 

Prob. 8, 1696: 33-35;  1708: 29-32.  Prob. 11, 1725: 68-75.  Section II, 1778: 68-74;  1803: 70-76;  1814: ??NYS;  1840: 34-36.  A discussion of geometric progression and a mention of  1, 2, 4, ...,  without any application to weighing.  1778 et seq. also mentions 1, 3, 9, ....

Prob. 12, vol. II, 1694: 18-19 (??NYS).  Prob. 12, 1696: 284 & fig. 131, plate 46, p. 275;  1708: 360 & fig. 26, plate 14, opp. p. 351.  Prob. 8, vol. II, 1725: 345‑348 & fig. 131, plate 46 (42).  Prob. 14, vol. I, 1778: 206‑207;  1803: 201-202.  Prob. 13, vol. II, 1814: 174-175;  1840: 90‑91.  Gives double and triple progressions.  Knobloch gives the 1694 citation.  The figure is just a picture of a balance and is not informative -- the same figure is also cited for other sets of weights.

Les Amusemens.  1749.  Prob. 8, p. 128.  Coins of value  1, 2, 4, 8, 15  to pay for a room at a rate of  1  per day for  30  days.

The Bile Beans Puzzle Book.  1933.  No. 42: Money juggling.  Place £1000 in 10 bags so any amount can be paid without opening a bag.  Solution has bags of: 

          1, 2, 4, 8, 16, 32, 63, 127, 254, 493.  I cannot see why the solution isn't: 

          1, 2, 4, 8, 16, 32, 64, 128, 256, 489.

 

          7.L.3. 1 + 3 + 9 + ... AND OTHER SYSTEMS OF WEIGHTS

 

          See MUS I 88-98;  Tropfke 633.

 

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 125ff., no. 43.  ??NYS -- Hermelink, op. cit. in 3.A, and Tropfke 634-635 say this gives  1, 3, 9, ..., 19683  =  39  to weigh up to 10,000.

Fibonacci.  1202. 

P. 297 (S: 420-421).  Weights  1, 3, 9, 27  and  1, 3, 9, 27, 81  'et sic eodem ordine possunt addi pesones in infinitum' [and thus in the same order weights can be added without end].

Pp. 310‑311 (S: 437).  Finds  1 + 2 + 6 + 18 + ... + 2*362  =  363  by repeated squaring to get  364  and then divides by  3.

Gherardi.  Libro di ragioni.  1328.  P. 53.  Weights  1, 3, 9, 27, 80  to weigh up through  120.

Columbia Algorism.   c1350.  Prob. 71, pp. 92‑93.  Weights  1, 3, 9, 27.

AR.  c1450.  Prob. 127, pp. 67 & 182.  1, 3, 9, 27.

Chuquet.  1484.  Prob. 142. 

1, 2, 7  to weigh up to  10  (English in FHM 225);

1, 2, 4, 15  up to  22;

1, 3, 9  up to  13;

1, 3, 9, 27  up to  40;  (original of this and the next case reproduced on FHM 226)

1, 3, 9, 27, 81  up to  121.  Knobloch says Chuquet gives a general solution, but I don't see that Chuquet is general.

Pacioli.  Summa.  1494.  Ff. 97r-97v, no. 34.  General discussion of  1, 3, 9, 27, 81, 243, ....  In De Viribus, c1500, F. XIIIv, item 85 in the Indice for the third part is: De far 4 pesi che pesi fin 40 (To make four weights which weigh to 40)  = Peirani 20,  but at the end Pacioli says this problem is in 'libro nostro', i.e. the Summa.  Cf Agostini, p. 6.

Cardan.  Practica Arithmetice.  1539.  Chap. 65, section 12, ff. BB.vii.r - BB.vii.v (p. 136).  Weights  1, 3, 9, 27, ....

Knobloch also cites:  Giel vanden Hoecke (1537);  Gemma Frisius (1540);  Michael Stifel (1553);  Simon Jacob (1565);  Ian Trenchant (1566);  Daniel Schwenter (1636);  Kaspar Ens (1628);  Claude Mydorge (1639);  Frans van Schooten (1657).

Tartaglia, 1556 -- see in 7.L.2.c.

Buteo.  Logistica.  1559.  Prob. 91, pp. 309-312.  Use of  1, 3, 9, 27, ...  as weights.  (Cited by Knobloch.)

John [Johann (or Hanss) Jacob] Wecker.  Eighteen Books of the Secrets of Art & Nature Being the Summe and Substance of Naturall Philosophy, Methodically Digested ....  (As:  De Secretis Libri XVII; P. Perna, Basel, 1582 -- ??NYS)  Now much Augmented and Inlarged by Dr. R. Read.  Simon Miller, London, 1660, 1661 [Toole Stott 1195, 1196];  reproduced by Robert Stockwell, London, nd [c1988].  Book XVI -- Of the Secrets of Sciences: chap. 19 -- Of Geometricall Secrets: To poyse all things by four Weights, p. 289.  1, 3, 9, 27;  1, 3, 9, 27, 81;  1, 3, 9, 27, 81, 243.  Cites Gemma Frisius.

Bachet.  Problemes.  1612.  Addl. prob. V & V(bis),  1612: 143-146;  as one prob. V, 1624: 215-219;  1884: 154‑156.  Weights:  1, 3, 9, 27, ...,  and the general case via the sum of a GP.  In the 1612 ed., Bachet only does the cases  40  and  121,  then does the general case.  Knobloch cites 1612, pp. 127 & 143-146, but p. 127 is Addl. prob. I, which is a Chinese Remainder problem.  He also says this is the first proof of the problem, excepting Chuquet, though I don't see such in Chuquet.

van Etten.  1624.  Prob. 53 (48), pp. 48‑49 (72).  1, 3, 9, 27;  1, 3, 9, 27, 81;  1, 3, 9, 27, 81, 243.  Henrion's Notte, pp. 20‑21, refers to Bachet and compares this with binary weights.

Ozanam.  1694. 

Prob. 8, 1696: 33-35;  1708: 29-32.  Prob. 11, 1725: 68-75.  Section II, 1778: 68-74;  1803: 70-76;  1814: ??NYS;  1840: 34-36.  A discussion of geometric progression and a mention of  1, 2, 4, ...,  without any application to weighing.  1778 et seq. also mentions 1, 3, 9, ....

Prob. 12, vol. II, 1694: 18-19 (??NYS).  Prob. 12, 1696: 284 & fig. 131, plate 46, p. 275;  1708: 360 & fig. 26, plate 14, opp. p. 351.  Prob. 8, vol. II, 1725: 345‑348 & fig. 131, plate 46 (42).  Prob. 14, vol. I, 1778: 206‑207;  1803: 201-202.  Prob. 13, vol. II, 1814: 174-175;  1840: 90‑91.  Gives double and triple progressions.  Knobloch gives the 1694 citation.  The figure is just a picture of a balance and is not informative -- the same figure is also cited for other sets of weights.

Les Amusemens.  1749.  Prob. 18, p. 140: Les Poids.  Weights  1, 3, 9, 27, 81, 243.

Vyse.  Tutor's Guide.  1771?  Prob. 2, 1793: p. 303; 1799: p. 316 & Key pp. 356-357.  Weights  1, 3, 9, 27.

Bonnycastle.  Algebra.  1782.  P. 202, no. 13.  1, 3, 9, 27, 81, 243, 729, 2187  to weigh to  29 hundred weight -- an English hundred weight is  112  pounds.  c= 1815: p. 230, no. 33.  1, 3, 9, 27, 81  to weigh to a hundred weight.

Eadon.  Repository.  1794.  Pp. 297-298, no. 1.  1, 3, 9, ..., 313.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  P. 50, no. 77: To find the least Number of Weights that will weigh from One Pound to Forty.  1, 3, 9, 27.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 24, pp. 20 & 78‑79.  1, 2, 4, 8, 16, ...  and  1, 3, 9, 27, 81, ....

Rational Recreations.  1824.  Exer. 22, p. 131.  1, 3, 9, 27.

Endless Amusement II.  1826? 

P. 105: To find the least Number of Weights that will weigh from One Pound to Forty.  = Badcock.

Prob. 24, p. 201.  To name five weights, which, added together, make 121 pounds; by means of which may be weighed any intermediate weight, excluding fractions.  1, 3, 9, 27, 81.  = New Sphinx, c1840, p. 137.

Young Man's Book.  1839.  P. 242.  To name five weights, ....  Identical to Endless Amusement II, p. 201.

Boy's Own Book.  1843 (Paris): 346-347.  "To find the least number of weights which will weigh any intermediate weight, from one pound to forty, exclusive of fractions.  Indicates that one can continue the progression."  = Boy's Treasury, 1844, p. 304. 

Family Friend 1 (1849) 150 & 178.  Problems, arithmetical puzzles, &c. -- 5.  How can one divide  40  lb into four weights to weigh every value from  1  to  40?  Proposer says he can't do it.  Usual answer, but one solver gives weights  6, 10, 11, 13.  However, the latter weights will not weigh  22, 25, 26, 31, 32, 33, 35, 36, 37, 38, 39.

Boy's Own Book.  To weigh from one to forty pounds with four weights.  1855: 392;  1868: 430.  1, 3, 9, 27.  No generalizations.

Magician's Own Book.  1857.  The mathematical blacksmith, p. 230.  1, 3, 9, 27  and this can be continued.  = Boy's Own Conjuring Book, 1860, p. 200.

Todhunter.  Algebra, 5th ed.  1870.  Miscellaneous Examples, no. 175, p. 559.  Show that  1, 1, 5, 5, 25, 25, 125, 125  can weigh any integral amount up to 312.  No solution given.

F. J. P. Riecke.  Op. cit. in 4.A.1, vol. 3, 1873.  Art. 3: Die Zauberkarten, p. 13.  Uses balanced ternary for divination.  See under 7.M.4

Mittenzwey.  1880. 

Prob. 107, pp. 22 & 74;  1895?: 124, pp. 26 & 76-77;  1917: 124, pp. 24 & 74-75.  Stone of weight  40  breaks into four parts which weigh up through 40.  Solution is a table showing how to weigh up through  20.  1895? adds solutions for weights  1, 2, 3, 34;  5, 9, 10, 16;  allowing several weighings, e.g.  4  is obtained by weighing out  2  twice.

1895?: prob. 125, pp. 26 & 77;  1917: 125, pp. 24 & 75.  Five weights to get through 121.

P. A. MacMahon.  Certain special partitions of numbers.  Quart. J. Math. 21 (1886) 367‑373.  Very technical.

P. A. MacMahon.  Weighing by a series of weights.  Nature 43 (No.1101) (4 Dec 1890) 113‑114.  Less technical description of the above work.

Lucas.  L'Arithmétique Amusante.  1895.  Pp. 166-168.  Notes that pharmacists, etc. use weights:  1, 1, 2, 5, 10, 10, 20, 50, 100, 100, 200, 500, 1000, 1000, 2000, 5000, ....  Discusses ternary.

Wehman.  New Book of 200 Puzzles.  1908.  P. 49.  1, 3, 9, 27, 81.

Ahrens.  MUS I.  1910.  Pp. 88-98 discusses this and some generalizations like MacMahon's.

In 2002, Miodrag Novaković told me that a student had told him how to determine twice as many integral weights with the same number of weights.  E.g., he used weights  2, 6, 18, 54  to determine integral weights  1, 2, ..., 81.  One can get exact balancing for the even values:  2, 4, ..., 80.  The odd values fall between two consecutive even amounts, so if a package weighs more than 6 but less than 8, we deduce it weighs 7!

 

          7.M.   BINARY SYSTEM AND BINARY RECREATIONS

 

            The binary system has several origins.

a)         Egyptian  &  Russian peasant multiplication.

b)         Weighing -- see 7.L.2.c.

c)         Binary divination -- see 7.M.4.

d)         The works below.

 

            See also:  5.E.2 for Memory Wheels;  5.F.4 for circuits on the  n‑cube;  5.AA for an application to card-shuffling;  7.AA.1 for Negabinary.

 

Anton Glaser.  History of Binary and Other Nondecimal Numeration.  Published by the author, 1971;  (2nd ed., Tomash, Los Angeles, 1981).  General survey, but has numerous omissions -- see the review by Knuth at Harriot, below, and MR 84f:01126.  He has no references to early Chinese material.

 

Shao Yung.  c1060.  Sung Yuan Hsüeh An, chap. 10.  Fu‑Hsi diagram of the 64 hexagrams of the I‑Ching, in binary order.  A version appears in Leibniz‑Briefe 105 (Bouvet) Bl. 27r/28r in the Niedersächsische Landesbibliothek, Hannover.  Needham, vol. 2, p. 341, notes that this had only been published in Japanese and Chinese up to 1956.  See Zacher & Kinzô below for reproductions.  Also reproduced in:  E. J. Aiton; Essay Review [of Zacher, below]; Annals of Science 31 (1974) 575‑578.

Chu Hsi.  Chou I Pen I Thu Shuo.  12C.  Fu‑Hsi Liu‑shih‑ssu Kua Tzhu Hsü (Segregation Table of the symbols of the Book of Changes) -- reproduced in Hu Wei's I Thu Ming Pien.  An illustration is given in Needham, vol. 2, fig. 41 = plate XVI, opp. p. 276 -- he says it is based on the original chart of Shao Yung and that Tshai Chhen (c1210) gave a simplified version.  Also in Kinzô and in Aiton & Shimao, below.  Shows the alternation of 0 and 1 in each binary place. 

Thomas Harriot.  Unpublished MS.  c1604.  Described by  J. W. Shirley; Binary numeration before Leibniz; Amer. J. Physics 19 (1951) 452‑454;  and by  D. E. Knuth; Review of 'History of Binary and Other Nondecimal Numeration'; HM 10 (1983)) 236‑243.  This shows some binary calculation.  Shirley reproduces BM: Add MSS 6786, ff. 346v‑347r.  Knuth cites 6782, 1r, 247r;  6786, 243v, 305r, 346v, 347r, 516v;  6788, 244v.

Francis Bacon.  Of the Advancement of Learning.  1605.  ??NYS.  Describes his binary 5‑bit coding.

Francis Bacon.  De augmentis scientarum.  1623.  ??NYS.  Full description of his coding.  He does not have any arithmetic content, so he is not really part of the development of binary.

John Napier.  Rabdologiae.  Edinburgh, 1617.  ??NYS.  Describes binary as far as extracting square roots.  William F. Hawkins; The Mathematical Work of John Napier (1550‑1617); Ph.D. thesis, Univ. of Auckland, 1982, ??NYS, asserts this is THE invention of the binary system.

G. W. Leibniz.  De Progressione Dyadica.  3pp. Latin MS of Mar 1679.  Facsimile and translation into German included in:  Herrn von Leibniz' Rechnung mit Null und Eins; Siemens Aktiengesellschaft, (1966), 2nd corrected printing, 1969; facsimile between pp. 20 & 21 and translation on pp. 42-47, with an essay by Hermann J. Greve: Entdeckung der binären Welt on pp. 21-31.  His first, unpublished, MS on the binary system, showing all the arithmetic processes.

H. J. Zacher.  Die Hauptschriften zur Dyadik von G. W. Leibniz.  Klosterman, Frankfurt, 1973.  Gathers almost all the Leibniz material, notably omitting the above 1679 paper.  He does reproduce the Fu‑Hsi diagram sent by Bouvet (cf. Shao Yung above).  However Leibniz's letter of 2 Jan 1697 to Herzog Rudolf Augustus, in which he gives his drawing of his plan for a medallion commemorating the binary system, is now lost, but it was published in 1734.

G. W. Leibniz.  Two Latin letters on the binary system, 29 Mar 1698  &  17 May 1698, recipient not identified, apparently the author of a book in 1694 which occasioned Leibniz's correspondence with him.  Opera Omnia, vol. 3, 1768, pp. 183-190.  Facsimile and translation into German included in:  Herrn von Leibniz' Rechnung mit Null und Eins; Siemens Aktiengesellschaft, 2nd corrected printing, 1966; facsimile between pp. 40 & 41 and translation on pp. 53-60.  On p. 26, it seems to say the second letter was sent to Johann Christian Schulenberg. 

G. W. Leibniz.  Explication de l'arithmètique binaire.  Histoire de l'Academie Royale des Sciences 1703 (1705) 85-89.  Facsimile and translation into German included in:  Herrn von Leibniz' Rechnung mit Null und Eins; Siemens Aktiengesellschaft, 2nd corrected printing, 1966; facsimile between pp. 32 & 33 and translation on pp. 48-52.  Illustrates all the arithmetic operations and discusses the Chinese trigrams of 'Fohy' and his correspondence with Father Bouvet in China.

G. W. Leibniz.  Letter of 1716 to Bouvet.  ??NYS -- cited in Needham, vol. 2, p. 342.  Fourth section is: Des Caractères dont Fohi, Fondateur de l'Empire Chinois, s'est servi dans ses Ecrits, et de l'Arithmétique Binaire.

Gorai Kinzô.  Jukyô no Doitsu seiji shisô ni oyoboseru eikyô (Influence of Confucianism on German Political Thought [in Japanese]).  Waseda Univ. Press, Tokyo, 1929.  ??NYS.  First publication of Leibniz's correspondence with Bouvet which led to the identification of the Fu Hsi diagram with the binary numbers.  Gives a redrawn Fu‑Hsi diagram and a segregation table.

E. J. Aiton & Eikon Shimao.  Gorai Kinzô's study of Leibniz and the I Ching hexagrams.  Annals of Science 38 (1981) 71‑92.  Describes the above work.  Reproduces Kinzo's Fu‑Hsi diagram and segregation table.

 

Ahrens.  MUS I.  1910.  24-104 discusses numeration systems in general and numerous properties of binary and powers of 2.

Gardner.  SA (Aug 1972)  c= Knotted, chap. 2.  General survey of binary recreations.  The material in the book is much expanded from the SA column.

 

          7.M.1.          CHINESE RINGS

 

          See MUS I 61-72;  S&B 104‑107, 111 & 135.

          See also 4.A.4, 11.K.1.

          In various places and languages, the following names are used:

Chinese Rings

Chainese Rings  [from www.tama.or.jp/~tane, via Dic Sonneveld, 13 Nov 2002]

Cardan's Rings,  but Cardan called it  Instrumentum ludicrum

Ryou-Kaik-Tjyo  or  Lau Kák Ch'a  =  Delay-guest-instrument

Kau Tsz' Lin Wain  =  Nine connected rings

Chienowa  =  Wisdom rings

Kyūrenkan  =  Nine connected rings

Lien nuan chhuan  [from www.roma.unisa.edu.au, via Dic Sonneveld, 13 Nov 2002]

Tarriers  or  Tarriours

Tiring Irons  or  Tyring Irons  or  Tarrying Irons

The Puzzling Rings

The Devil's Needle

Complicatus Annulis  [Wallis]

Baguenaudier  spelled various ways, e.g.  Baguenodier

Juego del ñudo Gordiano

Меледа  [Meleda]

Наран-шина  [Naran-shina] (stirrup ring toy)

Zauberkette

Magische Ringspiel

Nürnberger Tand

Grillenspiel

Armesünderspiel

Zankeisen

Nodi d'anelli

 

S. N. Afriat.  The Ring of Linked Rings.  Duckworth, London, 1982.  This is devoted to the Chinese Rings and the Tower of Hanoi and gives much of the history.

 

Sun Tzu.  The Art of War.  c-4C.  With commentary by Tao Hanzhang.  Translated by Yuan Shibing.  (Sterling, 1990);  Wordsworth, London, 1993.  In chap. 5: Posture of Army, p. 109, the translator gives:  "It is like moving in a endless circle"  In the commentary, p. 84, it says:  "their interaction as endless as that of interlocked rings."  Though unlikely to refer to the puzzle,, this and the following indicate that interlocked rings was a common image of the time.

Needham, vol. 2, pp. 189-197, describes the paradoxes of Hui Shih (‑4C).  P. 191 gives HS/8:  Linked rings can be sundered.  On p. 193, Needham gives several explanations of this statement and a reference to the Chinese Rings in vol. III, but he does not claim this statement refers to the puzzle.

Stewart Culin.  Korean Games.  Op. cit. in 4.B.5.  Section XX: Ryou‑Kaik‑Tjyo -- Delay Guest Instrument (Ring Puzzle), pp. 31‑32.  Story of Hung Ming (181‑234) inventing it.  (Wei-Hwa Huang says this is probably Kong Ming (= Zhuge Liang), a famous war strategist, to whom many inventions were attributed.)  States the Chinese name is  Lau Kák Ch'a  (Delay Guest Instrument) or  Kau Tsz' Lin Wain  (Nine Connected Rings).  Says there a great variety of ring puzzles in Japan, known as  Chie No Wa  (Rings of Ingenuity) and illustrates one, though it appears to be just 10 rings joined in a chain -- possibly a puzzle ring??  He says he has not found out whether the Chinese rings are known in Japan -- but see Gardner below.

Ch'ung‑En Yü.  Ingenious Ring Puzzle Book.  In Chinese:  Shanghai Culture Publishing Co., Shanghai, 1958.  English translation by  Yenna Wu,  published by Puzzles -- Jerry Slocum, Beverly Hills, Calif., 1981.  P. 6.  States it was well known in the Sung (960‑1279).  [There is a recent version, edited into simplified Chinese (with some English captions, etc.) by  Lian Huan Jiu,  with some commentary by  Wei Zhang,  giving the author's name as  Yu Chong En,  published by China Children's Publishing House, Beijing, 1999.]

The Stratagem of Interlocking Rings.  A Chinese musical drama, first performed c1300.  Cited in:  Marguerite Fawdry; Chinese Childhood; Pollock's Toy Theatres, London, 1977, pp. 70-72.  Otherwise, Fawdry repeats information from Culin and the story that it was used as a lock.

Needham.  P. 111 describes the puzzle as known in China at the beginning of the 20C, but says the origins are quite obscure and gives no early Chinese sources.  He also cites his vol. 2, p. 191, for an early possible reference -- see above.

Pacioli.  De Viribus.  c1500.  Ff. 211v-212v, Part 2, Capitulo CVII. Do(cumento), cavare et mettere una strenghetta salda in al quanti anelli saldi. dificil caso (Remove and replace a joined string a number of joined rings - a difficult thing).  = Peirani 290-292.  Dario Uri says this describes the Chinese Rings.  It is hard to make out, but it appears to have six rings.  Uri gives several of the legends about its invention and says Cardan called it Meleda, but that word is not in Cardan's text.  He lists 27 patents on the idea in five countries.

Cardan.  De subtilitate.  1550.  Liber XV.  Instrumentum ludicrum, pp. 294‑295.  = Basel, 1553, pp. 408‑409.  = French ed., 1556, et les raisons d'icelles; Book XV, para. 2, p. 291, ??NYS.  = Opera Omnia, vol. 3, p. 587.  Very cryptic description, with one diagram of a ring.

In England, the Chinese Rings were known as Tarriers or Tarriours or Tiring or Tyring or Tarrying Irons.  The OED entry at Tiring-irons gives 5 quotations from the 17C:  1601, 1627, 1661, 1675, 1690.

John II Tradescant (1608-1662).  Musæum Tradescantianum.  1656.  Op. cit. in 6.V.  P. 44: "Tarriers of Wood made like our Tyring-Irons."  (The following entry is: "Tarriers of Wood like Rolles to set Table-dishes on." -- I cannot figure out what this is.)

Gardner.  Knotted, chap. 2, says there are 17C Japanese haiku about it and it is used in Japanese heraldic emblems.

Kozaburo Fujimura  &  Shigeo Takagi.  Pazuru no Genryū (The Origins of Puzzles, in Japanese).  Daiyamondo Sha, Tokyo, 1975.  ??NYS -- information kindly sent by Takao Hayashi.  Chap. 9 is on the Chinese Rings. 

                    This says the oldest datable record in Japan is in  Osaka Dokugin Shū,  a book of haiku compiled in 1675.  A haiku of  Saikaku Ihara  is  "Chienowa ya shijōdōri ni nukenuran",  where the first word means 'wisdom rings' and denotes the Chinese Rings puzzle. 

                    Another poetry book,  Tobiume Senku,  of 1679 has  "Tenjiku shintan kuguru chienowa". 

                    The Chinese Rings was used as a family crest.  The first known reference is in the description of a kimono worn by a character in  Monzaemon Chikamatsu's  joruri,  Onna Goroshi Abura Jigoku,  first staged in 1721.  See next item for more details.

                    Another word for the puzzle is  kyūrenkan,  which is borrowed from Chinese and means 'nine connected rings'.  It is explained in a Chinese lexicon,  Meibutsu rokujō,  by  Choin Ito,  c1725.

                    The mathematician  Yasuaki Aida (1747-1817),  in his unpublished autobiography  Jizai Butsudan  of 1807, says he solved the puzzle when he was nine, i.e. c1756.

                    Yoriyuki Arima (1714-1783), another mathematician, treats the Chinese Rings in his mathematical work  Shūki Sampō,  of 1769.

                    Gennai Hiraga (1728-1779) unlocked a bag locked with a Chinese Rings belonging to Captain Jan Crans of the Dutch factory (i.e. trading post) in c1769.  This is related in  Genpaku Sugita's  Rangaku Kotohajime  of 1815.

                    Chienowa  is recorded in the 1777 Japanese dictionary  Wakun no Shiori.

Dictionary of Representative Crests.  Nihon Seishi Monshō Sōran (A Comprehensive Survey of Names and Crests in Japan),  Special issue of  Rekishi Dokuhon (Readings in History),  Shin Jinbutsu Oraisha, Tokyo, 1989, pp. 271-484.  Photocopies of relevant pages kindly sent by Takao Hayashi.  Crest 3447 looks like a Chinese Rings with five rings and 3448 looks like one with four rings, but both are simplified and leave out one of the bars.

John Wallis.  De Algebra Tractatus.  1685, ??NYS.  = Opera Math., Oxford, 1693, vol. II, chap. CXI, De Complicatus Annulis, 472‑478.  Detailed description with many diagrams.

Ozanam.  1725: vol. 4.  No text, but the puzzle with  7  rings is shown as an unnumbered figure on plate 14 (16).  Ball, MRE, 1st ed., 1892, p. 80, says the 1723 ed., vol. 4, p. 439 alludes to it.  The text there is actually dealing with Solomon's Seal (see 11.D) which is the adjacent figure on plate 14 (16).

Minguet.  1733.  Pp. 55-57 (1755: 27-28; 1822: 72-74; 1864: 63-65): Juego del ñudo Gordiano, ò lazo de las sortijas enredadas.  7  ring version clearly drawn.

Alberti.  1747.  No text, but the puzzle is shown as an unnumbered figure on plate XIII, opp. p. 214 (111), copied from Ozanam, 1725.

Catel.  Kunst-Cabinet.  1790.  Der Nürnberger Tand, p. 15 & fig. 41 on plate II.  Figure shows  7  rings, text says you can have  7,  9,  11  or  13. 

Bestelmeier.  1801.  Item 298:  Der Nürnberger Tand.  Diagram shows  6  rings, but text refers to  13  rings.  Text is partly copied from Catel.

Endless Amusement II.  1826?  Prob. 29, pp. 204-207.  Cites Cardan as being very obscure.  Shows example with  5  rings and seems to imply it takes 63 moves.

The Boy's Own Book.  The puzzling rings.  1828: 419‑422;  1828‑2: 424‑427;  1829 (US): 216-218;  1855: 571‑573;  1868: 673-675.  Shows  10  ring version and says it takes 681 moves.  Cites Cardan.

Crambrook.  1843.  P. 5, no. 9: Puzzling Rings, or Tiring Irons.

Magician's Own Book.  1857.  Prob. 45: The puzzling rings, pp. 279-283.  Identical to Boy's Own Book, except 1st is spelled out first, etc.  = Book of 500 Puzzles, 1859, pp. 93-97.  = Boy's Own Conjuring Book, 1860, prob. 44, pp. 243‑246.

Magician's Own Book (UK version).  1871.  The tiring-irons, baguenaudier, or Cardan's rings, pp. 233-235.  Quite similar to Boy's Own Book, but somewhat simplified and gives a tabular solution.

L. A. Gros.  Théorie du Baguenodier.  Aimé Vingtrinier, Lyon, 1872.  (Copy in Radcliffe Science Library, Oxford -- cannot be located by them.)  ??NYS

Lucas.  Récréations scientifiques sur l'arithmétique et sur la géométrie de situation.  Troisième récréation, sur le jeu du Baguenaudier, ...   Revue Scientifique de la France et de l'étranger (2) 26 (1880) 36‑42.  c= La Jeu du Baguenaudier, RM1, 1882, pp. 164‑186 (and 146‑149).  c= Lucas; L'Arithmétique Amusante; 1895; pp. 170-179.  Exposition of history back to Cardan, Gros's work, use as a lock in Norway.  He says that Dr. O.-J. Broch, former Minister and President of the Royal Norwegian Commission at the Universal Exposition of 1878, recently told him that country people still used the rings to close their chests and sacks.  RM1 adds a letter from Gros.

The French term 'baguenaudier' has long mystified me.  A 'bague' is a ring.  My large Harrap's French‑English dictionary defines 'baguenaudier' as "trifler, loafer, retailer of idle talk;  ring‑puzzle, tiring irons;  bladder‑senna", but none of the related words indicates how 'baguenaudier' came to denote the puzzle.  However, Farmer & Henley's Dictionary of Slang gives 'baguenaude' as a French synonym for 'poke', so perhaps 'baguenaudier' means a 'poker' which has enough connection to the object to account for the name??  MUS I 62-63 discusses Gros's use of 'baguenodier' as unreasonable and quotes two French dictionaries of 1863 and 1884 for 'baguenaudier' which he identifies as an ornamental garden shrub, Colutea arborescens L.

Cassell's.  1881.  Pp. 91-92: The puzzling rings.  = Manson, 1911, pp. 144-145: Puzzling rings.  Shows  7  ring version and discusses  10  ring version, saying it takes 681 moves.  Discusses the Balls and Rings puzzle.

Peck & Snyder.  1886.  P. 299: The Chinese puzzling rings.  9  rings.  Mentions Cardan & Wallis.  Shown in Slocum's Compendium.

Ball.  MRE, 1st ed., 1892, pp. 80-85.  Cites Cardan, Wallis, Ozanam and Gros (via Lucas).  P. 85 says:  "It is said -- though a priori the fact would have seemed very improbable -- that Chinese rings are used in Norway to fasten the lids of boxes, ....  I have never seen them employed for such purposes in any part of the country in which I have travelled."  This whole comment is dropped in the 3rd ed.

Hoffmann.  1893.  Chap. X, no. 5: Cardan's rings, pp. 334‑335 & 364‑367 = Hoffmann‑Hordern, pp. 222-225, with photo.  Cites Encyclopédie Méthodique des Jeux, p. 424+.  Photo on p. 223 shows The Puzzling Rings, by Jaques & Son, 1855‑1895, with instructions, and Baguenaudier, with box, 1880-1895.  Hordern Collection, p. 92, shows the Jaques example, an ivory example with an elaborate handle and another of ivory or bone, all dated 1850-1900.  I now have an example of the Jaques version which has rings coloured red, white and blue.

H. F. Hobden.  Wire puzzles and how to make them.  The Boy's Own Paper 19 (No. 945) (13 Feb 1896) 332-333.  Magic rings (= Chinese rings) with  10  rings, requiring 681 moves.  (I think it should be 682.)

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  "... it is said to be used by the Norwegians as a form of lock for boxes and bags ..."

Ahrens.  Mathematische Spiele.  Encyklopadie article, op. cit. in 3.B.  1904.  Note 60, p. 1091, reports that a Norwegian professor of Ethnography says the story of its use as a lock in Norway is erroneous.  He repeats this in MUS I 63.

M. Adams.  Indoor Games.  1912.  Pp. 337‑341 includes The magic rings.

Bartl.  c1920.  P. 309, no. 80: Armesünderspiel oder Zankeisen.  Seven ring version for sale.

Collins.  Book of Puzzles.  1927.  The great seven-ring puzzle, pp. 49-52.  Cites Cardan and Wallis.  Says it is known as Chinese rings, puzzling rings, Cardan's rings, tiring irons, etc.  Says  3 rings takes 5 moves,  5 rings takes 21  and  7 rings takes 85.

Rohrbough.  Puzzle Craft.  1932.  The Devil's Needle, p. 7 (= p. 9 of 1940s?).  Cites Boy's Own Book of 1863.

R. S. Scorer,  P. M. Grundy  &  C. A. B. Smith.  Some binary games.  MG 28 (No. 280) (Jul 1944) 96‑103.  Studies the binary representations of the Chinese Rings and the Tower of Hanoi.  Gives a triangular coordinate system representation for the Tower of Hanoi.  Studies Tower of Hanoi when pegs are in a line and you cannot move between end pegs.  Defines an  n‑th order Chinese Rings and gives its solution.

E. H. Lockwood.  An old puzzle.  With Editorial Note by H. M. Cundy.  MG 53 (No. 386) (Dec 1969) 362‑364.  Derives number of moves by use of a second order non‑homogeneous recurrence.  Cundy mentions the connection with the Gray code and indicates how the Gray value at step  k,  G(k),  is derived from the binary representation of  k,  B(k).  [But he doesn't give the simplest expression, given by Lagasse, qv in 7.M.3.]  This easily gives the number of steps.

Marvin H. Allison Jr.  The Brain.  This is a version of the Chinese Rings made by Mag-Nif since the 1970s.  [Gardner, Knotted.]

William Keister.  US Patent 3,637,215 -- Locking Disc Puzzle.  Filed: 22 Dec 1970;  patented: 25 Jan 1972.  Abstract + 3pp + 1p diagrams.  This is a version of the Chinese Rings, with discs on a sliding rod producing the interaction of one ring with the next.  Described on the package.  Keister worked on puzzles of this sort since the 1930s.  It was first produced by Binary Arts in 1986 under the name Spin Out.

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Protocol, pp. 163-166.  Gives seating and standing problems which lead to the same sequence of moves as for the Chinese rings, but one is in reverse order.

Anatoli Kalinin says that the Chinese Rings are a old folk puzzle called  Меледа  [Meleda],  especially popular among the Kalmyks near the Caspian Sea, where it is called  Наран-шина  [Naran-shina] (stirrup ring toy).  The name  Меледа  is derived from a verb which is no longer in Russian.

 

          7.M.2.          TOWER OF HANOI

 

            See MUS I 52-61, S&B 135.

            See also 5.F.4 for connection with Hamiltonian circuits on the n‑cube and 5.A.4 for the Panex Puzzle.

            All the following have three pegs unless specified otherwise.

 

The Conservatoire National des Arts et Métiers -- Musée National des Techniques, 292 Rue St. Martin, Paris, has two examples -- No. 11271 & 11272 -- presented by Edouard Lucas, professeur de mathématiques au Lycée Saint-Louis à Paris, in 1888.  The second is a 'grand modèle pour les cours publics' 1.05 m high!  Elisabeth Lefevre has kindly sent details and photocopies of the box, instruction sheet (one sheet printed on both sides) and an article.  Both versions have  8  discs.  I am extremely grateful to Jean Brette, at the Palais de la Decouverte, who told me of these examples in 1992.

                    The box is  157 x 180 mm  and has an elaborate picture with the following text:

                              La Tour d'Hanoï / Veritable casse‑Téte Annamite / Jeu / rapporté du Tonkin / par le Professeur N. Claus (de Siam) / Mandarin / du College / Li‑Sou‑Stian / Brevete / S. G. D. G.

          This cover is shown in Claus [Lucas] (1884) and Héraud (1903).  I will refer to this as the original cover.  (The Museum does not know of any patent -- they have looked in 1880-1890.  S. G. D. G.  stands for  Sans garantie du gouvernement.)  The bottom of the box has an ink inscription:  Hommage del'auteur   Ed Lucas   Paris  1888  -- but the date is not clearly legible on the photocopy.  Inside the cover, apparently in the same hand (that is, in Lucas's writing), is an ink inscription:

                                        La tour d'Hanoï, --

                                        Jeu de combinaison pour

                              expliquer le systeme de la numération

                              binaire, inventé par M. Edouard Lucas,

                              (novembre 1883).  --  donné par l'auteur.

                    The Museum describes the puzzle as  15 cm long  by  14.5 cm wide by  10 cm high.  There is no photo available, but the examples are shown in catalogues of 1906 and 1943.

                    The instruction sheet reads as follows.  

 

                                                                La Tour d'Hanoï  

                                                       Véritable casse‑tête annamite  

                                                           Jeu rapporté du Tonkin  

                                                  par le Professeur N. Claus (de Siam)  

                                                   Mandarin du Collège Li‑Sou‑Stian! 

                                                                          ....

                                                       Paris, Pékin, Yédo et Saïgon

                                                                          ....

                                                                        1883

                    The sheet mentions the Temple of Bénarès where there are  64  discs.  A prize of a million (= a thousand thousand) francs is offered for a demonstration of the solution with  64  discs!  The second sheet of the instructions gives the rules and the number of moves for  2, 3, ..., 8  discs and the general rule.  It also refers to RM.

Edward Hordern's collection has an example with the original instruction sheet, but in a simple box with just  'La Tour d'Hanoï'  in Chinese‑style lettering, not like the box described above.  Also, my recollection is that it is much smaller than the example above.  It has  8  discs.

G. de Longchamps.  Variétés.  Journal de Mathématiques Spéciales (2) 2 (1883) 286-287.  (The article is only signed G. L., but the author is further identified in the index on p. 290.  Copy provided by Hinz.)  Solves the recurrence relation  un = 2un-1 + 1,  u0 = 0.  Says he was 'inspired by a letter which we have recently received from professor N. Claus.'  Describes the Tower of Hanoi briefly and says the above solution gives the number of moves when there are  n  discs.

Henri de Parville.  Column:  Revue des sciences.  Journal des Débats Politiques et Littéraires (27 Dec 1883) 1-2.  On p. 2, he reports receiving an example in the post with a cover like that of the original.  Gives the Benares story.  Wonders who the mandarin could be and notes the anagrams on  Lucas d'Amiens  and  Saint-Louis.

N. Claus (de Siam) [= Lucas (d'Amiens)].  La tour d'Hanoï.  Jeu de calcul.  Science et Nature 1:8 (19 Jan 1884) 127-128.  Says it takes  2n - 1  moves  "que M. de Longchamps l'a démontré (1)."  "(1) Journal de mathématiques spéciales.".  Observes that each of the discs always moves in the same cycle of pegs and hence gives the standard rule for doing the solution, which is attributed to the nephew of the inventor, M. Raoul Olive, student at the Lycée Charlemagne.  Asks for the minimum number of moves to restore an arbitrary distribution of discs to a Start position.  Says this is a complex problem in general, depending on binary and refers to RM 1 for this idea. 

                    (This paper is not in Harkin's bibliography (op. cit. in 1).  Hinz, 1989, cites it.)

Henri de Parville.  Récréations Mathématiques:  La Tour d'Hanoï et la question du Tonkin.   La Nature (Paris) 12 (No. 565, part 1) (29 Mar 1884) 285‑286.  Illustration by Poyet.  Asserts Lucas is the inventor. 

R. E. Allardice & A. Y. Fraser.  La Tour d'Hanoï.  Proc. Edin. Math. Soc. 2 (1883‑1884) 50‑53.  Includes de Parville from J. des Débats.  Then derives number of moves.

Anton Ohlert.  US Patent 303,946 -- Toy.  Applied: 24 Jul 1884;  patented: 19 Aug 1884.  1p + 1p diagrams.  8  discs.  Ohlert is a resident of Berlin.

Edward A. Filene, No. 4 Winter St., Boston Mass.  Eight Puzzle.  Copyrighted in 1887.  ??NYS -- described and illustrated in S&B, p. 135.  8  disc advertising version.

Tissandier.  Récréations Scientifiques.  5th ed., 1888, La tour d'Hanoï et la question du Tonkin, pp. 223-228.  Not in the 2nd ed. of 1881 nor the 3rd ed. of 1883.  Essentially de Parville's article with same illustration, but introduces it with comments saying it has had a great success and comes in a box labelled:  "la Tour d'Hanoï, véritable casse-tête annamite, rapporté du Tonkin par le professeur N. Claus (de Siam), mandarin du collège Li-Sou-Stian".  This would seem to be the original box.

                    = Popular Scientific Recreations; [c1890]; Supplement: The tower of Hanoï and the question of Tonquin, pp. 852‑856. 

Anon.   Jeux, Calculs et Divertissements.   Récréations Mathématiques.   La Tour d'Hanoï.  Liberté (9 Dec 1888) no page number on clipping.  Gives the story of N. Claus and says it was invented by Lucas and comes in a box decorated with annamite illustrations.  This would seem to be the original box.

Lucas.  Nouveaux jeux scientifiques de M. Édouard Lucas.  La Nature 17 (1889) 301‑303.  Describes a series of games under the title of the next item, so the next may refer to the game or its instruction booklet.  A later version of the Tower of Hanoi is described on pp. 302‑303, having  5  pegs, and is shown on p. 301.  He says you can have  3, 4 or 5  pegs.

Lucas.  Jeux scientifiques pour servir à l'histoire, à l'enseignement et à la pratique du calcul et du dessin.  Première série: No. 3: La Tour d'Hanoï.  Brochure, Paris, 1889.  ??NYS.  (Not listed in BNC, but listed in Harkin, op. cit. in Section 1.  I wonder if these were booklets that accompanied the actual games??  Hinz says he has found some of these, but not the one on the Tower of Hanoi.  The booklet for La Pipopipette (= Dots and Boxes) is reproduced in his L'Arithmétique Amusante of 1895 -- see 4.B.3.)

Jeux Scientifiques de Ed. Lucas.  Advertisement by Chambon & Baye (14 rue Etienne-Marcel, Paris) for the 1re Serie of six games.   Cosmos.  Revue des Sciences et Leurs Applications 39 (NS No. 254) (7 Dec 1889) no page number on my photocopy.

B. Bailly [name not given, but supplied by Hinz].  Article on Lucas's puzzles.  Cosmos.  Revue des Sciences et Leurs Applications.  NS, 39 (No. 259) (11 Jan 1890) 156-159.  Shows 'La nouvelle Tour d'Hanoï', which has five pegs and 15? discs.  I need pp. 156‑157.

Alfred Gartner & George Talcott.  UK Patent 20,672 -- Improvements in Games and Puzzles.  Applied: 18 Dec 1890;  patented: 21 Feb 1891.  2pp + 1p diagrams.  Shunting puzzle equivalent to Tower of Hanoi with 7 discs.

Ball.  MRE, 1st ed., 1892, pp. 78-79.  "... described by M. Tissandier as being common in France but which I have never seen on sale in England."  Gives the Benares story from De Parville in La Nature.

Hoffmann.  1893.  Chap. X, no. 4: The Brahmin's Rings, pp. 333-334 & 361-364  = Hoffmann/Hordern, pp. 220-222, with photo.  Gives Benares story and then gives the problem with  8  discs, noting that it is made by Messrs. Perry & Co.  Photo on p. 221 shows an example, with box with instructions on the top, by Perry & Co., 1880-1900.  Hordern Collection, p. 90, shows a version with box, by R. Journet, 1905-1920.

Lucas.  L'Arithmétique Amusante.  1895.  La tour d'Hanoï, pp. 179-183.  Description, including Benares story.  Says the nephew of the inventor, Raoul Olive, has noted that the smallest disc always moves in the same direction.  Says the whole idea of the mandarin and his story was invented a dozen years ago at  56 rue Monge,  which was built on the site where Pascal died.  [I visited this site recently -- it is a hotel and they knew nothing about Pascal.  Another source says Pascal died at  67 Rue Cardinal Lemoine,  which is several blocks away, but also is not an old building.]

Ball.  MRE, 3rd ed., 1896, pp. 99-101.  Omits the Tissandier reference and says:  "It was brought out in 1883 by M. Claus (Lucas)."

A. Héraud.  Jeux et Récréations Scientifiques -- Chimie, Histoire Naturelle, Mathématiques.  Not in the 1884 ed.  Baillière et Fils, Paris, 1903.  Pp. 300‑301 shows the original cover.

Burren Loughlin  &  L. L. Flood.  Bright-Wits  Prince of Mogador.  H. M. Caldwell Co., NY, 1909.  The five shields, pp. 20-24 & 59.  Five discs.

Tom Tit??.  In Knott, 1918, but I can't find it in Tom Tit.  No. 165: The tower of Hanoi, pp. 382‑383.  Describes it with cards  A ‑ 10  on piles labelled with  J, Q, K.

Robert Ripley.  Believe It Or Not!  Book 2.  (Simon & Schuster, 1931);  Pocket Books, NY, 1948, pp. 52‑53.  = Believe It or Not!  Two volumes in one; (Simon & Schuster, 1934); Garden City Books, 1946, pp. 222-223.  = Omnibus Believe It Or Not!; Stanley Paul, London, nd [c1935?], pp. 256‑257.  The Brahma Pyramid.  Outlines the Benares story, says he didn't locate the temple when he was in Benares, but it 'really exists', and that it will take  264  moves, but he then writes out  264 ‑ 1.  He says the Brahmins have been at it for  3000  years!!

B. D. Price.  Pyramid patience.  Eureka 8 (Feb 1944) 5-7.  Straightforward development of basic properties.

Scorer, Grundy & Smith.  1944.  Op. cit. in 7.M.1.  They develop the graph of all positions of the Tower of Hanoi.

Donald W. Crowe.  The  n-dimensional cube and the tower of Hanoi.  AMM 63 (1956) 29-30.  Describes the connection with Hamiltonian circuits and binary ruler markings.

M. Gardner.  The Icosian Game and the Tower of Hanoi.  (SA (May 1957)) = 1st Book, chap. 6.  Describes Crowe's work.

A. J. McIntosh.  Binary and the Tower of Hanoi.  MTg 59 (1972) 15.  He sees the connection between binary and which disc is to be moved, but he wonders how to know which pegs are involved.  [This is a valid query -- though we know each disc moves cyclically, alternate ones in alternate directions, I don't know any easy way to translate a particular step number into the positions of all the discs -- Hinz (1989) gives a method which may be as simple as possible, but I have a feeling it ought to be easier.  Actually, I have now seen a fairly easy way to do this.]

Andy Liu & Steve Newman, proposers and solvers.  Problem 1169 (ii) -- The two towers.  CM 12 (1986) 179  &  13 (1987) 328‑332.  Three pegs, two identical piles of size  n  on two of them.  The object is to interchange the bottom discs and reform the piles (though the smaller discs may or may not be interchanged).  They find it takes  7*2n+1 ‑ 3n ‑ (10 or 11)/3  steps, depending on whether  n  is odd or even.

Andreas M. Hinz.  The Tower of Hanoi.  L'Enseignement Math. 35 (1989) 289-321.  Surveys history and current work.  50 references.  Finds many properties, particularly the average distance from having all discs on a given peg and the average distance between legal positions.  He also studies illegal positions.  Uses the Grundy, Scorer & Smith graph.  Gives general results, such as Schwenk's below.  He asserts that the minimality of the classic solution was not proven until 1981, but I think the classic method clearly implies the proof of its minimality.

Hugh Noland, proposer;  Norman F. Lindquist, David G. Poole & Allen J. Schwenk, solvers.  Prob. 1350 -- Variation on the Tower of Hanoi.  MM 63:3 (1990) 189  &  64:3 (1991) 199-203.  Three pegs,  2n  discs, initially with the evens on one peg and the odds on another.  How many moves to get all onto the empty peg?  Answer is  ë(5/7) 4nû.  Schwenk gives a solution for any starting position of  N  discs and shows the average number of moves to get to a single pile is  (2/3)(2N-1).

Andreas M. Hinz.  Pascal's triangle and the Tower of Hanoi.  AMM 99 (1992) 538-544.  Shows the Grundy, Scorer & Smith graph is equivalent to the pattern of odd binomial coefficients in the first  2n  rows and hence to Sierpiński's fractal triangle.  Gives some life of Lucas.

David Poole.  The bottleneck Towers of Hanoi problem.  JRM 24:3 (1992) 203-207.  Studies the problem when big discs can go on smaller discs, but not too much smaller ones.  11 references to recent work on variations of the classical problem.

Ian Stewart.  Four encounters with Sierpiński's gasket.  Math. Intell. 17:1 (1995) 52-64.  This discusses the connections between the graph of the Tower of Hanoi, the pattern of odd binomial coefficients, Sierpiński's gasket and Barnsley's iterated fractal systems.  Lots of references, including 11 on the Tower of Hanoi.

Vladimir Dubrovsky.  Nesting Puzzles -- Part I: Moving oriental towers.  Quantum 6:3 (Jan/Feb 1996) 53-59 & 49-51.  Outlines the history and theory of the Tower of Hanoi.  Misha Fyodorov, a Russian high school student, observed that the peg not used in a move always moves in the same direction.  Discusses Kotani's modification which prevents placing some discs on a particular peg.  Also discusses the Panex Puzzle -- cf Section 5.A.4.

Jagannath V. Badami.  Musings on Arithmetical Numbers  Plus Delightful Magic Squares.  Published by the author, Bangalore, India, nd [Preface dated 9 Sep 1999].  Section 4.15: The Tower of Brahma, pp. 123-124.  "The author has lived in Banaras for a number of years and does not find any basis for this legend."

David Singmaster.  The history of some combinatorial recreational problems.  Draft of a chapter for History of Combinatorics, ed. Robin J. Wilson.  Jan 2001.  This gives a detailed development of the distances  di  of a position from the three perfect positions and the complementary distances  d'i = (2n-1) - di,  leading to  THEOREM 4.  The Scorer, Grundy & Smith graph of positions in the Tower of Hanoi with  n  discs and with adjacency between positions one move apart, is isomorphic to the graph of triples of binary  n‑tuples  (d'0, d'1, d'2)  satisfying   (*')   Σi Dk(d'i)  =  1  for each  k,  considered as triangular coordinates in a triangle of edge length  2n-1  and with adjacency being adjacency in the lattice.

 

          7.M.2.a.       TOWER OF HANOI WITH MORE PEGS

 

Lucas.  Nouveaux jeux scientifiques ..., op. cit. in 4.B.3, 1889.  (See discussion in 7.M.2.)

Dudeney.  The Reve's Puzzle.  The Canterbury Puzzles.  London Mag. 8 (No. 46) (May 1902) 367‑368  &  8 (No. 47) (Jun 1902) 480.  = CP, prob. 1, pp. 24‑25 & 163‑164.  4  pegs,  8, 10 or 21  discs.

Dudeney.  Problem  447.  Weekly Dispatch (25 May, 15 Jun, 1902) both p. 13.  4  pegs,  36  discs.

Dudeney.  Problem 494.  Weekly Dispatch (15 Mar, 29 Mar, 5 Apr, 1903) all p. 13.  5  pegs,  35  discs.

B. M. Stewart, proposer;  J. S. Frame & B. M. Stewart, solvers.  Problem 3918.  AMM 48 (1941) 216‑219.  k  pegs,  n  discs.  General solution, but editorial note implies there is a gap in each solver's work.

Scorer, Grundy & Smith.  Op. cit. in 7.M.1.  1944.  They give some variations on the Tower of Hanoi with four pegs.

Doubleday - 2.  1971.  Keep count, pp. 91-92.  15  discs,  6  pegs  -- solved in  49  moves.

Ted Roth.  The tower of Brahma revisited.  JRM 7 (1974) 116‑119.  Considers  4  pegs.

Brother Alfred Brousseau.  Tower of Hanoi with more pegs.  JRM 8 (1975/76) 169‑176.  Extension of Roth, with results for  4  and  5  pegs.

The Diagram Group.  Baffle Puzzles -- 3: Practical Puzzles.  Sphere, 1983.  No. 10.  6 pegs, 15 discs.  Gives a solution in 49 moves.

Joe Celko.  Puzzle Column:  Mutants of Hanoi.  Abacus 1:3 (1984) 54‑57.  Discusses variants:  where a disc can only move to an adjacent peg in a linear arrangement;  with two or three colours of discs;  with several piles of discs;  where a disc can only move forward in a circular arrangement.

Grame Williams.  In:  Joe Celko; Puzzle column replies; Abacus 5:2 (1988) 70‑72.  Table of minimum numbers of moves for  k  pegs,  k = 3, ..., 8  and  n  discs,  n = 1, ..., 10.

Andreas Hinz.  An iterative algorithm for the Tower of Hanoi with four pegs.  Computing 42 (1989) 133-140.  Studies the problem carefully.  17 references.

A. D. Forbes.  Problem 163.2 -- The Tower of Saigon.  M500 163 (Aug 1998) 18-19.  This is the Tower of Hanoi with four pegs.  Quotes an Internet posting by Bill Taylor giving an algorithm and its number of moves up to 12 discs.  Asks if this is optimal.

 

          7.M.3.          GRAY CODE

 

          See Gardner under 7.M.

 

L. A. Gros.  Op. cit. in 7.M.1, 1872.  ??NYS.  (Afriat.)

J. Émile Baudot.  c1878.  ??NYS.  Used Gray code in his printing telegraph. (Described by F. G. Heath; Origins of the binary code; SA (Aug 1972) 76‑83.)

Anon.  Télégraphe multiple imprimeur de M. Baudot.  Annales Télégraphiques (3) 6 (1879) 354‑389.  Says the device was presented at the 1878 Exposition and has been in use on the Paris‑Bordeaux line for several months.  See pp. 361‑362 for diagrams and p. 383 for discussion.

George R. Stibitz.  US Patent 2,307,868 -- Binary Counter.  Applied: 26 Nov 1941;  granted: 12 Jan 1943.  3pp + 1p diagrams.  Has an electromechanical binary counter using the Gray code with no comment or claims on it.

Frank Gray.  US Patent 2,632,058 -- Pulse Code Communication.  Applied: 13 Nov 1947, patented: 17 Mar 1953.  9pp + 4pp diagrams.  Systematic development of the idea and its uses.

A. J. Cole.  Cyclic progressive number systems.  MG 50 (No. 372) (May 1966) 122‑131.  These systems are Gray codes to arbitrary bases -- e.g. in base  4,  the sequence begins:  0, 1, 2, 3, 13, 12, 11, 10, 20, 21, 22, 23, 33, 32, 31, 30, 130, 131, ....  For odd bases, the sequence is harder.  He gives conversion rules and rules for arithmetic.

J. Lagasse.  Logique Combinatoire et Séquentielle.  Maîtrise d'E.E.A.  C3 -- Automatique.  Dunod, Paris, 1969.  Pp. 14-18 discuss the Gray code (code réfléchi) stating that the Gray value at step  k,  G(k),  is given by  G(k)  =  B(k)  EOR  B(ëk/2û).  [I noted this a few years ago and am surprised that it does not appear to be old.  Gardner's 1972 article describes it but not so simply.]  My thanks to Jean Brette for this reference.

William Keister.  US Patent 3,637,216 -- Pattern-Matching Puzzle.  Filed: 11 Dec 1970;  patented: 25 Jan 1972.  Abstract + 4pp + 2pp diagrams.  This has a bar to remove from a frame -- one has to move various bits in the pattern of the Gray code (or similar codes) to extract the bar.  Made by Binary Arts since about 1986.

 

          7.M.4.          BINARY DIVINATION

 

            The classic cards for this process have the numbers  1 - 2n  on them, the  i-th card containing those numbers whose binary expression has a  1  in its  (i-1)-st place -- e.g. the first card contains all the odd numbers.  Then one adds up the smallest numbers, i.e.  2i-1  on the  i‑th card, on the chosen cards to get the number thought of.  If one replaces the numbers by holes in the corresponding positions, one can overlay the cards to read off the answer.  This takes a little more work though -- one has to have each card containing its set of holes to be used if it is chosen and also containing the complementary holes to be used if it is not chosen.  This can be achieved if the holes are centrally located -- then turning the card around produces the complementary set of holes.

            A very similar principle is used as a kind of logical device.  See:  Martin Gardner; Logic Machines and Diagrams; McGraw‑Hill, NY, 1958, pp. 117-124;  slightly extended in the 2nd ed., Univ. of Chicago Press, 1982, and Harvester Press, Brighton, 1983, pp. 117-124;  for discussion of this idea and references to other articles.

            A simple form of binary division is used to divine a card among sixteen cards arranged in two columns, but it is surprisingly poorly described.  This is related to the 21 card trick which is listed in 7.M.4.b.  I have only recently added this topic and may not have noticed many versions.

 

Pacioli.  De Viribus.  c1500.  Ff. 114r - 116r.  C(apitolo). LXIX. a trovare una moneta fra 16 pensata (To find a coin thought of among 16).  = Peirani 161-162.  Divides 16 coins in half 4 times, corresponding to the value of the binary digits.  Pacioli doesn't describe the second stage clearly, but Agostini makes it clear.

Bachet.  Problemes.  1612.  Prob. XVI, 1612: 87-92.  Prob. 18; 1624: 143-151; 1884: 72-83.  15 card trick.   His Avertissement mentions that other versions are possible and describes divining from sixteen cards in two columns and in four columns, but with no diagrams!

Yoshida (Shichibei) Kōyū (= Mitsuyoshi Yoshida) (1598-1672).  Jinkō‑ki.  2nd ed., 1634 or 1641??.  Op. cit. in 5.D.1.  ??NYS  Shimodaira (see entry in 5.D.1) discusses this on pp. 2-12 since there are several Japanese versions of the idea.  The Japanese call these Metsukeji (Magic Cards).  The binary version is discussed on pp. 4-7 where it is said that they are known since the 14C or even earlier.  The Japanese magic card shown on p. 6 has  1, 2, 4, 8, 16  associated with branchings on a picture of a tree used to divine one of the  21  characters written on the flowers and leaves.  The other kinds of magic cards are more complex, not involving binary, but just memorisation.  The recent transcription of part of Yoshida into modern Japanese does not include this problem.

Ozanam.  1725.  Prob. 39, 1725: 231-233.  Prob. 15, 1778: 164-165;  1803: 165-166;  1814: ??NYS.  Prob. 14, 1840: 74-75.  Sixteen counters being disposed in two rows, to find that which a person has thought of.  Similar to Bachet, but with some diagrams.

Ozanam-Hutton.  1814: 124-126;  1840: 64.  (This is an addition which was not in the 1803 ed.)  Six cards to divine up through 63. 

Endless Amusement II.  1826?  Pp. 180-181.  Sixteen Cards being disposed in Two Rows, to tell the Card which a Person has thought of.  c= Ozanam, with 'counter' replaced by 'card'.

Young Man's Book.  1839.  Pp. 202-203.  Identical to Endless Amusement II.

Crambrook.  1843.  P. 7, no. 5: A pack of cards by which you may ascertain any person's age.  Not illustrated, but seems likely to be binary divination -- ??

Magician's Own Book.  1857.  The mathematical fortune teller, pp. 241‑242.  Six cards each having  30  numbers used to divine a number up through  60.  Some cards have duplicate numbers in the  30th position.  = Boy's Own Conjuring Book, 1860, pp. 211-212.   = Illustrated Boy's Own Treasury, 1860, prob. 38, pp. 402 & 442.

Book of 500 Puzzles.  1859.  The mathematical fortune teller, pp. 55-56.  Identical to Magician's Own Book.  However, my example of the book has  61  in the last cell of the last card, but a photocopy sent by Sol Bobroff has this as a  41,  as do all other versions of this problem.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 563-VI, pp. 249-250: Überraschungen mittels sieben Zauberkarten.  Seven cards used to divine up to 100.

Adams & Co., Boston.  Advertisement in The Holiday Journal of Parlor Plays and Pastimes, Fall 1868.  Details?? -- photocopy sent by Slocum.  P. 5: Magic Divination Cards.  For telling any number thought of, or a person's age.  Amusing, curious, and sometimes "provoking."   Not illustrated, but seems likely to be binary divination -- ??

Magician's Own Book (UK version).  1871.  The numerical fortune teller, pp. 89-90.  Very similar to Magician's Own Book, pp. 241-242, with the same cards, but different text.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. X, 1884: 196-197.  Divine a number up to  100  with  7  cards.

F. J. P. Riecke.  Op. cit. in 4.A.1, vol. 3, 1873.  Art. 3: Die Zauberkarten, pp. 11‑13.  Describes  5  cards giving values up to  31,  with explanation.  Describes how to use balanced ternary to construct  7  cards giving values up to  22.  The limitation to  22  is due to the size of the cards -- the method works up to  40.

Mittenzwey.  1880.  Prob. 37, pp. 6-8;  1895?: 43, pp. 12-13;  1917: 43, pp. 11‑13.  Seven cards.  Calls them "Boscos Zauberkarten" -- Bosco (1793-  ) was a noted conjurer of the early 19C.

Hoffmann.  1893.  Chap. IV, no. 68: The magic cards, pp. 160‑161 & 216‑217 = Hoffmann‑Hordern, pp. 141-142, with photo.  Seven cards.  Photo on p. 141 shows an ivory set of six cards, 1850-1900, and two German sets of seven cards from a box of puzzles called Hokus Pokus:  Zauber‑Karten  and  Ich weiss wie alt du bist,  both 1870‑1890.  These have their own boxes or wrappers.  Hordern Collection, p. 73, shows the same(?) Zauber-Karten, with its box or wrapper, dated 1860‑1890, and fully spread out so the instruction card is legible.

Lucas.  L'Arithmétique Amusante.  1895.  L'éventail mystérieux, pp. 168‑170.  Shows five cards for divining  1  through  31  and notes it is based on binary.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 10:2 (Jun 1903) 140-141.  To tell a lady's age.  Six cards to divine up through 63.

Ahrens.  MUS I.  1910.  Pp. 39-40 does  5  cards.  Pp. 43-48 develops a set of four see-through templates.

M. Adams.  Indoor Games.  1912.  How to divine ages, pp. 349-350.

The Quaker Oats Age Prediction Cards.  (I have a facsimile of a set, made by Dave Rosetti for G4G5.  He says it was invented in 1921.)  8 cards with various holes and numbers for divining from 1 to 99.  You turn the card round if the number is not on it but put it on the pile.  At the end, you turn the pile over and the age and the slogan 'Never Known to Fail' can be read through the holes. 

                    Card 1 is just a viewing window with four windows that are five digits wide.  Cards 6, 7, 8 have numbers ingeniously arranged so that groups of five digits are in the window positions.  Reading two digits at a time, these groups give four possible values, hence 16 for each configuration of cards 6, 7, 8 -- except that when both 6 and 7 are up, then card 8 is not visible, and when cards 6 and 7 are turned, there are only four different values -- this results in just the 100 values: 00, 01, ..., 99.  Cards 2 and 3 select which window is open.  Cards 4 and 5 select which pair of digits in the window can be seen.

William P. Keasby.  The Big Trick and Puzzle Book.  Whitman Publishing, Racine, Wisconsin, 1929.  Magic columns, pp. 176-177.  Applies the idea with letters, so you add the column heads to get the number of the letter.

Rohrbough.  Brain Resters and Testers.  c1935.  How to Mystify People, pp. 10-11.  = Keasby, whom he cites elsewhere.

John Fisher.  John Fisher's Magic Book.  Muller, London, 1968.  Think‑a‑Drink, pp. 27-29.  5  cards with cut out holes to divine  32  types of drink.  The last card seems to need one side reversed.

 

          7.M.4.a.       TERNARY DIVINATION

 

            This is much less common than binary divination and I have just added it.  Gardner, op. cit. at beginning of 7.M.4, describes one set of triangular cards.

 

Riecke.  1873.  See in 7.M.4.

Martin Hansen.  Mind probe.  MiS 21:1 (Jan 1992) cover & 2-6.  Adapts binary divination to locate a number among  1 - 80  with four cards -- but each card has its numbers half in black and half in white.  Describes how to make  4 x 4  and  8 x 8  binary cards with holes so the chosen number will appear in the hole.  Adapts to ternary to produce triangular cards with holes so that the chosen number appears in the hole.  Hansen has kindly given me a set of these:  'The Kingswood Mathemagic Club's Window Cards'.  He also describes 'logic cards' which display the truth values of three basic quantities which are consistent with various statements -- these were previously described by Gardner and Cundy & Rollett -- see the Gardner item at the beginning of 7.M.4 for details and references.

 

          7.M.4.b.       OTHER DIVINATIONS USING BINARY OR TERNARY

 

            New section.  Discussion with Bill Kalush has revealed that the classic 21-card trick is based on the ternary system.  Another trick involving twice taking  3/2  of a number, which goes back to at least Pseudo-Bede, makes a little use of binary.

            The  21-card trick can also be done with  15  or  27  cards and it is easiest to explain for  27  cards.  One deals  27  cards out, face up, into three columns and asks the spectator to mentally choose one card and tell you which column it is in.  You pick up the three columns, carefully placing the chosen column in the middle of the other two, then deal out the deck into three columns again.  Ask the spectator to tell you what column the card is now in and pick up the cards again with the chosen column in the middle and deal them out again.  Repeat the whole process a third time, then deal out the cards face down and ask the spectator to turn over the middle card, which will be his chosen card.  The first process puts the chosen card among the positions whose initial ternary digit is a  1.  The second process puts it among the position whose ternary representation begins  11.  The third process puts it at the position whose ternary representation is  111.  For any number of cards which is a multiple of three, each process puts the chosen card in the 'middle third' of the previously determined portion of the deck and the trick works, though the patterns are less systematic than with  27  cards.  When the trick is done with, e.g.  24  cards, there is no middle card and you have to expose the chosen card yourself or develop some other way to do it.  One can also adapt the idea to other numbers of columns -- the two column case will usually be connected to binary and may be listed in 7.M.

            In the  3/2  method, you ask a person to think of a number,  x,  then take  3/2  of it (or itself plus half of itself).  If there is a half present, he is to round it up and let you know.  Do this again on the result.  Now ask how many nines are contained in his result.  You then tell the number thought of.  The process actually converts  x = a + 2b + 4c  to  3a + 5b + 9c,  where  a, b  =  0  or  1.  We have that   a = 1  iff  there is a rounding at the first stage   and   b = 1  iff  there is a rounding at the second stage   and the person then tells you  c.  We will denote this as PB2. 

            Sometimes the second stage is omitted or the division by two in it is omitted, which makes things simpler.  The latter case takes  a + 2b  to  6a + 9b  and the person gives you  b  and there is a remainder if and only if  a = 1.  We will denote this by PB1.

            I'm including some early examples of simple algebraic divination here.

 

Pseudo-Bede.  De Arithmeticis propositionibus.  c8C, though the earliest MS is c9C.

                    IN: Venerabilis Bedae, Anglo‑Saxonis Presbyteri.  Opera Omnia: Pars Prima, Sectio II -- Dubia et Spuria: De Arithmeticis propositionibus.  Tomus 1, Joannes Herwagen (Hervagius), Basel, 1563, Columns 133-135, ??NYS.  Folkerts says Hervagius introduced the title De Arithmeticis propositionibus.

                    Revised and republished by J.‑P. Migne as:  Patrologiae Cursus Completus: Patrologiae Latinae, Tomus 90, Paris, 1904, columns 665‑668. 

                    Critical edition by Menso Folkerts.  Pseudo-Beda: De arithmeticis propositionibus.  Eine mathematische Schrift aus der Karolingerzeit.  Sudhoffs Archiv 56 (1972) 22-43.  A friend of Bill Kalush has made an English translation of the German text, 1998?, 11pp.

                    See also: Charles W. Jones; Bedae Pseudepigrapha: Scientific writings falsely attributed to Bede.  Cornell Univ. Press & Humphrey Milford, OUP, 1939, esp. pp. 50‑53.

                    This has three divination problems and Folkerts says these are the first known western examples.

                    1.       Triple and halve, then triple and tell quotient when divided by nine and whether there is a remainder in this.  I.e. PB1.

                    2.       3/2  twice and tell if there is a rounding up at each stage as well as the quotient when divided by 9.  I.e. PB2.

                    3.       Divine a digit  a  from  a * 2 + 5 * 5 * 10,  where the operations are performed sequentially from left to right.  The result is  100a + 250.  This is not of the type considered in this section, but is the prototype of most later divination methods.

                    Folkerts mentions several later occurrences of these methods. 

                    (The fourth and last part of the text is probably slightly later in the 9C and describes adding positive and negative numbers in a way not repeated in the west until the 15C.)

Fibonacci.  1202.  Pp. 303-304 (S: 427-428).  Take  3/2  twice, i.e. PB2.

Folkerts.  Aufgabensammlungen.  13-15C. 

11 sources for  x * 2 + 1 * 5 * 10  =  100x + 50. 

Two sources for  x * 2 + 5 * 5 * 10  =  100x + 250.  Also cites Pseudo-Bede, Fibonacci and AR (see AR, pp. 122-124, 138 & 227-228 for examples and further references).  Cf Pacioli, Effect VIII.

21 sources for PB1 or PB2, with two simple variants. 

6 sources where the above ideas are extended to more values.  Cf section 7.AO for one form of this.

Chuquet.  1484.  Prob. 155.  English in FHM 230-231.  PB1.  Text only indicates what happens when  x  is even.  Marre notes that this appears in de la Roche, 1520: ff. 218v ‑ 219r; 1538: ff. 150v - 151r.  FHM say Chuquet gives an example, but it is not in Marre.

Pacioli.  De Viribus.  c1500. 

Effects I - VI, ff. 3v - 16v.  = Peirani 25-39.  Algebraic divinations of 2, 3, 3, 3, 4, 5  values which are parts of a given number.  E.g. Effect I divines  x, y  such that  x + y = a  from  a(a+1) ‑ 2x - ay.  Divide this by  a-1  to get  x + y/(a-1). 

Effects VII, IX, X.  Agostini's descriptions are very inadequate.

   Ff. 16v - 19v.  Septimo effecto trovare un Nů pensato into (Seventh effect to find a number thought of).  = Peirani 40-43.  PB2, using  1 + ½.  He does examples for all four cases of  a, b.

   Ff. 20v - 21v.  Nono effecto a trovare un Nů senza rotto (Ninth effect to find a number without a fraction).  = Peirani 45.  Take  3/2  twice, but rounding down.  This takes  4(c+1) - a - 2b  to  9(c+1) - 3a - 5b,  with  a = 1  iff  there is a rounding at the first stage   and   b = 1  iff  there is a rounding at the second stage.  Pacioli gives a simpler rule:  given  c,  form  4c  and add:  1 if there are roundings at both stages;  2 if there is rounding at only the second stage;  3 if there is rounding only at the first stage.  He does examples of all four cases of  a, b.

   Ff. 21v - 23r.  Decimo effecto de trovare un Numero senza rotto (Tenth effect to find a number without a fraction).  = Peirani 46-47.  PB1, clearly describing that the result is  6a + 9b.

Ff. 19v - 20v.  Octavo effecto quando el Nů fosse con R(otto) (Eighth effect when the number can be a fraction).  = Peirani 43-44.  Divines  x  from 

x * 2 + 5 * 5 + 10 * 10  =  100x + 350.  Cf Folkerts.

Effects XI - XXI, XXVII - XXXI, ff. 23v - 34v, 47r - 63v.  = Peirani 48-62, 77-97.  More complex algebraic divinations.  E.g. Effect XI divines a number  a  by asking he person to split it into two parts,  x,  y  (so that  x + y = a)  and compute  x2 + y2 + 2xy.

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 34-35, f. 67r-67v.  Take  3/2  twice, but he rounds down each time and he is not at all clear how the roundings relate to finding the number, and it is not nearly as elegant as when one rounds up.

Apianus.  Kauffmanss Rechnung.  1527.  Ff. M.vii.r - M.viii.r.  PB1.

Tartaglia.  General Trattato.  1556.  Book 16, art. 197-198, f. 263v-264r.  Divination by  1 + 1/2  twice and by  3/2  twice, i.e. PB2.

Recorde-Mellis.  Third Part.  1582.  Ff. Yy.ii.r - Yy.ii.v  (1668: 473-474).  PB2, done with one example of 7.

Prévost.  Clever and Pleasant Inventions.  (1584), 1998. 

Pp. 166-168.  PB2. 

Pp. 185-189.  Divination from  16  marked counters by a two column version of the  21‑card trick.  The method of rearranging the  16  counters is not entirely clear, but the principle is clearly explained.

Io. Baptiste Benedicti (= Giambattista Benedetti).  Diversarum Speculationum Mathematicarum, & Physicarum Liber.  Turin, (1580), Nicolai Bevilaquæ, Turin, 1585; (Venice, 1599).  [Rara 364.  Graves 141.f.16.]  Theorema CXVI, pp. 78-79.  PB2.

Bachet.  Problemes.  1612. 

Prob. I.  1612: 14-17;  1624: 53-55;  1884: 15-16.  PB1.  = Pacioli X, but with more explanation.

Prob. II.  1612: 17-27;  1624: 56-65;  1885: 17-22.  PB2.  = Pacioli IX, but with more explanation.

Prob. III.  1624: 66-74;  1885: 23-26.  Take  3/2  twice and then tell all but one digit.  The Advertissement in the 1624 ed. says this was invented by R. P. Jean Chastelier, S.J.  Not in the 1612,

Prob. XVI, 1612: 87-92.  Prob. 18,  1624: 143-151;  1884: 72-83.  15-card trick.   His Advertissement mentions that other versions are possible and describes a two column version.  Labosne adds several diagrams which make the process much clearer and discusses the general idea, illustrating it with 27 cards and then with 45 cards in 5 columns.

Hunt.  1631 (1651). 

          Pp. 217-218 (209-210): A sixth way to find out a number thought.  PB2.

          Pp. 221 (misprinted 212) - 222 (213-214): An eighth way to find out a number thought.

                    PB1.

Schott.  1674. 

Art. I, p. 57.  PB1.

Art. II, pp. 56-57.  PB2.

Ozanam.  1694.

Prob. 14 [part 9],  1696: 58;  1708: 52.  Prob. 17, part 10,  1725: 146.  Prob. 1, part 1,  1778: 139-140;  1803: 137;  1814: ??NYS;  1840: 62.  PB1.  In 1840, the algebraic proof is given.  In 1725: 170‑173,  he adds, as a remark to Prob. 17, PB2, with a detailed explanation.

Prob. 16,  1696: 63-64;  1708: 56-57.  Prob. 19,  1725: 154-156.  Prob. 1, part 3,  1778: 140-142; 1803: 138-139;  1814: ??NYS;  1840: omitted.  Take  1 + 1/2  twice but then subtract  2x  to get  r,  then divide  r  repeatedly by  2  until one gets down to  1.  Observing when he has to discard gives you the binary expansion of  r.  Subtracting  2x  yielded  r = a + b + c,  so  x =   4r ‑ (2b + 3c).

Prob. 31.  1696: 85;  1708: 76-77.  Prob. 35,  1725: 220-221.  21-card trick done with  36  cards.  He says the desired card will be in the middle of its row, i.e. in the 6th place! 

Prob. 15.  1778: 164-166;  1803: 165-166.  Prob. 14, 1840: 74-75.  Replaces the above with the  16  counter binary version as in Prevost and notes that one can use other powers of two.

Henry Dean.  The Whole Art of Legerdemain, or Hocus Pocus in Perfection.  11th ed., 1790?  ??NX -- seen at UCL Graves 124.b.36.  Pp. 89-90 (89 is misprinted as 87).  21 card trick.  "This trick may be done by an odd number of cards that may may be divided by three."

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 81-83, no. 129: To tell the Number a Person has fixed upon, without asking him any question.  A variant of the  3/2  method.  Pick a number from  1  to  15,  add  1  to it.  (Triple the number; if it is odd, add  1;  then halve it) thrice.  Finally, if the result is odd, add one, and again halve.  From observing when the person has to add one, you determine the number.  If   n + 1  =  8a + 4b + 2c + d,   where the coefficients are zero or one, the process is straightforward up to the third tripling which yields  54a + 27b + 15c + 9d  which is odd if and only if  b + c + d  is odd.  The next steps are no longer simply expressible in terms of the coefficients.  One has to add one at the third stage if and only if  n + 1 º 1, 2, 4, 7 (mod 8)  and one has to add one at the fourth stage if and only if  n + 1 = 1, 6, 8, 10, 11, 12, 13, 15.  He gives tables to determine  n  from the observations.

Manuel des Sorciers.  1825.  ??NX 

Pp. 47-48, art. 20.  15 card trick.

Pp. 60-61, art. 33.  16 card trick using awkward binary.

Boy's Own Book.  To tell a card thought of blindfold.  1828-2: 390-391;  1829 (US): 197;  1855 & 1859: 542-543;  1868: 640-641;  1880: 670-671;  1881 (NY): 196.  21 card trick.

Boy's Own Book.  1843 (Paris).

Pp. 396-397.  "A certain number of cards being shown to a person, to guess that which he has thought of."  Describes the trick with any multiple of three and says it is easier with an odd multiple such as 15, 21 or 27.  Remarks that once you know the card, you can shuffle them and pick the card from the pack.  c= Boy's Treasury, 1844, pp. 326-327: To guess the card thought of.  = de Savigny, 1846, p. 276-277: Deviner la carte pensée.  = The Secret Out, 1859, pp. 80-81: To Guess the Card Thought of.

P. 345.  "To tell the number thought of by a person."  PB1.  = Boy's Treasury, 1844, p. 302.  = de Savigny, 1846, pp. 289-290: Deviner un nombre pensé par une autre personne.

Magician's Own Book.  1857.  Which counter has been thought of out of sixteen?, pp. 226‑227.  Like Prevost but clearer.  Uses  16  cards and repeated halving to bring the thought of card to the top in four steps.  Says it can be extended to  32  cards.  = Boy's Own Conjuring Book, 1860, pp. 196-197.

Vinot.  1860. 

Art. XVII: Deviner le nombre pensé par une personne sans lui rien demander, pp. 34-36.  Do  3/2  twice, then subtract twice the original number -- this produces  a + b + c.  Have him then divide this repeatedly by two, taking the smaller halves -- this essentially gives you the binary digits for  a + b + c,  which you compute.  Then form  4(a + b + c)  which is  x + 3a + 2b,  so you can recover  x  by using the knowledge of when there was an odd division in the original  3/2-ings.

Art. XXIII: Deviner un nombre pensé, pp. 41-44.  Premier moyen.  PB1.

Art. XXXV: D'un nombre de cartes (15, 21 ou 27) deviner celle qui aura été pensée, pp. 54-55.  21 card trick clearly explained, using 21 as his example.

Indoor Tricks and Games.  Success Publishing, London, nd [1930s??].  The wonderful twenty-seven  &  Variants of the  27-card trick, pp. 47-51.  Uses ternary to locate and position a card among  27  -- but I find the description of how to perform the trick a bit cryptic.

 

          7.M.5.  LOONY LOOP  =  GORDIAN KNOT

 

            This is a configuration of wire loops attached to a rod with a loop of string to be removed.  The solution method uses the binary pattern of the Chinese Rings.  I have a note of an early 19C version, but no details -- ??

 

George E. Everett, via George Barker.  UK Patent 15,971 -- An Improved Puzzle.  Applied: 21 Sep 1891;  accepted: 24 Oct 1891.  1p + 1p diagrams.  Barker states that the invention was communicated by Everett, of Grand Island, Nebraska.  There is no indication of the binary pattern in the puzzle.

Though this puzzle does not appear in Hoffmann, 1893, Hordern has included it in a photo of Victorian puzzles omitted by Hoffmann on p. 256 of Hoffmann‑Hordern.  Unfortunately, this example has no date associated with it. 

                    It appears in Jaques puzzle boxes of c1900, named The Canoe Puzzle.  I have an example and Dalgety has several examples of these boxes with the puzzle and the solution which says that if the cord gets entangled, it can be cut and replaced!

M. Adams.  Indoor Games.  1912.  Pp. 337‑341.  The double link (= Loony Loop or Satan's Rings). 

Ch'ung‑En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1.  P. 21 shows a simple version.

 

          7.M.6.          BINARY BUTTON GAMES

 

            New section.

            The problems here are usually electronic puzzles with an array of lightable buttons which can take on two states -- lit and unlit.  Pressing a button toggles the lights on a certain pattern of buttons.  The earliest commercial example I know is the XL-25.  The box says Patent No. 122-8201 061.  However, there is an older similar electrical switching problem by Berlekamp -- see below.  There are also mechanical versions, such as Game Jugo from Japan (mid 1980s?), which has 15 petals such that turning one over turns over some others, and Orbik -- see below.  A number of further versions have appeared and I have seen some incorporated in game packages on computers.  Rubik's Clock is essentially the same kind of problem except that the states of the clocks and the possible turns are (mod 12) instead of (mod 2), though the main interest centres on whether a clock is correct or not.  Likewise Orbik is essentially the same, but with values (mod 4).

 

Elwyn Berlekamp.  Switching Game.  Devised and actually built at Bell Labs, c1970.  This is an  n x n  array of lights with  2n  switches which will invert the lights in a row or column.  Berlekamp's version is  n = 10.  See Fishburn & Sloane below.

Lazlo Meero.  The XL-25.   Exhibited at the London Toy Fair in 1983 and marketed by Vulcan Electronics of London.  The box says Patent No. 122-8201 061. 

David Singmaster.  The XL-25.  Cubic Circular 7/8 (Summer 1985) 39-42.  The XL-25 has a  5 x 5  array of buttons and two choices of toggle patterns -- a simple cross of  5  cells and the pattern of a cell and the cells which are a knight's move away.  In both cases, cells off the edge are simply lost.  I show that such problems with  n  buttons can be specified by an  n x n  input/output or transition binary matrix  A = (aij),  with  aij = 1  if pressing the  j-th button toggles the  i-th button;  otherwise  aij = 0.  If  x  is a binary column vector showing which buttons are pushed, then  Ax  is a binary column vector showing which buttons have been toggled.  From a random start, the object is to light all buttons.  So the general solution of the problem is obtained by inverting  A (mod 2).  For the  5 x 5  array, the knight version is invertible and I sketch its inverse.  But the cross version gives a matrix of rank  23.  I note that when the matrix is singular, the Gauss-Jordan elimination method for the inverse yields the null space and range and a kind of pseudo‑inverse, allowing for complete analysis -- in theory.  (More recently, I have seen that this is a proper pseudo-inverse and all the pseudo-inverses of  A  are computable by this process.)  Again I sketch the solution process.  (I now incorporate this idea in my teaching of linear algebra as a handout  "On trying to invert a singular matrix".)  I corresponded with Meero who had obtained some similar results and showed that each feasible pattern for the cross version could be obtained in at most  15  moves.  He also showed that if the input/output matrix  A  is symmetric and reflexive (i.e.  aii = 1),  then one can turn on all the lights, starting with them all off.  He studied the cross version on  n x n  boards up to  n = 100.  A friend of Meero studied fixed point patterns such that  Ax = x  and anti-fixed points where  Ax  is the vector complementary to  x.  Meero asserts the knight version has  A  invertible if  n ¹ 6, 7, 8 (mod 9).  I wondered what happens for cylindrical or toroidal boards, but made no attempt to study them.

In Spring 1986, I applied this method to a Japanese puzzle called  Game Jugo  which Edward Hordern had shown me.  This has  15  two-sided 'petals' around a centre which has four pointers.  When one pointer points to petal  1,  the others point to  4, 8, 11  and turning one of these over turns over the others in this set.  Since the sum of all rows of the transition matrix is the zero vector, it follows that the matrix is singular.  I found that it had rank  12,  so there are just  212 = 4096  achievable patterns, each of which has  23 = 8  solutions.  It turns out that there are three groups of  5  petals, e.g.  {1, 4, 7, 10, 13},  such that the sum of the turns in a group must be  0  for an achievable pattern.  From this, I showed that any achievable pattern could be reached in at most  6  moves, determined how many patterns required each number of moves and showed that the average number of moves was  4.6875.  I am very keen to get an example of this and/or its instructions (preferably with an English translation).

Donald H. Pelletier.  Merlin's magic square.  AMM 94:2 (Feb 1987) 143‑150.  Merlin is a product of Parker Brothers and provides several games, including a  3 x 3  binary button game.  If we number the cells   1, 2, 3;  4, 5, 6;  7, 8, 9;  then:  pressing  1  toggles  1, 2, 4, 5;  pressing  2  toggles  1, 2, 3;  pressing  5  toggles  2, 4, 5, 6, 8  (same as the cross on the XL-25).  The object is to light all but button  5.  He develops the binary vectors as above, finds that the transition matrix is invertible, computes the inverse and answers a few simple questions.

T. E. Gantner.  The game of quatrainment.  MM 61:1 (Feb 1988) 29-34.  Considers a game on a  4 x 4  field where a corner move reverses the six cells closest to the corner;  an edge move reverse the the neighbouring cells and a centre move reverse the cell and its four neighbours (i.e. the  +  pattern).  Sets up the matrix approach and shows the transition matrix is invertible, finding inputs which reverse just one cell.  Modifies the moves and finds versions of the game with ranks  12  and  14.

Wiggs, Christopher C.  &  Taylor, Christopher J. C.  US Patent 4,869,506 -- Logical Puzzle.  Filed: 29 Jul 1988;  patented: 26 Sep 1989.  Cover page + 2pp + 6pp diagrams.  This is the patent for what was marketed by Matchbox as Rubik's Clock in 1988.  The address of Wiggs and Taylor is just across the street from Tom Kremer's firm which is Rubik's agent.  This is essentially the same kind of problem except that the states of the clocks and the possible turns are  (mod 12)  instead of  (mod 2),  though the main interest centres on whether a clock is correct or not.  18  clock dials --  9  on front and  9  on back, both in a  3 by 3  pattern.  Four drive wheels on the edges, connected to the corner clocks, but their effects on other clocks are determined by the positioning of  4  buttons in the middle of the puzzle, giving  30  types of move.  The four corner front clocks are connected to the four corner rear clocks, so there are  14  independent motions to make and the input/output matrix is  30 by 14.

Daniel L. Stock.  Merlin's magic square revisited.  AMM 96:7 (Aug/Sep 1989) 608-610.  He gives an easy algorithm for solving the problem by doing edges, then corners, then middle.

P. C. Fishburn & N. J. A. Sloane.  The solution to Berlekamp's switching game.  Discrete Mathematics 74 (1989) 263-290.  They describe Berlekamp's game, with photo, as a coding theory problem.  The transition matrix  A  here is  2n x n2.  For any given initial state  x0,  consider all the states that can be achieved from it, say  S(x0)  =  { x0 + Ax | x Î Z22n }.  We might expect  S(x0)  to have  22n  states, but reversing all rows is the same as reversing all columns -- and there is no other dependence -- so there are  22n‑1  states.  Among all these states, there is one with a minimal number,  f(x0),  of lights turned on.  The covering radius  R  of the code formed by the  2n  rows and columns, considered as words in  Z2n^2,  is the maximum of these minimal numbers, i.e.  min { f(x0) | x0 Î Z2n^2 }.  These codes are called 'light‑bulb' codes and have been investigated since c1970 since they have the smallest known covering radius.  From our game point of view, the problem corresponds to finding the most-unsolvable position and  R  is a measure of unsolvability.  The values for  R  were known for  n £ 5.  The authors use extensive hand computing to extend this up through  n = 9  and then a lot of computer time to get to  n = 10.  The values of  R  for  n  =  1, .., 10  are:  0, 1, 2, 4, 7, 11, 16, 22, 27, 34.  That is, for the  10 x 10  game, there is an array of  34  turned-on lights which cannot be reduced to a smaller number of turned-on lights by any inversion of rows and columns.

Orbik.  Orbik is a ring of  12  wheels, each having  4  colours but just one colour can be seen through the top cover of windows.  There are three marks.  When one mark is at  1,  the others are at  4  and  8.  When the top cover is turned ahead, the marked wheels move forward one colour.  A backward turn leaves everything fixed, but moves the position of the marks.  Made by James Dalgety.  I believe it was Edward Hordern's exchange gift for a puzzle party, c1993. 

Edward Hordern.  Orbik.  CFF 29 (Sep 1992) 26-27.  ??NYR.  Orbik is a ring of  12  wheels, each having  4  colours but just one colour can be seen through the top cover of windows.  There are three marks.  When one mark is at  1,  the others are at  4  and  8.  When the top cover is turned ahead, the marked wheels move forward one colour.  A backward turn leaves everything fixed, but moves the position of the marks.

Ralph Gasser.  Orbik.  CFF 32 (Aug 1993) 26-27.  He counts both forward and backward turns and finds there are  60  antipodal positions requiring  54  moves to solve.  The shortest processes for moving a single wheel by  1, 2, 3  colours take  29, 28, 25  moves.  If a sequence of turns in the same direction is counted as a single move, there are  4  antipodal positions requiring  23  moves to solve and the single wheel processes take  9, 7, 9  moves.

Revital Blumberg, Michael Ganor & Avish J. Weiner.  US Patent 5,417,425 -- Puzzle Device.  Filed: 8 Apr 1994;  patented: 23 May 1995.  Cover page + 3pp + 1p correction + 4pp diagrams.  This is the patent for Lights Out, which is essentially identical to XL-25, but has some additional patterns.  Patents 5,573,245 & 5,603,500, granted to different groups of people, continue this.  No reference to Meero or any Hungarian patent, but cites Parker Brothers' Merlin as undated.

Dieter Gebhardt.  Cross pattern piling.  CFF 33 (Feb 1994) 14-17.  Notes that Dario Uri independently invented the XL-25 idea with the cross pattern -- he called it Matrix of Lights.  Gebhardt modifies the problem by making two ons remain on.  Thus his computation of  Ax  is an ordinary matrix product and he wants results with each entry the same.  If one thinks of the cross shapes as five cubes piled onto the board, the sought result is a uniformly stacked board of some height  h.  This also allows for some cell to be turned on several times.  Thus we are trying to solve  Ax = hJ,  where  J  is the vector of all  1s,  h  is a positive integer and  x  is a vector with non-negative integer entries.  Obviously the minimal value of  h  is wanted.  He determines solvability and all minimal solutions up to  8 x 8,  with  9 x 9  given as a contest.

Tiger Electronics, 980 Woodlands Parkway, Vernon Hills, Illinois, 60061, USA  &  Belvedere House, Victoria Avenue, Harrogate, UK.  Lights Out.  Model 7-574, 1995.  Essentially the same concept as the  XL-25  with its 'cross' pattern.  With lots of preprogrammed puzzles, random puzzles and option to input your own puzzles.  The longest solution is  15  moves, as found by Meero for the  XL-25.

Uwe Mèffert produced Orbix (or Light Ball) in 1995 for Milton Bradley.  I advised a bit on the design of the games.  This is a sphere with  12  light buttons in the pattern of a dodecahedron.  There are four different games.  The object is to turn all lights on, but in some games, one can also get all lights off.  However, only the first game is a linear transformation in the sense discussed above.  The later games have rules where the effect of a button depends on whether it is lit or not and even on whether the opposite button is lit or not.  Nonetheless all examples are solvable in  12  moves or less.

Edward Hordern.  What's up?  CFF 38 (1995) 38.  ??NYR.  Discusses Tiger Electronics'  Lights Out.

Dieter Gebhardt & Edward Hordern.  How to get the lights of "Lights Out" out.  CFF 39 (1996) 20-22.  ??NYR.  Sketches a solution.

Edward Hordern.  What's up?  CFF 41 (Oct 1996) 42.  Discusses Tiger Electronics'  Deluxe Lights Out  which has a  6 x 6  array with several options -- one can affect five lights in the form of a  +  or of a  x;  a button can have effect only if it is lit, or alternately lit/unlit.

 

          7.N.   MAGIC SQUARES

 

      4    9    2                    The  3 x 3  magic square is usually given in the form on the left.

      3    5    7          We denote each of the 8 possible forms by its top row.  I.e. this is the

      8    1    6          492  form.  All Chinese material seems to give only this form, called the

                              Lo Shu [Lo River Writing].

 

            7

            2                                    An unrelated diagram, shown on the left, is called the Ho Thu

8    3    5    4    9              diagram [River Plan].  See 7.N.5 for magic versions of this shape.

            1

            6

 

            Semi‑magic  denotes a square whose rows and columns add to the magic constant, but not the diagonals.

            Pandiagonal  means that the 'broken diagonals' also add to the magic constant.  Lucas called these  diabolic  and they are also called  Nasik,  as they were studied by Frost, who was then living in Nasik, India.

            Associated  or  complementary  means that two cells symmetric with respect to the centre add to  n2 + 1.

            See 7.AC.3 for related pan-digital sums.

 

            The early history of magic squares remains rather obscure.  In particular, the first four Chinese sources below are extremely vague!  Cammann‑4 argues that magic squares had powerful mystic meanings to the Chinese, Indians and Arabs, hence were not explicitly described in writing.  However, some modern scholars doubt if the Chinese had any magic square before 10C! -- cf Gardner, 1996.

            There are several surveys of some or all of the history of magic squares which I list first for later reference.  These provide many more references.

            17‑20C material has generally been omitted, but see Bouteloup.  Smith & Mikami discuss several workers in Japan, but I've omitted some of them.

 

                    SURVEYS

 

Wilhelm Ahrens - 1.  Studien über die "magischen Quadraten" der Araber.  Der Islam 7 (1917) 186‑250.

Wilhelm Ahrens - 2.  Die "magischen Quadrate" al-Būnī's.  Der Islam 12 (1922) 157‑177. 

Schuyler Cammann ‑ 1.  The evolution of magic squares in China.  J. Amer. Oriental Soc. 80 (1960) 116‑124.

Schuyler Cammann ‑ 2.  The magic square of three in old Chinese philosophy and religion.  History of Religions 1 (1961) 37‑80.  ??NYR

Schuyler Cammann ‑ 3.  Old Chinese magic squares.  Sinologia 7 (1962) 14‑53.

Schuyler Cammann ‑ 4.  Islamic and Indian magic squares I  &  II.  History of Religions 8 (1968‑69) 181‑209  &  271‑299.

Bibhutibhusan Datta & Avadhesh Narayan Singh.  Magic squares in India.  Indian J. History of Science 27:1 (1992) 51-120.  All references to Datta & Singh in this section are to this paper, not their book.

Menso Folkerts.  Zur Frühgeschichte der magischen Quadrate in Westeuropa.  Sudhoffs Archiv 65:4 (1981) 313-338. 

Heinrich Hermelink.  Die ältesten magischen Quadrate höher Ordnung und ihre Bildungsweise.  Sudhoffs Arch. 42 (1953) 199‑217.

Lam Lay Yong.  1977.  See under Yang Hui below.  Her commentary surveys the history.

Needham.  1958.  Pp. 55‑61.  See also:  vol. 2, 1956, pp. 393 & 442;  Vol. 5, Part IV, 1980, pp. 462-472.

Jacques Sesiano ‑ I  &  II.  Herstellungsverfahren magischer Quadrate aus islamischer Zeit (I)  &  (II).  Sudhoffs Arch. 64 (1980) 187‑196  &  65 (1981) 251‑265.

 

            A number of the entries in 5.A -- The Fifteen Puzzle -- refer also to magic squares.  See:  Loyd (1896);  Cremer (1880);  Tissandier (1880 & 1880?);  Cassell's (1881);  Hutchison (1891). 

            Some entries in 5.A and here give problems of sliding the Fifteen Puzzle into a magic square.  See:  Dudeney (1898);  Anon & Dudeney (1899);  Loyd (1914);  Dudeney (1917);  Gordon (1988) in 5.A  and  Ollerenshaw & Bondi below.

 

                    POSSIBLE EARLY REFERENCES

 

Anon.  Shu Ching.  c‑650.  Part V, book IV, The Great Plan -- commentary and book XXII, The Testamentary Charge.  IN:  J. Legge, trans.  The Texts of Confucianism, Translated: Part I.  The Shû King, the religious portions of the Shih King, the Hsiâo King.  OUP, 1879, pp. 138‑139 & 239.  P. 138 discusses the Lo Shu and says it does not occur.  On p. 139, we see:  "To [Yu], Heaven gave the Great Plan with its nine divisions ..."  Various commentators, from  Gan‑Kwo  on, have asserted that this was the Lo‑shu which appeared on the back of a tortoise in the river Lo.  Legge says there is no evidence to connect the Lo‑shu with the Great Plan and that the commentators are indulging in leger‑de‑plume.  P. 239 mentions the River Plan.

                    (See also:  J. Legge, trans.; The Chinese Classics, etc.; Vol. III -- Part II; Trübner, London, 1865; pp. 321‑325 & 554.  This gives the Chinese and the English, with extensive notes.)

                    At this time, the number 'nine' was used to describe the largest number and hence does not necessarily imply  32.

Anon.  Lun Yu (Confucian Analects).  c‑5C.  Book IX, Tsze Han; chap. VIII.   IN:  J. Legge, trans.  The Chinese Classics, etc. vol. 1, Confucian Analects, The Great Learning, and the Doctrine of the Mean.  Trübner, London, 1861, p. 83.  = The Life and Teachings of Confucius; Trübner, London, 1869, pp. 169‑170, ??NX.  Also in:  A. Waley; The Analects of Confucius; Allen & Unwin, London, 1949, p. 140.  "The river sends forth no map."

Chuang Tzu (= Kwang‑Sze).  The Writings of Kwang‑Sze.  c‑300.  Part II, sect. VII  = Book XIV, Thien Yu (The Revolution of Heaven).  IN:  J. Legge, trans.  The Texts of Tâoism.  OUP, 1891.  Vol. 1, p. 346.  Refers to "the nine divisions of the writing of Lo."

Anon.  Ta Chuan (= Hsi Tzhu Chuan) (The Great Commentary on the I Ching [= Yi Jing]).  c‑300?? (Needham, vol. 2, p. 307, says c‑100 and vol. 5:IV, pp. 462-463, says -2C)  IN:  J. Legge, trans.  The Texts of Confucianism, Part II: The Yî King.  OUP, 1882.  Appendix III, sect. 1, chap. 12, art. 73, p. 374 & note on p. 376.  [There is a 1963 Dover ed. of Legge's 1899 edition.]  Also as:  Part I, chapter IX -- On the Oracle.  IN:  The I Ching, translated by R. Wilhelm and rendered into English by C. F. Baynes, 3rd ed., 1968, Routledge and Kegan Paul, London, pp. 308‑310.  The text is: "The Ho gave forth the map, and the Lo the writing, of (both of) which the sages took advantage."  This occurs just after paragraphs on the origin of the hexagrams and legend says the Ho Thu inspired the creation of the 8 trigrams.  Legge says the original Ho Thu map was considered to be lost in the -11C and the earliest reconstruction of it was presented during the reign of Hai Zung in the Sung Dynasty (1101-1125).  The I Ching is often cited but only this later commentary mentions an association of numbers with concepts.  Later commentators interpret this association as referring to the Ho Thu and Lo Shu diagrams, though this is not obvious from the association -- the names were not associated with the diagrams until about the 10C -- see Xu Yiu below.  See Needham, vol. 2, pp. 393 & 442 for discussion of the interpolation of the diagrams into the I Ching.)

 

E. John Holmyard.  Alchemy.  Penguin, 1957, pp. 36-38, discusses magic squares in relation to Chinese alchemy and Taoism.  He says the Taoist emphasis on the number 5 is related to its being the central value of the order 3 magic square.  He says this relation has been studied by H. E. Stapleton, but gives no reference.  Stapleton says the square of order 3 was the ground plan of the Ming-Tang or Hall of Distinction.  This Hall was used for the proclamation of monthly ordinances and the Calendar (which was partly lunar and hence of variable length).  When in the Ming-Tang, the Emperor became the incarnation of the god and hence the ground plan became of major importance in Chinese alchemy.  Stapleton conjectures that the original numbering of the  3 x 3  array of rooms in the Hall may not have been magic, but would have had 5 in the middle and that the magic numbering may have occurred at some time and been recognised as having special properties.  Holmyard indicates the Ming-Tang arose about -1000.  All in all, these statements do not agree with most of the other material in this section and it would be good to locate the work of Stapleton (??NYS), which is presumably well-known to students of ancient chemistry/alchemy.

Needham, Vol. 5:IV, 1980, pp. 462-472.  Cites Stapleton on p. 462 and indicates his work is a bit cranky, but I haven't got the details yet.  He goes on to discuss why 9 was so important to the Chinese.  He describes the tour of the pole-star sky-god Thai I which went through the nine cells of the Lo Shu in the order: 5, 1, 2, 3, 4, 5, pause, 5, 6, 7, 8, 9, 5.  His fig. 1535 shows this from a Tang encyclopedia, though this has the  2 7 6  orientation of the square.  The Chinese could see Yin and Yang (= even and odd), the Four Seasons and the Five Elements, and the Nine Directions of space, all in the Lo Shu.  Consequently it was not revealed to the general public until the end of the Tang (618‑907).  Needham then discusses the influence of the Lo Shu on Arabic alchemical thought.

Nâgârjuna.  c1C.  Order 4 squares, including one later called Nâgârjunîya after him, described in a MS on magic called Kakşapuţa [NOTE:  ş,  ţ  denote  s,  t  with underdot.], nd.  ??NYS.  [A. N. Singh; History of magic squares in India; Proc. ICM, 1936, 275‑276.  Datta & Singh.]  Datta & Singh say Nâgârjuna gives several rules for forming magic squares of order 4, but all the examples given do not use consecutive values, much less the first 16 positive integers, e.g.

                    n-3,   1, n-6,   8;   n-7,   9, n-4,   2;     6, n-8,   3, n-1;    4, n-2,   7, n-9,

          which has magic constant  2n.  The Nâgârjuna square is the following, with an unrelated structure and with constant 100.

                    30, 16, 18, 36;   10, 44, 22, 24;   32, 14, 20, 34;   28, 26, 40,  6.

          There is also no reference to Nâgârjuna or his book.  Can anyone provide this?

Tai the Elder.  Ta Tai Li Chi  (Record of Rites).  c80.  Chap. 67,  Ming Thang.  ??NYS  (See Needham, p. 58.)  Chap. 8, p. 43 of  Szu‑pu ts'ung‑k'an  edition, Shanghai, 1919‑1922.  Describes the  492  form.  (See Cammann‑2.)  (Cammann‑1,  Lam  and  Hayashi  say this is the first clear reference.)

Anon.  I Wei Chhien Tso Tu.  c1C.  Chap. 2, p. 3a.  ??NYS  (Translated in Needham, p. 58.)

Anon.  Lî Kî.  c2C.  Book VII --  Lî Yun,  sect. IV.  IN:  J. Legge, trans.  The Texts of Confucianism, Part III: The  Li Ki,  I‑X.  OUP, 1885.  Pp. 392‑393.  "The Ho sent forth the horse with the map (on his back) 1.   1 The famous 'River Map' from which, it has been fabled,  Fû-hsî  fashioned his eight trigrams.  See vol. xvi, pp. 14-16."  This last reference is ??NYS.

Theon of Smyrna.  c130.  Part B:  Βιβλιov τα τησ εv Αριθμoσ Μoυσικησ θεωρηματα Περιεχov  (Biblion ta tes en Arithmos Mousikes Theoremata Periechon).  Art. 44.  IN:  J. Dupuis, trans.; Théon de Smyrne; Hachette, Paris, 1892; pp. 166‑169.  (Greek & French.)   Natural square -- often erroneously cited as magic and used to 'prove' the Greeks had the idea of magic squares.

Xu Yiu (= Hsu Yo  = Xu Yue).  Shu Shu Ji Yi  (= Shu Shu Chi I)  (Memoir on Some Traditions of Mathematical Art).  190(?).  ??NYS.  Ho Peng Yoke  [Ancient Chinese Mathematics;  IN:  History of Mathematics, Proc. First Australian Conf., Monash Univ., 1980; Dept. of Math., Monash Univ., 1981, pp. 91‑102], p. 94, says that this is the earliest Chinese text to give the order 3 square.

              The date and authorship of this is contentious.  Current belief is that this was written by  Zhen Luan  (= Shuzun)  in c570, using the name of  Xu Yue.  Li & Du, pp. 96‑97, say that this work first introduces the diagram.  The diagram was called the "nine houses computation".  The diagram was connected with the  Yi Jing  commentary in the 10C and then renamed  Lo Shu.  After the 13C, magic squares were called  zong heng tu  (row and column diagrams).

                    Needham, vol. 5:IV, p. 464, considers this as being c190, referring the Chen Luan as a commentator on it.  He calls the diagram "Nine Hall computing method".

Varahamihira (= Varāhamihira (II)).  Bŗhatsamhitā [NOTE:  ŗ  denotes  r  with an underdot  it and the  m  should have an underdot.].  c550.  Hayashi, below, cites a Sanskrit edition (NYS) and the following.

                    M. Ramakrishna Bhat.  Varāhamihira's  Bŗhat Samhitā [NOTE:  ŗ  denotes  r  with an underdot  it and the  m  should have an overdot.] with English Translation, Exhaustive Notes and Literary Comments.  2 vols, Motilal Banarsidass, Delhi, 1981‑1982.  Vol. II: Chapter LXXVII -- Preparation of perfumes, pp. 704-718.  On pp. 714-715 is the description of a  4 x 4  array:  

                    2, 3, 5, 8;   5, 8, 2, 3;   4, 1, 7, 6;   7, 3, 6, 1,   with some cryptic observations that any mixture totalling  18  is permitted, e.g. "by combining the four corners, or four things in each corner, or the central four columns, or the four central ones on the four sides."  As given, many of the groups indicated do not add up to  18,  nor do the columns.  However, the following article notes that the bottom row should read  7, 6, 4, 1!!  (Datta & Singh have this correct.)

                    This material is described and analysed in:  Takao Hayashi; Varāhamihira's pandiagonal magic square of the order four; HM 14 (1987) 159-166.  He gives the book's name as  Bŗhatsamhitā [NOTE:  ŗ  denotes  r  with an underdot  it and the  m  should have an underdot.]  and says the material is in Chapter 76 (Combinations of perfumes).  He gives the correct form of the array.  He notes that the array is a pandiagonal magic square with constant  18,  except the entries are  1, ..., 8  repeated twice.  Hayashi believes that Varāhamihira must have known one of the actual magic squares which yield this square when the numbers are taken  (mod 8).  He shows there are only  4  such magic squares, two of which are pandiagonal.  One of the pandiagonal squares is a rotation of: 

                     8, 11, 14,  1;    13,  2,  7, 12;     3, 16,  9,  6;    10,  5,  4, 15, 

          which he describes as the most famous Islamic square of order  4  described in Ahrens‑2.  Hayashi feels that order  4  squares must originate in India, contrary to Cammann's thesis.  (See also Singh, op. cit. at  Nâgârjuna,  above, and  Ikhwān al- Şafā’ [NOTE:  Ş  denotes  S  with an underdot.], below.)  Datta & Singh just say that Varāhamihira gives a magic square.)

                    Datta & Singh give the square beginning  2, 3, 5, 8,  and says that it is a special case of

                    n-7,   3, n-4,   8;     5, n-1,   2, n-6;     4, n-8,   7, n-3;   n-2,   6, n-5,   1.

          Taking  n = 9  gives Varāhamihira's square and taking  n = 17  gives the square with   8, 11, 14,  1  as right hand column.  One can also take the other set of the first eight integers as given, getting a square starting    2, n-6,   5, n-1.  Varāhamihira's square has many magic properties and he called it 'Sarvatobhadra' (Magic in all respects).  They say these properties are fully described by Varāhamihira's commentator Bhaţţotpala [NOTE:  ţ  denotes  t  with an underdot.] in 966.

Jabir ibn Hayyan  (= Jâbir ibn Hayyân  = Geber) (attrib.).  Kitâb al‑Mawâzin  (Book of the Balances).  c800.  ??NYS -- discussed in Ahrens-1.  The Arabic and a French translation are in:  M. Berthelot; La Chimie au Moyen Age: Vol. 3 -- "L'Alchimie Arabe"; Imprimerie Nationale, Paris, 1893.  The text is discussed on pp. 19-20, where he refers to the magic square of Apollonius, with a footnote saying 'De Tyane'.  The text is given on p. 118 (Arabic section) & 150 (French section).  Gives  3 x 3  square in form  492. 

                    "Here is a figure divided into three compartments, along the length and along the width.  Each line of cells gives the number  15  in all directions.  Apollonius affirms this is a magic tableau formed of nine cells.  If you draw this figure on two pieces of linen [or rags], which have never been touched by water, and which you place under the feet of a woman, who is experiencing difficulty in childbirth, the delivery will occur immediately." 

                    The French is also in Ahrens-1, who notes that the square does not appear in the few extant writings of Apollonius of Tyana (c100).

                    Hermelink mentions this as the earliest Arabic square, but gives no details.  Needham, vol. 5:IV, p. 463, says Cammann gave this, and dates this as c900.  Folkerts gives this as the first Arabic example.

                    Suter, pp. 3‑4, doesn't mention magic squares for ibn Hayyan, but this appearance is simply in a list of questions on properties of animals, vegetables and minerals, so hardly counts as mathematics.

                    Holmyard [op. cit. above, pp. 74-75] discusses the work of Kraus and Stapleton on Jabir.  Jabir considers the numbers  1, 3, 5, 8  as of great importance -- these are the entries in the lower left  2 x 2  part of his magic square.  These add to  17  and everything in the world is governed by this number!  He also attaches importance to  28  which is the sum of the other entries.  Holmyard asserts this magic square was known to the Neo-Platonists of 3C -- an assertion which I have not seen elsewhere.  Jabir uses ratios  1/3  and  5/8  extensively in his alchemical theories.

‘Ali ibn Sahl Rabbān al-Tabarī (d. 860).  Paradise of Wisdom.  This is a gynaecological text discovered by Siggel.  ??NYS -- described in Needham, vol. 5:IV, p. 463.  Example of a magic square used as a charm in cases of difficult labour.  Needham thinks this is the earliest Arabic magic square.

Tâbit ibn Qorra  (= Thâbit ibn Qurra).  c875.  This is the first reference to magic squares in Suter, on pp. 34‑38, but he seems to say that the work has not survived and Ahrens-1 confirms this.  Needham, vol. 5:IV, p. 463, says Cammann wonders if this ever existed.

Ikhwān al‑Şafā’ [NOTE:  Ş  denotes  S  with an underdot.].  Rasā’il  (Encyclopedia) (??*).  c983.  Cairo edition, 1928, p. 69.  ??NYS.  Paris MS Arabe 2304 (formerly 1005) of this is the work translated by F. Dieterici as:  Die Propaedeutik der Araber im zehnten Jahrhundert; Mittler & Sohn, Berlin, 1865;  reprinted as vol. 3 of F. Dieterici; Die Philosophie bei den Araben im X.Jahrhundert n. Chr.; Olms, Hildesheim, 1969.  Pp. 42‑44 (of the 1969 ed.) shows squares of orders  3, 4, 5 and 6.  The order  3  square is in the form  276.  The text refers to orders  7, 8 and 9  and gives their constants.  On p. 44, the translator notes that the Arabic text has some further incomplete diagrams which are not understandable.  Hermelink and Cammann‑4 say that the Cairo ed. is the only version to give these diagrams.  Ahrens-1 says it continues with a cryptic description of the use of a  9 x 9  square on two sherds, which have not been sprinkled with water, for easing childbirth.  The prescription has several more details than ibn Hayyan's.

              The Arabic text and rough translation are given in:  van der Linde; Geschichte und Literatur des Schachspiels; op. cit. in 5.F.1, vol. 1, p. 203.  This is a description of the  3 x 3  magic square, form  492  or  294,  in terms of chess moves.  Ahrens-1 says that Ruska tells him that much, if not all, of the magic square and adjacent material in Dieterici was added later to the Encyclopedia.  Ruska says there are many errors in the translation and Ahrens cites several further errors in nearby material.

                    Hermelink describes the methods and reconstructs the squares of orders  7, 8, 9  from the 1928 Cairo ed. and van der Linde.  Cammann‑4 says he obtained the same squares independently, but he doesn't agree on all the interpretations.  He feels there are Chinese influences, possibly via India, and gives his interpretations.

                    The square of order  4  is given by Hayashi, op. cit. above at Varāhamihira, as: 

           4, 14, 15,  1;     9,  7,  6, 12;     5, 11, 10,  8;    16,  2,  3, 13.  This is not pandiagonal.  The square of order  7  is doubly bordered -- the first such.

(Abû ‘Alî el‑Hasan ibn el Hasan) (the  Hs  should have dots under them) ibn el‑Haitam.  c1000.  ??NYS.  Suter, p. 93, says he wrote:  Über die Zahlen des magischen Quadrates.  He cites Woepcke, ??NYS, for MS details.  Ahrens-1 indicates that the work does not exist.

J. H. Rivett‑Carnac.  Magic squares in India.  Notes and Queries (Aug 1917) 383.  Quoted in:  Bull. Amer. Math. Soc. 24 (1917) 106,  which is cited by:  F. Cajori; History of Mathematics; op. cit. in 7.L.1, pp. 92‑93.  The square is in the ruins of a Hindu temple at Dudhai, Jhansi, attributed to the 11C.  It is  4 x 4,  and each  2 x 2  subsquare also adds to  34, but the full square is not given.

                    Cammann‑4, p. 273, says this is the same as the Jaina square at Khajuraho described below and cites the archaeological report, ??NYS.  He is dubious about the date.

(Muhammed ibn Muhammed ibn Muhammed, Abû Hâmid,) el‑Ġazzâlî  (= al‑Ghazzali).  Mundiqh.  c1100.  ??NYS -- described by Ahrens-1.  Ahrens cites two differing French editions which give  3 x 3  forms  492  and  294.  He says the latter is a transcription error.  Al‑Ghazzali's  text is very similar to  ibn Hayyan's,  though one translator says the cloths are moistened.  Ahrens discusses this point.  He says that amulets with this magic square, called 'seal of Ghazzali' are still available in the Middle East.

                    Lam, p. 318, cites this as an early Arabic magic square, but doesn't give details.  Suter, p. 112, doesn't mention magic squares.

Abraham ibn Ezra.  Sepher Ha‑Schem  (Book of Names), 12C,  and  Jesod Mora,  1158.   ??NYS -- both are described in:  M. Steinschneider; op. cit. in 7.B  and excerpted in the next item.  The material is art. 13, pp. 95++.  The  672  form is shown on p. 98.  Steinschneider, p. 98, also gives the  492  form and says it appears in  Jesod Mora,  described on pp. 99‑101.  Ahrens-1 only mentions that  Sepher Ha‑Schem  gives an order  3  square.

Abraham ibn Ezra.  Sêfer ha‑Echad.  c1150.  Translated and annotated by Ernest Müller as:  Buch der Einheit; Welt-Verlag, Berlin, 1921, with excerpts from:  Jessod Mora,  Sefer ha‑Schem,  Sefer ha-Mispar  and his Bible commentary. 

Sefer ha-Echad,  p. 25, has a reference to areas of squares which Müller thinks may refer to magic squares.

Sefer ha-Schem,  pforte VI, p. 49, discusses the order  3  square.  A note says to see Fig. 6, which appears on p. 80 and is the  492  form.  Müller's notes, p. 64, observe that the magic square of order  3  is essentially unique and makes some mystic comments about this.

Abraham ibn Ezra.  Ta'hbula.  c1150.  ??NYS.  Some source says this has magic squares, but Lévi's comments in 7.B indicate that this book is only concerned with the Josephus problem.  Steinschneider's description of  Tachbula,  pp. 123‑124 of the above cited article, makes no mention of a magic square.

Anon.  Arabic MS, Fatih 3439.  c1150.  ??NYS.  Described in Sesiano‑I.  Construction of squares of almost all orders.  Describes:  a method of  ibn al‑Haytham  (c1000) for odd orders;  a method of  al‑Isfarâ’inî  (c1100) for evenly even orders;  a method of  ibn al‑Haytham  for oddly even squares which only works for order  º 2 (mod 8).  Suter, p. 93, mentions  ibn al‑Haitam  -- see above, c1000.

Tshai Yuan‑Ting.  Lo‑Shu diagram, c1160.  ??NYS -- Biggs cites this as being in Needham, but the only references to Tshai in Needham refer to indeterminate analysis (p. 40) and geology (p. 599).  Paul Carus [Reflections on magic squares, IN: W. S. Andrews, op. cit. in 4.B.1.a, pp. 113-128, esp. p. 123] says that Ts'ai Yüan-Ting (1135-1198) gives the Lo-Shu diagram 'but similar arithmetical diagrams are traceable as reconstructions of primitive documents among scholars that lived' during 1101-1125.  Datta & Singh cite this and say this is the earliest Chinese interpretation of the Lo-Shu as a magic square.  This ignores  Tai the Elder,  I Wei Chhien Tso Tu,  and  Xu Yiu,  though the first two are a bit vague.

(Ahmed (the  h  should have an underdot) ibn ‘Alî ibn Jûsuf) el‑Bûnî, (Abû'l‑‘Abbâs, el‑Qoresî)  = Abu‑l‘Abbas al‑Buni  (??= Muhyi'l‑Dîn Abû’l-‘Abbâs al‑Bûnî  -- can't relocate my source of this form.)  Kitâb et‑chawâşs [NOTE:  ş  denotes an  s  with an underdot.]  (= Kitab al Khawass  or  Sharkh ismellah el‑a‘zam??)  (The Book of Magic Properties).  c1200.  Suter, p. 136, mentions magic squares.  ??NYS -- described in:  Carra de Vaux; Une solution arabe du problème des carrés magiques; Revue Hist. Sci. 1 (1948) 206‑212.  Construction of squares of all orders by bordering.  Hermelink refers to two other books of  al‑Buni,  ??NYS.

al‑Buni.  Sams al‑ma‘ârif  = Shams al‑ma‘ârif al‑kubrâ  = Šams al-ma‘ārif.  c1200.  ??NYS.  Ahrens-1 describes this briefly and incorrectly.  He expands and corrects this work in Ahrens-2, which mainly deals with  3 x 3  and  4 x 4,  the various sources and the accumulated errors in most of the squares.  He notes that a  4 x 4  can be based on the pattern of two orthogonal Latin squares of order  4,  and  Al-Buni's  work indicates knowledge of such a pattern, exemplified by the square (discussed by Hayashi under Varāhamihira, c550)

           8, 11, 14,  1;    13,  2,  7, 12;     3, 16,  9,  6;    10,  5,  4, 15   considered  (mod 4).  Al‑Buni  gives several  4 x 4's,  including that of  Ikhwān al‑Şafā’ (the  Ş  should be an  S  with a dot under it),  c983, which does not have the above pattern.  He also has Latin squares of order  4  using letters from a name of God.  He goes on to show  7  Latin squares of order  7,  using the same  7  letters each time -- though four are corrupted.  (Throughout, the Latin squares also have 'Latin' diagonals.)  These are arranged so each has a different letter in the first place.  It is conjectured that these are associated with the days of the week or the planets.  In Ahrens-1, Ahrens reported that he had recently been told that  Al-Buni  had an association of magic squares of orders  3  through  9  with the planets, but he had not been able to investigate this.  In Ahrens-2, he is clear that  al-Buni  has no such association -- indeed, there is no square of order  9  anywhere in the standard edition of the works of  al-Buni.  But Folkerts says such an association was made by the Arabs, perhaps referring to the Nadrûnî, below.  See 14C & 15C entries below.

                    Cammann‑4, p. 184, says this text is "deliberately esoteric ... to confuse people" and the larger squares are so garbled as to be incomprehensible.  On pp. 200‑201, he says this has the knight's move method for odd orders.  Later it was noted that any number could be in the centre and  1  was popular, giving the 'unit centred' square of symbolic importance.  These squares are also pandiagonal.  Al‑Buni  gives many variant  4 x 4  squares with the top row spelling some magical word -- e.g. one of the  99  names of God.  He mentions a "method of the Indians", possibly the lozenge method described in Narayana, 1356.

BM Persian MS Add. 7713.  1211?  Described in Cammann‑4, pp. 196ff.  On p. 201, Cammann says p. 23 gives unit centred squares of orders  5  &  9,  pp. 112‑114 gives a rule for singly even order and p. 164 has an order  20  square.  This also has odd order lozenge squares -- see Narayana, 1356.  It also has some examples of a form of the system of broken reversions.

Persian MS.  1212.  Garrett Collection, No. 1057, Princeton Univ.  See  Cammann‑1  &  Cammann‑4, p. 196.  ??NYS

                    Cammann‑4, pp. 196ff, says the above two MSS show new developments and describes them.  Diagonal rules for odd orders first appear here and give an associated square with centre  (n2 + 1)/2  which acquired mystic significance as a symbol of Allah.

(Jahjâ (the  h  should have a dot under it) ibn Muhammed ibn ‘Abdân ibn ‘Abdelwâhid, Abû Zakarîjâ Neġm ed‑dîn,)  known as  Ibn el‑Lubûdî  (= Najm al‑Din  (or  Abu Zakariya)  al‑Lubudi.  c1250.  Essay on magic squares dedicated to al‑Mansur.  ??NYS.  Mentioned in Suter, p. 146.

Yang Hui.  Hsü Ku Chai Ch'i Suan Fa  (= Xugu Zhaiqi Suanfa)  (Continuation of Ancient Mathematical Methods for Elucidating the Strange [Properties of Numbers]) (Needham, vol. 5:IV, p. 464, gives: Choice Mathematical Remains collected to preserve the Achievements of Old).  1275.  IN:  Lam Lay Yong;  A Critical Study of the  Yang Hui Suan Fa;  Singapore Univ. Press, 1977.  Book III, chap. 1, Magic Squares, pp. 145‑151 and commentary, pp. 293‑322.  This is the only source for older higher order squares in China.  (See Cammann‑1 and Cammann‑3 for details of constructions.)  Bordered squares of order  5  and  7.  Magic squares of orders  3  through  10, the last being only semimagic.  Methods are given for orders  3  and  4  only.  Gives some magic circles and other forms.  (Lam's commentary, p. 313, corrects the first figure on p. 150.  Lam also discusses the constructions.)

                    Li & Du, pp. 166‑167, say that  6 x 6  and  7 x 7  'central' (= bordered) squares arrived from central Asia at about this time.  The  6 x 6  example on their p. 172 and the  7 x 7  example on their p. 167 are bordered.

              Cammann‑4 says one of the order  8  squares is based on a Hindu construction.

Jaina square.  Inscription at Khajuraho, India.  12‑13C.  Ahrens-1 (p. 218) says it first appears as:  Fr. Schilling; Communication to the Math. Gesellschaft in Göttingen [Mitteilung zur Math. Ges. in Göttingen], 31 May 1904;  reported in:  Jahresber. Deutschen Math.‑Verein. 13 (1904) 383‑384.  (Schilling is reporting a communication from F. Kielhorn.)

                     7, 12,  1, 14;    2, 13,  8, 11;   16,  3, 10,  5;    9,  6, 15,  4.

          It is pandiagonal and associated.  Cammann‑4, p. 273, says this is the same as the square reported by Rivett-Carnac (qv above) and there claimed to be 11C?  Cammann feels it may derive from an Islamic source.  See also:  Smith History II 594  and  Singh, op. cit. at Nâgârjuna above.  Datta & Singh give this and date it as 11C.

Five cast iron plates with  6 x 6  magic squares, late 13C(??), were found at Xian in 1956.  The numerals are similar to East Arabic numerals so these reflect the Arabic influence on the Mongol dynasty.  Li & Du, p. 172, reproduces one.  This is on display in the new Provincial History Museum in Xian.  Jerry Slocum has given me a facsimile, in reduced size, of the same one.  Can any one supply more information about the others??

Datta & Singh give another Jaina square, from 'not later than the fourteenth century', 'probably a very old one'.  This is the following, but with all entries multiplied by five to give a magic constant of 170, which 'is closely connected with an ancient Jaina mythology'.

                     5, 16,  3, 10;    4,  9,  6, 15;   14,  7, 12,  1;   11,  2, 13,  8.

Folkerts discusses an anonymous and untitled astrological-magic treatise which appears to derive from the court of Alfonso the Wise in Madrid, where a similar work of al-Magriti (al-Mağrīţī [NOTE:  ţ  denotes  t  with an underdot.]) (10C) had been translated in the 13C and developed under the name Picatrix.  Picatrix refers to astrological amulets but gives no instance of a magic square on one.  Though there is no known direct connection, Folkerts considers this treatise to be in the tradition of the Picatrix and names it 'Picatrix-Tradition'.  He finds seven MSS of it, from early 14C onward.  This specifically associated magic squares and planets according to Folkerts' System I.

                    Folkerts discusses these associations and calls them Systems I and II.  For  n = 3, 4, ..., 9, they are as follows.

                    I:   Saturn,  Jupiter,  Mars,  Sun,  Venus,  Mercury,  Moon.

                    II:  Moon,  Mercury,  Venus,  Sun,  Mars,  Jupiter,  Saturn.

          System I is almost universal, only a 1446 Arabic MS and Cardano use system II. 

                    Not all writers use the same squares, but there are generally only two examples for each order.  Folkerts gives a table of these and which authors used which squares.  Basically, there are two sets of squares, one used by Picatrix-Tradition and Pacioli, the other by Agrippa and Cardano (in reverse).

                    Folkerts says this association was done by the Arabs, but Nadrûnî (cf below) is the earliest Arabic source I have.

                    In Codex Vat. Reg. lat. 1283 is a 13C(?) fragment with a  5 x 5  magic square associated to Mars.  Paolo dell'Abbaco's Trattato di Tutta l'Arte dell'Abacho, 1339, op. cit. in 7.E, has order 6 and 9 squares with their associations (sun and moon) given.  A 15C Frankfurt MS (UB, Ms. lat. oct. 231), has some examples and Paracelsus (1572) copied squares from different sources.

                    Folkerts then discusses various constructions due to al-Buni, Moschopoulos, al‑Haitham, etc.  The 15C Frankfurt MS is the first attempt at a theory, followed by Ries and Stifel. 

Μαvoυηλ Μoσχoπoυλoυ  (Manuel Moschopoulos -- variously spelled in Greek and variously transliterated).  c1315.  MS 2428, Bibliothèque Nationale, Paris.

Greek version with discussion in:  Siegmund Günther; Vermischte Untersuchungen zur Geschichte der mathematischen Wissenschaften; Teubner, Leipzig, 1876;  reprinted by Sändig, Wiesbaden, 1968.  Chap. IV: Historische Studien über die magischen Quadrate, pp. 188-276.  Section 5, pp. 195-203 is the Greek.

Greek and French in:  Paul Tannery; Le traité de Manual Moschopoulos sur les carrés magiques; Annuaire de l'Assoc. pour l'Encouragement des Études Grecques en France 20 (1886) 88‑118.  (= Mémoires (??*) Scientifiques, Paris, 1916‑1946, vol. 4, pp. 27?‑61, ??NYS.)  English translation by:  J. C. McCoy; Manuel Moschopolous's treatise on magic squares; SM 8 (1941) 15‑26. 

Gives diagonal rules for odd order and two rules for evenly even order.  These rules are sometimes attributed to Moschopoulos, but see the MSS at 1211? & 1212.  Various examples up through order  9.

Paolo dell'Abbaco.  Trattato di Tutta l'Arta dell'Abacho.  1339.  Op. cit. in 7.E.  B 2433, ff. 20v - 21r, gives order 6 and order 9 magic squares.  The latter may be associated with the moon, but my copy is not quite legible.  Dario Uri (email of 31 Oct 2001) said he had found this MS (B 2433), which has  6x6  and  8x8  magic squares, but the latter must be a misreading.

‘Abdelwahhâb ibn Ibrâhîm, ‘Izz eddîn el‑Haramî el (the  H  should have a dot under it) Zenġânî  = ‘Abd al‑Wahhâb ibn Ibrâhîm al‑Zinjânî.  Arabic MS,  Feyzullah Ef. 1362.  c1340.  ??NYS -- Described in Sesiano‑II.  Suter, p. 144, doesn't mention magic squares.  Construction of bordered squares of all orders.

Muhammad [the  h  should have an underdot) ibn Yūnis.  Compendium on construction of bordered magic squares in MS Hüsrev Pasa 257 in the Süleymaniye Library, Istanbul, ff. 32v-37v.  Translated and discussed in: Jacques Sesiano; An Arabic treatise on the construction of bordered magic squares; Historia Scientiarum 42 (1991) 13-31.  The actual MS was compiled in the 12th century of the Hegira, i.e. c18C), but the treatise is undated.  Sesiano compares the methods with other medieval Arabic material, e.g. al-Buni, al-Karagi, al-Buzjani, al‑Zanjani, so he seems to think it dates from a similar period.

Narayana Pandita (= Nārāyaņa Paņdita [NOTE:  ņ  denotes  n  with an overdot and the  d  should have an underdot.]).  Gaņita[NOTE:  ņ  denotes  n  with an underdot.]  Kaumudī (1356).  Edited by P. Dvivedi, Indian Press, Benares, 1942.  Part II: Introduction -- magic squares, pp. xv‑xvi (in English); Chap. 14: Bhadra gaņita [NOTE:  ņ  denotes an  n  with an underdot.], esp. pp. 384‑392 (in Sanskrit).  Shows orders  6, 10, 14.  Shows the  8  forms of order  3.  Obviously an extensive section -- is there an English translation of this material??.  (Editor refers to earlier sources:  Bhairava  and  Ŝiva Tāndava [the  d  should have a dot under it] Tantras,  ??NYS.  Cammann‑4 cites other MS sources.  Singh, op. cit. under Nâgârjuna, c1C, above, says this is the first mathematical treatment.  He says it  classifies into odd, evenly even and oddly even;  gives the superposition method of de la Hire;  gives knight's move method for  4n  and filling parallel to diagonal for odd, attributing both to previous authors.

                    Cammann‑4, pp. 274‑290 discusses this in more detail.  He gives another diagonal rule, sometimes beginning and ending at the middle of a side.  He then gives a quite different rule based on use of  x + y  with  x  =  0, n, 2n, ..., (n‑1)n,   y  =  1, 2, ..., n  with both sets of values cycling in the row, then reversing the  xs.  E.g., for  n = 5,  his first row of  y  values is:   4  5  1  2  3   and the second is:   5  1  2  3  4.   His first two rows of  x  values are:   15  20  0  5  10   and   20  0  5  10  15.   Reversing the  xs  and adding gives rows:   14  10  1  22  18   and   20  11  7  3  24.   This process gives a central lozenge (or diamond) pattern of the odds and has an extended knight's move pattern.  He extends this to doubly even squares.  He also gives the 'method of broken reversions' for singly‑even squares in three forms -- cf. C. Planck; The Theory of Reversions, IN:  W. S. Andrews, op. cit. in 4.B.1.a, pp. 295‑320.

                    Datta & Singh give a lengthy (51pp) description of Narayana's work, including about 19 other magic figures.  Many of Narayana's methods are novel.

Ahrens-1 gives references to further Arabic mentions of magic squares, usually as amulets, notably to  ibn Khaldun  (c1370).  He also gives many 14C and later examples of  3 x 3  and  4 x 4  squares, with rearrangement and/or constants added, used for magical purposes.

Nadrûnî.  Qabs al‑Anwâr.  pre-1384.  ??NYS -- described in Ahrens-1, but not mentioned in Ahrens-2.  Ahrens only knows of this from a modern article in Arabic.  This gives the association of planets by System I.  See Folkerts, above.

Arabic MS, 1446, ??NYS.  Discussed in Ahrens-1 and Ahrens-2, citing:  W. Ahlwardt; Verzeichniss der arab. Handschr. der Königl. Bibliothek zu Berlin; Berlin, 1891; Vol. III, pp. 505-506 (No. 4115).  This gives the System II association of planets with magic squares, later given by Cardan in 1539, with the unique addition of a  10 x 10  square for the zodiac coming after Saturn.

Dharmananda.  15C Jaina scholar.  Datta & Singh present his  8 x 8  square and say his method works for the evenly even case in general, extending Narayana. 

Sundarasūri.  c15C Jaina scholar.  Datta & Singh say he gives some novel methods, extending Narayana.

Jagiellonian MS 753.  15C Latin MS in Cracow.  Described in Cammann‑4, pp. 291‑297.  Earliest European set of magic squares of orders  3  through  9  associated with the planets in System I -- but see Folkerts above, c13C.  The order  4  square is Dürer's.  These squares later appear in Paracelsus.

Sûfî Kemal al‑Tustarî.  Ghayat al‑Murâd.  1448.  MS at Columbia.  ??NYS -- cited by Cammann‑4, p. 192.  On p. 196 Cammann says this represents a Persian Sufi tradition which was lost in sectarian warfare and the Mongol invasion.  On p. 201 he says this has a unit‑centred square of order  7.  On pp. 205‑206 are squares of orders  20, 29, 30.  He describes two bordering methods beyond  al‑Buni's.

Hindu square in a temple at Gwalior Fort, 1483.  Cited by Cammann‑4, p. 275, where the original source is cited -- ??NYS.

Pacioli.  De Viribus.  c1500.  Ff. 118r - 118v, 121r - 122v (some folios are wrongly inserted in the middle).  C.A. [i.e. Capitolo] LXXII. D(e). Numeri in quadrato disposti secondo astronomi ch' p(er) ogni verso fa'no tanto cioe per lati et per Diametro figure de pianeti et amolti giuochi acomodabili et pero gli metto (Of numbers arranged in a square by astronomers, which total the same in all ways, along sides and along diagonals, as symbols of the planets and suitable for many puzzles and how to put them ??).  Gives magic squares of orders  3  through  9  associated with planets in System I, usually attributed to Agrippa (1533), but see Folkerts, above.  Ff. 121v and 122r have spaces for diagrams, but they are lacking.  He gives the first two lines of the order 4 square as  16,  3,  2, 13;   5, 10, 11,  8;  so it must be the same square as shown by Dürer, below.

Albrecht Dürer.  Melencolia.  1514.  Two impressions are in the British Museum.  4 x 4  square with  15, 14  in the bottom centre cells.  Surprisingly, this is the same as the  4 x 4  appearing in  Ikhwān al‑Şafā’ [NOTE:  Ş  denotes  S  with an underdot.]  (c983), with the two central columns interchanged and the whole square reflected around a horizontal midline, i.e. 

          16,  3,  2, 13;   5, 10, 11,  8;   9,  6,  7, 12;   4, 15, 14,  1.  This is the same as that described by Pacioli.  There is some belief that the association with Jupiter relates to the theme of the picture.  This is the first printed  4 x 4  magic square.

Riese.  Rechnung.  1522.  1544 ed. -- pp. 106‑107;  1574 ed. -- pp. 71v‑72v.  Gives  3 x 3  square in  672  form and how to construct other  3 x 3  forms.  Also gives a  4 x 4  square, like Dürer's but with inner columns interchanged.

Riese.  Rechenung nach der lenge ....  Op. cit. under Riese, Die Coss.  1525.  ??NYR.  Cammann‑4, p. 294, says pp. 103r‑105v gives a diagonal rule for odd orders.  A quick look shows the material starts on p. 102v.

Cornelius Agrippa von Nettesheim.  De Occulta Philosophia.  Cologne, 1531, ??NYS.  Included in his Opera, vol. 1, and available in many translations.  2nd Book, Chap. 22.  Gives association of planets with magic squares in System I -- as previously done by Pacioli, c1500, but with different squares.  See Folkerts, above, and Cammann‑4, p. 293-294.  The squares do not appear in the 1510 draft version of this book.  Bill Kalush has kindly sent Chap. 22 from a 1913 English version, but it doesn't have any squares -- perhaps it was from the wrong Book??.  He gives each square twice, with Arabic and Hebrew numerals.  His  3 x 3  is the  492  version.  His  4 x 4  is the same as that of  Ikhwān al‑Şafā’ [NOTE:  Ş  denotes  S  with an underdot.],  c983.

Cardan.  Practica Arithmetice.  1539.

Chap. 42, section 39, ff. H.v.r ‑ H.vi.r (p. 55).  Gives association of planets with magic squares in System II.  He is almost unique in using this System, though his squares are the same as Agrippa's.  See comments under al‑Buni and Folkerts.

Chap. 66, section 72, ff. FF.v.r - FF.v.v (p. 157).  Shows how to construct a  5 x 5  magic square from the natural  5 x 5  array.

Michael Stifel.  Arithmetica Integra.  Nuremberg, 1544.  ??NYS -- discussed in Cammann‑4, p. 194.  Pp. 25‑26a shows some some bordered squares.  Consequently he is sometimes credited with inventing the concept, but see  Ikhwān al‑Şafā’ [NOTE:  Ş  denotes  S  with an underdot.]  (c983),  al‑Buni  (c1200),  Yang Hui  (1275),  ‘Abdelwahhâb  (c1340),  Sûfî Kemal al‑Tustarî  (1448)  and  ibn Yūnis,  above.

M. Mersenne.  Novarum observationum physico‑mathematicarum.  Paris, 1647.  Vol. 3, chap. 24, p. 211.  ??NYS.  States Frenicle's result. (MUS II #29.)

Isomura Kittoku.  Ketsugi-shō.  1660, revised in 1684.  ??NYS -- described in Smith & Mikami, pp. 65-77.  He gives magic squares of orders up to order 10.  The order 9 square contains the order 3 square, in the  618  form, in the top middle section.  He gives magic circles with  n  rings of  2n  about a central value of  1,  for n = 2 - 6.  The values are symmetrically arranged, so corresponding pairs add to  2n2 + 3  and each ring adds up to  n (2n2 + 3),  while each diameter adds to one more than this.  In the 1684 edition, he gives some magic wheels, but these are simply a way of depicting magic squares, though it is not clear where the diagonals are.

Muramatsu Kudayū Mosei.  Mantoku Jinkō‑ri.  1665.  ??NYS -- described in Smith & Mikami, pp. 79-80.  Gives a magic square of order 19.  Gives a magic circle of Isomura's type for  n = 8.  Smith & Mikami, p. 79, gives Muramatsu's diagram with a transcription on p. 80.  The central  1  is omitted and the corresponding pairs no longer add to  131,  but the pairs adding to  131  lie on the same radius.

Bernard Frénicle de Bessy.  Des Quarrez ou Tables Magiques,  including:  Table generale des quarrez de quatre.  Mem. de l'Acad. Roy. des Sc. 5 (1666‑1699) (1729) 209‑354.  (Frénicle died in 1675.  Ollerenshaw & Bondi cite a 1731 edition from The Hague??)  (= Divers Ouvrages de Mathématique et de Physique par Messieurs de l'Académie des Sciences; ed. P. de la Hire; Paris, 1693, pp. 423‑507, ??NYS.  (Rara, 632).  = Recueil de divers Ouvrages de Mathematique de Mr. Frenicle; Arkstèe & Merkus, Amsterdam & Leipzig, 1756, pp. 207-374, ??NX.)

                    Shows there are  880  magic squares of order  4  and lists them all.  Cammann‑4, p. 202, asserts that they can all be derived from one square!!

                    The list of squares has been reprinted in the following. 

M. Gerardin.  Sphinx‑Oedipe -- supplement 4 (Sep‑Oct 1909) 129‑154.  ??NX

K. H. de Haas.  Frenicle's  880  basic Magic Squares of  4 x 4  cells, normalized, indexed, and inventoried (and recounted as  1232).  D. van Sijn & Zonen, Rotterdam, 1935, 23pp.

Seki Kōwa.  Hōjin Yensan.  MS revised in 1683.  Known also as his  Seven Books.  ??NYS -- described in Smith & Mikami, pp. 116-122.  Describes how to border squares of all sizes.  Gives an easy method for writing down a magic circle of Isomura's type.

Thomas Hyde.  Mandragorias seu Historia Shahiludii, ....  (= Vol. 1 of De Ludis Orientalibus, see  4.B.5 for vol. 2.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  Prolegomena curiosa.  The initial material and the Prolegomena are unpaged but the folios of the Prolegomena are marked (a), (a 1), ....  The material is on  (d 4).v - (e 1).v, which are pages 32-34 if one starts counting from the beginning of the Prolegomena. 

                    Seems to believe magic squares come from Egypt and gives association of orders of squares with planets as in Pacioli and Agrippa, but he only gives one example of a magic square -- an  8 x 8  which is associated.

W. Leybourn.  Pleasure with Profit.  1694.  Prob. 10, pp. 4-5.  Gives the  294  form and then says that each line can be rearranged four ways, e.g.  294, 492, 924, 942.  He writes these out for all eight lines, but I can't see any pattern in the way he chooses his four of the six possible permutations.

Ozanam.  1694.  1696: Prob. 9: Des quarrez magiques, 36-41.  Prob. 9: Of magical squares, 1708: 33-36.  Prob. 12: Des quarrez magiques, 1725: 82-102.  Chap. 12: Des quarrés magiques, 1778: 217-244.  Chap. 12: Of magic squares:  1803: 211-240;  1814: 183-207 & 366-367;  1840: 94-105.  Extended discussion, but contains little new -- except some comments on Franklin's squares -- see Ozanam-Hutton (1803).  Associates squares with planets, as done by Pacioli.

Wells.  1698.  No. 119, pp. 209-210.  Studies the  3 x 3  square carefully, showing that the centre cell must be 5 and the sum of each pair of adjacent side cells is double the value in the opposite corner -- e.g.  9 + 7  is twice  8.  I don't recall ever seeing this result before.

Philippe de la Hire.  Sur les quarrés magiques.  Mémoires de l'Académie Royale des Sciences (1705 (1706)) 377‑378.  Gives a method for singly‑even squares, but it uses so many transpositions that it is hard to see if it works in general.  ??NYS -- described in Cammann‑4, p. 286.

Muhammed ibn Muhammmed.  A Treatise on the Magical Use of the Letters of the Alphabet.  Arabic MS of 1732,  described and partly reproduced in:  Claudia Zaslavsky; Africa Counts; Prindle, Weber & Schmidt, Boston, 1973; chap. 12, pp. 137‑151.  Several of his magic squares are deliberately defective, presumably because of the Islamic belief that only God can create something perfect.  I do not recall any other mention of this feature.

Minguet.  1733.  Pp. 169-172 (1755: 122-123; 1864: 158-160; not noticed in 1822, but probably about p. 180.)  Magic squares of order three with various sums, made by laying out cards.

Benjamin Franklin.  1736-1737.  Discovery of some large magic squares and circles.  He described these in letters to Peter Collinson whose originals do not survive.  I. Bernard Cohen [Benjamin Franklin  Scientist and Statesman; DSB Editions, Scribner's, 1975, pp. 18-19] dates them as above, citing Franklin's Autobiography, but my copy is an abridged edition without this material -- ??.  He also reproduces them.  They were first published in the following.

                    James Ferguson.  Tables and Tracts, Relative to Several Arts and Sciences.  A. Millar & T. Cadell, London, 1767.  Pp. 309-317.  ??NYS.  Ferguson may be indicating that he is the first person to whom Franklin showed them. 

                    B. Franklin.  Experiments and Observations on Electricity.  4th ed., London, 1769.  Two letters to Peter Collinson, pp. 350-355(??).  ??NYS, but reprinted in:  Albert Henry Smyth; The Writings of Benjamin Franklin; Vol. II, Macmillan, 1907, pp. 456‑461 and Plates VII (opp. p. 458) and VIII (opp. p. 460).

                    The squares are of order  8  and  16,  but are only semi-magic (see Ozanam-Hutton (1803) and Patel (1991)), and the circle has  8  rings and  8  radii.  Franklin said he could make these squares as fast as he could write down the numbers!

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.  Problem 4.  Asks for a  3 x 3  magic square.

Caietanus Gilardonus.  9 x 9  square on a marble plaque on the Villa Albani, near Rome, dated 1766.  The square and the accompanying inscription are given in:  E. V. R.; Arranged squares; Knowledge 1 (27 Jan 1882) 273, item 231.  These are also given by  Catalan; Mathesis 1, p. 151 (??NYS)  and  Lucas; L'Arithmétique Amusante; 1895; pp. 224-225.  Lucas says they were discovered in 1881, and that the villa is now owned by Prince Torlonia and is outside the Porta Salaria.

Catel.  Kunst-Cabinet.  1790.

Das grosse Zauberquadrat, p. 16 & fig. 33 on plate II.  49  numbered pieces to make into a magic square.

Das kleine Zauberquadrat, p. 16 & fig. 34 on plate II.  3 x 3.  Diagram is only semimagic.

Bestelmeier.  1801.

Item 441: Das arithmetische Zauber‑Quadrat.  9  numbered pieces to into a magic square.  Diagram is shown disarranged.

Item 961.a: Das Sonderbarste aller magischen Zahlen=Quadrate.  64  numbers to form into a magic square, with various groups of four to add up to half the magic constant.

Item 961.b: Gewöhnliche Zauberquadraten von  64, 48, 36, 25, 16, 9  Zahlen.

Ozanam-Hutton.  1803.  Chap. 12: Remarks, 1803: 237-240 & fig. 1, plate 4.  1814: 203-207 & corrections: 366-367 & fig. 1, plate 4 & additional plate 5.  1840: 104-105, with no figure.  The 1814 corrections note that Franklin's square is only semi-magic and gives another large example -- 1840 omits this.

Rational Recreations.  1824. 

Exer. 7, p. 51.  Order three square in the  834  form.  Calls  15  the 'product' of the entries  8, 3, 4.  Cf Boy's Own Book.

Exer. 9, pp. 53-54.  Natural and magic squares of order 5.

Exer. 24., p. 132.  Magic square of order 10.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 238-243, no. 907-926.  This gives lots of straightforward exercises -- e.g. find a  5 x 5  magic square with sum  96,  which he does by adding  6 1/5  to each entry of a normal example.

Young Man's Book.  1839.  P. 232.  The Magical Square.  "The Chinese have discovered mystical letters on the back of the tortoise, which is the common magical square, making each way 15, viz."  Gives the  294  form, while all Chinese forms have  492.  Pp. 236-238 is a straightforward section on Magic Squares.

Boy's Own Book.  1843 (Paris).

P. 342.  "The digital numbers arranged so as to give the same product, whether counted horizontally, diagonally, or perpendicularly."  Order 3 magic additive square, despite the title,  834  form.  = Boy's Treasury, 1844, p. 300.  = de Savigny, 1846, p. 290.  Cf Rational Recreations, 1824.

Pp. 342-343.  "Magic squares."  Constructs order 5 magic square from the order 5 natural square.  = Boy's Treasury, 1844, pp. 300-301.  = de Savigny, 1846, pp. 291-292.

P. 347.  "The figures, up to 100, arranged so as to make 505 in each column, when counted in ten columns perpendicularly, and the same when counted in ten files horizontally."  This is actually an associated magic square of order 10.  = Boy's Treasury, 1844, p. 305.  = de Savigny, 1846, p. 293.

Indoor & Outdoor.  c1859.  Part II, prob. 13: Franklin's magic square, pp. 132-133.  Gives Franklin's  16 x 16  square and states some of its properties, very similar to Ozanam‑Hutton.

Vinot.  1860.  Art. CLIV: Des carrés magiques, pp. 188-201.  On p. 190, he gives an association of squares with planets, as given by Pacioli, but with the addition of:  1 -- God;  2 -- matter.

Magician's Own Book (UK version).  1871.  The magic square oraculum, pp. 94-98.  This shows a square of order 11 and says it is "a magic square of eleven, with one in the centre", but it is not at all magic.  Initially it appears to be bordered, but it is a kind of arithmetical square.  1  is in the middle.  Then  2 - 9  are wrapped around the central square, going clockwise with the  2  above the  1.  Then  10 - 25  are wrapped around the central  3 x 3  area, with the  10  above the  2,  etc.  The 'oracle' consists of thinking of a number and consulting a list of fortunes, so the 'magic square' is never used!

Carroll-Wakeling.  c1890?  Prob. 8: Magic postal square, pp. 10-11 & 65.  The first nine values of postage stamps in Carroll's time had values  1, 2, 3, 4, 5, 6, 7, 8, 10  in units of half-pence.  But the total of the values in a magic square is three times the magic constant, and these values add up to 46.  So Carroll allows one of the values to be repeated and the ten values now have to be placed to make a  3 x 3  magic square.  Surprisingly, the value to be repeated is uniquely determined and there is just one such square.

T. Squire Barrett.  The magic square of four.  Knowledge 14 (Mar 1891) 45-47  &  Letter (Apr 1891) 71  &  Letter (Aug 1891) 156.  Says he hasn't seen Frenicle's list.  Classifies the  4 x 4  squares into  12  types, and obtains  880  squares, but doesn't guarantee to have found all of them.  The first letter notes that he erred in counting one type, getting  48  too many, but a friend has found  16  more.  Second letter notes that the missing squares have been found by another correspondent who has compared them with Frenicle's list and then Barrett corrects some mistakes.  The other correspondent notes that Frenicle had proceeded by a trial and error method and probably had found all examples.

Ball.  MRE, 1st ed., 1892, pp. 108-121: Chap. V: Magic squares.  Later editions amplify this material, but the material is too detailed and too repetitive to appeal to me at the moment.

Hoffmann.  1893.  Chap. IV, pp. 146 & 183‑191 = Hoffmann-Hordern, pp. 114-118, with photos on pp. 133 & 143.

No. 7: A simple magic square.  3 x 3 case.  Photo on p. 133 shows a wood circular board with a  3 x 3  array of holes and nine numbered pegs, by Jaques & Son, registered 1858.  The Hordern Collection of Hoffmann Puzzles, p. 69, shows the same puzzle, dated 1860-1890.

No. 8: The "thirty‑four" puzzle.  4 x 4  case.  Asserts that Heywood of Manchester publish a booklet, 'The Curiosities of the Thirty‑four Puzzle' which has instructions for obtaining all the solutions (i.e. of the  4 x 4  magic square), ??NYS.  Photo on p. 143 shows several ordinary fifteen puzzles and two definite thirty-four puzzles and one possible.  At the upper right is a box with "The Great American Puzzle  9  15 & 34  3 Games in One." -- I don't know what the game involving  9  can be -- ??  Below this is a solution sheet headed  "Novel and Exasperating Yankee Puzzles,  15 and 34."  -- The Hordern Collection of Hoffmann Puzzles, 70, shows this with the box which is to the lower left in this photo, which reads  "Perry & Co's Calculator Puzzles.  Two Games in One.  The third, possible, example is at the lower right and the box just has  "Number Puzzle", by McLoughlin Bros N.Y.  These are all dated 1879-1885.  Hordern Collection, p. 70, shows the instructions and the Perry example, dated 1880-1900.

No. 9: The "sixty-five" puzzle.  5 x 5  case.

Stewart Culin.  Chinese Games with Dice and Dominoes.  From the Report of the U. S. National Museum for 1893, pp. 489‑537.  On pp. 536-537, he discusses the Lok Shü, citing Legge.  He reports that the  618  version is popular as a charm with both Hindus and Moslems in India, while the Chinese  492  form is used in Tibet. 

Dudeney.  A batch of puzzles.  Royal Magazine 1:3 (Jan 1899)  &  1:4 (Feb 1899) 368-372.  The eight clowns puzzle.  = CP; 1907; prob. 81, pp. 128 & 126.  3 x 3  array with pieces  x 2 3;  4 5 6;  7 8 9  to be rearranged into a magic square, the blank being counted as a  0.  Answer is that this is impossible, but the clown marked  9  is juggling balls which make his number  .9,  i.e.  .9  recurring,  which is  1!

Dudeney.  The magic square of sixteen.  The Queen (15 Jan 1910) 125-126.  Good derivation of the  880  squares of order  4,  classified into  12  types.  A condensed version with some extra information is in AM, pp. 119-121.

Loyd.  Cyclopedia.  1914.  The 14‑15 puzzle in puzzleland, pp. 235 & 371.  = MPSL1, prob. 21, pp. 19‑20 & 128.  c= SLAHP: The "14‑15" magic square, pp. 17‑18 & 89.  Given the  15  Puzzle with the  14  and  15  interchanged, move to a magic square.  The blank counts zero, so the magic constant is  30.

Collins.  Book of Puzzles.  1927.  Magic squares and other figures, pp. 79-94.  Brief survey.  Gives a number of variant forms.

D. N. Lehmer.  A complete census of  4 x 4  magic squares.  Bull. Amer. Math. Soc. 39 (1933) 764-767.  Here he is dealing with semi-magic squares and then any permutation of the rows or columns or transposition of the whole array preserves the row and column sums.  Hence there are  2(n!)2  arrays in each equivalence class and he describes a normalized form for each class.  For the  3 x 3  case, there are  72  semimagic squares and one normalized form.  For the  4 x 4  case, there are  468  normalized forms and hence  468 x 2 x 242  =  539,136   semimagic squares.

D. N. Lehmer.  A census of squares of order  4,  magic in rows, columns, and diagonals.  Bull. Amer. Math. Soc. 39 (1933) 981-982.  Here he discusses Frenicle's enumeration of  880  4 x 4  magic squares and points out that there are additional equivalences beyond the symmetries of the square so that Frenicle only needed to find and list  220  squares.  Using ideas like those in his previous article, he finds  220  such squares, confirming Frenicle's result once again.

J. Travers.  Rules for bordered magic squares.  MG 23 (No. 256) (Oct 1939) 349‑351.  Cites Rouse Ball, MRE (no details), as saying no such rules are known.  He believes these are the first published rules.

Anonymous.  A Book of Fun with Games and Puzzles.  One of a set of three 12pp booklets, no details, [1940s?].  P. 7: Here is the magic square.  Gives a  4 x 4  square using the numbers  3, ..., 17  and asks for it to be dissected along the lines into four pieces which can be rearranged into a magic square.

D. H. Hallowes.  On  4 x 4  pan‑magic squares.  MG 30 (No. 290) (Jul 1946) 153‑154.  Shows there are only  3  essentially different  4 x 4  pan‑magic squares.

Ripley's Puzzles and Games.  1966.  P. 52.  The magic cube.  This actually shows only the front of a cube and really comprises three magic squares as the faces have no relation to each other.  What is really being attempted is to arrange the numbers  1, 2, ..., 27  into three  3 x 3  magic squares.  Then the magic sum must be  42.  The given arrangement has  22  of the  24  lines adding to  42  -- two of the diagonals fail.  ?? -- is such an arrangement possible?  Ripley's says the pattern adds to  42  in  44  directions -- apparently they count each of the  22  lines in each direction.

Gardner.  SA (Jan 1976)  c= Time Travel, chap. 17.  Richard Schroeppel, of Information International, used a PDP‑10 to find  275,305,224  magic squares of order  5,  inequivalent under the  8  symmetries of the square.  If one also considers the 'eversion' symmetries, there are  32  symmetries and  68,826,306  inequivalent squares.  (Gardner says there is an Oct 1975 report on this work by Michael Beeler, ??NYS, and gives Schroeppel's address:  835 Ashland Ave., Santa Monica, Calif., 90405 -- I believe I wrote, but had no reply??)

K. Ollerenshaw & H. Bondi.  Magic squares of order four.  Phil. Trans. Roy. Soc. Lond. A306 (1982) 443‑532.  (Also available separately.)  Gives a new approach to Frénicle's results.  Relates to Magic Card Squares and the Fifteen Puzzle.

Lee C. F. Sallows.  Alphamagic squares:  I  &  II.  Abacus 4:1 (Fall 1986) 28‑45  &  4:2 (Winter 1987) 20-29 & 43.  Reprinted in: The Lighter Side of Mathematics; ed. by R. K. Guy & R. E. Woodrow; MAA, 1994, pp. 305-339.  Introduces notion of alphamagic square -- a magic square such that the numbers of letters in the words for the numbers also form a magic square.  Simplest example is:   5, 22, 18;   28, 15,  2;   12,  8, 25.  He asserts that this appears in runes in an 1888 book describing a 5C charm revealed to King Mi of North Britain.  (This seems a bit far‑fetched to me or mi?)  Asks if there can be an alphamagic square using the first  n2  numbers and shows that  n ³ 14.  Notes some interesting results on formulae for  4 x 4  squares, including one with minimum number of symbols.  There was a letter and response in Abacus 4:3 (Spring 1987) 67-69.

Martin Gardner.  Prime magic squares.  IN: The Mathematical Sciences Calendar for 1988; ed. by Nicholas J. Rose, Rome Press, Raleigh, North Carolina, 1987.  Reprinted with postscript in Workout, chap. 25.  Says Akio Suzuki found a  35 x 35  magic square using the first odd primes in 1957.  I have a poster of this which Gardner gave me.  Offers $100 for the first  3 x 3  magic square using consecutive primes.  The postscript says Harry L. Nelson won, using a Cray at Lawrence Livermore Laboratories.  He found 22 examples.  The one with the lowest constant has smallest value  1,480,028,129  and the values all have the same first seven digits and their last three digits are:  129,  141,  153,  159,  171,  183,  189,  201,  213. 

Lalbhai D. Patel.  The secret of Franklin's  8 x 8  'magic' square.  JRM 23:3 (1991) 175-182.  Develops a method to make Franklin's squares as fast as one can write down the numbers!

Jacques Bouteloup.  Carrés Magiques, Carrés Latins et Eulériens.  Éditions du Choix, Bréançon, 1991.  Nice systematic survey of this field, analysing many classic methods.

Lee Sallows.  Alphamagic squares.  CFF 35 (Dec 1994) 6-10.  "... a brief synopsis of [his above article] which handles the topic in very much greater detail."

Martin Gardner.  The magic of  3 x 3.  Quantum 6:3 (Jan-Feb 1996) 24-26; with addendum in (Mar-Apr 1996).  Reprinted with a postscript in Workout, chap. 22.  Says that modern scholars doubt if the pattern in China is older than 10C!  Mentions his 1987 prize and Nelson's least solution.  Says Martin LaBar [CMJ (Jan 1984) 69] asked for a  3 x 3  magic square whose entries are all positive squares.  Gardner reiterates this and offers $100 for the first example -- in the postscript, he extends the prize to include a proof of impossibility.  He gives examples of  3 x 3  squares with various properties and the lowest magic sum, e.g. using primes in arithmetic progression.

Lee Sallows.  The lost theorem.  Math. Intell. 19:4 (1997) 51-54.  Gives an almost solution to Gardener's problem, but one diagonal fails to add up correctly.  Gives an example of Michael Schweitzer which is magic but contains only six squares.  Using Lucas' pattern, where the central number,  c,  is one third of the magic sum and two adjacent corners are  c + a  and  c + b,  he observes that these can be vectors in the plane or complex numbers, which allows one to correspond classes of eight magic squares with parallelograms in the plane.  This leads to perhaps the most elegant magic square by taking  c = 0,  a = 1,  b = i.

Kathleen Ollerenshaw, Kathleen  &  David S. Brée.  Most-perfect Pandiagonal Magic Squares  Their construction and enumeration.  Institute of Mathematics and its Applications, Southend-on-Sea, 1998.  A most-perfect square is one which is magic and pandiagoanal and all  2 x 2  subsquares have the same sum, even when the square is considered on a torus.  They find a formula for the number of these in general.  Summary available on  www.magic-squares.com  or  www.most-perfect.com .

Lee Sallows.  Email of 11 Jun 1998.  He asks if any set of 16 distinct numbers can produce more than  880  magic squares of order  4.  He finds that  -8, -7, ..., -2, -1, 1, 2, ..., 7, 8  gives  1040  magic squares.  He has not found any other examples, nor indeed any new examples with as many as  880,  though he has looked at other types of values, even Gaussian integers.

K. Pinn  &  C. Wieczerkowski.  Number of magic squares from parallel tempering Monte Carlo.  Intern. J. Modern Physics C 9:4 (1998) 541-546.  ??NYS - cited by Chan & Loly, below.  They estimate there are  1.77 x 1019  magic squares of order 6.

Harvey D. Heinz  &  John R. Hendricks.  Magic Square Lexicon: Illustrated.  Harvey D. Heinz, 15450 92A Avenue, Surrey, British Columbia, 2000.

Frank J. Swetz.  Legacy of the Luoshu - The 4000 Year Search for the Meaning of the Magic Square of Order Three.  Open Court, 2002.  ??NYS - cited by Chan & Loly, below.

Wayne Chan  &  Peter Loly.  Iterative compounding of square matrices to generate large-order magic squares.  Mathematics Today 38:4 (Aug 2002) 113-118.  Primarily they develop programs for doing compounding to produce very large squares.

Lee Sallows.  Christmas card for 2003: Geometric magic square.  Consider the magic square: 

          11  9  9  7;  6 10  8 12;  6  8 10 12;  13  9  9  5  with magic constant  36.  Each quadrant also adds up to  36.  Sallows uses sixteen polyominoes having these numbers of unit squares and arranged in this pattern so that each quadrant forms a  6 x 6  square.  These polyominoes can be assembled into many other squares.

 

          7.N.1 MAGIC CUBES

 

            Note.  Historically, a  k3  has been called magic when all the  3k2  lines parallel to the axes and the  4  space diagonals have the same sum.  But there are also  6k  2‑dimensional diagonals -- if these also have the same sum, we will say that the cube is perfectly magic.  Pandiagonal (= pan‑n‑agonal) refers just to the space diagonals.  Perfectly pandiagonal refers to all the diagonals.  In higher space, the simpler words refer to the  2n-1  'space' diagonals and perfect will include all the diagonals in intermediate dimensions. 

            A  k-agonal is a line which varies in  k  coordinates, so a  1-agonal is a row or column, etc.,  the  2-agonals of a cube include the diagonals of the faces,  while the  3-agonals of a  3‑cube are the space diagonals.

            Associated  or  complementary  means that two cells symmetric with respect to the centre add to  kn + 1.

 

Pierre de Fermat.  Letter to Mersenne (1 Apr 1640).  Oeuvres de Fermat.  Ed. by P. Tannery & C. Henry.  Vol. 2, Gauthier‑Villars, Paris, 1894, pp. 186‑194.  Gives a magic  43.

                    A shorter, undated, version, occurs in Varia Opera Mathematica D. Petri de Fermat, Toulouse, 1679;  reprinted by Culture et Civilization, Brussels, 1969; pp. 173‑176.  The version in the Oeuvres has had its orthography modernized.

               On p. 174 of the Varia (= p. 190 of the Oeuvres), he says:  "j'ay trouvé une regle generale pour ranger tous les coubes à l'infiny, en telle façon que toutes les lignes de leurs quarrez tant diagonales, de largeur, de longeur, que de hauteur, fassent un méme nombre, & determiner outre cela en combien de façons differentes chaque cube doit étre rangé, ce qui est, ce me semble, une des plus belles choses de l'Arihmetique [sic]..."  He describes a assembly of four squares making a magic cube.  [The squares are missing in the Varia.]  He says that the magic sum occurs on  72  lines, but it fails to have the magic sum on  8  of the  2‑agonals and all  4  of the  3‑agonals.

Lucas.  Letter.  Mathesis 2 (1882) 243‑245.  First publication of the magic  43  described by Fermat above.  Says it will appear in the Oeuvres.

E. Fourrey.  Op. cit. in 4.A.1.  1899.  Section 317, p. 257.  Notes that Fermat's magic cube has only  64  magic lines.

Lucas.  L'Arithmétique Amusante.  1895.  Note IV:  Section IV: Cube magique de Fermat, pp. 225‑229.  Reproduces the  43  from his Mathesis letter and gives a generalization by V. Coccoz, for which the same diagonals fail to have the magic sum, though he implies they do have the sum on p. 229.

 

Pierre de Fermat.  Letter to Mersenne, nd [Jun? 1641].  Opp. cit. above:  Oeuvres, vol. 2, pp. 195‑199;  Varia Opera, pp. 176‑178.  On p. 177 of the Varia (= p. 197 of the Oeuvres), he says:  "Pource qui est des cubes, je n'en sçay pas plus que Monsieur Frenicle, mais pourtant je puis les ranger tous à la charge que les Diagonales seules de quarrez que nous pouvons supposer paralleles à l'Horizon, seront égales aux côtez des quarrez, ce qui n'est pas peu de chose.  En attendant qu'une plus longue meditation découvre le reste, je dresseray celuy de  8.  10.  ou  12.  à ces conditions si Monsieur de Frenicle me l'ordonne."

Joseph Sauveur.  Construction générale des quarrés magiques.  Mémoires de l'Académie Royale des Sciences (1710 (1711)) 92‑138.  ??NYS -- mentioned by Brooke (below), who says Sauvier [sic] presented the first magic cube but gives no reference.  Discussed by Cammann‑4, p. 297, who says Sauveur invented magic cubes and Latin squares.  This paper contains at least the latter and an improvement on de la Hire's method for magic squares, but Cammann doesn't indicate if this contains the magic cube.

Charles Babbage.  Notebooks -- unpublished collection of MSS in the BM as Add. MS 37205.  ??NX.  See 4.B.1 for more details.  F. 308: Essay towards forming a Magick Cube, c1840??  Very brief notes.

Gustavus Frankenstein.  [No title].  Commercial (a daily paper in Cincinnati, Ohio) (11 Mar 1875).  ??NYS.  Perfect  83.  Described by Barnard, pp. 244‑248.

Hermann Scheffler.  Die magischen Figuren.  Teubner, Leipzig, 1882;  reprinted by Sändig, Wiesbaden, 1981.  Part III: Die magische Würfel, pp. 88‑101 & plates I & II, pp. 113 & 115.  He wants all  2‑  and  3‑agonals to add up to the magic constant, though he doesn't manage to construct any examples.  He gives a magic  53  which has the magic sum on  14  of the  30  2‑agonals and many of the broken  2‑agonals.  He also gives a  43  and a  53,  but I haven't checked how successful they are.

F. A. P. Barnard.  Theory of magic squares and of magic cubes.  Memoirs of the National Academy of Science 4 (1888) 209‑270.  ??NYS. Excerpted, including the long footnote description of Frankenstein's  83,  in Benson & Jacoby (below), pp. 32‑37, with diagrams of the result on pp. 37‑42.

C. Planck, on pp. 298 & 304 of Andrews, op. cit. in 4.B.1.a, says the first magic  63  was found by W. Firth of Emmanuel College, Cambridge in 1889.

Pao Chi-shou.  Pi Nai Sahn Fang Chi  (Pi Nai Mountain Hut Records).  Late 19C.  ??NYS -- described by Lam (in op. cit. in 7.N under  Yang Hui),  pp. 321-322, who says it has magic cubes, spheres and tetrahedrons.  See also Needham, p. 60.

V. Schlegel.  ??  Bull. Soc. Math. France (1892) 97.  ??NYS.  First magic  34.  Described by Brooke (below).

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Magic cubes, pp. 99-100.  33  magic cube with also the  6  2-dimensional diagonals through the centre having the same sum, so there are  37  lines with the magic sum.  43  magic cube -- this has the  3 x 16 + 4 = 52  expected magic lines, but he asserts it has  68  magic lines, though I can only find  52.  The perfect case would have  76  magic lines.

C. Planck.  Theory of Path Nasiks.  Privately printed, Rugby, 1905.  ??NYS.  (Planck cites this on p. 363 of Andrews, op. cit. in 4.B.1.a, and says there are copies at BM, Bodleian and Cambridge.)  The smallest Nasik (= perfectly pandiagonal)  kn  has  k = 2n.  If the cube is also associated, then  k = 2n + 1.  He quotes these results on p. 366 of Andrews and cites earlier erroneous results.  On p. 370 of Andrews, he says that a perfect  k4  has  ³ 8.

Collins.  Book of Puzzles.  1927.  A magic cube, pp. 89-90.  33,  different than that in Berkeley & Rowland, but with the same properties.

J. Barkley Rosser & Robert J. Walker.  MS deposited at Cornell Univ., late 1930s.  ??NYS.  (Cited by Gardner, loc. cit. below, and Ball, MRE, 11th ed., p. 220;  12th ed., p. 219.)  Finds a Nasik  83  and shows that Nasik  k3  exist precisely for the multiples of  8  and for odd  k > 8.

G. L. Watson.  Note 2100:  To construct a symmetrical, pandiagonal magic cube of oddly even order  2n ³ 10.  MG 33 (No. 306) (Dec 1949) 299‑300.

Maxey Brooke.  How to make a magic tessarack.  RMM 5 (Oct 1961) 40‑44.  Cites Sauvier and Schlegel.  Believes this is the first English exposition of Schlegel.  The  33  he develops is magic, but only the  6  2‑agonals through the centre have the magic sum.  The resulting  34  is magic but not perfect.

Harry Langman.  Play Mathematics.  Hafner, 1962.  ??NYS -- cited by Gardner below.  Pp. 75‑76 gives the earliest known perfect  73.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  The magic cube, pp. 157 & 199.  33  which is associated, but just the  6  2-agonals through the centre have the magic sum; the other  12  2-agonals do not.  The space diagonals also have the magic sum.

John Robert Hendricks.  The third‑order magic cube complete.  JRM 5:1 (1972) 43‑50.  Shows there are  4  magic  33,  inequivalent under the  48  symmetries of the cube.  None of these is perfect.  (The author has published many articles on magic cubes in JRM but few seem appropriate to note here.)

Gardner.  1976.  Op. cit. in 7.N.  Gives Richard Lewis Myers Jr.'s proof that a perfect  33  does not exist, and Richard Schroeppel's 1972 proof that a perfect  43  does not exist.  (Gardner says Schroeppel published a memorandum on this, ??NYS.)  Says that perfect cubes of edge  5, 6, 7  are unknown and gives a perfect, associated  83  found by Myers in 1970.  The Addendum in Time Travel cites Planck and Rosser & Walker for earlier  83  and says that many readers found a perfect  73  and refers to Langman.  Also  93,  113  and higher orders were found.

Johannes Lehmann.  Kurzweil durch Mathe.  Urania Verlag, Leipzig, 1980.  No. 8, pp. 61 & 160.  Arrange  0 - 15  on the vertices of a  24  hypercube so that each  2-dimensional face has the sum  30.

John R. Hendricks.  The perfect magic cube of order 4.  JRM 13 (1980‑81) 204‑206.  Shows it does not exist.

William H. Benson & Oswald Jacoby.  Magic Cubes -- New Recreations.  Dover, 1981.  Summarises all past results on p. 5.  There are perfect  n3  for  n ¹ 6 (mod 12),  n ³ 7,  except  n = 10.  It is not clear if they have proofs for  n º 4 (mod 8)  or  º 2 (mod 4).  They are unable to show the non‑existence for  º 5  (cf. p. 29).  There are pandiagonal  n3  for  ¹ 6 (mod 12),  n ³ 4,  though it is not clear if they have a proof for  º 2 (mod 4)  (cf. p. 102).

Rudolf Ondrejka.  Letter:  The most perfect (8 x 8 x 8) magic cube?  JRM 20:3 (1988) 207‑209.  Says Benson & Jacoby sketch a perfectly pandiagonal  83.  He gives it in detail and discusses it.

Joseph Arkin, David C. Arney & Bruce J. Porter.  A perfect  4‑dimensional hypercube of order 7 -- "The Cameron Cube".  JRM 21:2 (1989) 81‑88.  This is the smallest known order in four dimensions.

Allan William Johnson Jr.  Letter:  Normal magic cubes of order  4M+2.  JRM 21:2 (1989) 101‑103.  Refers to Firth, 1889, who is mentioned by Planck in Andrews.  Gives a program to compute  (4M+2)3  cubes and gives a  63  in base 10  &  base 6.

John Robert Hendricks.  The magic tessaracts of order  3  complete.  JRM 22:1 (1990) 15‑26.  Says there are  58  of them and gives some history.

Jacques Bouteloup.  Carrés Magiques, Carrés Latins et Eulériens.  Éditions du Choix, Bréançon, 1991.  Nice systematic survey of this field, analysing many classic methods.  Includes some material on magic cubes.

 

          7.N.2.          MAGIC TRIANGLES

 

            There are quite a number of possible types here and I have not been very systematic in recording them.

 

Frenicle de Bessy.  Letter to Mersenne, Mar 1640.  In:  Oeuvres de Fermat, op. cit. in 7.N.1, vol. 2, pp. 182‑185.  Discusses a magic triangle.

Scheffler.  Op. cit. in 7.N.1.  1882.  Part II: Das magische Polygon, pp. 47‑88 & Plate I, p. 113.  He considers nested  n‑gons with the number of numbers on each edge increasing  1, 3, 5, ...  or  2, 4, 6, ...,  such that each edge of length  k > 2  has the same sum and the diameters all have the same sum, though these sums are not all the same.  He develops various techniques and gives examples up to  26‑gons and  5‑level pentagons.

H. F. L. Meyer.  Magic Triangles.  In:  M. Adams; Indoor Games; 1912; pp. 357‑362.  He divides a triangle by lines so the triangle of order three has rows of  1, 3, 5  cells.  He gets some lines of  2  and of  3  to add to the same value, and then considers hexagons of six cells, but doesn't really get anywhere.

Peano.  Giochi.  1924.

Prob. 3, p. 2.  Gives triangle with sides  2, 5, 4;   4, 1, 6;   6, 3, 2.

Prob. 4, p. 2.  Gives triangle with sides  8, 1, 6, 5;   5, 4, 9, 2;   2, 3, 7, 8.  Gives a number of simple consequences of the magicness and also that the sum of the squares of the numbers on a side is  126.  [There are  18  magic triangles, but only this one has the sum of the squares constant.]

Collins.  Book of Puzzles.  1927.

Pp. 92-93: The magic triangle.  Consider a triangle with two points between the vertices.  Put the numbers  1 - 9  on the vertices and intermediate points so that the sum of the values on each edge is constant and the sum of the squares of the values is constant.  Gives one answer and is somewhat vague as to whether it is unique.  See Peano.

P. 93: A nest of magic triangles.  Says this occurs on a 1717 document of the Mathematical Society of Spitalfields.  Start with a triangle and join up its midpoints.  Repeat on the resulting triangle and continue to the fourth tie, getting five triangles with 18 points.  The values  1 - 18  are placed on these points to get various sums which are multiples of 19.

Collins.  Fun with Figures.  1928. 

The Siamese twin triangles, pp. 108-109.  Triangle with  4  cells along each side.  Place the digits  1  through  9  on the cells so each line adds to  20.  Two complementary solutions, with sums  19  &  21.  He gives a number of further properties about various sums of squares.

A magic hexagon within a circle, pp. 110-112.  This is really a pattern of six magic triangles like the above, with sums  20  and with sums all distinct, with a further property about sums of squares.

Perelman.  FFF.  1934.  1957: probs. 46-48, pp. 56-57 & 61-62;  1979: probs. 49-51, pp. 70‑71 & 77-78.  = MCBF, probs. 49-51, pp. 69-70 & 74-75.

49: A number triangle.  Triangle with  4  cells along each side.  Place the digits  1  through  9  on the cells so each line adds to  20 -- as in Collins, pp. 108-109.  One solution, but he notes that the two central cells in each line can be interchanged.

50: Another number triangle.  Same with total  17,  again one solution given.

51: A magic star.  Star of David with cells at the star points and the intersections.  Place numbers  1, .., 12  so each line of four and the six points all add to  26.  One solution given.

Ripley's Puzzles and Games.  1966.  Pp. 34-35, item 2.  P;ace the fifteen pool balls so each edge and the central three balls total to the same sum.  Ripley's gives examples with magic sum of 34, 35, 36.  One can construct examples with magic sum of 32, 33, ..., 39.  A sum of 40 initially seems possible but further analysis shows it is not possible.  If one rules out trivial rearrangements leaving the rows having the same elements, there are 2716 distinct solutions.  Each of these has 1296 trivial rearrangements.

Jaime Poniachik, proposer;  Henry Ibstedt, solver.  Prob. 1776 -- Connected differences.  JRM 22:1 (1990) 67  &  23:1 (1991) 74-75.  Triangular lattice with edge  2.  Place numbers  1, ..., 15  on the  6  points and  9  edges so that each edge is the difference of its end points.  19  solutions found by computer.  [Not sure where to put this item??]

 

          7.N.3.          ANTI‑MAGIC SQUARES AND TRIANGLES

 

            An antimagic  n x n  has its  2n+2  sums all distinct.  A consecutively antimagic  n x n  has its  2n+2  sums forming a set of  2n+2  consecutive integers.  Berloquin calls these heterogeneous and antimagic, respectively.  A heterosquare has all the  4n  sums along rows, columns and all broken diagonals being different.

 

Loyd Jr.  SLAHP.  1928.  Magic square reversed, pp. 44 & 100.  3  2  7;   8  5  9;   4  6  1  has all eight sums different.

Dewey Duncan.  ??  MM (Jan 1951) ??NYS -- cited by Gardner.  Defines a heterosquare as an arrangement of  1, 2, ..., n2  such that the  4n  sums along the rows, columns and all broken diagonals are all different.  (However, in the next item it appears that only the main diagonals are being considered??)  Asks for a proof that the  2 x 2  case is impossible and for a  3 x 3  example -- which turns out to be impossible.

John Lee Fults.  Magic Squares.  Open Court, La Salle, Illinois, 1974.  On p. 78, he asserts that Charles W. Trigg posed the problem of non‑existence of anti‑magic  2 x 2's  in 1951, that it was solved by Royal Heath and that Trigg gave the spiral construction for anti‑magics of odd order.  Unfortunately Fults gives no source, only noting that Trigg was editor at the time.  There is nothing in Heath's MatheMagic.  I now see that this is a corruption of the preceding item.  Madachy (see below at Lindon) refers to the problem in MM (1951) without further details, but restricted to just the main diagonals, and then says "An exchange of correspondence between Charles W. Trigg, then "Problems and Questions" editor for the magazine, and the late Royal V. Heath, ..., soon established some basic properties of potential heterosquares."

C. W. Trigg, proposer;  D. C. B. Marsh, solver.  Prob. E1116 -- Concerning pandiagonal heterosquares.  AMM 61 (1954) 343  &  62 (Jan 1955) 42.  The solution is also in:  Trigg; op. cit. in 5.Q; Quickie 160: Pandiagonal heterosquare, pp. 45 & 151.  There is no arrangement of  1, 2, ..., n2  such that the  4n  sums along the rows, columns and all broken diagonals are consecutive numbers.

Charles F. Pinska.  ??  MM (Sep/Oct 1965) 250-252.  ??NYS -- cited by Gardner.  Shows there are no  3 x 3  heterosquares, but gives two  4 x 4  examples.

Gardner.  SA (Jan 1961) c= Magic Numbers, chap. 2.  Notes that  1  2  3;   8  9  4;   7  6  5  is anti‑magic, i.e. all  8  sums are different, and it is also a rook's tour.  In Magic Numbers, Gardner says he had not known of anti-magic squares before seeing this one, but later discovered the Loyd example.  He summarises the knowledge up to 1971.

J. A. Lindon.  Anti‑magic squares.  RMM 7 (Feb 1962) 16‑19.  Summarised and extended in Madachy; Mathematics on Vacation, op. cit. in 5.O, (1966), 1979, pp. 101-110.  Author and editor believe this is the first such article.  Wants the  2n+2  sums for an  n x n  square to be all different and also to be a set of consecutive integers.  There are no such for  n = 1, 2, 3,  but they do exist for  n > 3.

M. Gardner.  Letter.  RMM 8 (Apr 1962) 45.  Points out his SA article and notes that  9  8  7,   2  1  6,   3  4  5  is even more anti‑magic in that the  8  lines and the  4  2 x 2  subsquares all have different sums.

Pierre Berloquin.  The Garden of the Sphinx, op. cit. in 5.N.  1981. 

Prob. 15: Heterogeneous squares, pp. 11 & 93.  Shows there is no antimagic square of order  2  and reports that a reader found  3120  inequivalent antimagic squares of order  3.  (The  8  symmetries of the square are the equivalences.)

Prob. 16: Antimagic, pp. 12 & 94.  Asserts there is no  3 x 3  consecutive antimagic square.  [Translation is unclear whether he is saying none is known or none exists.]  Gives examples of  4 x 4  and  5 x 5  consecutive antimagics and says there are  20  known examples for  4 x 4.

Prob. 148: Superior antimagic, pp. 81 & 184.  Gives a  6 x 6  consecutive antimagic square.

C. W. Trigg.  Special antimagic triangular arrays.  JRM 14 (1981‑82) 274‑278.  Says Gardner gives the following definition in a letter to Trigg on 22 Dec 1980.  Consider a triangular array of the numbers  1, 2, ..., n(n+1)/2.  We say this is anti‑magic if the sum of the three vertices and the  3(n‑1)  sums of rows of two or more, parallel to the sides, are all distinct.  Gardner then asks if these  3n‑2  sums can be consecutive.  Trigg asks when these sums can be in arithmetic progression and finds that this requires  n = 2, 3, 4, 10, 24  or  n > 99.  He resolves the existence problems for  n = 2, 3, 4.

M. Gardner.  Puzzles from Other Worlds.  Vintage (Random House), NY, 1984.  Problem 8: Antimagic at the number wall, pp. 19‑20, 96‑97 & 142‑143.  Notes his examples (which are complementary) are the only rook‑wise connected anti‑magic squares of order  3  and discusses anti‑magic triangles.

 

          7.N.4.          MAGIC KNIGHT'S TOUR

 

          Note: for the  8 x 8,  the magic constant is  260.

 

G. P. Jelliss.  Special Issue -- Magic Tours; Chessics 26 (Summer 1986) 113‑128  &  Notes on Chessics 26 (Magic Tours); Chessics 29 & 30 (1987) 163.  Says Kraitchik (L'Echiquier, 1926), ??NYS, showed  there is no magic tour unless both sides are even.  (Mentioned in his Math. des Jeux, op. cit. in 4.A.2, p. 388.)  Jelliss considers tours by other pieces including various generalized chess pieces.  He gives  8 x 8  magic king's and queen's tours.  Gives  97  inequivalent semi‑magic knight's paths of which  29  are tours.  These are derived from MSS of H. J. R. Murray at the Bodleian Library, Oxford.  He gives Beverley's square -- see below.  He says magic tours (paths?) exist on  8k x 8k  boards for  k ³ 2  and gives a  48 x 48  magic tour.  He attributes the method to Murray (Fairy Chess Review, Aug 1942).

                    The Notes report that one of the magic queen's tours was miscopied and that Tom Marlow has found that one of the  8 x 8  semi‑magic knight's tours is wrong.

 

William Beverley.  On the magic square of the knight's march.  (Letter of 5 Jun 1847.)  (The London, Dublin and Edinburgh) Philosophical Magazine & Journal (of Science) ?? (Aug 1848) 1-5.  Semimagic knight's path with diagonals of  210  and  282.  See:  G&PJ 1 (Sep 1987) 11  &  2 (Nov 1987) 17,  which describe what is known about the problem.  Tom Marlow reports that he has found that there are  101  examples, but he doesn't seem to consider the diagonals, so these are semimagic.  See also:  Murray, 1936, below.

C. F. de Jaenisch.  Op. cit. in 5.F.1.  1862.  Vol. 2, pp. 151-189.  ??NYS.  Semi‑magic knight's tour, with diagonal sums of  256  and  264.  [Given in Dickins, p. 27 -- which Dickins??.]

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Magic squares in chess -- The knight's tour, pp. 101-102.  Semi-magic knight's tour with diagonals of  264  &  256,  taken from a BM MS: Bibl. Reg. 13, A. xviii., Plut. xx. c. -- ??NYS, but same as de Jaenisch's example.  Notes that symmetric cells differ by  32.  Gives a cryptic argument that this solution can be used to produce examples starting at  48  of the cells of the board.

Ball.  MRE, 5th ed., 1911, p. 133 gives a magic king's tour, which is hence a magic queen's tour.  The 11th ed., 1939, p. 185 refers to Ghersi, below.

Italo Ghersi.  Matematica Dilettevoli e Curiosa.  2nd ed., Hoepli, 1921.  Pp. 320‑321, fig. 261 & 265.  Magic king's tour -- different than Ball's.

H. J. R. Murray.  Beverley's magic S-tour and its plan -- probs. 2106-2108.  Problemist Fairy Supplement (later known as Fairy Chess Review) 2:16 (Feb 1936) 166.  Discusses Beverley's 1848 semi-magic path and says the method leads to  28  solutions and many, including Beverley's example, have the property that each quarter of the board is also [semi-]magic.

H. J. R. Murray.  A new magic knight's tour -- Art. 68, prob. 5226.  Fairy Chess Review 5:1 (Aug 1942) 2-3.  A  16 x 16  semi-magic tour.  Cites Beverley and Roget.  Implies that Kraitchik asserted that no such  16 x 16  tours were possible.

Joseph S. Madachy.  Mathematics on Vacation.  Op. cit. in 5.O, (1966), 1979.  Pp. 87-89.  Order 16  magic knight's tour.

Stanley Rabinowitz.  A magic rook's tour.  JRM 18:3 (1985‑86) 203‑204.  Gives one.  Also gives Ball's magic king's tour.  Says the magic knight's tour is still unsolved.

David Marks.  Knight's Tours.  M500 137 (Apr 1994) 1.  Brief discussion of magic knight's tours, giving a semi-magic example due to Euler? and a magic example made up of  2  tours of  32  squares due to Roget.

 

          7.N.5.          OTHER MAGIC SHAPES

 

          See also 7.Q and 7.Q.1.

 

Yang Hui.  Hsü Ku Chai Ch'i Suan Fa.  Op. cit. in 7.N.  1275.  Book 3, chap. 1, Magic squares, pp. 149-151, with Lam's commentary on pp. 311-322.  This includes  6  magic 'circles' which are diagrams of overlapping circles of values such that each circle adds to a constant value.  In two cases, the centres of the circles are used and in one case some lines also give the same total.  Needham, pp. 60-61, and Lam's commentary describe later work by:  Ch'êng Ta-Wei  (1593),  Fang Chung-Thung  (1661),  Chang Ch'ao  (c1680),  Ting I-tung  (Sung dynasty),  Wang Wên-su  (Ming Dynasty)  and  Pao Ch'i‑shou  (late 19C), who has magic cubes, spheres and tetrahedrons -- see 7.N.1.

Kanchusen.  Wakoku Chiekurabe.  1727.  Pp. 5 & 18-21 show two kinds of magic circles.  The first has two rings of  4  and one in the centre so that each ring adds to  22  and each diameter adds to  23.  This is achieved by putting  1  in the centre and then symmetrically placing the numbers in the pairs  2-9, 3-8, 4-7, 5-6.  The second example uses the same pairing principle to give three rings of six, with  1  in the centre, so each ring adds to  63  and each diameter to  64.

See Franklin, c1750, in 7.N for a magic circle.

Curiosities for the Ingenious selected from The most authentic Treasures of Nature, Science and Art, Biography, History, and General Literature.  2nd ed., Thomas Boys, London, 1822.  Magic circle of circles, pp. 56-57 & plate IV, opp. p. 56.  12 - 75  arranged in  8  rings of  8  sectors, with another  12  in the centre, so that each ring and each radius, with the  12  in the centre, makes  360.

Rational Recreations.  1824.  Exer. 4, pp. 24-25.  = Curiosities for the Ingenious.

Manuel des Sorciers.  1825.  Pp. 82-83.  ??NX  4 x 8  semi-magic rectangle, associated.  The row sums are 132 while the column sums are 66.  I don't ever recall seeing a magic rectangle like this before.  For an  A x B  rectangle, we want  1 + ... + AB  =  AB(AB+1)/2  to be divisible by both  A  and  B  which holds if and only if  º B (mod 2).  Since it doesn't make sense to talk about diagonal sums, this can only give semi-magic shapes, hence they should be easier to produce.  I find essentially one solution for the  2 x 4  case.

The Secret Out.  1859.  The Twelve-cornered Arithmetical Star, pp. 374-375.  This is the case  n = 6  of the general problem of arranging  1, 2, ..., 2n  around a circle so that  ai + ai+1  =  an+i + an+i+1  for each  i.  This immediately leads to  ai ‑ an+i  =  ‑ (ai+1 ‑ an+i+1)  and this requires that  n  be odd.  The given solution fails to work at several points.  See Devi, 1976, and Singmaster, 1992, below.

Mittenzwey.  1880.  Prob. 101, pp. 20-21 & 72;  1895?: 117, pp. 25 & 74;                                         A

          1917: 117, pp. 23 & 71.  Gives the triangular form at the right and asks for                            B C

          S  =  A + B + D  =  A + C + F  =  D + E + F  =  B + C  =  B + E.  This easily                    D E F

          forces  D = A,  C = E = 2A,  B = A + F,  S = 3A + F.  The original pattern had  A = 4, F = 5, S = 17,  and asks for a solution with  S = 13.  The solution gives  A = 2, F = 7,  but there are five solutions corresponding to  A = 0, 1, 2, 3, 4.

Hoffmann.  1893.  Chap IV, no. 10: The 'twenty‑six' puzzle, pp. 146‑147 & 191 = Hoffmann‑Hordern, pp. 118-119, with photo on p. 133.  4 x 4  square with corners deleted.  Arrange  1 ‑ 12  so the  4  horizontal and vertical lines of  4  give the magic sum  26,  and so do the central  4  cells and hence the two sets of opposite outer cells.  Gives two solutions and says there may be more.  Photo shows a German version labelled with a large  26,  comprising a board with square holes and 12 numbered cubes, 1880-1895.

I. G. Ouseley.  Letters: Pentacle puzzle.  Knowledge 19 (Mar 1896) 63  &  (Apr 1896) 84.  Consider a pentagram, its five points and the five interior intersections.  Place the numbers  1  to  10  on these so that each line of five has the same sum and the five internal values shall add to the same sum, while the five outer values shall add to twice as much.  Second letter says it seems to be unsolvable and Editorial Note points out that the sum of all the numbers is  55,  which is not divisible by  3,  so the problem as stated is unsolvable.  [But what if we take the numbers  0  to  9 ??]

Pearson.  1907.  Part I, no. 35: A magic cross, p. 35.  Same pattern as Hoffmann, with numbers differently arranged.  Says there are  33  combinations that add to  26.

Williams.  Home Entertainments.  1914.  The cross puzzle, pp. 117-118.  Shape of the Ho Thu (the River Plan, see beginning of 7.N) to make have the same sum of  23  across and down.  [In fact one can have magic sums of  23, 24, ..., 27.]

Wood.  Oddities.  1927.  Prob. 8: A magic star, pp. 9-10.  Make an  8-pointed star by superimposing two concentric squares, one twisted by  45o.   There are  8  vertices of the squares and  8  points of intersection, so there are  4  points along each square edge.  Arrange the numbers  1 - 16  on these points so each edge adds up to the same value.  This forces the sums of the 4 vertices to be the same (which he states as a given) and the magic constant to be  34.  He gives one solution and says there are  18  solutions, which I have confirmed by computer -- doing it by hand must have been tedious or I have overlooked some simplifications.

Collins.  Book of Puzzles.  1927.

Pp. 74-75: The card addition puzzle.  If one views cards as  1 x 2  rectangles, then a  6 x 6  frame can be formed by having three vertical cards on each side and two horizontal cards filling in the top and bottom.  Arrange the cards  1 - 10  so that the total along each side is the same.  Note that this adds four cards along the top and bottom but three along the sides.  Gives one solution and implies it is unique, but I have found ten solutions.

Pp. 82-84: Adding, subtracting, multiplying, dividing and fractional magic squares.  For a subtraction square, if  a, b, c  are the elements in a row, then  c ‑ (b ‑ a)  =  a ‑ b + c  is a constant.  An easy solution is obtained by replacing the odd or the even terms of a magic square by ten minus themselves.  He asserts this idea was invented by Dudeney.  He says the multiplying magic square dates from the last quarter of the 18C but was first published by Dudeney.  For a division square, we have  c/(b/a)  =  ac/b  is a constant.  One can find an easy solution from a multiplication square.  He says this was invented by Dudeney.  Fractional squares are ordinary additive squares with constant of one.

Pp. 90-92: A magic circle.  This is basically an  8 x 8  (semi?) magic square spread into a circular pattern.  He uses the numbers  12 - 75  to get a sum of  348 and then sticks some values of 12 in so as to yield 360.

Pp. 93-95: A magic pentagon.  Five  4 x 4  squares skewed to rhombi and fit together at a point to make a five-pointed star.  Each rhombus has constant  162  and various sums add up to  324.

Meyer.  Big Fun Book.  1940.  Mathematical fun, pp. 98 & 731.  Make the                                    ABC

          figure at the right magical.  I've called this a Magic Hourglass -- see below.                              D  

          standard arguments show that the magic constant must be  12  and  D = 4.                            EFG

          The solution is essentially unique, with one horizontal line containing 

          7, 3, 2.

Anonymous.  The problems drive.  Eureka 12 (Oct 1949) 7-8 & 15.  No. 3.  Place the numbers  1, ..., 20  at the vertices of a dodecahedron so that the sum of the numbers at the corners of each face is the same.  Answer: it cannot be done!  [Similar argument shows that the only regular polyhedron that can be so labelled is the cube.  Then one sees that the sum on an edge is the same as on the edge symmetric with respect to the centre.  Then one finds that an edge must have the numbers  1, 8  on it and hence all the edges parallel to it have a sum of  9,  hence must have the pairs  2, 7;  3, 6;  4, 5.  Putting these vertically and putting  1  in the top face, we find the top face must contain  {1, 4, 6, 7}  and there are three distinct ways to place these.  The cubo-octahedron cannot be done, but I'm not sure about the rhombic dodecahedron.]

Ripley's Puzzles and Games.  1966.  P. 48.  Magic cross-cube.  Consider a  2 x 2 x 2  cube.  This has the numbers  1 - 12,  14 - 25  on the facelets so that each face totals  52  and the three facelets at each corner total  39.  Such an arrangement requires the total of all numbers be a multiple of  24,  but  1 + ... + 24  =  300  is an odd multiple of  12.  312  is the next multiple of  24  and leads to the numbers used. 

Doubleday - 2.  1971.  Ups and downs, pp. 121-122.  Gives the figure at                                 3    3    3

          right, with intermediate lines making three rows horizontally,                                              5 5 5  

          one row vertically and four rows diagonally.  Rearrange the                                           7    7    7

          numbers present so all these rows total  15.  I wondered if one

          can put the numbers  1, ..., 9  on this figure to get the same

          sum on all these lines.  However, the magic sum must then be  15  and the middle number of the top and bottom rows must occur in four different sums  15  and only the digit  5  occurs in four such sums.  Indeed, letting  B  be the middle digit in the top row and adding these four sums gives  4 x 15  =  45 + 3B,  so  B = 5.  Similarly the middle digit of the bottom row must also be  5.  So the magic figure is impossible.  Further, this argument works for Doubleday's problem, forcing the vertical line to be all  5s.  There are then just two solutions, depending on whether the top row is  3 5 7  or  7 5 3.  Doubleday gives one solution, saying you may be able to find others.

Birtwistle.  Math. Puzzles & Perplexities.  1971.                                                                         A       

          Pp. 11 & 13.  Consider the pattern at the right.  Place the digits                                         B       

          1, 2, ..., 8  so that the vertical and horizontal quadruples and the                                  E  F  G  H

          inner and outer circles (i.e.  A, H, D, E  and  B, G, C, F)  all have the                                C       

          same sum.  By exchanging  C, D  with  G, H,  we have a cube with four                            D       

          magic faces, but the pattern has more automorphisms than the cube.  One

          sees that  A + D = F + G,  B + C = E + H,  so these pairs can be interchanged.  Also, one can interchange  B  and  C,  etc.  We can then assume  A = 1,  B < C,  B < E < H,  F < G  and then there are  6  solutions.  Each of these gives  16  solutions with  A = 1  and hence  128  solutions allowing any value of  A.  Hence there are  768  solutions in total.  He gives one.

Shakuntala Devi.  Puzzles to Puzzle You.  Orient Paperbacks (Vision Press), Delhi, India, 1976.  Prob. 102: The circular numbers, pp. 65 & 125.  The case  n = 5  of the arithmetical star of The Secret Out, 1859.  Asks for and gives just one solution.  See Singmaster, 1992.

Gareth Harries.  Going round in triangles.  M500 128 (Jul 1992) 11-12.  Consider the lattice triangle of side two.  This has  4  triangles,  6  vertices and  9  edges.  Place the numbers  1, .., 15  on the vertices and edges so that each edge number is the difference of the numbers at its ends.  He says his computer found  19  solutions.  For the side one problem, there are just two solutions.

David Singmaster.  Braintwister: Correct sum, rounded up.  The Weekend Telegraph (27 Jun 1992) xxx  &  (4 Jul 1992) xxviii.  Based on the version in Devi, I asked for all the solutions for  n = 5  and  n = 4.  Using the argument I gave under The Secret Out, there are no solutions for  n = 4  or any even  n.  For  n = 5,  the common difference  d = │ai - an+i  must be either  1 or 5  and in either case, the antipodal pairs are determined.  Fixing  a1 = 1  gives  4!/2  distinct solutions for each value of  d,  where the divisor factors out mirror images.  In a note to my solution, which was not published, I showed that for general  n,  d  must be a divisor of  n,  and each such divisor gives  (n-1)!/2  distinct solutions.  Now that I have found the version in The Secret Out, I am somewhat surprised not to have found more examples of this problem.

David Singmaster.  Braintwister: Give the hour-glass some time.  The Weekend Telegraph (6 Feb 1993) xxxii  &  (13 Feb 1993) xxxvi.  The Magic Hourglass problem, as in Meyer, though I don't recall where I saw the problem -- possibly in one of Meyer's other books.  I recall that my source gave the value of  D  or  S.

Mirko Dobnik (Rošnja 5, 62205 Starše, Slovenia).  Letter and example of 31 Aug 1994.  Take a solid cube and make each face into a  2 x 2  array of squares.  He puts the numbers  1, 2, ..., 24  into these cells as shown below.  Each face adds up to  50  and numerous bands of  8  around the cube (at least  18)  add up to  100.  As he notes, the roundness of these values is notable.

                              10  13

                              11  16

                                1  22      8  17      3  24    20    5

                                4  23    19    6      2  21      7  18

                              14    9

                              15  12

Mirko Dobnik (Rošnja 5, 62205 Starše, Slovenia).  Letter and example of 24 Jan 1995.  Amends two faces of the above array to the following

                              10  13

                              11  16

                                1  22      8  19      3  24      6  17

                                4  23      5  18      2  21      7  20

                              14    9

                              15  12

 

          7.O.   MAGIC HEXAGON

 

          The unique solution is given at the right.  It is often presented                               18  17    3        

with a point up.  I will describe versions by saying which point or                               11    1     7   19    

edge value is up and whether it is a reflection or not.                                                9     6     5    2    16

E.g. This has 17 up.                                                                                                14    8     4   12    

                                                                                                                                 15  13  10        

 

[Ernst] von Haselburg (of Stralsund).  MS of problem and solution in the City Archives of Stralsund, dated 5 May 1887, with note of 11 May 1887 saying it was sent to Illustrierten Zeitung, Leipzig.  Xerox kindly provided by Heinrich Hemme.

[Ernst] von Haselburg (of Stralsund), proposer and solver.  Prob. 795.  Zeitschrift für mathematischen und naturwissenschaftlichen Unterricht 19 (1888) 429  &  20 (1889) 263‑264.  Poses the problem for side  3  hexagon.  Solution deals with the three sums of six symmetric points to show the central number is at most  8  and then finds only  5  is feasible and it gives a unique solution.  He has  18  up.  Reported by Martin Gardner, 1988, without a diagram;  see also Hemme, Bauch.

William Radcliffe.  38 Puzzle.  UK & US patents, 1896.  ??NYS -- cited by Gardner, 1984.  I couldn't find the US patent in Marcel Gillen's compilation of US puzzle patents, but a version is reproduced in Tapson and in Hemme, which has  3  up.

The Pathfinder.  (A weekly magazine from Washington, DC.)  c1910.  ??NYS  Trigg says Clifford Adams saw the problem here, but with a row sum of  35!

Tom Vickers.  Magic Hexagon.  MG 42 (No. 342) (Dec 1958) 291.  Simply gives the solution, with  13  up, reflected.

M. Gardner.  SA (Aug 1963)  = 6th Book, chap. 3.  Describes Clifford Adams' discovery of it.  Shows  15  up, reflected.

C. W. Trigg.  A unique magic hexagon.  RMM 14 (Jan‑Feb 1964) 40‑43.  Shows the uniqueness in much the same way as von Haselburg, but in a bit more detail.  Shows  15  up, reflected.

Ross Honsberger.  Mathematical Gems.  MAA, 1973.  Chap. 6, section 2, pp. 69-76.  Describes Adams' work, as given in Gardner, and outlines Trigg's proof of the uniqueness.  A postscript adds that a Martin Kühl, of Hannover, found a solution c1940, but it was never published and cites Vickers note.

M. Gardner.  Puzzles from Other Worlds.  Vintage (Random House), NY, 1984, p. 141.  Describes Radcliffe's work -- see above.  Gardner says he was a school teacher at the Andres School on the Isle of Man and discovered the hexagon in 1895 and patented it in the US and UK.  He shows the pattern with 15 up, reflected.

Frank Tapson.  Note 71.25:  The magic hexagon: an historical note.  MG 71 (No. 457) (Oct 1987) 217‑220.  Says he has a book of correspondence of Dudeney's containing:  two letters from Radcliffe in 1902;  Dudeney's copy of Frénicle's letter to Fermat (or Mersenne?? -- see 7.O.1 & 7.N.2)  and a reproduction of Radcliffe's published solution, labelled  'Discovered 1895   Entered at Stationers Hall 1896'.  He shows  13  up, reflected, and a reproduction of Radcliffe's reflected form, which is  3  up.  Radcliffe's letters refer to the similar problem discussed by Dudeney in Harmsworth's Magazine (Jan 1902) and London Magazine (Feb 1902) -- see under Dudeney in 7.O.1.  [These are the same magazine -- it changed its name.]

Martin Gardner.  Letter:  The history of the magic hexagon.  MG 72 (No. 460) (Jun 1988) 133.  Describes von Haselburg, with no diagram, as communicated to him by Hemme.

Heinrich Hemme.  Das Kabinett:  Das magische Sechseck.  Bild der Wissenschaft (Oct 1988) 164-166.  Shows the solution is unique.  Describes Adams's discovery and Gardner's article.  Says R. A. Cooper then discovered Vickers' note, then Tapson discovered Radcliffe's version in 1973.  Hemme says Tapson said Radcliffe was a teacher at the Andreas-Schule on the Isle of Man and got a UK patent.  Then in the mid 1980s, Ivan Paasche of Stockdorf saw the German translation of Gardner and recalled Haselburg's work which he was able to locate.  Paasche found that there was a 'Stadtbaurat' with an interest in mathematics named von Haselberg in Stralsund at the time.  Stadtbaurat has several meanings -- it could be a member of the local planning board or a city architect, but it also was an honorific for a distinguished architect.

Hans F. Bauch.  Zum magischen Sechseck von Ernst v. Haselberg.  Wissenschaft und Fortschritt 40:9 (1990) 240-242 & a cover side.  This is the first source to give von Haselberg's given name and to give a picture of him.  He was born in 1827 and died in 1905.  The original MS of the problem and solution have been located in the City Archives of Stralsund (see above) and Bauch sketches the original method -- there is a fair amount of trial and error -- and reproduces some of the MS.  There are a number of sub‑configurations which must add up to the constant  38  and Bauch shows these, including a Star of David configuration -- see 7.O.1.  Von Haselberg submitted his MS to the Illustrierten Zeitung of Leipzig, but they didn't use it.  He was 'Stadtbaumeister' (= City Architect) in Stralsund and restored the facade of the City Hall.  He also published a five volume work on local architectural monuments. 

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  A magic hexagon, pp. 151-152.  When he gave this problem in his puzzle column, a solution came in only two weeks later and a second solution arrived six weeks after that.  Neither solver seems to have used a computer.  He cites only Honsberger.  The first solver obtained a number of identities which would simplify the solution.

 

          7.O.1.          OTHER MAGIC HEXAGONS

 

Frenicle de Bessy.  Loc. cit. in 7.N.2.  1640.  Discusses a magic hexagon of White's form II (below).

I. G. Ouseley.  Letter:  The puzzle of "26".  Knowledge 18 (1895) 255-256.  Arrange  1 ‑ 12  on the points and crossings of a Star of David so that  the sum on each line  =  sum of vertices of each large triangle  =  26.  Says this has come from Mr. T. Ordish, who has arranged the numbers to sum to  26  in  30  different ways.  A number of these are consequences of the above conditions, e.g.  the sum on the inner hexagon  =  the sum of the vertices of a rhombus  which is 26,  as is the sum along certain angles,  but I can't understand what is meant by  '6  obtuse angles' or  '4  rhomboids'.  "I believe there are at least six ways ...."

                    In the next issue, p. 278, are three letters with a comment by Ouseley.  Un Vieil Étudiant sends four solutions with the sum of the vertices of the large triangles being  13,  and hence the sum on the inner hexagon being  52.  These are complementary to the solutions requested, though neither he nor Ouseley notes this.  These complementary solutions are rather easier to find though and I have indeed found six solutions, as does Ahrens below.  J. Willis sends one solution of the original form, possibly implying that it is unique.  T. sends one different solution.  Ouseley notes that in a solution, the value at a point is the sum of the values at the two opposite crossings.  ?? -- possibly more letters in the next issue.

                    In the next volume, pp. 35 & 84, are two notes saying that the publishers (T. Ordish & Co., London) and the proprietors (Joseph Wood Horsfield & Co., Dewsbury) of the puzzle have complained that the above notes are an infringement of their copyright in the "26" Puzzle and Knowledge apologizes for this.

Dudeney.  Puzzling times at Solvamhall Castle.  London Magazine 7 (No. 42) (Jan 1902) 580‑584  &  8 (No. 43) (Feb 1902) 53-56.  The archery butt.  = CP, prob. 35, pp. 60-61 & 187-188.  Hexagon of  19  numbers so that the  6  radii from the centre to the corners and the  6  sides each add to  22.  Problem is to rearrange them so each adds to  23.  Solution says one can get any number from  22  to  38,  except  30.

William F. White.  Op. cit. in 5.E.  1908.  Magic squares -- magic hexagons, pp. 187‑188.

I. Arrange  1 ‑ 12  on the points and crossings of a Star of David so that  the sum on each line  =  sum of vertices of each large triangle  =  sum on inner hexagon  =  sum of vertices on each parallelogram  =  26  -- i.e. the problem of Ordish/Ouseley.  Gives one solution.  Quotes Escott:  "There are only six solutions."  "The first appeared in Knowledge, in 1895, and the second is due to Mr. S. Lloyd."  [Error for Loyd??]

II. Arrange  1 ‑ 19  in a hexagon, consisting of six equilateral triangles "so that the sum on every side in the same".  This gives just  12  sums of three points, which is Dudeney's problem.  He gives solutions with the sum  =  22  and  23  and notes that subtracting from  20  gives sums  =  38  and  37.

Ahrens.  A&N.  1918.  Chap. XII: "Die wunderbare  26", pp. 133‑140.  Consider a hexagram (or six pointed star) formed from two triangles.  This has  12  vertices.  He finds  6  ways to place  1, 2, ..., 12  on these vertices so that each set of four along a triangle edge adds up to the same value (which must then be  26),  and the six corners of the inner hexagon also add to  26.  I.e. this is Ordish/Ouseley's problem, but with a slightly different statement of conditions.

                    On p. 134, a note says the puzzle "Wunderbare  26" is made and sold by Züllchower Anstalten, Züllchow bei Stettin, and they have registered designs 42,768 and 45,600 for it, though the 'Rustic  26' has already been on sale for many years.  S&B, p. 39, shows an English example with no identification.

Collins.  Fun with Figures.  1928. 

A nest of magic hexagons, pp. 109-110.  Central point surrounded by hexagons of  12, 24, 36  points.  The numbers  1, ..., 73  are placed on the points so the sides of each hexagon add up to  111, 185, 259  respectively  and  the diagonals and the midlines add up to  259.  In fact, opposite points in each hexagon add to  74  and the central value is  37.  Note that  111 = 3 x 37,  185 = 5 x 37,  259 = 7 x 37.

A magic hexagon within a circle, pp. 110-112.  This is really a pattern of magic triangles -- cf. 7.N.2.

Perelman.  1934.  See in 7.N.2 for the Star of David pattern with just one solution.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

Digital sum 1, pp. 29, 163 & 188.  Take  12  points on the vertices and midpoints of a hexagon, so each edge contains three points, together with a point in the middle of the hexagon.  Place the numbers  1, ..., 13  so each straight line of three points has the same sum.  Standard arguments show the central point must be  7  and the magic sum is  21.  Factoring out symmetry, there are four solutions.  He gives one.

 Digital sum 2, pp. 29, 163 & 188.  Magic star of David, as in the  26  Puzzle, but with no condition on the sum of the inner hexagon.  This gives many more solutions -- see Singmaster, 1998.  He gives one solution, which has the sum of the inner hexagon being  26.

 

 

 

 

Mathe mit Energie -- Energie mit Mathe.  Verlag Leipziger Volkszeitung, 1981.                               A     

          Magische Rosette, p. 34 & solutions p. 17.  Place numbers  1, ..., 13  on                         F L G B

          the centred Star of David pattern at the right so that each rhombus from                           K M H 

          the centre to an outer vertex has the same sum  35.  They give one                                 E  J  I C

          solution.  In 2000, I found three, after factoring out the symmetries of the                             D    

          figure.  A little work shows that the magic sum satisfies  20 £ S £ 36  and

          a simple program shows that  21 £ S £ 35  and for  S = 21, 22, ...,  there are  3, 5, 9, 20, 23, 18, 8, 26, 8, 18, 23, 20, 9, 5, 3  solutions, giving a total of  198  solutions.

Bell & Cornelius.  Board Games Round the World.  Op. cit. in 4.B.1.  1988.  Marvellous  '26', p. 79.  Says this was sold for  6d  by T. Ordish "probably in about 1920".  [Above we see that it was known in 1895.]  They quote the instructions from the box:  "each of the six sides as well as the six spaces around the centre total up to  26  with perhaps the finding of several additional 'twenty sixes'."  This is Ordish/Ouseley's problem.

Bauch.  Op. cit. in 7.O, 1990.  In this, he gives the Star of David problem, which is a subproblem of the Magic Hexagon.  He asserts it has  96  'classical' solutions, but he gives no discussion or reference and it is not clear if this is for all possible magic sums -- he shows one solution with magic sum  26  and the central hexagon having total  26.

David Singmaster.  Magic Stars of David and the  26  Puzzle.  Draft written in May 1998.  There are  960  solutions for a magic Star of David.  There is clearly an equivalence given by the symmetries of the regular hexagon, so these solutions fall into  80  equivalence classes.  Six of these are solutions of the  26  Puzzle (i.e. the sum of the central hexagon is also  26) and six are solutions of the complemented problem (i.e. the sum of the outer points is  26).  One can use complementation to reduce the number of classes to  40.  However, there are more symmetries of the magic Star of David when one ignores the additional constraint of the  26  Puzzle -- indeed the pattern is isomorphic to labelling the edges of a cube so the sum of the edges around each face is  26.  Hence one can use the symmetries of the cube to produce  20  equivalence classes of solutions, but these symmetries do not interact simply with the additional sums used in the  26  Puzzle.

 

          7.P.    DIOPHANTINE RECREATIONS

 

          See also 7.E, 7.R.1, 7.R.2, 7.U.

 

          7.P.1.           HUNDRED FOWLS AND OTHER LINEAR PROBLEMS

 

            See Tropfke 565, 569, 572 & 613.

            NOTATION:  n  for  p  at  a, b, c  means  n  items of three types, costing  a, b, c  were bought for a total of  p.  I.e. we want   x + y + z = n;   ax + by + cz = p,  with the conditions that  x, y, z  are positive (or non-negative) integers.

            (a, b)  solutions means  a  non‑negative solutions, including  b  positive solutions -- so  £ b.  I have checked these with a computer program.  I also have a separate numerical index to these problems which enables me to tell whether problems are the same.

            When there are just two types of fowl, one gets two equations in two unknowns, but I have generally omitted such problems except when they seem to be part of an author's development or they represent a different context.  See:  MS Ambros. P114;  Tartaglia 17‑25 & 26;  Hutton;  Ozanam-Hutton 9;  Williams;  Collins.

            Della Francesca gives two problems where the value of two different combinations of two fruits or three animals are given.  These are not Hundred Fowls Problems, but in the second, elimination of one unknown leads to one equation in two unknowns, just as the Hundred Fowls Problem does.  Pacioli gives a similar problem with no integral solutions.  I don't recall seeing other examples of this type of problem.

            The medieval problems of alligation are related, but the solution need not be integral.  See:  Fibonacci;  Lucca 1754;  Bartoli;  della Francesca;  Borghi;  Apianus;  Tropfke 569  for discussion and examples.  See:  Devi  for a modern version with integral solutions.  See  Williams  for a history of this aspect.

 

            Some monetary problems naturally occur here.  Paying a sum with particular values and a specified number of pieces is just our ordinary problem.  Paying a sum with particular values, without specifying the number of pieces, leads to one equation in several unknowns.  This is the same as asking for the number of ways to change the total value.  Perhaps more interesting are problems where one person has to pay a debt to another and they only have certain values, which leads to problems like  ax - by = c.

            Ordinary problem in monetary terms -- see:  Riese (1524);  Dodson (1747?);  Ozanam‑Montucla (1778);  Ozanam-Hutton (1803);  Hall (1846);  Colenso (1849);  Perelman (1934);  M. Adams (1939);  Depew (1939);  Hedges (1950?);  Little Puzzle Book (1955);  Ripley's (1966);  Scott (1973);  Holt (1977); 

            Paying a sum or making change:  ax + by + ...  =  n  -- see:  Simpson (1745 & 1790);  Dodson (1747? & 1753);  Euler (1770);  Moss (1773);  Ozanam-Montucla (1778);  Bonnycastle (1782 & 1815);  Hutton, 1798?;  Ozanam-Hutton (1803);  De Morgan (1831?);  Bourdon (1834);  Unger (1838);  Hall (1846);  Clark (1916);

            Paying a debt with limited values: usually  ax - by = c  -- see:  Euler (1770);  Ozanam‑Montucla (1778);  Bonnycastle (1782);  Stewart (1802);  Ozanam-Hutton (1803);  De Morgan (1831?);  Unger (1838);  Todhunter (1870);  McKay (1940); 

 

            McKay (1940);  Little Puzzle Book (1955)  use the context of buying postage stamps.

            G.F. (1993)  uses the context of wheels of vehicles.

            Simpson (1745);  Dodson (1747?);  Euler (1770);  Ozanam-Montucla (1778);  Bonnycastle (1782 & 1815);  Ozanam-Hutton (1803);  Bourdon (1834);  Todhunter (1870);  Clark (1904);  McKay (At Home Tonight, 1940)  are the only examples here which consider problems with one equation in two unknowns.

            Problems involving heads and feet of a mixture of birds and beasts:  Clark;  Williams;  Collins;  Ripley's.

            Impossible problems -- sometimes a problem is impossible only if positive solutions are required.  Abu Kamil;  Fibonacci 1202 & 1225;  Tartaglia;  Buteo;  Simpson;  Euler;  Ozanam‑Montucla;  Bonnycastle;  Perelman;  Depew;  Little Puzzle Book;  Scott;  Holt.

            I have recently realised that the relatively modern problem of asking how to hit target values to make a particular value is a problem of this general nature, especially if the number of shots is given.  E.g. a target has areas of value  16, 17, 23, 24, 39;  how does one achieve a total of  100?  These occur in Loyd, Dudeney, etc., but I haven't recorded them.  I may add some of them.

 

Zhang Qiujian.  Zhang Qiujian Suan Jing.  Op. cit. in 7.E.  468.  Chap. 3, last problem, pp. 37a ff (or 54 f ??).  ??NYS.   Hundred Fowls: 100 for 100 at  5, 3, ⅓.  (Cocks, hens, chicks.)  This has (4, 3) solutions -- he gives (3, 3) of them, but only states the relation between the solutions -- no indication of how he found a solution.  (Translation in Needham 121-122 (problem only),  Libbrecht 277,  Mikami 43,  Li & Du 99.)

Ho Peng Yoke.  The lost problems of the  Chang Ch'iu‑chien Suan Ching,  a fifth‑century Chinese mathematical manual.  Oriens Extremus 12 (1965) 37‑53.  On pp. 46‑48, he identifies prob. 31 of  Yang Hui  (below at 1275) as being from Chang (= Zhang).  The solution involves choosing one value arbitrarily.

Zhen Luan (= Chen Luan).  Op. cit. in 7.N, also called  Commentary on  Hsü Yo's  Shu Shu Chi I.  c570.  ??NYS.  [See Li & Du, p. 100.]  100 for 100 at  5, 4, ¼.  (Cocks, hens, chicks.)  This has (2, 1) solutions -- he gives (1, 1).  Mikami says Chen's method would give one solution to Chang's problem.  Libbrecht, pp. 278‑279, gives the method, which is indeed nonsense, and states that other scholars noted that Chen's method is fortuitous.  He says Chen also gives 100 for 100 at  4, 3, ⅓,  which has (3, 2) solutions, of which Chen gives (1, 1).

Liu Hsiao‑sun.  Chang ch'iu‑chien suan‑ching hsi‑ts'ao  (Detailed solutions of [the problems] in the  Chang ch'iu‑chien suan‑ching).  c600.  ??NYS.  Described in Libbrecht pp. 279‑280 as nonsense!!

Bakhshali MS.  c7C.  Sūtra C7 (VII 11-12).  Hayashi 648-650 studies this.  General discussion and first example are too mutilated to restore.

Example 2 (VII 11).  20 for 20 at  ?, 4, ½.  From the working, the lost coefficient satisfies  0 £ ? £ 15/32.  One solution is given, but only partly readable:  ?, 3, ?.  Not mentioned by Kaye.

Example 3 (VII 11-12).  20 for 20 at  3, 3/2, ½.  (Earnings of men, women, children.)  (3, 1) solutions, (1, 1) given:  2, 5, 13.  (Also in Kaye I 42; III 191, f. 58v.  A. K. Bag; Mathematics in Ancient and Medieval India; 1979, p. 92 gives just the problem.  Datta, p. 50, says the answer is mutilated, but Hayashi does not comment on this -- Kaye III 191 displays the answer given as  2, 5, 1...,  so the mutilation is pretty minimal.)

Li Shun‑fêng.  Commentary on  Chang chiu‑chien.  7C.  ??NYS.  Libbrecht (p. 280) describes his comments as "unmitigated nonsense".

L. Vanhée.  Les cent volailles ou l'analyse indéterminée en Chine.  T'oung Pao 14 (1913) 203‑210  &  435‑450.  On pp. 204‑210, he gives 24 problems in Chinese and French, but doesn't identify the sources!!

Alcuin.  9C.

Prob. 5: Propositio de emptore in C denarius.  100 for 100 at  10, 5, ½.  (Boars, sows, piglets.)  (1, 1) solution, which he gives.

Prob. 32: Propositio de quodam patrefamilias distribuente annonam. 20 for 20 at  3, 2, ½.  (Dividing grain among men, women, children.)  This has (2, 1) solutions -- he gives (1, 1).

Prob. 33: Alia propositio.  30 for 30 at  3, 2, ½.  (Like Prop. 32.)  (3, 1) solutions, (1, 1) given.

Prob. 33a (in the Bede text): Item alia propositio.  90 for 90 at  3, 2, ½.  (Like Prob. 32.)  (7, 5) solutions, (1, 1) given.

Prob. 34: Item alia propositio.  100 for 100 at  3, 2, ½.  (Like Prop. 32.)  (7, 6) solutions, (1, 1) given.

Prob. 38: Propositio de quodam emptore in animalibus centum.  100 for 100 at  3, 1, 1/24.  (Horses, cows, sheep.)  (2, 1) solutions, (1, 1) given.

Prob. 39: Propositio de quodam emptore in oriente.  100 for 100 at  5, 1, 1/20.  (Camels, asses, sheep.)  (2, 1) solutions, (1, 1) given.

Prob. 47: Propositio de episcopo qui jussit XII panes in clero dividi.  12 for 12 at  2, ½, ¼.  (Dividing loaves among priests, deacons, readers.)  (2, 1) solutions, (1, 1) given.

Prob. 53: Propositio de homine patrefamilias monasteri XII monachorum.  This problem seems to be a major corruption of the following.  12 for 204 at  32, 8, 4.  (Dividing eggs among priests, deacons, readers.)  This problem has one solution: 5, 4, 3  and it seems like the problem was last and some scribe has used the result to reformulate the problem as giving  204/12 = 17  to each.

Mahavira.  850.  Chap. III, v. 133, pp. 67-68 is related to this general type.  Chap VI, v. 143‑153, pp. 130‑135.

     Chap. III.

133:  x(3/4)(4/5)(5/6) + y(1/2)(5/6)(4/5) + z(3/5)(3/4)(5/6)  =  1/2.  This is:  x/2 + y/3 + 3z/8  =  1/2  and he arbitrarily picks two of the values, getting:  1/3, 1/4, 2/3.

     Chap. VI.

143:  complex triple version with prices also to be found.

146:  general method.

147:  72 for 56 at  ⅔, ¾, 4/5, 5/6.  (Peacocks, pigeons, swans, sârasa birds.)  (217, 169) solutions, (1, 1) is given.

150:  68 for 60 at  3/5, 4/11, 8.  (Ginger, long pepper, pepper.)  (1, 1) solutions, (1, 1) given.

151:  gives a complex method.

152:  100 for 100 at  3/5, 5/7, 7/9, 9/3.  (Pigeons, sârasa birds, swans, peacocks.)  (26, 16) solutions, (1, 1) given.  (See Sridhara & Bhaskara II below.)

Pseudo-Alcuin.  9C.  ??NYS -- cited by Hermelink, op. cit. in 3.A.

Sridhara.  c900.  V. 63‑64, ex. 78‑80, pp. 50‑52 & 95.

Ex. 78‑79.  Same as Mahavira's 152.  (Pigeons, cranes, swans, peacocks.)  (26, 16) solutions;  commentator gives (4, 4);  editor gives (16, 16).

Ex. 80.  100 for 80 at  (2, 3/5, ½).  (Pomegranates, mangoes, wood‑apples.)  Commentator gives (5, 5) solutions of the (6, 5).

Abu Kamil [Abū Kāmil Shujā‘ ibn Aslam ibn Muhammad (the  h  should have an underdot) ibn Shujā‘, al-Hāsib al-Mişrī].  Kitāb al‑ţara’if [NOTE:  ş,  ţ  denote  s,  t  with an underdot.] fi’l‑hisāb (the  h  should have an underdot)  (Book of Rare Things in the Art of Calculation).  c900.  Trans. by H. Suter as:  Das Buch der Seltenheiten der Rechenkunst von  Abū Kāmil el‑Mişrī;  Bibl. Math. (3) 11 (1910‑1911) 100‑120.  (I have a reference to an Italian translation by G. Sacerdote; IN:  Festschrift zum 80 Geburtstag M. Steinschneiders; Leipzig, 1896, pp. 169‑194, ??NYS.)  (Part is in English in Ore; Number Theory and Its History; 139‑140.)  Six problems of 100 for 100 at the following.

1.  5, 1, 1/20.  (Ducks, hens, sparrows.)  (2, 1) solutions, (1, 1) given.

2.  2, ⅓, ½.  (Ducks, doves, hens.)  (7, 6) solutions, he gives (6, 6).

3.  4, 1/10, ½, 1.  (Ducks, sparrows, doves, hens.)  (122, 98) solutions, he gives (96, 96) and an early commentator pointed out the missing two positive solutions.

4.  2, ½, ⅓, 1.  (Ducks, doves, larks, hens.)  (364, 304) solutions, he says (304, 304).

5.  3, ⅓, 1/20.  (Ducks, hens, sparrows.)  (1, 0) solutions, he says there is no solution.

6.  2, ½, ⅓, ¼, 1.  (Ducks, doves, ring‑doves, larks, sparrows.)  (3727, 2678)  solutions -- he says  2676  once and  2696  twice.  Suter notes that the Arabic words for  70  and  90  are easily confused.  Suter's comments say there are  2676  solutions.  Tom O'Beirne [Puzzles and Paradoxes; OUP, 1965; Chap. 12] discusses this and finds  2678  solutions -- he is apparently the first to find this number, but this probably appeared in his New Scientist column in 1960-1961, ??NYR.

The Appendix is a fragment of a commentary which the translator fills in to be Prob. A1:  400 for 400 at  1, 3, 2, 1/7.  (Doves, partridges, hens, sparrows.)  This has (1886, 1806) solutions. Suter estimates  c1700.  The commentator and Suter only check the case when the number of doves is divisible by  5.  This has (398, 342) solutions.  The commentator mentions  334  solutions and Suter says  341.

al‑Karkhi.  c1010.  Sect II, no. 10, p. 82.  Mix goods worth  5, 7, 9  to make one worth  8.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 110f., part IV, no. 20.  ??NYS - cited by Tropfke 614, who says it has  (cows, sheep, hens).

Bhaskara II.  Bijaganita.  1150.  Chap VI, v. 158‑159.  In Colebrooke, pp. 233‑235.  Same as Mahavira's 152.  (Doves, cranes, geese, peacocks.)  (This also has  5x + 8y + 7z + 92  =  7x + 9y + 6z + 62.  ??)

Fibonacci.  1202.  Chap. 11: de consolamine monetarum [on the alloying of monies], pp. 143‑166 (S: 227-257) deals with this problem in the more general form of combining metals.  (Other versions involve mixing spices, wines, etc.)  These lead to  a1x1 + ... + anxn  =  b(x1 + ... + xn),  often with  x1 + ... + xn  specified.  xi  is the weight and  ai  is the purity or gold content, etc., of the  i-th metal.  We are mixing the metals to produce a total of weight  x1 + ... + xn  with purity  b.  Hence there is no need to consider only integral values but he usually gives one (or a few) integral solutions.  I describe a few examples.  A  denotes the vector of  ai's.

Pp. 152‑154 (S: 240-242): De consolamine trium monetarum inter se [On the alloying of three monies].  A = (3, 4, 6),  b = 5.  Answers:  1, 1, 3  and  2, 5, 9.

Pp. 156‑158 (S: 245-247): De consolamine septum monetarum [On an alloy of seven monies].  A = (1, 2, 3, 5, 6, 7, 8),  b = 4.  Answer has fractions.

Pp. 158‑159 (S: 247-248) deals with 240 metals!

P. 160 (S: 249-250): De homine qui emit libras  7  trium canium per denarios  7 [On a man who buys 7 pounds of three meats for 7 denari].  7 for 7 at  3, 2, ½.  (Pork, beef, hyrax [an animal somewhat like a rabbit].)  This has (0, 0) solutions!  He gives  1, ⅓, 5⅓!

P. 165 (S: 256): De homine qui emit aves triginta trium generum pro denariis  30 [On a man who buys thirty birds of three kinds for 30 denari].  30 for 30 at  3, 2, ½.  (Partridges, doves, sparrows.)  = Alcuin 33.  He gives the (1, 1) answer.

P. 165 (S: 256-257): De eodem [On the same].  12 for 12 at  2, ½, ¼.  (Same birds.)  = Alcuin 47.  He gives the (1, 1) answer.

P. 165‑166 (S: 257): De eodem cum genera avium sint quattuor [On the same when there are four kinds of birds].  30 for 30 at  3, 2, ½, ¼.  (Partridges, doves, turtle doves, sparrows.)  (27, 19) solutions -- he gives (2, 2).

Pp. 322‑323 (S: 452-453) give some examples with two and four metals done by false position.

Fibonacci.  Epistola.  c1225.  In Picutti, pp. 332-336, numbers XI - XIII.  One of the problems is briefly mentioned in:  M. Cantor; Mathematische Beiträge zum Kulturleben des Völker; Halle, 1863; reprinted by Olms, Hildesheim, 1964; p. 345.  Surprisingly, none of these problems have appeared elsewhere!

P. 247: De avibus emendis secundum proportionem datam.  30 for 30 at  ⅓, ½, 2.  (Sparrows, turtledoves, pigeons.)  This has (3, 1) solutions -- he gives (1, 1).

Pp. 247-248: De eodem.  29 for 29 at  ⅓, ½, 2.  (Sparrows, turtledoves, pigeons.)  This has (2, 2) solutions -- he gives both.

P. 248: Item de avibus.  15 for 15 at  ⅓, ½, 2.  (Sparrows, turtledoves, pigeons.)  This has (0, 2) solutions -- he says "hoc esse non posse sine fractione avium demonstrabo."  By eliminating the first variable, he gets  y + 10z = 60  and he notes that  0, 10, 5  solves the problem.  "Sed si volemus frangere aves", one can have  9/2, 5, 11/2.  However he fails to find  9, 0, 6.

P. 248: above continued.  15 for 16 at  ⅓, ½, 2.  (Sparrows, turtledoves, pigeons.)  This has (1, 1) solutions which he gives.

P. 248: above continued.  30 for 30 at  ⅓, 2, 3.  (Sparrows, pigeons, partridges.)  This has (2, 1) solutions -- he gives (1, 1).

P. 249: above continued.  24 for 24 at  1/5, ⅓, 2, 3.  (Sparrows, turtledoves, pigeons, partridges.)  This has (6, 2) solutions -- he gives (2, 2).

Abbot Albert.  c1240.

Prob. 2, pp. 332‑333:  6 for 50 at  2, 9, 10.  (1, 1) solutions.  Actually he is using this for divination:  if  x + y + z = 6,  then the value of  2x + 9y + 10z  determines  x, y, z.  He gives a table of all the partitions of  6  into  3  non‑negative summands and computes  2z + 9y + 10z  for each.  This is more properly a problem for Section 7.AO.

Prob. 7, p. 334:  30 for 30 at  4, 2, ½.  (Geese??, ducks??, fig-peckers.)  (2, 1) solutions, he gives (1, 1).

Yang Hui.  Xu Gu Zhai Qi Suan Fa  (= Hsü Ku Chai Ch'i Suan Fa).  1275.  Loc. cit. in 7.N, pp. 165‑166, prob. 29‑31.

29.  100 for 100 at  5,3, ⅓.  (Same as Chang's problem, but only gives (1, 1) of the (4, 3) solutions.)

30.  100 for 100 at  7, 3, ⅓.  (Three types of tangerine.)  (6, 4) solutions, (1, 1) given.

31.  same as 30 in terms of wines and with different measures.

Gherardi.  Libro di ragioni.  1328.

Pp. 85‑86.  Chompera.  24 for 24 at  ¼, 2, 3.  (Sparrows, doves, geese.)  This has (1, 1) solutions, which he gives.

P. 86.  Chompera ucelli.  24 for 24 at  1/5, 1, 2, 3.  (Sparrows, thrushes, doves, geese.)  (1, 1) solution given of the (13, 6) solutions.

Lucca 1754.  c1330.

Ff. 9v‑10r, pp. 34‑35.  40 for 40 at  3, 2, ¼.  (Thrushes, larks, sparrows.)  (2, 2) solutions, both given.

Ff. 10r‑10v, pp. 35‑36.  100 for 100 at  3, 1, 1/20.  (Oxen, pigs, sheep.)  (2, 1) solutions, (1, 1) given.

F. 46v, pp. 96‑97.  60 for 600 at  7, 9, 11, 17.  (Mixing grain.) (314, 272) solutions, he gives one:  35, 5, 5, 15.

F. 46v, p. 97.  60 for 480 at  5, 9, 7.  (Mixing of metals.)  (16, 14) solutions, he gives one:  0, 30, 30.

F. 48v, p. 103.  775  for  162.75  at  .16,  .18, .20, .27, .31.  (Mixing of metals.)  This has  (3027289, 2966486)  solutions!  He gives one solution:  250, 150, 150, 100, 125.

(There are several similar problems here with solutions obtained by guessing.)

F. 56r, p. 125.  100 coins worth  2150  at  50, 33, 17, 25, 15.  (2160, 1536)  solutions, he gives one solution:  10, 10, 10, 10, 60.

Also the same with total value  3900.  (526, 388)  solutions, he gives one solution:  60, 10, 10, 10, 10.

F. 59v, pp. 135‑136.  24 for 24 at  ¼, 2, 3.  (Sparrows, doves, geese.)  (1, 1) solutions, which he gives.  = Gherardi, pp. 85‑86.

Munich 14684.  14C.  Prob. XI, p. 79.

12 for 12 at  2, ¼, ½.  (2, 1) solutions, he gives (1, 1).  = Alcuin 47.

20 for 20 at  2, ¼, ½.  (2, 2) solutions, he gives (1, 1).

Prob. XII, p. 79.  12 for 12 at  2, 1, ½, ¼.  (11, 4) solutions, he gives (1, 1).

Narayana Pandita (= Nārāyaņa Paņdita [NOTE:  ņ  denotes  n  with an overdot and the  d  should have an underdot.]).  1356.  Op. cit. in 7.N, p. 1, lines 2‑5, p. 93.  (Same as Mahavira's 152.)  ??NYS -- see Bag, op. cit. under Bakhshali MS, p. 92.

Folkerts.  Aufgabensammlungen.  13-15C. 

12 for 12 at  2, ½, ¼.  (Soldiers, girls, footsoldiers.)  16 sources.  (2, 1) solutions,  (1, 1) given.  = Alcuin 47.

12 for 12 at  2, 1, ½, ¼.  (Dividing a sum of money.)  9 sources.  (11, 4) solutions,  (1, 1) given.  = Munich 14684, XII.

20 for 20 at  2, ½, ¼.  (Birds or horses.)  8 sources.  (2, 2) solutions,  (1, 1) given.  = Munich 14684, 2nd problem.

20 for 20 at  2, 1, ½.  (Birds or horses.)  1 source.  (7, 6) solutions,  (1, 1) given.

20 for 20 at  3, 2, ½.  (Birds or horses.)  2 sources.  (2, 1) solutions,  (1, 1) given.  = Alcuin 32.

30 for 30 at  2, 1/2, 1/10.  (Birds.)  5 sources.  (2, 1) solutions,  (1, 1) given.  Cf AR 45.

30 for 30 at  3/2, 1, 1/2.  1 source.  (16, 14) solutions,  (1, 1) given. 

In no case is the solution method given.  Folkerts conjectures it was done by double false position.  He cites other material, all cited in this section.

Bartoli.  Memoriale.  c1420.  Ff. 89v - 95r (= Sesiano, pp. 134-135).  Eleven problems of alligation, many identical to Lucca 1754.

Jamshid al-Kāshī  = Ğamšīd ibn Mas‘ūd ibn Mahmūd (the  h  should have an underdot), Ġijāt ed-dīn al-Kāšī  = Ghiyāth al-Din al-Kāshī.  Miftāh al-hisāb (the  h  should have an underdot) (The Calculator's Key).  1426.  Ed. by A. Demerdash & M. H. Hifna (the  H  should have an underdot) ; Cairo, nd.  Facsimile and Russian translation by:  B. A. Rozenfeld', V. S. Segal' & A. P. Juškevič as:  Ključ Arifmetiki -- Traktat ob Okružnosti; Moscow, 1956.  ??NYS -- Hermelink, op. cit. in 3.A, says he gives a version with three fowls.

Provençale Arithmétique.  c1430.  Op. cit. in 7.E. 

F. 116r, p. 62.  12 for 12 at  2, 1, ½.  (Men, women, children.)  (5, 3) solutions, he gives (3, 3).

F. 117r, p. 63.  30 for 30 at  4, 2, ½.  (Costs of cloth.)  Same as Abbot Albert, p. 334.  (2, 1) solutions, he gives the positive one.  See Chuquet, 1484, prob. 83.

Pseudo-dell'Abbaco.  c1440.  Prob. 190, pp. 150‑151.  48 for 48 at  4, 2, ¼.  (1, 1) solution, given.

AR.  c1450.  Prob. 45, 81, 102, 119‑126, 309.  Pp. 40, 52, 60‑61, 66‑67, 137‑138, 175‑176, 221‑222.

45.  30 for 30 at  2, ½, 1/10.  (Pears, apples, nuts.)  (2, 1) solutions, (1, 1) given.  Vogel, p. 221, says this appears in Clm 8951, which is unpublished.  = Folkerts, Aufgabensammlungen, sixth example, where Clm 8951 is one of 5 sources cited.

81.  Mix three kinds of wax worth  43, 29, 22  to make wax worth 32.

102.  Mix wines worth  5, 6, 8  to make wine worth 7.

119.  100 for 100 at  10, 5, ½.  (Oxen, cows, sheep.)  (1, 1) solution, which is given.  = Alcuin 5.

120.  20 for 20 at  3, 3/2, ½.  (Oxen, sheep, geese.)  (3, 1) solutions, (1, 1) given.  = Bakhshali.

121.  12 for 12 at  2, ½, ¼.  (Knights, boys, girls eating bread.)  (2, 1) solutions, (1, 1) given.  = Alcuin 47.

122.  12 for 12 at  2, 1, ½, ¼.  (Knights, men, boys, girls dividing a bill?)  (11, 4) solutions, (1, 1) given.  = Munich 14684, XII.

123.  30 for 30 at  3, 2, ½.  (Men women, children.)  (3, 1) solutions, (1, 1) given.  = Alcuin 33.

124.  100 for 100 at  3, 2, ½.  (Men, women, children.)  (7, 6) solutions, (1, 1) given.  = Alcuin 34.

125.  90 for 90 at  3, 2, ½.  (Men, women, children.)  (7, 5) solutions, (1, 1) given.  = Alcuin 33A.

126.  100 for 100 at  3, 1, 1/24.  (Horses, oxen, sheep.)  (2, 1) solutions, (1, 1) given.  = Alcuin 38.

309.  same as 81, done in three different ways.

Correspondence of Johannes Regiomontanus, 1463?-1465.  Edited by Maximilian Curtze as:  Der Briefwechsel Regiomontan's mit Giovanni Bianchini, Jacob von Speier und Christian Roder.  Part II of: Urkunden zur Geschichte der Mathematik im Mittelalter und der Renaissance.  AGM 12 (1902).

P. 262, letter to Bianchini, nd [presumably 1464].  240 for  16047  at  97, 56, 3,  specifically ruling out fractions, but not describing anything bought.

Pp. 293 & 296, letter to von Speier, nd [apparently early 1465].  On p. 293, Regiomontanus invites his friend to a dinner of pheasants, partridges and wine (see Hermelink, op. cit. in 3.A).  P. 296, same problem with (pheasants, partridges, doves).

P. 300, letter from von Speier, 6 Apr 1465.  Gives the only answer:  114, 87, 39.

Benedetto da Firenze.  c1465.  Pp. 151‑152.  Part of the text is lacking, but the problem must be 60 for 60 at  6, ½, ⅓.  (Cows, calves, pigs.)  This has (2, 2) solutions, (1, 1) is given.

Gottfried Wolack.  1468.  Dresden MS C80, ff. 301'-303.  This is a MS of lectures given at Erfurt, 1467 & 1468,  Transcribed and discussed in:  E. Wappler; Zur Geschichte der Mathematik im 15. Jahrhundert; Zeitschrift für Math. und Physik -- Hist.-Litt. Abteilung 45 (1900) 47-56.  P. 51 has:  100 for 100 at  2, 1, ½.  (Men, women, children.)  There is an obscure calculation leading to dividing the people in the proportion  4 : 2 : 1,  so there are  57 1/7  men!  This may be getting confused with 7.G.2 -- Posthumous twins.

della Francesca.  Trattato.  c1480. 

Ff. 13r-13v (58).  Combining metals:  60 for 480 at  5, 7, 9.  (16, 14) solutions, (1, 1) given:  10, 10, 40.  = Lucca 1754, 46v, with different solution.

F. 14v (60).  Combining metals.  = Lucca 1754, 48v.

Ff. 14v-15r (60-61).  Combining grain, stated as:  30 for 300 at  7, 11, 13, 15, 16,  but he solves for prices  7, 9, 11, 13, 15, 16.  (1491, 289) solutions, but he gives a non-integral solution:  (60, 90, 10, 10, 10, 30)/7.

F. 16v (63-64)  = ff. 37v (98).  Two types of cloth.  7 for 100 at  12, 15.  Answer:  (16, 5)/3.  English in Jayawardene.

F. 19r (67-68)  = f. 44r (108).  Three melons less a watermelon are worth 6, while 7 melons and 10 watermelons are worth 28, i.e.  3m - w = 6,  7m + 10w = 28.  I include this as an early example of the use of a negative coefficient in such problems.  Answer:  (88, 42)/37.  English in Jayawardene.

Ff. 19v-20r (68-69).  7 geese, 6 hens and 8 partridges cost  240  but  3 geese, 5 hens and 10 partridges cost  480.  Though not really a hundred fowls problem, one elimination step leads to one equation in two unknowns just as the hundred fowls does.  (1, 1)  solution, given.

F. 20v (69-70).  100 for 100 at  ½, ⅓, 1, 3.  (Pearls, rubies, sapphires, balas-rubies.)  This has (276, 226) solutions -- he gives (1, 1):  8, 51, 22, 19.  English in Jayawardene.  Cf Pacioli, Summa, 17.

Luca Pacioli.  Aritmetica.  c1480.  ??NYS -- described in Sesiano.  F. 238r. 

10 for 10 at  2, 3.  Answer:  20, -10.

10 for 10 at  ½, ⅓.  Answer:  40, -30.

Chuquet.  1484.

     Triparty, part 1.  Of apposition and remotion.  English in FHM 88-90.  Several examples treated as problems in numbers -- no mention of buying anything.  Gives a method of finding one positive solution.

12 for 12 at  2, 1, ½.  = Prov. Arith. no. 1, which has (5, 3) solutions.

12 for 60 at  8, 5, 3.  (3, 2) solutions.

12 for 36 at  4, 3, 2.  (7, 5) solutions.

Two further examples are in the margin and FHM just gives the formulae and one solution.

12 for 35 at  6, 3, 1  {FHM has  6, 1, 1  but this doesn't agree with their answer}.  (2, 2) solutions.

20 for 20 at  6, 2, ½.  (1, 1) solutions

Chuquet says:  "This rule cannot be extended for the discovery of four or more numbers.  Also one should know that all such calculations have several answers, as many as you want, as appears [later in] this book.  Wherefore apposition and remotion is a science which has little to recommend it."

     Appendice.

Prob. 34.  15 for 160 at  (9, 13).  This is determinate and is only included as a lead-in to the next problem.

Prob. 35.  English in FHM 206-207.  15 for 160 at  (11, 13).  (Two kinds of cloth.)  Answer:  35/2,  -5/2.  Chuquet gives an interpretation of this.  See Sesiano, op. cit. in 7.R.  FHM says the "explanation fails to be entirely clear or convincing".

Prob. 40.  English in FHM 208.  Mixing of two kinds of waxes giving 100 for 1100 at  (9, 14).

Prob. 83.  English in FHM 213-214.  Reckoning which is done by the rule of apposition and remotion.  30 for 30 at  4, 2, ½.  Same as Abbot Albert, p. 334.  (2, 1) solutions.  He gives solutions  3,  3,  24  and  4,  ⅔,  25⅓  and says one can have as many as one wants.  This holds because the products are cloth!

Prob. 84.  Mentioned in passing on FHM 214.  20 for  20 x 20  at  30, 25, 16, 18.  (10, 5) solutions -- he gives (1, 1).

Borghi.  Arithmetica.  1484.  Ff. 93v-101v.  Several problems of alligation, getting up to mixing five grains of values  44, 48, 52, 60, 66  per measure to mix to produce a product of value  50.

Johann Widman.  Op. cit. in 7.G.1.  1489.  ??NYS.  Glaisher, pp. 14 & 121, gives:  F. 109v: mix wines worth  20, 15, 10, 8  to make one worth 12.  Gives one solution:  6, 6, 11, 11.

Pacioli.  Summa.  1494.

F. 105r, prob. 14.  3 hens, 4 partridges and 5 geese cost  72;  while  2, 5, 7  cost  94⅔.  That is:  3x + 4y + 5z = 72,  2x + 5y + 7z = 94⅔.  He gives one solution:  (4, 16, 28)/3  with no indication that the problem is indeterminate.  ⅓, 9, 7  is an easier solution.  There is no integral solution and the number of rational solutions is infinite!  Cf della Francesca 19v.

Ff. 105r-105v, prob. 17.  100 for 100 at  ½, ⅓, 1, 3.  (Sheep, goats, pigs, asses.)  This has (276, 226) solutions -- he gives (1, 1):  8, 51, 22, 19.  = della Francesca 20v, with different objects, but same solution.

F. 105v, prob. 18.  20 for 20 at  4, ½, ¼.  (Men women, children) eating at a tavern.  This has (2, 1) solutions -- he gives (1, 1).  (See H&S 93.)

Riese.  Rechnung.  1522.  1544 ed. -- pp. 104‑106;  1574 ed. -- pp. 70r‑71v.  The 1544 ed. calls this section 'Regula cecis oder Virginum';  the 1574 ed. calls it 'Zech rechnen'.  There is first a simple problem with only two types, hence determinate.

20 for 20 at  3, 2, ½.  (Men, women, girls drinking.)  (2, 1) solutions, (1, 1) given.  = Alcuin 32.

100 for 100 at  4, 3/2, ½, ¼.  (Oxen, pigs, calves, goats.)  (265, 222) solutions, (1, 1) given:  12, 20, 20, 48.  The 1574 ed. has a nice woodcut illustration.

Tonstall.  De Arte Supputandi.  1522.  P. 240.  Repeats Pacioli's prob. 14, except there is a misprint -- in the second case, he has  3, 5, 7  costing  94⅔. 

Riese.  Die Coss.  1524.  No. 67, p. 49.  100 for 460 at  3, 5.  (Coins.)

H&S 93 says a tavern version is in Rudolff (1526?).

Apianus.  Kauffmanss Rechnung.  1527. 

Ff. H.viii.r - J.ii.v  is  Regula virginum.  He describes how to eliminate the least valuable variable and then gives a feeble attempt at describing how to find a solution from the result.

[No. 1.]  26 for 88 at  6, 4, 2.  (Men women, girls.)  (10,8) solutions.  He gives  (3,3)  solutions and says more (all?) can be found.

[No. 2.]  20 for 20 at  2, 1, 1/2, 1/4.  (Men women, girls, children.)  (25,14) solutions.  He gives  (3,3).

[No. 3.]  100 for 200 at  4, 3, 5/2, 1.  (Nutmeg, cinnamon, cloves, saffron & pepper.)  (289, 256)  solutions.  He gives (1,1).

[No. 4.]  300 for 2000 at  24, 12, 8, 4.  (Ranks of soldiers.)  (2081, 1921)  solutions.  He gives (1,1).

Ff. J.vvi.r - K.viii.v  is  Regula Alligationis

Ff. K.viii.v - L.ii.r  is  Munzschlagen.   These two sections deal with mixing of wine, spices, metals, etc., getting up to seven types.

Sesiano cites a 16C MS Ambros. P114 sup which gives  40 for 100 at  1/5, 1/10.  Answer:  960, -920.

Cardan.  Practica Arithmetice.  1539.

Chap. 47, f. L.iiii.v (p. 71).  End mentions Pacioli prob. 17.

Chap. 66, section 35, f. DD.v.r (pp. 145-146).  100 for 100 at  3, 2, ½.  (Pigs, asses, cows.)  (1, 1) solution given.  = Alcuin 34.

Chap. 66, section 67, ff. FF.ii.v - FF.iii.v (pp. 155-156).  (67 is not printed in the Opera Omnia.)  100 for 100 at  3, ½, ⅓, 1/11.  (Turtledoves, thrushes, crested larks, sparrows.)  (18, 15) solutions, (1, 1) given.

Tartaglia.  General Trattato.  1556.  Book 16, art. 117‑129, pp. 254r‑255v  &  Book 17, art. 25, 26, 43, 44,  pp. 272v & 277r-277v.  18 versions.  The objects being bought are mostly not in the Italian dictionaries that I have consulted.  They are apparently 16C Italian, probably Venetian dialect.  Several Italian or Italian-speaking friends have helped to determine these -- my thanks to Jennifer Manco, Ann Maury, Ann Sassoon and especially Maria Grazia Enardu and a student of hers.

16-117.  60 for 60 at  4, 2, ½.  (Thrushes, larks, redstarts.)  (3, 2) solutions, he gives (1, 1).

16-118.  20 for 20 at  3, 2, ½.  (Partridges, pigeons, quails.)  (2, 1) solutions, he gives (1, 1).  = Alcuin 32.  He discusses fractional solutions and gives one, but says it is not really acceptable.

16-119.  20 for 240 at  18, 10, 3.  (Sorghum, bran, grape seeds.)  (2, 0) solutions.  He finds  5, 15, 0,  then rejects it and finds a fractional solution but refers back to the previous discussion.

16-120.  40 for 480 at  36, 12, 1.  (Curlews or siskins, stock-doves, starlings.)  (2, 1) solutions, he gives (1, 1).

16-121.  40 for 40 at  3, 2, 1/5.  (Blackbirds, larks, sparrows.) (1, 1) solutions, which he gives.

16-122.  31 for 31 at  3, 2, ⅓.  (Capons, ducks, thrushes.)  (1, 1) solutions, he gives (1, 1).

16-123.  100 for 100 at  3, 1, 1/20.  (Piglets, goats, weasels??)  (2, 1) solutions, he gives (1, 1).  = Lucca 1754, p. 10r.

16-124.  60 for 60 at  3, 1, 1/20.  (Same? animals.)  (2, 1) solutions, he gives (1, 1).

16-125.  100 for 100 at  3, 2, 1/20.  (Same? animals.)  (1, 1) solutions, which he gives.

(Probs. 126‑129 involve men, women, children eating.) 

16-126.  12 for 12 at  2, ½, ¼.  (2, 1) solutions, he gives (1, 1).  = Alcuin 47.

16-127.  15 for 15 at  4/3, 2/3, 1/3.  (I read the last  3  as a  2,  but the answer implies it is a  3.)  (3, 2) solutions, he gives (1, 1).

16-128.  18 for 18 at  2, 1, ½.  (7, 5) solutions, he gives (1, 1).

16-129.  20 for 20 at  4, ½, ¼.  (2, 1) solutions, he gives (1, 1).  = Pacioli 18.

17-25.  16 for 640 at  32, 50.  (Two types of cloth.)  This is determinate and he gives the solution.

17-26.  6 for 332 at  42, 66.  (Two types of cloth.)  Determinate -- he gives the solution.

17 -- 44-45.  See also Bachet, below. 

17-44.  100 for 100 at  3, 1, ½, ⅓.  (Asses, pigs, sheep, goats.)  Quoted from Luca Pacioli, p. 105.  (276, 226) solutions -- he gives one but notes that the method of double false position does not totally resolve the problem, but that it can be solved by a combination of trying and of mathematics, but this is too long to describe here and he reserves it for another time.  = Pacioli 17

17-45.  200 for 200 at  12, 3, 1, ½, ⅓.  (Mules, asses, pigs, goats, sheep.)  "Proposed to me in 1533 by a Genovese."  (8331, 6627)  solutions -- he gives one and indicates that more are possible, saying again that he will deal with this at another time.

Buteo.  Logistica.  1559. 

Prob. 66, p. 274.  50 for 160 at  7, 2, 4, 1.  (Partridges, thrushes, quails, fig-peckers.)  Cites Pacioli for similar problems.  Gives (2, 2) of the (163, 144) solutions.

Prob. 67, pp. 274-275.   3x + 9y + 2z = 50;   7x + 3y + 6z = 70.   Cites Pacioli for a similar problem.  This has no non-negative integral solutions.  Gives three positive solutions, choosing different values to be integral.

Baker.  Well Spring of Sciences.  1562?  Prob. 11,  1580?: ff. 194r-195r;  1646: pp. 305‑306;  1670: pp. 346-347.  20 for 240 at  20, 15, 8.  (Payments to men, women, children.)  (1, 1) solutions, which he gives.

Bachet.  Problemes.  1612.  Addl. prob. X, 1612: 164-172;  1624: 237-247;  1884: 172‑179.

1612 cites  Tartaglia,  Pacioli,  de la Roche,  etc.

41 for 40 at  4, 3, ⅓.  (1, 1) solution which he gives.

20 for 20 at  4, ½, ¼.  (2, 1) solutions -- he gives (1, 1).  = Pacioli 18.

He then describes the last two examples of Tartaglia.  For art. 44, he says there are 226 solutions and gives some.  For art. 45, my copy of Bachet has a defective type which made me think that the price of  3  was a  5,  but seeing Tartaglia has corrected this error.  Bachet says this has  6639  solutions and he gives the numbers for each given number of mules.  I find that for  7  mules, he has  571  instead of  570  and for  4  mules, he has  914  instead of  903,  which accounts for his extraneous numbers, but I cannot see why he might have miscounted.  Labosne adds a general argument for art. 44.

Book of Merry Riddles.  1629?  12 for 12 at  4, 2, ½, ¼.  (Capons, hens, woodcocks, larks).  (5, 1) solutions, with the positive one given.

John Wallis.  A Treatise of Algebra, both Historical and Practical.  John Playford for Richard Davis, Oxford, 1685.  (Not = De Algebra Tractatus.)  Chap. LVIII, pp. 216-218.  ??NX

20 for 20 at  4, ½, ¼.  (Geese, quails, larks.)  = Pacioli 18.  He gives the positive solution of the (2, 1) solutions.  Cites Bachet and gives some general discussion.

100 for 100 at  3, 1, ½, 1/7.  Says the solutions are in Bachet, but they are not.  (121, 81) solutions.

W. Leybourn.  Pleasure with Profit.  1694.  Chap. XIII: Of Ceres and Virginum, pp. 51-55.

Quest. 1, p. 51.  8 for 20 at  4, 2.  (Geese, hens).  This is determinate.

Quest. 2, pp. 51-53.  21 for 26 at  2, 1, ½.  (Men, women, children.)  Some general discussion.  From  x + y + z  =  N  and  ax + by + cz  =  P,  assuming  a > b > c,  he gets  (a‑c)x + (b-c)y  =  P - cN  and deduces that  x  £  (P-cN)/(a-c).  He then assumes  x @ y  and hence that  x  cannot be much less than  (P-cN)/(a-c+b-c)  and he later drops the 'much' from this.  There are  (6, 5)  solutions; he finds all of them, but rejects the case  5, 16, 0  as the problem says there are 'some children'.

Quest 3, p. 53.  30 for 900 at  60, 40, 20.  (Ministers, lame soldiers, poor tradesmen.)  From the above argument, he claims that  x ³ 3,  but then gives the solution  2, 11, 17  as though this showed that  x = 2  was impossible.  He finds  (6, 6)  of the  (8, 7)  solutions.

Quest 4, p. 54.  10 for 1000 at 50, 70, 130, 150.  (English, Dutch, French, Spanish) creditors.  Finds  (2, 2)  of the  (10, 4)  solutions and implies that all solutions are symmetric.

Quest 5, p. 55.  12 for 12 at 2, 1, ½, ¼.  (Different prices of loaves of bread.)  He finds  (2, 2)  of the  (11, 4)  solutions.

Anonymous proposer and solver.  Ladies' Diary, 1709-10  =  T. Leybourn, I: 5, quest. 8.  Mix wines worth  32, 20, 16  to make  56  worth  22.  I.e.  56 for 1232 at  32, 20, 16.  (12, 11) solutions of which the positive ones are given.

Adrastea, proposer; anonymous solver.  Ladies' Diary, 1721-22  =  T. Leybourn, I: 112-113, quest. 89.  24 passengers of four ranks pay £24, with their fares in the proportion  16 : 8 : 2 : 1.  The solutions depend on whether one takes fares as whole numbers of pounds, shillings or pence.  The cheapest fare,  R,  is readily seen to satisfy  15d £ R £ 1£.  The solution says a Mr. Evans collected 100 true answers.  I found  (1, 0)  solutions in pounds and  (89, 43)  solutions in shillings, then wrote a special program to solve the 24 cases which occur in pence, finding  (201, 101)  solutions.  Presumably Mr. Evans missed one of the positive solutions.  Using farthings gives just one more case with  (8, 3)  solutions.

Ozanam.  1725.  [This is one of the few topics where the 1725, 1778 and 1803 editions vary widely -- see each of them.]

Prob. 24, question 8, p. 180.  41 for 40 at  4, 3, ⅓.  (Men, women, children.)  = Bachet's first.  (1, 1) solutions which he gives.

Prob. 50, pp. 255‑256.  100 for 100 at  9, 1, ½, 3.  "Ce problême est capable d'un grand nombre de résolutions; ...."  It has  (73, 46)  solutions -- he gives  (3, 3).

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.  Problem 3.  20 for 20 at  4, ½, ¼.  (Pigeons, larks, sparrows.)  This has (2, 1) solutions, none given.  = Pacioli 18.

Simpson.  Algebra.  1745.  Section XIII, quest. 2-8, 10-12, pp. 170-181  (1790: prob. II‑VII, IX, XI‑XV, pp. 183‑200).

2.  9x + 13y = 2000.  (17, 17) answers which he gives.

3.  Pay £100 (= 2000s) in guineas (= 21s) and pistoles (= 17s), i.e.  21x + 17y = 2000.  (6, 6) solutions which he gives.

4.  Pay £100 (= 2000s) in guineas (= 21s) and moidores (= 27s), i.e.  21x + 27y = 2000.  But  2000  is not a multiple of three, so there are no solutions, as he notes.

5.  Buy sheep at  17  and oxen at  140  to cost  2000.  (1, 1) solution which he gives.

6.  Men pay  42,  women  16,  to make  396.  (1, 1) solution which he gives.

7 (1745 only).                                                             20 for 16 at  2, 1, ¼.  (Loaves of bread of different values.)  Gives (1, 1) of the (2, 2) solutions.  He gets the equation  y = 14 - 2x + (2-x)/3  and neglects to allow  x = 2  in it.

8 (1745 only).                                                             20 for 20 at  3, 2, ½.  (Expenditure of men, women, children.)  Gives (1, 1) of the (2, 1) solutions.  = Alcuin 32.

10 (1790: XI).                                                             5x + 7y + 11z = 224.  This has (72, 59) solutions.  He gives (60, 60), but he has two erroneous solutions for  z = 14  instead of one.

11 (1790: XII).                                                            17x + 19y + 21z = 400.  (13, 10) solutions.  In 1745, there is an algebraic mistake and he gets (9, 9) solutions.  This is corrected to (10, 10) in 1790.

12 (1745 only).                                                           Pay £20 (= 400s) in pistoles (= 17s), guineas (= 21s) and moidores (= 27s).  Gives (9, 7) of the (9, 7) solutions, which is the only time here that he gives solutions with zeroes and is unusual for the time.  Cf. prob. XIV in the 1790 ed.

  The following are in the 1790 ed.

VII  12 for 12 at  2, 1, ¼.  (Loaves of bread of different values.)  (2, 1) solutions, he gives (1, 1).

IX.  87x + 256y = 15410  --  find the least solution.  In fact, there is only (1, 1) solution.

XIII.  7x + 9y + 23z = 9999.  This has  (34634, 34365)  solutions -- he describes the  34365  positive solutions.

XIV.  Pay £1000 (= 20000s) in crowns (= 5s), guineas (= 21s) and moidores (= 27s),  i.e.  5x + 21y + 27z = 20000.  This has  (70734,  70395)  solutions.  He describes all  70734.  Cf. Euler II.III.9;  Bonnycastle, 1782, no. 16.

XV.  12x + 15y + 20z = 100001.  There are  (1388611, 1388611)  solutions which he describes.

Dodson.  Math. Repository.  (1747?); 1775.  He also has several mixture problems and some simple problems which are mentioned at the entry for vol. II, below.

P. 16, Quest. XLI.  Pay  £50 (= 1000s) with  101  coins worth  21/2 s and  5s.

P. 139, Quest CCXXIII.  20 for 20 at  4, ½, ¼  (geese, quails, larks).  There are  (2, 1)  answers.  He gives the positive one.  = Pacioli 18.

P. 140, Quest CCXXIV.  Pay  £100 (= 2000s) with guineas (= 21s) and pistoles (= 17s).  Gives all (6, 6) solutions.  = Simpson 3.

P. 154, Quest. CCXXXVIII.  41 persons spent 40s; men spent 4s, women 3s, children ⅓ s.  I.e.  41 for 40 at  4, 3, ⅓.  He gives the (1,1) answer.  = Bachet's first.

Pp. 331-336, Quest CCI.  Divide 200 into five whole numbers  a, b, c, d, e,  so that  12a + 3b + c + e/2 + z/3 = 200.  I.e.  200 for 200 at  12, 3, 1, ½, ⅓.  He finds 6627 answers.  There are  (8331, 6627)  answers.  = Tartaglia 45.

Les Amusemens.  1749.  Prob. 164, p. 310.  20 for 60 at  6, 4, 1.  (Men, women, valets.)  (3, 2) solutions -- he gives (1, 1).

James Dodson.  The Mathematical Repository.  Vol. II.  Containing Algebraical Solutions of A great Number of Problems, In several Branches of the Mathematics.  I.  Indetermined Questions, solved generally, by an elegant Method communicated by Mr. De Moivre.  II.  Many curious Questions relating to Chances and Lotteries.  III.  A great Number of Questions concerning annuities for lives, and their Reversions; wherein that Doctrine is illustrated in a Multitude of interesting Cases, with numerical Examples, and Rules in Words at length, for those who are unacquainted with the Elements of these Sciences, &c.  J. Nourse, London, 1753.  On pp. 1 - 63, he deals with 13 examples of determining the number of solutions of a linear equation in two (5 cases), three (6 cases) or four (2 cases) unknowns.  Because of the unusual extent of this, I will describe them all.  (I had to completely revise my programs for counting the number of solutions in order to deal with these.  The revision introduced double precision values for the counts and a way of computing the number of solutions for the two variable problem which speeded up the programs up by a factor of about 100, but one problem still took 2½ days!)  P. 1 has a subheading:  The Solution of indetermined Questions in Affirmative Integers, communicated by Mr. Abraham De Moivre, Fellow of the Royal Societies of London and Berlin.

Pp. 1-3, Quest. I.                                                         35x + 43y = 4000.  (3, 3)  solutions which he gives.

P. 4, Quest II.                                                              71x + 17y = 1005.  = p. 143, Quest CCXXVII of vol. I, 1775 version.  (1, 1)  solution which he gives.

Pp. 5-6, Quest III. 21x + 17y = 2000  -- pay  £100  with "guineas at  21,  and pistoles at  17  shillings each"  = Simpson 3  = P. 140, Quest. CCXXXVIII of vol. I, 1775 version.  (6, 6)  solutions, which he gives.

Pp. 6-9, Quest IV.       5x + 8y = 1989.  (50, 50)  solutions which he gives.

P. 9, Quest V.                                                              3x + 5y = 173.  (12, 12)  solutions which he gives.

Pp. 10-14, Quest. VI. 3x + 5y + 8z = 10003.  (417584, 416250)  solutions of which he gives the positive ones.

Pp. 14-22, Quest. VII. 3x + 5y + 19z = 13051.  (299440, 298204)  solutions of which he gives the positive ones.  He describes De Moivre's analysis and then does a different approach.

Pp. 23-34, Quest. VIII. 5x + 7y + 9z = 93256.  (13807365, 13801148)  solutions of which he gives the positive ones.  Again, he does this two different ways.

Pp. 34-35, Quest. IX. 3x + 5y + 9z = 1849.  (12710, 12546)  solutions.  He outlines the method, which leads to adding three arithmetic progressions, but he makes a simple error which leads to a considerably smaller number.

Pp. 36-37, Quest. X. 3x + 5y + 20z = 1849.  (5766, 5612)  solutions.  He outlines the method and it does yield the number of positive solutions.

Pp. 38-39, Quest. XI. 3x + 5y + 17z = 1849.  (6794, 6613)  solutions.  He outlines the method and it does yield the number of positive solutions.

Pp. 39-44, Quest. XII. 2x + 3y + 5z + 30w = 100003.  (185312986853, 185090752407)  solutions.  He sketches the method and says it all adds up to  160190378249  positive solutions.  I have not tried to locate his mistake, but one can estimate the number of solutions of  ax + by + cz + dw = e  as  e3/6abcd  which is  185201 x 106  here, so it is clear that his answer is wrong.  Indeed, even his final addition is incorrect -- I get  160412356049.  He finds that the number of solutions for  w = 3333, 3332, ...  is a quadratic and sums this by Newton's interpolation formula.  I did some calculations and found that the number of positive solutions for w = 3333, 3332, ...  is  1, 24, 77, 160, 273, ...,  whose first differences are  23, 53, 83, ...  and second differences are constant at  30.  Hence the total number of positive solutions is obtained by adding  3333  terms, which is given by   0  +  1 * 3333  +  23 * 3333·3332/2  +  30 * 3333·3332·3331/6   and this gives the answer I found previously.  For reasons which I haven't tried to determine, Dodson gets   0  +  1 * 3333  +  20 * 3333·3332/2  +  26 * 3333·3332·3331/6.

Pp. 46-63, Quest. XIV. 3x + 57 + 19z + 143w = 91306.  (3121604438, 3104216955)  solutions.  He outlines the method but I haven't tried to see if it does yield the number of positive solutions.

Euler.  Algebra.  1770.

     II.I, pp. 302‑310.

Art. 8: Question 5.  Men pay  19,  women  13,  to total 1000.  He gives the general solution and all (4, 4) solutions.

Art. 9: Question 6.  Buy horses worth  31  and oxen worth  21  to cost  1770.  General solution and all (3, 3) solutions.

Art. 15: Question 9.  Men pay  25,  women pay  16.  All together the women pay  1  more than the men.  General solution and first few examples.

Art. 16: Question 10.  Horses cost  31  and oxen  20.  All together, the oxen cost  7  more than the horses.  General solution and first few examples.

Art. 17‑19.  General rule for  bp = aq + n.  Applies to the above problems and Chinese Remainder problems.

     II.II, pp. 311‑317.

Art. 25: Question 1.  30 for 50 at  3, 2, 1.  (Men, women, children.)  Gives all (11, 9) answers.

Art. 26: Question 2.  100 for 100 at  7/2, 4/3, ½.  This has (4, 3) solutions.  He gives (3, 3) answers and mentions  0, 60, 40  as another answer.

Art. 27.  Discusses when such problems are impossible, e.g.  100 for 51 at  7/2, 4/3, ½.

Art. 28: Question 3.  Combine silvers of quality  7, 11/2, 9/2  ounces per marc (= 8  ounces) to produce  30  marcs of quality  6.  I.e.  x + y + z = 30  and  7x + 11y/2 + 9z/2  =  6 x 30  or  30 for 180 at  7, 11/2, 9/2.  He gives all (5, 3) solutions.

Art. 29: Question 4.  100 for 100 at  10, 5, 2, ½.  (Oxen, cows, calves, sheep.)  Gives all (13, 10) answers.

Art. 30: Question 4.  3x + 5y + 7z = 560  and  9x + 25y + 49z = 2920.  Gives all (2, 2) answers.

     II.III: Questions for practice, p. 321.

No. 4.  Old guineas worth  21½  shillings and pistoles worth  17s  to make 2000s.  Gives all (3, 3) solutions.

No. 5.  20 for 20 at  4, ½, ¼.  = Pacioli 18.  Gives (1, 1) of the (2, 1) solutions.

No. 7.  Can one pay £100 (= 2000s) with guineas (= 21s) and moidores (= 27s)?  = Simpson 4.

No. 8.  How to pay  1s  to a friend when I only have guineas (= 21s) and he only has louis d'or (= 17s)?  I.e.  21x ‑ 17y = 1  with  x, y  positive.  He gives the least solution.

No. 9.  Same as Simpson XIV.

Mr. Moss, proposer and solver.  Ladies' Diary, 1773-74  =  T. Leybourn, II: 374-376, quest. 658.  Pay  50£ (= 1000s)  with pistoles (17s), guineas (21s), moidores (27s) and six-and-thirties (36s).  (529, 412)  solutions, of which the positive ones are given.

Ozanam-Montucla.  1778.  [This is one of the few topics where the 1725, 1778 and 1803 editions vary widely -- see each of them.]

Prob. 9, 1778: 195.  Pay  2000  with  450  coins worth  3  and  5,  i.e. 450 for 2000 at  3, 5.

Prob. 13, 1778: 204-205.  This is a complicated problem on the number of ways to make change -- see third section below.  He breaks it up into silver and copper coins as given in the first two sections.

            Pay  60  with coins worth  60, 24, 12, 6.  There are (13, 0) solutions -- he says 13.

            Pay  6k  with coins worth  2, 1½, 1, ½, ¼,  for  k = 1, 2, ..., 10.  There are (155, 2),  (1292, 194),  (5104, 1477),  (14147, 5615),  (31841, 15236),  (62470, 33832),  (111182, 65759),  (183989, 116237),  (287767, 191350),  (430256, 298046) solutions.  He gives the number of non-negative solutions in each case, but he gives four wrong values:  62400,  183999,  287777,  430264.

            Pay  60  with coins worth  60, 24, 12, 6, 2, 1½, 1, ½, ¼.  There are  (1814151, 0)  solutions -- he says  1813899,  which corresponds to using all but the third of the above wrong values to do the final calculation.  Presumably the third wrong value is a misprint, rather than an error.

Prob. 23, 1778: 214-216.  A  must pay  B  31,  but  A  has only pieces worth  5  and  B  has only pieces worth  6,  i.e.  5x - 6y = 31.  Gives the general solution.

     The following are contained in the Supplement and have no solutions given.

Prob. 45, 1778: 433.  120 for 2400 at  12, 24, 60.  (Paying with coins.)  (21, 19) solutions.

Prob. 51, 1778: 434.  A,  B  and  C  have 100,  but nine times  A  plus  15  times  B  plus  20  times  C  equals  1500,  i.e. 100 for 1500 at  9, 15, 20.  Notes that this problem, and other similar problems, have several solutions and one should find them all.  This has (10, 9) solutions.

Prob. 52, 1778: 434-435.  120 for 20 at  3, 2, ½.  (Hares, pheasants, quails.)  [There must be a misprint here as this is clearly impossible.  Perhaps it should be 20 for 20,  which would  = Alcuin 32.  ??]

Prob. 57, 1778: 435.  Pay 24 livres with demi-louis, pieces worth  6  livres and pieces worth  3  livres.  [I think a demi-louis is  10  livres,  so this would be  10a + 6b + 3c = 24.]  This has (5, 0) solutions.

Prob. 58, 1778: 435-436.  10a + 6b + 3c + 2d + e + ½f = 24.  [This is like the preceding and I have assumed that the demi-louis is  10  livres.]  This has  (1178, 2)  solutions.

Bonnycastle.  Algebra.  1782.  Pp. 135-137 give a number of problems of finding some or all the integral solutions of one linear equation in two or three variables.

P. 135, no. 3.  Same as Simpson 2.  All answers given.

P. 136, no. 8.  Same as Simpson 10.  He says there are 60 solutions, but does not give them.

P. 136, no. 9.  Same as Simpson 4.

P. 136, no. 10.  Same as Simpson 3.

P. 136, no. 11.  Same as Bachet's first.

P. 136, no. 12 (1815: p. 158, no. 10).  Same as Euler II.III.8.  He gives the least solution.

P. 136, no. 14.  Pay £351 (= 7020s) with guineas (21s) and moidores (27s).  Asks for the fewest numbers of pieces, and implies that there are  36  answers, but there are  (38, 37)  answers.  See p. 206, no. 49.

P. 137, no. 15.  Mix wines worth  18, 22, 24  per gallon to make  30  gallons worth  20,  i.e.  30 for 600 at  (18, 22, 24).  (6, 4) answers -- he gives the  4  positive ones.

P. 137, no. 16.  Pay £100 (=2000s) in crowns (5s), guineas (21s) and moidores (27s).  He says there are  70734  answers.  This is intended to be Simpson XIV, but that had £1000 (!).  This has only  (725, 691)  solutions.

P. 206, no. 49 (in 1805, no. 48 in 2nd ed., 1788.) 

                                   "With guineas and moidores, the fewest, which way,

                                        Three hundred and fifty-one pounds can I pay?

                                          If paid every way 'twill admit of, what sum

                                    Do the pieces amount to? -- my fortune's to come."

This is the same problem as p. 136, no. 14, but here he says the answer is  9  guineas and  233  moidores (which was  257  in 2nd ed., 1788, but should be  253),  so he is ignoring the case with  0  guineas and  260  moidores -- or this is a misprint.  He says there are  37  solutions -- there are  (38, 37) solutions.

Hutton.  A Course of Mathematics.  1798?  Prob. 19,  1833: 221;  1857: 225.  Pay  £120 (= 2400s)  with 100 coins using guineas (= 21s) and moidores (= 27s), i.e.  100 for 2400 at  27, 21.

John Stewart.  School exercise book of 1801‑1802.  Described by:  W. More; Early nineteenth century mathematics; MG 46 (No. 355) (Feb 1962) 27‑29.  "Having nothing on me but guineas and having nothing on him but pistoles, I wish to pay him a shilling."  = Euler II.III.8.  Least solution given.

Ozanam-Hutton.  1803.  [This is one of the few topics where the 1725, 1778 and 1803 editions vary widely -- see each of them.]

Prob. 9, 1803: 192;  1814: 167;  1840: 86.  Pay  £2000  with  4700  half-guineas and crowns,  i.e.  4700 for 40000 at  10½, 5.  No solution.

Prob. 23, 1803: 209-210.  Prob. 22, 1814: 181-183;  1840: 94.  A  must pay  B  31,  but  A  has only pieces worth  7  and  B  has only pieces worth  5,  i.e.  7x - 5y = 31.  Gives the general solution.

     The following are contained in the Supplement and have no solutions given.

Prob. 45, 1803: 426;  1814: 361;  1840: 186.  Pay £100 (= 2000s) with guineas (= 21s) and pistoles (= 17s),  i.e.  21a + 17b = 2000.  = Simpson 3.

Prob. 51, 1803: 427;  1814: 362;  1840: 187.  Same as Ozanam-Montucla prob. 51.

Prob. 52, 1803: 427;  1814: 362;  1840: 187.  100 for 100 at  7/2, 4/3, ½.  (Calves, sheep, pigs.)  This has (4, 3) solutions.  = Euler II.II.26.

Prob. 57, 1803: 428;  1814: 362;  1840: 187.  Divide  24  into three parts  a, b, c  so that  36a + 24b + 8c = 516,  i.e. 24 for 516 at  36, 24, 8.  This has (3, 3) solutions.

Bonnycastle.  Algebra.  10th ed., 1815.

P. 158, no. 9.  Pay £20 (= 400s) with half-guineas (= 10½ s) and half-crowns (= 2½ s).  This gives  21x + 5y = 800.  (8, 7) solutions -- he says there are 7.

P. 158, no. 11.  Mix spirits worth 12, 15, 18 per gallon to make 1000 gallons worth 17.  This has no integral solutions -- he gives one fractional solution.

P. 228, no. 21.  Pay £100 (= 2000s) with 7s pieces and dollars (worth 4½ s).  I.e.  14x + 9y = 4000.  I find (31, 31) solutions, he says there are 21 -- a misprint?

P. 230, no. 34.  Spend 28s (= 336d) on geese worth 52d and ducks worth 30d.  (1, 1) solution, which he gives.

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Arts. 310-317, pp. 103-110, cover indeterminate problems in general. 

Art. 312, pp. 105-106.  Pay  2£ 16s  with crowns (= 5s) and three shilling pieces.  Describes the (4, 4) solutions.  Observes that negative solutions correspond to refunding.  Notes that paying in crowns (= 5s) and half-sovereigns (= 10s) is impossible.

Art. 314, pp. 107-108.  Discusses  4x + 9y + 10y  =  103  with no context.  Describes the (17, 13) solutions.

Art. 315, p. 108.  Discusses  14x + 11y + 9z  =  360;  x + y + z  =  30  without any context and describes how to solve such problems in general.  Gives (4, 3) of the (5, 3) solutions, omitting  18, 0, 12.

Bourdon.  Algèbre.  7th ed., 1834.

Art. 127, pp. 221-222.  Pay  78 fr.  with pieces worth  5 fr.  and  3 fr.  Sees that the number of  5 Fr.  pieces must be divisible by  3  and finds all (6, 5) solutions.

Art. 138, question 3, pp. 238.  = Euler II.II.28.

Art. 141, question 7, pp. 243-244.  = Euler II.II.29.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 145-172 & 259-262, nos. 541-645.  He treats the topic at great and exhausting length, starting with solving  ax = by,  then  x + y = c,  then  ax + y = c,  then  ax = by + 1  and  ax = by + c,  also phrased as  ax º c (mod b).  He then does Chinese Remainder problems, but with just two moduli.  He continues with  ax + by = c,  initially giving just one solution, but then asking how many solutions there are.  He varies this in an uncommon way -- solve  x + y = e  with  º a (mod b)  and  y º c (mod d).  He also varies the problem by letting  a, b  be rationals.  After 20 pages and 90 problems, he finally gets to one problem, no. 631, of the present type.  He then goes on to  ax + by = c (x + y),  but returns with 12 problems of our type, no. 634-645.  Of these, 640 and 645 have an answer with a zero, but he only gives the positive answers.  None of these problems are the same as any others that I have seen.

Pp. 153-154, no. 578.  Exchange fewer than  120  coins worth    for coins worth  5⅔  with  a value of  14  left over.  I.e.  1½ x  =  5⅔ y  +  14.  Finds two answers.

P. 163 & 261, no. 616.  Change  50  into coins worth  3/8  and  2/9.  (9, 8) answers -- he gives the positive ones.  No. 617 also deals with coins.

Pp. 165-166, no. 631.  30 for 105 at  5, 3, 2 (unspecified goods).  (8, 7) answers -- he gives the positive ones.

PP. 167-168, no. 634.  50 for 395 at  5, 7, 12 (numbers).  (6, 5) answers -- he gives the positive ones.

Thomas Grainger Hall.  The Elements of Algebra: Chiefly Intended for Schools, and the Junior Classes in Colleges.  Second Edition: Altered and Enlarged.  John W. Parker, London, 1846.

P. 132, ex. 4.  55 crowns (= 5s) and shillings make £7 3s.

P. 133, ex. 18.  25 coins worth 1/30 and 1/15 to make 1.

Pp. 232-233, ex. 4.  How many ways can one pay £100 with sovereigns (= £1) and half-guineas (= 21/40 £)?  He gives all (5, 4) answers.

P. 233, ex. 5.  = Simpson 4.

John William Colenso (1814-1883).  The Elements of Algebra Designed for the Use of Schools.  Part I.  Longman, Brown, Green, Longmans & Roberts, London, (1849), 13th ed., 1858 [Advertisement dated 1849].  Exercises 63, p. 114 & Answers, p. 14.

No. 11.  Change a pound into  18  coins comprising half‑crowns (= 2½ s), shillings (= 1s) and six pences (= ½ s).  I.e.  20 for 18 at  5/2, 1, ½.  (4, 4) solutions, all given.

No. 16.  40 for 40 at  5, 1, ¼.  (Calves, pigs, geese.)  (3, 2) solutions, (2, 2) given.

Family Friend (Dec 1858) 357.  Arithmetical puzzles -- 3.  Same as Alcuin's 39 with (oxen, sheep, geese).  I haven't got the answer.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-44, pp. 256 & 397: Arithmetisches Rätsel.  100 for 100 at  10, 3, ½.  (Geese, hares, partridges.)  (2, 1) answers, only (1, 1) given.

Todhunter.  Algebra, 5th ed.  1870.  Many straightforward examples, of which a I give a few.

Examples XLVI, no. 13, pp. 393 & 596.  Same as Hall, 1846, pp. 232-233.  Says there are (5, 4) answers.

Examples XLVI, no. 17, pp. 393 & 596.  Pay  10s 6d (=10½s)  with guineas (= 21s) to one who only has half-crowns (= 2½s).  Gives least solution.

Examples XLVI, no. 18, pp. 393 & 596.  Pay  44s  with sovereigns (= 20s) to one who only has francs (= 4/5 s).  Gives least solution.

Examples LV, no. 6, pp. 503 & 601.  Pay  £24 15s  using shillings and francs, where 26 francs equals 21 shillings.  (24, 23) solutions.  His solution is:  x = 26t;  y = 495 ‑ 21t.

Examples XLVI, no. 28, pp. 395 & 597.  Pay  £4 16s  with  16  coins,  using guineas (= 21s), crowns (= 5s) and shillings.  I. e.  16 for 96 at  21, 5, 1.  (4, 2) answers which he describes.

Examples XLVI, no. 34, pp. 395 & 597.  100 for 100 at  5, 1, 1/20.  (Oxen, sheep, ducks.)  = Alcuin 39.

Mittenzwey.  1880. 

Prob. 111, pp. 22 & 74;  1895?: 129, pp. 27 & 77;  1917: 129, pp. 24-25 & 75.  14 X + 9 Y = 391.  He gives one of the  (3, 3)  answers.

Prob. 121, pp. 25 & 76;  1895?: 139, pp. 29 & 79;  1917: 139, pp. 26 & 77.  100 for 100 at  1/2, 3, 10,  (Hares, does, stags.)  = Leske.  (2, 1) answers, only (1, 1) given.

1895?: 147, pp. 30 & 79-80;  1917: 147, pp. 28 & 77.  Reduces to  10 for 40 at  10, 5, 1.  (Payment for grades.)  (1, 1)  solution, which he gives.

Editorial answer to F. Chapman, a correspondent.  Knowledge 2 (17 Nov 1882) 409.  100 for 100 at  10, 3, ½.  (Bullocks, sheep, pigs.)  (2, 1) answers, only (1, 1) given.  c= Leske.

Hoffmann.  1893.  Chap. IV, no. 35: Well laid out, pp. 152 & 200‑201 = Hoffmann-Hordern, pp. 125-126.  21 for 24 at  2, 3, ½, 4.  (6, 2) solutions -- he gives (2, 2).

Clark.  Mental Nuts.  1904, no. 49;  1916, no. 63.  Turkeys and sheep.  "A drove of turkeys and sheep have 100 heads and feet.  How many are there of each?"  This leads to  3T + 5S = 100,  which has (7, 6) answers.  He says there are  6  answers and gives one example.  His   1897, no. 12;  1904, no. 20;  1916, no. 93   and   1904, no. 67   are ordinary versions, with only one positive answer in each case.

Pearson.  1907.  Part II, no. 155, pp. 144 & 222.  = Alcuin 39.  Only the positive answer is given.

Loyd.  Cyclopedia.  1914.  Sam Loyd's candy puzzle, pp. 121 & 354.  = SLAHP: Assorted postcards, pp. 45 & 101.  = Pacioli 18.  Only the positive solution is given.

Williams.  Home Entertainments.  1914.  The menagerie, p. 127.  Menagerie of birds and beasts has  36  heads and  100  feet.  I.e.  36 for 100 at  2, 4.

Clark.  Mental Nuts. 

1916, no. 5.  How old is dad.  Dad, Ma, Bro & I.  D + M + B + I = 83;  6D = 7M;  M = 3I.  This yields  15I + 2B = 166,  This has  (6, 5)  solutions, but only one is possible.

1916, no. 13.  Change a quarter.  How many ways can you change a quarter (= 25¢) using  1¢, 5¢, 10¢  coins?  (12, 2) solutions; he says  12.

Hummerston.  Fun, Mirth & Mystery.  1924.  The cinema puzzle, Puzzle no. 36, pp. 92 & 177.  How many ways to change  6d  using  6d, 3d, 1d, ½d, ¼d?  (67,0) solutions -- he says 67.

Collins.  Fun with Figures.  1928.  Four out of five have it, p. 183.  Hunter has shot birds and rabbits and has  36  heads and  100  feet.  I.e.  36 for 100 at  2, 4.

Loyd Jr.  SLAHP.  1928.  He gives a number of examples, usually with some extra feature.

Poultry profits, pp. 24 & 91.  100 for 100 at  .62, 1.02, 1.34.  If he makes profits of  .12, .22, .25  on each one, how does he maximize his profit?  (11, 10) solutions -- he says  9, 86, 5  is maximal, but  5, 95, 0  is better.

Easy come, easy go, pp. 25 & 91.  100 for 1000 at  100, 30, 5.  (2, 1) solutions -- he gives (1, 1).

Shooting mathematically, pp. 38 & 97.  Find least number  a + b + c  such that  6a + 12b + 30c  =  17 (a + b + c).

A puzzle in pants, pp. 42 & 99.  147 for 147 at  .49, .98, 2.45,  but he wants to maximize the minimum of the three numbers.  (37, 36) solutions in general.

An observing waiter, pp. 66 & 111.  This has men, women and couples, leading to  a + b + 2c  =  20,   .2a + .3b + 3c  =  20,    which has (1, 1) solutions which he gives.

Perelman.  FFF.  1934.  Hundred rubles for five.  1957: prob. 37, pp. 54 & 57-58;  1979: prob. 40, pp. 69‑70 & 72‑73.  = MCBF: prob. 40, pp. 67 & 70-71.  Magician asks for  20 for  500 or 300 or 200  at  50, 20, 5.  These are all impossible!

M. Adams.  Puzzle Book.  1939.  He has several straightforward problems, which I omit, and the following.  Prob. C.10, pp. 125 & 173.  Use  26d  in  florins (= 24d),  shillings (= 12d),  sixpence (= 6d),  pennies (= 1d)  and  half‑pences (= ½ d)  to measure  5⅝  inches.  The widths of the coins, in 16ths of an inch, are:  18,  15, 12, 19, 16,  respectively.  This leads to:   24a + 12b + 6c + d + e/2  =  26,   18a + 15b + 12c + 19d + 16e  =  90.  (1, 0) solution, which is given.

Depew.  Cokesbury Game Book.  1939. 

Change, p. 211.  Change a dollar into  50  coins.  I.e.  100 for 50 at  1, 5, 10, 25, 50.  (2, 0) answers, both given.  Ripley's and Scott, pp. 129-130 are the same problem but give only one answer.

Six bills, p. 217.  Pay  $63  with six bills, no  $1  bills used.  I.e.  63 for 6 at  2, 5, 10, 20, 50.  (1, 0) answer, given.

McKay.  At Home Tonight.  1940. 

Prob. 3: A mixed bag, pp. 63 & 75-76.  50 for 50 at  5, 1, ½.  (Sheep, lambs, bundles of rabbits, where a bundle is one item.)  This has (6, 5) solutions, but it is added that he got the same number of two items and this has (2, 1) solutions of which he gives the positive one.

Prob. 5: A question of change, pp. 63 & 77.  I want to pay a friend  6/6  (= 78d),  but I only have  4s (= 48d)  pieces and he only has  half-crowns (= 30d).  I.e.  78 = 48x - 30y.  Gives the smallest solution.

Prob. 22: Brown at the market, pp. 66-67 & 81.  Spend 300 at  35, 25, 16.  (Cows, sheep, pigs.)  This has (5, 2) solutions, he gives the positive ones.

Prob. 32: A postage puzzle, pp. 69 & 82.  Spend  2s (= 24d) on  2d  and 4½d  stamps.  (2, 1) solutions, he gives the positive one.

McKay.  Party Night.  1940.  No. 24, pp. 181-182.  6 for 6 at  2, 1, ½.  (Men, women, children eating loaves of bread.)  Gives (1, 1) of the (3, 1) solutions.

Sid G. Hedges.  The Book of Stunts & Tricks.  Allenson & Co., London, nd [c1950?].  Shilling change, p. 45.  Change a shilling into  12  coins without using pennies, i.e.  12 for 12 at  6, 3, ½, ¼.  (2, 1) solutions -- he gives (1, 1).

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  This has a number of examples which I include as illustrating mid 20C usage.

P. 6: Farm deal.  100 for 100 at  10, 5, ½.  (Cows, hogs, chickens.)  (1, 1) solutions, which he gives.  = Alcuin 5  = AR 119.

P. 19: Christmas savings.  100  coins make  500  cents, using  50, 10, 1.  (1, 1) solutions, which he gives.

Pp. 25-26: Arithmetic ability 1, 2, 3.   6  coins make  48, 52, 23  cents, using  25, 10, 5, 1.  There are  (1, 0),  (1, 1),  (1, 0)  answers, which he gives.

P. 41: Buying stamps.  19 for 50 at  1, 2, 3,  with the condition that there be more  1s  than  2s.  (4, 3) answers, of which (1, 1) satisfies the condition and he gives it.

Ripley's Puzzles and Games.  1966. 

Pp. 16-17, item 8.  Change a dollar into  50  coins, i.e.  50 for 100 at  1, 5, 10, 25, 50.  (The use of a dollar is obviously impossible.)    (2, 0) solutions, both with no  50s.  Only the solution with a  25  is given.  Cf Depew, p. 211.  = Scott, pp. 129‑130.

P. 23.  Chickens and sheep have  24  heads and  76  feet, i.e.  24 for 76 at  2, 4.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.

Coin-counting, pp. 105-106.  100 for 500 at  1, 10, 25, 50.  (Coins.)  There are (9, 8) solutions -- only the one with a zero is given.

Change of a dollar, pp. 129-130.  50 for 100 at  1, 5, 10, 25, 50.  As in Ripley's.  Cf Depew, p. 211.

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 143: The farmer and the animals, pp. 88 & 134.  Buy animals at  50, 40, 25, 10  (mules, sheep, goats, pigs)  to produce an average value of  30.  This is essentially an alligation problem as discussed under Fibonacci.  She gives one answer:  1, 1, 2, 1  and says "Other answers are possible" -- somewhat of an understatement since it has a three parameter set of solutions and two of the parameters can range to infinity!

Michael Holt.  Math Puzzles and Games.  Walker Publishing Co., NY, (1977), PB ed., 1983.  Hundred dollars for five, pp. 37-38 & 101-102.  10 for 500 at  10, 25, 50  -- making change, but requiring each type of coin to be used.  (1, 0) answers, hence impossible if each type of coin must be used.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Takeaway pay, pp. 83 & 133.  Nine employees of three types, earning  £5.00, £3.75 and £1.35  per hour earn  £333.60  in a shift of a whole number of hours.  How long was the shift?  This gives us   x + y + z  =  9,   500x + 375y + 135z  =  33360/n,   where  n  is the number of hours in a shift.  In the problem,  x  is given as  1,  but this information is not needed -- I have used this generalization for one of my problems.

Tom Bullimore.  Sherlock Holmes Puzzles.  (Originally: Baker Street Puzzles; Ravette; ©1992 with Knight Features.)  Orient Paperbacks (Vision Books), New Delhi. India, 1998.  No. 34, PP. 38 & 125.  2000 for 7500 at  3, 4, 5.  (Goblet, candlesticks, figurines.)  This would have (751, 749) answers, but he states that the sum of two of the numbers sold must be 506, which leads to just one solution.  [Question: can you find values for the sum of two of the numbers which give two or three solutions?]

G. F.  The crash.  Mathematical Pie, 129 (Summer 1993) 1022  &  Notes, p. 1.  Determinate version involving vehicles having  2, 4, 8  wheels  (motorcycles, cars, lorries),  giving   m + c + l  =  11,   2m + 4c + 8l  =  44,   m = 4.  Omitting the last equation would give us 11 for 44 at  (2, 4, 8),  which has (4, 3) solutions.

David Singmaster.  The hundred fowls, or how to count your chickens.  Accepted by Mathematics Review (Univ. of Warwick) for 1994, but it closed before the article was published.  General survey of the history.

J. Williams.  Mathematics and the alloying of coinage  1202-1700.  Annals of Science 52 (1995) 213-234 & 235-263.  ??NYS -- abstracted in BSHM Newsletter 29 (Summer 1995) 42 -- o/o.  Surveys the problem from Fibonacci through Kersey, etc. in the 17C.

Vladimir Dubrovsky.  Brainteaser B161 - Two-legged, three legged, and four legged.  Quantum 6:3 (Jan/Feb 1996) 15 & 48.  Room contains a number of people sitting on three-legged stools and four-legged chairs.  There are no spare seats and the total number of legs is  39.  This gives  x = y + z,  2x + 3y + 4z = 39.

 

          7.P.2.           CHINESE REMAINDER THEOREM

 

            See Tropfke 636.

            There are two forms of this.  The ancient Indians describe them as the residual pulveriser, where one is given the residues to several moduli, and the non-residual pulveriser, where one has  ax - c º 0 (mod b)  which is  ax - by = c.  The residual form is the classic Chinese Remainder Theorem, whose solution is generally found by reducing to the non-residual form, solved by the Euclidean algorithm.

            Notation.  Multiple congruences are written in an abbreviated notation.  E.g.  n  º  1 (mod 2)  and  n  º  2 (mod 3)  is written  n  º  1 (2),  2 (3)  or even more abbreviatedly as  n  º  -1 (2, 3).

            Standard problem types.

A‑k.      n  º  1 (2, 3, 4, 5, ..., k‑1),   0 (k).

          A-5.  See:  Tartaglia;  Baker;  Dilworth;  Jackson.

          A-7.  See:  Bhaskara I;  Ibn al-Haitam;  Fibonacci;  AR;  Benedetto da Firenze; 

                    della Francesca;  Chuquet;  HB.XI.22;  Pacioli;  Tagliente;  Cardan;  Tartaglia; 

                    Buteo;  Baker;  van Etten;  Ozanam, 1725;  Vyse;  Dodson;  D. Adams, 1801; 

                    Badcock;  New Sphinx;  Magician's Own Book;  Wehman. 

                              Solution is  301 + 420k,  but early examples give just  721.

          A-10.  See: Pacioli. 

          A-23.  See: Pacioli.

B-k.      n  º  0 (2, 3, 4, ..., k).

          B-9.  See:  Euler;  Bonnycastle 1782;  Jackson.

          B-10. See:  Ripley's.

C-k.      n  º  -1 (2, 3, ..., k).

          C-5.  See:  Mahavira.

          C-6.  See:  Ladies' Diary, 1748;  Vyse;  Bonnycastle 1782.

          C-9.  See:  Tartaglia;  W. Leybourn;  Carlile.

          C-10. See:  Pseudo-dell'Abbaco;  Muscarello;  Ripley's.

          C-20. See:  Gentlemen's Diary, 1747;  Vyse.

C.         n  º  ‑1 (3, 4, 5, 6).  See: Brahmagupta;  Bhaskara I ??;  Bhaskara II.

D-k.     n  º  ‑1 (2, 3, 4, 5, ..., k‑1),   0 (k).

          D-5.  See: Baker;  Illustrated Boy's Own Treasury.

          D-7.  See: Fibonacci;  Marliani;  Benedetto da Firenze;  Pacioli;  Ghaligai;  Cardan;  Tartaglia; 

                    Buteo;  Baker;  van Etten;  Ozanam-Montucla;  Dodson.

          D-9.  See: Pacioli.

          D-11.  See: Pacioli.

          D-23.  See: Pacioli.

E.         General result for  3, 5, 7.   See:  Sun Zi;  Fibonacci;  Yang Hui;  AR;  Pacioli; 

                    Tartaglia.

F.         General result for  5, 7, 9.  See:  Fibonacci;  Pacioli.

          Cases with for moduli  28, 19, 15  arise in computing the Julian year -- see: Simpson;  Dodson;  Bonnycastle, 1782;  Todhunter.

I have a separate index of problems.

 

Sun Zi  (= Sun Tzu).  Sun Zi Suan Ching  (Master Sun's Arithmetical Manual).  4C.  [This is not the famous general and writer of The Art of War, c-4C.]  Chap. 3, prob. 26, p. 10b: There is an unknown number of things.  ??NYS.  n  º  2 (3),  3 (5),  2 (7)  & problem  E.  Only the least answer is given.  (See Needham, pp. 34 & 119.  English in Needham 119,  Mikami 32  and  Li & Du 93;  Chinese and English in Libbrecht 269.)

The problem has been transmitted as a folk rhyme in China and Japan.  It is known as  Sun Zi Ge  (The Song of Master Sun) or  Han Xin Dian Bin  (General Han Xin's Method of Counting Soldiers).  (Han Xin  was a general of c‑200.)  The rhyme is cryptic but gives the three multipliers  70, 21, 15  for problem  E.  An English version of a 1592 version of the rhyme and of  Sun Zi's  text is in:  Li Wenlin and Yuan Xiangdong; The Chinese Remainder Theorem, pp. 79‑110 of  Ancient China's Technology and Science, op. cit. in 6.AN.  Li & Yuan also note that such problems arose and were undoubtedly solved in calendrical calculations in the 3C.

Li & Du, pp. 93‑94, discuss several Chinese versions over the next centuries, including the rhyme of Li & Yuan.  Li & Du's p. 94 mentions the calculation of the Da Ming Calendar by Zu Chongzhi in 462 which probably used  11  simultaneous congruences, but the method has not survived.  (Li & Yuan say it was  10  congruences.)

Aryabhata.  499.  Chap. II, v. 32‑33, pp. 74-84.  (Clark edition: pp. 42‑50.)  Rule for the residual form, i.e. the Chinese Remainder Theorem.  The text is rather brief, but Shukla makes it clear that it is giving the Euclidean algorithm for the problem with two residues.  Shukla does an example with three residues and gives the general solution, though the text stops with one solution.  Shukla gives an alternative translation which would apply to the non-residual form of the problem and notes that later writers recognised the relation between the two forms.  (See Libbrecht 229,  Datta & Singh II 87‑99 & 131‑133  and  Bag, op. cit. under Bakhshali MS in 7.P.1, pp. 193‑204.)

Brahmagupta.  Brahma‑sphuta‑siddhanta.  628.  Chap. XVIII, sect. 1, art. 7.  In Colebrooke, pp. 326‑327.  After some astronomical data, he gives Problem  C  as an example.

                    The earlier part of Section 1 (pp. 325-326) deals with the general rule and the later part (pp. 327-338) gives astronomical examples.

Bhaskara I.  Mahā‑Bhāskarīya.  c629.  Edited and translated by Kripa Shankar Shukla.  Lucknow Univ., 1960.  Chap 1, v. 41‑52, Sanskrit pages 7-9; English pages 29‑46.  These deal with the non-residual 'pulveriser' in its astronomical applications and it seems clearly illustrated.  Illustrative examples are provided by Shukla, from other works of Bhaskara, or from Chap 8 of stated, but not worked, examples to this work, e.g.   44789760000 x ‑ 101  =  210389 y  (chap. 8, no. 13).

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 32-33.  Sanskrit is on pp. 132-171; English version of the examples is on pp. 309-332.  Some further applications are discussed on pp. 191-196, 199-201, 332-334.  Bhaskara I gives 26 examples (and four further examples).  I will just give the simpler ones.  Shukla, pp. lxxxi-lxxxii, says Bhaskara I was the first to distinguish the residual and non-residual forms of the problem.  He also gives tables of solutions of  ax - 1 = by  for values of  a, b  of astronomical significance -- Shukla gives these in Appendix II, pp. 335-339.  Shukla asserts that the Indians were the first to find the general solution for the Chinese Remainder Problem and it was transmitted to China, c700.  However, see Li & Du above.

                    Ex. 1:  n  º  1 (5),  2 (7).

                    Ex. 2:  n  º  5 (12),  7 (31).

                    Ex. 3:  n  º  1 (7),  5 (8),  4 (9).

                    Ex. 4:  A-7.  First appearance; see beginning of section for list of later appearances.  Answer:  721.

                    Ex. 5:  8x + 6 º 0 (13).

                    Ex. 6:  11x - 3 º 0 (23).

                    Ex. 7:  576x - 86688 º 0 (210389). 

          All the later examples are similar, arising from astronomical calculations.

          (Bag, loc. cit., cites Examples 2 and 4 above and says  C  occurs, but it is not here.  Datta & Singh may imply that  C  occurs here.  Cf Brahmagupta.)  (See  Libbrecht 231‑232  and  Datta & Singh II 133‑135.)

Bhaskara I.  Laghu‑Bhāskarīya.  c629.  Edited and translated by Kripa Shankar Shukla.  Lucknow Univ., 1963.  Chap 8, v. 17-18, Sanskrit pages 26-27; English pages 99‑102.  This is a simplified version of his earlier Mahā‑Bhāskarīya, so it does not deal with the pulveriser, but these verses give some astronomical problems with number theoretic conditions which lead to uses of the pulveriser, e.g.  36641x - 24 º 0 (394479375).

Mahavira.  850.  Chap. VI, v. 121‑129, pp. 122‑123.  6 simple and 3 more complex examples, e.g. the following.

121.  n  º  7 (8),  3 (13).

122.  n  º  1 (2, 3, 4, 5).

123.  n  º  ‑1 (2, 3, 4, 5).

124.  n  º  1 (2),  0 (3),  3 (4),  4 (5).

127.  2n  º  3 (9),  3n  º  5 (11),  5n  º  2 (8).

Ibn al‑Haitam.  c1000.  ??NYS.  Problem  A‑7  (= Bhaskara I).  (English in Libbrecht, p. 234.)  (See also:  E. Wiedemann; Notiz über ein vom  Ibn al Haitam  gelöstes arithmetisches Problem; Sitzungsber. der phys. Soz. in Erlangen 24 (1892) 83.  = Aufsätze zur Arabischen Wissenschaftsgeschichte; Olms, Hildesheim, 1970, vol. 2, p. 756.)  Answer:  721.

Bhaskara II.  Bijaganita.  1150.  Chap. VI, v. 160 & 162.  In Colebrooke, pp. 235‑237 & 238‑239.

V. 160.  Problem  C  (= Brahmagupta).

V. 162.  n  º  1 (2),  2 (3),  3 (5).

Fibonacci.  1202.

Pp. 281‑282 (S: 402).  Problem  A‑7  (= Bhaskara I).  He says the answer is  301  but that one can add  420.  He doesn't mention this point in later problems.

Pp. 282‑283 (S: 402-403).  Problems  A‑11,  A‑23,  D-7,  D-10,  D-23.  These are all first appearances of these forms, and  A-23,  D-10,  D-23  never occur again, while  A‑11  only reappears in Tartaglia.

P. 304 (S: 428-429).  n  º  2 (3),  3 (5),  4 (7);  problems E  &  F.  (See Libbrecht, pp. 236‑238 for Latin and English.)

Chhin Chiu‑shao  (= Ch'in Chiu Shao  = Qin Jiushao).  Shu Shu Chiu Chang  (Mathematical Treatise in Nine Sections).  1247.  Complete analysis.  (See Libbrecht, passim.  See also Mikami 65‑69  and  Li & Yuan (op. cit under Sun Zhi).  (n  º  32 (83),  70 (110),  30 (135)  is given by Mikami 69.)

Yang Hui.  Hsü Ku Chai Ch'i Suan Fa.  1275.  Loc. cit. in 7.N, pp. 151‑153, problems 1‑5.

1.  n  º  2 (3),  3 (5),  2 (7)  & problem  E  (= Sun Zi).

2.  n  º  2 (3),  3 (5),  0 (7). 

3.  n  º  1 (7),  2 (8),  3 (9). 

4.  n  º  3 (11),  2 (12),  1 (13).

5.  n  º  1 (2),  2 (5),  3 (7),  4 (9).

Giovanni Marliani.  Arte giamata arismeticha.  In codex A. II. 39, Biblioteca Universitaria de Genova.  Van Egmond's Catalogue 139 dates it c1417.  Described and partly transcribed by Gino Arrighi; Giuochi aritmetici in un "Abaco" del Quattrocento  Il matematico milanese Giovanni Marliani.  Rendiconti dell'Istituto Lombardo.  Classe di Scienze (A) 99 (1965) 252‑258.  Prob. IV: D-7.

Pseudo-dell'Abbaco.  c1440. 

Prob. 114, p. 95.  n  º  ‑1 (2, 3, ..., 10)  (C-10).  Takes  10! ‑ 1.

Prob. 115, p. 96.  n  º  1 (2, 3, ..., 10).  Takes  7560 + 1  and says  75601  also works.

AR.  c1450.  Prob. 268, 311, 349.  Pp. 120‑121, 138‑139, 153, 181, 228‑229.

268.  Divinare.  Three moduli used for divination.  Gives the multipliers for triples;  3, 5, 7;  2, 3, 5;  3, 4, 5;  3, 4, 7;  2, 3, 7;  2, 7, 9;  5, 6, 7;  5, 8, 9;  9, 11, 13.

311.  Case  3, 5, 7  of prob. 268  = Problem  E.

349.  Problem  A‑7  (= Bhaskara I).  Answer:  721.

Correspondence of Johannes Regiomontanus, 1463?-1465.  Op. cit. in 7.P.1.

P. 219, letter to Bianchini, late 1463 or early 1464, question 8:  n  º  15 (17),  11 (13),  3 (10).

P. 237, letter from Bianchini, 5 Feb 1464.  Bianchini answers the above problem with  1103  and  3313  and says there are many more solutions, but he doesn't wish to spend the labour required to find more.  Curtze notes that Bianchini must not have understood the general solution.

P. 254, letter to Bianchini, nd [presumably 1464].  Notes that  1103  is the smallest solution and that the other solutions are obtained by adding the product of  17, 13 and 10,  namely  2210.  Curtze notes that Regiomontanus clearly understood the general solution.

P. 295, letter to von Speier, nd [apparently early 1465].  Prob. 6:  n  º  12 (23),  7 (17),  3 (10).

Benedetto da Firenze.  c1465.  Pp. 68‑69.  Problems  A‑7 (= Bhaskara I),  D-7 (= Fibonacci).  He indicates the general answers.

Muscarello.  1478.  Ff. 69r-69v, p. 180.  n  º  -1 (2, 3, 4, ..., 10)  (= Pseudo-dell'Abbaco).

della Francesca.  Trattato.  c1480.  F. 122v (261).  A-7.  Answer:  721.  English in Jayawardene.

Chuquet.  1484. 

Prob. 143.  English in FHM 227, with reproduction of original on p. 226.  Prob.  A‑7 (= Bhaskara I).  Gets  301  by trial and error and says there are other solutions, e.g.  721,  519841,  90601.  "Thus it appears that such questions may have several and divers responses."

Prob. 144.  n  º  2 (3, 4, 5, 6),  0 (7).  Mentioned on FHM 227, which erroneously implies  n  º  1 (3, 4, 5, 6).

HB.XI.22.  1488.  Pp. 52‑53 (Rath 247).  Prob. A‑7  (= Bhaskara I).  Answer:  721.  Editor notes that  301  is the smallest solution.

Pacioli.  De Viribus.  c1500.  Problems XXII - XXV. 

Ff. 34v - 36v.  XXII effecto atrovare un numero pensato non piu de 105 (XXII effect: to find a number thought of not larger than 105)  = Peirani 62-64.  Problem  E. 

Ff. 36v - 39r.  XXIII. effecto atrovare un Numero pensato non piu de 315 (XXIII effect: to find a number thought of not larger than 315)  = Peirani 64-67.  Problem  F.

Ff. 39r - 42r.  XXIIII. effecto 1 n ch' partito per 2.3.4.5.6. sempre avanzi 1o. et partito per .7. avanzi nulla (XXIIII effect: a number which divided by 2, 3, 4, 5, 6, always leaves remainder 1 and divided by seven leaves nothing)  = Peirani 67-71.  Problem  A‑7  (= Bhaskara I).  Discusses general solution.  Then does  A-10.

Ff. 42r - 44r.  XXV. effecto atrovare un Nů ch' partito in 2. avanza 1o. in .3 . 2. in .4 . 3. in 5 . 4. in 6. 5. in 7. nulla etc (XXV effect: to find a number which divided by 2 leaves remainder 1; by 3, 2; by 4, 3; by 5, 4; by 6, 5; by 7, nothing, etc.)  = Peirani 71-73.  Problem  D-7  (= Fibonacci).  Gives general solution.  Then tries to solve  D-9  as  2+1*3+2*4+3*5+4*6+5*7+6*8+7*9  =  725751,  but this is  º  ‑9 (2,3,...,8), 0 (9).  A marginal note, omitted by Peirani, is basically illegible in Uri's photo and the microfilm, but seems to  be noting that the answer is divisible by  3,  hence does not have remainder  2  when divided by  3.  In fact,  D-9  is inconsistent and unsolvable.  He then considers  D-11  and gets the answer  2519  and says one can determine a value whose multiples can be added to  2519  to get more solutions, but he doesn't compute this.  He then examines  D-23  and gets the minimal solution  4,655,851,199,  which is  20 * LCM (2, 3, ..., 22) - 1.  He then says that if one wants the remainder on division by  23  be other than  0,  and seems to say that if one takes the same condition, then one gets  698,377,681.  This is  3 * LCM (2, 3, ..., 22) + 1,  and is the solution of  A-23.

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 116, f. 57v.  Woman with basket of eggs -- problem  A‑7 (= Bhaskara I).

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 26, f. 66r.  Basket of eggs -- problem  D‑7 (= Fibonacci).

Cardan.  Practica Arithmetice.  1539.  Chap. 66, sections 63 & 64, ff. FF.i.v - FF.ii.r (pp. 154‑155).  Problems  A-7 (= Bhaskara I)  &  D-7 (= Fibonacci).

Tartaglia.  General Trattato, 1556, art. 146‑150, pp. 257v‑258v; art. 199, p. 264r. 

146.  Problem  A‑7  (= Bhaskara I).

147.  Problem  A‑5.  First appearance.

148.  Problem  A‑11  (= Fibonacci).

149.  n  º  ‑1 (2, 3, ..., 9).

150.  Problem  D-7  (= Fibonacci).

199.  Problem  E.

Buteo.  Logistica.  1559.  Prob. 70, pp. 279-280.  Problem  D-7 (= Fibonacci).  Discusses problem  A‑7 (= Bhaskara I)  and Cardan.

Baker.  Well Spring of Sciences.  1562?

Prob. 4,  1580?: ff. 190v-191r;  1646: pp. 300-301;  1670: pp. 342-343.  Probs.  A‑5 (= Tartaglia)  &  A‑7 (= Bhaskara I).

Prob. 5,  1580?: ff. 191v-192r;  1646: pp. 301-302;  1670: p. 342.  Problem  D‑5.  First appearance.

Prob. 6,  1580?: f. 192r;  1646: p. 302;  1670: pp. 343-344.  Problem  D-7 (= Fibonacci).

Bachet.  Problemes.  1612.  Prob. V: Faire encore le même d'une autre façon, 1612: 37-45.  Prob. VI, 1624: 84-93;  1884: 34‑37.  General solution for  3, 4, 5  used for divination.  Labosne adds case  2, 3, 5, 7  and a general approach.  1612 cites Forcadel, Gemma Frisius, Tartaglia, etc.

van Etten.  1624.  Prob. 51‑52 (46‑47), pp. 46‑48 (69‑71).  Problems  A‑7 (= Bhaskara I)  and  D-7 (= Fibonacci).  French ed. refers to Bachet for more detailed treatment.  Henrion's 1630 Notte to prob. 52, p. 18, says that Bachet has treated this problem. 

Seki Kōwa.  Shūi Shoyaku no Hō.  1683.  ??NYS -- described in Smith & Mikami, pp. 123‑124.  Studies  ax - by = 1.   n  º  1 (5),  2 (7)  (= Bhaskara I).   Then extends to any number of congruences.  Then does the system   35 n  º  35 (42),  44 n  º  28 (32)  and   45 n  º  35 (50).

W. Leybourn.  Pleasure with Profit.  1694.  Prob. 17, pp. 40-41.  Prob.  C-9.  Constructs a table of  2·n! - 1,  n = 2, ..., 9,  and says  2·9! - 1  is the least solution.  But he then gives  2519 + 2520k,  k = 0, ..., 7  and says these are some of the infinitely many  other solutions

Ozanam.  1694.  Prob. 23, 1696: 74-77;  1708: 65‑67.  Prob. 27, 1725: 188‑198.  Prob. 10, 1778: 195-198;  1803: 192-195;  1814: 167-169;  1840: 86-87.  1696 gives many examples, too numerous to detail, and some general discussion.  The following is common to all editions:   n  º  1 (2, 3, 5),  0 (7).  1725 has problem  A-7 (= Bhaskara I).  1778 et seq. has problem  D-7 (= Fibonacci)  and then notes that Ozanam would solve this as  119 (mod 5040)  rather than  119 (mod 420)  -- but the 1696 or 1725 ed. only have relatively prime moduli.

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.  Problem 2.  Problem  A-5  (= Tartaglia).

Simpson.  Algebra.  1745.  Section XIII.

Quest. 1, p. 170 (1790: prob. I, p. 182).  n  º  7 (17),  13 (26).

Quest. 9, pp. 175-176.  (1790: prob. VIII, p. 187).  n  º  19 (28),  15 (19),  11 (15).

1790: Prob. X, pp. 189‑190.  General case for  28, 19, 15.  This is for computing the Julian year.  Cf Dodson.

Gentlemen's Diary, 1747 -- see Vyse, below.  C-20.

Ladies' Diary, 1748 -- see Vyse, below.  C-6.

Les Amusemens.  1749.  Prob. 171, p. 318.  n  º  1 (2, 3, 4, 6),  4 (5),  0 (7).

Euler.  Algebra.  1770.

     II.I.

Art. 13, pp. 305‑306.  n  º  2 (6),  3 (13).

Art. 14: Question 8, pp. 306‑307.  n  º  16 (39),  27 (56).

Art. 20: Question 11, p. 310.  n  º  3 (11),  5 (19).

Art. 21: Question 12, pp. 310‑311.  n  º  3 (11),  5 (19),  10 (29).

     II.III: Questions for practice.

No. 10, p. 321.  Problem  B-9.

Vyse.  Tutor's Guide.  1773?  The following are in a supplement in the Key only, but some are referred to earlier sources.

Prob. 1, Key p. 361.  C-20.  Attributed to the Gentlemen's Diary, 1747.

Prob. 3, Key p. 362.  A-7 in verse.

Prob. 4, Key pp. 362-363.  C-6.  Attributed to the Ladies' Diary, 1748.

Dodson.  Math. Repository.  1775.

P. 142, Quest. CCXXVI.  A-7  (= Bhaskara I).

Pp. 142-143, Quest CCXXVII.  D-7  (= Fibonacci).

Pp. 148-149, Quest. CCXXXVI.  x  º  n (28),  m (19).  Then gives an Example: "The cycle of the sun 17; and the cycle of the moon 13; being given; to find the year of the Dionysian period?"  This is the case  n = 17,  m = 13  of the general problem.

Pp. 150-151, Quest. CCXXXVII.  x  º  n (28),  m (19),  p (15).  Then applies to Julian period where the first is the cycle of the sun, the second is the cycle of the moon and the third is the Roman indiction.  Cf Simpson.

Pp. 151-153.  He continues the discussion to find the general solution for any number of moduli which are prime to each other.  Then does  x  º  1 (2),  2 (3),  3 (5),  4 (7),  5 (11).

Bonnycastle.  Algebra.  1782.  Pp. 137-140 (c= 1815: pp. 159-162) discuss the general method and give the following examples and problems.

No. 1.  n  º  7 (17),  13 (26).  (1815: no. 1,  = Simpson).

No. 2.  n  º  3 (11),  5 (19),  10 (29).  (1815: no. 2,  = Euler).

No. 3.  n  º  7 (19),  13 (28).

No. 4.  n  º  2 (3),  4 (5),  6 (7),  0 (2).

No. 5.  n  º  6 (16),  7 (17),  8 (18),  9 (19),  10 (20).

No. 6.  Problem  B-9.  (1815: no. 6,  = Euler).

No. 7.  n  º  1 (2),  2 (3),  3 (5),  4 (7),  5 (11)  (= Dodson).

P. 205, no. 35 (in 1805; 34 in 1788).  n  º  -1 (6, 5, 4, 3, 2)  (C-6).

Carlile.  Collection.  1793.  Prob. LXVIII, pp. 39-40.  n  º  ‑1 (2, 3, 4, 5, ..., 9)  (= Tartaglia, C-9).  He gives the answer:  2519.

D. Adams.  Scholar's Arithmetic.  1801.  P. 208, no. 65.  A-7.  Answer:  721.

Bonnycastle.  Algebra.  10th ed., 1815.  Pp. 159-162 is similar to the 1782 ed., but has the following different problems.

No. 3.  n  º  2 (6),  3 (13)  (= Euler).

No. 4.  n  º  5 (7),  2 (9).

No. 5.  n  º  16 (39),  27 (56)  (= Euler).

No. 6.  n  º  5 (7),  7 (8),  8 (9).

No. 8.  n  º  0 (2, 3, 4, 5, 6),   5 (7).

P. 230, no. 35.  n  º  -1 (6, 5, 4, 3, 2).

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 49-50, no. 76: A quantity of eggs being broken, to find how many there were, without remembering the number.  Problem  A-7  (= Bhaskara I).

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.

No. 15, pp. 17-18 & 74.  Problem  A-5  (= Tartaglia).

No. 20, pp. 19 & 76.  Problem  B-9  (= Euler).

Rational Recreations.  1824.  Exer. 11, p. 56.  A-7 (= Bhaskara I).  Answer: 301.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 145-172 & 259-262, nos. 541-645.  He treats the topic of 7.P.1 and 7.P.2 at great and exhausting length, starting with solving  ax = by,  then  x + y = c,  then  ax + y = c,  then  ax = by + 1  and  ax = by + c,  also phrased as  ax º c (mod b).  On pp. 156-158 & 260, nos. 588-597, are Chinese Remainder problems, but with just two moduli, so I will not describe them.

The New Sphinx.  c1840.  No. 45, pp. 24 & 121.  Problem A-7 in verse, with bunch of walnuts.

Magician's Own Book.  1857.  The basket of nuts, pp. 245-246.  Problem  A-7 (= Bhaskara I),  general solution.  = Book of 500 Puzzles, 1859, pp. 59-60.  = Boy's Own Conjuring Book, 1860, p. 217.

Illustrated Boy's Own Treasury.  1860.  Prob. 14, pp. 428 & 431.  Problem  D-5, one solution  (= Baker).

Todhunter.  Algebra, 5th ed.  1870.  Examples XLVI, nos. 22-24, pp. 394 & 596.

No. 22.  n  º  21 (28),  17 (19).

No. 23.  n  º  -1 (3, 5, 7).

No. 24.  n  º  13 (28),  2 (19),  7 (15).  This is an example of calculating the Julian year, but he does not state this.

Wehman.  New Book of 200 Puzzles.  1908.  The basket of nuts, p. 55.  A-7 (= Bhaskara I)  = Magician's Own Book, except that a line of the statement has been dropped by the typesetter, making the problem unintelligible.

Ripley's Puzzles and Games.  1966. 

P. 15, repeated on p. 66.  Problem  C-10,  with solution  14,622,047,999.  Though a correct answer, I cannot fathom how this was obtained.

P. 16.  Problem  B-10,  asking for smallest example.  Answer is  2520,  which is the smallest positive example.

P. 68.  n º 1 (2, ..., 10).  Answer:  2521.

 

          7.P.3.           ARCHIMEDES' CATTLE PROBLEM

 

Archimedes?   Letter to Eratosthenes, c‑250. 

                    Greek text, with commentaries, first published by Gotthold Ephraim Lessing in Beiträge zur Geschichte und Literatur 1 (1773) 421‑???.

               Archimedes.  Opera Omnia.  Ed. by. J. L. Heiberg.  2nd ed., vol. II, Teubner, 1913.  Problema Bovinum, pp. 527‑534.  Heiberg gives the same classical references as Dijksterhuis (below), cites Lessing as the first editor of the problem (from Gud. Graec. 77, f. 415v), gives the later commentators and editors and says the problem also appears in Cod. Paris Gr. 2448, f. 57.  He then gives the Greek and a Latin translation.

               The first edition of Heiberg's edition is the basis of:

               T. L. Heath.  The Works of Archimedes (CUP, 1897)  +  The Method of Archimedes (CUP, 1912);  reprinted in one volume by Dover, 1953.  The Cattle‑Problem, pp. 319‑326, discusses the problem and the attempts at solving it.

               Dijksterhuis, p. 43, says Lessing's article occurs in the Zweiter Beitrag (1773), 2nd ed., Braunschweig, 1773 (??).  = Sämtliche Schriften; ed. by K. Lachmann, vol. IX, p. 285+; 3rd ed., much corrected by F. Muncker, Leipzig, 1897, vol. XIII (or 12??), pp. 99-115.

               English verse version in H. Dorrie; 100 Great Problems of Elementary Mathematics; Dover, 1965, pp. 5‑6.  Dorrie also cites the 19C historians on the question of authenticity.

                    Greek and English in SIHGM II 202-207.  Thomas notes that the epigram is unlikely to have been actually written by Archimedes.  SIHGM I 16-17, is a Scholium to Plato's Charmides 165 E which states "logistic ... treats on the one hand the problem called by Archimedes the cattle-problem" and Thomas gives some of the standard references in a note.

                    English translation in D. H. Fowler; Archimedes Cattle Problem and the Pocket Calculating Machine; Preprint, 1980, plus addenda.  (Based on SIHGM.)

Dijksterhuis.  Archimedes.  Op. cit. in 6.S.1.  P. 398 gives some classical references to the problem:  a scholium to Plato's Charmides;  Heron;  two references in Cicero -- all ??NYS.

T. L. Heath.  Diophantos of Alexandria.  Op. cit. as Diophantos.  1910.  Pp. 121‑124 discusses the problem.

 

J. F. Wurm.  Review of J. G. Hermann's pamphlet:  De archimedis problemate bovino; Leipzig, 1828.  In:  Jahn's Jahrbücher für Philologie und Pädagogik 14 (1830) 195-??.  ??NYS -- cited by Archibald.  Solves the easier interpretation, getting  5,916,837,175,686  cattle in all.

B. Krumbiegel.  Das problema bovinum des Archimedes.  Zeitschrift für Mathematik und Physik -- hist.-litterar. Abt. 25 (1880) 121-136.  ??NYS -- cited by Archibald.  Survey of earlier historical work.

A. Amthor.  ???.  Ibid., pp. 153-171.  ??NYS -- cited by Archibald.  Survey of the mathematics.

H. E. Licks.  Op. cit. in 5.A.  1917.  Art. 54: The cattle problem of Archimedes, pp. 33‑39. discusses the work of Amthor & A. H. Bell (AMM (May 1895), ??NYS), who started the calculation of the answers.  This is an abridgement of an article by Mansfield Merriman in Popular Science Monthly (Nov 1905) ??NYS, o/o, but omitting the author's name, which leads us to believe that H. E. Licks is a pseudonym of Mansfield Merriman.

[R. C. Archibald.]  Topics for club programs -- 14: The cattle problem of Archimedes.  AMM 25 (1918) 411-414.  He gives a detailed history, but some details vary from Dijksterhuis.  He cites Krumbiegel and Amthor as the basic works, Wurm as the first solver with the simpler interpretation and numerous other works.

H. C. Williams, R. A. German & C. R. Zarnke.  Solution of the cattle problem of Archimedes.  Math. Comp. 19 (1965) 671.  Plus comments by D. Shanks, ibid., 686‑687.  Describes full solution, but doesn't print it.

Harry L. Nelson.  A solution to Archimedes' cattle problem  &  Note.  JRM 13 (1980‑81) 162‑176  &  14 (1981‑82) 126.  First printed version of the solution --  206,545  digits.  The note clarifies the formulation of the problem.

 

          7.P.4.           PRESENT OF GEMS

 

Bakhshali MS.  c7C.

Kaye I 40-42 interprets all of the following as examples of the form described for the third problem, which has some connection with 7.R.1.

Kaye I 40-42; III 170-171, f. 3v.  Hoernle, 1886, p. 129; 1888, pp. 33‑34, and Gupta describe it fully.  Three men have  7  horses,  9  ponies,  10  camels.  Each puts in three animals which are then divided equally, which is the same as each giving one to both others.  Then they are equally wealthy.  Solution gives  42, 28, 24  as values of the animals, though  21, 14, 12  is the smallest integral solution.  The original wealths ("capitals") of each merchant are  294, 252, 240  and the final wealth is  262.  (Kaye III 170 has  242  for  240.)

Kaye I 41; III 171, f. 3r says there is a fragment of another problem of this type with values  4, 5, 6.

Kaye I 40-42, III 168-169, ff. 1r-2r, sutra 11.  Gupta says the MS is poor and Kaye has misinterpreted it.  Gupta interprets it as leading to  x1/2 + x2 + x3 + x4 + x5  =  h,  etc.,  where  h  is the price of a jewel.  The  i-th equation then reduces to  T  =  h + xi (i‑1)/i,  where  T  is the total wealth.  Kaye considers it as  T ‑ xi = h ‑ xi/i.  Gupta converts this to a present of gems problem, but since the multipliers are non-integral, it takes a little more work. Answer:  120, 90, 80, 75, 72;  377.  I can't see that Kaye treats this any differently.

Kaye III 170, f. 2v is another example, with three values and diagonal coefficients  ‑7/12, ‑3/4, -5/6  and answer:  924, 836, 798;  1095.

Mahavira.  850.  Chap. VI, v. 162‑166, pp. 137‑138.  Gupta (op. cit. under Bakhshali MS) says the rule given is similar to the Bakhshali rule.

164.  6, 7, 8  giving one each. 

165.  16, 10, 8  giving two each.

Sridhara.  c900.  Gupta (op. cit. under Bakhshali MS) says that Sridhara gives the same rule as the Bakhshali MS, but allowing  n  people.  This rule is quoted in the Kriyākramakari, a 1534 commentary on the Lilavati and Gupta quotes and translates it.  The Kriyākramakari also quotes Mahavira, without attribution.  Gupta cannot locate this rule in Sridhara's extant works.

Bhaskara II.  Lilavati.  1150.  Chap. IV, sect. IV, v. 100.  In Colebrooke, p. 45.  Also in his Bijaganita, chap. IV, v. 111, pp. 195‑196.  8, 10, 100, 5  giving one each to others and all are equal.

 

7.P.5  .         SELLING DIFFERENT AMOUNTS 'AT SAME PRICES' YIELDING THE SAME

 

            NOTATION:  (a, b, c, ...)  means the sellers initially have  a, b, c, ....   They all sell certain amounts at one price, then sell their remnants at a second price so that each receives the same amount.  Western versions give  a, b, c, ...  and sometimes the amount each receives.  The Indian versions give the proportion  a : b : c : ...  (by stating each person's capital, but not the cost price of the items; they invest their capital in the items and then sell them) and the larger price for selling the remnant of the items.  Further, the price for selling the first part of the items is the reciprocal of an integer.  (However the remnant price is sometimes a fraction.)  In both versions, the problem is indeterminate, with a  3  parameter solution set, but scaling or similarity or fixing the yield reduces this to  2.  There are also non‑negativity and integrality conditions.  The Indian version has infinitely many solutions, while the Western version gives a finite number of solutions.  I have recently found a relatively simple way to generate and count the solutions in the Western version, which is basically a generalization of Ozanam's example -- see my paper below.  The article by Glaisher discusses many of these problems.  As in 7.P.1,  (a, b)  solutions means  a  non-negative solutions of which  b  are positive solutions. 

            Versions where the earnings are different:  Ghaligai.

            See Tropfke 651.

 

                        Index of western versions.

 

(10, 20)                                      Abraham

(10, 30)                                      Fibonacci

(12, 32)                                      Fibonacci

(12, 33)                                      Fibonacci

(18, 40)                                      Labosne

(  7, 18, 29)                                 McKay

(  8, 17, 26)                                 Blasius

(10, 12, 15)                                 Labosne

(10, 16, 22)                                 Amusement

(10, 20, 30)                                 Pacioli

(10, 25, 30)                                 Ozanam

(10, 30, 50)                                 Munich 14684, Folkerts, Marliani, Provençale Arithmétique, Pseudo‑dell'Abbaco, Chuquet, HB.XI.22?, Widman, Demaundes Joyous, Tagliente, Ghaligai, Tartaglia, Jackson, Badcock, Rational Recreations, Boy's Own Book, Rowley, Hoffmann

(11, 33, 55)                                 Tartaglia

(16, 48, 80)                                 Tartaglia

(18, 40, 50)                                 Labosne

(19, 25, 27)                                 Williams & Savage

(20, 25, 32)                                 Bachet

(20, 30, 40)                                 Bachet, van Etten, Hunt

(20, 40, 60)                                 Tagliente

(27, 29, 33)                                 Leske, Mittenzwey, Hoffmann, Pearson

(30, 56, 82)                                 Widman

(31, 32, 37)                                 Labosne

(60, 63, 66)                                 Bath

(17, 68, 119, 170)                        Widman

(20, 30, 40, 50, 60)                      Dudeney

(305, 454, 603, 752, 901)             Widman

(20, 40, ..., 140)                          Glaisher, Gould

(10, 20, ..., 90)                            Tartaglia

 

Mahavira.  850.  Chap. VI, v. 102‑110, pp. 113‑116.  He gives a rule which gives one special solution of Sridhara's set of solutions.

V. 103.  Capitals:  2, 8, 36;  remnant price  6.

V. 104.  Capitals:  1½, ½, 2½;  remnant price  6.

V. 105.  Each receives  41;  remnant price  6.  What is the largest of the capitals?  (The other capitals are not determined.)

V. 106.  Each receives  35;  remnant price  4.  (Cf. v. 105.)

V. 108.  Capitals:  ½, ⅓, ¼;  remnant price  6/5.

V. 110.  Capitals:  ½, ⅔, ¾;  remnant price  5/4.

Sridhara.  c900.  V. 60‑62, ex. 76‑77, pp. 44‑49 & 94.  The verses are brief rules, which are expanded by editorial algebra, giving a one parameter family of solutions.

Ex. 76.  Capitals:  1, 3, 5  or  ⅓, ¼, ½;  remnant price  3.

Ex. 77.  Capitals:  3/2, 2, 3, 5;  remnant price  ½.

Bhaskara II.  Bijaganita.  1150.  Chap. 6, v. 170.  In Colebrooke, pp. 242‑244.  Capitals  6, 8, 100;  remnant price  5.  Solution given is  3294, 4392, 54900,  which is one solution from Sridhara's set of solutions, but not by the same method as Mahavira.  The method is not clearly described.  Bhaskara says:  "Example instanced by ancient authors ....  This, which is instanced by ancient writers as an example of a solution resting on unconfirmed ground, has been by some means reduced to equation; and such a supposition introduced, as has brought out a result in an unrestricted case as in a restricted one.  In the like suppositions, when the operation, owing to restriction, disappoints; the answer must by the intelligent be elicited by the exercise of ingenuity."

Fibonacci.  1202.  Pp. 298‑302 (S: 421-423): De duobus hominibus, qui habuerunt poma [On two men who had apples].  He clearly states that there are two forums where the same prices are different. 

(10, 30)  -- he gives  5  solutions, there are  (55, 36).

(12, 32)  -- he gives  6  solutions, there are  (78, 55).

(12, 33)  -- he gives  1  solution, there are  (78, 55).

            Glaisher, pp. 79-107, analyses this text in detail and finds that Fibonacci gives a reasonably general method which would give the  24  positive solutions with smaller price  1  in the first example.  He then considers the amounts received as being fixed.  He then permits the amounts received to differ, one receiving a multiple of what the other receives.  He also considers whether a solution exists for given amounts and prices and how to find solutions with one price given.

Abraham.  Liber Augmenti et Diminutionis.  Early 14C.  ??NYR -- cited by Tropfke 651.  (10, 20).  This has  (55, 36)  solutions.

Munich 14684.  14C.  Prob. XIII, pp. 79‑80.  (10, 30, 50).  Gives the solution with prices  1/7  and  3.  There are  (25, 16)  solutions.

Folkerts.  Aufgabensammlungen.  13-15C.  21 sources for  (10, 30, 50).  Says the problem goes back to Fibonacci, but Fibonacci only has examples with two vendors.

Giovanni Marliani.  Arte giamata arismeticha.  In codex A. II. 39, Biblioteca Universitaria de Genova.  Van Egmond's Catalogue 139 dates it c1417.  Described and partly transcribed by Gino Arrighi; Giuochi aritmetici in un "Abaco" del Quattrocento  Il matematico milanese Giovanni Marliani.  Rendiconti dell'Istituto Lombardo.  Classe di Scienze (A) 99 (1965) 252‑258.  Prob. V: (10, 30, 50).

Provençale Arithmétique.  c1430.  Op. cit. in 7.E.  F. 116v, pp. 62-63.  (10, 30, 50).  Gives the solution with prices  3  and  1/7  and then gives a solution with three prices!

Pseudo-dell'Abbaco.  c1440.  Prob. 101, pp. 85‑87.  (10, 30, 50).  Gives the solution with prices  3  and  1/7.

Chuquet.  1484.  English in FHM 227-228.  Prob. 145.  (10, 30, 50).  Same two solutions as in Provençale Arithmétique.  FHM says it appears in one of Dudeney's books, where he expresses "grave dissatisfaction with the answer".

HB.XI.22.  1488.  P. 54 (Rath 247).  Rath doesn't give the numbers and says it is similar to Munich 14684 and p. 73v of Cod. Vindob. 3029.  Glaisher dates Cod. Vindob. 3029 as c1480.

Johann Widman.  Op. cit. in 7.G.1.  1489.  F. 134v+.  ??NYS -- discussed by Glaisher, pp. 1‑18.  (10, 30, 50)  -- one solution with prices  3  and  1/7.  He then generalises this example to construct single solutions for other examples:  (30, 56, 82),  (17, 68, 119, 170),  (305, 454, 603, 752, 901),  which have  (225, 196),  (45, 35),  (11552, 11400)  solutions respectively.  Glaisher then describes several examples that Widman might have constructed.

Pacioli.  De Viribus.  c1500.  Ff. 119r - 119v. 

LXV. C(apitolo). D dun mercante ch' a .3. factori et atutti ma'da auno mercato con p(er)le.  (10, 20, 30).  Gives the solution with prices  1  and  1/6  and result  5,  selling  4, 2, 0  at the higher price.  There are  (25, 16)  solutions.

Ff. IIIv - IVr.  = Peirani 7.  The Index lists the above as Problem 69 and then gives the following.

            Problem 70: De unaltro mercante ch' pur a .3. factori et mandali a una fiera con varia quantita de perle' et vendano a medesimo pregio et portano acasa tanti denari al patrone uno quanto laltro (Of another merchant who sends three agents to a fair with varying numbers of pearls and they sell them at the same price and they each carry as many pence as the others to the master at home).

            Problem 71: De unaltro vario dali precedenti ch' pur a .3. factori con vari quantita de perle' pregi pari et medesimamente portano al patrone d(enari) pari (Of another variant of the preceding with three agents having various quantities of pearls at equal prices and likewise take as many pence to the master).

            Problem 72: De unaltro mercante ch' ha 4. factori ali quali da quantita varie di perli ch' amedisimi pregi le vendino et denari equalmente portino (Of another merchant who has 4 agents to whom he gives various numbers of pearls which they sell at the same prices and receive equal money).

            Problem 73: De un altro ch' pur a .4. factori con quanti(ta) varie di perle apari pregi et pari danari reportano a casa vario dali precedenti (Of another who sends 4 agents with varying numbers of pearls and they report back to the house the same prices and the same money, variation of the preceding).

Anon.  Demandes joyeuses en manière de quodlibets.  End of 15C.  ??NYS.  Selected and translated as: The Demaundes Joyous.  Wynken de Worde, London, 1511.  [The French had 87 demandes, but the English has 54.  This is the oldest riddle collection printed in England, surviving in a single example in Cambridge Univ. Library.  Often attributed to de Worde.  Santi 9 uses Yoyous and Wynkyn and list de Worde as author.]  Facsimile with transcription and commentary by John Wardroper, Gordon Fraser Gallery, London, 1971, reprinted 1976.  Prob. 50, pp. 6 of the facsimile, 26-27 of the transcription.  (10, 30, 50)  apples.  One solution with prices  3  and  1/7.

Blasius.  1513.  F. F.iii.r: Decimaquarta regula.  Selling eggs --  (8, 17, 26).   There are  (16, 9)  solutions.  He give one with prices  2  and  1/5  and each sold as many batches of  5  as possible.  Discussed by Glaisher.

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 115, ff. 57r-57v.  Women selling eggs --  (10, 30, 50).  One solution with prices  1/7  and  3.  Glaisher, below and in op. cit. in 7.G.1, cites  Hieronymus Tagliente and says the 1515 & 1527 editions give  (10, 30, 50)  with solution at prices  1/7  &  3,  and the 1525 ed. has  (20, 40, 60)  with solution at prices  3  and  1/7.  The latter has  (100, 81)  solutions.

Ghaligai.  Practica D'Arithmetica.  1521. 

Prob. 21, ff. 65r-65v.  (10, 30, 50).  One solution, with prices  1/7  and  3.  Glaisher says it is p. 66 (misprinted 64) in the 1548 ed.  (H&S 53 gives Italian from 1552 ed., but no solution.)  Ghaligai says the problem was known to Benedetto (da Firenze, flourished 1470s) and Ghaligai's teacher Giovanni del Sodo, as a problem outside of any rule, and Ghaligai labels it  "Ragione apostata" (Exceptional problem).

Prob. 22, f. 65v.  (10, 50)  with first making twice the second.  Solution with prices  13  and  1/7.

Tartaglia.  General Trattato, 1556, art. 136‑139, pp. 256r‑256v.

136.  (10, 30, 50)  yielding  10  each.  He gives the solution with prices  3  and  1/7.  There are  (25, 16)  solutions.

137.  (11, 33, 55)  yielding  11  each.  He gives the solution with prices  2  and  1/6.  There are  (30, 20)  solutions.

138.  (16, 48, 80)  yielding  16.  He gives the solution with prices  3  and  1/11.  There are  (64, 49)  solutions.

139.  (10, 20, ..., 90),  yielding  100.  He gives the unique positive solution.  There are  (3, 1)  solutions.

Bachet.  Problemes.  1612.  Prob. XXI, 1612: 106-115.  Prob. XXIV, 1624: 178-186;  1884: 122‑126.  He gives a new "general and infallible rule", which is fairly general -- Glaisher says it produces a selection of the solutions.  He applies the idea to  (20, 30, 40),  exhibiting  4  solutions.  It has  (100, 81)  solutions.  Cf. van Etten.  1612 also does  (20, 25, 32).

                    In the 5th ed., the general material is dropped and replaced by some vague algebra.  Labosne gives two solutions for  (18, 40),  but one of them uses fractions.  It has  (171, 36)  solutions.  He then considers  (18, 40, 50)  and gives one fractional solution -- there are  (3, 1)  solutions.  He then makes some discussion of  (10, 12, 15)  (which has  (7, 4)  solutions) and  (31, 32, 37)  (which has  (70, 60)  solutions).

van Etten.  1624.  Prob. 69 (62), pp. 64‑65 (90‑91).  (20, 30, 40).  Gives one solution with prices  3  and  1.  There are  (100, 81)  solutions.  Cf. Bachet.  Henrion's 1630 Notte, p. 22, states that Bachet found many other solutions and gives a solution with prices  2  &  7. 

Hunt.  1651.  Pp 282-283: Of three women that sold apples.  (20, 30, 40).  Gives one solution with prices  1  and  3.

Ozanam.  1694.  Prob. 24, 1696: 77-80;  1708: 68‑70.  Prob. 28, 1725: 201‑210.  Prob. 12, 1778: 199-204;  1803: 196-201;  1814: 170-174;  1840: 88-90.  (10, 25, 30).  Glaisher describes the material in the 1696 ed. and says that Ozanam first considers the same general form that Bachet considered and then applies it to the example.  Glaisher indicates that Ozanam's and Bachet's methods are essentially the same, but Ozanam certainly gets all solutions, while I am not sure that Bachet can do so.  1696 gives two solutions at prices  7  and  2  and at prices  6  and  1.  1725 et seq. gives a general method and finds all  10  solutions of the specific problem.  (Glaisher notes that the Remarques on pp. 203‑210 are new to the 1723 ed.  They give an algebraic form of the solution.)  1725 refers to the second part of Arithmétique Universelle, p. 456 (more specifically identified as by M. de Lagny [1660-1734] in 1778 et seq.), where  6  solutions are found.  1725 says there are  10  solutions and 1778 says de Lagny is mistaken -- but in fact, there are  (10, 6)  solutions and de Lagny probably meant just the positive ones.  1778 drops the preliminary general general form.

Amusement for Winter Evenings.  A New and Improved Hocus Pocus; or Art of Legerdemain: Explaining in a Clear and Comprehensive Manner Those Apparently Wonderful and Surprising Tricks That are performed by Slight of Hand and Manual Dexterity: Including Several Curious Philosophical Experiments.  M. C. Springsguth, London, nd [c1800 -- HPL], 36pp.  Pp. 22-23: Of three sisters.  (10, 16, 22)  sold at  7 a penny  and then a penny apiece, i.e. prices  1/7  and  1,  each earning  4.

Bestelmeier.  1801.  Item 718: Das Eyerverkauf.  Three women sell different numbers of eggs and make the same.  Further details not given.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 14, pp. 17 & 74.  (10, 30, 50).  Gives the solution with prices  3  and  1/7.

John Badcock.  Domestic Amusements, or Philosophical Recreations.  Op. cit. in 6.BH.  [1823].  P. 155, no. 192: Trick in reasoning.  (10, 30, 50)  -- gives the solution with prices  3  and  1/7.

Rational Recreations.  1824.  Exer. 21, p. 98.  (10, 30, 50)  eggs.

Boy's Own Book.  1843 (Paris): 341.  "Three country-women and eggs."  (10, 30, 50)  -- gives the solution with prices  3  and  1/7.  = Boy's Treasury, 1844, p. 299.  = de Savigny, 1846, p. 289: Les trois paysannes et les œufs.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-21, pp. 254 & 395-396: Die Blumenmädchen.  (27, 29, 33).  This has  (117, 100)  solutions -- she gives one with prices    and  1,  each earning  13.

Hugh Rowley.  More Puniana; or, Thoughts Wise and Other-Why's.  Chatto & Windus, London, 1875.  P. 179.  Man with three daughters and lots of apples --  (10, 30, 50).  Solution with prices  3  and  1/7.

Mittenzwey.  1880.  Prob. 122, pp. 25 & 76;  1895?: 140, pp. 29 & 79;  1917: 140, pp. 27 & 77.  Three sisters selling bunches of violets.  (27, 29, 33).  This has  (117, 100)  solutions -- he gives one with prices    and  1,  each earning  13.  = Leske.

Hoffmann.  1893.  Chap. IV, pp. 160 & 215-216 = Hoffmann-Hordern, p. 140.

No. 65: The three market‑women.  (27, 29, 33).  c= Leske.

No. 66: The farmer and his three daughters.  (10, 30, 50).  He gives the solution with prices  3  and  1/7.

Dudeney??  Breakfast Table Problems No. 328:  "How were the oranges sold".  Daily Mail (27  &  28 Jan 1905) both p. 7.  (20, 30, 40, 50, 60).  Gives the solution with prices  1  and  1/11.  There are  (45, 36)  solutions.

Pearson.  1907.  Part II, no. 37, pp. 121 & 199.  (27, 29, 33)  c= Leske.

J. W. L. Glaisher.  On certain puzzle-questions occurring in early arithmetical writings and the general partition problems with which they are connected.  Messenger of Mathematics 53 (1923-24) 1-131.  Discusses the versions in Blasius, Widman, Tagliente  and attempts to explain the methods used.  Unfortunately he is rather prolix and I often get lost in the many examples and special cases, but he seems to have general solutions.  On p. 12, he mentions that Tagliente's problem could be extended to  (20, 40, ..., 140)  -- cf. Gould below.  On p. 77, he says more results will appear in a later paper -- check index of Messenger??

A. A. Krishnaswami Ayyangar.  A classical Indian puzzle-problem.  J. Indian Math. Soc. 15 (1923-24) 214-223.  Responding to Glaisher, he analyses the Indian version, obtaining a simple complete solution with two parameters having an infinite range and a third parameter being bounded.  I found this a bit confusing since he sometimes uses price to mean the number of items per unit cost.  He says many of the solutions are not in Glaisher's system given on p. 19, but I can't tell if Glaisher intends this to be a complete solution.

Rupert T. Gould.  The Stargazer Talks.  Geoffrey Bles, London, 1944.  A Few Puzzles -- write up of a BBC talk on 10 Jan 1939, pp. 106-113.  Seven applewomen with  20, 40, 60, 80, 100, 120, 140.  One solution with prices  3  and  1/7  -- cf. Glaisher.  There are  (27, 21)  solutions.

McKay.  At Home Tonight.  1940.  Prob. 16: Extraordinary sales, pp. 65 & 81.  (7, 18, 29)  eggs.  This has  (12, 6)  solutions.  He asks for a solution where all make  10d.  He gives one solution with prices  1/4  and  3  and selling as many eggs in batches of  4  as possible.

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.  No. 75: A falling market, pp. 43 & 128.  Cauliflowers  (19, 25, 27)  with each making  85d,  both prices being integral and each sells some at the lower price, not to be less than  2d.  Actually there is only one solution with integral prices and each making  85d.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 45: Potatoes for sale, pp. 19 & 47.  (60, 63, 66).  This has  (900, 841)  solutions.  He wants solutions where all make  9/6 = 114d.  There are two solutions with integral prices, but he gives the solution with prices  2d  and  12/7 d,  i.e.  7 for 1s,  but he doesn't require maximum numbers of batches of  7  to be sold.

David Singmaster.  Some diophantine recreations.  In:  Papers Presented to Martin Gardner on the occasion of the opening of the exhibition: Puzzles: Beyond the Borders of the Mind at the Atlanta International Museum of Art and Design; ed. by Scott Kim, 16 Jan 1993, pp. 343-356  AND  The Mathemagician and Pied Piper  A Collection in Tribute to Martin Gardner;  ed. by Elwyn R. Berlekamp & Tom Rodgers;.  A. K. Peters, Natick, Massachusetts, 1999, HB, pp. 219-235.  Sketches some history, gives complete solutions for the Western and Indian cases (filling a gap in Ayyangar) and finds a new simple formula for the number of solutions in the Western case.

 

          7.P.6. CONJUNCTION OF PLANETS, ETC.

 

            See Tropfke 642. 

            Some overtaking problems in 10.A take place on a circular track and are related to or even identical to these problems.  In particular, if two persons start around an island of circumference  D, from the same point and in the same direction at rates  a, b,  this is the same as  O-(a, b; D)  of Section 10.A.  This is easily adapted to dealing with different starting points and going in opposite directions (which gives a meeting problem).  Clock problems, 10.R, are also related to these.

 

Sun Zi.  Sun Zi Suan Ching.  Op. cit. in 7.P.2.  4C.  ??NYS.  Sisters come home every  5, 4, 3  days, when do they all come together?.  (Mikami 33 gives English.)

Zhang Qiujian.  Zhang Qiujian Suan Jing.  Op. cit. in 7.E.  468,  ??NYS -- translated on p. 139 of:  Shen Kangsheng; Mutual‑subtraction algorithm and its applications in ancient China; HM 15 (1988) 135‑147.  Circular route of length 325.  Three persons start with speeds  150, 120, 90.  When do they all meet at the start?

Brahmagupta.  Brahma‑sphuta‑siddhanta.  628.  Several of the sections of Chap. XVIII discuss astronomical versions of this problem, but with complicated values and unclear exposition.

Gherardi.  Libro di ragioni.  1328.  P. 47.  Two men start going in a circuit.  One can do it in  4  days, the other in    days.  When do they meet again?  Same as  O-(1/4, 1/5½),  D = 1,  in the notation of Section 10.A.

AR.  c1450.  Prob. 148, pp. 72, 164‑165, 214.  Though titled 'De planetis' and described as conjunction by Vogel, this is really just an overtaking problem -- see 10.A.

H&S 74‑75 says sun and moon problems are in van der Hoecke and Trenchant (1556).

Cardan.  Practica Arithmetice.  1539.  Chap. 66, sections 20-24, ff. CC.vii.v - DD.i.r (pp. 142‑143).  Several versions concerning conjunctions of planets, including irrational ratios and three planets.  Examples with periods:  Ö7, Ö5;  Ö18, Ö30;  Ö8, Ö20;  1000, 999;  and  Ö5, Ö4, Ö3.  In section 23, he gives periods of Saturn and Jupiter as  30  &  12  years and periods of Jupiter and Mars as  144  &  23  months.  (H&S 75 gives English and some of the Latin.)

Cardan.  Opus Novum de Proportionibus Numerorum.  Henricpetrina, Basil, 1570, ??NYS.  = Opera Omnia, vol. IV, pp. 482-486.  General discussion and examples, e.g. with periods  2, 3, 7.

Wells.  1698. 

No. 107, pp. 206-207.  Two clock hands start together.  One circles in  1  day, the other in  3  days.  When do they meet?  Also does the general problem with periods  b, c,  b < c,  getting  bc/(c-b).

No. 108, p. 207.  Applies above to sun and moon to get synodic month.

Vyse.  Tutor's Guide.  1771?  Prob. 20, 1793: p. 79; 1799: p. 85 & Key p. 110-111.  Island  73  in circumference;  three persons set out in the same direction at speeds  5, 8, 10.  When do they all meet again?

Bonnycastle.  Algebra.  1782.  P. 86, no. 23.  Identical to Vyse.

Pike.  Arithmetic.  1788.  P. 353, no. 31.  Island  50  in circumference.  Three walkers start in the same direction at speeds  7, 8, 9.  When and where do they meet again?  = D. Adams; Scholar's Arithmetic; 1801, p. 208, no. 66.

Hutton.  A Course of Mathematics.  1798?  Prob. 37,  1833: 223;  1857: 227.  Identical to Vyse.

Kaida Anmuyo.  c1800.  Problem given on pp. 139‑140 of Shen Kangsheng, loc. cit. under Zhang Qiujian above.  Assume  365¼  degrees in a circle.  Five stars are in a line and travel at speeds of  28 13/16,  19 1/4,  13 5/12,  11 1/7,  2 7/9  degrees per day.  When do they meet at the starting point again?

D. Adams.  New Arithmetic.  1835.  P. 244.

No. 84.  Island of circumference  20  and three travellers set out from the same point in the same direction at rates  2, 4, 6.  When do they meet?

No. 85.  Same with just the travellers of rates  2, 6.  = O-(2, 6; 20)

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  3rd ed., 1850.  Op. cit. in 7.H.  P. 360, no. 48.  Three walkers start to circle an island of circumference  73,  at rates  6, 10, 16.  When do they meet again?

James B. Thomson.  Higher Arithmetic; or the Science and Application of Numbers; ....  Designed for Advanced Classes in Schools and Academies.  120th ed., Ivison, Phinney & Co, New York, (and nine copublishers), 1862.  Prob. 94, p. 308 & 422.  Same as Anon: Treatise.

Daniel W. Fish, ed.  The Progressive Higher Arithmetic, for Schools, Academies, and Mercantile Colleges.  Forming a Complete Treatise of Arithmetical Science, and its Commercial and Business Applications.  Ivison, Blakeman, Taylor & Co., NY, nd [but prefaces give: 1860; Improved Edition, 1875].  P. 418, no. 64.  Island  120  in circumference.  Seven men start walking around it from the same point at speeds  5, 25/4, 22/3, 33/4, 19/2, 41/4, 45/5  per day.  When are they all together again?  Answer: after 1440 days.

Mittenzwey.  1880.  Prob. 117, pp. 23-24 & 76;  1895?: 135, pp. 27-28 & 79;  1917: 135, pp. 25-26 & 76.  Seven colleagues return to a forest inn every  1, 2, 3, 4, 5, 6, 7  days.  when will they all return again?

Lemon.  1890.  The Maltese cross, no. 483, pp. 63 & 115.  Walkers complete  6, 9, 12, 15  circuits per hour -- when are they all again at start?

Hoffmann.  1893.  Chap. IV.

No. 38: When will they get it?, pp. 152 & 202 = Hoffmann-Hordern, p. 127.  Guests come to restaurant with periods  1, 2, ..., 7  days.  When do they all meet again?

No. 47: The walking match, pp. 154 & 207 = Hoffmann-Hordern, p. 130.  Four men walk around a track of length  1  with speeds  5, 4, 3, 2  per hour.  When do all meet at start again?

Clark.  Mental Nuts.  1897, no. 51.  When will we three meet again.  Three bicycle riders can ride around a one mile track in  2 1/2,  2 3/5,  3 1/4  minutes.  If they all start together, when will they all meet again at the starting point?

Dudeney.  "The Captain" puzzle corner.  The Captain 3:2 (May 1900) 97 & 179  &  3:4 (Jul 1900) 303.  No. 2: The seven money boxes.  Boy puts a penny in  i-th box on  i-th day, where day  1  is  1 Jan 1900.  When he has to put in seven pennies, he will then open them all up.  When is this and how much will he have?  Answer:  420 days  =  24 Feb 1901  and  £4 10s 9d.

Depew.  Cokesbury Game Book.  1939.  Bicycle racers, p. 221.  One can travel around the track in  6  minutes, the other in  9  minutes.  When are they together again?

 

          7.P.7. ROBBING AND RESTORING

 

            Men keep money together and divide it into amounts   x1,  x2,  ...   -- usually by robbing the common fund.  They put fractions  aixi  into a pool and divide the pool in proportion  b1 : b2 : ....  They then have money in the proportion  c1 : c2 : ...,  or actual amounts  d1, d2, ....

            I use  A  for  (a1, a2, a3),  etc.

 

Kurt Vogel.  Ein unbestimmtes Problem al-Karağī in Rechenbüchern des Abendlands.  Sudhoffs Archiv 61 (1977) 66-74.  Gives the history of this problem, particularly the transmission to Fibonacci via John of Palermo and the different methods of solving it -- Fibonacci gives three methods.  He mentions all the entries below except Gherardi and Calandri, which had not been published when he wrote, and Pacioli, which is sad because Pacioli is not very clear!

Jacques Sesiano, op. cit. under 7.R, 1985, discusses this problem along with the problems in 7.R.  He calls it "The disloyal partners".  He cites Abū Kāmil's Algebra, ff. 100r-101r, ??NYS, and says it is solved two ways there.  Martin Levey's 1966 edition of the Algebra does not give the problem and states that the unique Arabic MS ends on f. 67r, and uses a 'better' Hebrew text.  The Arabic MS actually continues and the third part of the book, ff. 79r-111r, was treated by Schub & Levey in 1968 & 1970.  Sesiano; Les méthodes d'analyse indéterminée chez Abū Kāmil; Centaurus 21 (1977) 89-105 is scathing about the work of Levey and of Schub & Levey, saying the Hebrew MS is third-rate, and the translators have made serious mathematical and philological errors.  Sesiano studies some of Abu Kamil's problems in this article, but unfortunately the problem of this section is not among them.

 

al‑Karkhi.  c1010.  Sect I, no. 45 & 47;  sect III, no. 6;  pp. 80‑81 & 90.

I‑45: Two men have  d1 = 40  and  d2 = 60.  From the common sum, they take  x1  and  x2  =  100 ‑ x1.  The first gives  a1 = 1/4  of what he took to the second and the second gives  a2 = 1/5  of what he took to the first.  Then they have correct amounts.  Answer:  (400, 700)/11.

I‑47: Usual version with two people.  A = (⅓, ¼),   B = (1, 2),   D = (30, 70).  Answer:  (156, 344)/5.

III‑6: Usual version:  A = (1/2, 1/3, 1/6),   B = (1, 1, 1),   C = (3, 2, 1).  Answer:  33, 13, 1.

Fibonacci.  Flos.  c1225.  Pp. 234-236: De tribus hominibus pecuniam comunem habentibus.  = Fibonacci, below, pp. 293-294.  = al‑Karkhi's III‑6.  Problem II in Picutti, pp. 310‑312. 

                    Cantor; op. cit. under Fibonacci in 7.P.1; 1863; p. 345, discusses the contest between Fibonacci and John of Palermo before Frederick II at Pisa in 1225 (or 1226) and says this problem is the third and last of the contest problems and this is how I read the text.  However Picutti has this as the second problem.  Licks, op. cit. in 5.A, says it was problem 5 in the contest.  Vogel doesn't mention which problem it was.

                    Leonardo says he later found three further methods of solution, which are "in libro uestro, quem de numero composui, patenter inserui".  Leonardo is here addressing the Emperor, so Vogel interprets 'libro uestro' as a book dedicated to the Emperor.  Vogel interprets this as referring to the material in Liber Abbaci, so I have now dated the next entry as 1228 rather than 1202, although there is no mention of the contest or the Emperor in Liber Abbaci.

                    The solution here is different than below and Vogel calls this Fibonacci's third method, the shortest and cleverest, and which Fibonacci described as "exceedingly beautiful"'.  Vogel notes the remarkable hybrid notations:  XXX3 for 33;  XXIII 1/1 for 23½;  X 1/1 for 10½  in this 15C MS.

Fibonacci.  1228 -- see above entry.  Pp. 293‑297 (S: 415-420).  Several versions.  He often notes that the values  xi  can be multiplied through by any value.

Pp. 293‑294 (S: 415-417).  A = (1/2, 1/3, 1/6),   B = (1, 1, 1),   C = (3, 2, 1).  Answer:  33, 13, 1.  (= Al‑Karkhi III‑6.)  Vogel calls this Fibonacci's first method and notes a minor variation, which Fibonacci may have intended as another method.  Vogel notes some typographical errors.

P. 294 (S: 417).  Same  A,  with   B = (3, 2, 1) = C.  Answer:  30, 15, 6.

P. 294-295 (S: 417-418.  A = (1/3, 1/4, 1/5),   B = (1, 1, 1),   C = (3, 2, 1).  Answer:  177, 92, 25.

P. 295 (S: 418).  Same  A,  with  B = (3, 2, 1) = C.  Answer:  162, 96, 45.

Pp. 295‑296 (S: 418-419).  Same  A,  with  B = (5, 4, 1),   C = (3, 2, 1).  Answer:  543, 296, 175.  In this problem he computes  360 - 360  as  0  and  0/2  as  0.

Pp. 296‑297 (S: 419-420).  A = (1/2, 1/3, 1/6),   B = (1, 1, 1),  C = (5, 4, 1).  Answer:  326, 174, ‑30.  "Quare hec questio non potest solui, nisi solvatur cum aliqua propria pecunia tercii hominis ..." [Therefore this problem can only be solved with some smaller amount of money for the third man ...]  He treats  ‑30  as a debt of the fund to the third man who steals nothing, thereby losing  30.  See Sesiano.

P. 297 (S: 420).  Four men.  A = (1/2, 1/3, 1/4, 1/5),   B = (1, 1, 1, 1),   C = (10, 9, 6, 5).  Answer:  1034, 666, 300, 190.

Pp. 335‑336 (S: 468-469).  Problem on pp. 293‑294 done by false position.  Vogel calls this Fibonacci's second method, but there is a typographical error citing p. 235f.

Gherardi. Libro di ragioni.  1328.  Pp. 54‑56.  A = (½, ⅓, ¼),   B = (1,1,1),   C = (6, 4, 3).  Answer is  240 : 93 : 44.  He gives  4/29  of these values by starting with  x1 + x2 + x3 = 52.

Lucca 1754.  c1330.  Ff. 60v‑61r, pp. 138‑139.  (= al‑Karkhi III‑6.)  Gives some explanation, but Vogel says only the beginning makes sense.

Columbia Algorism.  c1350.  Prob. 5, pp. 34‑35.  (= al‑Karkhi III‑6.)  Gives only the answer with no explanation.  Vogel's Introduction, p. 22, sketches the history.

Pacioli.  Summa.  1494.  Gives a number of versions.

F. 105r, prob. 11.  A = (½, ⅓),   B = (1, 1),   D = (100, 100).  Solution is  (600, 800)/7.

F. 157v, prob. 77.  = al-Karkhi III-6 with total  12.  He then sketches the solution with total  60, but immediately has wrong values.  He also seems to have changed some of the parameters -- his answers don't add up to  60,  and he gives final values in the ratio  6 : 4 : 3.

Ff. 157v-158r, prob. 78.  A = (1/4, 1/6, 1/5),   B = (9, 7, 4),   C = (6, 4, 3),  with total  12.  His working is correct until he has  2 158/1651  instead of  2 958/1651  for his unknown and then the erroneous value is used in the final steps.  Answer should be  (9639, 5526, 4650)/1651.

F. 158v, prob. 82' (the second problem numbered 82).  Seems to be a discussion of modification of prob. 78 to have   B = (1, 1, 1),   C = (1/2, 3/10, 1/7),  but he never uses the values in  C  and winds up giving the answers of prob. 77 for total of  564,  namely  396, 156, 12.

F. 158v, prob. 83.  A = (⅓, ¼),   B = (1, 1),   D = (15, 15).

Calandri, Raccolta.  c1495.  Same as Fibonacci, pp. 296‑297.  Calandri simply says it is "insolubile".

Tonstall.  De Arte Supputandi.  1522.  Pp. 244-245.  Same as Pacioli, prob. 11.

Cardan.  Practica Arithmetice.  1539.  Chap. 66. 

Section 90, ff. GG.vii.v - GG.viii.v (pp. 164-165).  Same as Fibonacci, p. 295, first problem.  Answer:  354, 184, 50.

Section 91, ff. GG.viii.v - HH.i.v (pp. 165-166).  A somewhat similar situation, where the first two take money leaving the third with  5.  Friend says the first is to give  10  and    of what he has left to the second and the second is then to give  7  and  ¼  of what he has left to the third to make  C = (3, 2, 1).  This is determinate.  Answer is  x = 172,  y = 39  and the total sum is  216.

Recorde.  Second Part.  1552.  Pp. 330-335: A question of partners, the ninth example.  A = (¾, ⅓)  or  (⅓, ¾),  the person with the larger  xi  to give back  ¾.  B = (1, 1),  D = (180, 120).  Solution:  (1680, 1620)/11.  The second person took the more money.

Buteo.  Logistica.  1559. 

Prob. 7, pp. 335-336.  A = (½, ¼),   B = (1, 1),   C = (1, 1).  He assumes total is  500,  then answer is  300, 200.

Prob. 8, pp. 336-337.  A = (½, ⅓, ¼),   B = (1, 1, 1),   D = (116, 116, 116).  Answer:  144, 108, 96.

Vogel says that Clavius; Epitome Arithmeticae; Rome, 1595, pp. 249-252, gives a simple example with two persons and that then the problem vanishes from the literature.

 

          7.Q.   BLIND ABBESS AND HER NUNS -- REARRANGEMENT ALONG SIDES OF A  3 x 3  SQUARE TO CONSERVE SIDE TOTALS

 

            This is a kind of magic figure, except that here we generally have repeated values. 

            There are three trick versions of 6.AO which might be classified here or in 7.Q.2. 

            (12,   4, 5) -- Trick version of a hollow  3 x 3  square with doubled corners, as in 7.Q:  Family Friend (1858),  Secret Out,  Illustrated Boy's Own Treasury.

            Van Etten and Mittenzwey are the only inverse examples, where the total number remains fixed but the number on each side changes.

 

Shihâbaddîn Abû’l‑‘Abbâs Ahmad [the  h  should have an underdot] ibn Yahya [the  h  should have an underdot] ibn Abî Hajala [the  H  should have an underdot] at‑Tilimsâni alH‑anbalî [the  H  should have an underdot].  Kitâb ’anmûdhaj al‑qitâl fi la‘b ash‑shaţranj [NOTE:  ţ  denotes a  t  with an underdot] (Book of the examples of warfare in the game of chess).  Copied by Muhammed ibn ‘Ali ibn Muhammed al‑Arzagî in 1446.  This is the second of Dr. Lee's MSS, described in 5.F.1, denoted Man. by Murray.  Described by Bland and Forbes, loc. cit. in 5.F.1 under Persian MS 211, and by Murray 207‑219.

                    Murray 280 says No. 46‑49 give the problems of arranging  32, 36, 40, 44  men along the walls and corners so the total along each edge is  12.

Pacioli.  De Viribus.  c1500. 

Ff. 117v - 118r.  .C(apitolo). LXXI. D(e). un quadro quale  .3.  per ogni verso Diametro elati et giontovi .3. doventa .4. per og' verso (Of a square which has 3 on every diagonal and side and adding 3 has 4 in every direction).  Says to start with one object in each cell of a  3 x 3  array, which has three on each line, then add one to the cells along a diagonal to get  4  in each line.  This gets  6  on that diagonal however, but he ignores that. 

Ff. IVv - Vr.  = Peirani 8.  The Index gives the above as Problem 88 and continues with the following.  Problem 89: De uno abate ch' tolse aguardar certo monasterio de monache in levante contandole sera e matina per ogni verso tante et pur daloro schernito desperato la bandona (Of an abbot who tries to guard a certain monastery of monks in the Levant by counting evening and morning the same on each side and how the sneering desperados abandoned it ??).

Hunt.  1631 (1651).  Pp. 264-266 (256-258).  General and guards.  24  guards become  20  then  28.

van Etten.  English ed., 1653, prob. 72: Of the game of square formes, pp. 124‑125.  24  men on sides of a fort, becoming  28  and  20.  Discusses case of  12  men making  3,  4  or  5  on a side.

Anon.  Schau‑Platz der Betrieger: Entworfen in vielen List‑ und Lustigen Welt‑Händeln.  Hamburg & Frankfurt, 1687, pp. 543‑545.  ??NYS  (A&N, p. 5.)

Ozanam.  1694.  Prob. 1, 1696: 1-2;  1708: 1‑2;  1725: 1-3.  Prob. 20, 1778: 172-174;  1803: 172-174;  1814: 151-153.  Prob. 19, 1840: 77‑78.  Blind abbess and  24  nuns with  9  on a side.  1696 gives three arrangements with  24, 28, 20  on a side.  1725 adds another arrangement with  32.  1778 says Ozanam has presented this in a rather indecent manner to excite the curiosity of his readers and adds arrangements with  36  and  18  on a side.  The last has  5  and  4  in the corners and none in the side cells, but can be done in other ways.  1803 drops the 'indecent' reference.

Dilworth.  Schoolmaster's Assistant.  1743.  Part IV: Questions: A short Collection of pleasant and diverting Questions, p. 168.  Problem 1.  General stationing guards around a castle, wanting  18  on a side, starting with  48  men and changing to  56,  then  40.

Les Amusemens.  1749.  Prob. 11, p. 131: Les rangs de Neuf.  Wine merchant with  32  bottles,  9  on a side, reduced to  28, 24, 20.

Catel.  Kunst-Cabinet.  Vol. 2, 1793.  Die Nonnenlist (The nuns' strategem), pp. 15-16 & fig. 251 on plate XII.  The diagram shows the eight outside cells with  5  spots in the form of a  5  on a die and one spot in the centre.  However, the text says there are  25  cones or pieces and one must read the instructions to learn the game.  The number of pieces seems peculiar and I'm not entirely sure this is our problem, despite its name.

Bestelmeier.  1801.  Item 191: Die Nonnenlist.  Picture is an obscure copy of Catel.  Text is copied from part of Catel, but says there are  15  pieces!

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 10-11, no. 20: The blind abbess and her nuns.  9  on a side, starts with  24  and changes to  28,  then  20.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.

No. 22, pp. 6 & 57.  18  men on a side of a castle, total =  48, 56, 40.

No. 29, pp. 7 & 58.  Blind abbess and nuns,  9  on a side, total ranging from  20  to  32.

No. 32, pp. 8 & 59.  Wine merchant and casks,  9  on a side, total of  32,  diminished to  20.

No. 36, pp. 9 & 59-60.  Blind abbess and nuns, equal numbers on each side, then with four extras, then with four gone away.  Solution starts with  24.

John Badcock.  Domestic Amusements, or Philosophical Recreations.  Op. cit. in 6.BH.  [1823].  Pp. 156-157, no. 194: Dishonest contrivance.  32  sheep, with  12  on a side, reduced to  28.

Rational Recreations.  1824.  Exer. 18, pp. 94-95: The convent.  24 increased to 28 then reduced to 20.

Manuel des Sorciers.  1825.  Pp. 75-78, art. 38.  ??NX  Blind abbess.  Gets totals of  32,  28,  24,  20.

Endless Amusement II.  1826? 

Pp. 108-109: Curious arithmetical question.  9  on each side, changed from a total of  24  to  20, 28, 32.

Prob. 32, pp. 209-210.  Uses a novel figure -- take a  3 x 3  square and draw the  1 x 1  cells in each corner, then diagonally connect the interior vertices of these to form an  X  in the central cell.  This gives  8  cells -- the four squares at the corners and four pentagonal shapes along the edges.  (See Wehman, p. 22.)  15  on each side, beginning with  40,  reduced to  36.  = New Sphinx, c1840, p. 139.

The Boy's Own Book.   The wine merchant and his clerk.  1828: 412;  1828‑2: 418;  1829 (US): 211;  1855: 565;  1868: 669.  32  bottles, reducing to  20.

The Riddler.  1835.  The wine merchant and his clerk, pp. 4-5.  Identical to Boy's Own Book.

Crambrook.  1843.  P. 10, no. 23: The Blind Abbess and her Nuns, a laughable trick.

Magician's Own Book.  1857.

The square of Gotham, pp. 229-230.  24  scholars,  9  on a side, changing to  20, 28, 32.  = Boy's Own Conjuring Book.

Prob. 24: The nuns, pp. 274 & 297.  24  nuns.  =  Book of 500 Puzzles, prob. 24.  = Boy's Own Conjuring Book, prob. 23.  c= Illustrated Boy's Own Treasury, prob. 29.

Landells.  Boy's Own Toy-Maker.  1858.  Pp. 149-150.  c= Magician's Own Book, prob. 24.

The Sociable.  1858.  Prob. 21: The blind abbot and the monks, pp. 292-294 & 309.  24  monks,  9  on a side, changed to  20, 28, 32, 36, 18.  = Book of 500 Puzzles, prob. 21. 

Book of 500 Puzzles.  1859. 

Prob. 21: The blind abbot and the monks, pp. 10-12 & 27.  As in The Sociable.

Prob. 24: The nuns, pp. 88 & 111.  Identical to Magician's Own Book, prob. 24.

Boy's Own Conjuring book.  1860.

The square of Gotham, pp. 199‑200.  Identical to Magician's Own Book.

Prob. 23: The nuns, pp. 236 & 261.  Identical to Magician's Own Book, prob. 24.

Illustrated Boy's Own Treasury.  1860.  Prob. 29, pp. 429 & 434.  Very similar to Magician's Own Book, prob. 24.

Vinot.  1860.  Art. LVIII: Les étrennes du Commissaire, pp. 75-76.  140 bottles of wine arranged  1, 34, 1  on each side.  Clerk steals four bottles sixteen times, reducing to  17, 2, 17.

The Secret Out (UK).  c1860.  Both of the following are presented with cards.

The unfaithful knave, pp. 4-5.  32  wine bottles,  9  on a side, reduced to  28, 24, 20.

The blind abbot and his monks, pp. 5-6.  24  monks,  9  on a side, changing to  20, 28, 32, 18.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?

Prob. 585-4, pp. 296 & 409: Der Kasten des Juweilers.  32  rings reduced to  28.

Prob. 585-6, pp. 296 & 409: Des Müllers Säcke.  32  sacks reduced to  28.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. VI, 1884: 189-191.  60  bottles diminished to  44.

Kamp.  Op. cit. in 5.B.  1877.  No. 16, pp. 325‑326.  Nine wine bottles on each side on the square.  Servant steals some.

Mittenzwey.  1880. 

Prob. 224, pp. 41 & 92;  1895?: 249, pp. 45 & 94;  1917: 249, pp. 40-41 & 89.  56  coins with  23  on a side, but with all corners the same and all edges the same, which gives a unique solution.

Prob. 225, pp. 41 & 92;  1895?: 250, pp. 45 & 94;  1917: 250, pp. 41 & 89.  32  flasks of wine with  9  on a side, reduced to 30, 28, 24, 20.

1895?: prob. 251, pp. 45 & 94;  1917: 249, pp. 41-42 & 90.  40  flasks of champagne,  11  on a side, reduced to 36, 32, 28, 24, 22.

Prob. 229, pp. 421 & 93;  1895?: 256, pp. 46 & 94;  1917: 256, pp. 42-43 & 90.  32  sacks,  12  on a side, reduced to  28.  = Leske, 585-6.

Prob. 230-232, pp. 42-43 & 93;  1895?: 257-259, pp. 46-47 & 94-95;  1917: 257-259, pp. 43 & 90.  Abbess and  24  nuns,  9  on a side, changed to  20, 28, 32.

Prob. 233, pp. 43 & 93;  1895?: 260, pp. 47 & 95;  1917: 260, pp. 43 & 91.  Fortress with  600  defenders,  200  on a side,  changed to  250,  then  300,  on each side.

Cassell's.  1881.  Pp. 98‑99: The twenty‑four monks.  = Manson, 1911, pp. 249-250.

Handy Book for Boys and Girls.  Op. cit. in 6.F.3.  1892.  Pp. 38-39: The counter puzzle.  "In an old book published over half a century ago, I came across this puzzle ...."  Rearranges  24  as  20, 28, 32. 

Somerville Gibney.  So simple!  V. -- A batch of match tricks.  The Boy's Own Paper 20 (No. 988) (18 Dec 1897) 188-189.  24  changed to  25, 20, 28, 32, 30.

Dudeney.  Problem 70: The well and the eight villas -- No. 70: The eight villas.  Tit‑Bits 33 (5 Mar 1898) 432  &  34 (2 Apr 1898) 8.  How many ways can numbers be placed in the  8  cells to make  9  along each side?  Answer is  2035.  He gives a general formula.

H. D. Northrop.  Popular Pastimes.  1901.  No. 3: The blind abbot and the monks, pp. 66-67 & 72.  = The Sociable.

Dudeney.  The monk's puzzle.  London Mag. 9 (No. 49) (Aug 1902) 89‑91  &  9 (No. 50) (Sep 1902) 219.  (= CP, prob. 17, pp. 39‑40 & 172‑173.)  How many ways can numbers be placed in the  8  cells to make  10  along each side?

Benson.  1904.  The dishonest servant puzzle, p. 228.  28  bottles,  9  on each edge, reduced to  24,  then  20.

Wehman.  New Book of 200 Puzzles.  1908. 

Pp. 20-21: The blind abbot and the monks.  = The Sociable.

P. 22: Fifteen "square" puzzle.  Uses the unusual figure of Endless Amusement II, prob. 32.  Starts with  5  marks in each cell so that it adds to  15  each way.  Remove four marks.  Solution is a bit unclear. 

M. Adams.  Indoor Games.  1912.  The cook and the jam, pp. 353-354.  36  jars of jam.

Blyth.  Match-Stick Magic.  1921.  Escaping from Germany, pp. 75-76.  32  with  9  on a side -- arranged  1, 7, 1  -- reduced to  28,  25  and  24.

Will Blyth.  Money Magic.  C. Arthur Pearson, London, 1926.  The stolen tarts, pp. 91-95.  9  on a side, start with  32,  reduce to  28, 24, 20.

Rohrbough.  Brain Resters and Testers.  c1935.  Ba Gwa, p. 18 (= pp. 18‑19 of 1940s?).  3 x 3  frame.  Two player game.  Start with  7  in each corner and  1  in each edge.  A player places a extra counter in the frame and the other tries to rearrange to preserve  15  in each outside row.  It says you can get  56  men on the board.  [I can get  60  if the corners can be empty.]

Jeffrey J. F. Robinson.  Musings on a problem.  MTg 37 (1966) 23‑24.  A farmer has  41  cows and wants to see  15  along each side of his house which is in the centre of a  3 x 3  array.  How many solutions are there if only  1, 2, ..., 8  different values can be used?  He finds all the solutions in some cases.

Ripley's Puzzles and Games.  1966.  Pp. 64-65, item 1.  16  pigs in  8  pens with  6  along each side.  Add four pigs. 

J. A. Dixon & Class 3T.  Number squares.  MTg 57 (1971) 38‑40.  Use the digits  1 ‑ 8  so the four side totals are the same.  They find that the sum can only be:  12, 13, 14, 15,  with  1, 2, 2, 1  solutions.  They then use  8  digits from  1 ‑ 9  and find  35  solutions.

 

          7.Q.1.          REARRANGEMENT ON A CROSS

 

            The counts from the base to the top and to the end of each arm remain constant though some (usually  2)  of the pearls or diamonds have been removed.  Trick versions with doubling up are in 7.Q.2.

            Versions with a  T  or  Y:  Secret Out;  M. Adams; 

 

Pacioli.  De Viribus.  c1500.  Ff. 116r - 117v.  Cap. LXX. D(e). un prete ch' in pegno la borscia del corporale con la croci de p(er)le al Giudeo (Of a priest who pledges to a Jew the burse of the corporale with a cross of pearls).  15  with three on each arm, one counts to nine from the base to each arm end.  This is reduced to  13.  Asks how one can add one pearl and produce a count of ten -- answer is to put it at the base.

Thomas Hyde.  Historia Nerdiludii, hoc est dicere, Trunculorum; ....  (= Vol. 2 of De Ludis Orientalibus, see 7.B for vol. 1.)  From the Sheldonian Theatre (i.e. OUP), Oxford, 1694.  De Ludo dicto Magister & Servus, pp. 234-236.  Cross with  28  thalers with  16  from the base to the end of each arm.  He points out that the picture has an error causing the count to the ends of the side arms to be  17.  Discusses general solution.  Says it is known to the Arabs of the Holy Land.

Les Amusemens.  1749.  P. xxvi.  13  markers reduced to  11.

Manuel des Sorciers.  1825.  Pp. 136-138, art. 16.  ??NX  Cross of coins reduced from  13  to  11.

Endless Amusement II.  1826?  Prob. 11, pp. 195-196.  15  diamonds reduced to  13.

The Boy's Own Book.  The curious cross.  1828: 414;  1828-2: 420;  1829 (US): 213;  1855: 568;  1868: 628.  13  markers, reducing to  11.

Nuts to Crack III (1834), no. 80.  The curious cross.  Almost identical to Boy's Own Book.

The Riddler.  1835.  The curious cross, p. 6.  Identical to Boy's Own Book.

Young Man's Book.  1839.  P. 60.  Easy Method of Purloining without Discovery.  Identical to Endless Amusement II, except with a title.

The New Sphinx.  c1840.  The curious cross, p. 143.  Same as Boy's Own Book, with a few words changed.

The Sociable.  1858.  Prob. 25: The dishonest jeweller, pp. 295 & 310.  15  diamonds, reducing to  13.  = Book of 500 Puzzles, 1859, prob. 25, pp. 13 & 28.  = Wehman, New Book of 200 Puzzles, 1908, p. 7.

The Secret Out.  1859.  The Dishonest Servant, pp. 78-80.  T shape.  16 coins reduced to 14.  Presentation of the problem is done with cards.

Boy's Own Conjuring Book.  1860.  Easy method of purloining without discovery, p. 295.  15  diamonds, reducing to  13.  Very similar to The Sociable.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 585-5, pp. 296 & 409: Juwelenkreuz.  18  jewels reduced to  16.

Mittenzwey.  1880.  Prob. 226, pp. 41-42 & 92;  1895?: 252-253, pp. 45-46 & 94;  1917: 252‑253, pp. 42 & 90.  15  jewels reduced to  13.  The 1895? addition has  10  reduced to  8.

Lucas.  La croix de perles.  RM2, 1883, pp. 134‑135.  c= Lucas; L'Arithmétique Amusante; 1895; pp. 10-11.  15  reduced to  13  and discussion.

Lemon.  1890.  The puzzling pearls, no. 535, pp. 69 & 117.  15  reduced to  13.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 136, no. 7.  Diamond cross reduced from  15  to  13.  No solution.

É. Ducret.  Récréations Mathématiques.  Op. cit. in 4.A.1.  1892?  P. 104: Les croix de jetons.  Rearrange a cross of  13  to  11.

H. D. Northrop.  Popular Pastimes.  1901.  No. 4: The dishonest jeweller, pp. 67 & 72.  = The Sociable.

M. Adams.  Indoor Games.  1912  The dishonest steward, pp. 22 & 24.  Rearrangement of a  Y.

Dudeney.  AM.  1917.  Prob. 423: The ruby brooch, pp. 144-145 & 249.  Complicated version.  Brooch is a circle with  8  radii and the owner used to count from the centre to the circle, along    of the circumference and then back in, getting  8  each time.  Dishonest jeweller reduces stones from  45  to  41  in a symmetric pattern of one in the middle, two on each arm and three on each arc.  What was it before?

Blyth.  Match-Stick Magic.  1921.  Counting the cross, p. 33.  Rearrange a cross of  15  to  13.

 

          7.Q.2.          REARRANGE A CROSS OF SIX TO MAKE TWO LINES OF FOUR, ETC.

 

            For the standard version, one has to put one marker on top of the one at the crossing.  See 6.AO and 7.Q for some similar trick versions.  The one which is closest to this section is:

(12,   7, 4) -- Trick version of a  3 x 3  square with doubled diagonal:  Hoffmann (1876),  Mittenzwey,  Hoffmann (1893), no. 8.

            See also in 6.AO, Hoffmann (1893), no. 9.

 

Les Amusemens.  1749.  P. xxv.  A cross of six -- four crossing three -- rearranged to count four both ways.

Blyth.  Match-Stick Magic.  1921.  Five by count, pp. 33-34.  A cross of seven -- five crossing three -- rearranged to count five both ways.

J. F. Orrin.  Easy Magic for Evening Parties.  Jarrolds, London, nd [1930s??].  The five puzzle, pp. 66-67.  As in Blyth.

Sid G. Hedges.  More Indoor and Community Games.  Methuen, London, 1937.  Penny puzzle, pp. 51-52.  Four pennies in the shape of a  T  or  Y  tetromino.  Make two rows of three.  Put one from an end of the row of three on the crossing.

Depew.  Cokesbury Game Book.  1939.  Seven coins, p. 223.  As in Blyth.

Ripley's Puzzles and Games.  1966.  Pp. 18-19, item 2.  An  L  with four in one leg and three in the other rearranged to have four on both lines.

 

          7.R.   "IF I HAD ONE FROM YOU, I'D HAVE TWICE YOU"

 

Jacques Sesiano.  The appearance of negative solutions in mediaeval mathematics.  Archive for the History of the Exact Sciences 32 (1985) 105-150.  In this, he discusses problems of the types given here and in 7.P.1, 7.R.1, 7.R.2 and 7.P.7.  He pays particular attention to whether the author discusses the problems in general or recognizes conditions for positivity or consistency, covering this in more detail than I do here.

 

            See Tropfke 609.

            In the medieval period, all these problems are extended in various ways to lead to quadratic and higher equations, but I think these become non-recreational although they certainly were not practical at the time, except for a few involving compound interest.

            I include a few examples of unusual cases of two equations in two unknowns here.

 

            NOTATION.

(a, b;  c, d) denotes the general form for two people.

            "If I had  a  from you, I'd have  b  times you."

            "And if I had  c  from you, I'd have  d  times you."

 I‑(a, b;  c, d;  ...) denotes the case for more people where 'you' means all the others.

II‑(a, b;  c, d;  ...) denotes the same where 'you' means just the next person, taken cyclically.

            With two people, there is no need to distinguish these cases.

            See Alcuin for an example where the second statement is interpreted as occurring after the first is carried out. 

            Sometimes  a = 0  -- see: Kelland (1839); Hummerston (1924).  If  a = c = ... = 0,  this can interpreted as a form of 7.R or 7.R.1 -- see: Dodson (1775).

            See Ghaligai, Cardan for versions with  "If I had  a  times yours from you, I'd have  b  times you", i.e.  x + ay = b(y - ay),  ....

            Versions giving  (x-a)/(y-a) = c;  (x+b)/(y+b) = d,  etc.  are forms of Age Problems and are generally placed in 7.X.  But see: Dodson.

            See Hall for a version where the first equation is  x + a = by.

            There are versions where one asks for a fraction  x/y  such that  (x+a)/y = c/d,  x/(y+b) = e/f.  These are forms of 7.R.1.  See: Dodson; Hall.

            Let  T  be the total of the amounts.  Then  I‑(a1,b1; a2,b2; ...)  with  n  people has  n  equations   xi + ai  =  bi(T ‑ xi - ai),   which can be rewritten as   xi + ai(1+bi)  =  bi(T - xi),   so we see that this is the same problem as discussed in 7.R.1 below where men find a purse, but with variable known purses,  pi = ai(1+bi).  We get  xi  =  biT/(1+bi) - ai.   Adding these for all  i  gives one equation in the one unknown  T,   T [Σ {bi/(1+bi} - 1]  =  Σ ai.

            For  II‑(a1,b1; a2,b2; ...),  systematic elimination in the  n  equations  xi + ai = bi (xi+1 ‑ ai)  leads to   x1 [b1b2...bn ‑ 1]   =  a1(b1+1) + a2b1(b2+1) + a3b1b2(b3+1) + ....  ,  and any other value can be found by shifting the starting point of the cycle.

            Verse versions:  Euclid;  Wingate/Kersey;  Ozanam;  Vinot; 

 

Euclid.  c-325.  Opera.  Ed. by J. L. Heiberg & H. Menge, Teubner, Leipzig, 1916.  Vol. VIII, pp. 286‑287.  Ass and mule in Greek and Latin verse.  (1, 2;  1, 1).  (Sanford 207 gives English of Clavius's 1605 version.  Cf Wingate/Kersey.)

Diophantos.  Arithmetica.  c250.  Book I.

No. 15. pp. 134‑135.  "To find two numbers such that each after receiving from the other may bear to the remainder a given ratio."  Does  (30, 2;  50, 3).

No. 18, pp. 135‑136.  "To find three numbers such that the sum of any pair exceeds the third by a given number."  E.g. "If I had 20 more, I'd have as much as you two."  Does with values  20, 30, 40.  This is like finding several purses -- see 7.R.1.

No. 19, pp. 136.  Same as no. 18, with  4  people.  Does with values  20, 30, 40, 50.

Metrodorus.  c510.  Art. 145‑146, p. 105.  (10, 3;  10, 5);  (2, 2;  2, 4).  Earliest example with non-integral answers.

Alcuin.  9C.  Problem 16: Propositio de duobus homines boves ducentibus.  (2, 1;  2, 2),  but the second statement in the problem is interpreted as happening after the first is actually carried out.  If a problem with parameters  (a, b;  c, d)  is interpreted this way, it is the same as our usual problem with parameters  (a, b;  c-a, d).

Mahavira.  850.  Chap. VI, v. 251‑258, pp. 158‑159.

253.  I‑(9, 2;  10, 3;  11, 5).

256.  I‑(25, 3;  23, 5;  22, 7).

al‑Karkhi.  c1010.  Sect. III, no. 5, p. 90.  II‑(1, 2;  2, 3;  3, 4;  4, 5).

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 145f., no. 7.  ??NYS -- cited by Tropfke 611.

Bhaskara II.  Bijaganita.  1150.  Chap. IV, v. 106.  In Colebrooke, p. 191.  (100, 2;  10, 6).

Fibonacci.  1202.  Pp. 190‑203 (S: 289-305).  Numerous versions, getting up to five people, some inconsistent examples and types where the second clause is  "I'd have  b1  times you plus  b2  more." 

Pp. 190 (S: 289-290).  (1, 1;  1, 10).

Pp. 190‑191 (S: 290-292).  (7, 5;  5, 7)  and general rules.

Pp. 191-192 (S: 292).  (6, 5¼;  4, 7⅔).

P. 192 (S: 292-293).  (7, 5, 1;  5, 7, 1)  -- denoting the extended type where the first says  "If I had  7  from you, I'd have  5  times you plus  1  more".  I.e. the first equation is  x + 7  =  5 (y ‑ 7) + 1.

Sesiano analyses the extended type on pp. 192-198 in detail.

Pp. 197‑198 (S: 298-300).  (7, 5, ‑1;  5, 7, ‑3).

Pp. 198‑199 (S: 300-301).  I‑(7, 5;  9, 7;  11, 7),  but he solves  I‑(7, 5;  9, 6;  11, 7).

Pp. 199‑200 (S: 301-302).  I‑(7, 5, 1;  9, 7, 1;  11, 7, 1),  but he solves as though the middle  7  is a  6.

Pp. 200‑201 (S: 302-303).  Types giving  x + y + 7  =  5 (z ‑ 7)  etc.

P. 201 (S: 303-304).  De eodem inter quattuor homines questio insolubilis [On the same with four men, an unsolvable problem].  This gives equations like  w + x + 7  =  3 (y + z - 7)  with coefficients   7, 3;  8, 4;  9, 5;  11, 6.  This is indeed inconsistent.

Pp. 202‑203 (S: 304-305).  I‑(7, 2;  8, 3;  9, 4;  10, 5;  11, 7),  but he solves with the last  7  as a  6.

Pp. 325‑326 (S: 455-456).  Problem of pp. 190‑191,  (7, 5;  5, 7),  done by false position.

Pp. 332‑333 (S: 463-465).  I‑(7, 4;  9, 5;  11, 6).

Pp. 344‑346 (S: 477-480).  I‑(7, 3;  9, 4;  11, 5)  done in two ways.

Lucca 1754.  c1330.  Ff. 29r‑30v, pp. 68‑70.  (12, 2;  17, 3).  II‑(15, 2;  18, 3;  21, 5).

Munich 14684.  14C.  Prob. XV, p. 80.   (1, 1;  2, 2)  and  (n, 1;  n, 2).

Folkerts.  Aufgabensammlungen.  13-15C.  Many sources.  Almost all have  (a, 1;  a, 2),  with the objects being exchanged being, gold, animals, coins, nuts, often noting that the answer is  a  times the answer for  a = 1.  Examples of  (a, 1;  a, d)  with  d = 3, 5, 9. 17.  An example of  (a, ½; a, d).  Folkerts cites Metrodorus, Alcuin, Fibonacci, AR, etc. 

                    Three sources of the following.  Ask a person to put the same amount of money into each of her hands.  Tell her to transfer  n  coins from the right hand to the left.  Now transfer enough from the left to double what is in the right hand.  This leaves  2n  in the left hand.

Giovanni di Bartolo.  Op. cit. in 7.H.  c1400.  He gives complex examples in probs. 10‑14, 54, 56, 57 on pp. 18‑27, 101‑107.  E.g. prob. 10, pp. 18‑21.  "If I had the square root of your money, I'd have  3  times you."  "And if I had the square root of your money, I'd have  4  times you."

Provençale Arithmétique.  c1430.  Op. cit. in 7.E.

F. 114r, p. 60.  (1, 1;  1, 2).

F. 114v,  p, 61.  (1, 2;  1, 4).  This has non-integral answers.

F. 115r, p. 61.  II-(1, 1;  1, 2;  1, 3).  Non-integral answers.

Pseudo-dell'Abbaco.  c1440.

Prob. 69, pp. 63‑65 with simple plate on p. 64.  (8, 2;  10, 3).

Prob. 126, pp. 100‑102 with simple plate on p. 101.  (3, 2;  5, 3).

AR.  c1450.  Prob. 138‑147, 220, 334‑336.  Pp. 70‑71, 102, 146‑147, 169‑171, 218.

138.  Regula augmentationis:  (1, 2;  1, 3);   (10, 2;  10, 3).

139‑147.  (3, 4;  4, 3);   (3, 3;  4, 4);   (3, 2;  5, 3);   (5, 2;  3, 3);   (1, 1;  1, 3);   (15, 2;  15, 3);   (3, 3;  3, 5) -- erroneous answer;   (1, 10;  1, 20);   (1, 1;  1, 2).

220.  same as 144.

334.  I‑(3, 2;  3, 4;  3, 10;  3, 100)  phrased as "If you each give me  1, ..."

335.  I‑(3, 2;  4, 4;  6, 7;  5, 9).

336.  I‑(2, 2;  2, 4;  2, 10). 

He then says that  I‑(1, 2;  1, 3;  1, 5)  and  I‑(1, 2;  1, 3;  1, 4)  are impossible.  However, I find solutions in both cases, though each person has  £ 1  which may be why AR is unhappy.  Vogel finds the same solutions that I do, but doubled because he reads the problems as  I‑(2, 2;  2, 3;  2, 5)  as in 334.  I don't read them that way, but the text and numerical layout are a bit inconsistent.

Benedetto da Firenze.  c1465.  Pp. 153.  (60, 6;  50, 13);   II‑(10, 2;  19, 4;  15, 9);   I‑(16, 2;  30, 8;  26, 7/2).

Muscarello.  1478.  Ff. 81v-82r, p. 198.  (2, 2;  2, 1).

della Francesca.  Trattato.  c1480. 

Ff. 43r (106-107).  The amount requested by the second two is a fraction of what the others have, so this is a mixed version of the problem leading to   x + 6 = 2(y+z‑6);  y + 2(x+z)/3 = 3(x+z)/3;  z + 3(x+y)/4 = 4(x+y)/4.  Answer:  (198, 90, 72)/7.

Ff. 121r-121v (258-259).  First says "If you give me that part of yours which is as  5  is to mine, then I will have  4  times you."  I.e.  x + 5y/x = 4(y-5y/x).  The second makes a similar statement.  The solution is obtained by a kind of false position, but I don't follow it.  There is an arithmetic error in the last line.

Ff. 121v-122r (259-260).  Two similar problems with a mixture of ordinary statements giving a fixed amount or a fixed part of the others money and statements as in the previous.  Again, I don't follow the solutions and the second leads to a quadratic.

Chuquet.  1484.  Prob. 57, 58, 59, 60.  Prob. 61‑79 extend in various ways.  English of 69, 70 78 in FHM 210-212.  78 is indeterminate.

57.  (7, 2;  9, 6).

58.  (20, 2;  30, 3).

59.  (1, 1;  1, 2).

60.  (2, 1;  3, 4).

61.  II-(3, 2;  4, 3;  5, 5).

62.  II-(2, 2;  2, 3;  2, 4).

63.  I-(7, 5;  9, 6;  11, 7).

64-76 lead to equations like   x + 7  =  5 (y - 7) + 1   or   5 (y + z - 7) + 1.

77.  II-(-9, 1/6;  -11, 1/7;  -7, 1/5) - i.e. the first equation is  y + z + 9  =  6 (x - 9).

78.  This has equations like   w + x + 100  =  3 (y + z - 100)   with coefficients  (3, 100;  4, 106;  5, 145;  6, 170), which is indeterminate.  "Thus it appears that such problems have a necessary answer for two by two, but for one by one they have whatever answer one desires."

79.  This has equations like   v + w + x + 7  =  2 (y + z - 7)   with coefficients  (7, 2;  8, 3;  9, 4;  10, 5;  11, 6).

Calandri.  Arimethrica.  1491.  F. 66r.  (20, 2; 30, 3).  This has the unusual feature that  x = y  and I do not recall any other such example.  The condition for  x = y  is more complex than one might expect:   a (b+1)/(b-1)  =  c (d+1)/(d-1).

Pacioli.  Summa.  1494.  He has numerous problems, sometimes mixing amounts and parts and sometimes mixing this topic with 7.R.1 and 7.R.2, often saying "that part of yours that 12 is to mine", i.e.  y(12/x)  {cf. della Francesca and 7.R.1}, and he often continues into problems where one gives the square root of what one has or says something about the square of an amount.

F. 105v, prob. 19.  Two men find two purses of values  p+10, p,  giving equations:  x + p+10 + 10  =  4 (y - 10),   y + p + 20  =  5 (x - 20).  He assumes the purses are worth  100  in total, so  p = 45,  p+10 = 55.  Answer:  (765, 690)/19.

Ff. 189r-189v, prob. 12.  x + 12  =  2 (y - 12),   y + x(12/y)  =  3 {x ‑ x(12/y)}

F. 191r-191v, prob. 24 & 25.  x + y(20/x)  =  y - y(20/x) + 28,   y + x(30/y)  =  x ‑ x(30/y) + 70.  Answer:  100, 120.  Prob. 25 is an alternative way to solve the problem.

F. 192v, prob. 29.   x + ⅓ (y + z)  =  96,   y + 60  =  2 (z + x - 60) - 4,   z + ¼ (x  + y) + 5  =  3 {¾ (x + y)} - 5.   Answer:  -79, 236, 289.  His algebra leads to  79 + x  =  0. 

F. 193v, prob. 34.   x + 6  =  2 (y + z - 6),   y + ⅔ (z + x)  =  3 {⅓ (z+x)},   z + ¾ (x + y)  =  4 {¼ (z + x)}.   Answer:  (198, 90, 72)/7.

Calandri, Raccolta.  c1495. 

Prob. 23, pp. 22‑23.  (20, 2;  30, 3).

Prob. 44, pp. 40‑41.  Three people -- first two as  (12, 2;  20, 3),  third says  "If I had  24  from you two, I'd have  3  times you plus the square root of what you have."

Hans Sachs (attrib.).  Useful Table-talk, or Something for all; that is the Happy Thoughts, good and bad, expelling Melancholy and cheering Spirits, of Hilarius Wish-wash, Master-tiler at Kielenhausen.  No publisher, place or cover, 1517, ??NYS -- discussed and quoted in: Sabine Baring-Gould; Strange Survivals  Some Chapters in the History of Man; (1892), 3rd ed., Methuen, 1905, pp. 220-223.  [Not in Santi.]  Baring-Gould, p. 221 has  (1, 2;  1, 1).

Ghaligai.  Practica D'Arithmetica.  1521.  He has a series of problems of this type, of increasing complexity, all involving men and money.  I omit the more complex cases.  He also uses parts as in Pacioli.

Prob. 1, f. 100r.  (10, 1;  20, 2).

Prob. 2, f. 100r.  (20, 2;  30, 3).

Prob. 3, f. 100v.  x + ¼y  =  y - ¼y,   y + ½x  =  4 (x - ½x) + 2.

Prob. 5, f. 101r.  x + 10  =  y - 10,   y + x(20/y)  =  3 {x - x(20/y)}.

Prob. 6, f. 101r.  x + 12  =  2 (y - 12),   y + x(12/y)  =  3{x - x(12/y)}.  = Pacioli 12.

Prob. 7, f. 101v.  x + y(6/x)  =  21,   y + x(3/y)  =  20,   given that  y(6/x) + x(3/y)  =  11.  [Without the extra condition, this gives a fourth order equation with solutions   x = 0, 12, 15 ± Ö6;   y = 0, 18, 11 -/+ 3Ö6,  though  0, 0  is indeterminate in the original equations.

Prob. 9, f. 102r.  Same as prob. 7, but the extra condition is replaced by   x + y  =  30.

Prob. 10, f. 102r-102v.  x + ry  =  2 (y - ry),   y + rx  =  5 (x - rx),  where  r  is an unspecified ratio.  This seems to be a unique version of this problem and both forms I and II lead to interesting solutions --  r  is determined by the coefficients  2, 5.  Cardan has a related version, but he gives a value of  r  which is inconsistent. 

            For the 'all others' (type I) version, we let  T = Σ xi.  Then the equations are:   xi + r(T - xi)  =  ai(1-r)(T - xi).  Since   xi + r(T - xi) + (1-r)(T - xi)  =  T,   we see that   xi + r(T - xi)  =  aiT/(1+ai),   so   T  =  (1+ai)[xi + r(T‑xi)]/ai =  (1+ai)(1‑r)(T-xi)  or   T - xi  =  T/(1+ai)(1-r),  assuming  1-r ¹ 0.  Adding these last equations gives  (n‑1)T  =  [T/(1‑r)] Σ 1/(1+ai).  Assuming  ¹ 0  gives us   1‑r  =  [1/(n‑1)] Σ 1/(1+ai).  Hence  r  is determined by the  ai's,  or else  T = 0  and/or  1-r = 0.  In fact  T = 0  holds if and only if  1-r = 0  or  all xi = 0.  When  1-r = 0,  the  xi  are arbitrary.  In either of these degenerate cases, the  ai  are arbitrary.

            For the 'next one' (type II) version, the equations are:   xi + rxi+1  =  ai(1‑r)xi+1,   or   xi  =  [(1-r)(1+ai) - 1] xi+1.   Multiplying these together, we find that the product of the factors must be  1  and this gives an  n-th order polynomial for  1-r.  Set   P(x)  =  Π [(1+ai)x - 1].   If we assume all  1+ai > 0,  then this has  n  positive roots and hence  P(x) = 1  occurs for some  x = 1-r  greater than the largest term  1/(1+ai).  If we further assume the  ai  are not too small, namely that  Π ai > 1,  then this product is  P(1)  and we hence know there is a point with  P(x) = 1  for some positive  x  less than  1,  so the corresponding  r = 1-x  is also between  0  and  1.  There may well be other suitable roots.  If some  1+ai < 0,  the situation is more complex and there need not be any positive roots.  Even if all  ai  are positive,  P(x) = 1  may only occur for  x > 1,  and hence  r < 0.  When  n  is even, the constant terms in the equation cancel and one can factor out  1-r  (since  1-r = 0  leads to  r = 1,  and an easy solution or inconsistency).  E.g., for  n = 2,  we get   1‑r  =  1/(1+a1) + 1/(1+a2).

Riese.  Rechnung.  1522.  1544 ed. -- p. 93;  1574 ed. -- pp. 62v‑63r.  (1, 1;  1, 3).

Riese.  Die Coss.  1524.  No. 25‑30, p. 44 & No. 63‑64, p. 49.

No. 25.  (1, 1;  1, 2).

No. 26.  (1, 1;  1, 3).

No. 27.  (1, 10;  1, 20).

No. 28.  (1, 2;  1, 5).

No. 29.  II‑(1, 1;  1, 2;  1, 3).

No. 30.  II‑(1, 1;  2, 2;  3, 3).

No. 63.  (1, 1;  4, 2).

No. 64.  (1, ½;  5, 3).

Cardan.  Practica Arithmetice.  1539.  Chap. 61.

Section 3, f. S.viii.v (p. 110).  (5, 4;  4, 4).

Section 4, f. S.viii.v (p. 110).  Variation leading to:   x + ½y  =  3 (y - ½y),   y + ½x  =  7 (x - ½x),   which he rightly states is impossible (unless  x = y = 0).

Section 7, f. T.i.v (p. 111).  Variation giving equations:   x + ½y + 2  =  9 (y - ½y - 2),   y + ⅓x + 3  =  3 (x - ⅓x - 3).

Recorde.  Second Part.  1552.  Pp. 322-324: The fourth example.  (2, 4;  3, 1).

Tartaglia.  General Trattato.  1556. 

Book 16, art. 12‑15, pp. 240v‑241v.  (6, 1;  9, 2).   (7, 2;  13, 3).   (30, 1;  30, 2).  I‑(16, 1;  24, 2;  33, 3).

Book 17, art. 23, 34, 35,  pp. 271v-272r  &  273v-274v.  II‑(32, 2;  38, 3;  50, 4;  76, 7).   (24, 2;  42, 3).   II-(34, 2;  52, 3;  80, 5).

Buteo.  Logistica.  1559.  Prob. 59, p. 264.  i-th  says  "I have  ai  times as much as the rest of you."  with   (ai)  =  (1, 1/2, 1/5).   This could be considered as  I-(0, 1;  0, 1/2;  0, 1/5).  This is indeterminate with general solution proportional to  (3, 2, 1).  He assumes  x = 24  and gets  y = 16,  z = 8.

van Etten.  1624.  Prob. 83 (76), parts a & c, pp. 90‑92 (134‑136).  Ass & mule --  (1, 2;  1, 1)  = Euclid.   (10, 3;  10, 5),   (2, 2;  2, 4)  = Metrodorus.

Hunt.  1651.  Pp. 280-281: Of the mule and the ass.  (1, 2;  1, 1).

Schott.  1674.  Ænigma V, pp. 553.  (1, 1; 1, 2)  =  Euclid.  Cites Euclid, Clavius and Lantz.

Wingate/Kersey.  1678?.  Quest. 39, pp. 502-503.  Ass and mule in Latin verse - cf Euclid.  (1, 2;  1, 1)

Edward Cocker.  Arithmetic.  Ed. by John Hawkins.  T. Passinger & T. Lacy, London, 1678.  [De Morgan states "I am perfectly satisfied that Cocker's Arithmetic is a forgery of Hawkins" and then spends several pages detailing this charge and showing that the book is a rather poor compilation from several better books.  However Ruth Wallis [Ruth Wallis; Edward Cocker (1632?-1676) and his Arithmetick: De Morgan demolished; Annals of Science 54 (1997) 507-522] has argued that De Morgan is wrong.  Inspection of a 1st ed. at the Graves collection and a 3rd ed., 1680, at Keele shows no noticeable difference in the texts other than resetting which makes the book smaller with time -- all the editions seen have the same 32 chapters.  The 1st and 3rd eds. seem to have identical pagination so I will not cite the 1680.]  1st ed., 1678 & 3rd ed., 1680, both T. Passinger & T. Lacy, London.  = 33rd ed., Eben. Tracy, London, 1715.  = Revised by John Mair; James & Matthew Robertson, Glasgow, 1787.  Chap. 32, quest. 4.  1678: p. 333;  1715: p. 215;  1787: p. 186.  (1, 5;  1, 1).

Wells.  1698.  No. 104, p. 206.  Ass & mule:  (1, 1;  1, 2);  (a, 1;  a, 2).

Ozanam.  1725.  De l'asne et du mulet, prob. 24, question 3, 1725: 176‑178.  Prob. 6, 1778: 189-190;  1803: 186-188;  1814: 162-163;  1840: 84-85.  (1, 2;  1, 1).  1725 gives two versions and solutions in Latin verse.  1778 et seq. gives just one version and solution, but with slight differences, and refers to the Metrodorus problems in Bachet's Diophantos, though these are not the numbers in Metrodorus.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XXIII, p. 93 (not in 1790 ed.).  A  says to  B,  if I had  a  of your money, I'd have as much as you together with half of  C,  etc., giving:   x + a  =  y-a + ½z,   y + b  =  z-b + ⅓x,  z + c  =  x-c + ¼y.  Finds general solution and does case  a = b = c = 5.

Walkingame.  Tutor's Assistant.  1751.  1777: p. 175, prob. 98;  1860: p. 184, prob. 97.  (4, 1;  4, 2),  flocks of sheep.

Mair.  1765?  P. 458, ex. 7.  Two men with money, like ass & mule:  (1, 5;  1, 1).

Euler.  Algebra.  1770.  I.IV.IV.612: Question 3, pp. 208‑209.  Ass and mule:  (1, 2;  1, 3).

Vyse.  Tutor's Guide.  1771?  Prob. 11, 1793: p. 130; 1799: p. 138 & Key p. 183.  (1, 1;  1, 2)  (= Euclid).

Dodson.  Math. Repository.  1775.

P. 8, Quest. XIX.  (1, 1;  1, 2)  (= Euclid).

P. 19, Quest. L.  Find  x/y  such that  (x+1)/y  =  1/3,  x/(y+1)  =  1/4.  [(x+1)/y  =  1/n;  x/(y+1)  =  1/(n+1)  has solution  x = n+1,  y = (n+1)2 - 1.  In general,  (x+1)/y  =  a/b,  x/(y+1)  =  c/d  gives  x = c(a+b)/(ad-bc),  y = b(c+d)/(ad-bc).  One would normally assume  a/b > c/d.  This really belongs in 7.R.1.  Cf Wolff in 7.R.3 for a different phrasing of the same problem.]

P. 31, Quest. LXXVI.  Find  x/y  such that  (x+4)/(y+4)  =  4/3,  (x-4)/(y-4)  =  3/2.  [In general  (x+A)/(y+A)  =  a/b,  (x-A)/(y-A)  =  c/d  has solution  x = A (2ac‑bc‑ad)/(bc-ad),  y = A (ad+bc-2bd)/(bc-ad).  One would normally assume  a/b < c/d.  This really belongs in 7.X.]

P. 46, Quest. XCIX.  w = (x+y+z)/2,  x = (w+y+z)/3,  y = (w+x+z)/4,  x = 14 + z. 

Eadon.  Repository.  1794.  P. 296, no. 8.  (5, 1;  5, 3).

D. Adams.  Scholar's Arithmetic.  1801.  P. 209, no. 2.  (1, 1;  1, 2).

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Art. 121, p. 32.  (10, 2;  10, 3).

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 117-119 & 256-257, nos. 443‑450.  (5½, 2;  6½, 5),  (13½, 7;  16½, 3),  (10, 4;  7½, 2),  (10, 4;  12, 3),  (20⅓, 3;  12⅔, 4),  (9½, 4;  15, 2),  (33½, 2;  16½, 3),  (6, 5;  4 2/5, 8).

Philip Kelland.  The Elements of Algebra.  A. & C. Black, Edinburgh, et al., 1839.  ??NX.  P. 134:  "A's  money or debt is  a  times  B's;  if  A  lose  £10  to  B,  it will be  b  times  B's."  (Also entered in 7.X.)

The New Sphinx.  c1840.  No. 46, pp. 24 & 122.  Women with baskets of eggs:  (1, 2; 1, 1).

Fireside Amusements.  1850: No. 2, pp. 101 & 180;  1890: No. 2, p. 96.  = New Sphinx. c1840.

The Family Friend (1856) 376,  Enigmas, Charades, &c.  176  Arithmetical Puzzle.  Standard  (1, 2;  1, 1)  given in a four stanza poem involving two costermongers with barrows of apples.  Signed  G. M. F. G.

Thomas Grainger Hall.  The Elements of Algebra: Chiefly Intended for Schools, and the Junior Classes in Colleges.  Second Edition: Altered and Enlarged.  John W. Parker, London, 1846.  P. 132, ex. 7.  Appears to be  (50, 2;  50, 1),  but it reads: "A says to B, if you give me £50, I shall have twice as much as you had; but if I give you £50, each will have the same sum."  The use of 'had' means the first equation is  x + 50  = 2y,  while the second equation is the usual  x - 50  =  y + 50.  Answer:  250, 150.

Magician's Own Book.  1857.  The two drovers, p. 246.  (1, 1;  1, 2).  = Book of 500 Puzzles, 1859, p. 60.  = Boy's Own Conjuring Book, 1860, p. 218.

Vinot.  1860.  Art. XLVI: L'Anesse et le Mulet, pp. 64-65.  Gives a French translation of the Latin verse.  (1, 1;  1, 2).

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-33, pp. 255 & 396.  (1, 1;  1, 2).  Notes that the solution to  (a, 1;  a, 2)  is just  a  times the solution of the original.

(Beeton's) Boy's Own Magazine 3:4 (Apr 1889) 175  &  3:6 (Jun 1889) 255.  (This is undoubtedly reprinted from Boy's Own magazine 1 (1863).)  Mathematical question 34.  (5, 1;  10, 2)  with postage stamps.

Mittenzwey.  1880. 

Prob. 11, 12, 14, pp. 2 & 59;  1895?: 12, 13, 15, pp. 8 & 63;  1917: 13, 12, 17, pp. 8 & 57.  (1, 2;  1, 1).  (3, 1;  3, 2).  (5, 2;  5, 1).

1895?: prob. 86, pp. 19 & 69;  1917: 86, pp. 18 & 65.  (5, 3;  2, 5).  1917 adds an algebraic solution.

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.

No. 21, pp. 156 & 333.  (20, 4;  20, 3/2).

No. 11, pp. 163 & 334.  II-(20, 4;  40, 4/9;  60, 7/4).  Answer:  (5620, 1880, 6700)/19

Hoffmann.  1893.  Chap. IV, no. 26: A rejected proposal, pp. 150 & 195 = Hoffmann-Hordern, p. 122.  (1, 3;  2, 1),  but the first person also says he already has twice the second, so this is an overdetermined problem.  (Hoffmann's other example, no. 25, is just the classic  (1, 2;  1, 1), which is not overdetermined, so I have omitted it here.)

Hummerston.  Fun, Mirth & Mystery.  1924.  Pocket money, Puzzle no. 4, pp. 21 & 172.  (0, 2;  ½, 3).

Sullivan.  Unusual.  1947.  Prob. 36: A problem old enough to be considered new.  (1, 1;  1, 2).

David Singmaster.  Some diophantine recreations.  Op. cit. in 7.P.5.  1993.  Sketches some history; finds condition for integer data in  (a, b; c, d)  to produce an integer solution, namely  (bd - 1)/(b+1, d+1)  divides  a + c  where  (b+1, d+1)  is the Greatest Common Divisor of  b+1  and  d+1.  A letter from S. Parameswaran interpreted the problem given in Alcuin as though the second statement was also made by the first animal.  This gives  x + a  =  b (y - a);  x + c  =  d (y - c)  and similar reasoning finds the integrality conditions for this variant.

David Singmaster.  A variation of the ass and mule problem.  CM 28: 4 (May 2002) 236-238.  The 2001 Maritime Mathematics Contest had a specific case of the following.  The first person says:  "If I had  a  from you, I'd have  b  times you, but if I gave  c  to you, I'd have  d  times you."  This leads to the equations:  x + a  =  b (y - a);  d (y + c)  =  x - c.  The integral  d  of the classic problem has been changed to  1/d  with  d  integral.  Note that  b > d  for reasonable solutions.  Reasoning similar to the previous article finds the condition for integer data to produce an integer solution and the condition is simpler than for that problem.

Tomislav Došlić.  Fibonacci in Hogwarts?  MG 87 (No. 510) (Nov 2003) 432-436.  Observes that Fibonacci does the case  (7, 5;  5, 7),  whose solution is non-integral:  (121, 167)/17.  He considers the case  (a, b;  b, a)  and finds there are  16  positive pairs  a, b  which give integral solutions.  If we set  a £ b,  these are   a, b  =  1, 2;  1, 3;  1, 5;  2, 2;  2, 3;  2, 8;  3, 3;  3, 7;  5, 8.

David Singmaster.  Integral solutions of ass and mule problems.  Gives a simpler solution for Došlić's result and finds all integral solutions when positivity is not required.

 

          7.R.1. MEN FIND A PURSE AND 'BLOOM OF THYMARIDAS'

 

            See Tropfke 604 & 606. 

            Algebraically, 7.R.1 and 7.R.2 differ only in signs.

            NOTATION.  Finding a purse has two forms.

I‑(a1, a2, ..., an)  --  i-th says "If I had the purse, I'd have  ai  times the rest of you".

II‑(a1, a2, ..., an)  --  i-th says "If I had the purse, I'd have  ai  times the  i+1-st person".

            There are two related problems which I call forms  III  and  IV.

III‑(a1, a2, ..., an)  -- the sum of all amounts except the  i‑th is  ai.  See the discussion under

            Iamblichus.  A number of problems in 7.H lead to this type of problem when one uses

            reciprocal times as work rates.

IV-(a1, a2, ..., an)  --  xi + xi+1 = ai.  (This is determinate only if  n  is odd.)

            For  n = 3,  types III and  IV  are the same, though the constants or the variables are taken in a different order, so that  III‑(a, b, c) = IV-(c, a, b)  if we keep the variables in the same order.

            I give answers as a list of the amounts, in order;  then the purse.

            Let  p  be the value of the purse and let  T  be the total of the amounts.  Then  I‑(a1,a2,...)  with  n  people has  n  equations   xi + p  =  ai(T ‑ xi),   so we see that this is the same problem as discussed in 7.R.2 below where men buy a horse, but with the value of the horse and the multipliers all negative, which makes this version have fewer sign complications in its solution.  Thus we get  xi  =  (aiT-p)/(1+ai).   Adding these for all  i  gives one equation in the two unknowns  T  and  p.  However, letting  C = T + p  leads to the simplest equation:   (n‑1)T  =  [Σ  1/(1+ai)] C.

            For  II‑(a1,a2,...),  systematic elimination in the  n  equations  xi + p = aixi+1  leads to   x1 [a1a2...an - 1]  =  p [1 + a1 + a1a2 + ... + a1a2...an],  and any other value can be found by shifting the starting point of the cycle.

            In either case, the solution can be adapted to variable purses -- see 7.R.  In some problems, gaining the purse is replaced by paying out, so the purse can be treated as having a negative value -- see Unger, 1838.

            Other versions of the problem has several horses and a saddle (or other equipage) or several cups and a cover.  The  i-th horse with the saddle is worth  ai  times the others or the next.  I don't seem to have entered any of these until recently.

            There are versions where one asks for a fraction  x/y  such that  (x+a)/y = b,  x/(y+c) = d,  where  a, c  are integral (possibly non-positive) and  b, d  are rationals.  This is a mixture between 7.R and 7.R.1, but fits better here as we can think of  a, c  as purses.  I will denote this as:

V-(a, b; c, d).  This has solution  x  =  β(a + αb)/(α - β);  y  =  (a + βb)/(α - β). 

See: Dodson;  Todhunter.

          Some of the cistern problems in 7.H are of type III

 

Diophantos.  Arithmetica.  c250.  Book I.

No. 16, p. 135.  "To find three numbers such that the sums of pairs are given numbers."  He does  III‑(20, 30, 40).

No. 17, p. 135 is the same for four numbers.  He does  III‑(22, 24, 27, 20).

No. 18 & 19 are like 3 and 4 men finding 3 and 4 purses -- see under 7.R.

No. 20, pp. 136‑137.  This is Type I with a purse of 0, i.e.  "I have  ai  times the rest of you".  He does this as  I‑(3, 4, a3)  since  a3  is determined by  a1  and  a2.

Iamblichus.  On Nicomachus's 'Introductio Arithmetica'.  c325.  Pp. 62‑63, ??NYS.  Partly given in SIHGM I 138‑141.  Describes the 'Bloom of Thymaridas' which has  n+1  unknowns  x,  x1,  ...,  xn  and we know   x + xi  =  ai   and   x + x1 + ... + xn  =  s.   Then  x  =  (a1 + ... + an ‑ s)/(n‑1).  (Iamblichus uses  n  unknowns.)  Heath (HGM I 94‑96) says Iamblichus continues and applies the Bloom to  I‑(a1, ..., an),  with integral  ai,  by letting  x  be the value of the purse and  s  =  (a1+1)...(an+1),  which yields   x + xi  =  sai/(ai+1).  (We can take  n-1  times the value of  s  to insure integer solutions.)  For rational  ai,  we let  s  be the  LCM  of the denominators of  ai/(ai+1).  Iamblichus gives the problems  I‑(2, 3, 4)  and  I‑(3/2, 4/3, 5/4).  See Chuquet for an indeterminate version of the Bloom.

                    This is closely related to problems like the following:   x + y = a,  y + z = b,  z + x = c,   i.e.  III‑(b, c, a) = IV-(a, b, c).  We set  T = x + y + z  and so   T ‑ z = a,  T ‑ y = b,  T ‑ x = c.  This is a case of the 'bloom' for  n = 3,  with   x = T,  x1 = ‑z,  a1 = a,  etc., and  s = 0.  In general, this gives us   x  =  T  =  (a1 + ... + an)/(n‑1).

Aryabhata.  499.  Chap. II, v. 29, pp. 71-72.  (Clark edition: pp. 40‑41.)  III‑(a1, ..., an).  Gives   T  =  (a1+...+an)/(n‑1).

Bakhshali MS.  c7C.  Kaye I 39-42, sections 78-79  and  Datta, pp. 45‑46  discuss two types of related systems for  n  amounts  x1, x2, ..., xn.  See the discussion under Iamblichus.

IV-(a1, ..., an).  When  n = 3,  this is equivalent to type III.  Kaye notes that n is always odd and says the following occur:  IV-(13, 14, 15) (Kaye III 166, f. 29r);  IV‑(16, 17, 18, 19, 20) (Kaye III 166-167, ff. 29v & 27v);  while the following are implied:  IV‑(9, 5, 8);  IV‑(70, 52, 66);  IV-(1860, 1634, 1722);  and possibly  IV‑(36, 42, 48, 54, 60).  Kaye's concordance (I 38-39) implies these examples should be in the same area of the text, but I can't find them in his Part III -- ??.  Also  III-(317, 347, 357, 362, 365) (Kaye I 40 (omitting the fourth equation); III 168‑169, ff. 1v-2r).

T ‑ xi  =  c ‑ dixi.   These are variations of Type III problems or of the Present of Gems problem, section 7.P.4.

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 29.  Sanskrit is on pp. 125-127; English version of the examples is on pp. 307-308.

                    Ex. 1:  III-(30, 36, 49, 50).

                    Ex. 2:  III-(28, 27, 26, 25, 24, 23, 21).

Mahavira.  850.  Chap. VI, v. 159, 233‑250, pp. 136‑137, 153‑158.

V. 159, pp. 136‑137.  III‑(22, 23, 24, 27).

V. 236, p. 155.  I‑(2, 3, 5).   Answer:  1, 3, 5; 15.

V. 239, p. 156.  i-th says  "If I had  bi  of the purse, I'd have  3  times the rest of you", with  B = (1/6, 1/7, 1/9, 1/8, 1/10).  Answer:  261, 921, 1416, 1801, 2109; 110880.

V. 242, p. 157.  i-th says  "If I had  bi  of the purse, I'd have  ai  times the rest of you", with  a1, b1;  a2, b2  =  2, ½;  3, ⅔.   Answer:  11, 13; 30

V. 244, p. 157.  I‑(2, 3).   Answer:  3, 4; 5.

V. 245, p. 157.  I‑(8, 9, 10, 11).   Answer:  103, 169, 223, 268; 5177.

V. 248, pp. 157‑158.  As in v. 242, with four men and   a1, b1;  ...  =  2, 1/5;  3, 1/4;  5, 1/2;  4, 1/3.   Answer:  356, 585, 445, 624; 14760.

V. 249.  As in v. 242, with   2, ¼;  3, ⅓;  4, ½.   Answer:  55, 71, 66; 876.

al‑Karkhi.  c1010.  Sect. III, no. 24‑25 & 29‑30, pp. 95 & 98.

24:  III‑(20, 30, 40).   (= Diophantos I 16.)

25:  III‑(30, 45, 40, 35).

29:  Three men find purses  (30, 40, 20).  i-th says:  "If I had the  i‑th purse, I'd have as much as all of you."

30:  same as 29 with four men and purses  (20, 30, 40, 50).

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 129, no. 46.  ??NYS -- Hermelink, op. cit. in 3.A, says this is a problem with two persons.  Tropfke 607 cites this with no details.

Fibonacci.  1202.  Pp. 212‑228 (S: 317-337), Chap. XII, part 4: De inventione bursarum [On the finding of a purse].  Many problems, going up to five men, 4 purses and an example with a negative solution.

Pp. 212‑213 (S: 317-318).  I‑(3, 4).   Answer:  4, 5; 11.

Pp. 213‑214 (S: 318-319).  I‑(2, 3, 4).   Answer:  7, 17, 23; 73.   (= Iamblichus's 1st.)

Pp. 214‑215 (S: 320).  I‑(3, 4, 5, 6).   Answer:  4, 67, 109, 139; 941.

Pp. 215‑216 (S: 320-322).  I‑(5/2, 10/3, 17/4, 26/5, 37/6).   Answer:  ‑49154, 30826, 89478, 131962, 163630, 1088894.   "... aut positio huius questionis indissolubilis est; aut primus homo debitum habebit ...." [... this posed problem will not be solvable unless the first man has a debit ...].  See Sesiano.

P. 216 (S: 322-323).  II‑(2, 3, 4).   Answer:  9, 16, 13; 23.

P. 217 (S: 323).  II‑(5/2, 10/3, 17/4).   He doesn't find the answer which is  284, 444, 381; 826.

Pp. 218‑220 (S: 325-327).  II‑(2, 3, 4, 5).   Answer:  33, 76, 65, 46; 119.

Pp. 220‑223 (S: 327-330).  Four men find four purses of values:  

p1, p2, p3, p4  =  p, p+3, p+7, p+13. 

i-th says  "If I had the  i-th purse, I'd have  ai  times the rest of you",  with  A  =  (a1, a2, a3, a4)  =  (2, 3, 4, 5).

P. 223 (S: 330-331).  Two men & purses of values:  p, p+13,  with  A = (2, 3).

Pp. 223‑224 (S: 331-332).  Three men & purses:  p, p+10, p+13,  with  A = (2, 3, 4).

Pp. 224‑225 (S: 332-333).  Four men & purses:  p, p+10, p+13, p+19,  with  A = (2, 3, 4, 5).

Pp. 225‑226 (S: 333-335).  Four men, one purse, giving  w + x + p  =  2y,  etc.

P. 227 (S: 335-336).  Three men, one purse, giving  x + y + p  =  2z,  etc.

P. 227 (S: 336).  Four men, one purse, giving  w + x + p  =  2y + z,  etc.

Pp. 227‑228 (S: 336-337).  Five men, one purse, giving  v + w + p  =  2 (x + y + z),  etc., with constants  2, 3, 4, 5, 6.   Answer:  22, -9, 57, 12, 71; 267  (the text has  7  instead of  71).  "... quare hec questio est insolubilis, nisi ponamus, secundum hominem habere debitum  9, ..." [... therefore this problem is unsolvable unless we put the second man to have a debit of 9, ...].  See Sesiano.

P. 284 (S: 405).  III‑(31, 34, 37, 27).

P. 285 (S: 405-406).  III‑(31, 34, 37, 39, 27).  On pp. 284-285 & 286, he also considers problems like  IV-(a, b, c, d)  (cf. Bakhshali) and notes that some are inconsistent (e.g. for  a, b, c, d  =  27, 31, 34, 37)  and others are indeterminate.

Pp. 302‑303 (S: 426-427).  Four men, one purse, giving equations   w + x + p  =  3/2 y,  etc., with constants  3/2, 9/4, 16/5, 25/6.   Answer:  8665, 5682, 12718, 10280; 4730.

Pp. 326‑327 (S: 456-458).  Problem of pp. 218‑220 done by false position.

Pp. 330-331 (S: 461-462).  III-(75, 70, 67, 64, 54, 50). 

P. 333 (S: 465).  Problem of pp. 214‑216 done by false position.

Pp. 346‑347 (S: 480-481).  Three men & purses of values  18, 16, 20,  giving   x' + 18  =  3y',   y' + 16  =  4z',   z' + 20  =  5x'.  By setting  

x' = x ‑ 11,  y' = y ‑ 7,  x' = z ‑ 9,  he converts to the ass & mule problem (7.R) on pp. 344‑346.

Pp. 349‑352 (S: 484-487).  Four men, one purse, giving   w + p  =  2 (x + y),   etc., with constants  2, 3, 4, 5.   Answer:  ‑1, 4, 1, 4; 11.  "... hec questio insolubilis est, nisi concedatur, primum hominem habere debitum, ..." [... this problem is not solvable unless it is conceded that the first man can have a debit, ...].  See Sesiano.

Fibonacci.  Flos.  c1225.  In Picutti, pp. 316-319, numbers IV-V.

Pp. 238-239: De quatuor hominibus et bursa ab eis reperta, questio notabilis.  Described as the second of the problems that Fibonacci sent to Frederick II.  Same as Fibonacci, pp. 349‑352, though he doesn't cite this.  "... hanc quidem questionem insolubilem esse monstrabo, nisi concedatur, primum hominem habere debitum: ...."  See Sesiano.

Pp. 239-240: De eadem re.  Does the same problem with multipliers  4, 5, 6, 7.   Answer:  -1, 6, 1, 6; 29.  Implies a general solution for multipliers   k ‑ 2, k ‑ 1, k, k + 1   is   -1, k, 1, k;  k2 - k - 1.

Jordanus de Nemore.  De Numeris Datis.  c1225.  Critical edition and translation by Barnabas Hughes.  Univ. of Calif. Press, Berkeley, 1981.  Prob. II‑24, pp. 150‑151.  General version of type  I  with purse given.  Example:  I‑(1/9, 1/3, 3/5, 1)  with purse  6.  Answer:  2, 14, 24, 34.

Ibn Badr  = Abenbéder  = Abu ‘Abdallah Muhammad (the  h  should have an underdot) ibn ‘Umar ibn Muhammad (the  h  should have an underdot).  c1225.  Arabic text with Spanish translation by José A. Sánchez Pérez as:  Compendio de Álgebra de Abenbéder; Centro de Estudios Históricos, Madrid, 1916.  Tercer problema análogo, pp. 109-111 (pp. 70-71 of the Arabic).   (4, 7).   Answer:  (8, 5; 27)/9.

BR.  c1305.  No. 61, pp. 84‑87.  (7, 11).  Answer given is  12, 8; 76,  but should be  8, 12; 76.

Gherardi.  Libro di ragioni.  1328.  P. 53.  3 men find a purse.  I‑(2, 3, 4).  Answer:  7, 17, 23; 73.

Lucca 1754.  c1330.  F61r, pp.139‑140.  II‑(2, 3, 4, 5).  Answer:  33, 76, 65, 46; 119.  (= Fibonacci, pp. 218‑220.)

Bartoli.  Memoriale.  c1420.  Prob. 4, f. 75r (= Sesiano, pp. 136-137 & 147.  I‑(2, 3, 4).   Answer:  7, 17, 23; 73.   (= Iamblichus's 1st.)

Pseudo-dell'Abbaco.  c1440.  Prob. 125, p. 100.  I‑(2, 3, 4).  Answer:  7, 17, 23; 73.  (= Iamblichus' 1st problem.)

AR.  c1450.  Prob. 113, 159, 227.  Pp. 63‑64, 76, 168‑169, 217.

113:  I‑(1, 2, 3),  answer:  2, 10, 14; 22.

159:  I‑(4, 10),  answer:  5, 11; 39.

227:  I‑(1, 2, 3),  answer:  1, 5, 7; 11.  Cf. 113.

Correspondence of Johannes Regiomontanus, 1463?-1465.  Op. cit. in 7.P.1.

P. 238, letter from Bianchini, 5 Feb 1464.  Query 8:  III‑(42, 54, 30).

Pp. 259 & 291, letter to Bianchini, nd [presumably 1464].  P. 259 gives the answer:  21, 9, 33.   P. 291 is part of Regiomontanus's working for this letter where he solves the problem by assuming the first has  x,  so the second has  30 - x  and the third has  54 - x,  hence   84 - 2x  =  42.

Benedetto da Firenze.  c1465.

P. 67:  III‑(12, 14, 10).

P. 68:  III‑(10, 8, 15, 12).

     Chap. 22:  "... huomini che ànno danari et trovano borse di danari", pp. 181‑183.

I‑(2, 3),  answer:  3, 4; 5.  (= Mahavira v. 244.)

II‑(2, 3, 4),  answer:  9, 16, 13; 23.  (= Fibonacci, p. 216.)

I‑(2, 3, 4),  answer:  7, 17, 23; 73.   (= Iamblichus's first.)

Muscarello.  1478.  Ff. 65r-65v, pp. 172-173.  Men find a purse,  I‑(3, 4, 5).  Answer:  7, 13, 17; 83.

della Francesca.  Trattato.  c1480. 

F. 42v (105-106).  This is a type I problem with four values and no purse, except the last value is specified.  The multipliers are  7/12 (given as  1/3 + 1/4),  9/20, 11/30  and  z = 8.  So the first equation is  w = (7/12)(x+y+z).  We could interpret  7z/12  as a purse of negative value, but it is easier to write the first equation as  w = (7/12)(T-w),  so  w = (7/19)T  which leads to 

T =  (7/19 + 9/29 + 11/41)T  + 8.  Answer:  (66584, 56088, 48488, 9568)/1196.

F. 120v (257).  III-(20, 30, 40).  Answer:  25, 25, 5.

Chuquet.  1484. 

     Triparty, part I.  FHM  80-81 & 83-85.  Sesiano cites pp. 642 & 646 of Chuquet, ??NYS. 

III-(120, 90, 80, 75).   Answer:  (5, 95, 125, 140)/3.   Not mentioned by Sesiano; English in FHM 80, where the next two cases are just given as formulae with solutions.

III-(120, 90, 80, 75, 72).   Answer:  (-43, 77, 117, 137, 149)/4.   Sesiano notes that this is the same as the solution of the problem in 7.R.2 in the Provençale Arithmétique, c1430, and in Pellos, 1492.

III-(120, 180, 240, 300, 360).   Answer:  180, 120, 60, 0, -60.   Sesiano notes these numbers occur in the solution of the problem on p. 641 -- see 7.R.2.

FHM 83 gives the next four just by formulae with solutions.

II-(2, 3, 4)  with purse worth  40.  Answer:  (280, 680, 920)/73.

I-(3, 4, 5, 6)  with purse worth  50.  Answer:  (200, 3350, 5450, 6950)/941.

I-(2, 3, 4, 5)  with purse worth  26.  Answer:  (‑78, 312, 546, 702)/123.  FHM puts down  +78  and hence misses this interesting example which Sesiano discusses.

Version giving equations like   w + z  =  2 (x + y)  =  26⅔,   but this is a bit far away from the problems of this section.

FHM 83-84: Indeterminate version of the Bloom of Thymarides:   w + x  =  17,  w + y  =  18,  w + z  =  19.   He solves by setting  w = 12,  and says any other value less than  17  can be used "wherefor such calculations may have innumerable responses."

     Appendice.

Prob. 81.  III-(19, 23, 30).   Answer:  17, 13, 6.

Prob. 82, English in FHM 212-213.  (3, 5).   Answers:  4, 6; 14  and  (60, 90; 210)/7.  Says "such questions do not have a necessary answer."

Borghi.  Arithmetica.  1491? -- this material is additional to the 1484 ed. and Rara indicates that the first ed. to have 100ff is the 4th of 1491.  The folio numbers are from the 1509 ed.

Ff. 99r-99v.  Men find a purse,  I-(3, 4).   = Fibonacci p. 212.

Ff. 99v-100r.  Men find a purse,  I-(2, 3, 4).   = Fibonacci p. 213.  = Iamblichus' 1st.

F. 100r.  I-(3, 4)  with purse worth  16.

F. 100r.  I-(2, 3, 4)  with purse worth  30.

Calandri.  Arimethrica.  1491. 

F. 65v.  III-(12, 14, 10).

F. 65v.  I-(6, 40).  Answer:  7, 41; 239.

Pacioli.  Summa.  1494.  Some of his problems mix this with 7.R.2.

F. 105v, prob. 19.  Two men find two purses of values  p+10, p,  giving equations:  x + p+10 + 10  =  4 (y - 10),   y + p + 20  =  5 (x - 20).  He assumes the purses are worth  100  in total, so  p = 45,  p+10 = 55.   Answer:  (765, 690)/19.   = Tonstall, pp. 245-246.

F. 190v, prob. 22.  I-(3, 4).   Answer:  4, 5; 11.

Ff. 190v-191r, prob. 23.   x + p  =  2y + 2,   y + p  =  4x + 2.   Answer:  12, 20; 30.

Ff. 192r-192v, prob. 28.  I-(2, 3, 4).   He assumes  p = 10  and gets answer:  (70, 170, 230; 730)/73.

F. 192v, prob. 30.  I-(3/2, 7/3, 15/4).  Assumes  p = 1  and gets answer:  (63, 177, 279; 621)/621.

F. 192v, prob. 31.  I-(2, 3, 4).  Doesn't observe that this  = prob. 28.  Answer:  (7, 17, 23; 73)/73.

F. 193v, prob. 39 & 40.  III-(35, 32, 27).   III-(122, 114, 106, 96).   Says one can deal similarly with more people.  Prob. 40 discusses the point further.

Ff. 193v-194r, prob. 41.  3  men find a purse and want to buy a horse, giving:  x + y + p/3  =  h,   y + z + p/5  =  h,   z + x + p/4  =  h.  If   T  =  x + y + z,   one gets   T + 47p/60  =  3h   and the solution space is actually two dimensional.  He assumes  p = 60,  h = 47  and this problem reduces to prob. 39.  Cf della Francesca 40v in 7.R.2.

Ghaligai.  Practica D'Arithmetica.  1521.  He gives several versions, of increasing complexity -- the later ones involve various numbers in geometric progressions or using roots and I omit these.

Prob. 13, f. 103r.  I-(6, 10).   Answer:  (42, 66; 354)/7.   He arbitrarily sets the first man's money at  6.

Prob. 14, ff. 103r-103v.  I-(4, 6),  but with total money  = 100.  Answer:  (100, 140, 460)/7.

Prob. 16, ff. 103v-104r.  Two men find purses worth  p  and  p+13,  with coefficients  2  and  3,   i.e.   x + p  =  2y,   y + p+13  =  3x.   Answer:  (339, 400; 461)/12.

Tonstall.  De Arte Supputandi.  1522. 

Quest. 42,  pp. 172-173.  III-(150, 240, 326).  Considers four person case and sketches general solution.

Pp. 245-246.  Same as Pacioli's prob. 19, but clearly says the two purses are worth  100.

Riese.  Die Coss.  1524.

No. 45, p. 46.  2 men find a purse,  (2, 5).

No. 65, p. 49.  2 men find a purse, leading to:   x + p + 1  =  b ‑ 1,   y + p + 4  =  3 (x ‑ 4).   He assumes  p = 2,  but the general solution is  b  =  2a ‑ 7,  p  =  a ‑ 9.

Apianus.  Kauffmanss Rechnung.  1527. 

F. M.iii.v.  I-(½, 3)  with  p = 30.  Answer:  60, 160.

F. M.iv.r.  III-(44, 36, 30).  Answer:  11, 19, 25.

Cardan.  Practica Arithmetice.  1539.  Chap. 66.

Section 97, ff. HH.v.r - HH.vi.r (p. 169).  III-(34, 73, 72, 88).

Section 98, ff. HH.vi.r - HH.vi.v (p. 169).  Men find a purse and buy a horse, giving:   x + y + p/2  =  y + z + p/5  =  x + z + p/3  =  h.   Answer:  6, 10, 15; 30, 31.

Recorde.  Second Part.  1552.  Pp. 320-322: A question of debt, the third example.  IV‑(47, 88, 71).  Answer:  15, 32, 56.

Tartaglia.  General Trattato.  1556.  Book 16, art. 28‑35 & 40, pp. 243r‑245r.

Art. 28.  I‑(2, 3, 4).   Answer:  7, 17, 23; 73.  (= Iamblichus' 1st.)

Art. 29.  I‑(3, 4, 5).   Answer:  7, 13, 17; 83.

Art. 30.  I‑(4, 6, 8).   Answer:  17, 53, 73; 487.

Art. 31.  Same as 30 with purse given as  100.

Art. 32.  Same as 30 with total given as  1200.

Art. 33.  II‑(2, 3, 4).   Answer:  9, 16, 13; 23.  (= Fibonacci, p. 216.)

Art. 34.  II‑(4, 5, 6).   Answer:  25, 36, 31; 119.

Art. 35.  II‑(2, 3, 4, 5).   Answer:  33, 76, 65, 46; 119.  (= Fibonacci, pp. 218‑220  &  Lucca 1754.)

Art. 40.  III‑(24, 28, 32, 36).

Buteo.  Logistica.  1559.  Prob. 9, p. 209-210.  III-(4900, 3760, 4660).   Remarks on the case with four people.

Schott.  1674. 

Ænigma X, p. 556.  Two cups and a cover (equivalent to a purse) worth  90.  (2, 3).

Ænigma II, p. 563.  I-(2, 3, 4)  with purses worth  136, 184, 176.  Answer:  24, 32, 48.

Wells.  1698.  No. 117, p. 208.  Two horses and equipage:  (1, 2)  with equipage (equivalent to a purse) worth  5. 

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. V, pp. 79-80 (1790: prob. VII, p. 80).  III-(17, 16, 15).   1745 gives two methods of solution;  1790 gives one.

Mair.  1765?  Pp. 458-459, ex. 8.  Two horses:  (2, 3)  with saddle worth  50.

Vyse.  Tutor's Guide.  1771?  Prob. 27, 1793: pp. 56-57; 1799: pp. 61-62 & Key p. 67.   III‑(33000, 30000, 32000, 28000, 25000).

Dodson.  Math. Repository.  1775.

P. 6, Quest. XV.  III-(30, 36, 40).

P. 19, Quest. L.  Find  x/y  such that  (x+1)/y  =  1/3,  x/(y+1)  =  1/4,  i.e. 

V-(1, 1/3; 1, 1/4).  [(x+1)/y  =  1/n;  x/(y+1)  =  1/(n+1)  has solution  x = n+1,  y = (n+1)2 - 1.  In general,  (x+1)/y  =  a/b,  x/(y+1)  =  c/d,  i.e.  V-(1, a/b; 1, c/d)    gives         x = c(a+b)/(ad-bc),  y = b(c+d)/(ad-bc).  One would normally assume  a/b > c/d.  Cf Wolff in 7.R.3 for a different phrasing of the same problem.]

P. 38, Quest LXXXIX.  I-(1, 2, 3)  with purse worth 55.  The context is three horses and a saddle.  The first horse, with the saddle, is worth as much as the other two horses, etc.  Answer:  5, 25, 35.

P. 64, Quest. CXX.  Multiplicative form of type III.  xy = 12,  xz = 18,  yz = 24.  He doesn't multiply the equations to get  (xyz)2 = 722.  Answer:  3, 4, 6.

P. 75, Quest. CXL.  Like the previous, with four variables.  wxy = 252,  wxz = 432,  wyz = 756,  xyz = 336.  Here he does multiply them together to find  wxyz = 3024.  Answer:  9, 4, 7, 12.

Hutton.  A Course of Mathematics.  1798?  Prob. 30,  1833: 222;  1857: 226.  Two horses with saddle.  I‑(2, 3)  with saddle worth  50.  Answer:  30, 40; 50.  (=  10 times Mahavira v. 244.)

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 115-117 & 256, nos. 431-442.  All of these have different purses except no. 435, which is  I-(½, ⅓)  with a purse of  -10,  i.e.  10  is taken away rather than gained.  No. 432 uses the context of ages.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 2, pp. 172-173 (1868: 184).  I-(1, 2, 3)  with purse of  11.  Answer:  1, 5, 7; 11.  See AR 113 & 227.

Todhunter.  Algebra, 5th ed.  1870.  Section XIII, art. 185; Example XIII, nos. 1, 3, 5,  pp. 99-100 & 578.

Art. 185.  V-(6, 3/4; -2, 1/2).  Answer:  9/20.

No. 1.  V-(3, 1; 2, 1/2).  Answer:  5/8.

No. 3.  V-(1, 1/3; 1, 1/4.  Answer:  4/15.  = Dodson, p. 19.

No. 5.  V-(36, 3; 5, 2).  Answer:  42/26.

Mittenzwey.  1880. 

Prob. 295, pp. 54 & 105;  1895?: 325, pp. 57 & 106;  1917: 325, pp. 52 & 100.  III‑(89, 82, 97),  in context of measuring pairs of edges of a triangle.  Answer:  45, 52, 37.

1895?: prob. 93, pp. 21 & 69-70;  1917: 93, pp. 19 & 66-67.  III-(23. 19, 10).  Answer:  3, 7, 16.  Algebraic solution by elimination given.

Clark.  Mental Nuts.  1897, no. 75.  The horses and saddle.  I-(1, 2, 3)  with the saddle acting as the purse and the total value of the horses and saddle being  220.  Answer:  (1, 5, 7; 11) * 55/6.  See AR 113 & 227; Parlour Pastime.

Haldeman-Julius.  1937.  No. 23: The coin problem, pp. 5 & 21.  Like two men finding a purse of 25.  First says: If I had the purse, I'd have  3  times what you have.  Second says: And if I had the purse, I'd have five less than half of what you have.  Answer:  230, 85.

Ken Russell & Philip Carter.  Intelligent Puzzles.  Foulsham, Slough, 1992.  Prob. 57, pp. 45 & Answer 75, p. 164: Money.  Jack has 75p and ¾ of what Jill has;  Jill has 50p and ½ of what Jack has.  This is a straightforward problem, but it is also a concealed version of  I‑(4/3, 2)  with a purse of 100.

 

          7.R.2. "IF I HAD  1/3  OF YOUR MONEY, I COULD BUY THE HORSE"

 

            See Tropfke 608.

            For related problems, see 7.H.4.

            NOTATION.  Again there are two forms.

I‑(a1, a2, ..., an)  -- i-th says "If I had  ai  of what the rest of you have, then I could buy the

            horse".

II‑(a1, a2, ..., an)  -- i-th says "If I had  ai  of what the i+1-st has, then I could buy the horse".

            The problems are the same when there are only two people and I will label two person cases as form I.

            The problem often replaces horse by house, ship, etc.  In some cases, the value of the horse, house, etc. is given.  In some cases the value is given with no reference to anything bought and I say  "with  h = ..."  to indicate the value.  Some problems have different values of horses.  If values are simply given, I say  "with  h1, h2 = ..."  or  "with  hi = ...".  See:  Fibonacci,  Gherardi,  Lucca 1754,  AR,  Benedetto da Firenze,  della Francesca,  Pacioli,  Riese: Coss,  Tartaglia,  Buteo,  Schott,  Simpson  for examples where there are different values of horses.

            Solution notation as in 7.R.1 with the horse last.

            Let  h  be the value of the horse and let  T  be the total of the amounts.  Then  I‑(a1,a2,...)  with  n  people has  n  equations   xi + ai(T ‑ xi)  =  h,   so   xi  =  (h - aiT)/(1-ai).   Adding these for all  i  gives one equation in the two unknowns  T  and  h.  However, letting  C = T - h  leads to the simplest equation:   (n‑1)T  =  [Σ 1/(1‑ai)] C.

            For  II‑(a1,a2,...),  systematic elimination in the  n  equations  xi + aixi+1 = h  leads to   x1 [1 + (-1)n+1a1a2...an]  =  h [1 - a1 + a1a2 - ... + (-1)na1a2...an],  and any other value can be found by shifting the starting point of the cycle.

            In either case, the solution can be adapted to variable purses -- see 7.R.  Taking negative values of  h  and all  ai  converts this into 7.R.1 -- men find a purse, which is slightly easier to deal with.

 

                        INDEX OF SOME COMMON TYPES.

            I have recently compiled this and was surprised to see how much repetition is present.

 

I‑(⅔, ¾).  Riese: Rechnung;  A Lover of the Mathematics;  Euler IV.618;  Vyse; 

I‑(½, ⅔).  Chiu Chang Suan Ching;  Sun Zi;  AR 176; 

I-(½, ⅓).  della Francesca 16r;  Peurbach;  Lacroix; 

I-(⅓, ¼).  al-Karkhi I-42;  Fibonacci 228;  BR 7;  AR 171-175, 177, 221; 

          della Francesca 36v;  Pacioli 190v;  Riese: Rechnung;  Schott 562-563; 

I-(1/3, 1/5).  Calandri;  Calandri: Raccolta; 

I‑(1/6, 1/7).  Benedetto da Firenze; 

 

I-(½, ⅔, ¾).  della Francesca 21v; 

I‑(½, ⅓, ¼).  Gherardi 46-47;  Lucca 1754 58r;  Munich 14684;  Folkerts; 

          Provençale Arithmétique;  AR 178, 224, 340;  Benedetto da Firenze; 

          della Francesca 17v-18r, 39r;  Chuquet;  Borghi;  Pacioli 105v-106r, 192r, 192v; 

          Tonstall;  Riese: Coss 122, 123;  Buteo 81;  Les Amusemens;  Euler III.19, IV.622; 

I‑(1/3, 1/4, 1/5).  Diophantos 24;  al-Karkhi III-26;  Fibonacci 245;  AR 341;  Buteo 192-193; 

          Schott 563; 

 

I-(2/3, 3/4, 4/5, 5/6).  Chuquet; 

I‑(1/2, 1/3, 1/4, 1/5).  Fibonacci 245-248.  de Nemore II-27;  Provençale Arithmétique; 

          Riese: Coss 124-126;  Pearson; 

I‑(1/3, 1/4, 1/5, 1/6).  Diophantos 25;  al-Karkhi III-27;

 

I-(1/2, 2/3, 3/4, 4/5, 5/6).  Chuquet; 

I‑(1/2, 1/3, 1/4, 1/5, 1/6).  Provençale Arithmétique;

 

 

II‑(3/4, 4/5, 5/6).  Riese: Coss 121; 

II‑(½, ⅔, ¾).  AR 180;  Chuquet;  Tartaglia 22; 

II‑(½, ⅓, ¼).  Lucca 1754 58r;  AR 179, 223, 339;  della Francesca 36v-37r;  Chuquet; 

          Riese: Rechnung;  Riese: Coss 31, 47, 120;  Peurbach;  Tartaglia 40, 41;  Buteo 81; 

          Euler IV.619-629; 

II‑(1/3, 1/4, 1/5).  Fibonacci 229;  BR 114;  AR 156, 157, 338;  della Francesca 17v;  Chuquet; 

II‑(1/3, 1/5, 1/4).  al-Karkhi III-33; 

 

II‑(1/2, 1/3, 1/4, 1/5).  de Nemore II-25;  Gherardi 45-46

II‑(1/3, 1/4, 1/5, 1/6).  Fibonacci 231-232.

 

II‑(1/2, 1/3, 1/4, 1/5, 1/6).  Fibonacci 327-329.

II‑(1/3, 1/4, 1/5, 1/6, 1/7).  Fibonacci 234.

 

II‑(2/3, 3/4, 4/5, 5/6, 6/7, 7/8, 8/9).  Riese: Coss 140; 

 

Buying a horse with a friend or a found purse.  Lucca 1754 61r-61v;  della Francesca 40v; 

          Pacioli 193v-194r;  Cardan; 

 

Jens Høyrup.  Sub-scientific mathematics: Undercurrents and missing links in the mathematical technology of the Hellenistic and Roman world.  Preprint from Roskilde University, Institute of Communication Research, Educational Research and Theory of Science, 1990, Nr. 3.  (Written for: Aufsteig und Niedergang der römischen Welt, vol. II 37,3 [??].)  P. 18 quotes Plato's Republic, 333b-c, "to buy in common or sell a horse" and feels this may be a reference to this type of problem.  (This seems a bit far-fetched as there are many simpler types of commercial problem which could be described by this phrase.)

Chiu Chang Suan Ching.  c‑150.  Chap. VIII.

Prob. 10, p. 86.  I‑(½, ⅔)  with  h = 50.  Answer:  37½, 25.

Prob. 12, p. 87.  x + 3z = 40,  z + 2y = 40,  y + x = 40.  Equivalent to  II‑(3, 2, 1)  with  h = 40.  Answer:  x, z, y  =  (160, 40, 120)/7.

Prob. 13, pp. 87‑88.  Equivalent to  II‑(6, 5, 4, 3, 2).  Answer given is  265, 76, 129, 148, 191; 721  and this is the least integral solution.

Diophantos.  Arithmetica.  c250.  Book I.

No. 22, p. 138:  "To find three numbers such that, if each give to the next following a given fraction of itself, in order, the results after each has given and taken may be equal."  However, he says that the equal amounts are the results after both giving and taking, i.e.   (1 ‑ b) y + ax  =  (1 ‑ c) z + by  =  (1 ‑ a) x + cz.   He does  a, b, c  =  1/3, 1/4, 1/5.   Answer:  6, 4, 5.

No. 23, pp. 138‑139 is the same with four people and numbers  1/3, 1/4, 1/5, 1/6.  Answer:  150, 92, 120, 114.

No. 24, p. 139:  "To find three numbers such that, if each receives a given fraction of the sum of the other two, the results are all equal."  Does  I‑(1/3, 1/4, 1/5).   Answer:  13, 17, 19; 25.

No. 25, pp. 139‑140.  Same with four numbers.  Does  I‑(1/3, 1/4, 1/5, 1/6).   Answer:  47, 77, 92, 101; 137.

Sun Zi.  Sun Zi Suan Ching.  Op. cit. in 7.P.2.  4C.  ??NYS.  Chap. III, no. 28.   I-(½, ⅔)  with  h = 48  -- like Chiu Chang Suan Ching, prob. 10.  (English in Lam & Shen, HM 16 (1989) 117.)

See 7.P.4 -- Bakhshali MS for a problem which is related, leading to   x1/2 + x2 + x3 + x4 + x5  =  h,  etc.,  where  h  is the price of a jewel.  Solution: 120, 90, 80, 75, 72; 377.  Also an example with three values and diagonal coefficients   -7/12,  -3/4,  -5/6   and solution:  924, 836, 798; 1095.

Sesiano cites Abū Kāmil's Algebra and al-Karajī's Kāfī, ??NYS.  Hermelink, op. cit. in 3.A, cites Kāfī  &  Beha-Eddin, ??NYS

al‑Karkhi.  c1010.

     Sect I, no. 42‑43, p. 80.

42:  x + ⅓y  =  20  =  y + ¼x.  I.e.  I-(⅓, ¼)  with  h = 20.  Answer:  (160, 180)/11.

43:  x + ⅓y + 5  =  20  =  y + ¼x + 6.  Equivalent to  I-(⅓, ¼)  with  hi = 15, 14.  Answer:  (124, 123)/11.

     Sect. III, no. 26‑27, 32‑35, pp. 95‑100.

26:  I‑(1/3, 1/4, 1/5)  with horse worth  20.   Answer:  (52, 68, 76)/5.   See Diophantos I 24.

27:  I‑(1/3, 1/4, 1/5, 1/6).   Answer:  47, 77, 92, 101; 137.   (= Diophantos I 25.)

32:  x  =  y + ⅓z,   y  =  z + ⅓x,   z  =  x + ⅓y.

33:  II‑(1/3, 1/5, 1/4).   Answer:  (44, 51, 50; 61)/11.

34:  = Diophantos I 22.

35:  = Diophantos I 23, but with answer divided by  23,  making  y = 4.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 133, no. 52.  ??NYS -- Hermelink, op. cit. in 3.A, mentions this without details.  Tropfke 609 cites this problem and also p. 150f., no. 13, as buying a horse.

Fibonacci.  1202.  Pp. 228‑258 (S: 337-372): chap. 12, part 5: De emptione equorum inter consocios secundum datam proportionem [On the purchase of horses among partners according to some given proportion].  Many examples, getting up to seven men, up to five horses, an inconsistent example and negative solutions.  I have omitted some of the variations and some of the more complex problems.  Some of the problems cited are followed by some discussion.

                    See:  K. Vogel; Zur Geschichte der linearen Gleichungen mit mehreren Unbekannten; Deutsche Mathematik 5 (1940) 217‑240;  for a thorough study of the problems on pp. 228‑258.  Some further versions occur on pp. 327‑349 (S: 458-484).  Vogel cites several sources, ??NYS:  Codex lat. Monacensis 14908 (1456)  and, for a simpler type of problem,  Mich. Pap. 620.

P. 228 (S: 337).  I‑(⅓, ¼).   Answer:  8, 9; 11.  Cf al-Karkhi I-42.

P. 229 (S: 338).  II‑(1/3, 1/4, 1/5).   Answer:  45, 48, 52; 61.

P. 231 (S: 341).  II‑(2/3, 4/7, 5/9).   Answer:  135, 141, 154; 229.

Pp. 231‑232 (S: 341-342).  II‑(1/3, 1/4, 1/5, 1/6).   Answer:  264, 285, 296, 315; 359.

Pp. 234 (S: 344-345).  II‑(1/3, 1/4, 1/5, 1/6, 1/7).   Answer:  1855, 1998, 2092, 2145, 2156; 2521.  The margin has  1815  for  1855  and  2256  for  2156.

Pp. 234‑235 (S: 345-346).  II‑(2/3, 4/7, 5/11, 6/13, 8/19).   Answer:  35435, 35313, 41712, 38643, 44057; 58977.

Pp. 235 (S: 346).  II‑(1/4 + 1/3,  1/5 + 1/4,  1/6 + 1/5,  1/7 + 1/6).  Answer:  176274, 200772, 205820, 238830; 293391.

Pp. 235‑236 (S: 346-347).  Two men, two horses of values  h1, h2  =  h, h+2.   i‑th says  "If I had  ai  of what the  i+1-st has, then I could buy the  i-th horse",  with  A = (⅓, ¼).  I.e.  II-(⅓, ¼)  with  hi = h, h+2.  Gives the first two solutions:  8, 12; 12  and  16, 21; 23.  Varies to  h2 = h+3  and gives the solution  20, 27; 29,   which is the third positive solution of the problem.

Pp. 236‑240 (S: 347-352).  Versions with  n  men and  n  horses,  n = 3, 4, 5.

Pp. 240‑242 (S: 352-354).  Four men, one horse, giving   w + ⅓ (x + y)  =  ...  =  h,   with constants  1/3, 1/4, 1/5, 1/6.   Answer:  187, 209, 247, 273; 339.

Pp. 242‑243 (S: 354-355).  Three men, one horse, giving   x + y + z/3  =  ...  =  h,   with constants  1/3, 1/4, 1/5.   Answer:  15, 16, 18; 37.  (He interchanges roles of  x  and  y.)

P. 243 (S: 355-356).  Same with four men and constants  1/3, 1/4, 1/5, 1/6.  Answer:  15, 18, 15, 20; 38.   The margin has  28  for  38.

Pp. 243‑244 (S: 356-357).  Five men, one horse, giving   v + w + x + ¼ y  =  ...  =  h.   Answer:  1218, 1295, 1200, 1260, 1365; 4028.   The margin has  1269  for  1260.

P. 245 (S: 357-358).  I‑(1/3, 1/4, 1/5).   Answer:  13, 17, 19; 25.   (= Diophantos I 24.)

Pp. 245‑248 (S: 358-360).  I‑(1/2, 1/3, 1/4, 1/5).   Answer:  1, 19, 25, 28; 37.

Pp. 248‑249 (S: 361-362).  I‑(2/5, 3/8, 4/11, 6/19).   Answer:  1774, 2047, 2164, 2614; 4504.   Margin has  2164  for  2614.  States a variation:   I‑(1/3 + 1/4,  1/4 + 1/5,  1/5 + 1/6,  1/6 + 1/7)   with answer:  1376, 54272, 76022, 87902; 128657.

Pp. 249‑250 (S: 362-364): Questio nobis proposita a peritissimo magistro musco constantinopolitano in constantinopoli [A problem proposed to us by a most learned master of a Constantinople mosque].  Buying a ship. 

I‑(1/5 + 2/3,  1/480 + 1/6 + 2/3,  1/638 + 1/6 + 2/3,  1/420 + 1/7 + 2/3,  1/810 + 1/27 + 1/10 + 2/3).  Answer:  3, 228, 231, 348, 378; 1030.

Pp. 250‑251 (S: 364-365).  Five men, giving   v + w + ½ (x + y + z)  =  ...  =  h,   with constants  1/2, 1/3, 1/4, 1/5, 1/6.   Answer:  58, 19, 148, 49; 163.

Pp. 251‑252 (S: 365-366): Questio insolubilis [An unsolvable problem].  Four men, giving equations   w + x + ½ (y + z)  =  ...  =  h,   with constants  1/2, 1/3, 1/4, 1/5.  This is inconsistent unless  h = 0.  He then varies the constants to  1/2, 3/7, 3/11, 5/13,   which gives answer:  5, 6, 7, 9; 19.

Pp. 252‑253 (S: 366-367).  Seven men, giving   t + u + v + ½ (w + x + y + z)  = ...  =  h,   with constants  ½, ⅓, ..., ⅛.   Answer:  507, 171, ‑9, 1347, 451, 131, 1431; 2349.  "... quare tercius homo habet debitum ipsos  9,  vel hec questio est insolubilis: sit itaque solubilis cum debito tercii hominis; ..." [... therefore the third man has a debit of 9 bezants, or this problem is unsolvable; ...].  He then varies the constants to  1/3, 1/4, ..., 1/9,  with answer:  1077, 717, 489, 1637, 997, 657, 1749; 3963.  See Sesiano.

Pp. 253‑254 (S: 367-368).  Two men and two horses of values  h1, h2 = h, h+2.  i-th says  "If I had  ai  of what the rest of you have, then I could buy the  i-th horse",  with constants  ⅓, ¼.  (This is the same as on pp. 235‑236 above, but only because  n = 2.)

P. 254 (S: 368).  Same with  3  men,  3  horses worth  h, h+2, h+4  and constants  1/3, 1/4, 1/5.  I.e.  I-(1/3, 1/4, 1/5)  with  hi = h, h+2, h+4.  Answer:  7, 13, 17; 17.

Pp. 254‑257 (S: 369-371).  Same with  4  men,  4  horses worth  h, h+3, h+7, h+12  and constants  1/3, 1/4, 1/5, 1/6.   I.e.  I-(1/3, 1/4, 1/5, 1/6)  with 

hi = h, h+3, h+7, h+12.  Answer:  (‑4, 13, 27, 41; 23)/2.   "Unde hec questio cum hiis iiii‑or positis residuis solui non potest, nisi primus homo haberet debitum." [Whence this problem with these IIII posed residues can be solved with the first man having a debit. (Sigler), but it would be more literal to have  Whence this problem with these IIII posed residues can not be solved unless the first man has a debit.].  He later gives another solution:  (82, 193, 265, 325; 343)/6,   by choosing a larger value for  h.  The margin has  32 1/2  for  32 1/6  which I have converted to  193/6.  See Sesiano.

Pp. 257‑258 (S: 371-372).  Four men, one horse, giving   w + x/2 + y/3 + z/4  =  x + y/3 + z/4 + w/5  =  ...  =  h.   Answer:  105, 168, 210, 240; 319.

Pp. 327‑329 (S: 458-460).  II‑(1/2, 1/3, 1/4, 1/5, 1/6)  done by false position.  Answer:  456, 530, 573, 592, 645; 721  -- the text has  529  instead of  592.  Sesiano notes that negatives are used in one of the false positions.

Pp. 334‑335 (S: 466-467).  Problem of pp. 245‑248 done by false position.

Pp. 336-338 (S: 469-470).  (⅓, ¼)  with  hi = 14, 17.   Answer:  (100, 162)/11.

P. 338 (S: 470-471).  II-(1/3, 1/4, 1/5)  with  hi = 14, 17, 19,  done by false position.   Answer:  (595, 777, 1040)/61.

Pp. 338-339 (S: 471-472).  I-(1/3, 1/4, 1/5)  with  hi = 14, 17, 19,  done by false position.   Answer:  (241, 594, 783)/50.

Pp. 347‑349 (S: 481-483).  Three men, one horse, giving   x + y/2 + z/3  =  y + z/4 + x/5  =  z + x/6 + y/7  =  h.  Answer:  1530, 3038, 3540; 4229.  Done two ways.

P. 349 (S: 484).  Four men, one horse, giving   w + x/2 + y/3 + z/4  =  x + y/4 + z/5 + w/6  =  y + z/6 + w/7 + x/8  =  z + w/8 + x/9 + z/10  =  h.  Answer:  8569848, 21741336, 26955060, 29657460; 35839901.

Fibonacci.  Flos.  c1225.  In:  Picutti, pp. 312-316 & 320-326, numbers III, VI & VII.

Pp. 236-238: De quinque numeris reperiendis ex proportionibus datis.  Five values:   v + (w + x + y)/3  =  w + (x + y + z)/3  =  ....   Answer:  7, 10, 19, 25, 28; 34.

Pp. 240-242: No heading -- paragraph begins:  Item de mode predicto extraxi ....  I‑(1/3, 1/4, 1/5)  with  hi = 14, 17, 19.   Answer:  (241, 594, 783)/50.

Pp. 242-243: De quatuor hominibus bizantios habentibus.  He refers to Liber Abaci, apparently to pp. 338-339.  I‑(1/2, 1/3, 1/4, 1/5)  with horses worth  33, 35, 36, 37.   Answer:  ‑3, 18, 25, 29.   "... hanc insolubilem esse sub posita conditione."  On p. 243, he states that if the values of the horses are  181, 183, 184, 185,   then the answer is  1, 94, 125, 141.  See Sesiano, who notes that for  hi = h, h+2, h+3, h+4,  the smallest positive integral solution is that given by Fibonacci.

Fibonacci.  Epistola.  c1225.  In Picutti, pp. 338-340, numbers XV & XVI.

Pp. 250-251: Modus alius solvendi similes questiones.  I‑(1/2, 1/3, 1/4, 1/5, 1/6)  with horses worth  12, 15, 18, 20, 23.   Answer:  (4938, 7428, 10161, 11268, 15760)/721.

Pp. 251-252: Investigatio unde procedat inventio suprascripsit.  II‑(1/2, 1/3, 1/4, 1/5, 1/6)  with horses worth  12, 15, 18, 20, 23.   Answer:  (‑5316, 1479, 4532, 6157, 7920)/394.   He says  "tunc questio esset insolubilis, nisi concederetur, primum habere debitum; quod debitum esset [5316/394]."  See Sesiano.

Jordanus de Nemore.  c1225.  Op. cit. in 7.R.1.

Prob. II‑25, p. 151.  General version of type II with value of horse given.  Example:  II‑(1/2, 1/3, 1/4, 1/5)  with horse worth  119.  Answer:  75, 88, 93, 104.

Prob. II‑26/28, pp. 152‑155.  General version of type I with value of horse given.  Example in II‑26:  I‑(3, 13/4, 25/7, 4)  with horse worth  28.  Answer:  1, 2, 3, 4.  Example in II‑27:  I‑(1/2, 1/3, 1/4, 1/5)  with horse worth  37.  = Fibonacci

                                                  245‑248.  Answer:  1, 19, 25, 25.  

II‑28 is II‑26  done in a different way.

BR.  c1305.

No. 6, pp. 26‑27.  Like Fibonacci's pp. 242‑243 with constants  1/5, 1/7, 1/9  and  h = 100.  Method is wrong, but the answer is right:  (2700, 2800, 3000)/61.

No. 7, pp. 26‑29.  Buying a ship,  I-(⅓, ¼)  with ship worth  100.  Answer:  (800, 900)/11.  Cf al-Karkhi I-24.

No. 8, pp. 28‑29.  Buying a business,  I-(2/3, 3/5)  with business worth  100.  Answer:  (215, 258)/9.

No. 50, pp. 66‑69.  Same as no. 6 with constants  (1/3, 1/4, 1/5)  and  h = 11.  Answer:  45, 48, 54.

No. 55, pp. 72‑75.  Buying a business worth  20,  I-(4/9, 12/35).  Answer:  (10500, 12420)/801.

No. 56, pp. 74‑77.  Same,  I-(12/35, 13/42),  with  h = 72.  Answer:  (69552, 73080)/1314.

No. 57, pp. 78‑81.  Same,  I-(71/105, 37/60),  with  h = 50.  Answer:  (102000, 120750)/3673.

No. 58, pp. 80‑83.  A  gives  B  7/12  of what  A  has, then  B  returns  9/20  of what he has, then both have  12.   Answer:  (288, 1032)/55.

No. 72, pp. 94‑97.  Same with constants  1/7, 1/4,  and both finish with  36.  Answer:  28, 44.

No. 114, pp. 130‑131.  (= Fibonacci, p. 229.)

Gherardi.  Libro di ragioni.  1328.

P. 42.  Same as Fibonacci, pp. 235‑236, with  hi = 10, 12,  A = (⅓, ¼).   Answer:  (72, 114)/11.

Pp. 45‑46.  Chopera.  II‑(1/2, 1/3, 1/4, 1/5).   Assumes house is worth  60.   Answer is  60/119  times de Nemore's II‑25.

Pp. 46‑47.  I‑(½, ⅓, ¼).   Answer:  5, 11, 13; 17.

Pp. 59‑60.  Three men and three horses:  I‑(1/3, 1/4, 1/5)  with  hi = 40, 47, 55.   Answer:  (702, 1602, 2289)/50.

Lucca 1754.  c1330.

F. 58r, p. 131.  Buying a horse.  II‑(½, ⅓, ¼).   Answer:  16, 18, 21; 25.

F. 58r, pp. 131‑132.  Buying a horse.  I‑(½, ⅓, ¼).   Answer:  5, 11, 13; 17.  Cf Gherardi, pp. 46-47.

F. 58v, p.132.  Two men.  First says:  "If you give me    of your money, then I can buy  20  horses."   Second says:  "If you give me  ¼  of your money, then I can buy  21  horses."  Answer:  156, 192  with horses worth  11  each.

Ff. 61r‑61v, p. 141.  Three men and a friend.  i-th says:  "If I had  ai  of our friend's money, I could buy the horse",  with  (ai) = (½, ⅓, ¼).  This has a two dimensional solution space.  He gives only  1/2,  5/2,  7/2  with friend having  12  and horse worth  13/2.

Munich 14684.  14C.  Prob. XVII, p. 80.  I‑(½, ⅓, ¼).   Answer:  10, 22, 26; 34.  Cf Gherardi, pp. 46-47.

Folkerts.  Aufgabensammlungen.  13-15C.  He calls it Sperberkauf (sparrow hawk purchase).  11 sources for  I‑(½, ⅓, ¼)  = Munich 14684 = Gherardi, pp. 46-47, none of which give a derivation.  It's not clear if  h = 34  is given in some cases.  Numerous other citations.

Provençale Arithmétique.  c1430.  Op. cit. in 7.E.  F. 100r-101v, pp. 49-53.

Three men buy a horse:  I‑(½, ⅓, ¼).   Answer:  5, 11, 13; 17.  Cf Gherardi, pp. 46-47.

Four men buy a horse:  I‑(1/2, 1/3, 1/4, 1/5)  (= Fibonacci, 245-248).  The answer here takes  C = 60,  which gives:  (5, 95, 125, 140; 185)/3

Five men buy a piece of cloth:  I‑(1/2, 1/3, 1/4, 1/5, 1/6).   Answer:  (‑43, 77, 117, 137, 149; 197)/4.  Sesiano, loc. cit. in 7.E, asserts  "Tel est le plus ancien exemple de l'acceptation d'une solution négative dans un texte mathématique, ...."  The text says simply  "restan  10  et  ¾  mens de non res."  Sesiano, loc. cit. at beginning of 7.R, gives this in English:  "... the first work in which a negative result is admitted without any restriction, ....  No interpretation whatsoever is given of the negative result.  The only hint at its peculiarity is the -- exceptional -- verification of the results: ...."

AR.  c1450.  Prob. 156, 157, 171‑181, 186, 221, 223, 224, 338‑341.  Pp. 74‑75, 82‑85, 87, 102‑103, 149‑150, 171‑173, 218‑219.

156:  Regula posicionum:  II‑(1/3, 1/4, 1/5)  with horse worth  100.  Makes two steps by false position and then gives up and goes to the next section.

157:  II‑(1/3, 1/4, 1/5).   Answer:  45, 48, 52; 61.  = Fibonacci, p. 229.

171:  I‑(⅓, ¼).   Answer:  8, 9, 11.   Cf Al-Karkhi I -42.  Vogel says this occurs in Ibn al‑Haitham, ??NYS.

172‑175:  same as 171, with various values of horse:  31, 100, 81, 1.

176:  I‑(⅓, ⅔)  with horse worth  30.  Cf Chiu Chang Suan Ching.

177:  173 redone.

178:  Pferd mit  3  an namen (Horse with  3  [men] without price).   I‑(½, ⅓, ¼).   Answer:  10, 22, 26; 34.  Cf Gherardi, pp. 46-47.

179:  same heading,  II‑(½, ⅓, ¼).   Answer:  16, 18, 21; 25.  Cf Lucca 1754 58r.

180:  Pferd mit  3  [und] mit namen (horse with  3  [men and] price).   II‑(½, ⅔, ¾)  with horse worth  100.   Answer:  (200, 200, 150)/3.

181:  3  men buy a fish, leading to   x + ¼ (y + z)   =   ½y + ¼ (x + z)  =  ⅓z + ¼ (x + y)  =  cost.   Answer:  1, 3, 9; 4  and its multiples.

186:  3  men want to buy horses worth  10,  20,  30,  as in Fibonacci, pp. 235‑240, with  A = (½, ⅓, ¼).   Answer:  (24, 52, 144)/5.

221:  same as 173.

223:  II‑(½, ⅓, ¼)  with horse worth  51.  See 179.

224:  same as 178, with answer:  5, 11, 13; 17.

338:  same as 157.

339:  same as 179, with comment that multiplying the solution is again a solution.

340:  same as 178.

341:  I‑(1/3, 1/4, 1/5).   Answer:  (65, 85, 95; 125)/2.  Cf Diophantos 24.

Dresden MS C80, 15C, has some of these problems.  ??NYR -- mentioned in BR, p. 157.

Benedetto da Firenze.  c1465.

     Pp. 155‑157. 

I-(½, ⅓)  with  hi = 60, 80.

I-(½, ⅓, ¼)  with  hi = 60, 50, 163/5. 

II-(½, ⅓, ¼)  with  hi = 60, 50, 40.

     Chap. 20:  "... uomini che ànno d. et vogliono chompare chavagli," pp. 168‑171.

I‑(1/6, 1/7).  Answer:  35, 36; 41.

II‑(1/4, 1/6, 1/8).  Answer:  152, 164, 174; 193.  Illustration on p. 169.

I‑(½, ⅓, ¼).  Answer:  5, 11, 13; 17.  Cf Gherardi 46-47.

Muscarello.  1478.  Ff 56v-58r, pp. 158-161.  Four men buying a house worth  100,  II‑(2/3, 5/8, 4/5, 7/10).   Answer should be:  (1250, 1575, 1160, 1425)/23,  but there are two copying errors in the MS.  The second answer is given as  66 11/23  instead of  68 11/23  and the fourth answer is given as  61 12/23  instead of  61 22/23.

della Francesca.  Trattato.  c1480. 

F. 16r (62-63).  I-(½, ⅓)  with horse worth  35.  Answer:  21, 28.  English in Jayawardene.

F. 17v (64-65).  II-(1/3, 1/4, 1/5)  with horse worth  30,  Answer:  (1350, 1440, 1560)/61.  English in Jayawardene.  Cf Fibonacci, p. 229;  AR 156, 157.

Ff. 17v-18r (65-66).  I-(½, ⅓, ¼)  with horse worth  30.  Answer:  (450, 990, 1170)/51.  He doesn't notice that the fractions in the answers can be reduced.  Cf Gherardi 46-47; Provençale Arithmétique.

F. 21r (70-71).  II-(1/3, 1/4, 1/5)  with  hi = 12, 15, 18.  Answer:  (510, 666, 996)/61.

F. 21v (71-72).  I-(½, ⅔, ¾)  with  h = 12.  Answer:  (60, 36, 12)/7.

Ff. 22r-22v (72-73).  Three men buying a horse.  Leads to equations: 

x + y + z/2  =  x + y/3 + z  =  x/4 + y + z  =  30.  Answer:  (240, 270, 360)/23

F. 36v (95-96).  I-(⅓, ¼)  with jewel worth  30.  Answer:  (240, 270)/11.  Cf al-Karkhi III-26.

F. 36v-37r (96).  II-(½, ⅓, ¼)  with horse worth  35.  Answer:  (112, 126, 147)/5.  Cf Lucca 1754 58r.

Ff. 37v-38r (98-99)  =  f. 21r.

F. 39r (100-101).  I‑(½, ⅓, ¼)  with horse worth  30.  = ff. 17v-18r.  Answer:  (150, 330, 390)/17.  Cf Gherardi 46-47.

F. 39v (101-102).  Three men buying a horse.  Leads to equations: 

x + y + z/3  =  x + y/4 + z  =  x/5 + y + z  =  h.  Answer:  30, 32, 36; 74.  He doesn't notice that this can be divided through by two. 

F. 40v (102-103).  3  men want to buy a horse and have a friend with  p  in his purse, giving:   x + y + p/3  =  h,   x + z + p/4  =  h,   y + z + p/5  =  h.  Answer:  12, 15, 20; 47  with  60  in the purse.  Cf Lucca 1754 61r-61v.

F. 41r (103-104).  x + (y + z)/3 + 1  =  14,   y + (z + x)/4 ‑ 2  =  17,  

z + (x + y)/5 + 3  =  19.   Equivalent to  II-(1/3, 1/4, 1/5)  with  hi = 13, 19, 16.  Answer:  197/50, 2244/150, 1833/150.

F. 42r (105).  x + y/2 + z/3 = 12,  x/3 + y + z/4 = 14,  x/4 + y/5 + z = 18.  Answer:  (456, 2040, 3384)/217.

Ff. 44r-44v (108-109).  x + y + z/2 + 1 = 20,  x + y/3 + z - 2 = 20,  x/4 + y + z + 3 = 20.  Answer:  (228, 84, 250)/23.

Chuquet.  1484.  Triparty, part 1.  Sesiano cites p. 641, ??NYS

I-(½, ⅓, ¼)  with object worth  30.  English in FHM 79.  Answer:  (150, 324, 390)/17.  Cf Provençale Arithmétique.  324/17  is  given as  19 1/17  in FHM -- the correct value is  330/17 = 19 7/17,  so the  7  has been misread as a  1  at some stage.

FHM 79-80 then simply states and discusses the next two, which are discussed by Sesiano.

I-(2/3, 3/4, 4/5, 5/6)  with horse worth  40.   Answer:  24, 16, 8, 0.

I-(1/2, 2/3, 3/4, 4/5, 5/6)  with horse worth  40.   Answer:  0, 20, 10, 0, ‑10.  Chuquet then explains how to deal with negatives and zero.

FHM 81 gives the English of a version with equations   x + y + ½z  =  ⅓x + y + z  =  x + ¼y + z  =  20.   Answer:  (180, 160, 240)/23.

FHM 81-83 then gives the next two by formulae with solutions, then then gives the English of the next.

II-(½, ⅓, ¼)  with object worth  20.   Answer:  (64, 72, 84)/5.  Cf Lucca 1754 58r.

II-(½, ⅔, ¾)  with object worth  30.   Answer:  (600, 540, 495, 500)/29.  Cf AR 180.

II-(1/3, 1/4, 1/5)  with object worth  20.   Answer:  (900, 960, 1040)/61.  Cf Fibonacci 229.

Borghi.  Arithmetica.  1484.  Ff. 113v-116v (1509: ff. 95v-98r).  I‑(½, ⅓, ¼)  with  h = 20.   Answer:  (100, 220, 260)/17.

Calandri.  Arimethrica.  1491.  F. 667v.  Two men buy a lamprey.  I-(1/3, 1/5).  Takes  h = 60  rather arbitrarily and gets answer:  (300, 360)/7.

Francesco Pellos.  Compendion de lo Abaco.  Turin, 1492.  ??NYS -- see Rara 50-52  &  both Sesiano papers mentioned at Provençale Arithmétique, c1430, above.  Sesiano says ff. 64v-65r translates the three problems in the Provençale Arithmétique and is the first printed problem with a negative solution.  He says it may have been composed c1460.  Smith doesn't mention either of these points.

Pacioli.  Summa.  1494.

Ff. 105v-106r, prob. 23.  I-(½, ⅓, ¼)  with h = 20.  Answer:  (100, 220, 260)/17.  Cf Gherardi 36-37.

F. 190v, prob. 21.  Two men find a purse, but leads to:   x + ⅓y  =  p,   y + ¼x  =  p,   so this is  I-(⅓, ¼).  He assumes   x + y + p  =  100   and gets answer:  (200, 225; 275)/7.  Cf al-Karkhi I-24.

Ff. 191v-192r,  prob. 26.   x + ½ (y + z)  =  90,   y + ⅓ (z + x)  =  84,   z + ¼ (x + y)  =  81.   I.e.  I-(½, ⅓, ¼)  with  hi = 90, 84, 81. 

Answer:  (576, 900, 1008)/17.

F. 192r, prob. 27.  Same as prob. 26 with all values equal to  50.  I.e.  I-(½, ⅓, ¼)  with  h = 50.  Answer:  (250, 550, 650)/17.  Cf Gherardi 36-37.

F. 192v, prob. 31.  I-(½, ⅓, ¼).   Notes that the common value,  h,   can be set, as in prob. 23 (cf. 7.R.1) to  50,  but when fractions appear, he converts to answer:  (5, 11, 13; 17).  Cf Gherardi 36-37.

F. 193r, prob. 35.   x + (y + z)/3 + 1  =  14,   y + (z + x)/4 ‑ 2  =  17,   z + (x + y)/5 + 3  =  19.   Equivalent to  II-(1/3, 1/4, 1/5)  with  hi = 13, 19, 16.  Answer:  (197, 748, 611)/50.  Cf della Francesca 41r.

F. 193v, prob. 37.   x + y/2 + z/3  =  12,   y + z/3 + x/4  =  15,   z + x/4 + y/5  =  20.   Answer:  (164, 810, 1677)/94.

Ff. 193v-194r, prob. 41.  3  men find a purse and want to buy a horse, giving:   x + y + p/3  =  h,   y + z + p/5  =  h,   z + x + p/4  =  h.   If  T = x + y + z,  one gets   T + 47p/60  =  3h   and the solution space is actually two dimensional.  He assumes  p = 60,  h = 47  and then this problem reduces to prob. 39 (cf in 7.R.1).  Cf Lucca 1754 61r-61v; della Francesca 40v.

Calandri, Raccolta.  c1495.  Prob. 25, pp. 24‑25.  I‑(1/3, 1/5)  with horse worth  60.  Answer:  (300, 160)/7.  Cf Calandri.

Riese.  Rechnung.  1522.  1544 ed. -- pp. 91‑92 & 94‑95;  1574 ed. -- pp. 61v‑62v & 63v‑64r.

I‑(⅓, ¼)  with horse worth  15.   Answer:  (120, 135)/11.  Cf al-Karkhi I-42.

I‑(⅔, ¾)  with horse worth  39.  Answer:  (52, 39)/2.

II‑(½, ⅓, ¼)  with cow(?) worth  200.   Answer:  64, 72, 84.  Cf Lucca 1754 58r.

Tonstall.  De Arte Supputandi.  1522.  Pp. 246-248.  I-(½, ⅓, ¼)  with common value  20.  Cf Gherardi 36-37.

Riese.  Die Coss.  1524.  Many examples, including the following.

No. 31, pp. 44‑45.  3  men buy a horse worth  100,  II‑(½, ⅓, ¼).  Cf Lucca 1754 58r.

No. 47, pp. 46‑47.  Same with horse worth  17,  I‑(½, ⅓, ¼).  Cf Lucca 1754 58r.

No. 48, p. 47.  3  men buy  3  horses leading to:   a + b + c/5  =  12,   a/2 + b + c  =  18,   a + b/3 + c  =  16.

No. 120, p. 56.  Same as no. 31.

No. 121, p. 56.  3  men buy horse worth  100,  II‑(3/4, 4/5, 5/6).   Answer:  (510, 520, 475)/9.

No. 122, p. 56.  Same,  I‑(½, ⅓, ¼).   Says his friend Hans Conrad learned this from a Dominican(?) monk named Aquinas.  Cf Gherardi 36-37.

No. 123, p. 56.  Same with horse worth  204,  I‑(½, ⅓, ¼).  Cf Gherardi 36-37.

No. 126, p. 58.  4  men buy a horse worth  37,  I‑(1/2, 1/3, 1/4, 1/5).   Answer:  1, 19, 25, 28.  Cf Fibonacci 245-248.

No. 124 & 125, pp. 56‑58.  Complex formulations leading to the same problem with horse worth  2701  and  14800.   No. 125 has the order of the constants reversed.

No. 140, pp. 60‑61.  7  men buy a horse worth  100,  II‑(2/3, 3/4, ..., 8/9).  Answer:  (27630, 27855, 24460, 27175, 22830, 27265, 21640)/462.

No. 143, pp. 61‑62.  4  men buy a horse worth  100  leading to  II‑(1/3, 1/5, 1/6, 1/8)  except the last man borrows from all the others.  Answer:  (9747, 11058, 11875, 9348)/707.

Apianus.  Kauffmanss Rechnung.  1527.  Ff. M.iii.r - M.iii.v.  x + y/3  =  y + z/2  =  z + 2y/3  =  30.  Answer:  (45, 45, 60)/2.  Because  x = y,  this is actually the same as  II-(1/3, 1/2, 2/3).

Georg von Peurbach.  Elementa arithmetices, algorithmus ....  Joseph Klug, Wittenberg, 1534, ??NYS.  (There were several previous editions back to 1492, with variant titles.  Rara 53‑54.  Glaisher, op. cit. in 7.G.1 under Widman, describes this extensively and gives the following on p. 97.  This edition is substantially better than previous ones, but Peurbach died in 1461!)

F. D.iii.verso.  I-(½, ⅓)  with horse worth  10.  Cf della Francesca 16r.

F. E.i.verso.  II-(½, ⅓, ¼)  with horse worth  100.  Cf Lucca 1754 58r.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 98, ff. HH.vi.r - HH.vi.v (p. 169).  Men find a purse and buy a horse, giving:   x + y + p/2  =  y + z + p/5  =  x + z + p/3  =  h.   Answer:  6, 10, 15; 30, 31.  Cf Lucca 1754 61r-61v; della Francesca 40v; Pacioli 193v-194r.

Tartaglia.  General Trattato.  1556.  Book 17, art. 8, 13, 14, 22, 32, 40, 41, pp. 268v, 269v, 271v, 273v, 275v-277r.

Art. 8.  I-(⅓, ¼)  with  hi = 14, 17.   Answer:  (100, 162)/11.

Art. 13.  II-(½, ⅓, ¼)  with  hi = 40, 52, 62.   Answer:  (584, 832, 1404)/25.

Art. 14.  II-(2/3, 5/8, 4/5, 7/10)  with house worth  200.  Answer:  (2500, 3150, 2320, 2850)/23.

Art. 22.  II-(½, ⅔, ¾).   Answer:  48, 48, 36; 72,   which is not in lowest terms.  Cf AR 180.

Art. 32.  I-(⅓, ¼)  with  hi = 32, 42.   Answer:  (216, 408)/11.

Art. 40.  I-(½, ⅓, ¼)  with  h = 40.   Answer:  (200, 440, 520)/17.  Cf Lucca 1754 58r.

Art. 41.  I-(½, ⅓, ¼)  with  h = 20.   Takes about two pages to finally get half of the preceding answer.  Cf Lucca 1754 58r.

Buteo.  Logistica.  1559. 

Pp. 189-190.  I-(½, ⅓)  with  hi = 30, 20.   Answer:  24, 12.

Pp. 190-192.  I-(1/3, 1/4, 1/5)  with  hi = 14, 8, 8.  I find this remarkable in that he uses three unknowns -- A, B, C -- and solves by systematic elimination.   Answer:  11, 4, 5.

Pp. 192-193.  I-(1/3, 1/4, 1/5).  He assumes the amount is  17  and gets:  5, 11, 13; 17.  Cf Fibonacci 229.

Pp. 193-196.  I-(1/2, 1/3, 1/4, 1/6)  with  hi =  17, 12, 13, 13.   Answer:  6, 4, 8, 10.

Prob. 81, pp. 289-291.  Problem with soldiers, equivalent to  I‑(1/2, 1/3, 1/4)  with amount  14280.   Answer:  4200, 9240, 10920.  Cf Gherardi 36-37.

Prob. 30, pp. 357-358.  Amounts desired from others are variable, giving   x + y/2 + z/3  =  14,   x/3 + y + z/4  =  13,   x/6 + y/8 + z  =  14.   Answer:  6, 8, 12.

Schott.  1674. 

Ænigma V, p. 555.  Two men want to buy a field worth  100,  giving  x + y/2 + 5  =  100  =  y + x/3.  Equivalent to  I-(1/2, 1/3)  with  hi = 95, 100.  Answer:  54, 82.

Ænigma I, pp. 562-563.  I-(1/3, 1/4)  with  h = 110.  Answer:  80, 90.  Cf al-Karkhi I‑42.

Ænigma III, p. 563.  I-(1/3, 1/4, 1/5)  with h = 100.  Answer:  52, 68, 76.  Cf Diophantos 24.

"A Lover of the Mathematics."  A Mathematical Miscellany in Four Parts.  2nd ed., S. Fuller, Dublin, 1735.  Part III, no. 45, p. 94.  Men buying a house worth 1200.  I-(⅔, ¾).   Answer:  800, 600.  Cf Riese: Rechnung.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XXV, pp. 95-96 (1790: prob. 34, pp. 94‑95).   II‑(1/2, 1/3, 1/4, 1/5)  with  hi = a, b, c, d.   Does example with values  357, 476, 595, 714  and answer:  190, 334, 426, 676.   1745 gives a second method.

Les Amusemens.  1749.  Prob. 172, pp. 319-320.  I‑(½, ⅓, ¼)  with  h = 40.   Solves with a general  h.  Cf Gherardi 36-37.

Euler.  Algebra.  1770.  I.IV.

III: Question for practice, no. 19, p. 205.  I‑(½, ⅓, ¼)  with horse worth  34.  Cf Gherardi 36-37.

IV.618: Question 5, pp. 212‑213.  I‑(⅔, ¾)  paying a debt of  29.  Cf Riese: Rechnung.

IV.619‑620: Question 6, pp. 213‑214.  II‑(½, ⅓, ¼)  buying a vineyard worth  100.  Cf Lucca 1754 58r.

IV.621, pp. 214‑215.  Gives general form of solution of type II problem, using  4  person case as an example.

IV.622: Question 7, pp. 215‑216.  Problem equivalent to  I‑(½, ⅓, ¼)  with  h = 901,  i.e.  53/2  times the problem on p. 205.  Cf Gherardi 36-37.

Vyse.  Tutor's Guide.  1771?  Prob. 13, 1793: p. 131; 1799: p. 138 & Key pp. 183-184.  I‑(⅔, ¾)  buying a horse worth  1200.  Cf Riese: Rechnung.

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Section 82, ex. 8, pp. 122-123.  Variant version -- he has  I-(1/2, 1/3)  with  h = 50,000  and then says the third could buy if he had  1/4  of the first's money.  I think this is a corruption of  I-(1/2, 1/3, 1/4).  Cf della Francesca 16r.

Pearson.  1907.  Part II, no. 139: The money‑boxes, pp. 141 & 218.  I‑(1/2, 1/3, 1/4, 1/5)  with  h = 740.  Cf Fibonacci 245-248.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 101, pp. 43 & 144: A garaging problem.  A variation on type IV.  a + b + c + d + e = 100;  a + b = 52;  b + c = 43;  c + d = 34;  d + e = 30.

 

          7.R.3. SISTERS AND BROTHERS

 

            New section.

            NOTATION:  (a, b)  means each boy has  a  times as many sisters as brothers, while each girl has  b  times as many brothers as sisters.  This only has integer solutions for the integer pairs:

     (a, b)           =  (1, 2),  (1, 3),  (2, 1),  (2, 2),  (3, 1),   with solutions: 

(Boys, Girls)     =  (4, 3),  (3, 2),  (3, 4),  (2, 2),  (2, 3).

            See Wolff for a related problem.

            See  Fireside Amusements;  Cutler  for a related trick problem.

 

Fireside Amusements.  1850: No. 47, pp. 114 & 181;  1890: No. 33, p. 102.  "Mr. Jones told another gentleman that he had six daughters, and each daughter had a brother; how many children had Mr. Jones?"  Cf Cutler, below.

Mittenzwey.  1880.  Prob. 15, pp. 2-3 & 59;  1895?: 17, pp. 8-9 & 63;  1917: 15, pp. 8 & 57.  Sisters and brothers.  (2, 1).

Peano.  Giochi.  1924.  Prob. 55, p. 15.  (1, 2).

King.  Best 100.  1927.  No. 39, pp. 19 & 47.  = Foulsham's, no. 11, pp. 7 & 11.  (2, 1).

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.  P. 108.  (2, 1).

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 28, pp. 193 & 203.  A man in an office says he has four times as many male colleagues as female colleagues.  One of the women answers that she has five times as many male colleagues as female colleagues.  This is  (1/4, 5)  in the above notation.  In general, this would give  m - 1 = af,  m = b (f - 1),  which has solution  f = (b + 1)/(b ‑ a).  See Dodson in 7.R for a different phrasing of the same problem.

Depew.  Cokesbury Game Book.  1939.  Sisters and brothers, p. 218.  (2, 1).

Owen Grant.  Popular Party Games.  Universal, London, nd [1940s?].  Prob. 8, pp. 36 & 50.  (2, 1).

John Henry Cutler.  Dr. Quizzler's Mind Teasers.  Greenberg, NY, 1944.  ??NYS -- excerpted in: Dr. Quizzler's mind teasers; Games Magazine 16:3 (No. 109) (Jun 1992) 47 & 43, prob. 27.  "Mr. and Mrs. Twichell have six daughters.  Each of the daughters has a brother.  How many persons are there in the entire family?"  Only  9  (counting the parents).  Cf Fireside Amusements, above.

Joseph Leeming.  Riddles, Riddles, Riddles.  Franklin Watts, 1953;  Fawcett Gold Medal, 1967.  P. 113, no. 40.  (1, 2).

Ripley's Puzzles and Games.  1966.  P. 12.  Man has two boys and a girl.  He wants to have 12 boys and for each boy to have a sister.  How many girls are needed?

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  The sum of the siblings, pp. 31 & 110.  (2, 1).

 

          7.R.4. "IF I SOLD YOUR EGGS AT MY PRICE, I'D GET ...."

 

          New section.

 

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XL, pp. 106-107 (1790: prob. LIII, p. 108).  Market women bring  x  and  y  eggs to market, with  x + y = 100 [c]  and sell at prices  A  and  B  such that they get the same, i.e.  Ax = By.  First says to the second:  "Had I brought as many eggs as you I should have received  18 [a]  Pence for them".  The other responds:  "Had I brought no more than you, I should have received only  8 [b]  Pence for mine".  I.e.  Ay = a,  Bx = b.  This gives   (x/y)2  =  b/a,   x  =  cÖb/[Öa+Öb]  =  bc/[b+Öab],   y  =  cÖa/[Öa+Öb].

Mittenzwey.  1880.  Prob. 123, pp. 25 & 76;  1895?: 141, pp. 29 & 79;  1917: 141, pp. 27 & 77.  Same as Simpson, with it being clearly phrased "If I had had your eggs and sold them at my price, ...", with  c = 110,  a = 250,  b = 360.  Answer is given as  50, 60,  but ought to be as  60, 50.  Here the prices also come out integral.

McKay.  At Home Tonight.  1940.  Prob. 11: Buying cows, pp. 64 & 78.  Farmers  A  and  B  each buy  £350  worth of cows.  If  A  had bought at  B's  price, he would have paid  £250.  What would  B  have paid if he bought at  A's  price?  Letting  x, y  be the numbers of cows and  A, B  be the prices, we have   Ax = By = 350,   Bx = 250,  and we want  Ay.  Then   y/x  =  7/5   and   Ay  =  Ax(7/5)  =  490.

 

          7.S.    DILUTION AND MIXING PROBLEMS

 

            See Tropfke 569.

            There are a number of problems of this sort.  One type is the same as the Hundred Fowls problem (7.P.1) where the solutions need not be integers.  Lucca 1754, c1330, has a number of these.  Here I consider only some of special interest and the following.

 

Recorde.  Second Part.  1552.  H&S quotes from the 1579 ed., f. Y.3.  "It hath great use in composition of medicines, and also in myxtures of metalles, and some use it hath in myxtures of wines. but I wshe [sic] it were lesse used therein than it is now a daies."  The 1668 ed., p. 295: The Rule of Mixture, has:  "And it hath great use in composition of Medicines, and also in mixtures of Metalls, and some use it hath in mixtures of Wines : but I wish it were less used therein then it is now-a-days."

 

          7.S.1. DISHONEST BUTLER DRINKING SOME AND REPLACING WITH WATER

 

            Dodson, Todhunter and Clark are problems to determine the amount taken off each time.

 

Papyrus Rhind, op. cit. in 7.C.  Prob. 71, p. 108 of vol. 1 (1927) (= p. 57 of 1978 ed.).  ¼  is poured off & replaced, what is the strength?  (H&S 85 quotes Peet's version.)

Bakhshali MS.  c7C.  Kaye I 48; III 201-202, ff. 12r-12v.  Man has bottle holding  4  prasthas of wine.  (Drinks  ¼  and refills with water)  four times.  How much wine is left?

Cardan.  Practica Arithmetice.  1539.  Chap. 66, sections 36 & 37, ff. DD.v.r - DD.v.v (p. 146).  Drink three pitchers and replace with water four times leaving wine of half strength.  Then the same for three times.

Tartaglia.  Quesiti, et Inventioni Diverse, 1546, op. cit. in 7.E.1, Book 9, quest. 18, pp. 102v‑103r.  (Remove 2 and replace)  thrice to halve strength.

Buteo.  Logistica.  1559.  Prob. 85, pp. 296-298.  Butler drinking some and replacing with water.  (7/8)5.  (H&S 85)

Trenchant.  Op. cit. in 7.L, 1566.  1578 ed., p. 297.  ??NYS.  (Remove  1/12th  and replace)  six times.  (H&S 85 gives French and English.  Sanford 209 gives English.)

Les Amusemens.  1749.  Prob. 176, pp. 326-327.  Sommelier drinks  6  pints from a cask of  360  and replaces with water three times -- how much wine has he drunk?

Dodson.  Math. Repository.  1775.  P. 76, Quest. CXLI.  Cask of 81 gallons.  (x  is drawn off and replaced by water) four times, leaving 16 gallons of wine in the mixture.  Gives a general solution.

Ozanam‑Montucla.  1778.  Prob. 21, 1778: 212-214;  1803: 207-209.  Prob. 20, 1814: 179‑181;  1840: 93.  Dishonest butler  (removes  1/100th  and replaces)  30  times.  Notes that it is easier to use logarithms.

Bullen.  Op. cit. in 7.G.1.  1789.  Chap. 38, prob. 33, p. 243.  Cask of  500  gallons;  (remove  1/10th  and replace with water)  five times.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 222, last problem, p. 364.  Barrel of  100  pints of wine.  One pint is drunk and replaced by water each day.  How much wine is left after  50  days?  When is the wine diluted to half its strength?  One third?  One quarter?

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  3rd ed., 1850.  Op. cit. in 7.H.  P. 358, no. 35.  Same as Bullen.

Vinot.  1860.  Art. LVI: Le Sommelier infidèle, pp. 73-74.  Barrel of  100.  Sommelier drinks  1  and replaces with water  30  times.  Computes the amount of water in the barrel each day.

Todhunter.  Algebra, 5th ed.  1870.  Examples XXXIII, no. 8, pp. 285 & 590.  From  256  gallons of wine, draw off  x  and replace with water, four times to leave only  81  gallons in the container.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  Prob. 98, p. 62.  Barrel of wine holding  210.  Remove  45  and replace with water three times.  Determine the amount of wine and water.

Mittenzwey.  1880. 

Prob. 76, pp. 15 & 66;  1895?: 83, pp. 19 & 69;  1917: 83, pp. 18 & 65.  Cask of  192 l  worth  4 per l.  Remove a quarter and replace with water.  What is the value of the result?  Repeat the dilution.  What is the value now and how much less is the cask worth than originally?

Prob. 105, pp. 21-22 & 73;  1895?: 122, pp. 26 & 75.  (In 1917, this was replaced.)  Remove  4  from a cask and replace with water, thrice.  The mixture now contains    more water than wine.  How big is the cask?  Letting  C  be the size of the cask, this leads to  [(C-4)/C]3 =  (C/2 - 5/4)/C  which is a cubic with one real root,  C = 16.  I am not surprised that this was dropped.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XLI: Le tonneau inépuisable, p. 183.  (Remove  1/100  and add water)  20  times -- how much is left?  Gives general solution.

Charles Pendlebury & W. S. Beard.  A "Shilling" Arithmetic.  Bell, London, 1899 -- my copy is 59th ptg, 1960 and says it was reset for the 49th ptg of 1944.  Examination Papers XIX, prob. 15, p. 192.  A vessel is 17% spirit.  When  10  has been drawn off and replaced with water, it is now  15 1/9 % spirit.  How big is the vessel?

Dudeney.  Weekly Dispatch (8 Feb 1903) 13.  (Remove  1/100th  and replace)  30  times.

Clark.  Mental Nuts.  1904, no. 39.  Find capacity of the keg.  (Fill a keg from a  20 gallon cask and then replace with water) three times to dilute the cask to half-strength.  How big is the keg?

 

          7.S.2. WATER IN WINE VERSUS WINE IN WATER

 

            Two containers, one of wine and one of water.  One puts an amount of water in the wine, stirs and then transfers the same amount of the mixture back to the water.  Is there now more water in the wine or wine in the water?

 

Todhunter.  Algebra, 5th ed.  1870.  Miscellaneous Examples, no. 187, pp. 560.  Vessels of size  A  and  B  containing wine and water respectively.  C  is taken from each and then put into the other.  This is repeated  r  times.  Determines the quantity of wine in the second vessel as  AB/(A+B) [1 - βr]  where  β  =  1 - C/A - C/B.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  A predecessor of this problem -- prob. 92, p. 61.  Vase  A  has  12  litres of wine and  4  litres of water.  Vase  B  has  8  litres of wine and  3  litres of water.  Take off  4  litres from each and then put the  4  litres from  A  into  B  and the four litres from  B  into  A.  Determine contents in each vase.

Viscount John Allsebrook Simon.  [Memory of Lewis Carroll.]  IN:  Appendix A of Derek Hudson; Lewis Carroll, An Illustrated Biography; Constable, (1954);  new ed., 1976, pp. 248‑249.  = Carroll-Wakeling II, prob. 34: Two tumblers, pp. 52 & 73.  50  spoonfuls of brandy and  50  spoonfuls of water -- transfer a spoonful back and forth.  He says Carroll posed this.  Mentioned in Carroll-Gardner, p. 80, who gives the full name.  The DNB says he entered Wadham College, Oxford, in 1892, and his Memory says he met Carroll then.  So this dates from ³1892, but Carroll could have been propounding it years before.

Ball.  MRE, 3rd ed., 1896, pp. 26‑27.  Water in wine versus wine in water.  He says this is a question "which I have often propounded in past years".  Not in the 1st ed of 1892.

W. P. Workman.  The Tutorial Arithmetic, op. cit. in 7.H.1.  1902.  Section IX, examples CXLV, prob. 35, pp. 427 & 544  (= 433 & 577 in c1928 ed.).  A  contains  11  pt water and  7  pt wine,  B  contains  5  pt water and  13  pt wine.  Transfer  2  pt from  A  to  B  and back.  Find changes of water and wine in both  A  &  B.  This is a precursor of the puzzle idea.

Pearson.  1907.  Part II, no. 18, pp. 117 & 194‑195.  Butter in lard versus lard in butter.

Loyd.  Cyclopedia, 1914, pp. 287 & 378.  Forty quarts milk and forty quarts water with a quart poured back and forth.  He says the ratios of milk to water are then  1 : 40  and  40 : 1,  which is correct, but isn't the usual question.

H. E. Licks.  Op. cit. in 5.A.  1917.  Art. 20, pp. 16‑17.  Water and wine.

Ahrens.  A&N, 1918, p. 89.

T. O'Conor Sloane.  Rapid Arithmetic.  Quick and Special Methods in Arithmetical Calculation Together with a Collection of Puzzles and Curiosities of Numbers.  Van Nostrand, 1922.  [Combined into: T. O'Conor Sloane, J. E. Thompson & H. E. Licks; Speed and Fun With Figures; Van Nostrand, 1939.]  Wine and water paradox, pp. 168‑169.

F. & V. Meynell.  The Week‑End Book.  Op. cit. in 7.E.  1924.  Prob. two: 2nd ed., p. 274;  5th?? ed., p. 407.  Find "proportion of the amount of water in  A  to the amount of milk in  B."

Peano.  Giochi.  1924.  Prob. 24, p. 7.  Water and wine.

Loyd Jr.  SLAHP.  1928.  Cheating the babies, pp. 40 & 98.  Two large cans with  10  gallons of milk and  10  of water.  Pour  3  gallons back and forth.  "Have I more milk in the water can than I have water in the milk can?"  He works out that each can has the proportion  7 9/13 : 2 4/13  [=  100 : 30  =  10 : 3].

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 25, pp. 17 & 190.  Whisky and water in equal amounts.  He asks about proportions rather than amounts.

Phillips.  The Playtime Omnibus.  Op. cit. in 6.AF.  1933.  Section XVII, prob. 17: The two vessels, pp. 56 & 238.  Same as in his Week‑End.

Abraham.  1933.  Prob. 26 -- Whiskey and water, pp. 10 & 24 (7 & 112).

Perelman.  FMP.  c1935?  Water and wine, pp. 215 & 218.

Phillips.  Brush.  1936.  Prob. D.5: Whisky and water, pp. 12 & 82.  Same as in his Week‑End.

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 36, pp. 194 & 204.  Man takes a cup of black coffee.  He drinks  1/6  of it and fills it up with milk.  He then drinks  1/3  of it and fills it up with milk.  He then drinks half of it and fills it up with milk.  Then he drinks the whole cup.  Has he drunk more milk than coffee or vice-versa?

McKay.  Party Night.  1940.  No. 6, pp. 176-177.  Water and wine.  "I have seen, after dinner, parties discuss the problem for a long time, ...."

Sullivan.  Unusual.  1943.  Prob. 21: Mixtures.  Gasoline and alcohol.

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 14-16;  1945: 13‑15;  1961: 23‑25.  Water and milk.

Gamow & Stern.  1958.  Gin and tonic.  Pp. 103‑104.

 

          7.T.    FOUR NUMBER GAME

 

            The game takes  (a, b, c, d)  to  (½a ‑ b½, ½b - c½, ½c ‑ d½, ½d ‑ a½).  See 7.BB for other iterated functions.

 

C. Ciamberlini & A. Marengoni.  Su una interessante curiosità numerica.  Period. Mat. (4) 17 (1937) 25‑30.  They attribute the problem to E. Ducci.

G [= J. Ginsburg].  Curiosa 30.  An interesting observation.  SM 5 (1938) 135.  Brief report on above article.

Benedict Freedman.  The four number game.  SM 14 (1948) 35‑47.  No references.  Obtains basic results for  n‑number version.

S. P. Mohanty.  On a problem of S. J. Bezuszka & M. J. Kenney on cyclic difference of pairs of integers.  Fibonacci Quarterly 19 (1981) 314‑318.  (This and three of its references are not in Meyers.)

Leroy F. Meyers.  Ducci's four‑number problem:  a short bibliography.  CM 8 (1982) 262‑266.  See Ludington-Young, below, for nine additional references.

M. Gardner.  Riddles of the Sphinx and Other Mathematical Puzzle Tales.  New Math. Library, MAA, 1987.  Prob. 29: Hustle off to Buffalo, parts 2‑5, pp. 134‑136, 151‑152, 160‑163.  Gives a proof that most quadruples converge to all zeroes and finds the quadruples that cycle.

Joseph W. Creely.  The length of a three-number game.  Fibonacci Quarterly 26:2 (May 1988) 141-143.  Solves three and two number versions.

Stanley Bezuszka with Lou D'Angelo & Margaret J. Kenny.  The Wonder Square.  Booklet 2, Boston College Math. Inst.  Motivated Math Project Activity.  Boston College Press, Chestnut Hill, Mass., 1976.  32pp.  Studies the process with various special numbers, e.g. progressions, Fibonacci, Tribonacci and figurate numbers.  The Tribonacci case produces starting sequences with length  n  for any  n.

Anne Ludington-Young.  The length of the  n-number game.  Fibonacci Quarterly 28:3 (Aug 1990) 259-265.  Obtains a bound and solves some cases.  Cites Meyers and 9 additional references.

 

          7.U.   POSTAGE STAMP PROBLEM

 

            What integers are non‑negative integral combinations of  a, b, ...?  In particular, what is the largest integer which is not such a combination?  This is well known for the case of two values, but remains unknown for more values.  From about 1960 onward, the case with two values frequently occurs in number theory texts and as a puzzle problem, but I haven't entered such appearances.

 

Dickson, vol. II, chap. II is uncharacteristically obscure about this.  It is generally attributed to Frobenius (1849-1917).

J. J. Sylvester.  Math. Quest. Educ. Times 41 (1884) 21.  ??NYS.  Solves the problem for two values.

Alfred Brauer.  On a problem of partitions.  Amer. J. Math. 64 (1942) 299‑312.  On p. 301, he says that some of the results are due to I. Schur in a lecture in 1935, others are due to himself and others are joint work.  He cites Sylvester, but doesn't mention Frobenius.

Alfred Brauer & B. M. Seelbinder.  On a problem of partitions II.  Amer. J. Math. 76 (1954) 343‑346.  "A number of years ago, the first of the authors studied together with I. Schur the following problem of Frobenius: ..."

Alfred Brauer & James E. Shockley.  On a problem of Frobenius.  J. reine angew. Math. 210 (1962) 215‑220.  "G. Frobenius, in his lectures, raised the following problem repeatedly ....  No result was obtained for many years.  In 1935, I. Schur proved in his last lecture in Berlin the following result ...."

Ernest S. Selmer.  On the linear diophantine problem of Frobenius.  J. reine angew. Math. 293/294 (1977) 1‑17.  Gives 25 references which he believes to be a complete list.  Cites Sylvester, but the next oldest are the 1942 and 1954 papers above.  The 1962 paper above is the first to mention Frobenius in the title.

 

          7.V.   XY  =  YX  AND ITERATED EXPONENTIALS

 

D. Bernoulli.  Letter to C. Goldbach, 29 Jun 1728.  In:  P. H. Fuss, ed. op. cit. in 5.F.1, vol. 2, p. 262.  xy = yx  in integers.

Goldbach.  Letter to D. Bernoulli, 31 Jan 1729.  Ibid., pp. 280‑283.  Reply to the above.  Setting  y = ax,  he gives an easy proof for the only integer solutions.  He says the fractional solutions are  (f/g)g/(f‑g).

L. Euler.  Introductio in Analysin Infinitorum.  Bousquet, Lausanne, 1748.  Vol. 2, § 519, pp. 295-296 & Tab. XXV, fig. 103.  = Introduction to the Analysis of the Infinite; trans. by John D. Blanton; Springer, NY, 1988-1990; Book II, pp. 339-340 & 489.  Gets  x  =  (1 + 1/n)n;   y  =  (1 + 1/n)n+1.

L. Euler.  De formulis exponentialibus replicatus.  Acta Acad. Sci. Petropol. 1 (1777(1778)) 38‑60.  = Opera Omnia (1) 15 (1927) 268‑297.  Iterated exponentials.

Ahrens.  A&N, 1918, pp. 76‑78, discusses the problem and notes that Goldbach's fractional solution is rational if and only if  g/(f‑g)  is integral, say  n,  which gives   x  =  (1 + 1/n)n.

C. A. B. S[mith].  5-minute problem.  Eureka 3 (Jan 1940) 4 & 24.  Letting  p, m, c  be the number of physicists, chemists and mathematicians at a lecture, he observes that  pm = mc,  cm = mp,  cp = pc  and  m > c.  One finds  p = c,  hence  cm = mc.

R. L. Goodstein.  Note 1725:  The equation  ab = ba.  MG 28 (No. 279) (May 1944) 76.  Quick derivation of general solution and the rational solutions.

Anonymous.  The problems drive.  Eureka 14 (Oct 1951) 12-13 & 22.  No. 5.  Find rational solutions of  pq = qp.

R. A. Knoebel.  Exponentials reiterated.  AMM 88 (1981) 235‑252.  Extensive history and bibliography.

F. Gerrish.  Note 76.25:  ab = ba:  the positive integer solution.  MG 76 (No. 477) (Nov 1992) 403.  Short note on the integer case with two recent references.

R. F. Churchhouse.  Solutions of the equation  xy = yx.  Bull. Inst. Math. Appl. 31:7/8 (Jul/Aug 1995) 106.  Easily finds the real solutions, then the rational and integer solutions.  Notes that  x = i,  y = -i  is a solution!

 

          7.W.  CARD PILING OVER A CLIFF

 

            Identical cards (or dominoes) of length  1  can be stacked to reach out from the edge of a cliff.  The simplest analysis shows that  n  cards can reach out  1/2 + 1/4 + 1/6 + ... + 1/2n  @  (log n)/2.  Some authors consider real dominoes which can be piled in several orientations.

 

J. G. Coffin, proposer.  Problem 3009.  AMM 30 (1923) 76.  Asks for maximum overhang for  n  cards.  (Never solved!)

Max Black.  Reported in: J. F. O'Donovan; Clear thinking; Eureka 1 (Jan 1939) 15 & 20.  Problem 1.  Asks for maximum extension from the bottom card with  52  cards, then for  n  cards.  Solution is  1/2 + 1/4 + 1/6 + ... + 1/(2n-2).  For  n = 52,  this is about  9/4.

A. S. Ramsey.  Statics.  2nd ed., CUP, 1941, example 4.68, pp. 47‑48.  Discusses equal spacing with a support at the outer end, e.g. a staircase.  (Is this in the 1st ed. of 1934??  My source indicated that Ramsey cited a Tripos exam.)

Heinrich Dörrie.  Mathematische Miniaturen.  Ferdinand Hirt, Breslau, 1943;  facsimile reprint by Martin Sändig, Wiesbaden, 1979.  Prob. 240, pp. 279-282.  Using cards of length  2,  he gets the extension  1/1 + 1/2 + 1/3 + ...  and shows the curve formed is  x = log (y/t),  where t is the thickness of the cards and  x  and  y  are measured from the outer end of the top card -- in the usual picture, they are both going negatively.

R. T. Sharp, proposer;  C. W. Trigg, solver.  Problem 52.  Pi Mu Epsilon J. ??  &  (April 1954) 411‑412.  Shows overhang approaches infinity, but the proposal asks for the largest overhang for  n  dominoes, which is not answered.  Notes that the dominoes can be angled so the diagonal is perpendicular to the cliff edge.  This is also in Trigg; op. cit. in 5.Q; Quickie 52: Piled dominoes, pp. 17 & 99, but it still doesn't answer the proposal.

P. B. Johnson, proposer;  Michael Goldberg, Albert Wilansky, solvers.  Problem E1122 -- Stacking cards.  AMM 61 (1954) 423  &  62 (1955) 123‑124.  Both show overhang can go to infinity.

Paul B. Johnson.  Leaning tower of lire.  Amer. J. Physics 23 (Apr 1955) 240.  Claims harmonic series gives greatest overhang!!

P. J. Clarke.  Note 2622:  Statical absurdity.  MG 40 (No. 333) (Oct 1956) 213‑215.  Considers homogeneous bricks with weight bounded above and below and length bounded below.  Then one can take such bricks in any order to achieve an arbitrarily large overhang.

Gamow & Stern.  1958.  Building blocks.  Pp. 90‑93.

D. St.P. Barnard.  Problem in The Observer, 1962.  ??NYS.

D. St.P. Barnard.  Adventures in Mathematics.  Chap. 8 -- The Domino Story.  (1965);  Funk & Wagnalls, NY, 1968, pp. 109‑122.  Gets  3.969  for  13  dominoes.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Overbalance, pp. 122-124.  Standard piling.  Shows the harmonic series diverges.

See A. K. Austin, 1972, in 5.N for a connection with this problem.

Stephen Ainley.  Letter:  Finely balanced.  MG 63 (No. 426) (1979) 272.  Gets  1.1679  for  4  cards.

[S. N. Collings].  Puzzle no. 47.  Bull. Inst. Math. Appl. 15 (1979) 268 & 312.  Shows that a simple counterbalancing scheme gets  m/2  for  2m ‑ 1  dominoes, so the overhang for  n  dominoes is at least  ½ log2(n+1).

R. E. Scraton.  Letter:  A giant leap.  MG 64 (No. 429) (1980) 202‑203.  Discusses some history, especially Barnard's problem.

Nick Lord, proposer;  uncertain solver.  Problem 71.E.  MG 71 (No. 457) (Oct 1987) 236  &  72 (No. 459) (Mar 1988) 54‑55.  Overhang can diverge even if the lengths converge to zero.

Jeremy Humphries, proposer;  various solvers.  Prob. 129.5 -- Planks.  M500 129 (Oct 1992) 18  &  131 (Feb 1993) 18.  Uniform planks of lengths  2, 3, 4.  Find maximum overhang.  Best answer is  3 5/18.

Unnamed solver, probably Jeremy Humphries.  Prob. 145.2 -- Overhanging dominoes.  M500 145 (?? 1995) ??  & 146 (Sep 1995) 17.  He uses approximately real dominoes of size  .7 x 2.2 x 4.4 cm.  Given three dominoes, how far out can you reach?  And how far away from the edge can one domino be?  Answer is basically independent of the shape.  Place one domino on its large face with its centre of gravity at the cliff edge and its diagonal going straightout.  For maximum reach, similarly place another domino with its centre of gravity at the far corner of the first domino.  To get a domino maximally far from the edge, place one on edge so its largest face is parallel to the cliff edge and with its centre of gravity at the far corner of the first domino.  In both cases, counterbalance by putting the other domino with its centre of gravity at the inner corner of the first domino.  So the answer to the first problem is the longest face diagonal and the answer to the second problem is half the longest face diagonal minus half the thickness.

 

          7.X.   HOW OLD IS ANN? AND OTHER AGE PROBLEMS

 

            The simplest age problems are 'aha' problems, like 'Diophantos's age', going back at least to Metrodorus and will not be considered here -- see Tropfke 575 for these.  More complicated, but still relatively straightforward age problems appear in various 16-19C arithmetic and algebra works, e.g.  Schooten;  Recorde;  Baker;  Cocker;  Pike;  American Tutor's Assistant;  Ainsworth & Yeats, 1854, op. cit. in 7.H.4;  Colenso.  I have included only a sample of these to show the background and a few earlier versions.  Simple problems of Form III then appear from the mid 18C and later in standard arithmetic books, and later in puzzle books like Hoffmann;  Pearson;  Loyd;  Dudeney  and  Loyd Jr.  These usually lead to two equations in two unknowns, a bit like 7.R, or one equation in one unknown, depending on how one sets up the algebra.  In the 19C, this problem was popular in discussions of algebra as the problem can have a negative solution, which means that the second time is before the first rather than after, and so the problem was used in discussions of the existence and meaning of negative numbers -- see:  Hutton, 1798?;  Kelland;  De Morgan (1831, 1836, 1840).  About 1900, we get the  "How old is Ann (or Mary)"  versions, forms I and II below.

            Form I:  "The combined ages of Mary and Ann are  44  years, and Mary is twice as old as Ann was when Mary was half as old as Ann will be when Ann is three times as old as Mary was when Mary was three times as old as Ann.  How old is Ann?"  Answer is  16½;  Mary is  27½.  See:  Pearson;  Kinnaird;  Loyd;  Bain;  White;  Dudeney;  Loyd Jr;  Grant;  Ransom;  Doubleday - 2.

            Form II:  "Mary is  24.  She is twice as old as Ann was when Mary was as old as Ann is now.  How old is Ann?"  Answer is  18.  See:  Haldeman-Julius;  Menninger;  Ransom;  Doubleday - 2.

            Other examples of this genre:  Baker;  Vinot;  Brooks;  Gibney;  Lester;  Clark;  M. Adams;  Meyer;  Little Puzzle Book;  Dunn.

            Form III‑(a, b, c).  X  is now  a  times as old as  Y;  after  b  years,  X  is  c  times as old as  Y.  I.e.   X = aY,   X + b  =  c (Y + b).   This can be rephrased depending on the time of narration -- e.g.  X  is now  c  times as old as  Y,  but  b  years ago,  X  was only  a  times as old as  Y.  See Clark for an equivalent problem with candles burning. 

            If the problem refers to both some time ago and some time ahead, it is the same as a form usually stated about increasing or decreasing both the numerator and denominator of a fraction  X/Y,  i.e.  (X-a)/(Y-a) = b;  (X+c)/(Y+c) = d.  See: Simpson; Dodson; Murray; Vinot.  I have seen other examples of this but didn't note them.

            Form IV-(a, b, c).  X  is now  a;  Y  is now  b;  when will (or was)  X  be  c  times as old as  Y?  I.e.  a + x  =  c (b + x). 

            General solution of Form III occurs in:  Milne;  Singmaster.

            General solution of Form IV occurs in:  De Morgan (1831).

 

Examples of type III.

 

            a            b            c

 

        10/7            2         4/3              Boy's Own Mag.

            2          12         8/5              Milne

            2          30         7/5              Dilworth

            2          60         5/4              Dilworth

         7/3            6        13/9              Milne

            3          10            2              Vinot

            3        10½            2              Walkingame, 1751

            3          14            2              Murray;  Hummerston

            3          15            2              Ladies' Diary;  Mair;  Vyse;  Amer. Tutor's Asst.;  Eadon;  Bonnycastle, 1815;  Child;  Walkingame, 1835;  New Sphinx;  Charades, etc.;  Pearson; 

            3          18            2              Dudeney

            4           -2            6              Unger

            4            4            3              Tate

            4            5            3              Young

            4          10         5/2              Tate

            4          14            2              Simpson

            5            5            3              Dudeney

            5            5            4              Hutton, c1780?;  Hutton, 1798?;  Treatise; 

            6            3            4              Unger;  Milne

            6          24            2              Clark, 1897

            7            3            4              Berkeley & Rowland

            7          15            2              Unger

            9            3            3              M. Adams

          14          10         5⅓              Unger

 

Examples of type IV.

 

            a            b            c

 

            a            b            n              De Morgan, 1931?

            a            b            3              De Morgan, 1831?

            a            b            4              Bourdon

          32            5          10              Perelman

          40            8            3              De Morgan, 1831?

          40            9            2              Young

          40          20            3              Kelland;  Carroll-Gardner

          42          12            4              Hutton, 1798?

          45          12            3              Hoffmann;  Dudeney

          45          15            4              Bourdon

          48          12            3              Vinot

          48          12            7              Vinot

          50          40            2              De Morgan, 1840

          54          18            2              Manuel des Sorciers

          56          29            2              De Morgan, 1836

          62          30            5              Colenso

          71          34            2              Hoffmann

          71          34            3              Hoffmann

See Young World for a variant form equivalent to 

          2        2        -20

 

Robert Recorde.  The Whetstone of Witte.  John Kyngstone, London, 1557.  Facsimile by Da Capo Press, NY  &  Theatrum Orbis Terrarum, Amsterdam, 1969.  A question of ages, ff. Gg.i.v - Gg.ii.r.  (The gathering number at the bottom of folio ii is misprinted G.)  Father and two sons.   B = A + 2.   C = A + B + 4.   A + B + C = 96.

Baker.  Well Spring of Sciences.  1562?  Prob. 1,  1580?: ff. 189r‑190r;  1646: pp. 297-299;  1670: pp. 340-341.  A  is  120.  B  says if he were twice his present age, he would be as much older than  A  as  A  is older than  B  is now.  C  says the same with three times his age,  D  with four and  E  with five.

Frans van Schooten Sr.  MS algebra text, Groningen Univ. Library, Hs. 443, c1624, f. 54r.  ??NYS -- reproduced, translated and described by Jan van Maanen; The 'double--meaning' method for dating mathematical texts;  IN: Vestigia Mathematica; ed. by M. Folkerts & J. P. Hogendijk; Rodopi, Amsterdam, 1993, pp. 253--263.  "A man, wife and child are together  96  years, that is to say the man and child together  2  years more than the wife and the wife with the child together  14  years more than the man.  I ask for the years of each."  Van Maanen feels that this problem probably describes the van Schooten family and it is consistent with the family situation for a period in 1623--1624, when the man was  41,  the wife  47  and the child  8,  thereby giving a reasonable date for the undated MS.  Van Schooten's two sons also used the problem in their works, but reversed the role of man and wife because it was not common for the wife to be so much older.

Edward Cocker.  Arithmetic.  Op. cit. in 7.R.  1678.  Several problems, of which the following are the most interesting.

Chap. 31, quest. 2.  1678: p. 325 (misprinted 305 in 1678);  1715: p. 210;  1787: p. 182.  B  says he is  3/2  A.  C  says he is twice  B.  A  says the sum of their ages is  165.

Chap. 32, quest. 2.  1678: pp. 332-333;  1715: p. 214;  1787: p. 186.  A  says he is  18.  B  says he is  A + C/2.  C  says he is  A + B.

Anonymous proposer and solver.  Ladies' Diary, 1708-09  =  T. Leybourn, I: 3, quest. 5.  [I have a reference to this as Question V and Leybourn also gives this number.]  III‑(3, 15, 2).  "A person remarked that upon his wedding day the proportion of his own age to that of his wife was as three to 1; but 15 years afterwards the proportion of their ages was as 2 to 1.  What were their ages upon the day of marriage."

                    Also reproduced in The Diarian Repository, 1774, Wallis 341.5 DIA, ??NX, but I copied it as follows.  [This seems to indicate that Leybourn has condensed the material -- ??] 

                    When first the marriage knot was tied

                              Betwixt my wife and me,

                    My age to her's we found agreed

                              As nine doth unto three;

                    But after ten and half ten years,

                              We man and wife had been,

                    Her age came up as near to mine,

                              As eight is to sixteen.

                    Now tell me if you can, I pray,

                              What was our age o' th' wedding day?

Dilworth.  Schoolmaster's Assistant.  1743. 

P. 92, no. 4.  A  is  20;  B  is  A + C/2;  C = A + B.

P. 167, no. 104.  III-(2, 30, 7/5).  After another  30  years, we have  III-(2, 60, 5/4),  when they died.  "I demand ... also the reason why the lady's age, which was continually gaining upon her husband's, should, notwithstanding, be never able to overtake it."

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XVI, p. 87 (1790: prob. XXII, pp. 85-86).  Essentially  III‑(4, 14, 2),  though the present time is halfway between the times involved in the problem -- i.e.  7  years ago, the father was  4  times the son's age;  in  7  years, he will be  2  times.

Les Amusemens.  1749.  Prob. 97, p. 239.  Father is three times the age of his son, when will he be only twice as old? -- i.e.  III-(3, b, 2)  -- with answer  b = Y,  the son's present age.

Walkingame.  Tutor's Assistant.  1751. 

1777, p. 82, prob. 5;  1835: p. 91, prob. 4;  1860: p. 11, prob. 5.  H  is  30.  K = H + L/4.  L = H + K.

1777, p. 177, prob. 120;  1860: p. 185, prob. 113.  (The 1835 edition differs -- see under 1835 below.)  III-(3, 10½, 2).

                                              When first the marriage knot was ty'd,

                                                                                    Between my wife and me,

                                                   Her age did mine as far exceed,

                                                                                    As three times three does three;

                                         But when seven years, and half seven years,

                                                                                    We man and wife had been,

                                                 My age came then as near to her's,

                                                                                    As eight is to sixteen.

Quest.  What was each of our ages when we married?

(I find it extraordinary that the man is younger in this version and in no other that I have seen.  The 1860 makes a few minor changes and puts it on four lines.  Wehman, New Book of 200 Puzzles, 1908, p. 51 is an incomplete copy of this, comprising the first two lines of a four-line version, with the answer.)

Mair.  1765? 

P. 454, ex. 5.  Same as Cocker, chap. 31, quest. 2.

P. 458, ex. 5.  Same as Cocker, chap. 32, quest. 2.

P. 143, no. 11.  III‑(3, 15, 2).

                                              When first the marriage-knot was tied

                                                                                    Betwixt my wife and me,

                                                    My age did hers as far exceed,

                                                                                    As three times three doth three;

                                                    But after ten and half ten years,

                                                                                    We man and wife had been,

                                                  Her age came up as near to mine,

                                                                                    As eight is to sixteen.

                                                Now, Tyro, skill'd in numbers, say,

                                                                                    What were our ages on the wedding-day?

                                                                                                Answer.

                                                 Sir, Forty-five years you had been,

                                                                                    Your Bride no more than just fifteen.

(Text copied from 2nd ed, ??NX, and differs very slightly from the 3rd ed.)

Vyse.  Tutor's Guide.  1771? 

Prob. 3, 1793: p. 128; 1799: pp. 135-136 & Key p. 177.  A = C + 4,  B = A + C + 9,  D = 45 = A + B + C.

Prob. 5, 1793: p. 128; 1799: p. 137 & Key pp. 180-181.  B = 5A = 7(A-4).

Prob. 8, 1793: p. 130; 1799: p. 138 & Key pp. 181-182.  III-(3, 15, 2).

When first the Marriage-Knot was ty'd

                                                        Betwixt my Wife and me,

My Age did her's as far exceed

                                                   As three Times three does three;

But when ten Years, and Half ten Years,

                                                      We Man and Wife had been,

Her Age came up as near to mine

                                                           As eight is to sixteen.

Now, tell me, I pray,

                                          What were our Ages on the Wedding Day?

Dodson.  Math. Repository.  1775.

P. 6, Quest. XIV.  A + B + 25  =  2A;  A - B - 8  =  B.

P. 15, Quest. XL.  III-(4, 14, 2)  stated at the middle of the 14-year period.

P. 31, Quest. LXXVI.  Find  x/y  such that  (x+4)/(y+4)  =  4/3,  (x-4)/(y-4)  =  3/2.  [In general,  (x-a)/(y-a)  =  r,  (x+b)/(y+b)  =  s  has solution  x = {br(s‑1) + as(r‑1)}/(r-s),  y = {b(s-1) + a(r-1)}/(r-s).  Cf Vinot, below.]

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?]  1804: prob. 2, p. 132.  III-(5, 5, 4).

Bonnycastle.  Algebra.  1782.  Prob. 11, p. 84 (1815: prob. 10, p. 105).  A = 2B;  B = 3C;  A + B + C = 140.

Pike.  Arithmetic.  1788.  P. 335, no. 7.  B  =  3A/2;   C  =  21 (A + B)/10;   A + B + C  =  93.

Anon.  The American Tutor's Assistant.  (1st ed. is unknown;  2nd ed., Philadelphia, 1791);  1810 ed., Joseph Crukshank, Philadelphia.  ??NYS -- quoted in:  Lucas N. H. Bunt et al.; The Historical Roots of Elementary Mathematics; Prentice-Hall, 1976; p. 33.  III‑(3, 15, 2).

                    When first the marriage knot was ty'd

                              Between my wife and me,

                    My age was to that of my bride

                              As three times three to three

                    But now when ten and half ten years,

                              We man and wife have been,

                    Her age to mine exactly bears,

                              As eight is to sixteen;

                    Now tell, I pray, from what I've said,

                              What were our ages when we wed?

          Answer:

                    Thy age when marry'd must have been

                              Just forty-five; thy wife's fifteen.

Eadon.  Repository.  1794.  P. 297, no. 16.  III-(3, 15, 2) in verse.

                    When first the marriage knot was ty'd

                              Betwixt my wife and me,

                    My age did her's as far exceed,

                              As three times three doth three;

                    But after ten, and half ten years,

                              We man and wife had been,

                    Her age came up as near to mine,

                              As eight is to sixteen.

                    Now tell me (you who can) I pray,

                              What were our ages on the wedding day?

Hutton.  A Course of Mathematics.  1798?  Prob. 1,  1833 & 1857: 80.  III-(5, 5, 4).  On  1833: 224-231;  1857: 228-235,  he has an extensive discussion  Remarks upon Equations of the First Degree  concerning possible negative roots and considers  IV‑(42, 12, 4)  whose solution is  -2.

Bonnycastle.  Algebra.  10th ed., 1815.  P. 104, no. 7.  III-(3, 15, 2).

L. Murray.  The  Young Man's  Best Companion, and  Book of General Knowledge;  ....  Thomas Kelly, London, (Preface dated 1814; BL has 1821), 1824.  P. 177, example 2.  III-(3, 14, 2)  stated at the middle of the 14 years.

Manuel des Sorciers.  1825.  P. 81, art. 41.  ??NX  IV-(54, 18, 2).

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Art. 113, p. 29.  IV-(40, 8, 3)  -- "in how many years hence will the father's age be just three times the son's?"  Then gives general solution of  IV-(a, b, 3).  Observes that  a = 40,  b = 18  leads to  -7  and that  a = 3b  leads to  0  and discusses the meaning of both situations.  Then solves the general case  IV-(a, b, n).

Child.  Girl's Own Book.  Arithmetical Puzzles.  1832: No. 8, pp. 171 & 179;  1833: No. 8, pp. 185 & 193 (answer numbered 6);  1839: No. 8, pp. 165 & 173;  1842: No. 8, pp. 283 & 291;  1876: No. 6, pp. 231 & 244.  III‑(3, 15, 2)  given in verse which is the same in the 1832, 1839 and 1876 eds.  (Except 1839 & 1842 have  her's  in line three.)

 

                    When first the marriage knot was tied

                    Between my wife and me,

                    My age exceeded hers as much,

                    As three times three does three.

 

                    But when ten years and half ten years

                    We man and wife had been,

                    Her age approached as near to mine

                    As eight is to sixteen.

 

          The 1833 ed has the first lines of the second verse garbled as:  But when the man and wife had been, / For ten and half ten years.

                    = Fireside Amusements; 1850: Prob. 7, pp. 132 &184.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 58, pp. 87-89.  Discusses the problem  IV-(a, b, 4)  and the significance of negative solutions, using  IV-(45, 15, 4)  as an example.

Walkingame.  Tutor's Assistant.  1835.  P. 180, prob. 59.  III-(3, 15, 2).

                    When first the marriage knot was ty'd

                              Between my wife and me,

                    My age did hers as far exceed,

                              As three times three does three;

                    But when ten years, and half ten years,

                              We man and wife had been,

                    Her age came then as near to mine,

                              As eight is to sixteen.

          = Depew; Cokesbury Game Book; 1939; Marriage problem, pp. 207-208.

Augustus De Morgan.  On The Study and Difficulties of Mathematics.  First, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  P. 53.  IV-(56, 29, 2)  -- "when will the father be twice as old as the son?"  Answer is  -2  and he discusses the meaning of this.

Unger.  Arithmetische Unterhaltungen.  1838. 

Pp. 108 & 255, no. 392.  III-(6, 3, 4).

Pp. 109 & 255, no. 402.  III-(7, 15, 2).

Pp. 110 & 255, no. 405.  III-(4, -2, 6),  phrased in terms of wealth and both give away  2.

Pp. 133 & 258, no. 508.  III-(14, 10, 5⅓).

Philip Kelland.  The Elements of Algebra.  A. & C. Black, Edinburgh, et al., 1839.  ??NX.

P. 15:  IV-(40, 20, 3).

P. 105:  "A's  money or debt is  a  times  B's;  if  A  lose  £10  to  B,  it will be  b  times  B's."  This is perhaps closer to the problems in 7.R.

Augustus De Morgan.  Negative and impossible numbers.  IN:  Penny Cyclopædia, vol. XVI, 1840, pp. 130-137.  ??NX  IV-(50, 40, 2)  -- "at what date is (was, or will be, as the case may be) the first twice as old as the second?" 

The New Sphinx.  c1840.  P. 142.  III‑(3, 15, 2).

                    When first the marriage knot was tied,

                              Betwixt my love and me,

                    My age did then her age exceed

                              As three times three doth three.

                    But when we ten and half ten years

                              We man and wife had been,

                    Her age came up as near to mine,

                              As eight is to sixteen.

                                        Solution. -- The man was 45, the woman was 15.

T. Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.

P. 37, no. 20.  III-(4, 4, 3).

P. 45, no. 11.  Couple are married for    of his life and  ¼  of hers.  Man is  8  years older than the wife, who survived him by  20  years.  How old were they at marriage?

P. 45, no. 12.  Man is  32  years older than  and  5  times as old as  his son.

P. 67, no. 6.  III-(4, 10, 5/2).

Phineas Taylor Barnum, c1848, is reputed to have given a friend a problem involving ages:  a man aged  30  with a child aged  1  is  30  times as old as his child, but in  30  years he is only twice as old, and in another  30  years he is only one-third older, ..., when does the child catch up with the father?  This is given in a cartoon biography of P. T. Barnum by Walt Kelly in the early 1930s -- reproduced in:  Outrageously Pogo; ed. by Mrs. Walt Kelly & Bill Crouch Jr; Simon & Schuster, NY, 1985, pp. 14-21.  Cf Abbot & Costello, 1941.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 210, No. 4.  III‑(3, 15, 2).  No solution.

                    When first the marriage knot was tied, betwixt my wife and me;

                    My age did hers as far exceed, as three times three does three.

                    But when ten years and half ten years, we man and wife had been,

                    Her age came then as near to mine, as eight does to sixteen.

John H. Boardman.  Arithmetic: Rules and Reasons.  Macmillan, Cambridge, 1850.  P. 98.  "A.  is now twice as old as  B.;  eight years ago he was three times as old, and one year more;  find the age of each."

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  3rd ed., 1850.  Op. cit. in 7.H.

P. 353, no. 14.  III-(5, 5, 4).

P. 355, no. 1.  "Your age is now  1/5  of mine;  but  4  years ago it was only  1/7  of what mine is now".

John Radford Young (1799-1885).  Simple Arithmetic, Algebra, and the Elements of Euclid.  IN: The Mathematical Sciences, Orr's Circle of the Sciences series, Orr & Co., London, 1854.  [An apparently earlier version is described in 7.H.]

No. 3, p. 176.  III-(4, 5, 3)  phrased as five years ago.

No. 10, p. 177.  IV-(40, 9, 2).

The Family Friend (1856) 298 & 329.  A  is now one-fifth the age of  B,  but five years ago,  A  was one-seventh of  B's  present age.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 3, p. 173 (1868: 184).  "The square of my age  60  years ago is double my present age."

Magician's Own Book.  1857.  December and May, p. 246.  X + Y = 100,  Y = 4X/9.  = Boy's Own Conjuring Book, 1860, p. 216.  = Illustrated Boy's Own Treasury, 1860, prob. 19, pp. 428 & 432, but this has no title. 

Charades, Enigmas, and Riddles.  1860: prob. 17, pp. 58 & 62;  1862: prob. 16, pp. 33 & 139;  1865: prob. 560, pp. 105 & 152.  III‑(3, 15, 2).  (1862 and 1865 differ very slightly in typography.)  This is essentially identical to Child.

                    When first the marriage knot was tied

                    Between my wife and me,

                    My age exceeded her's as much,

                    As three times three does Three:

                    But when Ten years and half ten years

                    We man and wife had been,

                    Her age approached as near to mine

                    As Eight is to sixteen.

Vinot.  1860. 

Art. LII: Les âges, p. 69.  I am twice the age you were when I was the age you are now.  When you are my age, the sum of our ages will be  63.

Art. LXI: Sur les âges, pp. 77-78.  IV-(48, 12, 3);  IV-(48, 12, 7).

Art. LXVI: Retrouver une fraction, pp. 82.  (x-3)/(y-3) = 1/4;  (x+5)/(y+5)  =  1/2.  Cf Dodson, above, for the general solution of this type of problem.

Art. LXVII: Sur les âges, p. 82.  III-(3, 10, 2)  phrased as five years ago and five years hence.

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Lots of examples.  I mention only those of some novelty to illustrate mid/late 19C texts.

1863 -- p. 129, no. 10;  1873 -- p. 158, no. 9.  A = 4B;  A + B = 25;  when will  A = 3B?

1863 -- p. 144, no. 13;  1873 -- p. 174, no. 4.  "Said  E  to  F,  my age is  5  years more than yours, but  4  years ago my age was  ½  of what yours will be  4  years hence;  what was the age of each?"

1873 -- p. 175, no. 14.  "Two years ago Mr. Smith was  5  times as old as his son John will be  2  years hence, and  3  years hence his age will equal  15  times John's age  3  years ago: required the age of each."

(Beeton's) Boy's Own Magazine 3:4 (Apr 1889) 175  &  3:6 (Jun 1889) 255.  (This is undoubtedly reprinted from Boy's Own Magazine 1 (1863).)  Mathematical question 33.  III‑(10/7, 2, 4/3)

Colenso.  Op. cit. in 7.H.  These are from the new material of (1864), 1871.

No. 23, pp. 190 & 215.  IV-(62, 30, 5).

No. 24, pp. 190 & 215.  I am  24  years older than my son.  When I am twice my present age, he will be  8  times his present age.

Lewis Carroll.  Letter of 6 Feb 1873 to Mary MacDonald.  c= Carroll-Gardner, p. 50.  Congratulates her sister Lilia "Lily" on becoming 21 and adds "Why, last year I was double her age!  And once I was three times her age, but when that was, I leave you to find out."  I.e.  IV-(40, 20, 3),  cf Kelland, 1839.

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.

No. 54, pp. 108 & 328.  III-(7/3, 6, 13/9).

No. 3, pp. 166 & 334.  Asks for general solution of  III-(a, b, c)  and for  III-(6, 3, 4).

No. 90, pp. 301 & 347.  III-(2, 12, 8/5).

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Pp. 117-118.  III‑(7, 3, 4).

Hoffmann.  1893.  Chap. IV, nos. 20 and 29, pp. 149 & 194 and 151 & 196 = Hoffmann‑Hordern, pp. 121 & 123.  Simple age problems:  IV-(71, 34, 3),  IV‑(71, 34, 2),  IV-(45, 12, 3)

Clark.  Mental Nuts.  1897, no. 39.  Father and child.  III-(6, 24, 2).

Somerville Gibney.  So simple!  The Boy's Own Paper 20 (No. 992) (15 Jan 1898) 252  &  (No. 993) (22 Jan 1898) 267.  "When you are as old as I am, I shall be twice as old as you were when I was as old as you are."  This only gives the ages being in the ratio of  5  to  4.

P. Holland Lester.  Some "B.O.P." puzzles.  The Boy's Own Paper 20 (No. 1017) (9 Jul 1898) 655.  "I am twice as old as you were when I was what you are;  when you are what I am our united ages will be  63."  = Vinot.  "You are twice as old as I was when you were what I am;  when you are twice what I shall be when you are twice what I am, our united ages will be  133." 

Parlour Games for Everybody.  John Leng, Dundee & London, nd [1903 -- BLC], p. 32: Nuts to crack, no. 8.

                    A bachelor tired of single life

                    Took to himself a charming wife;

                    The damsel's name was Mary Page,

                    The bachelor was three times her age.

                    But after fifteen years had flown

                    Her husband's age was twice her own;

                    Now tell me, gentle reader, pray,

                    Their ages on their wedding day?

Clark.  Mental Nuts.

1904, no. 28: The candle question.  1916, no. 97: The candles.  "Suppose two candles, one of which will burn 4 hours and the other 5 hours, are lighted at once.  How soon will one be three times the length of the other?"  (1916 has a shortened version.)  I assumed that one candle starts out as  5/4  the length of the other candle.  Then this is like our age problem  III-(5/4, b, 3),  where the problem is to determine  b  in terms of  X  or  Y.  Or we might say that we have  X = 4,  Y = 5,  3(X-b) = (Y-b).  In any case, I get  b = 7/2 = 3 1/2.  However, Clark's answer is  3 7/11  and this arises by considering both candles to be the same length initially, but made of different materials, so they burn at different rates.  Taking the candles to be of unit length, the lengths after  b  hours are  (1 ‑ b/4)  and  (1 ‑ b/5).  Setting three time the first equal to the second gives  b = 40/11 = 3 7/11.  However, this formulation cannot be interpreted as an age problem as the candles age at different rates!

1904, no. 52;  1916, no. 64.  Man and wife.  "A man is twice as old as his wife was when he was as old as she is now.  When she reaches his present age, the sum of their ages will be 100 years.  What are their ages?"

Pearson.  1907.

     Part I, no. 39, pp. 123‑124 & 186‑187.  III-(3, 15, 2)  in verse with verse solution, very similar to previous examples.

     Part II.

No. 13: A brain twister, pp. 116 & 193.  Form I.

No. 57, pp. 124 & 202.  4  persons.

No. 70: How old was John?, pp. 128 & 205.  One person like Diophantos' age.

No. 105: A delicate question, pp. 136 & 212.  One person, in verse, using a square root.

No. 117: When was he born?, pp. 137 & 214.  One person.

No. 124: Ask any motorist, pp. 138‑139 & 215.  Car and tyres.

No. 156, pp. 144 & 222.  Two people.

No. 168: Very personal, pp. 146 & 224.  Two people, verse problem and solution.

Wehman.  New Book of 200 Puzzles.  1908.  The marriage knot, p. 51.  Gives only half of the problem!  Seems to be copied from a four-line version of Walkingame.

Loyd.  How Old was Mary?  Cyclopedia, 1914, pp. 53 & 346.  (= MPSL2, prob. 10, pp. 8 & 123.)  Form I -- he says this is a companion to his 'famous problem "How old was Ann"'.  He gives other age problems. 

                    Tell mother's age, pp. 84 & 349‑350.  (= MPSL1, prob. 85, pp. 82 & 151.) 

                    Pp. 216 & 367.  (= MPSL2, prob. 106: How old is Jimmy, pp. 75 & 155.)

G. G. Bain.  An Interview with Sam Loyd, 1907, op. cit. in 1, p. 777.  Refers to "How old was Mary?", and gives form I as in the Cyclopedia with slightly different wording but the same illustration.

A. C. White.  Sam Loyd and His Chess Problems.  1913.  Op. cit. in 1.  P. 101 calls it "How old was Mary?" and gives form I as in the Cyclopedia.

Dudeney.  AM.  1917.  Numerous problems, including the following.

Prob. 43: Mrs. Timpkins age, pp. 7 & 152.  III-(3, 18, 2).

Prob. 45: Mother and daughter, pp. 7 & 152.  IV-(45, 12, 3).

Prob. 47: Rover's age, pp. 7 & 152-153.  III-(5, 5, 3)  concealed by saying the sister was "four times older than the dog", meaning five times as old.

Prob. 51: How old was Mary?, pp. 8 & 153.  Form I, attributed to Loyd.

Hummerston.  Fun, Mirth & Mystery.  1924.  A problem in ages, Puzzle no. 85, pp. 170 & 185.  III-(3,14,2).

M. Adams.  Puzzles That Everyone Can Do.  1931.

Prob. 170, pp. 66 & 153.  A is older than B by as many years as B is older than C.  C is half as old as B was when B was half as old as A is now.  In a year's time, the combined ages of C and B will equal that of A.

Prob. 246, pp. 93 & 166: Molly & Polly.  M will be three times P in three year's time.  M is three times as old as P will be when P is three times as old as she is now.  (Since the second statement simply says  M = 9P,  this problem is actually  III‑(9, 3, 3).)

James Joyce.  Ulysses.  (Dijon, 1922);  Modern Library (Random House), NY, 1934, apparently printed 1946.  P. 663 (Gardner says the 1961 ed. has p. 679).  Humorous calculation assuming the ratio of ages could remain fixed at  17  to  1.  The numbers become wrong after a bit.

Ackermann.  1925. 

Pp. 93‑94: different problem, due to C. V. Boys.

Pp. 98‑100: form I, ending  "What are their present ages?"

Pp. 100‑102: complex version due to A. Honeysett, with four adults and an unspecified number of children, which turns out to be two, one of which has age zero.

Collins.  Book of Puzzles.  1927.  How old is Jane?, pp. 73-74.  Form I with Jane and Ann. 

Loyd Jr.  SLAHP.  1928.  How old is Ann?, pp. 4 & 87.  Gives the "original wording" as form I.  He gives numerous other age problems.

Perelman.  1937.  MCBF.  The equation does the thinking, p. 244.  IV-(32, 5, 10).

Haldeman-Julius.  1937. 

No. 45: How old is Ann?, pp. 7 & 23.  Form II, with James and Ann.

No. 86: Problem in rhyme, pp. 10-11 and 25. 

I was twice as old as you are

The day that you were born;

You'll be what I was then

When fourteen years are gone.

Answer is  42  and  14.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939. 

Number, please!, pp. 20 & 210.  One person, giving   x + 3  =  3 (x - 3).

Pp. 27 & 210.  A variation of type III, giving   X + 2  =  3Y;   X + 8   =   2 (Y + 8).

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.  No. 88: The monkey's mother, pp. 50 & 132-133.  Monkey on a rope over a pulley with a weight on the other side, but the weight of the rope is equal to the age of the monkey's mother, who was twice as old as the monkey was ....  What was the length of the rope?

R. S. S[corer].  Problem for poultry farmers.  Eureka  4 (May 1940) 4  &  5 (Jan 1941) 15.  "The chicken was twice as old when when the day before yesterday was to-morrow to‑day was as far from Sunday as to-day will be when the day after to-morrow is yesterday as it was when when to-morrow will be to-day when the day before yesterday is to-morrow yesterday will be as far from Thursday as yesterday was when to-morrow was to-day when the day after to-morrow was yesterday.  On what day was the chicken hatched out?"  Solution is: Friday.

Bud Abbott and Lou Costello.  Buck Privates.  Universal Pictures, 1941.  Text and stills in Richard J. Anobile, ed.; Who's On First?  Verbal and Visual Gems from the Films of Abbott & Costello; Darien House, NY, 1972; Studio Vista, London, 1973.  Pp. 54-59.  "You're forty years old and you're in love with a little girl who's ten years old.  ...  You're four times as old ....  ... so you wait five years.  ...  You're only three times as old ....  So wait fifteen years more ....  You're only twice as old ....  How long do you have to wait until you and that little girl are the same age?"  Cf Barnum, c1848.

Clark Kinnaird.  Loc. cit. in 1.  1946.  Pp. 265‑266, says Loyd Jr. is best known for this problem (form I).  On p. 267, he says it  "made Sam Loyd [Jr.] famous, although he did not originate it .... Loyd based his version upon a similar poser which went around like a chain‑letter fad in the early years of this century ...."  He also says  'References to "How Old is Ann" have been found as far back as 1789 ....', but he doesn't give any such references!

Owen Grant.  Popular Party Games.  Universal, London, nd [1940s?].  Prob. 17, pp. 39-40 & 52.  Form I with Mary and Ann replaced by Smith and Robinson.

Meyer.  Big Fun Book.  1940.  No. 10, pp. 167 * 753.  "The sum of our ages is  22.  I shall be  7  times as old as you are now when I become twice your age." 

H. Phillips.  News Chronicle "Quiz" No. 5: Dickens.  News Chronicle, London, 1946.  Pp. 14 & 37.  III-(2, 20, 6/5)  phrased as seven years ago and thirteen years hence.

Karl Menninger.  (Mathematik in deiner Welt.  Vandenhoek & Ruprecht, Göttingen, 1954;  revised, 1958.)  Translated as:  Mathematics in Your World; G. Bell, London, 1961.  The complicated problem of Anne and Mary, pp. 65‑66.  Gives form II with no history.

William R. Ransom.  Op. cit. in 6.M.  1955.  How old is Ann?, p. 91;  Mrs. M. & Miss A., p. 92.  He first gives form II and says  "This problem raged throughout the United States in the early 1900's.  It was concocted by Robert D. Towne, who died at the age of 86 in 1952."  He then gives form I and says  "This is a much older problem, of the same type as our "How old is Ann?" which has circulated mostly in England."

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  P. 62: A problem of age.  "When I am as old as my father is now I shall be five times the age my son is now.  By then my son will be eight years older than I am now.  The combined ages of my father and myself total 100 years.  How old is my son?"

Young World.  c1960.  P. 53: A matter of time.  In two years,  X  will be twice as old as she was  20 years ago.

B. D. Josephson  &  J. M. Boardman.  Problems drive 1961.  Eureka 24 (Oct 1961) 20-22 & 24.  Prob. A.  "When A was three times as old as B was the year before A was a half of B's present age, B was 3 years younger than A was when B was two thirds of A's present age.  A's and B's ages now total 73.  How old are A and B?"

R. L. Hutchings  &  J. D. Blake.  Problems drive 1962.  Eureka 25 (Oct 1962) 20-21 & 34-35.  Prob. K.  "Tom is twice as old as Dick was when Tom was half as old as Dick will be when Tom is twice as old as Dick was when Tom was a year younger than Dick is now.  Dick is twice as old as Tom was when Dick was half as old as Tom was when Dick was half as old as Tom was two years ago.  How old are Dick and Tom?"

L. S. Harris  &  J. M. Northover.  Problems drive 1963.  Eureka 26 (Oct 1963) 10-12 & 32.  Prob. A.  "When A was half B's present age, B's age was the square of A's age when B was born; the sum of their ages is a perfect cube.  How old are they both now?  (Take (1 year)2 = 1 year.)"  Given solution is:  65, 60.  However the problem is quadratic and there is a second solution:  59 3/8,  65 5/8  for which A's age when B was born is  ‑6 1/4.

Doubleday - 2.  1971.

Passing years, pp. 15-16.  Form I.

It's not so simple!, pp. 47-48.  Form II.

Angela Fox Dunn.  Mathematical Bafflers.  Dover, 1980.  Selected from Litton's Problematical Recreations, which appeared from 1971, ©1964.  An age problem, pp. 18 & 42.  "Lottie and Lucy Hill are both 90 years old.  Mary Jones, on the other hand, is half again as old as she was when she was half again as old as she was when she lacked 5 years of being half as old as she is now."  The solution is 90, i.e. Mary is as old as the Hills!

David Singmaster.  Some diophantine recreations.  Op. cit. in 7.P.5.  1993.  Determines when integer data in  III-(a, b, c)  gives integer solutions.

 

          7.Y.   COMBINING AMOUNTS AND PRICES INCOHERENTLY

 

            Often called the applesellers' problem or the market women's problem.

            Two persons have the same number of items which they combine.  They average the number of items per unit cost rather than the cost per item.

            See Tropfke 652.

 

Alcuin.  9C.  Prob. 6: Propositio de duobus negotiatoribus  C  solidos communis habentibus.  Buy  125 & 125  at  5 for 2,  costing  100,  sell at  2 for 1  and  3 for 1, leaving  5 & 5  left over, so they received  100  and the leftovers are worth  4 1/6.

Ibn Ezra, Abraham (= ibn Ezra, Abraham ben Meir  = ibn Ezra, Abu Ishaq Ibrahim al‑Majid).  Sefer ha‑Mispar.  c1163.  Translated by Moritz Silberberg as:  Das Buch der Zahl  ein hebräisch-arithmetisches Werk; J. Kauffmann, Frankfurt a. M., 1895.  P. 42.  Buys  100 for 100,  sells  50 & 50  at  5 for 4  and  3 for 4,  makes  6⅔.  Silberberg's note 95, p. 107, says that the the same problem occurs in Elia Misrachi, c1500.

Fibonacci.  1202.  P. 281 (S: 402).  Buys  100 for 100  and sells  50 & 50  at  4/5  and  4/3.  What profit?

Folkerts.  Aufgabensammlungen.  13-15C.  Buy  1000  birds at  4 for 1.  Sell  500 & 500  at  5 for 1  and  3 for 1,  i.e. at the same average price, but he has a profit of  16  and  2  birds.  5 sources for similar problems.  Cites Alcuin and AR.

Provençale Arithmétique.  c1430.  Op. cit. in 7.E.  F. 113r, pp. 59-60.  Buy  60 for 24,  sell as  x & y  at  a & b  to make profit of  1.  How?  This is actually an indeterminate question, complicated by prices which are non-integral.  The author seems to give just one solution:  x = y = 30,  a = 1/2,  b = 1/3.

AR.  c1450.  Prob. 354, pp. 154, 183, 229‑230.  Buy  100  at  5 for 2,  sell  50 & 50  at  3 for 1  and  2 for 1,  making  1⅔.

Pacioli.  De Viribus.  c1500. 

Ff. 119v - 120r.  LXVI. C(apitolo). D. de uno ch' compra 60. perle et revendele aponto per quelli ch' gli stanno et guadago (Of one who buys 60 pearls and resells for exactly what they cost and gains).  = Peirani 155-156.  Buy  60  at  5 for 2,  sell  30 & 30  at  2 for 1  and  3 for 1.

F. IVr.  = Peirani 7.  The Index lists the above as Problem 74 and continues with Problem 75: De unaltro mercante ch' pur compro perle' .60. a certo pregio per certa quantita de ducati et sile ceve'de pur al medesimo pregio ch' lui le comparo et guadagno un ducato ma con altra industria dal precedente (Of another merchant who buys 60 pearls at a certain price for a certain quantity of ducats and resells them at the same price at which he bought them and gains a ducat but with different effort than the preceding).

Tartaglia.  General Trattato, 1556, art. 160‑162, p. 259v.

160:  buy  60 & 60  at  5 for 1,  sell at  2 for ½  and  3 for ½,  make 1.

161:  buy  42 & 42  at  7 for 12,  sell at  3 for 6  and  4 for 6,  make 3.

163:  buy  270 & 270  at  9 for 12,  sell at  4 for 6  and  5 for 6,  make 9.

Buteo.  Logistica.  1559.  Prob. 17, pp. 215-217.  Buy  60 for 24,  i.e. at  5 for 2.  How were they sold 'at the same price' to make  1?  Sell at  2 for 1  and  3 for 1  and gain  1.  Says this was proposed by Stephanus and that the apples aren't sold at the same price as they were bought.  (H&S 53 gives Latin.)

Ozanam.  1725.  Prob. 51, art. 1, 1725: 258.  Buy  20  at  5 for 2;  sell  10 & 9  at  2 for 1  and  3 for 1  to recover cost and have  1  left over.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XVII, pp. 87-88 (1790: prob. XXV, p. 87).  Buy  x & x  at  2 for 1  and  3 for 1,  sell at  5 for 2,  lose 4 -- what was  x?

Vyse.  Tutor's Guide.  1771?  Prob. 28, 1793: p. 57; 1799: p. 62 & Key p. 68.  Buy  120 & 120  at  2 for 1  and  3 for 1,  sell at  5 for 2,  lose 4  -- "Pray how comes that about?"

Dodson.  Math. Repository.  1775.  P. 24, Quest LXII.  Same as Simpson.

Bonnycastle.  Algebra.  1782.  P. 82, no. 9 (1815: pp. 101-102, no. 8).  Same as Simpson.

Pike.  Arithmetic.  1788.  P. 492, no. 9.  Same as Simpson.

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 55.  Buy  120 & 120  at  3 for 1d  and  2 for 1d.  Do I gain or lose by selling at  5 for 2d?  He finds a loss of  4d.  On pp. 152-153, the Editor discusses Alcuin, where pigs are sold at three different prices, but he knows no example with four or more different prices.  On p. 163, he mentions almost the same problem, with prices multiplied by  12.  [I think he has misinterpreted Alcuin -- the left over pigs are sold at the same prices as the others.]

Robert Goodacre.  Arithmetic  &  A Key to R. Goodacre's Arithmetic.  2nd ed., T. Ostell & C. Law, London, 1804.  Miscellaneous Questions, no. 128, p. 205 & Key p. 270.  Buy  30 & 30  at  3 for 1  and  2 for 1,  sell at  5 for 2.  Does he gain or lose?

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 21, pp. 19 & 76‑77.  Buy  120 & 120  at  2 for 1  and  3 for 1,  sell at  5 for 2  and discover a loss of  4.  c= Magician's Own Book (UK version), 1871, The costermonger's puzzle, pp. 38-39.

Boy's Own Book.  Profit and loss.  1828: 414;  1828-2: 418-419;  1829 (US): 212;  1843 (Paris): 348;  1855: 566;  1868: 669.  Buy  96 & 96  at  3 for 1  and  2 for 1,  sell at  5 for 2  giving  2  left over and a loss of  4  and says the loss is a fraction over  3½.  = Boy's Treasury, 1844, p. 306.  = de Savigny, 1846, p. 294: Profit et perte.

Augustus De Morgan.  Examples of the Processes of Arithmetic and Algebra.  Third, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Prob. 11, p. 28.  Buy  150  at  3 for 1  and  100  at  2 for 1,  sell at  5 for 2  leads to no loss.  But buying  150 & 150  at  3 for 1  and  2 for 1,  selling at  5 for 2  leads to a loss.  "What was the reason of this?"

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 13,  1857: 81.  Buy  180 & 180  oranges at  3 for 1d  and  2 for 1d.  Does he gain or lose by selling at  5 for 2d?  He finds a loss of  6d. 

Magician's Own Book.  1857.  The astonished farmer, p. 244.  Compare selling  30 & 30  at  3 for 1  and  2 for 1  versus selling  60  at  5 for 2,  which makes  1  less.  = Boy's Own Conjuring Book, 1860, p. 214.  = Illustrated Boy's Own Treasury, 1860, prob. 16, pp. 428 & 432. 

The Sociable.  1858.  Prob. 35: The market woman's puzzle, pp. 297 & 315.  Same as Jackson.  = Book of 500 Puzzles, 1859, prob. 35, pp. 15 & 33.  = Wehman, New Book of 200 Puzzles, 1908, p. 29.

Book of 500 Puzzles.  1859. 

Prob. 35: The market woman's puzzle, pp. 15 & 33.  Same as Jackson and The Sociable.

The astonished farmer, p. 58.  Identical to Magician's Own Book.

[Chambers].  Arithmetic.  Op. cit. in 7.H.  1866?  P. 250, quest. 6.  Buy  x & x  at  2 for 1  and  3 for 1.  Sell at  5 for 2  and lose  8d.  How many were bought?

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.  Buy  x  apples at  5 for 2;  sell half at  2 for 1,  the other half at  3 for 1;  make  1.

Cassell's.  1881.  Pp. 100‑101: The costermonger's puzzle.  Buy  120 & 120  at  2 for 1  and  3 for 1,  sell at  5 for 2  and lose  4.

Hoffmann.  1893.  Chap. IV, no. 6: A little miscalculation, pp. 146 & 182 = Hoffmann‑Hordern, p. 114.  Buy  120 & 120  at  4 for 1  and  6 for 1,  then sell at  10 for 2.

Loyd.  The lost cent.  Cyclopedia, 1914, pp. 39 & 344 (plus mention on p. 153).  = MPSL1, prob. 60 -- The missing pennies, pp. 58 & 142  = SLAHP: The apple mystery, pp. 44 & 100.  Calls it also the Covent Garden Problem.  Two applesellers with  x & x  being sold at  2 for 1  and  3 for 1.  They combine and sell at  5 for 2,  losing  7.  Further the proceeds are divided equally -- how much did the  2 for 1  lady lose?

Mittenzwey.  1917: 137, pp. 26 & 76.  Statement is rather vague, but the solution is the situation of Alcuin, who is cited.

Ahrens.  A&N, 1918, pp. 85‑87, discusses this problem, gives Alcuin's problem and the following.  Sell  30 & 30  at  1 for 5  (= 2 for 10)  and  3 for 10,  combine and sell  60  at  5 for 20,  losing  10  thereby.

Hummerston.  Fun, Mirth & Mystery.  1924.  The lost pound, Puzzle no. 35, pp. 91 & 177.  30 & 30  at  3 for 1  and  2 for 1,  combine and sell  60  at  5 for 2,  losing  1.

Dudeney.  PCP.  1932.  Prob. 17: The missing penny, pp. 18 & 129.  = 536, prob. 23, pp. 8 & 229.

McKay.  Party Night.  1940.  No. 26, p. 182.  "A man bought equal quantities of apples at  2  a penny and at  3  a penny.  He sold them all at  5  for twopence.  Did he gain or lose?"  Does with  30 & 30.

Sullivan.  Unusual.  1947.  Prob. 33: Another missing dollar.  Compare selling  30 & 30  at  3 for 1  and  2 for 1  versus selling  60  at  5 for 2,  which makes  1  less.

 

          7.Y.1.          REVERSAL OF AVERAGES PARADOX

 

            Example.  Player  A  gets  1 for 1  and  1 for 2  while player  B  gets  8 for 9  and  1 for 3.  A  has averaged better than  B  in each part, but overall  A  has  2  for  3  while  B  has  9  for  12  and has averaged better overall.  I have now computed small examples and the examples using the smallest integer values are:  2 for 1  and  2 for 4  versus  4 for 3  and  1 for 3  and  3 for 1  and  3 for 4  versus  4 for 2  and  1 for 2.  Allowing numerators of  0  gives a simpler? example:  2 for 1  and  1 for 3  versus  3 for 2  and  0 for 1.  By going up to values of  15,  one can have  A  being twice as good as  B  in each part but  B  is twice as good overall!.  E.g.  2 for 1  and  2 for 15  versus  13 for 13  and  1 for 15.

            This is sometimes called Simpson's Paradox.

            New section -- having now found Hummerston, it seems likely there are other early examples.

 

Hummerston.  Fun, Mirth & Mystery.  1924.  Bowling averages, Puzzle no. 77, pp. 165 & 184.  A & B  both have  42 for 90  (wickets for runs) and then  A  does much better, getting  6 for 54  while  B  gets  1 for 39,  but they have the same overall average.

Morris R. Cohen & Ernest Nagel.  An Introduction to Logic and Scientific Method.  Harcourt, 1934.  ??NYS -- cited by Newson below.  I have an abridged student's edition which doesn't seem to have the example described by Newson.

Rupert T. Gould.  The Stargazer Talks.  Geoffrey Bles, London, 1944.  A Few Puzzles -- write up of a BBC talk on 10 Jan 1939, pp. 106-113.  Cricket version.  A  takes  5  wickets for  30  runs,  B  takes  5 for 31.  Then  A  takes  3 for 12  and  B  takes  7 for 29.  But overall  B  is better as the totals are  8 for 42  and  12 for 60.  On p. 113, his Postscript gives another version due to a correspondent:  28 for 60  and  28 for 60  combined with  4 for 36  and  1 for 27  give the same total averages of  32 for 96  and  29 for 87.

E. H. Simpson.  ??  J. Royal Statistical Society B, 13:2 (1951) ??NYS -- cited by Newson below.

R. L. Bolt.  Class Room Note 19:  Cricket averages.  MG 42 (No. 340) (May 1958) 119‑120.  In cricket one says, e.g. that a bowler takes  28  wickets for  60  runs, but one considers the runs per wicket.  A  and  B  both take  28  wickets for  60  runs, then continue with  4 for 36  and  1 for 27.  A  seems to be better than  B,  but both have averaged  3  runs per wicket.  He gives a clear graphic explanation of such perplexities and constructs  10 for 26  and  5 for 44  versus  10 for 22  and  10 for 78  as an example where the second is better at each stage but worse overall.

Colin R. Blyth.  On Simpson's paradox and the sure-thing principle.  J. American Statistical Association 67 (Jun 1972) 364-381.  ??NYS -- cited by Gardner below.

Gardner.  SA (Apr 1976)  c= Time Travel, chap. 19.  Time Travel gives a number of more recent references up to 1985.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 18: Cricket commentary, pp. 16 & 75-76.  Two bowlers have each taken  10  wickets for  70  runs, then get  1 for 15  and  2 for 26.  The latter has done better in this last match, but is worse overall.

Clifford Wagner.  Simpson's paradox in real life.  American Statistician 36:1 (Feb 1982) 46‑48.  ??NYS -- cited by Gardner in Time Travel and by Newson below.

Donald Watson.  Note 72.23:  Combination of ratios.  MG 72 (No. 460) (Jun 1988) 126‑127.  Graphical and other analyses.

Nick Lord.  Note 74.11:  From vectors to reversal paradoxes.  MG 74 (No. 467) (Mar 1990) 55‑58.  He says the paradox is called "Simpson's reversal paradox", but gives no reference.  [Another source cites E. H. Simpson without reference.]  He discusses various interpretations of the phenomenon.

Graham Newson.  Simpson's paradox revisited.  MG 75 (No. 473) (Oct 1991) 290-293.  Cites the Simpson paper and some other recent papers, but with a lamentable lack of details.  He gives an example from "old SMP Puzzle Corners" which deals with cricket averages in 1906-07.  He quotes three examples from Wagner, one of which is taken from Cohen & Nagel referring to incidence of tuberculosis in 1910.

 

          7.Y.2.          UNFAIR DIVISION

 

New section.  The first problem might be considered as having some resemblance to 7.E and 7.J, but the second is novel.

 

Magician's Own Book.  1857.  The unfair division, p. 255.  Farmer is to give  2/5  of his yield to the landlord, but the farmer uses  45  bushels of the harvest before they can divide it.  He then proposes to give  18  bushels to the landlord and then divide up the rest.  Is this correct?  = Book of 500 Puzzles, 1859, p. 64.  = Boy's Own Conjuring Book, 1860, p. 225, which has some mathematical misprints.

Clark.  Mental Nuts.  1897, no. 30.  The gentleman and his tenant.  Landlord and tenant are sharing a harvest equally.  They are carrying away their shares each day.  One day they each have 20 bushels on their wagons when the tenant's wagon breaks down.  They shift the 20 bushels onto the landlord's wagon and take it all to the landlord's.  The landlord then says that the tenant should get an extra 20 bushels on the next day to compensate for the 20 taken to the landlord's.  Is that correct?

 

          7.Z.    MISSING DOLLAR AND OTHER ERRONEOUS ACCOUNTING

 

            The withdrawal version is the confusion caused by adding the amounts remaining, which has no meaning. 

            The missing dollar version mixes payments and refunds.  E.g. three people pay  $10  each for a triple room.  The landlord decides they were overcharged and sends  $5  back with the bellhop.  Perplexed by dividing  $5  by 3, he appropriates  $2  and refunds each person  $1.  Now they have paid  $9  each, making  $27,  and the bellhop has  $2,  making  $29 in all.  But there was  $30  originally.  What happened to the other dollar?

            I have just added the swindles.

 

Walkingame.  Tutor's Assistant.  1751.  1777: p. 177, prob. 118;  1835, p. 180, prob. 57;  1860: p. 185, prob. 116.  "If  48  taken from  120  leaves  72,  and  72  taken from  91  leaves  19,  and  7  taken from thence leaves  12,  what number is that, out of which, when you have taken  48, 72, 19, and 7,  leaves  12?"  Though this is not the same as the withdrawal problems below, the mixing of amounts subtracted and remainders makes me think that this kind of problem may have been the basis of the later kind.

Mittenzwey.  1880. 

Prob. 133, pp. 27-28 & 78;  1895?: 151, pp. 31 & 80-81;  1917: 151, pp. 28-29 & 78.  Barthel sees two boxes at a jeweller's, priced at  100  and  200.  He buys the cheaper one and takes it home, where he decides he really prefers the other.  He returns to the jeweller and gives him the box back and says that the jeweller already has  100  from him, which together with the returned box, makes  200,  which is the cost of the other box.  The jeweller accepts this and gives Barthel the other box and Barthel goes on his way.  Is this correct?

Prob. 134, pp. 28 & 78;  1895?: 152, pp. 31-32 & 81;  1917: 152, pp. 29 & 78.  Eulenspiegel (a common German name for a trickster) buys a horse for  60,  paying half and owing the rest, giving a written IOU.  After some time, the seller comes for his money.  Eulenspiegel argues that he cannot pay the man, because he has promised to owe him the money!  [The text is a bit unclear here - Eulenspeigel is a master of obfuscation!]

Lost Dollar Puzzle.  Puzzle trade card, c1889, ??NX -- seen at Shortz's.

Sigmund Freud.  Jokes and their Relation to the Unconscious.  (As: Der Witz und seine Beziehung zum Unbewussten, 1905); translated and edited by James Strachey; The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. VIII (The Hogarth Press and R&KP, 1960).  The Penguin Freud Library, vol. 6, edited by Angela Richards, Pelican, 1976, p. 94.  A gentleman entered a pastry-cook's shop and ordered a cake; but he soon brought it back and asked for a glass of liqueur instead.  He drank it and began to leave without having paid.  The proprietor detained him.  "What do you want?" asked the customer.  "You've not paid for the liqueur."  "But I gave you the cake in exchange for it."  "You didn't pay for that either."  "But I hadn't eaten it."

                    [The idea of the above is not always humorous.  When I was a student in Berkeley, I went into my bank in the middle of the day when it was quite busy and went to the first counter, near the door.  An elderly and rather shabby man was in front of me and handed the cashier some jumbled up one dollar bills, asking for a ten dollar bill.  She flattened them out and counted them and said there were only nine bills, setting them between her and the man.  The man apologised, fumbled in his pockets and came out with another one dollar bill, whereupon she put down a ten dollar bill between them.  The man then pushed the pile of ten ones and the ten toward her and asked for a twenty, which she gave him.  He took it and shuffled out the door, at which point the cashier screamed "I've been done!"  But by the time a manager got to the door, the man was nowhere to be seen.  Apparently this is a famous swindle.]

Cecil B. Read.  Mathematical fallacies.  SSM 33 (1933) 575‑589.  Gives a version with  $50  in the bank being withdrawn.  Withdraw  $20  leaving  $30;  withdraw  $15  leaving  $15;  withdraw  $9  leaving  $6;  withdraw  $6  leaving  $0.  But  $30 + $15 + $6 = $51.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  What happened to the shilling?, pp. 82 & 213.  Three girls paying  5s  each to share a room, landlord refunds  5s,  lift boy appropriates  2s.

Meyer.  Big Fun Book.  1940.  Where did the dollar go?, pp. 111 & 735.  3  men paying  $30.  = Jerome S. Meyer; Fun-to-do; op. cit. in 5.C; 1948; prob. 71: Where did the dollar go?, pp. 55-56 & 195.

Harriet Ventress Heald.  Mathematical Puzzles.  Booklet 171,  Educational Research Bureau, Washington, 1941.  Prob. 9, pp. 6‑7.  $50  being withdrawn from a bank.

Bud Abbott and Lou Costello.  Buck Privates.  Universal Pictures, 1941.  Text and stills in Richard J. Anobile, ed.; Who's On First?  Verbal and Visual Gems from the Films of Abbott & Costello; Darien House, NY, 1972; Studio Vista, London, 1973.  Pp. 47-51.  "Loan me $50."  "I've only got $40."  "Thanks, now you owe me $10."

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 8-9;  1945: 8;  1961: 18.  3  men paying a bill of  $30  or  30s.

W. A. Bagley.  Paradox Pie.  Op. cit. in 6.BN.  1944.  No. 12: The financier, p. 15.  Withdrawing  £100.

Leeming.  1946.  Chap. 3, prob. 5: What happened to the dollar?, pp. 19‑20 & 152.  3  men paying  $30.

Sullivan.  Unusual. 

1943.  Prob. 8: Balancing the checkbook.  Withdrawing  $50.

1947.  Prob. 32: Where is the dollar?  3  men paying  $30.

"Willane".  Willane's Wizardry.  Academy of Recorded Crafts, Arts and Sciences, Croydon, 1947.  Easy money, p. 47.  Withdrawing  £100.

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966.  Prob. 4/26 (b): The dishonest waiter, pp. 44 & 85.

 

          7.AA.          NEGATIVE DIGITS

 

            This is the use of  ‑5, ‑4, ..., 0, 1, ..., 5  instead of  0, 1, ..., 9.  See Knuth in 7.AA.1 for a general discussion of positional number systems, including negative digits, e.g. balanced ternary.

 

J. Colson.  A short account of negativo‑affirmative arithmetick.  Philos. Trans. Roy. Soy. 34 (1726) 161‑173.  He describes the use of negative digits quite clearly.  All work is done in the decimal system.  In concluding, he mentions "the several other species, as duodecimal, sexagesimal, centesimal, etc."

John Leslie.  The Philosophy of Arithmetic.  Constable, Edinburgh & Longman, London, 1817.  Pp. 33-34, 54, 64-65, 117, 150.  I have the 2nd ed, 1820, ??NYR.

Thomas Fowler (1777-??), of Great Torrington, Devon, built a calculating machine using base 3 with negative digits, i.e. using the digits  0, 1, -1 (written T).  It was made of wood,  6' x 3' x 1'.  He exhibited it in May 1840 at King's College London, where Babbage, De Morgan, Airy and others came to see it.  He proposed using decimal in a later machine because of the labour of converting to and from ternary, but he suggested using balanced decimal, which would still require conversions.  Memorial window in St. Michael's Church, [Great] Torrington, showing the machine.  Sadly, neither the machine nor any drawings for it survive, but a working partial model based on extant descriptions was built in Aug 2000 and is on display in the Torrington Museum and Archive.  [See:  John McKay & Pamela Vass; Thomas Fowler; www.thomasfowler.org.uk .  Mark Clusker; Thomas Fowler's ternary calculating machine: how a nineteenth century inventor's departure from decimal presaged the modern binary computer, BSHM Newsletter 46 (Summer 2002) 2-5.]

A. L. Cauchy.  CR 11 (1840) 789-798.  ??NYS.

Léon Lalanne.  CR 11 (1840) 903-905.  ??NYS -- described in Knuth, op. cit. in 7.AA.1.  Introduces balanced ternary.  Knuth cites some mid 20C discussions of the system as a possible system for computers.

J. Halcro Johnston.  The Reverse Notation, Introducing Negative Digits with  12  as the Base.  Blackie, London, 1937.  ??NYS.

C. A. B. Smith.  A new way of writing numbers.  Eureka  5 (Jan 1941) 7-9  &  6 (May 1941) 11.  General exposition, citing Cauchy, Johnson.  Discusses the method for base 6 and gives a  7 x 6  crossnumber puzzle in this system.

Cedric A. B. Smith.  Biomathematics.  (Originally by W. M. Feldman, 1923.  3rd ed by Smith, 1954.)  4th ed, in two volumes, Hafner, 1966 & 1969.  Chap. 23: Colson notation: Arithmetic made easy, pp. 611-624.  Gives an exposition of the idea.  In the Appendix: Tables, pp. 631-632 & 651-661, he gives Colson versions of tables of squares, common logarithms, sines and cosines, Woolf's function (2x ln x, used in testing contingency tables) and common logarithms of factorials

J. Halcro Johnston.  Two‑way arithmetic.  MiS 1:6 (Sep 1972) 10‑12.

C. A. B. Smith.  Looking glass numbers.  JRM 7 (1974) 299‑305.

E. Hillman, A. Paul & C. A. B. Smith.  History of two‑way numbers  &  Bibliography of two‑way numbers.  Colson News 1:4 (Dec 1984) 45‑46  &  47.  The bibliography lists 14 items which are all that are known to the authors.  Additional references in 2:1 (Mar 1985) 1.  [Smith produced about 12 issues of this newsletter devoted to unusual number systems.]

 

          7.AA.1.       NEGATIVE BASES, ETC.

 

          For binary, see 7.M.

          The Duodecimal Bulletin has regular articles discussing various bases.

 

Georg Cantor.  Zeitschrift für Math. und Physik 14 (1869) 121-128.  ??NYS -- Knuth, below, says this is the first general treatment of mixed base systems.

Vittorio Grünwald.  Intorno all'arithmetica dei sisteme numerici a base negativa.  Giornale di Matematiche di Battaglini 23 (1885) 203‑221 & 367.  ??NYS -- cited by Glaser, op. cit. in 7.M, pp. 94 & 109..

N. G. de Bruijn.  On bases for the set of integers.  Publ. Math. Debrecen 1 (1950) 232‑242.  ??NYS -- cited by Knuth & Gardner, below.  Knuth says it has representations with negative bases, but doesn't do arithmetic.

Donald E. Knuth.  Paper submitted to a Science Talent Search for high-school seniors in 1955.  ??NYS -- described in Knuth, below.  Discussed negative bases and complex bases. 

G. F. Songster.  Master's thesis, Univ. of Pennsylvania, 1956.  ??NYS -- Knuth, below, says it studies base  -2.

Z. Pawlak & A. Wakulicz.  Bull. de l'Acad. Polonaise des Sciences, Classe III, 5 (1957) 233‑236; Série des sciences techniques 7 (1959) 713-721.  ??NYS -- Knuth, below, cites this and the next item as the first mentions of negative base arithmetic in print.

Louis R. Wadel.  Letter.  IRE Transactions on Electronic Computers EC‑6 (1957) 123.  ??NYS -- cited by Knuth & Gardner, below.

W. Parry.  Acta Math. (Hungar.) 11 (1960) 401-416.  ??NYS -- Knuth says he treats irrational bases.

Donald E. Knuth.  The Art of Computer Programming:  Vol. 2: Seminumerical Algorithms.  Addison-Wesley, Reading, Massachusetts, 1969.  Section 4.1: Positional arithmetic, pp. 161-180 is an exposition of various bases.  The fact that any positive number can be used as base seems to first appear in Pascal's De numeris multiplicibus of c1658.  He suggested base  12.  Erhard Wiegel proposed base  4  from 1673.  Joshua Jordaine's Duodecimal Arithmetick, London, 1687, obviously expounded base  12.  Juan Caramuel Lobkowitz's Mathesis biceps 1 (Campaniae, 1670) 45-48 discussed bases  2, 3, 4, 5, 6, 7, 8, 9, 10, 12 and 60.  Charles XII of Sweden seems to have invented base  8  c1717 and also considered base  64.  John W. Nystrom developed base  16  in:  J. Franklin Inst. 46 (1863) 263-275, 337-348 & 402-407.  Knuth then discusses negative and complex bases -- see above items -- and describes bases  2i  and  i-1  and the use of balanced ternary and negative digits in general.

Gardner.  SA (Apr 1973)  c= Knotted, chap. 8.  Exposits negative bases, which seem to have been invented c1955 by Donald Knuth.  But the Addendum in Knotted cites Grünwald via Glaser.  Gardner cites several other articles.

Daniel Goffinet.  Number systems with a complex base: a fractal tool for teaching topology.  AMM 98 (1991) 249-255.  Explores the set of all sums of distinct powers of the base, using base  1/2,  i,  etc.

Fred Newhall.  History of the duo-decimal, base  12,  dozenal idea, chronologically.  The Duodecimal Bulletin 37;:2; (No. 73; (11*2) 4‑6 [i.e. 43:2 (No. 87) (1994) 4-6].  Outline chronology with 42 entries.

Paul Gee.  Mad Hatters maths!  MTg 174 (Mar 2001) 15.  Students asked about using base  -2  and then base  Ö2  and then base  π.

 

          7.AB.           PERFECT NUMBERS, ETC.

 

            This is too lengthy a subject to cover in detail here.  Below are a few landmarks.  See Dickson I, ch. I for an extended history.  Heath's notes summarise the history.  I have a separate file on the history of Mersenne and perfect numbers. 

            Let  Mn  =  2n - 1  and  Pn  =  2n‑1(2n-1).  Mn  is known to be prime, and hence  Pn  is perfect for  n =  2, 3, 5, 7, 13, 17, 19, 31, 61, 89, 107, 127, 521, 607, 1279, 2203, 2281, 3217, 4253, 4423, 9689, 9941, 11213, 19937, 21701, 23209, 44497, 86243, 110503, 132049, 216091, 756839, 859433.

            Paul Poulet (1918) found two amicable chains, one starting at 12496, and Henri Cohen found seven more, one of which starts at 14316 and has 28 links.

 

Euclid.  (The Thirteen Books of Euclid's Elements, edited by Sir Thomas L. Heath;  2nd ed., (CUP, 1925??);  Dover, vol. 2, pp. 278, 293-294 & 421-426.)

VII, def. 22.  "A perfect number is that which is equal to its own parts." 

IX, prop. 36.  "If as many numbers as we please beginning from an unit be set out continuously in double proportion, until the sum of all becomes prime, and if the sum multiplied into the last make some number, the product will be perfect."

In HGM I 74, he says that this is the earliest appearance of the concept of perfect number.

Marcus Vitruvius [Pollo].  De Architectura.  c-20.  Translated by Morris Hicky Morgan as: Vitruvius  The Ten Books on Architecture; Harvard Univ. Press, 1914 = Dover, 1960.  Pp. 73-75.  Rather general discussion of special numbers and mentions six as being perfect according to the mathematicians and gives some related properties of ten and sixteen.

Nicomachus of Gerasa (c100).  Introduction to Arithmetic.  Translated by Martin Luther D'Ooge, with studies in Greek Arithmetic by Frank Egleston Robbins and Louis Charles Karpinski.  University of Michigan Studies  Humanistic Series  Vol. XVI.  Macmillan, London, 1926.  (The translation was also separately printed with the same pagination, in:  The Classics of the St. John's Program; St. John's College Press, Annapolis, 1960; special edition of 250 copies for the College.)  Chap. XIV - XVI, pp. 207-212.  Defines abundant (which he calls superabundant), deficient and perfect numbers.  Gives  6, 28, 496, 8128.  He implies there is one in each range (i.e. with a given number of digits) and states that they end alternately in 6 and 8 and that Euclid's rule gives all of them, correctly noting that  Mn  must be a prime.  He seems to be the first to claim the perfect numbers alternately end in  6  and  8.

Theon of Smyrna (c125) defines abundant and deficient numbers.

Iamblichus.  On Nicomachus's Introduction to Arithmetic.  c325.  ??NYS -- cited in Dickson I 38.  Gives first four perfect numbers.  Earliest known reference to amicable numbers, attributed to Pythagoras, giving the first pair:  220, 284. 

There was Arabic interest in these numbers.  Thabit ibn Qurra (Thābit ibn Qurra) (836-901) gave a complex rule to produce amicable numbers, but apparently could not find any.   Ibn al-Banna (Ibn al-Bannā’) (1256-1321) discovered the second known pair of amicable numbers:  17296 and 18416,  apparently using Thabit ibn Qurra's rule.  [Q. Mushtaq & A. L. Tan; Mathematics: The Islamic Legacy; Noor Publishing House, Farashkhan, Delhi, 1993, pp. 93-94.]

Jordanus de Nemore.  De Arithmetica.  c1225.  An edition by Lefèvre d'Etaples was printed in Paris in 1496.  Book VII, prop. 53-60.  ??NYS -- described in: Nobuo Miura; Charles de Bovelles and perfect numbers; Historia Scientarum 34 (1988) 1-10.  He knows of the existence of odd abundant numbers, despite Dickson's assertion to the contrary.

Paolo dell'Abbaco.  Trattato di Tutta l'Arta dell'Abacho.  1339.  Op. cit. in 7.E.  B 2433, ff. 11r-11v discuss perfect numbers, mentioning only 6, 28 and 496.

Chuquet.  1484.  Triparty, part 1.  FHM 57-59 & 360 gives English and discussion.  Believes the pattern is regular, giving  Pn  for  n = 2, 3, 5, 7, 9, 11, 13  -- but the  5th  &  6th  of these are not perfect.

Pacioli.  Summa.  1494.  Ff. 6v-8v.  Discussion of perfect numbers.  Cites Euclid.  Says  3, 7, 31, 127  are prime and so  6, 28, 496, 8128  are perfect.  Says endings alternate.

Pacioli.  De Viribus.  c1500.  Ff. 44v - 47r.  XXVI effecto a trovare un nů pensato quando sia perfecto (26th effect to find a number thought of if it is perfect).  Gives the first five perfect numbers as  6, 28, 496, 8128, 38836.  The last is actually  4 · 7 · 19 · 73  and is so far wrong that I assumed that Peirani had miscopied it, but it is clear in the MS.  We do have  38836 = 76 M11,  so it seems Pacioli erroneously thought  M11 = 511  was prime, but the multiplication by  256  was corrupted into multiplication by  76,  probably by shifting the partial product by  2  into alignment with the partial product by  5.

Charles de Bovelles (Carolus Bovillus).  Liber de perfectis numeris.  This is a small treatise included in an untitled collection of his works, Paris, 1510 (reprinted Stuttgart/Bad Cannstatt, 1970), ff. 172r-180r.  ??NYS -- described in: Nobuo Miura; Charles de Bovelles and perfect numbers; Historia Scientarum 34 (1988) 1-10.  This is probably the first printed work on number theory, but it claims that  Pn  is prime when  n  is odd!!  Gives values of the 'perfect' number for  n = 1, 2, 3, 5, 7, 9, ..., 39.

Tonstall.  De Arte Supputandi.  1522.  Pp. 222-223.  Brief discussion of abundant, deficient and perfect numbers, only mentioning  6  and  28.

Cardan.  De Numerorum Proprietatibus.  ??, ??NYS.  = Opera Omnia, vol. IV, pp 2-4, sections 4 & 5.  (It is possible that this is the first publication of this item??)  Brief discussion, mentioning  6, 28, 496, 8128  and Euclid.

Cardan.  Practica Arithmetice.  1539.  Chap. 42, sections 2 & 3, ff. H.i.r - H.i.v (p. 52).  Mentions  220 & 284,  then similar to above.

Robert Recorde.  The Whetstone of Witte.  John Kyngstone, London, 1557.  Facsimile by Da Capo Press, NY  &  Theatrum Orbis Terrarum, Amsterdam, 1969.  Nombers perfecte, ff. A.iv.r - A.iv.v.  Discusses perfect numbers briefly and asserts  Pn  is perfect for   n  = 2, 3, 5, 7, 9, 11, 13, 15.

van Etten.  1624.  Prob. 70 (63), parts IIII & VII, pp. 66-67 (92-93).  Mentions perfect numbers and says they occur for  n = 2, 3, 5, 7, 9, 11, 13,  with  486  for  496.  He asserts the endings alternate between  6  and  28.  Henrion's 1630 Notte, pp. 22‑23, refers to Euclid, corrects  486  to  496,  and says someone has recently claimed  120  is a perfect number.  Deblaye, op. cit. in 1, copies  486  as  286!  The 1653 English ed copies  486,  gives  120816  for  130816,  extends the list with  n = 15, 17, 19  and asserts  n = 39  gives the  20th  perfect number.  Describes  220  and  284. 

M. Mersenne.  Letter to Descartes, 1631.  (??NYS -- cited in:  Ore, Number Theory and Its History, 95  and  Dickson I 33.)  Raises question of multiply perfect numbers and states(?)  120  is  3‑perfect.

Thomas Stanley.  Pythagoras.  The Ninth Part of The History of Philosophy, (1655-1662), collected ed., 1687, pp. 491-576.  Reprinted by The Philosophical Research Society, Los Angeles, 1970.  P. 552 discusses the Pythagorean attitude to perfect numbers.  "... nor without reason is the number  6  the foundation of generation, for the Greeks call it τελείov, we perfect; because its three parts,  1/6  and  1/3  and  1/2  (that is  1,  2,  and  3.) perfect it".

W. Leybourn.  Pleasure with Profit.  1694.  Observation 2, p. 3.  Essentially taken from the English ed. of van Etten, with the same numerical mistakes and a further mistake in copying the 20th case.  Mentions  220, 284.

Ozanam.  1694.  Prob. 5, quest. 17-19, 1696: 14-18;  1708: 13-15.  Prob. 8, quest. 17-19, 1725: 29-41.  Chap. 3, art. 11-12 & 15, 1778: 32-38;  1803: 35-40;  1814: 32-36;  1840: 19-21.  1696 says  2p‑1(2p-1)  is perfect if  2p - 1  is prime, but asserts this is true for  p = 11.  1696 also notes that  120  is  3-perfect and gives several amicable numbers.  1725 extends the remarks on amicable numbers.  1778 notes that  p = 11  fails, stating that Ozanam forgot that  2p - 1  had to be a prime, and gives the first  8  perfect numbers.  1778 also gives some amicable pairs. 

Manuel des Sorciers.  1825.  P. 86.  ??NX  Gives the "perfect numbers" corresponding to  p = 2, 3, 5, 7, 9, 11, 13,  giving  486  for  496  and saying the endings alternate between 6 and 8.  Probably copied from van Etten.  I've included this because it is surprisingly late to be so erroneous!

B. N. I. Paganini.  Atti della Reale Accademica delle Scienze di Torino 2 (1866‑1887) 362.  ??NYS.  Discovery of second smallest amicable  pair:  1184, 1210.

Pearson.  1907.  Part II: Amicable numbers, pp. 35‑36.  Asserts that  220, 284;  17296, 18416;  93 63584, 94 37056  are the only amicable pairs below  10 millions.

Alan Turing and/or colleagues was the first to use a computer to search for new Mersenne primes on the Manchester Baby in 1949, but it could not easily deal with numbers greater than  M353.

R. M. Robinson wrote a program to search for Mersenne primes using the Lucas-Lehmer test on the SWAC in late 1951/early 1952.  It was his first program.  On 30 Jan 1952. it was loaded and ran!  It discovered the 13th and 14th Mersenne primes:  M521  (at about 10:00 pm, taking about a minute) and  M607  (just before midnight).  M1279,  M2203  and  M2281  were found in the next months.  The program comprised  184  machine instructions on  24  feet of paper tape and would handle cases up through  2297.  It ran successfully on its first trial!  Lehmer was present when the program was tested on  M257,  which Lehmer spent some  700  hours in testing  c1932,  and the program confirmed this in a fraction of a second.  c1982,  Robinson ran his program on an early PC which only ran about twice as fast as the SWAC.

Alan L. Brown.  Multiperfect numbers -- cousins of the perfect numbers.  RMM 14 (Jan‑Feb 1964) 31‑39.  Lists all known  3‑,  4‑,  5‑perfects and the first  100  6‑perfects.

Elvin J. Lee  &  Joseph S. Madachy.  The history and discovery of amicable numbers -- Parts 1, 2, 3.  JRM 5 (1972) 77‑93, 153‑173, 231‑249.  Part 1 is the main history.  Parts 2 and 3 give all  1107  amicable pairs known at the time, with notes explaining the listings.

B. L. van der Waerden.  A History of Algebra.  Springer, Berlin, 1985.  Pp. 21‑23 describes the work of Tabit ibn Qurra (824?‑901) on amicable numbers and its development by Fermat, Descartes, Euler and Legendre.

Jan P. Hogendijk.  Thābit ibn Qurra and the pair of amicable numbers  17296, 18416.  HM 12 (1985) 269-273.  This pair is often named for Fermat, who first mentions it in Europe.  Thābit gives a general rule which would yield this pair as the second example, though he doesn't give the values.  Hogendijk analyses Thābit's work and concludes that he must have known these values.  In the same issue, a review by Hogendijk (pp. 295-296) mentions that the pair in question was known in 14C Persia and that the pair  9363584, 9437056,  usually ascribed to Descartes, was known c1600 in Persia.

Ettore Picutti.  Pour l'histoire des septs premiers nombres parfaits.  HM 16 (1989) 123-126.

 

          7.AC. CRYPTARITHMS, ALPHAMETICS AND SKELETON ARITHMETIC

            A skeleton problem shows all the working with most digits indicated by the same symbol, e.g.  *,  and only a few digits are left in place.

            A cryptarithm or alphametic usually shows just the data and the result with digits replaced by letters as in a substitution cipher.

            The opening section includes some miscellaneous numerical-alphabetical recreations which I haven't yet classified in subsections.

 

C. Dudley Langford.  Some missing figure problems and coded sums.  MG 24 (No. 261) (Oct 1940) 247‑253.  Lots of examples of various forms.

[J. S. Madachy?]  Alphametics.  RMM 6 (Dec 1961) 27,  7 (Feb 1962) 13  &  10 (Aug 1962) 11.  Historical comments.  Cites Berwick, Schuh, Dudeney, Minos, Hunter.  Says Strand Mag. (1921) is first division with letters instead of uniform  *.

"Fomalhaut".  Cryptophile cryptofile:  Cryptarithms.  World Game Review 8 (Jul 1988) 5‑12.  Survey of various forms of these puzzles and related books and magazines.

Graham Hawes.  Wordplay.  M500 116 (Nov 1989) 6‑7.  Using the numerological mapping  A = 1, B = 2, ..., Z = 26,  he finds two numbers, in British usage, whose numerological value is itself.  One is "two hundred and fifty one".  (The use of 'and' is British, but not American.)  He has found none in American usage.  Again in British usage, the sum  73 + 89 = 162   gives a correct sum for its numerological values:  166 + 116 = 282.

 

          7.AC.1.        CRYPTARITHMS:  SEND + MORE  =  MONEY,  ETC.

 

American Agriculturist (Dec 1864).  ??NYR -- copy sent by Shortz.   Multiplication problem where the letters for  1 ‑ 0  spell Palmerston.

Anon.  Prob. 82.  Hobbies 31  (No. 797) (21 Jan 1911) 395  &  (No. 800) (11 Feb 1911) 464.  Multiplication laid out:  PHSF * XV  =  HBAKF + OFPHF_  =  OHSBKF.  Solution:  7690 * 48  =  369120.

Loyd.  Cyclopedia.  1914. 

Alphabetical addition, pp. 233 & 370. 

 BOW + APPLE + CHOPS + HASHES + CHEESE + APPLES + EHW  =  PALEALE; B + LAY + TEN + DOZ  =  DNLL. 

Pp. 238 & 371.  = SLAHP: Masquerading digits, pp. 86 & 119.  JGDCH * IFABE  =  BIBDEB.

Smith.  Number Stories.  1919.  See 7.AC.2 for examples with full layout.

Dudeney.  Perplexities: Verbal arithmetic.  Strand Mag. (Jul 1924).  ??NYS.  SEND + MORE  =  MONEY;   EIGHT ‑ FIVE  =  FOUR;   TWO * TWO  =  THREE;   SEVEN/TWO  =  TWO  (with full division layout).

Dudeney.  Problem ?: The Arab's puzzle.  Strand Mag. (Early 1926?).  ??NX.  ABCD * EFGHI  =  ACGEFHIBD.

Loyd Jr.  SLAHP.  1928.  Kindergarten algebra, pp. 48 & 102.  AB * AB  =  CDDD.

MINOS [Simon Vatriquant].  Sphinx 1 (May 1931) 50.  Introduces word "cryptarithmie".  "A charming cryptarithm should  (1)  make sense in the given letters as well as the solved digits,  (2)  involve all the digits,  (3)  have a unique solution, and  (4)  be such that it can be broken by logic, without recourse to trial and error."  (Translation by C. W. Trigg in CM 4 (1978) 68.)

C. O. Oakley, proposer;  W. E. Buker, solver.  Problem E7.  AMM 39 (1932) 548 ??NYS  &  40 (1933) 176.  SEND + MORE  =  MONEY.  Editorial comment in solution cites L'Echiquier (June 1928) and Sphinx.

H. Phillips.  The Playtime Omnibus.  Op. cit. in 6.AF.  1933.  Section XVI, prob. 7: The money code, pp. 49‑50 & 234.  SEND + MORE  =  MONEY.

Rudin.  1936.  No. 84, pp. 28-29 & 92.  SEND + MORE  =  MONEY.

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 7, pp. 188 & 197-198.  SEND + MORE = MONEY.

M. Adams.  Puzzle Book.  1939.  Several straightforward examples and the following.  Prob. C.38: Economy, pp. 133 & 176.  SAVE + MORE  =  MONEY.  Four solutions given.

Alan Wayne.  The Cryptogram (c1945) (a US puzzle magazine), ??NYS.   Introduces 'doubly true additions', e.g.  SEVEN + SEVEN + SIX  =  TWENTY.  (See Trigg cited above at MINOS.)

Alan Wayne, proposer;  A. Chulick, solver;  editorial note by Howard Eves.  Problem E751 -- A cryptarithm.  AMM 54 (1947) 38  &  412‑414.  FORTY + TEN + TEN  =  SIXTY.  Editor cites Wayne in The Cryptogram for several others:  SEVEN + SEVEN + SIX  = TWENTY;   SEVEN + THREE + TWO  =  TWELVE;   TWENTY + FIFTY + NINE + ONE  =  EIGHTY.

Morley Adams.  Puzzle Parade.  Faber, London, 1948.

Chap. 3, no. 25: A wordy sum, pp. 46 & 52.  ONE + TWO + FIVE  =  EIGHT.  Answer starts "Here is one solution".

Chap. 9, no. 32: Simple as ABC, pp. 149 & 151.  ABC + ABC/5  =  CBA.  Answer is  495.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 27: The Cleveland butcher, pp. 34 & 186.  PORK/CHOP  =  C  >  2.  Unique answer is  9867/3289 = 3.  [Can there be answers with  C = 2??]

Anonymous.  The problems drive.  Eureka 17 (Oct 1954) 8-9 & 16-17.  No. 4.  ONE + TWO + FOUR  =  SEVEN.  One solution (not unique) given.

J. A. H. Hunter.  Fun with figures.  Globe & Mail (Toronto) (27 Oct 1955) 27  &  (28 Oct 1955) 29.  "It's just an easy alphametic today."   ABLE/RE  =  SIR  given in full diagram, determine the value of  MAIL  (=  8940).  Brooke (below), p. 45, reproduces Hunter's column.  Brooke says:  'Hunter received a letter from a reader referring to a "alphametical problem in which letters take the the place of figures".'

Anonymous.  Problems drive, 1957.  Eureka 20 (Oct 1957) 14-17 & 29-30.  No. 2.  THIS + IS = EASY  in base 7.

G. J. S. Ross  &  M. Westwood.  Problems drive, 1960.  Eureka 23 (Oct 1960) 23-25 & 26.  Prob. F.  TWO + TWO = FOUR.  Find the minimum base in which this holds, and a solution.

R. L. Hutchings  &  J. D. Blake.  Problems drive 1962.  Eureka 25 (Oct 1962) 20-21 & 34-35.  Prob. E.  THIS + ISSO = HARD  in base eight.  Four solutions.

Maxey Brooke.  150 Puzzles in Crypt‑Arithmetic.  Dover, (1963), 2nd ed. 1969.  On p. 4, he asserts that "Arithmetical Restorations" ... "were probably invented in India during the Middle Ages", but he gives no evidence on this point.

J. A. H. Hunter.  Note 3104:  CROSS + ROADS  =  DANGER.  MG 48 (No. 366) (Dec 1964) 433‑434.  This was posed by E. A. Maxwell in MG (Feb 1964) 114.  Hunter finds the unique solution.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971. 

Prob. 79: Very simple arithmetic again, pp. 39 & 85.  7 * DAYS = WEEK.  Says there are two solutions, to the best of his belief:  DAYS  =  1048,  1207.  This is correct, but there are also seven solutions with  D = 0.

Prob. 117: The right and the wrong of it, pp. 56 & 99.  WRONG + WRONG = RIGHT.  Gives two solutions.  See Singmaster, 1998, for discussion of all solutions.

Roy Childs.  Letter of 4 Aug 1999.  He has used a computer to study 'doubly-true additions' with sums up to  TWENTY.  In 14 cases, there are unique solutions, of which all but the first have sum  TWENTY. 

          THREE + THREE + TWO + TWO + ONE; 

          FIVE + THREE + THREE + THREE + THREE + THREE;  

          SEVEN + THREE + THREE + ONE + ONE + ONE + ONE + ONE + ONE + ONE;  

          SEVEN + THREE + THREE + TWO + TWO + ONE + ONE + ONE;  

          SEVEN + FIVE + TWO + TWO + ONE + ONE + ONE + ONE;  

          SEVEN + FIVE + TWO + TWO + TWO + ONE + ONE;  

          SEVEN + FIVE + FIVE + TWO + ONE;  

          SEVEN + SEVEN + TWO + ONE + ONE + ONE + ONE;  

          SEVEN + SEVEN + TWO + TWO + TWO;  

          SEVEN + SEVEN + SIX;  

          EIGHT + EIGHT + TWO + ONE + ONE;  

          ELEVEN + THREE + THREE + ONE + ONE + ONE;  

          ELEVEN + THREE + THREE + TWO + ONE; 

          ELEVEN + THREE + THREE + THREE

David Singmaster.  Two wrongs can make a right.  WRONG + WRONG = RIGHT.  I find  21  solutions, two of which have  O = 0  and two others have  I = 1.  Micromath 14:2 (Summer 1998) 47.  Brain jammer column, The Daily Telegraph, Weekend section, (20 Mar 1999) 27  &  (27 Mar 1999) 19. 

David Singmaster.  Three lefts make a right.  LEFT + LEFT + LEFT = RIGHT.  I find  27  solutions, four of which use only positive digits.  In no case does  I = 1.  Brain jammer column, The Daily Telegraph, Weekend section, (22 May 1999) 21  &  (29 May 1999) 19.

Victor Bryant.  On an episode of Puzzle Panel in 1999?, he asked: "How is  ONE + TWELVE = TWO + ELEVEN?"  Though it initially seems like an alphametic, it is actually an ingenious anagram. 

David Singmaster.  Letter to Victor Bryant, 27 Dec 2002.  I wondered if  ONE + TWELVE = TWO + ELEVEN  could be made into an alphametic.  This requires some repeated values as we have to have  TWE = ELE,  letter by letter, so  T = E,  W = L.  There are 133 solutions of the resulting alphametic, e.g.  047 + 797917 = 790 + 797174.  This seems to be the closest thing to a triply-true alphametic.  I thought Victor said that  ONE + TWELVE = TWO + ELEVEN  was unique, but there are six other such anagrams, such as  FOUR + SIXTEEN = SIX + FOURTEEN, though one might regard these as fairly trivial anagrammatically.  I've tried these examples to see if they give alphametics as above.  In all but one case, the lengths differ and this rapidly leads to a contradiction.  E.g., for the first case cited, we have to have  SIX = 999,  FOUR = 1000  and then the units digits lead to  X = R,  which is a contradiction.  (This is making the assumption that the numbers do not have leading zeroes.)  But for  FOUR + NINETEEN = NINE + FOURTEEN,  the units digits give us  R = E  and this forces  FOUR = NINE  and the problem reduces to  NINE + NINETEEN = NINE + NINETEEN  which is trivial, with  (10)7 = 10·9·8·7  solutions (this includes the cases with leading zeroes, but replacing the 10 by a 9 gives the number without leading zeroes).  So this isn't really satisfactory, but again it seems to be the best one can do.

                    We also have  21 + 32 = 22 + 31,  etc., as well as  20 + 31 = 21 + 30.  Consider the problem as being of the form  AC + BD  =  AD + BC,  where the first example above would be  21 + 32 = 22 + 31  or  A = TWENTY,  C = ONE,  B = THIRTY,  D = TWO.  Let │A│  be the number of letters in the English word for  A0,  │C│  be the number in  C,  etc., so  │B│ = 6,  │D│ = 3.  Let │A│  be the number of letters in the English word for  A,  etc.  We can assume  │D│ ³ │C│.  It is easily seen that any assignment of values to letters gives an alphametic solution when  │C│ = │D│.  But if  │D│ > │C│,  then we can get an alphametic if and only if  │A│ = │B│.  These alphametics will generally have some different letters having the same value.

                    There are also possibilities of the form  20 + 31 = 21 + 30, i.e.  │C│ = 0.  Similar analysis shows this gives an alphametic solution if and only if  │A│ = │B│.

                    More elaborately, we have  67 + 79 + 96 = 76 + 69 + 97  and  679 + 796 + 967 = 697 + 976 + 769.  We write this out as    SIXTYSEVEN + SEVENTYNINE +   NINETYSIX  = 

          SEVENTYSIX +   SIXTYNINE + NINETYSEVEN.  The 0-th, 1st, 2nd, 3rd and 4th columns from the right give no information, but the 5th column (appropriately!) gives us  Y + T + E = N + T + Y,  whence we must have  E = N.  The 6-th column gives  T + N + N = E + X + T,  but  E = N  forces  E = X.  Carrying on, we get  E = I = N = S = V = X  and both sides reduce to  EEETYEEEEE + EEEEETYEEEE + EEEETYEEE,  which has  (10)3  solutions.  Considering the hyphen,  -,  as a character, only shifts the argument a bit and one gets both sides reducing to  EEETY-EEEEE + EEEEETY-EEEE + EEEETY-EEE  with  (10)4  solutions.  In the second case, we get the same letter identifications and both sides of the problem reduce to  EEEHUEDREDEEEETYEEEEE + EEEEHUEDREDEEEEETYEEE + EEEEEHUEDREDEEETYEEEE,  with  (10)7  solutions.

 

          7.AC.2.        SKELETON ARITHMETIC:  SOLITARY SEVEN, ETC.

 

            Anonymous, Problems drive, 1957  is the only example I have seen of the inverse problem of starting with a known situation and finding a skeleton problem that gives it.

 

W. P. Workman.  The Tutorial Arithmetic, op. cit. in 7.H.1, 1902.  See also comments in Ackermann, under Berwick, below.  Chap. VI -- Examples XIX, probs. 30‑41, pp. 48‑49 & 503 (= 50‑51 & 529 in c1928 ed.).  Simple problems, e.g. prob. 30:  2982**  divided by  456  leaves remainder  1.  Probs. 31‑34 are skeleton multiplications;  35‑37 are skeleton divisions.

W. E. H. Berwick.  School World 8 (Jul & Aug 1906) 280 & 320.  ??NYS.  Division with seven sevens given:  7375428413 / 125473  =  58781.  Actually, there are  13  7s  in the layout, so not all the  7s  are shown.  Ackermann, below, says Berwick composed this, at age 18, after seeing some examples in Workman.

Pearson.  1907. 

Part II, no. 5: Fill in the gaps, pp. 114 & 191.  Division layout with some numbers given.

Part II, no. 11: Find the multiplier, pp. 115‑116 & 193.  Multiplication layout with some numbers given.

Anon.  Sol. 38.  Hobbies 30 (No. 756) (9 Apr 1910) 37 & 46.  (I don't have the problem proposal -- ??NYS.)  ...1 * 7.  =  6.... + .6.2._  =  ....5.  with solution  8061 * 78  =  628758.

???  J. Indian Math Club, 1910.  ??NYS -- cited by Archibald who says it has skeleton divisions with  4  digits given.

Smith.  Number Stories.  1919.  Pp. 111‑112 & 139‑140.  Two cryptarithmic multiplications and two cryptarithmic divisions, but with full layouts.

W. E. H. Berwick.  MG (Mar 1920) 43.  ??NYS -- cited by Archibald.  Four fours.

F. Schuh.  Een tweetal rekenkundige ardigheden.  Nieuw Tijdschrift voor Wiskunde 8 (1920‑1921) 64.  Skeleton division with no digits given, but the quotient has a repeating decimal and the divisor and dividend are relatively prime.  The problem is reproduced as Note 16, AMM 28 (1921) 278, signed ARC [= R. C. Archibald], with solution by D. R. Curtiss and comment by A. A. Bennett in AMM 29 (1922) 210‑213.  Bennett shows that relatively primality is not essential.  The problem and solution are given as Section 258: Repeating division puzzle, pp. 320‑322, in Schuh's The Master Book of Mathematical Recreations, Dover, 1968.  (Originally Wonderlijke Problemen; Leerzaam Tijdverdrijf Door Puzzle en Spel, Thieme, Zutphen, 1943.)  7752341 / 667334  =  11.6168830001168830001....

R. C. Archibald.  AMM 28 (1921) 37 -- sketches the history:  J. Indian Math Club;  Berwick's 'seven sevens' and 'four fours' cited to MG (Mar 1920) 43.  See The President, 1941.

W. E. H. Berwick.  Problem 555: The four fours (in Dudeney's column).  Strand Mag. (1921 or 1922?)  ??NYS.  Four fours given in the skeleton of  1200474 / 846  =  1419,  but there are other fours present and there are three other solutions.

Egbert F. Odling.  Problem 627: Solitary seven (in Dudeney's column).  Strand Mag. (Nov  &  Dec 1922), ??NYS.  Skeleton of  12128316 / 124  =  97809,  with only the  7  given.  Unique solution and the  7  only occurs once.  Repeated as Problem 1105: Skeleton sum; Strand Mag. (Jul 1932) 104 & (Aug 1932) 216.

Anon.  Note 671.  MG 11 (1922‑23) 338.  Gives Odling's problem and solution and some comments.

Ackermann.  1925.  Pp. 109‑115.  Discusses Berwick's problems, referring to MG of Mar 1920, Dec 1921 and Jan 1922 (??NYS) for four fours, five fives, three threes and six sixes.  Gives one of Smith's problems, Berwick's seven sevens and five fives, Schuh's problem (attributed to Ball) and Odling's problem.

Dudeney.  MP.  1926.  Prob. 70: The solitary seven, pp. 26‑27 & 118.  (= 536, prob. 144, pp. 43 & 259.)  Cites EFO [= Odling].  "It is the first example I have seen ... in which only one figure is given."

A. A. Bennett, proposer;  H. Langman, solver.  Problem 3212.  AMM 33 (1926) 429  &  34 (1927) 538-540.  Skeleton division   xxxxxxxxxxxxxccfx / xxxxabxxxx  =  xcxxxxx,  with numerous further positions given by definite letters.  Solution notes that some of the information is not needed, e.g.  f  can be replaced by  x.  Answer is   70900515872010075 / 68253968253  =  1038775.

Collins.  Fun with Figures.  1928.  Through a knot-hole, p. 189.  *7*9* / 215  =  1**   with some further figures given.

H. E. Slaught, proposer;  C. A. Rupp, solver.  Problem E1.  AMM 39 (1932) 489  &  40 (1933) 111-112.  Odling's 'Solitary Seven' problem.  Editor's note says it has appeared in Le Sphinx.

Perelman.  FFF.  1934.  Mysterious division & Another division.  1957: probs. 104 & 105, pp. 138 & 145;  1979: probs. 107 & 108, pp. 167 & 176.  = MCBF, probs. 107 & 108, pp. 168 & 178-179.  Berwick's 'four fours' and 'seven sevens', with all four solutions of the latter, rather poorly attributed to "the American publications School World (1906) and Mathematical Magazine (1920)".

Perelman.  FMP.  c1935?  Mysterious division, pp. 256 & 268‑269.  Skeleton of  11268996 / 124  =  90879,  with only the  7  given.  The quotient is unique, but there are  11  possible divisors:  114, 115, ..., 124,  of which  115, 116 and 120  give no other  7s  in the layout.

A. G. Sillito.  Note 1424:  Division without figures.  MG 23 (No. 257) (Dec 1939) 467‑468.  Two divisions like Schuh's:  16 / 41  and  81 / 91.

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940. 

No. 67: A skeleton square root, pp. 40 & 125.  A dotted diagram with two letters marked in, one  8  times, the other  4  times.

No. 87: A skeleton division, pp. 49 & 132.  = Odling, 1922.

The President.  Some missing-figure divisions.  Eureka 6 (May 1941) 21-24.  Studies skeleton divisions.  Cites 'the solitary seven' and 'the numberless decimal division' 'in one of Dudeney's Puzzle Books' -- presumably MP, 1926.  Then gives six problems, four given by Berwick in MG -- ??NYS.  I is Four-threes, with complete solution.  II is Three-threes, proposed by Berwick, the same as Four-threes with one three not given, which is a harder version for which he does not have a satisfactory solution, though the answer is the same as for case I.  III is Berwick's Four-fours, for which Berwick gives the complete solution so here the four answers are only stated.  IV is Berwick's Five-fives, again completely solved by Berwick with just the answer here.  V is Four-sixes, apparently novel, with complete solution.  VI is Seven-sevens, proposed by Berwick, with complete solution here.

Sullivan.  Unusual.  1947.  Prob. 37: Lost and found.  Skeleton division of  1089708 / 12  =  90809   with only the  8  of the quotient given.

P. L. Chessin, proposer;  ???, solver.  Problem E1111.  AMM 61 (Apr 1954) 712  &  ???.  ??NYS -- given in the Otto Dunkel Memorial Problem Book, ed. by Howard Eves and E. P. Starke, AMM 64:7 part II (Aug-Sep 1957) 6, where it is described as the most popular problem ever published in the AMM, with 70 solvers.  Also given by Gardner, SA (May 1959) = 2nd Book, chap. 14, prob. 5: The lonesome  8,  pp. 154-155 & 160‑161.  Skeleton division of   10020316 / 124  =  80809   with only the middle  8  of the quotient given.  Answer is unique.

William R. Ransom.  Op. cit. in 6.M.  1955.  Only one digit known, p. 134.  Skeleton of   11260316 / 124  =  90809  with only the  8  given.  Answer is unique and  8  only appears once.

Anonymous.  Problems drive, 1957.  Eureka 20 (Oct 1957) 14-17 & 29-30.  No. 9.  Gives complete layout of  1345 x 32  and asks to find five digits which would determine the rest.

G. A. Guillotte.  Note 2865:  Missing digits.  MG 43 (No. 345) (Oct 1959) 200.  Long division with  17  0s  specified.

B. D. Josephson  &  J. M. Boardman.  Problems drive 1961.  Eureka 24 (Oct 1961) 20-22 & 24.  Prob. E.  Full skeleton of  5980482 / 498  with a single  8  given in the last line.

Anonymous postcard to The Science Correspondent, "The Glasgow Herald", 8 May 1963, found in Prof. Lenihan's copy of Gardner's More Mathematical Puzzles and Diversions and given in Jay Books sale catalogue 129 (Feb? 1992) and 130 (Jun 1992).  Full skeleton of  1062 / 16 = 66.375   with no digits specified.  The solution is unique.

L. S. Harris  &  J. M. Northover.  Problems drive 1963.  Eureka 26 (Oct 1963) 10-12 & 32.  Prob. J.  Complete skeleton of  32943 / 139  with all six 3s given.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 69, pp. 70 & 104.  Same as Sullivan, 1947.

Birtwistle.  Math. Puzzles & Perplexities.  1971.

Seven and eight, p. 149.   **7** / **  =  8***,   full skeleton shown, but no other values given.  Unique solution is   99708 / 12  =  8309.

Long division, pp. 150, 185 & 197.  Skeleton of  123195 / 215  =  573  with four values given.

Letter division, pp. 151, 186 & 198.  Complete layout of a division with letters denoting the digits, as in a cryptarithm.  Two possible answers:  6420 / 20  =  321  or  6930 / 30  =  231.

On all fours, pp. 151 & 198.  Skeleton of  31666 / 142  =  223  with all fours given.

 

          7.AC.3.        PAN‑DIGITAL SUMS

 

            These are generally of the form   ABC + DEF  =  GHI.   The digits may be positive or  9  of the  10  digits.  One can also have a  10‑digital form, e.g.   ABC + DEF  =  GHIJ.

 

Dudeney.  Problem 64: The lockers puzzle.  Tit‑Bits 33 (18 Dec 1897  &  5 Feb 1898) 220  &  355.  = AM, prob. 79, pp. 14 & 156.  Find   ABC + DEF  =  GHI   using  9  of the  10  digits which have the least result, the greatest result and a result whose digits are distinct from the first two.  Answers:   107 + 249  =  356;   235 + 746  or  324 + 657   =  981;  134 + 586  =  720   or   134 + 568  =  702   or   138 + 269  =  407.

Dudeney.  AM.  1917.  Prob. 77: Digits and squares, pp. 14 & 155.  For  ABC + DEF  =  GHI,  he wants  DEF  =  2 * ABC,  so   GHI  =  3 * ABC,  using the  9  positive digits.  Says there are four solutions, the tops being   192,  219,  273,  327.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 89, pp. 38 & 142: Doubles and trebles.  Same as Dudeney, AM, prob. 77.

Morley Adams.  Puzzle Parade.  Op. cit. in 7.AC.1.  1948.  No. 12: Figure square, pp. 146 & 150.  As in Dudeney, AM, prob. 77.  Says there are four solutions, but wants the one with minimal  E.  Solution:   219 + 438  =  657.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 15: Skeleton addition, pp. 10 & 41.  Complete   ABC + DE7  =  GH8   using all nine positive digits.  Gets four forms by reversing  A  and  D  and/or  B  and  E.

Jonathan Always.  Puzzles to Puzzle You.  Op. cit. in 5.K.2.  1965.  No. 135: All the numbers again, pp. 42 & 90.  As in Dudeney, AM, prob. 77.  Gives one solution:   192 + 384  =  576.

Ripley's Puzzles and Games.  1966.  Arrange the nine positive digits in two columns with the same sum.  He forms an  X  shape with  5, 9, 1, 6, 2  down one line and  4, 8, 1, 7, 3  down the other.  Both 'columns' add to  23.  [Another solution is to invert the  6  or  the  9.]

Wickelgren.  How to Solve Problems.  Op. cit. in 5.O.  1974.  Integer‑path‑addition

          problem, pp. 130‑132.  Wants the  9  positive digits in a pan-digital sum, so the                     1 2 9

          resulting  3 x 3  array has each digit  i  horizontally or vertically adjacent                           +  4 3 8

          to the digit  i+1.  Says there is one solution, shown at the right.                                        =  5 6 7

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 41: The

          number and the square, pp. 31 and 107.  As in Dudeney, AM, prob. 77.  Gives all solutions.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 87: Three by three -- part two, pp. 61 & 120‑121.  Which distributions of the nine positive digits as  ABC, DEF, GHI  have the lowest sum and product?  147 + 258 + 369  =  774   gives the lowest sum and the digits in each position can be permuted -- e.g.  348 + 257 + 169  gives the same sum.  The lowest product is uniquely given by   147 x 258 x 369  =  13994694.

Johannes Lehmann.  Kurzweil durch Mathe.  Urania Verlag, Leipzig, 1980.  No. 14, pp. 39 & 139-140.   A + B + C  =  D + E + F  =  G + H + I   has just two solutions using the positive digits.  [Interestingly, one gets no more solutions using the ten digits.]

David Singmaster.  Determination of all pan-digital sums with two summands.  JRM 27:3 (1995) 183-190.  AB + CDE = FGHI  has no solutions with the nine positive digits and ten basic solutions using nine of the ten digits.  AB + CDEF = GHIJ  has nine basic solutions.  (Each basic solution gives four or eight equivalent solutions.)  ABC + DEF = 1GHI  has 12 basic solutions, which can be paired.  ABC + DEF = GHI  has 216 basic solutions, but 80 have  A = 0.  42 cases use the positive nine digits.  The 216 can be grouped into 72 triples and a canonical example is given for each triple.  The cases where the three terms form a simple proportion are listed.

 

          7.AC.3.a      INSERTION OF SIGNS TO MAKE 100, ETC.

 

            I include here problems like inserting  +  and  -  (and perhaps  ´  and  ¸)  signs into  12...9  to yield  100,  which I call  'insertion to make  100'.  This has two quite different sets of answers depending on whether the operations are carried out sequentially (as on an old calculator) or in algebraic order of precedence (as on a computer or modern calculator).  See also 7.AC.6 for similar problems.

 

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-16, pp. 253 & 395.  Make  100 from  1, 2, ..., 9, 0.  Answer:  9x8 + 7 + 6 + 5 + 4 + 3 + 2 + 1.  (0 isn't used, but could be added.)

Mittenzwey.  1880.  Prob. 139, pp. 30 & 79;  1895?: 159, pp. 33 & 82;  1917: 158, pp. 31 & 79.  Make  100  from  1, 2, ..., 9  using only multiplication and addition.  Same answer as Leske. 

Anon.  &  Dudeney.  A chat with the Puzzle King.  The Captain 2 (Dec? 1899) 314-320;  2:6 (Mar 1900) 598-599  &  3:1 (Apr 1900) 89.  Insert as few signs as possible in  12...9  to make  100.  Usual answer is  1 + 2 + 3 + 4 + 5 + 6 + 7 + (8x9)  (cf Leske),  but he gives  123 ‑ 45 ‑ 67 + 89.  Cf. AM, 1917.

Dudeney.  AM.  1917.  Prob. 94: The digital century, pp. 16-17 & 159-160.  Insert signs into  12...9  to make  100,  using:  (1) as few signs as possible;  (2) as few strokes as possible, with  -  counting as  1  stroke;  +,  ()  &  ´  counting as  2;   ¸  counting as  3.  He finds his 1899 result is best under both criteria.  Cf. Anon & Dudeney, 1899.

Hummerston.  Fun, Mirth & Mystery.  1924.  Century making, p. 66.  1 + 2 + 3 + 4 + 5 + 6 + 7 + (8x9)  (cf Leske).   12 + 3 - 4 + 5 + 67 + 8 + 9.   123 + 45 ‑ 67 + 8 - 9.   9x8 + 7 + 6 + 5 + 4 + 3 + 2 + 1  (cf Leske; Hummerston notes this is the reverse of his first example).   98 - 76 + 54 + 3 + 21.   Also the trick version:  15 + 36 + 47 = 98 + 2 = 100  (cf 7.A.6).

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 262, pp. 98 & 169: A number puzzle.  Insert 'mathematical signs' into  4 3 2 1  to make 100.  Answer:  Ö4 * [(3 + 2) / .1] .

Perelman.  1934.  See in 7.AC.6 for pandigital sum yielding  1  and  100.

McKay.  At Home Tonight.  1940.  Prob. 19: Centuries, pp. 66 & 80.  Insertion to make  100.  -1x2 - 3 - 4 - 5 + 6x7 + 8x9.   -1x2 -3 - 4 + 5x6 +7 + 8x9.   1x2x3x4 + 5 + 6 + 7x8 + 9.

Anonymous.  The problems drive.  Eureka 12 (Oct 1949) 7-8 & 15.  No. 6.  Insert symbols into  1 2 ... 9  to make  1000;  1001; 100.  Answers:  1234 - (5+6+7+8)x9;  (12 x 34) ‑ (5 x 6) + (7 x 89);  123 - 45 - 67 + 89.  "These solutions are not unique."

Anonymous.  Problems drive, 1958.  Eureka 21 (Oct 1958) 14-16 & 30.  No. 10.  Use  1, 2, 3, 4, 5,  in order to form  100;  3 1/7;  32769.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 48: A tricky problem, pp. 20 & 48.  "Can you replace the asterisks by the digits of the number  216345879  in this order so the resulting total is  100?   * + * + * + * + * + *"   Answer:  2/1 + 6/3 + 4 + 5 + 8 + 79.  His fraction bars are horizontal, but this problem seems a bit unreasonable to me.

Richard E. Bellman.  On some mathematical recreations.  AMM 69:7 (Aug/Sep 1962) 640‑643.  Develops a general theory for the number of ways  n  can be obtained by inserting  +  or  x  into   a1a2...aN   and for determining the minimum number of  +  signs occurring.  He computes the example of inserting into  12...9  to make  100  by use of recursion, finding that   1x2x3x4 + 5 + 6 + 7x8 + 9  =  100   has the minimal number of  +  signs, cf McKay.

Gardner.  SA (Oct 1962) c= Unexpected, chap. 15.  Insertion to make  100.  Cites Dudeney.  Asks for minimum number of insertions into  98...1  to  make  100.  Answer with four signs.

Gardner.  SA (Jan 1965) c= Magic Numbers, chap. 6.  Considers inserting  +  and  -  signs in  12...9  or  98..1  to yield  100.  Says he posed this in SA (Oct 1962) c= Unexpected, chap. 15 and many solutions for both the ascending and descending series were printed in Letters in SA (Jan 1963).  He gives a table of all the answers for both cases:  11  solutions for the ascending series and  15  solutions for the descending series.  He extends the problem slightly by allowing a  -  in front of the first term and finds  1  and  3  new solutions.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 10: Very simple arithmetic, pp. 14 & 62.  Insert signs into  1 2 3 4 5 6  to form an equation.  He gives 12 ¸ 3 ¸ 4 + 5  =  6.  I find  1 + 2 x 3 + 4  =  5 + 6.  which seems more satisfactory.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 65: Key to the problem, pp. 47-48 & 104.  Using a calculator, insert operations in  012...9  and  98...10  to produce  100.  Gives one example of each:  0 + 1/2 + 3x4x5 + 6 + 7 + 8 + 9;   9x8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 + 0 (cf Leske).

Steven Kahane.  Sign in, please!  JRM 23:1 (1991) 19-25.  Considers inserting  +  and  -  signs in  12...n  and in  n...21  to produce various results, e.g.  0,  n+1,  n,  among others.

Ken Russell & Philip Carter.  Intelligent Puzzles.  Foulsham, Slough, 1992. 

Prob. 102, p. 79 & Answer 30, p. 140: One hundred puzzles.  Three problems to insert the nine positive digits into formulae to make 100.  E.g.  + + + - - -  =  100  is solved as:  32 + 91 + 7 + 8 - 6 - 5 - 4.

Prob. 140, p. 106 & Answer 104, p. 181: Plus and Minus.  Find all ways to insert  +  and  -  signs into  1 2 ... 9  to yield 100.  Finds 12 ways, one with a leading minus, as given by Gardner, 1962.

 

          7.AC.4.        PAN‑DIGITAL PRODUCTS

 

Loyd.  Problem 36: Juggling with figures.  Tit‑Bits 32 (5 Jun  &  3 Jul 1897) 173  &  258.  10‑digital product with smallest result is   3907 * 4  =  15628.   Solution also gives the largest case:   9403 * 7  =  65821.   [I have verified that these are correct.]

Dudeney.  The miller's puzzle.  The Canterbury puzzles.  London Mag. 8 (No. 46) (May 1902) 367-371  &  8 (No. 47) (Jun 1902) 480-482.  = CP, prob. 3, pp. 26 & 164-165.  Find a solution of   A*BC  =  D*EF  =  GHI  using the  9  positive digits and which is closest to a given pattern.  Answer says there are four solutions, the closest is   2 * 78  =  4 * 39  =  156.

Dudeney.  CP.  1907. 

Prob. 93: The number blocks, pp. 139-140 & 238.  Find a solution of   AB*CDE  =  FG*HIJ.   Answer:  64 * 915  =  80 * 732.

Prob. 101: The three motor-cars, pp. 147-149 & 242-243.  Wants a solution of   AB*CDE  =  FGHIJ   such that  AB  divides  CDE.  Answer is   27 * 594  =  16038   and he says it is hard to show this is unique.  He says there are many solutions for   A*BCDE  =  FGHIJ,   e.g.   3 * 5694  =  17082.  = Wood, 1927, prob. 62.

Anon.  Prob. 50.  Hobbies 30  (No. 765) (11 Jun 1910) 228  &  (No. 768) (2 Jul 1910) 289.  Same pattern as in Dudeney, 1902, with different set-up.

M. Thié.  ??  Nouv. Ann. Math. (4) 11 (1911) 46.  ??NYS -- cited by Dickson I 463, item 62a.  Found examples with  9  positive digits like   12 * 483  =  5796.

T. C. Lewis.  ??  L'Intermédiaire des Math. 19 (1912) 26-27  &  187.  ??NYS -- cited by Dickson I 463, item 66.  Examples with the  10  digits, like   7 * 9403  =  65821   and   3 * 1458  =  6 * 0729.

Dudeney.  AM.  1917. 

Prob. 80: The three groups, pp. 14 & 156.  Cites Thié and his example as being an extension of CP, prob. 101.  Asks for solutions to Thié's form and to   A*BCDE  =  FGHI,   with the  9  positive digits, e.g.   4 * 1738  =  6952.  Answer gives  7  solutions in the first case and  2  in the second case.

Prob. 81: The nine counters, pp. 14 & 156.  Find solution of   AB*CDE  =  FG*HI,   with the  9  positive digits, such that the product is maximal.  Answer:  32 * 174  =  58 * 96  =  5568. 

Prob. 82: The ten counters, pp. 15 & 156.  Divide the  10  digits into two equal products giving maximum and minimum products.  Answers:   2 * 3485  =  1 * 6970  =  6970   &   64 * 915  =  80 * 732  =  58560.  = Wood, 1927, probs. 59 & 60.

Prob. 85: The cab numbers, pp. 15 & 157.  Find two numbers, using all  9  positive digits, whose product contains all  9  positive digits and is maximal.  He believes the maximum is   96 * 8745231  =  839542176.

Prob. 86: Queer multiplication, pp. 15-16 & 157.  Examples of  A*BCDEFGHI  =  abcdefghi,  where both sides use the  9  positive digits:   3 * 51249876   and   9 * 16583742.   Asks for a solution with   A = 6.  Answer:  6 * 32547891.

Peano.  Giochi.  1924.  Prob. 34, p. 9.  Notes   2 * 78  =  39 * 4  =  156.  (Cf Dudeney, 1902.)

Hummerston.  Fun, Mirth & Mystery.  1924.  Grand-dad's age, Puzzle no. 68, pp. 155 & 182.  A product of two 2-digit numbers contains the same digits and is the birth date of the grandfather, namely  21 * 87  =  1827.  [There are three other examples, but none has a product in the recent past:  15*93 = 1395;  27*81 = 2187;  35*41 = 1435.  After making the first number less than or equal to the second, eliminating leading and trailing zeroes in the factors, I find the following numbers of solutions for factors of  M, N  digits.   1, 1 : 0;   1, 2 : 3;   1, 3 : 7;   1, 4 : 36;   2, 2 : 4;   2, 3 : 41;   2, 4 : 170;   3, 3 : 119;   3, 4 : 972.  Note that  0 * 0 = 00,  etc. has been eliminated.  The solutions in the  1, 3  case are:  3*51 = 153;  6*21 = 126;  8*86 = 688.]

Wood.  Oddities.  1927.

Prob. 14: A problem in multiplication, p. 16.  A * BC  =  D * EF  =  GHI.  Gives four solutions:  2 * 78  =  4 * 39  =  156;   3 * 58  =  6 * 29  =  174  and their reversals (i.e.  4 * 39  =  2 * 78  =  156;   6 * 29  =  3 * 58  =  174) and implies there are no more.

Probs. 59 & 60: Number blocks  &  More number blocks, p. 47.  Same as Dudeney's AM prob. 82.

Prob. 62, pp. 47-48.  Same as Dudeney's CP prob. 101.

W. F. Cheney Jr, proposer;  Victor Thébault, solver.  Problem E13.  AMM 39 (1932) 606  &  41 (1934) 265‑266.  Two factor products using all the digits just once.  Gives all solutions without  0:  2 of form   A*BCED  =  FGHI  (confirming results of Buker in AMM 40 (1933) 559 ??NYS);  7  of form   AB*CDE  =  FGHI.  Gives some solutions with  0:  4  of form   A*BCDE  =  FGHIJ;  3  of form   AB*CDE  =  FGHIJ.

Perelman.  FFF.  1934.  Tricky multiplication.  1957: prob. 45, pp. 56 & 61;  1979: prob. 48, pp. 71 & 77.  = MCBF: prob. 48, pp. 69 & 74.  Gives all  9  solutions without  0  to Dudeney's AM prob. 80.

Victor Thébault, proposer;  G. H. Biucliu & L. Tits, solvers.  Mathesis 44 (1935) 205‑207.  ??NYS -- described in CM 9 (1983) 89.  All  94  solutions of   n*ABCDE  =  FGHIJ.

Jerome S. Meyer.  Arithmetricks.  Scholastic Book Services, NY, 1965.  No repeating digits, pp. 16-17.  Says   A*BCDEFGHI  =  abcdefghi   with  A = 9  has four solutions such that we also have   2*abcdefghi  =  abcdefghij.   81274365;  72645831;  58132764;  76125483 (where the last  3  is misprinted as  4).  I find seven other examples:  58463721;  57624831;  85372461;  72534861;  83257614;  82164735;  71465328.  However this doesn't include Dudeney's example in AM 86, because it doesn't satisfy the extra condition.

Charles L. Baker.  (Presumably in RMM.  ??NYS)  Reported in Madachy's Mathematical Recreations, op. cit. in 5.O, (1966), 1979, pp. 183‑185.  Confirms Thébault's (1934) and Perelman's results without  0  and presents all two‑factor products with  0:  13  of form   A*BCED  =  FGHIJ;  9  of form   AB*CDE  =  FGHIJ.

Charles W. Trigg, proposer;  Edward Moylan, solver;  David Daykin, commenter.  Problem 691 -- A product of integers.  MM 41:3 (1968) 158;  42:1 (Jan 1969) 44-45  &  42:2 (Mar 1969) 102-103.  Solve   a = 8b,   where  a  and  b  together use the  9  positive digits once each.  Must have the form   ABCDE  =  8*FGHI   and there are  46  solutions, all listed.  D. Sumner assumed that both  a  and  b  contained all nine positive digits and found a unique solution with  b = 123456789.  Daykin gives the number of solutions of the first problem in base  β,  with multiplier  m,  for  2 £ m < β £ 15,  and also considering the use of  0  as a digit and the use of just odd or just even digits.  The tables show surprising irregularity.  [Is this really surprising??]

Stewart Metchette.  A note on digital products.  JRM 10 (1977‑78) 270‑271.  Extends Thébault and Baker to three factor products and gives all of the following forms:  12  of form   A*BC*DE  =  FGHI;  10  of form   A*BC*DE  =  FGHIJ;  2  of form   A*BC*DEF  =  GHIJ.

Birtwistle.  Calculator Puzzle Book.  1978. 

Prob. 25, pp. 20 & 79.  Same as Dudeney's AM 82.

Prob. 87: Three by three -- part two, pp. 61 & 120-121.  See 7.AC.4 for the distribution of the nine positive digits as   ABC, DEF, GHI   with the lowest product, which is uniquely given by   147 x 258 x 369  =  13994694.

David Singmaster.  Work of 30 Jul 1998 in response to a letter of J. I. Collings.  There are 99 solutions of  AB*CDE = FG*HIJ,  with  A < F.  Of these,  35  have a leading zero.  Trailing zeroes lead to  13  pairs of related solutions, e.g.  23 * 760  =  76 * 230  =  95 * 184.  The largest value of the common product is  58560  =  64 * 915  =  80 * 732,  as given by Dudeney, AM, prob. 82.  The smallest common product is  3588  =  04 * 897  =  23 * 156,  while the smallest without a leading zero is  8262  =  18 * 459  =  27 * 306.  There are two cases with the same common product and further one of the two products is the same:  18 * 465  =  30 * 279  =  45 * 186.  Collings notes that there is only one solution with  E  or  J  being three and no leading zeroes

 

          7.AC.5.        PAN‑DIGITAL FRACTIONS

 

Mittenzwey.  1880.  Prob. 141  pp. 30 & 79;  1895?: 161, pp. 33 & 82;  1917: 161, pp. 31 & 79.  Use  1, 2, ..., 9  to form three fractions which add to one.  Solutions:  9/12 + 5/34 + 7/68  (cf Yoshigahara, below);  21/84 + 9/63 + 5/7;  21/48 + 7/36 + 5/9;  19/76 + 4/32 + 5/8,  etc.  [Are there really others?]

Pearson.  1907.  Part II: Juggling with the digits, pp. 40‑41.  Examples of   ABCD/EFGHI =  1/n   for  n = 3, 4, ..., 9.

Dudeney.  AM.  1917. 

Prob. 88: Digital division, pp. 16 & 158.  Gives  13458/6729 = 2.  Find solutions of   ABCDE/FGHI  =  n   for  n = 3, 4, ..., 9.  Also find the smallest solutions in each case -- e.g.  14658/7329  =  2  is a larger solution than the first example.

Prob. 90: The century puzzle, pp. 16 & 158-159.  Write a mixed number, using the  9  positive digits, equal to  100,  e.g.  91  5742/638.  Says Lucas found  7  ways, but he has shown that there are just  11  ways.  One of these has a single digit integer part -- find it.  Answer gives all  11  solutions. 

Prob. 91: More mixed fractions, pp. 16 & 159.  Says he has tried the same question with  100  replaced by other values and gives  12  values to try.  However, two of these are impossible.  He has found solutions for all values from  1  to  100,  except that  1, 2, 3, 4, 15, 18  are impossible, though  15  and  18  can easily be expressed if the integer part is permitted to be zero or if compound fractions are permitted, e.g.  3 (8952/746)/1.

Prob. 92: Digital square numbers, pp. 16 & 159.  Find largest and smallest squares using all  9  positive digits.  Answers:  923,187,456  =  303842;  139,854,276  =  118262.

Peano.  Giochi.  1924.  Prob. 35, p. 10.  Gives  6  solutions of   9  =  ABCDE/FGHIJ,   three of which have  F = 0.

Haldeman-Julius.  1937.  No. 112: Half problem, pp. 18 & 26.  Use the nine positive digits to make  1/2.  Answer is  7293/14586,  which is similar to Dudeney's prob. 88.

M. Adams.  Puzzle Book.  1939.  Prob. B.83: Figure juggling, part 2, pp. 78 & 107.  Asks for an example of using the nine positive digits to make  1/2  and remarks that solutions exist for  1/n  with  n = 3, ..., 9.  Gives same solution as Haldeman-Julius.

George S. Terry.  The Dozen System.  Longmans, Green & Co., NY, 1941.  ??NYS -- quoted in Underwood Dudley; Mathematical Cranks; MAA, 1992, p. 25.  Express unity as a sum of two fractions which contain all the digits once only, duodecimally.  E.g.   136/270 + 48χ/95ε  =  1   (χ = 10, ε = 11).  Says about five dozen.  Terry (or Dudley) says the decimal answer is about one dozen.

Ripley's Believe It or Not, 24th series.  Pocket Books, NY, 1975.  P. 76.  3/6  =  7/14  =  29/58  uses all nine positive digits.  [Are there other examples or other forms??]

Michael Holt.  Math Puzzles and Games.  Walker Publishing Co., NY, (1977), PB ed., 1983.  9  in ten digits, pp. 26 & 98.   ABCDE/FGHIJ  =  9   has six solutions, all given.

James W. Carroll.  Letter:  Computerizing Sam Loyd.  Games 7:5 (May 1983) 6.  ABCD/EFGHI  =  1/n   for  n = 2, 3, ..., 9  has  12, 2, 4, 12, 3, 7, 46, 3  solutions.

Nob Yoshigahara.  Puzzle problem used on his TV(?) program in Japan and communicated to me at 13th International Puzzle Party, August, 1993.  Use the nine positive digits to make   A/BC + D/EF + G/HI  =  1.  There is a unique solution:   5/34 + 7/68 + 9/12.

 

          7.AC.6.        OTHER PAN‑DIGITAL AND SIMILAR PROBLEMS

 

          See also 7.I and 7.I.1 for related problems.

 

The Family Friend (1856) 149 & 180.  Enigmas, Charades, &c.  87  Mathematical Puzzle.  "Take all the figures, (i.e., 1 2 3 4 5 6 7 8 9 0,) and place them in such a mode, that, when they are added up, they may be equal to 100."  Signed  S. W. S.  Answer is   76 + 3 10/5 + 8 + 9 4/2.

Magician's Own Book.  1857.  The united digits, p. 246.  "Arrange the figures  1  to  9  in such order that, by adding them together, they amount to  100."   15 + 36 + 47  =  98 + 2  =  100.   = Book of 500 Puzzles, 1859, p. 60.  = Boy's Own Conjuring Book, 1860, p. 216.  c= Parlour Games for Everybody.  John Leng, Dundee & London, nd [1903 -- BLC], p. 41.

Leske.  Illustriertes Spielbuch für Mädchen.  1864? 

Prob. 564-16, pp. 253 & 395.  Combine  1, 2, ..., 9  to make  100.  Answer:  1 + 3/6 + 27/54 + 98.

Prob. 564-19, pp. 253 & 395.  Add the nine digits to make  100.  Answer: 75 9/18  +  24 3/6.

Boy's Own Book.  The united digits.  1868: 429.  "The figures  1  to  9  may be placed in such order that the whole added together make exactly  100.  Thus --  15 + 36 + 47  =  98 + 2  =  100."

Hanky Panky.  1872.  The century of cards, p. 294.   15 + 36 + 47  =  98 + 2  =  100   given with cards.

Mittenzwey.  1880. 

Prob. 138, pp. 29-30 & 79;  1895?: 158, pp. 33 & 82;  1917: 158, pp. 31 & 79.  Use  1, 2, ..., 9  to make  100.  He gives:  95 + 3 + 1 + 6/7 + 4/28;   75 9/18  +  24 3/6  (cf Leske).  He says there are many such and many variations, e.g. allowing one repetition and making  1000.  [There are ways to make  1000  without any repetitions, e.g.  987 + 6 + 5 + 4 + 2 - 3 - 1.]

Prob. 140  pp. 30 & 79;  1895?: 160, pp. 33 & 82;  1917: 160, pp. 31 & 79.  Use  0, 1, ..., 9  to make  10.  Solution:  1 35/70  +  8 46/92. 

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. 4: Century addition, p. 4.  Use  1, ..., 9  to add up to  100.  Solutions are:   74 + 12 + 3  =  89 + 6 + 5  =  100;   same as Book of 500 Puzzles;   19 + 28 + 6  =  53 + 47  =  100.

Anonymous problem proposal with solutions by  K. K.;  R. Ichikawa;  S. Tamano  and two papers by T. Hayashi.  J. of the Physics School in Tokyo 5 (1896) 82, 99-103, 153-156 & 266-267, ??NYS  Abstracted in:  Yoshio Mikami, ed.; Mathematical Papers from the Far East; AGM 28 (1910) 16-20; as: A queer number.  In base  b,  we have  [12...b * (b‑2)] + b-1  =  b...21.  [I wonder about other solutions of   a * x + b  =  y,   where  a, b  are digits and  x, y  are pandigital expressions (either with the  9  positive digits or all  10  digits, either separately or together) or  y  is the reversal of  x,  etc.]

H. D. Northrop.  Popular Pastimes.  1901.  No. 17: Magical addition, pp. 69 & 74.  "Arrange the figures 1 to 9, so that by adding them together they will make 100.  How can this be done?"  Solution is:  15 + 36 + 47  =  98 + 2  =  100.  Cf Magician's Own Book.

T. Hayashi.  On the examination of perfect squares among numbers formed by the arrangements of the nine effective figures.  J. of the Physics School in Tokyo 5 (1896) 203-206, ??NYS.  Abstracted in:  Yoshio Mikami, ed.; Mathematical Papers from the Far East; AGM 28 (1910) 23-25.  Says Artemas Martin asked which squares contain all nine positive digits once each and that Biddle found  29  of these.  (Cites same J. 5 (1896) 171, ??NYS, for the solutions.)  [What about with all 10 digits?].

Clark.  Mental Nuts.  1897, no. 98;  1904, no. 42;  1916, no. 44.  One in addition.  "Place the figures 1 2 3 4 5 6 7 8 9 0 to add 100."  Answer:  50 1/2  +  49 38/76.

Ball.  MRE, 4th ed., 1905.  P. 14.  Use the  10  digits to total  1  -- a solution is  35/70 + 148/296  -- or to total  100  -- a solution is  50 + 49 + 1/2 + 38/76.  Use the  9  digits to make four numbers which total  100  -- a solution is  78 + 15 + 2Ö9 + 3Ö64.

Ball.  MRE, 5th ed., 1911.  Pp. 13-14.  Briefly restates the material in the 4th ed. as "questions which have been propounded in recent years.  ...   To the making of such questions of this kind there is no limit, but their solution involves little or no mathematical skill."

M. Adams.  Indoor Games.  1912.  A clever arrangement, p. 353.  Same as Boy's Own Book.

Ball.  MRE, 6th ed., 1914.  Pp. 13-14.  Restates the material in the 5th ed. as "... numerous empirical problems, ....  To the making of such questions there is no limit, but their solution involves little or no mathematical skill."

                    He then introduces the "Four Digits Problem".  "I suggest the following problem as being more interesting."  Using the digits  1, 2, ..., n,  express the integers from  1  up using four different digits and the operations of sum, product, positive integral power and base-10  (or also allowing iterated square roots and factorials).  With  n = 4,  he can get to  88  or to  264.  With  n = 5,  he can get to  231  or  790.  Using  0, 1, 2, 3,  he can get to  36  (or  40). 

Dudeney.  AM.  1917.  Prob. 13: A new money puzzle, pp. 2-3 & 148-149.  States largest amount of old English money expressible with all nine positive digits is  £98765  4s  3½d.  Asks for the smallest amount.  Answer:  £2567  18s  9¾d.

Ball.  MRE, 9th ed., 1920.  Pp. 13-14.  In the "Four Digits Problem", he considers  n = 4,  i.e. using  1, 2, 3, 4,  and discusses the operations in more detail.  Using sum, product, positive integral power and base-10  notation, he can get to  88.  Allowing also finitely iterated square roots and factorials, he can get to  264.  Allowing also negative integral indices, he can get to  276.  Allowing also fractional indices, he can get to  312.  He then mentions using  0, 1, 2, 3  or four of the five digits  1, ..., 5.

Ball.  MRE, 10th ed., 1922.  Pp. 13-14.  In the "Four Digits Problem", he repeats the material of the 9th ed., but at the end he adds that using all of the five digits,  1, ..., 5,  he has gotten to  3832  or  4282,  depending on whether negative and fractional indices are excluded or allowed.

Hummerston.  Fun, Mirth & Mystery.  1924.  Century making, p. 66. 

          15 + 36 + 47 = 98 + 2 = 100.

Wood.  Oddities.  1927.

Prob. 53: Can you do this?, p. 44.  Consider  1, 2, 3, 4  and  5, 7, 8, 9.  Rearrange the sets so both have the same sum!  He phrases it in terms of numbers on jerseys of football players.  Cf Morris, 1991.

Prob. 55: A matter of multiplication, p. 45.  AB*CDE  consists of the same five digits and  A = 1.  Answer is:  14 * 926 = 12964.  He says the only solutions when the condition  A = 1  is dropped are:  24 * 651 = 15624;  42 * 678 = 28476;  51 * 246 = 12546;  57 * 834 = 47538;  65 * 281 = 18265;  65 * 983 = 63895;  72 * 936 = 67392;  75 * 231 = 17325;  78 * 624 = 48672;  86 * 251 = 21586;  87 * 435 = 37845.  Cf Ripley below.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 94: The "100" problem, pp. 88 & 114.  Use  1, 2, ..., 9  to make  100.  Answer:  15 + 37 + 46  =  98  + 2  =  100.

Perelman.  FFF.  1934.  1957: prob. 99 & 101, pp. 137 & 144;  1979: probs. 102 & 104, pp. 166-167 & 174-175.  = MCBF, probs. 102 & 104, pp. 167 & 177-178.

102: One.  "Write one by using all the ten digits."   148/296 + 35/70.   Also  (123456789)0,  etc.

104: Ten Digits.  "Write 100 using all the ten digits.  How many ways are there of doing it?  We know at least four."   70 + 24 9/18 + 5 3/6   and three other similar answers.

See: Meyers in 7.I.1 for the largest integer constructible with various sets of numbers.

Haldeman-Julius.  1937.  No. 29: The 700 problem, pp. 6 & 22.  Arrange the ten digits so "they'll add up to 700."  Answer is:  102 4/8  +  597 3/6.

J. R. Evans.  The Junior Week‑End Book.  Op. cit. in 6.AF.  1939.  Prob. 32, pp. 264 & 270.  Find largest and smallest amounts in pounds, shillings, pence and farthings using the nine positive digits.  Solutions as in Dudeney.

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  P. 56: One hundred per cent.  Form  100  using all  10  digits, but not in order.  Gives just one solution.

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 27: Another sum of money, pp. 14 & 48.  Least amount of English money using all  10  digits.  Answer:  £20567  18s  9¾d.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  1, 2, 3 ... 100,  p. 77.  Use each of the  10  figures to make a total of  100.  Answer:  57 + 23  =  80   + 1 + 4 + 6 + 9  =  100.

Jonathan Always.  Puzzles to Puzzle You.  Op. cit. in 5.K.2.  1965.  No. 91: Monetary matter, pp. 30 & 82.  Find smallest amount, as in Dudeney.

Ripley's Puzzles and Games.  1966.  P. 14.  Says there are five examples of integers  A, B  such that  A + B  and  A x B  have the same digits:  9, 9;  3, 24;  2, 47;  2, 263;  2, 497. 

                    There are infinitely more examples.  One can show that one of  A, B  must be a single digit, say  A,  and that except for the special case  9, 9,  the number of digits in  A + B  must be the same as in  B,  say  n.  There are two special solutions with  n = 1,  namely  0, 0  and  2, 2.  The number of solutions for  n = 2, 3, 4, 5, 6, 7  is  2, 2, 8, 29, 184, 1142.  I have found solutions for each digit  A  greater than one, except for  A = 4, 7  and I searched up to nine digit numbers.

Ripley's Believe it or Not, 14th series.  Pocket Books, NY, 1968.  Unpaged -- about 85% of the way through.   AB * CDE   gives a product of five digits which is a permutation of ABCDE.  E.g.   14 * 926 = 12964.   Asserts there are just  12  "sets of figures in which the multiplicand and the multiplier reappear in the product."  Gives six examples.  Cf Wood, 1927.

Ripley's Believe it or Not, 15th series.  Pocket Books, NY, 1968.  Unpaged -- about 30% of the way through.  Gives  7  examples of the above situation, one of which was given above, saying  "The original figures reappear in the results ...."  Cf Wood, 1927.  They do not seem to have considered other forms, e.g. I find   3 * 51  =  153;   6 * 21  =  126;   8 * 86  =  688.

Doubleday - 3.  1972.  Sum total, pp. 129-130.  Put  1, 2, 3, 5, 6, 7, 8, 9  into two groups of four with the same sum.  Answer:  173 + 5  =  86 + 92.

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 20: Make a century, pp. 20 & 99-100.  Express  100  as a mixed fraction using the nine positive digits, e.g.  81  5643/297.   There are  10  of this form and one other:   3  69258/714.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Fodder for number crunchers, pp. 69 & 128.  Find the  10-digit numbers, using all  10  digits, which are divisible by  1, 2, ..., 18  [i.e. are multiples of  12252240].  There are four such and none is divisible by  19.  [I checked and found no smaller multiples of  12252240  with distinct digits.  I found that there are  94  multiples of  9!  (=  362880)  with distinct digits.  Of these, there are  2, 3, 16, 73  with  7, 8, 9, 10  digits.  (Any multiple of  10!  (=  3628800)  has its last two digits equal.)]

Scot Morris.  The Next Book of Omni Games.  Op. cit. in 7.E.  1991.  Pp. 55 & 192.  Form the digits into two sums:  2 + 6 + 7 + 9  and  1 + 3 + 4 + 5 + 8.  Make the sums equal by moving one number!  Cf Wood, 1927.

Nob Yoshigahara.  Puzzlart.  Tokyo, 1992.  Pandigital times, pp. 16 & 93.  Using two digits for month, day, hour, minute, second, one can use the nine positive digits as in  8:19:23:46:57.  This happens 768 times a year -- what are the earliest and latest such times.  He gets:  3:26:17:48:59  and  9:28:17:56:43  and I have checked this.  [?? -- what if we use all ten digits?]

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Four problems with whole numbers, pp. 127-130. 

                    1.  Find all integers such that the number and its square contain all nine positive digits just once.  Answers:  567  and  854.

                    3.  Find all integers such that its cube and its fourth power contain all ten digits just once.  Answer:  18.

Jamie & Lea Poniachik.  Cómo Jugar y Divertirse con su Inteligencia; Juegos & Co. & Zugarto Ediciones, Argentina & Spain, 1978 & 1996.  Translated by Natalia M. Tizón as: Hard-to-Solve Brainteasers.  Ed. by Peter Gordon.  Sterling, NY, 1998.  Pp. 15 & 71, prob. 21: John Cash.  Find a three digit number  abc  with  abc = 5*bc  and  bc = 5*c.

 

          7.AC.7.        SELF-DESCRIPTIVE NUMBERS, PANGRAMS, ETC.

 

          New section.  Are there older examples?

 

Solomon W. Golomb.  Shift Register Sequences.  Holden-Day, 1967.  ??NYS -- cited by a 1996 article, but I cannot locate the material in the revised ed., Aegean Park Press, Laguna Hills, California, 1982; perhaps it is in some other work of Golomb ??check.  Find a non-decreasing sequence of positive integers,  (ai),  i = 1, 2, ...,  such that  i  appears  ai  times.  Unique answer is:  1, 2, 2, 3, 3, 4, 4, 4, 5, 5, 5, 6, 6, 6, 6, 7, ....

Douglas Hofstadter.  SA (Jan 1982)  c= Metamagical Themas, Basic Books, NY, 1985, chap. 2, pp. 25-48.  In Jan 1981, he had discussed self-referential sentences, and readers sent in a number of numerical ones. 

          Jonathan Post:

                    This sentence contains ten words, eighteen syllables and sixty four letters. 

          John Atkins:

                    'Has eighteen letters' does. 

          Howard Bergerson:

                    In this sentence, the word and occurs twice, the word eight occurs twice, the word four occurs twice, the word fourteen occurs four times, the word in occurs twice, the word seven occurs twice, the word the occurs fourteen times, the word this occurs twice, the word times occurs seven times, the word twice occurs eight times and the word word occurs fourteen times.

          Lee Sallows:

                    Only the fool would take trouble to verify that his sentence was composed of ten a's, three b's, four c's, four d's, forty-six e's, sixteen f's, four g's, thirteen h's, fifteen i's, two k's, nine l's, four m's, twenty-five n's, twenty-four o's, five p's, sixteen r's, forty-one s's, thirty seven t's, ten u's, eight v's, eight w's, four x's, eleven y's, twenty-seven commas, twenty-three apostrophes, seven hyphens, and last, but not least, a single !

          Raphael Robinson asks to fill in the blanks in the following and says there are two

                              solutions:

                    In this sentence, the number of occurrences of  0  is  __,  of  1  is  __,  of  2  is  __,  of  3  is  __,  of  4  is  __,  of  5  is  __,  of  6  is  __,  of  7  is  __,  of  8  is  __,  and of  9  is  __.

          The supplemental material in the book includes

          J. K. Aronson:

                    'T' is the first, fourth, eleventh, sixteenth, twenty-fourth, twenty-ninth, thirty-third, ....

          See below for further material.

Douglas Hofstadter.  Metamagical Themas, Basic Books, NY, 1985, chap. 3, pp. 49-69.  In the Post Scriptum, pp. 68-69, he reports on Sallows' search for a 'pangram'.  Apparently the first of the type he wants is in Dutch, by Rudy Kousbroek and Sarah Hart:  Dit pangram bevat vijf a's, twee b's, ....  After some months search, Sallows' computer found: 

                    This pangram tallies five a's, one b, one c, two d's, twenty-eight e's, eight f's, six g's, eight h's, thirteen i's, one j, one k, three l's, two m's, eighteen n's, fifteen o's, two p's, one q, seven r's, twenty-five s's, twenty-two t's, four u's, four v's, nine w's, two x's, four y's, and one z. 

          He challenges people to compute a version starting: 

                    This computer-generated pangram contains ....

Douglas Hofstadter.  Metamagical Themas, Basic Books, NY, 1985, chap. 16 pp. 364-395.  In the Post Scriptum, pp. 387-395, he continues his discussion of the above material.  He notes that Robinson's problem is convergent in the sense that if one inserts a random sequence of numbers, then counts the occurrences of the numbers and uses the counts as a new number, etc., then this iterative process usually converges to a solution.  There are two solutions, but there is also a two term cycle and Hofstadter conjectures all initial values converge to one of these three situations.  Sallows' challenge was given in A. K. Dewdeney's Computer Recreations column (SA, Oct 1984) and Larry Tesler used an iterative program on it.  Tesler soon found a loop and modified the program a bit to obtain a solution: 

                    This computer-generated pangram contains six a's, one b, three c, three d's, thirty-seven e's, six f's, three g's, nine h's, twelve i's, one j, one k, two l's, three m's, twenty-two n's, thirteen o's, three p's, one q, fourteen r's, twenty-nine s's, twenty-four t's, five u's, six v's, seven w's, four x's, five y's, and one z. 

Lee Sallows.  In Quest of a Pangram.  Published by the author, Holland, nd [1985?].  Describes his search for a pangram.

Lee Sallows.  Reflexicons.  Word Ways 25 (1992) 131-141.  A 'reflexicon' is a list of numbers and letters which specifies the number of times the letter occurs in the list.  There are two in English. 

                    fifteen e's,  seven f's,  four g's,  six h's,  eight i's,  four n's,  five o's,  six r's,  eighteen s's,  eight t's,  four u's,  three v's,  two w's,  three x's.

                    sixteen e's,  five f's,  three g's,  six h's,  nine i's,  five n's,  four o's,  six r's,  eighteen s's,  eight t's,  three u's,  three v's,  two w's,  four x's.

                    He also discusses 'pangrams', which are sentences containing the above kind of information -- e.g.  This sentence contains one hundred and ninety-seven letters:  four a's, ....  The search for these is described in his booklet cited above.  He then discusses crosswords using the number names.

Tony Gardiner.  Challenge! What is the title of this article?  Mathematics Review 4:4 (Apr 1994) 28-29.  Following on a previous article in 4:1, he discusses self-describing sequences, where the description arises by reading the sequence.  E.g.  22  is read as  'two twos';  31 12 33 15  is  'three ones, one two, three threes, one five'.  He also mentions self-describing lists, e.g.  1210   contains  'one 0, two 1s, one 2, zero 3s'.

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Pp. 117-122.  Considers self-describing lists of length  n  and shows there are only the following:  1210,  2020,  21200  and, for  n > 6,  (n-4)2100...001000

Lee Sallows.  Problem proposal to Puzzle Panel, 18 Jun 1998.  "How many letters would this question contain, if the answer wasn't already seventy three?"

 

          7.AD. SELLING, BUYING AND SELLING SAME ITEM

 

            The problem is to determine the profit in a series of transactions involving the same item, but there is usually insufficient information.  But see Clark and Sullivan for an unusual answer.

 

Clark.  Mental Nuts.  1897, no. 1;  1904, no. 2;  1916, no. 24.  The horse question.  Sell a horse for $90,  buy back at $80,  resell at $100.  "What did he make on the transaction?"  Answer is $20,  but this assumes the horse had no initial cost.  If the item has no initial cost and the prices are  a, b, c,  then the gain is  a - b + c.  But if the question is asking for the profit, then the data are insufficient as the base cost is not given.

Loyd.  The trader's profit.  Cyclopedia, 1914, pp. 291 (no solution).  (= MPSL1, prob. 13 -- What was the profit?, pp. 12 & 125.)  Sells a bicycle at  50,  buys back at  40,  sells again at  45.  Lengthy discussion of various 'solutions' of  15, 5 and 10.  He says "the President of the New York Stock Exchange was bold enough to maintain over his own signature that the profit should be  $10."  Gardner points out that there isn't enough information.

H. E. Licks.  Op. cit. in 5.A.  1917.  Art. 25, p. 18.  Sells at  50,  buys back at  45,  sells at  60  -- what is the profit?  Author labels it impossible.

Smith.  Number Stories.  1919.  Pp. 126‑127 & 146.  Buys at  5000,  sells at  5000,  buys back at  4500,  sells again at 5500.

Loyd Jr.  SLAHP.  1928.  The used‑car puzzle, pp. 9‑10 & 88.  Sell a used car at  100,  buy back at  80,  resell at  90.  "This popped into my head one morning ...."  Gives arguments for profits of  30, 10 and 20.  Solution says the information is insufficient.

Collins.  Fun with Figures.  1928.  This sticks 'em up, p. 69.  Buys at  $55,  sells at  $55,  buys back at  $50,  sells again at  $60.

Rudin.  1936.  No. 160, pp. 57 & 113.  Buy for  $70,  sell for  $80,  buy back for  $90,  sell for  $100.

Sullivan.  Unusual.  1943.  Prob. 2: A business transaction.  Sell for  $4000,  buy back for  $3500,  sell again for  $4500.  Says the gain is  $5000,  composed of the initial  $4000  plus the gain on the buying and selling.  This is like Clark and markedly different than most approaches, which refer to profit.

Hubert Phillips.  Something to Think About.  Ptarmigan (Penguin), 1945.  Problem 12: Alf's bike, pp. 15 & 89.  More complex version with four persons and each person's percentage profit or loss given.

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  P. 40: Multiple choice (B).  Sell a cow for  $100,  buy back at  $90,  resell at  $120.  Says the profit is  $30.

 

          7.AD.1.       PAWNING MONEY

 

Viscount John Allsebrook Simon.  [Memory of Lewis Carroll.]  Loc. cit. in 7.S.2.  = Carroll-Wakeling II, prob. 33: Going to the theatre, pp. 51 & 73.  He says Carroll gave the following.  Man pawns  12d  for  9d  and sells the ticket to a friend for  9d.  Who loses and how much?  Simon relates that when he said that the friend lost  6d,  Carroll pointed out that pawnbrokers charge interest.  Mentioned in Carroll-Gardner, p. 80, who gives the full name.  The DNB says he entered Wadham College, Oxford, in 1892, and his Memory says he met Carroll then.  So this dates from ³1892.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  He doesn't have the problem, but on p. 155 he gives current pawnbrokers' charges.  These would be  ½d  for the ticket and  ½d  per  2s  or part thereof per month or part thereof.  So the above will cost at least  ¾d.

Smith.  Number Stories.  1919.  Pp. 131 & 148.  Pawn  $1  for  $.75  and sell ticket for  $.50.  Doesn't consider interest.

Lilian & Ashmore Russan.  Old London City  A Handbook, Partly Alphabetical.  Simpkin, Marshall, Hamilton, Kent & Co., London, 1924, p. 222.  Says the Bank of England once contemplated setting up a pawn business, to charge  1d  per  £1  per month.

R. Ripley.  Believe It Or Not!  Book 2.  Op. cit. in 7.J.  1931.  P. 143.  "A man owed  $3.00.  He had a  $2.00  bill, which he pawned for  $1.50,  and then sold the pawn ticket to another man for  $1.50,  who redeemed the  $2.00  bill.  Who lost?"  No answer given.

H. Phillips.  The Playtime Omnibus.  Op. cit. in 6.AF.  1933.  Section XVII, prob. 7, pp. 54 & 236.  Jones pawns  6d  for  5d,  then sells the ticket to Brown for  4d.  Who lost?  Doesn't consider interest.

John Paul Adams.  We Dare You to Solve This!.  Op. cit. in 5.C.  1957?  Prob. 52: Losers weepers, pp. 32 & 49.  Pawn  $5  for  $3,  sell ticket to a friend for  $3  and he redeems it.  Who lost and how much?  Doesn't consider interest.

 

          7.AE.           USE OF COUNTERFEIT BILL OR FORGED CHEQUE

 

            These problems give one or several transactions involving a bill or cheque which is then found to be counterfeit or forged.  Who loses and how much?  In the classic version of the hatter, the straightforward answer is that he loses the value of the counterfeit bill.  However, there are two values for the hat -- the sale price and the cost price, whose difference is the profit that would be made in a normal sale.  One can argue that one has only lost the cost price of the hat, so the loss is the value of the counterfeit bill less the expected profit.

 

Magician's Own Book.  1857.  The unlucky hatter, p. 245.  Man buys  $8  hat with counterfeit  $50  bill.  "... and in almost every case the first impression is, that the hatter lost  $50  besides the hat, though it is evident he was paid for the hat...."  = Book of 500 Puzzles, 1859, p. 59.  = Boy's Own Conjuring Book, 1860, pp. 215‑216, but this spells out  $  as dollars.. 

Lemon.  1890.  The unlucky hatter, no. 225, pp. 34 & 106.  Man pays for  $8  hat with counterfeit  $50  bill.  "In almost every case the first impression ... is that the hatter lost  $50  beside [sic] the hat..." 

Hoffmann.  1893.  Chap. IV, no. 43: What did he lose?, pp. 153‑154 & 205‑206 = Hoffmann‑Hordern, p. 129.  Man pays for hat with a counterfeit bill.  How much does the seller lose?  Answer says that "The reply of most people is, almost invariably, that the hatter lost [the change] and the value of the hat, but a little consideration will show that this is incorrect."  He then says that the seller loses the amount given in change less his profit on the goods sold;  "the nett value of the hat, plus such trade profit, being balanced by the difference ... which he retained out of the proceeds of the note."  This is wrong, as the sale price is part of the the refund that he has to make to the person who changed the note.  Hordern notes that Hoffmann is wrong and 'the reply of most people' is indeed correct.

Clark.  Mental Nuts.  1897, no. 2;  1904, no. 21;  1916, no. 7.  The shoe question.  Boy pays for  $4  pair of shoes with counterfeit  $10  bill.  Answer says he lost  $6  and the pair of shoes.

Dudeney.  Some much‑discussed puzzles.  Op. cit. in 2.  1908.  Dud cheque used to buy goods and get cash.  "Perpetually cropping up in various guises."

H. E. Licks.  Op. cit. in 5.A.  1917.  Art. 21, p. 17.  Use of a counterfeit bill.

Lynn Rohrbough, ed.  Mental Games.  Handy Series, Kit E, Cooperative Recreation Service, Delaware, Ohio, 1927.  Counterfeit Bill, p. 10.  Man buys $6 pair of shoes with a phoney $20 bill.  How much did seller lose?  No solution given.

Ahrens.  A&N, 1918, pp. 95‑96 gives such a problem.

Dudeney.  PCP.  1932.  Prob. 34: The banker and the note, pp. 21 & 131.  = 536; prob. 31: The banker and the counterfeit bill, pp. 10‑11 & 230.  Counterfeit bill goes in a circle so no one loses.

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 15, pp. 14‑15 & 188.  Forged cheque used to settle account and receive cash.

Robert A. Streeter & Robert G. Hoehn.  Are You a Genius?  Vol. 1, 1932; vol. 2, 1933, Frederick A. Stokes Co., NY.  Combined ed., Blue Ribbon Books, NY, 1936.  Vol. 1, p. 46, no. 10: "Brain twister".  Man owes me  40¢,  he gives me a knife worth at least  60¢  and I give him  20¢.  I then find the knife was stolen and I pay the owner  75¢,  which is its value.  How much have I lost?  Answer is  60¢.  This is based on my payment of  75¢  being a fair purchase, so my loses are the  40¢  debt and the  20¢  change, which are now irrecoverable.  However, if the first transaction is considered fair, then I've lost  75¢.  Further, I might consider the original debt as a past loss, which would reduce the present loss to  20¢  or  35¢.

Phillips.  The Playtime Omnibus.  Op. cit. in 6.AF.  1933.  Section XVII, prob. 5: Pinchem, pp. 53 & 236.  Identical to Week‑End.

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 3, pp. 187-188 & 196-197.  Sell bracelet worth  60  for phoney  100  bill, which is changed by the neighbouring shopkeeper.  Author says he gets many answers, including  140,  200  and even  340,  but the right answer is  100.

Depew.  Cokesbury Game Book.  1939.  Shoe dealer, p. 211.  Sell shoes worth  $8  for phoney  $20  bill.  Loss is  $12  plus value of shoes.

McKay.  Party Night.  1940.  No. 7, p. 177.  Man buys boots worth  15s  with a bad  £1  note.  Says he gets answers up to "35s  and a  15s  pair of boots".  Notes that the neighbouring grocer who changed the bill is irrelevant and the bootseller is simply  £1  out of pocket.  This ignores his profit on the boots. 

Meyer.  Big Fun Book.  1940.  No. 10, pp. 171 & 754.  Same as Streeter & Hoehn.

Doubleday - 1.  1969.  Prob. 8: Cash on delivery, pp. 15 & 157.  = Doubleday - 4, pp. 19-20.  Forged $5 bill used to settle a circle of debts.  Solution claims everything returns to the previous state, except for one stage -- this seems very confused and incorrect.

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 13: The counterfeit note, pp. 17 & 98.  Counterfeit note goes in a circle.  She claims all the transactions are invalid -- but the bill has simply allowed a circuit of debts to be cancelled and hence no one has lost and there is no reason to cancel anything.

 

          7.AF.           ARITHMETIC PROGRESSIONS

 

            See Tropfke 625.

            There are many problems which are based on this.  Some occur in 7.H.7 and 10.A.  Here I only include the most interesting.

 

Bakhshali MS.  c7C.  In:  G. R. Kaye, The Bakhshāli manuscript;  J. Asiatic Soc. Bengal (2) 8:9 (Sep 1912) 349‑361; p. 358  and in  Kaye I 43 & III 176-177, ff. 4r-5r, sutra 18  and in  Gupta.  Consider two APs  a, a + b, ...,  and  c, c + d, ...,  and suppose the sums after  n  terms are equal to  S.  In the notation of 10.A, this is O-(a, b; c, d).  Then  n = 2(a ‑ c)/(d - b) + 1.  Does examples with  (a, b; c, d)  =  (4, 3; 6, 1);  (2, 3; 3, 2)  and  (5, 6; 10, 3).

                    Kaye III 174, f. 4v  &  Gupta.  This is a problem of the same type, but most of it is lost and the scribe seems confused.  Gupta attempts to explain the confusion as due to using the data   a, b; c, d  =  3, 4; 1, 2,   with the rule   n  =  2(c‑a)/(b-d) + 1,   where the scribe takes the absolute values of the differences rather than their signed values.  In this way he gets  n = 3  rather than  n = -1.

Pacioli.  Summa.  1494.  F. 44v, prob. 32.  1 + 2 + ... + 10½.  He gets  10½ x 11½ / 2.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 182 & 263, no. 693.  Man digging a well  49  feet deep.  First foot costs  15,  but each successive foot costs  6  more than the previous.  Find cost of last foot and total cost.  So this is really an arithmetic progression problem, but I haven't seen others of those using this context.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  Prob. 550, p. 239.  How many edges and diagonals does a convex octagon have? 

(Beeton's) Boy's Own Magazine 3:6 (Jun 1889) 255  &  3:8 (Aug 1889) 351.  (This is undoubtedly reprinted from Boy's Own magazine 1 (1863).)  Mathematical question 59.  Seller of  12  acres asks  1  farthing for the first acre,  4  for the second acre,  16 for the third acre, ....  Buyer offers  £100  for the first acre,  £150 for the second acre,  £200  for the third acre, ....  What is the difference in the prices asked and offered?  Also entered in 7.L.

Perelman.  1937.  MCBF, A team of diggers, prob. 195, pp. 372-373.  A team can dig a ditch in  24  hours, but just one digger begins and then the others join in at equal intervals, with the work finished in one interval after the last man joined.  The first man works  11  times as long as the last man.  How long did the last man work?  Perelman finds this noteworthy (and I agree) because the number of men in the team cannot be determined!

 

          7.AF.1.        COLLECTING STONES

 

Alcuin.  9C.  Prob. 42: Propositio de scala habente gradus centum.  Computes  1 + 2 + ... + 100   as  100 + (1+99) + (2+98) + ... + 50.

Pacioli.  Summa.  1494.  F. 44v, prob. 31.  Collect  100  oranges.

Pacioli.  De Viribus.  c1500.  Ff. 122v - 124r.  C(apitolo) LXXIII. D(e). levare .100. saxa a filo (To pick up 100 stones in a line).  Wager on the number of steps to pick up  100  stones (or apples or nuts), one pace apart.  Gives the number for  50  and 1000  stones.

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 55, f. EE.iiii.r (pp. 150-151).  (The 55 is not printed in the Opera Omnia.)  Picking up  100  stones in a line.  (H&S 56-57 gives Latin with English summary.)

Buteo.  Logistica.  1559.  Prob. 87, pp. 299-300.  Ant collecting  100  grains.  (H&S 56.)

H&S 56 says this occurs in Trenchant (1566), ??NYS.

Baker.  Well Spring of Sciences.  1562?  ??check if this is in the Graves copy of the 1562/1568 ed.  Prob. 2,  1580?: f. 36r;  1646: p. 56;  1670: pp. 72-73.  100  stones.

van Etten.  1624.  Prob. 87 (84), part IV (8), p. 114 (184).  100  apples, eggs or stones.  Henrion's Notte, p. 38, observes that there are many arithmetical errors in prob. 87 which the reader can easily correct.

Ozanam.  1694.  Prob. 7, question 6, 1696: 53;  1708: 29.  Prob. 10, question 6, 1725: 65‑66.  Prob. 1, 1778: 64-65;  1803: 66-67;  1814: 59-60;  1840: 32.  100  apples, becoming stones in 1778 et seq.  1778 describes a bet based on this process versus a straight run of the same distance.

Wells.  1698.  No. 101, p. 205.  20  stones.

Dilworth.  Schoolmaster's Assistant.  1743.  P. 94, no. 6.  100  stones  2  yards apart.  Answer in miles, furlongs and yards.

Walkingame.  Tutor's Assistant.  1751.  Arithmetical Progression, prob. 3, 1777: p. 90;  1835: p. 98;  1860: p. 118.  100  eggs a yard apart.  Answer in miles and yards.

Edmund Wingate (1596-1656).  A Plain and Familiar Method for Attaining the Knowledge and Practice of Common Arithmetic.  ....  19th ed., previous ed. by John Kersey (1616-1677) and George Shell(e)y, now by James Dodson.  C. Hitch and L. Hawes, et al., 1760.  Quest. 44, p. 366.  100  stones a yard apart.

Mair.  1765?  P. 483, ex. II.  100  eggs a yard apart.  Answer:  5 miles, 1300 yards.

Euler.  Algebra.  1770.  I.III.IV: Questions for practice, no. 4, p. 139.  100  stones a yard apart.  Answer:  5 miles, 1300 yards.

Vyse.  Tutor's Guide.  1771? 

Prob. 3, 1793: p. 133; 1799: p. 141 & Key p. 186.  94 eggs.

Prob. 4, 1793: p. 133; 1799: p. 141 & Key p. 186.  100 stones.

Prob. 21, 1793: p. 138; 1799: p. 146 & Key pp. 188-190.  1000 eggs,  2 yards apart, gathered by ten men, each man to collect ten and then the next man to collect the next ten, etc.  Can they do it in 24 hours?  How far did each man run?

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?] 

1804: prob. 4, p. 125.  100  stones a yard apart.  Converts answer to miles.

1804: prob. 74, p. 140.  100  eggs a yard apart.  Again converts to miles.

Bonnycastle.  Algebra.  1782.  P. 60, no. 7 (1815: p. 76, no. 8).  100  stones, a yard apart.

Pike.  Arithmetic.  1788.  P. 221, no. 3.  Stones laid a yard apart over a mile, starting a yard from the basket.  Finds the travel is  1761  miles.

Bullen.  Op. cit. in 7.G.1.  1789.  Chap. 30, prob. 6, pp. 213‑214.  100  stones, at  2  yard intervals.  Converts to miles, furlongs and yards.

Eadon.  Repository.  1794. 

P. 235, ex. 4.  Collect  500  stones a yard apart.  This takes  142  miles and some, so he could sooner run from Sheffield to York and back, since they are  50  miles apart.

P. 374, no. 19.  1760  stones a yard apart.  In six days a man only manages to collect  769  of them.  How far has he gone and how much farther has he to go?

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995.  P. 100.  100  eggs a yard apart.  He also gives a variation:  100  sheep are priced in arithmetic progression, with the first costing  1s  and the last costing  £9 19s (= 199s);  what do the sheep cost all together?

Hutton.  A Course of Mathematics.  1798?  Prob. 7,  1833: 277;  1857: 313.  100  stones  2  yards apart.  Gives answer in miles and yards.

Manuel des Sorciers.  1825.  Pp. 83-84.  ??NX  120 stones 6 feet apart.

Endless Amusement II.  1826?  Pp. 115-116:  "If a hundred Stones ...."

Boy's Own Book.  The basket and stones.  1828: 176;  1828-2: 239;  1829 (US): 107;  1843 (Paris): 342;  1855: 394;  1868: 432.  100  stones a yard apart.  = Boy's Treasury, 1844, pp. 299-300.  = de Savigny, 1846, pp. 290-291: Le panier et les petites pierrés, using mètres instead of yards, except that it ends with '10,100 mètres, ou 21 kilomètres' -- ??

Bourdon.  Algèbre.  7th ed., 1834.  Art. 190, question 5, p. 319.  Loads of sand (or grit) have to be delivered, one load at a time, to  100  vehicles in a line,  6 meters apart, from a pile  40 meters from the end of the line.

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 24,  1857: 82.  100 eggs a yard apart.

Nuts to Crack XIV (1845), no. 76.  The basket and stones.  Almost identical to Boy's Own Book.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 197: "The following is a favorite old problem."  100  stones, a yard apart.

John Radford Young (1799-1885).  Simple Arithmetic, Algebra, and the Elements of Euclid.  IN: The Mathematical Sciences, Orr's Circle of the Sciences series, Orr & Co., London, 1854.  [An apparently earlier version is described in 7.H.]  No. 2, p. 228.  200  stones.

Magician's Own Book.  1857.  The basket and stones, pp. 246-247.  100  stones, one yard apart.  = Book of 500 Puzzles, 1859, pp. 60-61.  = Boy's Own Conjuring Book, 1860, p. 218.  = Indoor & Outdoor, c1859, part II, prob. 20, pp. 136-137.

Illustrated Boy's Own Treasury.  1860.  Prob. 30, pp. 429-430 & 434.  100  trees to be watered, five steps apart.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-32, pp. 255 & 396: Am Feste der bemalten Eier.  100  eggs, a fathom apart.  Says that the 'egg gathering' is a traditional race.

Todhunter.  Algebra, 5th ed.  1870.  Examples XXX, no. 23, pp. 265 & 588.  Basket at origin and  n-1  stones, with first stone at  1,  then second is  3  further,  then  5  further, the  7  further, ....  That is, the stones are at positions  1, 4, 9, 17, ...,  so the total distance is  2 [12 + 22 + ... + (n-1)2]  =  n(n-1)(2n-1)/3.

Daniel W. Fish, ed.  The Progressive Higher Arithmetic, for Schools, Academies, and Mercantile Colleges.  Forming a Complete Treatise of Arithmetical Science, and its Commercial and Business Applications.  Ivison, Blakeman, Taylor & Co., NY, nd [but prefaces give: 1860; Improved Edition, 1875].  Pp. 412-413.  Equivalent to a man picking up stones with the distance to the first stone being 5, the distance to the last stone being 25 and his total travel being 180.

M. Ph. André.  Éléments d'Arithmétique (No 3) a l'usage de toutes les institutions ....  3rd ed., Librairie Classique de F.-E. André-Guédon, Paris, 1876.  Probs. 551 & 552, p. 239, are similar to Bourdon.

W. W. Rouse Ball.  Elementary Algebra.  CUP, 1890 [the 2nd ed. of 1897 is apparently identical except for minor changes at the end of the Preface].  Prob. 38, pp. 347 & 480.  10  balls equally spaced in a row, starting  12 ft  from the basket.  A boy picks them up in the usual way and find he has travelled  ¼ mile.  What was the spacing?

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XXX: La course des œufs, pp. 110-111.  100  apples.  Says it is played on the beaches with eggs in good weather.

A. Sonnenschein & H. A. Nesbit.  The New Science and Art of Arithmetic For the Use of Schools.  A. & C. Black, London, 1903.  Pp. 342 & 489.  Potato race with  50  potatoes a yard apart to be retrieved.

Pearson.  1907.  Part II, no. 143: The stone carrier, pp. 142 & 219.  52  stones, with spacing  1, 3, 5, ..., 103,  to be brought to the first stone, yielding   2 (1 + 3 + 5 + ... + 103)  =  2 * 522.

Wehman.  New Book of 200 Puzzles.  1908.  An egg problem, p. 53.  100 eggs, a yard apart.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  More mileage possible, pp. 187-188.  Consider  10  points equally spaced on a line.  Salesman starts at one of them and visits each of them once.  What is the maximum mileage possible?  They get 49 miles by going  5 7 4 8 3 9 2 10 1 6.  Adapting my computer search mentioned below, I find this is indeed maximal and there are  1152  maximal solutions.  Though not the same as the above problems, this uses the same set-up, except one does not return to base after each visit.

David Singmaster.  Problem proposal sent to AMM on 22 Oct 2001.  General version of the Scotts' problem with a base office at position  0  and the salesman has to start and return to it.  Conjectures the maximum distance for  N  customers is  ë(N+1)2/2û,  which is verified by exhaustive search up to  N = 10,  but there are many such trips, e.g.  28,800  of them for  N = 10.  Adding  0  at each end of the Scotts' trip is an example of a maximal journey.  The referee produced an elegant proof which also gives the number of maximal journeys.  The editor suggested it would be more appropriate for MM and a revised version has been submitted there.

 

          7.AF.2.        CLOCK STRIKING

 

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 113, f. 56v.

Rudolff.  Künstliche rechnung, 1526, op. cit. in 7.L.2.b.  1540 ed., f. N.vii.v.  ??NYS.  (H&S 57 gives the German.)

Apianus.  Kauffmanss Rechnung.  1527.  How many times does a clock strike during  1  to  12?

H&S 57 says this is in Buteo (1559), ??NYS.

Wingate.  Arithmetic.  1629?  Cf Wingate/Kersey.  This item is in a section which Kersey did not revise.  P. 296 in the 1678? ed.  "How many strokes the Clock strikes betwixt midnight and noon."

Baker.  Well Spring of Sciences.  Prob. 1, 1670: p. 71, ??NX.  1 + ... + 12  strokes.

Wells.  1698.  No. 97, p. 204.  1 + ... + 12  strokes.

A Manual of Curious and Useful Questions.  MS of 30 Jan 1743(OS) owned by Susan Cunnington (??NYS) and described in her:  The Story of Arithmetic; Swan Sonnenschein, London, 1904, pp. 155-157.  "How many strokes do ye clocks of Venice (which go on to  24  o' th' clock) strike in the compass of a natural day?"  [Where is this MS??]

Dilworth.  Schoolmaster's Assistant.  1743.  P. 93, no. 1.  "How many strokes does the hammer of a clock strike in  12  hours?"

Walkingame.  Tutor's Assistant.  1751.  Arithmetical Progression, prob. 1, 1777: p. 90;  1835: p. 98;  1860: p. 118.   How many strokes in  12  hours?

Mair.  1765?  P. 483, ex. V.  How many strokes in  12  hours?

Euler.  Algebra.  1770.  I.III.IV: Questions for practice, no. 3, p. 139.  "The clocks of Italy go on to  24  hours:  how many strokes do they strike in a complete revolution of the index?"

Vyse.  Tutor's Guide.  1771? 

Prob. 1, 1793: p. 133; 1799: p. 141 & Key p. 186.  "How many strokes do the Clocks at Venice (which go on to  24  o'Clock) strike in the Compass of a natural Day?"

Page 2, 1793: p. 133; 1799: p. 141 & Key p. 186.  "How many Strokes does the Hammer of a Clock strike in  12  Hours?"

Pike.  Arithmetic.  1788.  P. 221, no. 2.  "It is required to find out how many strokes the hammer of a clock would strike in a week, or  168  hours, provided it increased at each hour?"

Bonnycastle.  Algebra.  1782.  P. 60, no. 5 (1815: p. 75, no. 5).  "How many strokes do the clocks in Venice, which go on to  24  o'clock, strike in the compass of a day?"  (1815 omits "the compass of".)

Hutton.  A Course of Mathematics.  1798? 

Prob. I-2,  1833 & 1857: 66.  "It is required to find the number of all the strokes a clock strikes in one whole revolution of the index, or in 12 hours?"

Prob. I-3,  1833 & 1857: 67.  "How many strokes do the clocks of Venice strike in the compass of the day, which go right on from 1 to 24 o'clock?"

Prob. 5,  1833: 277;  1857: 313.  "How many strokes do the clocks of Venice, which go on to  24  o'clock, strike in the compass of a day?"  See Bonnycastle, 1782.

D. Adams.  New Arithmetic.  1835.  P. 224, no. 12.  "How many times does a common clock strike in 12 hours?"

Hutton-Rutherford.  A Course of Mathematics.  1841?  Prob. 25,  1857: 82.  "The clocks of Italy go on to 24 hours; then how many strokes do they strike in one complete revolution of the index?"

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  P. 196.  "How many strokes does a common clock strike in the compas of 12 hours?"

[Chambers].  Arithmetic.  Op. cit. in 7.H.  1866?  P. 224, quest. 5.  "How many times does a common clock strike in a day?"  Answer:  156.

James Cornwell & Joshua G. Fitch.  The Science of Arithmetic: ....  11th ed., Simpkin, Marshall, & Co., London, et al., 1867.  (The 1888 ed. is almost identical to this, so I suspect they are close to identical to the 2nd ed. of 1856.)  Exercises CXXXVIII, no. 9, pp. 291 & 370.  "How many times does the hammer of a clock strike in a week?"

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XXXI: Les quatre cents coups, p. 111.  A clock that goes to  12,  strikes  78 + 78 = 156  hours in a day.  If it also strikes quarters, it makes  240  of those, totalling  396,  nearly  400  blows per day.

Charles Pendlebury & W. S. Beard.  A "Shilling" Arithmetic.  Bell, London, 1899 -- my copy is 59th ptg, 1960 and says it was reset for the 49th ptg of 1944.  Examination Papers III, prob. 181.  How many hour strokes does a clock make in a day?  How many in the year 1896? 

 

          7.AG. 2592

 

Dudeney.  AM.  1917.  Prob. 115: A printer's error, pp. 20 & 162.   ABCA  =  ABCA   has the unique solution  2592.

Hubert Phillips.  Question Time.  Op. cit. in 5.U.  1937.  Prob. 133: Phoney 'phone, pp. 87 & 219.   ABCD  =  ABCD.   Answer asserts that  2592  is the unique solution.

Donald L. Vanderpool.  Printer's "errors".  RMM 10 (Aug 1962) 38.  Extends Dudeney's examples, e.g.   25 * 25/31  =  25  25/31;   34 * 425  =  34425   (which can be multiplied by any power of  10).

M. H. Greenblatt.  Mathematical Entertainments, op. cit. in 6.U.2, 1968.  Expression for which  2,592  is a solution, pp. 15‑16.  He asserts that Dudeney's expression was discovered by his officemate, Larry, in response to a colleague, B. Rothlein, complaining that he could not remember his telephone number,  EVergreen 2592.  Rothlein had this telephone on 41st St., Philadelphia, during 1943‑47.  [A likely story??]

 

          7.AH. MULTIPLYING BY REVERSING

 

            For two-digit numbers, reversing and shifting are the same, but I will consider them here as the transformation  ab  to  ba  seems more like a reversal than a shift.

            I have just noted the items by Langford in 7.AR and the item by Meyer below which make me realise that this section is connected to 7.AR.  Meyer doubles the result of  1089  which comes up in 7.AR and gets  2178  which I remembered occurs in this section.  Upon investigating, we see Langford notes that   1089  =  1100 - 11   so that  k * 1089  =  kk00 ‑ kk  =  k,k‑1,9‑k,10-k   in base  10.  From this we see that  k*1089  is the reverse of  (10-k)*1089.  Now  10-k  is a multiple of  k  for  k = 1, 2,  but we get some new types of solution for  k = 3, 4,  namely:   7 * 3267  =  3 * 7623;   6 * 4356  =  4 * 6534.

            This pattern leads to all solutions in 7.AR and also leads to further solutions here, using   11000 ‑ 11 = 10989,  etc., but this does not seem to give a complete solution here.

 

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4r has entry 12:  "Given product of a number consisting of  n  figures mult. by same [some??] figures in an inverted order  query num"  Though a bit cryptic, this seems to refer to the problem of this section.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 136 & 258, no. 525.  ab  =  2⅔ * ba.  He adds a further condition, but this is not needed.

Sphinx.  1895.  Arithmetical, no. 217, pp. 33 & 104.   x2  =  4*(x reversed).

T. C. Lewis.  L'Intermédiaire des Mathematiciens 18 (1911) 128.  ??NYS. Dickson, Vol. 1, p. 463, item 70 says he discusses "number divisible by the same number reversed".

Ball.  MRE, 6th ed., 1914, p. 12.  Mentions the general problem and gives   8712  =  4*2178  and   9801  =  9*1089   as examples.  Gives four citations to L'Intermédiaire des Maths., none of them the same as the above!

R. Burg.  Sitzungsber. Berlin Math. Gesell. 15 (1915) 8‑18.  ??NYS. Dickson, vol. 1, p. 464, item 83, says he found those  N, base 10,  whose reversal is  kN, in particular for  k = 9, 4.

Wood.  Oddities.  1927.  Prob. 56: Wizard stunts, pp. 45-46.  Notes that  83 * 41096 = 3410968  and asks for more examples of  AB * n = BnA.  Finds  86 * 8 = 688  and then says the next example is 'infinitely harder to find':  71 * 16 39344 26229 50819 67213 11475 40983 60655 73770 49180 32787. 

Haldeman-Julius.  1937.  No. 102: Four-digit problem, pp. 17 & 26.  Asks for a four digit number whose reversal is four times it.  Answer:  2178.

Ball.  MRE, 11th ed., 1939, p. 13.  Adds some different types of examples.   312*221  =  68952;   213*122  =  25986.   See also 7.AJ.

G. H. Hardy.  A Mathematician's Apology.  CUP, 1940.  Pp. 44-45.  "8712  and  9801  are the only four-figure numbers which are integral multiples of their 'reversals': ....  These are odd facts, very suitable for puzzle columns and likely to amuse amateurs, but there is nothing in them which appeals to a mathematician."  He cites Ball for this information.

Morley Adams.  1948.  See item in 7.AC.1 for an example of   6 * ABC  =  5 * CBA.

K. Subba Rao.  An interesting property of numbers.  The Mathematics Student 27 (1959) 57‑58.  Easily shows the multiplier must be  1, 4 or 9  and describes all solutions.

Jonathan Always.  Puzzling You Again.  Tandem, London, 1969.  Prob. 32: Four different answers, pp. 23 & 80.   AB * 7/4  =  BA   has four solutions.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  A tram called Alec, pp. 89 & 137-138.   4 * TRAMS  =  SMART.

Jerome S. Meyer.  Arithmetricks.  Scholastic Book Services, NY, 1965. 

The  2178  trick, pp. 3-4.  Essentially the trick which gives  1089,  but he doubles the result at the end to get  2178.

Juggling numbers no. 2, pp. 83 & 88.  4 * ABCDE  =  EDCBA  with solution  21978.

 

          7.AH.1.       OTHER REVERSAL PROBLEMS

 

          New section.  There are many forms of this that I have not recorded before.

 

William Leighton, proposer;  Rich. Gibbons, solver.  Ladies' Diary, 1751-52  =  T. Leybourn, II: 49, quest. 338.  ABC  has digits in arithmetic progression, its value divided by the sum of its digits is 48  and  CBA - ABC = 192.  This is rather overdetermined as there are only two three digit numbers whose quotient by the sum of the digits is 48, namely  432  and  864. 

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Art. 114, pp. 29-30.  ab + 18 = ba  with  a + b = 6.

Todhunter.  Algebra, 5th ed.  1870.  Examples XIII, nos. 30-31, pp. 104 & 578.

No. 30.  ab = 3(a+b)  and  ab + 45 = ba.  In fact, the second condition is unnecessary!

No. 31.  ab = 7(a+b)  and  ab = 27 + ba.  Here both conditions are needed.

Haldeman-Julius.  1937.  No. 108: A problem in 2's, pp. 13 & 26.  x + 2  is reversal of  2x.  Answer:  47.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 37: In reverse, pp. 28 & 84.  Which two digit numbers squared are the reversals of the squares of their reversals?  These are given by  12  and  13.

 

          7.AI.            IMPOSSIBLE EXCHANGE RATES

 

F. & V. Meynell.  The Week‑End Book.  Op. cit. in 7.E.  1924.  2nd ed., prob. three, p. 274;  5th? ed., prob. six, pp. 407‑408.  US & Mexico value each other's dollar at  58/60  (i.e. at  4s 10d : 5s).

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 36, pp. 20‑21 & 192.  Two countries each value the other's money at  90%  of its own.  The solution says this appeared in the New Statesman and Nation in late 1931, ??NYS.

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 9;  1945: 8‑9;  1961: 18‑19.  As in Phillips.

W. A. Bagley.  Paradox Pie.  Op. cit. in 6.BN.  1944.  No. 3: South of the border, pp. 8-9.  US & Mexico each value other's dollar at  $1.05.

John Fisher.  John Fisher's Magic Book.  Muller, London, 1968.  Magical shopping, pp. 120‑121.  North and South Fantasi value each other's  £  as  19s  (i.e.  95%)  of its own.

 

          7.AJ.            MULTIPLYING BY SHIFTING

 

            That is, we want solutions of   k * a1 ... am b1 ... bn  =  b1 ... bn a1 ... am.   E.g., for  m = 1,   3 * 142857  =  428571;  for  n = 1,   4 * 025641  =  102564,   4 * 102564  =  410256,  5 * 142857  =  714285.   If  1/n  has a repeating decimal of period  n-1,  then all multiples of it by  1, 2, ..., n-1  are shifts of it -- e.g.   1/7  =  .142857142857...   gives   2 * 142857  =  285714,   etc.  Such an  n  must have  10  as a primitive root.

            Early versions of the idea are simply observations of the properties of  1/7  etc.

            Note that two-digit problems, i.e.  m = n = 1,  look more like reversals and are considered in 7.AH.

 

Charles Babbage.  The Philosophy of Analysis -- unpublished collection of MSS in the BM as Add. MS 37202, c1820.  ??NX.  See 4.B.1 for more details.  F. 4r is "Analysis of the Essay of Games".  F. 4r has entry 13:  "What number is that whose  6  first multiples are the same digits differently arranged".

Ainsworth & Yeats.  Op. cit. in 7.H.4.  1854.  Exercise XXXVI, pp. 73 & 176.

No. 9:  k = 3,  m = 1,  a1 = 1,  n = 5.

No. 10:  k = 3,  m = 1,  a1 = 2,  n = 5.

Birger Hausted.  ??  Tidsskrift for Math. 2 (1878) 28.  ??NYS -- cited by Dickson, vol. 1, pp. 170‑171, item 81.  Studies problem with all shifts of the same number, but starting with the case  n = 1.  He finds that   k * a1...amB  =  Ba1...am   gives   B/(10k‑1)  =  (a1...amB) / (10m+1 ‑ 1).   He allows  k  to be rational.

Dickson, vol. 1, pp. 174‑179, items 101, 102, 106, 114, 120, 137, 150.

Hoffmann.  1893.  Chap. IV.

No. 15: A peculiar number, pp. 148 & 192‑193 = Hoffmann-Hordern, p. 120.   2 * 142857  =  285714,   which has  m = 2.

No. 49: A peculiar number, pp. 155 & 208 = Hoffmann-Hordern, p. 131.   5 * 142857  =  714285;   3 * 142857 =  428571;   6 * 142857  =  857142.

L. E. Dickson.  ??  Quarterly J. Math.  27 (1895) 366‑377.  [Item 106 above.]  Shows that all  ks  are integers and  a1 ¹ 0  only for  142857  (in base  10).

Anonymous note.  J. of the Physics School in Tokyo 6? (1897?) ??NYS  Abstracted in:  Yoshio Mikami, ed.; Mathematical Papers from the Far East; AGM 28 (1910) 21; as: Another queer number.   3 * 526 31578 94736 84210  =  1578 94736 84210 52630,   which is not quite a pure shift of the first number, but would be if the final zeroes were dropped.  Also multiplication by  4  or  8  or division by  5  give similar results.

T. Hayashi.  On a number that changes its figures only cyclically when multiplied or divided by any number.  J. of the Physics School in Tokyo 6 (1897) 148-149.  Abstracted ibid., p. 21.  Notes that the above number is based on the form   10 Σ (10r)i,   which explains and generalizes its properties.

U. Fujimaki.  Another queer number.  J. of the Physics School in Tokyo 6 (1897) 148-149.  Abstracted ibid., pp. 22-23.  Notes that the above number has the form   (10m - 1) / n,  which also explains and generalizes its properties.

Pearson.  1907.  Part II, no. 31: A large order, pp. 120 & 197‑198.  n = 1,  m = 21,  k  =  b1  =  7.

Schubert.  Op. cit. in 7.H.4.  1913.  Section 16, no. 186, pp. 51 & 137.  k = 3,  m = 1,  a1 = 1,  n = 5.

Peano.  Giochi.  1924.

Prob. 39, p. 10.  Notes cyclic property of  142857.

Prob. 40, pp. 10-11.  Notes cyclic property of   (1018 - 1) / 19,  says an English book calls them phoenix numbers and says similar properties hold for  (10n-1 - 1) / n   for   n = 7, 17, 19, 23, 29, 47, 59, 61, 97, 109, 130, 149, ....

Ackermann.  1925.  Pp. 107‑108.  k = 7,  n = 1,  b1 = 7.

W. B. Chadwick.  On placing the last digit first.  AMM 48 (1941) 251-255.  n = 1.  Shows   b1 ³ k   and   N  =  a1 ... am b1  =  b1 (10m+1 - 1) / (10k-1).  Finds all solutions for   k = 2, 3, ..., 9.  Cites Guttman, AMM 41 (1934) 159 (??NYS) for properties of these 'cyclic numbers'.  An editorial note adds that the expression for  N  works for any base.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  No. 76: Mathematical whiz, pp. 59-60.  Discusses cyclic properties of the period of  1/17,  though he doesn't identify it as arising from  1/17.

J. Bronowski.  Christmas Teasers.  New Statesman and Nation (24 Dec 1949).  ??NYS.  Case  m = 1,  k = 3/2.

Anonymous.  The problems drive.  Eureka 14 (Oct 1951) 12-13 & 22.  No. 11.  Same as Bronowski.  Solution:  428571. 

J. H. Clarke.  Note 2298:  A digital puzzle.  MG 36 (No. 318) (Dec 1952) 276.  Solves Bronowski's problem, leading to   3 * 10n  º  2 (mod 17)   for an  n+1  digit answer.

D. E. Littlewood.  Note 2494:  On Note 2298:  A digital puzzle.  MG 39 (No. 327) (Feb 1955) 58.  Easy solution of Bronowski's problem, via  20/17.

E. J. F. Primrose.  Note 2495:  A digital puzzle.  Ibid., 58‑59.  Case   m = 1,  k = p/q.

R. Sibson.  Note 2496:  On Note 2298.  Ibid., 59.  Simple but fortuitous solution via  20/17.

R. L. Goodstein.  Note 2600:  Digit transfers.  MG 40 (No. 332) (May 1956) 131‑132.  Shows Littlewood's method gives an easier solution to Primrose's version.

Gardner.  SA (Jan 1961) c= Magic Numbers, chap. 2.  Mentions   4*102564  =  410256   and the general problem with  n = 1  and  k = b1.  Gives some references to the most general problem in various standard works and some journal references up to 1968.

Barnard.  50 Observer Brain-Twisters.  1962.  Prob. 23: A safe number, pp. 30, 62 & 85‑86.  Wants the smallest number with  n = 1  and  k = b1,  i.e. such that   b * a1a2...amb  =  ba1a2...am.   Finds it is  102564.

Charles W. Trigg.  Division of integers by transposition.  JRM 1:3 (1968) 180-182.  Considers the case  n = 1,  k = b1,  i.e. the same as Barnard.  He gives the smallest answers for bases  3, 4, ..., 12  and  k = 2, ..., b-1.  Cites some earlier versions not given above.

Steven Kahan.  k-transposable integers.  MM 49:1 (1976) 27-28.  Studies the case  m = 1  and shows that then only  k = 3  can work.  The basic solutions are  142857  and  285714  and all other solutions are obtained by repeating these, e.g.  142857142857.

Steven Kahan.  k-reverse-transposable integers.  JRM 9:1 (1976-77) 15-20.  Studies the case  n = 1  and finds all solutions.

Birtwistle.  Calculator Puzzle Book.  1978. 

Prob. 67: Moving around, pp. 49 & 105-106.   4 * abcde4  =  4abcde,  i.e.  k  =  b1  =  4,  m = 5,  n = 1.  Notes properties of  142857.

Prob. 78: Change around, pp. 55 & 113.  k = 2,  n = 1,  b1 = 5,  formulated as a division ignoring remainder.  Finds   52631/2  =  26315,   which is a bit of a cheat, but he notes that this is part of the decimal expansion of  1/19  =  .0526315789473684210526...   which would give a proper answer.

Warren Page.  A general approach to p.q r‑cycles.  In:  Warren Page, ed.; Two‑Year College Mathematics Readings; MAA, 1981, pp. 263‑274.  He considers rational  k  (=  his  q/p)  and given  n  (=  his  r).  He also studies the case where multiples of   a1...amb1...bn   have the same properties.  He gives tables of all solutions for  n = 1  and  n = 2.  11 references to similar work.

Anne L. Ludington.  Transposable integers in arbitrary bases.  Fibonacci Quarterly 25:3 (1987) 263-267.  Considers case  m = 1.  Cites Kahan's results.  Considers base  g  and shows that there is some  k  if and only if  g = 5  or  g ³ 7.  Shows that there are only a finite number of basic solutions for any  k  and hence for any  g  and describes how to find them.

Anne L. Ludington.  Generalized transposable integers.  Fibonacci Quarterly 26:1 (1988) 58‑63.  Considers general case with shift of  j  (=  m  in my notation) and arbitrary base  g.  Shows there is such a  k  for all  j ³ 0  if and only if  g = 5  or  g ³ 7.  For  g = 3, 4, 6,  there is such a  k  for all  j ³ 2.  For fixed  j,  there are only a finite number of solutions.

Keith Devlin.  Better by degrees -- Micromaths column in The Guardian (17 Nov 1988) 31.  Let  F, C  be the same temperature in Fahrenheit and Centigrade (= Celsius).  When can  F = a1b1...bn,   C = b1...bna1?   First answer is  527.

 

          7.AJ.1.         MULTIPLYING BY APPENDING DIGITS

 

          New section.

 

Tony Gardiner.  Challenge! What is the title of this article?  Mathematics Review 4:4 (Apr 1994) 28-29.  Find an integer  A  such that adding digits of  1  at each end multiplies it by  99.  If  A  has  n  digits, then this gives us   10n+1 + 10A + 1  =  99A,   so  A  =  (10n+1 + 1) / 89.  He leaves the rest to the reader.

 

          7.AK. LAZY WORKER

 

            A worker earns  a  for each day he works and forfeits  b  for each day he doesn't work.  After  c  days, he has gained  d.  This gives   x + y = c,  ax ‑ by = d   or  (a+b)x - bc = d.   This is a straightforward problem, giving   x = (bc + d)/(a + b),  y = (ac - d)/(a + b),  so I only give some early examples.  There are also many different forms of problem which give the same equations. 

            NOTATION:  we denote this by  (a, b, c, d). 

            PROBLEM -- for which integral  (a, b, c, d),  is it true that  x, y  are integral?  One can generate all quadruples  (a, b, c, d)  with integral solutions as follows.  Choose any  a, b, c, x  and set  d = (a+b)x - bc.  This is not the kind of solution of the Problem that I'd like to have, but it may be best possible since we have only one relationship.

            I have now computed  x  in each case and include it in the table below.  One can scale  a, b, d  by any factor, so we eliminate fractions in  a, b, d  and even make  GCD(a, b) = 1,  but I haven't done any scaling in the table.

            For general solutions, see:  Hutton, 1798?;  Lacroix;  De Morgan.

            See Tropfke 603.

 

            a            b            c            d              x                  Sources

 

                     1          30          54              24                Hutton, c1780?;  Hutton, 1798?; 

                            24                       27/4              Hutton-Rutherford

                            24          78              19                Hutton-Rutherford

            5            3          12          28              8                  Lacroix; 

            5            3          28            0              21/2              Riese, 1524;

            5            6          30          12              192/11          Chuquet; 

            5            9          30          10              20                Columbia Alg.; 

            5            9          30          15              285/14          Gherardi; 

            5          12          30          99              27                Robinson

            6            5          30            0              150/11          Muscarello; 

                            50          18              1222/49        Unger; 

            7            3         365            0              219/2            Dodson

            7            4          30            1              11                Fibonacci;  Columbia Alg.; 

            7            4          30          30              150/11          Fibonacci; 

            7            5          30            0              25/2              Riese, 1522; 

            9          11          30            0              33/2              Bartoli

                  6⅓          70         180              40                Unger; 

          10            4          30         132              18                BR; 

      10/30        6/30          30           -2              15/2              al-Karkhi; 

      10/30        6/30          30            0              45/4              al-Karkhi; 

      10/30        6/30          30            4              74/4              al-Karkhi; 

          10          12          30            0              180/11          Benedetto da Firenze;  Calandri, 1491; 

          10          12          40            0              240/11          AR;  Wagner; 

          10          14          20          15              295/24          Tartaglia; 

          11            8          36          12              300/19          Riese, 1524

          12            8         365            0              146               Schott

          12            8         390            0              156               Wells;  Vyse

          15            5          60         240              27                Todhunter

          16            8          12         126              37/4              Eadon; 

          16          15          30            0              450/31          della Francesca 17r

          16          20          30            0              50/3              Eadon; 

          16          24          36            0              108/5            Tartaglia; 

          16          24          36          60              231/10          Tartaglia; 

          18          16          30            0              240/17          Pacioli; 

          20            8          40         372              173/7            Hutton, 1798?

          20            8          40         380              25                Simpson;  Bonnycastle

          20          10          40         500              30                Vyse

          20          16          30            0              40/3              della Francesca 38r

          20          28          40            0              70/3              Borghi; 

          20          28          40          30              575/24          Borghi; 

          24          12          48         504              30                Bourdon

          25          30          40          65              23                Recorde

          30          15          40         660              28                Ozanam-Montucla

          40            5          60         980              256/9            Les Amusemens; 

 

            A vaguely related, but fairly trivial problem, is the following.  A man offers to work for a master for a time  T  with payment of a horse (or cloak) and  M  money.  After time  t,  the man quits and is paid the horse and  m  money.  Letting  r  be the rate per unit time and letting  H  be the value of the horse, this gives  rT = H + M,  rt = H + m,  which gives  r(T-t) = M-m.  della Francesca f. 43v (107) (English in Jayawardene) is an early example.

 

Muhammad (the  h  should have an underdot) ibn Muhammad (the  h  should have an underdot) ibn Yahyā(the  h  should have an underdot) al-Bŭzağānī Abū al-Wafā’  = Abū al-Wafā’ al‑Būzajānī.  Arithmetic.  c980.  Arabic text edited by  Ahmad Salim Saidan;  Arabic Arithmetic; Amman, 1971.  P. 353.  ??NYS -- mentioned by Hermelink, op. cit. in 3.A.

al‑Karkhi.  c1010.  Sect. I, no. 12‑14, p. 83.  (10/30, 6/30, 30, d)  with  d = 0, 4, ‑2.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 113, no. 23.  ??NYS -- Hermelink, op. cit. in 3.A, gives al‑Karkhi's problem and then says  "This problem occurs also in Ţabarī [NOTE:  Ţ  denotes  T  with an underdot.]  ...."  Tropfke 603 gives the same reference.

Fibonacci.  1202.

Pp. 160‑161 (S: 250-251): De laboratore laborante in quodam opere [On the work of a labourer on a certain job].  (7, 4, 30, 30).

Pp. 323‑324 (S: 453-454): De laboratore, question notabilis [Notable problem on a worker].  (7, 4, 30, 1).

BR.  c1305.  No. 51, pp. 68‑69.  (10, 4, 30, 132).

Gherardi.  Libro di ragioni.  1328.  Pp. 48‑49.  (5, 9, 30, 15)

Columbia Algorism.  c1350.

Prob. 64, pp. 85‑86.  (5, 9, 30, 0).

Prob. 65, pp. 86‑87.  (7, 4, 30, 1).

Dresden C80.  ??NYS -- asserted in BR, p. 158.

Bartoli.  Memoriale.  c1420.  Prob. 7, ff. 75r - 75v (= Sesiano, pp. 137 & 147).  (9, 11, 30, 0).

AR.  c1450.  Prob. 183, pp. 86, 176, 222‑223.  (10, 12, 40, 0).

Benedetto da Firenze.  c1465.  P. 88.  (10, 12, 30, 0).

Muscarello.  1478.  Ff. 73v-74r, p. 188.  Worker building a house, (6, 5, 30, 0).

della Francesca.  Trattato.  c1480. 

F. 17r (64).  (16, 15, 30, 0).  Answer:  450/31  days worked.  English in Jayawardene.

F. 38r (99).  (20, 16, 30, 0).  Answer:  40/3  days worked.

Wagner.  Das Bamberger Rechenbuch, op. cit. in 7.G.1.  1483.  Von Tagelohn oder Arbeit, pp. 98 & 216.  (10, 12, 40, 0).  (= AR.)

Chuquet.  1484.

Prob. 51, English in FHM 209.  (5, 6, 30, 0).

Prob. 52.  (5, 6, 30, 12).

Borghi.  Arithmetica.  1484. 

Ff. 111v-112r (1509: f. 94r).  (20, 28, 40, 0).

Ff. 112r-113r (1509: ff. 94v-95r).  (20, 28, 40, 30).

Calandri.  Arimethrica.  1491.  F. 69v.  (10, 12, 30, 0).

Pacioli.  Summa.  1494.  F. 99r, prob. 11.  (18, 16, 30, 0).

Riese.  Rechnung.  1522.  1544 ed. -- pp. 89‑90;  1574 ed. -- pp. 60v‑61r.  (7, 5, 30, 0).

Riese.  Die Coss.  1524.

No. 37, p. 45.  (5, 3, 28, 0).

No. 110, pp. 54‑55.  (11, 8, 36, 12).

Recorde.  Second Part.  1552.  Pp. 312-318: A question of Masonry, the first example.  (25, 30, 40, 65).

Tartaglia.  General Trattato.  1556.  Book 17, art. 38, 39, 42, pp. 275r-275v & 277r.

Art. 38.  (16, 24, 36, 0).

Art. 39.  (16, 24, 36, 60).

Art. 42.  (10, 14, 20, 15).

Schott.  1674.  Ænigma 1, p. 558.

Wells.  1698.  No. 113, p. 208.  (12, 8, 390, 0).

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. VII, pp. 80-81 (1790: prob. XIX, p. 84).  (20, 8, 40, 380).

Les Amusemens.  1749.  Prob. 112, p. 254.  (40, 5, 60, 960).

Vyse.  Tutor's Guide.  1771? 

Prob. 21, 1793: p. 79; 1799: pp. 85-86 & Key p. 111.  (12, 8, 390, 0).

Prob. 23, 1793: p. 80; 1799: p. 86 & Key pp. 111-112.  (20, 10, 40, 500).

Prob. 15, 1793: p. 131; 1799: p. 139 & Key p. 184.  (12, 8, 390, 0)  solved a different way.

Dodson.  Math. Repository.  1775.  P. 9, Quest. XXI.  (7, 3, 365, 0).

Ozanam‑Montucla.  1778.  Prob. 8, 1778: 194-195;  1803: 192;  1814: 166‑167;  1840: 86.  (30, 15, 40, 660).  1790 has  620  for  660,  apparently a misprint (check if this is in 1778 -- ??).

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?]  1804: prob. 3, p. 135.  (2½, 1, 30, 54).

Bonnycastle.  Algebra.  1782.  Prob. 6, p. 80.  Same as Simpson.

Eadon.  Repository.  1794. 

P. 297, no. 14.  (16, 8, 12, 126).

Pp. 374-375, no. 20.  (16, 20, 30, 0).  The answer has days worked and days away reversed.

Hutton.  A Course of Mathematics.  1798? 

Prob. 2,  1833 & 1857: 80.  (2½, 1, 30, 54).

Prob. 7,  1833: 212;  1857:: 216.  (20, 8, 40, 372),  then done in general.

Prob. 46,  1833: 224;  1857: 228.  Man to pay his friend  a  for each shot he misses and to receive  b  for each hit.  After  n  shots,  he owes  c,  where  c  may be positive, zero or negative.

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Section 15, pp. 28-31.  (5, 3, 12, 28)  and general solution.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 47, prob. 2, pp. 64-65.  (24, 12, 48, 504)  and the general problem.  On pp. 102-104, he discusses the problem algebraically and the meaning of the signs involved.

D. Adams.  New Arithmetic.  1835.  P. 246, no. 105.  (.75, .25, 50, 27.50).  Notes that if he worked every day, he would earn  37.50  and that he loses  1.00  from this for every day he didn't work.

Augustus De Morgan.  On The Study and Difficulties of Mathematics.  First, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Workman earns  a  each day he works, but has expenses of  b  every day.  After  m  days, he has earned  c.  This is equivalent to  (a-b, b, m, c),  giving  ax - bm = c.  Gives general solution and discusses problems with negative values.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 41-45 & 248, nos. 176-190.  These cover a variety of problems leading to the same equations.

Pp. 41-42, no. 176.  (9¼, 6⅓, 70, 180).

P. 45, no. 190.  Purely arithmetic formulation of  (6½, 5¾, 50, 18).

Hutton-Rutherford.  A Course of Mathematics.  1841?.  Prob. 43,  1857: 43.  (4½, 1½, 24, 78)  and  (4½, 1½, 24, 4½).

Joseph Ray (1807-1855).  Ray's Arithmetic, Third Book.  Practical Arithmetic, by Induction and Analysis.  One Thousandth Edition -- Improved.  Wilson, Hinkle & Co., Cincinnati & New York, ©1857, HB.  P. 268, no. 57.  (2, 1, c, 25)  where he works three times as many days as he idles.

Todhunter.  Algebra, 5th ed.  1870.  Section X, art. 172, p. 82.  (15, 5, 60, 240).

Horatio N. Robinson.  New Elementary Algebra: Containing the Rudiments of the Science for Schools and Academies.  Ivison, Bargeman, Taylor & Co., New York, 1875.  This has several versions of the problem, but on p. 77, probs. 77-78 use the following novel formulation: a boy is to deliver some glass vessels and will be paid  a  for each success and will forfeit  b  for each breakage; after  c  vessels, he has earned  d.  Does  (5, 12, 30, 99)  and the general case.

 

          7.AL.           IF  A  IS  B,  WHAT IS  C?

 

            Generally, this is done by proportion of some sort.  In the simplest case -- if  a  is  b,  what is  c?  -- the answer is  cb/a.  In the more complex case -- if  ab = c,  what is  de = f?  -- the answer is usually  f = cde/ab.  Many problems have  e  and  f  given and ask for  d,  which is then  d = abf/ce.  However, some authors say that all answers should be multiplied by  c/ab  and hence give  d = cf/abe,  which I will call inverted reasoning or the inverted answer -- see:  Benedetto da Firenze, p. 65;  The Sociable (& Book of 500 Puzzles);  Lemon;  Hoffmann;  Pearson;  Loyd.

 

Fibonacci.  1202.  P. 170 (S: 264).  He discusses this and says that it is just a proportion which is commonly stated in this way.  E.g. if  5  is  9,  what is  11?  He says  99/5.  Also, if  7  is  half  12,  what is half of  10?  He says  35/6.

Gherardi.  Libro di ragioni.  1328.  P. 17: Questi sono numeri.  "If  9  is the half of  16,  what part is  12  of  25?"  Answer is  8/9  of  12/25.

Bartoli.  Memoriale.  c1420.

Prob. 5, f. 75r (= Sesiano, pp. 137 & 147).  If  7  is  1/8  of  49,  what is  1/3  of  57?  He seems to say: "If  7  becomes  6 1/8,  what becomes  19?"  but he computes the inverse result.

Prob. 6, f. 75r (= Sesiano, pp. 137 & 147).  If  3  times  6  makes  17,  what will  7  times  8  be?  Obtains the normal answer.

AR.  c1450.  Prob. 303, 305, pp. 135‑136, 178, 225.

303:  If  4  is  ½  of  10,  what is    of  24?   Answer:  6 2/5  =  8 * 4/5.

305:  If  3 * 3  is  10,  what is  4 * 4?   Answer:  17 7/9  =  16 * 10/9.

Vogel says these problems also occur in Widman, 1489, ??NYS, and versions are in al‑Khowarizmi.  However al‑Khowarizmi's examples are straightforward rules of three, e.g.  "If you are told 'ten for six, how much for four?'...."

Benedetto da Firenze.  c1465.

P. 64.  If  3 * 3  =  10,  what is  10 * 10  by the same rule?  Answer:  100 * 10/9.

P. 65.  If  3  is half of  7,  what is the half of  7?   Answer:  7/6 * 7/2, by inverted reasoning.

The Treviso Arithmetic = Larte de labbacho.  Op. cit. in 7.H.  1478.  Ff. 31r‑33v (= Swetz, pp. 103‑109).  Three examples in abstract rule of three, e.g.  "If  8  should become  11,  what would  12  become?"   Answer:  12 * 11/8.

Chuquet.  1484.  Triparty, part 1.  FHM 73-74 gives a number of problems which are treated as proportions.

"If  3  times  4  will lead to  9,  what will  4  times  5  lead to?  ...  If  12  is worth  9,  what is  20  worth?" 

"If  7  is the  ½  of  12,  what is    of  9?"   Answer:  3½.

"If  2/3  will be  3/4  of  4/5,  what will be the  5/6  of  6/7?"   Answer:  50/63.

"If  7  were the  ½  of  12  one asks what part    would be of  9."   Answer:  3½.

"If  2/3  were the  3/4  of  4/5,  one asks what part would  50/63  of  6/7  be."  Answer:  5/6.

Pacioli.  Summa.  1494.

Ff. 99r-99v, prob. 13.  If  ½  of  5  is  3,  when is  ¼  of  x  equal to  5.  Gets  x = 16⅔.  Then does:  If  4  is  6,  what is  10?  He gets  (6/4)10  =  15  and then says that if one rephrases it, then one gets  (4/6)10  =  6⅔.  Then mentions:  If  ½  of  7  is  3,  what is    of  9?  -- no answer given.

F. 99v, prob. 14.  If  1/3  is  1/2  of  1/5,  when is  1/5  of  1/x  equal to  1/4?  The first statement says  1/5 = 2/3,  so  2/3x = 1/4,  giving  x = 8/3,  which is the direct answer.

F. 99v, prob. 15.  If    is  ½  of that number of which  5  is  ⅔,  of what number is  3  the  ½?  Takes  3/2  of  5  =    and then half of that, i.e.    and says  "If    is  3¾,  what is  3?"  Gets  45/14  and doubles it.

F. 99v, prob. 18.  If  3  is the  ½  of  7,  what part of  11  is  4?  Says "if  3  is  3½,  what is  4?"  and gets  14/3  which is  14/33  of  11.  Says this may not be the right way to do it.

Tonstall.  De Arte Supputandi.  1522.  Pp. 223-224.

If  4  is  6,  what makes  10?   (4/6)10  =  6⅔.

If  ½  of  5  is  3,  when is  ¼  of  x  equal to  5.  Gets  16⅔.

If  ½  of  7  is  3,  what part of  11  is  4?  His method would give  33/14,  but he has  33/13  due to an error.

Dilworth.  Schoolmaster's Assistant.  1743.  P. 157, no. 3.  "If the    of  6  be  3,  what will  ¼  of  20  be?"  Answer:  7½,  which is the direct answer.

Les Amusemens.  1749.  P. xxv.  See 7.AN for a problem that looks like it belongs here.

Walkingame.  Tutor's Assistant.  1751.  1777: p. 172, prob. 48;  1835: p. 178, prob. 27;  1860: p. 180, prob. 47.  Identical to Dilworth.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 2, pp. 15 & 71.  "If the half of five be seven, What part of nine will be eleven?"  Answer:  55/126  =  (5/2)(11/7)/9.

The Sociable.  1858.  Prob. 46: A dozen quibbles: part. 7, pp. 300 & 319.  "If  5  times  4  are thirty-three, what will the fourth of twenty be?"  Answer is    with no explanation, which is the inverted answer.  = Book of 500 Puzzles, 1859, prob. 46: part 7, pp. 18 & 37.  = Magician's Own Book (UK version), 1871, Paradoxes [no. 2], p. 37.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Arithmetical puzzles, no. 7, p. 174 (1868: 185).  "If six the third of twenty be, what is the fourth of thirty‑three?"  Gets   (6 * 33/4) / (20/3)  =  7 17/40.

Lewis Carroll.  Alice in Wonderland.  1865.  Chap. II.  In:  M. Gardner; The Annotated Alice; revised ed., Penguin, 1970, p. 38.  "Let me see:  four times five is twelve, and four times six is thirteen, and four times seven is -- oh dear!  I shall never get to twenty at that rate!"  Gardner's note 3 gives various explanations, the simplest of which is that  

          4 * 12  =  19   and Alice only knows tables up to 12 times.  Cf John Fisher; The Magic of Lewis Carroll; op. cit. in 1; pp. 34-35 & Carroll-Wakeling, prob. 7: Alice's multiplication tables, pp. 8-9 & 64-65.

Lemon.  1890.  Quibbles, no. 254(b), pp. 37 & 107 (= Sphinx, no. 453(b), pp. 63 & 113.)  "If five times four are thirty‑three, what will the fourth of twenty be?"  Inverted answer of    with no explanation -- see The Sociable.

Hoffmann.  1893.  Chap. IX, no. 34: A new valuation, pp. 320 & 327 = Hoffmann-Hordern, p. 212.  Identical to Lemon.  Answer is  8¼,  but he gives no reason.

Clark.  Mental Nuts.  1897, no. 76;  1904, no. 77;  1916, no. 73.  Suppose.  "Suppose the one-fourth of twenty was three, what would the one-third of ten be?"  Normal answer of  2.

H. D. Northrop.  Popular Pastimes.  1901.  No. 10: A dozen quibbles, no. 7, pp. 68 & 73.  = The Sociable.

Pearson.  1907.  Part II, no. 95, pp. 134 & 210.  Same as Lemon.

Loyd.  Cyclopedia.  1914.  Sam Loyd's perplexed professor, pp. 332 & 383.  = SLAHP: If things were different, pp. 56 & 106.  "If five times six were  33,  what would the half of  20  be?"  Answer:  "If five times six is  33  -- ten would naturally be  1‑3  of what  30  would be, viz:  11."  This is the inverted answer.  Loyd Jr. gives a bit more explanation.

Loyd.  Cyclopedia.  1914.  p. 317 (no solution).  As in Lemon.

Perelman.  MCBF.  1937.  Imaginary nonsense, prob. 143, pp. 243-244.  "What is the number  84  if  8 x 8  is  54?"  Finds that the base is  12,  so the answer is   8412  =  100.

Haldeman-Julius.  1937.  No. 63: The-what-is-it-problem, pp. 9 & 24.  "If one third of six be three, what will one third of 29 be?"  Answer is  10,  with no explanation.  This does not fit into either of the standard versions here, but would be the usual form if  29  were a misprint for  20.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.

Number, please! -- no. 3, pp. 95 & 214.  "If five times  8  made  60,  what would a quarter of  40  be?"  Answer is  15,  which is the inverted answer.

Is your brain working? -- no. 3, pp. 148 & 215.  "If five times four made thirty three,  What would a fifth of fifth be?"  Gives the answer  15½,  which is described as half of  33!  I.e. the answer is intended to be the inverted answer  16½.

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  P. 8: Mathematical "if".  "If a fourth of forty is six;  what is a third of twenty?"  Says four, basically by saying the result is  6/10  of the real result.

G. A. Briggs.  Puzzle and Humour Book.  Published by the author, Ilkley, 1966.  Prob. 1/9, pp. 12 & 71.  If a third of six were three, what would the half of twenty be?"  Answer is  15,  which is the normal answer, but he gives no reason.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  If, pp. 28 & 109.  "If a third of six was three, what would a quarter of twenty be?"  Takes  3/2  of the correct answer, as in the usual method.

 

          7.AM.          CROSSNUMBER PUZZLES

 

          New section -- there may be older examples, but see Barnard, 1963.

 

Dudeney.  Strand Magazine (1926).  ??NYS -- cited by Angela Newing.

Richard Hoadley Tingley.  Mathematical Cross Number Puzzle[s].  In:  S. Loyd Jr., ed.; Tricks and Puzzles; op. cit. in 5.D.1, 1927.  Pp. 103‑105 & Answers pp. 11-12.  Three examples with  13 x 13  frames.  Tingley's name is only on the first example.  "This puzzle is radically different from the usual type ....  We have named this new brain teaser "Cross Number Puzzle" ...."  I found several errors in the second example.

Dudeney.  PCP.  1932.  Prob. 175: Cross‑figure puzzle, pp. 48 & 148.  11 x 11  frame.  Erroneous set of clues and solution.  Corrected as Cross-number puzzle in the revised ed. of 1935?, pp. 48‑49 & 148.

Michael H. Dorey.  "Little Pigley" [or "Little Pigsby"].  1936.  This is also called "Dog's Mead" -- original ??NYS.  A 1939 version with this attribution and date are given in:  Tim Sole; The Ticket to Heaven and Other Superior Puzzles; Penguin, 1988, pp. 92 & 108.  A 1935 version is given in:  Williams & Savage, 1940, below.  A 1936 version is given in:  Philip Carter & Ken Russell; Classic Puzzles; Sphere, London, 1990, pp. 62-63 & 128.  A 1939 version is given as The Little Pigsby Farm Puzzle in:  David Ahl & Burchenal Green, eds.; The Best of Creative Computing, vol. 3; Creative Computing Press, Morristown, NJ, 1980, p. 177; no solution.

Phillips.  Brush.  1936.  4 x 4  numerical crosswords.

Prob. H.5, pp. 26 & 91.

Prob. J.5, pp. 35 & 96.

Prob. T.3, pp. 68‑69 & 115.

Jerome S. Meyer.  Fun for the Family.  (Greenberg Publishers, 1937);  Permabooks, NY, 1959.  No. 28: Family skeleton, pp. 41 & 240.  4 x 4  based on ages in a family.

Haldeman-Julius.  1937.  No. 132: Family skeleton problem, pp. 15 & 27.  Same as Meyer.  He says he got it from the Feb 1937 issue of College Humor.

E. P. H[icks].  &  C. H. B.  A mathematical crossword.  Eureka 1 (Jan 1939) 17  &  2 (May 1939) 28.  5 x 5  diagram with central square black and heavy division lines forming 18 lights.  Two verticals are four digit numbers and the excess single digits do not have down clues.  Some of the clues are pretty obscure -- e.g. 'Magazine without the printers' refers to the magazine Printers' Pie, so the solution is  3142.

S. E. W[ood].  A numerical square.  Eureka 3 (Jan 1940) 18.  6 x 6  array with some heavy division lines giving 18 lights.  I have not found a solution in this or following issues.

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.

No. 18: Mr Turtle, pp. 17 & 108.  4 x 4  square with  11  clues.

No. 51: Little Pigley Farm, 1935, pp. 32-33 & 119.  7 x 7  square with  21  clues and some more in the lead-in.

Anonymous.  Crosswords  Five hour or five minute puzzles.  Eureka  4 (May 1940) 16  &  5 (Jan 1941) 15.  8 x 8 array with 26 lights.

C. A. B. Smith.  A new way of writing numbers.  Eureka  5 (Jan 1941) 7-9  &  6 (May 1941) 11.  General exposition of the negative digit system -- see section 7.AA.  Discusses the method for base 6 and gives a  7 x 6  crossnumber puzzle in this system.

E. M. White.  A crossword in decimal.  Eureka  5 (Jan 1941) 20  &  6 (May 1941) 11.  9 x 9  array with  26  clues.

M. A. Porter.  Note 1982:  The missing clue.  MG 31 (No. 296) (Oct 1947) 237.  4 x 4  puzzle.

Anonymous examples in Eureka.

12 (Oct 1949) 4  &  13 (Oct 1950) 21.  6 x 6  array with 16 clues which are algebraic expressions in 10 unknowns.

The problems drive.  12 (Oct 1949) 7-8 & 15.  No. 2: Crux verbum.  4 x 4  array.  The simple vertical clues use Roman numerals and the solutions are Roman numerals using four symbols.  The horizontals are then proper Roman numerals and the problem is to give the Arabic forms of these.

13 (Oct 1950) 15  &  14 (Oct 1951) 23.  11 x 11  array with 21 clues.  This uses mathematical symbols as well as numbers, e.g. an answer is  b2+4=0,  where  b2  is one 'letter'.

14 (Oct 1951) 6  &  15 (Oct 1952) 16.  6 x 6  array with heavy division lines giving 24 lights, the clues being expressions in 25 variables (the alphabet omitting O).

15 (Oct 1952) 16.  No solution found.  3 x 3 x 3  version by 'Nero'.  Every  3 x 3  section contains all nine positive digits.  24 of the cells are labelled with the letters  a - y,  omitting  o.  Clues are of the form  abc = amw,  beh = ihg - fed,  etc.

17 (Oct 1954) 13.  No solution found.  Cross shaped, with four  2 x 3  rectangles on the edges of a  3 x 3.  12 lights with clues being expressions in the twelve values.  By 'Pythagoras'.

The Problems drive, 1956.  19 (Mar 1957) 12-14 & 19.  No. 2.  3 x 3  array with six clues, two each of 'prime number', 'perfect cube', 'perfect square'.

Problems drive, 1957.  20 (Oct 1957) 14-17 & 29-30.  No. 7.  4 x 4  array with eight clues, some going backward.

A. H. Barrass.  Numerical square.  Eureka 22 (Oct 1959) 13 & 22.  5 x 5  array with division lines giving 17 lights.  Each answer has the form  x2 ± y  where  x  is a positive integer and  y  is a positive prime.  Complex clues for the  x  and  y  values. 

Philip E. Bath.  Fun with Figures.  Op. cit. in 5.C.  1959.  No. 100: A number crossword, pp. 38 & 61.  5 x 5  array with  17  clues.

M. R. Boothroyd  &  J. H. Conway.  Problems drive, 1959.  Eureka 22 (Oct 1959) 15-17 & 22-23.  No. 5.  2 x 2 x 2  to be filled with eight distinct digits.  Eight clues saying the value is a square or half a square (where this can be the integer part of half an odd square).

G. J. S. Ross  &  M. Westwood.  Problems drive, 1960.  Eureka 23 (Oct 1960) 23-25 & 26.  Prob. C.  3 x 3  array with two opposite corners deleted, with six clues in base 7.

B. D. Josephson  &  J. M. Boardman.  Problems drive 1961.  Eureka 24 (Oct 1961) 20-22 & 24.  Prob. J.  3 x 3  array with opposite corners omitted, with six clues being algebraic expressions in four positive integer variables.

During 1960-1980, R. E. Holmes, "Rhombus", contributed 45 puzzles to The Listener.  Some, perhaps all, of these were formidable crossnumber puzzles.  I have been sent three examples of these, but there are no dates on them.  He also contributed at least one example to G&P, but my copy has no date on it.  Can anyone provide information about these puzzles or the setter?

P. E. Knight.  Oh, How I love thee, Dr. Pell.  In:  H. Phillips; Problems Omnibus II; Arco, London, 1962; pp. 163‑164 & 228‑229.  3 x 4  array, but complex, based on Pell's equation.

Barnard.  50 Observer Brain-Twisters.  1962.  Prob. 40: Crossnumber, pp. 46‑47, 65 & 98‑99.

D. St. P. Barnard.  Anatomy of the Crossword.  Bell, London, 1963.  On p. 30, he says  "The crossword ... has given rise recently to the Crossnumber Puzzle"  and he refers to his book above.

L. S. Harris  &  J. M. Northover.  Problems drive 1963.  Eureka 26 (Oct 1963) 10-12 & 32.  Prob. H.  3 x 3  array with centre omitted.  Clues are  9x2, 4x2;  4x2y, (3x+y)2,  where  x, y  are integers. 

Birtwistle.  Math. Puzzles & Perplexities.  1971.

Cross-number puzzle 1, pp. 91 & 192.  8 x 8  grid,  32  clues.

Cross-number puzzle 2, pp. 92 & 192.  7 x 7  grid,  24  clues.

Cross-number puzzle 3, pp. 93 & 192-193.  9 x 9  grid,  36  clues.

Cross-number puzzle 4, pp. 94 & 193.  8 x 8  grid,  26  clues.

Birtwistle.  Calculator Puzzle Book.  1978. 

Prob. 15, pp. 14-15 & 75.  6 x 6  grid,  16  clues.

Prob. 33, pp. 24-25 & 82.  8 x 8  grid,  28  clues.

Prob. 63, pp. 46-47 & 103.  8 x 8  grid,  20  clues.

Prob. 90, pp. 63-64 & 123.  6 x 6  grid,  16  clues relating to a story.

K. Heinrich.  Zahlenkreuzrätsel.  In:  Johannes Lehmann; Kurzweil durch Mathe; Urania Verlag, Leipzig, 1980; pp. 38 & 138.  6 x 6  grid with  24  clues.

 

          7.AN.          THREE ODDS MAKE AN EVEN, ETC.

 

Alcuin.  9C.  Prop. 43: Propositio de porcis.  Kill  300  (or  30)  pigs in three days, an odd number each day.  "[Haec ratio indissolubilis ad increpandum composita est.]  ...  Ecce fabula ....  Haec fabula est tantum ad pueros increpandos."  ([This unsolvable problem is set to cause confusion.]  This is a fable ....  This fable is posed to confuse children.)

Pacioli.  De Viribus.  c1500. 

Ff. 92v - 93v.  XLVII. C(apitolo). de un casieri ch' pone in taula al quante poste de d(ucati) aun bel partito (Chap. 47. of a cashier who placed on the table some piles of ducats as a good trick).  = Peirani 132-133.  Place four piles each of 

1, 3, 5, 7, 9  ducats.  Ask the person to take  30  ducats in  5  piles.  If he can do it, he wins all 100 ducats.  Discusses other versions, including putting  20  pigs in  5  pens with an odd number in each.  However, the Italian word for  20,  i.e.  vinti,  written  uinti,  can be divided into five parts as  u i n t i,  and each part is one letter.  Cites Euclid IX: 23.

Ff. 93v - 94r.  XLVIII. C(apitolo). ch' pur unaltro pone al quante altre poste pare bel partito (Chap. 48. about another who placed some other even piles, good trick).  = Peirani 133-134.  Place four piles each of  2, 4, 6, 8, 10  carlini (a small coin of the time) and ask the person to take  31  carlini in  6  piles.  Cites Euclid IX: 23.

F. IIIr.  = Peirani 6.  The Index lists the above as Problems 50 & 51 and lists Problem 52: Del dubio amazar .30. porci in .7. bote disparre (On the dubious placing of 30 pigs in 7 odd pens).

Part 3, F. 281v, no. 133: Dimme come farrai a partir vinti in 5 parti despare (Tell me how to divide 'vinti' into five odd parts) = Peirani 407.  Divides as  v.i.n.t.i,  and mentions dividing 20 into 7 pens.

W. Leybourn.  Pleasure with Profit.  1694.  ??NYS -- described in Cunnington, op. cit. in 7.G.2, 1904, p. 151  and  in De Morgan, Rara, p. 633.  "How can you put five odd numbers to make twenty?"  "Write three nines upside down and two ones."  De Morgan says he does not recall ever seeing this problem, that Leybourn considers the answer a fallacy, but that he thinks "the question more than answered, viz. in very odd numbers."

Les Amusemens.  1749. 

P. xxv. 

Quatre fois trois font quinze, il n'en faut rien rabattre;

Neuf cing et un font douze, et rien de surplus;

Deux sept et six font treize, et sur ce je conclus

Par ce juste calcul que tout ne fait que quatre.

            Solution is to take the number of letters in the words -- but  'et'  is printed like  '&'  which makes it a bit hard to recognize it as a two-letter word.

P. 52.  1o:  Exprimer un nombre pair par  3  impairs.  General solution:   a a/a  =  a+1.  Examples:  7 7/7,  21 21/21.  Pp. 53-54 give other problems but they are not relevant, e.g.  33 3/3  -- see entry under 7.I.

Philip Breslaw (attrib.).  Breslaw's Last Legacy.  1784?  Op. cit. in 6.AF.  1795: 78-81.  'How to rub out Twenty Chalks at five Times rubbing out, every time an odd one.'  Set out marks numbered  1  to  20.  Then erase the last four, which are those starting at  17,  which is odd, etc. 

Henri Decremps.  Codicile de Jérôme Sharp, ....  Op. cit. in 4.A.1.  1788.  Avant Propos, pp. 20-21.  Statement is the same as in Breslaw, but solution is not given.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Parlour magic, no. xx, p. 202 (1868: 203): How to rub out twenty chalks in five rubs, each time erasing an odd number.  "Begin at the bottom and rub out upwards, four at a time."  See Breslaw for clarification.

Magician's Own Book.  1857.  How to rub out twenty chalks at five times, rubbing out every time an odd one, p. 239.  = Boy's Own Conjuring Book, 1860, How to rub twenty chalks at five times rubbing out, every time an odd one, pp. 205‑206.  Cf Breslaw.

Magician's Own Book (UK version).  1871.  The Arabian trick, p. 313.  "To take up twenty cards, at five times, and each time an odd-numbered one", he lays out twenty cards, 1 ‑ 10, 1 - 10  (considered as  1 - 20)  and proceeds as in Breslaw.

Don Lemon.  Everybody's Pocket Cyclopedia.  Revised 8th ed., 1890.  Op. cit. in 5.A.  P. 136, no. 3.  "Place  15  sheep in  4  pens, so there will be the same number of sheep in each pen."  Though not of the same type as others in this section, and no solution is given, I think the solution is similar to other solutions here, namely to put the pens concentrically around the inner pen and put all the sheep inside the inner pen.

Pearson.  1907.  Part II.

No. 111, pp. 137 & 213.  Five odd figures to make  14.  Gives   1 + 1 + 1 + 11  =  14   and   1 + 1 + 1 + 1  =  4  with another  1  makes  14.

No. 115: What are the odds?, pp. 137 & 214.  Place  20  horses in three stalls with an odd number in each.  Puts  1, 3, 16  and says  16  is "an odd number to put into any stall".

Part III, no. 61: The shepherd's puzzle, p. 61.  Put  21  sheep in  4  pens, an odd number in each.  Uses concentric pens.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 27:1 (Nov 1911) 89: The Ass-tute farmer.  Farmer has 17 asses and a friend bets he can't put an odd number in each of four stalls.  Farmer puts 7, 5, 3, 2.  After inspecting the first three stalls, the friend says the farmer has lost, but the farmer says to go into the last stall to check, whereupon he shuts and locks the door, announcing there are three asses in the stall!

Loyd.  Cyclopedia.  1914.  The pig sty problem, pp. 37 & 343.  = MPSL2, prob. 7, pp. 6‑7 & 123.  = SLAHP: Pigs in pairs, pp. 51 & 104.  = Pearson's shepherd's puzzle.

Loyd.  Cyclopedia.  1914.  A tricky problem, p. 38.  = SLAHP: Torturing Dad, pp. 75 & 115.  Five odd figures to make  14.  Gives Pearson's first solution.

Smith.  Number Stories.  1919.  Pp. 126 & 146.  Put  10  pieces of sugar in three cups so each cup has an odd number.  Put  7  &  3  and put one cup inside another cup.

Blyth.  Match-Stick Magic.  1921.  The twenty game, p. 79.  As in Breslaw, etc., but more clearly expressed:  "The matchsticks have now to be removed in five lots, ....  Each time ... the last of the group must be an odd number."

Hummerston.  Fun, Mirth & Mystery.  1924.  Puzzle no. 64, pp. 149 & 182.  "How would you arrange twenty horses in three stalls so as to have an odd number of horses in each stall?"  Arranges as  1, 3, 16 -- "sixteen is a very odd number of horses to put into any stall."

Wood.  Oddities.  1927.  Prob. 50: The lumps of sugar, pp. 42-43.  Ten lumps of sugar into three cups so each cup contains an odd number of lumps.  Confusing solution, but lets one cup be inside another and lists all 15 possible solutions.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  Number, Please!, pp. 20 & 210.  Use the same odd figure five times to make  14.   1 + 1 + 1 + 11.

McKay.  Party Night.  1940.  Pill-taking extraordinary, p. 152.  "A man had a box holding  100  pills.  He took an odd number of pills on each of the seven days of the week, and at  the end of the week all the pills were gone.  How could he manage that?"  1  on each of the first  6  days and  94  on the last -- "you must admit that  94  is a very odd number of pills to take on any day."

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 17: Pigs and pens, pp. 27 & 184.  Put nine pigs in four pens, an odd number in each pen.  Three pens of three with a big pen around them all.

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  P. 19: Pigs in pens.  Same as Meyer.

John Paul Adams.  We Dare You to Solve This!.  Op. cit. in 5.C.  1957?

Prob. 40: A pen pincher, pp. 24 & 40.  9  pigs in four pens, an odd number in each.  Three pens of three, contained in one large pen.

Prob. 169, pp. 60-61 & 120.  21  lambs in four pens, each with an odd number.  Same solution idea as in prob. 40.

Gibson.  Op. cit. in 4.A.1.a.  1963.

P. 68: Odd figuring.  Use seven odd figures to add up to  20.  Answer:  13 + 3 + 1 + 1 + 1 + 1.

P. 69 & 74: Cross-out groups.  Make a row of  20  labelled marks and cross out marks in consecutive groups so that the last mark is odd each time.  Easy -- just mark backward!

Doubleday - 1.  1969.  Prob. 71: Ups and downs, pp. 87 & 170.  = Doubleday - 5, pp. 97-98.  Labourer has to carry 85 bricks up to a bricklayer in a hod that can hold 16, but he is instructed to always carry an odd number.  How does he do it in six journeys?  Solution is to carry up 15 on five trips and then bring one down and carry up the remaining 11.  To me, the return stages should all be alike and one could do this by carrying up and carrying down one five times, then carrying up the remaining 15.  Of course, the last stage cannot be expected to be the same as the others -- indeed he may be sent off on some other task.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  Three short problems -- no. 1, pp. 41-42.  Have five odd numbers add to  14.   11 + 1 + 1 + 1  =  14.

 

          7.AO. DIVINATION OF A PERMUTATION

 

            There are simpler versions of this used to divine three numbers, e.g. to locate a ring on a person, finger and digit -- a common one uses:  x *2 +5 *5 +10 +y *10 +z  =  350 + 100x + 10y + z  -- cf Fibonacci, Tagliente (1515).  The operations are performed from left to right, corresponding to the instructions given.  I will make no attempt to trace this very common but very dull type of problem, but see 7.M.4.b, especially the Folkerts entry which cites six early sources for this type of divination.

            If the  ai  are a permutation of  1, 2, ..., n,  the method of interest forms   P  =  a1m1 + a2m2 + ... anmn   for multipliers  mi.  By appropriate choice of the  mi,  the value of  P  determines the permutation.  Generally  P  is subtracted from some convenient constant.  Sometimes the solution uses a division to yield  a1  and  a2.  Some formulae work even if  ai  are not a permutation, but are digits or dice values.  If we have a permutation, one can ignore  an  since it is determined by the others, i.e. one can let  mn = 0.

            NOTE.  The standard form of this problem has  n  objects permuted among  n  people.  The permutation  (ai)  can be viewed in two ways.

            1.         The more natural view is that  ai = j  means the  i-th person has object  j.  Then the multipliers are associated with the people.

            2.         The inverse view is easier to implement and hence much more common in this problem.  We view  ai = j  as meaning that the  i-th object is with person  j.  Then the multipliers are associated with the objects.

            The most common form of the problem is with three people and  24  counters.  You give the people  1, 2, 3  counters to start and leave the other  18  counters on the table.  Let them secretly permute three items, say  A, B, C  among themselves.  You tell the person with object  A  to take as many counters as he already has; the person with object  B  to take twice as many; the person with object  C  to take four times as many.  In our inverse view, we are taking  P = a1 + 2a2 + 4a3  from  18.

            NOTATION:   M  =  (m1, m2, ..., mn)   denotes this process.

            A sequence of multipliers,  M,  is suitable if the products  P  are all distinct.  In the case that the  (ai)  are a permutation, it is easy to see that the following processes preserve the suitability of  M.  (1) Permuting the  mi.  (2)  Shifting all the  mi  by a constant.  (3) Multiplying all the  mi  by a non-zero constant.  Thus we can arrange the  mi  in ascending order and make  m1 = 0  and  m2 = 1.  So for the case  n = 3,  any sequence can be brought to the form  (0, 1, m3).  By subtracting from  m3  and scaling, we see that this sequence is equivalent to  (0, 1, m3/(m3-1)).  A little work shows that either  m3 ³ 2  or  m3/(m3-1) ³ 2,  so we can assume  m3 ³ 2.   m3 = 2  gives an unsuitable  M,  but  m3 > 2  always gives a suitable  M.  So the simplest possible case is  (0, 1, 3),  which is equivalent to the most common case  (1, 2, 4)  and to  (3, 4, 6).  Cases equivalent to  (0, 1, 4),  (0, 1, 6)  and  (0, 1, 8)  also occur.

            For the case of permutations, subtracting  P  values from some constant  S  is equivalent to changing  mi  to   S/Σai  -  mi.

            As in the Josephus Problem, mnemonics were constructed.  For the case  n = 3,  objects were labelled by the vowels  a, e, i  and mnemonics were constructed consisting of words (or phrases) with these three vowels in all six permutations (or having just the first two vowels of each permutation).  See Bachet for an example.  Gardner also gives mnemonics using consonants.

            See Meyer for a slightly more complex multiplication process, which can be reduced to the usual form.

            See Tropfke 648‑651.

 

                        COMMON VOWEL MNEMONICS -- note that spelling and layout vary.

 

Angeli, Beati, Taliter, Messias, Israel, Pietas.  Baker;

Angeli  Beati  Pariter  Elias  Israel  Pietas.  Hunt.

Anger,  fear,  pain, may  be hid  with a  smile.  Magician's Own Book; Boy's Own Conjuring Book;

Aperi,  Premati,  Magister,  Nihil,  Femina,  Vispane,  Vispena.  Minguét.

Aperì  Prelati  Magister  Camille  Perina  Quid habes  Ribera.  Alberti 76-77;

Avec  éclat  L'Aï  brillant  devint  libre.  Labosne, under Bachet; Lucas;

Brave  dashing  sea, like a  giant  revives  itself.  Magician's Own Book; Boy's Own Conjuring Book;

Graceful  Emma,  charming she  reigns  in all  circles.  Magician's Own Book; Boy's Own Conjuring Book;

Il a  jadis  brillé  dans ce  petit  État.  Lucas;

James  Easy  admires now  reigning  with a  bride.  Magician's Own Book; Boy's Own Conjuring Book;

Pallētis  Evandri  Sanguine  Feritas  Imane (the  m  should have an overbar) Vigebat.  Schott.

Par fer  César  jadis  devint  si grand  prince.  Bachet; Ozanam 1725, prob. 46; Alberti 76-77; Les Amusemens; Hooper; Magician's Own Book; Boy's Own Conjuring Book; Boy's Own Book; Magician's Own Book (UK version); Lucas;

Pare ella ai segni; Vita, Piè.  Alberti 76-77.

Salve  certa  anima  semita  vita  quies.  Van Etten; Schott (with one respelling); Ozanam 1725, prob. 46; Alberti 76-77; Manuel des Sorciers; Endless Amusement; Parlour Pastimes; Magician's Own Book; Boy's Own Conjuring Book; Magician's Own Book (UK version);

Take her  certain  anise  seedlings  Ida  quince.  Magician's Own Book (UK version);

 

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  P. 109, part IV, no. 17.  ??NYS - described by Tropfke 648.  As in Fibonacci, below, with  a = 18  and  M = (2, 17, 18).

Fibonacci.  1202. 

Pp. 304-306 (S: 429-431).  Several simple divinations, e.g.  x*2+5*5+10+y*10+z  =  350 + 100x + 10y + z.

Pp. 307‑308 (S: 433-434).  If   a1 + a2 + a3  =  a,   he uses  M = (2, a‑1, a)  and subtracts from  a2  to obtain   (a‑2) a1 + a2.   He also uses  M = (2, k, k+1)  and subtracts from a(k+1)  to get  (k‑1) a1 + a2.   These work whether the  ai  are a permutation or not.  He uses  a = 6, k = 9,  hence  M = (2, 9, 10).  For permutations (?) of  2, 3, 4,  he uses  a = 9, k = 9.  No mnemonics.

Abbott Albert.  c1240.  Prob. 2, pp. 332‑333.  M = (2, 9, 10).  He then subtracts from  60  which produces  8a1 + a2.  This is identical to Fibonacci's first example, but Albert gives a complete table of all the partitions of  6  into  3  non‑negative summands and computes  2x + 9y + 10z  for each, showing that  P  determines the  ai  if   a1 + a2 + a3  =  6,   whether it is a permutation of 1, 2, 3 or not.  No mnemonics.  This can also be viewed as a form of 7.P.1.

BR.  c1305.

No. 38, pp. 56‑59.  Determine values of three dice from  a + b + c  and  2a + 8b + 9c,  so this is essentially Fibonacci's second case with  k = 8.

No. 39, pp. 58‑59.  Same using  a + b + c  and  3a + 9b + 10c.

No. 100, pp. 118‑119.  Determine values of two dice from  a + b  and  2a + 10b.

Munich 14684.  14C.  No. IX.  This is obscure, but is repeated more clearly as No. XIX.  M = (2, 9, 10).  IX is followed by a two line verse.  Curtze could make no sense of it, but I wonder if it might be a mnemonic??

Folkerts.  Aufgabensammlungen.  13-15C. 

Discusses various simple divinations, citing Fibonacci.

He cites 18 sources using  a = 6, k = 9  in Fibonacci's first form.  For a permutation of three things, one can take  M = (m-n, m-1, m).  Subtracting from  am  leaves  na1 + a2.  This is equivalent to using  (0, 1, n).  10 sources for divining a permutation of  2, 3, 4  using  m = 10, n = 8.  Cites several older versions.

AR.  c1450.  Prob. 269, p. 122, 180‑181, 227‑228.  M = (2, 9, 10).  Then   60 ‑ P  =  8a1 + a2.   Vogel cites several earlier appearances, but mentions no mnemonics.  ??NYS.

Chuquet.  1484.  Prob. 159.  M = (1, 2, 4).  Starts with  24,  but  6  are used to label the people, so this is really subtracting from  18.  Table, but no mnemonic.  FHM 232.

Pacioli.  De Viribus.  c1500. 

Ff. 76v - 77v.  XXXV effecto de saper trovare .3. varie cose divise fra .3. persone et .4. divise fra .4. et de qua(n)te vorrai etc. (35th effect to know how to find 3 different things distributed among 3 persons and 4 among 4 and as many as you want etc.)  = Peirani 112-114.  Gives  12, 24, 36  counters to three people and says the person with the first object is to discard 1/2 of his pile, the person with the second is to discard 2/3 of his pile and the person with the third object is to discard 3/4 of his pile.  What is left takes on the values  23, 24, 25, 27, 28, 29.  The method is taking  M = (6, 4, 3),  which is equivalent to  (0, 1, 3).  There is no mention of doing it with four persons.

Ff. 78v - 79r.  XXXVII. commo el mode precedente se po far con fave et quartaruoli etc. (37th.  How the preceding method can be done with beans or farthings etc.)  = Peirani 115.  Same as the above.

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 142, f. 63v.  x*2+5*5+10+y*10+z  =  350 + 100x + 10y + z.  Cf Fibonacci.

Cardan.  Practica Arithmetice.  1539.  Chap. 61, section 18, f. T.iv.v (p. 113).  Mentions divination of a permutation of three things by use of  18  counters, so this is probably  M = (1, 2, 4),  subtracting from  18,  as in Chuquet, Baker, etc.

Baker.  Well Spring of Sciences.  1562?  Prob. 4,  1580?: ff. 197r‑198r;  1646: pp. 310-312;  1670: pp. 354-355.  M = (1, 2, 4),  subtracting from  18.  Vowel mnemonic:  Angeli, Beati, Taliter, Messias, Israel, Pietas  with explanatory table.

Recorde-Mellis.  Third Part.  1582.  Ff. Yy.v.r - Yy.v.v  (1668:. 477-478: Another [divination] of things hidden.  M = (1, 2, 4).  No mnemonics.

John Wecker.  Op. cit. in 7.L.3.  (1582), 1660.  Book XVI -- Of the Secrets of Sciences: Chap. 20 -- Of Secrets in Arithmetick: To discover to one a thing that is hid, pp. 289‑290.  M = (1, 2, 4).  No mnemonics.  Cites Gemma Frisius, ??NYS.

Prévost.  Clever and Pleasant Inventions.  (1584), 1998. 

Pp. 177-180.  M = (1, 2, 4),  subtracting from 18.  No mnemonics. 

Pp. 180-182.  M = (1, 2, 4),  subtracting from 23.  No mnemonics. 

Bachet.  Problemes.  1612.  Prob. XXII: De trois choses et de trois personnes proposées deviner quelle chose aura été prise par chaque personne, 1612: 115-126.  Prob. XXV,  1624: 187-198;  1884: 127‑134.  M = (1, 2, 4).  Vowel mnemonic:  Par fer  César  jadis  devint  si grand  prince.  Gives a four person version,  S = 78,  M = (1, 4, 16, 0), referring to Forcadel as giving an erroneous method.  Labosne says Diego Palomino (1599) has studied the four person case.  [This must be Jacobo Palomino; Liber de mutatione aeris in quo assidua et mirabilis mutationis temporum historia cum suis caussis enarratur. -- Fragmentum quodam ex libro de inventionibus scientiarum; Madrid, 1597 or 1599, ??NYR.]  Labosne adds some explanation, another mnemonic:  Avec  éclat  L'Aï  brillant  devint  libre,  and expands on Bachet's work on the case of four objects, but eliminates reference to Forcadel. 

van Etten.  1624.  Prob. 8 (8), pp. 9‑11 (19‑22).  M = (1, 2, 4).  Vowel mnemonics:  Salve  certa  anima  semita  vita  quies;   Par fer  Cesar  Iadis  Devint  si grand  Prince.  Henrion's 1630 Notte, pp. 10‑11, says that Bachet has extended it to  4  objects.

Hunt.  1631 (1651).  Pp. 255-261 (247-253).  Usual form with mnemonic:  Angeli  Beati  Pariter  Elias  Israel  Pietas.  Then does another version.

Schott.  1674.  Art. V, p. 58.  M = (1, 2, 4).  Vowel mnemonics:  Salve  Certa  Animæ  Semita  Vita  Quies;  Pallētis  Evandri  Sanguine  Feritas  Imane (the  m  should have an overbar) Vigebat.

Ozanam.  1694.  Prob. 28, 1696: 83;  1708: 74.  Prob. 32: 1725: 217-218.  Prob. 10, 1778: 154-155;  1803: 154-155;  1814: 136-137.  Prob. 9, 1840: 70.  M = (6, 4, 3).  No mnemonics.

Ozanam.  1725.  Prob. 46, 1725: 250-253.  Prob. 12, 1778: 158-161;  1803: 159-161;  1814: 140-142.  Prob. 11, 1840: 72.  M = (1, 2, 4).  Par fer  Cesar  jadis  devint  si grand  Prince   and   Salve certa anima semita vita quies.  1778 et seq. has  César  and  animæ.

Minguet.  1733.  Pp. 176-180 (1755: 127-129; not noticed in 1822; 1864: 164-166).  M = (1, 2, 4)  as in Chuquet.  Vowel mnemonic:  Aperi,  Premati,  Magister,  Nihil,  Femina,  Vispane,  Vispena.

Alberti.  1747.  Part 2, p. ?? (69).  M = (6, 4, 3)  translated from Ozanam, 1725, prob. 32.

Alberti.  1747.  Part 2, pp. ?? (76-77).  M = (1, 2, 4)  as in Ozanam, 1725, prob. 46, but he first gives an Italian mnemonic:  Aperì  Prelati  Magister  Camille  Perina  Quid habes  Ribera.  He explains the usage using  4 = Camille  as example, but later notes that  4  never occurs!  Then gives  Salva certa anima semita vita quies;  Perfer Cesar Jadis devint sigrand Prince;  Pare ella ai segni; Vita, Piè.

Les Amusemens.  1749.  Prob. 13, pp. 134-135: Les trois Bijoux.  M = (1, 2, 4).  Par fer,  César  jadis  devint  si grand  Prince.

Hooper.  Rational Recreations.  Op. cit. in 4.A.1.  1774.  Recreation IX: The confederate counters, pp. 34-36.  M = (1, 2, 4).  Par fer  Cesar  jadis  devint  si grand  prince.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 85-86, no. 131: Three persons having each chosen privately one out of three things, to tell them which they have chosen.  Salve  certa  anima  semita  vita  quies.

Manuel des Sorciers.  1825.  ??NX 

Pp. 35-37, art. 12.  Salve  Certa  Anima  Semita  Vita  Quies.

Pp. 38-39, art. 13.  Extends to four objects.  Start with  88  counters and give  1, 2, 3, 4  to the people, leaving  78.  Then tell them to take  1, 4, 16, 0  times more, so this is the same as Bachet's four person version.

Endless Amusement II.  1826? 

P. 115:  "Three things being privately distributed to three Persons, ...."  c= Badcock.

P. 179:  "Three Cards being presented to Three Persons, ...."  M = (6, 4, 3)  using the inverse permutation.  No mnemonic.

Young Man's Book.  1839.  Pp. 198-199.  Identical to Endless Amusement II, p. 179.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Parlour magic, no. xii, pp. 196-198 (1868: 209-210): How to discover the possessors of any articles taken from the table during your absence.  Salve  cesta  animæ  semita  vita  quies.

Magician's Own Book.  1857.

Prob. 36: The infallible prophet, pp. 23-24.  Mnemonics:  Salve  certa  animae  semita  vita  quies;   Par fer  Cesar  jadis  devint  si grand  prince.

The three graces, pp. 218-221.  Mnemonics:  James  Easy  admires now  reigning  with a  bride;   Anger,  fear,  pain, may  be hid  with a  smile;  Graceful  Emma,  charming she  reigns  in all  circles;   Brave  dashing  sea, like a  giant  revives  itself;   Salve ....

Boy's Own Conjuring Book.  1860.

Prob. 35: The infallible prophet, pp. 35-36.  Identical to Magician's Own Book.

The three graces, pp. 188-190.  Identical to Magician's Own Book.

Vinot.  1860. 

Art. XXXII: Trois cartes déterminées étant prises par trois personnes, deviner celle que chaque personne aura prise, p. 51.  Essentially  M = (6, 4, 3)  set up by giving them  12, 24, 36  and using  (1/2, 1/3, 1/4).

Art. XXXIV: Les trois bijoux, pp. 53-54.  Usual  (1, 2, 4)  versions starting with  24  counters, hence forming  18 - P.

Boy's Own Book.  Divination by cards.  1868: 637-638.  Basically  M = (1, 2, 4)  but set up as  (2, 3, 5).  Vowel mnemonic:  Par fer  César  jadis  devint  si grand  prince.

Magician's Own Book (UK version).  1871.  To tell which article each of three persons took, pp. 35-36.  Mnemonics:  Take her  certain  anise  seedlings  Ida  quince;  Salve  certa  animæe  servita  vita  quies;   Par-fer  Cæsar  jadis  devint  si-grand  prince.

Hanky Panky.  1872.  A new three-card trick, pp. 256-257.  M = (6, 4, 3),  which is equivalent to  (4, 2, 1).

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. VII, 1884: 192-193.  M = (6, 4, 3),  which is the same as  (4, 2, 1).  No mnemonics.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Simple Tricks by Calculation No. I: To tell three persons which card each one has chosen, pp. 32-33.  M = (4, 2, 1).

É. Ducret.  Récréations Mathématiques.  Op. cit. in 4.A.1.  1892?  Pp. 107-?? (didn't photocopy the following pages): Les trois Bijoux.  Seems to be usual form.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. VIII-X, pp. 21-25.  M = (1, 2, 4)  with result subtracted from  18:   Par fer  César  jadis  devint  si grand  prince;   Avec  éclat  l'Aï  brillant  devint  libre.   M = (0, 1, 3)  with result subtracted from  12:   Il a  jadis  brillé  dans ce  petit  État. 

Meyer.  Big Fun Book.  1940.  The picture card trick, pp. 509-510.  Uses  K, Q, J.  Forms  P  =  (2a1-1)m1 + (2a2-1)m2 + (2a3-1)m3   for multipliers  (mi) = (2, 3, 6).  Then examines the remainder of the deck, which has  49 - P = R  cards.  2 - [R/11]  is the person holding the  J.  A counting process, which is erroneously described, gives the person with the  Q  as congruent to  R + [R/11] (mod 3).  In fact  3 - R(mod 3)  is the position of the  K.

Gardner.  MM&M.  1956.  The purloined objects, pp. 57-59.  Gives several mnemonics using consonants.  E.g. when the objects are  Toothpick, Lipstick and Ring,  use:  tailor  altar  trail  alert  rattle  relate.  If the objects are denoted by  A, B, C,  use:  Abie's  bank  account  becomes  cash  club;  or if they are denoted  Small, Medium, Large,  use:  Sam  moves  slowly  (since)  mule  lost  limb.

 

          7.AP.           KNOWING SUM VS KNOWING PRODUCT

 

            Two persons are told the sum and product of two integers.  They then have some conversation such as:  "I don't know what the numbers are."  "I knew that."  "I now know what they are."  "So do I."  What were the numbers?  This seems to be a recent problem and I have only a few references.  There are older versions, often called census-taker problems, where one knows the sum and product of three numbers (usually ages), but needs more information (often whether there is an oldest, which eliminates twins).

            Thanks to Leroy F. Meyers for many of these references.

 

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.  No. 96: The church afloat, pp. 53 & 135.  Three ages:  product = 840,  sum is twice the curate's age.  This is insufficient, but whether the eldest is older or younger than the vicar is sufficient to decide.

Lester R. Ford, proposer.  Problem E776.  AMM 54 (1947) 339 & 55:3 (Mar 1948) 159-160.  ??NYS/NX -- proposal is quoted and the location of the solution given in the Otto Dunkel Memorial Problem Book (= AMM 64:7, part II, (Aug-Sep 1957) 61 & 89) and the location is more specifically given by Meyers & See.  Four families of different sizes, not enough to form two baseball teams (i.e. the total is  < 18), and the product of the numbers is the host's house number.  The guest says he needs more information -- does the smallest family have just one child?  The host's answer allows him to determine the numbers -- what were they?

Meyers writes that he first heard of the census-taker problem in 1951.

AMM problem E1126, 1954?.  ??NYS.

W. A. Hockings, proposer; A. R. Hyde, solver.  Problem E1156 -- The dimensions of Jones's ranch.  AMM 62 (1955) 181 (??NX) & 63:1 (Jan 1956) 39-42.  This is a continuation of E1126 and involves four men and their ranches.

Hubert Phillips.  My Best Puzzles in Mathematics.  Dover, 1961.  Prob. 87: The professor's daughter, pp. 48 & 101.  Youngest daughter is at least three.  The product of their ages is  1200  and the sum is ten less than the wife's age.  Visitor computes and then makes two wrong guesses as to the age of the youngest daughter.  How old is the wife?  The fact that the visitor made two wrong guesses means there must be at least three sets of ages, all  ³ 3,  with product  1200  and the same sum, and indeed there is just one such situation, and this has three daughters.  Allowing younger children permits some more complicated possibilities since we have 

                              1 + 3 + 20 + 20  =  2 + 2 + 10 + 30  =  3 + 16 + 25  =  4 + 10 + 30 

          and               2 + 2 + 15 + 20  =  4 + 15 + 20  =  5 + 10 + 24  =  6 + 8 + 25. 

          (Phillips had most of these problems in his newspaper and magazine columns so it is likely that this will turn up in the period 1930-1950.)

M. H. Greenblatt.  Mathematical Entertainments, op. cit. in 6.U.2, 1968.  Chap. 1: "Census-taker" problems, pp. 1-7.  Says he believes these problems came from some wartime project at MIT.  Discusses three similar types.

                    1: The neighborhood census.  Product of three ages is  1296  and the sum is the house number.  Census taker asks if any of them are older than than the informant, who says 'no' and then the census taker knows the ages.

                    2: The priest and the banker.  Product of three ages of ladies with the banker is  2450  and the sum is the same as the priest's age.  The priest says this is insufficient and asks if any of the ladies is as old as the banker.  When he says 'no', the priest knows all the ages.  This even determines the banker's age!

                    3: The three Martians.  Product of three ages is  1252  and the sum is the age of the informant Martian's father.  The interrogator says this is insufficient and asks if any of the Martians is as old as the informant.  When he answers 'no', the interrogator knows the ages.  When you start on the problem, you find that  1252 = 22·313  is not a suitable number.  However, the problem shows a drawing of the Martian and one see he only has three fingers on each hand!  Interpreting  1252  as a base  6  number gives the decimal  320  and the problem proceeds as before.

Gardner.  SA (Nov & Dec 1970)  c= Wheels, Chap. 3, prob. 10: The child with the wart.  Supplied by Mel Stover.  The product of the ages of three children is  36  and the sum is the questioner's house number.  When the questioner says the information is insufficient, the father says the oldest child has a wart, which is sufficient to determine the ages.  See Meyers & See for a generalization.

A. K. Austin.  A calculus for know/don't know problems.  MM 49:1 (Jan 1976) 12-14.  He develops a set-theoretic calculus for systematically solving problems involving spots on foreheads, etc. (see 9.D), including problems similar to the present section.  He gives the following problem considered by Conway and Patterson, but apparently unpublished.  Two persons each have a card on their back bearing a positive integer, visible to the other person, but not to the person.  They are told that the sum of the two integers is  6  or  7.  They are then asked, in turn, to state whether they know what their own integer is.  If they both have  3,  what is the sequence of responses?  Austin finds there are five 'no' answers, then the answers alternate  yes, no.

David J. Sprows, proposer;  Problem Solving Group, Bern, solver.  Prob. 977 -- Mr. P. and Ms. S.  MM 49:2 (Mar 1976) 96 & 50:5 (Nov 1977) 268.  P & S are given the product and sum of two integers  a, b,  greater than one, but the sum is  £ 100.  P says he doesn't know the numbers.  S says she knew that.  P replies that he now knows the values and S responds that so does she.  The unique answer is  a, b = 4, 13.  Editor cites the above AMM problems, though they are not quite the same type.  See the discussion below.

Gardner.  SA 241:6 (Dec 1979) 20-24.  Problem 1: The impossible problem.  Gives a version of Sprows' problem and says the problem was sent by Mel Stover and had been circulating for a year or two.  This version assumes the numbers are greater than  1  but at most  20,  which gives the unique solution  4, 13.  However Gardner asserts that the solution remains the same if the bound is increased to  100 -- but I think there are further answers, e.g.  4, 61,  see the discussion below.  Stover says a computer program has checked and found no further solutions up to  2,000,000  and it may be that there is no further solution when the upper bound is removed -- this is a mistake of some sort, see the discussion below.  Further, Kiltinen & Young say they had a letter from Gardner conjecturing that there are infinitely many solutions.

John O. Kiltinen  &  Peter B. Young.  Goldbach, Lemoine, and a know/don't know problem.  MM 58:4 (Sep 1985) 195-203.  This discusses the Sprows problem, without the bound on  S  and various generalizations involving more stages of conversation.  This leads to use of Goldbach's Conjecture and a conjecture of Lemoine that every odd number  ³ 7  can be expressed as  2p + q,  where  p, q  are odd primes.

Friedrich Wille.  Humor in der Mathematik.  Vandenhoek & Ruprecht, Göttingen, (1984), 3rd ed., 1987.  Paul und Simon, pp. 62 & 121‑122.  This has solution  4, 13.  Note to the solution says there was much correspondence after the 1st ed., leading to a new solution and the computation of further examples with  Sum =  17,  65,  89,  127,  137,  163,  179,  185,  191,  233,  247,  269,  305.

 

                    I recently was sent a version of the Sprows problem, without a bound on  S,  by Adrian Seville and solved it before checking this section.  Dr. S (= Simon) and Dr. P (= Paul) are given the sum and product of two integers greater than  1.  Dr. S says he doesn't know the numbers, but he can tell that Dr. P cannot determine the numbers.  After a bit, Dr. P says he now knows the numbers.  After a bit more, Dr. S says he now knows the numbers also.  The version I received implies there is a unique solution, but there are more.  Wille's discussion notes that the possible values of the sum, after Dr. S's statement are the  {odd composites + 2},  which much simplifies the analysis.  I found some further values of  Sum:  343,  427,  457 and then extended considerably, finding:  547,  569,  583,  613,  637,  667,  673,  697,  733,  757,  779,  787,  817,  821,  853,  929,  967,  977,  989,  997,   giving  36  values less than  1000  and there are another  42  solutions between  1000  and  2000.  In looking at the earlier solutions, it appeared that one of the two numbers was always a power of 2, with the other number being odd, but for  S = 757,  P = 111756,  one has the numbers being  202  and  556.  Five more counterexamples to the initial appearance occur up to  2000.

                    While doing this investigation, I wondered what would happen if the condition 'greater than  1'  was reduced to  'positive'.  After some calculation, it became apparent and is provable that all solutions have  P = S - 1  with the numbers being  1  and  S - 1.  Solutions occur for  S =  5,  9,  10,  16,  28,  33,  34,  36,  46,  50,  52,  66,  78,  82,  88,  92,  96,   giving  17  values less than  100,  and there are another  87  values between  100  and  1000,  making a total of  104  values less than  1000,  which is where I stopped.  If one assumes Goldbach's Conjecture, or part of it, one can show there are infinitely many solutions of the form  S = pq + 1,  where  p, q  are primes such that  p + q = r + 1  for a prime  r.  However, I have not been able to see if there are infinitely many solutions of the original form of the problem.

                    I later was referred to the MM problem 977 and was surprised to see that the solution is unique when  S £ 100  is imposed -- offhand, one might expect  S = 65 and 89  to also be solutions in this case, but the additional knowledge about  S  affects the intermediate stages of the deduction.  Rerunning an early version of my program with additional printout, I find that one needs  S < 107  to prevent  S = 65  being a solution.  I then found Gardner's 1979 version and examination of the same data indicates that  S = 65  is a solution when the numbers are bounded by  100.  All in all, I find the presence of a bound has subtle effects and I find it unsatisfying.

 

Tim Sole.  The Ticket to Heaven and Other Superior Puzzles.  Penguin, 1988.  Standard and Poor, pp. 116 & 126‑128.  Solution is  2, 15.

Leroy F. Meyers & Richard See.  The census-taker problem.  MM 63:2 (Apr 1990) 86-88.  The product of the three ages in the next house is  1296  and the sum is the house number.  The census-taker then asks if any of those people is older than the respondent.  When he says 'yes', the census-taker says he knows the ages.  This is a slight variant on Greenblatt's type 1.  Authors investigate what other values,  N,  can be used instead of  1296.  That is, we want  N = ABC = DEF  with  A + B + C = D + E + F  with just two such triples  A, B, C;  D, E, F.  They determine the simple forms of such  N  and list the  45  examples  £ 1296.  The first few are:  36,  40,  72,  96,  126,  176,  200,  234,  252,  280,  297,  320,  408,  520,  550,  576,  588,  600.  They give a number of references to similar questions.

Nicolas Guerrero.  Problem 184.7 -- Deux nombres.  M500 184 (Feb 2002) 25.  The problem is given in French, but I will translate and paraphrase.  Dr. P is given the product and Dr. S is given the sum of two integers between 2 and 100.  (This is ambiguous -- the range might be  [2, 100]  or  [3, 99].)  Dr. P says he doesn't know the numbers.  Dr. S then says he doesn't know the numbers.  Dr. P then says he knows them, and Dr. S then says he knows them. 

ADF [Tony Forbes].  Editorial remark on Problem 184.7.  M500 188 (Oct 2002) 26.  His citation is erroneously to M500 185, p. 25.  He says he has not been able to obtain a satisfactory answer, in particular his answer is not unique. 

 

                    My analysis gives four answers for  S, P  in the range  [2, 100]: 

          a, b, S, P  =  2, 6, 8, 12;  4, 19, 23, 76;  4, 23, 27, 92;  7, 14, 21, 98.  However, it is clear that the intention is for just  a, b  to be in this range and this leads to so many possible values of  P  that a computer must be used.  My program finds two solutions: 

          a, b, S, P  =  2, 6, 8, 12;  84, 88, 172, 7392.  Adapting the program to the range  [2, 200],  the larger solution disappears, but is replaced by six solutions with  P  larger than  30,000.  I suspect that the proposer of the problem may have intended the earlier problem of Sprows, but either misheard or misremembered the exact details.

                    If one takes range  [3, 99]  for all the variables,  the calculation is a little simpler, and there are three solutions:  a, b, S, P  =  3, 8, 11, 24;  9, 9, 18, 81;  9, 11, 20, 99.  But there is an extra feature -- the values  a, b, S, P  =  6, 13, 19, 78;  8, 11, 19, 88  lead to the above sequence of statements except that Dr. S's last statement is that he still doesn't know the numbers, but we know his sum!

 

          7.AQ. NUMBERS IN ALPHABETIC ORDER

 

            I recall the problem of figuring out the reason for the sequence:  8, 5, 4, 9, 1, 7, 6, 3, 2  from 1956 or 1957.

 

??  Alphabetic Number Tables, 0 ‑ 1000.  MIT, Cambridge, Massachusetts, 1972.  ??NYS.

"Raja" [= Richard & Josephine Andree].  Puzzle Potpourri #3.  Raja Books, Norman, Oklahoma, 1976.  No. 17: volumes  1, 2, ..., 12  shelved alphabetically.

Harvey & Robert Dubner.  A tabulation of the prime numbers in the range of one to one‑thousand, in English and in Roman numerals, in alphabetic order.  JRM 24 (1992) 89-93.

 

          7.AR.           1089

 

            The basic process is to take a three digit number, find the difference between it and its reversal, then add that result to its reversal and one gets  1089.  The items by Meyer and Langford note that  9 * 1089 = 9801,  so this is connected to 7.AH.

            I now think that this ought to have originated from the fact that a number minus its reversal is divisible by nine, cf Berkeley & Rowland below and Section 7.K.1, but the monetary version seems to have arisen first.  I have a note that I have seen a c1881 reference to the monetary version of this problem, but I cannot find it.

 

Prof. Orchard, proposer;  Prof. Anderson, Rev. H. Sewell, et al, solvers.  Prob. 10441.  Mathematical Questions with Their Solutions from the "Educational Times" [generally known as Educational Times Reprints] 53 (1890) 78-79.  Prove and generalise the fact that the process done with old English money gives  £12 18s 11d.  Solution says that if the multipliers of the units are  m  and  n  so that  a, b, c  denotes  amn + bm + c,  then the result of the calculation is  m, n-2, m-1.  The proposal makes no mention of the problems when the number of pounds is  12  or more or when the numbers of pounds and pence are equal and assumes the reversal is less than the original, but the solver assumes that  a < m  to make the reversal be an ordinary sum of money and then that  c < a  "to make the subtraction possible".  Mr. Davis notes that this trick has been "current in well-informed City-circles for some months."  If you perform it more than once, instruct the victim to add some convenient amounts to the result.

Don Lemon.  Everybody's Scrap Book of Curious Facts.  Saxon, London, 1890.  Curious arithmetical puzzle, pp. 302-303.  Quotes a letter to the London Globe from a 'constant reader' asking for an explanation for the monetary version, assuming you start with less than  £11 19s 11d  and the number of pence is less than the number of pounds.

Ball.  MRE, 1st ed., 1892. 

P. 9.  Monetary version getting  £12 18s 11d.  Cites Educational Times Reprints, giving the same examples, but specifically says the amount must be less than £12 and the number of pounds cannot be the same as the number of pence.  He says to take the difference of the number and its reversal.  At the end, he states: "The rule can be generalized to cover any system of monetary units", but he doesn't give the general solution.  This version is dropped in the 12th ed., 1974.

P. 13.  Notes that a number minus its reversal is divisible by  9.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Card Puzzles No. X: A subtraction sum, p. 10.  A  3-digit number minus its reversal has the form  a9c,  where  a + c = 9.

Somerville Gibney.  An arithmetical flourish for drawing-room shows.  The Boy's Own Paper 15 (No. 734) (4 Feb 1893) 299.  Monetary version, getting  £12 18s 11d.  "An arithmetical curiosity which is quite new ...."

Somerville Gibney.  Re The arithmetical flourish for drawing-room shows.  The Boy's Own Paper 15 (No. 750) (27 May 1893) 556.  Explains why it works.  Correspondents from various places do it:  in francs and centimes, getting  99 99;  in thalers and grosschen  (30  grosschen = a thaler), getting  29 29;  and  in avoirdupois, getting  28cwt 2qr 27lb.

Ball‑FitzPatrick.  1st ed., 1898, pp. 14‑15.  Shows the property of  1089.  A remark of the translator shows the answer in base  b  is   (b‑1) * (b+1)2, which is 

          b b2 + (b-2) b + (b-1).

Carroll-Collingwood.  1899.  P. 269 (Collins 194).  = John Fisher; The Magic of Lewis Carroll; op. cit. in 1; pp. 244-245.  = Carroll-Wakeling II, prob. 25: Pounds, shillings and pence, pp. 41 & 71,  Monetary version, with number of pounds 'not more than twelve', giving  £12 18s 11d.  Carroll assumes the reversal is less than the original number and says it works "whatever numbers may have been selected."  Mentioned in Carroll-Gardner, p. 76.  Neither Collingwood nor Wakeling nor Gardner note the difficulty when numbers of pounds and pence are equal nor the question of whether the reversal is actually smaller than the original, but Fisher does, and also mentions  1089.  Collingwood says he believes this was invented by Carroll and this seems reasonable, but see Orchard et al, 1890, above.

Clark.  Mental Nuts.  1904, no. 81; 1916, no. 54.  Mental telegraphy.  As in Berkeley & Rowland.

Ball.  MRE, 4th ed., 1905, p. 9.  Adds a section on  1089  just before the monetary version.

Laisant.  Op. cit. in 6.P.1.  1906.  Chap. 18: Opérations curieuses: No. 1, p. 43.  1089.  Notes that the two end digits must be different.

E. N. Barisien.  ??  Suppl. al Periodico di Mat. 13 (1909) 20-21.  ??NYS -- cited by Dickson I 462, item 53, but the interesting material is cited in the following.

E. Nannei.  ??  Suppl. al Periodico di Mat. 14 (1910/11) 17-20.  ??NYS -- cited by Dickson I 462, item 55, as treating a problem of Barisen.  Takes a  6-digit number, subtracts its reversal and then adds the reversal of that to itself.  Result is one of  13  values:   0,  9900,  ...,  1099989.

M. Adams.  Indoor Games.  1912.  Magic money, 354-355.  Monetary version getting  £12 18s 11d.

Clark.  Mental Nuts.  1916, no. 53.  Foretelling the answer.  Gives the rule and an example, saying the answer will always be  1089,  though he doesn't make any restriction on the original number.

Dudeney.  AM.  1917.  Prob. 33: A puzzle in reversals, pp. 5 & 151.  Says most people know that the money process leads to  £12 18s 11d  -- provided you start with less than 12 in the  £  place and the number of pounds exceeds the number of pence.  If the number of pounds can be  12  or more, then what is the smallest sum for which the process fails and the largest sum for which it works?

T. O'Conor Sloane.  Rapid Arithmetic.  Quick and Special Methods in Arithmetical Calculation Together with a Collection of Puzzles and Curiosities of Numbers.  Van Nostrand, 1922.  [Combined into: T. O'Conor Sloane, J. E. Thompson & H. E. Licks; Speed and Fun With Figures; Van Nostrand, 1939.]  A mystery in money, pp. 178-180.  Uses US money to get  $10.89  and also gives the English result  £12 18s 11d.

Peano.  Giochi. 1924.

Prob. 48, p. 12.  1089.

Prob. 49, pp. 12-13.  As in Berkeley & Rowland.

Hummerston.  Fun, Mirth & Mystery.  1924. 

An overheard conversation, p. 40.  After getting  1089,  he asks for it to be reversed and to multiply these, then do some other operations to get an 8-digit number when is translated to a message by  1 = A, 2 = B,  ....

Two parlour tricks (second part), p. 146.  Tell someone to take a number of three distinct digits, reverse it and subtract the smaller from the larger.  Divide the result by nine and have him tell you one of the digits.  You divine the other, which is in fact the same!  This is based on the easily verified result that when  C < A,  ABC ‑ CBA  =  99(A-C)

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 60: Mental arithmetic trick, pp. 51-52.  Monetary version.

Kraitchik.  Math. des Jeux.  Op. cit. in 4.A.2.  1930.  Chap. IV, no. 7, p. 55.  Notes that the first result is a multiple of  99,  but doesn't find  1089.  Not in his Math. Recreations.

Perelman.  1934??  Guessing a number without asking anything.  FFF (1934).  Not in 1957 ed.  1979: prob. 14, pp. 33-35.  = MCBF (1937), prob. 14, pp. 32-33.  Basic property of  1089.

Rohrbough.  Brain Resters and Testers.  c1935.  Brain Teaser, p. 15.  Assumes digits are in descending order.

Freda Holmdahl, ed.  The Rainy Day Book.  Nelson, Edinburgh, 1936.  Figure tricks, pp. 196‑197.  Monetary version.

Depew.  Cokesbury Game Book.  1939.  Descending numbers, p. 205.  1089.

"Willane".  Willane's Wizardry.  Academy of Recorded Crafts, Arts and Sciences, Croydon, 1947.  Curious figures, p. 45.  Version with English money, with  £  less than  12.

Anonymous.  Problems drive.  Eureka 18 (Oct 1955) 15-17 & 21.  No. 9.  Usual 1089 problem.

P. M. Seeviour & M. Keates.  Note 3077:  Fibonacci again.  MG 48 (No. 363) (Feb 1964) 78‑79.  Cites Kraitchik, Math. des Jeux.  Determines all the possible results of the process for  n‑digit numbers in base  b,  as being  (b2 ‑ 1)  times certain  n-digit numbers formed of  0s  and  1s,  corresponding to the borrows in the subtraction.  They find the number  t(n)  of such values is given by  t(2n)  =  t(2n+1)  =  F2n+1  where  Fm  are the Fibonacci numbers.  They do this via the case when the first and last digits are distinct for which it is stated that a certain recurrence holds, but this is not immediate -- I spent about half a day before I could see it and it would still require some work to make a careful proof.  For convenience, I list the first solutions for the case of different end digits.

                    n = 2.  99.

                    n = 3.  1089.

                    n = 4.  10989,  9999,  10890.

                    n = 5.  109989,  99099,  109890.

                    n = 6.  1098900,  1099989,  1090089,  1099890,  1089990,  999999,  990099,  991089.

                    n = 7.  10998900,  10999989,  10891089,  10999890,  10890990,  10008999,  9900099,  9901089.

C. Dudley Langford.  Note 3102:  A party puzzle extended (Cf Note 3077).  MG 48 (No. 366) (Dec 1964) 432.  If  a + b = 10,  then  1089*a  and  1089*b  are reversals.  1089  can be replaced by  99 * 11...1  and he asserts that these, up to trailing  0s,  are the numbers obtained in Note 3077.  However, his numbers are:  99,  1089,  10989,  109989,  1099989, ..., 109...989, ...,  and these do not immediately yield the above solutions.

Jerome S. Meyer.  Arithmetricks.  Scholastic Book Services, NY, 1965.  The  2178  trick, pp. 3-4.  Gets  1089,  but he doubles the result at the end to produce  2178.

Michael Holt.  What is the New Maths?  Anthony Blond, London, 1967.  Pp. 88-89.  Gives a four digit version, saying to choose the least significant two digits as  < 5  and the most significant as  ³ 5,  then the result is always  10890.

Walter Gibson.  Big Book of Magic for All Ages.  Kaye & Ward, Kingswood, Surrey, 1982.  Magic Math  &  Super Math, pp. 108-109.  After doing  1089,  he considers doing the same process with five digit numbers.  For convenience, he assumes the digits are all distinct.  Then only two possible answers can arise:  99099  or  109890.  No proof given.

Paul Swinford.  The Wondrous World of Numberplay & Wordplay.  A Lecture by Paul Swinford.  Published by the author, 1999.  Pp. 1-4.  Covers 1089 with a variety of ways to use the result.  On pp. 3-4, he considers four digit versions and is unhappy to often get  10890.  He modifies the process so that one only exchanges the end digits and finds that this always leads to  10989.  (For a  k-digit number, the process will lead to  1099..989,  with  k-3  9s  in the middle.)

 

          7.AS.           CIGARETTE BUTTS

 

            New section.  There must be older examples.  The usual form uses  b  butts to make a cigarette and the person has found  B = b2  butts which give him  b+1  smokes.  The Scotts are the only ones to realise that  B  can more generally be such that  b-1  divides  B-1,  giving  (B‑1)/(b-1)  smokes.  I have just found the general result that the number of smokes is  ë(B‑1)/(b-1)û.  Taking  b = 2  gives the result that a knockout tennis tournament with  B  players has  B-1  matches.

            On 15 Nov 2001, Willy Moser mentioned that if the tramp borrows a butt, he gets  ëB/(b-1)û  smokes with a butt left over to return to his friend!  He thought this was well known, but I'd never heard or seen it before

 

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 192, pp. 74 & 158: Scrooge the miser.  Starts with  125  cigarettes, but saves all his butts, using  5  to make a cigarette, so he gets  31  further cigarettes.

Rudin.  1936.  No. 49, pp. 16 & 85.  36  butts, using  6  to make a cigarette.

J. R. Evans.  The Junior Week‑End Book.  Op. cit. in 6.AF.  1939.  Prob. 47, pp. 266 & 272.  Tramp has  49  butts and uses seven to make a cigarette.  How many cigarettes can he make?

Harold Hart.  The World's Best Puzzles.  Home Recreation Library, NY, 1943.  Smoke gets in your eyes, pp. 22 & 54.  125  butts and five make a cigarette.

Leopold.  At Ease!  1943.  Iffs and Butts, pp. 10-11 & 196.  36  butts, six butts to a cigarette.

Sullivan.  Unusual.  1943.  Prob. 9: A bum cigarette.  As in Rudin.

J. S. R. Cameron.  A tennis problem.  Eureka  10 (Mar 1948) 18  &  11 (Jan 1949) 31.  Complicated system of byes in a tennis tournament with 100 players.  Solution notes that the system of byes is irrelevant!  This must have been known much earlier.

Jonathan Always.  Puzzles to Puzzle You.  Op. cit. in 5.K.2.  1965.  No. 131: Cigarettes this way, pp. 41 & 89.  As in Evans.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  Hard times made easier, pp. 35-36.  41  butts, six butts to a cigarette.

David Singmaster.  Cigarette butts.  MS 31:2 (1998/9) 40.  The number of smokes is  ë(B‑1)/(b-1)û.  Taking  b = 2  gives the result that a knockout tennis tournament with  B  players has  B-1  matches.

 

          7.AT.           BOOKWORM'S DISTANCE

 

          New section.  This must appear in the 19C??

 

Loyd.  Cyclopedia.  1914.  Pp. 327 & 383.  = SLAHP, Timing a bookworm, pp. 69 & 112.  2  volumes.

Dudeney.  AM.  1917.  Prob. 420: The industrious bookworm, pp. 143-144 & 248-249.  3  volumes.

Peano.  Giochi.  1924.  Prob. 10, p. 4.  3  volumes.

King.  Best 100.  1927.  No. 3, pp. 8 & 39.  = Foulsham's, no. 1, pp. 5 & 10.  3  volumes.

William P. Keasby.  The Big Trick and Puzzle Book.  Whitman Publishing, Racine, Wisconsin, 1929.  Pp. 37 & 63.

Streeter & Hoehn.  Op. cit. in 7.AE.  Vol. 1, 1932, p. 300, no. 10:  "Brain twister".

Haldeman-Julius.  1937.  No. 33: The bookworm problem, pp. 6 & 22.  Three volumes.

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  Bookworm's pilgrimage, pp. 163 & 215.  4  volumes.

McKay.  At Home Tonight.  1940.  Prob. 18: The bookworm, pp. 66 & 79-80.  2  volumes.

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.  P. 148, prob. 4.  2  volumes.

Leopold.  At Ease!  1943.  Way of a worm, pp. 9 & 195.

Owen Grant.  Popular Party Games.  Universal, London, nd [1940s?].  Prob. 10, pp. 37 & 50.  3  volumes.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Second, pp. 145 & 196.  4  volumes.

 

          7.AU.           NUMBER OF CUTS TO MAKE  N  PIECES

 

            New section.  A typical problem is that it takes twice as long to make three pieces as it does to make two pieces.  This will include other situations where one has to change from  n  to  n-1.

            If one can overlay the material, then one can reduce the number of cuts.  If one can fold the material, then one can reduce to one cut as in 6.BG.

            I've just added the problem of cutting a cube into 27 pieces here -- this must be much older??

            See also 7.AV.

 

Alcuin.  9C.  Prob. 15: De homine.  "How many furrows has a man made ..., when he has made three turnings at each end of the field?"  The Alcuin text gives six, but Bede, Folkerts and Folkerts & Gericke give seven.

Child.  Girl's Own Book.  Arithmetical puzzles.  1832: No. 10, pp. 171 & 179;  1833: No. 10, pp. 185 & 193 (answer numbered 8);  1839: No. 10, pp. 156 & 173;  1842: No. 10, pp. 284 & 291;  1876: No. 8, pp. 232 & 244.  "If you cut thirty yards of cloth into one-yard pieces, and cut one yard every day, how long will it take you?"  The 1839, 1842 & 1876 texts omit the hyphen in one-yard.  I didn't copy the exact text from the 1832.

                    = Fireside Amusements, 1850, Prob. 8, pp. 132 & 184.

The Sociable.  1858.  Prob. 46: A dozen quibbles: part 9, pp. 300 & 318.  "If you cut thirty yards of cloth into one yard pieces, and cut one yard every day, how long will it take?"  = Book of 500 Puzzles, 1859, prob. 46: part 9, pp. 18 & 36.  c= Depew; Cokesbury Game Book; 1939; Cutting up, p. 216.  c= Magician's Own Book (UK version), 1871, Paradoxes [no. 5], p. 38.  = Wehman, New Book of 200 Puzzles, 1908, p. 50.

Hoffmann.  1893.  Chap. IX, no. 24: The draper's puzzle, pp. 318 & 326 = Hoffmann-Hordern, p. 211.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XII: La coupe du tailleur, p. 26.  How long to cut a piece of length  16  into lengths of length  2?  [This makes the problem more deceptive.]

H. D. Northrop.  Popular Pastimes.  1901.  No. 10: A dozen quibbles, no. 9, pp. 68 & 73.  = The Sociable.

Benson.  1904.  The tailor's puzzle, p. 227.

Hummerston.  Fun, Mirth & Mystery.  1924.  Some queer puzzles, Puzzle no. 76, part 4, pp. 164 & 183.  Cut twenty yards into yard lengths. 

King.  Best 100.  1927.  No. 24, pp. 15 & 44.  Cut  90  yards of cloth into  90  lengths.

Meyer.  Big Fun Book.  1940. 

Sissy's scissors, pp. 120 & 739.  Three sheets of paper,  20 x 40,  to be cut into  2 x 20,  without folding.  He does it in  19  cuts by stacking the three sheets and cutting the side of  40  into  20  pieces.  One can easily reduce this to  10  by first cutting the  20 x 40  into two  20 x 20,  stacking them and then cutting the stack into  10  parts.  One can reduce further by cutting the  10  edge in the middle, getting  5, 5.  Stack these and cut the  5  into  1, 2, 2  with two cuts.  Then stack the  2s  and cut once more.  This takes a total of  5  cuts, which seems minimal to me.  There is a problem in that the last  2 x 10  pack is  24 sheets thick! 

No. 9, pp. 162 & 752.  Draw four equidistant horizontal lines and then four equidistant verticals.  How many squares are formed?  This gives a  3 x 3  array of squares, but he counts all sizes of squares, as in 5.X.2, getting  9 + 4 + 1 = 14.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Fifth, pp. 58 & 190.  How many cuts to make  27  cubes from a cube?

Putnam.  Puzzle Fun.  1978.  Nos. 93-96: Cube and  27  cubes, pp. 13 & 37.  How many cuts to make  27  cubes from a cube?  Other questions deal with colourings.

 

          7.AV.           HOW LONG TO STRIKE TWELVE?

 

          New section.  I have now started adding rate problems of this type.

          See also 7.AU.

 

King.  Best 100.  1927.  No. 36, pp. 19 & 47.  If a clock takes  8  sec to strike  8,  how long does it take to strike  12?

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 186, pp. 72 & 157: Big Ben.  "If it takes Big Ben ten seconds to strike five o'clock, how long will it take to strike twelve o'clock?"

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  Is your brain working? -- no. 2, pp. 148 & 215.  If Big Ben takes  4  seconds to strike  4,  how long does it take to strike midnight?

Depew.  Cokesbury Game Book.  1939.  The clock, p. 217.  If it takes  7  seconds to strike  7,  how long does it take to strike  11.

Sullivan.  Unusual.  1943.  Prob. 25: Caution.  If it takes  8  minutes to pass  8  stop lights, how long does it take to pass  12?

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 25: The patriot's paradox, pp. 30-31.  60  shots in an hour is not the same as one shot per minute, since the latter requires  61  shots to be fired in a  60  minute interval from the first shot.

Ripley's Puzzles and Games.  1966.  P. 70.  "If 10 sheep jump over a fence in 10 minutes  how many will jump over in an hour?"  Answer: 55.

Barbara Lee.  Six short problems.  M500 162 (?? 1998)  ??NYS -- cited by solver.  A, J. Welton.  Solutions to 'Six short problems'.  M500 163 (Aug 1998) 12-13.  No. 3: "A clock strikes six in 5 seconds.  How long does it take to strike twelve?"  "Six hours and five seconds."

 

          7.AW.          28/7  =  13

 

            There are three (or more) ways of demonstrating this result -- e.g.  7  doesn't go into  2,  but it goes into  8  once, leaving us with  21  and  7  goes into  21  3  times.  The other ways are based on multiplying  13  by  7  and on  adding seven  13s.  This is the basis of a vaudeville routine which was used in the Abbott and Costello movie 'In the Navy', 1949.  I can only recall seeing it in the following.

 

Rohrbough.  Brain Resters and Testers.  c1935.  Tricks for the Toastmaster (no. 2), pp. 12-13.

W. A. Bagley.  Paradox Pie.  Op. cit. in 6.BN.  1944.  No. 2: Financial skullduggery, pp. 7‑8.

Bud Abbott and Lou Costello.  In the Navy.  Universal Pictures, 1941.  Text and stills in Richard J. Anobile, ed.; Who's On First?  Verbal and Visual Gems from the Films of Abbott & Costello; Darien House, NY, 1972; Studio Vista, London, 1973.  Pp. 128‑139.

William R. Ransom.  Op. cit. in 6.M.  1955.  Fake arithmetic, pp. 9‑11.  Shows there are  22  examples of this phoney arithmetic.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  The Christmas club, p. 62.

 

          7.AX.          SUM  =  PRODUCT,  ETC.

 

            The archetypal example is to determine  A, B  such that  A + B  =  A * B,  usually with some additional condition.  I have not noticed early examples of such problems until recently reading Riese's Die Coss and later Muscarello, but I recall other later examples such as Briggs & Bryan.

 

Muscarello.  1478.  Ff. 69r-69v, p. 180.  A + B  =  A * B.  He gives examples:  6 & 6/5;  7 & 7/6;  8 & 8/7;  and says it continues.

Calandri.  Arimethrica.  1491.  F. 97v.  A + B  =  A * B.  He finds  5/2, 5/3.

Riese.  Die Coss.  1524.

No. 38, p. 45.  A + B  =  A/B,  with  A/B  =  3/2.

No. 41, p. 46.  A ‑ B  =  A/B,  with  A/B  =  4/3.

Ozanam.  1694.  Prob. 5, question 14, 1696: 13;  12.  Prob. 8, quest. 14, 1725: 27-28.  A + B  =  A * B.  Gives  (a+b)/a  and  (a+b)/b,  e.g.  5/2  &  5/3.

Eadon.  Repository.  1794. 

P. 299, no. 8.  A + B  =  A * B.  Gives  B  =  A/(A-1).

P. 299, no. 9.  A - B  =  A2 - B2.  Gives  B = 1-A.

P. 299, no. 10.  A + B  =  A2 - B2.  Gives  B = A-1.

Manuel des Sorciers.  1825.  ??NX  P. 85.  A + B  =  A * B.

Walter Taylor.  The Indian Juvenile Arithmetic ....  Op. cit. in 5.B.  1849.  Pp. 193-194.  Solves  A + B  =  A * B  as  A = (a+b)/a,  B = (a+b)/b  for any  a, b.

Mittenzwey.  1880.  Prob. 80, pp. 15 & 66;  1895?: 88, pp. 20 & 69;  1917: 88, pp. 18 & 66.  A + B  =  3 (A - B);  AB  =  3 (A + B).  1917 gives an algebraic solution.

Briggs & Bryan.  The Tutorial Arithmetic, Part II.  Op. cit. in 7.H.  1898.  Exercises X, prob. 16, pp. 124 & 579.  A + B  =  A * B  =  A2 ‑ B2.

C. T. C. Wall.  The (σ, x, p) problem.  Eureka 20 (Oct 1957) 20-25.  This starts with finding  x, y, p, q  such that  x + y = pq,  xy = p + q.  There are several special solutions:  x, ‑1, 1-x, -1;  x, -x, 0, -x2  and two sporadic solutions:  3, 2, 5, 1;  2, 2, 2, 2.  He then considers more variables such that the symmetric functions of one set are the same as for the other set, but in reverse order.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. e, pp. 189-190.  A - B  =  A/B.

Jerome S. Meyer.  Arithmetricks.  Scholastic Book Services, NY, 1965.  Pp. 35-36.  Considers all four cases.

          A * B  =  A + B.    (n+1)/n  and  n+1.

          A * B  =  A - B.    n  and  n/(n+1).

          A / B  =  A + B.    n + 1/(n+2)  and  (n+1)/(n+2).

          A / B  =  A - B.    n + 1/(n-2)  and  n-1.

Bronnie Cunningham.  Funny Business.  An Amazing Collection of Odd and Curious Facts with Some Jokes and Puzzles Too.  Puffin, 1978.  Pp. 37 & 142.  A - B  =  A/B  = 4.

 

          7.AY.          SUM OF POWERS OF DIGITS

 

            A  PDI  (Perfect Digital Invariant) is an  n  digit number which is equal to the sum of the  k-th powers of its digits.  If  k = n,  it is called a  PPDI  (PluPerfect Digital Invariant).  More generally, one can consider the function  fk(N) = sum of the  k-th powers of the digits of  N  (in base  10  usually)  and study its iterates.  It is not hard to see that  fk(N) < N  from some point on, so the iterates must lead to cycles.  E.g., for  k = 3,  we have: 

160 ® 217 ® 352 ® 160.

            See Reviews in Number Theory, sections A62 and A64 for related material.  See 7.BB for other iterated functions of integers.

 

F. Hoppenot.  Courrier du "Sphinx".  Sphinx 7:4 (Apr 1937) 72.  Gives  153, 371, 407, 8208, 9474  as PPDIs.

M. Fistié.  Courrier du "Sphinx".  Sphinx 7:5 (May 1937) 87.  Adds  370  to Hoppenot's list and says these are all the PPDIs for  k = 3.

Ball.  MRE, 11th ed., 1939, p. 13.  Gives the four PPDI's for  k = 3:  153, 370, 371, 407,  citing Sphinx.

G. H. Hardy.  A Mathematician's Apology.  CUP, 1940.  "There are just four numbers (after  1)  which are the sums of the cubes of their digits ....  These are odd facts, very suitable for puzzle columns and likely to amuse amateurs, but there is nothing in them which appeals to a mathematician."  He cites Ball for them.

H. Steinhaus.  One Hundred Problems in Elementary Mathematics.  (As:  Sto Zadań, PWN -- Polish Scientific Publishers, Warsaw, 1958.)  Pergamon Press, 1963.  With a Foreword by M. Gardner, Basic Books, NY, 1964.  Prob. 2: An interesting property of numbers, pp. 11-12 & 55-58.  Considers iterates of  f2(n).  He shows this always leads to a cycle and find there are two cycles,  1  of length  1  starting from  1  and  1  of length  8  starting from  4.

Kiyoshi Iseki and coworkers extended Steinhaus's idea to  k = 3, 4, 5  in 5 papers during 1960‑1963.  See Leveques's Reviews in Number Theory, A64-12 -- A64-16 for details.  For  k = 3,  there are  9  cycles  (5  of which have period  1:  1  and the four PPDI's for  k = 3).  For  k = 4,  there are  6  cycles  (4  of which have length  1:  1,  1634,  8208,  9474).  For  k = 5,  there are  15  cycles, the longest having length  28.  There are  5  cycles of length  1:  1,  54748,  93084,  92727,  194979.

Max Rumney.  Digital invariants.  RMM 12 (Dec 1962) 6-8.  Quotes Hardy.  General introduction to the ideas.  Interested in repeating the process of summing the  n-th powers of an  n-digit number, so his PDI is what is now called a PPDI.  He also considers cases where the process cycles through a few values, e.g. for the sum of the cubes of the digits, we get the cycle:  133,  055,  250,  133.  There are such cycles starting with  133,  136,  217,  919.

Gardner.  SA (Jan 1963) c= Magic Numbers, chap. 3.  Surveys what is known with numerous references up through 1973.  In particular, D. St. P. Barnard showed that the number of PPDIs is finite by noting that  61 * 961 < 1060,  so that a PPDI must have  < 61  digits. 

Harry L. Nelson.  More on PDI's.  Publication UCRL-7614, Univ. of California.  1 Dec 1963.  ??NYS -- cited by Schwarz and others below, who say it proves and improves Barnard's result and that it makes the distinction between PDI and PPDI as now generally used.

Azriel Rosenfeld, proposer;  Nathan J. Fine, solver.  Problem E1651 -- Sums of squares of digits.  AMM 71 (1964) 90  &  1042-1043.  "Prove that no multidigit integer is equal to the sum of the squares of its digits."

Benjamin L. Schwarz.  Finiteness of a set of self-generating integers.  JRM 2:2 (Apr 1969) 79‑83.  Proves Barnard's result and improves it to  < 60  digits.

Benjamin L. Schwarz.  Finite bounds on digital invariants -- some conjectures.  JRM 3:2 (Apr 1970) 88-92.  General discussion of whether the number of PDIs is infinite.  He gives a heuristic argument that there are infinitely many.

Joseph S. Madachy.  Some new narcissistic numbers.  Fibonacci Quarterly 10:3 (Apr 1972) 295-298.  "[R]eports on various narcissistic numbers other than digital invariants."  Gives various forms and some references.

Benjamin L. Schwarz.  Self-generating integers.  MM 46:3 (May 1973) 158‑160.  Gives a simple condition for functions of the digits so that the number of PPDIs (= SGIs in his notation) must be finite.  This applies to the sum of the factorials of the digits.

Lionel E. Deimel Jr. & Michael T. Jones.  Finding pluperfect digital invariants:  techniques, results and observations.  JRM 14:2 (1981‑82) 87-108.  Cites work from 1962.  All PPDIs for bases  2  through  10  were found, using months of Data General MV/8000 time, checking primality for bases  < 10.  There are  89  PPDIs in base  10  (but this includes  0  as an example for exponent  1,  though it is not clearly a  1-digit number, so perhaps  88  is a better count). 

Mary T. Whalen & Gordon L. Miller.  Prime PPDI's.  JRM 25:2 (1993) 118‑123.  Says PPDIs are also known as Armstrong numbers.  The present authors have checked the above results and looked for examples of primes in base  10.  The base  10  results took a DEC 6310 just over  14  days and three primes (beyond the exponent  1  cases) were found.  Results were checked with Mathematica on a NeXT.  They give  88  PPDIs for base  10.

Tony Forbes.  Recurring digital invariants.  M500 165 (Dec 1998) 4-7.  After some comments and results on PDIs in base  10  and loops in the iteration process, he decides to examine base  3  and finds there are many examples and gives some very large examples.

 

          7.AZ.           DIVINATION OF A PAIR OF CARDS FROM ITS ROWS

 

            A spectator mentally chooses a pair from ten pairs of cards.  You then lay them out in a  5  by  4  array and ask him to tell you in which rows his pair appears, whereupon you point out his pair.  This is done by laying out the cards in a particular pattern traditionally given by the Latin words:  MUTUS  DEDIT  NOMEN  COCIS.  In fact, such patterns are easily found for  r(r+1)/2  pairs laid out in  r  rows -- but finding a memorable pattern takes some doing.  Since the person's response can be either one row or two rows, the number of responses is  r + BC(r, 2)  =  r(r+1)/2  =  BC(r+1, 2).

            Unger,  Secret Out,  Ball,  Ahrens  and  Indoor Tricks and Games  deal with triples.  With  r  rows, there are  r + BC(r, 2) + BC(r, 3)  =  r (r2 + 5)/6  possible responses.  However, this does not always give a possible situation.  E.g., for  r = 2,  there are  3  possible responses, but one cannot layout  3  triples in two rows, and the same problem occurs when  r  is even.  For  r = 3,  there is an easy pattern:  AAADDEG  BBBDFFG  CCCEEFG  and I imagine it can be easily done whenever  r  is odd.  However, we are getting to more responses than we can handle and it might be easier to use fewer of them.  See Unger for an approach.

            This is actually a combinatorial problem and I will probably shift this section into Chapter 5.

 

Bachet.  Problemes.  1612.  Prob. XVI: De plusieurs cartes disposées en divers rangs deviner laquelle on aura pensée: 1612: 87-92.  Prob. 18, 1624: 143‑151;  1884: 72-83.  This problem covers several card divinations.  The present topic has been added to "cette seconde [impression de ce livre]".  He discusses the problem for  r  rows and shows the simplest pattern, but gives no mnemonics.  For  10  pairs, the simplest pattern has rows:  AABCD,  BEEFG,  CFHHI,  DGIJJ.

Wingate/Kersey.  1678?.  Prob. 5, pp. 539-542.  Does  10  and  15  pairs as in Bachet.

Hooper.  Rational Recreations.  Op. cit. in 4.A.1.  1774.  Recreation XXVII: The ten duplicates, pp. 70-71.  MUTUS  DEDIT  NOMEN  COCIS.

Endless Amusement I.  c1818.  P. 106: The ten duplicates.  As in Hooper.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 122-123, no. 184: The ten duplicates.  c= Endless Amusement I.

Rational Recreations.  1824.  Feat 38, p. 168: Cards in couples.  MUTUS DEDIT NONEM [sic] COCIS.

Manuel des Sorciers.  1825.  Pp. 48-51, art. 21.  ??NX  MUTUS  NOMEN  DEDIT  COCIS.  Extends to 15 pairs, in a different way than Bachet, by bordering the usual array of letters with a left hand column  5, 4, 3, 2, 1  and adding a bottom row which continues from the  1  just added with  1, 2, 3, 4, 5.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 233-235, nos. 897-901.  He first gives the 10 pair case, using  ABCBD  EFAGE  HGHIC  KFKID,  which is isomorphic to  MUTUS  NOMEN  DEDIT  COCIS,  but he never gives any mnemonics.  He then asks which other arrays can be done and sees that  r  rows can be used with  r(r+1)/2  pairs.  He gives a scheme for  r = 3:  ABAC  BDDE  EFCF,  and a scheme for  r = 5:  ABCDEA  BFGHFI  KGLLCM  ONNDHK  EIOPMP.  He then considers triples and gives a scheme for 8 triples:  AAA123  BBB124  CCC143  DDD423.  This has the feature that the responses are either one row or three rows, giving   r + BC(r, 3) = r (r2 ‑ 3r + 8)/6  responses.  He shows an easy method for extending a scheme to the next number of rows getting  AAA123567  BBB124589  CCC143860  DDD423907  EEE567890  with 15 triples.  He notes that one may want to rearrange the columns to make it less obvious when the cards are laid out.  He really should have started with  AAA1  BBB1  CCC1  and then constructed  AAA123  BBB124  CCC134  DDD234  in order to make the construction clearer.

Young Man's Book.  1839.  Pp. 205-206.  The Ten Duplicates.  Identical to Endless Amusement I.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Parlour magic, no. xxii, pp. 202-203 (1868: 213-214): The ten duplicates.  MUTUS  DEDIT  NOMEN  COCIS.

Magician's Own Book.  1857.  No. 21: The pairs re-paired, pp. 65-66.  MUTUS  DEDIT  NOMEN  COCIS  ["Mutus gave a name to the Coci," a people who have yet to be discovered.]  = Boy's Own Conjuring Book, 1860, p. 73.

The Secret Out.  1859. 

To Reveal a Person's Thoughts: an excellent Card Trick, pp. 87-89.  The UK version: To Tell Cards Thought Of, p. 4, is much briefer.  MUTUS  DEDIT  NOMEN  COCIS.

To Guess which of Twenty-Four Cards have been Noted, pp. 82-83.  Extension to eight triples using  LIVINI  LANATA  LEVETE  NOVOTO.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 3, no. 9: The pairs repaired, pp. 83-84.  MUTUS  DEDIT  NOMEN  COCIS.

Boy's Own Book.  The ten pairs of cards distinguished.  1868: 638.  NOMEN  MUTUS  DEDIT  RORIS.

Hanky Panky.  1872.  To tell two cards out of twenty, p. 254.  MISAI  TATLO  NEMON  VESUL.

Berkeley & Rowland.  Card Tricks and Puzzles.  1892.  Simple Tricks by Calculation No. XII: To name several different pairs of cards chosen by one or more persons, pp. 44-46.  SHEER  CHAFF  USUAL  COLOR;   MUTUS  DEDIT  NOMEN  COCIS;   MISAI  TATLO  NEMON  VESUL (where  V = U).

Ball.  MRE.  1st ed., 1892.  Determination of a selected pair of cards out of  ½n(n+1)  given pairs, pp. 98-101.  Gives general rule for pairs.  MATAS  DEDIT  NOMEN  COCIS.  Mentions  8  triples:  LANATA  LEVETE  LIVINI  NOVOTO,  as in Secret Out.

Ahrens.  MUS I.  1910.  Pp. 148-152.  MUTUS  DEDIT  NOMEN  COCUS.  Discusses generalizations, including examples with triples, etc.  Cites Unger, 1838, for the first idea of finding triples and quotes the mnemonic from Ball.

Indoor Tricks and Games.  Success Publishing, London, nd [1930s??]. 

The wandering pairs, pp. 38-40.  MUTUS  DEDIT  NOMEN  COCIS.

The re-united triplets, pp. 40-41.  LIVINI  LANATA  LEVETE  NOVOTO.

Phyllis Fraser & Edith Young.  Puzzles Quizzes and Games.  Bantam Books, NY, 1947.  No. 2: Card trick, pp. 97-99.  THIGH  ATLAS  GOOSE  BIBLE.

Stewart Judah.  The divertisements of Stewart Judah.  The New Phoenix, No. 319 (Nov 1954) 83.  UNDUE  GOANO  TETRA  RIGID.

Will Dexter.  The Illustrated Book of Magic Tricks.  Abbey Library, London, nd, but Introduction dated 1957.  The Romans had a word for it!, pp. 73-74.  DAVID  LOVEL  IN YON  ABBEY;   MUTUS  NOMEN  COCIS  DEDIT ("Mutus gave a name to Cocis").

Martin Gardner.  Mutus Nomen.  Pallbearers Review 5:7 (May 1970) 338.  Reprinted in his:  Martin Gardner Presents; Richard Kaufman & Alan Greenberg, 1993, pp. 237-238.  For 30 cards:  LIVELY  RHYTHM  MUFFIN  SUPPER  SAVANT.  For 20 cards:  BIBLE  ATLAS  GOOSE  THIGH.

Walter Gibson.  Big Book of Magic for All Ages.  Kaye & Ward, Kingswood, Surrey, 1982.  Twenty Cards  &  New Deal Twenty Card Trick, pp. 162-165.  Mentions  MUTUS ...  and says there are many English possibilities and gives  RUFUS  STEEL  TIARA  FOLIO.  The New Deal version uses a more complicated laying-out process starting with ten piles of two using  MAGIC  STORE,  then laying out according to  AGORA  METER  SCOTS  GIMIC.

Karl Fulves.  Mutus Nomen.  The author, 1998.  A 90pp booklet on variations of this classic trick.  ??NYR.

At a lunch in 2002, Rob Eastaway did the trick and said he used  BIBLE  GOOSE  ATLAS  THIGH  and Jeremy Wyndham said he used  DAVID  LOVEL  IN YON  ABBEY.  I had never met anyone who used anything other than  MUTUS  ...  before!

 

          7.BA.           CYCLE OF NUMBERS WITH EACH CLOSER TO TEN THAN THE PREVIOUS

 

            New section, due to discovering Lewis Carroll's version.  I have no idea if this is older than Carroll.

 

Lewis Carroll.  A Tangled Tale.  (1885) = Dover, 1965.  The pigs.  Knot VIII: De omnibus rebus, pp. 52-57 & 132‑134.  (In the answers, this part of the Knot is denoted §1. The pigs.)  "... place twenty‑four pigs in those four sties, so that, as she goes round the court, she may always find the number in each sty nearer to ten than the number in the last."  Answer:  8, 10, 0, 6  -- the key is that "nothing is nearer ten than  10".

 

          7.BB.           ITERATED FUNCTIONS OF INTEGERS

 

            See 7.T and 7.AY for some special forms.

            Let  f(n)  be a function of an integer  n.  The problems in this section concern the  behaviour of the sequence of iterations of  f:  n, f(n), f2(n) = f(f(n)), f3(n), ....  There are a great number of behaviours that can occur.

            1.         The sequence can diverge to infinity.

            2.         The sequence can converge to a fixed point.

            3.         The sequence can become cyclic.

            4.         The sequence can behave chaotically.

Further the behaviour may be different for different starting points. 

            If  f(n) > n  in all cases, then the iteration will diverge to infinity and is not very interesting.  Hence we are normally only interested in functions which have  f(n) < n  for some non-zero proportion of the integers.

            If  f(n) £ n  in all cases, then the sequence must converge to a fixed point, though different starting points may lead to different fixed points.  Then the problem is to identify the fixed points and the sets they attract.

            So the most interesting cases have  f(n)  sometimes greater and sometimes smaller than  n.  Two classic examples are the following.

            The proper sum of divisors,  Σ(n),  is  σ(n) - n,  where  σ(n)  is the classic number theoretic function of the sum of all divisors.  The iterates of this have been studied since the 19C as the fixed points are the perfect numbers and 2-cycles are amicable numbers -- see 7.AB.  Some larger cycles have been found -- I recall a 5 and a 17 cycle.  No provably infinite sequence has been found.

            The Syracuse function is the following.  If  n  is even, halve it;  if  n  is odd,  form  3n + 1.  Some of the sequences have been computed for hundreds of thousands of terms without termination.

            Problems of this sort are genuine mathematical recreations but are a bit too elaborate for me to include here.  What I will include are functions whose sequences are always finite.  These generally are of two types.  Functions such that  f(n) < n  for all  n  above some limit -- see, e.g., 7.AY, and functions such that  f(n)  has the same number of digits as  n.  If  f(n)  is the sum of some function  g(ai)  of the digits  ai  of  n,  set  M = max {g(0), g(1), ..., g(9)}.  Then for  10d-1 £ n < 10d,  we have  f(n) £ (d+1)M  and this will be less than  n  from some point on.

 

Joseph S. Madachy.  Mathematics on Vacation.  Op. cit. in 5.O, (1966), 1979.  Narcissistic numbers: Digital invariants, pp. 163‑165.  Based on Rumney and Nelson.  Carries on to consider other types of 'narcissistic numbers' which are numbers which are some function of their digits -- various forms are discussed on pp. 165-177.

George D. Poole.  Integers and the sum of the factorials of their digits.  MM 44:5 (Nov 1971) 278-279.  The only integers equal to the sum of the factorials of their digits are:  1,  2,  145,  40585.  He has to check up to  2,000,000  before a general method can be used.

                    In Dec 2001, Al Posamentier mentioned the following cycle of this function: 

          169 ® 363601 ® 1454 ® 169.  I essentially repeated Poole's work, but found five numbers which are equal to the sum of the factorials of their digits, excluding leading zeroes:  0, 1, 2, 145, 40585.  Zero is exceptional -- I view it as having no digits, so the sum involved is empty!  The largest number for which this function is larger than the given number is  1,999,999  and there are  208,907  such numbers.  Using this, one has a limited search to find all the cycles - there are two other cycles, both 2-cycles:  871 « 45361  and  872 « 45362.  Cf Schwartz and Kiss, below.

Benjamin L. Schwarz.  Self-generating integers.  MM 46:3 (May 1973) 158‑160.  Gives a simple condition for functions of the digits so that the number of PPDIs (= SGIs in his notation) must be finite.  This applies to the sum of the factorials of the digits.

Peter Kiss.  A generalization of a number-theoretic problem [in Hungarian, with English summary].  Mat. Lapok 25:1-2 (1974 (1977)) 145-149.  ??NYS -- described in Reviews in Number Theory A62-201F (= MR 55, #12612).  An apparent English translation is: A generalization of a problem in number theory; Math. Sem. Notes Kobe Univ. 5:3 (1977) 313-317; ??NYS - described ibid. A62-201G (= MR 57, #12362).  He demonstrates that if  f(n)  is the sum of any function of the digits of  n,  then the iteration of  f  must lead to a cycle -- as sketched above.  He finds all the cycles for three cases, including for the factorial function.  The review mentions Steinhaus (cf section 7.AY) and Poole.

Roger Cook.  Reversing digits.  MS 31:2 (1998/9) 35-37.  The initial problem was to take a four digit number, arrange the digits in ascending order and in descending order and subtract the first from the second.  In all cases except  aaaa,  the iterations converge to  6174  (7641 - 1467 = 6174).  He discusses other number of digits and other bases.

 

          7.BC.           UNUSUAL DIFFICULTY IN GIVING CHANGE

 

            The typical problem here is that a customer offers a note to pay a bill and the shopkeeper says he can't give change, but when the customer offers a larger note, then the shopkeeper can give change. 

            Some of the problems in 7.P.1 could be considered as falling into this topic.

            New section.  I have seen a 1935 version somewhere.

 

Carroll-Collingwood.  c1890  Prob. 2, pp. 317-318.  = John Fisher; The Magic of Lewis Carroll; op. cit. in 1; p. 79.  = Carroll-Wakeling II, prob. 31: Coins, pp. 49 & 72-73.  A man has a half-sovereign (10s), a florin (2s) and a 6d and wants to buy goods worth 7s 3d.  However the shopkeeper has a crown (5s), a shilling (1s) and a penny (1d), so cannot make change.  A friend comes into the shop and he has a double florin (4s), a half-crown (2s 6d), a fourpenny piece (4d) and a threepenny bit (3d).  Can they sort out the payment?  I consulted a dictionary and found that the double-florin was first minted in 1887, which puts a lower bound on the date of this problem, so I have given the date as c1890 rather than c1890?

Clark.  Mental Nuts.  1916, no. 4.  The conductor's money.  Passenger tenders a $1 note to pay 5¢ (= $.05) fare, but conductor cannot give change, but he can give change for a $5 note.  The answer is that he can give a $2.50 gold piece, a $2 note and 45¢ in coins.

David Singmaster.  No change!  Written up as a problem for my Telegraph column in summer 1999.  Shopkeeper cannot give change for a £5 note, but can for two £5 notes.  How can this happen?  How about in the US?

 

 

 

 

 

 

 

 

 

 

8.       PROBABILITY RECREATIONS

 

            Most of the recreations in probability are connected with some paradoxical feature.  A good exposition of most of these appears in the following.

 

Gábor J. Székely.  Paradoxes in Probability Theory and Mathematical Statistics.  Akadémiai Kiadó, Budapest and Reidel, Dordrecht, 1986.  [Revised translation of:  Paradoxonok a Véletlen atematikában; Műszaki Könvkiadó, Budapest, nd.]  Translated by Márta Alpár and Éva Unger.  ??NYR.

 

          8.A.   BUFFON'S NEEDLE PROBLEM

 

R. E. Miles & J. Serra.  En Matiere d'introduction.  In:  Geometrical Probability and Biological Structures: Buffon's 200th Anniversary.  Lecture Notes in Biomathematics, No. 23, Springer, 1978, pp. 3‑28.  Historical survey, reproduces main texts.

 

Buffon.  (Brief commentary).  Histoire de l'Acad. des Sci. Paris (1733 (1735)) 43‑45.  Discusses problem of a disc meeting a square lattice and then the stick (baguette) problem, but doesn't give the answer.

Buffon.  Essai d'arithmétique morale, section 23.  1777.  (Contained in the fourth volume of the supplement to his Histoire Naturelle, pp. 101‑104??)  = Oeuvres Complètes de Buffon; annotated by M. Flourens; Garnier Frères, Paris, nd [c1820?], pp. 180‑185.  (Also in Miles & Serra, pp. 10‑11.)

Laplace.  Théorie Analytique des Probabilitiés.  1812.  Pp. 359‑360.  ??NYS.  (In Miles & Serra, p. 12.)  3rd ed, Courcier, Paris, 1820, pp. 359-362.  Finds the answer to Buffon's problem with needle length  2r  and line spacing  a.  Then solves the case of two perpendicular sets of lines, with possibly different spacings.

M. E. Barbier.  Note sur le problème de l'aiguille et le jeu du joint couvert.  J. Math. pures appl. (2) 5 (1860) 273‑286.  Gives result for arbitrary curves and considers several grids.  Also gives his theorem on curves of constant width.

M. W. Crofton.  On the theory of local probability.  Philos. Trans. Roy. Soc. 158 (1869) 181‑199.  (Excerpted in Miles & Serra, pp. 13‑15.)

A. Hall.  On an experimental determination of  π.  Messenger of Mathematics 2 (1873) 113‑114.  ??NYS.

Tissandier.  Récréations Scientifiques.  1880?  2nd ed., 1881, 139-145 describes the result and says 10,000  tries with a  50 mm  needle on a floor with spacing  63.6 mm,  produced  5009  successes giving  π = 3.1421.

                    =  5th ed., 1888, pp. 204-208.  c= Popular Scientific Recreations, 1890? pp. 729‑731, but the needle is  2 in  on a floor of spacing  2½ in  and  5009  is misprinted as  £5000 (sic)  but  5009  is used in the calculation.

J. J. Sylvester.  On a funicular solution of Buffon's "Problem of the needle" in its most general form.  Acta Math. 14 (1890‑1891) 185‑205.

N. T. Gridgeman.  Geometric probability and the number  π.  SM 25 (1959) 183‑195.  Debunks experimental results which are often too good to be true, although they are frequently cited.

J. G. L. Pinhey.  The Comte de Buffon's paper clip.  MG 54 (No. 389) (Oct 1970) 288.  Being caught without needles, he used paperclips.  He derives the probability of intersection assuming a paper clip is a rectangle with semi‑circular ends.

Jack M. Robertson & Andrew F. Siegel.  Designing Buffon's needle for a given crossing distribution.  AMM 93 (1986) 116‑119.  Discusses various extensions of the problem.

 

          8.B.   BIRTHDAY PROBLEM

 

            How many people are required before there is an even chance that some two have the same birthday?

 

George Tyson was a retired mathematics teacher when he enrolled in the MSc course in mathematical education at South Bank in about 1980 and I taught him.  He once remarked that he had known Davenport and Mordell, so I asked him about these people and mentioned the attribution of the Birthday Problem to Davenport.  He told me that he had been shown it by Davenport.  I later asked him to write this down.

George Tyson.  Letter of 27 Sep 1983 to me.  "This was communicated to me personally by Davenport about 1927, when he was an undergraduate at Manchester.  He did not claim originality, but I assumed it.  Knowing the man, I should think otherwise he would have mentioned his source, ....  Almost certainly he communicated it to Coxeter, with whom he became friendly a few years later, in the same way."  He then says the result is in Davenport's The Higher Arithmetic of 1952.  When I talked with Tyson about this, he said Davenport seemed pleased with the result, in such a way that Tyson felt sure it was Davenport's own idea.  However, I could not find it in The Higher Arithmetic and asked Tyson about this, but I have no record of his response.

Anne Davenport.  Letter of 23 Feb 1984 to me in response to my writing her about Tyson's report.  "I once asked my husband about this.  The impression that both my son and I had was that my husband did not claim to have been the 'discoverer' of it because he could not believe that it had not been stated earlier.  But that he had never seen it formulated."

I have discussed this with Coxeter (who edited the 1939 edition of Ball in which the problem was first published) and C. A. Rogers (who was a student of Davenport's and wrote his obituary for the Royal Society), and neither of them believe that Davenport invented the problem.  I don't seem to have any correspondence with Coxeter or Rogers with their opinions and I think I had them verbally.

Richard von Mises.  Ueber Aufteilungs‑ und Besetzungs‑ Wahrscheinlichkeiten.  Rev. Fac. Sci. Univ. Istanbul (NS) 4 (1938‑39) 145‑163.  = Selected Papers of Richard von Mises; Amer. Math. Soc., 1964, vol. 2, pp. 313‑334.  Says the question arose when a group of  60  persons found three had the same birthday.  He obtains expected number of repetitions as a function of the number of people.  He finds the expected number of pairs with the same birthday is about  1  when the group has  29  people, while the expected number of triples with the same birthday is about  1  when there are  103  people.  He doesn't solve the usual problem, contrary to Feller's 1957 citation of this paper.

Ball.  MRE, 11th ed., 1939, p. 45.  Says problem is due to H. Davenport.  Says "more than  23"  and this is repeated in the 12th and 13th editions.

P. R. Halmos, proposer;  Z. I. Mosesson, solver.  Problem 4177 -- Probability of two coincident birthdays.  AMM 52 (1945) 522  &  54 (1947) 170.  Several solvers cite MRE, 11th ed.  Solution is  23  or more.

George Gamow.  One, Two, Three ... Infinity.  Viking, NY, 1947.  = Mentor, NY, 1953, pp. 204‑206.  Says  24  or more.

Oswald Jacoby.  How to Figure the Odds.  Doubleday, NY, 1947.  The birthday proposition, pp. 108-109.  Gives answer of  23  or more.

William Feller.  An Introduction to Probability Theory and Its Applications: Vol 1.  Wiley, 1950, pp. 29-30.  Uses an approximation to obtain  23  or more.  The 2nd ed., 1957, pp. 31‑32 erroneously cites von Mises, above.

J. E. Littlewood.  A Mathematician's Miscellany.  Op. cit. in 5.C.  1953.  P. 18 (38) mentions the problem and says  23  gives about even odds.

William R. Ransom.  Op. cit. in 6.M.  1955.  Birthday probabilities, pp. 38‑42.  Studies usual problem and graphs probability of coincidence as a function of the number of people, but he doesn't compute the break‑even point.  He then considers the probability of two consecutive birthdays and gets upper and lower estimates for this.

Gamow & Stern.  1958.  Birthdays.  Pp. 48‑49.  Says the break‑even point is "about twenty‑four".

C. F. Pinzka.  Remarks on some problems in the American Mathematical Monthly.  AMM 67 (1960) 830.  Considers number of people required to give greater than  50%  chance of having  3, 4 or 5  with the same birthday.  He gets  88, 187, 314  respectively, using a Poisson approximation.  He gives the explicit formula for having  3  with the same birthday.

Charlie Rice.  Challenge!  Op. cit. in 5.C.  1968.  Probable probabilities, pp. 32-36, gives a variety of other forms of the problem. 

A  chooses five letters of the alphabet;  B  tries to guess at least one of them in five guesses.  Author says odds are two to one in favor of  B,  though I get four to one.  This is the same as getting five distinct items from a set of  21.

Get people to think of cards.  For  9  or more people, the probability of two the same is  @ .52;   for  11,  @ .68;   for  12,  @ .75.

Get people to count their change.  For a moderate number of people, it is likely that two have the same amount (or the same number of coins).  Likewise, with a moderate number of people, two are likely to have fathers (or mothers) with the same given name.

He gives magic numbers, i.e. the size of the set to be selected from, for different sizes of group in order that a duplication is more likely than not.  For  6  people, the magic number is  23;   for  8,  43;   for  12,  99.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.  No. 58: The birthday puzzle, pp. 76, 116-117 & 122.  Asks for probability of a shared birthday among  30  people.  Says the problem was introduced by Gamow.  Answer is  .70.  He adds that  24  or more will give better than even odds and then asks how many people are necessary for one of them to have his birthday on a given day -- e.g. today.  Here the answer is  253.

E. J. Faulkner.  A new look at the probability of a coincidence of birthdays in a group.  MG 53 (No. 386) (Dec 1969) 407‑409.  Suggests the probability should be obtained by the ratio of the unordered selections, i.e.  Prob. (r distinct birthdays)  =  BC(365, r)/BC(364+r, r),  rather than  P(365, r)/365r.  But the unordered selections with repetitions are not equally likely events -- see Clarke & Langford, below.  For his approach, the breakeven number is  r = 17.

Morton Abramson & W. O. J. Moser.  More birthday surprises.  AMM 77 (1970) 856‑858.  If a year has  n  days,  k ³ 1  and  p  people are chosen at random, what is the probability that every two people have birthdays at least  k  days apart?  For  n = 365,  k =   1,   2,   3, 4, 5, 6, 7, 8, 9, 10,  the breakeven numbers are  p = 23, 14, 11, 9, 8, 8, 7, 7, 6,   6.

L. E. Clarke & Eric S. Langford.  Note 3298 -- I & II:  On Note 3244.  MG 55 (No. 391) (Feb 1971) 70‑72.  Note the non‑equally likely events in Faulkner.

W. O. J. Moser.  It's not a coincidence, but it is a surprise.  CM 10 (1984) 210‑213.  Determines probability  P  that in a group of  k  people, at least two have birthdays at most  w  days apart.  This turns out to have a fairly simple expression.  To get  P > .5  with  w = 0  requires  k ³ 23,  the classical case.  With  w = 1,  it requires  ³ 14, ....

Tony Crilly & Shekhar Nandy.  The birthday problem for boys and girls.  MG 71 (No. 455) (Mar 1987) 19‑22.  In a group of  16  boys and  16  girls, there is a probability greater than  ½  of a boy and a girl having the same birthday and  16  is the minimal number.

Roger S. Pinkham.  Note 72.25:  A convenient solution to the birthday problem for girls and boys.  MG 72 (No. 460) (Jun 1988) 129‑130.  Uses an estimate to obtain the value  16  of Crilly and Nandy.

M. Lawrence Clevenson & William Watkins.  Majorization and the birthday inequality.  MM 64:3 (1991) 183-188.  Do the numbers necessary for  P > .5  get bigger if birthdays are not random?  Answer is "no" and it is a result in majorization theory, but they give an elementary treatment.

S. Ejaz Ahmed  &  Richard J. McIntosh.  An asymptotic approximation for the birthday problem.  CM 26:2 (Apr 2000) 151-155.  Without using Stirling's approximation, they show that for a calendar of  n  days and a desired probability  p,  0 < p < 1,  then the minimum class size to produce a probability  ³ p  of two people having the same birthday is asymptotically  Ö[2n log {1/(1-p)}].  For  p = .5,  n = 365,  this gives  22.494.

 

 

            See also 6.BR, esp. the Mathematical Log article and my comments.  I have a large number of similar results, mostly by myself, in a file.  I estimate there are about  20  possible answers, ranging from  0  to  1.

 

J. J. Sylvester.  On a special class of questions in the theory of probabilities.  Birmingham British Association Report (1865) 8.  = The Collected Mathematical papers of James Joseph Sylvester, (CUP, 1908); Reprinted by Chelsea, 1973, item 75, pp. 480-481.  ??NYS, but Guy (below) reports that it is a discursive article with no results.  Attributes problem for three points within a circle or sphere to Woolhouse but feels the problem is not determinate.

C. Jordan.  1872‑1873.  See entry in 8.G.

E. Lemoine.  1882-1883.  See entry in 8.G.

L. Carroll.  Pillow Problems.  1893.  ??NYS.  4th ed., (1895).  = Dover, 1958.  Problem 58, pp. 14, 25, 83‑84.  Prob (acute) = .639.

C. O. Tuckey.  Note 1408:  Why do teachers always draw acute‑angled triangles?  MG 23 (No. 256) (1939) 391‑392.  He gets  Prob(obtuse)  varying between  .57  and  .75.

E. H. Neville.  Letter:  Obtuse angling -- a catch.  MG 23 (No. 257) (1939) 462.  In response to Tuckey, he shows  Prob(obtuse) = 0  and deduces that  Prob(acute) = 0 (!!!).

Nikolay Vasilyev.  The symmetry of chance.  Quantum 3:5 (May/Jun 1993) 22-27 & 60-61.  Survey on geometric probability.  Asks for the probability of an acute triangle when one takes three points at random on a circle and gets  ¼.

Richard K. Guy.  There are three times as many obtuse-angled triangles as there are acute-angled ones.  MM 66:3 (Jun 1993) 175-179.  Gives 12 different approaches, five of which yield  Prob(acute) = ¼,  with other values ranging from  0  to  .361.  He tracked down the Sylvester reference -- see above.

In Mar 1996, I realised that the two approaches sketched in 6.BR give probabilities of acuteness as  0  and  2 - π/2  =  .429.

 

          8.D.   ATTEMPTS TO MODIFY BOY‑GIRL RATIO

 

          This is attempted, e.g. by requiring families to stop having children after a girl is born.

 

Pierre Simon, Marquis de Laplace.  Essai Philosophique sur les Probablitiés (A Philosophical Essay on Probabilities).  c1819, ??NYS.  Translated from the 6th French ed. by F. W. Truscott & F. L. Emory, Dover, 1951.  Chap. XVI, pp. 160‑175, especially pp. 167‑169.  Discusses whether the excess of boys over girls at birth is due to parents stopping having children once a son is born.

Gamow & Stern.  1958.  A family problem.  Pp. 17‑19.

 

          8.E.    ST. PETERSBURG PARADOX

 

Nicholas Bernoulli.  Extrait d'une Lettre de M. N. Bernoulli à M. de M... [Montmort] du 9 Septembre 1713.  IN:  Pierre Rémond de Montmort; Essai d'analyse sur les jeux de hazards.  (1708);  Seconde edition revue & augmentee de plusieurs lettres,  (Quillau, Paris, 1713 (reprinted by Chelsea, NY, 1980));  2nd issue, Jombert & Quillau, 1714.  Pp. 401-402.  Quatrième Problème & Cinquième Problème, p. 402.  [See note to Euler; Vera aestimatio sortis in ludis; op. cit. below, pp. 459‑461.]  In the 4th problem, he proposes pay-offs of  1, 2, 3, 4, ...  if the player first throws a six with a die on throw  1, 2, 3, 4, ...  and asks for the expectation.  He does not compute it, but I get  6.  In the 5th problem, he asks what happens in the same situation if the pay-offs are  1, 2, 4, 8, ...  or  1, 3, 9, 27, ...  or  1, 4, 9, 16, ...  or  1, 8, 27, 64, ...  etc.  Again, he doesn't give and results, but the first two give divergent series, while the later two are convergent.

Daniel Bernoulli.  Specimen theoriae novae de mensura sortis.  Comm. Acad. Sci. Imp. Petropol. 5 (1730-31(1738)) 175-192, ??NYS.  IN: Die Werke von Daniel Bernoulli; ed. by L. P. Bouckaert & B. L. van der Waerden; Birkhäuser, 1982; pp. 223-234 and notes by van der Waerden, pp. 195 & 197-200.  English translation in Econometrica 22 (1954) 23-36, ??NYS.

Lewis Carroll.  Lionel Tollemache, Reminiscences of Lewis Carroll, Literature (5 Feb 1898), ??NYS, quoted in Carroll-Wakeling II, prob. 28: A good prospect, pp. 44 & 72.  (Tollemache was at Balliol College, Oxford, in 1856-1860.  He then entered Lincoln's Inn, London, so this must refer to c1858. 

                    He says that Carroll gave the problem with a coin and pay-offs of  0, 1, 3, 7, 15, ...  if the player first throws a tail on throw  1, 2, 3, 4, 5, ....  Neither Tollemache nor Wakeling give any reference to any other version of the problem.  I would compute the expectation as 

                    Σn=0 (2n-1)(1/2)n+1  =  0/2 + 1/4 + 3/8 + 7/16 + 15/32 + ....

          Wakeling does it by rewriting the sum as 

                    1/4 + (1+2)/8 + (1+2+4)/16 + (1+2+4+8)/32 + ...,  and regrouping as: 

                    1/4 + 1/8 + 1/16 + 1/32 + ... + 2/8 + 2/16 + 2/32 + ... + 4/16 + 4/32 + ... + 8/32 + ...  =  1/2 + 1/2 + 1/2 + 1/2 + ...  and says "Hence, his prospects are ½d. for every successive throw of a head."  Basically, making the payoff be a sum means that the expectation is a sum over half a quadrant.  The normal process would be to first sum the finite rows and see that each is at least  1/4.  Wakeling is first summing the columns and finding each column sum is  1/2.  I wonder if Carroll carefully chose the peculiar pay-offs in order to make the latter process work.

Leonhard Euler.  Vera aestimatio sortis in ludis.  [Op. postuma 1 (1862) 315-318.]  = Op. Omnia (I) 7 (1923) 458-465.

Tissandier.  Récréations Scientifiques.  1880?  2nd ed., 1881, p. 140 gives a brief unlabelled description, saying this "problème de Pétersbourg" was discussed by Daniel Bernoulli in "Mémoires de l'Académie de Russie".  Not in the 5th ed. of 1888.

Tissandier.  Popular Scientific Recreations.  1890?  Pp. 727-729 discusses the idea and says D. Bernoulli presented his material on this in "Memoires [sic] de l'Académie de Russie".  This is somewhat longer than the material in the 2nd French ed of 1881.

Anon. [presumably the editor, Richard A. Proctor].  Strange chances.  Knowledge 10 (Oct 1887) 276-278.  Brief discussion of "the famous Petersburg problem".

C. S. Jackson.  Note 438:  The St. Petersburg problem.  MG 8 (No. 116) (Mar 1915) 48.  Notes that the value of the game is not more than  log2 of the bank's funds.

Dan Pedoe, The Gentle Art of Mathematics, op. cit. in 5.C, 1958, p. 55, says D. Bernoulli published it in the Transactions (??) of of the St. Petersburg Academy.  ??NYS.

Pedoe, ibid., p. 57, also says that Buffon tested this with  2048  games and he won  10,057  in them.

Jacques Dutka.  On the St. Petersburg paradox.  Archive for the History of the Exact Sciences 39 (1988) 13-39.  ??NYR.

Nick Mackinnon  &  5Ma.  Note 74.9:  A lesson on the St. Petersburg paradox.  MG 74 (No. 467) (1990) 51‑53.  Suppose a maximum is put on the payment -- i.e. the game stops if  n  heads appear in a row.  How does this affect the expected value?  They find  n = 10  gives expected value of  6.

 

          8.F.    PROBLEM OF POINTS

 

            A game consists of  n  points.  How do you divide the stake if you must quit when the score is  a  to  b  to ....?  This problem was resolved by Fermat and Pascal in response to a question of the Chevalier de Méré and is generally considered the beginning of probability theory.  The history of this topic is thoroughly covered in the first works below, so I will only record early or unusual occurrences. 

            NOTATION:  Denote this by  (n; a, b, ...).

 

                    GENERAL HISTORIES

 

Florence Nightingale David.  Games, Gods and Gambling.  The origins and history of probability and statistical ideas from the earliest times to the Newtonian era.  Griffin, London, 1962.

Anthony W. F. Edwards.  Pascal's Arithmetical Triangle.  Griffin & OUP, London, 1987.  This corrects a number of details in David.

David E. Kullman.  The "Problem of points" and the evolution of probability.  Handout from talk given at MAA meeting, San Francisco, Jan 1991.  Uses  (6; 5, 3)  as a common example and outlines approaches and solutions of:  Pacioli (1494) -- 5 : 3;  Cardan (1539) -- 6 : 1;  Tartaglia (1539) -- 2 : 1;  Pascal (1654) -- 7 : 1;  Fermat (1654) -- 7 : 1;  Huygens (1657) -- like Pascal.  He also describes the approaches of Pascal, Fermat and Huygens to the three person case and generalizations due to Montmort (1713) and de Moivre (1718). 

 

Pacioli.  Summa.  1494. 

Ff. 197r-197, prob. 50 (?-not printed).  (6; 5, 2).  Discusses it and says there are diverse opinions.  His discussion is confusing, but he divides as  5 : 2  (should be  15 : 1).  Also says losing cardplayers at  (5; 4, 3)  offer to divide in the ratio  3 : 2  (should be  3 : 1).  His discussion is again confusing, but may be saying that one more game makes either a win or an even situation, and averaging these gives the right division as  3 : 1

F. 197v, prob. 51.  (6; 4, 3, 2).  Divides as  4 : 3 : 2  (I get  451 : 195 : 83). 

Oystein Ore.  Pascal and the invention of probability theory.  AMM 67 (1960) 409‑419.  Ore says Pacioli is the first printed version of the problem.  He translates parts of the texts.  Ore says he has seen the problem in Italian MSS as early as 1380, but he doesn't give details (???).  He opines that the problem is of Arabic origin.  He discusses Cardan and Tartaglia and gives some examples from Forestiani, 1603 -- (8; 5, 3),  (14; 10, 8, 5)  (??NYS).  But there is no proper mathematics until Pascal & Fermat.

Calandri, Raccolta.  c1495. 

Prob. 12, pp. 13‑14.  (6; 4, 3).  Divides in ratio  3 : 2,  but says this may not be the exact truth.  (Answer should be  11 : 5.)

Prob. 43, pp. 39‑40.  (3: 2, 1, 0).  He says there are two ways to do this, based on the numbers of points won or needed to win.  He then says  3/7  of the game has been played and distributes  3/7  of the stake in the proportion  2 : 1 : 0  and then distributes the remaining  4/7  equally, giving a final distribution in the proportion  10 : 7 : 4.  (Should be  19 : 6 : 2.)

Cardan.  Practica Arithmetice.  1539.

Chap. 61, section 13, f. T.iii.r (p. 113).  (10; 9, 7).  He divides as   1 + 2 + 3 : 1.  (10; 3, 6)  --  divided as   1 + 2 + 3 + 4  :  1 + ... + 7.   Section 14, f. T.iii.v (p.113) may be discussing this problem. 

Chap. 68 (ultimo), section 5, ff. QQ.iv.r - QQ.viii.r (p. 214).  This chapter is on errors of Pacioli.  Mentions  (6; 5, 2)  and many other examples.

Tartaglia.  General Trattato, 1556, art. 206, pp. 265r‑265v.  (6; 5, 3).  Criticises Luca del Borgo (= Pacioli) and gives another method.

Giovanni Francesco Peverone.  Due brevi e facili trattati, il primo d'Arithmetica, l'altro di Geometria ..., Gio. Tournes, Lyons, 1558.  Reissued as: Arithmetica e geometria del Sig. Gio. Francesco Peverone di Cuneo, ibid., 1581.  ??NYS -- described in: Livia Giacardi & Clara Silvia Roero;  Bibliotheca Matematica  Documenti per la Storia della Matematica nelle Biblioteche Torinese;  Umberto Allemandi & C., Torino, 1987, pp. 117-118.  They say he includes correct solutions of some problems of games of chance, in particular the 'divisione della posta', i.e. the problem of points.

Ozanam.  1694.  Prob. 10, 1696: 41-52, esp. 45-50;  1708: 37-48, esp. 42‑45.  Prob. 13, 1725: 123‑130.  Prob. 3, 1778: 117-121;  1803: 116-120;  1814: 102-106;  1840: 54-55.  Discusses the problem in general and specifically  (3; 1, 0),  but 1778 et seq. changes to  (3; 2, 1)  and adds reference to Pascal and Fermat.

Pierre Rémond de Montmort.  Essai d'analyse sur les jeux de hazards.  (1708);  Seconde edition revue & augmentee de plusieurs lettres,  (Quillau, Paris, 1713 (reprinted by Chelsea, NY, 1980));  2nd issue, Jombert & Quillau, 1714.  Avertissement (to the 1st ed.), pp. xxi-xxiv and (to the 2nd ed.) xxv-xxxvii discusses the history of the problem, the work of Fermat and Pascal and de Moivre's assertion that Huygens had solved it first.

Chr. Mason, proposer; Rob. Fearnside, solver.  Ladies' Diary, 1732-33  =  T. Leybourn, I: 223, quest. 168.  (15; 10, 8, 5).  I haven't checked the solution, but the procedure is correct and another solver got the same results.  Editor cites De Moivre.

Pearson.  1907.  Part II, no. 98, pp. 134 & 210-211.  (3; 2, 1).  He divides correctly as  3 : 1.

Hummerston.  Fun, Mirth & Mystery.  1924.  Marbles, Puzzle no. 26, pp. 71 & 176.  (4; 3, 2).  He divides correctly as  3 : 1.

 

          8.G.   PROBABILITY THAT THREE LENGTHS FORM A TRIANGLE

 

          See also 8.C.

 

E. Lemoine.  Sur une question de probabilités.  Bull. Soc. Math. France 1 (1872‑1873) 39‑40.  Obtains  ¼  by considering that the stick can be broken at  m  equidistant points and then letting  m  increase.

? Halphen.  Sur un problème de probabilités.  Ibid., pp. 221‑224.  Extends Lemoine to  n  pieces, getting  1 ‑ n/2n‑1,  by an argument similar to homogeneous coordinates and by integration.

Camille Jordan.  Questions de probabilités.  Ibid., pp. 256‑258 & 281‑282.  Generalizes to find the probability that  n  of the  m  parts, into which a line is broken, have length  > a.  He finds the probability that four points on a sphere form a convex spherical quadrilateral.  Pp. 281‑282 corrects this last result.  [Laquière; Note sur un problème de probabilité; ibid. 8 (1879-80) 79-80 gives a simple argument.]

M. Laquière.  Rectification d'une formule de probabilité.  Bull. Soc. Math. France 8 (1879-80) 74-79.  Treats the first problem of Jordan.  Observes that Jordan's formula can give a probability greater than one!  Says Jordan has a confusion between  'some n'  and  'a given n'  and he gives a corrected version. 

E. Lemoine.  Quelques questions de probabilités résolues géométriquement.  Bull. Soc. Math. France 11 (1882-83) 13-25.  Refers to his article in vol. 1 and the many resulting works.  Takes a point in a triangle and asks for the probability that the three lengths to the sides form a triangle (an acute triangle).  Then says that breaking a stick corresponds to using an equilateral triangle, giving probabilities  ¼  and  log 8 - 2  =  .0794415....  Makes various generalizations.  Takes a point,  M,  in an equilateral triangle  ABC  and asks the probability that  MA, MB, MC  form an acute triangle, getting  4 - 2π/Ö3  =  .3718....

E. Fourrey.  Curiositiés Géometriques, op. cit. in 6.S.1.  1907.  Part 3, chap. 1, section 5: Application au calcul des probabilitiés, pp. 360‑362.  Break a stick into three pieces.  Gets  P = ¼.  Cites Lemoine.

G. A. Bull.  Note 2016:  A broken stick.  MG 32 (No. 299) (May 1948) 87‑88.  Gets Halphen's result by using homogeneous coordinates.

S. Rushton.  Note 2083:  A broken stick.  MG 33 (No. 306) (Dec 1949) 286‑288.  Repeats Bull's arguments and then considers making one break, then breaking the larger piece, etc.  (It's not clear if he takes the longer of the two new pieces when trying for a  4‑gon.)  Says that the only example of this that he has seen is Whitworth's DCC Exercises in Choice and Chance, 1897, exer. 677, ??NYR, which has  Prob(triangle) = ⅓.  Author says this is wrong and should be  2 log 2 ‑ 1 = .386...  He gets a solution for an  n‑gon.

D. N. Smith.  Letter:  Random triangles.  MiS 19:2 (Mar 1990) 51.  Suggests, as a school project, generating random triangles by rolling three dice and using the values as sides.

Joe Whittaker.  Random triangles.  AMM 97:3 (Mar 1990) 228-230.  Take a stick and break it at two random points -- or -- break once at random and then break the longer part at random.  Prob(triangle) = ¼  in the first case and appears to be    in the second case, but the second analysis assumes an incorrect distribution.  Correcting this leads to   Prob(triangle)  =  2 log 2 - 1  =  .38...,  as in Rushton.

 

          8.H.   PROBABILITY PARADOXES

 

          8.H.1.          BERTRAND'S BOX PARADOX

 

J. Bertrand.  Calcul des Probabilités.  Gauthier‑Villars, Paris, 1889.  Chap. I, art. 2, pp. 2‑3.

Howard P. Dinesman.  Superior Mathematical Puzzles.  Op. cit. in 5.B.1.  1968.  No. 26: Mexican jumping beans, pp. 40-41 & 96.  Deranged matchboxes of red and black beans -- see 5.K.1.  The problem continues by unlabelling the boxes -- if you draw a red bean, what is the probability that the other bean in the box is red?

Nicholas Falletta.  The Paradoxicon.  Doubleday, NY, 1983;  Turnstone Press, Wellingborough, 1985.  Probability paradoxes, pp. 116‑125, esp. pp. 118‑121, which describes:  a three‑card version due to Warren Weaver (1950),  the surprise ace paradox of J. H. C. Whitehead (1938)  and  a three prisoner paradox.

Ed Barbeau.  The problem of the car and goats.  CMJ 24:2 (Mar 1993) 149-154.  A version of the problem involving three doors with a car behind one of them appeared in Marilyn vos Savant's column in Parade magazine and generated an immense amount of correspondence and articles.  This article describes  4  (or more?) equivalents and gives 63 references, including Bertrand, but not Dinesman, Falletta, Weaver or Whitehead.

 

          8.H.2.          BERTRAND'S CHORD PARADOX

 

J. Bertrand.  Op. cit. in 8.H.1.  1889.  Chap. I, art 5, pp. 4‑5.  Gets answers  ¼,    and  ½.

F. Garwood & E. M. Holroyd.  The distance of a "random chord" of a circle from the centre.  MG 50 (No. 373) (Oct 1966) 283‑286.  Take two random points and the chord through them.  This gives an expected distance from the centre of  .2532.

 

          8.I.     TAKING THE NEXT TRAIN

 

            This is the problem where a man can board equally frequent trains going either way and takes the next one to appear, but finds himself going one way more often than the other.  Why? 

            New section.

 

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.  No. 42: Bus times, pp. 28 & 116.  Two bus lines running the same route equally often, but it is twice as likely that the next bus is Red rather than Green.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 14: The absent-minded professor, pp. 25 & 184.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Fifth, pp. 146 & 197. 

 

          8.J.    CLOCK PATIENCE OR SOLITAIRE

 

            The patience or solitaire game of Clock has  13  piles of four face-down cards arranged with  12  in a circle and the  13th pile in the centre.  You turn up a card from the top of the  13th pile -- if it has value  n,  you place it under the  n-th pile and turn up the card from the top of that pile and repeat the process.  You win if you turn over all the cards.  The probability of winning is precisely  1/13  since the process generates an arbitrary permutation of the  52  cards and is a win if and only if the last card is a  K  (i.e. a 13).  In response to a recent question, I looked at my notes and found there are several papers on this.

 

John Reade, proposer;  editorial solution.  Problem 46.5 -- Clock patience.  M500 46 (1977) 17  &  48 (1977) 16.  What is the probability of winning?  Solution says several people got  1/13  and the problem is actually easy.

Anon. proposal  &  solution,  with note by David Singmaster.  Problem 11.3.  MS 11 (1978/79) 28  &  101.  (a)  What is the probability of winning?  Solution as in my comment above.  (b)  If the bottom cards are a permutation of  A, 2, ..., K,  then the game comes out if and only if this is a cyclic permutation.  Singmaster notes that the probability of a permutation of  13  cards being cyclic is  1/13,  so the probability of winning in this situation is again  1/13.

Eric Mendelsohn & Stephen Tanny, proposers;  David Kleiner, solver.  Problem 1066 -- The last  1.  MM 52 (1979) 113  &  53 (1980) 184-185.  Generalizes to  k  copies of  L  ranks.  Asks for probability of winning and for a characterization of winning distributions.  Solution is not very specific about the distributions, just saying there is a correspondence between the cards turned up and the original piles.

T. A. Jenkyns & E. R. Muller.  A probabilistic analysis of clock solitaire.  MM 54 (1981) 202‑208.  Says the expected number of cards played is  42.4.  They consider continuing the game after the last  K  by restarting with the first available unturned card.  They call this the 'second play' and allow you to continue to third play, etc.  They determine the expected numbers of cards turned in each play and also generalize to  m  cards of  n  ranks.  They show that the number of plays is determined by the relative positions of the last cards of each rank and show that the probability that the game takes  p  plays, but fail to note that this is  │s(n,p)│/n!,  where  s(n,p)  is the Stirling number of the first kind, so they rederive a number of properties of these numbers.

Michael W. Ecker.  How to win (or cheat) in the solitaire game of "Clock".  MM 55 (1982) 42-43.  Shows that whether you can win is determined by the bottom cards of each pile.  Though these values are not a permutation, one can define  'f-cycles' and a distribution comes out if and only if every  f-cycle contains the initial value.

 

          8.K.   SUCKER BETS

 

            This covers situations where the punter can not easily tell if the bet is reasonable or not.  These are often used to lure suckers, but they have also been historically important as incentives to develop probabilistic methods.  As an example, the Chevalier de Méré knew that the probability of throwing one six in four throws of a die was better than even, but he thought this should make throwing a double six in  24  throws also better than even and it is not.  See the histories cited in 8.F for more on the early examples of these problems.  Of course lotteries come into this category.  See 8.L for some other forms.

 

The classic carnival game of Chuck-a-Luck is an excellent example of this category.  This is the game with three die.  You bet on a number -- if it comes up once, you win double your bet (i.e. your bet and the same again);  if it comes up twice, you win triple;  if it comes up thrice, you win quadruple.  The relative frequency of  0, 1, 2, 3  of your number is  125, 75, 15, 1,  so the return in  216  throws will be  ‑125 + 75 + 30 + 3  =  ‑17,  giving a  7.9%  profit to the operator.

Collins.  Fun with Figures.  1928.  How figures can cheat -- The tin-horn gambler, pp. 34-35.  Six dice, each marked on just one face.  Gambler bets  $100  to  $1  that the punter will not get the six marked faces all up in  20  throws.  Collins asserts that the probability of winning is   20/66  =  1/2332.3,   so the fair odds should be  $2332  to  $1.  This is not quite right.  The true probability is   1 - (1 - 1/66)20  =  .00042858  =  1/2333.27.  The fair odds depend on whether the winnings include the punter's  $1  or not -- in this case, they do not, so the odds according to Collins ought to be  $2331.3  to  $1,  or more exactly  $2332.27  to  $1.

In the 1980s, I saw stands at school and village fairs, where one gets one throw with six dice for  £.50.  If one gets six  6s,  one wins a new Rover.  Since  66 = 46656,  the promoter gets an average of  £23,328  for each car, which was a tidy profit.

Gardner.  Nontransitive dice and other probability paradoxes.  SA (Dec 1970).  Extended in Wheels, chap. 5.  Consider two red and two black cards.  Choosing two, what are the odds of getting two of the same colour?  Typical naive arguments get    or  ½,  but the true answer is  ⅓.  In the Addendum in Wheels, S. D. Turner describes the version with  R  red cards and  B  black cards.  The probability of choosing two of the same colour is then  [R(R-1) + B(B-1)]/[(R+B)(R+B-1)]  which is always less than  ½. 

 

          8.L.    NONTRANSITIVE GAMES

 

            In the simplest form, we know that  A  can defeat  B  and  B  can defeat  C,  but this does not imply that  A  can defeat  C,  so non-transitivity is quite common in real game playing.  Indeed, in the classical game of Rock, Scissors, Paper, the game situation has  C  defeating  A.  Similar phenomena occur in preferences, particularly voting and loving.  This section will generally deal with mathematical versions, particularly where the game seems fair, but making a later choice than your opponent gives you an edge.  Hence these versions can be used as the basis of sucker bets.

 

In about 1932, the following golf scores were sent to The Scotsman.

                    A   4 5 5 6 9 4 5 4 4 ‑ 46

                    B   5 6 6 7 3 5 6 3 5 ‑ 46

                    C   6 7 7 8 6 3 3 3 3 ‑ 46

          A  beats  B  by 4 and 3,  B  beats  C  by 5 and 4,  C  beats  A  by one hole.  A correspondent, J. C. Smith, suggested the following series for three holes.

                    A   1 2 3 ‑ 6   A  one up on  B.

                    B   2 3 1 ‑ 6   B  one up on  C.

                    C   3 1 2 ‑ 6   C  one up on  A.

          Similar results can be obtained for four men playing four holes, and so on. 

                    Reported by J. W. Stewart as Gleaning 854: Golf Scores; MG 16 (No. 218) (May 1932) 115.

Walter Penney, proposer and solver.  Problem 95 -- Penney-ante.  JRM 2:4 (Oct 1969) 241  &  7:4 (Fall 1974) 321.  Opponent picks a triple of heads and tails, then you pick a triple.  A coin is thrown until the triple occurs.  If he chooses  HHH  and you choose  HTH,  show your probability of winning is  3/5.

Walter Penney and David L. Silverman, proposers and solver.  Problem 96 -- Penney-ante.  JRM 2:4 (Oct 1969) 241  &  8:1 (1975) 62-65.  As above, but the opponent and you both pick a triple without the other's knowledge.  Elaborate analysis is required to obtain the optimal mixed strategy which guarantees you a probability of  ½.

Gardner.  Nontransitive dice and other probability paradoxes.  SA (Dec 1970).  Extended in Wheels, chap. 5.  Describes Bradley Efron's sets of 4 nontransitive dice which give the second chooser a    chance of winning.  Efron says it had been proven that this is the maximum obtainable with four dice.  For three dice, the maximum is  .618,  but this requires dice with more than six faces.  As the number of dice increases, the maximum value approaches ¾.  The Addendum in Wheels describes numerous variants developed by magicians and mathematicians.

Gardner.  Nontransitive paradoxes.  SA (Oct 1974)  c= Time Travel, chap. 5.  Discusses voting paradoxes and describes examples of nontransitive behaviour back to mid-20C.  Describes Penney's game, giving it with pairs first and showing that for each choice by the opponent, you can pick a better choice, with probability of winning being at least  ⅔.  The bibliography in Time Travel is extensive and Gardner notes that some of the items give many further references.

Richard L. Tenney & Caxton C. Foster.  Non-transitive dominance.  MM 49:3 (May 1976) 115-120.  Gives three dice with odds for the second chooser being  5/9.  Gives proofs for the results stated by Efron in Gardner, above, observing that they arise from results known about voting paradoxes.

 

 

 

9.       LOGICAL RECREATIONS

 

          Many combinatorial recreations can be considered as logical.

 

          9.A.   ALL CRETANS ARE LIARS, ETC.

 

Diogenes Laërtius.  3C.  De Clarorum Philosophorum Vitis, Dogmatibus et Apophtegmatibus, II, Life of Euclides.  Ed. by C. G. Cobet; Paris, 1888, p. 108, ??NYS.  Translated by C. D. Yonge; Bell, London, 1894, pp. 97‑98.  Translated by R. D. Hicks; Loeb Classical Library; vol. 1, pp. 236‑237.  Refers to Eubulides (c‑330) as the source of "The Lying One" or "The Liar" -- Ο Ψεθδoμεvoσ (O Pseudomenos).  According to:  I. M. Bochenski; Ancient Formal Logic; North Holland, Amsterdam, 1951, p. 100;  Eubulides also invented:  "the swindler",  "the concealed",  "the heap" (how many grains make a heap?),  "the Electra"  and  "the horned" (equivalent to "Have you stopped beating your wife?").

Bochenski, Ancient Formal Logic, pp. 101‑102, says the liar paradox was unknown to Plato, but is quoted by Aristotle in his  Libro de Sophisticis Elenchis 25, 180 b 2‑7, ??NYS, (See:  M. Wallies, ed.; Topica cum Libro De Sophisticis Elenchis; Leipzig, 1923; ??NYS;  and:  Ethica Nicomachea, H3, 1146 a 21‑27; ed. by Fr. Susemihl; Leipzig, 1887; ??NYS.)  It is also given in  Cicero; Ac. Pr. II, 95, 96 (=? Topica, 57); ??NYS (In:  G. Friedrich, ed.; Opera Rhetorica; Leipzig, 1893; ??NYS)  and in many later writers.

Athenaeus Naucratica.  c200.  The Deipnosophists, Book 9 (end of c.64).  Translated by C. D. Yonge, Bohn, London, 1854, vol. 2, p. 633.  Epitaph of Philetas of Cos (c‑340/c‑285).  "Traveller, I am Philetas; the argument called the Liar and deep cogitations by night, brought me to death."  Sadly, there is no indication where he died or was buried.  (Bochenski; Ancient Formal Logic,; p. 102 gives the Greek of Athenaeus.  I. M. Bochenski; History of Formal Logic; corrected ed., Chelsea, 1970, p. 131; gives the English.)

The Stoics.  c‑280.  Bochenski; Ancient Formal Logic; pp. 100‑102 says they invented several paradoxes, including "the crocodile" who takes a baby and says he will return it if the mother answers his question correctly.  He then asks  "Will I return the baby?"  She answers  "No".

Anon.  History of the Warring States.  [The Warring States period is ‑475/‑221 and this history may be ‑2C.]  The Elixir of Death.  Translated in:  Herbert A. Giles; Gems of Chinese Literature; op. cit. in 6.BN, p. 43.  Chief Warden swallows an elixir of immortality which he was supposed to convey to the Prince.  The Prince orders the Warden's execution, but the Warden argues that if the execution succeeds, then the elixir was false and he is innocent of crime.  The Prince pardons him.

Tung‑Fang So (‑2C, see Giles, ibid., p. 77) is said to have been in the same situation as the Warden and argued:  "If the elixir was genuine, your Majesty can do me no harm;  if it was not, what harm have I done?"

St. Paul.  Epistle to Titus, I, 12.  c50?  "One of themselves, even a prophet of their own, said, The Cretans are always liars, ....  This witness is true."

M. Cervantes.  Don Quixote.  1605.  Book II, chap. 51.  Translated by Thomas Shelton, 1612‑1620, reprinted by the Navarre Society, London, 1923, vol. 2, pp. 360‑362.  Sentinel paradox: truthtellers pass; liars will be hanged.  "I will be hanged."

Henri Decremps.  Codicile de Jérôme Sharp, ....  Op. cit. in 4.A.1.  1788.  Avant Propos, pp. 19-20: Sentinel paradox: truthtellers pass; liars will be thrown in the river.  "You will throw me in the river."  Author says he will give the answer in another volume.

Henri Decremps.  Les Petits Aventures de Jerome Sharp.  Professor de Physique Amusante; Ouvrage contenant autant de tours ingénieux que de leçons utiles, avec quelques petits portraits à la maniére noire.  Brussels, 1789;  also 1790, 1793.  Toole Scott records an English edition, Brussels, 1793.  Sentinel paradox.  ??NYS.  Cited by Dudeney; Some much‑discussed puzzles; op. cit. in 2; 1908; as the first appearance of this paradox.

The Sociable.  1858.  Prob. 43: The Grecian paradox, pp. 299 & 317.  Protagoras suing his pupil who had promised to pay for his tuition when he won his first case.  = Book of 500 Puzzles, 1859, prob. 43, pp. 17 & 35.  c= Magician's Own Book (UK version), 1871, pp. 26-27.  c= Mittenzwey, 1895?: prob. 157, pp. 33 & 81;  1917: 157, pp. 30 & 79,  which claims the teacher wins.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-21, pp. 253 & 395.  Form of the sentinel paradox.  To enter a garden, one must make a statement;  if true, one pays 3 marks;  if false, one pays 6 marks.  "I will pay 6 marks."

Carroll-Wakeling II.  c1890?  Prob. 32: Bag containing tickets, pp. 50 & 73.  This is one of the problems on undated sheets of paper that Carroll sent to Bartholomew Price.  Wakeling reproduces the MS. 

                    A bag contains 12 tickets, 3 marked 'A', 4 'B', 5 'C'.  One is drawn in the presence of 12 witnesses of equal credibility: three say it was 'A', four 'B', five 'C'.  What is the chance that it was 'A'?

          There is no answer on the Carroll MS.  Wakeling gives an answer.

                              Let the credibility of a witness be "a" when telling the truth.  Hence, the credibility of a witness when telling a lie is "1 - a".

                              If it was  A,  then 3 tell the truth, and 9 lie; hence the credibility is 3 in 12, or 1 in 4.

                              Therefore the chance that it is  A,  and no other, is:

                                        3/12 x 1/4  +  4/12 x 3/4  +  5/12 x 3/4  =  5/8

          I cannot see how this last formula arises.  Wakeling writes that he was assisted in this problem by a friend who has since died, so he does not know how the formula was obtained. 

                    Assuming  a  is the probability of a person telling the truth, this probability depends on what the chosen ticket is and is not really determined by the information given -- e.g., if the ticket was  A,  then any value of  a  between  0  and  1  is possible.  The value  3/12  is an estimate of  a,  indeed the maximum likelihood estimate.  If  k  of the  12  people are telling the truth, I would take the situation as a binomial distribution.  There are  BC (12, k)  ways to select them and the probability of having  k  liars is then  BC (12, k) ak (1-a)12-k.  Now it seems that Bayes' Theorem is the most appropriate way to proceed.  Our basic events can be denoted  A, B, C  and their a priori probabilities are  3/12, 4/12, 5/12.  Taking  a = k/12,  the a posteriori  probabilities are proportional to 

          k/12 x BC (12, k) (k/12)k ({12-k}/12)12-k,  for  k = 3, 4, 5.  Dropping the common denominator of  1213,  these expressions are  6.904,  8.504,  10.1913  times  1012.  Dividing by the total gives the a posteriori probabilities of  A, B, C  as

           27.0%,  33.2%,  39.8%.

                    This is really a probability problem rather than a logical problem, but it illustrates that the logical problem seems to have grown out of this sort of probability problem.

Lewis Carroll.  Diary entry for 27 May 1894.  "I have worked out in the last few days some curious problems on the plan of 'lying' dilemma.  E.g. 'A says B lies;  B says C lies;  C says A and B lie.'  Answer: 'A and C lie;  B speaks truly'.  The problem is quoted in Carroll-Gardner with his discussion of the result, pp. 22-23.  Gardner says this was printed as an anonymous leaflet in 1894.

                    Carroll-Wakeling.  Prob. 9: Who's telling the truth?, pp. 11 & 65.  Wakeling says "This is a puzzle based on a piece of logic that appears in his diary.

                    The Dodo says the Hatter tells lies.

                    The Hatter says that the March Hare tells Lies.

                    The March Hare says that both the Dodo and the Hatter tell lies.

                    Who is telling the truth?"

                    In a letter of 28 May 2003, Wakeling quotes the text from the Diary as above, but continues with it:  "And today 'A says B says C says D lies;  D says two lie and one speaks true.'  Answer: 'D lies; the rest speak truly.'  Wakeling adds that the Diary entry for 2 Jun 1896 (not given in the Lancelyn Green edition) says:  "Finished the solution of the hardest 'Truth-Problem' I have yet done",  but Carroll gives no indication what it was.

Lewis Carroll.  The problem of the five liars.  In his unpublished Symbolic Logic, Part II.  He was working on this after Part I appeared in 1896 and he had some galley proofs when he died in 1898.  Published in Lewis Carroll's Symbolic Logic, ed. by William Warren Bartley III; Clarkson N. Potter, NY, 1977,

Pp. 352‑361, including facsimiles of several letters to John Cook Wilson (not in Cohen).  Each of five people make two statements, e.g. A says "Either B or D tells a truth and a lie;  either C or E tells two lies."  When analysed, one gets contradictions because a form of the Liar Paradox is embedded.

Pp. 423-444 is a survey of logical paradoxes with some variations by Carroll.

Cesare Burali Forti.  Una questione sui numeri transfiniti.  Rendiconti del Circolo Matematico di Palermo 9 (1897) 154-164.  ??NYS.  This was the first published antinomy of modern set theory.  The set of all ordinal numbers is itself an ordinal!  However, Cantor had observed the paradox in 1895 and communicated it to Hilbert in a letter in 1896. 

Irving Anellis.  The first Russell paradox.  Paper given at AMS meeting, Chicago, Mar 1985.  ??NYS -- abstract given in  HM 12 (1985) 380.  Says it is usually believed that Russell discovered his paradox in Jun 1901, but he sent a version of it to Couturat on 8 Dec 1900 (unpublished MS in the Russell Archives, McMaster Univ.).

Gregory H. Moore.  A house divided against itself: The emergence of first-order logic as the basis for mathematics.  IN:  Esther R. Phillips, ed.; Studies in the History of Mathematics; MAA, 1987, pp. 98-136.  On pp. 114-115, he dates Russell's paradox to May 1901 and says Russell wrote about it to Frege on 16 Jun 1902.  The first publications are in Russell's Principles of Mathematics and Frege's Fundamental Laws, vol. 2, both in 1903.

B. Russell.  The Principles of Mathematics.  CUP, 1903.  ??NYS -- cited in Garciadiego.  He discusses Russell's paradox and also Cantor's paradox of the greatest cardinal and Burali Forti's paradox of the greatest ordinal.  I won't consider these much further, but this may have inspired the development of the more verbal paradoxes described in this section.

G. G. Berry.  Letter to Russell on 21 Dec 1904.  In the Russell Archives, McMaster University.  Quoted in Garciadiego.  "... the least ordinal which is not definable in a finite number of words.  But this is absurd, for I have just defined it in thirteen words."  The paradox of Jules Richard (late Jun 1905) is very similar and similar versions were found by J. König and A. C. Dixon about the same time, though these all use Zermelo's well-ordering axiom.  Sometime earlier, Berry had introduced himself to Russell with a note saying  "The statement on the other side of this paper is true"  with the other side reading  "The statement on the other side of this paper is false",  and consequently is also considered the inventor of the "visiting card paradox".

B. Russell.  Les paradoxes de la logique.  Revue de Métaphysique et de morale 14 (1906) 627‑650.  ??NYS -- cited by Garciadiego.  First publication of a modified version of Berry's paradox.

B. Russell.  Mathematical logic as based on the theory of types.  Amer. J. Math. 30 (1908) 222‑262.  On p. 223, he first gives Berry's paradox:  "the least integer not nameable in fewer than nineteen syllables".  He also reformulates König & Dixon as "the least indefinable ordinal".

Kurt Grelling & Leonard Nelson.  Bemerkungen zu den Paradoxien von Russell und Burali‑Forti.  Abhandlungen der Fries'schen Schule (NS) 2 (1908) 301‑344.  ??NYS.  Grelling's paradox:  "Is heterological heterological?"

A. N. Whitehead & B. Russell.  Principia Mathematica.  CUP, 1910.  Vol. 1, pp. 63‑64.  Discusses several paradoxes and repeats Berry's paradox.

F. & V. Meynell.  The Week‑End Book.  Op. cit. in 7.E.  1924.  2nd ed., prob. five, p. 275;  5th? ed., prob. nine, p. 408, gives Russell's paradox as a problem -- and gives no solution!

Hummerston.  Fun, Mirth & Mystery.  1924.  The bridge, pp. 68-69.  Sentinel paradox -- "I am going to be hanged on that gallows!"  Author says "It is impossible to answer ... satisfactorily.  Perhaps the best plan is to throw the varlet in the river."

John van Heijenoort.  Logical paradoxes.  Encyclopedia of Philosophy 5 (1967) 45‑51.  Excellent survey of the paradoxes of logic and set theory, but only a mention of pre‑19C paradoxes.

Alejandro R. Garciadiego.  The emergence of some of the nonlogical paradoxes of the theory of sets, 1903-1908.  HM 12 (1985) 337-351.  Good survey.  He has since extended this to a book:  Bertrand Russell and the Origins of the Set-theoretic "Paradoxes"; Birkhäuser, 1992, ??NYS

At the 19th International Puzzle Party in London, 1999, Lennart Green told the story of a friend of his who was such a failure in life that he decided to wrote a book on "How to be a Failure".  But if this failed, he would be successful and if it succeeded, he would have failed.

 

          9.B.   SMITH -- JONES -- ROBINSON PROBLEM

 

          See also 5.K.2 for a special form of these problems.

 

Dudeney.  PCP.  1932.  Prob. 49: "The Engine‑Driver's Name", pp. 24 & 132.  = 536; prob. 521: "The Engineer's Name", pp. 214 & 411.  The driver is Smith, but the other two names are not determined.

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 39, pp. 22 & 39.  Same as Dudeney.

Phillips.  Brush.  1936.  Prob. K.2: The Engine‑driver, pp. 36 & 96.  Same as Dudeney.

Rudin.  1936.  No. 183, pp. 65 & 119.  Similar to Dudeney, but Americanized, somewhat simplified(?) and asking for the brakeman's (= guard's) name.

Haldeman-Julius.  1937.  No. 4: Robinson-Smith-Jones problem, pp. 3 & 20.  Similar to Dudeney, but Americanized differently than in Rudin, asking for the engineer's name.  Says it was sent by J. C. Furnas.

James Joyce.  Finnegans Wake.  Viking Press, NY, 1939.  P. 302, lines 23-24:  "Smith-Jones-Orbison".

J. G. Oldroyd.  Mathematicians in the army.  Eureka  5 (Jan 1941) 6  &  6 (May 1941) 10.  Six men of different ranks from three different schools, colleges and faculties (i.e. subjects).

Irving Adler.  Thinking Machines.  Dobson, London, 1961.  Pp. 111-116: Who is the engineer?  Essentially identical to Rudin, but asks for the engineer's name.  Gives a systematic solution via boolean algebra.

Doubleday - 2.  1971.  Flight plan, pp. 153-154.  Same as Dudeney, slightly reordered.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  Who's who?, pp. 86 & 134-135.  Spaceship version, similar to Dudeney, but more precise.

 

          9.C.   FORTY UNFAITHFUL WIVES

 

          I now realise that this is an extension of 9.D.

          See Littlewood, 1953, in 9.D.

 

Gamow & Stern.  1958.  Forty unfaithful wives.  Pp. 20‑23.  (Communicated by V. Ambarzuminian.)

Michael Spivak.  Calculus.  2nd ed., Publish or Perish.  ??date, place.  P. 35, probs. 27 & 28. 

Uri Leron & Mike Eisenberg.  On a knowledge-related paradox and its resolution.  Int. J. Math. Educ. Sci. Technol. 18 (1987) 761-765.

Ed Barbeau.  Fallacies, flaws and flimflam.  CMJ 22:4 (Sep 1991) 307.  Gives a brief discussion and the reference to Spivak and to Leron & Eisenberg.  The paradox has to do with what information has been provided by the stranger.

 

          9.D.   SPOTS ON FOREHEADS

 

          See also 9.C.  7.AP is somewhat related.

 

William Wells Newell.  Games and Songs of American Children.  Harper and Brothers, (1883);  2nd ed. 1903;  reprinted with Editor's Note of 1883 and new Introduction and Index, Dover, 1963.  Chap. IX, No. 77: Laughter games, pp. 136‑137.  "In a Swiss game, ....  Each child pinches his neighbor's ear;  but by agreement, the players blacken their fingers, keeping two of the party in ignorance.  Each of the two victims imagines it to be the other who is the object of the uproarious mirth of the company."  The Notes on p. 274 indicate that this probably comes from:  E. L. Rochholz; Alemannisches Kinderleid und Kinderspiel, Leipzig, 1857, ??NYS

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 13, pp. 14 & 188.  Two boys fall down, one gets a dirty face, the other washes his own face.

Phillips.  The Playtime Omnibus.  Op. cit. in 6.AF.  1933.  Section XVII, prob. 11: Odd, pp. 55 & 237.  Identical to Week‑End.

W. E. Buker, proposer;  Robert Wood and O. B. Rose, solvers.  Science Question 686.  SSM 35 (1935) 212  &  429.  3 persons.

A. A. Bennett, proposer;  E. P. Starke and G. M. Clemence, solvers.  Problem 3734.  AMM 42 (1935) 256  &  44 (1937) 333‑334.  n  persons with smudges on foreheads.  Says the  3  person case was suggested by Dr. Church of Princeton and cites the Buker problem.

Phillips.  Brush. 1936.

Prob. A.2: The green crosses, pp. 1‑2 & 73.  Three men with green crosses.

Prob. R.1: The roof, pp. 58 & 112.  Same as in Week‑End, above.

Rudin.  1936.  No. 145, pp. 51-52 & 110.  3  persons with black crosses on foreheads, who whistle when they see a black cross.

Haldeman-Julius.  1937.  No. 76: The circle problem, pp. 10 & 24.  Three persons with red crosses, who put up their hands if they see a red cross.  "This problem promises to become famous.  It has been going the rounds during the past few weeks -- ....  We printed it several years ago, but believe it deserves reprinting."

McKay.  At Home Tonight.  1940.  Prob. 36: The five disks, pp. 70 & 83.  Three mandarins and five disks, two black and three white.  Emperor puts white on each forehead.  Answer argues using the number of blacks.

M. Kraitchik.  Mathematical Recreations.  Op. cit. in 4.A.2.  1943.  Chap. 1. 

Prob. 3: The problem of the three philosophers, p. 15.  Three painted faces. 

Prob. 4, pp. 15‑16.  3 white and 2 black discs --  three whites placed on backs.  All realize simultaneously. 

   (Neither is in Math. des Jeux.)

Harold Hart.  The World's Best Puzzles.  Op. cit. in 7.AS.  1943.  The problem of the marked foreheads, pp. 23 & 55.  Three students with blue and green crosses.

Leopold.  At Ease!  1943.  Short cut to chevrons, pp. 23-24 & 199.  Three men with smudged foreheads.

A. K. Austin.  A calculus for know/don't know problems.  MM 49:1 (Jan 1976) 12-14.  He develops a set-theoretic calculus for systematically solving problems involving spots on foreheads, etc., including problems with knowing sums, see 7.AP.  His typical problem has a man with four red and three blue stamps and he sticks three on the foreheads of two boys, telling them they each have at least one red.  The first says he doesn't know what he has; then the second says he doesn't know; then the first says he does know.  What stamps did the first have?

Leeming.  1946.  Chap. 3, prob. 9: The three small boys, pp. 21‑22 & 153‑154.  Three boys with smudged foreheads.

Henry Cattan.  The Garden of Joys.  (An anthology of oriental anecdotes, fables and proverbs.)  Namara Publications, London, 1979.  How he won the office of Grand Vizier, pp. 107‑109  &  note 89 on p. 114.  Same as Kraitchik's problem 4, but with the simpler solution based on symmetry.  The note says:  "This story is anonymous and was heard by the author in Palestine."  A letter from the author says he heard the story in this form in Palestine before 1948.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 51: Red hats and green hats, pp. 46 & 190-191.  2  red hats &  3  green hats.  Answer gives the symmetry solution and the logical solution without clearly recognizing the distinction.

Max Black.  Critical Thinking.  1952.  Op. cit. in 6.F.2.  Prob. 11, pp. 12 & 432.  Three sons with white marks.

J. E. Littlewood.  A Mathematician's Miscellany.  Op. cit. in 5.C.  1953.  Pp. 3‑4 (25‑26).  Three dirty faces.  Mentions that this can be extended to  n  dirty faces, which "has not got into the books so far as I know".  This may be the origin of 9.C?

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  Pp. 34-35: Clean and dirty.  Two men fall through a roof.  Man with clean face goes to wash.

T. J. Fletcher.  The  n  prisoners.  MG 40 (No. 332) (May 1956) 98‑102.  Considers Kraitchik's problem with  n  persons,  n  white discs and  n‑1  black discs.  Also studies various colour distributions and assignments, including  < n  white discs and use of three colours.

Gamow & Stern.  1958.  Three soot‑smeared faces.  Pp. 77‑79.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Hats Off!, pp. 108 & 193.  A, B, C  are seated in a row, so  B  can see  A  while  C  can see both  B  and  A.  They all know that there are three white and two red hats in a bag.  A hat is taken out and put on A's  head, but he can't see it.  Similarly, hat are taken and put onto  B  and  C.  C  is now asked it he knows what colour his hat is and replies that he does not.  B,  having heard  C's  response, is asked if he knows what colour his hat is and he replies that he does not.  Is  A, having heard these, able to know what colour his hat is?  Extends to various other combinations and to  n  people.  The answer is "The answer to all the questions is, yes, it is possible."

 

          9.E.    STRANGE FAMILIES

 

            Complicated questions of kinship have arisen due to religious taboos on incest.  Most religions have a list of kinship relations which are not permitted to marry.  These get a bit more involved than I want to go into.  See the items in the first section below for some typical material.

            The second section deals with marrying a deceased wife's sister.

            The third section deals with general strange families riddles and puzzles, but 'That man's father ...' are collected in 9.E.1.

            Ripley's Believe It or Not! books give a number of examples of strange families.  I will enter these under the date of the persons involved.  BION-xx  denotes the  xx-th series of Believe It or Not!

            Problems of this type are generally put in the form of a riddle, and many of these are collected in the following.

 

Mark Bryant.  Dictionary of Riddles.  Routledge, 1990.  (Based on his Riddles Ancient and Modern; Hutchinson, 1983.)

 

                    GENERAL STUDIES OF KINSHIP RELATIONS

 

J. Cashdan & Martin D. Stern.  Forbidden marriages from a woman's angle.  MG 71 (No. 456) (1987).  ??NYS -- cited by Stern, 1990.

Martin D. Stern.  Consanguinity of witnesses -- a mathematical analysis.  Teaching Mathematics and Its Applications 6:2 (1987).  ??NYS -- cited by Stern, 1990.

Martin D. Stern.  Mathematical motivation through matrimony.  MM 63:4 (Oct 1990) 231‑233.  ??NYS -- reproduced in Robert L. Weber; Science with a Smile; Institute of Physics, Bristol, 1992, pp. 314-318.  Presents a notation for kinship relations and uses it to see that the table of prohibited degrees of marriage given in the Book of Common Prayer is symmetric with respect to sex and hence there are no unexpected prohibitions.  However, the Jewish restrictions on marriage and on testimony by consanguineous relatives are not symmetric -- cf the above items.

Marcia Ascher.  Ethnomathematics.  Op. cit. in 4.B.10.  1991.  Chapter Three: The logic of kinship relations, pp. 66-83.  Gives a number of folk puzzles and then analyses several complicated kinship systems.  Some references.

Martin Stern.  Discrete avoidance of marital indiscretion.  Mathematics Review (Univ. of Warwick) 2:3 (Feb 1992) 8-11.  He presents a notation for kinship relations and uses it to describe the prohibited relations in Christian, Jewish and Islamic traditions.  The Jewish prohibitions are not symmetric between male and female.

Helen Cooper.  A little more than kin [Review of Elizabeth Archibald; Incest and the Medieval Imagination; OUP, 2001]; The Times Literary Supplement (26 Oct 2001) 27.  This notes that the taboos on incest are very variable.  Egyptian Pharaohs indulged in brother-sister marriages and the Macedonians did not understand why Oedipus was upset when he discovered he was married to his mother.  Medieval Christianity extended the family to include godparents, who were spiritual siblings, and even in-laws of in-laws, as well as all the relatives of lovers, since sex made the lovers 'one flesh'.  On the other hand, Henry VIII married his deceased brother's wife and later divorced her in order to marry the sister of his mistress.  Archibald notes that since Christ is God, Mary is "Maid and mother, daughter of thy son!" [Chaucer's translation of Dante].  The medieval legend of Judas makes him a version of Oedipus -- he unknowingly killed his father and married his mother.  The medieval period introduced the double incest version where the son of a brother-sister relation is sent away and returns as an adult and unknowingly marries his mother.  In 13C French versions of the Arthurian legend, Mordred, the nephew of Arthur, is made into his son, the result of a liaison between Arthur and his half-sister, who did not know of their relationship.  Mordred attempted to marry Arthur's wife.  Luther urged that if a wedded couple were later discovered to be brother and sister or half-sister, or even mother, that the knowledge should be suppressed lest it drive them to the ultimate sin of despair.

 

                    DECEASED WIFE'S SISTER, ETC.

 

E. S. Turner.   Roads to Ruin -- The shocking history of social reform.  Michael Joseph, London, 1950.  Chap. 5: Two wives, one mother-in-law, pp. 98-121.  This surveys the British preoccupation with the legality of marrying a deceased spouse's sibling.  Since a couple were considered to become 'one flesh' (Ephesians 5:31), such a marriage was considered incestuous by the Church.  Leviticus 18:6 & 16 were interpreted as prohibiting such marriage, but Leviticus 18:18 was interpreted as saying that the previous verses stated that a man should not have sisters as wives at the same time [which is the Islamic interpretation], while Deuteronomy 25:5-10 not only permits, but even commands, that a man should marry his brother's widow. 

                    The English preoccupation with the problem dates from Henry VIII's marriage to his brother's widow, Catherine of Aragon.  [In fact, he then divorced her to marry his mistress's sister.]  This particular question is mentioned in Shakespeare's Henry VIII and the general question is the basis of Hamlet, whose mother marries her dead husband's brother.  There was at least one execution, in early 18C Scotland, of a woman who had sex with her sister's widower.

                    Up to 1835, marriage to a deceased wife's sister was permitted, but it could be voided and the children declared bastards, if an action was brought.  But if an action was brought and dropped, further actions were prevented.  In 1835, the Duke of Beaufort, who had married his deceased wife's half-sister, persuaded the Lord Chancellor to introduce a bill to legitimize such marriages up to date.  The Bishops managed to amend this to prohibit such marriages in the future.  However, such a couple could go to Europe to be married and such marriages remained legal in places like Jersey, though they were not legitimate in England.  The Catholic Church generally gave dispensation for such marriages.  From 1841 onwards, bills to remove the prohibition were introduced in almost every Parliament.  Marriage to a deceased husband's brother or to a deceased spouse's nephew/niece was not sufficiently common to be considered by the reformers.  The question was mentioned by Gilbert & Sullivan (near the end of the first act of Iolanthe (1882), the Queen, referring to Strephon, says  "He shall prick that annual blister, / Marriage with deceased wife's sister;").  The journal Moonshine commented:  "To be able to marry two wives at the cost of but one mother-in-law is something to fight for."  In 1906, the Colonial Marriages Bill legitimized such marriages made in the colonies.  In 1907, the Deceased Wife's Sister Bill was passed.  Canon Law was later changed to accept this.  One man who had married his deceased wife's sister sued a Canon who refused him Communion and won, with his win being confirmed by the Court of Appeals and the House of Lords in 1912.  However, marriage to a divorced wife's sister was not permitted while the ex-wife lived.  Marriage to a deceased husband's brother was permitted in 1921.  A number of other marriages were permitted in 1921 and all these acts are consolidated in the Marriage Act of 1949.  Turner is not clear whether marrying a divorced spouse's sibling was permitted, and I don't know the further history.  A 2000 article says marrying a deceased wife's aunt or niece was permitted by the Marriage Act of 1931.

A discussion of Strawberry Hill House says marriage to a deceased husband's brother was prohibited by the 1835 act, so that Frances, the widow of John Waldegrave, had to go to Scotland to marry his half-brother George Waldegrave, 7th Earl Waldegrave.

Susan Kelz Sperling.  Tenderfeet and Ladyfingers.  Viking, NY, 1981, p. 98-99.  She gives some details of the Hebrew view.  The Hebrew law of yibbum declares that if a man dies without heir, his brother or nearest relative is obliged to marry the widow (i.e. she is marrying her dead husband's brother).  However, he could decline the duty by a ceremony called halitzah, as specified in Deuteronomy, by putting on a special shoe which the widow removed and then she spat in front of him to break the contract.  This takes place in Ruth, allowing her to marry Boaz.

BION-11 cites an American example where a woman successively married the widowers of two of her sisters.

William Holman Hunt, the Victorian painter, married his deceased wife's sister in 1866, in Switzerland [Judi Culbertson & Tom Randall; Permanent Londoners; Robson Books, London, 1991, p. 140].

See Dudeney, AM, prob. 52, below, for a complication of this situation.

Haldeman-Julius.  1937.  No. 88: Marriage problem, pp. 11 & 25.  How can a man have married his widow's sister?  (Also entered under General Family Riddles.)

Mindgames.  Frontiers (a UK popular science magazine) No. 1 (Spring 1998) 107.  "My wife's sister is also her cousin.  How can this be?"  Solution is that her father married his dead wife's sister and had another daughter.

 

                    GENERAL FAMILY RIDDLES

 

The riddles which the Queen of Sheba proposed to Solomon are not recorded in the biblical account of their meeting (I Kings 10 & II Chronicles 9), which would be c-960.  Josephus' History of the Jews only mentions that Hiram and Solomon traded riddles, without giving any of them.  Bryant, p. 19, says the Queen's riddles are given are given in the 2nd Targum to the Book of Esther and elsewhere in the rabbinical literature.  The Targums are commentaries on biblical books, created after the Babylonian Captivity of ‑587/-538 and written down from 100 onwards.  One of these is a strange family riddle which occurs in the first few entries below.  If this is really due to the Queen of Sheba, or even actually in the Targums, it would be by far the earliest strange families riddle known.  A variant of the riddle is given by Yachya Ben Sulieman, c1430, qv below.  Ms Zimmels at the library of the London School of Jewish Studies told me that there is an 1893 German translation: Targum Shennai(?) zum Buch Esther and that the riddles occur in Ginzberg.  However, Rappoport gives more precise information. 

Louis Ginzberg.  The Legends of the Jews.  Translated from the German Manuscript.  Vol. IV: Bible Times and Characters  From Joshua to Esther.  Jewish Publication Society of America, Philadelphia, (1913), 5th ptg, 1947, pp. 142-149.  He gives 22 riddles.  P. 146, no. 2: 'Then she [the Queen of Sheba] questioned him [Solomon] further: "A woman said to her son, thy father is my father, and thy grandfather my husband; thou art my son, and I am thy sister."  "Assuredly," said he, "it was the daughter of Lot who spake thus to her son."'  However, Ginzberg gives no source or date for this.

Angelo S. Rappoport.  Myth and Legend of Ancient Israel.  Vol. III.  Gresham Publishing Co., London, 1928.  The riddles of the Queen of Sheba, pp. 125-130.  P. 127: 

                    'Said she: "I will ask thee another question.  A woman once said unto her son: Thy father is my father, thy grandfather my husband; thou art my son but I am thy sister."

                    To which Solomon made answer: "It must surely have been one of Lot's daughters who thus spoke to her son."

                    The similarity of the text with Ginzberg's makes it clear that they are both taken from the same source.  Fortunately Rappoport is specific as to his sources.  He says the second Targum to Esther (citing Targum Sheni to the Book of Esther; ed. P. Cassel, Leipzig, 1885; ed. E. David, Berlin, 1898) contains three riddles (the last three in Ginzberg) and then says that the Midrash Mishle, or Midrash to the Proverbs (citing Midrash Mishle, ed. S. Buber, Vilna, 1893  and  A. Wünsche, Midrash Mishle, Leipzig, 1885), gives four riddles, which are the first four in Ginzberg, hence include our riddle.  For our riddle, he also gives another reference: J. Lightfoot, Horæ Hebraica, Rotterdam, 1686, II, 527; see also Yalkut, II, §1085.  After these four riddles, he says the Midrash Hachefez (ed. and translated by S. Schechter, Folklore, No. 1, pp. 349-358) gives 19 riddles, which are the first 19 of Ginzberg, so again include our riddle.  However, Rappoport gives no indication of the dates of these Midrashs.

The Exeter Book Riddles.  8-10C (the book was owned by Leofric, first Bishop of Exeter, who mentioned it in his will of 1072).  Translated and edited by Kevin Crossley-Holland.  (As: The Exeter Riddle Book, Folio Society, 1978, Penguin, 1979.)  Revised ed., Penguin, 1993. 

No. 43, pp. 47 & 103.  Body and soul both have the earth as their mother and sister.  Their mother because they are made from dust; their sister because all are made by the same heavenly father.

No. 46, pp. 50 & 104.

A man sat sozzled with his two wives,

his two sons and his two daughters,

darling sisters, and with their two sons,

favoured firstborn; the father of that fine

pair was in there too; and so were

an uncle and a nephew.  Five people

in all sat under that same roof.

            The solution is given in Genesis 19:30-38, which describes Lot and his two daughters who bore sons by him.  "The first use of this incestuous story for the purpose of a riddle is attributed to the Queen of Sheba; she tried it on Solomon."  Cf above entries and Yachya Ben Sulieman, c1430, below.

Alcuin.  9C. 

Prob. 11: Proposito de duobus hominibus singulas sorores accipientibus.  Two men marry each other's sister. 

Prob. 11a (in the Bede text): Propositio de duobus hominibus singulas matres accipientibus.  Two men each marrying the other's mother.  This is the classic  "I'm my own grandfather"  situation.

Prob. 11b (in the Bede text): Propositio de patre et filio et vidua ejusque filia.  Father and son marrying daughter and mother.  This is like 11a.

Abbot Albert.  c1240.  P. 335. 

Prob. 11.  Two widows and sons marry.  This is the same as Alcuin/Bede 11a.  He says the sons of the unions are each other's paternal uncles.

Prob. 12.  Two widowers and daughters marry.  This is the same as the previous except for a sex change.  Latin distinguishes paternal uncle from maternal uncle.

Prob. 13.  Complex situation with a man and three wives.

Cooper (above under General Studies of Kinship Relations) says the medieval period introduced the double incest situation, where the son of a brother-sister relation is sent away and returns as an adult and unknowingly marries his mother.  Cf Pacioli, c1500, and the Martham tombstone of 1730 for a triple incest.

Dialogue of Salomon and Saturnus.  14C.  ??NYS.  Given in Bryant, p. 12.  "Tell me, who was he that was never born, was then buried in his mother's womb, and after death was baptised?"  Answer: Adam.  Cf:  Adevineaux Amoureux, 1478;  Vyse, 1771?, prob. 2.

In about 1380, the Duke of Gloucester, uncle of Richard II, opposed the marriage of his brother's son to his wife's younger sister.  [John Kinross; Discovering Castles  1. Eastern England; Shire Discovering Series No. 23, 1969, p. 12.]

Yachya Ben Sulieman.  Hebrew text, c1430.  ??NYS.  Quoted in Folk‑Lore (1890) ??NYS.  Quoted in Tony Augarde, op. cit. in 5.B, p. 3.  A riddle attributed to the Queen of Sheba.  "A woman said to her son, thy father is my father, and thy grandfather my husband;  thou art my son, and I am thy sister."   "Assuredly,"  said he [Solomon],  "it was the daughter of Lot who spake thus to her son."  Bryant, no. 1116, pp. 259 & 346 gives the same wording, with an extra level of quotation marks, and attributes it to the Queen of Sheba with no further details.  Cf Queen of Sheba and Exeter Book above.

Adevineaux Amoureux.  Bruges, 1478.  ??NYS -- quoted by Bryant, no. 6, pp. 67-68 & 333.  "Je fus nez avant mon pere / Et engendré avant ma mere, / Et ay occis le quart du monde, / Ainsi qu'il gist a la reonde, / Et si despucelay ma taye. / Or pensez se c'est chose vraie."  (Bryant's translation: "I was born before my father, begotten before my mother and have slain a quarter of the world's population.  How can this be?"  Answer: Cain.  Cf:  Dialogue of Salomon and Saturnus, 14C;  Vyse, 1771?, prob. 2.

Chuquet.  1484.  Prob. 166.  Same as Alcuin/Bede 11a.  FHM 233 mentions it briefly without giving the relationships.

In 1491, the 14 year-old Duchess Anne of Brittany married Charles VIII, King of France in 1491.  This was slightly complicated because both of them were married already, indeed Charles was married to the daughter of Anne's husband, the future Emperor Maximilian of Austria, so he was marrying his own step-mother-in-law.  Fortunately, as was often the case in those days, both marriages were unconsummated -- indeed the couples had probably not yet seen each other and such proxy marriages were more like engagements -- so a little influence at Rome got both marriages dissolved.  Somewhat surprisingly, as the marriage was more or less forced by Charles' siege of Rennes, the couple got on very well and developed a definite affection. 

Pacioli.  De Viribus.  c1500.  Part 3. 

F. 263v, fourth item.  = Peirani 377.  Mother says her son is also her nephew and her brother.  A man impregnated his mother and this yielded a daughter, who is the speaker.  The man then impregnated her and this produced a son.  He is the son of her brother, hence her nephew, and the son of her father, so her brother.

F. 266r, middle.  = Peirani 383.  Two widows marrying the other's son.  = Alcuin 11a.  A woman is carrying the son of the other couple, so it is her grandson and the brother of her husband, as well as the son of her mother-in-law.

This is followed by a variation which I cannot understand.

F. 267v, first item.  = Peirani 385.  Tomb holds mother and son, man and wife, sister and brother, but only two people.  Son is offspring of father-daughter incest and later marries his mother.

F. 287v, no. 191.  = Peirani 416.  Two fathers and two sons are only three people.

Ff. 291r - 291v, no. 218.  = Peirani 421-422.  Quotes a Roman epitaph from S. Bartolomeo which seems to say that Hersillus is buried with Maralla who was his mother, sister and wife.  Pacioli's comment seems irrelevant.

Ivan Morris.  Foul Play and Other Puzzles of all Kinds.  (Bodley Head, London, 1972);  Vintage (Random House), NY, 1974.  Prob. 21: No incest, pp. 39 & 93.  Quotes Dudeney (??NYS) who gave an authentic 1538 epitaph describing the situation of Alcuin/Bede 11a with each couple having a child.

Tartaglia.  General Trattato, 1556, art. 135, p. 256r.  2  fathers and  2  sons make only  3  people.

16C(?) riddle in Mantuan dialect.  Given in:  Franco Agostini & Nicola Alberto De Carlo; Intelligence Games;  (As:  Giochi della Intelligenza; Mondadori, Milan, 1985);  Simon & Schuster, NY, 1987; p. 69.  Two fathers and two sons make three people.  The discussion is a bit unclear as to the date of this riddle.

Book of Merry Riddles.  1629? 

The 71 Riddle, pp. 42-43.  Two ladies and two boys are met and they say as follows.

                                               The sons of our sons they be certain,

                                              Brothers to our husband they be I wis,

                                            And each of them unto the other Uncle is

                                              Begotten and born in wedlock they be,

                                           An we are their mothers we tell you truly.

I.e., the same as Alcuin/Bede 11a.

The 73 Riddle, p. 44.  2  fathers and  2  sons make only  3  people.

Tombstone in the church at Martham, Norfolk.  1730.  ??NYS -- quoted in a letter from Judith Havens (Norwich, Norfolk) in Challenging Centipede; The Guardian, section 2, (1 Dec 1994) 6.

                              "Here Lyeth / The Body of Christ. / Burraway, who departed / this life ye 18 day / of October, Anno Domini 1730 / Aged 59 years / And their Lyes / Alice, who by hir Life / was my Sister, my Mistress, / My Mother, and my Wife. / Dyed Feb ye 12, 1729 / Aged 76 years."

                              This was a response to a vague description of the epitaph in:  Centipede; Famous last words; The Guardian, section 2 (24 Nov 1994) 4.  There it is stated that Burraway was the result of an incestuous union between a man and his daughter.  The baby was sent away to be brought up and years later happened to return to his native village where he met a older woman and became her lover, then her husband!  (Shades of Oedipus!)  Consequently he was his own uncle, step-father and brother-in-law!  Cf Cooper, above for the idea of double incest -- this seems to be a triple incest.

                    If Burraway's union with his sister/mother/wife produced a child, then it would have only three grandparents and six great-grandparents (as is the case for the offspring of any half-siblings).

Vyse.  Tutor's Guide.  1771? 

Prob. 1, 1793: p. 305; 1799: p. 318 & Key p. 360.  "Suppose two Women, and each a Son, were walking together, and were met by another Person, who asked the Boys in what Relation they stood to each other?  They replied, We are Sons and Grandsons by the Fathers; Brothers and first Cousins by the Mothers; who also are Aunts to each of us.  This Combination of Kinship once happened, but in what Manner?  See Gen. xix. ver. 31."  This is the same situation as in the Exeter Book Riddles, no. 46, above.

Prob. 2, 1793: p. 305; 1799: p. 318 & Key p. 360.  "Who was he that was begot before his Father, born before his Mother, and had the Maidenhead of his Grandmother?"  The answer notes that Adam was made "of the Dust of the Ground", etc. and then runs:  "Now Abel ... was murdered by his Brother Cain; therefore he got the Maidenhead of his Grandmother (the Earth); and was got before his Father (Adam, who was made of the Earth, therefore was not begotten; and was born before his Mother (Eve), who was made of Man, therefore was not born."  I think the meaning is that Abel was buried so he was the first man in the Earth.  Cf:  Dialogue of Salomon and Saturnus, 14C;  Adevineaux Amoureux, 1478.

Jackson.  Rational Amusement.  1821.  Arithmetical Puzzles.  No. 6, pp. 2 & 52.  Father, mother, son, grandson, brother and daughter comprise only  3  people in the situation of Alcuin/Bede 11a when a couple has a son.

Curiosities for the Ingenious selected from The most authentic Treasures of Nature, Science and Art, Biography, History, and General Literature.  2nd ed., Thomas Boys, London, 1822.  Singular intermarriage, p. 100.  Man and daughter marry daughter and father.  "My father is my son, and I am my mother's mother; / My sister is my daughter, and I'm grandmother to my brother."

Richard Breen.  Funny Endings.  Penny Publishing, UK, 1999, p. 25.  Gives the following.

                    "Here, beneath this stone,  /  Lie buried alone  /  The father and his daughter,  / 

                    The brother and his sister,  /  The man and his wife,  /  And only two bodies.  //

                              Work it out...

                    Early 19th century, Erfunt Cemetery, Germany."

          I suspect 'Erfunt' is a misprint for 'Erfurt' and I'm a bit suspicious as to the authenticity of this, but it's a good puzzle problem.

Judge Leicester King (1789-1856) of Akron, Ohio, and his son married sisters, so he was his son's brother-in-law.  BION-11.

Illustrated Boy's Own Treasury.  1860.  Arithmetical and Geometrical Problems, No. 32, pp. 430 & 434.  Will involving  6  relatives who turn out to be just  3  due to the situation of Alcuin/Bede 11a.

Charades, Enigmas, and Riddles.  1862: prob. 40, pp. 138 & 144-145;  1865: prob. 584, pp. 110 & 157-158.  "How can a man be his own grandfather?"  Mother and daughter marry son and father.  Mother and son produce a son, Tom.  Notes that perhaps Tom is only his own grandfather-in-law.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 564-12, pp. 253 & 395.  25  relationships among only  7  people.  Answer is a couple, with their son and his wife, with their son and two daughters.  This omits a grandchild, so there really ought to be  26  relationships here.  Dudeney, AM 54, gives the same grouping, but with a different list of 23 relationships.  However neither counts the relationships properly -- e.g. both count  4 children  and  2 sons and 2 daughters.  I find 23 reasonable relationships -- Dudeney's 23 less  4 repeated children  plus  2 husbands and 2 wives  that he omitted.  Leske has the husbands and wives, but omits the grandmother and a grandchild.

                    In theory,  n  people can have  n(n-1)  possible relationships, but not all of these relationships have distinct names.  E.g. in the above, the son of the first couple is a son to both parents, so two distinct relationships are both denoted by 'son'.  However, in the classic problem of man, son and grandson, we actually have  2 fathers, 2 sons, 1 grandfather and 1 grandson,  giving a full  6  relationships among just  3  people.  Extending this to a string of  n  generations gives the full  n(n-1)  relationships among  n  people.  One might ask if one can compact this a bit by using fewer generations for the  n  people.  E.g., Leske's problem has  3  generations.  So I pose the following problem:  for  n  people in  g  generations, how many of the  n(n-1)  relationships are distinctly named in English (where 'distinctly named' is a bit vague!).  I now realise that the same relationship may have different names, indeed several different names!  See Alcuin, above, and Carroll, below.

Mittenzwey.  1880. 

Prob. 25, pp. 3 & 59;  1895?: 30, pp. 9 & 63;  1917: 30, pp. 9 & 58.  Two fathers and two sons are three people.

Prob. 26. pp. 3-4 & 59;  1895?: 31, pp. 10 & 63;  1917: 31, pp. 9 & 58.  26  relationships among only  7  people, as in Leske.

E. S. Turner.  Op. cit. above, p. 109.  [Retold in his:  Amazing Grace; Michael Joseph, London, 1975; pp. 279-280.]  The 7th Duke of Marlborough described to the House of Lords, c1880, the supposedly real case of a father and son marrying a daughter and mother.  The son was his own grandfather and he became so confused that he committed suicide.  In a footnote, Turner quotes a letter in the Welwyn News-Chronicle of 1949 from a man who married his step-mother's sister, i.e. widower and son married sisters.

J. M.  Letter:  Genealogical puzzle.  Knowledge 3 (6 Jul 1883) 13, item 865.  +  Answer to genealogical puzzle in our last.  Ibid. (13 Jul 1883) 29.  Two unrelated persons have the same brother.  Editorial note to the Answer says there are several ways to solve the puzzle -- how many?

Lewis Carroll.  A Tangled Tale.  (1885) = Dover, 1965.  The dinner party.  Knot II: Eligible apartments.  Pp. 7-8 & 84-85.  (In the answers, this part of the Knot is denoted §1. The dinner party.)  = Carroll-Wakeling II, prob. 41: Who's coming to dinner?, pp. 59 & 75.  A man's:  father's brother-in-law;  brother's father-in-law;  father-in-law's brother  and  brother-in-law's father  are all the same person.  This involves several marriages between cousins. 

                    But brother-in-law denotes both  sister's husband  and  wife's brother!  With an earlier marriage between cousins, we can have this person being both the man's father's wife's brother and father's sister's husband.  His wife's brother's father is just his father-in-law, but another marriage between cousins makes this person also the man's brother-in-law's father-in-law.

E. W. Cole.  Cole's Fun Doctor.  The Funniest Book in the World.  Routledge, London  &  E. W. Cole, Melbourne,  nd [HPL dates it 1886 and gives the author's name as Arthur C. Cole].  P. 57:  A smart cut-out  &  Genealogy  are two stories of widow and daughter marrying son and father.  = Alcuin 11b.

Lemon.  1890.  How is this?, no. 725, pp. 83 & 123.  33  relatives who are only  8  people.

Hoffmann.  1893.  Chap. IX, no. 42: The family party, pp. 321 & 328‑329 = Hoffmann‑Hordern, p. 214.  A man is  his father's brother‑in‑law,  his brother's father‑in‑law,  his father‑in‑law's brother‑in‑law  and  his brother‑in‑law's father‑in‑law!

C. C. Bombaugh.  Facts and Fancies for the Curious.  Lippincott, 1905, ??NYS.  A Mr. Harwood and John Cosick, both widowers, married each other's daughter, at Durham in eastern Canada.  Quoted in:  George Milburn; A Book of Puzzles and Brainteasers; Little Blue Book No. 1103, Haldeman‑Julius, Girard, Kansas, nd [1920s?], pp. 33‑34.

Pearson.  1907.

Part I, no. 25: A family party, pp. 120 & 184.  Two widows and sons marry and each couple has a daughter, causing  24  apparent people to be only  6.

Part I, no. 36: Quite a family party, pp. 123 & 186.  Same as Hoffmann.

Part I: A table of affinity, p. 127.  Reports that M. de Lesseps and his son were to marry sisters and discusses the complications that would ensue.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 26:6 (Oct 1911) 569.  The mean Duke.  Same as Hoffmann.

Henry Edwards Huntington (1850-1927) married Arabella Duval Huntington (  -1924), the widow of his uncle Collis P. Huntington (  -1900) in 1913.  Henry and Arabella were the same age.

Dudeney.  AM.  1917.  Several examples, including the following.

Prob. 52: Queer relationships, pp. 8 & 153.  Discusses two brothers who married two sisters.  One man and one woman died and the survivors married and had a child.  The man married his deceased wife's sister which was legal, so he is married to the woman and his child is legitimate.  But the woman married her deceased husband's brother, which was not legal at the time, so she is not married to the man and her child is illegitimate!!  Mentions a man who married his widow's sister and a man who has a nephew, but the nephew is not the nephew of his sister.

Prob. 54: A family party, pp. 8 & 153.  23  apparent people are only  7.  See Leske.

Prob. 55: A mixed pedigree, pp. 8-9 & 153.  A  is  B's  father's brother-in-law,  brother's father-in-law  and  father-in-law's brother.

Prob. 56: Wilson's poser, pp. 9 & 153.  A  is  B's  uncle and nephew.  This is due to two men marrying the mother of the other and both couples producing -- the results are  A  and  B.

Ahrens.  A&N.  1918.  Pp. 105‑122, esp. 111‑122.  Describes an 11C report of a person with only three grandparents, but it turns out to be erroneous.  However, Cleopatra had only three grandparents, since her parents were half‑siblings -- I had believed they were full siblings which would give her only two grandparents.  See also the 1730 entry about Christopher Burraway.

                    For convenience in the following, let  n‑parents denote one's ancestors  n  generations back, so  1‑parents are parents,  2‑parents are grandparents,  3‑parents are great‑grandparents, etc.  Normally one has  2n  n‑parents. 

                    Prince Don Carlos of Spain (1545‑1568) had only  4  3‑parents,  6  4‑parents,  12  5‑parents  and  20  6‑parents.  Ahrens also gives more extended examples, e.g. the  12  generations of ancestors of Kaiser Wilhelm II comprise only  1549  people instead of the expected  8190,  and one person occurs in  70  places.

Smith.  Number Stories.  1919.  Pp. 114‑116 & 142.  Tartaglia's problem.  Also a more complex version where  27  apparent people are only  7.

Ackermann.  1925.  Pp. 92‑93.  Three mothers, each with two daughters, require only  7  beds.

Loyd Jr.  SLAHP.  1928.  A puzzling estate, pp. 68 & 111.  Three fathers and three sons are only four people.

Collins.  Fun with Figures.  1928.  He was his own grandfather, pp. 231-232.  Claims to quote a Pittsburgh newspaper story of a resident who committed suicide when he found he was his own grandfather.  Son and father married widow and daughter.

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937. 

Prob. 4, pp. 188 & 197.  'My grandfather is only five years older than my father.'

Prob. 5, pp. 188 & 197.  'My father and my grandfather are twins.'

Haldeman-Julius.  1937.  No. 88: Marriage problem, pp. 11 & 25.  How can a man have married his widow's sister?  (Also entered under Deceased Wife's Sister.)

Depew.  Cokesbury Game Book.  1939. 

Three ducks, p. 201.  Two fathers and two sons make three people.

Separate Room, p. 216.  Three mothers, each with two daughters, make seven people.

Joseph Leeming.  Riddles, Riddles, Riddles.  Franklin Watts, 1953;  Fawcett Gold Medal, 1967. 

P. 50, no. 5: "If your uncle's sister is not your aunt, just what relation is she to you?

P. 53, no. 32: "If Dick's father is Tom's son, what relation is Dick to Tom?"

P. 91, no. 3: "A doctor had a brother who went out West.  But the man who went out West had no brother.  How can this be?"

P. 91, no. 4: "Two men, with their two wives and two sons, are related to each other as follows: The men are each other's fathers and sons, their wives' fathers and husbands, and their children's fathers and grandfathers.  The women are the children's mothers and sisters; and the boys are uncles to each other.  How can this be?"  Same as Abbot Albert, prob. 12.

P. 92, no. 8.  Schoolteacher and his daughter, the minister's wife and the minister are just three people.

P. 109, no. 11: "Sisters and brothers have I none, but that man's father is my father's son.  Who am I looking at?"

P. 113, no. 44: "What relation is that child to its father who is not its father's own son?"

P. 113, no. 47: "Two Indians are standing on a hill, and one is the father of the other's son.  What relation are the two Indians to each other?"

P. 153, no. 30: "It wasn't my sister, not my brother,  But it still was the child of my father and mother.  Who was it?"

Ripley's Believe It Or Not!  6th series, Pocket Books, NY, 1958.  P. 145.  Jacob van Nissen, of Zwolle, Holland, and his son married a girl and her mother.

W. Leslie Prout.  Think Again.  Frederick Warne & Co., London, 1958.  Catch Quiz, No. 6, pp. 12 & 115.  "Said one boy to another:  "My mother's sister is your sister's mother."  What relation were the two boys?"

Kathleen Rafferty.  Dell Pencil Puzzles & Word Games.  Dell, NY, 1975.  Noodle Nudger 3, pp. 106 & 127.  "Mary's husband's father-in-law is Mary's husband's brother's brother-in-law, and Mary's sister-in-law is Mary's brother's stepmother.  HOW COME?"  "Mary's father married her husband's sister."

Scot Morris.  The Book of Strange Facts and Useless Information.  Doubleday, NY, 1979, p. 98.  In order to give his divorced mother some benefit from his father's estate, Robert Berston adopted his mother in 1967.

Patrick Donovan.  Peculiar People.  Fontana, 1984.  P. 76 reports that Dave Woodhouse of Wolverhampton divorced his wife and married her mother in 1983 at a double wedding where his ex-wife was also married to a new man.

Ripley's Believe It or Not! - Strange Coincidences.  Tor (Tom Doherty Associates), NY, 1990, p. 21.  George Clark Cheever, of Warsaw, Indiana, three of his sons and one of his daughters all married siblings.

Marcia Ascher.  Ethnomathematics.  Op. cit. in 4.B.10.  1991.  Chapter Three: The logic of kinship relations, pp. 66-83.  Gives a number of folk puzzles and references. 

                    Two mothers and two daughters are three people (Brazil). 

                    Two brothers say  "My brother's son is buried there";  third brother says  "My brother's son is not buried there"  (Ireland).

                    Who is the sister of my aunt, who is not my aunt? (Puerto Rico).

                    His mother is my mother's mother-in-law (Russia).

                    Who is my mother's brother's brother-in-law? (Wales).

"Smallweed" diary column.  The Guardian (10 Apr 1993) 18.  Bill Wyman of the Rolling Stones married and later divorced Mandy Smith.  Mandy's mother Patsy is about to marry Stephen Wyman, Bill's son from his first marriage.

Item on front of Society section of The Guardian (2 Apr 1997) 1.  "A 46-year-old Bedfordshire woman has married the father of an 18-year-old girl who eloped with her husband."

David Singmaster.  Some years before I read a description of some English aristocrat which said that he only had four grandparents while most people had eight.  This was later corrected -- he had only four great-grandparents while most people had eight.  Of course this puzzled me for a bit and I worked out a reasonable way this could have happened.  I later realised that there is another way it could have happened, though I think the second method is slightly less likely.  I posed the problem of finding both ways as a Brain Jammer: Four Great-Grandparents; Weekend Telegraph (12 Dec 1998) 21 & (2 Jan 1999) 17.  At the time I didn't know of any real examples, but Ahrens (1918) says Prince Don Carlos of Spain (1545‑1568) had only four great-grandparents.  I also used this on a Puzzle Panel program.  When cousins marry, their children have six great-grandparents.

Mindgames.  Frontiers (a UK popular science magazine) No. 1 (Spring 1998) 107.  "My wife's sister is also her cousin.  How can this be?"  Solution is that her father married his dead wife's sister and had another daughter.

 

          9.E.1. THAT MAN'S FATHER IS MY FATHER'S SON, ETC.

 

            New section -- I have just found the 17C example, but there must be other and older examples??  But see Proctor, 1883.  The problem continues to perplex people -- see Fairon, 1992.

            For the 'Blind beggar' type of problem see: Boy's Own Book, 1828; Rowley, 1866; Rowley, 1875; Neil, 1880s; Lemon, 1890; Clark, 1897; Home Book, 1941; Charlot, c1950s.  Again one thinks this should be older.

            Some of the items given above, could go in here.

 

Tomé Pinheiro da Veiga.  Fastiginia o fastos geniales.  (This work is a chronicle of courtly life in Castilla, from 1601 to 1606.  Translated & edited by Narciso Alonso Cortés.  Imprenta del Colegio de Santiago, Valladolid, 1916, pp. 155b-156a, ??NYS.)  French translation in: Augustin Redondo; Le jeu de l'énigme dans l'Espagne du XVIIe siècle.  Aspect ludique et subversion; IN: Les Jeux à la Renaissance; Actes du XXIIIe Colloque International d'Études Humanistes (Tours, 1980); J. Vrin, Paris, 1982; pp. 445-458, with the problem being on p. 453.  There are two brothers, born of the same father and mother.  One is my uncle, but the other isn't.  How is this possible?

The Book of Merry Riddles.  London, 1629.  ??NYS -- Santi 235 gives this and says it is reprinted in J. O. Halliwell, The literature of the sixteenth and seventeenth centuries, London, 1851, pp. 67‑102, ??NYS, and as pp. 7-29 in Alois Brandl, Shakespeares Book of Merry Riddles und die anderen Räthselbücher seiner Zeit, Jahrbuch der deutschen Shakespeare-Gesellschaft 42 (1906) 1-64, ??NYS.   Bryant, pp. 100-102, quotes from: A Booke of Merrie Riddles, Robert Bird, London, 1631 and says it is also known as Prettie Riddles.  Santi 237 gives Booke of Merrie Riddles, London, 1631, and says it is reprinted as pp. 53-63 in Brandl, ??NYS.  Santi 307 gives The Booke of Merry Riddles, London, 1660, reprinted by J. O. Halliwell in 1866 in an edition of 25 copies, of which 15 were destroyed!, ??NYS.  In Bryant, no. 275, pp. 102 & 336: "I know a child borne by my mother, / naturall borne as other children be, / that is neither my sister nor my brother. / Answer me shortly: what is he?"

Boy's Own Book.  Conundrums.

"What kin is that child to its own father who is not its father's own son?"  1828: No. 70, pp. 432 & 440.  1828‑2: No. 70, pp. 436 & 444;  1829 (US): No. 70, pp. 238 & 258;  1855: No. 102, pp. 583 & 595.  (I think this is the forerunner of the Blind Beggar problem.)

"What relation is your uncle's brother to you who is not your uncle?"  1828: No. 77, pp. 433 & 440.  1828‑2: No. 77, pp. 437 & 444;  1829 (US): No. 77, pp. 238 & 258;  1855: No. 109, pp. 584 & 595. 

The Riddler.  1835.  Conundrums Nos. 66 and 73, p. 16, no answers in my copy.  These are identical to Nos. 70 and 77 in Boy's Own Book.

Boy's Own Book.  1843 (Paris): 437 & 441, no. 11.  Woman says: 'Your mother was my mother's only daughter.'  = Boy's Treasury, 1844, pp. 425 & 429.  = de Savigny, 1846, pp. 354 & 358, no. 8: 'sa mère à lui est la seule fille de sa mère à elle', which doesn't seem quite right to me.

Child.  Girl's Own Book.  1848: Enigma 46, p. 236;  1876: Enigma 37, p. 199.  "His mother was my mother's only child." 

                    = Fireside Amusements, 1850: No. 36, pp. 111 & 181; 1890: No. 21, p. 99.

Fireside Amusements. 

1850: No. 17, pp. 134 & 184;  1890: No. 16, p. 110.  "If Dick's father be John's son, what relation is Dick to John?"

1850: No. 29, pp. 135 & 185;  1890: No. 25, p. 110.  "What kin is that child to its own father, who is not its own father's own son?"  c= Boy's Own Book.

1850: No. 89, pp. 138 & 186;  1890: No. 25, p. 110.  "What relation is your uncle's brother to you who is not your uncle?"  = Boy's Own Book.

Charades, Enigmas, and Riddles.  1860: prob. 176 & 177, pp. 21 & 44; 

                    1862: prob. 176 & 177, pp. 73 & 111.

Prob. 176.  "If your uncle's sister is not your aunt, what relationship does she bear to you?"

Prob. 177. 

My mother had a child, my very own mother,

It was not my sister nor yet was it my brother;

If you are as clever as I fancy you to be,

Pray tell me what relation that child was to me.

Boy's Own Conjuring Book.  1860.  P. 381.  "I've no sister or brother; / You may think I am wild; / But that man's mother / Was my mother's child."

Hugh Rowley.  Puniana or Thoughts Wise and Other-wise  A New Collection of the Best Riddles, Conundrums, Jokes, Sells, etc, etc.  Chatto & Windus, London, 1866.

P. 36.  Brothers  A  and  B  were walking.  ""I must speak to those children," said A; "they are my nephews and nieces."  "Ah!" said B, "as I have no nephews and nieces, I shall walk on."  How was this?"

P. 88.  "What kin is that child to his own father who is not his own father's son?"

P. 88.  "If Dick's father is Tom's son, what relation is Dick to Tom?"

P. 88.  "Who were your grandfather's first cousin's sister's son's brother's forefathers?  Why, his aunt's sisters, of course."  (This is non-logical, being a pun on ancestors, but it illustrates that the idea of such problems must have been well known.)

P. 137.  A fiddler said his brother played the double-bass, but the double-bass player denied having a brother.

Hugh Rowley.  More Puniana; or, Thoughts Wise and Other-Why's.  Chatto & Windus, London, 1875. 

P. 28.  "A blind beggar had a brother, who died and went to heaven.  What relation was the blind beggar to the person who went to heaven?"  Answer is 'sister', but the posing of the question is defective -- it should include the assertion that the person who went to heaven had no brother.

P. 134.  "What relation is that child to its own father, who is not its own father's son?"

P. 217.  "That gentleman's mother is my mother's only child."

P. 231.  "What relation is your father's only brother's sister-in-law to you?"

P. 231.  "Brothers and sisters have I none, but this man's father is my father's son."

[Richard A. Proctor]  Letters received and short answers.  Knowledge 3 (26 Oct 1883) 264.  Answer to Harry.  "Sisters and brothers have I none   But that man's father is my father's son."  Implies that the puzzle is not well known.  Recalls it being posed on a ship and distracting all the passengers for a day.

E. W. Cole.  Cole's Fun Doctor.  The Funniest Book in the World.  Routledge, London  &  E. W. Cole, Melbourne,  nd [HPL dates it 1886 and gives the author's name as Arthur C. Cole].  P. 329: A riddle.  "His mother was my mother's only child."

James Neil [= "A Literary Clergyman"].  Riddles:  And Something New About Them.  (Lang Neil & Co., London, nd [1880s?];  Simpkin Marshall & Co., London);  Village Games, London, 1993.  General Riddles: Relationship, p. 28.

"A blind beggar had a brother, the brother died, deceased had no brother.  What relation was the blind beggar to deceased?"

"What relation is that child to its own father who is not its own father's own son?"

"If your uncle's sister is not your aunt, what relation is she to you?"

Lemon.  1890. 

Do you see it?, unnumbered section after no. 80, pp. 15:  "That gentleman's mother is my mother's only child."

Conundrums, no. 142(a), pp. 23 & 102 (= Sphinx, no. 470(a), pp. 65 & 113): child "who is not his own father's son."

Fireside Amusements -- A Book of Indoor Games.  Op. cit. in 7.L.1.  1890?  P. 99, no. 21.  "His mother was my mother's only child."

William Crompton.  The odd half-hour.  The Boy's Own Paper 13 (No. 657) (15 Aug 1891) 731-732.  A true friend.  "If your uncle's sister is not your aunt, what relationship does she bear to you?"

Bennett Coll.  Prove it!  The Idler 2 (1892-1893, probably Dec 1892) 510-517.  Man in front of a portrait says  "Sisters and brothers have I none;  That man's father is my father's son."  Says the portrait is himself!  Observes that this leads to his father being his own son and being the father of his father.  Describes the difficulties people have in trying to see this answer [not surprisingly].  Various other solutions given:  grandfather,  brother,  uncle on the mother's side.

Hoffmann.  1893.  Chap. IX, no. 25: The portrait, pp. 318 & 326 = Hoffmann-Hordern, p. 211.  "Uncles and brothers have I none, But that man's father is my father's son."  He notes "This venerable puzzle forms the subject of a humorous article, entitled "Prove It," in a recent number of the Idler.  Its most amusing feature is that the writer has himself gone astray, ...."  [I'm not sure whether Coll has gone astray or is using the error to generate humour??]

W. H. Howe.  Everybody's Book of Epitaphs Being for the Most Part What the Living Think of the Dead.  Saxon & Co., London, nd [c1895] (reprinted by Pryor Publications, Whitstable, 1995).  P. 165 has the following entry.

          "In Llanidan Churchyard, Anglesea:--

                    Here lies the world's mother,

                    By nature my Aunt -- sister to my mother,

                    My grandmother -- mother to my mother,

                    My great grandmother -- mother to my grandmother,

                    My grandmother's daughter and her mother."

          Could this be a real case of 'I'm my own grandmother'??

Clark.  Mental Nuts. 

1897, no. 3.  The beggar.  "A beggar had a brother, the brother died and the man who died had no brother."

1897, no. 16.  The man in jail.  "Brothers and sisters have I none, but that man's father is my father's son." 

Somerville Gibney.  So simple!  The Boy's Own Paper 20 (No. 992) (15 Jan 1898) 252.  'That very old catch -- "If Dick's father is Tom's son,  What relation is Dick to Tom?"'

Dudeney.  "The Captain" puzzle corner.  The Captain 3:1 (Apr 1900) 1 & 90  &  3:3 (Jun 1900) 193 & 279.  No. 3: Overheard in an omnibus.  "Was that your father."  "No, that gentleman's mother was my mother's mother-in-law."   Essentially the same as:  AM; 1917; Prob. 53: Heard on the tube railway, pp. 8 & 153;  "That gentleman's mother was my mother's mother-in-law, but he is not on speaking terms with my papa."

Hummerston.  Fun, Mirth & Mystery.  1924.  Grandfather's problems: The portrait, p. 68.  "Sisters and brothers have I none,  But that man's father is my father's son."

James Joyce.  Ulysses.  (Dijon, 1922);  Modern Library (Random House), NY, 1934, apparently printed 1946.  P. 692 (Gardner says the 1961 ed. has p. 708; this is about 4/5 of the way between the start of Part III and Molly's soliloquy).  "Brothers and sisters had he none, Yet that man's father was his grandfather's son."  This is given as a quotation, while Bloom is looking in a mirror -- otherwise it could be a cousin.  [Given in Bryant, no. 782, pp. 194 & 342.]

Streeter & Hoehn.  Op. cit. in 7.AE.  Vol. 2, 1933, p. 16, no. 10: "Brain twister".  "My son's father is your father's only child.  What relative of yours am I?"

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 33, pp. 194 & 204.  'This man's mother is my mother's only child.'

Haldeman-Julius.  1937.  No. 44: Portrait problem, pp. 7 & 22.  Woman points to a man's portrait and says to her brother: "The man's mother was my mother's mother-in-law."  Answer is that she is his daughter, but she might be his step-daughter.

McKay.  Party Night.  1940.  No. 7, p. 175.  "Brothers and sisters have I none;  yet this man's father was my father's son."

Meyer.  Big Fun Book.  1940.  No. 5, pp. 175 & 756.  "My father is the brother of your sister.  What relative am I of yours?"  Answer is  nephew,  but  son  is also possible.

The Home Book of Quizzes, Games and Jokes.  Op. cit. in 4.B.1, 1941.

P. 149, prob. 8.  "Sisters and brothers I have none, but that man's father is my father's son."

P. 149, prob. 10.  "A beggar's brother died.  But the man who died had no brother."

John Henry Cutler.  Dr. Quizzler's Mind Teasers.  Greenberg, NY, 1944.  ??NYS -- excerpted in: Dr. Quizzler's mind teasers; Games Magazine 16:3 (No. 109) (Jun 1992) 47 & 43, prob. 14, with additional comments in Ibid. 16:4 (No. 110) (Aug 1992) 4 and 16:6 (No. 112) (Dec 1992) 4.  "What relation is a man to his mother's only brother's only niece?"  Answer is her brother, but comments point out that she could be his cousin, i.e. his mother's sister's daughter, or even a kind of cousin-in-law, i.e. his mother's brother's wife's sibling's daughter.

Yvonne B. Charlot.  Conundrums of All Kinds.  Universal, London, nd [c1950?].

          P. 77: "If your aunt's brother is not your uncle, who is he?"

          P. 82: "What kin are those children to their own father who are not their own father's sons?"

Hubert Phillips.  Party Games.  Witherby, London, 1952.  Chap. XIII, prob. 3: Photograph, pp. 204 & 252‑253. 

                    "Though sons and brothers have I none,

                    Your father was my father's son." 

          Solution says this "is my own invention".

See Ascher in 9.E for some examples.

Iona & Peter Opie.  I Saw Esau:  The Schoolchild's Pocket Book.  (Williams & Norgate, London, 1947.)  Revised edition, Walker Books, London, 1992, ??NX.  No. 42, p. 45, just gives the rhyme;  illustration on p. 44;  answer on p. 144 just gives the answer, with no historical comments.

Ripley's Believe It or Not!, 8th series.  Pocket Books, 1962, p. 26.  Jimmy Burnthet, of Emmerdale, England, courted a girl for 40 years.  She broke off the engagement and married his nephew, whereupon he married her niece.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 16, pp. 25 & 88.  "The father of the person in the portrait is my father's son, but I have no brothers or sons."

Ripley's Puzzles and Games.  1966. 

P. 12.  "What relation is your mother's brother's brother-in-law to you.  Answer: Your father".  There are several more answers.  My mother's brother is my maternal uncle.  He could have several sisters whose husbands are his brothers-in-law -- one is my father, the others are brothers-in-law of my mother.  Also, he could be married and any of his wife's brothers are also his brothers-in-law.

P. 13.  "Moab and Ben-am-mi were brothers yet cousins and their father was their grandfather."  These are the sons of Lot by his daughters.

P. 13.  Bible puzzle.  "Two people died who were never born" -- Adam and Eve.  "Two people were born who never died" -- Enoch and Elias both just disappeared!  "The oldest man who ever lived died before his father did" -- Methusaleh, the son of Enoch.

Pat Fairon.  Irish Riddles.  Appletree Press, Belfast, 1992;  Chronicle Books, San Francisco, 1992;  pp. 27 & 60.  "Brothers and sisters have I none  But this man's father  Was my father's son."  Answer is "Oneself".

 

          9.E.2. IDENTICAL SIBLINGS WHO ARE NOT TWINS

 

            Two siblings are born on the same day to the same parents but are not twins.  New section.  This must be older than the example below.

 

Harold Hart.  The World's Best Puzzles.  Op. cit. in 7.AS.  1943.  The problem of the two students, pp. 4 & 50

 

          9.F.    THE UNEXPECTED HANGING

 

Nicholas Falletta.  The Paradoxicon.  Op. cit. in 8.H.1.  1983.  Pp. 162‑163 relates that during World War II, Swedish Radio announced there would be an unexpected civil defence exercise next week.  Lennart Ekbom, a Swedish professor of mathematics, noted the paradoxical nature of this and discussed it with his students.

D. J. O'Connor.  Pragmatic paradoxes.  Mind 57 (1948) 358‑359.  Discusses several other paradoxes, e.g. "I remember nothing", then the unexpected blackout exercise.

L. Jonathan Cohen.  Mr. O'Connor's "Pragmatic paradoxes".  Mind 59 (1950) 85‑87.  Doesn't deal much with the unexpected blackout.

Peter Alexander.  Pragmatic paradoxes.  Mind 59 (1950) 536‑538.  Also doesn't deal much with the unexpected blackout.

Michael Scriven.  Paradoxical announcements.  Mind 60 (1951) 403‑407.  "A new and powerful paradox has come to light."  Entirely concerned with the unexpected blackout and considers the case of only two possible dates.

Max Black.  Critical Thinking.  1952.  Op. cit. in 6.F.2.  Prob. 1, pp. 156 & 433.

Gamow & Stern.  1958.  The date of the hanging.  Pp. 23‑27.

M. Gardner.  SA (Mar 63) = Unexpected, chap. 1, with an extensive historical addendum and references.

Joseph S. Fulda.  The paradox of the surprise test.  MG 75 (No. 474) (Dec 1991) 419-421.

 

          9.G.   TRUTHTELLERS AND LIARS

 

            I have just started to consider problems where a number of statements are given and we know at least or at most some number of them are lies. 

            Find correct answer in one question from a truthteller or liar:  Goodman;  Gardner;  Harbin;  Rice;  Doubleday - 2;  Eldin.  See:  Nozaki  for a generalization.

            Problem with three truthtellers or liars and first one mumbles:  Rudin;  Haldeman-Julius;  Depew;  Catch-My-Pal;  Kraitchik;  Hart;  Leopold;  Wickelgren.

 

Magician's Own Book.  1857.  P. 216.  A  lies  1/4  of the time;  B  lies  1/5  of the time;  C  lies  1/7  of the time.  "What is the probability of an event which  A  and  B  assert, and  C  denies?"  Answer is  140/143,  but I get  2/3.  = Book of 500 Puzzles, 1859, p. 54. 

Chas. G. Shaw.  Letter:  The doctrine of chances.  Knowledge 7 (27 Feb 1885) 181, item 1620.  Says Whitaker's Almanac for this year, under The Doctrine of Chances, gives the following problem with a wrong answer.  A  lies  1/4  of the time;  B  lies  1/5  of the time;  C  lies  1/6  of the time.  What is the chance of an event which  A  and  B  assert, but  C  denies?  Whitaker and I get  (3/4)(4/5)(1/6) / [(3/4)(4/5)(1/6)  +  (1/4)(1/5)(5/6)]  =  12/17,  but Shaw claims  19/24  by asserting that the probability of an event when  A  and  B  testify to it ought to be   1 ‑ (1/4)(1/5)  =  19/20   instead of   (3/4)(4/5)  =  3/5,   He then says this leads to  19/24  by modifying the above formula, but I can't see how this can be done.

Lewis Carroll.  Diary entry for 27 May 1894.  "I have worked out in the last few days some curious problems on the plan of 'lying' dilemma.  E.g. 'A says B lies;  B says C lies;  C says A and B lie.'  Answer: 'A and C lie;  B speaks truly'.  The problem is quoted in Carroll-Gardner with his discussion of the result, pp. 22-23.  Gardner says this was printed as an anonymous leaflet in 1894.

                    Carroll-Wakeling.  Prob. 9: Who's telling the truth?, pp. 11 & 65.  Wakeling says "This is a puzzle based on a piece of logic that appears in his diary.

                    The Dodo says the Hatter tells lies.

                    The Hatter says that the March Hare tells Lies.

                    The March Hare says that both the Dodo and the Hatter tell lies.

                    Who is telling the truth?"

                    In a letter of 28 May 2003, Wakeling quotes the text from the Diary as above, but continues with it:  "And today 'A says B says C says D lies;  D says two lie and one speaks true.'  Answer: 'D lies; the rest speak truly.'  Wakeling adds that the Diary entry for 2 Jun 1896 (not given in the Lancelyn Green edition) says:  "Finished the solution of the hardest 'Truth-Problem' I have yet done",  but Carroll gives no indication what it was.

Lewis Carroll.  The problem of the five liars.  In his unpublished Symbolic Logic, Part II.  He was working on this after Part I appeared in 1896 and he had some galley proofs when he died in 1898.  Published in Lewis Carroll's Symbolic Logic, ed. by William Warren Bartley III; Clarkson N. Potter, NY, 1977, pp. 352‑361, including facsimiles of several letters of 1896 to John Cook Wilson (not in Cohen).  Each of five people make two statements, e.g. A says "Either B or D tells a truth and a lie;  either C or E tells two lies."  When analysed, one gets contradictions because a form of the Liar Paradox is embedded.

Carroll-Wakeling II.  c1890?  Prob. 32: Bag containing tickets, pp. 50 & 73.  This is one of the problems on undated sheets of paper that Carroll sent to Bartholomew Price.  Wakeling reproduces the MS. 

                    A bag contains 12 tickets, 3 marked 'A', 4 'B', 5 'C'.  One is drawn in the presence of 12 witnesses of equal credibility: three say it was 'A', four 'B', five 'C'.  What is the chance that it was 'A'?

          There is no answer on the Carroll MS.  Wakeling gives an answer.

                              Let the credibility of a witness be "a" when telling the truth.  Hence, the credibility of a witness when telling a lie is "1 - a".

                              If it was  A,  then 3 tell the truth, and 9 lie; hence the credibility is 3 in 12, or 1 in 4.

                              Therefore the chance that it is  A,  and no other, is:

                                        3/12 x 1/4  +  4/12 x 3/4  +  5/12 x 3/4  =  5/8

          I cannot see how this last formula arises.  Wakeling writes that he was assisted in this problem by a friend who has since died, so he does not know how the formula was obtained. 

                    Assuming  a  is the probability of a person telling the truth, this probability depends on what the chosen ticket is and is not really determined by the information given -- e.g., if the ticket was  A,  then any value of  a  between  0  and  1  is possible.  The value  3/12  is an estimate of  a,  indeed the maximum likelihood estimate.  If  k  of the  12  people are telling the truth, I would take the situation as a binomial distribution.  There are  BC (12, k)  ways to select them and the probability of having  k  liars is then  BC (12, k) ak (1-a)12-k.  Now it seems that Bayes' Theorem is the most appropriate way to proceed.  Our basic events can be denoted  A, B, C  and their a priori probabilities are  3/12, 4/12, 5/12.  Taking  a = k/12,  the a posteriori  probabilities are proportional to 

          k/12 x BC (12, k) (k/12)k ({12-k}/12)12-k,  for  k = 3, 4, 5.  Dropping the common denominator of  1213,  these expressions are  6.904,  8.504,  10.1913  times  1012.  Dividing by the total gives the a posteriori probabilities of  A, B, C  as  27.0%,  33.2%,  39.8%.

                    This is really a probability problem rather than a logical problem, but it illustrates that the logical problem seems to have grown out of this sort of probability problem.

A. C. D. Crommelin.  Problem given in an after-dinner speech, reported by Arthur Eddington in 1919.  ??where, ??NYS -- quoted in:  Philip Carter & Ken Russell; Classic Puzzles; Sphere, London, 1990, pp. 50 & 120-121.  Four persons who tell the truth once with probability  1/3.  If  "A  affirms that  B  denies that  C  declares that  D  is a liar, what is the probability that  D  was speaking the truth?"

Collins.  Fun with Figures.  1928.  The evidence you now give, etc., etc., pp. 22-23.  Three witnesses who tell the truth  1/3, 1/5, 1/10  of the time.  First two assert something which the third denies.  What is the probability the assertion is true?  Asserts it is  9  to  8,  which I also get.

Nelson Goodman.  The problem of the truth-tellers and liars.  Anonymous contribution to the Brainteasers column, The Boston Post (Jun 1929).  ??NYS -- described in an undated letter from Goodman to Martin Gardner, 1960s?, where he says he 'made it up out of whole logical cloth' and submitted it to the paper.

H. A. Ripley.  How Good a Detective Are You?  Frederick A. Stokes, NY, 1934, prob. 22: An old Spanish custom.  King will present the Princess's suitor a choice of two slips, one marked 'win', the other 'lose'.  The king is determined to double-cross the suitor so he has both marked 'lose'.  But the suitor realises this, so when he picks a slip, he drops it in the fire and then asks the King to reveal the other slip!

Rudin.  1936. 

No. 43, pp. 14-15 & 84.  9  statements with only  3  correct.

No. 85, pp. 29 & 92-93.  11  statements with at least  7  lies.  Makes a table to solve it.

No. 200, pp. 72 & 122.  Three truthtellers or liars.  First one is inaudible.  Second says the first claims to be a truthteller.  Third says the second is lying.  Author adds that there is just one liar and determines which each is.

Haldeman-Julius.  1937.  No. 5: Noblemen-huntsmen problem, pp. 3-4 & 20.  Noblemen tell the truth and huntsmen lie.  Same as Rudin, no. 200.  Says the problem was sent in by J. C. Furnas.

Hubert Phillips.  Question Time, op. cit. in 5.U.  1937. 

Prob. 10: Red and blue, pp. 6 & 178.  Involves truthtellers, liars and alternators. 

Prob. 25: Tom, Dick, and Harry, pp. 14 & 181.  Involves truthteller, liar and alternator.

Depew.  Cokesbury Game Book.  1939.  Mixed blood, p. 202.  Same as Rudin 200.

Catch-My-Pal cards, issued with The Hotspur, a magazine published by D. C. Thomson, Dundee, in 1939.  The cards are unnumbered, but the sixth in my set has three truthtellers or liars being asked what they are.  The reply of the first is inaudible.  The second says: He says he's a truthteller and he is a truthteller and so am I.  The third says: I'm a truthteller and they're liars.  Solution finds they are truthteller, truthteller, liar (TTL).  This assumes that all parts of the statements are true or false simultaneously.  If one assumes that a conjunctive statement is a lie if any one part is a lie, then there are more possibilities -- I get TTL, TLL, LLT, LLL.

M. Kraitchik.  Mathematical Recreations, 1943, op. cit. in 4.A.2, chap. 1, prob. 2, pp. 14‑15.  Truthtellers and liars.  (Not in Math. des Jeux.)  Same as Rudin 200, but doesn't give number of liars so only determines which the second and third are.

Harold Hart.  The World's Best Puzzles.  Op. cit. in 7.AS.  1943.  The problem of the nobles and the slaves, pp. 11 & 51.  Similar to Rudin, but doesn't say how many liars, but the statements are more elaborate so all can be determined.

Leopold.  At Ease!  1943.  Simpletons and liars, pp. 6-7 & 194.  Similar to Rudin.

Hubert Phillips.  Something to Think About, op. cit. in 7.AD, 1945.

Prob. 83‑85: Crazy island problems, pp. 51‑54 & 110‑112.

Prob. 83 involves three truthtellers or liars, not like Rudin.

Prob. 84 involves three truthtellers or liars or alternators. (Not the same as either problem in his Question Time, above.)

Prob. 85 involves three truthtellers or liars or Minimums, who tell the truth at most a third of the time.

Hubert Phillips.  Hubert Phillips's Heptameron.  Eyre & Spottiswoode, London, 1945.  Day 1, prob. 25: Crazy island, pp. 18 & 231.  Same as prob. 83 in Something to Think About.

Leeming.  1945.  Chap. 3, prob. 17: Which was the officer?, pp. 25‑26 & 155‑156.  Two truthtellers and a liar.

Gardner.  SA (Feb 1957) c= 1st Book, chap. 3, problem 4: The fork in the road.  Truthteller or liar.  The book version includes a number of letters and comments.  I have photocopies from Gardner's files of letters from people who claim to have invented this problem -- only one of these seemed reasonable -- cf Goodman above.  Other material ??NYR -- DO.

N. A. Longmore, proposer;  editorial solution.  The oracle of three gods.  RMM 4 (Aug 1961) 47  &  5 (Oct 1961) 59.  Truthteller, liar and alternator.

Robert Harbin.  Party Lines.  Op. cit. in 5.B.1.  1963.  The road to freedom, p. 30.  Truthteller or liar.

Charlie Rice.  Challenge!  Op. cit. in 5.C.  1968.  Prob. 8, pp. 22-23 & 55.  Truthteller or liar.

F. W. Sinden.  Logic puzzles.  In:  R. P. Dilworth, et al., eds.; Puzzle Problems and Games Project -- Final Report; Studies in Mathematics, vol. XVIII; School Mathematics Study Group, Stanford, Calif., 1968; pp. 197‑201.  The District Attorney, pp. 200‑201.  Two truthtellers and a liar -- determine which in two questions.

Doubleday - 2.  1971.  Truth will out, pp. 151-152.  Truthteller or liar.

Peter Eldin.  Amaze and Amuse Your Friends.  Piccolo (Pan), London, 1973.  No. 34: Where am I?, pp. 79 & 106.  You are on an island of truthtellers or an island of liars.  Determine which in one question.

Wickelgren.  How to Solve Problems.  Op. cit. in 5.O.  1974.  Pp. 36‑37.  He uses 'truar' for truthteller.  From statements by three truars or liars, you can deduce the number of each, though you can't tell which is which!!

Rowan Barnes-Murphy.  Monstrous Mysteries.  Piccolo, 1982.  Tollimarsh Tower, pp. 14 & 57.  Two monster guards, one truar and one liar, and you have one question.  You discover one is asleep and the other says: "It doesn't matter that he's asleep, he always tells people to do  A."  Do you do  A  or  not A?

Shari Lewis.  Abracadabra!  Magic and Other Tricks.  (World Almanac Publications, NY, 1984);  Puffin, 1985.  Free choice, p. 22.  Truthteller and liar have distributed items  A  and  B.  You want to determine who has which item with one question.  You ask "Did the liar take  B?"  If the person answers 'yes', he has item A; if 'no', he has item B.

Akihiro Nozaki.  How to get three answers from a single yes‑no question.  JRM 20:1 (1988) 59‑60.  You have to ask a truthteller or a liar which of three roads is correct.  The author's question results in neither being able to answer in the third case.  He suggests extensions.

Ken Weber.  More Five-Minute Mysteries.  Running Press, Philadelphia, 1991.  No. 13, pp. 59-61 & 189-190.  You have gotten lost near the border between the truthtellers and the liars and you come out in an open area where there is a border marker with a guard on each side, but you cannot tell which side is which.  Further the guards are walking back and forth and exchanging positions, so you are not sure if they are on their correct sides of the border or not.  You can approach one of the guards and ask one question to determine which side you are on.  You ask "Are we in your country?"  If he answers 'yes', you are in truthtelling country.

 

          9.H.   PRISONER'S DILEMMA

 

Charles Babbage.  On the Economy of Machinery and Manufactures.  (1832, ??NYS);  4th ed., (1835), reprinted by Augustus M. Kelley, NY, 1971. Section 348, p. 289.  "... both parties are often led to adopt arrangements ... at variance ... with the true interests of both."

Frederick Winslow Taylor.  The Principles of Scientific Management.  (1911);  Harper & Brothers, NY, 1923.  P. 10.  Speaking of employers and employés, he says "that perhaps the majority on either side do not believe that it is possible so to arrange their mutual relations that their interests become identical."

Merrill M. Flood & Melvin Dresher.  c1950.  ??NYS -- details.  They identified the paradox, but I have no reference to any publication.

Robert Axelrod.  The Evolution of Cooperation.  Basic Books, NY, 1986.  p. 216, ??NYS.  "The Prisoner's Dilemma game was invented in about 1950 by Merrill Flood and Melvin Dresher, and formalised by A. W. Tucker, shortly thereafter."

Keith Devlin.  It's only a game.  The Guardian, second section (17 Nov 1994) 12-13.  Says Tucker invented the dilemma in 1950.

Sylvia Nasar.  Albert W. Tucker, 89, pioneering mathematician.  New York Times (27 Jan 1995) ??  Asserts Tucker invented the dilemma when teaching game theory to psychology students at Stanford in 1950.

 

          9.I.     HEMPEL'S RAVEN PARADOX

 

Carl G. Hempel.  Studies in the logic of confirmation.  (Mid 1940s?).  Reproduced in:  M. H. Foster & M. L. Martin, eds.; Probability, Confirmation and Simplicity; Odyssey Press, NY, 1966, pp. 145‑183.  ??NYS.

 

          9.J.    FALLEN SIGNPOST

 

          How do you use a fallen signpost to find your way?

 

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 9:5 (Mar 1903) 490-491  &  10:1 (May 1903) 50-51.  The sense of direction.

King.  Best 100.  1927.  No. 22, pp. 14 & 44.

H. A. Ripley.  How Good a Detective Are You?  Frederick A. Stokes, NY, 1934, prob. 30: Class day.

Bernd Plage.  Der umgestürzte Wegweiser.  Verlag von R. Oldenbourg, München, 1944.  Prob. 1: Der umgestürzte Wegweiser, pp. 11-12 & 67.  (Kindly sent by Heinrich Hemme.)

John Paul Adams.  We Dare You to Solve This!.  Op. cit. in 5.C.  1957?  Prob. 180: On the right track?, pp. 67 & 121.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 84: A resourceful motorist, pp. 41 & 87.

 

          9.K.   CARROLL'S BARBER PARADOX

 

            Martin Gardner asked me to look up some of these items as he is doing a section on it in a book about Carroll which will be much more detailed than the following, citing numerous other discussions.  This presumably refers to the now-appeared Carroll-Gardner, 1996.

 

Carroll-Wakeling.  c1890?  Prob. 16: Going out, pp. 20-21 & 67-68.  This is on one of the undated sheets Carroll sent to Bartholomew Price.  Wakeling notes that it is an early version of the Barber shop paradox and says that Mark Richards has pointed out a mistake in Carroll's solution.

Lewis Carroll.  Diary entry for 31 Mar 1894.  Says he has just had a leaflet "A Disputed Point in Logic" printed containing the problem that he and John Cook Wilson "have been arguing so long."  ??NYS -- quoted in Carroll-Gardner.  Gardner says the pamphlet was revised in Apr 1894.

Lewis Carroll.  A logical paradox.  Mind (NS) 3 (No. 11) (Jul 1894) 436-438.  Gardner says Carroll reprinted this as a pamphlet.

Alfred Sidgwick.  "A logical paradox".  Mind (NS) 3 (No. 12) (Oct 1894) 582.

W. E. Johnson.  A logical paradox.  Mind (NS) 3 (No. 12) (Oct 1894) 583.

Lewis Carroll.  Diary entry for 21 Dec 1894.  Not in Lancelyn Green.  Discusses the problem and seems to recognise the distinction between material and strict implication.  ??NYS -- quoted in Carroll-Gardner.

Alfred Sidgwick & W. E. Johnson.  "Hypotheticals in a context".  Mind (NS) 4 (No. 13) (Jan 1895) 143-144.

E. E. C. Jones.  Lewis Carroll's logical paradox (Mind, N.S., 3).  Mind (NS) 14 (No. 53) (Jan 1905) 146-148.

W. [= John Cook Wilson, according to Gardner].  Lewis Carroll's logical paradox (Mind, N.S., 3 and 53, P. 146).  Mind (NS) 14 (No. 54) (Apr 1905) 292-293.  He admits that Carroll had been right all along.

R. B. Braithwaite.  Lewis Carroll as logician.  MG 16 (No. 219) (Jul 1932) 174-178.  Brief discussion and solution of the paradox.

Warren Weaver.  Lewis Carroll: Mathematician.  Op. cit. in 1.  1956.  Discusses the paradox.  Alexander B. Morris's letter says the paradox is not real.  Weaver's response discusses this and other unpublished letters, saying he is not sure if the paradox is resolved.

Carroll-Gardner.  1996.  Pp. 67-71 discusses this in detail, citing a number of other references, including John Venn, but he only gives the page numbers of an article in Symbolic Logic.

Lewis Carroll discusses this in his unpublished Symbolic Logic, Part II.  He was working on this after Part I appeared in 1896 and he had some galley proofs when he died in 1898.  Published in Lewis Carroll's Symbolic Logic, ed. by William Warren Bartley III; Clarkson N. Potter, NY, 1977.  Bartley includes all eight known versions. 

 

 

10.     PHYSICAL RECREATIONS

 

            See also 7.S and 7.Y.

            I will collect here some material on physics toys in general.

 

Christian Ucke.  Physics toys for teaching.  IN: H. Kühnelt, ed.; Interdisciplinary Aspects of Physics Education; proc. of conf. at Altmünster, Austria, 1989; World Scientific, Singapore, 1990, pp. 267-273.  1: Some new and not so well known literature about physics toys gives 12 references.  2: Presenting a database about physics toys.  3. Physics toy experiments with PET bottles.  16 references at the end, some repeating those in part 1.

 

          10.A. OVERTAKING AND MEETING PROBLEMS

 

            See Tropfke 588.

            Note.  Meeting problems include two pipe Cistern Problems, 7.H.  Overtaking problems include Snail in Well problems without end effect, 10.H, and Cisterns with one inlet and one outlet, 7.H.  Many of the Indian versions involve gaining or losing wealth rather than covering distance.  Versions going around a circle or an island are related to Conjunction of Planets, 7.P.6, and to problems of Clock Hands meeting, 10.R. 

            In the 17, 18 and 19 C, this problem was often discussed in relation to negative numbers as a change in the relative values leads to a negative solution -- cf: Clairaut;  Manning;  Hutton, 1798?;  Lacroix;  De Morgan, 1831? & 1836; 

 

            NOTATION.  There are five types of meeting (M) and overtaking (O) problems which recur frequently with slight variations.  I have recently converted all problems to this notation and I hope I have done it correctly.  Tropfke 590 gives a more extended classification which includes motion on a right triangle (see 6.BF.5) and on a circle (see 7.P.6, though some occur here) and alternating motion (see 10.H), but doesn't distinguish between problems where times are given and those where rates are given.

            M-(a, b).   Two travellers can cover a route in  a, b  (usually days).  They start at opposite ends at the same time toward each other.  When do they meet?  This is identical to the cistern problem  (a, b)  of 7.H.  Sometimes, the distance apart is given and the point of meeting is also wanted.  If the distance is, say  100, I will then say  "with  D = 100".  This corresponds to asking how much each pipe contributes to a cistern of capacity  D.  Sometimes, one starts later than the other.  If, say the first starts  2  later, I will say  "with the first delayed by  2".  This corresponds to opening one pipe later than the other.  This is the version of Tropfke's  I B a  where times are given.

            MR-(a, b; D).   The same problem except that  a, b  are the actual rates of the two travellers and hence  D  must be given.  This corresponds to a simple form of cistern problem which does not have the characteristic feature of giving the times required to do the entire task.  This is the version of Tropfke's  I B a  with rates given.

            MR-(a, b; c, d; D).   The same problem except that the travellers travel in arithmetic progressions, so this gives: 

            a + a+b + ... + a+(n-1)b  +  c + c+d + ... + c+(n-1)d  =  D.

Hence   MR-(a, 0; c, 0; D)  =  MR-(a, c; D).  The value of  d  is often  0.  One can interpret this as a cistern problem as for  MR-(a, b; D),  but it is even harder to imagine a pipe increasing its rate in arithmetic progression that to imagine a traveller doing so.  (An additional difficulty is that the traveller is usually viewed discretely while a pipe ought to be viewed continuously.)  This is the version of Tropfke's  I B b  with an arithmetic progression specified.

            O-(a, b).   Two travellers start from the same point at rates  a, b,  with the slower starting some time  T  before the faster, or they start at the same time at rates  a, b,  with the slower starting some distance  D  ahead of the other.  I.e. the slower has a headstart of time  T  or distance  D.  When does the faster overtake the slower?  This corresponds to a cistern with rates given, so that  O-(a, b)  with headstart  D, which is the same as  MR-(a, -b; D),  is a cistern problem with one inlet and one outlet.  When  a > b,  then this corresponds to a full cistern of size  D,  inlet rate  a  and outlet rate  b.  The case  a < b  is most easily viewed by negating the amount done, which interchanges inlet and outlet, and taking an empty cistern.  Hound and hare problems are basically of this form, with a headstart of some distance, but usually with rates and distances complicatedly expressed.  This is Tropfke's  I C a.  Sometimes the rates are not given explicitly, so I assume the first has the headstart.

            O-(a, b; c, d).  Two travellers start from the same point at the same time, but in arithmetic progressions.  When do they meet again?  This gives us: 

            a + a+b + ... + a+(n-1)b  =  c + c+d + ... + c+(n-1)d.  Some versions of this are in 7.AF.

If the first has a headstart of time  T,  then we either increase the first  n  by  T  or decrease the second  n  by  T,  depending on which number of days is wanted.  This is Tropfke's  I C b  with an arithmetic progression specified.  Sometimes the first has a headstart of distance  D.  Occasionally it is the second that has the headstart which is denoted by negative values of  T  or  D.

 

            Snail in the well problems without end effect (see 10.H) are special cases of meeting problems, usually  MR-(a, 0; D)  (Tropfke's  I A c).  When there are approaching animals, then it may be  MR(a, b; D)  or  O‑(a, b)  with headstart  D  (Tropfke's  I B c).

 

            Hound and hare problems.  Here one is often only given the ratio of speeds,  r,  so one can determine where the hare is caught, but not when.  In this case, the problem is  O-(a, ra)  with some headstart and one asks for the distance to overtaking, but not the time.  See:  Chiu Chang Suan Ching;  Zhang Qiujian;  Alcuin;  Fibonacci;  Yang Hui;  BR;  Bartoli;  Pseudo-dell'Abbaco;  AR;  The Treviso Arithmetic;  Ulrich Wagner;  HB.XI.22;  Calandri, c1485;  Calandri, 1491;  Pacioli;  Tagliente;  Riese;  Apianus;  van Varenbraken;  Cardan;  Buteo;  Gori;  Wingate/Kersey;  Lauremberger;  Les Amusemens;  Euler;  Vyse;  Hutton, c1780?;  Bonnycastle (= Euler);  King;  Hutton, 1798?;  D. Adams, 1801;  De Morgan, 1831?;  Bourdon;  D. Adams, 1835;  Hutton‑Rutherford;  Family Friend (& Illustrated Boy's Own Treasury);  Anon: Treatise (1850);  Brooks;  Clark;  Haldeman-Julius (in verse).

                        Cases where leaps differ in both time and distance:  Pacioli?;  Apian;  Cardan;  Wingate/Kersey;  Lauremberger;  Euler; Bonnycastle;  King;  Hutton, 1798? (= Lauremberger);  De Morgan, 1831?;  Bourdon (= Lauremberger);  Brooks;  Todhunter;  Mittenzwey;  Clark (= Lauremberger); 

 

            General versions.  Newton.

            Versions with geometric progressions.  Chiu Chang Suan Ching;  della Francesca;  Chuquet;  Pacioli;  Cardan.  See also 7.L.

            Versions with sum of squares.  Simpson.

            Circular versions.  Aryabhata(?);  AR;  Wingate/Kersey;  Vyse;  Pike;  Anon: Treatise (1850);  Todhunter;  Perelman.  See also 7.P.6, where problems with more than two travellers in a circle occur.

            Versions using negatives.  Clairaut;  Manning.

 

Chiu Chang Suan Ching.  c‑150.  (See also Vogel's notes on pp. 126‑127.)

     Chap. VI.

Prob. 12, p. 61.  O-(60, 100),  D = 100.  (Mikami 16 gives English.)

Prob. 13, pp. 62‑63.  Slower starts  10  ahead.  After the faster goes  100,  he is  20  ahead.  When did he overtake?

Prob. 14, p. 63.  Hare starts  100  ahead of hound.  Hound runs  250  and is then  30  behind.  How much further for hound to overtake?  (Swetz; Was Pythagoras Chinese?; p. 21, gives this with numbers  50, 125, 30.)

Prob. 16, p. 64.  Slower (the guest) goes  300  per day.  Faster (the host) starts after    day, catches the slower and returns after  ¾  of a day.  How fast is the faster?

Prob. 20, p. 66.  Birds can traverse a distance,  M-(7, 9).

Prob. 21, pp. 66‑67.  M-(5, 7)  with first delayed by  2.

     Chap. VII.

Prob. 10, p. 74.  Melon and gourd growing toward each other,  MR‑(7, 10; 90).

Prob. 11, pp. 74‑75.  Rush grows  3, 3/2, 3/4, ....  Sedge grows 1, 2, 4,. ...  When are they equal?  Solution of  2 6/13 = 2.4615  is obtained by linear interpolation between  2  and  3  days.  (Correct solution is  log2 6 = 2.5848.)  (English in Mikami 18.)

Prob. 12, pp. 75‑76.  Two rats separated by a wall  5  thick.  One gnaws  1, 2, 4, ....  The other gnaws  1, 1/2, 1/4, ....  When do they meet?  Solution of  2 2/17 = 2.1176  obtained by linear interpolation between  2  and  3.  (Correct answer is   log2 (2 + Ö6)  =  2.1536,   though Vogel has  8  instead of  6  in the radical.

Prob. 19, p. 79.  Two horses travel  193, 206, 219, 232, ...  and 97, 96 ½, 96, 95 ½, ....  They set out for a city  3000  away.  The faster gets there and then returns.  When does he  meet the slower?  Solution of  15 135/191 = 15.7068  obtained by linear interpolation between  15  and  16.  Correct answer is   (‑227 + Ö147529)/10  =  15.7095.

Zhang Qiujian.  Zhang Qiujian Suan Jing.  Op. cit. in 7.E.  468,  ??NYS.  Analogous to above hare and hound with values  37, 145, 23.  (English in  Sanford 212,  Mikami 41  and  H&S 74.)

Aryabhata.  499.  Chap. II, v. 30-31, pp. 72-74.  (Clark edition: pp. 40-42.)

V. 30.  One man has  a  objects and  b  rupees, another has  c  objects and  d  rupees, and they are equally wealthy.  What is the value of an object?  I.e. solve  ax + b = cx + d.  Though simple, this is the basis of many of the problems in this section, as discussed by Gupta (op. cit. under Bakhshali MS).  Clark gives an example with   a, b, c, d  =  6, 100, 8, 60.

V. 31.  Shukla translates this as describing simple meeting and overtaking problems,  MR-(a, b; D)  and  O-(a, b) with headstart D.  Clark translates it as having two planets separated by  D = s1 - s2  travelling with velocities  v1  and  v2,  so they meet in time   t  =  (s1 - s2) / (v1 ± v2).   This is discussed in Gupta, loc. cit.

Bakhshali MS.  c7C. 

Kaye I 43-44 describes the following types in general terms and indicates that general solutions are stated in the MS.  O-(a, b),  T = t.  O-(a, b; c, d).  O-(a, b; c, 0).  A problem equivalent to  MR(a, b; 2D).  Kaye I 49-52 describes problems of earning and spending which are equivalent to the following.  O-(a, b), headstart  2D.  O‑(a, b)  headstart  D.  Cistern problems with several inlets -- see 7.H.

Kaye III 171, f. 3r, sutra 15  &  Gupta.  General rule for  O-(a, b),  headstart  D.  (Gupta cites Aryabhata as giving the same rule.)

Kaye III 175, f. 4r  &  Gupta.  Examples of the previous with   a, b, D  =  5, 9, 7  and  18, 25, 8*18.  (The second is really an example with the first starting  T = 8  days ahead.)

Kaye III 189, f. 83r.  O-(3/2, 2),  D = 9.

Kaye III 216, ff. 60r-60v, sutra 52  &  Gupta.  Same as f. 3r, sutra 15, except that distance is replaced by wealth.  Example has a man with wealth  30  earning at rate  5/2  and spending at rate  9/3.  This can be considered as  O‑(5/2, 9/3),  D = 30.  This is also very like the simple versions of a snail climbing out of a well and could be considered as  MR(9/3 - 5/2, 0; 30).

Kaye III 217-218, ff. 60v-61v, sutra 53  &  Gupta  &  Hoernle, 1886, p. 146; 1888, p. 44.  Two men earn at rates  a, b,  but the first gives  c  to the second at the beginning.  When are they equally rich?  Examples with   a, b, c  =  5/3, 6/5, 7;   13/6, 3/2, 10.  This is equivalent to  O-(a, b),  D =  2c.  See also: G. R. Kaye;  The Bakhshāli manuscript;  J. Asiatic Soc. Bengal (2) 8:9 (Sep 1912) 349‑361;  p. 360, Sutra 53 and example, but he has a  +  for a  -  in his formula.

Kaye III 186, f. 31r  &  Gupta.  Another example of the above, with  a, b, c  =  7/4, 5/6, 7.

Kaye III 172, f. 8r  &  Gupta.  O-(a, b; c, 0), T = -e,  i.e.  a + (a+b) + ... + (a+(n‑1)b)  =  c(n+e).  In general this is a quadratic problem -- see below -- but when  e = 0,  a factor of  n  cancels leaving a linear problem.  Example:  O‑(2, 3; 10, 0). 

Kaye III 174-175, ff. 7v & 4r  &  Gupta.  O-(3, 4; 7, 0);  O‑(1, 2; 5, 0).

Kaye III 173, f. 9r  &  Gupta.  O-(1, 1; 10, 0).

Kaye III 176-177, ff. 4.r-5.r, sutra 18  &  Gupta  & in:  G. R. Kaye, The Bakhshāli manuscript;  J. Asiatic Soc. Bengal (2) 8:9 (Sep 1912) 349‑361; p. 358.  O‑(4, 3; 6, 1);  O-(2, 3; 3, 2)  and  O-(5, 6; 10, 3).  See also 7.AF.

Kaye III 174, f. 4v  &  Gupta.  This is a problem of the same type, but most of it is lost and the scribe seems confused.  Gupta attempts to explain the confusion as due to using the data   a, b; c, d  =  3, 4; 1, 2,   with the rule   n  =  2(c‑a)/(b-d) + 1,   where the scribe takes the absolute values of the differences rather than their signed values.  In this way he gets  n = 3  rather than  n = -1.

Kaye III 173, f. 9v.  Travellers set out at rates  a, b  to a destination  D  away.  The faster, on his return, meets the slower -- when?  This is equivalent to  MR‑(a, b; 2D).  Example with  a, b, D = 1, 6, 70.

Kaye III 190, f. 53v.  MR-(96/18, 27/108; 9).

Kaye I 44-46; III 177-178, ff. 5r-6v.  O-(3, 4; 5, 0),  T = -6,  which leads to a quadratic, but has answer  5.  O-(5, 3; 7, 0),  T = -5  is also given by Hoernle, 1886, pp. 128‑129; 1888, p. 33, and Datta, p. 42.  This also leads to a quadratic, but the answer is  (7 + Ö889)/6  and the MS gives an approximation for the root.  I have recorded that Kaye also has  O-(3, 4; 5, 0)  with no delay, but I can't find this.

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 30-31.  Sanskrit is on pp. 127-132; English version of the examples is on pp. 308-309.

                    V. 30, Ex. 1:  7x + 100  =  9x + 80.

                    V. 30, Ex. 2:  8x + 90  =  12x + 30.

                    V. 30, Ex. 3:  7x + 7  =  2x + 12.

                    V. 30, Ex. 4:  9x + 7  =  3x + 13.

                    V. 30, Ex. 5:  9x - 24  =  2 x + 18.

                    V. 31, Ex. 1:  MR-(3/2, 5/4; 18).

                    V. 31, Ex. 2:  O-(3/2, 2/3),  D = 24.

Anania Schirakatzi (= Ananias of Shirak).  Arithmetical problems.  c640.  Translated by:  P. Sahak Kokian as:  Des Anania von Schirak arithmetische Aufgaben; Zeitschrift für d. deutschösterr. Gymnasien 69 (1919) 112-117.  See 7.E for description.

Prob. 8.  Messengers,  O-(50, 80),  T = 15.

Prob. 16.  One mason working faster than the other:  O‑(140, 218),  T = 39.  [The 140 is misprinted as 218!]

Mahavira.  850.  Chap VI, v. 320‑327, pp. 177‑179.

V. 320:  O-(3, 8; 21, 0).

V. 321:  MR-(5, 3; 216).

V. 323:  O-(4, 8; 10, 2).

V. 325:  O-(5, 8; 45, -8).

V. 327:  O-(9, 13),  D = 100.

Alcuin.  9C.  Prob. 26: Propositio de campo et cursu canis ac fuga leporis.  Hound catching hare, hound goes  9  to hare's  7,  hare has  150  head start.  =  O-(7, 9),  D = 150.  (H&S 72 gives Latin and English.)  The actual rates are not given, only their ratio.

Sridhara.  c900. 

     V. 65‑67(i), ex. 81‑83, pp. 52‑53 & 95.  The verses give rules for various cases.

Ex. 81‑82.  O-(8/(5 ‑ ½),  3),  T =  6 ‑ ¼.

Ex. 83.  Slower goes  2,  faster goes  8  and then returns from  100  away.  When and where do they meet?

     V. 96‑98, ex. 111‑112, pp. 78‑79 & 96.

Ex. 111:  O-(3, 1; 10, 0).

Ex. 112:  O-(a, b; c, 2),  T = 6.  Determine  a, b, c  so that they will meet twice.  Answer assumes  a = 1,  b = 6  and first meeting is at  n = 10,  (which gives  c = 67)  and then asserts there is a second meeting at  n = 18.

al‑Karkhi.  c1010.  Sect. I, no. 5‑9, p. 82.

5:  O-(1, 1; 10, 0).

6:  O-(1, 1; 11, 0),  T = -5.

7:  O-(1, 2; 10, 0).

8:  O-(2, 2; 10, 0).

9:  10 + 15 + 20 + ... + 5(n+2)  =  325.  This could be considered as  MR‑(10, 5; 0, 0; 325).

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  Pp. 103f.  ??NYS -- quoted in Tropfke 593.

No. 4.  O-(6, 9),  T = 4.

No. 5.  O-(1, 1; 30, 0).

No. 6.  M-(5, 7).

Ibn Ezra, Abraham (= ibn Ezra, Abraham ben Meir  = ibn Ezra, Abu Ishaq Ibrahim al‑Majid).  Sefer ha‑Mispar.  c1163.  Translated by Moritz Silberberg as:  Das Buch der Zahl  ein hebräisch-arithmetisches Werk; J. Kauffmann, Frankfurt a. M., 1895.  P. 56.  Brothers meeting,  MR-(17, 19; 100).  (H&S 72 gives English.)  Silberberg's note 121 (p. 109) says a similar problem occurs in Elia Misrachi, c1500.

Fibonacci.  1202.  He has many examples.  I give a selection.

P. 168 (S: 261-262).  O-(1, 1; 20, 0).

P. 168 (S: 262).  O-(1, 2; 21, 0).

P. 168 (S: 262).  O-(2, 2; 30, 0).

Pp. 168-169 (S: 262).  O-(3, 3; 60, 0).

P. 169 (S: 262).  O-(5, 5; 60, 0).

P. 169 (S: 263).  O-(3, 3; 10, 0).  This has a non‑integral solution.  He computes  n = 5 2/3  days from the equation and then considers the travel on the  6th day to be at constant rates  10  and  18,  with the first starting  5  ahead, so the overtaking is at  5 5/8  days.

Pp. 177‑178 (S: 274):  De duobus serpentibus [On two serpents].  Serpents approaching,  MR‑(1/3 ‑ 1/4,  1/5 ‑ 1/6;  100).

Pp. 179‑180 (S: 276):  De cane et vulpe [On a dog and a fox].  O-(6, 9),  D = 50.

P. 182 (S: 280):  De duabus formicis quorum una imittatur aliam [On two ants, one of which follows the other].  An ant pursuing another.  O‑(1/3 ‑ 1/4,  1/5 ‑ 1/6;  100).

P. 182 (S: 280):  De duabus navibus se se invicem coniungentibus [On two ships that meet].  Two ships approaching,  M‑(5, 7).

Yang Hui.  Supplements to the Analysis of the Arithmetical Rules in the Nine Sections  (=?? Chiu Chang Suan Fa Tsuan Lei).  1261.  Repeats first problem of the Chiu Chang Suan Ching.

BR.  c1305.

No. 24, pp. 42‑43.  Ship goes  380,  starting  24  days after another and overtakes in  85  days.  How fast does the slower ship go?  This is  O-(a, 380),  T = 24  -- determine  a  such that the solution is  85.

            General form:  a, b  =  velocities of slower and faster ships;  n  =  time the faster ship sails;  n + T  =  time the slower ship sails, i.e. the slower ship has  T  days headstart or the faster ship is delayed by  T.  This gives  a(n+T) = bn.  The above problem has   b = 380,  T = 24,  n = 85   and asks for the other value, namely  a.  I will denote this by   (a, 380, 24, 85),   etc.

No. 41, pp. 60‑61.  (20, 25, 4, n)  =  O-(20, 25),  T = 4.

No. 42, pp. 60‑61.  (20, b, 4, 16).

No. 43, pp. 62‑63.  (20, 25, d, 16).

No. 44, pp. 62‑63.  (a, 50, 5, 10).

No. 45, pp. 62‑63.  (15, 20, 10, n)  =  O-(15, 20),  T = 10.

No. 46, pp. 64‑65.  MR-(31½, 21; 105).

No. 87, pp. 106‑107.  O-(24, 30),  T = 4.

No. 88, pp. 106‑107.  Hare is  40  leaps ahead but hound's leap is  13/11  of hare's.

No. 94, pp. 114‑117.  MR-(18, 22; 240).

Gherardi?.  Liber habaci.  c1310.

P. 144.  O-(1, 1; 25, 0).

P. 144.  Couriers,  M-(20, 30),  D = 200.

Lucca 1754.  c1330.  F. 59r, p. 134.  Couriers,  M-(20, 30)  with  D = 200,  though this is not used.

Munich 14684.  14C.  Prob. VIII & XX, pp. 78 & 81.   Discusses  O‑(1, 1; k, 0).

Folkerts.  Aufgabensammlungen.  13-15C.  18 sources for  O-(1, 1; c, 0).  Cites AR for extensions;  Chiu Chang Suan Ching, Alcuin, Fibonacci.

See Smith, op. cit. in 3.

Bartoli.  Memoriale.  c1420.  Prob. 28, f. 78r (= Sesiano, pp. 144 & 149-150.  Fox is  121  (fox-)steps ahead of a dog.  9  dog-steps  =  13  fox-steps.  He computes  9/13  of  121.  Sesiano notes that this assumes the fox stands still and determines how many dog-steps they are apart.  The correct answer, which assumes that both steps take the same time, is that the dog gains  4  fox-steps  for every  9  steps  he makes,  so he has to make  121 · 9/4  =  272 1/4  dog-steps,  which are equal to  121 · 13/4  =  393 1/4  =  121 + 272 1/4  fox-steps.

Pseudo-dell'Abbaco.  c1440. 

Prob. 91, p. 78 with plate on p. 79 showing hound chasing fox holding a chicken.  Fox is  40  fox‑steps ahead and  3 dog‑steps = 5 fox‑steps  (assuming both steps take the same time).  I have a colour slide of this.

Prob. 108, pp. 89‑91 with plate on p. 90.  M-(8, 5).  Asks for time to meet, but also gives  D = 60.

Prob. 118, pp. 96‑97.  O-(2, 2; 30, 0).  = Fibonacci, p. 168.

Prob. 119, pp. 97.  O-(3, 3; 18, 0).

AR.  c1450.  Several problems with arithmetic progressions which I omit.

Prob. 33, p. 37, 164‑165, 177, 223.  Hound and hare.  Hare  100  ahead and goes  7  for each  10  of the hound.

Prob. 148:  De planetis, pp. 72, 164‑165, 214.  Though described as conjunction by Vogel, this is really just the discussion of the general overtaking problem on a circle.  The text gives a general solution.

Benedetto da Firenze.  c1465.  P. 67.   O-(1, 1; 30, 0);   O-(1, 2; 30, 0);   O-(3, 3; 60, 0).

The Treviso Arithmetic = Larte de Labbacho.  Op. cit. in 7.H.  1478.

Ff. 54v‑55v (= Swetz, pp. 158‑160).  Messengers,  M-(7, 9)  with  D = 250.  (English is also in  Smith, Source Book I 12  and  Isis 6 (1924) 330.)

Ff. 55v‑56r (= Swetz, pp. 160‑161).  Hare & hound, rates  6  and  10,  starting  150  apart.  (English also in Isis 6 (1924) 330.)

Muscarello.  1478. 

F. 73v, p. 189.  Hare is  70  leaps in front of a hound.  Each hound leap is  7/5  as big as a hare's, but in the same time.  Says the hound catches the hare after  175  leaps.

F. 79v, p. 196.  Couriers,  M(70, 80).

Ff. 81r-81v, pp. 196-197.  Couriers,  O-(1, 2; 15, 0).

Ff. 82r-82v, pp. 198-199.  Couriers,  O-(2, 2; 15, 0).

della Francesca.  Trattato.  c1480.  F. 40r (102).  25 + 25n = 1 + 2 + 4 + 8 + ... + 2n-1 = 2n - 1.  The exact solution is  7.785540889.  He does linear interpolation on the seventh day, getting  7 73/103  =  7.7087378641.  English in Jayawardene.

Ulrich Wagner.  Das Bamberger Rechenbuch, op. cit. in 7.G.1.  1483.

Von Wandern, pp. 112 & 223,  O-(1, 1; 6, 0).

Regel vom Hasen, pp. 113 & 223‑224.  Hare goes  12  for hound's  15  and starts  100  ahead.

Chuquet.  1484.  See also 7.L for a problem with geometric progression.

Prob. 22.  English in FHM 204.  Meeting problem,  M-(7, 9).  He erroneously takes  7  and  9  as rates.

Prob. 98-108, FHM 220-221, are problems involving arithmetic progressions, usually overtaking problems, often with a fractional number of terms.  101 is a cask draining.

Prob. 98.  O-(1, 1; 11, 0).

Prob. 99.  O-(1, 1; 10, 0),  T = -3.

Prob. 105, English in FHM 220-221.  O-(3, 3; 14, 0).  Chuquet says another author, Berthelemy de Romans, gets  8 4/13  instead of  8 1/3.  FHM note that this assumes the speed is constant each day, while Chuquet assumes the speed increases through the day.

Prob. 117.  x + (x+1) + (x+2) + ... + (x+5)  =  30.

Prob. 127.  M-(7, 8).  See FHM 204 for comparison with prob. 22.

Borghi.  Arithmetica.  1484. 

Ff. 109v-109r (1509: f. 92r).  Two ships going between Venice and Padua,  MR‑(5, 8; 2400).  (H&S 74 gives Italian.)

F. 109r (1509: ff. 92r-92v).  O-(1, 1; 16, 0).

Calandri.  Aritmetica.  c1485. 

F. 92r, p. 184.  Ships meeting between Livorno and Marseilles --  M(7, 4).

F. 92v, p. 185.  Hare  3000  in front of hound who goes  8  while the hare goes  5.

HB.XI.22.  1488.  P. 52 (Rath 247).  (Rath says it is similar to Alcuin, but with different numbers.)

Calandri.  Arimethrica.  1491.  F. 97v. 

F. 65r.  O-(3, 3; 60, 0).

F. 66r.  Hare is  3000  hare leaps ahead of a hound, but the hounds step's are  8/5  as long as the hare's and take the same time.

F. 69r.  Ships meeting between Pisa and Genoa.  (3, 5).  Same woodcut as used for cistern problems -- see 7.H.

Pacioli.  Summa.  1494.

Ff. 39r-40r, prob. 1-6  =  the six problems on Fibonacci 168-169.

F. 41r, prob. 13.  Man travelling.  25+x + 2(25+x) + ... + 2x‑1(25+x)  =  300,  i.e.  (2x ‑ 1)(25 + x)  =  300.  Pacioli sets   y = x - 3.   After  3  days, he has gone  196 + 7y  and he travels  224 + 8y  on the  4th day.  Pacioli interpolates both linearly to get   196 + 7y + y(224 + 8y) = 300,  obtaining  x = 3.44341....  I get  x = 3.52567....

Ff. 41r-41v, prob. 14.  100  =  1 + 2 + 3 + ... + a  =  1 + 3 + 5 + ... + 2b-1  =  2 + 4 + 6 + .. + 2c  =  4 + 8 + 12 + ... + 4d.  When should each start so they all arrive at once?  (H&S 73 gives Italian and English and H&S 74 says he gives a problem involving flying about the earth, but I haven't found it.)

F. 41v, prob. 16-18.  These are examples of the general form treated in BR, i.e.  (a, b, T, n)  gives the equation   a(n+T)  =  bn.

     Prob. 16.  (30, 35, 5, n)  =  O-(30, 35),  T = 5.

     Prob. 17.  (32, b, 6, 25).

     Prob. 18.  (a, 35, 4, 20).

F. 41v, prob. 19.  O-(1, 1; 25, 0).

F. 41v-42r, prob. 20.   25n  =  1 + 2 + ... + 2n-2.   He gets  8 73/103 = 8.70874...  by interpolation on the  8-th day.  I get  8.78554....  = della Francesca.

F. 42r, prob. 21.  M-(4, 5)  with  D = 100.

Ff. 42v-43r, prob. 27.  Hound and hare.  Hare is  60  ahead and  5  hound leaps  =  7  hare leaps.  Says the problem is very unclear.  Assumes the  60  are hare leaps and interprets the relation as saying that  7  hare leaps are the same as  5  hound leaps and leaps take the same time.  Later, he supposes the  60  are hound leaps and converts this to  84  hare leaps.  See Cardan, whose description implies both the lengths and times of the leaps differ.

     Swetz, op. cit. as Treviso Arithmetic in 7.H, p. 244, says this is the first listing of  variants of the problem -- but we've seen lots of examples before and many of Pacioli's are taken from Fibonacci.  Swetz gives examples from  Wagner,  Calandri (1491)  and  Köbel (1514).

     PART II.

F. 64, prob. 80.  Overtaking on a circular island with speeds  1, 1/12.  Equivalent to  O‑(1, 1/12)  with headstart of distance 1.

F. 64, prob. 81.  Meeting on a circular island with speeds  1, 1/12.  Same as  M(1, 12).

Blasius.  1513.  F. F.iii.v: Decimasexta regula.  O-(10, 13),  T = 9.

Köbel.  1514.  ??NYS -- given in H&S 73.  O-(10, 13),  T = 9.

Tagliente.  Libro de Abaco.  (1515).  1541. 

Prob. 111, ff. 55v-56r.  Couriers meeting between Rome and Venice --  M-(17, 20).

Prob. 114, ff. 56v-57r.  O-(1, 1; 30, 0).

Prob. 118, f. 58r.  Ships meeting between Venice and Candia --  M‑(3, 5).

Prob. 122, ff. 59r-59v.  Hound chasing goat.  Goat is  50  leaps ahead and hound makes  7  to goat's  5.

Ghaligai.  Practica D'Arithmetica.  1521. 

Prob. 17, f. 64r.  O-(1, 1; 20, 0).

Prob. 18, f. 64r.  M-(4, 5)  with  D = 80.

Tonstall.  De Arte Supputandi.  1522. 

Quest. 35, p. 167.  O-(20, 33),  T = 6.

Quest. 37, pp. 168-169.  Couriers,  M-(3, 5).

Quest. 38, p. 169.  Couriers,  M-(⅓, ½).

Riese.  Die Coss.  1524.

No. 50, p. 47.  Messengers,  MR-(10, 8; 40).

No. 51, p. 47.  O-(6, 9),  D = 100.

No, 109, p. 54.  Messengers,  MR-(7, 9; 300).

No. 133, p. 59.  Fox is  300  fox leaps ahead of hound.  Each hound leap is  31/20  of a fox leap and they leap at the same rate.

Apianus.  Kauffmanss Rechnung.  1527. 

F. M.v.r.  Couriers between Leipzig and Venice.  M-(18, 24).

F. M.viii.v.  O-(7, 9),  D = 64.

Ff. M.viii.v - N.i.r.  Messengers between Prague and Vienna.  MR-(7, 9; 33).

F. N.i.r.  Hare is 30 leaps ahead of a hound.  Hound makes  8  leaps to hare's  6,  assumed same size.  I.e.  O-(8, 6),  D = 30.

Ff. N.i.r - N.i.v.  Hare is  50  hare leaps ahead of a hound.  Hare makes  4  leaps while hound makes  3,  but 2  hound leaps equal  3  hare leaps in length.  This can be viewed as  O-(8, 9),  where the units are  hare leaps per period in which a hare makes  8  leaps, with a head start of  50  hare leaps.

van Varenbraken.  1532.  ??NYS -- cited by the editor of King, 1795, p. 154.  Hound and hare --  O-(12, 15),  D = 200.

Cardan.  Practica Arithmetice.  1539.  Chap. 66.

Section 7, ff. BB.viii.v - CC.i.r (p. 137).   10n  =  1 + 4/3 + (6/5)(4/3) + (4/3)(6/5)(4/3) + (6/5)(4/3)(6/5)(4/3) + ....

Section 11, ff. CC.ii.v - CC.iii.r (p. 138) (iii  is misprinted  ii).  Hound & hare.  Hare is  60  hound leaps ahead.  Hound makes  63  leaps to hare's  100,  but  61  hound leaps equal  140  hare leaps.  Says Pacioli does it wrong, but I don't see that this is the same as Pacioli's problem??  [The phrasing of the problem is awkward.  H&S 72 gives the Latin and a mistranslation of it.]

Section 12, f. CC.iii.r (p. 139) (iii  is misprinted  ii).  O-(5, 3; 20, 0).

Section 13, ff. CC.iii.r - CC.iii.v (p. 139) (iii  is misprinted  ii).  1 + 2 + 4 + 8 + 16 + ... + 3 + 4 + 6 + 9 + 13 + ...  =  330.  He interpolates between the  7th  and  8th  days to get  7 95/137 = 7.69343,  while the correct(?) answer is  7.83659.  Says Pacioli could not do this.

Section 14, ff. CC.iii.v - CC.iiii.r (p. 139) (iii  is misprinted  ii).  Birds flying around the world.  1 + 2 + 3 + 4 + ... + 1 + 8 + 27 + 64 + ...  =  44310.  Seems to get  20½  days, but  20  is the answer.  (H&S 74 gives Italian and says similar problems are in Pacioli and Stifel.)

Section 15, ff. CC.iiii.r - CC.iiii.v (pp. 139-140).  

1 + 2 + 4 + 8 + ...   =   2 + 4 + 6 + 8 + ....

Section 66, ff. FF.ii.r - FF.ii.v (p. 155).  (66 is not printed in the Opera Omnia).  O‑(3, 5; 30, 0).

Section 154, ff. MM.iiii.r - MM.iiii.v (p. ??).  An overtaking problem, but with times for the entire route given which can be considered as  M-(10, -7)  with delay of  2,  giving   (n+2)/10  =  n/7.  Described in H&S 73.

Buteo.  Logistica.  1559. 

Prob. 21, p. 219.  A simple hound and hare problem,  O‑(6b/5, b),  D = 1.

Prob. 28, pp. 229-230.  States  O-(1, 1; 22, 0),  but does  O‑(1, 1; 18, 0).

Prob. 29, pp. 230-231.  O-(1, 1; 20, 0).

Prob. 30, p. 231.  O-(2, 2; 15, 0).

Prob. 31, pp. 231-233.  O-(24, -1; 10, 2).  Finds common distance of  199⅜  by linear interpolation on  11th  day.

Prob. 32, pp. 233-234.  M(3, 2)  with  D = 200  given.

Baker.  Well Spring of Sciences.  1562?  ??check if this in the Graves copy of the 1562/1568 ed.) 

Prob. 3,  1580?: ff. 36v-37r;  1646: pp. 62-63;  1670: pp. 76-77.  O-(1, 1; 8, 0).

Prob. 4,  1580?: ff. 37r-37v;  1646: pp. 63-65;  1670: pp. 77-78.  MR-(1, 1; 2, 2; 200).  He notes that they meet when the first has done  200/3 = 66 2/3,  which is  2/3  on the 12th day.  He assumes they travel at constant speed on each day, so they meet after  11 1/18 = 11.0555....  [18  is misprinted  28  at one point.]  But if the velocity is increasing linearly, so that the distance covered to time  t  is  t(t+1)/2,  then the problem leads to a quadratic with solution  t = 11.0578....

Gori.  Libro di arimetricha.  1571.

F. 72r (p. 78).  Deer is  60  ahead of a dog, which goes  7  for every  4  of the deer, i.e.  O-(4, 7),  D = 60.

F. 72v (p. 78).  Fox is  80  ahead of a dog, which goes  7  for every  5  of the fox, i.e.  O‑(5, 7),  D = 80.

F. 81v (p. 79).  Couriers between Paris and Siena meeting --  M(20, 24)  with  D = 800,  but not used.

F. 82r (p. 79).  Couriers meeting --  MR(40, 50, 1000).

F. 71r (p. 79).  O-(1, 1; 30, 0).

F. 81v (p. 79).  O-(1, 1; 40, 0).

van Halle.  1568.  ??NYS -- cited by the editor of King, 1795, p. 155.  O‑(11, 0; 1, 1).

Io. Baptiste Benedicti (= Giambattista Benedetti).  Diversarum Speculationum Mathematicarum, & Physicarum Liber.  Turin, (1580), Nicolai Bevilaquæ, Turin, 1585; (Venice, 1599).  [Rara 364.  Graves 141.f.16.]  This has a number of overtaking and meeting problems, but he makes diagrams showing the sums of the arithmetic progressions involved.

Theorema CVI, pp. 68-69 (misprinted 70-71).  O-(4, 1; 1, 1)  with diagram showing  1 + 2 + ... + 7 = 4 x 7.

Theoremae CVII & CVIII, pp. 69 (misprinted 71) - 70.  O-(k, 0; 1, 2)  with diagrams showing  1 + 3 + ... + 15  =  8 x 8  and  1 + 3 + ... + 13  =  7 x 7.

Theoremae CIX & CX, pp. 70-71.  O-(k, 0; 2, 2)  with diagrams showing  2 + 4 + ... + 14  =  7 x 8  and  2 + 4 + ... + 16  =  8 x 9.

Theorema CXI, p. 71.  O-(k, 0; 3, 3)  with diagram showing  3 + 6 + ... + 21  =  7 x 12.

Theorema CXII, pp. 71-73.  O-(11, 0; 3, 3),  which does not have an integral solution.  He draws a diagram of  3 + 6 + ... + 21  and tries to make it into a rectangle of side  11,  but sees this does not work.  He then sees the answer is between  6  and  7  days, so he does linear interpolation over this period, getting  6.3  days instead of the correct  6 1/3.  He then uses a simple triangle to show this interpolation process.

Theorema CXIII, pp. 73-74.  O-(a, b)  with first having headstart of time  T  done in general.  Does  O-(20, 25),  T = 8.  Gives a diagram of  a x n+T  =  b x n.  There is an Appendix to this Theorema on pp. 74-75 which applies it to Jupiter overtaking Saturn but I can't make it out.

Theorema CXIIII, pp. 75-77.  M-(9, 11)  starting  D = 400  apart.  Then does the problem in general, giving a diagram which I can't follow.

Theorema CXV, pp. 77-78.  MR-(10, 15; 100),  then  MR-(a, b; D)  in general.

Wingate/Kersey.  1678?. 

Quest. 14, pp. 485-486.  O-(14, 22),  T = 8.

Quest. 15, pp. 486-487.  O-(9, 7)  on a circular island of circumference  36.

Quest. 16, p. 487.  O-(8, 0; 1, 1).

Quest. 17, pp. 487-488.  M-(8, 6; 140).

Quest. 18, pp. 488-489.  M-(a, a + 5/2; 100)  takes time  8.

Quest. 19, pp. 489-490.  M-(11/2, 17/3; 134)  going opposite ways round a circular island.

Quest. 23, pp. 491-492.  Hare is  100  hare-leaps ahead of a hound.  Hare takes five leaps while the hound takes four, but three hound leaps equal four hare leaps.

Edward Cocker.  Arithmetic.  Op. cit. in 7.R.  1678.  Chap. 10, quest. 32.  1678, p. 181;  1715: p. 121;  1787: p. 106.  This problem is attributed to Moor's or More's Arith. cap. 8, qu. 7.  1678 states  O-(40, 50),  T = 3,  giving answers  12  days and  600  miles.  1715 states  O-(40, 50),  T = 3,  giving answers  32  days and  600  miles.  1787 states  O-(48, 50),  T = 3,  giving answers  12  days and  600  miles.  I am surprised at the misprints here.

Wells.  1698. 

No. 109, p. 207.  Meeting:  MR-(1, 3; 50).

No. 116, p. 208.  Overtaking:  O-(6, 10),  T = 8.

Peter Lauremberger (Petrus Laurembergus).  Institutiones Arithmeticæ ....  4th ed., Joh. Lud. Gleditsch, Leipzig, 1698.  P. 196, prob. 12.XII.  Fox is  60  hare-leaps ahead of a hound.  She makes  9  leaps while the hound does  6,  but  7  hare-leaps are as long as  3  hound-leaps.

Isaac Newton.  Arithmetica Universalis, 1707.  ??NYS.  English version:  Universal Arithmetic, translated by Mr. Ralphson, with revisions and additions by Mr. Cunn, Colin Maclaurin, James Maguire and Theaker Wilder; W. Johnston, London, 1769.  (De Morgan, in Rara, 652‑653, says there were Latin editions of 1722, 1732, 1761 and Raphson's English editions of 1720 and 1728, ??NYS.)  Resolution of Arithmetical Questions, Problem V, pp. 180‑184.  Begins with  MR(7/2, 8/3; 59)  with second delayed by 1.  Then "The same more generally" does  O(c/f, d/g)  with second having a headstart of distance  e  and either a headstart or delay of time  h  and  MR(c/f, d/g; e)  with second having either a headstart or delay of time  h.  Then does example:  O(13, 1)  with second starting distance  90  ahead, but first delayed by  3  days.  Then repeats original example from general viewpoint.

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790).

Prob. XIX, p. 89 (1790: prob. XXXI, p. 92).  O-(28, -2; 20, 0).

Prob. L, pp. 113-115.  MR-(40, -2; 20, 2; 360).

1790: Prob. LXIII, pp. 114‑116.  MR-(60, -5; 40, 5; 500).

Prob. LI, pp. 115-116 (1790: prob. LXIV, pp. 116‑117).  Overtaking.  8 + 12 + ... + (4+4n)  =  1 + 4 + ... + n2.

Alexis-Claude Clairaut.  Élémens d'algèbre.  1746.  ??NYS -- cited by Tom Henley.  Discusses the relation between overtaking and meeting problems and the use of negative rates or negative headstarts to make them algebraically the same.

Les Amusemens.  1749. 

Prob. 113, p. 255.  O-(2½, ½; 7, 0).

Prob. 114, p. 256.  O-(5, 1; 2, 0).

Prob. 115, p. 257.  Meeting:   4n + n+2 + n+4 + n+6 + n+8  =  104.

Prob. 116, p. 258.  O-(10, 5; 30, 0).

Prob. 117, p. 259.  Cat and mouse,  O-(3, 5),  D = 23.

Euler.  Algebra.  1770.  I.IV.III: Questions for practice.

No. 16, p. 205.  Privateer and prey,  O-(18, 20),  D = 8.  [The numbers  8  and  18  are interchanged in the text.]

No. 25, p. 206.  Hare  50  hare leaps ahead of greyhound.  Hare makes  4  leaps to greyhound's  3,  but hare's leaps are only    as long.

Vyse.  Tutor's Guide.  1771? 

Prob. 57, 1793: p. 68; 1799: p. 74 & Key p. 99.  O-(22, 32),  T = 4.  Misprint in solution.

Prob. 60, 1793: p. 69; 1799: p. 74 & Key P. 100.  MR-(x, x+2½; 135)  with meeting after  8.

Prob. 65, 1793: p. 69; 1799: p. 75 & Key p. 101.  MR-(2, 3; 170)  with second delayed by 8.

Prob. 67, 1793: p. 70; 1799: p. 75 & Key p. 102.  Hound & hare,  O-(21, 15),  D = 96.  The actual rates are not given, only their ratio, so one can determine where the hare is caught, but not when.

Prob. 12, 1793: p. 78; 1799: p. 84 & Key p. 109.  A  and  B  start to circle a wood of circumference  135,  starting at opposite sides and going in the same direction.  A  goes  11/2,  B  goes  17/3.  When do they meet?  O‑(11/2, 17/3),  D = 67½.

Prob. 13, 1793: p. 186; 1799: p. 198 & Key p. 241.  Two rowers who can row at  5  set out towards each other at points  34  apart on a river flowing  2½.  Though this appears to belong in Section 10.G, it is simply  MR-(2½, 7½; 34).

Prob. 6, 1793: p. 189; 1799: p. 201 & Key p. 245.  Hare starts  5 rods  and  34 sec  before a greyhound and goes at  12 mph,  while the hound goes  20 mph.  Using  320 rods to the mile, this is  O-(16/15, 16/9)  where the slower has a headstart of  T = 34  and  D = 5.

Dodson.  Math. Repository.  1775.

P. 1, Quest. I.  MR-(8, 7; 150)

P. 2, Quest. V.  O-(30, 42),  T = 4.

P. 25, Quest LXIV.  Same as Euler's no. 25.

P. 57, Quest. CX.  MR-(7/2, 8/3; 59)  with second starting one unit of time later.

Pp. 58-59, Quest CXI.  General solution of  O-(a, b)  with one having a headstart and starting before or after the other.  Does  O-(10/3, 5/2),  D = -59,  T = 4;  then  O‑(10/3, 5/2),  D = -59,  T = -4.

P. 178, Quest XXIV.  O-(8, 0; 1, 2).

P. 179, Quest XXVI.  MR-(40, -2; 20, 2; 360).

P. 183, Quest. XXX.  8 + 12 + 16 + ...  =  1 + 4 + 9 + ...

Pp. 191-192, Quest XLII.  1 + 2 + 3 + ...  +  1 + 8 + 27 + ...  =  462.

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?] 

1804: prob. 68, p. 139.  Hare starts  40  yards before a hound and goes at  10  mph.  The hound doesn't see the hare for  40  seconds, and then goes at  18  mph.

1804: prob. 69, p. 139.  Exeter is  130  miles from London.  A  sets out from Exeter at  8  am going  3  mph;  B  sets out from London at  4  pm going  4  mph.  Where do they meet?

1804: prob. 71, p. 140.  Lincoln is  100  miles from London.  Travellers set out at the same time and meet after seven hours, when they find that  A  has gone    mph faster than  B.  What are their rates?

Bonnycastle.  Algebra.  1782. 

P. 84, no. 5.  MR-(8, 7; 150).

P. 85, no. 19 (1815: p. 107, no. 31).  Same as Euler's no. 25.

Pike.  Arithmetic.  1788.  P. 350, no. 17.  Circle  268  in circumference.  Two men start at ends of a diameter and go in the same direction at rates  22/2  and  34/3.  When and where do they meet?   I. e.  O-(34/3,  22/6),  D = 134.

Eadon.  Repository.  1794. 

P. 80, no. 36.  O-(30, 42),  T = 4.

P. 80, no. 37.  MR-(23, 31; 162).

P. 235, ex. 3.  5 + 9 + 13 + ... + (5 + 13x4).  Find total travel.

John King, ed.  John King  1795  Arithmetical Book.  Published by the editor, who is the great-great-grandson of the 1795 writer, Twickenham, 1995. 

P. 109.  A meeting problem, but the times to meeting and the meeting point are given, so it reduces to:   a + a+2 + a+4 + a+6 + a+8  =  50  =  b + b+3 + b+6 + b+9.

Pp. 117-118.  Hare is  144  hare leaps ahead of a grayhound.  Hare makes  4  leaps while grayhound makes  3,  but grayhound leaps are  3/2  as big.

Thomas Manning.  An Introduction to Arithmetic and Algebra.  2 vols., Nicholson, Lunn & Deighton, Cambridge, 1796 & 1798.  Vol. 1, pp. 208-210.  O-(5/2, 7/2),  T = 2.  Then considers what would happen if the second traveller went slower than the first [more simply, suppose the second traveller had the headstart, i.e.  T  is negative].  This gives a negative solution and he interprets this as that they must have met before the starting time.  He gives a general solution and discussion of the problem.  ??NX.

Hutton.  A Course of Mathematics.  1798? 

Prob. 5,  1833: 210-211;  1857: 214-215.  Hare is  60  hare-leaps ahead of a greyhound.  She makes  9  leaps while the hound does  6,  but  7  hare-leaps are as long as  3  hound-leaps.  = Lauremberger.

Prob. 9,  1833: 213-214;  1857: 217-218.  O-(31½/5, 22½/3),  T = 8.  Then does  O‑(m/t, m'/t')  with first having headstart of  T. 

Prob. 20,  1833: 221;  1857: 225.  M-(8, 7, 150).

Remarks upon Equations of the First Degree,  1833: 224-231;  1857: 228-235,  is an extensive discussion concerning possible negative roots and considers  O-(m, n)  with the second having a headstart of distance  D.  When  m < n,  he says the directions must be reversed.

D. Adams.  Scholar's Arithmetic.  1801.

P. 134, no. 6.  O-(20, 25),  T = 5.

P. 208, no. 64.  Hare is 12 rods (= 198 ft) ahead of a hound and goes for 45 sec before the hound starts, running at 10 mph.  Hound then starts at 16 mph.  How long until the hound catches the hare and how far does the hound go?  = O‑(44/3, 352/15),  D = 858,  using ft and sec.

Robert Goodacre.  Op. cit. in 7.Y.  1804.  Miscellaneous Questions, no. 125, p. 205 & Key p. 269.  A  goes  4  mi/hr for  7  hrs each day.  B  starts a day later at  5  mi/hr for  8  hours each day, both starting at the same time each morning.  When does  B  overtake?

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Sections 64-75, pp. 94-110  plus  Addition, pp. 359-360.  Discusses general problems  MR-(a, b; D)  and  O-(a, b)  and considers negative solutions and what happens when the divisor is zero!

Augustus De Morgan.  Arithmetic and Algebra.  (1831 ?).  Reprinted as the second, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  Art. 116, pp. 30-31.  Hare is  80  hare-leaps ahead of a greyhound.  Hare makes  3  leaps for every  2  of the hound, but a hound-leap is twice as long as a hare‑leap.  Then considers a hound-leap as  n/m  of a hare-leap.  Takes  n/m = 4/3  and finds a negative solution which he discusses.  Takes  n/m = 3/2  and finds division by zero which he interprets as the hound never catching the hare.

Bourdon.  Algèbre.  7th ed., 1834.

Art. 47, prob. 3, pp. 65-66.  Same as Hutton, 1798?, pp. 210-211, with greyhound chasing a fox.  = Lauremberger.

Art. 190, question 6, p. 319.  O-(10, 0; 3, 2).

D. Adams.  New Arithmetic.  1835. 

P. 243, no. 81.  O-(6, 8)  with headstart of distance  24.

Pp. 243-244, no. 82.  Hare is distance  12  and time  5/4  ahead of a hound.  Hare runs  36  and  hound runs  40.  When and where does the hound catch the hare?

Augustus De Morgan.  On The Study and Difficulties of Mathematics.  First, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  P. 30 mentions  O-(2, 3),  T = 4  and  O-(a, b) with second delayed by time  T.  Pp. 37-39 discusses the general courier problems  O-(m, n)  with second having headstart of distance  D  and  MR-(m, n; D).  He considers different signs and sizes, getting six cases.

Unger.  Arithmetische Unterhaltungen.  1838.  Pp. 135 & 258, nos. 515 & 516.  MR‑(5/4, 3/5; 57)  with first delayed by time  2½.  The second problem asks what delay of time for the first will make them meet at the half-way point?

Hutton-Rutherford.  A Course of Mathematics.  1841?

Prob. 7,  1857: 81.  Two persons on opposite sides of a wood of circumference  536,  start to walk in the same direction at rates  11, 11⅓.  How many times has the wood been gone round when they meet?  =  O-(11, 11⅓),  D = 268.  Answer is the number of times the faster has gone round.

Prob. 23,  1857: 82.  MR-(3, 4; 130),  T = 8.

Prob. 37,  1857: 83.  Hare starts  40 yd  and  40 sec  in front of a hound.  Hare goes  10 mph (=  44/9 yd/sec)  and hound goes  18 mph (= 44/5 yd/sec).  I.e.  O-(44/9, 44/5),  D = 40,  T = 40.

Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.  P. 46, no. 15.  Man makes a journey at  4  mph and returns at  3  mph, taking  21  hours in total.  How far did he go?

Family Friend 1 (1849) 122 & 150.  Arithmetical problems -- 1.  "A hare starts  40  yards before a greyhound, and is not perceived by him till she has been up  40  seconds:  she gets away at the rate of  10  miles an hour and the dog pursues her at the rate of  18  miles an hour:  how long will the course last, and what distance will the hare have run?"  = Illustrated Boy's Own Treasury, 1860, Prob. 4, pp. 427 & 431.  = Hutton, c1780?, prob. 68.

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  3rd ed., 1850.  Op. cit. in 7.H.

Pp. 198-199, no. 117.  Hare is  50  springs ahead of a hound.  Their springs are of equal length, but the hound makes  27  while the hare makes  25.  How many springs does the hare make before being overtaken?  Answer is misprinted  675  instead of  625.

P. 360, no. 47.  = Vyse, prob. 12.

P. 360, no. 48.  See in 7.P.6.

John Radford Young (1799-1885).  Simple Arithmetic, Algebra, and the Elements of Euclid.  IN: The Mathematical Sciences, Orr's Circle of the Sciences series, Orr & Co., London, 1854.  [An apparently earlier version is described in 7.H.]

No. 5, p. 177.  O-(9, 5),  T = 10.

No. 5, p. 216.  A  and  B  travel  90  miles.  A  goes  1  mile per hour faster than  B  and arrives  1  hour before him.  What were their speeds?

No. 8, p. 216.  Same as last with data  150,  3,  8⅓.

No. 4, p. 228.  MR-(1, 1; 20, -2; 165).

Vinot.  1860.  Art. LXIII: Problème du Renard et du Lévrier.  Fox is  72  (fox-)leaps in front of a greyhound.  Fox makes  9  leaps to hound's  5,  but fox-leaps are  3/7  the size of hound-leaps. 

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Many examples.  I mention just one example.

                    1863 -- pp. 132-13, no. 19;  1873 -- p. 161, no. 18.  "E  takes  60  steps before he is overtaken by  D;  how many steps does  D  take to catch  E,  provided  E  takes  4  steps while  D  takes  3,  and  5  of  D's  equal  8  of  E's,  and how far ahead was  E  when they started?"

Todhunter.  Algebra, 5th ed.  1870.  Several examples, including the following.

Examples X, no. 26, pp. 86 & 577.  Hare is 80 hare leaps ahead of a greyhound who starts after her.  Hare does three leaps for every two of the hound, but hare leaps are half the size of hound leaps.  How many leaps does the hare make before it is caught?

Examples XXX, no. 53, pp. 268 & 589.  O-(5, 0; 3, ½)  with first having a headstart of time  4½.

Examples XXXIII, no. 7, pp. 285 & 590.  O-(1, 1; 12, 0)  with first having headstart of time 5.  [He says the second overtakes the first after the first travels  36,  i.e. after time  8,  but the first then overtakes the second after time  15!]

Miscellaneous Examples, no. 77, pp. 550 & 605.  A  and  B  start on a circular walking race.  After 30 min,  A  has done three circuits and  B  has done  4½.  When do they meet again?  This is just  O-(6, 9)  with the first having a headstart of distance  ½.

Mittenzwey.  1880.  Prob. 78, pp. 15 & 66;  1895?: 85, pp. 19 & 69;  1917: 85, pp. 18 & 65.  Hare is  50  leaps ahead and makes  6  leaps while the hound makes  5,  but  7  houndleaps are as long as  9 hare leaps.  How leaps does the hare make before being caught?

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.  No. 4, pp. 166 & 334.  General problem:   ax  =  b (n - x).   Solves for  n ‑ x.

Clark.  Mental Nuts.  1897, no. 46;  1904, no. 62.  The fox and the hound.  Fox is  60  (fox‑)leaps in front of a hound.  Hound takes  6  leaps to fox's  9,  but the hound leaps are  7/3  times as long as the fox's.  (= Lauremberger)

Dudeney.  Weekly Dispatch (17 May 1903) 13  &  (14 Jun 1903) 16.

Perelman.  1937.  MCBF.  At the cycle track, prob. 150, pp. 255-256.  Circular track of circumference  170m.  When cyclists are going in opposite directions, they meet every  10  sec;  when going the same direction, the faster passes the slower every  170  sec.

Haldeman-Julius.  1937.  No. 109: Hare and hounds problem, pp. 13 & 26.  Gives a verse version -- the only one I have seen -- and says the problem 'is about 150 years old.'

                    As I was walking o'er my forest grounds

                    Up jumped a hare before my two greyhounds.

                    The distance that she started up before

                    Was fourscore rods, just, and no more.

                    My dogs did fairly run

                    Unto her 16 rods just 21.

                    Now I would have you unto me declare

                    How far they ran before they caught the hare.

          This is  O-(a, [21/16]a)  with headstart  D = 80.

C. Dudley Langford.  Note 1558:  A graphical method of solving problems on "Rate of Work" and similar problems.  MG 25 (No. 267) (Dec 1941) 304-307.  +  Note 2110:  Addition to Note 1558:  "Rate of Work" problems.  MG 34 (No. 307) (Feb 1950) 44.  Uses a graph to show  (a, b)  cistern problems as meeting problems.  Also solves problems  (A, x)  in  B  and  (a, -b),  the latter appearing as an overtaking problem.  The Addition gives a clearer way of viewing  (a, ‑b)  problems as overtaking problems

Gamow & Stern.  1958.  Pp. 9‑10, 59‑63.  Elevator problem.

 

          10.A.1.        CIRCLING AN ARMY

 

            The linear form has an army of length  L  moving with velocity  v.  A rider goes at velocity  V  from the rear to the front and then back to the rear, reaching the rear when it has advanced  d.  How far,  D,  does he go?  Since one version gives an erroneous answer, I will give the basic equations.  Let  t  be the time for the rider to reach the front of the army and  T  be the time to get from the front back to the rear.  We then have:  Vt = vt + L;  v(t+T) = d;  d = V(t-T);  D = V(t+T).  Here may be insufficient equations to determine all the values (as also occurs in 10.A.3), but the value of  D  can be found as   D  =  L + Ö(L2 + d2).   Note that  D/d = V/v.  Other versions of the problem can be solved, sometimes more simply, e.g.  L = (V2 - v2)d/2vV;  if  V = rv,  then  L = (r2-1)d/2r.

            If the army has a width  W,  and the rider goes across the moving army at each end, the situation is more complex -- see the first example and Loyd.

 

J. Gale, proposer;  Joseph Edwards, Jr.  &  Mr. Coultherd, solvers.  Question III.  A Companion to the Gentleman's Diary; ... for the year 1798, pp. 59-60  &  The Gentleman's Mathematical Companion, for the year 1799, pp. 16-17.  Wagoner walking around his wagon and team while it is travelling.  L = 20 yd,  W = 4 yd.  He can walk  V = 4 mph  and he walks  74⅔ yd  in his circumambulation (hence taking  T = 7/11 min).  A complication arises as to how he goes crossways.  The proposer says he turns at right angles and passes  2 yd  clear at the front and at the back.  (He also says the walker passes  2 yd  clear on each side, but he never gives a width, taking the distance between the two side paths as  4,  which I have accounted for by taking  W = 4.  In some cases, we can account for the  2 yd  at each end by taking  L = 24.)  How fast,  v,  is the team going?  Both solvers get  v = 2 mph, but I find neither is viewing the problem correctly and neither has correctly formulated the problem he has described! 

                    Edwards starts in the middle, 2 in front of the horses, and says the the walker must go 2 to the left, 24 back and 2 to the right to get the the middle in back.  He then says the time required is  28/(V+v).  But this assumes the crossways motion is at the same relative speed as the lengthwise relative motion and seems definitely incorrect to me.  Edwards then says that a similar argument gives the time in the other direction as  28/(V-v).  Setting the sum of these equal to  T  does give  v = 2.

                    Coultherd starts in the middle, 2 behind the wagon, but I will rephrase it to parallel Edwards' solution.  It takes the man  2/V  to go  2  to the left.  During this time, the wagon moves ahead  2v/V,  so he is now only  2 - 2v/V  in front.  When he gets to the back, he only needs to be  2 - 2v/V  behind the wagon when he turns to cross, so that he will be exactly 2 behind at the middle.  So he must make a relative motion of  24 ‑ 4v/V  at the relative velocity  V + v.  Similarly, for the forward trip, he makes relative motion  24 + 4v/V  at relative velocity  V - v.  Adding the times for these to  8/V  for the crossways trips and setting equal to  T,  I get  v = 2.056 mph, which seems to be the correct answer.  Coultherd forgets to account for the  2 - 2v/V  at the end of his first lengthwise trip and confuses distance travelled at  V  and at  V-v,  which simplifies his algebra to the same final equation as Edwards. 

                    I am rather surprised at the basic errors in both solutions.

Clark.  Mental Nuts.  1897, no. 88;  1904, no. 98;  1916, no. 99.  A West Pointer.  Column is  25  miles long.  Courier goes from the rear to the front and returns to the rear and sees that he is now where the front of the column was when he started.  I.e. his trip takes the same time as the time the column moves 25 miles.  How far did he go?  Answer is  60 miles, 1876 feet,  which is correct.  Here  L = 25 = d,  so  D = 60.36.

Loyd.  The courier problem.  Cyclopedia, 1914, pp. 315 & 382.  (= MPSL2, prob. 146, pp. 103 & 167‑168, with solution method provided by Gardner -- Loyd only gives the values "following the rule for solving puzzles of this kind".)  Army  50  miles long.  Rider goes from back to front to back of the army in the time it moves forward  50  miles.  Here  L = 50 = d,  so  D = 120.71  -- Loyd says "a little over 120 miles".  Note this is the double of Clark's problem. 

                    Loyd also extends to the case of a square army and Gardner gives a solution, assuming the courier goes across the ends on an angle with velocity  V,  so his crossways velocity is  w  =  Ö(V2-v2).  The total time of his trip is then  T  =  L/(V‑v) + L/(V+v) + 2W/w.  Assuming  that the army advances its own length in this time, i.e.  vT = L,  and setting  x = V/v  leads to  L  =  L/(x‑1) + L/(x+1) + 2W/Ö(x2-1).  Assuming  W = L  simplifies the expression, but it remains a fourth degree equation.  Gardner says the only relevant solution is  x = 4.18112+.  The courier's distance is  VT = xvT = xL = 209.056+.

Abraham.  1933.  Prob. 67 -- The column of troops, pp. 33 & 44 (19 & 116).  Rider circling army -- same as first part of Loyd.

Haldeman-Julius.  1937.  No. 74: Train problem, pp. 10 & 24.  Brakeman walks from rear of a train going at  V = 27 mph,  to the front, thereby passing a point in time  T = 2½ minutes  earlier than he would have.  How long,  L,  is the train?  Though this seems like a problem of this section, in fact it simply says the train takes  2½ minutes  to pass a given point and  L = VT.

Perelman.  1937.  MCBF.  Reconnaissance at sea, prob. 149 a & b, pp. 253-255.  Squadron moving and a reconnaissance ship is sent ahead.  Part  a  gives distance to go ahead and asks how long it will take.  Part  b  gives time for reconnaissance and asks when the reconnaissance ship will turn back.

McKay.  At Home Tonight.  1940.  Prob. 17: The orderly's ride, pp. 65 & 79.  Same as first part of Loyd with  1  mile army.

William R. Ransom.  Op. cit. in 6.M.  1955.  An army courier, p. 103.  Same as first part of Loyd with  25  mile army, but says it takes a day and asks how fast the courier rides.

G. J. S. Ross  &  M. Westwood.  Problems drive, 1960.  Eureka 23 (Oct 1960) 23-25 & 26.  Prob. G.  v = 15,  d + L = 23,  V = 60,  find  L.  At first I thought the result was wrong, but the phrasing is a bit different than usual as the back of the army is  23  from where the front finishes.

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. m, pp. 189 & 191-192.  Courier going from back to front and then back again at three times the speed of the army.  Where is he when he gets to the back again?

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 76, pp. 78 & 107.  L = 100,  V = 3v,  find  d.

Julius Sumner Miller.  Millergrams.  Ure Smith, Sydney, 1966.  Prob. 25, pp. 25 & 70.  Army is 3 mi long, officer starts at back, goes to front and returns, reaching the back when it has advanced 4 miles.  How far did he go?  L = 3,  d = 4,  giving  D = 8 miles.  He says: "you can get the right answer by erroneous logic", but he doesn't explain how to get the answer!

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Procession, pp. 43, 167 & 189.  = Birtwistle; Calculator Puzzle Book; 1978; prob. 55, pp. 38-39 & 99.  Procession    miles long going at  2  mph.  Marshall starts at head, walks to the back and then forward, reaching his starting point when half the army has passed.  What is his speed?  He continues on the head and returns to the same point.  Where is the end of the procession when he gets back?

 

          10.A.2.        NUMBER OF BUSES MET

 

          New section -- I haven't done much on this yet.  Kraitchik has several examples.

 

Mittenzwey.  1880.  Prob. 75, pp. 14-15 & 66;  1895?: 82, pp. 19 & 69;  1917: 82, pp. 18 & 65.  Daily trains crossing America in  7  days.

[Richard A. Proctor].  Letters received and short answers.  Knowledge 3 (26 Oct 1883) 264.  Answer to Harry.  Recalls it being posed on by the captain's wife, Mrs Cargill, on the S. S. Australasia, but no date is clear.  Trains going between New York and San Francisco taking  7  days.  How many does one meet on such a trip?  Says he gave the wrong answer!

[Richard A. Proctor].  Editorial gossip.  Knowledge 3 (23 Nov 1883) 318.  Gives a careful answer to the problem stated above.

L. Carroll.  A Tangled Tale.  (1885) = Dover, 1958. 

Knot III, pp. 13‑18, 90‑95.  Circular railway with trains going at different frequencies in the two directions.  How many are met on a round trip in each direction?

Knot V, pp. 27‑28.  Comment on Knot III problem.

Knot VIII, pp. 52‑53, 55‑57, 132‑134.  Buses in both directions pass depot every  15  minutes.  Walker starts from depot at same time as a bus and meets a bus in  12½  minutes.  When is he overtaken by a bus?

Laisant.  Op. cit. in 6.P.1.  1906.  Chap. 49: Du Havre à New-York, pp. 123-125.  Definitely asserts that this problem was posed by Lucas at a meeting 'longtemps déjà'.  Boats leave every noon each way between le Havre and New York and take exactly seven days to make the trip -- how many do they pass?

Pearson.  1907.  Part II, no. 106, pp. 136 & 212‑213.  Tube trains run every  2  minutes.  How many are met in a  30  min journey?  Answer:  30.

Peano.  Giochi.  1924.  Prob. 11, p. 4.  Buses take  7  minutes from the centre to the terminus.  Buses leave from the centre and from the terminus every minute.  How many buses are met in going one way?  Observes that it takes  14  buses to run such a service and so one bus meets the other  13.

McKay.  At Home Tonight.  1940.  Prob. 14: Passing the trains, pp. 65 & 79.  12  hour journey and trains start from the other end every hour.

Sullivan.  Unusual.  1943.  Prob. 24: Back in the days of gasoline, tires, and motorists.    hours from Chicago to Indianapolis, buses leaving every hour [on the hour].

Anonymous.  The problems drive, 1956.  Eureka 19 (Mar 1957) 12-14 & 19.  No. 10.  Circular line with trains departing each way every 15 minutes, but the east bound trains take 2 hours for the circuit while the west bound ones take 3 hours for a circuit.  Starting after one leaves the station, how many trains does one see?

Doubleday - 1.  1969.  Prob. 36: Traveler's tale, pp. 48 & 162.  = Doubleday - 5, pp. 55-56.  Trains going between Moscow and Paris, taking seven days.

[Henry] Joseph & Lenore Scott.  Master Mind Brain Teasers.  1973.  Op. cit. in 5.E.  How many buses, pp. 7-8.

 

          10.A.3.        TIMES FROM MEETING TO FINISH GIVEN

 

Simpson.  Algebra.  1745.  Section XI (misprinted IX in 1790), prob. XLVI, pp. 110-111 (1790: prob. LIX, pp. 111‑112).  Travellers set out from each of two cities toward the other, at the same time.  After meeting, they take  4  and  9  hours to finish their journeys.  How long did they take?  He gives a general solution -- if  x  is the time before meeting and  a, b  are the times from meeting to finishing, then  x2 = ab.  [I have seen a 20C version where only the ratio of velocities is asked for -- indeed I used it in one of my puzzle columns, before I knew that the times could be found.]

Dodson.  Math. Repository.  1775.

P. 67, Quest. CXXV.  Travellers set out from London and York at the same time.  When they meet, they observe that  A  had travelled  30  miles more than  B  and that  A  expected to reach York in  4  days and  B  expected to reach London in  9  days.  What is the distance between London and York?

P. 68, Quest. CXXVII.  Travellers set out from London and Lincoln at the same time.  When they meet, they observe that  A  had travelled  20  miles more than  B  and that  A  travelled in  6⅔  days as much as  B  had gone and  B  expected to get to London in  15  days.  What is the distance from London to Lincoln?

Ozanam‑Montucla.  1778.  Supplement, prob. 42, 1778: 432;  1803: 425;  1814: 360;  1840: 186.  Couriers set out toward one another from cities 60 apart.  After meeting, they take  4  and  6  hours to reach their destinations.  What are their velocities?

Thomas Grainger Hall.  The Elements of Algebra: Chiefly Intended for Schools, and the Junior Classes in Colleges.  Second Edition: Altered and Enlarged.  John W. Parker, London, 1846.  P. 136, ex. 38.  Simpson's problem with cities of London and York and times of 9 and 16.

T. Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.  P. 89, no. 4.  Same as Simpson.

Todhunter.  Algebra, 5th ed.  1870.  Examples XXIV, nos. 20 & 22, pp. 211-212 & 586.

No. 20.  Simpson's problem between London and York, with times  16  &  36.

No. 22.  =  Dodson, p. 67.

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.  No. 132, pp. 306 & 347.  Travellers between London and York reach their destinations  25  &  36  hours after meeting.  How long did each take?

W. W. Rouse Ball.  Elementary Algebra.  CUP, 1890 [the 2nd ed. of 1897 is apparently identical except for minor changes at the end of the Preface].  Ex. 5, p. 231.  A  and  B  are  168 miles  apart.  Trains leave each end for the other starting at the same time.  They meet after  1 hour 52 minutes.  The train from  A  reaches  B  half an hour before the other reaches  A.

Haldeman-Julius.  1937.  No. 133: Racers' problem, pp. 15 & 28.  Racers on a circular track start in opposite directions.  After meeting, they take  4  and  9  minutes to pass the starting point, but they continue.  When do they meet each other at the starting point?

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. b, pp. 188-190.  Trains start toward each other at  7  am;  one takes  8  hours, the other takes  12.  When do they meet?

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 37: Walking all the way, pp. 29 & 106-107.  Similar to Simpson, but they start at given times and the time of meeting is given and they get to their destination at the same time.  This is thus the same as Simpson if it is considered with time reversed.  The elapsed times to meeting are  25/12  and  49/12  hours.

Birtwistle.  Calculator Puzzle Book.  1978. 

Prob. 60: Meeting point, pp. 42 & 102.  Travellers start at the same time from opposite ends of a journey.  When they meet, they find that the first has  11 1/5  hours to go, while the other has  17 1/2.  Also the first has travelled  7  miles further than the second.  How long is the journey?

Prob. 71: Scheduled flight, pp. 50-51 & 108.  Planes  A  and  B  start toward each other at  550  mph.  Five minutes later, plane  C  starts from the same place as  A  at  600  mph.  It overtakes  A  and then meets  B  36  minutes later.  When does  A  land?

 

          10.A.4.        THE EARLY COMMUTER

 

            A man is usually met by a car at his local train station, but he arrives  A  minutes early and begins walking home.  The car meets him and picks him up and they arrive home  B  minutes early.  How long was he walking?  The car's trip is  B/2  minutes shorter each way, so the commuter is met  B/2  minutes before his usual time and he has been walking  A - B/2  minutes.  New section -- there must be older examples.

 

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 155, pp. 61 & 151: Catching the postman.  Man driving 10 mph usually overtakes the postman walking 4 mph at the same point every morning.  One morning, the man is four minutes late and he overtakes the postman half a mile beyond the usual point.  Was the postman early or late, and by how much?  (Note that the man is four minutes late starting, not in overtaking.)

Meyer.  Big Fun Book.  1940.  No. 10, pp. 162-163 & 752.  A = 60,  B = 10.

Harold Hart.  The World's Best Puzzles.  Op. cit. in 7.AS.  1943.  The problem of the commuter, pp. 8 & 50.  A = 60,  B = 20.

The Little Puzzle Book.  Op. cit. in 5.D.5.  1955.  Pp. 59-60: An easy problem.  A = 60,  B = 20,  he walks at  4  mph.  How fast does the chauffeur drive?

Nathan Altshiller Court.  Mathematics in Fun and in Earnest.  Op. cit. in 5.B.  1961.  Prob. f, pp. 189 & 191.  A = 60,  B = 16.  The time of arrival is also given, but is not needed.

Liz Allen.  Brain Sharpeners.  Op. cit. in 5.B.  1991.  The Commuter's tale, pp. 87-88 & 136.  A = 60,  B = 10.

 

          10.A.5.        HEAD START PROBLEMS

 

            Doubleday - 2 gives a typical example.  New section -- I have seen other examples but didn't record them.

            Gardner, in an article:  My ten favorite brainteasers  in Games (collected in Games Big Book of Games, 1984, pp. 130-131) says this is one of his favorite problems.  ??locate

 

Todhunter.  Algebra, 5th ed.  1870.  Examples XIII, no. 23, pp. 103 & 578.  In racing a mile (= 1760 yds),  A  gives  B  a headstart of  44 yd  and wins by  51 sec;  but if  A  gives  B  a headstart of  75 sec,  A  loses by  88 yd.  Find the times each can run a mile.

Charles Pendlebury.  Arithmetic.  Bell, London, (1886), 6th ed, 1893.  Section XXXV (b): Races and games of skill, pp. 267-268, examples XXXV (b) & answers, p. viii.  Does an example:  A  can give  B  10  yards in  100  yards and  A  can give  C  15  yards in  100  yards.  How much should  B  give  C  in  150  yards?  He gives  11  similar problems.  On pp. 363-364 & answers, part II, p. xix, he gives some further problems, 45-48.

W. W. Rouse Ball.  Elementary Algebra.  CUP, 1890 [the 2nd ed. of 1897 is apparently identical except for minor changes at the end of the Preface].

Ex. 8, p. 135.  "A  can row a mile in  ¾ of a minute less time than  B.  In a mile race,  B  gets  250 yards head start, and lose by  14 yards."  Determine their times to row a mile, assuming uniform speed.

Prob. 37, pp. 161 & 467,  If  B  gets a  12 sec headstart in a mile race, he loses by  44 yards.  If he gets a  165 yards headstart, he wins by  10 sec.  Determine their times for a mile.

Lewis Carroll.  Letter of 8 Apr 1897 to Enid Stevens.  = Carroll-Wakeling, prob. 25: Handicaps, pp. 33 & 71.  A  loses 10 yards in every 100 against  B,  while  B  gains 10 yards in every 100 against  C.  What handicaps should  B  give  A  and  C  in a quarter mile (= 440 yards) race?  Mentioned in Carroll-Gardner, pp. 50-51  Cohen's comment is that is that this is too ambiguous to have a precise answer, but such problems were common in late Victorian times and Wakeling gives an answer (but no method) which seems correct to me and in line with other problems of the time.  However, Gardner does not comment on Cohen's remark and does not give an answer.  Letting  A, B, C  denote the uniform speeds of the runners, Wakeling and I would interpret the problem statements as:  A = 90/100 · B  and  B = 110/100 · C.  The ambiguity is that one could read the statements as saying  A = 100/110 · B  and  B = 100/90 · C,  but I think the first interpretation is more natural.  Either case permits a clear answer.  Cohen says it is clear that  A = 90/100 · B,  but is unsure whether  B = 100/90 · C  or  B = 110/100 · C,  but the first case makes  C = A,  which makes the problem much less interesting.

Charles Pendlebury & W. S. Beard.  A "Shilling" Arithmetic.  Bell, London, 1899 -- my copy is 59th ptg, 1960 and says it was reset for the 49th ptg of 1944.  Examination Papers XIV, prob. 11, p. 187.  A runs at  12 1/2 mph  and B runs at  12 7/16 mph.  If A gives B ten yards headstart, when will he overtake him.  Who is ahead when A has run a mile?

Hummerston.  Fun, Mirth & Mystery.  1924.  Speed, Puzzle no. 18, pp. 49 & 174.  Cook can run  100  yards in  12  sec.  Brown can give Cook a  10  yd headstart and finish even with him.  Adams can give Brown a  5  yd headstart and finish even.  Adams gives Cook a  15  yd headstart and Cook loses  1/20  of a sec in a bad start.  Who wins and by how much?  He notes that one must assume uniform speeds.

Collins.  Fun with Figures.  1928.  The 100-yard dash, pp. 24-26.  = Hummerston with  Adams, Brown, Cook  replaced by  Bob, Jack, Bill.

Doubleday - 2.  1971.  Running commentary, pp. 59.  In a  100  yard dash,  A  gives  B  a head start of  25  to make an even race and  B  gives  C  a head start of  20.  How much head start should  A  give  C  to make an even race?

 

          10.A.6.        DOUBLE CROSSING PROBLEMS

 

            New section.  I recall Loyd and/or Dudeney have versions.  Also, the problems in 10.A with the faster meeting the slower on the return trip are related.

 

Mittenzwey.  1880.  Prob. 296, pp. 54 & 105;  1895?: 326, pp. 57 & 106;  1917: 326, pp. 52 & 100.  Field in the shape of a right triangle.  Runners whose speeds are in the ratio  13 : 11  start at the right angled corner along different legs.  They first meet at the midpoint of the hypotenuse and then at  60m  from the starting point.  As far as I can see, this leads to a quadratic.

Wood.  Oddities.  1927.  Prob. 61: The errand boys, p. 47.  Alan goes from  A  to  B  and back while Bob goes from  B  to  A  and back, both starting at the same time.  They first cross at  720  from  A  and then at  400  from  B.  How far is it between  A  and  B?  Gives a general formula:  If the two crossings occur at distances  a, b,  then the width is  3a - b.  He asserts that Alan goes faster, but the ratio of velocities is  a/(2a-b)  which is  9/13  in this case.

Philip Kaplan.  More Posers.  (Harper & Row, 1964);  Macfadden-Bartell Books, 1965.  Prob. 73, pp. 76 & 106.  Two swimmers in a pool, like above with distances  40,  20.

 

          10.A.7.        TRAINS PASSING

 

            New section.  If a fast train, going at velocity  V,  takes time  T  to overtake a slow train, going at velocity  v,  and it takes time  t  for the trains to pass when meeting, then the lengths of the trains cancel out and one gets  V/v  =  (T + t)/(T - t).

 

Haldeman-Julius.  1937.  No. 28: Speed problem, pp. 6 & 22.  T = 35,  t = 3,  V = 38.

 

          10.A.8.        TOO SLOW, TOO FAST

 

            New section.  If I go at rate  v1,  I arrive  t1  too late; if I go at rate  v2,  I arrive  t2  too soon.  I do not recall seeing anything like this before.

 

Charles Pendlebury & W. S. Beard.  A "Shilling" Arithmetic.  Bell, London, 1899 -- my copy is 59th ptg, 1960 and says it was reset for the 49th ptg of 1944.  Examination Papers XI, prob. 185, p. 185.  If a man goes at 4 mph, he arrives 5 min late; if he goes 5 mph, he arrives 10 min early.  How far is he going?

The general solution is easily found to be  D = v1v2(t1+t2)/(v2-v1).  However, I wondered what was the correct rate,  v,  so that I arrive on time.  This leads to the following.

                    (t1+t2)/v  =  t1/v2 + t2/v1   or   1/v  =  t1/(t1+t2)·1/v2  +  t2/(t1+t2)·1/v1,  so  v  is a weighted harmonic mean of  v1  and  v2.  On the other hand, we find the time to arrive is  t = D/v  =  v1/(v2-v1)·t1  +  v2/(v2-v1)·t2,  which is a weighted mean of  t1  and  t2. 

                    The only type of problem that I can recall that leads to similar means is the classic problem:  If I make a journey at 30 mph and return at 20 mph, what is my average speed.  In general, if I travel  di  at speed  vi,  what is my average speed?  Letting  ti  be the time at speed  vi,  we have  ti = di/vi,  so the average speed is given by 

                    (d1+d2)/v  =  t1 + t2  =  d1/v1 + d2/v2   or  

                    1/v  =  d1/(d1+d2)·1/v1  +  d2/(d1+d2)·1/v2,  so  v  is again a weighted harmonic mean of  v1  and  v2. 

 

          10.B. FLY BETWEEN TRAINS

 

            There are two trains  d  apart, approaching at rates  a, b.  A fly starts at one, flies to the other, then back to the first, then back to the second, etc., flying at rate  c.  How far does he go?

            NOTATION.  We denote this problem by  (a, b, c, d).  For overtaking problems, we let  b  be negative.  I need to check for details.

 

Laisant.  Op. cit. in 6.P.1.  1906.  Chap. 53: Le chien et les deux voyagers, pp. 132-133.  A  is  8  km ahead and goes at  4  km/hr.  B  starts after him at  6  km/hr.  A dog starts at one and runs back and forth between them at  15  km/hr until they meet.  I.e. (4, ‑6, 15, 8).  Notes that the distance the dog travels is independent of where he starts and that the travellers could be meeting rather than overtaking.

Dudeney.  Problem 464: Man and dog.  Strand Mag. (Jul 1919).  ??NX.

Dudeney.  Problem 643: Baxter's dog.  Strand Mag.  (1924?).  ??NX.  A  goes at rate  2  and has an hour's head start.  B  goes at rate  4  and dog starts with him at rate  10.  I.e. (4, ‑2, 10, 2).

G. H. Hardy.  Letter of 5 Jan 1924 to M. Riesz.  In:  M. L. Cartwright.  Manuscripts of Hardy, Littlewood, Marcel Riesz and Titchmarsh.  Bull. London Math. Soc. 14 (1982) 472‑532.  (Letter is on p. 502, where it is identified as Add. MS. a. 275 33, presumably at Trinity College.)  Says it defeated Einstein, Jeans, J. J. Thomson, etc.  Fly between cyclists  (10, 10, 15, 20).  "One thing only is necessary: you must not know the formula for the sum of a geometrical progression.  If you do, you will take  15-20  minutes: if not,  2  seconds."

Ackermann.  1925.  Pp. 116‑117.  Couple walking up a hill.  Their dog, who is twice as fast as they are, runs to the top and back to them continually.  (a, 0, 2a, d) -- he only says the couple start  1/4  of the way up the hill.

Dudeney.  Problem 754: The fly and the motor‑cars.  Strand Mag. (Jun 1925)  ??NX.  (?= PCP 72.)

H&S 53, 1927, says this is 'a modern problem'.

M. Kraitchik.  La Mathématique des Jeux, 1930, op. cit. in 4.A.2, chap. 2, prob. 17, p. 30.  (15, 25, 100, 120).  (Identified as from L'Echiquier, 1929, 20, ??NYS.)  I can't find it in his Mathematical Recreations.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 181, pp. 71 & 156: Little Red Riding Hood.  Little Red Riding Hood with her dog and Grandma set out at the same time to meet: (2, 2, 8, 6).

Dudeney.  PCP.  1932.  Prob. 72: The fly and the motor‑cars, pp. 28 & 86.  = 536; prob. 86: The fly and the cars, pp. 26 & 243.  (50, 100, 150, 300).

Phillips.  Week‑End.  1932.  Time tests of intelligence, no. 38, pp. 21 & 193.  Fly between cyclists.  (10, 15, 20, 60).

Abraham.  1933. 

Prob. 11 -- The fly and the cyclists -- A problem in convergent series?, pp. 6 & 23 (4 & 111).  (10, 10, 15, 20).

Prob. 63 -- The escaping prisoner, pp. 21 & 28 (16 & 115).  Warders going  4  after an escapee going  3  with  3/2  headstart.  A dog runs back and forth at  12.  I.e.  (4, ‑3, 12, 3/2).

Streeter & Hoehn.  Op. cit. in 7.AE.  Vol. 2, 1933, p. 29: Aeroplane dilemma.  Destroyer going  25  overtaking battleship going  20  with headstart of  30.  Plane flies back and forth at  90.  I.e.  (25, ‑20, 90, 30).

Phillips.  Brush.  1936.  Prob. G.21: The busy fly, pp. 20 & 87.  Same as in Week‑End.

J. R. Evans.  The Junior Week‑End Book.  Op. cit. in 6.AF.  1939.  Prob. 37, pp. 265 & 270.  Fly between cyclists,  (10, 10, 15, 20).

Haldeman-Julius.  1937. 

No. 73: The walking dog problem, pp. 9-10 & 24.  Man walking home with dog running between home and man.  (2, 0, 4, 2).

No. 149: Suicide of a bird, pp. 16-17 & 28.  Bird between two trains,  a = 52, b = 30, c = 60,  but  d  is not given.  Instead, he says the trains collide after half an hour, thereby making the problem pretty trivial and making the values of  a  and  b  unnecessary.

McKay.  At Home Tonight.  1940.  Prob. 20: The fluttered pigeon, pp. 66 & 80-81.  Pigeon between walkers --  (3, 3, 21, 30).  Gives solution by adding a GP and the easy solution.

Sullivan.  Unusual.  1943.  Prob. 4: A busy bee.  Bee between motorists -- (10, 10, 15, 20).

L. Lange.  Another encounter with geometric series.  SSM 55 (1955) 472‑476.  Studies the series involved.

William R. Ransom.  Op. cit. in 6.M.  1955.  The bicycles and the fly, pp. 22‑23.  Studies the series.

Eugene Wigner.  In the film:  John von Neumann, MAA, 1966, he relates this as being posed by Max Born to von Neumann, involving a swallow between bicyclists.  He says it was a popular problem in the 1920s.

Paul R. Halmos.  The legend of John von Neumann.  AMM 80 (1973) 382-394.  Gives the fly between two cyclists puzzle and story on pp. 386-387.

David Singmaster.  The squashed fly -- (60, 40, 50, 100)  with fly starting on first train.  Used in several of my series.

On training a fly to fly right.  Los Angeles Times (21 Dec 1987) Section CC Part II.

A very fly braintwister.  Special Holiday Edition of The Daily Telegraph:  The Great British Summer (Aug 1988) 15 & 10.

Flight of fancy.  Focus, No. 2 (Jan 1993) 63 & 98.

The squashed fly.  Games & Puzzles, No. 11 (Feb 1995) 19  &  No. 12 (Mar 1995) 41.

Herbert R. Bailey.  The girl and the fly: A von Neumann Legend.  MS 24 (1991/92) 108-109.  Cites Halmos, but gives a version with a girl walking toward a wall.  Finds the relevant series.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995. 

Chap. 2, prob. 10: The crazy dog (or the problem that didn't fool John Von Neumann), pp. 16 & 108.  Dog between cyclists; usual solution and some comments.  Refers to MTr (Nov 1991) -- ??NYS.

Chap. 11, prob. 7: The problem that didn't fool Von Neumann, pp. 95 & 183.  Does it by summing one series.

 

          10.C. LEWIS CARROLL'S MONKEY PROBLEM

 

            A monkey and a barrel of equal weight are on the ends of a rope over a pulley.  The monkey starts to climb the rope -- what happens?

 

Stuart Dodgson Collingwood.  The Life and Letters of Lewis Carroll.  T. Fisher Unwin, London, 1898.  Pp. 317-318.  Referring to Dec 1893, he says "Mr. Dodgson invented a new problem to puzzle his mathematical friends with, which was called "The Monkey and Weight Problem."  A rope is supposed to be hung over a wheel fixed to the roof of a building; at one end of the rope a weight is fixed, which exactly counterbalances a monkey which is hanging on to the other end.  Suppose that the monkey begins to climb the rope, what will be the result?" 

Carroll.  Letter of 19 Dec 1893 to Price.  Discussed and reproduced for the first time in Edward Wakeling: Lewis Carroll and the Bat; ABMR (Antiquarian Book Monthly Review) 9:2 (No. 99) (Jul 1982) 252-259.  Wakeling has sent me a copy of this.  Carroll starts: "Many thanks for your solution of the "Monkey & Weight" Problem -- It is the reverse of the solution given me by Sampson", and discusses consequences of Sampson's argument.  As a postscript, Carroll states: "I own to an inclination to believe that the weight neither rises nor falls!"

Carroll.  Diary entry for 21 Dec 1893.  In Roger Lancelyn Green's edition, p. 505.  Quoted in Collingwood, above; in Carroll-Wakeling II, below; and slightly differently in Wakeling's 1982 article, above.  "Got Prof. Clifton's answer to the "Monkey and Weight Problem."  It is very curious, the different views taken by good mathematicians.  Price says the weight goes up, with increasing velocity.  Clifton (and Harcourt) that it goes up, at the same rate as the monkey; while Sampson says it goes down."

Carroll-Wakeling II, prob. 9: The monkey and weight problem, pp. 15-16 & 66.  Quotes the problem and the Diary entry.  Identifies the people, who were all Oxford academics:  Robert Bellamy Clifton, Professor of Experimental Philosophy;  Bartholomew "Bat" Prices -- cf Carroll‑Wakeling in Common References;  Augustus Vernon Harcourt, Lee's Reader in Chemistry, at Christ Church;  Rev. Edward Francis Sampson, assistant mathematical tutor to Carroll at Christ Church.  Wakeling notes that most modern mathematicians and scientists agree with Clifton and Harcourt.

Carroll.  Letter of 23 Dec 1893 to Mrs. Price.  Discussed and partially reproduced in Wakeling's 1982 article, above.  In a PS, he asks her to remind Price to return Sampson's proof, summarises the various solutions received and concludes the paragraph with  "And my present inclination is to believe that it goes neither up nor down!!!

Carroll-Collingwood.  1899.  Pp. 267‑269 (Collins: 193-194).  Repeats material from The Life and Letters.  Then includes a letter from Arthur Brook, arguing that 'the weight remains stationary'.  Collingwood discusses it and sides with Sampson.

Pearson.  1907.  Part II, no. 11: A climbing monkey, pp. 9 & 188.  Asserts that the weight goes up, but the monkey does not!

Dudeney.  Some much‑discussed puzzles.  Op. cit. in 2.  1908.  Cites the diary entry from Collingwood's Life and Brook's comment in the Picture Book.  Says mechanical devices have been built.

Loyd.  Lewis Carroll's monkey puzzle.  Cyclopedia, 1914, pp. 44, 344‑345 (erroneous solution).  (= MPSL2, prob. 1, pp. 1‑2 & 121.)

William F. Rigge.  The climbing monkey.  SSM 17 (1917) 821.  (Refers to L'Astronomie (Jul 1917) ??NYS.)  Asserts that the weight remains fixed and that he made a clockwork climber and demonstrated this.

Wilbert A. Stevens.  The monkey climbs again.  SSM 19 (1919) 815.  Assert's Rigge's climber was too light to overcome friction and that both should ascend together.

William F. Rigge.  The monkey stops climbing.  SSM 20 (1920) 172‑173.  Says he increased the speed of his monkey and that now it and the weight go up together.

Editors, proposers;  E. V. Huntington & L. M. Hoskins, solvers.  Problem 2838.  AMM 27 (1920) 273‑274  &  28 (1921) 399‑402.  Proposal quotes Carroll and cites Collingwood & SSM.

Ackermann.  1925.  Pp. 1‑2.  Says they would go up together, but the rope moves to the side of the monkey and so the weight will rise faster.  Cites Carroll.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927. 

P. 83 & Answers p. 6: Balanced swings.  Two equally weighted swings and children connected by ropes over pulleys.  One child starts swinging.  What happens to the other?  Answer says the centrifugal force would pull the other child up.

P. 104 & Answers p. 12: The climbing monkey.  [Unsigned item -- possibly by Loyd Jr??]  Refers to L'Astronomie (Jul 1917), ??NYS.  Gives Carroll's problem.  Suggests making the rope a loop so movement of the rope doesn't cause an imbalance.  Answer says they go up together, but if the inertia of the rope is considered then the monkey will rise faster than the weight.  If the monkey lets go and recatches the rope, then things are again symmetric until one considers the inertia of the rope which will cause the barrel to fall less rapidly and hence the process will cause the barrel to rise.

Jerome S. Meyer.  Fun-to-do.  Op. cit. in 5.C. 1948.  Prob. 6: How good are you in physics?, part 2, pp. 19-20 & 181-182.  Carroll's problem, with no reference to Carroll.  Answer says the weight will go up.

Warren Weaver.  Lewis Carroll: Mathematician.  Op. cit. in 1.  1956.  Mentions the problem.  A. G. Samuelson's letter says a modern version has a mirror at the other end and asks if the monkey can get away from his image.  He says the monkey and the mirror will behave identically so he cannot get away.  Weaver's response is that this is correct, though he was saying that the behaviour of the weight cannot be known unless you know how the monkey climbs.

H. T. Croft  &  S. Simons.  Some little naggers.  Eureka 23 (Oct 1960) 22 & 36.  No. 5.  Answer says they go up together.

Carroll-Gardner.  1996.  Pp. 23-24.  Quotes the problem and gives the accepted solution.  Says there is a demonstration at the Chicago Museum of Science and Industry.

 

          10.D. MIRROR PROBLEMS

 

          10.D.1         MIRROR REVERSAL PARADOX

 

            Why does a mirror reverse right and left, but not up and down?  This is a perennial problem in Notes and Queries type columns.

 

Alice Raikes.  Letter.  The Times (22 Jan 1932).  ??NYS -- cited in:  Michael Barsley; The Left Handed Book; (Souvenir, London, 1966);  Pan, 1989, pp. 199‑200.  Quoted in:  Roger Lancelyn Green; Alice -- an excerpt from his:  Lewis Carroll, Bodley Head, 1960;  IN:  Robert Phillips, ed.; Aspects of Alice; (1971;  Gollancz, London, 1972);  Penguin, 1974, pp. 52-53.  Quoted in:  Florence Becker Lennon; Escape through the looking-glass -- an excerpt from her:  Victoria Through the Looking-Glass, 1971;  IN:  Robert Phillips, ibid., pp. 108-109.

                    Miss Raikes was one of Carroll's girl friends.  She relates that Carroll put an orange in her right hand and then asked her to stand in front of a mirror and say which hand the reflection had the orange in.  She said the left hand and Carroll asked her to explain.  She finally said  "If I was on the other side of the glass, wouldn't the orange still be in my right hand?"  Carroll said this was the best answer he had had and later said it was the idea for "Through the Looking Glass".  Green dates this as Aug 1868 and says it took place when Carroll was visiting his uncle Skeffington Lutwidge at his house in Onslow Square, London.

Dudeney.  PCP.  1932.  Prob. 327: Two paradoxes, pp. 112‑113 & 526.   = 536, prob. 526, pp. 216‑217 & 412.

Gardner.  Left and right handedness.  SA (Mar 1958) = 1st Book, chap. 16.

Eric Laithwaite.  Why Does a Glow-worm Glow?  Beaver (Hamlyn), London, 1977.  Pp. 54‑56: When you look into a mirror, your left hand becomes your right, so why doesn't your head become your feet?  He gives the correct basic explanation but then introduces an interesting complication.  If you hold a rotating wheel with its axis parallel to the mirror, the image appears to be rotating in the opposite direction to the real wheel.  Now turn the wheel so the axis is perpendicular to the mirror and the image now appears to be rotating in the same direction as the real wheel!!  [What do you see when you're halfway in the turning process??]

Gardner.  The Ambidextrous Universe.  2nd ed., Pelican (Penguin), 1982.  Pp. 6‑9 & 22‑26 surveys the question and cites three 1970s serious(!) philosophical articles on the question.

Richard L. Gregory.  Mirror reversal.  IN: R. L. Gregory, ed.; The Oxford Companion to the Mind; OUP, 1987, pp. 491-493.  ??NX.  Gives the basic explanation, but seems unhappy with the perceptual aspects.  I would describe it as making heavy weather of a simple problem.

Don Glass, ed.  Why You Can Never Get to the End of the Rainbow and Other Moments of Science.  Indiana Univ Press, Bloomington, Indiana, 1993.  (Adapted from scripts for the radio series A Moment of Science, WFIU, Bloomington, Indiana.)  A mirror riddle, pp. 32-33.  States that a mirror usually reverses right and left, but a reflecting lake (or a flat mirror) reverses up and down.  Says that the explanation is that a mirror reverses clockwise and anticlockwise, but this is basically inadequate.

In c1993, there was correspondence in the Answers column of The Sunday Times.  Collected in: The Sunday Times Book of Answers; ed. by Christopher Lloyd, Times Books, London, 1993.  (The column stared in Jan 1993, but 70% of the book material did not appear in the paper.)  Pp. 63-67.  One correspondent said he asked the question in New Scientist almost 20 years ago.  One correspondent clearly states it reverses front and back, not right and left.  Another clearly notes that the appearance of reversing right and left is due to the bilateral symmetry of our bodies, so we turn around to consider the mirror image.

In c1994, there were several letters on the problem in The Guardian's Notes & Queries.  These are reproduced in: Joseph Harker, ed.; Notes & Queries, Vol. 5., Fourth Estate, London, 1994, pp. 178-180  and in: Joseph Harker, ed.; The Weirdest Ever Notes & Queries; Fourth Estate, London, 1997, pp. 140-142.  R. Thomson clearly sees that right and left are not reversed, but front and back are.  Richard L. Gregory has a slightly confusing letter but adds that the problem goes back to Plato and that he has given the history and solution in his Odd Perceptions (Routledge, 1986, ??NYS) and in the Oxford Companion to the Mind -- see above.

Erwin Brecher.  Surprising Science Puzzles.  Sterling, NY, 1995.  The mirror phenomenon, pp. 16 & 80.  "Why does a mirror reverse only the left and right sides but not up and down?"  Gives a nonsensical answer: "Left and right are directional concepts while top and bottom, or up and down, are positional concepts." and then follows with an unreasonable analogy to walking over the North Pole.

Seckel, 2002, op. cit. in 6.AJ, fig. 17, pp. 26 & 44.  Straightforward discussion of the paradox.

 

          10.D.2.        OTHER MIRROR PROBLEMS

 

Richard A. Proctor.  Our puzzles.  Knowledge 10  (May 1887) 153  &  (Jun 1887) 186-187.  Prob. XXVIII: how to see yourself properly  --  use two mirrors at right angles.  Prob. XXIX: in a fully mirrored room, what do you see when you look into a corner?  --  yourself inverted.  Prob. XXX: in the same room, how many images of yourself do you see?  --  26.

 

          10.D.3         MAGIC MIRRORS

 

            These are oriental (Chinese or perhaps Japanese) polished discs which cast reflections containing a pattern.  I first came across them in one of R. Austin Freeman's detective stories and I was kindly brought one from China a few years ago.  Since about 1991, they have been made in and exported from China and are commercially available.  A fine example, with explanation, is in the Museum of the History of Science in Oxford, but it is not illuminated.  Basically, the pattern is hammered on the disc and this causes molecular changes which remain even when the surface has been made apparently smooth.  Apparently other methods of producing a difference in metallic structure have been used.  Sometimes the pattern is also made in relief on the opposite side of the disc, and sometimes a different pattern is made.

            Peter Rasmussen and Wei Zhang have sent a bundle of material on this, ??NYR.

 

Aignan Stanislas Julien.  Notice sur les miroirs magiques des Chinois et leur fabrication.  CR, 1847?  Separately printed or extracted from the journal, Bachelier, Paris, 1847.  15pp.  ??NYS -- seen in a dealer's catalogue.

John Timbs.  Things Not Generally Known,  Familiarly Explained.  A Book for Old and Young (spine says  First Series  and a note by a bookdealer on the flyleaf says  2 vol.).  Kent & Co., London, (1857?), 8th ed., 1859.  Chinese magic mirrors, p. 114.  Says the reflected pattern is in relief on the other side of the disc.  He quotes an explanation given by 'Ou-tseu-hing' who lived between 1260 and 1341 and who worked out the process by inspecting a broken mirror.  He says that the disc with the relief pattern is made by casting, in fine copper.  The pattern is then copied by engraving deeply on the smooth side and the removed parts are filled with a rough copper, then the disc is fired, polished and tinned.  The rough copper produces dark areas in the reflection.

R. F. Hutchinson.  The Japanese magic mirror.  Knowledge 10 (Jun 1887) 186.  Says he has managed to fulfil a boyhood longing and obtain one.  Describes the behaviour and asks for an explanation.

J. Parnell.  The Japanese magic mirror.  Knowledge 10 (Jul 1887) 207.  Says he studied it some 20 years earlier and cites:  The Reader (1866);  Nature (Jul 1877)  and  a paper read to the Royal Society by Ayrton & Perry in Dec 1878  -- all ??NYS.  Says the mirror is somewhat convex and the picture lines are slightly flatter, so the reflection of the picture is brighter than of the surrounding area.

The Japanese mirror.  Cassell's Magazine (Dec 1904) 159-160.  Short note in the Flotsam and Jetsam section.  Says they are bronze covered with a mercury amalgam with a raised pattern on the back, whose image is reflected by the front.  Says it "is due to inequalities of convexity on the face, caused by the pattern on the back dispersing the sunlight more or less."

R. Austin Freeman.  The Surprising Adventures of Mr. Shuttlebury Cobb.  Story VI: The magic mirror.  The series first appeared in  Pearson's Magazine:  (1 Jun 1913) 438-448,  (15 Jun 1913) 565-574,  (1 Jul 1913) 748-757   and  Red Magazine:  (15 Jul 1913) ??,  (1 Aug 1913) ??,  (15 Aug 1913).  Collected as a book:  Hodder & Stoughton, London, 1927,  with Story VI on pp. 231-281.  "... product of Old Japan, ...."  Says the device on the back is cast as a dark shape with a bright halo.  Says the design is formed by chasing or hammering the lines, which makes them harder than the surrounding bronze (or similar metal), so when polished, they project slightly and produce an image in the reflection.  Says there is an Encyclopedia Britannica article on the subject, but it is not in my 1971 ed.

R. Austin Freeman.  The magic casket.  Pearson's Magazine (Oct 1926) 288-299, ??NYS.  Collected in:  The Magic Casket; Hodder & Stoughton, London, 1927 and reprinted numerous times (I have 5th ptg, 1935), pp. 7-41.  Collected in:  Dr. Thorndyke Omnibus (variously titled); Hodder & Stoughton, London, 1929, pp. 398-425.  Says the phenomenon was explained by Sylvanus Thompson.  Says the polished lines project slightly and produce dark lines with a bright edges in the reflection.

Will Dexter.  Famous Magic Secrets.  Abbey Library, London, nd [Intro. dated Nov 1955].  P. 56.  Describing a visit to the premises of The Magic Circle, he says:  "Here are Japanese Magic Mirrors -- a whole shelf of them.  Made of a secret bronze alloy, ..., they have a curious property.  ...  Why?  Well now, people have written books to explain this phenomenon, ....  Some other time we'll talk about it...."

 

          10.E.  WHEEL PARADOXES

 

Clark.  Mental Nuts.  1897, no. 36;  1904, no. 44;  1916, no. 46.  The wheel.  "Does the top go faster than the bottom?"  Answer is: "Turning on ground, yes; on shaft, no."

See Laithwaite, 1977, in 10.D.1 for a combined wheel and mirror paradox.

 

          10.E.1.         ARISTOTLE'S WHEEL PARADOX

 

            Wheels of different sizes joined concentrically and rolling on two tracks at different heights.  At first, it appears that they each roll the same distance and hence must have the same circumference!

 

Aristotle (attrib.).  Mechanical Problems.  c-4C?  In:  Aristotle -- Minor Works.  Trans. by W. S. Hett; Loeb Classical Library, 1936, pp. 329‑441.  The wheel paradox is section 24, pp. 386‑395.

Heron (attrib.).  Mechanics (??*).  c2C?  Ed. by L. Nix & W. Schmidt.  Heronis Opera, vol. II, Teubner, Leipzig.  Chap. 7.  ??NYS.  (HGM II 347‑348.)

Cardan.  Opus Novum de Proportionibus Numerorum.  Henricpetrina, Basil, 1570, ??NYS.  = Opera Omnia, vol. IV, pp. 575-576.

Galileo.  Discorsi e Dimostrazione Matematiche intorno à Due Nuove Scienze Attenenti alla Mecanica & i Movimenti Locali.  Elzevirs, Leiden, 1638.  Trans. by S. Drake as:  Two New Sciences; Univ. of Wisconsin Press, 1974;  pp. 28‑34 & 55‑57.  (English also in:  Struik, Source Book, pp. 198‑207.)

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 59-60;  1945: 56‑57;  1961: 63‑64.  Footnote on p. 248 (1945: 229;  1961: 228) only dates it back to Galileo.

Israel Drabkin.  Aristotle's wheel:  notes on the history of a paradox.  Osiris 9 (1950) 162-198.

 

          10.E.2.         ONE WHEEL ROLLING AROUND ANOTHER

 

            How often does a wheel turn when it is rolled around another?  This is a well-known astronomical phenomenon -- the solar day and the sidereal day are different.  This is closely related to Section 10.E.3, qv.

 

Gardner.  Some mathematical models embedded in the solar system.  SA (Apr 1970) = Circus, chap. 16, Solar system oddities.  In 1806, a reader's letter to SA asked "How many revolutions on its own axis will a wheel make in rolling once round a fixed wheel of the same size?"  The editors replied "One", which started an enormous correspondence.  In vol. 18 (1868) 105-106, they printed a selection of letters, but more continued to come until in Apr 1868, they announced they were dropping the topic, but would continue it in a new monthly, The Wheel.  This was advertised as being available in the 23 May issue of SA.  In Carroll-Gardner, p. 46, Gardner indicates he has never seen a copy of The Wheel.

Clark.  Mental Nuts.  1897, no. 38;  1904, no. 46;  1916, no. 48.  The cog wheels.  "Suppose two equal cog wheels or coins (one stationary), how many turns will the other make revolving around it?"  Answer is: "Two full turns."

Pearson.  1907.  Part II, no. 58: The geared wheels, pp. 58 & 172.  10  tooth wheel turning about a  40  tooth wheel.

Dudeney.  AM.  1917.  Prob. 203: Concerning wheels, pp. 55 & 188.

McKay.  Party Night.  1940.  No. 23, p. 181.  Two equal circles, one rolling around the other.

W. T. Williams & G. H. Savage.  The Penguin Problems Book.  Penguin, 1940.  No. 80: Revolving coins, pp. 46 & 129.  Equal coins, then rolling coin of half the diameter.

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 55-57;  1945: 52‑54;  1961: 60‑62.  Also considers movement of a slab on rollers.

Gardner.  Coin puzzles.  SA (Feb 1966)  =  Carnival, chap. 2, Penny Puzzles.  Gives the basic problem and the elegant generalization for rolling around an arbitrary ring of coins of the same size.

 

          10.E.3.         HUNTER AND SQUIRREL

 

This is closely related to Section 10.E.2, qv.

 

Gardner.  Some mathematical models embedded in the solar system.  SA (Apr 1970) = Circus, chap. 16, Solar system oddities.  He says there have been many, mostly self-published, booklets arguing that the moon does not rotate and that several are included in De Morgan's Budget of Paradoxes of 1872. 

Jelinger Symons.  Letter: The moon has no rotary motion.  The Times (8 Apr 1856).  ??NYS -- discussed by Edward Wakeling in his edition of Lewis Carroll's Diaries, because Carroll responded and made an entry on 8 Apr.  Discussed in Carroll‑Gardner, pp. 45-46.  Carroll "noticed for the first time the fact that though [the moon] only goes 13 times round the earth in the course of the year, it makes 14 revolutions round its own axis, the extra one being due to its motion round the sun."  There were numerous responses to Symons' letter, and The Times printed seven of them on 9 Apr, but not Carroll's.  Symons riposted on 14 Apr, Carroll responded again, with a brief diary entry, and a reply appeared on 15 Apr, signed  E.B.D., but this is not known to be a pseudonym of Carroll and if it were, Carroll would most likely have recorded it in his diary.

Bubbenhall.  Letter:  A puzzle.  Knowledge 3 (9 Feb 1883) 91, item 719.  "A squirrel is sitting upon a post and a man is standing facing the squirrel, the squirrel presently turns round and the man moves round with it, always keeping face to face.  When the man has been round the post has he been round the squirrel?"  Proctor was editor at the time -- could he have written this letter??

Richard A. Proctor.  Editorial comment.  Knowledge 3 (9 Mar 1883) 141-142.  Says the hunter does go round the squirrel and that the problem is purely verbal.

W. Smith.  Letter:  The squirrel puzzle.  Knowledge 3 (4 May 1883) 268, item 807.  Disagrees with above, but not very coherently.  Proctor's comments do not accept Smith's points.  "In what way does the expression going round an object imply seeing every side of it?  Suppose the man shut his eyes, would that make any difference?  Or, suppose the man stood still and the squirrel turned round, so as to show him every side -- would the stationary man have gone round the squirrel?"

Clark.  Mental Nuts.  1897, no. 22;  1904, no. 15;  1916, no. 21.  The hunter and the squirrel.  Here the squirrel always keeps on the opposite side of the tree from the hunter.  The 1897 answer is: "No; he did not."  The 1904 ed extends this by: "They travel on parallel lines and do not change their relative position." and the 1916 abbreviates the answer as: "No; they travel on parallel lines, don't change relative position."

Pearson.  1907.  Part I: Round the monkey, p. 126.  Says R. A. Proctor discussed this some years ago in "Knowledge".

William James.  Pragmatism.  NY, 1907.  ??NYS.

Dudeney.  Some much‑discussed puzzles.  Op. cit. in 2.  1908.  'The answer depends entirely on what you mean by "go around."'

Loyd.  Cyclopedia.  1914.  The hunter and the squirrel, p. 61.  c= SLAHP: The hunter and the squirrel, p. 9.  Loyd Jr. says it is a "hundred‑year‑old question".

Collins.  Fun with Figures.  1928.  Monkey doodles business, pp. 232-234.  Monkey on a pole as in Bubbenhall.  "It is really a matter of personal opinion, ...."  Quotes Proctor's comments of 4 May 1883 with some minor changes.

Harriet Ventress Heald.  Op. cit. in 7.Z.  1941.  Prob. 36, p. 17.  Answer says it depends.

Robin Ault.  On going around squirrels in trees.  JRM 10 (1977‑78) 15‑18.  Cites James.  Develops a measure such that if the hunter and the squirrel move in concentric circles of radii  a  and  b,  then the hunter goes  a/(a+b)  around the squirrel and the squirrel goes  b/(a+b)  around the hunter.

 

          10.E.4.         RAILWAY WHEELS PARADOX

 

            New Section.  Although I learned this years ago and have used it as a problem, I don't recall seeing it in print before acquiring Clark in Aug 2000.

 

Clark.  Mental Nuts.  1897, no. 49.  Argument.  "The wheels of a locomotive are fixed fast on the axle.  The outer rail in a curve is the longest.  How do the outer and inner wheels keep even in rounding the curve?"  "By the bevel of the wheels and sliding."  His 'bevel' refers to the fact that the wheels are tapered, being smaller on the outside.  Centrifugal force on a curve causes the locomotive to move toward the outside of the curve so the outer wheel has a larger effective circumference than the inner wheel.

David Singmaster.  Wheel trouble.  Problem used as Round the bend, Weekend Telegraph (26 Nov 1988)  =  Wheel trouble, Focus, No. 9 (Aug 1993) 76-77 & 90.

                              Car enthusiasts will know that the rear axle of a rear-wheel-drive vehicle has a differential gear to allow the wheels to turn at different speeds.  This is essential for turning because the outside wheel travels along a circle of larger radius and hence goes further and turns more than the inner wheel.  But a railroad car has rigid axles, with the wheels firmly attached at each end.  How can a railroad car go around a bend?

                              One must look closely at the wheels of a railroad car to see the answer.  The wheels are tapered, with the larger part toward the inside of the car, and the rails are slightly rounded.  Consequently, when the car goes around a bend, its inertia (generally called centrifugal force) causes it to move a bit outward on the rails.  The outer wheel then rides out and up, giving it a larger radius, while the inner wheel moves in and down, giving it a smaller radius.

                              The diagram shows an exaggerated view of the wheels and axle.

 

                  /|                         |\

                 / |                         | \

                |  |                         |  |

                |  |                         |  |

                |  |                         |  |

                |  |                         |  |

                |  |                         |  |

                 \ |                         | /

                  \|                         |/

 

 

          10.F.  FLOATING BODY PROBLEMS

 

            Of course, the original floating body problem is Archimedes' testing of Hieron's crown.  I have only included a few examples of this -- it is fairly widely available.

 

Archimedes.  On Floating Bodies, Book I.  In:  T. L. Heath, The Works of Archimedes, ..., op. cit. in 6.AN, 1897 & 1912, pp. 253-262.  On pp. 258-261, Heath describes how Archimedes probably analysed Hieron's crown.

Marcus Vitruvius [Pollo].  De Architectura.  c-20.  Translated by Morris Hicky Morgan as: Vitruvius  The Ten Books on Architecture; Harvard Univ. Press, 1914 = Dover, 1960.  Pp. 253-254.  Describes Hiero's crown problem and says Archimedes simply measured the amount of water displaced by the crown and equal weights of gold and silver.

 E. J. Dijksterhuis.  Archimedes.  Op. cit. in 6.S.1.  1956.  Pp. 18-21 discusses the problem, noting that the object was actually a wreath, stating that the oldest known source is Vitruvius (late 1C) and giving several versions of the method thought to have been used by Archimedes.

[I've recently read Vitruvius and his description of Archimedes' work says he measured the displacements when the crown and equal weights of gold and silver were placed in water.  I'll expand on this when I find my photocopy.]

Isaac Disraeli.  Miscellanies of Literature.  Preface dated 1840 -- my copy is: New edition, revised;  Ward, Lock, London, 1882 (date of publisher's catalogue at end).  P. 211 relates that Charles II, when dining with the Royal Society "on the occasion of constituting them a Royal Society", asked what would happen if one had two pails of water of equal weight and put fish in one of them -- "he wanted to know the reason why that pail with such addition, should not weigh more than the other ...."  This produced numerous confused explanations until one member burst into laughter and denied the fact.  I find the phrasing confusing -- it seems that he wanted to know why the pails remained of equal weight.  However, there is a possible way that the pails would remain of equal weight -- if both pails were full to the brim, then the insertion of fish would cause an equal weight of water to overflow from the pail.  Disraeli is not specific about dates -- there are two basic dates of the beginning of the Society.  It was founded on 28 Nov 1660 and it was chartered on 15 Jul 1662, with a second charter on 22 Apr 1663.  The 1662 date seems most likely.

Dudeney??  Breakfast Table Problems No. 334: Water and ice.  Daily Mail (3  &  4 Feb 1905) both p. 7.  Ice in a full glass of water.  "..., what volume of water will overflow when the ice melts?"

Ackermann.  1925. 

Pp. 62‑63.  Barge in a canal going over a bridge.  How much more weight is on the bridge?

Pp. 94‑95.  Ice in full vessel of water.

Dudeney.  Problem 1060: Up or down?  Strand Mag. (Jun? 1931).  ??NX.  Boat full of iron in a reservoir.

Perelman.  FMP.  c1935?

Which is heavier?, p. 114.  Bucket of water versus equally full bucket with wood floating in it. 

Under water, pp. 199 & 202.  Submerge a balanced balance with stone on one side and iron on the other.

W. A. Bagley.  Paradox Pie.  Op. cit. 6.BN.  1944. 

No. 90: Supporting the ship, p. 60.  Ship in canal going over a bridge.

No. 91: A n'ice question, p. 60.  Ice in glass.

John Henry Cutler.  Dr. Quizzler's Mind Teasers.  Greenberg, NY, 1944.  ??NYS -- excerpted in: Dr. Quizzler's mind teasers; Games Magazine 16:3 (No. 109) (Jun 1992) 47 & 43, prob. 7.  Balance two bowls of water on a scales.  Add some goldfish to one of them.  Do the bowls still balance?

W. T. Williams & G. H. Savage.  The Third Penguin Problems Book.  Penguin, Harmondsworth, 1946.  Prob. 79: Aquatics, pp. 38 & 117.  Boat with iron weight in a bathtub.

J. De Grazia.  Maths is Fun.  Allen & Unwin, London, 1949 (reprinted 1963).  Chap. I, prob. 7, Cobblestones and water level, pp. 12 & 111.

Gamow & Stern.  1958.  The barge in the lock.  Pp. 104‑105.  Barge full of iron in a lock.

H. T. Croft  &  S. Simons.  Some little naggers.  Eureka 23 (Oct 1960) 22 & 36. 

No. 1.  Ordinary milk bottle with cream on top of the milk.  When shaken, they claim the pressure on the bottom becomes greater, but no reason is given and this seems wrong to me -- ??

No. 2.  Ice cube in a full glass of water.

No. 3.  Boat on a lake.

David Singmaster.  Any old iron?  Barge full of iron in the lock.  What happens to the boat when the iron is thrown into the lock?  Appeared as:  Watertight problem.  The Weekend Telegraph (9 Jun 1990) XXIV  &  (16 Jun 1990) XXIV. 

About 1994, there was some correspondence -- possibly in the Guardian's Notes & Queries column -- about barges in a canal on a viaduct.  Apparently Telford's 1801 Pontcysyllte viaduct on the Shropshire Union Canal at Chirk, near Llangollen, is  1007  ft long, but has a notice restricting the number of barges on it to three, though a barge is about  70  ft long.  Responses indicated that the reason for the restriction may have been wave problems.

Erwin Brecher  &  Mike Gerrard.  Challenging Science Puzzles.  Sterling, 1997.  [Reprinted by Goodwill Publishing House, New Delhi, India, nd [bought in early 2000]].  Pp. 33 & 73: Water level.  Discusses the problem of a boat in a pool and throwing a brick overboard.  Asks what happens if the boat springs a leak and slowly sinks?

 

        10.G.                             MOTION IN A CURRENT OR WIND

 

            Simple problems of this type are just variations of meeting problems -- see Vyse and Pike, etc. -- but they only seem to date from the late 18C.

            The comparison of up and down stream versus across and back is the basis of the Michelson-Morley experiment and the Lorentz-Fitzgerald contraction in relativity, so this idea must have been pretty well known by about 1880, but the earliest puzzle example I have is Chapin, 1927.

            More recently, I posed a problem involving travel uphill, downhill and on the level and I have now seen Todhunter.  I will add such problems here, but I may make a separate subsection for them.

 

Vyse.  Tutor's Guide.  1771?  Prob. 13, 1793: p. 186; 1799: p. 198 & Key p. 241.  Two rowers who can row at  5  set out towards each other at points  34  apart on a river flowing  2½.  Though this appears to belong here, it is simply  MR-(2½, 7½; 34).

Pike.  Arithmetic.  1788.   P. 353, no. 33.  Two approaching rowers, starting  18  apart and normally able to row at rate  4, but the tide is flowing at rate  1½.  I. e.  MR‑(2½, 5½; 18).  (Sanford 218 says this is first published version!)

D. Adams.  New Arithmetic.  1835.  P. 244, no. 86.  Two boats, with normal speeds  8   start to meet from  300  apart on a river flowing at rate  2.  I. e.  MR-(6, 10; 300). 

I imagine this appears in many 19C texts.  I have seen the following. 

T. Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.

Pp. 45-46, no. 14.  Rower goes  30  miles down and back in  7  hours.  He can row  2  miles upstream in the time he can row  5  miles downstream.  Find his rates.

P. 86, no. 11.  Rower goes  a  mph with the tide and  b  mph against the tide.  What is the rate of the tide?

Colenso.  Op. cit. in 7.P.1.  1849.

Exercise 47, no. 10, p. 85 & Answers, p. 11.  Rower goes  20  miles and back in  10  hours.  He goes  3  miles downstream in the time he goes  2  miles upstream.  How long does he take each way?

Miscellaneous examples, no. 336, p. xix & Answers, p. 23.  Crew rows    miles down and back up a river in  100  minutes, where the current is  2  miles per hour.  What is their rate in still water?

Edward Brooks.  The Normal Mental Arithmetic  A Thorough and Complete Course by Analysis and Induction.  Sower, Potts & Co., Philadelphia, (1858), revised, 1863.  Further revised as:  The New Normal Mental Arithmetic: A Thorough and Complete Course by Analysis and Induction;  Sower, Potts & Co., Philadelphia, 1873.  Several examples of the following type.

                    1863 -- p. 136, no. 4;  1873 -- p. 146, no. 4.  Steamboat goes  12  in still water; current is  4.  Goes down river and returns in  6  hours.  How far did it go?

Todhunter.  Algebra, 5th ed.  1870. 

Examples X, no. 21, pp. 85 & 577.  A crew can row  9 mph  in still water.  In a river, they find it takes them twice as long to go up a distance as to come back down.  How fast is the river?

Examples XXIV, no. 4, pp. 209 & 586.  A boat goes  3½ mi  down river and returns in  1 hr 40 min.  If the river goes at  2 mph,  how fast is the boat?

Miscellaneous Examples, no. 58, pp. 549 & 604.  A  to  B  is  7½ mi  with some uphill, some downhill and some level (or flat).  Man's walking speeds on these sections are  3, 3½, 3¼ mph  respectively.  It takes him  2 hr 17½ min  to go and  2 hr 20 min  to return.  I find that the general solution for the flat distance  S,  given the total distance  L,  speeds  U, D, F  and times  T1, T2,  is given by  S [2/F ‑ 1/U - 1/D]  =  T1 + T2 - L(1/U + 1/D). 

Miscellaneous Examples, no. 88, pp. 551 & 605.  Road from  A  to  B  comprises  5  uphill, then  4  level, then  6  downhill.  Man walks from  A  to  B  in  3 hr 52 min (i.e. 3:52)  and from  B  to  A  in  4:00.  He then walks halfway from  A  to  B  and returns in  3:55.  Find his three rates.

Colenso.  Op. cit. in 7.H.  These are from the (1864), 1871 material.  Let  V  be the velocity of the rower and  T  the velocity of the tide.

No. 6, p. 187.  V + T  =  5/3 * (V ‑ T)   and   V + T + ½  =  2 (V ‑ T ‑ ½).

No. 25, pp. 190 & 215.  V = 1/9  mile/minute,  V ‑ T = 1/14  mi/min.  What is  V + T?

Horatio N. Robinson.  New Elementary Algebra: Containing the Rudiments of the Science for Schools and Academies.  Ivison, Blakeman, Taylor & Co., New York, 1875.  Prob. 90, p. 305.  Man can row  15  miles down river in    hours, but requires    to row back.  What are the rates of the rower and of the river?

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.  No. 89, pp. 301 & 347.  Steamboat goes  10  in still water, current is  4,  boat goes down and returns in  10  hours -- how far did it go?

Briggs & Bryan.  The Tutorial Algebra, Part II.  Op. cit. in 7.H.  1898.  Exercises X, prob. 24, pp. 125 & 580.  Stream goes  4.  Man rows up and back and takes  39  minutes longer than in still water.  With a second rower, they can go  3/2  as fast as the single man.  They do the same trip in only  8  minutes longer than in still water.

W. P. Workman.  The Tutorial Arithmetic, op. cit. in 7.H.1.  1902.  Section IX, art. 240, example 3, p. 410 (= 416 in c1928 ed.).  Boat goes  4  mi in  20  min in still water and in  16  min with the tide.  How long against the tide?

Richard von Mises.  c1910.  Described in: George Pólya; Mathematical Methods in Science; (Studies in Mathematics Series, vol. XI; School Mathematics Study Group, 1963); reprinted as New Mathematical Library No. 26, MAA, 1977;  Von Mises' flight triangle, pp. 78-81.  How can one determine the airspeed of a plane from the ground speed when the wind is unknown?  "It was solved by Von Mises some fifty years ago; this I well remember as I heard it from him at that time."

Schubert.  Op. cit. in 7.H.4.  1913.  Section 17, no. 99, pp. 64‑65 & 140.  Steamship covers  60  km in  4  hours going upstream and in  3  hours going downstream.

Loyd.  Cyclopedia.  1914.  Riding against the wind, pp. 199 & 365.  = MPSL2, prob. 49, pp. 34 & 137.  = SLAHP: Wind influence, pp. 38 & 97.  Against and with the wind.

Loyd.  Cyclopedia.  1914.  The Santos‑Dumont puzzle, pp. 202 & 366.  Against and with the wind.

Peano.  Giochi.  1924.  Prob. 14, pp. 4-5.  Two ships travel  6000  miles and return.  The first goes at  8  mph and returns at  12  mph;  the second goes  10  mph both ways.  Which is faster?

Ackermann.  1925.  Pp. 77‑81.  Determine speed of wind using sound echoes.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927. 

P. 92 & Answers p. 9.  Flyers go from  A  to  B  and back, one in still air, the other in a steady wind.  Who is faster?  Answer notes that the naive view is that the wind helps and hinders equally, but one only needs to consider a wind as fast as the flyer to see that it is not equal and that for a lesser wind, the flyer is hindered for longer than he is helped.

Prob. 3, p. 98 & answers pp. 10-11: The problem of the swimmers.  Both swim a mile, one up and down stream, the other across and back.  Who is faster?  Answer gives detailed calculation.

Harriet Ventress Heald.  Op. cit. in 7.Z.  1941.  Prob. 48, pp. 21‑22.  Current of  1  mph.  Man rows up in  3  hours & back in  2  hours.  How far did he go?

D. A. Hindman.  Op. cit. in 6.AF.  1955.  Chap. 16, prob. 23: The floating hat, pp. 260‑263.  Hat falls off rower going upstream and is picked up on his return.

Gamow & Stern.  1958.  Boat and bottle.  Pp. 100‑102.  Bottle falls off boat going upstream and is picked up later.

David Singmaster.  Racing along.  Used in my puzzle columns.  Test track has one mile stretches of level, uphill, level, downhill, then two mile stretches, then three mile stretches, until a total of 60 miles.  For optimum economy, a car goes  30 mph  uphill,  60 mph  downhill and  40 mph  on the level.  How long does it take to cover the track?

                    Fuel on the hill.  Weekend Telegraph  (17 Dec 1988) xv  &  (24 Dec 1988) x.

                    G&P 16 (Jul 1995) p. 26.  (Solution never appeared as this was the last issue.)

 

          10.H. SNAIL CLIMBING OUT OF WELL

 

            A snail is at the bottom of a well which is  D  deep.  He climbs  A  in the day and slips back  B  in the night.  How long does it take to get out?  The earlier versions had serpents, snakes and lions.  The 'end effect' is that when the snail gets to within  A  of the top, he doesn't slip back.  The earlier versions did not have this clear and thus are just equivalent to meeting problems,  MR-(A, 0; D)  or  MR-(A, B; D)  or  O-(A, B) with headstart  D.  See 10.A:  Bakhshali ff. 60r-60v;  al-Karkhi no. 9(?);  Fibonacci pp. 177-178 & 182 (S: 274 & 280). 

            It appears that the  problem grows out of the usage of unit fractions such as  1/2 - 1/3  to specify rates.  At first this just meant  1/6  per day, but then it began to be interpreted as going ahead  1/2  followed by retreating  1/3,  resulting in the 'end effect'.  The earliest to treat the end effect clearly seems to be c1350. 

            For convenience, let the net gain per day be  G = A - B.  The solution is to take the least  N  such that  NG + A ³ D,  i.e.   N  =  é(D‑A)/Gù,   then interpolate during the daytime of the  (N + 1)‑st day,  getting  (D - NG)/A  of the day time on the  (N + 1)-st day as the time of meeting. 

            When we have approaching animals, going  +a, -b; +c, -d,  set  A = a + c,  B = b + d,  so  G = a - b + c - d.  If we are considering meeting without an end termination such as the cat eating the mouse, and  A ³ G,  then one can have multiple meetings.  The last night on which a meeting occurs during the retrogression is the last  M  such that  MG ³ D,  i.e.   M  =  ëD/Gû.   Thus there will be   2 (M - N) + 1   meetings, though this is reduced by one if  (D - A)/G  is an integer and by one if  D/G  is an integer.  (When both are integers, one reduces by two.)  Simple modifications deal with the cases of negative  B,  say if a snail continues going up at night, but at a different rate, and the case with  A < B. 

            Versions with two approaching animals:  Fibonacci,  Muscarello,  Chuquet,  Borghi,  Pacioli,  Tagliente,  Ghaligai,  Buteo,  Tartaglia.  Chuquet and Buteo treat the end effect clearly and Pacioli and Tagliente almost get it.

            More complex versions:  Mahavira, Wood.

            See Tropfke 588.  Tropfke 589 classifies these as  I A c (one person),  I B c (two persons meeting),  I C c (two persons overtaking).

 

Bakhshali MS.  c7C. 

See 10.A for ff. 60r-60v.   v  =  + 5/2 - 9/3,   D = 30.

Kaye I 51; III 222, f. 20v.   v  =  + 1/2 + 1/18 - 1/21,   D = 360.

Kaye I 51; III 225, f. 36v.  Seems to say   v  =  (1/2*3 + 1/3 ‑ 1/4)/(1/2*3)  ‑  (1/2*5)/(3/8),   D = 108,  but the problem is not complete.

Bhaskara I.  629.  Commentary to Aryabhata, chap. II, v. 26-27.  Sanskrit is on pp. 115-122; English version of the examples is on pp. 304-306.  The material of interest is example 4.   v  =  + 1/2 - 1/5,   D = 480.

Chaturveda.  860.  Commentary to Brahma‑sphuta‑siddhanta, chap. XII, section 1, v. 10.  In Colebrooke, pp. 283‑284.   v  =  + 4/5 ‑ 1/60,   D = 76,800,000.

Mahavira.  850.  Chap. V, v. 24‑31, pp. 89‑90.

v. 24:   v  =  (1/5 ‑ 1/9) / (3/7);   D = 4 * 99 2/5.

v. 25:   rate  =  (5/4) / (7/2) ‑ (5/32) / (9/2);   total = 70.

v. 26:   v  =  (3/10) / (11/2) ‑ (2/5) / (7/2);   D = 399/2.

v. 28:   lotus growing in well with outflow and evaporation of water and turtle pulling down lotus;   v  =  (5/2) / (3/3) + (6/5) / (4/3) + (5/2) / (3/2) ‑ (21/4) / (7/2);   D = 960;  problem has  1  for  4/3,  but  4/3  is needed for the given answer.

v. 31:  snake going into hole while growing;  v = (15/2) / (5/14) ‑ (11/4) / (1/4);   D = 768.

Sridhara.  c900.  Ex. 32‑33, pp. 24 & 93.

Ex. 32.   v  =  ½(1+¼)(1‑⅓)(1+½) / 6(1/5)(1/9)(⅓)(1+¼)    2(1‑⅓) / (1+½);   D = 100.

Ex. 33:   rate  =  (8 ‑ ½) / (1 + ⅓) ‑ ½;   total = 100.

Tabari.  Miftāh al-mu‘āmalāt.  c1075.  Pp. 103f.  ??NYS -- quoted in Tropfke 593.  No. 7.  "A boat comes forward 18 parasangs per day and goes backward 12 parasangs per day.  It comes and goes for 40 days.  How many days does it come and how many days does it go?"  This is not clear -- I wonder if 'days' in the last sentence should be 'parasangs' -- ??

Fibonacci.  1202. 

P. 177 (S: 273): De leone qui erat in puteo [On the lion who was in a pit].  v = 1/7 ‑ 1/9;  D = 50.  No alternation.  H&S 63 gives Latin and English.  H&S 64 claims that this is an example of day and night alternation, but this is not in the text.

Pp. 177-178 (S: 274): De duobus serpentibus [On two serpents].  Tower is 100.  One comes down  1/3 - 1/4,  other goes up  1/5 - 1/6.  No end effect.

Columbia Algorism.  c1350.  Prob. 67, pp. 88‑89.  Pigeon going down tower.  Rate  =  ⅔ ‑ (⅓ + ¼),   D = 10 (but Cowley says 50).  These are alternatively day and night, and the end effect is clearly treated to get  112  full days and one more daytime.  (Cowley 399.)

Pseudo-dell'Abbaco.  c1440.  Prob. 191, pp. 151‑153 with fine plate on p. 152.  Serpent in well of depth  30.  Goes  +2/3  in day and  ‑1/5  at night.  Gives simple answer and then carefully analyses the end effect to get  63  full days plus  9/10  of daytime.  Says that some solve it erroneously.  I have a colour slide of this.

AR.  c1450.  Prob. 65, pp. 47, 177, 224.  Tower  10  high.  Dove flies up    in the day and drops back  ¼ + ⅓  at night.  Answer is  120  days, which ignores the end effect.  Should be  112½,  as in the Columbia Algorism.

Muscarello.  1478. 

F. 73r, p. 187-188.  Cat climbing a tower  30  high, going  +½, ‑⅓.  End effect not treated, so he gets  180  days.

Ff. 74r-74v, pp. 188-189.  Cat at bottom of a tower going  +⅓, ‑¼.  Mouse at the top going down  ½  and back up  ⅓.  When do they meet?  Again, the end effect is not treated and he gets  240  days.

Tommaso della Gazzaia.  Liber geometriae.  Manuscript C.III.23, Biblioteca Communali di Siena.  15C??  F. 169r.  ??NYS -- quoted in Franci, op. cit. in 3.A, p. 29.  Serpent going  +⅓ -¼  up a tower  30  high.  End effect ignored.

Chuquet.  1484.  Mentioned on FHM 204 as 'the frog in the well'.  All these treat the end effect clearly.

Prob. 125:  +15, ‑10,  to go  100.  Takes  17  natural days and one 'artificial' day.

Prob. 126:  +15, ‑9,  to go  100.  Takes  15  natural days and    of an artificial day.

Prob. 128:  two travellers  100  leagues apart travelling  +12, ‑7  and  +10, ‑6.  Answer:  9  days and  9  nights with  19/22  of a day without the night.

Borghi.  Arithmetica.  1484. 

F. 110r (1509: f. 92v).  Two brothers paying a debt from earnings less expenses:  rate  =  2/3 ‑ 2/5 + 3/4 ‑ 1/2;   total = 700.  (H&S 64 gives Latin.)

Ff. 110r-110v (1509: ff. 92v-92r).  Sparrow hawk at bottom and dove at top of a tower  60  high.  Hawk goes  + ⅔ - ½;  dove goes  + ¾ - ⅔,  approaching in day and retreating at night.  Ignores end effect and gets  240  days.  With end effect, one gets  235  days plus  15/17  of the daytime.

Calandri.  Arimethrica.  1491. 

F. 66r.  Traveller goes  +4  in the day and  -3  in the night to go  20.  Gets  17  days, meaning  16  whole days and a daytime.

F. 71v.  Serpent in well.  +1/7  in the day and  -1/9  in the night to go  50.  He doesn't consider the end effect, so gives  1575  instead of  1572½.)  Nice woodcut, reproduced in Smith, History II 540 and  Rara 48.  Same as Fibonacci, p. 177. 

F. 72r.  Two serpents on a tower of height  100  going  +1/3, -1/4;  +1/5, -1/6,  clearly distinguished as day and night, but he ignores the end effect.  Woodcut showing one serpent (or dragon) at bottom of tower.

F. 72v.  Two ants meeting from  100  away going  +1/3, -1/4;  +1/5, -1/6,  distinguished as day and night, but he ignores the end effect.  Woodcut showing two ants.

F. 73r.  Two ants are at distances  D  and  D + 100  from a pile of grain.  Going  +1/3, ‑1/4;  +1/5, -1/6,  distinguished as day and night, gets them to the pile at the same time.  He ignores the end effect.  Woodcut showing ants by a pile.

F. 73v.  Cat and squirrel in tree of height 26.  They go  +1/2, -1/3;  +1/4, -1/5  clearly distinguished as day and night.  Ignores end effect and gets  120,  but then adds one to get  121  for no clear reason.

Pacioli.  Summa.  1494. 

F. 42r, prob. 22.  Two ants  100  apart approaching at rates  1/3, ‑1/4  and  1/5, -1/6,  clearly stated to be day and night.  He says they approach  7/60  per whole day and then computes  (100 - 7/60) / (7/60)  and adds  1  to get  762 6/7  days.  In the numerator, he seems to have used  88 53/60  instead of  99 53/60.  His process is close to the general solution with end effect, but he should subtract  8/15 = 1/3 + 1/5  instead of  7/60  and he should interpolate on the last day.

Ff. 42r-42v, prob. 23.  Cat & mouse at bottom and top of a growing and shrinking tree originally  60  high.  Mouse descends  1/2  per day and returns  1/6  at night.  Cat goes  +1, -1/4.  Tree grows  +1/4, -1/8  between them.  He says the net gain between them is  23/24  per whole day and computes  (60 - 23/24) / (23/24)  and adds  1  to get  T = 62 14/23.  He asserts the tree has grown  T/8  -- i.e. he is still thinking of overall rates rather than considering the alternation properly.  (Sanford 207‑208.  H&S 64‑65 gives Italian and English.)

F. 42v, prob. 24.  +2, -1  to go  10.  Obtains answer of  10  and says it isn't right and should be  8  whole days and a daytime, which he calls  9  days.  He also does  +3, -2  and gets  7  whole days and a daytime.

Blasius.  1513.  F. F.ii.r: Secunda regula.  +50, -19  to reach  992.  No end effect considered, even though the rates are in the day and in the night.

Tagliente.  Libro de Abaco.  (1515).  1541.  Prob. 121, f. 59r.  Cat and mouse on tree which is  26¾  tall.  Cat goes  +1/2, -1/3 (misprinted  2/3);  mouse goes  +1/4, -1/5.  Seems to say they meet on the  120th  day, but it should be the  121st  day.

Ghaligai.  Practica D'Arithmetica.  1521.  Prob. 19, f. 64v.  Two ants  100  apart going toward a pile of grain.  Further goes  +7, -4;  other goes  +5, -3  during day, night.  How far is the pile if they get there at the same time?  He says they take  100  days, but this ignores the end effect -- they meet at the end of the  99th  daytime.  This is equivalent to a snail going  +2, -1  up a wall of  100,  (H&S 65 gives Italian and English.)

Riese.  Rechnung.  1522.  1544 ed. -- pp. 107‑109;  1574 ed. -- pp. 72v‑73r.  The 1574 ed. calls it Schneckengang.  +4⅔, ‑3¾,  D = 32.  Treats end effect clearly and says it was first done correctly by Hansen Conrad, Probierer zu Eissleben (?= assayer at Eissleben).  He discusses how to convince people of the end effect.

Tonstall.  De Arte Supputandi.  1522.  Quest. 36, p. 168.  +70, -15  clearly stated to be day and night, to go  4000,  but he ignores the end effect.

Riese.  Die Coss.  1524.  No. 142, p. 61.  +4¼, ‑3⅓,  D = 32.  Treats end effect properly.  Says the Nurmbergk Rechenmeister N. Kolberger got it wrong and that Hans Conradtt got it right.

Giovanni Sfortunati.  Nuovo lume.  Venice, 1545.  F. 88r.  ??NYS -- described by Franci, op. cit. in 3.A, p. 41.  Solves a problem with end effect and says that Borghi, Pacioli and Calandri have done it wrong.

Christoff Rudolph.  Künstlich rechnung mit der ziffer und mit den zalpfennigen ...  Auffs new wiederumb fleissig ubersehen und an vil arten gebessert.  Nuremberg, 1553, 1561.  ??NYS -- quoted by:  Grosse; Historische Rechenbücher ...; op. cit. in 7.H under Faulhaber, p. 28.  Von einem schnecken.  Snail climbing  + 7, ‑2  to get out of a well  20  deep.  End effect clearly treated.  [This is a revision of the 1526 ed., ??NYS, which may have the problem?]

                    J. De Grazia; Maths is Fun; p. 12, attributes the problem to Christoff Rudolph (1561).

Buteo.  Logistica.  1559.  Prob. 33, pp. 234-237.  Problem of ships with oscillating winds.  They start  20000  apart.  First goes  +1200  per day and  -700  per night.  Second starts on the first night and goes  +1400  per night and  -600  per day.  He first solves without end effect and then end effect, getting  14  days +  12  hours of day +  10 2/7  hours of the night.  (H&S 65 gives Latin & English.)

Gori.  Libro di arimetricha.  F. 73r (p. 80).  Height  50,  rates  +⅓, ‑ ¼.  Notes end effect, but then forgets to add the last day!

H&S 64 says examples with alternating motion are in:  Fibonacci (1202),  Columbia Algorism (c1350),  Borghi (1484),  Calandri (1491),  Pacioli (1494?),  Tartaglia  and  Riese.  Also that examples with two animals approaching are in:  Fibonacci (two ants),  Borghi (1484, hawk & dove),  Pacioli (1494?, both types)  and  Tartaglia (both types).

Faulhaber.  Op. cit. in 7.H.  1614.  ??NYS -- quoted by Grosse, loc. cit. in 7.H under Faulhaber, p. 120.  No. 12, p. 212.  Worm climbs  +¾, ‑⅓  up a tree  100  high.  Gives answer with end effect.

Dilworth.  Schoolmaster's Assistant.  1743.  P. 166, no. 91.  +20, -15  to go  150.  No end effect considered so answer is  30  days.

Walkingame.  Tutor's Assistant.  1751.  1777: p. 172, prob. 47;  1860, p. 180, prob. 46.  Snail going  +8, -4  to get up a May pole  20  high.  Answer is  4  days, presumably meaning  3  whole days and a daytime.

Vyse.  Tutor's Guide.  1771?  Prob. 27, 1793: p. 40; 1799: pp. 43-44 & Key p. 39.  Same as Walkingame, but answer is done step by step and says 'the fourth Day at Night'.

D. Adams.  Scholar's Arithmetic.  1801.  P. 209, no. 1.  Frog in well  +3, -2  to go  30.  No answer.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 5, pp. 15 & 72.  Snail going  +8, -4  to get up a maypole 20 high.  Treats end effect properly but states the answer is  4  days.

Family Friend 1 (1849) 150 & 178.  Problems, arithmetical puzzles, &c. -- 2.  Snail on wall.  +5, -4  to reach  20.  End effect clearly treated.  = The Illustrated Boy's Own Treasury, 1860, Arithmetical and Geometrical Problems, No. 24, pp. 429 & 433.

Magician's Own Book.  1857.  The industrious frog, p. 234.  + 3, ‑ 2  up a well  30  deep.  End effect treated.  = Boy's Own Conjuring Book, 1860, p. 200.

Charades, Enigmas, and Riddles.  1860: prob. 24, pp. 59 & 63;  1862: prob. 25, pp. 135 & 1865: prob. 569, pp. 107 & 154.  Snail going  +5, -4  up a wall  20  high.  Answer is  16  days, which considers the end effect, but doesn't describe the final daytime well.

Bachet-Labosne.  Problemes.  3rd ed., 1874.  Supp. prob. II, 1884: 182.  +9, -5  to cover  173.  Gives simple answer and then considers end effect.

Mittenzwey.  1880. 

Prob. 136, pp. 29 & 78;  1895?: 154, pp. 32 & 81;  1917: 154, pp. 29-30 & 78.  Locomotive goes  +11, -7  in alternate hours.  When does it get to  255  away.  Gives trivial approach and then asks why it is wrong.

1895?: prob. 18, pp. 9 & 63;  1917: 18, pp. 8 & 57.  Snail going up a wall of height  40,  going  +8, -5  in day and night.  1895? just states the answer; 1917 gives a brief comment, referring to prob. 154.

Lucas.  L'Arithmétique Amusante.  1895.  Prob. XI: La ballade de l'escargot rétrograde, pp. 25-26.  +5, -2  going up a tree of height  9.  Also quoted in Laisant; op. cit. in 6.P.1; 1906; p. 125.

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 9:6 (Apr 1903) 544-546.  "A frog was trying to get up a slippery bank twelve feet high.  In the first twelve hours he climbs eight feet, but in the next twelve hours he loses four feet.  How long will he be in reaching the top?"  "Why, having lost his four feet, how could he get to the top at all?"

Clark (1897, 1904 & 1916),  Pearson (1907),  Loyd (Cyclopedia, 1914),  Ahrens (A&N, 1918),  Loyd Jr. (SLAHP, 1928)  all have versions.

Wood.  Oddities.  1927.

Prob. 65: A snail's journey, p. 51.  Usual problem with  A, B, H  =  3, 2, 12.

Prob. 66: another snail, pp. 51-52.  A, B, H  =  3, 2, 20,  but the snail has to get to the top and down the other side.  He asserts the level speed of the snail would be  3 + 2 = 5 per day  and hence the downward speed is  5 + 2 = 7 per day.  Hence, at the end of the daylight of the 18th day, he has reached the top and then slides down 2 on the other side in that night.  It then take two more whole days to reach the bottom, making 20 whole days for the trip.

M. Adams.  Puzzles That Everyone Can Do.  1931.  Prob. 197, pp. 76 & 158: Seaside problem.  Deck-chair man earns £6 when it is sunny but loses £3 when it rains.  The weather alternates, being sunny on the first day.  When is he £60 ahead?

Stephen Leacock.  Model Memoirs and Other Sketches from Simple to Serious.  John Lane, The Bodley Head, 1939, p. 290.  In a sketch on quizzes, he has the following.

                    "Sometime we drop into straight mathematics, which has the same attraction as playing with fire: for example: -- If a frog falls into a sand-pit twenty feet deep and gets up the side in jumps two feet at a time, but slips back one foot on the sand while taking his breath after each jump, how many jumps would it take him to get out of the pit?

                    There, be careful with it.  Don't say you can do it by algebra -- that's cheap stuff -- and anyway you can't."

Evelyn August.  The Black-Out Book.  Op. cit. in 5.X.1.  1939.  The man who hoarded petrol, pp. 154 & 215.  Tank holds  8  gallons, man puts in  2  gallons every day and  1  gallon leaks out every night;  when is it full?

 

          10.I.   LIMITED MEANS OF TRANSPORT -- TWO MEN AND A BIKE, ETC.

 

            The men have to get somewhere as quickly as possible and have to share a vehicle which may be left to be picked up or may return to pick up others.

 

Laisant.  Op. cit. in 6.P.1.  1906. 

Chap. 51: Deux cyclistes pour une bicyclette, pp. 127-129.  Graphic solution, assuming walking speeds and riding speeds are equal, but notes one can deal with the more general problem with a little more mathematics.

Chap. 52: La voiture insuffisante, pp. 129-132.  Two couples, but the car can only carry two persons besides the driver.  Graphic solution, again assuming walking speeds and driving speeds are equal, but making an estimate for calculational convenience.  He also states the exact solution.

Loyd.  Tandem puzzle.  Cyclopedia, 1914, pp. 322 & 382 (erroneous solution).  = MPSL2, prob. 123, pp. 88 & 160‑161, with solution by Gardner.  Three men and a tandem bike.

Loyd Jr.  SLAHP.  1928.  A tandem for three, pp. 52 & 104.  Like Loyd's but with different data.

Dudeney.  PCP.  1932.  Prob. 75: A question of transport, pp. 29 & 136.  = 536, prob. 89, pp. 27 & 244.  Twelve soldiers and a taxi which can take four of them.

Haldeman-Julius.  1937.  No. 64: Tandem for three problem, pp. 9 & 24.  Tom, Dick and Harry can walk  3, 4, 5  mph.  They have a tandem bike and any one or two of them can ride it at  20 mph.  They want to go  43 1/3  miles as quickly as possible.  Clearly the slowest boy should stay on the bike at all times, so this is really two men and a taxi. 

Gaston Boucheny.  Curiositiés & Récréations Mathématiques.  Larousse, Paris, 1939, pp. 77‑78.  Two men and a bike.

R. L. Goodstein.  Note 1797:  Transport problems.  MG 29 (No. 283) (Feb 1945) 16‑17.  Graphical technique to solve general problem of a company of men and a lorry.

William R. Ransom.  Op. cit. in 6.M.  1955.  A ride and walk problem, pp. 108‑109.  Same as Dudeney.

Karl Menninger.  Mathematics in Your World.  Op. cit. in 7.X.  1954??  A bit of 'hitch hiking', pp. 100‑101.  Three men and a motorcyclist who can carry one passenger.  Graphical method, but he neglects to consider that the passengers can be dropped off before the goal to walk the remaining distance while the cyclist returns to pick up the others.

Doubleday - 2.  1971.  Dead heat, pp. 79-80.  Two men and a pony, but the pony is assumed to stay where it is left, so this is like two men and a bike.  If the pony could go back to meet the second traveller, then this would be two men and a motorcyclist.

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 103: Two men and a bicycle, pp. 50 & 94.  Journey of 25, men can walk at rates 3 and 4 or either can ride the bike at rate 7.

David Singmaster.  Symmetry saves the solution.  IN:  Alfred S. Posamentier & Wolfgang Schulz, eds.; The Art of Problem Solving: A Resource for the Mathematics Teacher; Corwin Press, NY, 1996, pp. 273-286.  Men and a vehicle, pp. 282-285.  Uses appropriate variables to make the equations more symmetric and hence easily solvable, even with different speeds.

 

          10.J.  RESISTOR NETWORKS

 

          Find the resistance between two points in some network of unit resistors.

 

E. E. Brooks & A. W. Poyser.  Magnetism and Electricity.  Longmans, Green & Co., London, 1920.  Pp. 277‑279.  ??NYS.

Anon.  An electrical problem.  Eureka 14 (Oct 1951) 17.  Infinite square lattice of one ohm resistors along edges.  Asks for resistance from  (0, 0)  to  (0, 1)  and to  (1, 1).  Solution to be in No. 15, but isn't there, nor in the next few issues.

Huseyin Demir, proposer;  C. W. Trigg, solver.  Problem 407 -- Resistance in a cube.  MM 33 (1959‑60) 225‑226  &  34 (1960‑61) 115‑116.  All three inequivalent resistances for a cubical network are found, i.e. the resistances between points which are distance  1, 2, 3  apart.  Trigg says  distance = 1 and 3  cases are in Brooks & Poyser.

Gardner.  SA (Dec 1958) = 2nd Book, p. 22.  Gives cube solution and cites Trigg's reference to Brooks & Poyser.

B. van der Pol & H. Bremmer.  Operational Calculus -- Based on the Two‑sided Laplace Integral.  2nd ed., Cambridge Univ. Press, 1959.  ??NYS.  'Very last section' obtains formulae for the resistance  R(m, n)  in the infinite plane network between  (0, 0)  and  (m, n).  These are in terms of Chebyshev polynomials or Bessel functions.

Albert A. Mullin & Derek Zave, independent proposers;  A. A. Jagers, solver.  Problem E2620 -- Symmetrical networks with one‑ohm resistors.  AMM 83 (1976) 740  &  85 (1978) 117‑118.  All regular polyhedra and the  n‑cube, but only between furthest vertices.  (Editorial note says the cubical case first occurs in Coxeter's Regular Polytopes, but I can't find it??  It also erroneously says Gardner gives other solutions.)

D. C. Morley, proposer;  Friend H. Kierstead Jr., solver.  Problem 529 -- Hyper‑resistance.  JRM 9 (1976‑77) 211  &  10 (1977‑78) 223‑224.  4‑cube, resistance between points  3  apart.

David Singmaster, proposer;  Brian Barwell, solver.  Problem 879 -- Hyper‑resistance II.  JRM 12 (1979‑80) 220  &  13 (1980‑81) 229‑230.  n‑cube, furthest vertices.  What happens as  n  goes to infinity?

David Singmaster, proposer;  B. C. Rennie, partial solver.  Problem 79‑16 -- Resistances in an  n‑dimensional cube.  SIAM Review 21 (1979) 559  &  22 (1980) 504‑508.  In an  n‑cube, what are the resistances  R(n, i)  between points  i  apart?  Results are only known for  i = 1, 2, 3, n‑2, n‑1, n.  Other solvers considered the infinite  n‑cubical lattice and obtained a general result in terms of an  n‑fold integral, including  R(n, 1) = 1/n.

P. Taylor & C. Feather.  Problems drive 1981.  Eureka 44 (Spring 1984) 13‑15 & 71.  No. 1.  Find all resistances for regular polyhedral networks.

P. E. Trier.  An electrical resistance network and its mathematical undercurrents.  Bull. Inst. Math. Appl. 21:3/4 (Mar/Apr 1985) 58‑60.  Obtains a simple form for  R(m, n)  (as defined under van der Pol) which involves a double integral.  He evaluates this explicitly for small  m, n.  The integral is the same as that of the other solvers of my problem in SIAM Review.

P. E. Trier.  An electrical network -- some further undercurrents.  Ibid. 22:1/2 (Jan/Feb 1986) 30‑31.  Letter making a correction to the above and citing several earlier works (McCrea & Whipple, 1940;  Scraton, 1964;  Hammersley, 1966 -- all ??NYS) and extensions:  an explicit form for  R(m, m),  the asymptotic value of  R(m, m)  and extensions to three and  n  dimensions.

 

          10.K. PROBLEM OF THE DATE LINE

 

          A man who circles the earth gains or loses a day. 

          I include some related problems here.  See also 6.AF.

 

E. John Holmyard.  Alchemy.  Penguin, 1957, p. 119, says Roger Bacon (1214-1294 (or 1292)) foresaw circumnavigation, but doesn't indicate if he recognized the date problem.

Nicolas Oresme.  Traitié de l'espere.  c1350.  ??NYS -- described in:  Cora E. Lutz; A fourteenth‑century argument for an international date line; Yale University Library Gazette 47 (1973) 125‑131.  Chap. 39.  Three men.  One circles the world eastward in  12  days, another westward and the third stays at home.  He computes their effective day lengths.

                    Lutz describes the occurrence of the problem in other of Oresme's writings.

                    Quaestiones supra speram (c1355),  where the travellers take  25  days and Oresme suggests "one ought to assign a definite place where a change of the name of the day would be made".

                    His French translation of Aristotle's  De caelo et mundo  as:  Traitié du ciel et du monde (1377),  where they take  9  days.

Kalendrier des Bergers; 1493.  = The Kalendayr of the Shyppars (in a Scottish dialect); 1503.  = The Shepherds' Kalendar; R. Pynson, London, 1506.  ??NYS.  Described in:  E. G. R. Taylor; The Mathematical Practitioners of Tudor & Stuart England; (1954); CUP for the Inst. of Navigation, 1970; pp. 11‑12 & 311.  Three friends, one stays put, others circle the earth in opposite directions.  When they meet, they disagree on what day it is.

Antonio Pigafetta.  Magellan's Voyage: A Narrative Account of the First Circumnavigation ....  Translated and edited by R. A. Skelton, 1969.  Vol. I, pp. 147‑148.  ??NYS -- quoted by Lutz.  When they reached Cape Verde in 1522, a landing party was told "it was Thursday, at which they were much amazed, for to us it was Wednesday, and we knew not how we had fallen into error."

Cardan.  Practica Arithmetice.  1539.  Chap. 66, section 34, ff. DD.iiii.v ‑ DD.v.r (p. 145).  Discusses ship which circles world three times to the west, and also mentions going east.  (H&S 11 gives Latin.  Sanford 214 thinks he was first to note the problem.)

Cardan.  De Rerum Varietate.  1557, ??NYS.  = Opera Omnia, vol. III, pp. 239-240.  Liber XVII.  Navis qua orbem cirūambivit.  Describes Magellan's circumnavigation, giving dates and saying they had lost a day.

van Etten.  1624.  Prob. 91 (96), part IV (7), p. 141 (231‑232).  "How can two twins, who are born at the same time and who die together, have seen a different number of days?"  The English edition comments on a Christian, a Jew and a Saracen having their Sabbaths on the same day.

"A Lover of the Mathematics."  A Mathematical Miscellany in Four Parts.  2nd ed., S. Fuller, Dublin, 1735.  The First Part is:  An Essay towards the Probable Solution of the Forty five Surprising PARADOXES, in GORDON's Geography, so the following must have appeared in Gordon.

Part I, no. 8, pp. 8-9.  Two men born & dying together, but living different numbers of days.  Gives various explanations, e.g. differing calendars, living in the Arctic and going around the earth.

Part I, no. 9, p. 9.  Two nearby places have their dates different.  Again, this can be due to different sabbaths:  Christians on Sunday, Graecians on Monday, Persians on Tuesday, Assyrians on Wednesday, Egyptians on Thursday, Turks on Friday, Jews on Saturday.  Better, Macao and the Philippines differ by a day since one was colonized from the west and the other from the east.

Ozanam‑Montucla.  1778.  Vol. III, prob. 5, 1778: 32-33;  1803: 34-35;  1814: 33-34;  1840: 415-416.  Two men, born & dying together, but one living one day, or even two days, more than the other.

Philip Breslaw (attrib.).  Breslaw's Last Legacy.  1784?  Op. cit. in 6.AF.  1795: pp. 36-38.

                    Geographical Paradoxes.

Paradox IV, p. 36.  How can there be two persons who were born and died at the same time and place, but one lived several months longer than the other? 

Paradox VI, pp. 37-38.  A man sees one day which is the longest day, the shortest day and a day whose day and night are equal.  Answer: he crosses the equator on 22 Jun.

Carlile.  Collection.  1793. 

Prob. CXIV, p. 68.  Two children who were born and died at the same time, "yet one was several months older than the other."  One lived within the Arctic Circle, at 73o22', where days are three months long!

Prob. CXV, p. 68.  Two sailors meet and find their calendars are off by a day.  They have gone to New Zealand, one eastward and the other westward.

Jackson.  Rational Amusement.  1821.  Geographical Paradoxes.

No. 6, pp. 35-36 & 102-103.  Two nearby places in Asia whose reckonings differ by a day.  Solution is either Christians and Jews, whose sabbaths are a day different, or Macao and the Philippines, which differ by a day since the Portuguese came to Macao from the west and the Spanish came to the Philippines from the east.

No. 8, pp. 36 & 103-104.  Children born and dying at the same times, but one lives longer than the other.  Solution is either due to difference between lunar and solar calendars or due to one sailing round the world.

No. 47, pp. 47 & 114.  How can Christians, Jews and Turks all have their sabbath on the same day.  Christian sails east and the Turk west.

Endless Amusement II.  1826?  Pp. 74-75:  "How Two Men may be born on the same Day, die at the same Moment, and yet one may have lived a Day, or even two Days more than the other."  Notes that this can give three Thursdays in one week.

Carroll.  The Rectory Umbrella.  c1845??  Difficulties: No. 1.  In:  The Rectory Umbrella and Mischmasch; (Cassell, London, 1932), Dover, 1971, pp. 31‑33.  Also in:  Carroll-Collingwood, pp. 4‑5 (Collins: 11-12).  Carroll-Gardner, pp. 81-82, says he published a letter on this in the Illustrated London News of 18 Apr 1857.  In 1860, he later gave this as a lecture to the Ashmolean Society at Oxford. 

Warren Weaver.  Lewis Carroll:  Mathematician.  Op. cit. in 1.  1956.  Mentions Carroll's interest in the problem.

J. Fisher.  Where does the day begin?  The Magic of Lewis Carroll.  Op. cit. in 1, pp. 22‑24.  This discusses several of Carroll's versions of the problem.

Vinot.  1860.  Art. CXXIV: La semaine des trois jeudis, p. 145.  How to have three Thursdays in a week.

Jules Verne.  Le tour du monde en quatre-vingts jours (Around the World in Eighty Days).  1873.  Features the gain of a day by going eastward.

In 1883 and 1884, the Rome and Washington Conferences for the Purpose of Fixing a Prime-Meridian and a Universal Day adopted the Greenwich Meridian and the basic idea of time zones, which implied the acceptance of the International Date Line.  The Philippines, having been colonized from the New World, had to skip a new to conform with its Asian neighbours, but I don't know when this happened.  When Alaska was purchased in 1867, it had to drop 12 days to convert to the Gregorian calendar and then had to have one eight-day week to conform with the rest of the New World.

Clark.  Mental Nuts.  1897, no. 89.  Where is the day lost?  Travel from San Francisco to Pekin at the speed of the sun and one finds it is a day later.  Answer says "New day begins at 180th meridian, which is midway in the Pacific Ocean."

A. M. W. Downing.  Where the day changes.  Knowledge 23 (May 1900) 100-101.  Observes that the precise location of the Date Line has not yet been fixed.  Gives four versions on a map, the last two of which only differ at a few islands.

F. &. V. Meynell.  The Week‑End Book.  Op. cit. in 7.E.  1924.  2nd. ed., prob. eight, pp. 276‑277;  5th? ed., prob. twelve, p. 410.  What happens if you go around the world from west to east in  24  hours, or in less than  24  hours?  Suggests that in the latter case, you get back before you started!

McKay.  Party Night.  1940.  No. 36, p. 184.  Two airmen circle the earth in opposite directions, both taking  14  days.  Which gets back home first?  Answer is that the eastbound one gets back two days sooner because the  14  days are considered as days viewed by the airmen.

A. P. Herbert.  Codd's Last Case and Other Misleading Cases.  Methuen, London, 1952.  Chap. 14: In re Earl of Munsey:  Stewer v. Cobley -- The missing day case, pp. 72-83.  (This probably appeared in Punch, about 7 Dec 1949.)  Reprinted in:  More Uncommon Law; Methuen, 1982, pp. 74-83.  Lord Munsey left property to his great-nephew "if he has attained the age of 21 before the date of my death", otherwise the property went to Lord Munsey's brother.  The great-nephew's birthday was 2 May.  On 1 May, Lord Munsey was on a cruise around the world and the ship crossed the Dateline at about noon, going westward, so the Captain declared that the next day would be 3 May at midnight, but Lord Munsey expired just after midnight.  So who inherits?

Jonathan Always.  Puzzles for Puzzlers.  Tandem, London, 1971.  Prob. 2: Another birthday poser, pp. 11 & 60.  Four persons, born on different dates, lived to be fifty years of age, but never saw their fiftieth birthday.  One was born on 29 Feb.  "Another crossed the International Date-line travelling eastward, so that he gained a day and never arrived at his fiftieth birthday at all however long he lived."  This confuses me -- crossing the Dateline goes back a day, so it would seem his fiftieth birthday would simply be delayed by at most a day.  Always continues: "The third crossed the International Date-line travelling westwards on the eve of his fiftieth birthday, thus losing a day, a died a few hours later."  I think he has confused things.  If a man travels eastwards across the Dateline on the day before his 50th birthday, then he goes back a day and hasn't yet arrived at his 50th birthday, though he is now 50 years old.  If he dies before midnight, then that fits the problem.  If another man travels westwards across the Dateline at midnight on the eve of his fiftieth birthday, he goes ahead two days and really loses a day, namely his fiftieth birthday, which he never saw.  The fourth person was blind!

Shakuntala Devi.  Puzzles to Puzzle You.  Op. cit. in 5.D.1.  1976.  Prob. 138: The Sabbath day, pp. 86 & 134.  In order for a Moslem, a Jew and a Christian to have their Sabbath at the same time, send the Moslem around the world to the west and the Christian around the world to the east.  When they meet again they will all have the same Sabbath day!

David Singmaster.  Letter [on the International Date Line].  Notes and Queries column, The Guardian, section 2 (20 Dec 1995) 7.  Reprinted in Guardian Weekly (7 Jan 1996) 24.  Sketches the history from Oresme onward and notes the anomalies that Alaska and the Philippines essentially crossed the Date Line.

Ed Barbeau.  After Math.  Wall & Emerson, Toronto, 1995.  Double Christmas, pp. 153-154.  Straightforward problem illustrating that flying east from Australia, one may return to the previous day.  Here the heroine gets 36 hours of Christmas, in two disconnected parts!  [Indeed if one just crosses the date line from west to east at midnight, one gets 48 hours of the same day.]

 

          10.L.  FALLING DOWN A HOLE THROUGH THE EARTH

 

            The angular frequency of the oscillation is  Ö(g/R)  where  g  is the acceleration of gravity at the surface and  R  is the radius of the earth.  Taking a mean radius of  3956.665 mi  and  g = 32.16 ft/sec2  gives half-period of  42.20 min  =  42 min 12 sec.

 

Hesiod.  The Theogony.  c-700.  IN: The Homeric Hymns and Homerica, translated by H. G. Evelyn White; Harvard Univ. Press, 1959, p. 131, lines 724-725.  ??NYS -- information sent by Andrew Simoson [email of 25 Feb 2003].  This claims that a brazen anvil dropped from the Earth's surface will reach Tartarus in nine days.

Plutarch.  Plutarch's Moralia, vol. XII, translated by H. Cherniss & W. C. Helmbold.  Harvard Univ. Press, 1957, pp. 65-67.  ??NYS -- cited by Simoson.  Plutarch notes that if the centre of the earth is at a person's navel, then both his head and his feet are pointing up.  Simoson quotes Plutarch as: "Not that ... masses ... falling through the ... earth stop when they arrive at the centre, though nothing encounter or support them; and, if in their downward motion the impetus should carry them past the centre, they swing back again and return of themselves?"

Galileo.  Dialogo ... sopra i due Massimi Sistemi del Mondo ... (Dialogue Concerning the Two Chief World Systems).  Gio. Batista Landini, Florence, 1632.  Translated by Stillman Drake; Univ. of Calif. Press, Berkeley, 1953; pp. 22‑23, 135‑136, 227 & 236.  Asserts the object will oscillate.  No mention of air resistance.

van Etten.  1624.  Prob. 91 (88), part II (2), p. 139 (220).  Says a millstone dropped down such a hole at  1  mile per minute will take more than    days to reach the centre, where "it would hang in the air".

Ozanam.  1694.  Prob. 7, corollary 3, 1696: 218;  Prob. 7, Remark, 1708: 312.  Prob. 7, part 8, corollary 3, 1725: vol. 2, 151-152.  Considers falling down a well to the centre of the earth but uses a hypothetical constant value of  g.  Then considers a tube through the earth and says the object will oscillate, but air resistance will slow it down to rest at the centre.

Euler.  A Physical Dissertation on Sound.  1727?  ??NYS -- described by Truesdell in his Introduction to Euler's Algebra, p. xiv.  Annex announces the solution of the problem of oscillation through a hole in the earth.

Euler.  Algebra.  1770.  I.III.X.501, p. 163.  How far would an object fall in a hour under constant  g  as at the earth's surface?  39272  miles!

Euler.  Letters to a German Princess.  ??NYS -- Simoson cites an Arno Press, 1975, reprint of the 1833 ed.  Vol. 1, letter L, pp. 178-182.  Simoson quotes Euler as: "You will remember how Voltaire used to laugh at the idea of a hole reaching to the centre of the earth, ... but there is no harm is supposing it, in order to discover what would be the end results."

Ozanam‑Montucla.  1778.  Vol. IV, prob. 9, 1778: 41-42;  1803: 42-43;  1814: 34-35;  1840: 616-617.  First finds the time to reach the centre if  g  is constant, namely  19  minutes.  Then considers that gravity will decrease and quotes a result of Newton to find the time to the centre is  21' 5" 12"'.

John Baines, proposer;  Wm. Rutherford;  N. J. Andrew & George Duckett;  independent solvers.  Question (23).  The Enigmatical Entertainer and Mathematical Associate for the Year 1830; ....  Sherwood & Co., London.  No. III, 1829.  This has two separate parts with separate pagination.  The second part is The Mathematical Associate and the problem is on pp. 36-37 in the Answers to the Questions Proposed Last Year.  "If a hole were bored through the earth, parallel to the equator, in lat. 20o, and a heavy body let fall into it from the surface, it is required to determine its velocity at any point of its descent, taking into account the variation of gravity, but abstracting all resistance."

Lewis Carroll.  Alice in Wonderland.  Macmillan, London, 1865.  Chap. I, pp. 27-28 in Gardner's Annotated Alice, below.  "I wonder if I shall fall right through the earth!"

Lewis Carroll.  Sylvie and Bruno Concluded.  Macmillan, London, 1893.  Chap. 7, pp. 96‑112, esp. pp. 106‑108.  Discusses trains using straight holes, not through the centre.  Cf Carroll-Gardner, pp. 7-8, where Gardner notes that Carroll frequently uses the value 42 which is the half-period in minutes.

Martin Gardner.  The Annotated Alice.  Revised ed., Penguin, 1970.  Chap. I, note 4 (to the line given above), pp. 27‑28.  Describes Carroll's interest in the problem.  Says it interested Plutarch, Bacon and Voltaire and that it had been resolved by Galileo (see above).  Gardner also cites C. Flammarion, Strand Mag. 38 (1909) 348; ??NYS. 

Clark.  Mental Nuts.  1897, no. 53.  Matters of gravity.  "Suppose you drill a hole through the earth and drop an iron ball in it, where will the ball go?"  Answer is "Centre of earth."

Pearson.  1907.  Part II, no. 79: Dropped through the globe, pp. 130 & 207.  Says it will oscillate, but air friction will cause it to come to rest at the centre.

Ackermann.  1925.  Pp. 60‑61.  Similar to Pearson.

Collins.  Fun with Figures.  1928.  A hole through the earth, p. 203.  Says it will oscillate like a pendulum and if air is present, it will slow down and stop at the centre.

W. A. Bagley.  Puzzle Pie.  Op. cit. in 5.D.5.  1944.  No. 46: Down Under, pp. 51-53.  Various discussions of what happens to a person or a cannonball falling through the earth.  Seems to think a man would turn over so that he would be rightside up at the other side??  Says the oscillation will continue with diminishing periods (presumably meaning amplitude) until the person is stuck at the centre.  A cannonball would burn up from friction.

R. E. Green.  A problem  &  H. Martyn Cundy.  A solution.  Classroom Note 178:  Quicker round the bend!  MG 52 (No. 382) (Dec 1968) 376‑380.  Green notes the well known fact that the time to fall through a straight frictionless hole is independent of its length -- about  42  minutes.  He asks what path gives the least time?  Cundy says it is a straightforward application of the calculus of variations.  He finds a solution for  θ  as a function of  r,  in terms of  R,  the radius of the earth, and  m,  the distance of closest approach of the curve to the centre.  Also   m/R  =  1 ‑ 2a/π,  where  2a  is the central angle between the ends of the tube.  The straight through time is  π Ö(R/g).  The shortest time is  π Ö{(R2‑m2)/Rg}.

K. E. Bullen.  The earth and mathematics.  MG 54 (No. 390) (Dec 1970) 352‑353.  A riposte to Classroom Note 178, pointing out Saigey's result of c1890, that  g  increases as you start down a hole because the interior of the earth is denser than the surface.  More recent theoretical and practical work indicates  g  is essentially constant for at least the first  2000  km.

H. Lindgren.  Classroom Note 250:  Quicker round the bend (Classroom Note 178).  MG 55 (No. 393) (Jun 1971) 319‑321.  Shows the optimal curve is a hypocycloid and rephrases the time required.  If  d  is the surface distance between the ends and  C  is the earth's circumference, the minimal time is  Ö{d(C‑d)/Rg}.  He cites 1953 and 1954 papers which treat the problem in general.

Erwin Brecher.  Surprising Science Puzzles.  Sterling, NY, 1995.  Hole through the earth, pp. 20 & 88.  Asks a number of straightforward questions and then asks whether a ball will take more or less time to fall through a hole in the moon.  He says it will take about  53 min  on the moon -- I get  54.14 min.

In 2000 or 2001, Tim Rowett asked me the following.  If an apple could be dropped from a point on the earth's orbit, but only attracted by the sun, how long would it take to reach the sun?  This is complicated by gravity varying and I found it quite awkward to do, having to make some approximations to get a time of 64.4 days.  After seeing Simoson's article, below, I asked him if his work would deal with the problem and he pointed out that the formula for falling to the centre of the earth considered as a point mass (p. 349, case 3) can be used if one adapts the parameters appropriately.  After a little conversion, this is identical to the formula I obtained.  He computes  64.57 days,  or  63.89 days  to the surface of the sun, the difference being due to his using more accurate values for the astronomical constants.

Andrew J. Simoson.  The gravity of Hades.  MM 75:5 (Dec 2002) 335-350.  Considers the problem for several models of the earth.  Cites Plutarch, Halley, Euler and discusses models from Greek mythology, Dante, Hooke (a multi-layer, onion-like, earth), Halley (a hollow earth), etc.  He finds the time to the centre is about  21.2 min.  He finds the time to the centre for a constant acceleration model is about  19.0 min -- cf Ozanam and Ozanam-Montucla.  For the best known results on the earth's mass distribution, he finds about  19.2 min.  If all the earth's mass is at the centre, he gets about  15.0 min  and observes this is the model of the the earth's mass distribution which gives the fastest time (p. 349, case 3).  He then poses a problem of falling down the  z-axis to a galaxy in the  x-y plane, as Satan may have done when cast out of heaven, thereby estimating the Miltonian distance between Heaven and Hell.

 

          10.M. CELTS  =  RATTLEBACKS

 

            When rotated, these objects stop and then start turning in the opposite direction.  The word 'celt', with a soft 'c', so it sounds like 'selt', means a stone hand axe, chisel or similar primitive implement -- see 1910 below.  I once heard that the phenomenon was discovered by anthropologists examining handaxes and that they used the spin as a form of classification.

 

The OED entry for Celt is long and not definitive.  The word 'celte' appears in the Vulgate translation of the Bible and is understood to mean some kind of tool, but others feel it is a miscopying -- it seems to be 'certe' in some manuscripts.  By 1700, it was considered a proper Latin word and was adopted by British archaeologists for primitive tools.

Chambers's Encyclopedia.  Revised edition, W. & R. Chambers, London & Edinburgh, 1885.  Vol. II, p. 711.  About a column on celts.  "CELT (Lat. celtis (?), a chisel), the name by which certain weapons or implements of the early inhabitants of Western Europe are known among archaeologists."

G. T. Walker.  J. Walker says his investigations occur in old books on rotational mechanics in the chapter on asymmetrical tops.  ??NYS

G. T. Walker.  On a curious dynamical property of celts.  Proc. Camb. Philos. Soc. 8 (1892/95) 305‑306.  (Meeting of 13 May 1895.)  ??NX

G. T. Walker.  On a dynamical top.  Quarterly J. Pure & App. Mathematics 28 (1896) 175‑184.  (& diagrams??)  ??NX

Harold Crabtree.  An Elementary Treatment of the Theory of Spinning Tops and Gyroscopic Motion.  Longmans, Green & Co., London, 1909, ??NX;  2nd ed, 1914.  Pp. 7 & 54 and Plate I.  (I have a citation to these pages in the 1st ed and the material is on the same pages in the 2nd ed.  The Preface to the 2nd ed says changes occur elsewhere.)  Says the term comes from Latin, celtis, a chisel, and uses it for the stone axes in general.  Shows and discusses three examples.  The first is an ordinary axe.  The second is a two-way version -- when spun in either direction, it will reverse.  The third is a one-way version -- when spun in one direction, it continues, but when spun in the other direction, it reverses.  Cites Walker, 1896.  The plate has photos of the three examples.

Encyclopedia Britannica.  11th ed., 1910.  Entry for Celt.  ??NYS -- quoted by Bürger, below.  "CELT, a word in common use among British and French archeologists to describe the hatchets, adzes or chisels of chipped or shaped stone used by primitive man.  The word is variously derived from the Welsh celit, a flintstone (that being the material of which the weapons are chiefly made, though celts of basalt, felstone and jade are found);  from being supposed to be the implement peculiar to the Celtic peoples;  or from a Low Latin word celtis, a chisel.  The last derivation is more probably correct."

Andrew Gray.  Treatise of Gyrostatics and Rotational Motion.  (1918);  Dover, 1959.  ??NYS.

Charles W. Sherburne (3409 Patton Ave., San Pedro, California, 90731, USA).  US Design 210,947 -- Scientific demonstration toy.  Filed: 12 Nov 1995;  patented: 7 May 1968.  1p.  This simply says it 'shows my new design' and there is no indication of the skew curvature of the bottom.  Sherburne has published material claiming that the rattleback demonstrates the failure of Newton's laws and that it is the shape of Noah's Ark!

Karl Magnus.  The stability of rotations of a non-symmetrical body on a horizontal surface.  Festschrift Szabo, Berlin, 1971, pp. 19-23.  ??NYS -- cited by Bürger.  This determines which direction of rotation is stable.  I don't know if it deals with 'both-way reversing' examples.

Jearl Walker.  The mysterious "rattleback": a stone that spins in one direction and then reverses.  SA 241:4 (Oct 1979) 144‑150.  Reprinted with extra Note and recent references in:  Jearl Walker; Roundabout -- The Physics of Rotation in the Everyday World; Freeman, NY, 1985; Chap. 6, pp. 33‑38 & 66.  Cites Crabtree and G. T. Walker.  Discusses work of Nicholas A. Wheeler and of A. D. Moore.

Allan J. Boardman.  The mysterious celt.  Fine Woodworking (Jul/Aug 1985) 68‑69.  Describes how to make celts.

Hermann Bondi.  The rigid body dynamics of unidirectional spin.  Proc. Royal Soc. London A 405 (1986) 265‑274.  Analyses the dynamics and shows that the phenomenon occurs even without friction.  Only cites G. T. Walker, Quarterly J.  The Cavendish Laboratory has made a fine steel model with adjustable weights which Bondi has seen make five reversals.

Wolfgang Bürger.  A Celtic rocking top.  English version of leaflet to accompany the plastic version of the celt distributed by Nixdorf Computers.  Nd [probably late 1980s].  Cites Walker, 1896, and quotes Encyclopedia Britannica, 11th ed., 1910, for the term as quoted above.  He conjectures that the spinning property may have been discovered by an archaeologist and he gives a myth that such spinning was used by ancient priests to determine guilt or innocence.  (Frame-ups were common even then.)  Until recent realization that these objects were man-made, they were the subject of superstitions throughout the world.  He gives a short discussion of the physics/geometry involved and says that since 1980, nine scientific papers have tried to analyse the motion and that it was the subject of a German Jugend forscht (Young Researchers) prize winning project in 1985.

 

          10.M.1.        TIPPEE TOPS

 

          The physics of this is hard and I will only give some general articles.

 

Helene Sperl.  German Patent 63261 -- Wendekreisel.  Patented: 7 Oct 1891;  published: 12 Jul 1892.  1p + 1p diagrams.  Several slightly diferent shapes.  Diagram is reproduced by Kuypers & Ucke, below.

Harold Crabtree.  An Elementary Treatment of the Theory of Spinning Tops and Gyroscopic Motion.  Longmans, Green & Co., London, (1909, ??NX);  2nd ed, 1914.  Pp. 4-5 describes some rising tops.  One of these is a loaded sphere, which seems to have the same inverting properties as the Tippee Top.  Appendix IV: The rising of a spinning top, pp. 145-158, is new in the 2nd ed and discusses the loaded sphere on pp. 155-156. 

Gwen White.  Antique Toys and Their Background.  Op. cit. in 5.A.  1971.  P. 45.  "An interesting little top known as a 'tippe top' came in 1953 ....  It was a great commercial success for the British Indoor Pastimes Company."

D. G. Parkyn.  The inverting top.  MG 40 (No. 334) (Dec 1956) 260‑265.  Cites 4 papers in 1952.

Leslie Daiken.  Children's Toys Throughout the Ages.  Spring Books, London, 1963.  [This may be a reprint of an earlier publication??]  P. 38.  "... the most recent craze, invented by a Swede.  Made from plastic, and known as the Tippy Tap [sic], this type will turn upside down and spin on its head!"

Jearl Walker.  Amateur Scientist columns.  SA (Oct 1979) -- op. cit. in 10.M  and  The physics of spinning tops, including some far‑out ones.  SA 244:3 (Mar 1981) 134‑142.  Reprinted with extra notes in:  Roundabout, op. cit. in 10.M, chap. 6 & 7, pp. 33‑44 & 66, esp. pp. 37‑44.  15 references on p. 66, not including Parkyn.

Wolfgang Bürger.  Elementary dynamics of simple mechanical toys.  Mitteil. der Ges. f. Angew. Math. und Mechanik 2 (Jul 1980) 21-60.  (Reproduced in: Spielzeug-Physik; Bericht Nr. 98, Akademie für Lehrerfortbildung Dillingen, 1986, pp. 159-199.)  Pp. 49‑52 (= 188-191) discusses the Tippe Top, noting that Fraülein Sperl's explanation in her patent 'was far from being correct.'  Says the dynamics is difficult and cites 1952 & 1978 articles on it.  J. L. Synge asserted that friction was not necessary for the turning over motion, but Bürger shows it is essential.

Friedhelm Kuypers & Christian Ucke.  Steh' auf Kreisel!  Physik in unserer Zeit 25:5 (Sep 1994) cover & 214-215.  The German names are Stehaufkreisel and Kippkreisel.  Describes Sperl's patent and modern work on the mechanics involved -- "it is not so easy as we first believed".

 

          10.N. SHIP'S LADDER IN RISING TIDE

 

            Water is touching the bottom rung of a rope ladder hanging over the side of a ship.  The tide is rising at a known rate.  How many rungs will be covered after some time?

 

Phillips.  Brush.  1936.  Prob. O.1: The ship's ladder, pp. 49 & 106.

Haldeman-Julius.  1937.  No. 42: The Queen Mary problem, pp. 7 & 22.

Morley Adams.  The Children's Puzzle Book.  Faber, London, 1940.  Prob. 174: The ship's ladder, pp. 55 & 78.

Shirley Cunningham.  The Pocket Entertainer.  Blakiston Co., Philadelphia, and Pocket Books, NY, 1942.  Prob. VI: The rope ladder, pp. 72 & 222.

Harold Hart.  The World's Best Puzzles.  Op. cit. in 7.AS.  1943.  The rope ladder, pp. 31 & 59.

Leopold.  At Ease!  1943.  Of time and tide, pp. 9-10 & 195.

Sullivan.  Unusual.  1943.  Prob. 3: Time the tide.

Leeming.  Op. cit. in 5.E.  1946.  Chap. 3, prob. 40: Rising tide, pp. 36 & 162.

 

          10.O. ERRONEOUS AVERAGING OF VELOCITIES

 

            See also 7.Y which involves erroneous averaging of unit costs.  There must be earlier examples than I have here.

 

H. A. Ripley.  How Good a Detective Are You?  Frederick A. Stokes, NY, 1934, prob. 42: Class day.  Average  10 mph  and  50 mph  over the same distance.

Dr. Th. Wolff.  Die lächelnde Sphinx.  Academia Verlagsbuchhandlung, Prague, 1937.  Prob. 17, pp. 190 & 201.  Going and returning at  100  km/hr versus going at  120  and returning at  80.

Harriet Ventress Heald.  Op. cit. in 7.Z.  1941.  Prob. 31, pp. 15‑16.  Man goes  30  mph for a mile.  How fast must he go for a second mile in order to average  60  mph overall?

Sullivan.  Unusual.  1943.  Prob. 22: Don't get caught trying it.  If you are going  60  mph, how much faster do you have to go to save a minute on each mile?

E. P. Northrop.  Riddles in Mathematics.  1944.  1944: 11-13;  1945: 10‑12; 1961: 20-22.  Same as Heald with rates  15, 30.  Relates to airplane going with and against the wind.

Leeming.  Op. cit. in 5.E.  1946.  Chap. 5, prob. 20: Sixty miles per hour, pp. 59‑60 & 178.  Does  30  mph for  10  miles.  How fast for the next  10  miles to average  60  mph?

Sullivan.  Unusual.  1947.  Prob. 31: A problem without a title.  Same as Heald.

Birtwistle.  Calculator Puzzle Book.  1978.  Prob. 40: Motoring problem, pp. 30 & 87-88.  Same as Heald with rates  15, 30.

 

          10.P.  FALSE BALANCE

 

Solomon (or The Preacher).  c-960.

Proverbs 11:1.  "A false balance is abomination to the Lord:  But a just weight is his delight."

Proverbs 16:11.  "A just weight and balance are the Lord's:  All the weights of the bag are his work."

Proverbs 20:10.  "Divers weights, and divers measures,  Both of them are alike abomination to the Lord."

Proverbs 20:23.  "Divers weights are an abomination unto the Lord:  And a false balance is not good."

Hosea 12:7.  "He is a chapman, the balances of deceit are in his hand."

Aristotle.  Mechanical Questions.  c-340.  ??NYS -- cited by van Etten.

Pappus.  Collection.  Book 3.  c320.  ??NYS - cited by W. Leybourn.

Muhammed.  Koran.  c630.  Translated by J. M. Rodwell, Everyman's Library, J. M. Dent, 1909.  Sura LXXXIII -- Those who stint: 1-3.  "Woe to those who STINT the measure:  Who when they take by measure from others, exact the full;  But when they mete to them or weigh to them, minish --".  (I saw the following version on a British Museum label, erroneously attributed to Sura LXXX:  "Woe be unto those who give short measure or weight.")

Cardan.  De Subtilitate.  1550, Liber I, ??NYS.  = Opera Omnia, vol. IV, p. 370: Modus faciendi librā, que pondera rerum maiora quàm sunt[?? -- nearly obliterated in the text I have seen] ostendat.  Describes a scale with arm divided  11 : 12.

John Wecker.  Op. cit. in 7.L.3.  (1582), 1660.  Book XVI -- Of the Secrets of Sciences: Chap. 20 -- Of Secrets in Arithmetick: Fraud in Balances where things heavier shall seem to be lighter, p. 293.  Says such fraud is mentioned by Aristotle.

van Etten.  1624.  Prob. 54 (49), pp. 49‑50 (73‑74).  Mentions Aristotle's mechanical questions and cites Archimedes' law of the lever.  Discusses example with arm lengths  12 : 11

W. Leybourn.  Pleasure with Profit.  1694.  Tract. IV, pp. 2-3.  Cites Solomon and Pappus' Collections, Book 3.  Discusses arm lengths  11 : 10.

Ozanam.  1694.  Prob. 4, 1696: 275-276 & fig. 131, plate 46.  Prob. 4 & fig. 26, plate 14, 1708: 351‑352.  Vol. II, prob. 7, 1725: 339‑340 & fig. 131, plate 46.  Vol. II, prob. 3, 1778: 4-5;  1803: 4-6;  1814: 3-5;  1840: 196-197.  Construct a balance which is correct when empty, but gives dishonest weight.  This can be detected by interchanging the contents of the two pans.  Hutton notes that the true weight is the geometric mean of the two weights so obtained, and that this is close to the average of these two values.  Illustrates with weights  16  and  14.  The figure is just a picture of a balance and is not informative -- the same figure is also cited for various sets of weights.

Vyse.  Tutor's Guide.  1771?  Prob. 1, 1793: p. 303; 1799: p. 316 & Key p. 356.  Cheese weighs  76  in one pan and  56  in the other.  States the general rule with no explanation.

Jackson.  Rational Amusement.  1821.  Curious Arithmetical Questions.  No. 8, pp. 16 & 72.  Cheese weighs  16  on one side and  9  on the other.  Says the answer is the mean proportional.  = Illustrated Boy's Own Treasury, 1860, prob. 12, pp. 428 & 431. 

Julia de Fontenelle.  Nouveau Manuel Complet de Physique Amusante ou Nouvelles Récréations Physiques ....  Nouvelle Édition, Revue, ..., Par M. F. Malepeyre.  Librairie Encyclopédique de Roret, Paris, 1850.  Pp. 407-408 & fig. 146 on plate 4 (text erroneously says V): Balance trompeuse.  A bit like Ozanam, but doesn't indicate the true weight is the geometric mean.  Figure copied from Ozanam, 1725.

Magician's Own Book.  1857.  The false scales, p. 251.  Cheese weighs  9  on one side and  16  on the other.  Says the true weight is the mean proportional, hence  12  here.  = Book of 500 Puzzles, 1859, p. 65.  = Boy's Own Conjuring Book, 1860, p. 223.  Almost identical to Jackson.

Hoffmann.  1893.  Chap. IV, no. 94: The false scales, pp. 169 & 229 = Hoffmann-Hordern, p. 154.  On one side a cheese weighs  9  and on the other it weighs  16.  Answer notes that the true weight is always the geometric mean.  Almost identical to Jackson and Magician's Own Book.

Clark.  Mental Nuts.  1904, no. 66;  1916, no. 84.  The grocer puzzled.  Weights of  8  and  18.  Answer says to solve  8 : x :: x : 18.

Briggs & Bryan.  The Tutorial Algebra -- Part II.  Op. cit. in 7.H.  1898.  Exercises X, pp. 125 & 580.  Weighing one way gains an extra  11%  profit, but weighing the other way gives no profit at all.  What is the legitimate profit?

Pearson.  1907.  Part II, no. 72, pp. 128 & 205.  Same problem as Hoffmann.  Answer says to take the square root of  9 x 16.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927.  Prob. 1, p. 98 & Answers p. 10: The druggist's balance.  One arm is longer than the other, but the shorter is weighted to give balance when the scales are empty.  He uses the two sides equally -- does he gain or lose (or come out even)?  If the lengths are  L  and  l,  then balancing against weight  W  will result in  WL/l  and  Wl/L  equally often and the arithmetic mean of these is greater than their geometric mean of  W,  so the druggist is losing.  Answer only does the example with  L = 2l.

Loyd Jr.  SLAHP.  1928.  The jeweler's puzzle, pp. 21‑22 & 90.  More complex version.

Kraitchik.  La Mathématique des Jeux.  Op. cit. in 4.A.2.  1930.  Chap. II, prob. 26, p. 34.  = Mathematical Recreations; op. cit. in 4.A.2; 1943; Chap. 2, prob. 54, pp. 41‑42.  Merchant has a false balance.  He weighs out two lots by using first one side, then the other.  Is this fair on average?

 

          10.Q. PUSH A BICYCLE PEDAL

 

            Holding a bicycle upright, with the pedals vertical, push the bottom pedal backward.  What happens?

 

Pearson.  1907.  Part II, no. 17: A cycle surprise, pp. 14 & 189.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927.  P. 99 & Answers p. 11.  Pull the bottom pedal forward.  What happens?  What is the locus of the pedal in ordinary travel?  Answer says it is a cycloid but it is actually a curtate cycloid.

W. A. Bagley.  Paradox Pie.  Op. cit. in 6.BN.  1944.  No. 60: Another cycling problem, p. 46.  Uses a tricycle -- which simplifies the experimental process.

David E. Daykin.  The bicycle problem.  MM 45:1 (Jan 1972) 1.  Short analysis of this 'old problem'.  No references.

 

          10.R. CLOCK HAND PROBLEMS

 

            These are questions as to when the hands can meet, be opposite, be interchangeable, etc.  There are also questions with fast and slow clocks.  Many examples are in 19C arithmetic and algebra books and in Loyd, Dudeney, etc.  I am somewhat surprised that my earliest examples are 1678? and 1725, as clocks with minute hands appeared in the late 16C.  The problems here are somewhat related to conjunction problems -- see 7.P.6. 

            See also 5.AC for digital clocks, which are a combinatorial problem.

            These are related to Section 7.P.6.

 

Wingate/Kersey.  1678?.  Quest. 20, p. 490.  Clock with an hour hand and a day hand, which goes round once every 30 days.  They are together.  When are they together again?  In 30 days, the faster hand must pass the slower 59 times, so the time between coincidences is  30/59  of a day.

Ozanam.  1725.  Prob. 11, question 3, 1725: 76‑77.  Prob. 2, 1778: 75-76;  1803: 77-78;  1814: 69;  1840: 37.  When are the hands together?  1725 does it as a geometric progression, like Achilles and the tortoise.  1778 adds the idea that there are  11  overtakings in  12  hours, but this does not appear in the later eds.

Les Amusemens.  1749.  Prob. 122, p. 264.  When are clock hands together?

Vyse.  Tutor's Guide.  1771?  Prob. 7, 1793: p. 304; 1799: p. 317 & Key pp. 357-358.  When are the hands together between  5  and  6  o'clock?

Dodson.  Math. Repository.  1775.

P. 147, Quest. CCXXXIV.  When are hour and minute hands together?

P. 147, Quest. CCXXXV.  When are hour, minute and second hands together?

Charles Hutton.  A Complete Treatise on Practical Arithmetic and Book-keeping.  Op. cit. in 7.G.2.  [c1780?]  1804: prob. 35, p. 136.  When are watch hands together between  4  and  5?

Pike.  Arithmetic.  1788.  P. 352, no. 27.  When are the clock hands together next after noon?

Bonnycastle.  Algebra.  1782.

P. 86, no. 22.  When are the hands next together after  12:00?

P. 201, no. 1 (1815: p. 226, no. 1).  When are the hands together between  8:00  and  9:00?

Eadon.  Repository.  1794.  P. 195, no. 13.  When do the hands meet after noon?

Hutton.  A Course of Mathematics.  1798?  Prob. 36,  1833: 223;  1857: 227.  When do the hands next meet after noon?

Robert Goodacre.  Op. cit. in 7.Y.  1804.  Miscellaneous Questions, no. 128, p. 205 & Key p. 270.  When are clock hands next together after  12:00?

Silvestre François Lacroix.  Élémens d'Algèbre, a l'Usage de l'École Centrale des Quatre-Nations.  14th ed., Bachelier, Paris, 1825.  Section 82, ex. 6, p. 122.  When are clock hands together after noon?  Notes that this is related to overtaking problems.

Bourdon.  Algèbre.  7th ed., 1834.  Art. 57, prob. 10, p. 85.  When are clock hands together?

D. Adams.  New Arithmetic.  1835.  P. 244, no. 83.  When are hands next together after 12:00?  Observes that the minute hand gains  11  spaces (i.e. hour marks) per hour on the hour hand.

Augustus De Morgan.  Examples of the Processes of Arithmetic and Algebra.  Third, separately paged, part of: Library of Useful Knowledge -- Mathematics I, Baldwin & Craddock, London, 1836.  P. 90.  How much time elapses between times when the hands are together?  Suppose one hand revolves in  a  hours, the other in  b  hours -- how long between conjunctions?  In the latter case, how long does it take the faster hand to gain  p/q  of a whole revolution on the other?

Hutton-Rutherford.  A Course of Mathematics.  1841?

Prob. 11,  1857: 81.  When are clock hands next together after 12:00?

Prob. 19,  1857: 82.  When are clock hands together between 5 and 6 o'clock?

T. Tate.  Algebra Made Easy.  Op. cit. in 6.BF.3.  1848.

Pp. 46-47, no. 19.  When are the hands together between  2:00  and  3:00?

P. 46, no. 20.  Same between  3:00  and  4:00?

Anonymous.  A Treatise on Arithmetic in Theory and Practice: for the use of The Irish National Schools.  3rd ed., 1850.  Op. cit. in 7.H.

P. 356, no. 4.  When are hands together between  5  and  6?

P. 360, no. 46.  When are hands together between  1  and  2?

Dana P. Colburn.  Arithmetic and Its Applications.  H, Cowperthwait & Co., Philadelphia, 1856.  Miscellaneous Examples.

No. 12, p. 359.  When are hands together after  12:00?

No. 13, p. 359.  When are hands opposite after  12:00?

No. 32, p. 361.  When is the hour hand as much after  12  as the minute hand is before  1?

Todhunter.  Algebra, 5th ed.  1870. 

Examples X, no. 35, p. 87 & 577.  Between  1:00  and  2:00,  when is the minute hand one minute in front of the hour hand?

Miscellaneous Examples, no. 17, pp. 546 & 604.  When are clock hands together between 9:00 and 10:00?

Colenso.  Op. cit. in 7.H.  These are from the (1864), 1871 material.

No. 3, p. 186.  When are hour and minute hands at right angles, or coincident, between  5:00  and  6:00?

No. 21, pp. 190 & 215.  When are hour and minute hands  162o  apart between  11:00  and  12:00?

Problem 203.  Zeitschrift für mathematische und naturwissenschaftlichen Unterricht 15 (1874) 197-198.  Interchanging clock hands.  Taken from Journ. élém., which must be a French journal, but I don't recognise it.  Shows that interchanging hands gives a legal appearance when the hour hand is at  60k/143 minutes,  k = 0, ..., 143.  Copy sent by Heinrich Hemme.

William J. Milne.  The Inductive Algebra ....  1881.  Op. cit. in 7.E.

No. 95, p. 140.  When are hands together after  2:00?

No. 96, pp. 140 & 332.  When are hands together after  5:00?

No. 97, pp. 141 & 332.  When are hands together after  8:00?

No. 98, pp. 141 & 332.  When are hands at  180o  after  4:00?

No. 99, pp. 141 & 332.  When are hands at  180o  after  5:00?

No. 100, pp. 141 & 332.  When are hands at  90o  after  6:00?

No. 101, pp. 141 & 332.  When are hands at  90o  after  8:30?

No. 104, pp. 303 & 347.  When are hands at  180o  after  11:00?

Carroll-Wakeling.  1888 to 1898.  Prob. 14: Looking-glass time, pp. 17 & 66-67.  This was on one of the undated typed sheets Carroll sent to Bartholomew Price.  "A clock face has all the hours indicated by the same mark, and both hands the same in length and form.  It is opposite to a looking-glass.  Find the time between 6 and 7 when the time as read direct and in the looking-glass shall be the same."  This seems to be the first example of looking at the mirror image of a clock.  This can also occur by reading a transparent clock from the wrong side.  Mentioned in Carroll-Gardner, p. 53.  Wakeling has sent me a copy of the typescript.

Anon.  Prob. 80.  Hobbies 31  (No. 795) (7 Jan 1911) 350  &  (No. 798) (28 Jan 1911) 412.  Three pendulum clocks, supposed to have one second swings.  One gains 5 minutes per day, the second loses ten minutes per day and the third is correct.  Assuming we start all three pendula at the same time at the same end of their swing, when will they all be together at this point again?  Answer: 9 min 36 sec  = 1/150 of a day.

Ernest K. Chapin.  Loc. cit. in 5.D.1.  1927.  Prob. 5, p. 89 & Answers p. 8.  Can the three hands make equal angles?  [Last line of the answer has slipped into the next column.]

W. B. Campbell, proposer;  W. E. Buker, solver.  Problem ??.  AMM 41 (Sep 1934) 447  &  42 (Feb 1935) 110-111.  Interchanging clock hands.  ??NYS - information sent by Heinrich Hemme.

Perelman.  1937.  MCBF.  Interchanging the hands of a clock, prob. 141, pp. 235-238.  Says the following was posed to Einstein by his friend A. Moszkowski:  at what times can one exchange the hour and minute hands and get a valid position of the hands?  There are  143  solutions, of which  11  are the positions where the hands coincide, which Perelman discusses as prob. 142, pp. 238-239.

Sullivan.  Unusual.  1943.  Prob. 1: Watch and see.  How many places do hands meet?

Anonymous.  Problems drive, 1957.  20 (Oct 1957) 14-17 & 29-30.  No. 10.  B reckons A's clock gains one minute per day.  C reckons B's clock gains one minute per day.  They synchronise watches.  Assuming rates are constant, when do all three agree again?

Pierre Berloquin.  The Garden of the Sphinx, op. cit. in 5.N.  1981.

Prob. 11: What coincidences?, pp. 10 & 91.  When are the hour, minute and second hands together?

Prob. 12: Meetings on the dot, pp. 10 & 92.  At what other times are the three hands closest together?

Jamie & Lea Poniachik.  Cómo Jugar y Divertirse con su Inteligencia; Juegos & Co. & Zugarto Ediciones, Argentina & Spain, 1978 & 1996.  Translated by Natalia M. Tizón as: Hard-to-Solve Brainteasers.  Ed. by Peter Gordon.  Sterling, NY, 1998.  Pp. 14 & 70, prob. 18: What time is it -- IV.  Hour hand is on a minute mark and the minute hand is on the previous mark.

??, proposer; Ken Greatrix & John Bull, separate solvers, with editorial note by ADF [Tony Forbes].  Problem 168.1 -- Clock.  M500 168 (?? 1999) & 170 (Oct 1999) 11-14.  Hour hand is 3" long; minute hand is 4".  When are the ends 'travelling apart at the fastest speed'?  Bull notes that the problem can be interpreted in two ways.  A: 'When is the relative velocity of the ends the greatest?', which occurs when the hands are at 180o.  B: 'When is the distance between the ends increasing most rapidly?', which occurs when the angle between the hands has cosine of  3/4.  Greatrix did the first case in general, finding the maximum occurs when the cosine equals  h/m,  where  h, m  are the lengths of the hour and minute hand.

John Conway.  Naming and understanding the PENTAgonal polytopes; When are the three hands equally spaced?  4pp, handwritten, Handout at G4G5, 2002.  Basically the hands are never equally spaced, but they are close at  9:05:25  (and at the mirror image time).  When are they closest?  This depends on how you measure 'closeness'.  He finds there are just five reasonable answers within a second of this time.  There is a sixth answer but it is nearly two seconds away.  The most interesting solution is at  9 hr 5 min 25/59 sec  when the angle between the minute and second hands is  120o,  the angle between the second and hour hands is  120o + ε  and the angle between the hour and minute hands is  120o - ε,  where  ε = 10¢ 10² 10 ....

 

          10.S.  WALKING IN THE RAIN

 

W. A. Bagley.  Paradox Pie.  Op. cit. in 6.BN.  1944.  No. 93: Better take a brolly, p. 61.  Says the faster you go, the wetter you get.

A. Sutcliffe.  Note 2271:  A walk in the rain.  MG 41 (1957) 271‑272.

C. O. Tuckey.  Note 2384:  A walk in the rain [Note 2721].  MG 43 (No. 344) (May 1959) 124‑125.

M. Scott.  Nature.  ??NYS -- reported in This England 1965-1968, p. 70.  "When walking into the rain one should lower the head and walk as fast as possible.  When the rain is coming from behind one should either walk forward leaning backwards, or backwards leaning forwards, at a deliberate pace."

David E. Bell.  Note 60.21:  Walk or run in the rain?  MG 60 (No. 413) (Oct 1976) 206-208.  "... keep pace with the wind if it is from behind; otherwise, run for it."

D. R. Brown.  Answer to question.  The Guardian (2 Apr 1993) 13.  Cites Bell.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995.

Chap. 4, prob. 5: Rainy day on the carousel, pp. 27 & 113.  How to hold your umbrella when riding on a carousel in vertical rain.

Chap. 4, prob. 9: The May Day parade, pp. 31 & 116-117.  How to point a cannon on a truck so that vertical rain does not touch the sides of the barrel.

Chap. 4, prob. 11: Boris and the wet basketball -- reference frames and fluxes, pp. 33 & 118-120.  Sliding a sphere in the rain, it gets wetter faster, the faster it goes.  However the total wetness is reduced by going faster because it takes less time to get home.

Chap. 4, prob. 12: Chicken feed, pp. 34 & 120.  Boris is wheeling a wheelbarrow with a bucket of chicken feed in it when it starts to rain vertically.  Assuming the bucket is level, should he run?  The flux of rain entering the bucket is constant, so running reduces the time in the rain and hence the amount of rain which gets into the bucket.

Chap. 4, prob. 13: Fluxes and conservation laws (or it always helps to run in the rain), pp. 34 & 121-122.  Is there a contradiction between the rate results in prob. 11 and prob. 13?  No, because the effective cross sections behave differently.

Erwin Brecher  &  Mike Gerrard.  Challenging Science Puzzles.  Sterling, 1997.  [Reprinted by Goodwill Publishing House, New Delhi, India, nd [bought in early 2000]].  Pp. 38 & 76‑77: Lingering in the Rain.  Starts with no wind and wearing a hat, but says he ran home and the rain stopped when he got there.  So the wetness is due to moving into the rain and he says it makes no difference if the rain continues, but since the rain stopped, walking would have been drier.  Then says that if he has no hat, his head gets wetter the longer he is in the rain.  Then says if there is a head wind, one gets wetter the longer one is out, but if there is a tail wind, the best strategy is to move as fast as the wind.

 

          10.T.  CENTRIFUGAL PUZZLES

 

Fred Swithenbank.  UK Patent 11,801 -- An Improved Toy or Puzzle, applicable also as an Advertising Medium.  Applied: 21 May 1913;  completed: 21 Nov 1913;  accepted: 1 Jan 1914.  2pp + 1p diagrams.  Shows two ball version and four ball round version.

I have acquired an example in wood with a celluloid(?) top with the following printed on the sides.   'Zebra' Grate Polish.  Get one ball in each hole at the same time.   'Brasso' Metal Polish.  Patent No. 11801/13.  I have seen others of similar date advertising:  "Swan" Pens / "Swan" Inks;  Hoffmann Roller Bearings.  Both have the same instruction and patent no.

Slocum says that he has the same Zebra/Brasso version and another version from R. Journet, called Spoophem, with the same patent number and that a round version "The Balls in the Hole Puzzle" is in Gamage's 1913 catalogue.  He cites Maxwell's patent, below.

James Dalgety.  R. Journet & Company.  A Brief History of the Company & its Puzzles.  Published by the author, North Barrow, Somerset, 1989.  P. 13 mentions the Spoophem puzzle, patented in 1913.

Slocum.  Compendium.  Shows 4 ball Centrifugal Puzzle from Johnson Smith 1919 & 1929 catalogues.  Shows 2 ball Spoophem Puzzle from the latter catalogue.

Western Puzzle Works, 1926 Catalogue.  No. 72: Four Ball Puzzle.

William R. Maxwell.  US Patent 1,765,019 -- Ball puzzle device.  Filed: 6 Apr 1929;  patented: 17 Jun 1930.  2pp + 1 p diagrams.  Shows 2-ball and 4-ball versions.

L. Davenport & Co. and Maskelyne's Mysteries.  New Amusement Guide [a novelties catalogue].  Davenport's, London, nd [Davenport's identifies this as being in the period 1938-1942], p. 6, shows a Boat Puzzle, which is a 2 ball version..

 

          10.U. SHORTEST ROUTE VIA A WALL, ETC.

 

            I have just added this.  I imagine there are ancient versions of this problem.  See also 6.M and 6.BF.3 for some problems which use the same reflection principle.  Basically, this will cover two-dimensional cases where the reflection process is very physical.

 

Ozanam‑Montucla.  1778.  Prob. 24 & fig. 36, plate 5, 1778: 305;  1803: 300-301;  1814: 256;  1840: 130.  Run from  A  to  B,  touching the wall CD.

Ozanam‑Montucla.  1778.  Vol. II.  Prob. 35: Du jeu de billard & figs. 34 & 35, plate 7, 1778: 58-62;  1803: 63-66;  1814: 52-55;  1840: 222-224.  Use cue ball to hit another after hitting one or two cushions.

Birtwistle.  Math. Puzzles & Perplexities.  1971.  Playground circuit, pp. 142-143.  In a rectangular playground with a point marked in it, what is the shortest route from the point to all four walls and back to the point?  He draws it and says it's a parallelogram, but doesn't see that the length is twice the diagonal of the rectangle, independently of the starting point.

 

          10.V. PICK UP PUZZLES  =  PLUCK IT

 

            I have recently added these.  The puzzle comprises a ball which is about half into a hole and the object is to remove it without moving the hole.  It can be done with the fingers, but one can also use the Bernoulli or Venturi effect!

 

Western Puzzle Works, 1926 Catalogue.  No. 11: Pick up ball.  Negro head.

 

          10.W. PUZZLE VESSELS

 

            S&B 140-141 gives an outline of the ancient history.  Vases which fill from the bottom were found in Cyprus, by General Louis Palma di Cesnola, who was American Consul there for ten years from about 1865.  He discovered the Treasure of Curium.  He shipped back a large amount which forms the Cesnola Collection at the Metropolitan Museum of Art, NY.  Such a vase is named an 'askos' by archaeologists.

            There are examples or descriptions of such jugs in ‑5C Greece, in the works of Hero and Philo of Alexandria in 1C, and in 13C France.  From the 16C, they are common throughout Europe and China, and the bottom filling 'Cadogan' teapot goes back to perhaps c1000 in China.  I have now looked at some of the literature and have realised that this topic covers a much larger range of types than I had initially thought.  In Banu Musa, there are about a hundred types.  Some of these are common in later works, e.g. van Etten, Ozanam, etc. have casks which pour different liquids from the same spout.  I don't know if I will try to include all these later versions.  Archaeologists refer to these as 'trick vessels' or 'trick vases'.

            A Tantalus, greedy, justice, temperance or dribble cup has a siphon such that if it is filled above a certain point, the siphon drains the cup, usually into the drinker's lap!

            In response to an exchange on NOBNET in Feb 1999, Peter Rasmussen kindly sent a page and a half of bibliography on Chinese puzzle vessels.  A number of his entries are auction, sale or collection catalogues which apparently only show one or two items, so I won't list them here, but I have included the more general books at the end of this section.

            Norman Sandfield has sent a draft copy of his A Monograph on Chinese Ceramic Puzzle Vessels (Antique, Vintage and Contemporary) Featuring a Classification System and Descriptive Inventory of more than 60 different puzzle vessels, dated 20 Jan 2000, with 3pp of bibliographical material and a 4pp version of this section, 25pp in all.  He also sent further material: 3pp of time line and a 7pp extended version of his bibliographical material incorporating this section, dated 27 Jan 2000.  He has more recently sent An Annotated Bibliography on Chinese and Non-Chinese Puzzles Vessels (Over 155 records) and Museums with Puzzle Vessels (over 30 Museums), 25pp, dated 4 Jun 2002.  I will not try to copy all these references.  Most Chinese examples are 'Cadogan' wine pots (= Magic Wine Pot) or 'Tantalus' Cups (= Justice Cups -- this is the literal translation of the Chinese name).  The Chinese did not adopt teapots until the Ming Dynasty (1368-1644) and they then adapted the Cadogan style to teapots.  The Justice Cup traditionally has a base to hold the spillage, sometimes inside it as a mystery to the drinker.  Sandfield feels there is a difference between the ordinary Justice Cup and the Dribble Cup, but it may just be a difference in the outlet hole size.  (The modern Dribble Glass has cut decorations, one or several of which go through the glass so that it dribbles when you tilt it to drink, but this is unrelated to the present topic.)

 

                    CYPRUS

 

General Louis Palma di Cesnola was the American Consul in Cyprus for ten years from about 1865 (his commission was one of the last documents signed by Lincoln).  He collected antiquities enthusiastically and discovered the Treasure of Curium.  He shipped a large amount back to the Metropolitan Museum of Art in New York.  He describes his adventures and collecting in: Cyprus: Its Ancient Cities, Tombs and Temples; New York, 1877; but there is no mention of puzzle vessels there.  He also wrote  A Descriptive Atlas of the Cesnola Collection of Cypriote Antiquities in the Metropolitan Museum of Art -- ??NYS, vol. 2 appeared in 1894.  Karageorghis cites pl. XCV.813 for Myres 518.

John L. Myres.  Handbook of the Cesnola Collection of Antiquities from Cyprus.  The Metropolitan Museum of Art, New York, 1914.  Items 518, 519, 930, 931, pp. 67-68 & 113.  The first two and the last are illustrated with photos.  The first two are described as 'animal-headed vases' from the Early Iron Age (c‑1000, with the second being later than the first) and their trickiness is only mentioned in the index.  The latter two are in a group of Spout Vases with Modelled Head and in these two examples the woman's head "blocks the apparent neck of the vase, and the real opening, through which the vase is filled, is inside the foot, which communicates with a long tube running up the middle of the inside of the vessel after the manner of the neck of a modern safety ink-pot.  These trick vases are not common, ....  Both these examples are as early as the sixth century, and 931 may go back to the seventh."  (These are centuries BC.)  These were fairly common at Kourion (= Curium) and at Marion, but rare in other parts of Cyprus.

Einar Gjerstad.  The Swedish Cyprus Expedition: Vol. IV, Part 2: The Cypro-geometric, Cypro-archaic and Cypro-classical Periods.  The Expedition, Stockholm, 1948.  This has several sections - the drawing numbers refer to the section Pottery Types, drawn by Bror Millberg, following p. 545, labelled Fig. I, Fig. II, ....  The individual items on the page have two numbers, one being its number on the page, e.g. 1), 2), ..., at the upper left of the item; the other being the number within its category, e.g. 1, 2, ..., below the item -- I will use the first.  The sources of the pots drawn are in a preliminary section with capital Roman numeral page numbers.  The author is interested in pottery styles and rarely gives dates or even periods.  There is a Relative Chronology for pottery on pp. 186-206 and an Absolute Chronology on pp. 421‑427.  Date ranges below are deduced from these chronologies.

                    Pp. 52-53: White Painted II Ware (-950/-850), says that the askos, Fig. XV, 3, in the Cyprus Museum, B. 1933, is a trick vessel.  As with several other entries, it says to see Fig. XXXVI, 9, which is Myres 519, with a cross-sectional view showing how it works.

                    P. 60: Bichrome I Ware (-1050/-850), says that the askos, Fig. VIII, 24 (= Myres 518), at the Metropolitan, is a trick vase and says to see Fig. XXXVI, 9.

                    Pp. 62-64: Bichrome IV Ware (-600/-475), says that the askos, Fig. XXXVI, 9 (= Myres 519), at the Metropolitan, is a trick vase.

                    Pp. 69-71: Black-on-Red II (IV) Ware (-700/-475), says that 'the spout-jug with the neck in the shape of a human protome', Fig. XXXIX, 15, in the BM (Brit. Mus. Cat. Vases I:2, C 882), is a trick vase and says to see Fig. XXXVI, 9,

                    P. 73: Bichrome Red I (IV) Ware (-700/-600), says the spout-jug with human protome, Fig. XLII, 5 (= Myres 931) in the Metropolitan is a trick jug similar to the item above.  He also says that the askos, Fig. XLII, 9 in the BM (Corp. Vas. Ant., Great Britain 2, Brit, Mus. 2, II Cc, Pl. 10.24. (Gr. Brit. 63) 'has a low foot, but apart from that its shape is identical with the corresponding type of Black-on-Red II (IV).  It is a trick vase (cf. Fig. XXXVI, 9)."  However, the material on Black-on-Red II (IV) is on pp. 69‑71 and it does not explicitly say there that the askos, Fig. XXXIX, 19 in the BM (Corp. Vas. Ant., Great Britain 2, Brit, Mus. 2, II Cc, Pl. 13.12. (Gr. Brit. 57)) is a trick vase, but says the remaining items are similar to Bichrome IV Ware and that seems to imply this is, like XXXVI, 9, a trick vase.

Vassos Karageorghis.  Ancient Art from Cyprus  The Cesnola Collection in the Metropolitan Museum of Art.  The Museum, New York, 2000.  P. 80, #126: Trick vase in the form of an askos.  = Myres 518  = Gjerstad Fig. VIII, 24.  A bottom-filling (Cadogan) pot with a goat's head from c-1050/c-950, with colour picture.  He says it is Cypro-Geometric I, White Painted I Ware and is a "trick vase".

 

 

Joseph Veach Noble.  Some trick Greek vases.  Proc. Amer. Philosophical Soc. 112:6 (Dec 1968) 371-378.  He describes and illustrates a number of types of trick vases from Athens:  false bottoms (-4C);  dribble vases (which dribble wine over the guest) (-5C);  a covered drinking cup which fills from the bottom like a Cadogan teapot (‑5C);  vases with concealed contents which pour when the host releases fingers from holes (-5C).  In each case he cites museums and extended descriptions.  One reference gives a list of covered cups.

Jasper Maskelyne.  White Magic.  Stanley Paul, London, nd [c1938], p. 110.  Discusses puzzle pitchers, saying  "On the site of a Greek temple at Athens, excavators a few years ago discovered a magic pitcher which was famous in Greek legend in trials for witchcraft."  He describes the standard hollow handled pitcher.

Walter Gibson.  Secrets of the Great Magicians.  (Grosset & Dunlap, 1967);  Collins, Glasgow, 1976.  Describes several ancient devices based on siphons. 

Pp. 13-14: The tomb of Belus.  This was encountered by Xerxes, c‑480, and was a vessel which maintained its level when liquid was poured into it.

Pp. 18-19: The ever-full fountain.  This was described by Hero and is a vessel which maintains its level when water is drawn from it.

Pp. 95-96: The Bundar boat.  This is a device which has intermittent flow which was used by Indian magicians, apparently already in use when Westerners got to India.

John Timbs.  Things Not Generally Known,  Familiarly Explained.  A Book for Old and Young (spine says  First Series  and a note by a bookdealer on the flyleaf says  2 vol.).  Kent & Co., London, (1857?), 8th ed., 1859.  Hydrostatic wonders, p. 111, mentions some classical examples as:  "The magic cup of Tantalus, which he could never drink, though the beverage rose to his lips; the fountain in the island of Andros, which discharged wine for seven days, and water during the rest of the year; the fountain of oile which burnt out to welcome the return of Augustus from the Sicilian war; the empty urns which, at the annual feast of Bacchus, filled themselves with wine, to the astonishment of the assembled strangers; the glass tomb of Belus, which, after being emptied by Xerxes, would never again be filled; the weeping statues of the ancients; and the weeping virgin of modern times, whose tears were uncourteously stopped by Peter the Great when he discovered the trick; and the perpetual lamps of the magic temples, -- were all the obvious effects of hydrostatical pressure.  -- North British Review, No. 5."

Banu Musa  = Banū Mūsā bin Shākir  (Sons of Moses),  but largely the work of Ahmad  = Abu-l-Hasa [the  H  should have an underdot] Ahmad [the  h  should have an underdot] ibn Mūsa.  Kitāb al-Hiyal [the  H  should have an underdot].  c870.  Translated and annotated by Donald R. Hill as:  The Book of Ingenious Devices; Reidel, 1979.  Describes 103 devices, most of which are trick vessels, as well as fountains, etc.  E.g. Model 1:  "We wish to explain how a beaker is made in which a quantity of wine is poured, and if [a measure] of wine or water is added to it all its contents are discharged."  This is a Tantalus cup.  Model 4:  "We wish to make a jar with an open outlet: if water is poured into it nothing issues from the outlet, and if pouring is stopped the water issues from the outlet, and if pouring is resumed [discharge] ceases again, and if pouring is stopped the water discharges, and so on continuously."  Models 12 - 15 have finger holes under the handle allowing the pourer to produce various effects.

Sandfield says the earliest Chinese puzzle vessels are in the Xian Museum and are dated to either the Song (951-960) or Northen Song (960-1127 or 1279).  (My Chinese chronology has Song being 960-1279, with Northern Song being 960-1127 and Southern Song being 1127-1279.)  Sandfield says they visited the Yaozhou Kiln Museum which apparently has been making Celadon Magic Teapots ever since that time!  [See his Annotated Bibliography items 6 & 7.]

Al-Jazari  = Badī‘al-Zamān Abū al-‘Izz Isma‘il ibn al-Razzāz al-Jazarī.  Kitāb fī ‘rifat al-hiyal (the  h  should have an underdot) al-handasiyya.  c1204.  Translated and annotated by:  Donald R. Hill as:  The Book of Knowledge of Ingenious Mechanical Devices; Reidel, Dordrecht, 1974.  Introduction, pp. 3‑12, and Conclusion, pp. 279-280, include comparisons with other works.  Not much of interest to us, except for the following.

Category II, chapter 5, pp. 110-114, 221, 256-257, 272.  "It is a pitcher for wine which is used in carousals, into which water and wines of [different] colours are poured; it has a valve from which each colour is drawn separately."

Category III, chapter 1, pp. 127-129, 259, 272.  "A pitcher from which hot water, cold water and mixed water is poured."

In the Historical Museum at Cologne is the 'goblet of Albertus Magnus', which has a false bottom which allowed him to introduce antimony and make the drink emetic.  He lived 1193-1280, so this might date from c1250.  Described and illustrated in:  Edwin A. Dawes; The Great Illusionists; Chartwell Books, Secaucus, New Jersey, 1979, p. 21.

Ashmolean Museum, Oxford, item 1921-202, presently in Room 4, in the case devoted to Decorative Techniques, item 14.  Late 13C, different than the Exeter jug below.  Described, with photo and photo of a cut-away model, in Crossley, below.  The same item is illustrated by a drawing in the following entry.

Jeremy Haslam.  Medieval Pottery in Britain.  Shire, Aylesbury, (1978);  2nd ed., 1984.  On pp. 19, under: Some regional types; Oxford region; Thirteenth and fourteenth centuries, he mentions, among characteristic items of the Oxford region in the late 13C, "'puzzle‑jugs' decorated with applied scales, strips of red-firing clay and stags' heads (13, 10)".  Fig. 13 on p. 47, Oxford, Fourteenth- and Fifteenth-Century Vessels, item 10, is "Puzzle-jug, glazed green, with app. scales, red strips, face masks and deer head over spout."  The drawing is by the author from the above mentioned item in the Ashmolean Museum, Oxford.

The Royal Albert Museum, Queen Street, Exeter, EX4 3RX, has a notable example of a puzzle jug, c1300, from SW France or the Saintonge region of west France, described as the finest piece of imported medieval pottery in England.  A postcard of it is available from the Museum.  Crossley, below, says it is late 13C from the Saintonge and was excavated in Exeter in 1899.  He gives a cross-sectional drawing and cites a 1988 excavation report.  In fact, the puzzle aspect is quite simple -- the upper level is connected to the base via the handle which leaves room for three levels of decoration, showing some apparently unclothed bishops inside, then some ladies leaning out of windows and then some musicians serenading outside!  It is depicted and described in: John Allan & Simon Timbs; Treasures of Ancient Devon; Devon Books (Devon County Council), Tiverton, 1996, pp. 34-35. 

A 16C German puzzle glass is shown and described in:  Franz Sales Meyer.  (Ornamentale Formenlehre; 1888.  Translated as:  A Handbook of Ornament; Batsford, 1894.  Reprinted as:  Meyer's Handbook of Ornament; Omega Books, London, 1987, pp. 351 & 353, plate 207 (on p. 351), fig. 10.  "To those times also belonged: Puzzle-glasses from which the liquor had to be sucked-out at the end of the handle (fig. 10): ...."  The figure shows a tallish vessel apparently with a closed top through the hollow handle runs to the bottom of the vessel and the bottom of the handle bends out to provide a sucking point.  The interior tube is marked with dotted lines.

Cardan.  De Rerum Varietate.  1557, ??NYS.  = Opera Omnia, vol. III, pp. 252.  Liber XVIII.  Urcei qui non se mergunt.  Apparently a vase which one fills with water and then pours wine out of.

The Percival David Foundation of Chinese Art, London, has a late 16C Tantalus Cup (labelled Mazer Cup, item PDF A721) with a statue in the middle.

Prévost.  Clever and Pleasant Inventions.  (1584), 1998.  Pp. 41-44.  A Tantalus cup.

John Bate.  The Mysteries of Nature and Art.  In foure severall parts.  The first of Water works.  The second of Fire works.  The third of Drawing, Washing, Limming, Painting, and Engraving.  The fourth of sundry Experiments.  (Ralph Mab [sic], London, 1634.)  The second Edition; with many additions into every part.  Ralph Mabb, London, 1635.  (3rd ed., Andrew Crooke, London, 1654.)  [BCB 20-22.  Toole Stott 81-83.  HPL [Bate] RBC has 2nd & 3rd eds.]  The first part has several examples, notably the following.

                    P. 2.  How to make a conceited pot, which being filled with water, will of it selfe runne all out; but not being filled will not run out.

                    P. 3.  Another conceited Pot out of which being first filled with wine and water, you may drinke pure wine apart, or faire water apart, or els both together.

Louis L. Lipski.  Dated English Delftware: Tin-glazed Earthenware, 1600-1800.  Sotheby Publications, London  &  Harper & Row, Scranton, Pennsylvania,  1984.  ??NYS -- cited and quoted by Sandfield.  Their earliest upright puzzle jug is no. 1009, dated to 1653.

Edgar Gorer  &  J. F. Blacker.  Chinese Porcelain and Hard Stones, Illustrated ....  Quaritch, London, 1911.  Vol. 1, plate 135 and preceding page.  ??NYS -- cited by Rasmussen and quoted by Sandfield.  Two illustrated examples of puzzle jugs, i.e. 'how-to-pour puzzles', dated from the Kang-he (= K'ang Hsi) period, 1662-1722.  ??check dates.

Ozanam.  1694.  Probs. 14 & 26: 1696: 286 & 294 & figs. 137 & 138, plate 47 & fig 146, plate 49;  1708: 362-363 & 370-371 & figs. 33 & 34, plate 15 & fig. 44, plate 18.  Probs. 18 & 30: 1725: vol. 2: 389-390 & 404 & figs. 137 & 138, plate 47 & fig. 146, plate 49.  Probs. 4 & 5: 1790: vol. 4: 33-35 & figs. 5-7, plate 1;  1803: 34-36 & figs. 5-7, plate 1;  1840: 613‑614.  Three forms of Tantalus cups.  In 1790, it is called Tantalus and a figure of Tantalus is put in the cup so that when the liquid approaches him, it runs out.  Some of these are such that they pour out when the cup is tilted.

Patrick J. Donnelly.  Blanc de Chine: The Porcelain of Tehua in Fukien.  London, 1969.  Pp. 95, 345, Appendix 5: N86 & N69.  ??NYS -- cited and quoted by Sandfield.  Describes some tantalus cups, where the siphon is concealed in a figure, dated 1675‑1725.  The T'ao-ya says the figure represents Lu Hung-chien (= Lu Hong Jian), a 16C author on porcelain whose name was Hsiang Yuan-pien Tzu-ching and that the vessel is called a 'quiet mind' dish.  Augustus the Strong (of Saxony?) had an example before 1721.  There is an example in the Princessehof Museuem, Leeuwarden, The Netherlands.  Other versions mentioned in Sandfield have naked men or women in the cups.

The Percival David Foundation of Chinese Art, London, has a 'peach shaped wine pot', i.e. a Cadogan pot, from the Kangxi period, 1662-1722, item PDF 826.

The Victoria and Albert Museum, London, has an example from the K'ang Hsi period (1622‑1722).  ??NYS -- cited and quoted by Sandfield.  ??check dates.

Alberti.  1747.  Art 31, pp. ?? (131-132) & fig. 57 on plate XVI, opp. p. ?? (130).  Vase pours different liquors.

Tissandier.  Récréations Scientifiques.  1880?;  2nd ed., 1881, pp. 327-328 describes 'vases trompeurs', with illustrations on pp. 324-325.  Says they were popular in the 18C and earlier and the illustrations are of examples in the Musée de la Manufacture de Sèvres.  Not in 5th ed., 1888.

Tissandier.  Popular Scientific Recreations.  1890?  Pp. 65-67, with illustrations on pp. 67-69, describes "vases of Tantalus" -- cups which cannot be filled too full -- and then gives the material from the 2nd French ed., 1881.

The Ashmolean Museum, Oxford, Room 47 has a fine example of a Lambeth puzzle jug dated 1745.

The Rijksmuseum, Amsterdam, Room 255 has a fine example of a Delft puzzle jug dated 1768.

Henry Pierre Fourest.  Delftware: Faience Production at Delft.  (Translated from La Faiences de Delft.)  Rizzoli, NY, 1980.  P. 136, no. 130.  ??NYS - cited and quoted by Sandfield.  Shows a Delft puzzle jug from second half of 18C.  In a footnote, the author says: "This very old common form was copied from the Chinese for the East India Company."

Sandfield says the earliest English Cadogan teapots were made at Swindon, Yorkshire, after 1806 and they became popular after 1810.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  Pp. 130-131, no. 195 & Frontispiece fig. 10: Hydraulic Experiment called Tantalus's Cup.

Rational Recreations.  1824.  Experiment 66, pp. 122-123.  Tantalus cup.

The Boy's Own Book.  1828: 446.  Tantalus's cup.  With statue of Tantalus inside.

An exhibit in the Bramah Tea & Coffee Museum, London, says Mrs Cadogan brought the first examples, of what became Cadogan teapots, to England about 1830, but see above.

Young Man's Book.  1839.  Pp. 172-173.  Tantalus's Cup.

The Secret Out.  1859. 

The Magic Coffee-Pot, pp. 346-347.  Pours coffee, milk or both.

A Vessel that will let Water out at the Bottom, as soon as the Mouth is uncorked, p. 371 (UK: p. 181).  Small holes in the bottom of a corked bottle.

G. J. Monson-Fitzjohn.  Drinking Vessels of Bygone Days  from the Neolithic age to the Georgian period.  Herbert Jenkins, London, 1927. 

                    On pp. 42-43, he discusses and illustrates fuddling cups and pot crowns.  He says fuddling cups, tygs and puzzle pots dates from Elizabethan times.  Says fuddling cups have three to six cups and there is an example with six in the British Museum, dated 1790, with the inscriptions: "My friend is he that loves me well, but who he is I cannot tell".  Says Drake, Frobisher and Raleigh used or owned examples.

                    A pot crown has four cups mounted on a ring, with four tubes rising to a central spout.  A maiden wore this on her head and the object was to drain the cups by sucking on the spout, which is not a puzzle, but the maiden could tilt her head to prevent anyone but her chosen suitor managing to drink.

                    On p. 49-50, he describes tygs and posset pots as simple large drinking vessels.

                    On p. 58, he illustrates and describes an ordinary puzzle jug.  An example in the BM bears the following verse:  "Here, gentlemen, come try your skill  /  I'll hold a wager if you will  /  That you don't drink this liquor all  /  Without you spill or let some fall."  He says there is a good example in the Unicorn Hotel, Ripon,. and that an example, dated 1775, has the following:  "God save the King I say  /  God bless the King I pray  /  God save the King."

                    On pp. 103-104, he describes and illustrates the 17C 'Milkmaid Cup' of the Vintners' Company.  When inverted, the milkmaid's skirt is a large cup.  A smaller second cup is supported on pivots between her upraised arms.  When inverted, this second cup is also filled and the drinker must drain the larger cup without spilling any from the swinging cup.

Robert Crossley.  The circulatory systems of puzzle jugs.  English Ceramic Circle Transactions 15:1 (1993) 73-98, front cover & plate V.  Starts with medieval examples from late 13C -- see above.  Says the traditional jug with sucking spouts is first known from late 14C England, though it probably derives from Italy but no early examples are known there.  Classifies puzzle jugs into 10 groups, subdivided into 20 types and into 40 variations.  Some of these are only known from one example, while others were commercially produced and many examples survive.  64 B&W illustrations, 4 colour illustrations on the cover and plate V.  Many of the illustrations show the circulation systems by cut-away drawings or photos of cut-away models, followed by photos of actual jugs.  72 references, some citing several sources or items.  The Glaisher collection at the Fitzwilliam Museum, Cambridge, provided the most examples -- 13.  (I wrote to Crossley in about 1998, but had no answer and the Secretary of the English Ceramic Circle sent me a note about a year later saying he had died.)

                    In the same issue of the journal, pp. 45, 48 & plate IIb show 6 examples of Cadogan teapots.

The Glaisher Collection of pottery in the Fitzwilliam Museum, Trumpington Street, Cambridge, CB2 1RB; tel: 01223-332900, has the largest number of puzzle vessels on display that I know of.  I found 22 on display, including one specially made for Glaisher.  James Whitbread Lee Glaisher (1848-1928) was a mathematician of some note at Trinity College, Cambridge.  His collection included over 3000 items.  Many of the items on display are described in Crossley's article, but Crossley mentions six items of the Glaisher Collection which are not on display.  I have prepared a list of the items on display and the further six items.

                    The collection includes several 'tygs', which are large cups with several handles.  Some of these are specifically described as puzzle vessels.  In some other cases one cannot see if a tyg is a puzzle tyg or not, but these cases are all included in Crossley.  [Crossley, pp. 92-93.]

                    The collection includes some 'fuddling cups' which are multiple interconnected small cups which either spill on you if you don't use them correctly or cause you to drink several cupfuls instead of one -- when this is not expected, you get fuddled!  One example has ten cups in a triangular array.  [Crossley, pp. 91-92] mentions these briefly.  The collection is described in the following.

Bernard Rackham.  Catalogue of the Glaisher Collection of Pottery & Porcelain in the Fitzwilliam Museum, Cambridge, 2 vol., CUP, 1935;  Reprinted: Antique Collectors' Club, Woodbridge (Suffolk), 1987.  ??NYS -- found in BLC.

J. F. Blacker.  Chats on Oriental China.  T. Fisher Unwin, Ltd., London, 1908, 406 pp.  See pp. 272‑273.  Cadogan pots.  (??NYS -- from Rasmussen.)

R. L. Hobson.  Chinese Pottery and Porcelain.  Funk & Wagnalls, New York, 1915.  Reprinted by Dover, 1976.  See pp. 276, 278.  Tantalus cups, Cadogan pots.  (??NYS -- from Rasmussen.)

R. L. Hobson & A. L. Hetherington.  The Art of the Chinese Potter from the Han Dynasty to the end of the Ming.  Benn, London  & Knopf, New York,  1923, 20[sic ??] pp.  Reprinted as: The Art of the Chinese Potter: An Illustrated Survey, Dover, 1982, 137 pp.  See Plates 128 & 149.  Tantalus cups.  Cadogan pots.  (??NYS -- from Rasmussen.)

R. L. Hobson; Bernard Rackham & William King.  Chinese Ceramics in Private Collections.  Halton & Truscott Smith, London, 1931, 201 pp.  See pp. 109‑110.  Tantalus cups.  (??NYS -- from Rasmussen.)

D. F. Lunsingh Scheurleer.  Chinese Export Porcelain: Chine de Commande.  Pitman, New York, 1974, 256 pp.  See pp. 94‑95, 215, plate 105.  Puzzle jugs.  (??NYS -- from Rasmussen.)

Miriam Godofsky.  The Cadogan pot.  The Wedgewoodian (Oct 1982) 141-142.  ??NYS -- cited and quoted by Sandfield.  Three illustrations and mentions of several in museums.

Tassos N. Petris.  Samos  History - Art - Folklore - Modern Life.  Toubis, Athens, 1983, pp. 32 & 66-68] says the local potters at Mavratzaioi, Samos, were still making puzzle jugs (maskara bardak) and tantalus cups (dikia koupa), but "the secrets of making these vessels are now known to only a few, and it must be regarded as a dying trade."

C. J. A. Jörg.  Interaction in Ceramics: Oriental Porcelain & Delftware.  Hong Kong Museum of Art, Hong Kong, 1984, 218 pp.  See pp. 78‑79 and 162‑163, nos. 36 and 115.  Puzzle jugs.  (??NYS -- from Rasmussen.)

Nob Yoshigahara.  Puzzlart.  Tokyo, 1992.  Puzzle Collection, pp. 50-57 shows many examples and a photo of Laurie Brokenshire, all in colour.

Lynn Pan.  True to Form: A Celebration of the Art of the Chinese Craftsman.  FormAsia Books Ltd., Hong Kong, 1995, 148 pp.  See p. 18.  Cadogan pots.  (??NYS -- from Rasmussen.)

Fang Jing Pei.  Treasures of the Chinese Scholar.  Weatherhill, New York, 1997, 165 pp.  See pp. 116‑117.  Cadogan pots.  (??NYS -- from Rasmussen.)

Rik van Grol.  Puzzling China.  CFF 47 (1998??) 27-29.  ??NYR -- cited and quoted by Sandfield. 

Franz de Vreugd.  Oriental puzzle vessels.  CFF 49 (Jun 1999) 18-20.  ??NYR-- cited by Sandfield.  12 puzzle vessels in colour.

 

          10.X. HOW FAR DOES A PHONOGRAPH NEEDLE TRAVEL?

 

            New section.  I have just seen a recent version of this and decided it ought to be entered.  There must be examples back to the early part of this century.

 

Meyer.  Big Fun Book.  1940.  No. 2, pp. 173 & 755.

Young World.  c1960.  P. 29.  "How many grooves are there in a long playing record?  Just one long one."  [Actually there are two!]

The Diagram Group.  The Family Book of Puzzles.  The Leisure Circle Ltd., Wembley, Middlesex, 1984.  Problem 95, with Solution at the back of the book. 

 

          10.Y. DOUBLE CONE ROLLING UPHILL

 

W. Leybourn.  Pleasure with Profit.  1694.  Tract. III, pp. 12-13: "A Mechanical Paradox: or, A New and Diverting Experiment.  Whereby a Heavy Body shall by its own Weight move up a sloping Ascent.  Written by J. P."  Nice drawing.  No indication of who J. P. is and he is not one of the publishers or the additional author.

Andrew Q. Morton.  Science in the 18th Century.  The King George III Collection.  Science Museum, London, 1993.  P. 33 shows the example in the George III Gallery and says it is c1750 and that the idea was invented at the end of the 17C and was popular in 18C lectures on mechanics.

Henk J. M. Bos.  Descriptive Catalogue  Mechanical Instruments in the Utrecht University Museum.  Utrecht University Museum, 1968, pp. 35-37.  Describes several examples, saying they are described in the classic experimental mechanics texts of the early 18C, citing 's Gravesande, Desaguliers, Nollet, Musschenbroek.  Item M 5 was bought in 1755. 

                    The Museum also has two oscillatory versions where the cone seems to roll over a peak and down the other side, then back again, ....  Item M 7 is first mentioned in an inventory of 1816.

Ozanam-Montucla.  1778.  Vol. II, prob. 26:  1790: 45-46 & figs. 22-24, plate 5;  1803: 49-50 & figs. 22-24, plate 5;  1840: 216-217.

Catel.  Kunst-Cabinet.  1790.  Der bergangehende Kegel, p. 12 & fig. 4 on plate I. 

Bestemeier.  1801.  Item 40 -- Der Berggehende Kegel.  Copies most of Catel.  Diagram is copied from Catel, but less well done.

Gardner D. Hiscox.  Mechanical Appliances  Mechanical Movements  and  Novelties of Construction.  A second volume to accompany his previous Mechanical Movements, Powers and Devices.  Norman W. Henley Publishing Co, NY, (1904), 2nd ed., 1910.  Item 931, p. 372, is an attempt to use this device as a perpetual motion by having the rails diverging where it goes uphill and parallel where it goes downhill in alternate sections.  Patented in 1829!

Magician's Own Book (UK version).  1871.  A double cone ascending a slope by its own weight, pp. 136-137.  The picture of the cone looks more like an octahedron!

Will Goldston.  Tricks & Illusions  for amateur and professional conjurers.  Routledge  &  Dutton,  9th ed (revised), nd [1920s?], pp. 28-29: The magic cone.

 

          10.Z.  THE WOBBLER

 

            This is formed by slotting two discs on radial lines and fitting them together.  When the distance between centres is  Ö2  times the radius, then the centre of gravity at two obvious positions is at the same height and the object rolls rather smoothly with a 'wobbly' motion.  The basic result is that the centre of gravity stays at constant height as it rolls.  Similar shapes are included here.

 

Gardner D. Hiscox.  Mechanical Appliances  Mechanical Movements  and  Novelties of Construction.  A second volume to accompany his previous Mechanical Movements, Powers and Devices.  Norman W. Henley Publishing Co, NY, (1904), 2nd ed., 1910.  Item 355: The pantanemone, p. 144.  Take a disc and cut it along a diameter.  Twist one half by 90o about the diameter perpendicular to the cut, so the planes of the semi-circles are at right angles, looking  bit like the Wobbler.  Mount the pieces on an axle at 45o to each plane -- in practice this requires some guy wires between the pieces.  He asserts that this rotates in a wind from any direction, except perpendicular to the axle, so it can be used as a stationary windmill, giving 60 days more work per year than conventional windmills.

 

A. T. Stewart.  Two-circle roller.  Amer. J. Physics 34:2 (Feb 1966) 166-167.  Shows the basic result and gets the height of the centre of gravity for any spacing.  Shows examples made from slotted plastic discs.  I learned of Stewart's work in 1992 from Mike Berry.  In a letter of 27 Oct 1992 to Berry, he says the problem is difficult and little known.  It took him "a dozen pages of cumbersome trigonometry".  He offered $5 to anyone who could derive the result on one page and only two people ever collected.

Anthea Alley.  Rocking Toy  1969.  This is an example of the wobbler made from 48" diameter disks, 1" thick, with holes about 12" diameter in the middle.  This is shown, with no text, in a photograph on p. 108 of:  Jasia Reichardt, ed.; Play Orbit [catalogue of an exhibition at the ICA, London, and elsewhere in 1969-1970]; Studio International, 1969.

Paul Schatz.  Rhythmusforschung und Technik.  Verlag Freies Geistesleben, Stuttgart, 1975.  In this he considers a number of related ideas, but his object -- the Oloid -- has the distance between centres being the radius.  This does not roll smoothly on the level, but he would roll it down a slope -- see Müller.

Paul Schatz.  Swiss patent 500,000.  (??NYS -- cited in Müller.)

Georg Müller, ed.  Phänomena -- Eine Dokumentation zur Ausstellung über Phänomene und Rätsel der Umwelt an der Seepromenade Zürichhorn, 12 Mai - 4 November 1984.  Zürcher Forum, Zürich, 1984.  P. 79 shows the Oloid rolling down a slope.

David Singmaster and Frederick Flowerday.  The wobbler.  Eureka 50 (1990) 74-78.  Flowerday developed the idea in the 1980s, but I can't recall if he developed it from Schatz's work.  This paper proves the basic result and discusses unsolved problems such as the path of the centre of gravity.  For the Oloid, the distance between the two contact points is of constant length.  Stewart's paper was not known when this was written.

Christian Ucke & Hans-Joachim Schlichting.  Wobbler, Torkler oder Zwei-Scheiben-Roller.  Physik in unserer Zeit 25:3 (1994) 127-128.  Short description, mentioning Flowerday and citing Stewart and Schatz.  Says they are sold in Germany and Switzerland as Wobbler or Go-On.  Shows a version using slotted plastic discs called Rondi.  Instructions on how to make one.

Christoph Engelhardt & Christian Ucke.  Zwei-Scheiben-Roller.  Preprint of 3 May 1994, to appear in Mathematisch-Naturwissenschaftlicher Unterricht (1995).  Basic description and derivation of the basic result.  Shows this also works for elliptical discs!  Attempts to find the path of the centre of gravity.  Singmaster & Flowerday was not known when this was written.

 

          10.AA.        NON-REGULAR DICE

 

            This deals with determining the probability of the various faces of a die which is not a regular polyhedron.  The immediate approach is a simple geometric model -- the probability of a face should be proportional to the solid angle subtended by that face viewed from the centroid.  However, this fails to agree with reality and a number of authors have attempted to explain the real situation by more complex modellings of the physical situation.  Note that a coin or cylinder is a non-regular die, so the phenomenon of a coin landing on edge is included here.  I have included everything I know of here as there has been considerable interest in this problem in recent years.  Ricky Jay recently asked about the history of loaded or mis-spotted dice and we would appreciate references on this -- it is commonly asserted that mis-spotted dice occurred in the 11C, but I cannot relocate my source for this and there may not have been very specific -- cf Jay, 2000.

 

Scott Beach.  Musicdotes.  Ten Speed Press, Berkeley, California, 1997, p. 77.  Says Jeremiah Clarke (c1674-1707), the organist of St. Paul's Cathedral and a composer best known for the Trumpet Voluntary (properly the Prince of Denmark's March, long credited to Purcell) became enamoured of a lady above his station and was so despondent that he decided to commit suicide.  Being somewhat indecisive, he threw a coin to determine whether to hang himself or drown himself.  It landed on the ground and stuck on edge!  Failing to recognise this clear sign, he went home and shot himself!  (The text is given as a Gleaning: A loss of certainty, submitted by me, in MG 66 (No. 436) (1982) 154.)

P. S. de Laplace.  Théorie Analytique des Probabilités.  3rd ed, Courcier, Paris, 1820.  P. 359 remarks that determining the probabilities for a cuboid is beyond analytic techniques.

J. D. Roberts.  A theory of biased dice.  Eureka 18 (Nov 1955) 8-11.  Deals with slightly non-cubical or slightly weighted dice (e.g. due to the varying number of pips).  He changes the lengths by  ε  and ignores terms of order higher than first order.  He uses the simple geometric theory.

J. R. Probert-Jones.  Letter to the Editor.  Eureka 19 (Mar 1957) 17-18.  Says Roberts' article is the first to treat the problem quantitatively.  States that about 1900, Weldon made two extensive trials on dice.  Weldon's work is not known to have been published, but Pearson cited his results of  26,306  throws of six dice in the paper in which he introduced the  χ2  test.  The analysis showed the results would occur with fair dice about once in  20,000  trials, but using Roberts' estimate of the bias of dice, the result would occur about once in every two trials.  Edgeworth describes Weldon's other trials, which were a bit more elaborate.  Assuming fair dice, the probability of the results was  .0012,  while Roberts's estimate leads to probability  .067.  Both trials show that fairness of ordinary dice is not reasonable and Roberts' estimates are reasonable.

L. E. Maistrov.  Probability Theory.  A Historical Sketch.  Academic, 1974.  ??NYS -- Heilbronner says he measured ancient dice at Moscow and Leningrad, finding them quite irregular -- the worst cases having ratios of edge lengths as great as 1.2 and 1.3.

Robert A. Gibbs, proposer;  P. Merkey & Martin Berman, independent solvers.  Problem 1011 -- An old dice problem.  MM 50:2 (Mar 1977) 99  &  51 (1978) 308.  Editor says it is an old problem which might be of interest to new readers.  Can one can load a pair of dice so each value has equal probability of occurring.  I.e. each of  2, 3, ..., 12  has probability  1/11.  Solution notes refer to  Problem E 925, AMM 57 (1951) 191-192  and  E. J. Dudewicz & R. E. Dann; Equally likely dice sums do not exist; Amer. Statistician 26 (1972) 41-42,  both ??NYS.

Scot Morris.  The Book of Strange Facts and Useless Information.  Doubleday, 1979, p. 105.  Says  6  is the most common face to appear on an ordinary die because the markings are indentations in the material, making the six side the lightest and hence most likely to come up.  He says that this was first noticed by ESP researchers who initially thought it was an ESP effect.  The effect is quite small and requires a large number of trials to be observable.  (I asked Scot Morris for the source of this information -- he couldn't recall but suspected it came from Martin Gardner.  Can anyone provide the source?)

Frank Budden.  Note 64.17: Throwing non-cubical dice.  Math. Gaz. 64 (No. 429) (Oct 1980) 196-198.  He had a stock of 15mm square rod and cut it to varying lengths.  His students then threw these many times to obtain experimental values for the probability of side versus end.

David Singmaster.  Theoretical probabilities for a cuboidal die.  Math. Gaz. 65 (No. 433) (Oct 1981) 208-210.  Gives the simple geometric approach and compares the predictions with the experimental values obtained by Budden's students and finds they differ widely.

                    Correspondence with Frank Budden led to his applying the theory to a coin and this gives probabilities of landing on edge of  8.1%  for a UK 10p coin and  7.4%  for a US quarter.  [And  9.5%  for a US nickel.]

Trevor Truran.  Playroom: The problem of the five-sided die.  The Gamer 2 (Sep/Oct 1981) 16 & 4 (Jan/Feb 1982) 32.  Presents Pete Fayers' question about a fair five-sided die and responses, including mine.  This considered a square pyramid and wanted to determine the shape which would be fair.

Eugene M. Levin.  Experiments with loaded dice.  Am. J. Physics 51:2 (1983) 149-152.  Studies loaded cubes.  Seeks for formulae using the activation energies, i.e. the energies required to roll from a face to an adjacent face, and inserts them into an exponential.  One of his formulae shows fair agreement with experiment.

E. Heilbronner.  Crooked dice.  JRM 17:3 (1984-5) 177-183.  He considers cuboidal dice.  He says he could find no earlier material on the problem in the literature.  He did extensive experiments, a la Budden.  He gives two formulae for the probabilities using somewhat physical concepts.  Taking  r  as the ratio of the variable length to the length of the other two edges, he thinks the experimental data looks like a bit of the normal distribution and tries a formulae of the form  exp(-ar2).  He then tries other formulae, based on the heights of the centres of gravity, finding that if  R  is the ratio of the energies required to tilt from one side to another, then  exp(-aR)  gives a good fit.

Frank H. Berkshire.  The 'stochastic' dynamics of coins and irregular dice.  Typescript of his presentation to BAAS meeting at Strathclyde, 1985.  Notes that a small change in  r  near the cubical case, i.e.  r = 1,  gives a change about  3.4  times as great in the probabilities.  Observes that the probability of a coin landing on edge depends greatly on how one starts it - e.g. standing it on edge and spinning it makes it much more likely that it will end up on edge.  Says professional dice have edge  3/4"  with tolerance of  1/5000 " and that the pips are filled flush to the surface with paint of the same density as the cube.  Further, the edges are true, rather than rounded as for ordinary dice.  These carry a serial number and a casino monogram and are regularly changed.  Describes various methods of making crooked dice, citing Frank Garcia; Marked Cards and Loaded Dice; Prentice Hall, 1962, and John Scarne; Scarne on Dice; Stackpole Books, 1974.  Studies cuboidal dice, citing Budden and Singmaster.  Develops a dynamical model based on the potential wells about each face.  This fits Budden's data reasonably well, especially for small values of  r.  But for a cylinder, it essentially reduces to the simple geometric model.  He then develops a more complicated dynamical model which gives the probability of a 10p coin landing on edge as about  10-8.

John Soares.  Loaded Dice.  (Taylor Publishing, Dallas, 1985);  Star (W. H. Allen), 1988.  On p. 49, He says "Aristotle wrote a scholarly essay on how to cheat at dice."  In an Afterword by George Joseph, pp. 243-247, on p. 246, he describes 'flats': "Certain sides of the dice are slightly larger (Flat) 1/5,000th to 1/10,000th of an inch."

David Singmaster.  On cuboidal dice.  Written in response to the cited article by Heilbronner and submitted to JRM in 1986 but never used.  The experimental data of Budden and Heilbronner are compared and found to agree.  The geometric formula and Heilbronner's empirical formulae are compared and it is found that Heilbronner's second formulae gives the best fit so far.

                    I had a letter in response from Heilbronner at some point, but it is buried in my office.

Joseph B. Keller.  The probability of heads.  AMM 93:3 (Mar 1986) 191-197.  Considers the dynamics of a thin coin and shows that if the initial values of velocity and angular velocity are large, then the probability of one side approaches  1/2.  One can estimate the initial velocity from the amount of bounce -- he finds about  8 ft/sec.  Persi Diaconis examined coins with a stroboscope to determine values of the angular velocity, getting an average of  76π rad/sec.  He considers other devices, e.g. roulette wheels, and cites earlier dynamically based work on these lines.

Frank H. Berkshire.  The die is cast.  Chaotic dynamics for gamblers.  Copy of his OHP's for a talk, Jun 1987.  Similar to his 1985 talk. 

J. M. Sharpey-Schafer.  Letter: On edge.  The Guardian (20 Jul 1989) 31.  An OU course asks students to toss a coin 100 times and verify that the distribution is about 50 : 50.  He tried it 1000 times and the coin once landed on edge.

D. Kershaw.  Letter: Spin probables.  The Guardian (10 Aug 1989) 29.  Responding to the previous letter, he says the probability that a tossed coin will land on edge is zero, but this does not mean it is impossible.

A. W. Rowe.  Letter.  The Guardian (17 Aug 1989) ??  Asserts that saying the probability of landing on edge is zero admits 'to using an over-simplified mathematics model'.

K. Robin McLean.  Dungeons, dragons and dice.  MG 74 (No. 469) (Oct 1990) 243-256.  Considers isohedral polyhedra and shows that there are 18 basic types and two infinite sets, namely the duals of the 5 regular and 13 Archimedean solids and the sets of prisms and antiprisms.  Then notes that unbiased dice can be made in other shapes, e.g. triangular prisms, but that the probabilities are not obvious, citing Budden and Singmaster, and describing how the probabilities can change with differing throwing processes.

Joe Keller, in an email of 24 Feb 1992, says Frederick Mosteller experimented with cylinders landing on edge 'some time ago', probably in the early 1970s.  He cut up an old broom handle and had students throw them.  He proposed the basic geometric theory.  Keller says Persi Diaconis proposed the cuboidal problem to him c1976.  Keller developed a theory based on energy loses in rolling about edges.  Diaconis made some cuboidal die and students threw them each 1000 times.  The experimental results differed both from Diaconis' theory (presumably the geometric theory) and Keller's theory.

Hermann Bondi.  The dropping of a cylinder.  Eur. J. Phys. 14 (1993) 136-140.  Considers a cylindrical die, e.g. a coin.  Considers the process in three cases: inelastic, perfectly rough planes;  smooth plane, for which an intermediate case gives the geometric probabilities;  imperfectly elastic impacts. 

In late 1996 through early 1997, there was considerable interest in this topic on NOBNET due to James Dalgety and Dick Hess describing the problem for a cubo-octahedron.  I gave some of the above information in reply.

Edward Taylor Pegg, Junior.  A Complete List of Fair Dice.  Thesis for MSc in Applied Mathematics, Dept. of Mathematics, Univ. of Colorado at Colorado Springs, 1997.  40pp.  He started with the question of finding a fair five-sided die and tried to find one in the form of a triangular prism as well as a square pyramid.  He considers the simple geometric model and notes it cannot work as some polyhedra have unstable faces!  Asserts that an isohedron is obviously fair.  [A polyhedron is isohedral if any face can be mapped onto any other by a symmetry (i.e. a rotation or a rotation and reflection) of the polyhedron.]  Notes that the method of throwing a die affects its fairness and says this shows that isohedra are the only fair dice.  He lists all the isohedra -- he says there are 24 of them and the two infinite families of dual prisms (the bipyramids) and dual antiprisms (the trapezohedra), but I only see 23 of them in his lists).  He explicitly describes a fair but unsymmetrical cardboard tetrahedron, but it seems he has to cut a circular hole in one face or weight the face.  Develops an energy state model and finds that the probability of a US nickel landing on edge is  3/10,000, but then revises his model and gets  15/100,000,000  -- Budden's 1981 method gives  9.5%.  20 references, but he was unaware of the above literature and the references are mostly to polyhedra and their groups.  Pegg also sent a graph of Frank Budden's data versus the mathematical model - presumably one of Pegg's models.

[A. D. [Tony] Forbes.]  Problem 171.1 - Cylinder.  M500 171 (Dec 1999) 9.  Asks for the shape of cylinder such that the probability of landing on the side is 50%.

David Singmaster.  Re: Problem 171.1 -- Cylinder.  M500 173 (Apr 2000) 19.  Sketch of history of the problem.

David Singmaster.  Non-regular Dice.  M500 174 (Jun 2000) 12-15.  The material of this section up through 1997, though I have since amended some of it.  With a prefatory note by ADF that he had no idea the problem had any history when he posed it.

Gordon Alabaster.  Re: Problem 171.1 -- Cylinder.  M500 174 (Jun 2000) 16-17.  Gives a simple physical argument that 50% should occur when the cross-section is square.

Colin Davies.  Re: Problem 171.1 -- Cylinder.  M500 176 (Oct 2000) 22.  Differs with Alabaster's analysis and notes the the way in which the cylinder is tossed has a major effect.

Ricky Jay.  The story of dice.  The New Yorker (11 Dec 2000) 90-95.  A set of 24 loaded or mismarked dice was found in a container, dated late 15C, in the Thames in 1984, apparently ditched by an early Tudor gambler to avoid being caught out; these are now in the London Museum.  Two 19C histories of gambling state that loaded dice were discovered at Pompeii or Herculaneum, but neither gives a reference or details of the dice and they give different sites.  The earliest discussion of false dice in English is in Roger Ascham's Toxophilus of 1545.  The Mirror of Saxony, a 13C legal compendium, specifies that users of false dice could have their hands cut off and the makers of false dice could have their eyes put out.

 

          10.AB.         BICYCLE TRACK PROBLEMS.

 

            There are three different problems involved here. 

            First, does the front wheel wear out more rapidly than the rear one?  Why?

            Second, ignoring the first point, can you determine from bicycle tracks which one is front and which is rear?

            Third, can you determine which way the bicycle was going? 

This section was inspired by running across several modern items and recalling Doyle's article.

 

I have recently read that the front wheel of a bicycle wears out faster than the rear wheel because the front wheel travels further -- on a curve, front wheels travel in an arc of larger radius than rear wheels, and even in fairly straight travel, the front wheel oscillates a bit to each side of the line of travel.  In addition, front wheels are often turned when the vehicle is at rest.  However, I cannot relocate my source of this, though I recall that the answer simply said the front wheel travels farther with no further explanation.

Yuri B. Chernyak & Robert M. Rose.  The Chicken from Minsk.  BasicBooks (HarperCollins), NY, 1995.  Asks why the front tires on a car wear out faster than the rear tires and says that proper turning requires the front wheels to turn by different amounts and that this, with some other undiscussed reasons, leads to the front wheels being set slightly out of parallel, which causes the extra wear.  The solution concludes: "The perfect suspension, which would turn the wheels at exactly the proper angles, has yet to be devised." 

(On the other hand, a cross-country cyclist recently told me that his rear tire wears out faster.)

See Gerrard & Brecher in Section 6.Y for a somewhat related problem.

 

From the fact that the front wheel makes a more sinuous path, or that it is the outer track on a curve, or that the rear track goes over the front track, or, perhaps, that it makes a shallower track, we can tell which of the two tracks is the front wheel, but Doyle, below, does not refer to this point.

 

A. Conan Doyle.  The Adventure of the Priory School.  In this, Holmes says he can tell which way a bicycle was going from its tracks on a pathway. 

A. Conan Doyle.  The truth about Sherlock Holmes.  The National Weekly (= Collier's Weekly) (29 Dec 1923).  Reprinted in: The Final Adventures of Sherlock Holmes; ed. by Peter Haining; W. H. Allen, London, 1981, pp. 27-40 in the PB ed., esp. p. 38.  See also: The Uncollected Sherlock Holmes; ed. by Richard Lancelyn Green; Penguin, 1983, pp. 305-315, esp. pp. 313-314, which gives a longer version of the article that appeared as Sidelights on Sherlock Holmes; Strand Mag. (Jan 1924) and is basically a part of a chapter in Doyle's autobiography which he was writing in 1923.

                    Doyle recalls the bicycle track episode in "The Adventure of the Priory School" and says that a number of letters objected to this, so he went out and tried it, finding that he couldn't tell on the level, but that "on an undulating moor the wheels make a much deeper impression uphill and a more shallow one downhill; so Holmes was justified of his wisdom after all." 

Ruth Thomson  &  Judy Hindley.  Tracking & Trailing.  Usborne Spy Guides, 1978, Usborne, London, pp. 44-45.  This says: "The front wheel of a cycle makes a loopy track as the cyclist turns it from side to side to keep his balance.  As he goes faster he turns it less, so the loops are flatter.  The narrow end of the loops point in the direction where the cyclist is heading."  When I first read this, I thought that one could tell the direction from the fact that the loops get flatter as the cycle goes downhill, but the track going uphill will look similar - the cycle travels faster at the bottom then at the top.  I am not convinced that 'the narrow end of the loops' works -- see my analysis below.

Joseph D. E. Konhauser, Dan Velleman & Stan Wagon.  Which Way Did the Bicycle Go? ...and Other Intriguing Mathematical Mysteries.  MAA, Dolciani Math. Expos. 18, 1996, prob. 1, pp. 1 & 63-64.  This is a careful treatment of determining which way the bicycle was going from the geometry of the tracks in general, but I have found there is a much simpler solution in ordinary cases. 

                    Consider when the bicycle is going essentially straight and begins to turn.  Both wheels move off the straight route onto curves, so the front wheel will have gone a bit further (namely the distance between the axles) along the straight route than the rear one did, so the outer track, which is made by the front wheel, has a short straight section at the beginning of the turn.  When the bicycle completes its turn and both wheels are now going straight, the front wheel is the same distance ahead, so the rear wheel makes a bit of a straight track before meeting the track of the front wheel.  So the inner track, made by the rear wheel, has a short straight section at the end of the turn.  Knowing this, one can tell which way the cycle was going from examination of one end of a turn, provided the track is distinct enough.

 

          10.AC.         ROBERVAL'S BALANCE.

 

            This is a mechanism commonly used in pan balances but if one extends part of it outward, then it exhibits the paradoxical behaviour that the position of a weight doesn't affect the equilibrium, apparently in violation of the law of the lever.  Imagine a rectangle with pivoted corners.  Let the long edges be horizontal and the short edges be vertical.  Attach the midpoints of the long edges to an upright, so these can pivot.  As the rectangle pivots the short ends will remain vertical.  Now attach horizontal rods to these ends.  As the rectangle pivots, these remain horizontal.  If you hang equal weights on these rods, the whole thing balances, regardless of where the weights are positioned on these rods.

 

Nouvelle maniere de Balance inventée par M. de Roberval, Professor Royal des Mathématiques dans l'Université de Paris.  Journal des Sçavans (10 Feb 1670).  ??NYS - cited and described in: Henk J. M. Bos; Descriptive Catalogue  Mechanical Instruments in the Utrecht University Museum; Utrecht University Museum, 1968, pp. 37-38.

 

          10.AD.        POUND OF FEATHERS.

 

            New section.  The usual question is: which is heavier -- a pound of feathers or a pound of gold?  This has the trick answer dependent on the different pounds used to weigh these materials.  However, I have recently seen the 1850 & 1930 items below and decided to add this section.

 

Fireside Amusements.  1850: No. 108, pp. 138 & 187?;  1890: No. 77, p. 114.  "Which is heaviest -- a pound of lead, or a pound of feathers?"  "Both the same."

Clark.  Mental Nuts.  1897, no. 69;  1904, no. 79;  1916, no. 75.  Which weighs the most?  "A pound of feathers or a pound of gold."  Answer: "Feathers, 7000 grains; gold, 5760."  (The editions vary slightly.)

Heinrich Voggenreiter.  Deutsches Spielbuch  Sechster Teil: Heimspiele.  Ludwig Voggenreiter, Potsdam, 1930.  P. 108: Was ist schwerer?  'Which is heavier, a pound of feathers or a pound of lead.'  There is no answer or explanation.  'Schwer' has several other related meanings, especially 'strong' and 'difficult'.

Hummerston.  Fun, Mirth & Mystery.  1924.  Problem, Puzzle no. 63, pp. 148 & 182.  "If one pound of potatoes balances with a box containing one pound of gold  What is the weight of the box?"  Correctly notes that Avoirdupois and Troy ounces are different, so one has to go to grains of which there are  5760  in a Troy pound (1 Troy oz = 480 grains) and  7000  in an Avoirdupois pound (1 Avoirdupois oz = 437.5 grains).  So the box must weigh  1240  grains  =  2.58333... oz  =  2 oz 11 pennyweight 16 grain  Troy  =  2.83429... oz  =  2 oz 13 dr 9 17/32 grain  Avoirdupois.

 

          10.AE.         JUGGLING OVER A BRIDGE

 

          New section.

 

Doubleday - 1.  1969.  Prob. 10, Over the limit, pp. 17 & 157-158.  = Doubleday - 4, pp. 21‑22.  A man of weight 145 finds three balls, each weighing 2 and wishes to carry them across a bridge over a ravine.  But the bridge can only carry a weight of 150.  How does he do it?  Solution says to juggle the balls, so the man is only holding at most two balls at a time.  He says he was once challenged about this by someone who claimed that air pressure and humidity invalidated the method.

David Singmaster.  Problem proposal 78.B.  MG 78 (No. 481) (Mar 1994) 112.  Shows that the above is a fatal delusion, as the average force, on the juggler, of a ball being juggled is its weight.

 

11.     TOPOLOGICAL RECREATIONS

 

            Many of the puzzles described here have the common characteristic that a loop of string is entangled in some object and the entangled string has to be worked through a number of holes in order to remove the string or to release a ring, etc.

            In 11.I, the end of the loop is worked through holes until it can be looped around the other end of the string which has an obstructive object on it.  Alternatively, the loop can be passed round the object containing the holes.  Which is easier depends on the relative sizes of the two objects involved.

            In 11.A, the other end of the loop is inaccessible and the end of the loop is then passed around the object, which is equivalent to passing it over the other end of the loop.  

            In 11.E, the other end of the loop is inaccessible and the end of the loop must be partly passed over the deformable object to allow the obstruction to pass through the deformed object.  11.I, 11.A and 11.E are thus all based on the reef or square knot and topologically equivalent.  Some of the trick purses in 11.F use this idea.

            In 11.B, the basic process is obscured by using people so that one does not readily see the necessary path.  11.H are 11.H.1 are essentially the same as this, both in their topology and their obscuring process.  The wire puzzle called The United Hearts, Cupid's Bow, etc. is isomorphic to this.

            In 11.C, the basic process is obscured by using a flexible object which is deformed to act as the loop.

            In 11.F, the basic process is again obscured, this time by the fact that the holes do not appear to be part of the puzzle and by the fact that one does not remove the loop, but instead a ring is released.

            11.D is somewhat similar, but the process of moving the end of the loop is quite different and the object is to move objects along the string, so this is basically a different type of puzzle.

            7.M.5 is in this general category, but the systematic binary pattern of disentanglement makes it quite different from the items below.

            James Dalgety has shown me some examples of Puzzle Boxes from John Jaques & Son, London, c1900, which contain many of these puzzles.  The decorative features on these are very similar to those in Hoffmann's illustrations.  It is clear that Hoffmann (or his artist) drew from such examples (Jaques are not known to have published any illustrations of these puzzles), so Jaques must have been producing them in the 1880s.  I will note 'drawing based on Jaques' puzzle' in the entry for Hoffmann for such puzzles.

 

          11.A. SCISSORS ON STRING

 

Ozanam.  1725.  Vol. IV, prob. 35, p. 437 & fig. 42, plate 18 (error for 13) (15).

Minguet.  1733.  Pp. 108-109 (1755: 76-77; 1822: 83-84 & 127-128; 1864: 72-73 & 107-108).  Somewhat similar to Ozanam.

Alberti.  1747.  Art. 35, p. 209 (110) and fig. 43, plate XII, opposite p. 212 (110).  Taken from Ozanam.

Manuel des Sorciers.  1825.  Pp. 210-211, art. 25.

de Savigny.  Livre des Écoliers.  1846.  P. 265: Le nœud des ciseaux.

The Sociable.  1858.  Prob. 40: The scissors entangled, pp. 298 & 316.  "This is an old but a capital puzzle."  Says the ends are held in the hand, but figure shows them tied to a post.  = Book of 500 Puzzles, 1859, prob. 40, pp. 16 & 34.  = The Secret Out, 1859, pp. 238‑239: The Disentangled Scissors, but says the ends 'are held by the hand or tied firmly to a post ...', and with a diagram for the solution.  See Magician's Own Book (UK version) for a clearer version.  = Wehman, New Book of 200 Puzzles, 1908, p. 44. 

Indoor & Outdoor.  c1859.  Part II, p. 129, prob. 9: The scissors entangled.  Almost identical to The Sociable, but the figure omits the post and the problem statement starts with 56. -- apparently the problem number in the source from which this was taken.

Magician's Own Book (UK version).  1871.  The liberated prisoner, pp. 211-212.  Shows a prisoner chained in this manner, but the diagram is too small to really see what is going on.  Then says it is equivalent to the scissors problem, which is clearly drawn and much bigger than in The Sociable.  The explanation is clearer than in The Sociable.

Tissandier.  Récréations Scientifiques.  1880?  2nd ed., 1881, gives a brief unlabelled description on pp. 330-331, with figure copied from Ozanam on p. 328. 

                    5th ed., 1888, La cordelette et les ciseaux, p. 259.  Based on Ozanam, copying the diagram.

                    The index of the English ed. has a reference to this, but the relevant pages 775‑776 have become the title for the Supplement!  This is included on p. 84 of Marvels of Invention -- cf Tissandier in Common References.

Anonymous.  Social Entertainer and Tricks (thus on spine, but running title inside is New Book of Tricks).  Apparently a compilation with advertisements for Johnson Smith (Detroit, Michigan) products, c1890?  P. 38b: The scissors trick.

Hoffmann.  1893.  Chap. X, no. 46: The entangled scissors, pp. 355 & 393 = Hoffmann‑Hordern, p. 254.

A. Murray.  Tricks with string.  The Boy's Own Paper 17 or 18?? (1894??) 526-527.  Well drawn.

Devant.  Tricks for Everyone.  Op. cit. in 4.A.1.  1910.  The scissors trick, pp. 35-36.  Simple version.

M. Adams.  Indoor Games.  1912.  The tailor's scissors, pp. 28‑30.

Williams.  Home Entertainments.  1914.  The entangled scissors, pp. 111-112.

Hummerston.  Fun, Mirth & Mystery.  1924.  The entangled scissors, p. 127.

Collins.  Book of Puzzles.  1927.  The dressmaker's puzzle, pp. 21-22.

J. F. Orrin.  Easy Magic for Evening Parties.  Op. cit. in 7.Q.2.  1930s??  The scissors puzzle, pp. 36-37.

 

          11.B. TWO PEOPLE JOINED BY ROPES AT WRISTS

 

            This is isomorphic to 11.K.8.  See von Hartwig, Goldston and Svengarro for one person versions.

 

Ozanam.  1725.  Vol. IV, prob. 38, p. 438 & fig. 45, plate 18 (error for 13) (15).

Minguet.  1733.  Pp. 110-111 (1755: 77-78; 1822: 129-130; 1864: 108-109).  Similar to Ozanam.

Alberti.  1747.  Art. 38, p. 212 (111) and fig. 46, plate XII, opposite p. 212 (110).  Taken from Ozanam.

Family Friend 2 (1850) 267 & 353.  Practical Puzzle -- No. IX.  = Illustrated Boy's Own Treasury, 1860, Practical Puzzles, No. 37, pp. 402 & 442.

Parlour Pastime, 1857.  = Indoor & Outdoor, c1859, Part 1.  = Parlour Pastimes, 1868.  Mechanical puzzles, no. 10, pp. 180-181 (1868: 191-192). 

Magician's Own Book.  1857.  No. 11: The handcuffs, p. 11.  = The Secret Out, 1859, pp. 248‑250, but with a few changes of words, a diagram for the solution and more elegant drawing.  = Boy's Own Conjuring Book, 1860, p. 23.

Leske.  Illustriertes Spielbuch für Mädchen.  1864?  Prob. 177, p. 96: Die verschlungenen Schnüre.

Anonymous.  Every Little Boy's Book  A Complete Cyclopædia of in and outdoor games with and without toys, domestic pets, conjuring, shows, riddles, etc.  With two hundred and fifty illustrations.  Routledge, London, nd.  HPL gives c1850, but the text is identical to Every Boy's Book, whose first edition was 1856, and which has not yet been entered.  In 4.A.1, I've guessed this book may be c1868.  Pp. 360-361: The handcuffs.

Magician's Own Book (UK version).  1871.  The prisoner's release, pp. 209-211.  Adds that one can also intertwine the two cords in the form of a square or reef knot which allows a simpler disentanglement.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 4, no. 13: The handcuffs, p. 97.

Cassell's.  1881.  P. 94: The prisoners' release puzzle.  = Manson, 1911, 143-144

Tissandier.  Récréations Scientifiques.  1880?  2nd ed., 1881, brief unlabelled description on p. 330 with figure copied from Ozanam on p. 328. 

                    5th ed., 1888, Les deux prisonniers, pp. 257-258.  Based on Ozanam, copying the diagram. 

                    The index of the English ed. has a reference to this, but the relevant pages 775‑776 have become the title for the Supplement!  This is included on p. 84 of Marvels of Invention -- cf Tissandier in Common References.

Anonymous.  Social Entertainer and Tricks (thus on spine, but running title inside is New Book of Tricks).  Apparently a compilation with advertisements for Johnson Smith (Detroit, Michigan) products, c1890?  P. 14a: Slipping the bonds.

Richard von Hartwig.  UK Patent 3859 -- A New Game or Puzzle.  Applied: 27 Feb 1892;  accepted: 2 Apr 1892.  1p + 1p diagrams.  One man, the other loop being tied to a tree.

Hoffmann.  1893.  Chap. X, no. 36: Silken fetters, pp. 349‑350 & 390 = Hoffmann-Hordern, pp. 246-247.

A. Murray.  Tricks with string.  The Boy's Own Paper 17 or 18?? (1894??) 526-527.

M. Adams.  Indoor Games.  1912.  Release the prisoners, pp. 28‑29.

Will Goldston.  The Young Conjuror.  [1912 -- BMC];  2nd ed., Will Goldston Ltd, London, nd [1919 -- NUC],  Vol. 1, pp. 34-39: Three Malay rope tricks.  No. two is the present section, with one person having his wrists tied and a string looped around and held by another person.  Goldston thanks E. R. Bartrum for the text and illustrations.

Prof. Svengarro.  Book of Tricks and Magic.  I. & M. Ottenheimer, Baltimore, 1913.  Rope trick, p. 15.  As in Goldston, with wrists tied by a handkerchief and then a rope looped around it.

Williams.  Home Entertainments.  1914.  The looped chains, pp. 109-110.  Jailer tries to secure prisoner by this method.

J. F. Orrin.  Easy Magic for Evening Parties.  Op. cit. in 7.Q.2.  1930s??  The magic release (no. 1), pp. 26-27.

McKay.  Party Night.  1940.  How did it get there?, p. 150.  This is an alternate method which gives a 'knot' between the two strings.  It is most easily described from the undone state and the second loop is most easily visualised as a ring.  Form a bight in the string and pass it through the ring, then pass it under the loop around one wrist, over the hand and back under the loop.  This leaves the bight around the wrist below the loop.  Now just lift it off the hand and the string will be knotted to the ring.

 

          11.C. TWO BALLS ON STRING THROUGH LEATHER HOLE AND STRAP  =  CHERRIES PUZZLE

 

            The basic version has a leather strap with two longish cuts allowing the central part to flex away from the rest of the strap.  There is a hole or two holes at the bottom of the strap.  A string with balls at each end comes through the hole(s) and around the central part.  The balls are larger than the hole(s), but the central part can be brought through the hole(s) to form a loop big enough to pass a ball through.  (I have just noted that two hole versions occur, but I haven't checked all versions.)

            The balls were often called cherries and even drawn as such.  I wonder if the puzzle originally used a pair of joined-together cherries?? 

            An equivalent version has a slit in a card (sometimes tubular), producing a thin part on which hangs the string with two balls with a ring or cylinder about the double string.  A variation of this has a doubled paper or leather object such as a pair of boots attached at the tops, with a ring or just a paper loop or annulus around it.  The key to these versions is folding the card so the thin bit can be brought through the ring, cylinder or loop. 

            A version with names like Key, Heart and Arrow has a card heart with slits and a card arrow as the doubled object and the key acting as the ring.  SEE:  Girl's Own Book; Secret Out; Magician's Own Book (UK);

 

Pacioli.  De Viribus.  c1500. 

Ff. 210r - 210v, Part 2, Capito. CIIII. Do(cumento). cavare' et mettere' .2. cirege' in una carta tramezzatta (To remove and replace two cherries in a cut card).  = Peirani 288-289.  Description clearly shows there is one hole, but Dario Uri illustrates this with the picture from Alberti which has two holes.

Ff. 215v - 216v, Part 2, Capitulo. CX. Do(cumento). uno bottone' de un balestro. o vero doi cirege' de un botone' et valestro (A button from a (cross)bow or two cherries from a button and bow).  = Peirani 296-297.  Dario Uri translates 'balestro' as 'flexible stick' and illustrates this with Alberti's Fig. 37.

Schwenter.  1636.  Part 10, exercise 30, p. 411.  Version using a card.

Witgeest.  Het Natuurlyk Tover-Boek.  1686. 

          Prob. 18, pp. 160-161.  Elaborate card version.

          Prob. 19, pp. 162-163.  Cherries, with two holes.

Ozanam.  1725.

Vol. IV, prob. 30, p. 434 & fig. 36, plate 11 (13).  Version using a tube.

Vol. IV, prob. 33, p. 436 & fig. 39, plate 12 (14).  "On peut passer des queues de Cerises dans un papier, ...."  Two holes.

Minguet.  1733.  Pp. 112-113 (1755: 79; 1822: 131-132; 1864: 110-111).  Cherries version.  Similar to Ozanam, prob. 33.  Two holes and one hole are both shown.

Alberti.  1747.  Loc. cit. in 11.A.

Art. 30, pp. 204‑205 (108) and fig. 37, plate X, opp. p. 206 (between pp. 108 & 109)  Version using a tube.  Taken from Ozanam.

Art. 33, pp. 207-208 (109‑110) and fig. 40, plate XI, opp. p. 210 (109).  "Si possono passare dai gambi di cerase in una carta....".  Two holes.  Taken from Ozanam.

Catel.  Kunst-Cabinet.  1790.  Die verbundenen Kirschen, pp. 13-14 & fig. 18 on plate I.

Bestelmeier.  1801.  Item 273: Der verschlungenen Kirschen.  Copies part of Catel's text.

Manuel des Sorciers.  1825.  Pp. 182-183: Le jeu des cerises.

The Boy's Own Book.

The cherry cheat.  1828: 418;  1828-2: 423;  1829 (US): 215;  1855: 570;  1868: 672.

The card puzzle.  1828: 422‑423;  1828-2: 427-428;  1829 (US): 219;  1843 (Paris): 437 & 441, no. 12;  1855: 574;  1868: 676.  c= Magician's Own Book, 1857.  = Wehman, New Book of 200 Puzzles, 1908, pp. 38-39. 

Child.  Girl's Own Book.  Heart, dart, and key.  1833: 138-139;  1839: 122-123;  1842: 203‑204.  A variation of the card version with the key as the ring.  Cf: Secret Out: Magician's Own Book (UK).

Nuts to Crack II (1833), no. 92.  The card puzzle.  Almost identical to Boy's Own Book.

Crambrook.  1843.  P. 4, no. 16: Cherry Cheat Puzzle.  Check??

Family Friend 2 (1850) 208 & 239.  Practical Puzzle, No. VII.  Repeated as Puzzle 10 -- The button puzzle in (1855) 339 with solution in (1856) 28 .  = The Illustrated Boy's Own Treasury, 1860, Practical Puzzles, No. 43, pp. 403 & 442.  Identical to Magician's Own Book, prob. 11.  Cherries puzzle using buttons.

Magician's Own Book.  1857. 

Prob. 11: The button puzzle, pp. 269 & 294.  Identical to Family Friend with slight changes of wording.  = Wehman, New Book of 200 Puzzles, 1908, p. 15. 

Prob. 19: The card puzzle, pp. 272-273 & 296.  Identical to Boy's Own Book card puzzle, except the answer is split from the problem, but the problem refers to the figures which are now in the answer!

Book of 500 Puzzles.  1859. 

Prob. 11: The button puzzle, pp. 83 & 108.  Identical to Magician's Own Book.

Prob. 19: The card puzzle, pp. 86-87 & 110.  Identical to Magician's Own Book.

The Secret Out.  1859.  Key, Heart, and Arrow, pp. 390-391.  As in Girl's Own Book, but with much better pictures and clearer text.  Cf Magician's Own Book (UK).

Boy's Own Conjuring Book.  1860.

Prob. 10: The button puzzle, pp. 230 & 257.  Identical to Magician's Own Book.

Prob. 18: The card puzzle, 234 & 259.  Identical to Magician's Own Book.

Magician's Own Book (UK version).  1871. 

The string and button puzzle, p. 326.

Key, heart, and dart, pp. 232-233.  As in The Secret Out, the text is not as detailed.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 10: The button puzzle, pp. 29 & 31.

Hanky Panky.  1872.  The undetachable cylinder, pp. 125‑126.  Card version.

Martin Appleton Wright.  UK Patent 7002 -- Improved Advertisement Cards.  Dated: 28 Apr 1884 and 30 Apr 1884.  1p + 2pp diagrams.  Card versions.

Gaston Tissandier.  Jeux et Jouets du jeune age  Choix de récréations  amusantes & instructives.  Ill. by Albert Tissandier.  G. Masson, Paris, nd [c1890].  No. 10-11: Le problème des cerises, pp. 40-41, with elegant coloured plate.

William Crompton.  The odd half-hour.  The Boy's Own Paper 13 (No. 657) (15 Aug 1891) 731-732.  The slippery buttons.

Handy Book for Boys and Girls.  Op. cit. in 6.F.3.  1892.  Pp. 57-59: The boot puzzle.  Card version with a pair of boots.

Tom Tit, vol. 3.  1893.  Le jeu de la fève, pp. 225-226.  = K, no. 171: The bean trick, pp. 394‑395.  Card version, using a bean pod to make all the parts.

Hoffmann.  1893.  Chap. II, no. 24: The ball and three strings, pp. 34‑35 & 59‑60 = Hoffmann-Hordern, pp. 38-39, with photos.  This is a more complex puzzle, but based on the same principle.  A string goes around other strings, through a ball and then has ends separately knotted, so you have to bring the other strings through the ball in order to release the string.  Drawing based on Jaques' puzzle.  Photos on p. 39 show The Three-String Ball Puzzle, with instructions, by Jaques & Son, in its original state and with the strings through the ball, 1870-1900.

Benson.  1904.  The ball and strings puzzle, p. 217.  As in Hoffmann.

Anon.  Triangular string puzzle.  Hobbies 31 (No. 793) (24 Dec 1910) 302-303.  Same as Hoffmann, with the strings inside a triangular prism

Williams.  Home Entertainments.  1914.  String and button puzzle, p. 112.  Cherries puzzle using buttons.

Hummerston.  Fun, Mirth & Mystery.  1924.  The buttons puzzle, Puzzle no. 73, pp. 160 & 183.

Collins.  Book of Puzzles.  1927.  The bachelor's button puzzle, pp. 22-23.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 88: The button trick, pp. 80-82.  This is a card version, but with a string and buttons.  The card has two long parallel cuts, as in the leather version, but it then has two slots off to the side that the string goes through.  One has to fold the card to superimpose these two slots, then fold it again to bring the thin strip to these slots.

 

          11.D. SOLOMON'S SEAL

 

            See S&B, p. 114.  Slocum and Gebhardt have pointed out that there are two approaches to this problem, particularly in the two ball case, depending on how the ends of the string are attached to the board.  If the string is passed through a hole and a knot is tied in the string, rather than tying the string to the board, then one can partially undo the central loop by passing it through the end hole and around the end of the string -- repeating with a slight change completely undoes the central loop.  This approach makes the problem really a form of 11.I.  The only example in print that I have seen is in Hobbies, 1910.  However, I think that in most cases with a knot in the string, the indicated size of the hole in the board is too small to permit a loop to pass through and this method is not possible. 

            I have recently acquired an example of the Waterloo Puzzle, produced by Jaques, c1900, and seen other examples at James Dalgety's.  This has seven holes in a bone plank.  There are four small holes at the corners and three larger holes on the midline, say A, B, C in order.  Two strings run lengthwise between pairs of end holes, but one dips into the centre hole B.  The small holes seem too small to admit another string, and the dip makes a very small loop on the other side.  (The other string has some slack in it and probably could also be brought through to make a double dip.)  A loop starts around both upper strings, goes down through hole A, through a ring, through the small dip (which is barely big enough for two strings), through another ring and up through hole C and around both strings.  I cannot decide if the string has been broken and then erroneously restrung.  Dalgety has provided a copy of the instructions which say the object is to remove the rings.  The instructions also state that the loop has to be passed through each of the end holes to pass around the ends of the strings -- at present, there is not enough slack to do this and it would be difficult to get that much string into the small holes.

 

Pacioli.  De Viribus.  c1500.  Ff. 206v-207r, Part 2, Capitulo. CI. Do(cumento) un altro filo pur in .3. fori in la stecca con unambra. per sacca far le andare' tutte in una (Another string also through three holes in the stick with one bead per loop, make them go onto one (loop)).  = Peirani 283‑284.  The problem titles vary between the actual problem and the Table of Contents and the latter shows that 'unambra' should be 'una ambra' -- Peirani has given it as un'ambra.  Sacca means pocket or bay or inlet and it seems clear he means a loop which has that sort of shape.  Ambra is amber, but seems to mean an amber bead here.  Sadly there is no picture though the text refers to one.

Schwenter.  1636.  Part 10, exercise 27, pp. 408-410.  With two rings.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 43, pp. 33-34.  Clearly taken from Schwenter.

Ozanam.  1725.  Vol. IV, prob. 40, pp. 439‑440 & fig. 47, plate 14 (16).  Le Sigillum Salomonis, ou Sceau de Salomon -- version with four rings.

Alberti.  1747.  Art. 40, pp. 214 (112) and fig. 48, plate XIII, opposite p. 214 (111).  Il Sigillum Salomonis, o Sigillo Salomone -- version with 4 rings.  Taken from Ozanam.

Catel.  Kunst-Cabinet.  1790.  Die Salomonsringe, pp. 14-15 & fig. 24 on plate I.  Version with 4 rings.  Describes how to solve it.

Bestelmeier.  1801.  Item 214: Die Salomons‑Ringe.  Version with 3 rings.  Brief text.

Boy's Own Book.  1843 (Paris): 438 & 442, no. 17: The bead puzzle.  "This puzzle may be procured at many toy-shops."  = Boy's Treasury, 1844, pp. 426 & 430, no. 14.

Family Friend 3 (1850) 30 & 61.  Practical puzzle -- No. XI.  Love's Puzzle with two hearts.  = The Sociable, 1858, Prob. 24: Love's Puzzle, pp. 294 & 310.  = Book of 500 Puzzles, 1859, prob. 24, pp. 12 & 28.

Magician's Own Book.  1857.  Prob. 37: The string and balls puzzle, pp. 277-278 & 301.  Two balls.  = Boy's Own Conjuring Book, 1860, prob. 36, pp. 240-241 & 265.  = Wehman, New Book of 200 Puzzles, 1908, p. 9. 

Book of 500 Puzzles.  1859. 

Prob. 24: Love's Puzzle, pp. 12 & 28.  As in Family Friend.

Prob. 37: The string and balls puzzle, pp. 91-92 & 115.  Identical to Magician's Own Book.

The Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 5: Love's puzzle, pp. 396 & 436.  Identical with Family Friend.

Magician's Own Book (UK version).  1871.  The puzzle of Cupid, p. 227.  Two hearts.  Diagram is hard to make out.

Elliott.  Within‑Doors.  Op. cit. in 6.V.  1872.  Chap. 1, no. 11: The string and balls, pp. 29 & 31.

Cassell's.  1881.  P. 90: The string and balls puzzle.  = Manson, 1911, p. 147.  Version with 2 balls.

Hoffmann.  1893.  Chap. II, no. 13: The two balls, pp. 27‑28 & 53‑54 = Hoffmann-Hordern, p. 28, with photo.  Photo on p. 28 shows an ivory example, 1850-1900.  Hordern Collection, p. 22, shows a different example, apparently in ivory, 1870-1900.

Burnett Fallow.  An ingenious bead puzzle.  The Boy's Own Paper 15 (No. 755) (1 Jul 1893) 638.  Shows two loop version but notes it can be extended.

Anon.  Hexagonal wood puzzle.  Hobbies 31 (No. 791) (10 Dec 1910) 261-262 & 279.  Two loop version, but asks to undo the central loop from the wood, as discussed in the introduction to this section.

Collins.  Book of Puzzles.  1927.  The string and ball puzzle, pp. 23-24.

James Dalgety.  Email of 3 Sep 1999.  Reports that his father saw the puzzle in use by Inuits in the Canadian Arctic or Greenland in c1930, but his family lost the walrus ivory (or bone) and leather examples that his father brought back.  Also that a collection of topological puzzles from near Lake Tanganyika, gathered in the 1920s and now in the Horniman Museum, London, does not contain an example of Solomon's Seal.

R. P. Lelong.  Casse-tête guerzé.  Notes Africaines 22 (Apr 1944) 1.  ??NYS -- cited and described by Béart.  Says M. Gienger found the variant with an extra ring encircling both loops in the forest of the Ivory Coast in 1940, named  kpala kpala powa [body of a toucan]  or  kpa kpa powa [body of a parrot].

Paul Niewenglowski.  Bulletin de l'IFAN [Institut Français d'Afrique Noire] 14:1 (Jan 1952).  ??NYS -- cited and described by Béart.  Describes his invention of an interesting, rather simpler, variant as a result of seeing a standard version from Béart.

Charles Béart.  Jeux et jouets de l'ouest africain.  Tome I.  Mémoires de l'Institut Français d'Afrique Noire, No. 42.  IFAN, Dakar, Senegal, 1955.  Pp. 413-418 discusses and carefully illustrates several versions.  The standard version, but with several beads on one loop, is called pèn and is common in the forests of Guinea and Ivory Coast.  Describes variants of Gienger/Lelong and Niewenglowski. 

Ch'ung‑En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1.  P. 28 shows a version, called Double Coin Ring Puzzle.

Fred Grunfield.  Games of the World.  Ballantine, NY, 1975.  On p. 267, he calls this "African String Game", but gives no reference.  Probably based on Béart.

Pieter van Delft & Jack Botermans.  Creative Puzzles of the World.  Abrams, New York, 1978.  African ball puzzles.  "It was once used in magic rites by tribes living in the jungles of the Ivory Coast.  The puzzle is still used for amusement in this part of Africa, not only by the people who inhabit the remote outlying areas but also by city dwellers.  ...  The puzzles were not restricted to this part of Africa.  Variations may be found in Guinea, and some ... were made in China."  No reference given, but I suspect it must come from Béart, although this is not listed in their bibliography.  My thanks to Mark Peters for the reference to van Delft and Botermans.

 

          11.E.  LOYD'S PENCIL PUZZLE

 

            See S&B, p. 114.  I have seen it claimed that the phrase 'to buttonhole someone' derives from the use of this.

 

Devant.  Tricks for Everyone.  Op. cit. in 4.A.1.  1910.  The flexible pencil, pp. 13-15.  No history.

Will Goldston.  More Tricks and Puzzles without Mechanical Apparatus.  Op. cit. in 6.AK.  1910?.  The pencil, loop and buttonhole, pp. 69‑71.

W. P. Eaton.  Loc. cit. in 1.  1911.  Gives Loyd's narration of the invention of this for John A. McCall, President of the New York Life Insurance Co.

A. C. White.  Sam Loyd and His Chess Problems.  1913.  Op. cit. in 1.  P. 103.  Quotes from Eaton.

A. B. Nordmann.  One Hundred More Parlour Tricks and Problems.  Wells, Gardner, Darton & Co., London, nd [1927 -- BMC].  No. 87: Latch key trick, pp. 79-80 & 111.  This is the only version I have seen using something other than a pencil.  It has the advantage that a key has a loop at the end to tie the loop of string to, but the buttonhole will have to be large!

Rohrbough.  Puzzle Craft.  1932.  The Buttonholer, p. 4 (= p. 4 of 1940s?).

Abraham.  1933.  Prob. 165 -- Pencil and buttonhole, pp. 77 (49).

J. F. Orrin.  Easy Magic for Evening Parties.  Op. cit. in 7.Q.2.  1930s??  Looping the loop, pp. 34-36.  No mention of Loyd.

Slocum.  Compendium.  Shows Magic Coat Pencil from Johnson Smith 1937 catalogue.

Depew.  Cokesbury Game Book.  1939.  Lapel needle, p. 167.  No mention of Loyd.

"Willane".  Willane's Wizardry.  Academy of Recorded Crafts, Arts and Sciences, Croydon, 1947.  The looped pencil, pp. 10-11.

Gardner.  SA (Nov 1971) = Wheels, Chap. 12.

 

          11.F.  THE IMPERIAL SCALE

 

          This could be combined into 11.I.

 

Catel.  Kunst-Cabinet.  1790.  Das einfache Ringspiel, p. 13 & fig. 39 on plate II. 

Bestelmeier.  1801.  Item 199: Das einfache Ringspiel.  Copies part of Catel's text.

The Boy's Own Book.  The scale and ring puzzle.  1828: 424‑425;  1828‑2: 429;  1829 (US): 220-221;  1843 (Paris): 437 & 442, no. 13;  1855: 575;  1868: 677-678.  In the 1843 (Paris), the engraved heading of the section, p. 435, shows the puzzle.  Boy's Treasury, 1844, omits the puzzle, but copies this engraving on p. 424.  = de Savigny, 1846, pp. 354 & 358, no. 10: Le problème de la balance et de l'anneau.

Nuts to Crack III (1834), no. 82.  The scale and ring puzzle.  Almost identical to Boy's Own Book.

Crambrook.  1843.  P. 4, no. 8: Imperial Scale.  Check??

See Bogesen, 6.W.2, for an actual example, mid 19C.

Magician's Own Book.  1857.  Prob. 15: The scale and ring puzzle, pp. 270-271 & 295.  Identical to Boy's Own Book, but with an illogical break between puzzle and solution.  = Book of 500 Puzzles, 1859, pp. 84-85 & 109.  = Boy's Own Conjuring Book, 1860, prob. 14, pp. 232‑233 & 258.  = Magician's Own Book (UK version), 1871, pp. 229‑230.  c= Wehman, New Book of 200 Puzzles, 1908, p. 23, which omits the solution, which is not really needed. 

F. Chasemore.  Loc. cit. in 6.W.5.  1891.  Item 1: The balance puzzle, p. 571.

Hoffmann.  1893.  Chap. II, no. 19: The imperial scale, pp. 31‑32 & 56‑57 = Hoffmann‑Hordern, pp. 32-33, with photo.  Drawing based on Jaques' puzzle.  Photo on p. 32 shows The Imperial Scale Puzzle, with instructions, by M.M.M.M., London, in the "St. Dutius" Series of Novelties, 1870-1900.  Hordern Collection, p. 27, shows another version without instructions.

William Hollins.  UK Patent 21,097 -- An Improved Puzzle.  Applied: 23 Sep 1896;  accepted: 5 Dec 1896.  1p + 1p diagrams.  Identical to all the above!!

 

          11.G. TRICK PURSES

 

van Etten.  1624.  Prob. 60 (55), p. 55 & fig. opp. p. 33 (p. 80).

Schwenter.  1636.  Part 15, exercise 18, p, 545.  Almost identical to van Etten.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 33, pp. 25-26.  Similar to van Etten, but with a more decorative purse.

Ozanam.  1725.  Vol. IV.

Prob. 32, pp. 435‑436 & fig. 38, plate 12 (14).  This has a sliding slit leather tab.

Prob. 39, p. 439 & fig. 46, plate 14 (16).  This is based on trick stitching, a concealed running stitch.

Prob. 42, pp. 440‑441 & fig. 49, plate 15 (17).  This is like the van Etten example.

Alberti.  1747.

Art 32, pp. 206-207 (109) & fig. 39, plate XI, opp. p. 211 (109).  Taken from Ozanam, prob. 32.

Art. 39, pp. 212-213 (111‑112) & fig. 47, plate XIII, opp. p. 214 (111).  Based on trick stitching, taken from Ozanam, prob. 39.

Art. 42, pp. 215‑217 (112‑113) and fig. 50, plate XIIII, opp. p. 218 (112).  Taken from Ozanam, prob. 52.

Catel.  Kunst-Cabinet.  1790. 

Die Jägertasche, p. 21 & fig. 35 on plate II.  Like Ozanam, prob. 42.

Die Vexierbörse, p. 21 & fig. 22 on plate I.  Apparently a trick stitch as in Ozanam, prob. 39.

Bestelmeier.  1801.  Item 387: Ein Zauberbeutel.  Copies Catel's Jägertasche.  Like Ozanam, prob. 42.

Child.  Girl's Own Book.  1833: 219-220;  1839: 198-199;  1842: 316.  The miser's purse, apparently a trick stitch as in Ozanam, prob. 39 ??

Crambrook.  1843.  P. 3.  The Miser's Purse;  The Ring Purse.  These might be trick items??

The Sociable.  1858.  Prob. 15: Puzzle purse, pp. 291 & 307.  Like Ozanam, 1725, prob. 52 but with a second similar locking mechanism.  = Book of 500 Puzzles, 1859, prob. 15, pp. 9 & 25. = The Secret Out, 1859, pp. 372-373: The Magic Purse.

The Illustrated Boy's Own Treasury.  1860.  Practical Puzzles, No. 44: Puzzle purse, pp. 403 & 442‑443.  Identical to The Sociable.

Peck & Snyder.  1886.

P. 246: No. 161 -- A magical "pouch".

P. 247: No. 168 -- Grandfather's purse.

  Both are shown in Slocum's Compendium.

Hoffmann.  1893.  Chap. II, no. 42: Puzzle purse, pp. 43 & 68 = Hoffmann-Hordern, p. 54.  This is Ozanam's prob. 32 and finally made it clear to me how it worked. 

Burnett Fallow.  How to make the "B.O.P." puzzle purse.  The Boy's Own Paper 16 (No. 771) (21 Oct 1893) 43-45.  As in Ozanam's no. 32.

Jaques puzzle boxes of c1900 included The Secret Purse, based on the trick stitching of Ozanam prob. 39.

Will Goldston.  Tricks & Illusions  for amateur and professional conjurers.  Routledge  &  Dutton,  9th ed (revised), nd [1920s?], pp. 136-137.  Uses the trick stitching of Ozanam prob. 39 to make a sack large enough to hold a person, who easily escapes from it behind a screen and then tightens up the stitching before coming out.

Davenport's catalogue, op. cit. in 10.T, c1940, p. 3.  Tantalizing pants.  They are based on a trick stitch like Ozanam prob. 39.  Described as the very latest novelty.  Remnants of this stock, made in Japan, are still available from Davenport's as Puzzle pants.

In 1996, Harriet Hall gave me an example of Ozanam's prob. 32, made from deerskin, labelled:  Genuine Indian Handicraft  Native Renaissance II  Deseronto, Ontario  Made in Canada.  She obtained it from Puzzletts in Seattle.

 

          11.H. REMOVING WAISTCOAT WITHOUT REMOVING COAT

 

Philip Breslaw (attrib.).  Breslaw's Last Legacy.  1784?  Op. cit. in 6.AF.  1795: 128-129: 'To pull off any Person's Shirt, without undressing him, or having Occasion for a Confederate.'

Manuel des Sorciers.  1825.  Pp. 132-133, art. 9: Enlever la chemise à quelqu'un sans le déshabiller.

The Boy's Own Book.  1828: 362.  To pull off a person's shirt without undressing him.

Magician's Own Book (UK version).  1871.  To take a man's vest off without removing his coat, pp. 239-240.  The fact that this English book uses the American term 'vest' makes me suspect this is taken from a US source.

Cassell's.  1881.  P. 95: To take a man's waistcoat off without removing his coat.  = Manson, 1911, pp. 148-149: Waistcoat puzzle.

Anonymous.  Social Entertainer and Tricks (thus on spine, but running title inside is New Book of Tricks).  Apparently a compilation with advertisements for Johnson Smith (Detroit, Michigan) products, c1890?  P. 78: How to remove a man's shirt without taking off his coat or vest.

H. D. Northrop.  Popular Pastimes.  1901.  No. 20: The vest puzzle, pp. 69 & 74-75.  "How can you take a man's vest off with out removing his coat?"  Solution is = Cassell's.

Devant.  Tricks for Everyone.  Op. cit. in 4.A.1.  1910.  The waistcoat trick, pp. 106-109.

Hugh Shaw.  The Helping Hand: How to take a man's waistcoat off without removing his coat.  Hobbies 32 (No. 821) (8 Jul 1911) 303. 

Games and Tricks -- to make the Party "Go".  Supplement to "Pearson's Weekly", Nov. 7th, no  year indicated [1930s??].  Removing a waistcoat without the coat, p. 7.

Foulsham's New Party Book.  Foulsham, London, nd [1950s?].  P. 49: Removing a boy's waistcoat without taking off his coat.

Alfred C. Gilbert.  Gilbert  Knots and Splices with Rope-Tying Tricks  Explains methods of knot tying and reveals rope tricks made famous by great artists.  A. C. Gilbert Co, New Haven, Connecticut, 1920, pp. 79-82.  Turning vest inside out without removing coat and with the wrists securely tied. 

Gardner.  MM&M.  1956.

Reversing the vest, pp. 86-87.  Reverse a waistcoat (= vest) while the wearer has his hands clasped, whether he is wearing a coat or not.  In theory, you don't even need to unbutton the waistcoat.  Cf Gilbert, 1920.

Removing the vest, p. 87.  Remove a waistcoat without removing the coat.

 

          11.H.1.        REMOVING LOOP FROM ARM

 

Family Friend 3 (1850) 341 & 351.  Practical puzzle.  No. XXII.  Loop around arm with hand in waist coat pocket.  Remove the loop without moving the hand.  Posed in verse.  = Magician's Own Book, 1857, prob. 4: The endless string, pp. 267 & 292.  = Book of 500 Puzzles, 1859, pp. 81 & 106.  = Boy's Own Conjuring Book, 1860, prob. 5, pp. 229 & 255, but this has a different picture.  = Illustrated Boy's Own Treasury, 1860, no. 20, pp. 399 & 438, with slightly abbreviated answer.

Magician's Own Book (UK version).  1871.  The string without an end, pp. 204-205.  Only shows a loop of string and the problem is a bit unclearly set.  Solution is similar to Family Friend.

J. F. Orrin.  Easy Magic for Evening Parties.  Op. cit. in 7.Q.2.  1930s??  Another magic release (no. 4), pp. 30-31. 

McKay.  Party Night.  1940.  Removing the string, pp. 147-148.  As in Family Friend.  Notes that many people will fail to do it because they put their hand in their trousers pocket!

Gardner.  MM&M.  1956.  The puzzling loop, p. 86 & fig. 41 on p. 90.  Coat off, loop around arm with hand in waistcoat pocket.  Loop must be big enough to pass round his chest.

 

          11.I.   HEART AND BALL PUZZLE AND OTHER LOOP PUZZLES

 

            The ALLIANCE or VICTORIA puzzle has two boards, each with two holes, and a length of string with a largish loop at each end.  To disentangle, a loop has to be worked along the string and taken around the entire board at the other end (or, equivalently, around the first board).  One can also use a large single loop of string.  There are also versions with three holes in each board.  One can also have one board with the free ends of the string tied to something else.  See most entries below.

 

            The HEART AND BALL puzzle has a heart-shaped piece with several holes in it and a loop of string which goes through the holes and then loops around itself.  The other end of the loop has a ball on both strings of the loop and a large knot to prevent it coming free.  One has to work the first end of the loop along the string until it comes out along the free end of the loop when it can be passed around the free end and worked back to release the entire string and hence the ball.  A simple version of this has just three holes in a straight line on a board and is often sold with the Alliance puzzle.

            SEE:  Boy's Own Book;  Nuts to Crack;  Crambrook;  Family Friend, 1850;  Illustrated Boy's Own Treasury;  Magician's Own Book (UK);  Hoffmann;  Benson;  Slocum;  Collins; 

 

            The BOARD AND BALL puzzle has two holes in a board.  A string goes through both holes, crosses itself and then each end comes back through its hole and is terminated with a bead and a knot.  A largish bead is located on one of the loops through the holes.  To remove, this loop is put through the other hole and passed over the end and brought back, which allows everything to come apart.  The alternate approach to the Solomon's Seal described at 11.D is a variant of this approach. 

            SEE:  Family Friend, 1858;  Secret Out;  Illustrated Boy's Own Treasury; Magician's Own Book (UK);  Anon: Social ...; 

 

            The CHINESE LADDER has several crossbars with holes and a long string winding through the holes and through intermediate discs, with a needle attached to the end of the string.  One threads the needle back through all the holes, keeping the end of the loop from pulling through, then threads back through all the ladder holes, avoiding all the disc holes and again keeping the loop from pulling through.  When completed, pulling the end causes all the discs to fall off, but the string is still in its original place on the ladder.  Making another pass through the discs but not the holes has the effect that when you pull the string free, all the discs are on it. 

            SEE:  Adams & Co.;  Hoffmann;  Benson;  Slocum, Compendium, c1890;  Williams;  Collins.

 

Pacioli.  De Viribus.  c1500. 

Ff. 206r - 206v, Part 2, (C)apitulo. C. Do(cumento) cavare una stecca. de un filo per .3. fori (To remove a stick from a cord through 3 holes).  = Peirani 282-283.  On f. 206r is a marginal drawing clearly showing the string through three holes in one stick, but this is not reproduced in the transcription.

Ff. 207v - 208v, Part 2, Capio. CII. Do(cumento) unaltro speculativo cavar doi botoni di una stenga fessa nel mezzo et sce'pia in testa (Another speculation - remove two buttons from a string divided in the middle and halved at the ends).  = Peirani 284-286.  Dario Uri says this is describing a version of the Alliance or Victoria puzzle with four holes in each button.

Ff. 209r - 210r, Part 2, Capo. CIII. Do(cumento) legare con la sopra detta strenga fessa. doi sola. de carpe' ambe doi. a uno modo. bella cosa (To tie two shoe soles together into one with the above mentioned doubled string -- a beautiful thing).  = Peirani 286-288.  Pacioli says this is "quasi simile alla precedente".  Dario Uri illustrates this with the Alliance or Victoria puzzle from Alberti.

Ff. 213r - 215r, Part 2, Capitulo. (C)VIII. Do(cumento). Cavare' uno anello grande fore' de doi legati a una bacchetta per testa (Remove a large ring from two tied to a stick by the ends ??)  = Peirani 292-295.  Dario Uri says this is a version of this idea and illustrates it with an unidentified picture.

Cardan.  De Rerum Varietate.  1557, ??NYS.  = Opera Omnia, vol. III, pp. 245-246.  Liber XIII.  Ludicrum primum & Ludicrum secondum.  The first is the Alliance or Victoria puzzle, with one diagram showing it apart.  Second is the version with three holes in each piece, with one diagram, again showing it apart. 

Prévost.  Clever and Pleasant Inventions.  (1584), 1998. 

Pp. 133-134.  Alliance Puzzle with two holes in each board.

Pp. 134-136.  Alliance Puzzle with three holes in each board.

Schwenter.  1636.  Part 10, exercise 29, pp. 410-411.  Two hole version.

John Wecker.  Op. cit. in 7.L.3.  1660.  Book XVIII -- Of the Secrets of Sports: The first Pastime and the second Pastime, p. 338.  Taken from and attributed to Cardan, with the same diagrams.  (I don't know if this material appeared in the 1582 ed.??)

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 44, pp. 35-36 is a two hole version taken from Schwenter.

Ozanam.  1725.  Vol. IV, prob. 31, p. 435 & fig. 37, plate 11 (13).  This is like the Alliance puzzle in Hoffmann, below.

Alberti.  1747.  Art. 31, pp. 205-206 (108‑109) & fig. 38, plate X, opp. p. 206 (between pp. 108 & 109).  Taken from Ozanam.

Catel.  Kunst-Cabinet.  1790.  Das Scheibenspiel, p. 15 & fig. 23 on plate I.  The Alliance puzzle, looking much like one of Ozanam's figures. 

Bestelmeier.  1801.  Item 291: Die verschlungenen Bretchen.  Alliance puzzle.  The diagram is very obscure and the description is too brief to be certain, but I now see it is copied from part of Catel's text, so this is definitely the Alliance.

Boy's Own Book.  The heart and ball puzzle.  1828: 424;  1828‑2: 428;  1829 (US): 220;  1843 (Paris): 438 & 442, no. 16;  1855: 574‑575;  1868: 675-676.  = Wehman, New Book of 200 Puzzles, 1908, p. 18.  In the 1843 (Paris), the engraved heading of the section, p. 435, shows the puzzle.  Boy's Treasury, 1844, omits the puzzle, but copies this engraving on p. 424.  = de Savigny, 1846, pp. 355 & 358, no. 12: Problème du cœur et de la boule.

Nuts to Crack II (1833), no. 91.  The heart and ball puzzle.  Almost identical to Boy's Own Book.

Crambrook.  1843.  P. 4, no. 9: Heart and Ball Puzzle.  Check??

Family Friend 3 (1850) 300 & 331.  Practical puzzle, No. XX.  Drawing shows the artist didn't understand the puzzle at all.  = Magician's Own Book, 1857, prob. 16: The heart puzzle, pp. 271-272 & 295.  = Book of 500 Puzzles, 1859, pp. 85-86 & 109.  = Boy's Own Conjuring Book, 1860, prob. 15, pp. 233 & 258. 

Family Friend (Dec 1858) 359.  Practical puzzles -- 3.  Board and ball, but no name given to it.  I don't have the solution.

The Secret Out.  1859. 

The Board and Rings, pp. 389-390.  Alliance Puzzle with two holes on each board formed by attaching rings at each end of the boards.  Cf Hoffmann.

The Board and Ball, pp. 382-383.  Same picture as Family Friend, but more text.  Diagram for the solution.

Illustrated Boy's Own Treasury.  1860.

Prob. 36, pp. 401‑402 & 441.  Identical to Family Friend, 1850, but omitting some comments in the problem.

Prob. 45, pp. 403 & 443.  Identical to Family Friend, 1858, with slightly more detailed text, but not as much as in The Secret Out.

Adams & Co., Boston.  Advertisement in The Holiday Journal of Parlor Plays and Pastimes, Fall 1868.  Details?? -- photocopy sent by Slocum.  P. 6: Chinese Ladder Puzzle.  Made of ivory and silk.  Ornamentally carved.  Probably the same as Hoffmann 17, below -- ??

Magician's Own Book (UK version).  1871.

The puzzle of Cupid: Variation. -- With beads or balls, pp. 227-228.  Board and ball puzzle.

The heart puzzle, pp. 228-229.  Diagram as in Family Friend, but text is different.

The board and rings, pp. 230-231.  A version of the Alliance puzzle.  It uses boards with screw-eyes in each end rather than holes and uses a long single loop of string.  Cf Secret Out.

Anonymous.  Social Entertainer and Tricks (thus on spine, but running title inside is New Book of Tricks).  Apparently a compilation with advertisements for Johnson Smith (Detroit, Michigan) products, c1890?  P. 68: The board and ball.  Same as Family Friend, 1858.

Hoffmann.  1893.  Chap. II.

No. 11: The heart puzzle, pp. 26 & 52‑53 = Hoffmann-Hordern, pp. 26-27, with photo.  Drawing based on Jaques' puzzle.  Photo on p. 27 shows The Heart Puzzle, with instructions, by M.M.M.M., London, in the "St. Dutius" Series of Novelties, 1880-1900.  Hordern Collection, p. 20, looks like the same example, without the instructions.

No. 12: The Alliance (otherwise known as the Victoria) puzzle, pp. 26‑27 & 53 = Hoffmann-Hordern, pp. 27-28, with photo.  Photo on p. 27 shows The Alliance Puzzle, with instructions, by M.M.M.M., London, in the "St. Dutius" Series of Novelties, 1881-1884.  Hordern Collection, p. 21, looks like the same example, without the instructions.

No. 14: The Ariel puzzle, pp. 28 & 54 = Hoffmann-Hordern, pp. 28-29.  Photo on p. 28 shows an example, 1881-1884.  Hordern Collection, p. 23, is a different example.

No. 15: The pen and wheel, pp. 28‑29 & 54 = Hoffmann-Hordern. pp. 29-30.

No. 17: The Chinese ladder, pp. 30 & 55‑56 = Hoffmann-Hordern, pp. 31-32, with photo.  "It is said to be a genuine importation from China."  Photo on p. 31 shows an example, 1881-1884.  Drawing based on Jaques' puzzle.  Hordern Collection, p. 25, looks like the same example, with the same dating.

At the end of the book (p. 396) is an advertisement for J. Bland's Magical Palace, showing a heart puzzle, an Imperial scale (11.F), Chinese rings (7.M.1) and handcuff puzzle (11.K).  This is not included in Hoffmann-Hordern.

A. Murray.  Some simple puzzles.  The Boy's Own Paper 17 or 18?? (1894??) 46.  Well drawn.

Benson.  1904.

The heart puzzle, p. 210.  (= Hoffmann, p. 26.)

The pen and wheel puzzle, pp. 210‑211.  (= Hoffmann, pp. 28-29.)

The Alliance puzzle, p. 215.  (= Hoffmann, p. 26.)

The Chinese ladder puzzle, p. 216.  (= Hoffmann, p. 30.)

Slocum.  Compendium.  Shows Chinese ladder from Joseph Bland catalogue, c1890, and heart and string puzzle from Johnson Smith 1929 catalogue.

Williams.  Home Entertainments.  1914.  Another string and buttons puzzle, pp. 112-114.  Chinese ladder.  Get all the buttons out of the ladder and onto the thread.

"Toymaker".  The New-century Cross Puzzle.  Work, No. 1394 (4 Dec 1915) 158 (or 153??).  Like half of a Victoria or Alliance Puzzle with two holes in the arms of a cross.

"Toymaker".  The "Wheel of Fate" Puzzle.  Work, No. 1467, (28 Apr 1917) no page number on the photocopy from Slocum -- ??  (= Hoffmann, pp. 28-29.)

Collins.  Book of Puzzles.  1927. 

Pp. 24-26: The Egyptian heart.  Heart and ball form.

Pp. 27-29: The Chinese ladder puzzle.  Description is a bit cryptic, but he winds up with the string only partly on the ladder and all the discs together on the string.

Charles Béart.  Jeux et jouets de l'ouest africain.  Tome I.  Mémoires de l'Institut Français d'Afrique Noire, No. 42.  IFAN, Dakar, Senegal, 1955.  Pp. 418-419.  Describes and illustrates a one-board, two-hole, version called djibilibi or jibilibi from Fonta-Djallon.

 

          11.J.  MÖBIUS STRIP

 

Lorraine L. Larison.  The Möbius band in Roman mosaics.  Amer. Scientist 61 (1973) 544‑547.  Describes and illustrates a Roman mosaic in the Museum of Pagan Art, Arles, France, which has a band with five twists.  No date given.

At the Möbius Conference at Oxford, in 1990. it was stated that the strip appears in Listing's notes for 1858, apparently a few months before it appears in Möbius's notes.

Walter Purkert.  Die Mathematik an der Universität Leipzig von ihrer Gründung bis zum zweiten Drittel des 19. Jahrhunderts.  In:  H. Beckert & H. Schumann, eds.; 100 Jahre Mathematisches Seminar der Karl‑Marx‑Universität Leipzig; VEB Deutscher Verlag der Wissenschaften, Berlin, 1981; pp. 9‑39.  On p. 31, he says that Möbius's Nachlass shows that he discovered the strip in 1858.

Chris Pritchard: Aspects of the life and work of Peter Guthrie Tait, pp. 77-88  IN: James Clerk Maxwell Commemorative Booklet; James Clerk Maxwell Foundation, Edinburgh, 1999.  On p. 81, he says that Listing found the Strip in Jul 1858 and brought it to public attention in 1861, while Möbius found it around Sep 1858 but did not publish until 1865.

J. B. Listing.  Der Census räumlicher Complexe.  Abh. der Ges. der Wiss. zu Göttingen 10 (1861) 97-180.  This appeared as a separate book in 1862.  ??NYS -- cited by M. Kline, p. 1164.

A. F. Möbius.  Über die Bestimmung des Inhaltes eines Polyeders.  Königlich Sächsischen Ges. der Wiss. zu Leipzig 17 (1865) 31-68.  = Gesammelte Werke, Leipzig, 1885-1887, vol. 2, pp. 473-512.  ??NYS -- cited by M. Kline, p. 1165.  He also considers multitwisted bands.

Tissandier.  Récréations Scientifiques.  5th ed., 1888, Les anneaux de papier, pp. 272-273.  Illustration by Poyet.  Shows and describes rings with  0, 1, 2  twists.  Not in 2nd ed., 1881.  I didn't see whether this was in the 1883 ed.

                    = Popular Scientific Recreations; [c1890]; Supplement: The paper rings, pp. 867‑869.

                    Gardner, MM&M, 1956, p. 70 says the earliest magic version he has found is the "1882 enlarged edition of" Tissandier.  This may be Popular Scientific Recreations, but I don't see any date in it and the Supplement contains material that was not in the 1883 French ed. -- cf. comments in Common References.

P. G. Tait.  Listing's Topologie.  Philosophical Mag. (5) 17 (No. 103) (Jan 1884) 30‑46 & plate opp. p. 80.  This is based on Listing's Vorstudien zur Topologie (1847) and Der Census räumlicher Complexe (1861).  Section 8, pp. 37-38, discusses strips with twists, noting that an odd number of half-twists gives one side and one edge.  If the odd number  m  of half-twists is greater than  1,  then cutting it down the middle gives a knotted band with 2m half-twists.  He says this was the basis of a pamphlet which was popular in Vienna a few years ago, which showed how to tie a knot in a closed loop.  Pritchard [op. cit. above, p. 83] says Tait originally did not credit either Listing or Möbius for these results, and in an article in Nature in 1883, he noted that these properties were common currency among conjurers for some time as alluded to in Listing's Vorstudien.

Anon. [presumably prepared by the editor, Richard A. Proctor].  Trick with paper bands.  Knowledge 11 (Jan 1888) 67-68.  Short description, based on La Nature, i.e. Tissandier, with copy of the illustration, omitting Poyet's name.

J. B. Bartlett.  A glimpse of the "Fourth Dimension".  The Boy's Own Paper 12 (No. 588) (19 Apr 1890) 462.  Simple description.

Lewis Carroll.  Letter of  Jun 1890 to Princess Alice.  Not in Cohen.  In 1890, Carroll met Princess Alice (whose father, Prince Leopold had been a student at Christ Church and had been enamoured of the original Alice, then aged 18, but the Queen prevented such a marriage), then age 6 and became friends.  This letter has the plan of a Möbius strip.  This letter was advertised for sale by Quaritch's at the 2001 Antiquarian Book Fair in London.  Carroll refers to it in his letter of 12 Aug 1890 to R. H. Collins and Cohen's note quotes Sylvie and Bruno Concluded, qv below, and explains the object.

Tom Tit, vol. 3.  1893.  See entry in 6.Q for a singly-twisted band.

Lewis Carroll.  Sylvie and Bruno Concluded.  Macmillan, 1893.  Chap. 7, pp. 96‑112, esp., pp. 99‑105.  Discusses Möbius band ("puzzle of the Paper Ring"), Klein bottle and projective plane.  Quoted, with extended discussion in John Fisher; The Magic of Lewis Carroll; op. cit. in 1; pp. 230-234.  Cf Carroll-Gardner, pp. 6-7

Lucas.  L'Arithmétique Amusante.  1895.  Note IV:  Section II: No. 3: Les hélices paradromes, pp. 222-223.  Attributes the ideas to Listing's Vorstudien zur Topologie, 1848.  Says this is the basis of a lengthy memoir sold at Vienna some years ago showing how one can make a knot in a closed loop -- cf Tait.  Gives basic results, including cutting in half.

Herr Meyer.  Puzzles.  The Boy's Own Paper 19 (No. 937) (26 Dec 1896) 206.  No. 3: Paper ring puzzle.  Asks what happens when you cut into halves or thirds after one or two or more twists.

Somerville Gibney.  So simple!  The hexagon, the enlarged ring, and the handcuffs.  The Boy's Own Paper 20 (No. 1012) (4 Jun 1898) 573-574.  Cuts Möbius band in half twice, then does the same with doubly twisted band. 

Hoffmann.  Later Magic.  (Routledge, London, 1903);  Dover, NY, 1979.  The Afghan bands, pp. 471‑473.  Cuts various strips in half.

                    Gardner, MM&M, 1956, p. 71, says this is the first usage of the name Afghan Bands that he has found.

C. H. Bullivant.  The Drawing Room Entertainer.  C. Arthur Pearson, London, 1903.  Paper rings, p. 48.  Cuts various rings in half.

Dudeney.  Cutting-out paper puzzles.  Cassell's Magazine ?? (Dec 1909) 187-191  &  233-235.  Calls it "paradromic ring" and says it is due to Listing, 1847.  Probably based on Lucas.

Devant.  Tricks for Everyone.  Op. cit. in 4.A.1.  1910.  Curious paper patterns, pp. 20-21.  Cutting rings with  0, 1, 2  half-twists in half.

Anon. [H. W. R.]  Games and Amusements.  Ward, Lock & Co., London, nd [c1910??].  The mysterious paper bands, p. 128.  Describes bands with  1, 2, 3  twists and cutting them in half.

Williams.  Home Entertainments.  1914.  The magic paper rings, pp. 104-106.  Rings with  0, 1, 2  half-twists to be cut down the middle.  Good diagram.

Lee de Forest.  US Patent 1,442,682 -- Endless Sound Record and Mechanism Therefor.  Filed: 5 Oct 1920;  patented: 16 Jan 1923.  3pp + 2pp diagrams.  No references and no mention of any previous forms.

Hummerston.  Fun, Mirth & Mystery.  1924.  The magic rings, p. 91.  0, 1, 2  half-twists, each cut in half.

William Hazlett Upson.  A. Botts and the Moebius strip.  Saturday Evening Post (?? 1945).  Reprinted in:  Clifton Fadiman, ed.; Fantasia Mathematica; Simon & Schuster, NY, 1958; pp. 155-170.

Owen H. Harris.  US Patent 2,479,929 -- Abrasive Belt.  Applied: 19 Mar 1949;  patented: 23 Aug 1949.  2pp + 1p diagrams.  Same comments as on de Forest.

William Hazlett Upson.  Paul Bunyan versus the conveyor belt.  ??, 1949.  Reprinted in:  Clifton Fadiman, ed.; The Mathematical Magpie; Simon & Schuster, NY, 1962; pp. 33‑35.

Gardner.  MM&M.  1956.  The Afghan Bands, pp. 70-73 & figs. 9-14, pp. 74-77.  Describes several usages as a magic effect.  Cites Tissandier and Hoffman, cf. above.

James O. Trinkle.  US Patent 2,784,834 -- Conveyor for Hot Material.  Applied: 22 Jul 1952;  patented: 12 Mar 1957.  2pp + 1p diagrams.  Same comments as on de Forest, except that 5 references are mentioned in the file, but not in the patent itself.

James W. Jacobs.  US Patent 3,302,795 -- Self-cleaning Filter.  Filed: 30 Aug 1963;  patented: 7 Feb 1967.  2pp + 2pp diagrams.  "This invention relates to dry cleaning apparatus and more particularly to a self-cleaning filter element comprised of an endless belt having a half twist therein."  The diagram does not show the twist very clearly.  Same comments as on de Forest, except that the Examiner cites three patents.

Making resistors with math.  Time (25 Sep 1964) 49.  Richard L. Davis of Sandia Laboratories has made a Möbius strip noninductive resistor.  This has metal foil on both sides of a nonconductive Möbius strip with connections opposite to one another.  The current flows equally both ways and passes through itself.  He found the inductance as low as he had hoped, but he is not entirely clear why it works!  Gardner, below, describes this and also cites Electronics Illustrated (Nov 1969) 76f, ??NYS.

Jean J. Pedersen.  Dressing up mathematics.  MTr 60 (Feb 1968) 118-122.  Describes garments with two sides and one edge or one side and one edge or two sides and no edges!!

Gardner.  The Möbius strip.  SA (Dec 1968) = Magic Show, chap. 9.  Describes the above mentioned patents and inventions and numerous stories and works of art using the idea.

Ross H. Casey.  US Patent 4,161,270 -- Continuous Loop Stuffer Cartridge having Improved Moebius Loop Tensioning System.  Filed: 15 Jul 1977;  patented: 17 Jul 1979.  2pp + 3pp diagrams.  This is actually only for an improvement in the idea.  "Typically, a cubically-shaped wire form or a plurality of guides are used to effect a Moebius twist in the continuous loop.  The invention includes an improved Moebius loop device and tensioner in the form of an easily constructed planar triangular-shaped device to effect a Moebius twist and tensioning in a continuous loop."  Basically the loop folds around three edges of a triangle.  Cites several earlier patents, but Joe Malkevich says none of them relate to the Möbius idea.

In the early 1990s, Tim Rowett found a German making a stainless steel strip with a double twist in it which could be manoeuvred into a double Möbius strip which appeared to be cut in half through the thickness of the strip and which sprang apart when released.  In fact it can also be seen as the result of an ordinary cutting of a Möbius strip in half.  I cannot recall seeing this behaviour described anywhere, though I imagine it is well known.  The process is shown clearly in the following.

Jean-Pierre Petit.  Gémellité Cosmique.  Text for the month of Juin.  Mathematical Calendar:  Tous les mois sont maths! for 1990 produced by Editions du Choix, Bréançon, 1989.

Scot Morris.  The Next Book of Omni Games.  Op. cit. in 7.E.  1991.  Pp. 53-54 describes Jacobs' 1963 patent.  He says that David M. Walba and colleagues at Univ. of Colorado have synthesised "the first molecular Möbius strip", a molecule called trisTHYME and that they have managed to cut it down the middle!!

 

          11.K. WIRE PUZZLES

 

            Wire and string puzzles are difficult to describe.  Only a few were illustrated before 1900.  S&B, p. 90, says they first appeared in the 1880s (when wire became common), though some are older -- see 7.M.1, 7.M.5, 11.A, 11.C, 11.D, 11.F, 11.I, 11.K.7 and possibly 11.B, 11.E, 11.H.  Ch'ung-En Yü's Ingenious Ring Puzzle Book (op. cit. in 7.M.1) implies that wire and ring puzzles, besides the Chinese Rings, were popular in the Sung Dynasty (960-1279) but I have no confirmation of this assertion.  See the entry under Stewart Culin in 7.M.1 for a vague reference to Japanese ring puzzles called  Chiye No Wa.  This section will cover the various later versions, but without trying to describe them in detail.  I have separated the Horseshoes (11.K.7) and the Caught Heart (11.K.8), as they are so common.

            Wire puzzles were included in general puzzle boxes by 1893 -- see the ad at the end of Hoffmann mentioned in 11.I.  By  1912, they were being sold in boxes of just wire puzzles.  Six boxes of wire puzzles are offered in Bartl's c1920 Zauberkatalog, p. 305.  Wire puzzles are a major component of the Western Puzzle Works, 1926 Catalogue.

            See S&B, pp. 88‑115.

 

Peck & Snyder.  1886.  P. 245, No. A -- The puzzle brain links.  11 interlocked links.  Not in Slocum's Compendium.

Slocum.  Compendium.  Shows Egyptian Mystery from Joseph Bland's catalogue, c1890.

Herr Meyer.  An improved ring puzzle.  In:  Hutchison; op. cit. in 5.A; 1891; chap. 70, section III, pp. 573‑574.  Folding ring on loop on loop on loop on bar.

Hoffmann.  1893.  Chap. VIII: Wire puzzles, pp. 302‑314 = Hoffmann-Hordern, pp. 198-206, with six photos.

No. 1: The united hearts.  (Isomorphic to 11.B.)  Photo on p. 199 shows an undated example.

No. 2: The triangle.  Photo on p. 199 shows an undated example.

No. 3: The snake and ring (= Ring and spring puzzle).  Cf 11.K.1.  Photo on p. 199 shows L'Anneau Prisonnier, with box, 1891-1905.  Hordern Collection, p. 80, shows the same, dated 1880-1905.

No. 4: The hieroglyph.  Photo on p. 201 shows an undated example.  Hordern Collection, p. 81, shows two similar puzzles, La Balance de Thémis and Le Bon Geôlier, dated 1890-1910. 

No. 5: The interlaced triangles.  Photo on p. 201 shows an undated example.

No. 6: The double bow and ring (= Horseshoes puzzle).  Photo on p. 201 shows an undated example.  Hordern Collection, p. 82, shows two versions, La Question du Divorce and Les Anneaux Diaboliques, dated 1880-1905.

No. 7: The Egyptian mystery.  Photo on p. 202 shows a version, with box and instructions, by J. Bland, 1855-1892.  Hordern Collection, p. 83, shows a different version, just of the puzzle.

No. 8: The ball and spiral.  Photo on p. 203 shows an undated example.  Hordern Collection, p. 84, shows what seems to be the same example.

No. 9: The Unionist puzzle (the solution is labelled The two horse-shoes, but Hoffmann-Hordern changes it to agree with the problem).  Photo on p. 205 shows Les Crochets du Diable, with box, 1890-1905.  Hordern Collection, p. 85, shows the same and Comment Vous Separer, with box, both dated 1890-1910.

No. 10: The Eastern question (= Intertwined question marks  = Double witch key).  Photo on p. 205 shows an example, with instructions, by Hamley Bros., 1890‑1900.  Hordern Collection, p. 86, shows two examples of Question Romaine, with their boxes, by W. S. and R. T., both Paris, 1890-1910.

No. 11: The handcuff puzzle (four interlocked broken rings). Photo on p. 205 shows an example by Jaques & Son, 1860-1900.  Hordern Collection, p. 87, shows a version with instructions and a three-ring version, both dated 1880-1890 -- this was included in Jaques puzzle boxes, c1900 (cf discussion at beginning of Section 11).

No. 12: The Stanley puzzle.  Photo on p. 206 shows an example, dated 1890.

See also the J. Bland ad on p. 396.

Montgomery Ward & Co.  Catalog No 57, Spring & Summer, 1895.  Facsimile by Dover, 1969, ??NX.  P. 237, item 25479, is a collection of 8 wire puzzles.  The Chilian Puzzle is shown.  Cf 1903.

Ernest C. Fincham.  Street Toys.  Strand Mag. 10 (1895) 765-773.  Shows:  bicycle puzzle (a circular version of Hoffmann No. 2);  "Ally Sloper" puzzle (a caught heart, cf 11.K.8);  gridiron puzzle (basically an elaborated version of the caught heart, cf 11.K.8);  magnetic puzzle (involves placing three needles by use of a magnet);  key puzzle (a cast puzzle of two interlocked keys);  oriental ring (ordinary four strand puzzle ring, cf 11.K.4).

H. F. Hobden.  Wire puzzles and how to make them.  The Boy's Own Paper 19 (No. 945) (13 Feb 1896) 332-333. 

Gridiron and tongs (similar to Hoffmann no. 2). 

Hourglass and ring (= Horseshoes  = Hoffmann no. 6). 

Cupid's bow (similar to Hoffmann no. 1). 

Saltspoon and eggwhisk (= Hoffmann no. 8). 

Magic rings (= Chinese rings) with  10  rings, requiring  681  moves.  (I think it should be  682.)

Montgomery Ward & Co.  Catalogue.  1903.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 118.  Shows one puzzle (Chilian Puzzle) from a set.  Cf 1895.  Slocum's Compendium also shows this.

Benson.  1904.  Chap. VIII: Wire puzzles, pp. 236‑240.

The interlaced ring puzzle.  (= Hoffmann no. 11.)

The two hearts.  (= Hoffmann no. 1.)  (Isomorphic to 11.B.)

The bow and ring puzzle.  (= Hoffmann no. 6.)

The two crooks.  (= Hoffmann no. 10.)

The "X" puzzle.  (= Hoffmann no. 7.)

The spiral and ring.  (= Hoffmann no. 3.)  Cf 11.K.1.

The triangular maze.  (= Hoffmann no. 2.)

The ball puzzle.  (= Hoffmann no. 8.)

The mysterious loop.  (= Hoffmann no. 4.)

M. Adams.  Indoor Games.  1912.  Pp. 337‑341.

The gridiron and tongs.  (c= Hoffmann no. 2.)

The hourglass and ring.  (= Hoffmann no. 6  = horseshoes puzzle.)

Cupid's bow.  (c= Hoffmann no. 1.)  (Isomorphic to 11.B.)

The saltspoon and eggwhisk. (= Hoffmann no. 8.)

The magic rings.  (= Chinese rings, 7.M.1.)

The heart and link.  (= Cupid's bow, c= Hoffmann no. 1.)  (Isomorphic to 11.B.)

The gridiron and shovel  (c= Cupid's bow, c= Hoffmann no. 1.)  (Isomorphic to 11.B.)

The double gridiron.

The cross and double links.

The tandem and luggage.

The double link.  (= Loony Loop or Satan's Rings.)  See 7.M.5.  This has a loop of string and is isomorphic to the Chinese rings, 7.M.1.

Bartl.  c1920.  Vexier- und Geduldspiele, pp. 305-312, shows 43 wire puzzles.

Western Puzzle Works, 1926 Catalogue.  Shows about 65 wire puzzles.

Collins.  Book of Puzzles.  1927.  Some good metal puzzles, pp. 52-53.  Shows:  The devil's ring;  The teaser;  The three-in-one rings;  The link and rings;  The union puzzle.  Pp. 44-54 cover other wire puzzles:  The Twin Hearts (= Hoffmann no. 1, 11.K.8);  The eternal triangle (= Hoffmann no. 2);  The bead and spiral (= Hoffmann no. 8);  The snake and ring (= Hoffmann no. 3, 11.K.1);  The great seven-ring puzzle (= Chinese rings, 7.M.1);  The eastern question (= Hoffmann no. 10, Double witch key, 11.K.6).

Ch'ung-En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1. 

Pp. 18-19 discusses the 'horseshoes puzzle' which is called 'Jade Interlocked Ring Puzzle' and is "the simplest and easiest puzzle of the Incomplete Ring Type".  Yü then discusses more complex versions of the puzzle.

P. 21 shows the interlocked nails puzzle.

P. 26 shows a single heart, similar to Hoffmann 1, but with a doubled second part, called Recessed Handle Ring Puzzle.  (Isomorphic to 11.B.)

Richard I. Hess.  Compendium of Over  7000  Wire Puzzles.  5th ed., published by the author, Rancho Palos Verdes, Mar 1991, 259pp.  7090  wire, entanglement and cast iron puzzles classified in 17 categories.  There is an 53 page index and then 205 pages of reduced photocopies of the actual puzzles.  The pictures come out almost as good as drawings.  Some of the more obvious combinations do not have pictures.  The naming of the puzzles has a certain poetry about it.  A118:  Type 2 trapeze with baffle, horseshoes, lock and 1-ring key.  A226:  Cascaded double compound trapeze with baffled heart.  B053:  Hong Kong house with semicircles (Hard 2-story).  C226:  Bug (3-ring) with triple cross.  D105:  Triple Finnish diddle.  (Previous editions:  2nd ed, 1982, c500 puzzles;  3rd ed, 1985, c1400 puzzles;  4th ed, 1988, c2600  puzzles.)

 

          11.K.1.        RING AND SPRING PUZZLE

 

          A ring is on a spring with sealed ends.

 

I have seen this illustrated in Wizard's Guide, a catalogue of magic apparatus by W. J. Judd, 1882.  ??NX.

Montgomery Ward & Co.  Catalogue.  1886.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 24.  Slocum's Compendium also shows this.

Peck & Snyder.  1886.  P. 162: Universal pocket puzzle.

Hoffmann.  1893.  Chap. VIII, no. 3: The snake and ring (= ring and spring puzzle) = Hoffmann-Hordern, p. 199.  Photo on p. 199 shows L'Anneau Prisonnier, with box, 1891-1905.  Hordern Collection, p. 80, shows the same, dated 1880-1905.

Benson. 1904.  Chap. VIII: Wire puzzles, pp. 236‑240.  The spiral and ring.  (= Hoffmann's no. 3.)

I have seen a French example, called L'anneaux prisonnier, dated 1900-1920, but see above.

Slocum.  Compendium.  Shows many later examples:  1913, 1915, 1915, 1919, etc.

Bartl.  c1920. 

P. 307, no. 61: Der Ring der Nibelungen.

P. 307, no. 63: Spirale mit Nadel.  This has a rod down the middle of the spring.

P. 310, no. 87: Ring mit Kette.

Western Puzzle Works, 1926 Catalogue.  No. 9: Down and Out.

Collins.  Book of Puzzles.  1927.  The snake and ring puzzle, pp. 48-49.

 

          11.K.2.        STRING AND SPRING PUZZLE

 

            A loop of string goes through the spring which has a few turns and long tails so the string doesn't come off obviously.

 

Western Puzzle Works, 1926 Catalogue.  No. 1877: Loop and Chain.

Slocum.  Compendium.  Shows Magic Chain from Johnson Smith 1929 catalogue.

 

          11.K.3.        MAGIC CHAIN  =  TUMBLE RINGS

 

            One holds the two top rings and releases the upper one which appears to drop to the bottom.

 

A magic chain.  The Boy's Own Paper 12 (No. 581) (1 Mar 1890) 351.  Good picture by Poyet, so this ought to be in Tissandier or Tom Tit, but I haven't seen it.

Der Gute Kamerad.  Kolumbus‑Eier.  1890.  ??NYS, but reproduced in Edi Lanners' 1976 edition, translated as:  Columbus' Egg; Paddington Press, London, 1978; The magic chain, pp. 176‑177, with good illustration on p. 177 -- an enlargement of Poyet's picture with his name removed.

Bartl.  c1920.  P. 308, no. 71: Konsilkette.

Western Puzzle Works, 1926 Catalogue.

No. 12: Drop Rings Illusion.  15¢.  Picture shows  20  rings.

No. 334: Drop Ring illusion, chain of  35  Rings.  15¢.  However the picture shows  26  rings??

Davenport's catalogue, op. cit. in 10.T, c1940, pp. 8 & 29, calls it The Wizard's Chain.

Gardner.  SA (Aug 1962).  = Unexpected, chap. 13.  Calls it "tumble rings".  Describes the chain with a good diagram for making your own, but gives no indication of its history.

 

          11.K.4.        PUZZLE RINGS

 

William Jones.  Finger-Ring Lore.  Chatto & Windus, London, 1877.  Pp. 313-321 discusses gemmel or gemmow rings with two or three parts.  He cites Herrick -- "a ring of jimmals",  quotes Dryden's play Don Sebastian describing a two part ring,  describes a three-part jointed ring,  describes a two part ring excavated in 1800 (medieval??),  describes 'a plain geemel wedding-ring' given by the Prince Regent to Mrs. Fitzherbert,  describes the five-link wedding ring of Lady Catherine Gray,  illustrates a 15C gemmel ring with a head of Lucretia of the type mentioned in Twelfth Night II.v,  illustrates Sir Thomas Gresham's betrothal ring of 1544,  describes and illustrates several other examples -- 16C, 13C.  On pp. 321-322, he mentions an exhibition of some puzzle rings by Rev. John Beck at a meeting of the Archaeological Institute in Mar 1863.  These had  7, 4 and 9  parts.

Ernest C. Fincham.  Street Toys.  Strand Mag. 10 (1895) 765-773.  Shows:  Oriental ring (ordinary four strand puzzle ring).

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  Says he believes it is of Indian origin.

George Frederick Kunz.  Rings for the Finger.  J. B. Lippincott, Philadelphia, 1917;  reprinted by Dover, NY, 1973 (with the two colour plates done in B&W). 

                    Pp. 218‑233 discusses puzzle rings.  The earliest forms were gimmel rings which had two or three parts which could be separated for use at betrothal, with the parts rejoined at the wedding and given to the bride.  These are known from the 16C -- e.g. the plate facing p. 219 shows a 16C German example from the BM and Sir Thomas Gresham's betrothal ring (c1540) is similar.  P. 219 gives a quote from Dryden's play Don Sebastian which describes a two part ring.  The plates opp. pp. 220, 221 & 230 show other examples including a three part one from 17C.

                    Kunz says "the so‑called 'puzzle ring' ... was derived from the East."  The plate facing p. 220 shows examples of a three part ring and the common four part Middle Eastern ring, in gold, from the 17C in the BM.

                    The plate opp. p. 233 shows a six part gold betrothal ring from the Albert Figdor collection, Vienna.  This forms a simple chain of six rings.

Stewart Culin.  Korean Games.  Op. cit. in 4.B.5.  Section XX: Ryou‑Kaik‑Tjyo -- Delay Guest Instrument (Ring Puzzle), pp. 31‑32.  Says there are many Japanese ring puzzles, called Chiye No Wa, and shows one which seems to be  10  rings linked in a chain -- possibly the simple type of puzzle ring??

Slocum.  Compendium.  Shows The Lady's and Gentleman's "Wonderful Ring" from Joseph Bland's catalogue, c1890.  This has four parts which form a simple chain.  The Compendium also shows Puzzle Ring from Johnson Smith 1929 catalogue, which is the classic Turkish or Middle Eastern four part ring.

 

          11.K.5.        RING MAZES

 

            These are plates with holes and perhaps raised sections.  An open ring must be removed by working it from hole to hole.

 

S&B 92 says a version was sold by Hamleys in 1879 and appeared as The Boston In-and-Out Puzzle in 1880-1885 and as The Queen's Jubilee Puzzle in 1887.

Peck & Snyder.  1886.  P. 250: no. 188 -- The order of stupids.

Hoffmann.  1893.  Chap. X, pp. 353-354 & 392 = Hoffmann-Hordern, pp. 251-252, with photos.

No. 42: The Conjurer's Medal.  Photos on p. 251 show an example from 1880-1890 and The Queen's Jubilee Puzzle, 1887.  Hordern Collection, p. 96, shows a version labelled  Boston  In and Out,  dated 1880-1895.

No. 43: The Maze Medal.  Photo on p. 252 shows an example dated 1888.  Hordern Collection, p. 97, shows the same.

Western Puzzle Works, 1926 Catalogue.

No. 52: Spider Web.

No. 123: Boxing the Check.  (I'm not certain from the picture that this is a ring maze??)

Dudeney.  Great puzzle crazes.  Op. cit. in 2.  1904.  "... the 'Conjurer's Medal,' that came out some years ago ...."  Medal with five holes.

 

          11.K.6.        INTERLOCKED NAILS, HOOKS, HORNS, ETC.

 

            I have just added this.  The puzzle has two interlocked objects.  One type has nails bent around a  270o  twist -- see S&B 96-97.  A variation, called Wishbone Puzzle, has longer tails -- see S&B 97.  A variation has one of the nails made longer with twists at each end, sometimes called Tangle Twister -- see S&B 96.  A variation has two circular bits with tails, sometimes called Double Witch Key -- see S&B 97 & 102.  Another type has two S-shaped pieces or two J-shaped pieces (hooks) -- see S&B 95.

 

Hoffmann.  1893.  Chap. VIII, no. 10: The Eastern question, p. 307 = Hoffmann-Hordern, pp. 204-205.  Interlocked circular bits with tails  = Double witch key.  Photo on p. 205 shows an example, with instructions, by Hamley Bros., 1890‑1900.  Hordern Collection, p. 86, shows two examples of Question Romaine, with their boxes, by W. S. and R. T., both Paris, 1890-1910.

Burnett Fallow.  How to make an ingenious link puzzle.  The Boy's Own Paper 16 (No. 777) (2 Dec 1893) 143.  = Hoffmann no. 10.

Benson. 1904.  Chap. VIII: Wire puzzles, pp. 236‑240.  The two crooks.  (= Hoffmann no. 10.)

Walter S. Jenkins.  US Patent 969,481 -- Puzzle.  Filed: 16 Mar 1908;  patented: 6 Sep 1910.  3pp + 1p diagrams.  Interlocked nails.

Bartl.  c1920. 

P. 309, no. 76: Hexen-Schlüssel  = Double witch key.

P. 311, no. 91: Wodanschlüssel  = Hooks. 

P. 311, no. 98.  A type of Tangle twister.

P. 311, no. 99.  Interlocked nails.

P. 311, no. 100.  Double witch key.

Western Puzzle Works, 1926 Catalogue.

No. 10: Link the Link.  Two hooks.

No. 102: Nails.

No. 71: Elk Horns.  Patented.

No. 1978: [unnamed, circular bits with tails, sometimes called Question Marks or Double Witch Key.]

No. 211: Tantalizer -- like nails but one piece is twisted at both ends.

Collins.  Book of Puzzles.  1927.  The Eastern question, pp. 53-54.  As in Hoffmann, with very similar diagram.

S&B 95 calls the interlocked hooks, Loop the Loop or The Devil's Keys, and shows it, but gives no date.  S&B 96-97

Ch'ung-En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1.  P. 21 shows the interlocked nails puzzle.

 

          11.K.7.        HORSESHOES PUZZLE

 

          See S&B 99.

 

Magician's Own Book (UK version).  1871.  The ring and wire-loop puzzle, p. 113.  With elongated horseshoe parts.

Hoffmann.  1893.  Chap. VIII: Wire puzzles, no. 6: The double bow and ring (= Horseshoes puzzle) = Hoffmann-Hordern p. 201.  Photo on p. 201 shows an undated example.  Hordern Collection, p. 82, shows two versions, La Question du Divorce and Les Anneaux Diaboliques, dated 1880-1905.

H. F. Hobden.  Wire puzzles and how to make them.  The Boy's Own Paper 19 (No. 945) (13 Feb 1896) 332-333.  Hourglass and ring. 

Benson. 1904.  Chap. VIII: Wire puzzles, pp. 236‑240.  The bow and ring puzzle.  (= Hoffmann no. 6.)

M. Adams.  Indoor Games.  1912.  Pp. 337‑341.  The hourglass and ring.  (= Hoffmann 6.)

Bartl.  c1920. 

P. 307, no. 62: Hufe an einer Kette.  Here the horseshoes are joined by bits of chain.

P. 308, no. 66: Vexierlyra.  Here the horseshoes are in a 'lyre' shape.

P. 310, no. 88: Hufeisen-Vexierspiel.

Ch'ung-En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1.  Pp. 18-19 discusses the "horseshoes puzzle" which is called "Jade Interlocked Ring Puzzle" and is "the simplest and easiest puzzle of the Incomplete Ring Type".  Yü then discusses more complex versions of the puzzle.

 

          11.K.8.        THE CAUGHT HEART

 

          This puzzle is isomorphic to 11.B.

 

Hoffmann.  1893.  Chap. VIII: Wire puzzles, pp. 302‑314.  No. 1: The united hearts.

Ernest C. Fincham.  Street Toys.  Strand Mag. 10 (1895) 765-773.  Shows:  "Ally Sloper" puzzle (a Caught heart)  and Gridiron puzzle (basically an elaborated version of the Caught heart).

H. F. Hobden.  Wire puzzles and how to make them.  The Boy's Own Paper 19 (No. 945) (13 Feb 1896) 332-333.  Cupid's bow (similar to Hoffmann 1).

Benson. 1904.  Chap. VIII: Wire puzzles, pp. 236‑240.  The two hearts.  (= Hoffmann no. 1.)

M. Adams.  Indoor Games.  1912.  Pp. 337‑341. 

Cupid's bow.  (c= Hoffmann 1.)

The heart and link.  (= Cupid's bow, c= Hoffmann 1.)

The gridiron and shovel  (c= Cupid's bow, c= Hoffmann 1.)

Bartl.  c1920.  P. 308, no. 65: Vexierherz.

Collins.  Book of Puzzles.  1927.  The twin hearts puzzle, pp. 44-45.  As in Hoffmann, with similar diagram.

Ch'ung-En Yü.  Ingenious Ring Puzzle Book.  1958.  Op. cit. in 7.M.1.  P. 26 shows a single heart, similar to Hoffmann 1, but with a doubled second part, called Recessed Handle Ring Puzzle. 

 

          11.L.  JACOB'S LADDER AND OTHER HINGING DEVICES

 

            I am now realising that a number of the objects in 6.D, namely the tetraflexagons and the trick or magic books, are just extended forms of the Jacob's ladder.  See also Engel in 6.X.  The Chinese wallet has two boards with this kind of hinging so it can open on either side, giving different effects. 

 

Pacioli.  De Viribus.  c1500.  Ff. 229r - 229v, Part 2, Capitolo. CXXXII. Do(cumento). del solazo puerile ditto bugie (On the childish recreation called deception?).  = Peirani 316‑317.  Uses two tablets and three leather straps.  Describes how to use it to catch a straw.  His references show that it is called 'calamita di legno' (calamity of wood (or magnet of wood -- depending on whether the Italian is  calamità  or  calamita) in Italian.

Bernardino Luini.  "A Boy with a Toy" or "Cherub with a Game of Patience".  Proby Collection, Elton Hall, Peterborough, Cambridgeshire.  The painting is 15" by 13" (38 cm by 33 cm).  Luini was a fairly well known Lombard follower of Leonardo, born c1470 and last known in 1533.  I have found no indication of the date of the work, but the middle of his working life is c1510.  The figure is reproduced in: Angela Ottino della Chiesa; Bernardino Luini; Electa Editrice, Milan, 1980, item 59, but the quality is not good.  The painting is described in: G. C. Williamson; Bernardino Luini; George Bell and Sons, London, 1900, pp. 104-105.  Williamson says the tapes holding the boards together are red and are apparently holding a straw, but he doesn't seem to recognize the object.

                    F. Bartolozzi made a nice engraving of this in 1795, attributing the painting to da Vinci, as was generally believed at the time.  James Dalgety has an example and I have a photo of it.

Gustav Pauli.  Ausstellung von Gemälden der Lombardischen Schule im Burlington Fine Arts Club  London  April - Juni 1898.  (Schluss).  Zeitschrift für bildene Kunst (NF) 10 (1898-1899) 149.  The Luini was on display and the author describes the toy as a 'Taschenspielerstückchen', a little juggler's trick -- but recall that juggler was long a synonym for magician -- with two boards which allow one to vanish the straw.

Bernardino Licino (attrib.).  "Portrait of a Man with a Puzzle".  Picture Gallery, Hampton Court Palace, London, Richmond upon Thames, London.  Licinio was a Venetian painter born about 1485 and last known in 1549.  The painting is described and illustrated (in B&W) in John Shearman; The Early Italian Pictures in the Collection of Her Majesty The Queen; CUP, 1983, item 141, plate 124.  It is very similar to another portrait known to be by Licinio and dated 1524, so this is probably c1524 and hence a bit later than the Luini.  Shearman cites the Luini painting and Pauli's notice of it.  The description says the binding tapes are red, as in the Luini, and both show something like a straw being trapped in the wallet, which suggests some connection between the two pictures, though it may just be that this toy was then being produced in or imported to North Italy and was customarily made with red tapes.  On the toy is an inscription: Carpendo Carperis Ipse (roughly: Snapping snaps the snapper), but Shearman says it definitely appears to be an addition, though its paint is not noticeably newer than the rest of the painting.  Shearman says the toy comprises 'three or more rectangles ...', though both paintings clearly show just two pieces.  My thanks to Peter Hajek who reported seeing this in an email of 22 May 1998

Prévost.  Clever and Pleasant Inventions.  (1584), 1998.  Pp. 136-140.  Chinese wallet.

Schwenter.  1636,  Part 15, exercise 27, pp. 551-552.  "Ein einmaul zu machen."  Chinese wallet.  I can't find Einmaul in my dictionaries.

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 66, pp. 49-50.  Chinese wallet.

Peter Haining.  Moveable Books  An Illustrated History.  New English Library, 1979.  Pp. 12‑13 shows The Unique Click Tablets of c1790 which is a Jacob's Ladder of six pieces with a two part picture glued to each face, described as 'possibly unique'.  The pictures are uncoloured and similar to those occurring in chapbooks and/or hornbooks of the period.  (Thanks to Christopher Holtom for this information.)

de Savigny.  Livre des Écoliers.  1846.  P. 272: Les deux planchettes.  Shows 3½ parts and indicates one can continue beyond four.  "Tous les écoliers connaissent le jeu des planchettes ...."

Leslie Daiken.  Children's Toys Throughout the Ages.  Spring Books, London, 1963.  Plate 6 on p. 24 shows "Hand‑operated game of changing pictures, c. 1850" which clearly shows the "Jacob's ladder" hinging with four parts.

An example with nine blocks with coloured pictures, possibly home-made, mid 19C, was advertised by a book dealer in 2003.

Edward Hordern's collection has an example with four panels from c1854.

Hanky Panky.  1872.  The magic pocket-book (Die zwei magichen [sic] Brieftaschen), pp. 270-272.  This seems to be a form of this mechanism similar to the Chinese wallet.

H. F. Hobden.  Jacob's ladder, and how to make it.  The Boy's Own Paper 12 (No. 592) (17 May 1890) 526.  Says he doesn't know why it is called Jacob's ladder, but that it has been popular "for a number of years".  Suggests at least  7  blocks, preferably  12.

Hutchison.  Op. cit. in 5.A.  1891.  Chap. 71: Jacob's ladder, and other contrivances.  Section I -- Jacob's ladder, pp. 583‑585.  Shows it clearly.  Suggests use of  12  parts, or at least  7.

Axel O. Sodergren.  US Patent 778,282 -- Folding Picture-Album.  Applied: 28 Mar 1904;  patented: 27 Dec 1904.  2pp + 2pp diagrams.  Ordinary Jacob's Ladder used to display pictures.

Davenport's catalogue, op. cit. in 10.T, c1940, pp. 8 & 21, shows the two part version called The Wonderful Magic Book, now often sold as the Chinese Wallet.

Norman F. Rutherford.  US Patent 2,245,875 -- Toy.  Applied: 4 Dec 1939;  patented: 17 June 1941.  4pp + 7pp diagrams.  "... a development of and improvement upon the toy known for years under the name of Jacob's ladder."  Starts with two block versions, then shows 8 blocks formed into a ring and a number of shapes it can form.  Then considers three blocks in an L with hinges in two directions, using with square or hexagonal blocks.  Then considers five blocks in a W, or closed into a four block ring and the seven block extension of this idea, closed into a six block ring, and the nine block version closed into an eight block ring.  then considers the six hexagon version in a ring.  All in all, a pretty direct ancestor of Rubik's Magic puzzle.

 

          11.M. PUZZLE BOXES

 

            A six piece burr with identical flat notched pieces and no key piece is sometimes assembled by forcing together (perhaps after steaming) to make an unopenable money box.  Cf 6.W.5.  Vesta boxes were small boxes to hold matches and were popular from c1850 to c1920.  Some of these had trick openings and form a distinct class of puzzle boxes.

 

Catel.  Kunst-Cabinet.  1790.  Das Vexierkästchen, p. 21 & fig. 21 on plate I.  Figure just shows a box.  Text says the cover is "made like a see-saw" and one presses firmly on one side and the other lifts up.

Bestelmeier.  1801.  Item 208: Eine Kästchen, welches man ohne das Geheimniss zu wissen nicht öfnen kan. 

Jerry Slocum, compiler.  Match box puzzles and related trick boxes from catalogs, books and patents.  Jan 1993.  12pp, with 25 entries during 1843-1912, some entries being lists of many items.  Includes Crambrook, Taylor, Peck & Snyder (only items 35 & 122), Hoffmann (items 30-39), below.

Crambrook.  1843.  P. 3 lists 24 types of Puzzling Boxes.  ??

1-4:  Brass Boxes with Dial-locks or Clock-faces.

5-6:  Common Oblong Snuff Box.

7-8:  Flat box-root [sic] Snuff Box.

9:  Mahogany double-lidded box.

10:  Waterloo Boxes, round reeded.

11:  Large Shaving ditto.

12:  Tin Pricker Box.

13:  Secret Cigar Case, in shape of a book.

14:  Larger ditto, for a Case for a Pack of Cards.

15-17:  Magic Cigar Case.

18:  Japan tinplate Oval Box.

19:  Brass Oblong Box.

20-22:  Indian Ball Puzzles.

23:  French Cube Box.

24:  The Puzzling Dice Box.

The 'Ne Plus Ultra' match or snuff box has a lid which opens by pressing on one edge.  Slocum, above, says the earliest known example is in silver, produced by Alfred Taylor, silversmiths of Birmingham, and hall-marked 1864.

Roger Fresco-Corbu.  Vesta Boxes.  Antique Pocket Guides, Lutterworth Press, Guildford, Surrey, 1983.  He mentions several trick boxes. 

P. 7, fig. 3.  Silver box, 1869, opens by sliding the stud on the front.

P. 9, fig. 4 & p. 13.  S. Basnett.  UK Patent 1212 (26 Jan 1887).  Catch on the front allows inner case to pivot out sideways.  Figure shows a silver example made by Basnett.

P. 34, fig. 48 & p. 35.  Fur covered boxes usually opened by pressing on a stud on the front, concealed by the fur.  Figure shows an example from the 1880s.

P. 35, fig. 50 & p. 36.  An ivory example, undated, but 2nd half of 19C, from the continent, that opens by pressing the stud on the front.

Peck & Snyder.  1886.

P. 129, no. 8: The deceitful tobacco box.

P. 230, no. 20: The Indian puzzle ball.

P. 230, no. 21: The enchanted tea chest.

P. 231, no. 35: The magic fusee box.  (Gravity lock.)

P. 241, no. 122: The Not for Joseph snuff or fusee box.

P. 251, no. 203: The wonderful secret ball.

Montgomery Ward & Co.  Catalogue.  1889.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 27.  Puzzle Locomotive Savings Bank.  Does not open until it is full of money.

Marshall Field & Co.  Catalogue.  1892.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 66. 

The Dairy.  A money bank with a puzzle padlock.

Presto Trick Bank.  Has a trick drawer in which you put a coin which vanishes when you close the drawer.  Doesn't really seem to be a puzzle.

Hoffmann.  1893.  Chap. 2, pp. 37-68 = Hoffmann-Hordern, pp. 44-54, with photos, shows several items which could be included here. 

No. 30: The Psycho Match-box has a gravity lock.  Photo on p. 45 shows a wood version, with instructions, by Gamage, 1860-1890, and a brass version, 1860‑1890.  Hordern Collection, p. 32, shows a wood version with 'edelweiss' pattern, 1860-1900, in two states.

No. 31-37 are other match-boxes.  Photos on pp. 46 & 49.  Hordern Collection, pp. 33‑34.

No. 38-39 are snuff boxes.  Photos on p. 51.  Hordern Collection, p. 35.

No. 40-41 are puzzle balls.  Photos on pp. 53-54.  Hordern Collection, p. 36.

Carl P. Stirn.  1894  Trade Price List  Toys, ....  Reproduced in: Ronald S. Barlow, ed.; The Great American  1879-1945  Antique Toy Bazaar; Windmill Publishing, El Cajon, California, 1998.  The Magic Bank for dimes, on p. 99 of Barlow.  I believe this is a cylindrical device which will not open until it is filled with dimes -- I have several examples.

James Scott.  Chinese puzzles, tricks, and traps.  Strand Mag. 20 (No. 120) (Dec 1900) 715‑720.  Figure 7 shows an object that looks like a three-piece burr but which is a puzzle box made of cardboard or thin wood.  It opens by sliding one stick and then pressing its end, when its sides are seen to be hinged and they flex outward.  Figures 8‑10 show an ivory globular trinket casket, which has three orthogonal rods crossing in the centre.  These have to be turned and slid in the right order to open the box.

Montgomery Ward & Co.  Catalogue.  1903.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 120. 

Little Gem Pocket Savings Bank.  Opens when $5 of dimes have been inserted.

Gem Pocket Savings Bank.  Opens when 50 pennies have been inserted.

Butler Brothers (now City Products Corp.).  Catalogue.  1914.  Reproduced in: Joseph J. Schroeder, Jr.; The Wonderful World of Toys, Games & Dolls  1860··1930; DBI Books, Northfield, Illinois, 1977?, p. 179.  "Gem" Pocket Money Banks.  Types which open when 20 nickels or 50 dimes are inserted.

M. Adams.  Indoor Games.  1912.  A puzzle matchbox, pp. 260-261.  Simply concealed slide.

Cecil Henry Bullivant.  Every Boy's Book of Hobbies.  T. C. & E. C. Jack, London, nd [c1912]. 

Pp. 46-51: How to make a school box with secret compartments.  Simple chest with false bottom and hidden release. 

Pp. 52-55: How to make a puzzle box.  Small box with a gravity lock.

Bartl.  c1920.  P. 306, no. 25 is very similar to Hoffmann's Psycho match-box.

Butler Brothers.  [Toy catalogue], 1921.  Reproduced in: Ronald S. Barlow, ed.; The Great American  1879-1945  Antique Toy Bazaar; Windmill Publishing, El Cajon, California, 1998.  "Gem" Pocket Savings Banks, on p. 186 of Barlow.  One of these may be the same as the item mentioned under Stirn, above.

 

          11.N. THREE KNIVES MAKE A SUPPORT

 

            New section.  The pattern of interlocking knives involved is also used in basketry, woven fences, latticework, etc.  Indeed, if they are beams with notches, this gives frameworks which can roof a space much wider than any available beam, e.g. Wren's ceiling of the Sheldonian Theatre, Oxford, and this is used in puzzle pot stands.

            Four knives versions:  Magician's Own Book; Hoffmann; Blyth; Collins; Bile Beans; Doubleday;

 

Pacioli.  De Viribus.  c1500.  Ff. 228r - 228v, Part 2, Capitolo. CXXIX. Do(cumento). atozzare .iij tagli de coltelli insiemi (Join together three blades of knives).  = Peirani 315.  Pacioli says this was shown to him on 1 Apr 1509 (Peirani has misread this as  isog) by "due dorotea veneti et u perulo 1509 ad primo aprile ebreo".  Peirani transcribes  u  as  un  but Dario Uri thinks it is the initial of Perulo's given name.  I wonder if 'dorotea' might refer to some occupation, e.g. nuns at S. Dorothy's Convent.  In Vienna, the Dorotheum is a huge public auction house where estates are auctioned off.  The word 'ebreo' means 'Hebrew', but I cannot see what it refers to.

Cardan.  De Subtilitate, Book 17, 1550.  = Op. Omnia III, p. 629.

Prévost.  Clever and Pleasant Inventions.  (1584), 1998.  PP. 19-20.

van Etten.  1624.  Prob. 6 (6), p. 7 (15‑16).  Henrion's 1630 Notte cites Cardan.

Hunt.  1631 (1651).  Pp. 278-279 (270-271).

Schwenter.  1636.  Part 15, exercise 7, pp. 536-537.  Three knives.

John Wecker.  Op. cit. in 7.L.3.  1660.  Book XVIII -- Of the Secrets of Sports: Sticks that mutually support one the other, p. 345.  Taken from and attributed to Cardan, with very similar diagram.  (I don't know if this material appeared in the 1582 ed.??)

Witgeest.  Het Natuurlyk Tover-Boek.  1686.  Prob. 74, pp. 56-57.  Three pipes shown.  Text refers to sticks, etc.

Ozanam.  1694.  Prob. 16: 1696: 287-288 & fig. 140, plate 47.  Prob. 16 & fig. 35, plate 15, 1708: 364.  Prob. 20 & fig. 140, plate 47 (45), 1725: vol. 2, 392-393.  Prob. 47 & fig. 50, plate 11, 1778: vol. 2, 87;  1803: vol. 2, 93-94;  1814: vol. 2, 76;  Prob. 46, 1840: 235.

Badcock.  Philosophical Recreations, or, Winter Amusements.  [1820].  P. 15, no. 25 & Frontispiece fig. 2: To place three sticks, or tobacco pipes, upon a table in such a manner, that they may appear to be unsupported by any thing but themselves.

Rational Recreations.  1824.  Feat 16, p. 64.  Three knives or tobacco pipes.

Endless Amusement II.  1826? 

Pp. 58-59: To arrange three sticks that shall support each other in the air.

Prob. VI, p. 193.  "The annexed figure explains a most ingenious device for forming flat roofs or floors, of pieces of timber, little more than half the length of such roof or floor.  This plan is well known to architects; and is particularly mentioned in Plot's Natural History of Oxfordshire, from which we have taken this figure.  ...."  = New Sphinx, c1840, pp. 132-133, but omitting  '; and is particularly mentioned in Plot's Natural History of Oxfordshire, from which we have taken this figure'.

Boy's Own Book.  1828.

The bridge of knives.  1828: 338;  1828‑2: 346;  1829 (US): 153;  1843 (Paris): 393, Bridge of knives;  1855: 485;  1868: 620.  = de Savigny, 1846, pp. 267-268: Pont de couteaux.

The toper's tripod.  1828: 338;  1828‑2: 352;  1829 (US): 159;  1843 (Paris): 391, The tobacco pipe jug stand;  1855: 485;  1868: 621.  = de Savigny, 1846, p. 260: Le plateau avex les pipes.

                   The 1843 (Paris) are c= the versions in Every Little Boy's Book, c1856.

Nuts to Crack II (1833).

No. 94.  The bridge of knives.

No. 95.  The toper's tripod.

Julia de Fontenelle.  Nouveau Manuel Complet de Physique Amusante ou Nouvelles Récréations Physiques ....  Nouvelle Édition, Revue, ..., Par M. F. Malepeyre.  Librairie Encyclopédique de Roret, Paris, 1850.  P. 408 & fig. 147 on plate 4: Disposer trois batons ....  Figure copied from Ozanam, 1725.

Magician's Own Book.  1857. 

P. 186: The toper's tripod.  Use three pipes to support a pot of ale.  = Boy's Own Conjuring Book, 1860, p. 162, with different illustration.

P. 187: The bridge of knives.  = Boy's Own Conjuring Book, 1860, p. 163 with redrawn illustration.

Anonymous.  Every Little Boy's Book  A Complete Cyclopædia of in and outdoor games with and without toys, domestic pets, conjuring, shows, riddles, etc.  With two hundred and fifty illustrations.  Routledge, London, nd.  HPL gives c1850, but the text is clearly derived from Every Boy's Book, whose first edition was 1856.  The material here is in the 1856 ed of Every Boy's Book (with J. G. Wood as unnamed editor), not yet entered, and later editions, but with different text and pictures.  These are essentially the same as the versions in Boy's Own Book, 1843 (Paris).

                    P. 339: The tobacco pipe jug stand.

                    P. 351: Bridge of knives.

Magician's Own Book (UK version).  1871.

To make a seat of three canes, p. 123.  Says this can also be done with three knives or "rounders" bats or long pipes (called the Toper's Tripod).

The puzzle bridge, p. 123.  Stream 15 or 16 feet across, but none of the available planks is more than 6 feet long.  He claims that one can use a four plank version of our problem to make a bridge.  See the discussion in 6.BD.

Tissandier.  Récréations Scientifiques.  1883?  Not in the 2nd ed., 1881.  I didn't see if these were in the 3rd ed., 1883.  5th ed., 1888.  Illustrations by Poyet.

Poser un verre sur trois bâtons ayant chacun une extrémité en l'air, pp. 42-43.  Cites and quotes Ozanam.

La carafe et les trois couteaux, p. 43.  The knives are resting on three glasses.

Tissandier.  Popular Scientific Recreations;  Supplement.  1890?  Pp. 798-799.  To poise a tumbler upon three sticks, each one of which has one end in the air.  The water-bottle and the three knives.  = (Beeton's) Boy's Own Magazine 3:6 (Jun 1889) 249-251.

Der Gute Kamerad.  Kolumbus‑Eier.  1890.  ??NYS, but reproduced in Edi Lanners' 1976 edition, translated as:  Columbus' Egg; Paddington Press, London, 1978.  Uses Poyet's illustrations.

The balancing goblet, p. 22.  c= Tissandier.

The floating carafe, pp. 22-24.  c= Tissandier.

Hoffmann.  1893.  Chap. X, pp. 350-351 & 390-391  = Hoffmann/Hordern pp. 248-249, with photo.

No. 38: The cook in a difficulty (La question de la marmite).  Four iron staffs to support a stewpan.  Photo on p. 249 shows Question de la Marmite, with box and instructions, by Watilliaux, 1874-1895.

No. 39: The Devil's bridge (Le pont du diable).  Three planks to make a bridge joining three points.  Photo on p. 249 shows Le Pont du Diable, with box instructions and solution, by Watilliaux, 1874-1895.  Hordern Collection, p. 95, shows La Passerelle de Mahomet, with box and solution, dated 1880-1905.

Blyth.  Match-Stick Magic.  1921.

Three-way bridge, p. 52.  Three matchsticks on three teacups or tumblers.

Four-way bridge, pp. 52-53.  "rather more stable".

Collins.  Book of Puzzles.  1927.  The bridge builder's puzzle, pp. 41-42.  Three and four match bridges over goblets.

The Bile Beans Puzzle Book.  1933.  No. 15.  Use four matches to support a wine glass above four wine glasses.

Ripley's Believe It or Not!, 16th series, 1971, unpaginated, next to last page, shows six strips forming the sides of a hexagon and then extending, so each strip goes  over, under, over, under,  four other strips, forming six versions of the three knife configuration.

Doubleday - 2.  1971.  What a corker!, pp. 17-18.  Four matches have to support a cork off the table, and the match heads must not touch the table.

 

          11.O. BORROMEAN RINGS

 

            New section. 

            The Borromean rings occur as part of the coat of arms of the Borromeo family, who were counts of the area north of Milan since the 15C.  The Golfo Borromeo and the Borromean Islands are in Lago Maggiore, off the town of Stresa.  In the 16 and 17C, the Counts of Borromeo built a baroque palace and gardens on the main island, Isola Bella.  The Borromean rings can be seen in many places in the palace and gardens, including the sides of the flower pots!  Although the Rings have been described as a symbol of the Trinity, I don't know how they came to be part of the Borromean crest, though the guide book describes some of the other features of the crest.  (Thanks to Alan and Philippa Collins for the information and loan of the guide book.)  Perhaps the most famous member of the family was San Carlo Borromeo (1538-1584), Archbishop of Milan and a leader of the Counter-Reformation, but he does not seem to have used the rings in his crest.

 

Clarence Hornung, ed.  Traditional Japanese Crest Designs.  Dover, 1986.  On plates 10, 20, 24 & 39 are examples of Borromean rings.  On plate 10, the outer parts of the rings are split open.  On plates 20 and 23, the Borromean rings are in the centre of an extended pattern.  On plate 39, three extra rings are added, giving the Borromean ring pattern four times.  These designs have no descriptions and the only dating is in the Publisher's Note which says such designs were common in the 12C-17C.

Claude Humbert, ed.  1000 Ornamental Designs for Artists and Craftspeople.  Ill. by Geneviève Durand.  [As: Ornamental Design; Office du Livre, Fribourg, Switzerland, 1970.]  Dover, 2000.  Design 260 on p. 73 is like the pattern on plate 10 of Hornung, but more opened.  It is only identified as from Japan.

Pietro Canetta.  Albero Genealogico  Storico Biografico della nobile Famiglia Borromeo.  1903.  This is available at:  www.verbanensia.org/biographica/Canetta%20‑%20albero%20genealogico.htm .  This says it is copied from a manuscript of the archivist Pietro Canetta, with a footnote: Il Bandello, p. 243, vol. VIII.  I suspect that this refers to a publication of the MS.

                    This simply says that the three rings represent the three houses of Sforza, Visconti and Borromeo which are joined by marriages.  [Thanks to Dario Uri [email of 17 Jul 2001] for this source].

Takao Hayashi has investigated the Borromean Rings in Japan and the following material comes from him.

Dictionary of Representative Crests.  Nihon Seishi Monshō Sōran (A Comprehensive Survey of Names and Crests in Japan),  Special issue of  Rekishi Dokuhon (Readings in History),  Shin Jinbutsu Oraisha, Tokyo, 1989, pp. 271-484.  Photocopies of relevant pages kindly sent by Takao Hayashi. 

                    The Japanese term for the pattern is 'mitsu-wa-chigai mon' (three-rings-cross crest). 

                    Crest 3077 is called 'mitsu-kumi-kana-wa mon' (three-cross-metal-rings crest) and is a clear picture of the Borromean Rings with thin rings.

                    Crest 3083 = 3930 is the Borromean Rings with moderately thick rings and an extra rounded triangular design behind it.

                    Crest 3114 is the example on Hornung's plate 10.

                    Crests 3638-3644 are Borromean Rings with open rings, almost identical to the example in Humbert.

                    Crest 3850 is the Borromean Rings with the rings being wiggly diamond shapes.

                    Crest 3926 is called 'mitsu-wa-chigai mon' (three-rings-cross crest) and is a clear picture with moderately thick rings. 

                    The pattern usually occurs in a group of interlinked rings patterns called 'wa-chigai mon' (rings-cross crest), starting with two interlocked rings in various surrounds (crests 3921-3925), then going to three rings (crests 3926-3931, all containing the Borromean Rings) and then more rings, but without the Borromean property, though some are partially Borromean in that some of the rings are not really interlocked and will be released when another ring is removed.  3931 is the example on Hornung's plate 10 without the surrounding circle.  3939 is on Hornung's plate 39. 

                    Crests 4058 and 4060 are Borromean Rings where the rings are hexagons.

                    Crest 4284 is the Borromean Rings where the rings are squares.

Hayashi says the patterns may have developed from a linear pattern of interlocked or overlapping circles used as early as the 8C or 9C.  But the Borromean Rings do not appear before the Edo Period (1603-1867).

Kansei Chōshū Shokafu.  This is a book of family records compiled in 1799-1812.  ??NYS -- information supplied by Hayashi.  Three families use the Borromean Rings as a crest:  the Tanabes (from c1630);  the Fushikis (from c1700);  the Kobayashis (from c1700).

P. G. Tait.  Letter to J. C. Maxwell in Jun 1877.  Quoted on p. 81 of Chris Pritchard: Aspects of the life and work of Peter Guthrie Tait, pp. 77-88  IN: James Clerk Maxwell Commemorative Booklet; James Clerk Maxwell Foundation, Edinburgh, 1999.  Tait says he is confused about the diagram which he draws, but does not name, of the Borromean rings: "This is neither Knot nor Link.  What is it?"

Birtwistle.  Math. Puzzles & Perplexities.  1971. 

Ring-ring, pp. 135-136 & 195.  Three Borromean rings.

String-ring, pp. 136 & 195.  If the rings are loops of string, find other ways to join them so that all three are joined, but no two are.  Find a way to extend this to  n  loops.

 

          11.P.  THE LONELY MONK

 

            New section.  I know of earlier examples from perhaps the 1950s, but the problem must be much older.

            A monk starts at dawn and walks to top of a mountain to meditate.  Next day, at dawn, he walks down.  Show that he is at some point at the same time that he was there on the previous day.  This can be approached in two ways. 

            Let  D  be the distance to the top of the mountain and let  d(t)  be his position at time  t  on the second day minus his position at time  t  on the first day.  Then  d(dawn) = D  while  d(evening) = -D,  so at some time  t,  we must have  d(t) = 0.

            Equivalently, draw the graphs of his position at time  t  on both days.  The first day's graph starts at  0  at dawn and goes up to  D  at evening, while the second day's graph begins at  D  at dawn and goes down to  0  at evening.  The two graphs must cross.

            This is an application of either the Intermediate Value Theorem of basic real analysis or of the Jordan Curve Theorem of topology. 

 

Ivan Morris.  The Lonely Monk and Other Puzzles.  (Probably first published by Bodley Head.)  Little, Brown and Co., 1970.  (Later combined into the Ivan Morris Puzzle Book, Penguin, 1972.)  Prob. 1, pp. 14-15 & 91.  Monk leaves his mountaintop retreat to go down to the village at 5:00 one morning and starts back at 5:00 the next morning.  Is there always a place which he is at at the same time each day?  A note says the problem is based on an idea of Arthur Koestler.

Yuri B. Chernyak & Robert S. Rose.  The Chicken from Minsk.  BasicBooks, NY, 1995.  Chap. 5, prob. 9: Another triumph of central planning, pp. 43-45 & 129-130.  This is a complicated version of the problem.  There are two roads from  A  to  B,  such that for each point on the first road, there is a point at most 20 m away of the second road.  This is verified by sending two cars along the two roads, attached by a 20 m phone line.  Is it possible to start trucks of width 22 m simultaneously from  A  to  B on one road and from  B  to  A  on the other road?  The solution uses a graphical method, plotting the distance from  A  of the first vehicle going along the first road, versus the distance from  A  of the second vehicle, going along the second road.  If the distance along the longer road is  D,  the verification cars give a graph starting at  (0, 0)  and ending at  (D, D),  while the trucks give a graph starting at  (0, D)  and going to  (D, 0).

 

          11.Q. TURNING AN INNER TUBE INSIDE OUT

 

          New section.

 

Gardner.  SA (Jan 1958)  c= 1st Book, Chap. 14: Fallacies.  Says that the process was illustrated in SA (Jan 1950) and a New Jersey engineer sent in an inner tube which had been turned inside out.  Gardner then describes and illustrates painting rings in both directions, one inside, the other outside so they are interlinked at the beginning.  He then draws the inner tube apparently inverted, with the rings unlinked and asks for the resolution of this paradox.  The solution is given in the Addendum: "the reversal changes the 'grain,' so to speak, of the torus.  As a result, the two rings exchange places and remain linked."  Several readers made examples using parts of socks.

Victor Serebriakoff.  (A Second Mensa Puzzle Book.  Muller, London, 1985.)  Later combined with A Mensa Puzzle Book, 1982, Muller, London, as: The Mensa Puzzle Book, Treasure Press, London, 1991.  (I have not seen the earlier version.)  Problem P13: The Yonklowitz diamond, pp. 164-166, & answer A22, pp. 237-238.  After some preliminaries, he asks three questions.

                    A)  Can you pull an inner tube inside out through a small hole?

                    B)  What is the resulting shape?

                    C)  If you draw circles in the two directions, one inside and one outside so they are linked at the beginning, are they linked at the end?

                    His answers are:  A) yes;  B) the result has the same shape;  C) the circles become unlinked.  The last two are wrong.  The shape changes -- effectively the two radii change roles.  When this is carefully seen, the two circles are seen to still be linked.

 

          11.R. STRING FIGURES

 

            This topic has a vast literature and I intended to omit it as it is very difficult to summarise.  However, I have recently acquired the books of Haddon which are an excellent source, so I have included here just a few books with extensive bibliographies which will lead the reader into the literature.  I have just learned of: 

            International String Figure Association, PO Box 5134, Pasadena, California, 91117, USA.  Web: http://members.iquest.net/~webweavers/isfa.htm .  They publish Bulletin of the International String Figure Association (formerly without the International).  [SA (Jun 1998) 77.]

            String figures seem to be relatively late in getting to Europe.  The OED cites  1768, 1823, 1824, 1867, 1887  for Cat's Cradle.  Magician's Own Book and Secret Out are the earliest examples I have where the process is explicitly described, but I have only started looking for such material.

 

Abraham Tucker.  The Light of Nature Pursued.  7 vols.   Vol. I, (1768); reprinted 1852, p. 388.  ??NYS -- quoted in the OED.  "An ingenious play they call cat's cradle; ...."  This is the first citation in the OED.

Charles Lamb.  Essays of Elia: Christ's Hospital.  1823, p. 326.  ??NYS -- quoted in the OED.  "Weaving those ingenious parentheses called cat-cradles."  A popular book says Lamb is describing his school days of 1782.

Child.  Girl's Own Book.  Cat's Cradle.  1833: 76;  1839, 63;  1842: 57.  "It is impossible to describe how this is done; but every little girl will find some friend kind enough to teach her." 

Fireside Amusements.  1850: No. 31, p. 95;  1890: No. 31, p. 71.  As a forfeit, we find: "Make a good cat's cradle."

Magician's Own Book.  1857.  No. 9: The old man and his chair, pp. 8-10.

The Secret Out.  1859.  The Old Man and his Chair, pp. 231-236.  Like Magician's Own Book, but many more drawings and more detailed explanation.

Caroline Furness Jayne.  String Figures.  Scribner's, 1906, ??NYS.  Retitled: String Figures and How to Make Them;  Dover, 1962.  97  figures given with instructions;  another 134 figures are pictured without instructions.  55  references.

Kathleen Haddon [Mrs. O. H. T. Rishbeth].  Cat's Cradles from Many Lands.  Longmans, Green and Co., 1911.  16 + 95 pp, 50 string figures and 12 tricks, 14 items of bibliography.

Kathleen Haddon [Mrs. O. H. T. Rishbeth].  Artists in String.  String Figures: Their Regional Distribution and Social Significance.  Methuen, 1930.  Discusses string figures in their cultural setting, describing five cultures and some of their figures.  41  string figures given with instructions, and some variants.  P. 149 says "descriptions of over eight hundred figures have been published and many more have been collected."  The Appendix on p. 151 gives the numbers of figures classified by type of objects represented and location, with a total of  1605  figures plus  101  tricks.  Two bibliographies, totalling  116  items.  (An abbreviated version, called String Games for Beginners, containing  28  of the figures and omitting the cultural discussions, was printed by Heffers in 1934 and has been in print since then, recently from John Adams Toys.)

Alex Johnston Abraham.  String Figures.  Reference Publications, Algonac, Michigan, 1988.  31  figures given with instructions plus a chapter on Cat's Cradle.  156  references.

 

          11.S.  PUZZLE KNIVES

 

          New section, inspired by finding Moore.

 

Mr. X [cf 4.A.1].  His Pages.  The Royal Magazine 26 (Jul-Oct 1911).  Each issue offers puzzle penknives as prizes.

Simon Moore.  Penknives  and other folding knives.  Shire Album 223, Shire, 1988.  Pp. 5-6 describe and illustrate several examples which he says became popular in the early 17C.  Most of these are of the type where the two sides of the handle have to be separated, then one side turns 360o to bring the blade out 180o and bring the handle back together.  There are a few 18C and 19C examples, but they were basically superseded by the spring-back knife in the late 17C.  On p. 10, he says the 'lockback' knives, with a mechanism to prevent accidental closure, were made from the late 18C.

                    A mid 17C example has lugs between the two sides of the handle which are locked by a pin, whose head is concealed by a false rivet near the pivot.  When the rivet is moved, the pin can be removed.

                    An example dated 23 Oct 1699 has two dials which have to be set correctly to release the locking lugs.