You are on page 1of 335



English translation
A rnold Dresden
with additions of the author

Fourth edition


Hardcover edition published throughout the world, exclusive of
North America, by Kluwer Academic Publishers, Spuilboulevard 50,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands

Paperback edition published throughout the world by The Scholar's
Bookshelf, 51 Everett Drive, P.O. Box 179, Princeton Junction,
New Jersey 08550, United States of America

Hardcover edition published in North America by
The Scholar's Bookshelf

Copyright © 1975 by Noordhoff International
Softcover reprint of the hardcover 4th edition 1975
Publishing, a division of Kluwer Academic Publishers,
Dordrecht, The Netherlands

All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted, in any
form by any means, electronic, mechanical, photocopying,
recording or otherwise without the prior permission of the
copyright owner.

First edition 1954
Second edition 1%1
Third edition 1969
Fourth edition 1975
Fifth edition 1988

Kluwer Academic Publishers

ISBN-13: 978-94-010-7115-4 ~ISBN-13: 978-94-009-1379-0
DOl: 10.1007/978-94-009-1379-0

First Scholar's Bookshelf hardcover printing, 1988
First Scholar's Bookshelf paperback printing, 1988


Soon after the publication of
my"Ontwakende W etenschap"the need for an English translation
was felt. We were very glad to find a translator fully familiar with
the English and Dutch languages and with mathematical terminol·
ogy. The publisher, Noordhoff, had the splendid idea to ask
H. G. Beyen, professor of archeology, for his help in choosing
a nice set of illustrations. It was a difficult task. The illustrations
had to be both instructive and attractive, and they had t~ illustrate
the history of science as well as the general background of ancient
civilization. The publisher encouraged us to find better and still
better illustrations, and he ordered photographs from all over
the world, with never failing energy and enthusiasm. Mr. Beyen's
highly instructive subscripts will help the reader to see the inter·
relation between way of living, art, and science of the ancient world.

Thanks are due to many correspondents, who have suggested
additions and pointed out errors. Sections on Astrolabes and
Stereographte Projection and on Archimedes' construction of
the heptagon have been added. The sections on Perspective and
on the Anaphorai of Hypsicles have been enlarged.

In the second English edition I have incorporated an important
discovery of P. Huber, which sheds new light upon the role of
geometry In Babylonian algebra (see p. 73). The section on Heron's
Metrics (see p. 277) was written anew, follOWing a suggestion of
E. M. Bruins.

Zurich. 1961 B. L. VAN DER WAERDEN


Why History of Mathematics?

Every one knows that we are living in a technological era. But it is not often
realized that our technology is based entirely on mathematics and physics.
When we ride home on the streetcar in the evening, when we turn on the
electric light and the radio, everything depends on cleverly constructed physical
mechanisms based on mathematical calculations. But more than that! We owe
to physics not only these pleasant articles of luxury, but, to a large extent, even
our daily bread. Apart from the fact that our grains come to us, chiefly by steamer
from overseas, our own agriculture would be far less productive without artificial
fertilizers. Such fertilizers are chemical products, and chemistry depends on
physics. l
But science has not brought us blessings only. The destructive armaments which
mankind uses at the present time to knock its own civilization to pieces are also
products to which the development of mathematics and physics have inevitably
Our spiritual life is also influenced hy science and technology, in a measure but
rarely fully understood. The unprecedented growth of natural science in the 17th
century was followed ineluctably by the rationalism of the 18th, by the deification
of reason and the decline of religion; an analogous development had taken place
earlier, in Greek times. In a similar manner, the triumphs of technology in the
19th century were followed in the 20th by the deification of technology. Unfor-
tunately, man seems to be overly inclined to deify whatever is powerful and
These considerations indicate that science has put its stamp on the whole of
our life, material and spiritual, in its beneficent and in its evil aspects. Science
is the most significant phenomenon of modern times, the principal ingredient of
our civilization - alas!
But if this be true, then the most important question for the history of culture
is: How did our modern natural science come about?
It will be conceded that most historical writings either do not consider this
question at all, or else deal with it in a very unsatisfactory manner. For example,
which are the histories of Greek culture that mention the names of Theaetetus
and of Eudoxus, two of the greatest mathematicians of all times? Who realizes
that, from the historical point of view, Newton is the most important figure of
the 17th century?
Every physicist will admit that the mechanics of Newton are the foundation of
modern physics. Every astronomer knows that modern astronomy begins with
Kepler and Newton. And every mathematician knows that the largest domain of
1 The ruder should bur in mind that this book was addressed originally to the public of the N,·thcrl.nds.

vith the laws of the lever. a number of scattered communications of Plato. to 200 B. as a result of the indefatigable industry of Otto Neugebauer and his collaborators. the greatest genius of antiquity. In short.C. we know now the mathematical cuneiform texts. For this reason.C. succeeded in discovering the hidden algebraic element in Greek mathematics and in demonstrating its connection with Babylonian algebra. The History of Greek Mathematics. In the first place. No . the "Theorem of Pythagoras" In recent times however more light has penetrated into the darkness. And Newton's integral calculus can be understood only as a continuation of Archimedes' determination of areas and volumes. The history of mechanics as an exact science begins '. Without the conic sectIOns of Apollonius. He placed the crown on the task of renovating antique astronomy. of mechanics and of astronomy. to which terrestrial as well as celestial objects are subject. the first three of these four centuries were enveloped in twilight. the laws of hydrostatics and the determinatIOn of mass centers by Archimedes. because we possess only two original texts from this period: the fragment con- cerning the lunules of Hippocrates and that of Archytas on the duplication of the cube. Pappus. all the developments which converge in the work of Newton. contain little more about this period than speculations concerning things of which we really do not know anything. To this can be added two brief fragments of Archytas. whereas Archimedes. and a self- contradictory set of Pythagorean legends. begun by Copernicus and Kepler.4 PREFACE modern mathematics. but especially on the earliest history of arithmetic and of algebra. following in the tracks of Zeuthen. would have been impossible. The work of Newton can not be understood without a knowledge of antique science. Without the stupendous work of Ptole- my. He discovered a general method for solving all problems of differentiation and integration. such as. Kepler's Astronomia Nova. the older works. covers the four centuries from 600 B. Proclus and Eutocius. is Analysis. Neugebauer. such a~ Cantor's Geschichte der Mathematik. and hence the mechanics of Newton. which completed and closed antique astronomy. which has its roots in the Differential and Integral Calculus of Newton. had not progressed beyond special methods for particular problems. Thus the work of Newton constitutes the foundation for by far the greater part of modern exact science. It was Newton who discovered the fundamental laws of motion. Until recently."1hose of mathematics. for example. begin in Greece. his deVelopment of the law of gravitation is equally unthinkable. from fhales to Apollonius. Aristotle. the part most important for physics. Newton did not create in a void. which Newton knew thoroughly. which have thrown an entirely new light not only on the Theorem of Pythagoras.

established the penodical "~ellen und Studien zur Geschichte der Mathematik. In 1927. a specifically Greek character. how. We shall see moreover that the mathematical method of proof served as a prototype for Plato's dialectiCS and for Aristotle's logic. This work. and even to religion (Pythagoras!). As to these latter writers. through the work of Plato's friends Theaetetus and Eudoxus. Reidemeister and others to obtain highly important results. reciprocally. These different things have not as yet been brought together in a book. Aristarchus. and there are excellent works on Euclid. and. Stenzel and Toeplitz. with Neugebauer. we shall therefore be able to confine our- selves to the most interesting and the most important things. a cuneiform text. to learn more about Greek mathematics by means of an analysis of Plato. had probably better be called Babylonian". . Eva Sachs had rescued the excellent mathematician Theaetetus from oblivion. mathematics was brought to the state of perfection. PREFACE 5 longer does the history of algebra begin with Diophantus.C. beauty and exactness. Mathematics is a domain of intellectual activity. proves to be largely a compilation of mathematical fragments. but also to architecture and technology..C. an excellent book by O. how. It was their purpose to get to know more about the philosophy of Plato by an analysis of the fundamental concepts of Greek mathematics. by dusting them off and then replacing them in the mathe- matical historical environment from which they had originally come. in 1937 Neugebauer wrote: "What is called Pythagorean in the Greek tradition. Neugebauer on "Vorgriechische Mathe- matik". it is true. showed that he was entirely right. we shall pick tidbits here and there from their works and we shall try to serve them in as tasteful a manner as possible. mathematics was brought to higher and ever higher development and began gradually to satisfy the demands of stricter logic. which we admire in the elements of Euclid. The history of mathematics should not be detached from the general history of culture. written about 300 B. in the Pythagorean school and outside. This method has enabled Becker. it starts 2000 years earlier in Mesopotamia. And. A second new impulse came from philosophically oriented philology. We have. as to arithmetic. intimately related not only to astronomy and mechanics. Archimedes and Apollonius. At an earlier date. and. concerning "Pythagorean numbers". it has become possible to obtain a considerably clearer picture of Greek mathematICS of the years 500---300 B. But the principal purpose of the present book is to explain clearly how Thales and Pythagoras took their start from Babylonian mathematics but gave it a very different. quite diverse in calibre and quite varied in age. to philosophy. By carefully taking these fragments apart. dis- covered in 1943. Astronomie und Physik". Another very fertile method was the analysis of the Elements of Euclid.

The naive reader may take the use of such a method for granted. Neugebauer has translated and published all mathematical cuneiform texts. Not only is it more instructive to read the classical authors themselves (in translation if necessary). Now Cantor reasons as follows: these right angles must have been constructed by the rope-stretchers. even if. The Egyptian mathematical texts have all been translated into English or German. so as to enable the reader to judge the conclusions for himself. I have checked all the conclusions which I found in mo- dern writers. The facts are the following: "rope-stretchers" took part in laying out an Egyptian temple. Therefore the Egyptians must have known this triangle. as is my case. To avoid such errors. 4 and 5. For example. It is the intention to make this book scientific. This is not as difficult as might appear. it also gives much greater .. good translations exist in French. The plan of this book. right angles. But . German and English. and one is not a classical philologist. forming a right triangle. of all these. one cannot read either the Egyptian characters or the cuneiform symbols. This will become very clear in Chapter VII when the mathematics of Alexandria is compared with that of the classical period. and who is interested in the history of mathematics. Plato. . and that they used it for laying out right angles. and the angles at the base of temples and pyramids are nearly always. but that repeated copying made it a "universally known fact". For reliable translations are obtainable of nearly all texts. very accurately. It is to be scientific in the sense that it is to be based on a study of the sources and that its conclusions are to be supported by arguments. Only in a few doubtful cases it became necessary to consult the Greek text. and still more in Chapter VIII in the discussion of the causes for the decay of Greek mathematics. 4 and often has it been sinned against! How frequently it happens that books on the history of mathematics copy their assertions uncritically from other books.. Is this not incredible? Not that Cantor at one time formulated this hypothesis. This is nevertheless the fact. Euclid. How much value has this statement? None! What is it based on! On two facts and an argument of Cantor. and I (Cantor) can not think of any other way of constructing a right angle by means of stretched ropes than by using three ropes of lengths 3. rather than modern digests. In 90 %of all the books. Archimedes. without consulting the sources! How many fairy tales circulate as "universally known truths" ! Let us quote an example. one finds the statement that the Egyptians knew the right triangle of sides 3.6 PREFACE Political and social conditions are of very great importance for the flowering of science and for its character. but at the same time accessible to anyone who has learned some mathematics in school and in college.

particularly the Pythagorean mathematics. In Chapter VII. Archytas. In Chapter VIII. An analysis of Book X of the Elements and a reconstruction of the mathematical work of Theaetetus. The cause of the decay of Greek mathematics. the scarcity of source material made this impossible. Of the Egyptian and Babylonian mathematicians not even the names are known. . but to verify everything. Connections between the Babylonian and the Greek mathematics. If my book should lead the reader to do this. Acknowledgements. It is also possible to get an impression of the character of Thales. The feeble logic of Archytas of Taras. I thank Dr. Professor Freudenthal and especially Dr. Freudenthal's interpretation of a Babylonian textbook. For this reason. Brinkman. Theaetetus ana Archimedes can be made to stand out clearly. Reconstruction of the Pythagorean theory of numbers from the arithmetical books of the Elements. Dijksterhuis. In Chapter VI. What is new in this book. A hypothesis of Freudenthal on Indian number symbols. I advise the reader emphatically not to accept anything on my say-so. The history of the Delian problem. but striking personalities such as Pythagoras. Eudoxus and Eratosthenes. In Chapter IV. it will fully have accom- plished its purpose. Mathematics and the theory of harmony in the Epinomis. A new way of looking at the mathematics of Thales. In Chapter III. actually and according to the dialogue Platonicus. In Chapter V. who have read the manuscript critically and have made numerous useful remarks. The irrationality proofs of Theodorus of Cyrene. In some cases. I have tried to consider the great mathematicians as human beings living in their own environment and to reproduce the impression which they made on their contemporaries. In Chapter II. PREFACE 7 enjoyment.

I thank the publishers for the generous way in which they have taken all my wishes into account. In conclusion. Mr. I am very much obliged to Mr. assisted by his excellent draughtsman. Bousche.8 PREFACE I thank the many others who have helped with brief observations or with tech- nical advice. . Wijdenes who has taken care of the diagrams in his well-known careful manner.

. . . 48 . 34 What could the Greeks learn from the Egyptians? 35 CHAPTER II. . . . . . 19 Natural fractions and unit fractions. . . 40 Oldest Sumerian period (before 3000 B. . . . . . . . 27 Applied calculations . . . The Egyptians 15-36 Chronological Summary. . 18 Division . 7 CHAPTER I. . .. . . . . . 26 "Aha·calculations" . . . '>-8 Why History of Mathematics. . square roots and cube roots. . . . . 31 Inclination of oblique planes 31 Areas . . . . . . . . . . . digits and the art of computing. . . . . . . .C. . . . . . . . . . 32 Area of the hemisphere. . . . . . 45 Counting boards and counting pebbles. . 17 Multiplication. 37 How did the sexagesimaI system originate? . . . . . . . Number systems. . . . .40 . . . 6 What is new in this book . . . . . . . 7 Acknowledgements. 47 Calculation with fractions. 29 The development of the computing technique 30 Hypothesis of an advanced science . 37-61 The sexagesimal system. . . TABLE OF CONTENTS PREFACE . 15 The Rhind papyrus . . . . . . . 17 The technique of calculation .C. .) . . 3 The History of Greek. . . . . . . . . . 20 Further relations between fractions . 21 Duplication of unit fractions 22 Division once more. . . . . . . . 40 Sumerian technique of computation. . . . . . . . . 19 Calculation with natural fractions. . . 26 Complementation of a fraction to 1. . . . 15 The Egyptians as the "inventors" of geometry . 4 The plan of this book . . 23 The red auxiliaries . . . . . . 33 Volumes . . . . Mathematics. . 42 Normal table of inverses . 42 Table of 7 and of 16. . . . . . . . . . . . . 31 The geometry of the Egyptians. . 16 For whom was the Rhind papyrus written? 16 The \. . . .lass of royal scribes. 43 Squares. . . . . . . 23 The (2 : n) table . 44 The Greek notation for numbers. . . . . . . . . .) 40 Later Sumerian period (about 2000 B.

342) 72 A lesson-text (MKT II. . . _ _ . 87 From Thales to Euclid. p. . . . . no. . . . . . . . . . . 176 and 178) . . . . . . . 68 Quadratic equations (MKT III. . . 63 First example (MKT I. . . . . . . . . . _ ... 92 The travels of Pythagoras 94 Pythagoras and the theory of harmony. . . . . . . . . p. 55 Where does the zero come from? . 323) . . . . . 73 Babylonian geometry. . 68 Fifth example (MKT III. . . . . 8. . 71 Geometrical proofs of algebraic formulas? 71 Ninth example (MKT I. . 57 The abacus of Gerbert. 70 Seventh example (MKT I. . no. . . . 97 . . . . . . . . 54 Aryabhata and his syllable-numbers . ..10 TABLE OF CONTENTS Sexagesimal fractions . . . . _ . Kharosti and Brahmi . . . . . . 63 Second example (MKT I. . 62 Babylonian algebra . . . p. 83 Thales of Milete . . 77 Plimpton 322: Right triangles with rational sides. .. . . . . pp. . . . 85 Prediction of a solar eclipse. . 82-104 Chronological summary. .. . . . . p. 154). . . . . 80 GREEK MATHEMATICS CHAPTER IV. p. . . 53) . . . . . . . 58 CHAPTER III. . . . . . . 18). 80 Summary . 53 The invention of the positional system. . . . . 14) . 56 The triumphal procession of the Hindu numerals . . . . . 39) . . 113). . . . . . 65 Third example (MKT I. . . . . 96 Perfect numbers . . . . 82 Hellas and the Orient . . 76 Babylonian theory of numbers . . . 9. . 75 Volumes and areas .. 62-81 Chronological summary. . . 6).. . . p. . . .. . . . . p. . . p. 75 The "Theorem of Pythagoras" (MKT II. .. . . . . 51 Number systems. . . . . Babylonian mathematics __ . . 90 Pythagoras of Samos. . . 75 Frustra of cones and of pyramids (MKT. p. The age of Thales and Pythagoras. 95 Pythagoras and the theory of numbers . . . . 53 Poetic numbers . . 485) . . . p. 53 The date of the invention . . . . . 204). . . 99) _ . . . . . . . . . . 86 The geometry of Thales . . . . p. . . . . 66 Fourth example (MKT I. . .. 280). . p. p. . . 69 Sixth example (MKT III. . . . . . 78 Applied mathematics . . . . . . . 50 Hindu numerals. 77 Progressions (MKT I. . . . 70 Eighth example (MKT I.. . . 63 Interpretation. .. . . .

. . . . . . 127 Democritus of Abdera . . . 179 Eudoxus as an astronomer . . 138 The duplication of the cube 139 Theodorus of Cyrene. 137 Cone and pyramid . . . . 150 The style of Archytas . . . . . . . . . . . 104 CHAPTER V. . 98 Pythagoras and geometry. . . 106-147 Hippasus . . . . 129 Squaring the circle. . . . . . . . . . 98 Figurate numbers. 108 The theory of numbers." . . 131 Solid geometry in the fifth century. 146 Hippias and his Quadratrix. . . 108 The theory of the even and the odd. . 102 Antique measuring instruments. . 162 Theaetetus . . . . 137 Plato on solid geometry . .::ritus. The century of Plato. . 175 Eudoxus of Cnidos. . . . . 130 Antiphon . . . . . 1-46 The main lines of development. . . . . . . . . . . 159 According to Menaechmus . . . . . . . . and Perspective. 125 Lateral and diagonal numbers. . . . . . . . . 1-41 Theodorus on higher curves and on mixtures. . . . . . 136 Demo. . . 155 The duplication of the cube . . . . . 165 Analysis of Book X of the Elements . . . 126 Anaxagoras of Clazomenae . 149 Tht: duplication of the cube . . 100 The astronomy of the Pythagoreans 101 Summary . 141 Theodorus and Theaetetus . 130 Hippocrates of Chios. . . . . 180 . . . 128 Oenopides of Chios. . . . . . . . 146 CHAPTER VI. . . . . . . The golden age . 152 Book VIII of the Elements. . . 102 The tunnel on Samos. . . . . . . . . . . . . . 110 The solution of systems of equations of the first degree 116 geometry: . TABLE OF CONTENTS 11 Amicable numbers . . 108 Proportions of numbers . . . . . . . . 118 Why the geometric formulation? . 153 The Mathemata in the Epinomis . . . . . 168 The theory of the regular polyhedra . . . . . 14&-200 Archytas of Taras . . . . . 106 The Mathemata of the Pythagoreans . . . . . 116 Geometnc Algebra . . . . 173 The theory of proportions in Theaetetus . .

. . . . 199 Lost geometrical writings. . . 238 Conica . . 211 The works of Archimedes . . . . . . . . . . . . 226 The other works of Archimedes 227 Eratosthenes of Cyrene . 190 Dinostratus. . . . . . . 193 On the rotating sphere. . . . . . . . 230 Duplication of the cube. . . . . . . . . . . . . . 208 Archimedes as an astronomer . 223 On conoids and spheroids . 220 On sphere and cylinder II . .C.12 TABLE OF CONTENTS The exhaustion method. . . . . 228 Life . . . . . . . . 202 Archimedes' measurement of the circle . . . 189 Menaechmus . . 246 Conjugate diameters and conjugate hyperbolas. . . . 235 The trisection of the angle . . . 194 On the rising and setting of stars 195 Euclid . . . . . . . . . . . . . . . . 222 On spirals . 208 Stories about Archimedes. . . . . . . . . . . 212 The quadrature of the parabola. . 216 On sphere and cylinder I. 223 The notion of integral in Archimedes. . 211 The "Method" . . . 200 Euclid's work on applied mathematics 200 CHAPTER VII. . . . . . . . 224 The book of Lemmas. . . . . . . . . . . 225 The construction of the regular heptagon. . . 184 The theory of proportions . . 195 The "Elements" 196 The "Data" . . . 206 Archimedes. . . 245 The derivation of the symptoms according to Apollonius. The Alexandrian Era (330-200 B. 228 Chronography and measurement of a degree . . . . . . . . . . . . . . . .30 Theory of numbers. 237 The theory of the epicycle and of the excenter. . . . . . 236 The duplication of the cube in Nicomedes 236 Apollonius of Perga. . . . . . . . . . . . . . . . . . . . . . . . . . 231 Medieties . . . . . . . . . . . 201-263 Aristarchus of Samos. 241 The ellipse as a sfction of a cone according to Archimedes 243 How were the symptoms derived originally? . 231 Nicomedes . 204 Tables for the lengths of chords . . 245 A question and an answer . . .). . . . 2. 187 Theaetetus and Eudoxus . . . 198 On the division of ligures. . 248 . . . . . 240 The conic sections before Apollonius. 191 Autolycus of Pitane.

. . . 286 A porism of Euclid. . . . . . . . . . . . 267 1. . . . . . . .. . 269 Anaphora . . . . . . 260 The sixth. . 258 Loci involving 3 or 4 straight lines. . . . . . . . . . 289 Theon of Alexandria . 271 Plane trigonometry. . . 259 The fifth book . 278 Diophantine equations 279 The precursors of Diophantus 279 Connection with Babylonian and Arabic algebra . Diodes . Hypsicles . 266 The commentaries of Pappus of Alexandria. 268 Isoperimetric figures. . 276 Metrics . . . 274 Transversal proposition. . . . . . . .. . . 269 3. . . . . . . . 261 Further works of Apollonius. . 287 The theorem on the complete quadrangle. 264 The inner causes of decay . . . . .. . . 285 Pappus of Alexandria. . . . . . . . 252 Cones of revolution through a given conic . . 267 The epigones of the great mathematicians. . 270 History of trigonometry . 265 1. . . . . . 269 The fourteenth book of the Elements. 290 Hypatia . . . . . . . . . . . Proclus Diadochus. . . 278 Arithmetica. . . . . _ .. .. 251 The two-tangents theorem and the transformation to new axes. . . 280 The algebraic symbolism 281 From Book II . seventh and eighth books . . The difliculty of the written tradition . . . 256 The second book. 249 The equation referred to the center . . . 284 From Book V . 274 Menelaus . 290 The Athens school. 258 The third book. . . . . The difliculty of geometric algebra. 268 2. . 267 The cissoid. . . The decay of Greek mathematics. . 285 From Book VI . . 275 Heron of Alexandria . . . . . . 277 Diophantus of Alexandria. TABLE OF CONTENTS 13 Tangent lines. Zenodorus. . . . 291 . . . .. . . . . 271 Spherical trigonometry . . . . . . . . . . . . . . . 265 2. . . " . . . . . . . . 283 From Book IV . . . 288 Theorem of Pappus. 290 Isidore of Milete and Anthemius of Tralles . 282 From Book III . . . . . . 264-291 External causes of decay . . . . . . . . 261 CHAPTER VIII.

" The rope-stretchers ("harpedonaptai"). to my thinking. Gandz. because there the priestly class was allowed leisure.. it became necessary for the purpose of taxation. • See S. Qutlltn ••J StyJi•• z".oa.000 2000-1800 Literature Rhind papyrus and Moscow The Middle Kingdom The goldsmith's art papyrus I Star calendars on sarcophagi 1700 The Hyksos I Ahmes copies the Rhind domination I papyrus 1600-1100 I New theology (Echnaton) Very primitive astronomy The New Kingdom Architecture (Senmuth's tomb) I Sculpture 3OOB. . Aristotle writes (Metaphysics AI) 1: "Thus the mathematical sciences originated in the neighborhood of Egypt. to determine how much land had been lost.C.D. The largest and most famous of these is 1 All references to classical writers." Herodotus. G. mentioned here and in the ""qud. I In view of the fact that the mathematical abilities of the Egyptians are so highly praised by the Greeks. it is certainly worth while to have a-rook at the Egyptian mathematical texts.When the Nile had flooded an agricultural tract.p"n. For instance. B I. not even the so-called rope-stretchers among the Egyptians. to whom Democritus here refers. Alexandria as center of Highest development of Hellenism Greek art and science Greek science Rise of astrology Egyptian arithmetic and astronomy remain very primitive The Egyptians as "inventors" of seometry The Greeks generally assumed that mathematics had its origin in Egypt. Ha'T". whose principal measuring instrument is every- where the stretched cord. 255./ue In Mathe_iI:. CHAPTER I THE EGYPTIANS Chronolosical Summary General History I History of Civilization I History of Science 3000 Menes Hieroglyphics I Number symbols to The Old Kingdom I Pyramids 100. the Greeks learned the art of measuring land" (Herodotus II 109).300 A. follow the translation of the corrcaponding author in Loeb's Classical Library. . "from this."hi. looked at the more prac- tical side of the matter. who knew Egypt better. are probably surveyors. unless otherwise noted. And Democritus writes: "No one surpasses me in the con- struction of lines with proofs. Di.

primarily the people concerned with the applications. the geometrical problems 10 our texts are entirely designed for the applications. which provide exercises in the difficult art of calculation with fractions . or the volume of a barn for storing grain. named after Mr. in order to apply these to various practical problems with which the officials of the great state had to deal. knowledge of all secrets . But it soon becomes evident that we shall not witness the revelation of the origin of things."'. such as the distribution of wages among a number of laborers.. . Rhind.. the conversion of different measures for grains. H. In view of this. " he undertakes to teach. but we are concerned with calculating the area of a piece of land. 1923. As we shall see. Calculation with fractions. There is therefore reason to expect that we may become acquain- ted through this text.' Papyr. the calculation of the amount of grain needed for the production of a given quantity of bread or of beer. 1 This papyrus was written during the period of the domination of Egypt by the Hyksos (after 1800 B. taught systematically in the Rhind 1 See the excellent edition of T. the calculation of areas and volumes. but not of the priest.. London. Aristotle starts fro~ an entirely wrong presupposition. that theoretical interests led them to devote their leisure to mathematics.) but. as Greek scholars did in his day? Or was it. Weare not asked to prove or to construct something. The papyrus starts in a very promising way: "Complete and thorough study of all things. there was no well- organised estate of priests. A. Accordingly. who cultivated geometry? . or of the official who has to erect a granary. E. The Rhi. Problems of this kind are the concern of the surveyor..a'.. the period of our texts... insight into all that exists.d Ma. Among these are also found however some purely theoretical questions. also belong to the Middle Kingdom. one is led to ask For whom was the Rhind papyrus written? Is Aristotle right in asserting that the pri"'Sts were the actual carriers of mathe- matical development. Peet. as its writer assures us. who bought this text in Luxor and then willed it to the British Museum. All other texts that have mathematical contents and that are known to us.16 CHAPTER I The Rhind papyrus.C. the task of the priests was usually carried on by lay- men. but that we shall merely be initiated into the secrets of numbers and into the art of calculating with fractions. During the Old Kingdom of the Egyptians (3200-2000) and even dur10g the Middle Kingdom (2000-1800). with the principles of the mathematical knowledge of this period. . as Herodotus and Democritus thought. along with their ordinary occupations. it derives from a prototype which dates from the Middle Kingdom (2000---1800).

fro".. is applied in texts on economics." .. and the task of teaching you to perform it falls on my shoulders" . 1 Who needs such calculations? It is The class of the royal scribes We get a good idea of the problems of these officials from the papyrus Anastasi I.t""S . i'iIew Jersey. . how many bricks are needed for a certain structure. A (building-)ramp is to be constructed.".. e. It seems clear that this papyrus was intended for use in a school for scribes. 1 Griffith. 55 cubits wide. etc. Plate VUI. p.. you. The writer of the letter reproaches him with his inability to solve these problems without his help. p. Tilt ""'" SC. ""fo.. They all put their trust in you and say 'You are the clever scribe. who has to calculate. Himrtic P. The number system of the Egyptians is as simple and as primitive as that of the Romans. and filled with reeds and beams. and the scribes are all asked together. etc.". Neugebauu. which is indeed the first topic dealt with in the Rhind papyrus. often they gradually increase in difficulty. and you say 'reckon it out'. London 1893..g.. 730 cubits long. Let us now have a closer look at the papyrus. 60 cubits high at its summit. with a batter of twice 15 cubits and its pavement 5 cubits. 79. You are given a lake to dig. my friend! Decide for us quickly!' Behold your name is famous . it is a strictly decimal system. such as the Kahun thousand. how many soldiers are needed for digging a ditch. who are his Royal Scribe" . 30 cubits in the middle.. "I will cause you to know how matters stand with you. THE EGYPTIANS 17 papyrus. when you say 'I am the scribe who issues commands to the army' . The quantity of bricks needed for it is asked of the generals. without one of them knowing anything. we have to familiarize ourselves with The technique of calculation. 15. To start with.. You come to me to inquire concerning the rations for the soldiers. You are deserting your office. 1952. • O. One has obviously to think of the person addressed as some one who occupies an important position as a "scribe of the army". "I cause you to be abashed(?) when I disclose to you a command of your lord. containing 120 compartments.. In hieroglyphics: I = one therefore III n = ten nn nn therefore ~ = one hundred 1 . KIIh. "the clever scribe who is at the head of the troops. Answer us how many bricks are needed for it?" I The problems in the Rhind papyrus are of exactly the same general character.nd Gvroh. Princeton. in which a scribe ridicules another-one for his lack of skill.

but in hiuatic 1Cript. No. it was taught as "Egyptian calculation". Even during the Middle Ages. which can readily be changed into h. "mediatio".glyphiat. linn /4 48 &4 nn 1111 nn9 ~ IIInnnnn III nonn 1\" .. was taught .g. But very peculiar is Multiplialtion This is accomplished by doubling and adding the results. As an example we quote from the Rhind papyrus (Pert edition).1 1 12 IIII~ 4:t "Z.6. as. Duplication is a special case of addition. There is no difficulty in adding these numbers. The result 144 is accompanied by the hieroglyph dmd. where 16 x 16 is calculated as follows: / 1 16 /10 160 /5 80 sum 256 This Egyptian method of multiplication is the foundation of the entire art of calculation.. in the Kahun papyrus. which represents a scroll with a seal. in the Greek schools.' The ~Pyrua itadf is not written in hieroglyphics.d .. The numbers which are to be added. "duplatio" was looked upon as an independent operation. of hundreds. are indicated by an inclined line on the right (placed on the left in the "translation"). so that this presents no difficulties. It must be very ancient.- 1111 /8 96 Sum 144 441 d.. it is only necessary to count the numbers of units. without change. No. but it has been able to maintain itself.. etc. the multiplication of12 by 12. of tens...". e.. sometimes the lO-fold multiple is halved. 1111/ . Afong with it.. into the Hellenistic period.18 CHAPTER I An numbers that arose could be represented by placing these symbols in a row. one frequently multiplies by 10.. . . And. 32.. 2 24 IIllnn . To proceed more rapidly. first in hieroglyphics (to be read from right to left) 1 and then in modern no- tation lin 2. the process of halving." a Four times 12 and eight times 12 are added to produce 12 times 12..

"the fourth part" %. In addition. do not exist in Egyptian mathematics. As the technique of calculation developed. In modern life. says No. In Greek one also speaks of "the two parts" % . it is indicated as such by the symbol of the sr. It should not take him long to familiarize himself with it and thus to acquire the point of view of the Egyptians.roll. necessary to make a whole out of the two parts is "the third part". The stock exchange calculates also with an eighth or a sixteenth of a percent. one rarely uses fractions except %. a mastery of the Egyptian process for division is absolutely indispensable. The French language still has a special word for a third (tiers). The interested reader is therefore re- quested to take a pencil and to carry out a few divisions in accordance with the Egyptian method. The Egyptian was not able to write any other fraction. "Multiply 80 (or literally: add beginning with 80) until you get 1120". he took recourse to fractions. Ancient Egypt had special words for Yz. In the first place. % and %. as we know them. the set of fractions was extended to include the unit fractions (fractions which have 1 as a numerator). The complement. 69 in the Rhind papyrus. but formulated in- versely. which occur in daily life and which are designated by specific names. "the third part" Ya "the three parts" % . which were the natural fractions. For a fuller understanding of what follows. and the solution looks exacdy like a multi- plication: 1 80 /10 800 2 160 /4 320 sum 1120 We would write the result in the form 1120 : 80 = 14. %. and the expression "percent". But what did the Egyptian do when the division did not go "evenly"? He did just what we do. ¥S. Worthy of notice is the verbal expression for % which means literally "the two parts". 1120 is the result. except those which have been men- tioned. Natural fractions and unit fractions Fractions with numerator and denominator. there is a limited number of natural fractIOns. But for the Egyptians. Ye and Ye may be counted among the natural fractions. THE EGYPTIANS 19 Diflision is considered by the Egyptians as a kind of multiplication. It presents quite naturally a concrete image: three parts and then a fourth part . % and %.

: % = % + ¥S.. Ii for Ijnand 3 for %. 1916. are constantly applied in the Rhind papyrus. Also A.. which combines with the four other parts to complete the unit.'tch. fifth. <lMa. Consequently they did not obtain a convenient notation for fractions which are not unit-fractions. B I.> The ancient symbol for % was replaced later on by the partition =x = Y2+~. 5 •• dtr Am"" O!!. Scbrift.. • Se. Philologically it does not make sense to speak of two fifths.g. because it had a name of its own. 3+3=3- Further relations. and not reminiscent of our numera- tors and denominators. K. it is necessary to introduce a notation for fractions based on the Egyptian notation. Tht Rhiod Mallttmalical Papp. Aril'-'il: uruI Recht. and a hieroglyphic of its own c. The fraction % is the only one which did not need to be reduced. Thus we shall write. viz. because there is only one fifth part. B. fourth. as Neugebauer does. uruI "'" fUr'odm V~lm ••11 S!"".. Simple examples are the following cruce.. deduced from the last.. "the two parts". In this representation the fifth part is the last part..l.g. The Egyptians did not succumb to the temptation of this philological contra- diction.bauer. "..20 CHAPTER I combine to make the whole. . 3. supplied with a special symbol r = part: and they reduce other fractions to unit fractions.n der wissenscbaftlicben G... etc. Ohio 1929.cilaft. Sttasaburg. N. 1" 301. e. 1 Analogously we can explain our use of the words third. Seth•• v •• ZtJd•• . They represent the unit-fractions by writing the denominator... Calculation with natural fractions is very simple for anyone who has mastered a few simple formulas. Am"'. hoi .n. 6 + 6 + 6 = 2. to be found in the "London leather-scroll" 3: (Plate 3) 6 + 6 = 3.u lmottt isl. I O.. which follow immediately from the meaning of the fractions. e.Uen und Studien.. . In order to follow Neugebauer 2) in his interpretation of Egyptian calculation with fractions from their own point of view.II ZAhI""'ltr.

This art is called ··ideoplastic". (Photo Lehnert 6. In relief art and in painting. for this "ideoplastic" conception is joined with a sharp observation of nature (see next plate). Painting of the tomb of Djeserkere. The mathematical knowledge of the Egyptians. Thebes. Cairo) . Upper part: surveying. I. goes along with a form of reproduction in art. 38). concerned with practical matters. Egypt. because the Egyptian does not picture things as he sees them ("" physioplastic") but as he knows them to be.the largest number of exact data concerning reality. Yet not entirely so. because this enables him to give. the artist spreads everything as much as possible in the plane. Up to a certain point.sonb (No. 18th dynasty (1567-1310 B. this art must be "read" .Landrock. PLATE 1 PL.C. or rather as perfect a picture of things as possible. The New Kingdom.). attempting to give which aims at giving as complete.

A writer. (Phol<> Ali"". it seeks to perpetuate it.C.) . Louvre. This art is not abstract in the sense that it considers the external appearance oflife as unimportant. ebony and rock-cystal. To approximate nature as closely as possible. fifth dynasty. This is shown not only by the plasticity of the body. On the contrary.. about 2500 B. The statue is full of spirit and "magic" life. 2. the eyes have been formed from various materials: white quartz. Limestone statue from Egypt. PLATE 2 PL. but also and especially by the almost frightening "realism" of the look. Paris. The Old Kingdom.

. p. containing simple relations between fractions. Archeo!. ~ pyrus near the Rammesseum at Thebes. Nevertheless. Glanville. Egypt.> contents were disappointing: every line gave a simple relation between fractions.PL. the British Museum. the tti (.tst. such as jj + IS = il.hUHssStSchicht. 3. R. tIer ii8YPtiscne.ung.C. 232. 13 (1927) p. At first. 359. d. The scroll is said to have been found together with the Rhind pa. See S. the leather scroll has proved valuable inasmuch as it supplies a key for the understanding of the first stages in calculation with fractions. The date is about 1700 B. Q!Jellen u. and B. L.z leather roll i. The unrqlling of the hardened scroll was a clever piece of work of modern chemistry. Die E. The IftIltltttMti". Journ. Studien Gesch. Math. . B 4.. etc. 5 + l!ii = 4.Waorden. Leather scroll in the British Museum (BM 10250). K. v. Bruchrech.

papyrus In London. .C. 4th dynasty (about 2600 B. Old seum (about 1480 B. Paris. early 18th dynasty). 411.. writing is a simplification of hieroglyphic writing.C.). Louvre. Hieratic writing. Hieratic Kingdom. it already came into use during the Old Kingdom. British Mu- from the Mastaba of Meri. PLATE 4 PL. Egyptian stele with hieroglyphic writing PL. "lb.

21 + 42 = 14. there results the very important formula ~ 3=~+~ Again adding 6 to both sides. 15 + 30 = 10. 4. we do not first find 8 + 3 + 2. to double 4 + ~ + 4 in Rhind No. when in the first step of the division of 2 by 41. 96 + 192 = 64.: 9 + 18 = 6. 24 + 48 = 16. If 6 is added to both members of (1). which follows from 6 + ~ + ~ by means of (4). 61 we find 3 of IT is 22 + 66. similarly (1) gives 6+ 12 = 4. Formufa (3) is applied especially when it is required to determine two thirds of a unit fraction. No. By halving. . For example. in Rhind. thirds and sixths which every Egyptian computer had to know by heart. by dividing by 2. 3. we obtain two equivalent representations of 6/8: ~ ~+~=3+~ Finally. etc. the Rhind papyrus takes them for granted. 12 + 24 = 8. (2) ~+~+6=1. 18 + 36 = 12. formula (3) of the leather scroll leads to ~= 4+ 12. this is at once written in the form 6 + 3 + 6. THE EGYPTIANS 21 (1) ~ + 6 =~. by use of (5). That the Egyptians used this method systematically becomes clear from the following sequence found in the leather scroll and obtained from (1) upon division by 3. addition of 6 to both sides of (2) gives ~ 3+~=1+~ These are the rules for the calculation with halves. Again. 31. Further relations between fractions are obtained from (1)-(5). we have to determine one half of13 + 3. and writing the result from right to left. 48 + 96 = 32. but immediately 9 + 6. For example. 45 + 90 = 30. 30 + 60 = 20. etc.

And we have already seen how to double a fraction whose denominator is divisible by 3. Obviously. we obtain 9+ 9= 6 + 18. Our reply would be that we have to accept the fact that the Egyptians were not like our modern mathematicians. multiplication depends upon duplication. whose order of magnitude is readily kept in mind and which are dealt with easily. He obtains expressions which are easily managed and which contain but few unit-fractions with larger denominators. for him. indeed. on the model of 9 + 9 = 6 + 18. 4 3+ 9. At this point the reader may well ask: why not leave well enough alone and be satisfied with 9 + 9? The answer to this question is found by considering what would happen in the further development of the calculation.. 6. For. etc. one would get '9 + '9 + 9 + 9. and then divide by 3. An example IS found in the fact that for thousands of years Egyptian documents continue . namely.22 CHAPTBR I Moreover. 15 + 15 = 10 + 30. etc. This Duplication of unit fractions is of great importance for the Egyptian computer. strictly observing ancient traditions and adhering to current notations. the same method makes it possible to double a unit-fraction whose denominator is divisible by 3. as shown above. 5. there is little of value in such a senseless accumulation of unit fractions. who immediately introduce an abbreviation for every expression that occurs frequently. Doubling a fraction with even denominator is very easy. if we write (3) in the form :3 + 3 = 2 + 6. 21 + 21 = 14 + 42. by using something like our numerators and denominators. like 3. 3. 2 6 + 18. a third duplication would lead to 9 + 9 + 9 + 9 + 9 + 9 + '9 + 9. The Egyptian proceeds however as follows: 1 9. it is done by simply halving the denominator. To this one might make the rejoinder that the Egyptians might have thought of a shorter way to write 9 + 9 and 9 + 9 + 9 + 9. In doubling 9 + 9. They were extremely conservative. in addition to natural fractions. 8 :3 + 6 + 18.

3.rrus. 6 . The (2 : n)-table.v. The sequence of fractions 3. 3 + 30 each. we obtain exactly the 3 we need. For th~ Egyptians this was not the case. 7.. 3. Now we take one half of 8. . By adding together the fourth and the eighth. 21) 4 : 15 1 8 /1 3 1 15 / ~ 164 2 6 10 1 +2 2 / 4 12 / "5 3 / 4 2 3 2 / 15 / 8 1 / 3 1 ~otient: 2 + +8 4 ~otient: 5 +3 ~otient: "5 + 15.. 3 + "5 + 30 each. like 5 or '7. but to obtain it from 3. the (2 : n)-table answers the quest~on as to the duplication of fractions. In No. we take first 1/10' this is doubled and then increased by 1/u. although Upper. 25 where we take first the double and the quadruple. or 9 loaves among 10 men. 6.. which was raised above. 8. since the problem of duplicating 5 is totally different from the division problem which amounts to calculating with 5 . had evidently become entrenched in an established tradi- tion. and again. 24) 19 : 8 (No. and again. and then % and Ya of 3. we lack 3. 25) 16: 3 (No. Doubling 8 gives 16. In No. n. one is asked "to operate with 8 so as to obtain 19". THE EGYPTIANS 23 to speak of "the two kingdoms".. The ~OIU~l?n is based on a principle. d. The following three are divisions from the Rhind papyrus. . It is a peculiarity of the Egyp- tian calculators. It is customary to use. This is seen in No. we have to consider Division once more... What was done when the denominator was not divisible by 2 or by 3? To understand this. 3 + 10 + 30 each. 8. perhaps in order to reach smaller numbers more rapidly. not to write down 3 first. the other sequence 2 by n. employed here.and Lower-Egypt had long ago been united into one empire. Thus the required sum 3 + 1 = 4 is obtained in an elegant manner. which is as simple as it is fertile: To calculate 2 . The EglPtian answers are as follows: 2 + 10 each. Perhaps it appears to us as a matter of course that 1/5 and 2 : 5 are the sam! thing. 21. instead of the sequence of fractions 2. 24. which do leave a remainder: (No. 4. At the beginning of the Rhind pae. Perhaps the reader will find it of interest to use this method to answer questions 3 to 6 of the Rhind papyrus! The problems are: to divide 6.

28 4 /4 28. while the 3- sequence gives 2: 11 = 6 + 66. The oldest section contains the denominators which are divisible by 3. indicates where 28 comes from. It turns out that only in these two cases the 2-sequence produces a simpler result than the j-sequence. 4. the duplication appears to have been obtained by actually carrying out the division of 2 by n. the further details being shown in an auxiliary column. without exception.. . . its separate parts were computed by different methods. The text exhibits the divisions more or less explicitly. a fifteenth is 3. viz. lead to the result 2: 11 = 8 + 22 + 88. as in the following examples (2 : 5 and 2 : 7). for the (2 : n)-table of the Rhind papyrus. nor when this happened. . For instance. 12. which. . they all proceed according to the same rule: 2: 9 = 6 + 18 2: 15 = 10 + 30 2 : 21 = 14 + 42. 8. 2iI is 4. In the other cases (certainly from n = 11 on). What part is 2 of 51 3" is 1 + 1 15 is 3.. In our "translation" the red symbols have been printed in bold-face type.. is obviously to be preferred. Computation: 1 5 j 3+3 / 3 1+j / rs 3 What part is 2 of 71 4 is 1 + i + 4. In this manner the work proceeds. but the division by 7 and 13 employs only the 2-sequence (2.). The calculations which have been reproduced here certainly tell their own story. in calculating 2 : 11. In these cases the division 2 : 3k is simply a confirmation of a known result. 6.23. i. It certainly occurred long before the era of our texts. The result of the division is therefore 3 + 15. which includes all the odd numbers from n = 3 to n = 101. It was therefore indeed a new idea to duplicate ii by dividing 2 by n.e. the use of the 2-sequence would. was not constructed all at one time. these add up to 2. from 4 x 7. In dividing 2 by 5. having fewer terms and smaller de- nominators. 17. the number 4. 11. the terms 3 and 15 are clearly visible because they are written in red.24 CHAPTER I until you obtain 2. In the case 2 : 7. a third of 5 is 1 + 3.29 and a few of the larger integers. We do not know in whose brain this thought arose for the first time.e. the lsequence is used. Computation: 1 7 ~ 3+~ 1 7 / 4" 1 + ~ + 4" 2 14 /4. placed in front of 28. the sequence of fractions 3..9. i. 3.

35 proceeded quite differently. the form of presentation changes. 292 + 365 2:29=24+ 58+174+2. the method of calculation changes. which does not include divisors that are divisible by 3. another idea is introduced. all of which follow the rule 2 : 3k = 2k+ 6k. and halving of the result shows that 1/20 of 31 is 1 + 2 + 20. since division of 31 by 10. THE EGYPTIANS 25 The results of the divisions 2 : n are summarized in the following table. as is seen from the following examples: What part is 2 of 31? 20 is 1 + 2 + 20. While up to this point. what is more important. for the leather scroll has the relation . How did the calculator hit upon the idea that this requires 4 + 5? It checks. all divisions were carried out by means of the 2-sequence and the 3-sequence. 124 is 4. the calculations are given in abbreviated form.32 2: 77= 44+308 2: 79 = 60 + 237 + 316 + 790 2 : 31 = 20 + 124 + 155 2: 83 = 60 + 332 + 415 + 498 2 :35= 30+ 42 2: 85 = 51 + 255 2 : 37 = 24 + ill + 296 2: 89= 60+ 356+ 534+ 890 2 : 41 = 24 + 246 + 328 2: 91 = 70 + 130 2 : 43 = 42 + 86 + 129 + 301 2: 95 = 60 + 380 + 570 2 : 47 = 30 + 141 + 470 2: 97 = 56 + 679 + 776 2 : 49 = 28 + 196 2 : 51 = 34 + 102 2 : 101 = 101 + 202 + 303 + 606 Beginning with 2 : 31. Computation: 1 31 20 1+2+20 /4 124 4 /5 155 5 What part is 2 of 35? 30 is 1 + 6. But. 42 is :3 + 6 6 7 5 Computation: The start of the computation of 2 : 31 is easy to account for. the divisions 2 : 31 and 2 : . 2: 5 = 3 + 15 2: 53 = 30 + 318 + 795 2: 7= 4+ 28 2: 55= 30+ 330 2:11= 6+ 66 2: 59 = 36 + 236 + 531 2 : 13 = 8+ 52 + 104 2: 61 = 40 + 244 + 488 + 610 2: 17 = 12 + 51 + 68 2: 65 = 39 + 195 2 : 19 = 12 + 76 + 114 2: 67= 40+335+536 2 : 23 = 12 + 276 2: 71 = 40 + 568 + 710 2: 25 = 15 + 75 2: 73 = 60 + 219 -I. 155 is 5. This fraction is to be increased so as to produce 2.

In the calculatlOn of 2 : 35. In modem notation. The red auxiliaries. But these modem ideas should not simply be read into the old texts. viz. the solution 4 + 5 is not at all obvious. we proceed to consider these more fully. Hence 5 + 15 is the desired complement. reproduced above. Something further about these auxiliary numbers and about the role they play. these numbers 6. auxiliary numbers occur whose significance must not be underestimated. these considerations lead to 2 12 7 5 1 1 35 = 210 = 210 + 210 = 30 + 42 which corresponds to the Egyptian division 2:35=30+42. Calculate with 15 till you find 4 (i. divide 4 by 15) 1 15 10 1+2 / 5 3 / 15 1 sum 4. . 21 we find: How is 3 + 15 complemented to I? 10 1 sum 11. It is not surprising therefore that such complementations receive special attention in the Rhind papyrus. This is followed by a cheek. It appears therefore that these auxiliary numbers are to be looked upon as the numerators of fractions which have been reduced to a common denominator. In our modem arithmetic. 7 and 5 would appear as numerators when the three unit-fractions are reduced to the common denominator 210. and 1 + 2 + 4 + =I lS of course equal to 2. the sum 2 + 20 has to be complemented to 1. and therefore proportional to 35. Complementation of a fraction to 1 is a problem which occurs over and over again in the E4Nptian divisions. In the division of 2 by 31. In No. is learned from the complementary calculations (~ekem-calculations) of the Rhind papyrus.26 CHAPTER I 20 + 5 = 4. 30 and 42 under which they are placed. given above. 30 and 42. They are inversely proportional to the numbers 35. the red (bold-faced) 6 and the numbers 7 and 5 which follow it.e. How does one obtain t~e fractions needed to increase 1 + 2 + 20 to 2? This requires a new procedure ViZ. remainder 4.

together 15. does not occur. can be carried out. together 5". As a rule. the smallest of the fractions is taken as a new unit and the other fractions are then expressed in terms of it. the fourth part is 3. It appears however from No. A "numerator". It is more accurate to describe the state of affairs as follows: Jr/hen a somewhat complicated sum of fractions has to be compared with another such sum. When the given fractions are reduced to the "denominator" 45. the denominator of the smallest fraction. of this you must take the fourth part. No. in our case 45. even simpler and very readily applicable: The transformation of the siflen quantities to the auxiliaries is accomplished by multiplyins by the larsest of the Jenom. By means of it. 23 that our interpretation of the auxiliaries as "numerators" is not entirely valid. "hau". the sum of the numerators is + (11 + 4) + (5 + 2 + 8) + (4 + 2) + (1 + 2) 1 = 23 + 2 + 4 + 8. but it is readily supplied. or has to be complemented to 1 . Sum 1. 3 = 12. we lack 6 + 8. The Egyptian word 'h'. division of 6 + 8 by 45 gives the quotient 9 + 40. and the imerse process is carried out by Jiflision by this number. finally a multiplication. 26: "A quantity and a fourth part of it give together IS". 4 . An understanding of this simple rule gives complete control of the calculations by means of auxiliaries. that "counts" the number of forty-fifths. incorrectly. It is clear that the method followed here is that of the "false assumption" : . These aha-calculations are quite like our linear equations in one unknown. is taken as the new denominator. and at the present time somewhat less incorrectly as "aha"./15 to obtain the total of 16/16 • The calculation of 4. Or. The required "quantity" is therefore 12. Then the division 15 : 5 = 3 is carried out.nators (in our case 45 j. for fractions appear there as auxiliaries: 4 + 8 + 10 + 30 + 45 has to be complemented to 1 11+4 5+2+8 4+2 1+2 1 The required complement is 9 + 40. no matter how complicated. a col- lection. when they are reduced to the denominator 15. Toob- tain 30. "Aha-calculations" . in the sense of an integer. THE EGYPTIANS 27 The red auxiliaries 10 and 1 are the numerators of the fractions :3 and IS. This calculation with auxiliaries completes and puts the crown on the Egyptian computation technique. means a quantity. The Egyptian solution begins as follows: "Calculate with 4. Since their sum is II. that used to be pronounced./16 then follows. namely 1. A simple example is found in Rhind. every division. there is lacking 4. Check: 4 + 8 + 9+ 10 + 30 + 40 +45+ :3 11+4 5+2+8 5 4 + 2 1 + 2 1 + 8 1 15. The computation is not given.

B 1. take a fourth part. But I must not fail to mention that Neugebauer considers tbe "false assumption" as a legend.J Shu/ito. Neullebauer. equal 1. Divide 1 by 3 + 2 + 18: 3+ 2+ 18 2 1+ 2+ 4+ 36 / 4 2+ 4+ 8+ 72 8 4+ 8+ 16 + 144 16 8+ 16 + 32 + 288 J32 16 + 32 + 64 + 576 Now the auxiliaries enter the field: the sum 2 + 4 + 8 + 72 + 16 + 32 + 64 + 576 must be complemented to 1. then it is shown by means of auxiliaries that their sum is indeed 1. 1 Frequently the quantity is at first taken to be 1. I go three times in a bushel. since tbe division 15: (1 + ") can very easily be carried out witbout using them. . and a ninth of 4 + 32 are determined. a third. On the other hand. viz. in our case 4. and my ninth part are added to me and I come out entirely (i. as a calculation with aUXIliaries: 1 + 4" 4+1=5 (0.e. the bushel is entirely filled). Hence the sum of all the fractions is already 1 . 9 + 9 = 6 + 18). e. hence the quantity has to be multiplied by 15 : 5 = 3. three times 4 + 32. so that the result of the division is 4 + 32. together 5") differently. Butthe required result is 15.28 CHAPTER I one starts with an arbitrarily chosen number as the required quantity. The auxiliaries show that the last five fractions add up to 8 exactly: 72 16 32 64 576 sum a- 8 36 18 9 1 72.g. my third part and a third of my third part. a third of a third. 333). and that he interprets tbe beginning of the computation (""calculate witb 4. Four and a fourth part offour give 5. because this makes the computation of the fourth part easy. To check this result. Who says this? Answer: 1 1 2 2 :3 :3 :3 of his :3 9 his 9 9 sum 3 + 2 + 18 (because. 37. in Rhind. there seems to be little justilication for tbe auxiharies.. No. The group of aha-calculations also includes the first problem of the Berlin 1 The above explanation seems to me to be the simplest one. Qwiltll .

have together an area of 100. Take the square root of the given number 100. such as.Schackenburg.. The reciprocal quotient is the grain-density or the "strength" of the loaf of bread or the jug of beer. that is to saL:2 + 4. the "pesu-calculations" which are concerned with determining the amounts of grain needed for making beer or bread. Multiply :2 + 4 by itself. Other problems are concerned with the reduction of bushels to other units of 1 Schack. whose solution requires even the extraction of square roots. but we can easily make a guess at the rest: 8· 1 = 8 and 8· (2 + 4) = 6 are the sides of the required squares. whose side is % (in the text. Thus we have the following equation quantity of grain x pesu = number of loaves uugs of beer) pesu = number / amount of grain. 38 (1900). they bear witness to the purely theoretiql interests of the Egyptian computers. e.g. The aha-calculations constitute the climax of Egyptian arithmetic. 138 and 40 (1900). It is perfectly clear that the solution starts with a false assumption: "Take a square of side 1.herable. and take % of 1. that of the other as:2 + 4. THE EGYPTIANS 29 papyrus 6619 1 . 65. with their primitive and laborious computing technique. it is 10. They have ob- viously been set up by people who enjoyed pure calculations and who wanted to drill their pupils on really hard problems. Much space is taken up in the Rhind papyrus by Applied calculations. Ag. etc. From here on the text becomes undecie. if it were not for complications arising from the non-equivalence of different grains. arithmetic strives for its highest development. Show me how to calculate this".. as the side of the other area. p. p. The technical term "pesu" = "baking value" designates the number of loaves of bread or the number of jugs of beer that can be made from a bushel Of grain. Hence. then the addition of the areas gives 1 + 2 + 16. The Egyp- tians could not possible get beyond linear equations and pure quadratics with one unknown. Take the square root of this. The text is as follows: "A square and a second square. . How many times is 1 + 4 contained in 10? 8 times". :2 + 4) of that of the first square. this gives :2 + 16. Z. it is 1 + 4. if the side of one of the areas is taken to be 1. These simple relations would suffice to solve all pesu-problems. Like every art. The aha-calculations are not based on practical problems. The problems can then be complicated still further by trading loaves of one "baking value" against those of another value. this re- qUires the determination of reduction-coefficients which the computer has to know.

the Egyptians. a symbol for a million existed. Such an extension became a necessity at a time when an empire had to be organised. certainly disposed only of a re- stricted number of integers. Their use made it possible to carry out every . " Egyptian multiplication is markedly a written operation. rules for the duplication of the unit fractions 3n were derived. the history of the technique of computing begins with the extension of this primitive arsenal of numbers in both directions. the distribution of wages. Next came the recognition of simple relations among the natural fractions. for obviously to calculate the amount of grain necessary for an army. it extends over the entire period from the invention of the art of writing to the Middle Kingdom. %. It is possible to indicate the successive phases of this development. It must have proceeded as follows: From multiplication we come naturally to division which is nothing but inverse multiplication. this led to the first third of the (2 : n).30 CHAPTER J measure. Multiplication is for him a kind of addition. He writes fractions as sums of unit-fractions. beginning with .. or to determine the quantity of money and of material required for the construc- tion of the pyramids. A decisive step was the discovery that every unit fraction ncan be duplicated by division of 2 by n. The development of arithmetic therefore took place in a very slow tempo. and of a similarly restricted number of "natural fractions": Yz. 5. the need for a method of checking was recognized. a stretch of many centuries. Indeed. the technical term for multiplying is "add. Ys. and taxes collected. a procedure to compare different sums of unit-fractions. etc. multiplications were required. but not to give their dates. which were now accessible. like other nations. by division by 3. Neither can it be much younger than the art of writing.000. with the calculations of quantities of food. Initially. This problem was solved by the intro- duction of auxiliary numbers. Additional rules resulted from halving.table. Let us now try to survey The det!elopment of the computing technique and to assign it a place in the history of civilization. During the Old Kingdom.. and so forth. in particular it became essential to be able to complement such sums to 1. From 3 = 2 + 6. we already find the symbol for 100. fractions and operations on fractions are needed. became more and more com· plicated. an army administered. on a monument of the first dynasty. But to carry out a division. The dominant element in the thought of the Egyptian calculator is addi- tion. Ye. Ye. It can not antedate the notation for numbers therefore. %. thus the division problem led to a further development in the cal- culation with fractions. but during the New Kingdom it disappeared. As the calculations. This primitive stage belongs however to pre-history. %. sufficient for daily life.

SO~. only approximate formulas are used. Gillam. v. the calculator has to know the rules on which the calculations depend. Quell. because we can account for all the essential points without it. Trans. ') O. While the other arithmetical problems are concerned with the calculation of wages. This quotient is now multiplied by 7. Waerden. O. it would be without value for the explanation of arithmetic. or inclination. pp. when astrology flourished. Brussels. 250 the height.ftique au Moyen Empire. B. It is possible to understand this entire development without taking recourse to the Hypothesis of an advanced science. Vol"n.n. Amer. 1927. no matter in how sophisticated a manner it was formulated. but merely applied arithmetic. 180 : 250 = 2 + 5 + 50 of a cubit. it becomes very clear that. Calculation: half of 360 = 180. Under the influence of Greek writers. We know absolutely nothing of such an advanced science. 360 is the side of the base. B 4. But a systematic derivation of these rules occurs nowhere. the geometric problems ask for the determination of areas and volumes. the levels of Egyptian geometry and astronomy are frequently placed too high. ))6 and 782. because a cubit is 7 hand's breadths. Egyptian astronomy lagged far behind Greek and Baby- lonian astronomy. Neugebauer and A. Rhind No. p. because frequently. even in the Roman period. it can not occur any- where. 2) What about geometry however? The geometry of the Egyptians. The analogy between the geometrical problems and the beer· and bread- exercises comes out clearly especially in the calculation of The inclination of oblique planes. d. )8). Royal Acad. tell me the inclination (skd). Now that we know a little more about Egyptian astronomy. PhiiOl. It becomes clear from the calculation that the saykad. 32 (1942). 36: "example for the calculation of a pyramid." The word skd might be pronounced as "saykad". 209. . to solve every aha-problem. Neugebauer.n und Studi. Weare going to show that Egyptian geometry is not a science in the Greek sense of the word. is the num· ') O. as does Gillain 1) for example. In all these cases. the quality of bread or of beer. as for instance in the case of the area of a circle. L. etc. Amsterdam 50 (1947). L'arithm. of which our texts would then merely represent the "first phase".. THE EGYPTIANS 31 division. Result: the inclination is 5 + 25 hands. Proc. p.

The base of the triangle is halved. To determine the area of a circle.thema. Fig. This corresponds to a very good approximation . who had reached a much hIgher stage of mathe- matical development.. In all these problems the difficulty does not lie in the geometry. = 3. the Egyptians square 8/9 of the diameter. . or of beer.ll. This is also the value given by Yitru- vius. 1 The quadrangles are treated as follows: half the sum of two opposite sides was multiplied by half the sum of the other two sides. Areas of triangles. Bulin. preuu. philo•. "in order tu make the triangle square" and is then multiplied by the height.II.tia Papyrus II. but in the calculation. Akad. The Jews even held thiS as a 1 upslus. 69. Rhind 51: Area ofa tri&nfle From A. In a similar way. This formula is obviously incorrect. hist. Abttilung. The Babylonians. Wists. 1. Moreover there are reduc- tions to other area-units. Abh.32 CHAPTER I ber of hand's breadths by which the inclined plane departs from the vertical for a rise of one cubit. 1855. In the deed of gift of the temple of Horus in Edfu appear the areas of a large number of triangles and quadrangles.1605 . rectangles and trapezoids are determined by use ofthe correct formulas. p. The Rhind M. and it is found again in the Chinese literature.. This means that the saykad of an inclined plane is the exact analogue of the pesu of a loaf of bread.. B. 4'(8/9)1 = 3.. always used. the sum of the parallel sides of a trapezoid is halved and then multiplied by the height. It is a great accomplishment of the Egyptians to have obtained such a good approximation. it gives the correct result only if the quadrangle is approximately a rectangle. --. Chace.

\57.tischer Papyr. to something quite ordinary. 1 S.(1-1/9)2. • W. are told a basket (of 4Y2) in diameter by 4Y2 in depth. Es entsteht '3 + 6 + 18. becomes Q =2x. the writer of the Mishnat ha-Middot 1 had the courage. 154..di .. p.1/9)(1 .tology.. and the formula which is used. 0 lass du mich wissen seine (Ober)flache! Berechne du 11. QudlCII u. 48. 17. TIlt Mis1w. about 150 A. 1/9)2x. A \ (\930). this would Indeed give the correct formula for the area of a sphere. Quell. . In Struve's translation. viz. The genius of the Egyptians would have been wonderful and indeed incom- prehensible. THE EGYPTIANS 33 sacred number.. Es entsteht 32. to give the value 3 1 / 7 and to offer a different explanation of the Bible-text (V4). es ist seine (Ober)flache. he inserts the words "of 4Y2" and translates as follows: "whe~. problem 10 of this papyrus 2 is as follows: Form der Berechnung eines Korbes. 23). Du hast richtig gefunden. both of which have the value 4Y2. Jo. which would have antedated Archimedes by more than a thousand years. p. which Struve had not made quite clear.1/9)2 is the Egyptian value for n/4. p. Siehe. 2Q = nx 2 . Peet.. W. wenn man dir nennt einen Korb mit einer Miindung zu 4Y2 in Erhaltung. von 9. then tell me the area.r. • T. 4Y2 mal. 3 the basket can also be taken to be a half-cylinder. Struv<. Mathe .d St. A 2. if they had succeeded in obtaining the correct formula for the Area of the hemisphere. According to Peet. Berechne du den Rest von dieser 8 nach diesen '3 + 6 + 18. this reduced this astounding accom- plishment. Es entsteht 1. von 8..d Studi. Instead of one number x = 4Y2. Gandz. But disappointment followed close upon amazement.Middot. as was thought to have been the case for some years.. x and since (1 .al of Egyptian Arclw.D. y..s des Muse. Rechne du mit 7 + 9. authorized by the Bible (1 Kings VII. Berechne du II... who edited and published the Moscow papyrus. Moskau. wei! ja der Korb die Halfte eines Eies ist.. Expressed in modern !Ovmbols. Berechne du den Rest als 8.t /w.. the diameter being taken as x = 4Y2. there are now given two numbers x and y. the cal- culation proceeds accordmg to the formula Q = (1 . It is true that Rabbi Nehemiah. E. This creates a totally different state of affairs. Es entsteht 7 + 9. on the authority of Struve. Peet begins by straightening out the gram- matical structure. but the Talmud restored the canonical value 3.

we obtain the correct formula for the lateral area {J . The further development of this idea can be found in Neugebauer'S beautiful book. indeed most of the problems are concerned with grain barns. p. but not with general quantities. It is not to be supposed that such a formula can be found empirically.b)S • 11/3.ik. of beams and of cylinders were of course determined by multiplying the area of the base by the height. a rectangular parallelopiped.. how? By dividing the frustrum into 4 parts. Neugebauer. 1 Fig. 5). viz. one finds.... He takes the ~'basket:' to be one of those dome-like barns. But can one justify the assumption that the Egyptians were able to make such an algebraic transformation? They were able to calculate with concrete numbers. Volumes of cubes. by means of the formula (1) V = (as + ab + b2) • h13.". Neugebauer suspected that (1) came from (2) by means of an algebraic trans- formation. Neugebauer has given a somewhat different interpretation. M. which we know from Egyptian lllustratlons and looks upon the calculation as an approximation. Neugebauer. as found in the Moscow papyrus (Plate 5a).. Fig. 136. two prisms and a pyramid (see Fig. 2. This leads us to wonder whether in I O. Struve.b)h + {a . Fig. where h is the height and a and b the sides of the lower and upper base.. An outstanding accomplishment of the Egyptian mathematics is found however in the entirely correct calculation of the volume of the frustrum of a pyramid with square base. The principal difficulty in those things arises from the relations between one another of different units of measure for volumes and for quantities of grain. y. 3.. 4. (nxI2)y = 2x . CHAPTER I If we interpret now y as the height of a haif cylinder and x as the diameter of its circular base. the volume of a pyramid being assumed as known. n/4 .Stitchi"". V. . the formula (2) V = blh + b(a . Peet.. It must have been obtained on the basis of a theoretical argument.

. The solid that is obtained in this way. the middle one has a base ab and the upper one a base b2 • ~~ ~~ This derivation of the formula does not transcend the level of Egyptian mathe- matics.. Cassina. but having 1/3 of its original height (Fig.. In the papyri one does not find a trace of "the construction of lines with proofs". however much one may appre- ciate particular d.. one cannot escape a feeling of disappointment at the general mathematical level. Whichever one of these hypotheses is adopted. and I definitely do not assert that the Egyptians actually proceeded in this manner. .. 22 (1942). The two prisms of Fig.. For example. Moreover. Per. Then the upper third of the first of these blocks is removed and placed on top of the second one (Fig. we must suppose that the Egyptians knew how to determine the volume of a pyramid. 7). pp. each of which has the height h/3.od.. Cassina 1 has suggested another derivation of 'the formula for the special case (and this is indeed the only case dealt with in the papyrus) in which the area of the upper base is one half of that of the lower base.. Ma. we should not a priori eliminate a possible effect of Babylonian algebra. I certainly do not want to tell a fairy tale. 5 are changed to rectangular blocks of half the height. in which Democritus is said to surpass even the Egyptian rope stretchers. 6). There are indeed other possibilities. we return to the question: What could the Greeks learn from the Egyptians? Looking at Egyptian mathematics as a whole. Having reached the end of our study of Egyptian mathematics. the lower one of these layers has a base equal to a2 . there are only rules for calculation without any motivation. can be divided into 3 horizontal layers. 1-19. let us assume that one of the edges is perpendicular to the base. the pyramid is also transformed into such a (4a serial. THE EGYPTIANS 35 this case Egyptian arithmetic was influenced by Babylonian algebra. But. Or should we suppose that (1) was obtained from (2) by a geometric argument? One might imagine the following deduction: For convenience. 1 M.

which they then developed further. as Struve does. characteristic of Egyptian mathematics are the elaborate calcula- tions with fractions. 7. 1 Set K. Sittung. whether one considers elementary texts or more advanced ones. could be abstracted by a competent historian from a handbook for engineers. In the second place. the Egyptians knew a great deal more than is shown in the texts which we have and that the Greeks were familiar with such additional material. We would not be able to demonstrate the validity of such hypotheses. But calcu- lation is not the same as mathematics. just as well as from a.. which can not serve as a basis for higher algebra. Since we know then the actual historical foundations on which the fine edifice of Greek mathematics was erected. and the treatment of geometry as applied arithmetic. how th does one prove this? • One might suppose. 357-472. a basis for Greek mathematics. . The Greeks may also have taken from th the Egyptians the rules fo.buichte der bayerischen Akademit. which are characteristic of Western mathematics. (ni«hi.36 CHAPTER I It is certain that from the Egyptians. and indeed did supply. But. and they would in no way serve our purpose. the general character of the mathematics which a people has at its command. that Fig. lAsiJlik. remains the same. volume of the Mathematische Annalen. 1936. Vogtl. But for the Greeks such rules did not constitute mathe- matics. the Greeks learned their multiplication and their computations with unit-fractions. Literal calculation. the determina- tion of areas and volumes. differential calculus and coordinate geometry. they merely led them to ask. we do not need to set up hypotheses concerning a lost Egyptian higher mathematics. the Akhmen papyrus of the Hellenistic period 1 gives evidence of this. Against this assump- tion there are two arguments: in the first place.a. we shall show that Babylonian mathematics could very well supply. pp. decimal fractions. Munchen.

They knew how to solve systems of linear and quadratic equations in two or more unknowns! This would have been impossible for the Egyptians. This shows itself most clearly in algebra. 48. not.e. 38. which had also invented the cuneiform script. every subtraction of fractions.C. Without mastery of these fundamental operations. because their complicated fraction-technique made a difficult problem out of every division. which enabled them to calculate with fractions as easily as with integers. mathematics can not get beyond a certain low level. usually mono- syllabic. 38/601-24/60 2 • and looked for the square root in a table: VO. i. 24. CHAPTER II NUMBER SYSTEMS. of course. remain unchanged. If the number was not a perfect square.. combining. dominated southern Mesopotamia during the third millenium B. How would t~ be able to determine the square root of a sum of unit-fractions like 2 + 10 + 25? The Babylonians however wrote this sum as a sexagesimal fraction: 0. which flourished around 3000 B. an excellent sexagesimal notation for whole numbers and fractions. or perhaps a few centuries later. The Babylonians had. in the sense that a good number system leads automatically to a high development in mathe- matics. This made possible their highly developed algebra.C. the Sumerians. The Sumerian language is neither Indo-Germanic nor Semitic. and of the related arithmetical techniques. who . The Sumerian civilization was taken over by the semitic Akkadians. This remarkable cultural group. The stems of words. 24 = 0. but rather that a good notation and a convenient manipulation of the four fundamental operations are necessary conditions for the development of mathematics. The sexasesirnal system was taken over by the semitic Babylonians from their predecessors.The oldest Sumerian cuneiform texts date from the first dynasty of Ur. 38. but other invariant syllables are placed in front or behind to determine the grammatic function of the word. DIGITS AND THE ART OF COMPUTING In this chapter we shall give a brief survey of the number systems and the number notations in the principal cultural periods. not mflectlve but agglutinate i. We shall see t~at these notations and these techniques are of very great importance for the development of mathematics.e. as we shall see. but of a totally different type. they would simply use an approximate value.

numbers under 60 are written in the ordinary decimal notation... Thus the symbol for 10 can also mean 10 X 60. but we shall transcribe them sexagesimally and 1 For Babylonian cultural history see. V .the number 9. Neugebauer. Paris. Chiera.S)"'im. Next E. I and 11 (Heidelberg 1921-25). M. There is nothing special in all this.S)"'io. Thus in '1 <<'0/ . 1922. Babyloni. we shall not put the sexagesimal numbers that occur in the texts in the decimal system. Waerden. Thus we see that the value of a symbol depends. the great lawgiver and ruler of the first Babylonian dynasty could call himself "king of Sumer and Akkad. Die Keilsch. the symbol for 10 also 10/60 or 10/602 • The fractions Y2 = 30/60. the lower powers at the end. Meissner. d. hence the word positional notation. 414 and the literature cited there. and with the multiples of 60 we start counting anew. The symbol for 1 can also mean 1/60 or 1/602 . 1904. upon the position of the symbol in the number. . 8. Fot cuneiform script. Thus the symbol «< W represents 35. For the chronology.38 CHAPTER II dwelt farther north. But. In the course of time the Semites became more and more dominant. the symbol for 1 also 60 X 60 or even an arbitrarily high power of 60. for example. and about 1700 Hammurabi.20 and 12 respectively: «< « <yv. Lux 10 (1948). the simple vertical wedge has the value 1. the symbol m. as in our modern number system. I. 60 + 24 = 84 or 1 + 24/60 = 84/60. Fossey. Following Neugebauer. it corresponds to the Egyptian system or to the familiar Roman numerals. 1938. Stylus for cuneiform script were produced by pressing a sharpened From O..later abbreviated to 9. and B." 1 In the Sumerian-Babylonian notation. Both signs Fig. L.bort Ex Orient.the first vertical wedge is worth 60 times as much as each of the four small ones. 8). fractions are also written in the sexagesimal system. Meissner."utl d'A. the wedge b with two ends the value 1O. f••. But now comes the remarkable thing.i[t (Sammlung Goschen). Vorgriechische Mathematik stylus into a clay tablet (see Fig.. see B. the number 60 is again represented by the symbol for 1. see Ch. more than this. IoSi. The value of the symbol can be. in the first place B. v.olt on clay. Chicago. Ya = 20/60 and 1/5 = 12/60 are then represented by the symbols for 30. Thq . • nd A. p. the higher powers of 60 are placed at the beginning.

as if they were whole numbers. in this way it became almost equivalent to our decimal system. also at the end of a number. The number just presented will therefore be written as 1. e. 3. DIGITS AND THE ART OF COMPUTING 39 separate the different powers of 60 by commas. 4 ~ 3604. Compare.30. It is true that in practtce the lack of distinction between the symbols for 1 and for 60 is not serious. for example v~ '0/ = 1. but this is lcss important because he hardly ever needs large integers. NUMBER SYSTEMS.g.D. we know very well that It does not stand for 30 cents!).). With these tables he can also multiply sexagesimal fractions. But for the Romans.g. 30. This gave the "finishing touch" to the sexagesimal positional system. and. be- cause the order of magnitude is usually known from the context (when we see the number 30 on a dress in a shop-window. 60 + 30. we treat the tens and the hundreds as if they were units moved one or two places to the left.. Still more so is the fact that the notation does not enable us to distinguish between 1. likewise. 30 = 63Yz and 1. or we shall use a semi-colon. to serve the purpose of the modern decimal point. 3. Do we not calculate with decimal fractions as if they were whole numbers. To overcome this drawback. and hence for our minutes and seconds. uses the symbol 0 for ZHO. It is true that Ptolemy wrote the whole numbers in the decimal system and only the fractions sexageslmally. CC x LX is something very different from II X VI. 30 = 17/120 . When the context shows that we have to multiply or divide by a definite power of 60. and he can ignore the entire muddle of calculating with common frac- tions. but in theoretical problems it can be very unpleasant.3. 3. we shall supply one or more zeros at the end (e. and then put the decimal point in the proper place in the result? Babylonian positional notation also had disadvantages. the Babylonian only requires tables from 1 X 1 to 59 X 59. 24. because there is no cypher. . And 1. but this is of minor importance for us now. The systematic positional notation has enormous advantages in the technique of computation. In practice these numbers are of course split into tens and units.602 + 3 . a multiplication in the modern notation with such a calculation in Roman numerals! To compute 243 x 65. a separate sign was introduced later on for the empty place between two digits. We can limit ourselves to the multiplication tables from 1 X 1 to 9 X 9. e. 0. 1.g. The enormous superionty of the sexagesimal fractions in computation was responsible for their use by the astronomers. 30 and 1. who made all his computatIOns in the sexagesimal system. 1. or this number multiplied or divided by a power of 60. O. The Greek astronomer Ptolemy (150 A. 0). 30 will mean 1 .

By simplification of the symbols (3600 = 1. The fact that 1 and 60 are represented by the same symbol. Why was 60 considered as the "big unit" and represented by the symbol for 1? 3. and by a systematic continuation of the powers of 60 in both directions.3600 3 The last of these symbols was called sar-gal. What led to the choice of 60 as the next stage after 1 and 1O? 2. it is also understandable why 10 .e. They had the following symbols: Oldest Sumerian period (before 3000 B. The large circle for sar 3600. i. was scratched in with a sharp stylus. The following sets of symbols are found.) &>ij <U <W D 0 0 [Q) 0 @ tf I t:Ii. big sar. I ~ 2 Y < Y 1< f:1¢1 ¢>' 2 "3 10 60 600 3600 10. which terminated the number system. is the fundamental principle of the positional notation. the stylus was held obliquely. 60 was the next step in the scale and why 60 .C. 60.) DODIQ)O 10 60 600 3600 The symbols for 1. for 1. for 10 perpendicularly. On this supposition. etc. What led to the representation of fractions in the sexagesimal system by taking 1/60 as the "small unit"? The second and third questions are more important than the first. Later on the symbols were separated into wedges which were pressed into clay by means of the sharpened stylus. Th:cee questions remain: 1.).3600 60.C. The symbol for 600 is a combination of those for 10 and 60. the sar terminated the old number system. side by side.<40 CHAPTER II How did the sexasesimal system originate? Originally the Sumerians did not have a systematic positional system for all powers of 60 and their multiples. in the Later Sumerian period (about 2000 B. one obtains quite naturally the consistent sexagesimal notation of the Babylonian texts. Apparently 60 was looked upon as the "big unit". 10 and 60 were made with the cylindrical lower end of a round stylus. because it . The symbol for 60 is sometimes a little larger than that for 1.

the choice of 60 is a historical accident.. Abl. and also monetary units. 10. 4 of his excellent book. and by 60 among the Sumerians. p. 1 mana = 60 bushel. as has already been said. and to Chapter 3. 1/12 of the as. But what answer does he give then to our second question? None at all. 2. . and this one is 1/60 of the first. du 'Y"'''''' """S. the development went in the opposite direction: the use of 60 as a big unit was carried over from pure numbers to the monetary units. which appears everywhere (probably because we all have 10 fingers). p. when it comes to our third question. not invasely. Es9uisse d'unt hi. In the domain of pure numbers. . 13 (1927). 10. Perhaps. in which it was important that one half and one third of the larger measure be simple multiples of the smaller one. is strictly reciprocal: one is 60 times the other.Dangin. and the English three score for 60). one would not be led to consider 1/60 as a new unit. .toi. Ya and 1/6 in terms of the new unit 1/ 12 . F. F. . DIGITS AND THE ART OF COMPUTING 41 is not the magnitude of the basis of the number system which matters. the reader IS referred to Neugebauer's article. a part in the choice of 60 was played by an old normalization of measures.ft der Wimnscha{ten. I. mentioned above.. The next step of great importance consists in carrying over this notation to non-denominate numbers. Thureau. As is well known. viz. were arranged sexagesimally: 1 talent = 60 mana..imal. Again... but in the monetary realm one would.. . In our own terminology. NUMBER SYSTEMS. we really do not know. 2 According to him. But. 1 O. this question is the least important of the three. and only there. In this domain. the uncia was originally a weight. Pans. these monetary units were adopted by the Greeks. 11. is followed by 20 among the Celts (remember the French numerals quatre-vingt and quatre-vingt-dix. But. the relation of a larger unit to a smaller one. when one monetary unit is sixty times another.. Thureau-Dangin. 1932. than denoting the larger unit again by 1 and designating its multiples by the ordinary symbols for 1. For a further discussion of ancient normalizations of units. In a certam sense. as Neugebauer 1 assumes. Gmllsclu. by 100 among the Egyptians. Vorgriechische Mathematik. dlungtn. one has the Roman custom of expressing the common fractions Y:!. Analogously. par. Thureau-Dangin also thinks that the origin of sexagesimal fractions must be looked for in metrology. Giittlngen. does not consider this hypothesis valid. Neugebauer answers the second question by the very plausible statement that the source of 60 as the "big unit" must be looked for in metrology. one-twenty is frequently used for one dollar and twenty cents.. but rath~r the principle of the positional notation and its extension to fractions.. perhaps by 50 among the Teutonic peoples. Nothing is more natural. the uncia. Neegebauer. Frankness compels us to say that the master of modern assynologists. Weights.

20 Tf~~ nr~ a-ra 30 3. 40 n:~ W4 a-ra SO 5. Frequently. . 2. 53. to form a large combination-table. 16. 1 O. .53. A separate table contains the multiples of a single number. . 48 1 : 1. from which the Sumerian number systenl was deduced. 33.15. 20 .SO a-ra 50 13. Neugebiuu. 0. 6. A ·3. 40 a-ra 40 ll.. 38.40 lf~-« rr~ a-ra 20 2. 58.53. 7. we must first have a look at the short tables of inverses.20 1 : 1. 13. One of them begins as follows: 1: 1 = 1 1: 1. 24. Before we can understand the arrangement of such a combination table.20 = 59. Bulin. 56. There are large tables of inverses from the times of the Seleucids (311-1 B. 1.. .0. a-d.53. 30 1: 1.43. 6..45 =59. dating from the time of Shulgi (about 2000). 45 = 58.42 CHAPTER II Sumerian technique of computation. The latter tables appear singly or in combination. several of these small tables are combined with a table of inverses and a table of squares..0.16. 52. 33.13 a-ra 19 5.40.20 Obviously. 19. Qwllta rwl SluJi. Two examples: Table of 7 and of 16. were tables of inverses (ljx) and multiplication tables.59. 34. The most ancient Sumerian texts. 30 a-ra 30 8. In this way it continues throughout 7 pages of Neugebauer's "Mathematische Keilschriftentexte" 1 up to 1:3 = 20. means "times".). 20 rf:. oS. VJ rr~ f . 7 a·ra 1 7 a·ra 1 16. 10. probably designed for the use of astronomers.C. 20 lf~~ ~4' a-ra 40 4. 20 a-ra 20 5.40 If ~r rr <fi a·ra 2 14 a-ra 2 33.:r m <<r a-ra 3 21 a-ra 3 SO Ii. 1935. It .33. lf~<i IT<m a-ra 19 2. 40. SO.

For example. 40. one is told (not in terms of a general formula.30. 21 44.24 1 1 8 7 30 27 2. DIGITS AND THE ART OF COMPUTING 43 ° The older tables of inverses are not so extensive. 4. This also accounts for the arrangement in order of decreasing magnitude. The matter can also be stated as follows: the combined tables of inverses and of multiplications is used to multiply numbers. 24 45 20 9 5 2. Normal table of inverses 1: 2 =30 16 3. 30 1. 20 3 20 18 3. 30 2 40 16. 15 44. 30 7 3 1. 3 and 5 and which have therefore reciprocals that can be expressed as finite sexa- gesimal fractions. but also to represent common fractions as sexagesimal fractions. and the others. 9. but especially for multiplications of the form a . . 12 SO 10 6 32 1. 5. 40 6 10 25 2. 20 25 12. 20 45 15 4 40 1. 40 1. b-1 . 30. Thus we see that the multiplication table served not only for ordinary multi- plications a . 6.20 1.45 45 1. we first locatel: 8 = 0. 50 which are to be expected in any ordinary multiplication table. 30 3. 20 1. are inverses of simple numbers. in all of the following numbers: so 24 12 6. 15 48 12 5 36 1. for divi- sions a: b. it is indeed the order of the reciprocals in the table of inverses. The mathematical texts fully confirm this interpretation. 8. 40 36 16 7.e. 12 5 12 24 2. 40 2. 20. b. i. 30 10 6 2. 10. 26. Whenever a division a : b is to be carried out in these texts. to write % in sexagesimal form. which contain only factors 2. but for definitely specified numbers): calculate the reciprocal b.30 54 1. NUMBER SYSTEMS. 26. 40 8 4' 1. 15 What determined the choice of these numbers? Most of them occur also in the normal table of inverses. 6. 40 18 8. 30 1.20 48 1. 30 30 15 7. but also tables for the multiples of several other numbers of two and three digits. 15 4 15 20 3 SO 1. except the number 7. 30 48 22. 13. 20 4. 52. 2 1. 45 1. 3. They usually contain the reciprocals of those integers between and 81. 4 56. 7.40 30. 40 The corresponding multiplication tables contain not only tables for the multi- ples of 2.7.l and multiply it bya.22. 12 3.30 in the table of inverse and then the product of this result by 3 in the multiplication table: 0. 15 9 6.

.e.1 1: 1 = 1 1: I. In the same way. Neugebauer and A. 16. 1 1 2 a-ra 2 4 3 a-ra 3 9 59 a-ra 59 58. in which this word occurs as well: 2-e 1 ba-sl 12-e 2 ba-si 36-e 3 ba-si. etc. p.: of 1 is 1 the root) 4-e 2 ib-si 9-e 3 ib-si. the four rational operations can be carried out rapidly without further thought. Whenever a division did not go without remainder. A systematic use of the advantages of the posi- tional notation avoids all messing about with fractions. From a short table of squares.1. 1 = 59. The text is reproduced in O. 1 YBC orands for Y. but more generally "root of an equation". The word ba-si however does not only mean cube root. Conn .59 1: 1. are used for the solution of pure cubic equations Xl = a. Indeed there are also tables for the equation xl(x + 1) . such as supplements most multiplication tables: 1 a·d. a. 2 = 58. 52 etc. An ancient Babylonian text. approximations were used. Babvlonian Collection. 1 one can of course immediately derive a table of square roots: l-e 1 ib-si (i. New Huen. square roots and cube roots.• 1945. YBC 10529 1. Sacha. l-e 1 ba-si 8-e 2 ba-si 27-e 3 ba-si. Mill"'· _iarl C_i/_ Texts. CHAPTER II It appears therefore that the Sumerian-Babylonian calculation tables were arranged in a very useful manner. 3. gives approximate values for the reciprocals of all numbers from 40 or 50 to 80 in the following manner: 1:59 =1. tables for cube roots. etc. In Babylonian mathematics these short tables are used in the solution of qua- dratic equations.

O!Iellen und Studien Al. Mul. j. Berlin Museum (V AT 12593)_ The tablet was found in Fara.W.. You get 56.L \. and the multiplication of 28 by 2 to the left of the drawing. Sumerian clay tablet with num- ber signs. PL. in Southern Mesopo- tamia. Perepelkm.I I I 1111. 2. The text says: (1) Add together this 16 (2) with this 8and this 4.. The hieroglyphic text IS read from right to left. (3) You get 28.. Agove. PLATE 5 II I lI\.\1\ 1111(l~'III"'1 . You have found right. Papyrus des Museums in Moskau.7. . Math.. (6) Behold: it is 56. You get 2. 4. Later texts were written in lines from 1. executed by j. 9. 40. the hieratic text. both reproduced fromW. The columns should be read from bottom to top. the an- cient Shurupak. Struve. the lower length 4 below.. Two Columns of the Moscow Papyrus. At the top of the last column we see the numbers 10 and 20 (Hat circular im- pressions). 8.. Compute (4)'/ •. The middle column shows the numbers 1.rl ' '(... .. c=~11 l in eo a IIMO II 'C7 'C:7 I I I1f1<=>!J! I I II 0::>. See p.(5) tiply 28 by 2... of 6. below. . 5a. containing the computation of the volume of a truncated pyramid with sides of 2 and 4 cubits and a height of6 cubits. 51) the text was probably written in the beginning of the third millennium B_C. the hieroglyphic transcription.:1 ::: rn.l 1 : : V1Ll\{'O~~--: I 1IIIfl~:::l::::&' ~. 5. left to right. The upper length 2 with its square 4 is written above the drawing.. According to Neugebauer (Vor- griechische Mathe". 3.tik p. 1111 I IIIVY1R • I I I ~llof 111(1-«» o::>fII e I 11 1f1'C7 1111('1 1IIIo.. 'C7 III 11<:::>tJII' 08iii=' . ~~"i"':: :R~+?~6t PL. 5b. r . the height 6 and the volume 56 within.

50 by 0. At the bottom crated under Adadnirari III (810-782 the old number signs are found (see p. Athens National Museum. Limestone statue (height 1. monetray units. . Salamis. Nabu. 46) and some signs for special B. probably from the 4th century B.C).C. Marble counting (1. PLATE 6 PL. 6b. 377).65 m). 60. conse· This may not bave been actually used as a counting board. god of Babylonian science. British Museum (D. found on the island of PL.75 m).

as the average of a = 1 ..Sachs. Texts. The Greeks (Archimedes. Leiden.Itt. Thureau-Dangin. In most remote antiquity. r. the same which turn up in later Greek arithmetic as "paramekepipedoi" numbers.25 and 2 : a = 1 ... Set Plate 8b. New Haven. as is seen by squaring a + bj2a. gives the general approximation-formula ~--. if a is too small. 24.42. Tnc'" """h. 2ja will be too large. Bccker.e.. accompanied by l-laI. appro- ximation was used."d S. Heron) knew also approximations for v'2 from below. The Greek notation for numbers. 1945. a + bjia. one finds 2ja = a/a. DIGITS AND THE ART OF COMPUTING -45 Moreover the same word ba-si is found in VAT 1 8521. Then...di . .. p. 3 (Plate 8b) Application of the same method of the arithmetical mean to the root of the sum a2 + b.". 'fa.24. and vice versa. The approximations obtained in this way are always too large. 123. to these we shall return at a later point. For this text Sf< F. 10. 1938.. '0.1) occur here. 181. so that the closer approximation is 1'/11 = 1. For instance.i.1 To determine the square root of a number which is not a perfect square. 51. B 4. p. This approximation frequently occurs in Babylonian texts. C. with the meaning: root of the equation x 8(x . It is worthy of notice that the numbers n2(n + 1) and n2(n . and even continued-fraction approximations ("lateral. as compared with the excellent Babylonian notation. "less 1".!o. 25.. 43. the very close approximation v'2 = 1. i.. A better approximation is then obtained by taking the arithmetic mean Y2(a + 2/a). The following method is also used frequently by the Greeks: Suppose that a is a first approximation for v'2. A repetition of the process gives.21 . . babyl. which is also found in Babylonian texts. was really a retrogression. If one starts with a = lY2.1) = a.. Quillt• . which resembles the well-known Roman numerals: li6rHiXiMr 5 10 50 iOO 500 1000 5000 10' 5 10 1 VAT stands for Vorder·Asiatische Textsammlung Bcrliner Museum." • Neugebauer. This is found in the recently discovered cuneiform text YBC 7289. they had a notation. .and diagonal-numbers").. NUMBER SYSTEMS.

. Archimedes and Apollonius used still different notations. M are of course the initial letters of the Greek words for 5. . ~ (900 = m= Sampi) 1000 . Numbers beyond the myriad M = 10' were designated by use of the symbol M. 1936.-Naturwias. one finds tables in which first 1 is multiplied by all the 37 numbers a. Tannery. 100. then 2 by these same numbers.) had already abandoned this simple notation for sums. In place of M one could also write XE'. an accent was added at the end.l. Until the time of Plato's friend Archytas (390 B. . Compare K. Sitzungsbcr. But even Euclid (300 B.... e.9 cx:. e_<. as in our algebra. .... Later on. alphabetic notation was. Undoubtedly they are bucd-on much older Greek sources.. When Euclid wants to add two numbers. a briefer. 10' = 108 • I The letters of Rbabdas were edited by P. a. For higher powers of M.i. In our schools we only have to learn the multiplication tables from 1 X 1 to 9 X 9. the Greek number symbols were about equally troublesome.. letters were used for indeterminates.1.. It did not leave the letters available for indeterminates or for unknowns. = 10'.. M."" (6=<. 90 1. t. 0: T.J. 9000.aTE' = 1305. .).. If supplemented by a sign for multiplication. and so on up to 10' . M PY' = 250 043 .islii.aTE or .A.. .. 900 e. this system might have provided an effective notation for theoretical arithmetic.-y. e.. X. p.~. I think. in Archytas. '\4'. 10.P. such as .ad.p. n. IV. introduced: 1.T").<p. scieroliMuu. .1. .. Abt. Gritdoiscltt u.a. p.g..: . . . Alr. " •••• . 1000 and 10 000.'j (90=(j=Koppa) 100 . Bayer..O.~. but in the 14th-century Greek arithmetic of Nikolaos Rhabdas I.).fl.X. or a dash was placed over them. 61-198. he represents them by means of line segments AB and BI. starting with 2..C. FLI represented for instance the sum of the numbersF and .1.C.• Miin- chen (Math. =Vau) 10 .o.46 CHAPTER II The letters II. p. etc.K. (accent at lower left ) To distinguish numbers from words. The use of letters for specified numbers was not advantageous for the develop- ment of algebra. ..00. and denotes the sum by AI.u. H. probably. to avoid confusion with the alphabetical nl!mber symbols. a minus sign and a symbol for fractions. . For purposes of calculation. Vogel.. 357..v.

e. There was known also "Egyp- tian multiplication' . 1 000 300 25 This method was called "Greek multiplication". A number consisting of more letters was split and written as a sum of numbers re- presented by a single letter. In Polybius (V 26) we 1 See the examples of Pappu. then multiply these radices and finally determine the power of 10. etc. e v . 265 btt a~e times 265 .aY"E aXE together 70225. In the "Sand-reckomer". indicate how general was the use of the "Abacus". To avoid this unmanageable table and to require only the use of the ordinary multiplication tables from 1 X 1 to 9 X 9. ~ a MM. those in the next to the last row 10. were aligned in parallel rows on a counting board (or perhaps on an ordinary table on which lines had been drawn). p. the third the product. a~E I. finds their pro- duct 6. There were separate rows for special monetary units.oii . then. the first two contain the factors. Remarks. The example 265 X 265 is found in Eutocius: Tn M.102 = 10'. by the rule of Archimedes. MuncheR: 1936.e. such as the obols. 100 X 400 = 40000./J. the counting board. 200 or 2000 by their "Pythmen" or radix. The pebbles. 102 . DIGITS AND THE ART OF COMPUTING 47 It is all arranged in three columns. so that the product turns out to be 6 myriads. Written computations were obviously rather complicated.YXT 12000 3600 300 op. familiar to us from Chapter I. by which this product has to be multi- plied. one would have to replace factors like 20. in antiquity. in Vogel. one determines first the radices 2 and 3 for a' and T' respectively./:J . 1 All this is very cumbersome! Numbers represented by a single letter could be multiplied in this manner.a 40000 12000 1000 - M. Archimedes gives for this purpose a rule which is equivalent to our formula For instance. i. scattered in the works of various writers. Sitzungsber. to use this rule for the calculation of a' times T' (200 times 300). 393. in this case 2. .by means of continued doubling. NUMBER SYSTEMS.g. e. moreover paper was expensive. The pebbles in the last row had the value 1. and therefore one took recourse in practice to Counting boards and counting pebbles. called Psephoi..- M .

" "Combining" apparently re- fers to the addition of fractions. and the combining and splitting of fractions.. the counting board was in general use in Western Europe. even to-day Chinese and Russian merchants follow the Roman exa llple in using the abacus for calculating.woIDsi.... to add 3 and Ya and lIlt and 1/. a marginal note. or rather beads. which can be pushed back and forth along wires.4.WMa/i. tI· A. by reducing to a com- mon denominator...... at the will of the reckoner are now worth a copper and now worth a talent". can be learned from the Akhmlm papyrus I. for example: "The seventeenth part of 3 is 1/11 + 1/17 + 1/61 + l/ea ... which Rhabdas performs. 20 and for various values of m. Among the Romans we find..e. because it was no longer needed.. 1. Prom McruUngcr.48 CHAPTER II find: "those in the courts of kings ..ft . For example.".. tit It. ." We get to know the notation for fractions from the works of the later mathe- I J. describes in the following words the most important aspects of arithmetic or "logistics": "The so-called Greek and Egyptian methods for multiplying and dividing. added to 1/.. M . Bainet.. The c:baoccIlor ot the uu. this gives 14 + 3 + 1 = 18 forty- seconds.. What is meant by the splitting of M.ury OD the Dui_vue (piau 7) 39 sevenths. Pari... this leads to Pig. combi- ned with 3 units. For these. .. in Plato's Charmides. &lido OD the COUDting board the IUClI that bave b«D ra:eivcd. Up to the late Middle Ages. 1892... i. Le P. besides the large counting board...rapiu •• c... . Many peoples have adopted this convenient instrument.. patterned after the Salamis counting board. A scholium. exactly as we do it. lIlt to three forty- seconds. It is also significant that the common verb for "to calculate" is Psephizein derived from the word Psephos the counting pebble. or 3 sevenths. Mi"i. . in which the fractions min are split into unit fractions for n = 3. (Plate 6a) also the small hand-abacus with a restricted num- ber of pebbles. 9.. 9. fue.. Zahlwort unci Zifcr. It was not until modern numbers were introduced that the abacus gradu- ally disappeared from the scene.i. are in truth exactly like counters on a counting board. Calculation lIIilh fractions.a' he changes Ya to fourteen forty-seconds.

of the second that the third is of the fourth.luns Erzhtrzog Rainer. 11. Def. which. or. "Parts" means a number of n-th parts. in Plato's own words. but rath~r because one did not wish to know them. Vogel. Def. trans- lated into the terminology of fractions. Heron.. according to the Egyptian example.). For example. Wien 1932.C.c/I< Aufsabtnsammlung im Panru. when it measures the sreater. Folge. Furthermore. follow the translation given by T. here and in the sequel. Elnal'. Heath. In Book VII of Euclid's Elements. traces are found of an ancient technique of fractions. 3. Why should they? They were not bound to a frozen tradition. to y'. N UMBER SYSTEMS. visible things are divisible. These definitions introduce therefore arbitrary fractions.. For... But parts when it does not measure it. but not mathematical units. Fractions were scorned and left to the merch- ants. Neue S<rie. so it was said. "the experts in this study" were absolutely opposed to divid- ing the unit (The Republic. = 3/S' According to Vogel.C.d'. or the same parts. It is true that they frequently operated (especially in the later papyri which were subject to a strong Egyptian influence) with sequences of unit-fractions.t/. y" or something similar. 20. 1 has also the inverted form. One application is found in the definition of proportionality. in which the copyists may have altered the notation. they operated with ratios of integers. Numbers are proportional when the first is the same multiple.. e." Pan"'" . Milleilu"s. the word TO Tf!tTO" could be written out in full.g. Nevertheless. Frequently the denominator was placed above the numera- tor: . p. TIrr . we find in VII 34-36 a development of the properties of the LCM. still shorter. as we shall see later. 4. This was not however because they were not known. and his contemporary Eratosthenes uses the fraction 11/83 to designate the inclination of the ecliptic.. Gustinger und K. L. A papyrus from the 1st century A. such as 2/3 + lito + 1/30 . was produced before 400 B. how fractions are reduced by determining the GCD of numerator and denominator and dividing it into both.. Instead of operating with fractions. . A number is a part of a number.It. I All refuences to Euclid's Elements. from papyri like those already mentioned... the unit was indivisible and. 3 fifths. but they could also replace this by 4/6.D. for. 1926. Then follows an explanation of the reduction of ratios to lowest terms or. From this we see that the Greeks did not restrict themselves to unit·fractions. 19996.. Eine "momelri. this notation came into use as early as the time of Archimedes (3rd century B. or it might be abbreviated to TO y . DIGITS AND THE ART OF COMPUTING 49 maticians (Archimedes. I.a. The determ- 1 H. Official Greek mathematics before Archimedes does not have any fractions at all. "Part" means here n-th part. or the sallie part. where n is an integer. Archimedes proves that the perimeter of the circle lies between 3 10/71 times and 3 1 / 7 times the diameter. most reliably.". second edition. VinJobon. WI of E. the "Indian notation" ~ = 3/S' which led to our notation for fractions. Cambridge. For Ya. 525E). because these are originals and not later copies. the less of the greater. Diophantus) and. according to Plato. we find the following definitions 2 : Def.

The method for estimating is quite remarkable. There was no danger of confusion with the number symbol 0 = 70..IIt l'AIIIIIJ8UI•• Tome 2. which corresponds to the musical interval of the fourth. Theon does not mul· 1 The actual tide of this utronomical. For 47°42'40". etc. Ptolemy writes ". A Rome. etc.. which means "one third additional" (1 %). and then he subdivides each part according to the same scheme. we would try whether it goes 60 times. the third part remains". seconds. etc. In modern symbols: 1515°20'15" : 25°12'10" = 0007'33".50 CHAPTER II ination of the LCM is important for reducing fractions to the lowest common denominator. difficulties in the calculation with fractions can not have been an obstacle in the way of the development of mathematics. • P.m. is called the "~itriton". the circle is divided. In the Almagest. Book K. to which corresponds the ratio 9: 8 is called the "epogdo-on".".e. " • TWo. on the Babylonian pattern. to a common denominator.{J' . an abbreviation of ovlJb = nothing. The terminology of the ratios of numbers in the Pythagorean theory of harmony also recalls the fact that these ratios are originally fractions..tandard treatise is Syntaxis mathematib.".. The most ancient occurrence of fractions is in Homer's Iliad. One begins byes" 'llating the number of units in the quotient. from remotest antiquity. Ptolemy is a virtuoso in computing with these sexagesimal fractions."".C "'P' p. reduction to lowest terms. they had mastered the opera- tions on fractions. when Ptolemy considers a circle. c. etc. into 360 degrees. multiply 25° 12'10" by 60 and see whether the result exceeds 1515°20'15". extracts square roots. the Greeks have known fractions and that in the 5th century. in contrast with the Egyptians. We conclude from tnis that.. of' alpes 'X' es" by 'XB t. For this reason we have to be grateful to his comm otator Theon of Alexandria I for giving an example of a sexagesimal divi- sion. For them therefore. at the latest. he divides and multiplies them. The word Almagest is an Arabic corruption of Meplc Syntaxis or Megiste with the Arabic anicle AI. For zero he has the symbol 0. 253: "Two parts of the night are past. he usually divides the diameter into 120 parts. It is not known who was the first to do this.-. and the whole tone. i. that is why the Greek astronomers adopted the Babylonian sexagesimal fractions. t/·A"'-otlm. Roma 1936. because numbers beyond 60 do not occur in counting minutes. But other units are also divided in the same way. . without wasting a word on the technique of these calculations. because the superb "Almagest" 1 of Claudius Ptolemy has cast the works of his predecessor into oblivion. viz. indeed more effective than ours. 11/8 . Sexasesimal fractions. Common fractions were too awkward for astronomical calculations. The ratio 4 : 3. •. each degree into 60 minutes and each minute into 60 seconds. According to our school books.

Prince Gautama 1 See. Dividing 190' 15" by 7' gives more than 25°12'10"..g.liCJ I. Now 60. he knew the Babylonian eclipse observations and lunar periods. but it is certain that the conquerors constituted the highest ca<. let us say between 2000 and 1400 B. That is to say. particularly in terribly large numbers. but he divides by 60. he divides the dividend by the estimated initial digit of the quotient. introduced by the Dutch engineer and mathematician Simon Stevin in his work "De Thiende". The following scene occurs in the book Lahtavistara 1. those of the warriors and the brahmans. i. etc. Does the reader see the advantage in this method of calculation? Our method would require the multiplication of the entire big number 25° 12' 10" by 7 or 8. so that we have to take 7'. Ptolemy's predecessor.60. Hipparchus calculated tables of chords and I can not ima- gine that these were arranged in any other way than sexagesimally. But we observe that in the Buddhistic period. also used sexa- gesimal fractions for his calculations. 10. e.e. . they preferred the sword to the pen. Indeed. _Jrnr. thus we have to take 60. Sexagesimal fractions never disappeared entirely from astronomy.C). taken from Menninger's excellent little book. Where do our Arabic numerals 0 1 2 3 -4 5 6 7 8 9 come from? Which people invented our excellent decimal positional system? Arabic and Persian arithmetic books are unanimous in ascribing the invention of the nine digits to the Hindus. NUMBER SYSTEMS. Hindu numerals. a remainder of 829"50'" is obtained. nobility and clergy. Division of 1515° by 60 gives a quotient in excess of 25° 12' 10". subjugated the population and introduced the caste system. I assume that Hipparchus (150 B. if we look at the forms of the digits in Fig. they show an interest in numbers. 10" are successively subtracted from 190' 15". In prehistoric times. but they did vanish from other fields of exact science.e. DIGITS AND THE ART OF COMPUTING 51 tiply by 60.12' and 60·10" are successively subtracted from 1515°20' 15". which in turn derive from the Hindu numerals. 12' and 7' . This was the endpoint of a develop- ment which began with the introduction of the Hindu-Arabic numerals. since he was thoroughly familiar with Babylonian astronomy. 7' .C.25°. These people did not pay much attention to science. this leaves 190' 15". i. Like Ptolemy. but division by 61 results in an answer that is too small. But let us start at the beginning. We do not know whether the separation of the castes was as sharp then as it is now. 25°. in the last six cen- turies B. Datta and Singh. we see at a glance that our numerals come from the West-Arabian ones. Why? Because they were supplanted by the decimal fractions. Hislory of Hirui. but Theon only requires that 190 be divided mentally by 7 or 8. we estimate the 7' of the quotient. but division by 8' gives less. Then 7' .tes. the Aryans pene· trated India. In the same manner.C.

Beyond this there are 8 other series. According to an arith- metic book. these same words ayuta and niyuta have other values.a.e." "How then do the numbers beyond the 11.10" = 1063• But in most arithmetics. 10. Zahlwort und Zifer.:'(. wrestling. one hundred ayutas niyuta. ko# is a hundred times one hundred thousand (sata sata sahassa). swimming and arithmetic. Cent. Grnalogy of our digits.0# contiI!ue by hundreds?" "One hundred kotis are called aJllta. From Menningt'. Now the great mathematician Arjuna questions him.329 running.-q-:-Ib-'/\-S-'-'.'lr.>ks the prince Dandapani for the hand of his daughter Gopa.ara. Westarablc (Gobar) Eastarablc 11. were constructed in stages to dazzling heights. one hundred niyutas lr.. But Buddha has not yet reached the end: This is only the first series. viz. 10' and 105•. They are pure fantasies which. through 23 stages.. he wins with flying colors. he says. "Oh. . 16.. like Indian towers. It is clear that these numerals were never used for actual counting or for calcul- ations. Cent.wara . Of course. so that the largest number mentioned by Buddha is 107 . In this manner Buddha continues.~-.. (Diirer) Fig. p. one hundred karikaras . young man. archery. do you know how the numbers beyond thdoJi continue by hundreds?" "I know it. (Apices) I':' J~ If 16""-89" I Ilt j +!I6"'r89 0 15.-. Cent. 52 CHAPTER 11 (Bhudda) ~. Ii-. He is now required to compete with five other suitors in writing.

. shown in fig. which already appear in the third cen· tury B.dica II. DIGITS AND THE ART OF COMPUTING 53 Number symbols: Kharosti and Brahmi. When did they invent it? This is a difficult question to answer! On a procla- mation of gift of the year 595 A. The Kharosti are shown in fig.) on. number symbols are also found in books. 12. the Arabs themselves included. look at those for 6.C. 11. the year Samvat 346 is expressed by means t EpigraphiCQ I. NUMBER SYSTEMS. 9! Some kind of "denominate positional system" begins to emerge in the numbers beyond 100. It is seen that a symbol for zero is still absent and that there is no positional system. supplemented by a zero. 4. For the Hindus alone.C. 12 Brahmi numbers. were to be used to repre- sent all numbers. the written numerals appear in the general wntten language like strange ducks in a pond. 8. For all other peoples who have adopted the "Arabic" numerals. Kharosti numbers. such as Kharosti and Brahmi. 19. and which. as in I /I III X IX IIX XX ? Roman numerals. these were 1 2 3 4 5 6 8 10 repeated as many times as necessary. The symbol for 60 is in no way related to that for 6.= . Nevertheless the system " holds promise. p. 1 2 3 4 5 6 7 8 9 Each of the numbers less than 10 CL 0 .D. Different types of script were in use at that time. later on. From the time of the great Buddhist king Asoka (3rd century B. The date of tile iuvention. The invention of the positional system. There were spe- cial symbols for I. As Datta and Singh justly observe.~ ~ tr 1 ') germs of the later development. the symbol for 100 or for 1000 is combined with the numeral which indicates the number of hundreds or of thousands. it contains the . they are the same numerals from 1 to 9 inclusive. The steps 4 and 20 also appear elsewhere. 11. The symbols themselves have a striking resem- Fig. each having its own number symbols. 7. they are not strange. 10. blance to the Arabic numerals. not by two or three as in the 7 2H )f 100 I 5H 7 4T 'tt 1000 ?OT r Kharosti system. this alone suffices to show that the Hindus have to be recognized as the inventors of our positional system. thus they do not present J 7H JJ) JJ)) AI 7" anything of special interest. I. 20 so 60 70 100 200 Much greater interest attaches to the Fig.r 'A J -I ~ Cl) e 10 20 30 40 50 60 70 80 90 is represented by a single symbol.20. taken once again from Menninger. Brahmi numbers.

i"". because there is only one moon. the number 1021 might be written as follows: sasi-paksa-kha-eka moon-wings-hole-one 120 1 In this manner the Indian astronomers learned an entire table of sines by heart in verse form. in which such a table of sines occurs in verse. It is as if we were to say "muses" for 9. because mythology knows 3 fires. 646.. etc. For 2. and you will find tha~ it does not go very easily. "arms" or "wings".. they used.hernDlics. This is what happened in Arabia and in Eutrope. Professor Gonda.. To prove this statement. they have no synonyms. or almost none. or "brothers". for understandable reasons.. "fire". they began with the units.1 Soti. 4 and 6. Tht: most ancient work known to us... declares that 16 of them are not genuine. Even though there may be some falsi- fications among them. everyone would at once know what is meant.. Journal of the Asiatic Society of Bengal.. This seems to me like a very high percentage.lo/ lhe A. 141. In their History of Hindu mathematics. for with such matters. }. and that they become known in wider circles only very gradually. followed by the tens. substituted other words for the numerals. or a table of sines in a rhyth- mical and rhymed form? Try it. dating from the years 595. according to its position. Datta and Singh list 30 inscriptions on which appear dates and other numbers. . too prosaic. for instance. 1907. Have you ever tried to write a table of logarithms. written in the decimal system. In the opposite direction therefore from the one used at present. . Numerals are too rigid. Notes •• I. p. that is to say. Ori"". Zeros were also mentioned. it is usual that at first they are used only by scholars. Kaye 1. Kaye. That is why the Hindu astronomers. 674. or forty..". who lists 18 of them.. Poetic numbers. CHAPTER 11 of the Brahmi digits 3. or "graces" for 3. up to 972. or four hundred. the Sanskritist of the University of Utrecht. sees no reason to consider these inscriptions as not genuine. the nine digits and the zero were generally known. we can still conclude that in the seventh century. for 3. But. 482. \II. I A tnnslation with excellent commentary is ginn by Burgess and Whitney.g. they wrote "eyes". e.di .. they wrote Sasi. at the latest. In place of 1. was certainly known among Hindu astronomers and computers in the 6th cen- tury. R. have the values four. The principle of the positional system. .. The invention itself must have taken place before 600. be- cause Rama had 3 brothers. who wrote in verse form. moon. is the famous Surya-Siddhanta 2 which has remained the standard work on 1 G." 6 (1860). etc. we must take a closer look at the number notation which these astronomers and computers used. such pro- clamations were often falsified. p. the principle that the same digit can.. For a large number.. etc.

1930. . All this makes it highly probable that around 500. this Bhaskara lived around 520..".g. Aryabhala and his syllable-numbers.. He knew the table of sines. missing places are simply omitted. the same syllables which can designate 4. the most ancient astronomer and com- puter whose date we can establish. indicate the powers of 100. e.'" A.. who lived about 537. not a positional one. to which Bhattotpala had acc. Bhaskara PuliSa-Siddhinta. they began with the units. but Thibaut (As".. . .. can also have the value 40 or 400. when Varaha·Mihira wrote his compendium in which he quotes from 5 astronomical handbooks (Siddhantas). followed by the tens.. c = 6. A zero is not needed in this system. . it has the further advantage 0 leaving the poet greater freedom in the choice of syllables and thus enabling him better to meet metrical requirements.u. According to Datta and Singh. &ci.. The vowels a. S = 3. etc. The 25 consonants from k to m have the values 1-25. the quotation ia from the origin.f ArytJhlfllta. but. Among these is also the Pulisa-Siddhanta. etc. 3 For. • See W. Phil.i•• Grundriss Indo-Ar. according to his own statement. 40. which is positional and has a zero. Conclusion: We do not know exactly when the word-numbers were first used. 100. The remaining 8 constants from y to h have the values 30. i... Like Aryabhata. I Pargiter..... 1 According to Datta and Singh (His. the astronomers and com- puters introduced the positional system..ntas. which is ascribed to the first centuries of our era by Pargiter 2. 1 These also appear in the Agni-purana.D.) says that the two puli§a-Siddh$. This work existed as early as the first half of the sixth century. introduced an imrroved system. he begins with the units. As an example. Chicago. which impresses one as more primitive than the SCirya-Siddh. but they were certainly in circulation about 500 A. Clark. according to Datta and Singh. at least in part. DIGITS AND THE ART OF COMPUTING 55 Hindu astronomy until the present day. .. The first to reverse the order (as far as we know) was Jinabhadra Gani. = 5.. to a later date by others. but in which word-numbers already occur (judging from a quotation by Bhattotpala). upon which VarU-a-Mihira could draw. p. a religiOUS compendium. u. 254. iota.-blr. he reached the age of23 in 499.1 Royal Asi. E. . were totally dil£erent from the origin. a pupil of Aryabhata.. Originally. we give in Aryabhata's notation the number 5775 33 36 (the number of revolutions of the moon in 432.Jv _heutics I). NUMBER SYSTEMS. l' etc. those with i the hun- dreds and thousands. ch = puli§a-Siddhinta. but he used a different notation for numbers.". 1902.000 years): cayagiyiiluSuchlf 63335775 The syllables which contain an a denote the units and tens. About this time lived also Aryabhata. but his system is actually positional.ory of Hi."i." .

Groningen (Noordhoff). with which we have already become acquainted. to make something out of nothing by giving it a name and inventing a symbol for it. More significantly. 13. then again retrogres· sing. Freudenthal's hypothesis reduces therefore to the fol- lowing: Before becoming subject to the Greek influ- ence.iolUll scieNCe. Could this not also be a result of the Western influence? The Babylonians and the Greeks always started with the largest units. abounds in Greek terms. center or distance from center. 19996. positional system. planetary motions systematically in terms cf epicycles and eccentrIC c. see p.rclca. If a planet (or the sun) describes an epicycle. In a later period there was a symbol for an absent digit in the interior of a number symbol. the Surya Siddhanta. says Halsted. The Babylonian sexagesimal system was imperfect because the zero was lacking. 5000 year. there is something else which indicates a foreign influence. there was no difference between 60 and 1 or 1/60 . To these arguments of Freuden-thal may be added the fact that the Hindu arithmetic books write fractions just as the later Greek papyri (such as Vindobon. is of course derived from XEvT{!O'II. . without a fraction line. from 200 to 600.. Kendra. as if it had an irregular orbit. He points out that the Hindus became acquainted with Greek astronomy during the same period. etc. It was the Greek astro~omers who completed the system by adding a 0 for zero. which connects most naturally with the language. 1 Is there any connection between the Babylonian system with the Gre~k round zero and the Indian decimal system with its identical round zero? Freudenthal thinks there is. It is a stroke of genius. The Neo-Pythagorean Iamblichus also knew the Zero. At the time the digital notation is introduced. the theory of the Surya Siddhanta is to a large extent based on the Greek theory of the epicycles. appeared in the order units. viz. lipta (minute) from le. According to Freudenthal. Freudenthal. The versified numbers.terna. the numerator above the de- nominator. it has two simultaneous circular motions. this order is suddenly reversed. the Indian astronomers quite naturally became acquainted with the sexagesimal positional system and the zero. tens. an abbreviation for ov!5iv = nothing. "It is like coining the Nirvana into dynamos". of i. the Hindus had a versified. note 27. 1 H. 2 An epicycle is a small circle whose center describes a larger circle about the earth as a center.U Fig. The great Greek astronomers A pollonius and Ptolemy exp~~d . 1946.. 49). Their standard astronomical work. On this account it appears. etc. as seen from the earth. Epicycle. as was pointed out by Freudenthal in his masterful inaugural address.. sometimes advancing. 2 Along with Greek astronomy and trigonometry. but not for its absence at the end. when the decimal positional system came into use in India.-'tTOV.56 CHAPTER II Where does the zero come from? The zero is the most important digit.

It is quite possible that things went in this way. Two of the greatest of Arab astronomers and mathematicians. became the world language of the Moslem empire. concern for safety and for welfare. to return to Mecca triumphantly shortly thereafter. . are found in exactly the same forms in the Hindu arithmetic books. Euclid. Princely stipends attract Jewish. a center of Gtaeco-Roman and of Semitic culture. on the contrary. who worshipped the stars. They had the digits 1-9 and similar symbols for 10. The pagans had to be converted. The rational operations with integers and fractions. wrote the first Arabic book on algebra. When in 622 Mohammed Hed to Medina. AI-Ma'mun.. 20. In 635 the Caliph Omar movcs his seat to rich Damascus. Later on his name . Along with Greek astronomy. called Alkhwarizmi. The prophet dies in 632.. the Caliph AI-Mansur erects the fabled city of Baghdad not far from the ruins of Seleucla and of Babylon. the Arabs conquered in less than a century the richest part of the world from the Indus to the Pyrenees. This detracts in no way from the honor due to the Hindus. Arabic. Muhammed ben Musa. because they believed in one God. In the year 145 after the Hegira (766 of our era). these are the characteristics of the Arab dominion. Syrian and Persian scholars and artists. a Siddhanta. But the ancient cultures of the subjected peoples are not destroyed. The triumph of the Hindu numerals. They amalgamated this positional system with their own. established in Baghdad an academy.orescent civilization of Islam. Tabit ben ~rra and al-Battani. the sacred language. NUMBER SYSTEMS. Barely 10 years after the founding of Baghdad. there began not only the Mohammedan era. an astronomical observatory and a library. In this library. Tolerance. which was immediately trans- lated into Arabic and became very popular under the title "Sindhind". With the numbers themselves. astronomy and astrology has been preserved in Arab literature. they are absorbed in the o. It is not surprising therefore that much of Babylonian algebra. the Persian army in 637. The law gave religious freedom to Jews and to Christians. the Hindus became acquainted with the sexagesimal system and the zero. to their own Brahmin digits 1-9. it is they" who developed the most perfect notation for numbers. they adjoined the Greek 0 and they adopted the Greek-Babylonian order. United and strong in the faith of one God. DIGITS AND THE ART OF COMPUTING 57 arranged decimally and starting with the lowest units. but also an entirely new period in the history of our culture. the nurture of the arts and the sciences. they reached us by way of the Arabs. But the works of the great Greek scientists were also zealously translated. wise government. a grandson of AI-Mansur. as they are taught in our schools. even came from the pagan sect of Sabians. a Hindu brought to the Caliph AI-Mansur an astronomical treatise. astronomy and philo- sophy. the Byzantine army is defeated in 635. Ptolemy and Aristotle became the great authorities for geometry. known to us.

Abac:ua of Gerbert. which was. He judges Hindu astronomy to be "superior to that of the Greeks and the Babylonians". The abacus of Gerbert. The East-Arabic digits are still in use to-day in Turkey. becauseotherwisuhere would be fewer places.-Khwarizmi wrote a small work on Indian calculation. from the introduction to his Algebra. discs. so that the second might be mistaken for the ftrst. and the East-Arabic (see Fig. He travelled much. but in the 12th century. and even at present the word algorithm is used to designate a process of calculation. Gerbert. carefully described by one of his pupils (see ago Fig. which is extant. In Spain he became acquainted with the Gobar numerals. they are there called "Indian digits". because they speak Greek. that the Western world became acquaint- ed with the Hindu-Arabic numerals. Let us see now. his connections and his gifts enabled him to rise to the highestecclesia- <V ® 00 tical dignity. We distinguish the West- Arabic or Gobar digits. after all. 14). nothing but a distillation of Greek astronomy! Indeed. The same aJ. his own invention. The little circle has to occupy the position. was born in Auvergne in 940. It is clear that both were derived from the Brahmi- digits. Concerning the zero. by means of which they carried out all calcul- ations. In place of the pebbles. 10. The original is lost. It is by means of this Latin opusculum. the later Pope Sylvester II. AI-Khwarizmi was not the first in Arabic culture who knew the Hindu numerals and Hindu arithmetic. Al-Khwarizmi already knew two forms of the digits. The Syrian bishop Severus S~bokht. An extreme example of overvalu- ation of Hindu astronomy. how the Hindu-Arabic numerals reached the North. He invented a new type of p. that they have attained the extreme limits of science" and who ignore "that there are others who know something". he used "Apices". the English monk Adelhard of Bath (or some one else) translated it into Latin.58 CHAPTER II was corrupted to Algorithmus. so that the place does not remain empty. then write the little circle. al-Khwarizmi writes as follows: When (in subtraction) nothing is left over. abacus. It becomes clear that Greek science was beyond the understanding of al-Khwarizmi. Of modest descent. who lived in the first century after the Hegira. In Rheims he taught mathematics and he wrote a small book on cal- culating with the abacus. 14. speaks with deep admiration of the arithmetic of the Hindus and of the nine digits.) Our digits derive from the Gobar digits which were used in Moorish Spain. and he is scornful of conceited people "who think. Arabia and Egypt. carrying numbers which have a remarkable appearance .

. Indeed on the abacus. i. ~ ~ " 5l. the abacus does not require it. The people did not want to have anything to do with them. . 15. This it was impossible to comprehend. so masterfully portrayed in Keller's Der Grune Heinrich (I. serve as multipliers." rAil Iff G1I. indeed the cypher par excellence. like the word "zany" with us and "Vocativus" with Fritz Reuter. Nevertheless it is a cypher. They were guest performers on the abacus. It is the first appearance of the 9 digits on this side of the Pyrenees. The new digits were used especially in the monasteries tor the determination of Ea. there appeared in Germany f'i 10 1'1 1. this time in the Latin translation of al-Khwarizmi's arithmetic. one operates much more easily with pebbles or sign. just as the old Jewess. In the 12th century.0 i'':J f{ an extract from the Algorithm book from o t '? 18 i'~ J-O which an extract is reproduced in Fig. In Latin the zero is called "Nulla figura" . it was by way ~ p .. dem Leyen zu gutt ulld nutz (dem die Ziffernzahle am ersten zu lernen schwer) durch die gewohnlich teutsch Zal geordnet.t. NUMBER SYSTEMS. ~ I sees the digits from 1 to 9 with their mul. all these Greek classics were first translated into Latin .e. "Algorismus- cifra" was a word of obloquy in France." The little book went through many editions. DIGITS AND THE ART OF COMPUTING 59 and look like fantastic cabbalistic signs. Some one in the 15 th century spoke of "a symbol that merely causes trouble and lack of clarity".. But when one takes a good look at them (see Fig.ter. "Ich habe disz rechenbiichlein. they pre- ferred to calcufate with the familiar Roman numerals. printed entirely in "German". The captions at the head of the "'"16 j!1J '1~ . but soon dis- appeared from the stage. It was a mistake on Gerbert's part to use the Arabic numerals on the abacus. For that matter.".. The zero is nothing.7. their real destiny was to make the abacus superfluous. sIr columns are the digits from 2 to 9 which Fig. yet it decuples the number in front of it. the reckoning master Kadel published an arithmetic. e '7 lR ?t r~ . It was particularly the zero which seemed queer to people. This book had been studied industriously in Spain ever since the 10th century. ? of Spain. Going from top to bottom. 8 'i' 16 q In 1143. 10) one recognizes the West-Arabic Gobar numerals! The zero is lacking. 6). Roman numerals. that we first became acquainted a ~ Jl f- with Euclid and Aristotle. Medieval multiplication table with Arabic numerala..1 ? t8 tiples. p.S' from the Arabic. one 93 16 fl I. As late as 1514. ~ ~? P.\ l. for the word cypher is derived from the Latin cifra = zero (the Arabic al!?ifr = the empty). the times were not ripe for them. But they had little success." 7 15.9 S I. the nine digits and the zero appear once more..

60 CHAPTER II But. and they used them in their book- keeping. Italy led the way. the city of Florence issued an ordinance which prohibited the writing of numbers in columns. in- Fig. Leonardo's book scored a great success. From the "Marguerita philooophica oo of Gregor Reisch. cluding the abacus of Gerbert. Why? The answer is supplied by a Venetian treatise on book-keeping. in the end. known as Fibonacci. there appeared an excellent arithmetic. But in 1299. In 1202. as in every domain. The large merchants realized very soon the great advantages of the new numbers. as well as the use of Hindu numerals. Leonardo employs systematically and exclusively the "ligurae Indorum" which he had learned as a boy from a Moorish teacher. the Liber Abaci of Leonardo of Pisa. 16. Pytugoru and Boethiua. Even in . as inferior in comparison with the "modus In- dorum". Here. He considers all other methods of calculation. it is so easy to change a 0 to a 6 or a 9. It was not so easy to falsify Roman numerals. the Hindu numerals were bound to win out.

somewhat showy in character and not attaining the pure beauty of the Attic vases of the 5th century. violently agitated persons. To his left and right. 9 in the text). in the National Museum in Naples. representing the tributary provinces. the king's treasurer is seen at his counting board (see also Fig. Large volutecrater (height 1. against the Greeks.30 m) from Canosa. addressing the king. In the central may be seen from his gesture . too heavily taxed. A bodyguard stands behind the king. The person who stands on a platform.7.) .against the dangers of the expedition.. before his famous expedition of 490 B. PLATE 7 PI.C. third century B. Darius sits in the midst of his counsillors. warns him .C. South Italian·Greek art. Below. (Photo AU"". the painting represents the council of war of King Darius.

See p.24. 78-80. in all other cases bend d have no factor in common.51. PLATE 8 PL. in decreasing order. The next 2 columns give "width" band ··diagonal" d as whole numbers. See pp. in line 15 a common factor 2. 45. Plimf. containing a list of right. In line 11 the numbers band d have a common factor 15. several columns are broken off. band d.ton Library. The heading of this column is unintel· ligible. h. New York. under the diagonal its length 42.1 0 is written. Pt. On the left. Cuneiform text from the Yale Babylonian Collection.Babylonian cuneiform text. Sb. Columbia University. Old.25. Plimpton 322. Sa. YBC 7289. A square with its diagonals. On the diagonal the ratio 1. . The last column contains only the running integers from 1 to 15. The nrst preserved column gives the ratios d'/h'.arlg cd triangle with rational sides. The side is 30 (the number is written above the left upper side).35..

the defenders of the old abacus. the first German arithmetic of Ulrich Wagner appeared in Ntirnberg.e. In these days. north of the Alps. A picture in an arithmetic of 1504 represents this quarrel in a very amusing way: On the right sits Pytha- goras. still calculated with the abacus and used Roman numerals? . i. in front of his abacus. the mayor of Frankfurt cautions the clerks to use these numerals sparingly. All these books for popular use teach "calcul- ation on lines and with numerals". NUMBER SYSTEMS. It was followed by many others. saldo. practically everybody. along with all such Italian expressions as conto. the number symbols reach us from Italy. there were violent controversies between the "abacists". i. taken to be its inventor. the use of the abacus and the use of Arabic numerals. Dame Arithmetica plays the role of the judge. It took a few more centuries. the most famous one is that of Adam Riese. lombard and bankrupt.e. which he was supposed to have invented. not to use them too much. Pythagoras' sour looks suggest that he is the loser. 500 years ago. DIGITS AND THE ART OF COMPUTING 61 1594 an Antwerp canon warns merchants not to use numerals in contractS or in drafts. Who would believe that. but the cause of the Hindu numerals is now definitely settled. In 1482. In 1494. the left shows Boethius in front of a disc with Arabic numerals. during the 15th century. and the "Algorithmici". Finally. giro.

1500-1250 Babylon under Series of astrological omens: Primitive astronomical calcul- the rule of the Cassites "Enuma Anu Enhl" ations.3000 Sumerian city states Cuneiform writing Sexagesimal system High level of culture 2800-1800 Semitification Tables for dividing and multi· plying First Babylonian dynasty Cultural 60wering Phenommal6owering of alge- 1700 Hammurabi Legislatio. CHAPTER III BABYLONIAN MATHEMATICS Chronolosical Summary. New period of 60wering of Observation of moon and End of the Assyrian empire arts and sciences planets. New Babylonian empire of the Qaldeans 580 Nebukadnezar II 540 Cyrus. founder of the Babylonian religion not af. administration bra and geometry. Dated observations of eclip- Babylon mical "Era Nabonassar" ses in Babylon. General History History of Civilization History of Science . Exten- 247 Beginning of the Arsacid sive calculation tables. king of Beginning of the astrono.alendar periods Periods of planets 333 Alexander the Great Hellenism Flowering of astronomy. Lu- 311 Beginning of the era of Birth-horoscopes Dar and planetary tables. Observations of he- liacal rising of fixed stars. Obser- of justice vations of Venus. 747. 729 The Assyrian Tiglatpi- lesar II ascends the throne of Babylon as Pulu_ 722 Sargon II Assyrian royal palaces Astronomical compendia: 700 Sanhenb Court astrologers I·NAM-GIS·HAR and mul 650 Assurbanipal ubrary of Assurbanipal APIN of BabYlonian origin. the Scleucids Revival of algebra. Nabonassar. copied in Assyria about 700 612 Destruction of Niniveh. Increased accuracy of ob- Persian empire fected servations of the Zodiac. era . 500 Darius C.

12 is the actual width. the sums (result:) 15 length 3. 15 x 12 = 3. Interpretation. from 14. The first lines formulate the problem: 2 equations with 2 unknowns.15 3.3. (given:) 27 and 3. 1 Neugebauer.30). All the text translations have been taken from the monumental work of O.30 2 + 27 = 29. Required length. BABYLONIAN MATHEMATICS 63 Babylonian Algebra. which has been added to 27. Neugebauer.30. New Haven 1935.0. I have multiplied 15 length by 12 width. Textes mathematiques babyloniens.30 x 14. Sublract 2. 14. Berlin 1935). each represented by a symbol.30 = 3.30 .30. 14. the width. The square root of 0. Moreover.30 = 14 width.12 = 3 3. I have added length and width: 27. 15 . old Babylonian (i. 113. 12 width One follows this method: 27 + 3. Mathematical Cuneiform Texts.0 area.0 area.3 = 3.30 + 0. 183 was the result). length and width.y = 183 x +y = 27.0 + 3 = 3. Then I added to the area. are the things we can dig out of the texts themselves. us and sag. We can therefore safely put the problem in the form of 2 algebraic equations: (1) xy + x . begins as follows: Length. It seems best to begin with a characteristic old-Babylonian cuneiform text. from Senkereh. thus obtaining the area. p. and by Neugebauer-Sachs.30 = 15 length 14. A 3.30. 1 Text AO 8862. The Sumerian symbols are dealt with as our algebraic symbols x and y. supplemented by Thureau-Dangin. the excess of the length over the width: 3. instead of indulging in extended preliminary considerations. 15 is 0. Mathematische Keilschrifttexte (~ellen und Studien. width and area.e.15.3. I have multiplied length and width.30 = 0.15 . .3. width. they possess the same advantages of remaining unchanged in declension. Take one half of 29 (this gives 14. the only things we know about what went on in the minds of our mathematical colleagues of Hammurabi's day. For indeed. Leiden 1938. from the Hammurabi dynasty). First example. MKT (Mathematisdl< Ke:uchrifttextt) I.3 (i.e.

64 CHAPTER III AO 8862.. 17.3 in line 6. Fig.. . in tina 6 to 29. MKT II Table 15. which occur in the foUowing translation.. copied by Neugebauer. otarting with 3. AO 8862. The ruder will have little dil6c:ulty in beling aU the numbe.

amounts indeed to application of the formulas (4). in place of the actual width y: j = y + 2. p. The transformation from (1) to (2) is indicated very succintly in the text by means of the two short lines: 27 + 3. in numbers. an analogous system. In modern algebraic symbolism. this recipe may be described as follows: The solution of the system of equations (3) xy' = P x+. This change does indeed simplify the problem. besides a system of the form (3). i. This follows a fixed recipe.y = J. Another old-Babylonian text (VAT 6598) contains. y = . The solution is as follows: x = 11' + J/2 (6) Y = 1I'-J/2 11' = V(C"7J/-"=27--:)2-+-P~. 2. the equations in x and y' are: (2) xy' = 183 + 27 = 210. The preceding sentence indicates that the author. 280). What the Babylonians do. BABYLONIAN MATHEMATICS 65 The last 4 lines of the text merely verify that the resulting numbers x = 15 and y = 12 indeed satisfy (1). wishing to simplify the pro· blem.=a IS x=Y2a+1I' (4) j=Y2a-1I' 11' = V(%a)2 . step by step. x + j = 27 + 2 = 29. Second example (MKT I. Then follows the solution of the simplified system (2). which appears again and again in other texts. in which the difference is given instead of the sum: (5) xy = P x .30 2 + 27 = 29. The first of the equations (2) is found immediately by adding the two equations (1). Where did the Babylonians get this method of solution? We can only guess at the answer.P. But we can say what they did not do: they dId not use the Arabian . But they do not give these formulas. each of which illustrates the same method of cal- culation.e. they merely give one example after another.3 = 3. has introduced a new variable y'. .

From a second bUr I have harvested 3 gur of grain.z . if they had. I gur = 5. p. in which x .0 15. But in the present . units.0 SAR.x or from x . and the second field 3 gur = 15. In the terminology of Thureau-Dangin. 323).%d. The areas of the two fields were together 30. practically more convenient. But instead. after having found x by solving a quadratic equation. usually proceeds in the same way. a late Greek writer on Arithmetica.20 more than that of the second. than with the processes of the classical Greek writers. to substitute in the first and then to solve for y. In our cuneiform texts we find the same method applied. x + .0. the SAR and the sila are the "scholar's units".w.d. he very often sets { x=%a+z or {x=z+%d .0 SAR. e.0 The Babylonians were fully able to solve the second equation for x. Indeed this is what they did in another problem of the same text. they found x in the form %a + w. they did not do this. square). while bur and gur are larger.0 sila per 1 bUr = 30.0 SAR. The first field yields 4 gur = 20. Call the unknown areas (expressed in SAR) x and y. How large were the fields? For a full understanding of the calculation which follows. As Gandz has observed.0' = . they would have determined y from a . but also to linear equations. = %a .0 sila per 30. and that I bur = 30. not only to quadratic equations. but in the cases we have just discus- sed. Frequently his methods have more in common with Babylonian algebra.g.30. Per bUr (surface unit) I have harvested 4 gur of grain. Then we have to solve two equations with 2 unknowns: 20.66 CHAPTER III method of determining one of the unknowns from one of the equations (5) and substituting in the other.0) is measured in SAR (1 SAR = 12 yds.0 820 (7) 3O. When Diophantus wants to determine two unknowns x and y. Diophantus.Ox. As we shall see in the sequel. of which the sum a or the difference d is given. and y as %a . The Arabians used to reduce every algebraic problem to one equation with one unknown. in terms of which mathematical calculations are carried out. in case (4). = z .y was given instead of x + y.= 30.. and then expresses in terms of z all further conditions which x and y have to satisfy. the Babylonians also knew how to eliminate by substitution.0 sila. and the difference in yields in terms of sila. the very remarkable old-Babylonian text VAT 8389: Third example (MKT I. one has to know that the combined area of the two fields (30. Let us take. The yield of the first field was 8.

0-5. if each of the fields had an area of 15.0 and again 15.30 for 1 SAR. BABYLONIAN MATHEMATICS 67 problem they followed a different path. So 20.0 is multiplied by 20. Then they calculate what the yield would be. the yield of the first field for 1 SAR.40 more and the second 0 .0. Subtract this 5. Thus take 15. 10. i.30.0 is the area of the first field. the text says.0. the reciprocal of 30. We do not know whether the Babylonians reasoned exactly as our teacher. since 5. if each of the fields had an area of 15.e. Hence the yield of a field of 15.40 x 15. into two parts: 15. so that 5.20. "Keep 5. the difference in yield would be 2.10 to obtain 5.50.0 SAR.0 = 10.0.0 + 5. I don't know the reciprocal of 1. the text says: 8.0 SAR.0 = 10.0 and we find "the wrong yield of grain" 0. "Subtract".10. when the Babylonians were gIven that . Everything is elaborately worked out in great detail.20.20 .30 = 2. the first would produce 0 . It has to be 8.50. the sum of the areas.10. Hence the first area must be 15. the second 10.0 and add it to the other. In the same manner one finds for the second field the "wrong yield of grain" of 0.0. the difference in yield would be 10.0 SAR.50 in mind".30 less. But it is given that the difference is -7. just as Diophantus. It is concluded that.0 = 20. At least we see that.0 and the second 15.30 = 5. which Neugebauer gives on page 334.0 that of the other. but I belIeve that their thought process is expressed better by this primitive argu- ment than by the elaborate algebraic transformation. This has to be taken 5.40 + What must I multiply by from one of the areas of 15. added to the first field and subtracted from the second. For every unit of area.40 + 0. and hence a yield of 7.0. The first is 20.50. "Keep this in mind". An elementary school teacher might explain the procedure to the children as follows: If each of the fields had an area of 15.0.30 = 1. They begin by dividing the total area into two equal parts: Divide 30.30 for 15.0 times to obtain exactly 5. says the text and then continues the calculation as follows: 0.SO has to be added.0 SAR is 0. so that the difference would be increased each time by 0.0 x 1 :10 = 5.5O? Take 5.30 = 1 .

Fourth example (MKT I. then the value of 1/1 is determined from the equation xy = (h + 1II)(1r _ III) = h2 _ 1112 = P. p. Fig.b2 • which can be solved for b2 (9) bl = a2 .b)(a + b) = a2 . Fifth example (MKT III.b)t = a2 1 lab + b2 . All of this would be totally inexplicable. the inclination a-b I1h \ b {3= .68 CHAPTER III x +y= 2h.. 14). one obtains 4{1S = (a . the product P is given as well as the sum. The formulas (] ]) (el I. e. and 112) (" . so that In this ways they could therefore derive formula (4). For example. no.b)2 = . of which are given the base a. 4{1S. they set y = h + III. 154).2 _ 2ab i b2 . 8. y=h-III and then tried to determine 1/1. unless the special product (8) were known.. The derivation of (6) is entirely analogous. 18. as soon as the formula ~ 0-111~+~=~-~ was known to them. The text actually calculates b2 by use of (9). After multiplying 2{1 by 2S. If. BM (British Museum) 85 194 requires the construction of the pro- file of a dike in the form of an isosceles trapezoid. Another problem in the same text takes b as given and computes a from (10) a2 = b2 + 4{1S.- 2Ir a and the area S = a ~ b_h. p.g. It becomes clear also from other texts that the special product (8) was known.

such as 1. It follows that the method of elImination. How may the Babylonians have obtained the solution of the quadratic equations (16) x 2 ± ax = b? I believe it lIkely that they proceeded just like the Arabians did and lIke we do. 6. c = 25.20.30 and 14. Quadratic equations (MKT III.26.45. viz. by changing the left side to a perfect square: (x± 72a)2 = b + (%a)2.30. you add to 14. then the quadratic equation (14) IS solved by use of the correct formula: (15) x = a. 15. b = 5 x 0.35. in which a= 1 + 0. The text first calculat~s the 3 coefficients a.402 = 1 . and 30 is (the side of) the square. This leads to a quadratic equation (14) ax 2 +2bx=c for x. BABYLONIAN MATHEMATICS 69 must also have been known to the Babylonians. c.30 the 0.30. b.52 = 25.25 .40 = 3.25. 15. For the old-Babylonian text BM 13901 contains the following problem: I have added the areas of my two squares: 25. formula (11) has to be used: .1 ('vac + b 2 . in the first equation. Divide 1 into two parts.40. (The side of) the second square is 'Is of that of the first plus 5 GAR. x 2 + (2/3)x = 0.30 x 0. p.402x 2 + 2.30. described above.X = 14.0. llx 2 + 7x = 6. x 2 . x 2 + X = 0. 6) The beginning of the same text exhibits simpler examples for the solution of quadratic equatIOns.15 has the root 29.40. (0. You add to 29. That is to say: (13) x2 + y2 = 25. 2. . obtained from the second equation. was used and that the formula (11) was known.0. b) and finally y = (%)x + 5 is determined. 0.30 = 0.30 which you have multiplied by itself. the coefficient (of x).4Ox + 5)2 = 0.25 Y = (2/a)x + 5 In order to substitute the value of y. The following solution is given of 2: Take 1. 7.5x + 52.

thus reducing this case to (5).y = 10. 31 problems. x and y are expressed in terms of z. the solutions are s~£plied in other texts. 9. Se1Ienth example (MKT I. 18) Our texts also contain systems of equations with 3 or more unknowns. in the same text BM 13 901.20. In other cases. In the following problem (no. Entirely analogously.4O. All this indicates that Babylonian algebra was taught systematically in the schools by the aid of sets of problems. y.40. Cun.z = 10. which would give us the difference a of x and y and the product xy = x(x± a) = b. the text G of Neuge- bauer-Sachs. not even for us! Let us look for instance at YBC 4697 (Yale Babylonian Collection). in the form (18) x =y!a +111. y=III-%d. elimination was not used. for example. Texts (p. this correction is then found from the other condition which x· and y have to satisfy.70 CHAPTER III Another possibility. x. Such eliminations were obviously like rolling off a log for the old algebraists. contains. There a re also texts which contain merely long series of problems witho ut solu- tions. we set both x and y equal to %a. of which the first 8 ar£ solved in text H and the last 10 in text J. the same text solves the system (17) x2 + yl = S = 21. plus or minus a correction. 00). y-%a-III. Math. problem 02: (21) Ij3. we find: x 2 + y2 + Z2 = 23. The method of solution is again the same. although a less probable one. instead. Sometimes.(x+y)-O.. the two unknowns were de- termined in parallel manner. 485) Sometimes. x-y-d=lO. x . p. 9). . and a quadratic equation is obtained for z.y is given instead of x + y: (19) xl+y2-S=21. xy=1O.1(x-y)2=15. no. E.g. lII=vy~-(%a)2 We have here the same idea that we met before: when x + y = a is given. Sixth example (MKT III. would be to introduce a second unknown y = x ± a.O. the solution is: (20) X =111 +%d. p. the solution of such a Babylonian problem is far from easy.

believed by Neugebauer (p. EifJhth example (MKT I. BABYLONIAN MATHEMATICS 71 Neugebauer thought at first that this would lead to a cubic equation for x or y. From these equations v2 is readily eliminated. Indeed we know that tables existed. we quote problem 22 of the text BM 85200. then (21) reduces to (23) 0. of course. p.40v 2 = 15. 23.0.40u . by using a table. By multiplying both sides by 12 the author obtains 2. as simple for cube roots as for square roots. which gave the "roots" n for numbers of the form n2(n + 1).30. But.7. without any further ado. vi = to. leads to a mixed cuibc x 2(12x + 1) = 1 . if we set (22) x = U + v. u2 .30 = 0. the Babylonians were therefore able to solve mixed cubics of the form x 2(px + 1) = V. the pure cubic 12. 1 . 193) to belong to a later date than the texts which have been discussed thus far.lab + b2? . and from this he obtains. as readily as pure quadratics or pure cubics.45. Where did he get this result? From a table.e. the use of a table is.30 is solved.12. y= U .x3 = 1 . By means of their tables. 12x = 6. no. i.o. leaving a quadratic equation for u. to show that the cube root of 1/12. In this problem. if one starts from the Babylonian idea that the unknowns are equal to one half of their sum plus or minus a correction. 204) But even cubic equations did not frighten the Babylonians.30 is 0.b)2 = a2 . Geometrical proofs of algebraic formulas? How may the Babylonians have obtained formulas like (a . The next problem. of course. (12x)2(12x + 1) = 4. v.b)(a + b) = a2 .b2 • (a + b)2 = a2 + lab + b2 (a . As a n example.

72 CHAPTER III We don't know. between them. The thought processes of the Babylonians were chiefly algebraic. Perhaps they derived them by the use of diagrams. It is given that . this becomes clear from their own terminology. a square as the area of a square.0 '1 -'2 r~ Yl = d = 20. 19. (. geometrically nonsensical. It is true that they illustrated unknown numbers by means of lines and areas.-61_ (. Just let us look at the Ninth example (MKT I. It is certain that they interpreted a product as the area of a rectangle. two parts by a line parallel to the base. Fa = LI = 7. such as are found in Euclid and in Arabic writers: b • IE] Fig. a- Fig. In the old-Babylonian text VAT 8512. 342). that the proportionality (3) Y2 : Yl = X : (h . the qllestion which the Babylonians asked. + b). Even in problems which were formulated in geometrical terms. b Hence we have three unknowns: the dividing line x and the Fig.. whose base b is given equal to 30. Neither did the Babylonians hesitate to multiply two areas. + b)1 =. . xl must hold.Y2Yr = LI. from Larsa. p." + 1M + 61. 20. heights Yl and y..-b) (. (2) 11. but they always re- mained numbers. Through the geometric exterior.y are calmly added. is divided into y. But we must guard against being led astray by the geometric terminology. the trapezoid· Fl of height Yl and the triangle FI of height y. we have the relations: (1) %11(X + h) . This is shown at once in the first example. the algebraic kernel is always visible. the following problem is proposed: A triangle. 21. viz.. in which the area xy and the segment x . . something the very careful Euclid would never do.11 = d From the diagram we see furthermore. was always to calculate something. never to con- struct or to prove something.

131).e(ys . A satisfactory explanation of the solution (4) was found by P. yJ = LI . the given base b is determined anew by subtracting Lljli from Lljli+b. If in the translation of the first Ime we follow Neugebauer (MKT II. but in lines 2-8 Thureau. The formulas for x. 39). Llj(Y2b1-xl). is c. but probably they were not found by pure algebra. . Next. which halves the area of a trapezoid. Hence we have x=z-e= VY2(a l +el )-e which explains formula (4).. 6770. An example of such a derivation was found by H. that almost all the texts merely contain problems and solu- tions. BABYLONIAN MATHEMATICS 73 The solution proceeds as follows: First x is calculated by (4) x = VY2{(Lllli + b)1 + (Lllli)l} . .Dangm. p.eli. The difference of the areas of these parts is (P1 . The solutions are given in the form of recipes. old-Babylonian from Uruk is a "lesson-text". b given by the formula Fig. intcrprctJlIlJ1I lwrl'. c If we want to make this difference zero. p. we have to put c = Lljli. Huber (Isis 46. Now the Babylonians knew that the line z parallel to the sides a and c. thus obtaining a trapezoid with parallel sides c and a = c + b. p. 22 Zl = Y2(a l + eI) (see MKT I. without saying how they were obtained. But these recipes must have been derived in some way and the teachers must certainly have told their pupils how they could solve equations. (6) YI = Y1 + li. A lesson-text (MKT II. Frcudrnthal for his pnl1lission ttl puhli . c Suppose we add a rectangle having the same height Y1 + )'2 to the triangle. divides the trapezoid into two parts. and how they could express one unknown in terms of the others. p. prolonged.O. but no derivations. we 1 lowe thanks to my friend Profc:!!o~or H. 104).LIlli. The heights Y1 and Y2 are found from (5) Y1 = (b-x) .PI) . are algebraically correct. Freudenthal in a text which had remained unintelligible to Neugebauer and Thureau-Dangin. 1 A. The line x. 39) It is to be regretted. The idea of this solution IS geometrical. not algebraic. h hi. Y1 and y.

But we must recognize that formula (7) is mathematically equivalent to (8). because u in the first line. and because iku (a measure of area) is translated as if the text had asa (area). Whether he expresses himself in a grammatically correct manner or not. M. It is usually so in mathematical texts. if Thureau-Dangin's translation of lines 2-8 is correct. NOIIKUes JA. width and diagonal times length. The product from the (square of the length. Palais Jc D~~ouvcrtc.2FJ. I can not judge how much importance is to be attributed to these philological objections. Bruins. that the meaning of the words has to be determined in relation to the mathematical content. we can not get rid of the fact that the sum of the length and width. the formula is d = %[(1 + b + d)2 . From this you subtract 1. 19). A free translation of BM 34568 gives: Length. area and diagonal of the rectangle respectively. in the interpretation of a text. There is another example of a general rule for calculation in Neugebauer's MKT III (p. The result is the diagonal. is equal to the area. although it can mean "and". the leading investigators give precedence to mathematical correctness. entirely in abstract form. The product you take twice. What reo mains times a half take. You form the reciprocal. let them be equal. With the product that you have taken You multiply and The width it gives you.x x-I Dr. Bruins 1 rejects this interpretation on philological grounds. lIniv. width and diagonal take. Conf~rC'nccs. F and d designate length. If I. You in your procedure. then (7) must also be accepted as giving the factually correct interpretation of the writer's intention. Je Paris. width and) diagonal you subtract. If this translation is taken as the point of departure. The words in parentheses do not occur in the text. (The reciprocal of) length. which the author had in mind. width and diagonal you must multiply by the half. formula (8) is stated in the text. b. . 11 (1951).9lltS bab. The following lines give the solution. without numbers: I (8) y =--. Serie D. l+b+J E. width. "0. can not indicate addition."u"rlts Sut Its fIfIItWtndt.ll1l1ienJlfS. And. The area times 2 take.74 CHAPTER III obtain the following meaning for these lines: Length and width as much as area. Neugebauer and Thureau- Dangin have added them in order to obtain the correct formula. One sees again that. In the first line the problem is formulated: (7) x + y = xy.

the perimeter 61." was used. A rounding off like this occurs not infrequently. Babylonian Geometry. pp.54. 12. It has been shown that the Egyptians had a better approximation for x. (a . an operation. one finds 1. instead of producing the result 21 . By adding this 0 . It has been mentioned above that the Babylonians knew how to calculate the area of a triangle and of a trapezoid. This amount is multiplied by the height 18. the calculation is based on ( 1) which is a correct formula for the volume of a frustrum of a pyramid with square bases. whose height is given and whose bases are squares with sides a = 10 and b = 7.45. We have already discussed proportionalities related to parallel lines (see Example 8).45 as representing Ya' (a~r Thureau-Dangin considers it to be %. According to Neugebauer. which is not entirely clear. The volumes of prisms and of cylinders were determined by multiplying the area of the base and the height.. According to Thureau-Dangin however.45. but we do have them for Frustra of cones and of pyramids (MKT I. BABYLONIAN MATHEMATICS 75 I am indebted to Dr.45 to 1. Volumes and areas. that might be interpreted perhaps as 0. leads to a number .13. The first step is to show that e~ by = 1 . Bruins for calling my attention to this formula. Unfortunately we have no examples of cones and of pyramids. Neugebauer interprets 0.b = 3.30. the nonsensical formula (2) V = {e . The multiplication is wrong. .12. 176 and 187) In the old-Babylonian text BM 85 194 (Plate 9) (see also the fourth example). The same text also deals with a frustrum of a pyramid. the text gives 22. 15. b) \ a ~ b}. They took the area of a circle of radius r to be equal to 3r2. the volume of a frustrum of a cone is determined by use of the wrong formula %height x sum of the bases.b).. 15 and that a . From this. .

BM 34 568.6 =0. raised by Neugebauer. ~fter 310 B.15. The Seltucida were the successor of Alexander the Grat. Another difficulty arises viz. 23. widths and diagonals of = 2. 53) The old-Babylonian text BM 85 196 contains the following nice problem (no. 9): A patli (beam?) of length 0. The upper end has slipped down a distance 0.30--0.30 and one leg equal to 0.t. throughout one and one-half millenia. This would mean that the work is based on the formula (4) which is indeed wrong. in which occurs. is too small for the calculation of (a-:r in the text. is shown by a text from the era of the Seleucids I. using "Pythagoras" and found to be b =vd2 . of which are given the hypothe- nuse equal to d = 0. 22): A reed stands against a wall. among a number of other small problems concerning lengths.45 is an error of calcula- tion and should be replaced by e. BM 85196 and BM 85210. How far did the lower end move? The problem amounts to the consideration of a right triangle. that in two other texts. which are closely related to BM 85 194.30 (stands against a d-h wall). The "Theorem of Pythagoras" (MKT II. how high the wall? 1 i. is that the space bols. Fig.24.76 CHAPTER III An objection to his own interpretation. Both difficulties disappear if we suppose that 0.C. which precedes the number 45. If[ go down 3 yards (at the top). The other side is determined quite properly. and which contains a few illegible sym- Ys. . the wrong formula (3) V = %(a 2 + b2)h is used. but which agrees with (3). p. rom the fact. How long is the reed.6. the (lower) end slides away 9 yards. the following little exercise (MKT III.hi. The faithfulness with which the Babylonians preserved. the tradition of the theorem of Pythagoras. p.

(British Museum) . A part of this problem is treated on p. PLATE 9 PL. fourth example. 75). The volume of the frustrum is determined by multi plying the height by half the sum of the upper ~nd lower areas (see p. 9. BM 85.194. The fourth problem. illustrated by a drawing. walls. is concerned with a circular wall. There are problems on dams. The 14th problem is concerned with a frustrated cone. water·clocks and excavations. wells. 68. This side of the tablet contains 16 problems with solutions. Old Babylonian cuneiform text.


PL. 10. The "seven sages" (no/ Plato teaching geometry). A mosaic from Torre Annunziata near Pompeii. in the
National Museum in Naples; first century B.C. From the mosaic one sees how the "seven sages" were honored in
Roman times as the representatives of wisdom and science. The distribution of names among the seven figures is
uncertain. The third from the left. pointing to the celestial sphere. is identified by some as Thales. The figure to
the left of the sun·dial could be the inventor Anaximander (or Solon').
(Photo AI."",.)


The diagram is ~xactly the same. This time it is given that b = 9 and d ----'- h = 3,
while hand d are required. The solution is
J = Jf2(92 + 3 2) = 15
h = VJ2 - b2 = 12
The other problems of the same text give d and h, or d + hand b, or again
d + hand d + b, etc. The last and most complicated problem gives d + h + b
and dh. As always, the real difficulty in these problems is algebraic in cha.racter.
The only geometric property, used over and over again, is the "Theorem of
Pythagoras" .
It would be wrong to suppose that in all these problems the ratios of width,
height and diagonal are 3 : 4 : 5. One finds other ratios, such as
5: 12 : 13. 8: 15 : 17. 20: 21 : 29.
These examples of right triangles with rational sides, except the last, are also
found in the writings of Heron of Alexar.dria.
Where did the Babylonians get such examples? This question leads us to the
domain of the
Babylonian Theory of Numbers.
The Babylonians were not only virtuosos in algebra, they had also gone far in
Arithmetic. At the end of his "Mathematische Keilschrift-Texte", Neugebauer
expresses the hope that we might get to know quite a great deal of elementary
number theory, and he conjectured that we would more properly have to call
"Babylonian" many things which the Greek tradition had brought down to us as
"Pythagorean" .
ThiS hope and this conjecture have been brilliantly realized by the later dis-
covery of the text "Plimpton 322". But let us first summarize what other texts
have revealed about Babylonian Theory of Numbers.

Progressions (MKT I, p. 99)
The summing of arithmetic progressions was like rolling off a log for our
Babylonian calculators. This is shown by many little problems about the distri-
bution of a SUIT. of money among a number of brothers, according to an arithmetic
In text AO 6484, we also find the summation of a geometrical progression whose
ratio is 2:
1 + 2 + 4 + ... + 2 9 = 2 9 + (2 9 - 1).

The same te"t computes the sum of the squares of the integers from 1 to 10.
according to the formula
12 + 22 + .. + /1 2 = (I .1/3 + /1.2/3)(1 + 2 + ... + /I).


The text carries a late date, but nevertheless it resembles the old-Babylonian
texts very closely. We still find the solutions carried out step by step for concrete
numerical cases. Two problems are concerned with the Pythagorean Theorem,
four with systems of equations of the type
xy=1. x+y=a.
These texts and other similar ones indicate, that the tradition of Babylonian
algebra was carried on uninterruptedly from the days of the flowering of old-
Babylonian science well into the Hellenistic period.

Plimpton 322: Right triangles with rational sides.
We come now to the very remarkable old-Babylonian text "Plimpton 322",
discovered and published by Neugebauer and Sachs in their Mathematical
Cuneiform Texts (1945).
This text is the right-hand part of a larger tablet with several columns. The
last column contains nothing but the numbers 1, 2, ... , 15. The two preceding
columns refer, according to the legend at the head of the columns, to the
"width" and the "diagonal"; these numbers do indeed satisfy the relation
d2 _ b2 = h2.

in which the height h is always an integer, having only factors 2, 3 and 5. Un-
doubtedly these heights were found in an earlier column. The numbers are the

h b d number

2.0 1.59 2.49 1
57.36 56.7 1.20.25 2
1.20. 0 1.16,41 1.50.49 3
3.45.0 3.31.49 5. 9. 1 4
1.12 1.5 1,37 5
6.0 5.19 8. 1 6
45.0 38.11 59. 1 7
16. 0 13.19 20.49 8
10.0 8. 1 12.49 9
1.48.0 1.22.41 2.16. 1 10
1. 0 45 1.15 11
'10.0 27.59 48.49 12
4.0 2.41 4.49 13
45.0 29.31 53.49 14
1.30 56 1.46 15
Where did these numbers come from and by which principle were they
ordered? The preceding column throws some light on these questions. For it
conta.ins the squares tit jh2.


These decrease fairly steadily from nearly 2 to slightly more than 'fa: Sub-
tracting 1, we obtain
J2/h2-1 ~ (d 2 _h2)/h2 ~ b2/h2.

It is a plausible conjecture that the column preceding this one, contained the
fJ = b/h and <5 ~ d/h.

This conjecture can also be formulated by saying that, first, h was taken to be
equal to 1. The problem would then be, to construct right triangles, whose sides
1, fJ, <5 are measured by rational numbers. and such that the relation
<5 2 - fJ2 ~ 1
is satisfied. This condition can be written in the form
(<5 + fJ)(<5 - fJ) ~ 1.
This means that the sum <5 + fJ and the difference <5 - fJ are reciprocals. Setting
<5 + fJ ~ at = p/q.
we find <5 - fJ = at-1 = q/p.
so that

If now one sets
h = 2pq.
so as to obtain integers, one finds immediately

These are the well-known modern formulas for right triangles with integral
sides, also used frequently by Diophantus.
The way in which the numbers fl and <5 were derived here, is a typically Baby-
lonian procedure. For it reduces to the problem of constructing right triangles. of
which are given one leg, taken as 1. and the sum of the other two sides, fl + <5 = at.
But, we already know that this problem was formulated and solved in the text
BM 34568 (see the discussion of the "Pythagorean Theorem" above).
We find support for our hypothesis that the numbers h, b, d were obtained
from 1, fl, <5 in line 11, on which the numbers 1, fl, <5 have been preserved. If mul-
tiplied by 2pq = 4, the simpler numbers h = 4, b = 3, d = 5 would have been
P. Huber has discussed. in L'Enseignement mathematlque 3 (1957), p. 19,
the errors In Plimpton 322. He ~ucceeded In explaining the errors by assuming


that fJ was calculated from (J by the formula
(3= Vb' - L
This hypothesis also explains why a column for 152 = d21h 2 preceeds the columns
for band d.
Problems of this character are very closely related to Greek arithmetic, especi-
ally to the fart
that is traditionally ascribed to Pythagoras and the Pythagoreans.
We shal not discuss
ApplietJ nulthematics
such as compound interest, problems related to excavations and constructions, to
alloys. etc. The technical terms which are involved are rather difficult to under-
stand; moreover, these matters are of little importance for the history of mathe-

I. Algebra and Arithmetic.
The standard forms. which the Babylonians could solve with ease and to which
they tried to reduce the solution of all algebraic equations. are the following:
A. Equations with one unknoflltl. 1
(Al) ax= b
(A2) Xl= a
(M) Xl + ax = b
(A4) xl - ax = b
(AS) x3 = a
(A6) XI(X + 1) = a

B. Systems of equations with 2 unknowns.
(Bl) x + J = a. XJ b =
(B2) x - J = a. XJ b =
(B3) x + J = a. xl + J2 = b
(84) x - y = a. xl + J' = b
Furthermore. the following formulas were known:
(el) (a + b)S = a l + lab + bl
(C2) (a + b)(a - b) = at - bl
(Rl) 1 + 2 + 4 + ... + 2" = 2" + (2" - 1)
(lU) 11 + 22 + 31 + ... + nl = (1/3 + 2/3. n)(l + 2 + 3 + ... + n);
and also the summation of arithmetical progressions.
Pythagorean numbers Xl + y2 = Zll were found from the formulas
(R3) zlJ = Yz(ot+IC'j. xlJ = V(zlJ)l-l
Compare s. Gandz. Osiris J (1938). p. 405 on the absence of the third form of the quadratic eqUAtion


II. Geometry.
The Theorem of Pythagoras. .
A formula for the line z halving the area of a trapezOId.
The area of a triangle and of a trapczOld.
The area of a circle 3r2 (poor approximation).
Perimeter of a circle 6r (poor approximation).
Volume of a prism and of a cylmder.
Volume of a frustrum of a cone Y2(3R2 + 3r2)h (wrong)
Volume of a frustrum of a pyramid with square bases
(MI) Y2(a 2 + b2) • h (wrong)
It is possible that in one case the correct formula

(M2) { (a ~ bf + 1/3(a; bf}. h
was used; more probably it was the formula

(M3) {(a~bf+(a;br}·h.
which is equivalent to (Mi).

Peloponnesian war Herodotus Anaxagoras Oenopides 430 B. Migration . Heron 1 A. Thales 550 B.C. 450 B.C.C. Augustus Neo-Pythagoreans 100 A. Archimedes 240 B.D. Thucydides Theodorus 370 B. Atomists 430 B. Menaechmus Aristotle Dinostratus Eudemus Autolycus 333 B.C.C.C.D. Eudoxus Callippus Heraclides of Pontus Hicetas 350 B.D.C. The beginning of Anaximenes the Persian empire 500 B. Ptolemy Neo-Platonists: 250 A.C. Aristarchus Hellenism Stoics 250 B. Anaxagoras Pythagoreans 420 B. Diophantus Proclus 320 A.C.D.C. C.C. Pappus 400 A.C. Persian wars The Eleatics 500 B.C. Thales the New Babylo.C. Hippocrates Democritus 410 B. Eratosthenes Nicomedes 210 B.C.D.-350 B. GREEK MATHEMATICS C H A PTE R IV THE AGE OF THALES AND PYTHAGORAS Chronolosical Summary Philosophers and Mathematicians and General History Historians Astronomers 610 B.C.C.C.D. Archytas Pbto Theaetetus 370 B. Epaminondas Socrates t399 390 B.C.D. Hipparchus 60 B.C. Alexander the 300 B. Menelaus 150 A. Pythagoras man empire Anaximander Anaximander 540 B.C. Pericles 450 B.C.C. Julius Caesar 60 A. Hippasus 480 B. Apollonius 150 B.C.C. The beginning of Milesian school: 585 B. Euclid Great 280 B. Ionian revolt Heraclitus 500 B.C.

Alyattes and Croesus. Less obvious. had annexed most of the Ionian coastal towns. become evident to anyone who has looked through the first book of the Histories o! Herodotus. His successor Sanherib defeated the lonians after they had landed in Cilicia. had crumbled. but looks upon them rather as an iHlecdotal way of saymg that Oriental clements were recognized in their theories. even then they prove enough. In 709.. In the mean time. They attempted to find another route for commetclal communication with Babylon. In his posthumous dialogue Epinomis.> of the kings of Lydia and Persia? They had abundant opportunity to get well acquainted with oriental culture. whatever Greeks acquire from foreigners is finally turned by them into something nobler" (987 E). are the connections with the Assyrian empire. which had oppressed the eastern peoples for so many years (consult the Bible!). Did they not dwell near the boundaries of the great oriental empires. received presents from 7 city-kings on Cyprus. they succeeded however in establishing a trade-center there. on the Black Sea. Thales and Pythagoras. The Greek oty of Naucratls. but nevertheless unmistakable. founded during the reign of Psammetichus (663-609) even received under Amasis (569-525) a trade monopoly for the whole of Egypt. at the terminals of old trade routes with Mesopotamia.. In the middle of the 7th century. . Democritus and Eudoxus. been subject. Even if one does not accept these reports of travels as historical facts. had they not. The Greeks themselves declare unanimously that they found in Egypt and in Babylon the material for their geometry and their astronomy. politically and economically. Numerous Greeks lived m the Nile delta. Moreover. they established the commercial towns of Sinope and T rapezus. as follows: " . when Gyges. The powerful military empirp. Plato sketches very strikingly the rela- tion of the Greeks to the old cultures of the Orient. for many years. Sargon II. As if the Hellenes were so narrow-minded as not to recognize the elements of value in an alIen culture! It is certainly not accidental that the lonians were the first torch-bearers of Greek civilization. Later on. put a barrier across the Ionians' access to the hinter- land. The close ties which linked Ionia and Asia Minor. who expanded this empire to Syria. after the later kings of Lydia. THE AGE OF THALES AND PYTHAGORAS 83 Hellas and the Orient. of the Assyrians. the route through ASia Minor became again accessible to the lonians. It was a serious blow to loman commerce. the founder of the Lydian empire (680--652). Only a few of the modern philolo- gists absolutely refuse to recognize that the Greeks have taken anything essential from the Orient. In alliance with the Medes. And the relations between Hellas and Egypt are within easy reach. all of them are reported to have travelled to Egypt and to Babylonia. important political changes had taken place. This is also applicable to the exact sciences.

m tIer Himmdsk.:s also at this time the Pythagoras migrated from Samos to Croton. the Gnomon and the Polos. Many Ionians left the country. In 612 the Medes destroyed Nineveh. Ein bGbyloniscb. loco cit. had liberated Babylon from Assyrian domI- nation.3/12 . Media and the "New-Babylonian empire". CHAPTER IV N abopolassar. king of the Chaldeans. .. 40 (192"'). 1 See E.ndt. &bylonian Astronomy III. 12/12 ofthe period of daylight for various times of the year.. are also derived from Ba- bylon. 1 The statement c:oncerning the 12 hours. mentions copper and iron from Ionia. expressed in terms of astronomical time-units (beru and us). The battle at the Halys was brought to a sudden stop by the solar eclipse of 585. bearing the date 551. Alyattes of Lydia marched to meet them with Ionian soldiers. Journal of Ncar Eastern Studies 10 (1951) p. for instance the brother of the poet Alcaius. The center of gravity of the world of mathematics and philosophy moved from Ionia to Italy. It w. which had come to the aid of his opponent Croesus. 3 The political equilibrium was disturbed again. 2 The names of the signs of the Zodiac. . The Polos was probably a sundial in the form of a hemi- sphere. gives a table for the duration of 1/12 . When they attempted to advance further to the West. for example. 192. the entire hinterland became accessible to the Greeks. that they were un- willing to continue the fight. I See van der Waerden. took part in battles under his leadership. The soldiers were frightened to such an extent. which had been predicted by Thales of Milete "Day turned into night". A cunei- form text from Uruk. Since that time three great empires maintained a balance of power: Lydia. The Ionian cities. p. Weidner. There is no doubt as to the correctness of the statement. in Italy. the Phocaeans.. A peace was concluded between the contending forces. p. F. but he also nurtured the cultural development of his country. as far as it relates to the Gnomon. for another text from Nineveh. established the town of Elea. American Journal for Semitic Languages.s Komptndi. 25· • See Weidner. consisting of a vertical bar. with which the Greeks became acquainted around the year 550 through Cleostratus of Tenedos. A statement of Herodotus (II 109) shows that there was also a cultural exchange with Greece. and the 12 hours of the day came to Greece from Babylon. which throws its shadow on a horizontal disc. The powerful Chaldean Nebu- kadnezar. Greek noblemen. which was destined to play such an important part in the history of philosophy. . 198. writes Herodotus. The Gnomon is also a sun-dial. in which Cilicia and Babylonia were included. founded by the kings of Chaldea. when Cyrus subjected the entire Orient to Persian dominiltion about the year 540. had to pay heavy tributes (plate 11). Cultural and commercial communications were opened in both directions. is also correct.2/12 . for a cuneiform text exists which contains a table for the lengths of the shadows of the bar at the various times of day. in which the Greeks divided the day from sunrise to sunset. king of Babylon (604-562) was not only a great general.

Manitius. upon his request. later on. 32 (1942).. THE AGE OF THALES AND PYTHAGORAS 85 But it was not long before the Persian empire reestablished economic and cul- turallinks with the Greeks. the most important center of Ionian culture. O. which Ptolemy could use 300 years later. 1 The sculptor Telephanes of Phocia worked for Darius and for Xerxes. Neugebauer. 506 (Heiberg). 407. Soc . Pannekoek and B.C. . Qutllen undStud". And the first Ionian natural philosopher. when the Babylonians were already in pos- session of a highly developed algebra and geometry. Phil.lso A. during the era of the Seleucids.m~t. Without these carefully dated observations. lived at the court of Darius. Callisthenes. the later flowering of Babylonian theoretical astronomy. these sages were not scholars. which the Greeks had not yet acquired? The natural point at which fruitful contacts between East and West coulJ take place at the beginning of the 6th century. Kugler. would have been impossible. 35 (1930) 1. v. The Greek physician Democedes from Croton and. Hipparchus (150 B. IS Thales of Milete. W. Neugeba"cr. a pupil of Aristotle. Programm Gymnasium Hei!. Kreuz.'" Mondrrchnuns (1900) and Sterntunik unJ Sterndi. L. Babylonian observations. practically without corrections 2 From all this we see that even during the period of flowering of their own astro- nomy.. cd. Geminus dis- cusses a method of the Chaldeans for the calculation of the velocity of the moon. lawgivers and moralIsts. a Greek astronomer of the 3rd century calculates the times of rising and setting of the signs in the Babylonian manner. Sayings. the successors of Alexander the Great 2. An. 1 F. II 12. ' And in his Isagoge. p. although not entirely voluntanly. I Compare F. The Greeks also showed much interest in these observa- tions. T.ans. Babylonian stellar rituals continued to exist.. Compart O. I (1940). • Simphcii in Arist. because Greek geometry of the sphere was not yet able to solve this problem. was the flourishing commercial town of Milete on the coast of Asia Minor. the Greeks were glad to learn from the Babylonians in any respect In which the latter had advanced beyond them. The observation of the moon and the planets of the Babylonian priest-astronomers. but statesmen. were continued systematically during the Persian regime. Der Bu. Am. sent his uncle Aristotle. 193 and p. Hypsicles.phoricu •. In general. Histiaeus.) makes use of Babylonian observations and periods of the moon. Konig. Ionian artisans and artists took part in the construction of the palace of Darius. p. Might this not be applicable as well to the initial period of Greek mathematics. . J Hypsicles. the tyrant of Milete. It was they who. . Babylon. Dresden 1888. 251.. they did not interfere with the cultures and the religions of subject peoples (a reference to the Bible is again in order here)..""." in Babel I . d. Thales was the first of the "seven sages" (plates 10 and 12). Mitteilungen Vorderasiatische Gesellschaft. The great Persian kings Cyrus and Darius (plate 12) were very tolerant.gbau xu Susa. the first mathematician and also astro- nomer. who accompanied Alexander the Great to Babylon. B 4 (1938). X. De CAtlo c. p. such as the celebrated Delphic "Know thyself" were ascribed to them. Wacrden in Eude.

alien to this world. He asked for the origin of things. • R.ut he did not observe what was in front of his own feet". But the wisdom of Thales had a more universal. but also a statesman and a man of practical judgment.C. r. a more philosophical character. . it is true that this was still aristocratic. Another example of his practical wisdom is furnished by his advice to naviga- tors to depend upon Ursa Minor rather than upon Ursa Major. Thus he was not only a theoretician and a philosopher. besides Herodotus. to predict a solar eclipse? Such a feat requires the experience of more than forty years. It is a fairy tale of modern times that. How was it possible for Thales. For we know from the letters of Assyrian court astrologers of about 700 that these had predicted solar and lunar eclipses (with varying success) 1. no matter how one proceeds. Xenophanes voiced his admiration of Thales for this prediction.'. It is re- ported that he made a great deal of money in an oil speculation and that he had a new river bed made to facilitate the crossing of Croesus' army. Thales had used the "Saros". On the contrary. It is well known that he taught that everything originates in water. It is not possible for one man alone to gather this experience.d . and that a. This points in the first place to the astronomers of Mesopotamia.C. but it was no longer feudal. The prediction of Thales fits in very well with this series of predictions. laid the foundations for the new political organization of the Greek cities.t. It would be a mistake to draw from this anecdote the conclusion that Thales was a scholar. of tltt ""'S. Thompson. The conclusion is inescapable that he must have drawn upon the experience of Oriental astronomers. But he advised his fellow citizens against an alliance with Croesus. But Thales had no Greek predecessors. day was suddenly turned into night and that Thales had predicted this event to the Delians for that year.olog".86 CHAPTER IV after the collapse of the old feudal system. during the battle on the Halys. of Ninew:b otul &by/on (1900). comely Thracian slave-girl laughed at him. saying: "he wanted to know what happens in the heavens. Herodotus reports (see p. According to Diogenes Laertius. and later on he attempted to establish a federation of Ionian towns with Teos as the capital. Prediction of a solar eclipse. Plato relates that Thales fell mto a well while looking at the stars. 84) that. some sort of absent-minded professor. is the first Greek astronomer. Thus. according to all our sources. who. a "sage" in the complete ancient sense.. he was at the very center of the intense Ionian life of his time.'" . we have the older witness Xenophanes for this accomplishment. At present it is generally agreed that this event refers to the solar eclipse of 585 B.. TItt 'eport.. the I8-year period which had been known to the Babylonians since about 400 B. in making it. C.

in which a side and two angles are equal.. 275). In Cone constructs a line CE perpendicular to CA. with D and B.l. Besides several other theorems. is indeed used in the I S~C r.. in order to demonstrate the validity of his method for deter- mining the distance between two ships at sea.kiu. while the chance of a solar eclipse occurs 23Y2 months after a total lunar eclipse. i)". sachs. p..I. 3. start from the approximate relation: 51 draconitic lunar periods = 47 synodic months. Leipzig 92 (194:0). (Proclus. 4. p. 374). Then CE has the same length as AB. referred to by Eudemus. L.. The congruence theorem (4). tr.'1r (Proclus. and one extends it to a point E. He was the first to prove that a circle is divided into two equal parts by its diameter (Proclus.. the most anCient method that has been brought down to us is that of the Roman surveyor Marcus Junius Nipsius.. edited by M Steck.. indeed very crrr-+-HL. 2.'raIH5Q:1't' 1I. angles are equal. the prediction of Thales indicates that he was acquainted with Babylonian astronomy.'nt Jr. According to Eudemus. the possibility for the repetition of a lunar eclipse exists 47 months after a total lunar eclipse.. ot'll. 341). with the remark that. Thales had to B make use of this congruence theorem (Proclus. Akad.n.+-.2 who obtained it from the History of Mathematics of Eudemus. itself unfortunately lost: 1. Wiss . paper V. he called these angles not equal. BLr. the com- mentator of the first b00k of Euclid's Elements. he had obtained the equality of the base angles in an isosceles triangle. of arbitrary length and determines its midpoint D. Schonberger. p. 409). How might Thales have determined the distance between ships at sea? According to Tannery. p. primitive.E site to that of AB. Indeed. who gives the following.1it [-'/loft>rn. . Halle 19n. rule: A In order to find the distance from A to the inaccessible point B. 24. Did Thales also know Babylonian mathematics? The following information concerning Thales is given by Proclus. According to this relation. in ancient fashIOn. 1 Whatever may actually -have happened. he discovered that when two straight lines inter- sect.iochu. was ascribed by Eudemus to Thales. but simii. The geometry of Thales. LukItJkomm. In a direction oppo. collinear Fig.LHCd bv P. P". THE AGE OF THALES AND PYTHAGORAS 87 I prefer to think that Thales. one erects in the plane a perpendicular AC to AB. as well as the ancient Babylonians. The congruence proposition concerning two triangles. . a considerable lunar eclipse could be seen 23Y2 months before the eclipse of Thales.

this proposition is related to certain calculations concerning chords and their apothems. that he would undertake to prOfle such obvious things as the equality of the parts into which a diameter divides a circle. Pamphile. Heath. We are hardly l'ustified therefore in simply ignoring his judgment that the geometry of Tha es was constructed logically in the same way as that of the later mathematicians . the eminent English historian of. says that Thales was the first to construct a circle about a right triangle and that. in reality. as well as proposition (3). Furthermore. for instance. even though they come from the best source. but also their external form. A History of Gmi . therefore Thales must have known this theorem and have formulated it explicitly. Heath. as reported by Diogenes Laertius. he even knew the terminology which Thales used for the equality of angles. the very opposite of what Eudemus explicitly says in (I)! The first objection to this view is that Eudemus not only knew the results of the mathematics of Thales. is ascribed to Thales. at least to a certain extent. It is thought that the structure of old Thales' mathematics can not have been so strictly logical. . which has just occupied the stage. 1921. stands and falls with the view that Thales stands at 1 T. that he incorrectly assumed for the mathematics of Thales the same external structure as had been given to the mathematics of his own time (about 330-300). in which every proposition was derived by strictly logical steps from previous propositions. The matter must be looked at from a broader point of view.and that evidently is his judgment because otherwise he would not have drawn the conclusion (4) . that even in Euclid this proposition is not proved. On the other hand. he sacrificed a bull. depends upon the congruence theorem mentioned in (4). definitions and axioms. also known to Thales.88 CHAPTER IV proof of this rule. have reasoned as follows: Purely logically. Thales did perhaps apply this congruence proposition without being aware of it. concerning the equality of vertical angles. 1 It has been held that the statements of Eudemus should be accounted for in this way. but that he established them empirically. Thus Eudemus might. and the statements of Proclus and of Pamphile must be studied more closely. It is thought that. but this relation is of course in no way sufficient to prove that Thales knew Babylonian mathematics.and certainly not his explicit statement that Thales had proved proposition (1). the measurement of the distance between ships at sea. which occur in Babylonian mathematics. the proposition that an angle inscribed in a semicircle is a right angle. in honor of this discovery. Some people even believe that Thales did in no way prove his discoveries. Hence. Greek mathematics. The absolute accuracy of statements (1) and (4) has been drawn into question. ob- serves in this connection. Only on the basis of better knowledge can one fairly correct an antique historian.. p. 131. It is therefore possible that this was Thales' method.. the entire criticism of the statements of Eudemus-Proclus.thematics I (Oxford).

2. of having introduced proof into geometry. the honor of having developed a logical structure for geometry. was no longer known. the base angles of an isosceles triangle are equal. partly right and partly wrong. they form part of a systematic. on the contrary. At the time of Thales. the Egyptian and the Babylonian mathematics had long been qead wisdom. This knocks out every resaon for refusing Thales credit for the proofs and for the strictly logical structure which Eudemus evidently attributes to him. from which Greek mathematics was constructed. The rules for computing could be deciphered and shown to Thales. instead of belonging to the early discoveries of mathematics. how do I divide a trapezoid into two equal parts by means of a line parallel to the bases? These are indeed the questions with which the Egyptian and the Babylonian texts are concerned. the disjecta membra can be dug out of the remnants of the old civilizations. It follows that we have to abandon the traditional belief that the oldest Greek mathematicians discovered geometry entirely by themselves and that they owed hardly anything to older cultures. is the advance by means of demonstration from theorem to theorem. at least 1200 years earlier. and it is Thales to whom it is due. the incorrect ones? Obviously. what is characteristic and absolutely new in Greek mathematics. a belief which was tenable only as long as no- thing was known about Babylonian mathematics. a diameter divides a circle into two equal parts. How was Thales to discriminate between the exact. in Babylon. shows that. This in no way diminishes the stature of Thales. The material. THE AGE OF THALES AND PYTHAGORAS 89 the beginning of ancient mathematics. etc. according to Eudemus and it is exactly at the beginning of such a logical system. one is occupied with questions such as these: how do I calculate the area of a quadrangle. but the train of thought which underlay them. the volume of a pyramid. by fitting them into a logically connected system! This is exactly what he did.)2. But . At the start. was not new. he must have discovered the theorems empirically. of the old mathematics. Evidently. The reasoning is as follows: since he was the first. logically unconnected. while the Egyptians asserted that it is (8/e . From the Babylonians he might hear that the area of a circle is 3. in the first excitement of discovery. 2. of:\ circle. the correct recipes for computation. his genius receives only now the honor that is due to it. logical exposition of mathematics.. the length of a chord. by proving them. It is only later on that the question arises: How do I prove all of this? This question becomes a central one. Indeed. and the approximate. are communi- cated to a younger generation of keenly interested foreigners. A closer look at the propositions which are ascribed to Thales. Greek geometry has had this character from the beginning. especially when the result. but. that one may expect to find such Irish bulls as: vertical angles are equal. But we know now that mathematics does not start with Thales.

Thus Leon was able to develop Elements. These men assembled in the Academy and conducted their investigations in common.) to Euclid (300 B. who came after them. Younger than Leodamas were Neoclides and his pupil Leon. aided a great deal in the development of geometry and of the other mathematical disciplines. In his commentary on the first book of Euclid's Elements (Friedlein.): Thales traveled to Egypt and brought geometry to Hellas. Sometimes he treated questions in a more general manner.. the Neo-Platonist Proclus. without concrete representation. Through his enthusiastic interest. nor doubts as to the correctness of the conclusions that have been acquired. he made many discoveries himself. pp. who lived in the same period. perfected geometry still further. Menaechmus. Hippocrates was indeed the first of whom it is recorded that he compiled Elements. transformed this science into a free form of education. 65-68). who brought a late Indian summer to Plato's Academy. Archytas of Taras and Theaetetus of Athens. not only carried out researches in accordance with Plato's diree. And Athenaeus of Cyzicus. . gives a rapid survey of the history of geometry from Thales (600 B. Eudoxus of Cnidos. one of Plato's pupils and led by him to an intere<. joined three more to the three mean proportionals and continued the researches on the section. the thinking that does not tolerate obscurities.D. who discovered the quadrature of the lunules.t in mathematic>. and he could formulate restrictions as to the possibility or impossibility of solving a given problem. of the regular polyhedra). became famous in other parts of mathematics. enlarged for the brst time the number of so·called general theorems. Hippocrates of Chios. making use of analysis. Following him Mamercus. Prom Thales to Euclid. because he put together Elements admirably and succeeded in generaliz· ing special propositions. one of Plato's friends. it is indeed well known that he filled his writings with mathematical arguments and that he used every opportunity to arouse admiration for mathematics among those who were devoting themselves to philo· sophy. in many other things he showed his successors the road to the principles.C. as well as in the other sciences. begun by Plato. but especially in geometry. also mentioned by Plato as famous mathematicians in the Rivals. who added many things to what was known before them. brother of the poet Stesichorus. Hermotimus of Colophon continued the investigations. who came after him.90 CHAPTER IV the style in which the edifice was erected. Later on. and his brother Dinostratus. and Hippias of Elis says that he acquired reputation as a geometer. he examined this discipline from its first principles and he endeavoured to study the propositions. Plato. it bears WItness to the clear thinking of the Greeks. a pupil of Eudoxus and a member of Plato's circle. and so did the somewhat younger Oenopides of Chios. After him Anaxasoras of Clazomenae dealt with many questions in geometry. he discovered many of the propositions of the Elements and developed a part of the theory of geometrical loci. Theudius of Magnesia was reputed to be excellent in mathematics.C. about 450 A. Pythasoras. occupied himself with geometry. a little younger than Leon. became famous geo· meters.e. sometimes in a more intuitive way. and Theodorus of Cyrene. In his time lived also Leodamas of Thasos. Amyclas of Heraclea. was new. better prepared as to the number of propositions and the use of proved propositions. He also discovered the theory of irrationals (or of proportions) and the construction of the cosmic solids (i. on terms of friendship with Plato's circle. by purely logical thinking. started by Eudoxus and Theaete· tus. Philippus of Mende. who increased the number of theorems and arranged them in a more scientific system.

. and supplied irrefutable proofs of the things which had not been proved strictly by his predecessors. Ptolemy asked him whether there was not a shorter route through Geometry than by way of the Elements. But unfortunately. and that he replied. exc:ept Eudemus. the pupil of Aristotle. This man lived in the days of Ptolemy I. and older than Eratosthenes and Archimedes. not much is said about these mathematicians. the value of these statements is very doubtful. mentions Euclid. the Catalogue serves a special interest. who is anxious to give fullest possible recognition to the merits of Plato and of his school. what it tells about Plato is not new. The statements about Pythagoras must also be viewed in that light. This "Catalogue of geometers" is obviously largely an extract from Eudemus' History of Mathematics. What is said about Plato and his pupil Philippus of Mende. does certainly not come from Eudemus. must stem from Eudemus and therefore ments our confidence. as a whole. that there is no royal road to geometry. shortly after Philippus of Mende and before Euclid? Eudemus is an excellent source. completed many of the results of Theaetetus. did not satisfy him and that he went to other. according to Archime- des. For who could have been meant. by "those who recorded events". Not much younger than these men is Euclid who composed the Elements. It is possible that what Proclus (or whoever wrote the Catalogue before him) found in Eudemus concerning Pythagoras. we can say: What the Catalogue says about Pythagoras is un- reliable. Democritlls is not mentioned at all. who was not as fervent an admirer of Plato as the Neo-Platonist Proclus. Unfo:tunately. he found the volume of the pyramid). THE AGE OF THALES AND PYTHAGORAS 91 tions. In the absence of knowledge of the sources. to the Neo-Pythagorean Iamblichus. We must therefore try to find better sources of information concerning the development of geometry after Thales. in order to be sure not to undervalue the famous and honored Pythagoras. the excerpt contains im- portant additions which can not have been taken from Eudemus. more doubt- ful sources (e. It is also strikmg that the entire catalogue bears the stamp of a Platonist. This is probably due to the fact that Plato considered the influence of Democritus pernicious and would best have liked to burn his works. who were not also philosophers or legendary figures. about Euclid. but also undertook to do things which could. contribute to Plato's philosophy. as stated somewhere by Eratosthenes. for. obviously does not come from the shop of Eudemus. This indicates that. In sUlllmary. the passage at the end. these were contemporaries. On the other hand. but the references to real mathemati- cians before Euclid. It is also reported that at one time. in his judgment.g. although he also played an important role in the develop- ment of geometry (he wrote several mathematical works and. Thus he was younger than the pupils of Plato. who came imme- diately after Ptolemy 1. in which he collected many of the discoveries of Eudoxus. For example. It is up to this point that the history of this science was carried by those who recorded events. For Archimedes. a fanciful and muddle-headed writer) for supplementary matenal.

in the neighbourhood of Archimedes.e. Dijkoterhuis.. In Amsterdam a street is named after him.g. is stronlgy advised to read the Ele- ments for himself. e. It will be seen that the manner in which these Elements have been brought together from a variety of separate fragments. F'OS_1t dtr Vo. Euclidu (Hist. The Pythagoreans. • See e.. Pythagoras of Samos.ttnl boob of Euclid's Elements.ding to a legend. The reader who wishes to become acquainted with the special character of Greek mathematics." x. 241 and 243 (1933- 37). Thaer. leads to important conclusions concerning the origin of these fragments. ' Pythagoras was also known as a performer of miracles. TIre thi. his name immediately makes one think of the famous "Theorem of Pythagoras". concerning the lunules of Hippocrates. Pythagoras himself was looked upon by his contemporaries in the very first place as a religious prophet. risen from tire dead. B 7. Did. Bib!. we are restricted to scattered notices in various later writers. Dt Element.. e.. 1926). Did. said: "Stop the beating. in which Pythagoras. for in this dog lives the soul of my friend.. the mathematical passages in the writings of Plato and Aristotle. 240. who were held up to ridicule on the stage. for example. (plate 13) It was quite different in antiquity. in Greek or in one of the excellent modern translations.. i. which will be dis- cussed when we come to Hippocrates of Chios.g . and. • See.. I and III 1929-30) give a 8ummary of the text and important commentaries.opl"". coming across a scene in which a little dog was being thrashed. were presented as superstitious..92 CHAPTER IV Concerning the sixth century. dtr=lt . I This looks more like religious ritual than like mathematics. teacher of wisdom. At the present time Pythagoras is thought of primarily as a mathematician.sov. F'OS . All kinds of wonderful tales concerning him were in circulation.. who had. In German: C. with extensive commentaries: T. 235.. there is moreover an important supplementary fragment of Eudemus. Moreover.1t dtr Vorsovlllil:tr. especially for the mathematics of the Pythagoreans and of the contemporaries of Plato. Heath.i!:. there is another fragment of Eudemus about the duplication of the cube by Ar- chytas. F'OS_1t dtr Vorsomti!:. the Elements of Euclid are an extremely important source. as filthy vegetarians. He tells a story.. Herodotus calls him "an important sophist". 3 but not as mathematicians. • H. . 1908 (2nd edition. I recognize him by his voice".1e E... pokes fun at the Pythagorean doctrine of the transmigration of souls. Ost•• ld's Klassi!:.• that the calf of one of his legs was 1 In English. Newton and Copernicus. WisstrUcluift. Relative to the fourth century. ex. Pythosoras A 1-2. He also tells that the Pythagoreans did not bury their dead in woollen clothing. Cambridge. and he relates that Zalmoxis. The much-traveled poet-singer Xenophanes. dccoJ.. wet... 1 The other sources will now be discussed in chronological order.. Did. had been a pupil of Pythagoras. furthermore. as.g." PY'hosorriSCM Sch.. For the fifth century. J. the saint of the Getae. . The two volumes in Dutch of E. 236.

sic!) and of his slave (E~po. being surrounded by the shiny black background. Italy. The archaic profuseness of ornamentation has been abandoned. Attic Amphora. The slave holds two burning torches. The vase is a splendid specimen of the severe style which dominated Greek art during the first half of the 5th cen- tury B. the last king of Lydia (560-546 B. in one of his odes.).. his life was spared by Cyrus. and the inner drawing and the contours have been indicated with black lines. Bacchilides relates that Croesus ended his own life..: Cheerful) are written on the vase.. While the light red-brown figures and pyre represent the original surface.C.) was ruined by the oncoming power of the Persians. 1l. The technique of representing figures in red began about 530 B. with which he lights the pyre. 17). from Vulci.C. According to Herodotus. (cf. the artist painted the Barnes blazing up from the pyre and also the burning torches with thin paint over red and black. Plates 7. Croesus. the king of the Persians. Paris The names of the king (KeoeO'o •.475 B. Louvre.C. but the very decorative efect of the ornaments and of the graceful folds recall the archaic art. with figures ill red. . PLATE 11 PL.C. This is the version of the saga followed by the painting on this vase. (5()(). Croesus on the pyre.

difficult to date.kings (the nine leaders of the revolt after the death of Cyrus) stand forth the symbol of Ahuramazda. (British Museum) PL. .). Roman copy from the 2nd century A.D.C). 12a. Schefold. King Darius of Persia as conqueror. Relief on the rock of Behistun (520/ 19 B. probably Thales of Milne (632-546 B. hagen). Copen. Portrait of a philosopher. It can certainly not be said that we have here an authentic portrait of the philosopher.C.C. According to K. 12b. The name Thales is merely a plausible hy- pothesis. of a famous original. It is an "image" of the sage of which the sryle indicates that is is not earlier than the 4th century B. the palm branch on the herma is an indication for Thales. because the poet Callimachus said of him: "The victory is to Thales". PLATE 12 PL. In the center above the nine conquered liar. Marble herma (Ny Carlsberg Glyptothek.

the world is plurality and it consists of contrasting elements. the order of Pythagoreans. 1926. what distinguishes the Pythagoreans from all others. P. the initiates of this order were allowed to hear the voice of the Master behind a curtain. But.these three are indissolubly united according 1 For these legends. It is harmony which restores unity to the contrasting parts and which moulds them into a cosmos. Ascetic. which later on spread from Croton to a lIlumber of Greek cities in Italy and which seems to have played an important role in the political life of these cities. he established a strict regime for his followers and founded a brotherhood of be- lievers. Orphic prophets roamed through Italy and Greece. Liege 1922. .Jur~:cs Je III le~wJe Je l'ythasort.. but Heraclitus spoke of "a lot of knowledge without intellect". namely by means of mathematics. 2 He preached the immortality of the soul. After the soul had been freed from earthly blemishes. he be- comes himself divine and immortal. For the religious phi1osorh~' of Pythagoras . 1940 . I\. lfaaditus H 10. and that he was seen at two places at the same 1930. This purification and the initiation into the mysteries of har- mony and of numbers. Paris.. BOYJIKC. Pytha- goras. A more tranquil way to attain immortality existed through initiation into the mysteries of Demeter and of Persephone in Eleusis.'r". There was a rebirth in God. God is unity. the river rose out of its bed. Dd.lIlJ his foUowt. Mathematics formed a part of their religion. ' In this penod there were many mystery-rites. THE AGE OF THALES AND PYTHAGORAS 93 of gold. The cult of Dio- nysus swept men and women into its wild ecstasy. it consists of numerical ratios. Ln S. would enable the soul to approach the Divine and thus to escape the circular chain of rebirths. but only after some years. Their doctrine proclaims that God has ordered the ufilverse by means of numbers. (Om1'. harmony and numbers . Harmony is divine. it could aspire to unity with the Divine." 1 Was Pythagoras a mathematician. a philosopher.lttc. Whosoever acquires full understanding of this number· harmony. t'ssai SliT III pvljtiquc Pyrha8uricicnne. PytluJsorean Politics in Southun Italy. . a saint or a charla- tan? He had some of the qualities of each of these. when their souls had been further purified by music and by living in purity in accordance with the regulations. and thus eternal life was attained. 3 After a testing period and after rigorous selection. 3 A. sec P. vegetarianism and common owner- ship ofgoods occur also in other sects. When he crossed a small stream. which promised their followers eternal life. For his adorers he was the personification of the highest divine wisdom. were they allowed to see him. IS the road along which they believe the elevation of the soul and the unipn with God to take place. The Pythagoreans thus have punfication and initiation in common With several 0ther mystery-rites. monastic living. 2 lJicls. Levy. \' Fritz. New York. Music. Lc wiu dtj muses chez If'S philosoph" Brees. a prophet. All such initiations began with ritual purifications.lrc especially Is. Fragmeflte Jt!r VorsukraJika. greeted him and said: "Hail.

All three are among the essential elements of the Pythagorean system of education and of its path for the elevation of the soul..A) : A = (B . Btihefte xU! . says that he journeyed to Egypt. According to Heraclides of Pontus. with indication of the sources. it was believed that Pythagoras had made extensive journeys to practically all Oriental countries. the travel records show that. • This proportionality plays an important role in the Pythagorean theory of music (ste my arucle In Hermes 78). thing was believed without any hesitation. . Oritn.I. 2 Older writers merely mention. upon an old tradition recorded by Aristotle and his pupil Aristoxenus. For this reason it is not surprising that. . The two means may be defined (in modern notation) by R=A+B H= 2AB 2 ' A + B' or (in classical form) by: R -A = B-R. that he came in contact with "Zaratas. even in antiquity. the Chaldean" (i. Mathematics and number mysticism mingle fantastically in the Pythagorean doctrine. Iambli- chus (Introductio in Nicomachi Arithm. tasies."he Phik"ophit. Hopfner. This was followed by a hypercritical tendency. is found in Th.94 CHAPTER IV to the doctrine of the Pythagoreans.) also reports that he learned from the Babylonians the "golden proportionality" A:H=R:B. Pythagoras said that "Beatitude is the know- ledge of the perfection of the numbers of the soul".H) : B. Cambyses. as well as by historians like Dicaearchus and Timaeus of Taurome- nion towards the close of the fourth century.e. which pushed all these later tales aside as being entirely unreliable fan. 3 1 A good survey of the travel accounts. (H . The story then relates that. The orator Isocrates. already in antiquity. 1 Mystery-rituals and number mysticism come from the orient. It is easy to prove the golden proportionality A : H = R : B by use of either of these definitions. in which Hand R are. in Egypt. since the work of Delatte and Rostagni. the harmonic and the arithmetic mean of A and B. according to Iambli- chus.en Oritn' 4 (1925).lischt ouuJ B. it was from this mystical doctrine that the exact science of the later Pythagoreans the late-antique tradition is again given greater credit. Here the priests initiated him into the mysteries. to a large extent. respectively. Nevertheless. It has been recognized that these accounts rest. whom many later writers copy. Pythagorean and Oriental wisdom were con- sidered as being related. The travels of Pythagoras. At all accounts. he remained there for seven years. the Persian conqueror. during which time he learned from the Magi the theory of numbers. I Frequently Pythagoras is represented wearing an oriental turban. Zarathustra) or that he received lessons from the Chaldeans. At the present time. Of how much value are these various statements? There was a time when every. made him a prisoner and carried him to Babylon. . the theory of music and the other sciences.

as an Ionian. J.orur.' Tetractys. Pythagoras and the theory of harmony. Gaudentius records the following history of this instrument: Pythagoras stretched a string over a straight edge and divided it into twelve parts. i. Pythagoras must have learned quite a few things from the Babylonians. the ratios 3 : 2 and 4 : 3 correspond to the intervals of the fifth and the fourth. become entirely understandable. 3. as well as the veneration of mystics. did he. the tone produced is an octave higher. . 3. when he had said I. 4. Pythagoras urged his followers to practice on "the mono- chord". in the ratios 2 : 1. a fifth or a fourth higher. not stand much closer to the source of this old wisdom than the Pythagoreans in Italy? Not only in pure mathematics. and proclaimed such an omnium-gatherum to his followers as divine wisdom in a prophetic manner. A very old saying runs: "What is the oracle of Delphi? The tetractys! For it is the scale of the sirens". then Heraclitus' ridicule. L. D. or "Everything is ordered by numbers". 2.e. is 10. the source and the root of eternal nature". ••• it I. 4. The numbers I. 2. 3 : 2 or 4 : 3. Pythagoras interrupted him as follows: "Do you see? What you take to be 4. Pythagoras asked some one to count. sacred numbers and geometrical calculations.c lIarmiJltj~lt'hrt JeT Pyzhop. and our oath". but also in the theory of music and in astronomy. All these •• sayings 2. he obtained tones which were an octave. 4 themselves constituted the famous "tetractys". Hcrmf's 78 (1943). it confirmed their general thesis "Everything is number". if Pytha- goras gathered into one lump. 3. v. of •he consonant intervals. such as Empedocles.nd these forms of oaths are characteristic of antiquity. Similarly. Geometrically the tetractys was represented by the "perfect triangle". all kinds of half-assimilated learning about the gods and the stars.t:~ms therefore necessary to ascribe to Pythagoras himself • • • • d!. arithmeti- cally by the "triangular number" 1 + 2 + 3 + 4 = to.crJCIl. 25. For the Pythagoreans pledged them- selves by oath to "him. I am convinced that it was Pythagoras himself who transmitted Baby- lonian scholarship. who had entrusted to our soul the • tetractys. or to 8 or 9. On his deathbed. According to Lucian. \\'J.. 1 When a string or a /lute is shortened to half its length. It was of eminent importance for the Pythagoreans to have learned that the most important consonant intervals could be obtained in this manner by ratios of the numbers I. the triangular numbers and the numerical ratios Fig. a perfect triangle. 1 Sec B. about musical scales. When he shortened the string from 12 to 6. THE AGE OF THALES AND PYTHAGORAS 95 In our further account we shall have occasion to point out a number of inst~nces of the connections between Babylonian mathematics and the science of the Pytha- goreans. 2. Did Heraclitus not speak of "a lot of knowledge without intelligence"? This con- temptuous remark cannot refer to a logically constructed theory of numbers and a geometry such as we find in the writings of the later Pythagoreans. But.

such as 12 : 9 = 8 : 6 for the fourth and 12 : 8 = 9 : 6 for the fifth. as is done in the Babylonian 1 Aristotle:. Metaphysics A 5. These things had of old played an important role among the Babylonians and the Magi. All these writers obtain the middle terms 9 and 8 as the arithmetic and the harmonic means of the extreme terms 12 and 6. 32/27 : 9/8 = 256/243. What was the empirical significance of these numbers? Apparently the Pythagor. Every magician utilizes the magic power of words and of numbers. these two being the only two numbers representing areas. the whole tone (9 : 8) and the major semi-tone or the "leimma" (256: 243).. was taken as a symbol of marriage. but inversely proportional. as the Master had taught. lucky numbers. 8. which is the double of a square. and the number 5. For 17 lies exactly halfway between 16. also merits confidence. The even numbers were called feminine. so do mysticism and number mysticism. . the "plane-number" xy be equal to the perimeter xy = 2x + 2y. for these ratios can be obtained by successive division from those for the octave (2 : 1). every superstitious person knows sacred numbers. Then one can express the unknown y in terms of x. For example.s. Pythagoras and the Theory of Numbers. 12 are found again in practically all Pythagorean and Neo-Pythagorean writers on music. they called 10 a perfect num- ber. etc. or their tensions or velocities. and the number 18. Usually the number 12 was assigned to the highest note and 6 to the lowest note. for which the perimeter (of the recto angle) equals the area. the fifth (3 : 2) and the fourth (4 : 3): 3/2: 4/3 = 9/8. taken from Plutarch (Isis and Osiris 42) has some mathematical interest: The Pythagoreans also have a horror for the number 17.. sec also Dids. as Aristotle informs us 1. and among the Pythagoreans as well. Interpretation: Let x and y be integers which measure the sides of a rectangle and let the area. The tradition which credits Pythagoras with the calculation of the intervals of the diatonic scale.ul~ n "I.:ans did not care very much whether they represented the lengths of the strings. and so forth.etlte. Pylll"'S'If~i. 9.. 4/3 : 9/8 = 32/27. the odd ones masculine. and. The most important thing was that the correct ratios of the harmonic intervals appeared. The following example. abstract concepts such as "justice" were identified with definite numbers. As magic and number magic belong together. which is a square. these numbers being not proportional to the lengths of the strings.d't· Sd. the sum of the first masculine and the first feminine number. they looked upon even and odd as the roots of all things.96 CHAPTER IV The same 4 numbers 6.. F.

x = 6. It looks to me as if Pythagoras. ] = 6 xy = 18. we shall draw chiefly on Euclid. 110). about ratios and the distinction between multiple and epimoric 2 ratios. This gives the two possibilities mentioned by Plutarch.g. In his Arithmetic Theology. 8 and 9 of the Elements).2 = 2 + x . Another source for the Pythagorean theory of numbers is found in the three arithmetical books of Euclid (Books 7.2' In order that y be an integer. but that they would be bored by prosaic proofs. In the discussion of proportionali. the Prophet. clothed his wisdom in a mysterious oracular form. like Nicomachus of Gerasa (100 A. about square. + 1 to. he makes nevertheless a much more primitive impression. and that only much later. or x . E.) and Iamblichus (300 AD. 1 Englisb translation with extensive commentary by Martin Lutber d'Ooge. The purpose of the layman's Introduction to Arithmetic of Nicomachus 1. In a very entertaining way he tells about triangular numbers. is to explain in a manner intelligible to everyone. about prime numbers. he was well aware of the fact that his readers wanted to be initiated into the mystery of numbers.A is a part of A and of 8. For this reason we use Nicomachus as our source for the theory of numbers of Pythagoras and of his immediate followers. such as 6 = 1 +2 + 3. rectangular and polygonal numbers.e. the Pythagoreans made the theory of numbers into an exact science. about geometric progressions. i. but never accompanied by proofs. xy=16. so that we must have 2 ~ 1. • The word "epimoric" is introduced as tbe Englisb equivalent of tbe Greek hnpt)(!LOa (superparticularis). But these are purely scientific. + 1) : •. when a number equals the sum of its proper divisors. Petfect numbers. New York 1926. Iamblichus expatiates broadly on the mystical and divine significance of numbers.. x-2 must be a divisor of 4.D. tpe wonderful and divine properties of numbers. Although Nicomachus lived four centuries after Euclid. or x-2=2. He knew his public. if A and 8 are multiples of 8-A. the ratio of.ies in the next Chapter (see p. .. we shall see that Archytas proved that every epimoric ratio can be reduced to the form (. He is much closer to the original number-mysticism of Pythagoras and his school. which would more- over deprive these things of a large part of their mystery. A and 8. about gnomonic numbers and spatial numbers. x = 3. xy = 18. everything is carefully and neatly proved. are in "epimoric ratio". x=4.2 = 4. 74): 2x 4 Y = x . Neo-Pythagoreans. etc. In the next chapter.) revel in this kind of number-mysticism. all illustrated by numerous examples. :Ie . there is nothing left of the mystery. is an epimoric ratio. in which we shall be concerned with the mathematics of the later Pythagoreans. THE AGE OF THALES AND PYTHAGORAS 97 lesson text AO 6770 (see p.. It was considered as something very remarkable by the Pythagoreans. if the difference 8 . ]=4. ] = 3. Two numbers.

+ 2"-1 = 2" . 2 2(23 . E. of which the Babylonians had constructed tables. hence 4 X 7 = 28 is a perfect number. Dickson. The proof makes use of the formula for the sum of a geometric progression 1 + 2 + .. The numbers mentioned by Nicomachus are 2(22 . each of which equals the sum of the proper divisors of the other: 1 + 2 + 4 + 5 + 10 + 11 + 20 + 22 + 44 + 55 + 110 = 284. pentagonal numbers.496 and 8128.. he said "a second I" and he mentioned the amicable numbers 284 and 22O.. and he considers. History of the Theory of Numbers. to the squares . 1 + 2 + 4 + 71 + 142 = 220. which is already found in Pythagorean texts. ± 1). among the parallelopiped numbers. +2"=1 is a prime number.The Babylonians also enjoyed such tricks based on the interchange of numbers... Nicomachus gives the four instances 6. then 2"p is a perfect number. 1 + 2 + 4 = 7 is a prime number. Similar interest is attached to Amicable "umbers like 220 and 284. • • •• • • • • • • • • • • • • • • • • • • • • • • • • • • • •• • • • •••• ••• triangular square rectangular pentagonal number number number Dumber Fig. When Pythagoras was asked what a friend is. square numbers . especially those of the form "1(.98 CHAPTER IV They called such numbers "perfect". 2'(26 . just as the Babylonians did..1. It is therefore very probable that Pythagoras knew this formula. The next perfect number is 2 11(218 ..2. 28. . etc. This leads us to one of the favorite topics of Nicomachus: Fisurate "umbers. abc. 26. He knew triangular numbers.1 and the cubes ~. For example.1). In Nicomachus one finds many more links with Babylonian arithmetic than in Euclid.1).1).1) and 21(27 . the largest known one is 2111(2117 . For further details. see L. I.1). rectangular numbers "("+ 1). which is proved in Euclid (Elements IX 36): When the sum 1+2+21+ . He pays especial attention. He also gives the following general rule.1).

Pythagoras has been given credit for a rule for determining numerical solutions of the indeterminate equation (5) x2 + y2 = Z2.. one concludes • • that the number of dots in a square is the sum of "gnomonic • • • • numbers".e.. Cantor. By repeating this division.. thus obtaining a square of side z = Yz(m 2 + 1). of odd numbers 1 + 3 + 5 + . 487). but nevertheless also a consequence of the Fig.. THE AGE OF THALES AND PYTHAGORAS 99 The marvellous properties of these numbers may be read in Nicomachus (in the translation of d'Ooge. Z = Yz(m 2 + 1). see p.. p.. one observes that a square may be divided into a smaller square and a carpenter's square or "gnomon'. is the following rule for the sums of one. (2) 1 + 3 + 5 + . . (3) 2 + 4 + 6 + . The area of the first square. The special case m = 3 leads to the right triangle whose sides are 3. one obtains readily a rule for the sum of cubes. They follow from the summation of simple arithmetical progressions such as (1) 1 + 2 + 3 + . of the gnomon and of the final square are equal exactly to x 2 . 97). three .1) = n2 (square number). To a square whose side is x = Yz(m2 . + n = Yzn(1I + 1) (triangular number).. An analogous formula for 12 + 22 + .. For example. found In Nicomachus: 1 = 13 3+ 5=23 7+9+11=33 From this. known to the Roman surveyors 1 (undoubtedly from a Greek source) 13 + 23 + 3 3 + . according to Vitruvius. i and Z2. so that (5) follows.1).. It is hardly thinkable that all these coincidences between the Babylonian and the 1 See M. It can be verified by means of a diagram that these values do Indeed satisfy (5). 559.2) = Yzn(3n . two. in the form x = Yz(m 2 .. y = m. to deduce (2) from a diagram... + n3 = {Yzn(n + 1)}2. Mathematik I.. p. We have already observed that the Babylonians knew a more general formula for the determination of right triangles with integral sides. successive odd numbers. • • •• Less immediate. where m is an odd integer (Proclus in Euclid I.1) (pentagonal number). + 2n = n(n + 1) (rectangular number). 5. Zl.. (4) 1 + 4 + 7 + . 4. formula for the sum of an arithmetic progression.. discover· ed by Pythagoras. + n2 we have already encountered in a Babylonian text. + (2n . i.1). + (3n .. Geschichte de. we add a gnomon of width 1.

C. But a scholium in the thirteenth book of the Elements says that the Pythagoreans knew only three regular polyhedra. his disciples conti- nued the investigation in a more systematic manner and built them into a logically consistent system.ii. 2. and saying that he sacrificed an ox in honor of the discovery". 26 (1897). pp.100 CHAPTER IV Pythagorean mathematics are only a matter of chance. the cube. That is about all we can properly say about it. and that it was Theaetetus who first discovered the other two. 5. Plutarch quotes a distich: "When Pythagoras discovered his famous figure. SiIZ. made of soapstone. Epicurum IX. It is quite possible that Pythagoras became acquainted with the theorem in Babylon. Proclus says.. of cattle especially. when he hears the name Pythagoras. But the entire story is an impossible one. Proclus' catalogue also ascribes to Pythagoras the construction of the regular polyhedra. ~a?t. 4. it is given greater credit nowadays than the catalogue. • Plutarch. Pythagoras and Geometry. 26. • See F. of the famous theo- rem about the square of the hypothenuse? Alas. It has been thought. But this ob- jection loses force now that we have found it applied even 1200 years earlier in the cuneiform texts. . I In another place. . the tetrahedron and the dodecahedron.".". dating from before 500 B. Who does not think. 4. VllI. and he says that the figure in question is either that of the square on the hypothenuse or that of the adaptation of areas. p. we may find some of them referring this theorem to Pythagoras. Qv«Jlione. because Pythagoras was strongly op- posed to the killmg and sacrificing of animals. Apparently Pythagoras has taken from the Babylonians various remarkable things about numbers and about their mystical significance. viz. The diagonals of such a pentagon form a star-pentagon. . 1 In his commentary on this proposition (Euclid I 47).i. 3 But Vitruvius is of the opinion that the bull fell victim to the discovery of the right triangle whose sides are 3. that the bull was sacrificed in connection with the problem of constructing a figure of the same area as another figure and similar to a third one. I CoUected by AUman. for which he sacrificed a bull". was found near Padua. eo". Precisely because this scholium directly contradicts the tradition which used to ascribe to Pythagoras anything that came along. An Etruscan dodecahedron. the tetrahedron and the dodecahedron. Win. 10 Euclid (1889). the proofs of the connection be- tween the two are extremely doubtful.z Non posse suaviter vivi sec. Gmk seO. that the theorem can not have been known in the days of Pythagoras.b"" &yo Akcul. very indefinitely "If we listen to those who wish to recount ancient history. The faces of a dodecahedron are regular pentagons.It is therefore quite possible that Pythagoras was acquainted with the cube. during the first stages of the development of geometry. the same Plutarch says however."S. Lindemann. As we will see in the next chapter.I'] from T". 625-768.

and the man <\ B was richly rewarded. leads to the quadratic equation x2 = a(a -x). As a matter of fact. . They contain among other things. The proportionality BF: BK = BI: BA. the statement that the numbers are produced from the unit and the inde- finite duality. Compare j. and he knew the proper motions of sun. and Plato replaced the unlimited by the indefinite duality of "large and 1 See Diogenes Laertius VIII 25. moon and planets to be opposite to the diurnal motion of the fixed stars. lying on his deathbed in a foreign country and unable to pay the man who had taken care of him until the end.x) : x = x : a. served as a distinctive mark among the Pytha- goreans.x Fig. Again we do not know whether the Pythagoreans actually constructed the star-pentagon in such a way. con· cerning this much-discussed frdgment. the symbol of health. Py'''''soreans and Elfa. is to the larger one KB = x as the larger piece is to the whole segment AB = a. But Aristotle asserts explicitly. they certainly had the knowledge for doing it. or (a . E. We know still less about The astronomy of the Pythagoreans. another point of contact between Babylon and the most ancient Pythagorean mathematics. that the shorter piece AK = a .j" (Cambridge 1948\. According to Alexander Polyhistor 1. the Pythagoreans could make use of the fact. p. By no means can we put unlimited faith in the compilations of Alexander.e. 28. a Pythago- rean did come past many years later. i. placed at the center of the cosmos. that this idea is Platonic and not Pythagorean. The Pythagoreans generated the numbers from the unit and the unlimited. so that any Pythagorean who might ever pass the house. The correctness of this statement follows at once from the similar triangles AlB and KFB. that each of these 5 lines divides every other one in mean and extreme ratio. 160. which he derives from certain "Pythagorean hypomnemata". One of them. which the Pythagoreans knew well how to solve by their method of "adaptation of areas". Little more is to be said about Pythagorean geometry. THE AGE OF THALES AND PYTHAGORAS 101 This figure. would make inquiries. advised this man to paint the star-pentagon on the outside wall of his house. we are already acquainted with the fact that the Babylonians also knew how to solve this equation. to be discussed in the next chapter. Pythagoras assumed the earth to be spherical in shape. RAven. The star-pentagon is also found on Babylonian drawings. For the construction of the star-pentago!1. Pcntagramma.

with ideas closely related to those of Pythagoras..'" A .OS . son of Naustrophus. First of these is the double·mouthed channel pierced for a hundred and fifty fathoms through the base of a high hill.. WaS therefore written after Plato and by some one who was not able to discrimi- nate between the doctrine of Pythagoras and that of Plato. F. but mathematics is not involved. Albro. Furthermore.. The Roman architect Vitruvius describes very carefully how to ~ui1d a hall of columns. Certamly not a reliable witness! Nonetheless. Herodotus (III 60) describes this work as follows: I have written thus at length of the Samians. exactly as described by Hero- dotus: 1 Kilometer in length. next to nothing about his theory of numbers. a Megarian .. In about 530 Eupalinus constructed an aqueduct. this view is already found among the Babylonians. It is possible to build beautifully and on a large scale even without mathematics . in a state of good preservation. There is one case however in which we know something about the mathematical preparation of a Greek building construction. we know something about the theory of music of Pythagoras. also speaks of the motion of the planets from West to East. The compilation on which Polyhistor draws. as be- comes clear from his terminology. The designer of this work was Eupalinus.. because they are the makers of the three greatest works to be seen in any Greek land. they found the tunnel. Summary.te.. . at the request of the powerful tyrant Polycrates straight across the lime-stone of the mountain Castro on the island of Samos. Diels. the writer was also influenced by the ideas of the Stoa. nothing at all about his geometry. who built their marvelous temples in Ionia and in Southern Italy .were also skilled in geometry? Speaking frankly: we do not know. namely in the case of The tunnel on Samos. with a deep ditch con- 1 H. 7 feet high and wide. When German archaeologists looked for antiquities on Samos in 1882. contrary to that of the fixed stars. 1 Indeed.102 CHAPTER IV small" (Aristotle. Metaphysics A 6). eight feet high and eight feet wide. a little younger than Pythagoras. still less about his astronomy and.the temple of Artemis in Ephesus was one of the seven wonders of the world .witness the accomplishments of the Romans.. Taking it all in all. wherethrough the water coming from an abundant spring is carried by its pipes to the city of Samos. and throughout the whole of its length there runs another channel twenty cubits deep and three feet wide. the statement concerning the proper motions can be credited. the whole channel is seven furlongs long. particularly since Alcmaeon. properly speaking. A sorry result! Are there no other sources of information for the development of mathematics in these ancient days? Is it not necessary that the architects..

oe~ line. . 15): "To cut through a ···i N mountain ABrLJ in a straight Hr--. Consequently the ratio BN : NLJ is known. with vertical vents for changing the air and for cleaning away rubble. Now. used by the working men to place their lamps. D. with two sights. and with niches. Mitt.bricius. by means of I \jB the dioptra. 2 Eupalinus did much better: his tunnel was essentially a straight line. Athells 9 (188-1). Then LJN can be determined from EZ. it was probably dug because the drop which had originally been planned. _ _ _ _ _ _-. He and KM. Evs. 12. similarly BN can be found from BE. The hypothenuses of these triangles 1 See E. 1 (Plate 14a) But now we come to the most important point: apparently the digging of the tunnel proceeded from both ends. to sight a right angle in a plane. p. until Fig. the result was a zigzag tunnel. . a horizontal bar mounted so as to rotate. his workers had to keep check on the direction in which the work was proceeding in a very primitive way. "He A draws an arbitrary line BE in r the plane. Heron gives for this ratio the value 5 : 1.r----A posed the following problem (no. If this takes place at M. Then he pro. the openings B and LJ of . then MLJ wil be perpendicular to KA. 165. next ZH perpendicular to O'. i.--:. met at the center with an error of about 30 feet horizontally and 10 feet vertically. F. eK and ML1 by addition and subtraction.. Platon tI rOrient. Sec J.e. which made it possible e. turned out to be too small. archiiol. the tunnel being given. THE AGE OF THALES AND PYTHAGORAS 103 taining the pipes. Bldez. and thus successively He.). ZH. Inst. The ditch has a depth of 7 feet at the upper end and of 27 feet at the lower end. 29.D.g.S EZ.. How could this be accomplished? The answer is found in the "Dioptra" of Heron of Alexandria (60 A. by means of vertical shafts from the top. twice as long as the distance between the ends. Brussels 1945. he constructs right triangles BOB and LJPll whose legs have the same ratio 5 : 1. p. 8K. KA. the ponit LJ is sighted in a right angle. A splendid accomplishment! I When king Ezechias of Judaea had a similar aqueduct constructed through the rocks near Jerusalem about the year 700. Now he moves the Z E dioptra along the line KA. the perpendicular 'I EZ. then.. K.. Heron first describes a dioptra.. The working men approaching from two directions. Now drop the perpendicular LJN from LJ on EB.

about thirty years before the construc- tion of the tunnel on Samos. among other things. "Free education" is to be understood in this connection as the kind of education which suits a free 1 See Pauly. Heron concludes..e.IOSi. u recoDltructed by Schone Pythagoras "transformed mathematics &om Huon's own description. which belonged there according to Vitru- vius. Ja l'. dioptra. wooden celestial spheres were manu- factured. into a free education". which says that Fig. the sighting instrument must have been a dioptra. R. we are frequently inclined to think of the ancient Greeks as being more primitive than they actually were." In no.. Eupalinus must also have been able to determine differences in altitude. A knowledge of these things makes it possible to form a clearer picture of scien- tific life in the sixth century than is pos- sible merely on the basis of the vague re- ports concerning Thales and Pythagoras..alenzJ1'''l''''di. It is quite likely that he did this like Heron and as we still do it: proceeding from point to point with vertical measuring rods and a horizontal sight. 1 This happened around 560. . "then the laborers will meet. article H. a vertical sundial. Huon'. supposed to be p:ectilinear. 16. There were no telescopes.Wissowa.".104 CHAPTER IV then give the directions in which the digging must take place. We can also better appreciate now the Proclus tradition. i. Antique measuring instruments. Anaximander also was commis- sioned to place a gnomon on the market place in Sparta. Heron also deals with the problem of making vertical shafts. AI'". We must remember that Anaximander of Milete. had a workshop in which. Such shafts are indeed present on Samos. to meet the tunnel..... Heron obtains a level by means of communicating tubes.. the pupil of Thales. "If the tunnel is dug in this manner".ssi"lot. 30. supplied of course with all the cleverly constructed hourlines and monthlines. The reader may think that this instru- ment is too clever for the era of Eupali- nus.

. THP AGE OF THALES AND PYTHAGORAS 105 man. and the training of an instrument maker in the workshop of Anaximander undoubtedly involved astronomy. but also for the sake of the practical applications. But Pythagoras freed mathematics from these practical applications. as a way to approach the eternal Truth. in contrast with training for a trade. such as Eupalinus. Surveyors and architects. The Ionians cultivated mathematics not only for its inherent interest. had to know something about geo- metry. The Pythagoreans pursued mathematics as a kind of religious contemplation.

a scientific one and a mystical-reli. Athens. were at a premium in wide circles. the most interesting figure in this school is Hippasus. The contrasts between democratic Athens and the military aristocracy of Sparta became more and more pronounced. This twofold aspect was bound to lead to a serious dilemma after his death. It was not until the fourth century. divine beings. The ancient religion and ethics were ridiculed by the intellectuals. the era of the widest unfurlment of power of the Athenian empire. such as that of the atomist Democritus. tossed to and fro in the whirl of the atoms. the century of Plato. Sophocles and Euripides. Anaxagoras and Democritus held that the sun and the moon were not living beings. And. democracy was introduced over the entire range of this power. this century was the period of the Enlightenment (Aufklarung). is the golden age of Hellenic culture. The sophists taught that all truth and all values are only relative. At the beginning of the fifth century. a lover of wisdom. In philosophy. and of its decline. divine revelations. . The Pythagorean doctrine ha. carrying out perfect circular motions as a result of their perfection and their divine insight. until at last the frightful Pelopon- nesian war brought about the end of the power of Athens and thus of the greatness of Hellas. Pythagoras called himself a philosopher. from which Athens and Sparta emerged as victors. which adhered for instance to the belief that the planets were animated. not a sage therefore. never equalled masterworks. but he was also a prophet whose teachings were looked upon as infallible. but a seeker after wisdom. the age also of the great tragic poets Aeschylus. that philosophy and mathematics reach:!d their zenith. which had been completely destroyed. We can follow the growth which took place in the fifth century and which was preparatory to the period of flowering. In contrast to this view stood the religious-mystical metaphysics of the Pytha- goreans. Thus the aristocratic Pythagoreans were divested of their power and driven out of Italy (about 430). Materialistic systems. It is the century during which the glorious temples of the Acropolis arose. gious one. which led to the great Persian wars. of the great historians Herodotus and Thucydides. The fifth century begins with the unsuccessful revolt of the Ionian cities in Asia Minor against the Persian domination. but inert rocks. was now splendidly rebuilt and Athenian naval power was established. when sculpture produced its ever admired. For mathe- matics. of rationalism. CHAPTER V THE GOLDEN AGE The fifth century. the age of Pericles. this growth occurs chiefly in the Pythagorean school.d two aspects.

about this we do know something. as a punishment for his sacrilege. reason should be ranged above revelation.rfllonlCs ArchytJ. in the manner of a genuine philosopher? In that case. Or should one search for the truth. as the supreme judge in all scientific questions.t hIS predecessors ignored the chromatic and diatonIC scales. He constructed the circumscribed sphere of a dodecahe- dron. he developed the theory of the harmonic mean. according to his opponents. 1 Arlstoxcnos tells us in his H. empty and half-filled. Pythagorean orgies there are still mentioned. Aristoxenes (a pupil of Aristotle. disciples of Philolaus and Eurytus. His followers called themselves "Mathematikoi". This conflict did indeed break out soon after the death of Pythagoras. concerning numbers. A third group. but we know from Ptolemy's HJ. there is something else. as is customary among sects of one religion. who was led thereby to experiment with vases. whose work will be discussed in the next chapter. 1 Another group must have remained in Italy. One of the most important and most interesting figures among these is Archytas of Taras. their mutal relations and their relations to the two old sects are not entirely clear. there were various groups of Pythagoreans. more important than such names is the development of mathematics itself in the school of the Pythagoreans. to tlie three consonant intervals. Anyone who wants to advance in science can not look upon this discipline as a secret accessible only to the initiates. It is thus inevitable that he should come into conflict with the pledge of silence.ltcd these SCJlcs. One of these groups was driven out of Italy around 430 by the advancing democracy. octave. is formed by the "so-called Pythagoreans in Italy". friend and contemporary of Plato. he discussed the theory of musical ratios with the musi- cian Lasos.umonics thJ. it would not be the only lie of which Aristoxenes is guilty. Around 350. A century after Hippasus. THE GOLDEN AGE 107 Should the divine revelations of Pythagoras. Later on he lost his life in a shipwreck. But. mention is still made of the astronomers Ecphantus and Hicetas of Syracuse who taught that the earth revolves on its axis. re- peatedly referred to by Aristotle. relying upon one's own thoughts. the double octave and the fifth beyond the octave. No names of Mathematikoi are known between Hippasus and Archytas. he added two more. the "ipse dixit" be looked upon as the source of all knowledge. for late in the 4th century. who strictly observed the sacred tenets and who faith- fully transmitted the" Akousmata". fifth and fourth. He has to take into account the results of others and should not conceal his own views. in opposition to the "Akousmatikoi". And. Hippasus made bold to add several novelties to the doctrine of Pythagoras and to communi- cate his views to others. the sacred sayings. . about 320) says that he had known the last remnant of this group. The two sects entered into vicious disputes. which appears to carryon the tradition of the "Mathematikoi". We are not certain of the truth of this.s cakul. harmony and the stars? In this case there was little sense in attempting to perfect this absolute wisdom by one's own studies. These and other similar indiscretions caused a split: Hippasus was expelled.

is: The theory of the even and the odd. they are: 21. clumsy for· mulations. Becker 2. it will also be prime to the double of it. Geometry (Geometria) and Astronomy (Astrologia). A number that results from (repeated) duplication of 2. to intesral positive numbers (quantities). is even times even and also even times odd. 34. A sum of an even number of odd numbers is even. . Four "mathemata". I 0 Becker. A sum of even numbers is even. 1 Thc Greeks rxcludc cvcn unity from the numbers because unity is not a quantity. 32. 22. Odd times even is even 29. were known to the Pythagoreans: Theory of Numbers (Arithmetica). and indeed typically Pythagorean mathematics. according to Greek usage. The oldest bit of the theory of numbers that we can reconstruct. p. Odd less even is odd. subjects of study. together with IX 36.108 CHAPTER V The Mathemata of the Pythasoreans. 1 The theory of numbers is there- fore the theory of naturai numbers.'"' SttuIim. 26. Astronomy will be discussed fully else- where. is exclusively even times even. 25. such as "if. and which constitutes. 23.. Theory of Music (Harmonica). we find a sequence of pro- positions (21-34). the word "numbers" always refers. a piece of archaic. 30. In abbreviated form. Every even number that does not belong among those mentioned in 32 and 33. Que/1m . If an odd number measure an even number. W c shall take no notice of these quibbles and we shall simply count 1 among the numbers. This compeL. B 4. A sum of an odd number of odd numbers is odd. it will also measure the half of it. Even less odd is odd. For the theory of music. according to O. If we turn to the ninth book of Euclid's Elements.. which have no connection whatever with what precedes. 27. is a number or 1 . Even less even is even. Here and in the sequel. We begin therefore with The theory of numbers. If an odd number be prime to any number. If a number has its half odd. 31. 533. 33. Odd less odd is even. it is even times odd only. I refer to my memoir in Hermes 78 (1943). 24.•.. i. 28.e. Odd times odd is odd.

(Ph. Bronze head of Pythagoras (?).". the most important Pythagorean mathematician and musicologist (see p. Naples. this would explain the turban. AI'""rl) . though so. from the Villa dei Papiri in Herculaneum.mewhat iaealized. portrait. probably from the end of the 4th century B. PLATE 13 PL. it is also possible that the head represents Archytas of Taras. Copy of a Greek original. If this be the case we bave probably a real. Aelianus relates that Pythagoras wore oriental dress.C.. 13. 149). Museo Nazionale. As Schefold points out.

he tunnel ond ditch. southern lun- nd . w. The upper line across the mountain represents the original tunnel. PLATE 14 PL.s m~de to join the.. Fabricius. 14•. ross·section P. PL. Deutsches Archaol. Castro. 14t. ground.. Athens. the lower one the ditch with greater fall. From E. Pt. Mitt. Castro on the island of Sarnos with the conduit of Eupalinos. 14b. Mt. How the nonhern !Unnel of . 1itI. Institut. the altitudes enlarged twofold. . 9 (1884).pLon. Cross-section of Mt.

2. In cular. we find an amusing allusion to the philosophy of the Pythagoreans.. + 2" is a prime number. The following dialogue takes place: "When there is an even number present or.. which asserts that numbers of the form 2"(1 + 2 + 22 + . Euclid gives a somewhat different proof." (Diels. in a different form in the work of Nicomachus of Gerasa. One grows. even unity . It becomes clear that the antithesis even· odd plays a very fundamental role in Pythagorean metaphysics. and we will be still different in future.. For the Pythagoreans. unbounded odd . so that by the same argument we are never the same. Plato always defines Arithmetica as "the theory of the even and the odd". probably to establish the connection with the Theory of Numbers developed in Books VII-IX. will the measure continue as before?" "Of course not".. "Well then. but there is no doubt that the original proof depended on the theory of the even and the odd. .gth to a cubit or wants to cut off a piece.. or perhaps even earlier). look at people. and some one wants to add a pebble or to take one away. they are constantly subject to change. Their point of view is expressed by Aristotle (Metaphysics AS) in the folle. THE GOLDEN AGE 109 The climax of the theory is proposition 36. + 2") are perfect.. .. which is found. for all I care.. that is certainly different from what has changed. 2np. Number develops from Unity . another one perhaps gets shorter. Unity consists of both . Others of the same school say that there are 10 principles. Fragmente der Vorsokratiker. In a fragment of the old comic poet Epicharmus (around 500 B. proposition 30 seems to have been introduced especially to make possible the proof that 2"p has no other divisors than 1. which they group in pairs: bounded . 2" and p. do you think that the number remains un· changed?" "God forbid'" "And when some one wants to add some lel. and at which Plato hints here and there. Epicharmus A 2). plurality male .. parti. You and I are also different people from what we were yesterday. Numbers constitute the entire heaven. But whatever is changeable in character and does not remain the same.. 2p.C. 22 . etc. even and odd are not only the fundamental concepts of arithmetic.. female.wing words: The elements of number are the even and the odd .. but indeed the basic principles of all nature.. provided p= 1 + 2 + . Propositions 32-34 are related to the classification of even numbers into even times even and even times odd. .. an odd number. Becker has shown that this proposition can be derived from 21-34 alone.

To justify this conclusion. we obtain ml = 4hz and hence nl = 2hz. was known before his time. came later than the theory of the even and the odd. On the other hand. Hence m is even. (stated more accurately: of the sides of squares of areas 3. But it is time to return to the mathematics of the Pythagoreans! The only place at which the theory of the even and the odd is applied in the Elements themselves. not in a philosophical discussion.. It is my judgment that this entire book should be attributed to the Pythagoreans before Archytas. Aristotle repeatedly alludes to this proof. if m were odd. II V. let one half of m be equal to h. Q!ite likely the first speaker owes the other some money and wants to prove by a philosophical argument. Rostagni. so that a repetition of the earlier reasoning shows that n is also even. The systematic organization of the theory of the ratios of numbers and of divisibility. but. butthe tacitly omitt: ed v'2. Theodorus of Cyrene (around 430) proved the irrationality of v'3. From AC : AB = m : n follows AO : . at the end of Book X. and m is even. The conclu- sion would then be that the number n is both even and odd.. Therefore nl is even. we will reproduce this proof in abbre- viated form. but a totally different person! Epicharmus is poking fun therefore at the disputes of the philosophers of his day. If the diagonal AC and the side AB of a square ABCD are commensurable. we would conclude from Proposition 29 that ml is also odd. from m = 2h. we must remember that this conversation takes place in a comedy. v'5. But. why should he speak in his first sentence of the antithesis even-odd? Epichar- mus lived in Sicily. that the irrationality of v'2. came from them. to v' 17. this leads to ml = 2nz. and the Pythagoreans played a big role throughout Southern Italy in this period. of the diagonal of the unit square. 1 Compa« A. found in Book VII of the Elements. so that mt is even. for.. since AO = 2A. 5. and their theory of the even and the odd gave the Pythagoreans the means to prove the irrationality of v'2. etc. .ABa = ml : nl."IIOf'IJ (1924) . i.1 For. it is therefore highly probable that the proof given in the preceding paragraph. Proportions in numbers. According to Plato. is the proof of the incommensurability of the side and the diagonal of a square.B2. it follows that n is odd. Since m and n have no common factor. Pappus states that the theory of irrationals started in the school of Pytha- goras. From this Zeuthen draws the conclusion. reduced to lowest form be m : n. It follows that m is even. let their ratio.110 CHAPTER V To appreciate this properly.e. it is necessary to take first a closer look at the Theory of Numbers of Archytas himself. that it is not he who has borrowed the money. 17). But who are the philosophers he is thinking of? Obviously the Pythagoreans. this is impossible. so the audience probably understood the allusion."" di Pi.

so many will also fall in continued proportion between the numbers which have the same ratio with the original numbers. THE GOLDEN AGE 111 In the theory of music of Archytas. This operation corresponds to our multi- plICatIOn of fractIOns. I assert now that D must be unity. The combination of two ratios can a : band c : d ca? al~vays be ob~jned by reducing them to the form at . then. 3 Two numbers are said to be in "multiple ratio" if one of them is a multiple of the other. . The sequel indicates that the equality of C and E is assumed. be and be : bd. that at the time of Archytas. then D is also a divisor of E and hence of C. Tannery and Heath rightly conclude from this. at the beginning of Euclid's "Sectio canonis". there must have existed in the school of Pytha. Proposition B is proved in the Sectio Canonis by appeal to another proposition from the Elements. D + E exceeds C by unity. which is a divisor of D + E itself and also of C. because the smallest of the numbers whose ratio equals their ratio. Between two numbers in epimoric ratio 1 there exists no mean proportional.e. 97 for the meaning of this term. Euclid's Elements VII 22). however many numbers fall between them in continued proportion. a mean proportional between the two original numbers A and B can not exist. There· fore D is unity. but this is impossible. The Latin text simply writes DE.. Furthermore. around 400. then the ratio itself is a multiple ratio. which we shall call A and B. viz.e. B. are relatively prime (cf. and the first measure the last. 2 The "combination" of t!. at the end of the proof a proposition from Book VIII is applied.' Then D + E exceeds the number C by an amount D. if one has a geometric progression of which the given numbers are the ex· treme terms). This proof quotes almost word for word a proposition from Book VII.e ratios a:b and b : c is the ratio a : c. and let C and D + E be the smallest numbers in this ratio.: VIII 8. The least numbers of those which have the same ratio with them are prime to one another. VIZ. i. For.g. The proof of Proposition A in Euclid's Sectio Canonis is essentially the same as that of Archytas. t For the sake of clarity I have written 0 + E. i. the following two propositions fiom the theory of numbers. When the "combination" of a ratio with itself2 is a "multiple ratio" 3. lt becomes evident that Archytas' theory of music presupposes quite a bit of the theory of numbers. If between two numbers there fall numbers in continued proportion with them (i. Consequently. viz. since their ratio equals that of C and D + E. Boethius says that Archytas' proof of proposition A ran as follows: Let A : B be the given epimoric ratio. and hence a mean proportional between C and D + E can not be found. If there be as many numbers as we please in continued proportion.e. 1 See footnote 2 on p. playa leading role: A.: VII 22. These two propositions are found. VIII 7. It follows that D is a common divisor of C and D + E. suppose that D is a number greater than 1 and a divisor of D + E. e. it will measure the second also.

e. For example. To find numbers in continued proportion.. the extremes of them are prime to one another.. Math. E . something similar to Books VII-IX of Euclid's Elements. AB. as many as may be prescribed.. Propositions VIII 7 and 8 both depend necessarily upon VIII 3. s« my article "Arit..F).'. AJ3I. . If as many numbers as we please in continued proportion 1 be the least of those which have the same ratio with them a. an analogous definition is given for "space numbers" A . Multiply these by A and the last also by B. in its lowest terms. . Ann. two mean proportionals. between two cubes. it is exactly those Platonic dialogues. a "treatise of some kind on the Ele- ments of Arithmetics".:b =b:c _"d . B and C· D are called similar if their "sides" are proportional. • i.ik tier p. in the Timaeus it is stated that between two squares there is one mean proportional. /31. .'Msmer". u _. we obtain then a geometric progression of 3 terms Aa. in which the largest number of Pytha- gorean ideas have been used (Timaeus and Epinomis). if. e. The solution of this problem proceeds as follows: Let the given ratio. h.. • For a mort detailed elaboration. AlB. that there are no smaller numbers . in particular to Archytas and his circle. This proposition in turn depends on VII 27 and on VIII 2. Continue in this manner. and the Epinomis speaks of similar plane and space numbers. and the least that are in a given ratio. Indeed. Multiply A and B. that contain numerous allusions ~o propositions and concepts of Book VIII.'. 130 (1948).. The logical genealogy of Archytas' propositions A and B is the following S 1 i. (Two "plane numbers" A . We have already seen that Archytas quotes VII 22 almost word for word. C and D . B . I am now going to inquire more closely what must certainly have been contained in these Pythagorean Elements of Arithmetic. This construction of geometric progressions must therefore certainly belong to the arithmetic of the Pythagoreans.". The contents of book VIII must therefore be credited to the Pythagoreans. 127.. p. be A : B. But Book VIII is based on Book VII. . /31. each by themselves and by each other..112 CHAPTER V goras some kind of "traite d'arithmetique". which have the same rati. The same pattern A B An is also found in Nicomachus with an application to the theory of music. The same is true of the rest of Book VIII.. b'.e. thus getting a progression of -4 terms: AS.

in numbers VISIOns. of VII 4-10. THE GOLDEN AGE 113 VII 1. If a number of numbers are • VIII 8. • VI I 27. ratio are relatively • VI I 26. If a is relatively prime to b. .. To find numbers in con· tinued proportion and the small- est in a given ratio. Propositions about proportions and multi- plication of numbers. Prime to c and d. I then the extre~es are relatively 1L-________ ~p~r_Im__e ____.successive di. The G. VII 33.O. If a and b are relatively VII 22. VII 3. The smallest Prime to c. . If a number of numbers are in continued proportion and are the smallest in a given ratio. The least numbers which have a bitrary number of num. H • V II 24. then it also measures the second. VIII 3. as many numbers in continued pro- portion between e and f.C. The G.~ numbers by. If a : b = e : f and if there in continued proportion and the are numbers in continued propor- tion between a and b. VII 20. • VII 20._--~ • VIII 7. To find for an ar. given ratio a : b.C. Propositions Oef. • VIII 2. several numbers of I VII 11-19. are contained an equal bers. numbers In a given vely prime to c.-* on "part" and "parts". then also a2 to b2 .. then it is itself a multiple moric ratio. then there are first measures the last. til to b3. Proportion. then ab is rela- tively prime to cd. then ab is also relati. ratio. Relatively prime numbers are the r least of those which have the same ratio as they have . no mean proportional is possible.O.. etc. 2. the least which have number of times in a and b. the same ratios as they have • VII 21.. If a and b are relatively pnme. Proposition B: If a ratio leads to a multiple ratio when it is multiplied Proposition A: Between two numbers in epi- by itself.

5). all of Book VII. 2) is also impossible. since.g. which includes detailed proofs of the most obvious things. definitions were given. such as the evenness of the sum of an even number of even numbers (IX 21). '" the same part. or the n-th part or m n-th parts. For it was customary for the Pythagoreans to prove number- theoretic propositions in a very careful. in which. or the third part. In Book VII. for example. . Strictly speaking.. again in the proof. But. and the strict proofs found in Book VII are of later date. in which even the most trivial syllogisms are worked out punctiliously. without proof. m and n being num- bers in the Greek sense. in Boethius' account of Archytas. of the irrationality of the diagonal. We shall see that. thus leaving 3) only! 1) is impossible. obviously because he would otherwise have been compelled to recast the entire I e. later on. It already occurs in Hippocrates of Chios. no matter whether one takes the m-fold.. which they needed for their developments. The last of these possibilities appears to be the most probable in view of the fact that. in sharp contrast to the muddy structure of Book VIII with its roundabout proofs. Book VII is constructed in such a strictly logical fashion that no pebble can be removed from it. such a definition makes sense only for commensurable magnitudes.g. one always obtains a rational ratio. the conclusion "if ml is even. as self-evident. finally it is seen in the elaborate proofs'of Book VIII itself. This can be seen in Archytas' elaborate proof of Proposition A. This definition is an ancient one. if they are the same part of their circles. applied to geometric magnitudes. he calls two circular segments proportional to their circles.114 CHAPTER V This analysis shows that in the proofs of VIII 2-3. But we can go further and eliminate possibilities 2) and 1) entirely. the double. '" the same parts. is involved. step-by-step manner. 3) Book VII was in apple-pie order at the time of Archytas. it can be seen in the Pythagorean theory of the even and the odd. its unnecessary repetitions and its logical errors. mentioned earlier. then m is also even" is very properly obtained by an indirect proof. Euclid retained however the old definition VII 20. or two thirds. 7-8 and of Propositions A and B. For the theory of yroportions in Book VII is developed from Definition 20: Numbers are prop"'tiona when the first is the same multiple. Def. which are applicable also to incommensurable magnitudes. to VII 27 and 33. 2) Propositions VII 21-22 and 27 were originally proved differently and Book VII represents a later recasting of these proofs. so that he could simply refer to propositions VII 21-22 and 27. how is Book VII historically related to propositions VIII 2-3. 7-8 and A-B of Archytas and his school? Three possibilities suggest themselves: 1) The school of Archytas originally accepted propositions VII 21-22 and 27. the definition of Eudoxus used in Book V of the Ele- ments (V. e. VII 22 is quoted word for word. such as a half or a third. 1 of the second that the third is of the fourth.

including the calculation with fractions.D. by means of successive subtractions. commercial calculations had of course to use them. Plato says: "For you are doubtless aware that experts in this study. cd. In the Republic (525E). in use in the Pythagorean school. always on guard lest the one should appear to be not one but a multiplicity of parts". but a piece of ancient mathematics. but rather on the theory of proportions and on the reduction of a ratio to lowest terms. but. it is certainly reduced as a body and divided into parts which are smaller than the body itself. if anyone attempts to cut up the 'one' in argument.GM. as well as that of relatively prime numbers are developed without the use of the factori- zation into prime numbers. In Book VII occur some propositions on prime numbers.ticarum. 18 . this equivalent is found in the ratio of numbers. It follows that Book VII is not a later reconstruction.. it is due to its strictly logical structure. The em- phasis is not on the divisibility properties of numbers. Theon of Smyrna explains this further 1: "When the unit is divided in the domain of visible things. Reduction of fractions to a common denominator brings the concept of the L. The entire development rests on the "Euclidean algo- rithm" for determining the G. comes the simplification of a ratio of numbers to smallest terms which is investigated theoretically in Book VII. and the L. but if you mince it up. Fractions do not occur in the official Greek mathematics before Archi- medes. they multiply. the question arises whether it is not possible to create a mathematical equivalent of the concept fraction and thus to establish a theoretical foundation for computing with fractions. p. The reason why fractions were eliminated from the theory is the theoretical indivisibility of unity. laugh at him and refuse to allow it. just as he discriminates between practical and theoretical arithmetic (Gorgias 1 Theoni. Hiller (1878).GM . but it is increased in numbers. Another argument in favor of this conclusion is that Book VII is a well-rounded whole without traces of later revisions. Plato distinguishes practical from theoretical logistics.GO. Euclid did not find anything to correct. Thus we have acquired an important insight: Book VII was a textbook on the ele- ments of the Theory of Numbers. It is not an accident that this book has been preserved. and this is also discussed in Book VII. neither could he make a change. The theory of the G. because many things take the place of one". short of tearing the whole thing apart and building it anew . It is probable that it was the calculation with fractions which led to the sett-ing up of this theory. THE GOLDEN AGE 115 book so as to adapt it to the new definition. In place of the reduction of fractions to lowest terms. but it does not contain the proposition that every number can be written as the product of primes in one and only one way. Indeed. which plays such an important role in our modern ele- mentary Theory of Numbers. on the stage. is called "lo- gistics" by the indication of its firm structure. such as occur in other books. The art of computing. in practice.C. SmyrlUl<i Expo'itio mum ""'the"". When fractions are thus thrown out of the pure theory of numbers.

calculation. We see from this that the Pythagoreans. and practical logistics.e. p.. who atrributes this solution to the old Pythagorean Thymaridas. i. 0« J. In further explanation.. in particular with fractions. counting. Apparently therefore. l Ancient arithmetic also includes something which is thought of at present as a topic in algebra. He contrasts these theoretical sciences with practical arithmetic. Tannery recalls (La geometrie grecque. the Pythagoreans lost their money.. Tannery has called attention to the following remarkable passage in Iamblichus' Pythagorica Vita: In the following manner the Pythagoreans explain how it came about that geometry was made known publicly: Through the fault of one of their number. it was decided to allow him to earn money with Geometry - thus Geometry came to be designated as 'The Tradition of Pythagoras". how systems of the form X + y = Ot(z + u). X +z = P(u + y). yj"hi. Theoretical logistics deals especially with the study of numbers in their mutual ratios. occupied them- selves with the solution of systems of equations with more than one unknown.e. The solution is of course (a 1 + . ._l) . exactly the sort of thing treated in Book VII. how much each amounts to in every individual case". i. 18. The solution of the special system of equations x + Xl -. x + u = r(Y + z) can be reduced to the form (1). LosiSlik IINI die I!. Tannery does not believe 1 For a further discussion of tbe Greek concept of number. shows furthermore.. Klein.atmItwos J" AIp6rw. Geometry. and the theory of the ratios of numbers (Books VII. while theoretical arithmetic is concerned with "the even and the odd. namely The solution of systems of equations of the first degree. Que'' ' I11III StwIj"" B 3. VIII) as a part of theo- retical logistics. Plato considers the old Pytha- gorean theory of the even and the odd (Book IX) as belonging to theoretical arith- metic.l = S (1) x+xl=al is known by the name "flower of Thymaridas". The Neo-Pythagorean Iamblichus. S X= n-2 ... 81) that the Pythagoreans had common ownership of goods.116 CHAPTER V 451 A-q. + X n . p. After this misfortune. like the Babylonians.X~ + . Di.a.. + a.

but he thinks (and very rightly) that this legend must nevertheless contain a core of truth. Therefore the sum of the three angles of a triangle is equal to two right angles. It is very likely that this proposition was related to their investigation of the regular polyhedra. two 'B' . equal in area to a second hgure and similar to a third. They cer- tainly knew the star-pentagon. or whether it has to do with the adaptation of areas. The last sentence of the quoted fragment indicates that a textbook on Geometry must have existed with the title "The Tradition of Pythagoras". the alter· nate interior angles are equal. but also the one on alternate interior angles formed by parallel lines It is probable that the Pythagoreans had a theory of regular polygons. It seems probable to me. it became quite feasible to earn money with science. Tannery supposes that the statements of Eudemus concerning the geometry of the Pythagoreans must also have been taken from this book. the square and the hexagon. wealthy people took pride in having their sons taught by the best and the most famous sophists. so that L LI AB = B r LABrandLEAr=LArB. Commentary on Euclid. Indeed.g. or perhaps with the problem of constructing a plane figure. Draw through A a line LI E / / Br. the Pythagoreans tried to increase their revenues in this manner. i. the triangle. It is only necessary to assume that these things occurred in "The Tradition of Pythagoras". there existed a great thirst for knowledge. It is therefore quite possible that. referred to. and the dodecahedron. that these three subjects were not arbitrarily thrown together. but that . the "teachers of wisdom" received excellent pay. According to Eude- K mus. doubted whether the proposition. viz. that in every triangle the angle . right angles will be equal to the sum of the three angles of the triangle. This is the proof of the Pythagoreans" (Proclus. Then L LlAB + L BAr + L rAE.e. the Pythagoreans discovered the proposition. This shows that the geometry of the Pythagoreans was constructed logically and that they knew not only the proposition concerning the angle sum of a tri- angle. the irrational) were certainly found only much later. and proved it as follows: "Let ABr be a triangle. some kind of written course of lectures from which the Pythagoreans made money. I 32).AddLBArtoboth F' 31 sides.If A E sum is equal to two right angles. This also explains how it happened that later writers attributed to Pythagoras all kinds of geometric discoveries. is the one about the hypothenuse. We recall that Plutarch. speaking of the distich "When Pythagoras dis- covered his famous figure. although some of these (e. and they also knew that there are only three regular polygons whose angles can fill the space about a point o in the plane. The sophists. THE GOLDEN AGE 117 in the original secrecy of the mathematical sciences (I do). after the middle of the 5th century. Since Br and LI E are parallel. for which he sacrified a bull". in their distress.

Euclid's proof of the proposition x2 = ab proceeds as follows: Xl = r2 . But what does Plutarch mean by the "application of areas"? This very important subject deserves separate consideration . alb + . ) = ab + DC + . + s) = ab: S2 = (r . Fig. as an auxiliary. by means of a semi-circle as illustrated in Fig. according to Eudemus. We a have here. In II 7. so to speak. 32. can be reduced to the special case of constructing a square equal in area to a given rectangle (Euclid.. Archytas uses twice. Let us start with the last one. or the source upon which he draws. instead of "the product Fig.118 CHAPTER V Plutarch. e. When one opens Book II of the Elements. II 4 corres- ponds to the formula b (a + b)Z = a2 + b2 + 2ab. the same con- struction is needed in the "application of areas". and one of them be cut into any number . they are directly connected with one another. 33. And.. I assume that this proof was taken from "The Tradition of Pythagoras"..b)2. one recognizes the analogous formula for (a . as will be seen presently. the redans'e contained by the tlllO straisht lines is equal to the redans'es contained by the uncut straisht line and each of the sesments. is applied here. the start of an algebra textbook. dres- sed up in geometrical form. II 14).[ill bed of sesments whate1ler. II 2 and 3 are special cases of this proposition. As a matter of fact. In his duplication of the cube. a b deration are always line segments.. II 1: "If there be two straisht lines. So.g. corresponds to the formula Fig. 34. The general problem of constructing a polygon similar to a given polygon and equal in area to another one. + . This amounts to the construction of the mean proportional between the base and the altitude of the rectangle. which we shall discuss later on.s)(r We see that the "Theorem of Pythagoras" r2 = xl + S2. one finds a sequence of propositions. which is. •• Geometric Alsebra". . The magnitudes under consi. 34. which are nothing but geometric formulations of algebraic rules. the construction of the mean proportional x = vi ab. got these three important propo- sitions from "The Tradition of Pythagoras". The proof can be read off immediately from Fig. a find of the Pythagoreans.. 32.

When Cantor (Geschichte der Mathematik I. the line of thought is contrary to that of the Greeks. whenever they simplify the pre- sentation. The greater part of the theory of polygons and polyhedra is based on this method. Qyite properly. occur as well. except in operations on in- tegers and on simple fractions. Similarly. which can not immediately be reformulated in the Greek terminology. 3rd or 4th edition. e. For our purpose it is not necessary to use the abbreviations O(a. consistently avoid such expressions. In words: If x and yare two mean proportionals between a and b. . Theaetetus in the 4th century. THE GOLDEN AGE 119 ab". a2 : ab = a: b. everything is translated into geometric terminology. T(a) and A(a. then the cube on a is to the cube on x as a is to b. but besides these and alternating with them. xy = ab from the continued proportion. y2 = bx. etc. and then to write x3 = axy = a2b. x!I can still be interpreted geometrically. viz. we shall put proportionalities into a modern dress. provided we take good care. the result x 2 = ay. a2 and a : b. root extraction. such arithmetic ex- pressions as multiplication. one finds numerous applications of this" algebra" The line of thought is always algebraic. the formulation geometric. thus leading to a3 : x:J = a3 : a2b = a : b. Let us now return from our digression to the geometric algebra of Book II. Zeuthen speaks in this connection of a "geometric algebra". It is proper to derive x 2 = ay. b). The Babylonians also used the terms "rectangle" for xy and "square" for XZ. if we reconvert the derivations into algebraic language and use modern notations. as the volume of a cube. But since it is indeed a translation which occurs here and the line of thought is algebraic. the entire theory of conic sections depends on it. Presently we shall make clear that this geometric algebra is the continuation of Babylonian algebra.g. not to use algebraic transformations. on the other hand.g. b) introduced by Dijksterhuis for ab. p. thus obtaining x:J = a2b. The Greeks. Archimedes and Apollonius in the 3rd are perfect virtuosos on this instrument. one speaks of "the rectangle formed by a and b". there is no danger of misrepresentation. and then x' = al y2 = a2bx. 213) derives e. but not xl. from the proportionality a: x = x :y = y : b. of "the square on a". "the rectangle formed by a and b" by the modern symbols a2 and ab. Throughout Greek mathematics. From now on we shall therefore quite coolly replace expressions such as "the square on a". and in place of a2 .

but in a remarkable double form. since Euclid avoids the subtraction of areas. if the line given in II 6 is denoted by 2b.b)(a + b) or. by (4) (-y- . and y to denote the extension itself. the redangle contained by the whole with the added straight line and the added straight line to- gether with the square on the half is equal to the square on the straight line made up of the half and the added straight line. If a straight line be biseded and a straight line be added to it in a straight line. Designating the line given in II 5 by 2a and the segment between the points of division by = (a . ~ ~=~-~~+~+~ It does actually occur. and one half of it plus the extension by a. Propositions II 5. then this proposition again leads to formula (3). for. -. x y 7 b • b y y • V • Fig. . Fig. 35. Then II 5 can be expressed by means of the formula (3) (X+y)2_ 2 -xy + (x_y)2 2' i. is used to designate the sum of the line given in II 6 and its extension.e.120 CHAPTER V As the next formula one might well expect (1) a2 . 36. But. if x. why should two propositions be given for one formula? Let us try something else! We call x and y the two unequal parts into which the line 2a is divided in II 5. we may express II 5 by formula (2). x 2. 6 are formulated as follows: II 5.)2_ (x-y- X But. If a straight line be cut into equal and unequal segments. then this proposition is expressed by the same formula (2). II 6.. the redangle contained by the unequal segments of the whole together with the square on the straight line between points of sedion is equal to the square on the half. But it can not have been the sole purpose of the two propositions to give formula (2) a geometric dress and to prove it in that way.l-_1)2 -_xy.

A Fig. if the difference is given. II 5 and II 6 are not propositions. which are quite openly formulated as problems: OZ TI------t=---nr. this can be transformed into a rectangle Xl by removing a piece (ay in the first case and by in the second) on the right side and placing it on the left side. when their sum x + y or their difference x -. y2 + ay = a2 or y(y + a) = a2 • This last equation calls for the construction of two segments y and y + a. VI 28. then II 5 is applied to determine half the difference %(x . The applications in the Elements themselves are consistent with this view.e. formed by the entire line and one of the parts is equal to the square on the other part".y. are known. 37.29. the problem is solved by use of II 6. of which the difference a and the product a2 are given. This gives again two propositions for one formula. at bottom. In the terminology of area calculations. To apply to a given line AB a parallelogram (AlI). In II 11 a line has to be divided into mean and extreme ratio. THE GOLDEN AGE 121 In both cases the proof i& the same: the difference of the squares on a and on b is a "gnomon". II 5 calls for the construction of two segments x and y of which the sum and product are given. This leads to the equation y2 = ala _ y). In the Data (84 and 85). in the Elements themselves. the problem receives the following form: "To divide a given straight line in such a way. This interpretation of 115. the way in which propositions II 5 and II 6 are applied. that. Why? What was the line of thought of the man who formulated the propositions in this way? The answer is found by following out. I.y) = b. x and y. that the rectangle. and also the area of the rectangle formed by them. are given. Since the difference is given. but solu- tions of problems. 6 as solutions of problems is raised beyond all doubt by the generalizations VI 28. equal in area to a . When the sum x + y = 2a is given. while in II 6 the difference and the product are given. We see therefore. then II 6 is used so that half the sum is found from (3). a carpenter's square. and in Euclid's other works. the question is considered how to prove that two seg- ments.

29. If the base and altitude of the required parallelogram are denoted by x and y. we have. 6 and VI 28. The solutions of problems (5) and (6).1. one finds y.Y2d. Figures 37 and 38 show clearly what is meant.v'(Y2s) 2 . In case (6). Case (5) is treated analogously.y = 20.122 CHAPTER V given rectilinear figure r. Analogously we explain VI 29. When a = Y2(x . or BE. 2 (see the end of Chapter III) and their solutions (Bl) {x+y=s {x =Y2s+v'(Y2s)2-F xy = F y = Y2s .1. by means of (3) (x.1. If it is changed into a square. lacking a parallelogram (Bll) similar to a given (parallelo- gram) . To apply to a given line AB a parallelogram (AE). indicated in the generalizations VI 28. 38. shows clearly that these are entirely analogous to II 5. must have the same area as the given polygon r. 6. In the most important applications the given parallelogram is a square. The required parallelogram Bll. x is obtained.F. y) 2.y) is added to this. In VI 28 the parallelogram All (Euclid designates parallelograms by indicating two opposite vertices) Fig. similar to a given (paral- lelogram) . is exactly the Same as that supplied by propositions II 5.~ F + at. the given area (of r) by F and the given line to which the rectangle has to be applied by 20. (B2) {x . equal in area to a rectilinear figure and in excess by a parallelogram. y) 2= xy + (x. must then be a square. x . using II 14. A comparison with the two normalized Babylonian problems B 1. VII 29. . while the parallelogram Bll on the remaining piece of the given line must be similar to the given parallelogram.y = d {x = v'F + (Y2d)2 + Y2d xy = F Y = v'F + (Y2d)2 . x + y = 20. then the side of this square is the half sum Y2(x + y). this area is now known. then the conditions in VI 28 are (5) xy = F. and in VI 29 (6) xy = F. When F and 2a are given. when it is subtracted. 29.

x(a-x) = F. are called in Greek Elleipsis and Hyperbole. the Greeks reduced them to one of the forms xIx + a) = F. If a straight line be cut into equal and unequal segments. In the first and third cases. . must also have been familiar to the Pythagoreans. so says Eudemus. the application to a line of a parallelogram of given area is called Parabole. THE GOLDEN AGE 123 Deficiency and excess. For the solution of quadratic equations. the peripathetic. an important construction (X 33--35) depends on the elliptic application of areas. In their formulation these propositions are quite similar to II 5. with excess and with defect. 29 are applied. with an excess or a deficiency of given form (e. Later on these names were carried over to the three conic sections. so that we have an application with a square in excess. it is always necessary to change a given area into a square (since a square root has to be extracted). However.g. ax_yx2 = F. We have already seen that they are dealt with twice in the Elements. in view of the fact that in both adapt- ations. which deals with the theory of irrational segments. a rectangle). Proclus has the following to say about the history of these adaptations: These things. II 10. are discoveries of the Pythagorean muse. we have two line segments. ax = F. which are then solved by means of the application of areas. 6: II 9. When the term in xl has the coefficient )' = p : q. of which the difference and the product are given. which solves this problem by use of the Theorem of Pythagoras. and again in the Data. which is lacking and which is left over. their deficiency (elleipsis) and their excess (hyperbole). this important part of geometric algebra is a discovery of the Pythagoreans. thus leading to an application with deficiency of a square. in Book X. proposition II 14.x. x and a . the square on the whole with the added straight line and the square on the added straight line 1 It will be made clear later on (in the discussion of Apollonius) what these adaptations have to do with conic sections. the parallelogram of given shape. namely the application of areas (parabole). there are two line segments. in the second case. of which the sum and the product are known. xIx-a) = F. But. It occurs a second time that there are two propositions which express the same algebraic formula. l The solution of problems (5) and (6) plays an extremely important part in Greek mathematics. moreover. )'Xl . )'Xl + xa = F. the squares on the unequal segments of the whole are double of the square on the half and of the square on the straight line between the points of section. In his commentary on Euclid I 44. the more general applications VI 28. and a straight line be added to it in a straight line. x and x ± a. If a straight iine be bisected. According to Eudemus therefore.

The fact that all four of the normalized forms of systems of equations. ~ellen und Studien B4. y) 2}. a special case is the famous "duplication of the cube". What could only be surmised before. Thus we conclude. which are called "Arithmoi paramekepipedoi" by Nicomachus (see Becker. The pure quadratic xl = F amounts to the trans- formation of a given area into a square ~I 14). the problem is: to t determine x and y when x + y and x2 + are given. Hence we are concerned with the solution of the systems of equations (7) x + y = S. y = %s . to which we shall return later on.(d/2)2 + d/2.d/2. are taken over by Euclid. The solutions. which we have found in the cuneiform texts. 181). in geo- metric form. x 2 + y2 = F.3 = V. in geometric formulation. x 2 + y2 = F.y and x2 + t.124 CHAPTER V both together are double of the square on the half and of the square described on the straight line made up of the half and the added straight line as on one straight line. 10. The ancients were concerned with this problem as well. without ex- ception. with their solu- tions. The proofs make clever use of the "Theorem of Pythagoras".(S/2)2. poses the problem of constructing a cube of given volume. For II 9. used in constructing their mathematics.(d/2)2 . Both propositions lead to the formula x2+ y2 {(x. . The pure cubic x. y) 2+ (x. without excess or deficiency. Apparently the Pythagoreans formulated and proved geome- trically the Babylonian rules for the solution of these systems. = 2 Both can be taken to be solutions of problems.required x and y. gives clear evidence of the derivation of the geometric algebra of Book II from Babylonian algebra. (8) x. It is out of the question to attribute this to mere chance. For II 10: given x . has now become certainty. indicated in II 9. The mixed cubics x2(x + 1) = V and XZ (x . p. amount of course to the Babylonian solu- tions (9) x = %s + Y(F/2) . the Pythagoreans in particular. and (10) x = YF/2 . left their trace in the arithmetic and the seometry of the Pythasoreans.(S/2)2. to the simple application of an area to a line.y = d. that all the Babylonian normalized equations have. We observe moreover that the simple linear equation ax =F leads. namely that the Babylonian tradition supplied the material which the Greeks. and the numbers in these tables are exactly the spatial numbers nl(n ± I).1) = V were solved by the Babylonians with the aid of tables.Y(F/2) . Y = YF/2 .

whose work was.) attains such heights is the fact that the architects Ictinus and Callicrates. says that the inward incline of the temple columns and the leaning forward of the structural elements above the columns serve to counterbalance an optical illusion. Less certain is the explanation of the slop- ing of the columns. PLATE 15 PL. which he explains mathematically. can be interpreted as a deliberately applied device to correct the optical illusion of a hollow. theoretical knowledge of optical phenomena and of persrective. 15.C. 5. which occurs also in other Doric temples . and indeed remarkable. (backward) and of the wall of the inner sanctuary (backward). the Roman architect (de Architectura III.~li""ri) . (Plwlo . the abacuses of the columns (forward). Vitruvius. The ancients themselves tell us that Phidias was familiar with optics and geometty. The Parthenon. the architrave and the tr'glyphs (backward) the cornice and the acroteria (forward). Such arching. by the slight arching of the stylobate (the upper one of the three steps on which the columns stand). among other things. That the architects 0 the Parthenon had a knowledge of optical phenomena is shown. applied to this building their quite advanced. seen from the South·East. according to Plutarch. supervised by Phidias.. 13). Not the least reason why the beauty of the Parthe· non (built in 447-438 B.

(National Museum in Athens). Height of the entire statue 2.14 m. The "Poseidon of Cape Artemision". Splendid work of one of the foremost bronze·founders of Greek antiquity. practically intact.). PLATE 16 PL. according to some Zeus (about 465-460 B.C. representing a deity. . Greek bronze original in severe style (compare Plate 17). probably Poseidon. 16. The statue was raised from the sea at Euboea.

and in Plato's image of the cave. neither by an integer. Their logical rigor did not ev~n allow them to admit fractions. not in terms of numbers. the CCjuJtlon x~ = 2 can not be solved. show this clearly. Nowadays we say that the length of the diagonal is the "irrational number" v'2. as well as numerous other immortal works of art. when the side is chosen as the unit of length. which turned them away from numbers. Indeed this is not difficult to find: it is the discovery of the irrational. why did they put it in geometric form? Was it their delight in the tangible and the visible. Anthmos means quantity. as Aristotle says (Metaphysics AS).it was parti- cularly the Pythagoreans who laid the foundations of geometric algebra. as Pappus tells us.onal to the side of the square. they had a very clear understanding of the ratio of the diar. numbers were "the rock-bottom of the 'entire universe". But it IS solvable 111 the domain of segments: indeed the . and they were ablc to prove rigorously that this ratio can not be expressed in terms of integers.tion. In the domain of numbers. its length can not be expressed. thcy replaced them by ratios of integers. But all of this is insufficient to account for the complete elimination of algebra. to occupy themselves with figures in- stead? Unquestionably.mJ. With "the diagonal itself". nor by a fraction. Engineers and natural sCientists have always done this. They had no scruples in adding the area of a rectangle to its base. not even in that of ratios of numbers. and we feel superior to the poor Greeks who "did not know irrationals". they calmly accepted an approximation. the world was made "by imitation of numbers" . And there is ample further evidence in the marvelously vivid descriptions of Homer. actually originated in the Pythagorean schoo!. the diagonal can not be measured. For the Babylonians. according to the reliable report of Eudemus. Would these worshippers of numbers have solved quadratic equations. When they could not determine a square root exactly. as Plato expresses it. The diagonal of the square is not commensurable with the side. But for the Pythagoreans. but by means of segments and areas. they attained these views exactly because they applied themselves to mathematics. which. But this means that. THE GOLDEN AGE 125 Why the seometric formulation? Why did the Greeks not simply adopt Babylonian algebra as it was. purely for the delight in the visible? This is hard to believe. But the Greeks were con- cerned with exact knowledge. not with an acceptable approx. And. As we shall see later on. That they did not consider v'2 as a number was not a result of ignorance. But the Greeks knew irrational ratios very well. every segment and every area simply represented a num- ber. the heavens were for them "harmony and numbers". therefore whole number. but of strict adherence to the definition of number. We must remember that. there must have been another push towards the geometrisation of algebra. The sculpture reproduced on plate 16.. the enjoyment of what can be seen was a strong motive power in the Greeks.

In the Republic Plato speaks of the number 7 as the "rational diagonal". if we set a1 = dx = I. and a new diagonal by adding twice the lateral unit to the diagonal unit. can indeed be expressed by the formula (20 + d)2 + J2 = 20 2 + 2(a + d)S If now cJI = 202 ± 1. then (2) holds for n = 1. This follows from the identity (2) d. = 2(a Thus. then subtraction of this relation from the preceding one gIves (20 + d)B + d)2± 1.. which he attributes to the Pythago- reans and which. one lateral unit and one diagonal unit. Side. indeed.." Thus we obtain the side-number 2 and the diagonal-number 3. In explanation of this passage. It is therefore logical necessity.lry on Plato's Republic II. but with the opposite sign. . etc. was proved by use of II 10.. then it also holds for n + 1. we meet with again in Theon of Smyrna and in Iambli- chus. chapters 23 and 27.. Now let two units be taken. We see from this that the Pythagoreans knew the principle of mathematical induction. : dft is an approximation for the ratio of the side of a square to its diagonal. hence it holds for n = 2. Geometric algebra is valid also for irrational seg- ments and is nevertheless an exact science. In this way. Proposition II 10. con- nected with the side 5. From here on we proceed in similar manner with the numbers found thus far. we get 2 +3 = 5. + d. we have to pass from the domain of numbers to that of geometric magnitudes. in order to obtain exact solutions of quadratic equations.and diagonal-numbers hint at the fact that the ratio a. etc.+1 = 20 . The names side.1 ±1 which. Consequently.and diagonal-numbers". and that they applied geometric algebra to pro- blems of the theory of numbers. + d.... then a new side is formed by adding the diagonal unit to the lateral unit. which was quoted above. Proclus gives the following definition of "side. not the mere delight in the visible..2=20 . d.. if (2) is valid for a particular value of n.126 CHAPTER V diagonal of the unit square is a solution. according to Proclus I. But. unity is potentially a side as well as a diagonal.and diasonal-numbers. 1 Comment. "As the source of all numbers. according to the formulas (1) an+l = a... certainly in essence. which compelled the Pythagoreans to transmute their algebra into a geometric form. 2 x 2 +3 = 7.

then the smaller of these magnitudes is again sub- tracted from the larger one. was proposed and solved by the Babylonians. Now we are going to take a look outside this school. then. For example. In Book VII of the Elements. THE GOLDEN AGE 127 But how did they get the recursion formula (I)? I venture the following con- jecture: Greek mathematics knew the method of successive subtractions (antanairesis) for determining the greatest common measure of two commensurable magnitudes a and b: the smaller one. 39). in case it does exist. say a. the process continues ad infinitum. The problem of approximating to the ratio of the diagonal and the side by means of rational numbers. is subtracted from the larger one. a" and b" .a.a A b' E a' B follows Fig. and. the process leads ultimately to two equal magnitudes c = d. Again and again it becomes apparent that there were excellent number·theoreticians in the Pytha- gorean school. in the form of the successive side· and diagonal. and from b' = a -a'. moreover they developed a scientific theory concerning these approximations and they proved the general proposition by complete induction. If ap- plied to two incommensurable magnitudes.a = AD = DE = EB = a' can now be subtracted from a = AB. Now. then one can subtract a from b once (see Fig. (3) b = Formulas (3) already present the same form as the recursion formulas (1). leaving a remainder b a b' = AE. if a is the side and b the diagonal of a C square. AnaxafJoras of Clazomenae is known chiefly as a natural philosopher. a = a' + b'o 2a' + b'. Repetition of this same process of subtraction gives again a smaller side a" and a smaller diagonal b". thus giving two new magnitudes a and b . choosing a'" as the unit of length.numbers.D. to arbitrary magnitudes to decide whether a common measure exists and to determine it. . He was held in high esteem in Athens. If a common measure exists. 39. say a'" and b'" has become too small to be observed and if one approximates by setting a'" = b'''. But the Pythagoreans carried this old problem infinitely farther than the Babylonians. this method is applied to numbers for determining the G. The remainder b . which equal the greatest common measure. If the process is continued until the difference between. They found a whole set of approximations of indefinitely increasing accuracy. a' = b . etc. a' and b'.C. a' and b' are again the side and the diagonal of a smaller square. and finally a and b are represented by means of (3). at the beginning of Book X.

that is why they fall behind the stars in the diurnal revolution. When the popularity of Pericles was on the wane. according to Anaxagoras. shortly before the Peloponnesian war. One is explained on the basis of the vortex motion. Anaxagoras taught that the moon receives its light from the sun and he gave the correct explanation of solar and lunar eclipses. and especially the moon. Instead of resolving the motion of the sun in a diurnal motion. his political opponents. move away from the equator. along the ecliptic. it is resolved into two compo- nents. sun and moon. which are nearer to the earth. He removed from the system the reason (vovu) which had. each of these components is accounted for in its own way. as that of all the planets. according to him it even exceeded the Peloponnesus in size. He reasoned as follows: the planets. appro- ximately. who taught us the proper motion of the planets in a direction opposite to that of the fixed stars. warms and softens the air. instead of moving in parallel circles. and Alcmaeon. this point of view is extremely sterile. The fundamental importance of the ecliptic as the sun's orbit.) Astronomically. But what is the reason that. The moon turns about more frequently because it can not master the cold air (while apparently the sun. from its own heat. then turn about and again approach the equator? For this Anaxagoras has a mechanistic explanation as well (Hippolytus. does not appear at all. It is purely accidental in this theory that these two mutually perpendicular motions have exactly the same period. along with the fixed stars. accused Anaxagoras of atheism because he had taught that the sun and the stars were dead glowing rocks. nevertheless his influence on the development of astronomy was not wholly favorable. This was due chiefly to his view that the celestial bodies are inanimate objects which do not move according to mathematical laws. set everything in . parallel to the equator. because they are driven back by the cold air in the North. sometimes to the North. In this respect therefore. probably to annoy him. Democritus of Abdera developed these ideas in a still more materialistic direction. viz. Anaximander who discovered the inclination of the ecliptic a century before Anaxagoras. and an annual motion. one year. enlightened ideas. the other by means of the compressed air in the North. then again to the South. There is no doubt therefore that Anaxagoras had merits as an astronomer. are pulled along less rapidly than the fixed stars. in which the sun lags one lap behind the stars each year. Refutatio I 8): Sun and moon make a tum about.. but are dragged along by the vorticial motion of the ether.g. He was condemned and compelled to leave Athens. one parallel and the other perpendicular to the equator. as the fixed stars do. and. were ahead of Anaxagoras with his modern. in opposite direction. He also estimated the size of the sun. e.128 CHAPTER V Pericles was one of his pupils and remained on intimate terms with him.

He is best known as astronomer. a common multiple therefore of the periods of all the plantes (including sun and moon). rational living beings which are enabled by their reason to describe their eternal circles according to mathematical rules. lay behind the stars. equal to 2 sidereal periods. Anaxagoras attempted the quadrature of the circle. but I do not know whether it has any connection with this. Proclus informs us that Oenopides was the first to deal with this problem because he considered the construction to be useful for astronomy. and made everything depend upon pure chance. This does not necessarily mean that the concept "long year" has the same meaning for Oenopides that it has for Plato. we can certainly conclude that Oenopides occupied himself with periods of revolution or with calendar periods and that he made a determination either of the position or of the division of the ecliptic. ()enopides of (. is not quite clear. on the authority of Eudemus. because it stimulates man to seek for himself an understanding of these rules and to find these circles. he in- vented "the cincture of the zodiac circle and the long year". according to Proclus. Should this be accepted? The Babylonians had a period of Saturn of 59 years. It is said that.hios was somewhat younger than Anaxagoras. such as those of 8 and of 19 years. In connection with the well-known construction for dropping a perpendicular (Euclid I 12). according to Eudemus. we shall revert to the treatises on perspective of Anaxagoras and Democritus. Proclus holds (I 23). The "long year" is used by Plato to designate a long period. such as Aetius. while' in prison. and especially the moon. On the basis of this stateml:lIl it ha~ sometimes been collduded that. consider the long year of Oenopides as similar to calendar· periods. Theon of Smyrna records that. but rather to the school of the Pythagoreans. He also thinks that the planets. who assigned to the planets fixed orbits in the Zodiac. he must therefore have reached the highest point of his life shortly after 450. This point of view is much more fertile for astro- nomy. THE GOLDEN AGE 129 order and had started the vortex motion. who assumed that the planets were dragged along without any order and tossed about by wild vortices. that he also discovered the method of transferring a given angle. after which all the planets return to their initial positions. They believe that Oenopides took the long year to be a period of 59 years and the ordinary year as consisting of 365 and 22/59 days. In con- trast to this the Pythagoreans held that the planets are divine. From the statement of Eudemus. animated. He did not belong therefore to the followers of Anaxagoras. Later doxogra- phers. because they are nearer to us and are therefore less powerfully dragged along by the vortex motion in the sky. What is meant by the cincture of the zodiac circle. at the . upon pressures and impulses. When we come to a discussion of solid geometry.

or that they were drawn simply by use of a carpenter's square. According to Xenophon (Memorabilia 16). while he was in prison. geometry was still in a very primitive state. It would seem to me preferable to assume that. who says: "With the straight ruler I set to work To make the circle four-comered. an Athenian. Which. occu- pied himself with the quadrature of the circle. the man is a Thales'''. not only as judged by the number of pro- positions that were known. In "The Birds". Converging to its center h'ke a star. an epic poet and a sophist. the need for these constructions had not been felt. although not accurate. right angles were construct- ed in a half circle. but also by the demands of rigor in proof. It has already been mentioned that. an interpreter of signs. before Oenopides." ''Venly.130 CHAPTER V time of Oenopides. the quadrature of the circle of the sophist Antiphon was intended to be serious. . which could only make him and his imitators unhappy. that it could be made a source of amusement in the theatre? Meton's quadrature was only a jest.ce. and that it was Oenopides who invented the con- structions. because even such simple constructions were not yet known. usually called Suidas. Is it not marvelous that a scholary problem was so popular in Athens at that time. It seems clear how the circle is here "made four-comered" by means of two streets which meet at right angles in the center. He also told him that it was not sensible to reveal his theories without charge. he introduces the astronomer Meton. On the contrary. that in his time geo- metry had reached a very high level. using the proposition of Thales. sends Forth its rays to all sides in a straight line. who was not much younger than Oenopides. He was called word-cook. In its center will be the market pla. that for instance. Altogether this problem was very popular towards the end of the 5th century. Antiphon reproached Socrates for his simple way of living. and drives Meton away with blows. But. contains the following note: Antiphon. It seems to me that such a conclusion is entirely incorrect. the leader of the birds. scoffs Pisthetaerus. Anaxagoras. in view of their application in astronomy. The comic poet Aristophanes even made a joke about it. although only orbicular. The late compendium Suda. we shall see from the quadrature of the lunules of Hippocrates. Squaring the circle. Into which all the streets will lead.

1 See F. . he obtains a polygon of double the number of sides. 1 Antiphon is of course quite right in thinking that by this method the circle can practically be squared. not a clever merchant. according to Aristotle. We can get some idea of the way in which he attacked this problem from a statement which Simplicius. He believed that. gets from his teacher Alexander of Aphrodisias. perhaps a triangle or a square. According Fig. formed by the semicircles on the Quod.. to this Alexander. He inscribes some polygon in the circle. by continuing in this way. The diffesence. and since the polygon that has been obtained is to be considered as equal to the circle. he concluded that the circle can also be squared.. is contrary to the principles of geo- metry. Afterwards (probably around 430) he came to Athens. In'the Physica (185a). because it is not based on the recognized principles of geometry. a hexagon or an octagon. Dtr &ich. the most learned and most reliable among the com- mentators of Aristotle.. The commentators (Simplicius. 41. Since for any polygon. on account of their small size.. he became very famous as a geometer. whose sides. Equally right are his opponents in saying that the assertion that a polygon can coincide with the circle. des Si"'1'/iciw iiher die Leipzig 1907. Fig. he would ultimately obtain a polygon. by pirates). Themistius and Joannes Philoponius) then explain how Antiphon proceeded. His method has been preserved due to the fact that Aristotle criticised it casually. The most famous among the geometers of the 5th century. for he allowed himself to be relieved of his money by crooked tax officials in Byzan- tium (according to others. one can construct a square of equal area. THE GOLDEN AGE 131 This typical representative of the guild of sophists thought that it is quite easy to square the circle.. Aristotle says that it is not even neces- sary to refute Antiphon's quadrature of the circle.. He started as a merchant. Hippocrates of Chios (not to be confused with the still more famous physician Hippocrates of Cos). By bisecting the arcs. Hippocrates began with an isosceles right triangle and proved that the sum of the areas of the two lunules. but. It is probable that Eudemus is tbe source for aU of tbem. would coincide with the circumference.'. 40. Rudio. also occupied himself with the quadrature of the circle.s in tbe statements of tbe three commentaton are insignificant.

and to eliminate the additions of Simplicius. since similar segments are segments which form the same part of the circle. v.IJU" Leipzig 1907. After having proved this. D.132 CHAPTER V right sides and by the semi-circle circumscribed about the triangle. and by constructing on the base a circular segment similar to the 1 O. p. It is still an open question whether Hippocrates actually proved this rigorously. Rudio. A•• il'luJ • •• d de. that unaerlies the Pythagorean theory of num- bers: four magnitudes are proportional if the first is the same part or the same multiple of the second that the third is of the fourth. bounded by arcs of circles. was first formulated by Hippocratus and his explanation was considered to be in good order.. copied word for word by Simplicius.. that similar segments of circles are in the same ratio as the squares of their bases. Rudio.r lltricbt du Simp/i. Hippocrates uses here the same concept of proportionality. is equal to the area of the triangle. according to his own statement. it would also be possible to "square" the semicircle and hence the circle. He demonstrated this by showing first that the squares of the diameters have the same ratio as the circles. I E. Groningcn 1929. 411. He considered as the foundation and as the first of the propositions which serve his purpose. B•• de.on Euclidi. .btr die Q-l. Diels and Usener. this de- finition is applicable only in the case of rational ratios. This led Hippocrates to study the squaring of lunules. p. Heiberg and finally Becker 1 have tried to restore the original text of Eudemus. Elemenw" . Tannery. We shall see that the proof in Euclid XII 2 comes from Eudoxus. I. is equal to the area of the trapezoid. See also the charming little book of F. Hippoc.d Studi . but to Hippocrates or to Eudemus. AHmann. he raised first of all the question how to square a lunule whose exterior boundary is a semicircle. Hippocrates had not yet arrived at a rigorous treatment of irrational ratios. J. Becker. The manner in which he squared such lunules can be learned from a famous fragment. He proved that the sum of the areas of a semicircle on one of the sides of the hexagon and of the threE lunules formed by the semicircles on the sides of the hexagon and by the semi- circle circumscribing the trapezoid. from the History of Mathematics of Eudemus. Dijksterhuis.i. Let us therefore attack the matter and study it. Therefore. with the addition of a few clarifying references to Euclid. It has been shown convincingly by Dijksterhuis 2 that this explanation is not due to Simplicius. The text conti- nues: For the ratio of the circles is the same as that of similar segments. if it were possible to "square" the three lunules. Next Hippocrates took an isosceles trapezoid formed by the diameter of a circle and three consecutive sides of an inscribed regular hexagon. considered as remarkable figures on account of their connec- tion with the circle. He accomplished this by circumscribing a semicircle about an isosceles right triangle. . The purified text begins as follows: The squaring of the lunules. Quellm u. This quadrature of the two lunules agrees with the first of the three quadratures in the fragment of Hippocrates that will be discussed presently. 34. Strictly speaking.

he proved it on the basis of a construc- tion like the following. Simplicius draws the accompanying explanatory diagram. squaring menlioned by Alexander. ALl2> AJ'2 + FLl2 ~ 2AB2. and let F Ll bisect the (line) BK perpendicularly. Hippocrates applies the following pro- positions: The square of a side (ALl) of a triangle (AFLl). with the diagonal. Therefore the segment in which it is inscribed exceeds a semicircle. Let there be a circle of diameter AB and center K. In modern symbolism the argument proceeds as follows: BLl2 ~ 3AB2. it can be squared. He circumscribed a circle about the trapezoid. opposite an obtuse angle (F) is greater than the sum of the squares of the other two sides. then the segment is greater than a semicircle. by taking a semicircle as the external boundary of the lunule. subtends this longest side. when the angle inscribed in a segment of circle is acute. For the line which subtends two sides of the trapezoid must of necessity have a square which is more than twice as great as that on the remaining (side). is added to both. 43. it follows. while the fourth side. when Fig. in both cases the external boundary of the lunule is a semicircle and the internal boundary a quadrant of a circle. BLl2 ~ 2AB2 + AB2 <ALl2 + AB2. the longer of the two parallel sides. which lies above the segment on the base. he could readily square the lunule. And. Because the segment on the base is equal to (the sum of) the two (segments) on the other (sides). Next he started from an (external boundary) greater than a semicircle. Observe that this squaring agrees with the first Fig. Since it has now been shown that the lunule is equal to the triangle. But then the angle which stands on the longest side is acute. Eudemus leaves the squaring of the lunules to the reader. 42. had a square equal to three times that of any of the others. while the squared longest side of the trapezoid is less than the diagonal and that one of the other sides which. then the opposite angle is acute. Obviously he is familiar with the concept of an "angle inscribed in a segment of circle" and he knows that all angles inscribed in the same arc are equal. and constructed on the largest side a segment. and that a smaller angle corresponds to a smaller arc. And let EZ lie between this (perpendicular bisector) and the circle. and this is the external boundary of the lunule. directed towards B. while its square is one-and-one half times as great as that of the radius. similar to the segments which each of the other three sides cut from the circle. Thus. He constructed a trapezoid. THE GOLDEN AGE 133 segments cut off by the right sides. the proof is similar to that of the previous case. But supposing that it were less than a semicircle. The lunule obviously equals the area of the trapezoid. of which three sides were equal to each other. And. By drawing the diagonal of the trapezoid one sees that the segment in question is greater than a semicircle. . Apparently. that the lunule is equal to the triangle. the square of a side is less than the sum of the squares of the other two sides. when the part of the triangle.

It is not entirely clear how Hippocrates concludes that KBI is greater than 2BZ2. since each of the two on the inside is 3/2times as great as the outside ones. which occur so frequently in Greek mathematics. EKI. If this is so. the lunule consists of the three segments and the rectilinear figure. it follows that the square on KE is more than twice as great as that on KZ. he proves as follows: Since the square on the line EZ is 8/2times as great as that on the radii. He proves that the outer boundary of this lunule is leiS than a semicircle. They require the construction of a line segment (EZ) of given length. The square on the line EZ is 1/. except for the two segments. The manuscripts add the explanation: "because the angle at Z is larger". the two segments being equal to the three.134 CHAPTER V Hippocrates applies here one of the "Neusis" constructions. He knew therefore how to construct the mean proportional x between two given lines a and b. HZB and KZB). therefore. then a circle win circumscnbe the trapezoid EKBH. Then it is clear that the extension of BZ win pass through E and that BH will be equal to EK. and he also knew that the square xl is equal to the rectangle abo Now let EH be drawn parane1 to AB and let K be joined to the points E and Z. composed of the three trio angles (EZK. and if the rectilinear figure is itself obtained by adding the two segments but removing the three. For. edge". so that the segment in which it is inscribed is less than a semicircle. "Neusis" means inclination.d curves (the line rA and the circle) and H whose extension is to pass through a given point (S). the lunule will be equal to the rectilinear figure. therefore . whose terminal points have to lie on given straight lines or .or whether Fi +I he used a so-called "adjustable straight. i. and the square on the line KB is more than twice as great as that on BZ. And the lunule that results win be equal in area to the rectilinear figure. And that this angle is obtuse. it follows that EZ > EK. And the trapezoid wlll have in its interior a circular segment circumscnbed about the triangle EZH. It is certain that Hippocrates must have constructed in advance the length EZ. by showing that the angle inscribed in the outer segment is obtuse. which is the mean proportional between the radius AK and 8/•. then. And let the line towards Z meet the line EH in H. so that L EZK is acute and L KZB leads to a qua· r dratic equation in BZ = x . Therefore the square 011 EZ is greater than those on EK and KZ together. If. It seems probable that the line of thought was as follows: from EZa = 8/a . an let the lines joining B to Z and to H also be drawn. We do not know whether he carried out this construction by means of compasses and straight edge . the segments which EZ and ZH cut off from the recti· linear figure inside the lunule. are equal to the segments outside the rectilinear figure. AK. g. Therefore the angle at K is obtuse. a straight edge on which the distance EZ was marked off and which was allowed to slide past and rotate around B until the endpoints E and Z were on the circle and the perpendicular bisector rA. times as great as that on EK. .

therefore the angle at K is obtuse. such as the fact that the square on the diagonal equal three times the square on the side. and the con- struction of the regular hexagon. KB and KF be the extended from the center until they meet e the outer circle in H. to the triangle. or greater or less than a semicircle. must have contained a large part of Books III and IV of Euclid. which deal with the circle and the inscribed polygons. and let the square of the diameter of the outer circle be 6 times as great as that of the inner circle. There- fore the lunule HeI will be as much less than the triangle designated by the same letters as the seg- ments cut off from the inner circle by the sides of the hexagon (are together). He is familiar with the concept of similarity and he knows that the areas of similar figures are proportional to the squares of homologous sides. And when the hexagon is added to each of these. The Elements of Geometry. whether the outer boundary was a semicircle. which in tum is six times as great as that on AB. he knows how to circumscribe a circle about a triangle and he knows that a circle can be circumscribed about an isosceles trape- zoid. The lunule and the segments cut off by the hexagon are together equal Fig. which he has written according to the Catalogue of Proclus. which already were "baby food" for the Pythagoreans. He knows . and 1. we see in the first place that Hippocrates mastered a considerable number of propositions from elementary geometry. THE GOLDEN AGE 135 KJ3a> KZI + BZ2 = 2BZ·. To square a lunule together with a circle. Since the square on the line HI is three times as great as that on the side H8 of the hexagon. Following the text. he proceeded as follows: Let there be two circles with centers at K. it follows that the segment con- structed on HI is equal to the sum of the segments H cut off from the outer circle by the lines He and eI. In this manner Hippocrates squared every lunule. the argument proceeds now as follows: KEI>2BZ2. Hippocrates obviously knows also the properties of the regular hexagon. as well as the con- tents of Books I and II. together with (all the segments cut off) from the inner circle by the sides of the hexagon. Looking at all these developments as a whole. And let a segment I be descnbed about the line HI similar to the seg- ~----:=-:~ ment cut off by He. He must therefore have known also that the side equals the radius. let a hexagon be inscribed in the inner circle and let the lines KA. Bydetermining the areas of these rectilinear figures. Hippocrates knows the relation between inscribed angles and arcs. it follows that the triangle and the hexagon are together equal to the sum of this lunule and the inner circle. 45. one can therefore also square the circle plus the lunule. EZ2 = 3/2 • EKI = EK2 + %EK2> EK2 + KZZ.

. Still more important for our evaluation of the mathematical level. are on a level with modern estimations of remainder terms. (1) a : x = x : y = y : b. 136 CHAPTER V not only the Theorem of Pythagoras for the right triangle (I 47). We also find the statement in Vitruvius that he left a treatise on the subject. This leads us naturally to Solid Geometry in the fifth century and Perspective. if (1) holds then the cube a3 would have to x" the same ratio as a has to b. durmg the "critical period" of modern mathematics. as we have developed before. these estimations were possible because Hippocrates had shown how to give exact proofs of inequalities.. reached dur- ing the second half of the fifth century and of Hippocrates in particular. i. he used divergent series in his calculations. and the estimation of remainders were recognized. he wants to prove this rigorously. Furthermore. 1M.e. that the necessity of the "epsilon tics". to construct a square with the same area. By means of this he knows how to construct lines whose squares have the ratio 3 : 2 or 6 : 1 to the square on a given line. 1 : 2) to that of a given cube. During the beginning of the fourth century. Hippocrates is not satisfied merely to construct the lunules and to conclude from the drawings that the external boundary is greater than or less than a semicircle. In the next chapter we shall return to the history of the problem and to the reliability of the tradition. Vitruvius reports that Agatharchus was the first to paint perspective wings for the performances of the tragedies of Aeschylus. The estimations of Eudoxus. how lines in a natural relation to each other could be made to correspond to the sharp view and the dispersal of the rays. is the excellent demonstrative technique and the high requirements of rigor demanded in the proofs. It is an established tradition 1 that Hippocrates was the first to recognize that the problem would be solved if one could construct two mean proportionals x and y between two given line segments a and b whose ratio is the given ratio. who showed that the difference between the inscribed and the circumscribed polygons of a circle can be made arbitrarily small. the calculations in terms of E and IS. he is able to square an arbitrary rectilinear figure. Hippocrates must also have occupied himself with the duplication of the cube.and acute-angled triangles ~I 12-13). It was not until later. Euler did not worry much about the convergence or divergence of his infinite series. hut also its ge- neralization for obtuse. But.g. For. without any scruples and obtained correct results. i. more exactly with the problem of constructing a cube whose volume has a given ratio (e. Greek mathematics passed through a similar "critical period".e.. "This also led Democritus and Anaxagoras to write on the same subject. The operation with inequalities is a very late achievement of modern mathematics.dis Opnam. 1 See Eutocius in Archi.

partly as receding and partly as ad- vancing" (Vitruvius. is equal to Ya of the volume of the cylinder.e. Persia. The dispersal of rays. It is a simple matter to divide a three-sided prism into three tetrahedra. 46. From this we conclude that. 2. they would have based such rules on the study of solid geometry.. not even the so-called rope- stretchers (probably surveyors) in Egypt. Archimedes says (in the Introduction to the "Method"). or of a pyramid. and on "spreadings" (ekpetasmata). in which all lines correspond to the center of a circle. praefatio). Cone and pyramid. and so that the things pictured on the plane fronts might appear.l See also Plate 19. even in India and Ethiopia. There are several mathematical titles among his extant works: On the tangency of circle and sphere. hedron). which are. but on what argument did he base his conviction that this formula. I This explanation of the I"ord is taken from the Geography of Ptolemy (VII 7). Babylonia.g. must have referred to the rays converging in the eye from the points in space and which cut the image plane in the image points. on geometry. and. such as would probably have been found in the work of the painter Agatharchus. in order that definite images on the painted wings might create the appearance of buildings. It is possible that Demo- critus had learned the formula for the volume of a pyramid in Egypt. with the same base """ and the same height. THE GOLDEN AGE 137 after a definite point had been selected as center. .pter I. Democritus had indeed acquired a reputation as a geometer. but that he did not prove it rigorously "-_ (this was first done by Eudoxus). solid geometry must have reacht:d a stage of development which made possible the formulation and the solution of problems which arise in perspective. Anaxagoras and Democritus were highly respected as men of science. then it is not possible that these would have contained merely practical rules for drawing. no one excelled him. 2 He takes pride in the fact that in the composition of lines with proofs. then it holds also for an arbitrary pyramid. e. on mapping a spherical surface on a plane. and the corresponding one for the cone. around 450. 1 In Cb. on numbers. or the prism. a limiting process then furnishes the proposition for the cone also. It is reported that he has been in Egypt. mentioned by Vitruvius. i. Vitruvius de6nes scene-painting as a representation of front and sides. that parallel lines must be drawn so as to converge to a point. If they have written monographs on perspective. are indeed correct? If the proposition is valid for a three-sided pyramid (tetra- Fig. that Democritus was the discoverer of the fact that the volume of a cone. De Architectura VII. according to some. 2. on irrational line-segments.

This. it is. and if we consider furthermore the large amount of solid geometry presupposed in the solution of this problem by Ar- chytas. then all sections will be equal. but this is entirely non- sensical. The right way is next in order after the second dimension to take the third. inasmuch as no city holds them in honour. but this subject. these inquiries are languidly pursued owing to their difficulty. We see therefore that in the 5th century. is not easy to find. as prims (or cylinders). A commtiOla'Y "" Plato'. and then. " Plato on Solid Geometry. but of its state in 422. equal or unequal? If they were unequal" (and. Since even now. I suppose. and continuous and strenuous investigation would bring out the truth. approxi- mately. This would at any rate explain the statement of Plutarch that Democritus raised the following question: "If a cone is cut by surfaces parallel to the base. It is possible that Democritus did something like this. Taylor 1 looks upon this passage not as a complaint of the condition of solid geometry around 374. does not appear to have been in- vestigated yet. as things are now. of the poor state of development of solid geometry. He writes (Republic. and that Hippocrates had already reduced the problem of the duplication of the cube to that of constructing two mean propor- tionals between two given segments. (Glaucon) Why. Oxford 1928. It follows therefore that each of these tetrahedra is equal to Ys of the prism. if the slices are considered as cylinders).138 CHAPTER V two by two. who is indispensable for success and who. According to Cavalieri. these specialists would accept advice. first. solid geometry was an active field of study of various mathematicians. have equal volumes. lightly esteemed as they are by the multitude and hampered by the ignorance of their students as to the true reasons for pursuing them. made up of equal circles. to be discussed presently. "then the cone would have the shape of a staircase. the investi- gators need a director. seekers in this field would be too arrogant to submit to his guidance.. E. provided one can prove that pyramids of equal bases and equal heights. And secondly. the year in which the dialogue takes place. To this Dijksterhuis raises the valid 1 A. (Socrates) There are two causes of that. is the enlarging of cubes and of everything that has depth. but if they were equal. if he could be found. we went on to solids in revolution before studying them in themselves. 528B): (Socrates) After plane surfaces. then how are the sections. when Plato wrote The Republic. yes. Tima"". to begin with. But if the state as a whole should join in superintending these studies and honour them. and the cone will look like a cylinder. in The Republic. then it becomes really very difficult to account for Plato's complaint. we might add mentally. they nevertheless in the face of all these obstacles force their way by their inherent charm and it would not surprise us if the truth about them were made apparent. Socrates. If we add the remark that the Pythagoreans knew three of the five regular polyhedra. Taylor. equal to each other in base and in height. and by considering these slices. . it is possible to convince oneself of the equality in volume of two pyramids (or cones) with equal bases and heights by slicing them by means of planes parallel to the bases.

the character of actuality. concerned with the government of his Academy. Di. this gives the "duplication of the cube". We are compelled to assume therefore that Plato took solid geometry in a more restricted sense. as developed in Book XI of the Elements. We have already seen that Hippocrates of Chios reduced the general problem of increasing the volume of a cube in a given ratio to that of constructing two mean proportionals between two given line segments. Plato(?). Obviously. about which Anaxagoras and Democritus had written books. with Eva Sachs 1. which are not themselves similar. the construction of a cube whose ratio to a given cube is the same as that of two given line segments. For the ratio 2 : 1. But what does Plato mean here by solid geometry. that these things do "not appear to have been investigated yet". Menaechmus. Heron. In the supplement to the Laws of Plato.e. i. Archytas of Taras discovered an extremely ingenious three-dimensional solution of this problem. Toeplitz. Qlglb vruI SItulinl B -4. I First.isstdlt. fiinf pl4l. the universally known Delian problem. p.. solutions were given by Eudoxus. p. counted as one of Plato's works but not published until after his death. Apol- lonius. known as the Epinomis. 334 and O. known to us chiefly from Eutocius' commentary on Archimedes. or. Philon of Byzantium. Later.. In the next chapter I shall discuss the history of this problem in some detail. so to speak. that Plato is here. in his own words. A few of the main points will be touched upon now. I See O. Berlin 1917. _bt_i"bt Epi . Di. THE GOLDEN AGE 139 objection that in that case. The duplication of the cube. the theory of the "sterea". Two numbers ab and cd 1 Eva Sachs. with the concurrence of"btn Korpcr. it is incomprehensible that Glaucon can say. He wants to lead the mathematicians in the Academy to a more systematic cultivation of solid geometry under his direction. the spatial bodies? Does it include the entire theory of planes and lines. One of the famous problems of Greek mathematics is the enlarging of a cube in a prescribed ratio. Diodes. plane geometry is defined here as the science. and it appears from the last sentence of the quotation (perhaps added later) that he succeeded in this. found by Democritus. Is it conceivable that Plato. the problem of the two mean proportionals is indeed considered as the problem of solid geometry. great importance was attributed to this problem. would have spoken so caustically about an undesir- able state of affairs which had existed 50 years earlier and which had been im- proved long since? We prefer to assume. the volume of the pyramid. for the sake of the histo- rical accuracy of the picture. . This more restricted sense is indicated in Plato's own words: "the enlarging of cubes". Qlglb "lid Studi" B 2. the theory of per- spective. would be hard to account for.0 forth? If so. and 1'. Sporus and Pappus.. Nicomedes. 191. which marks the passage. Eratosthenes. which teaches us how to make similar two numbers. Becker.

or { x . when Plato was making his plans for The Republic. and hence how to construct two cubes the ratio of whose volumes is equal to that of two arbitrary integers. which teaches us how to make similar in this sense two numbers which. to change an area into a square (TeT/?aywlIlCEtJI). at least not yet known in Athens. No~. consisting of three factors. how did it come about that Plato and his mathematical friends considered the problem of the enlargement of the cube as of such extreme importance? Let us first have a look at the state of plane geometry. the al'phcatlOn of an area F to a line a. this is evidently the problem of solid geometry. if their sides for a continued proportion: a:J=b:c=c:j. But plane geometry shows how to construct a mean proportional between any two lines. d respectively) are there called similar if their sides are proportional a: b = c: J. And now solid geometry is defined as "the new art. the enlargement of a cube in a given ratio is also the outstanding problem of solid geometry.g an arbitrary rec~ihn~ area F into a square amounts to solving the pu~e quadratiC x" = F. According to Euclid VIII 18. And now it also becomes clear why he can write that these things do "not appear to have been investigated yet". In parti- cular therefore. band c. But. and was then led to the more optimistic tone noticeable at the end of the passage quoted above. hence. In this new connection two numbers abc and del are called similar. shows that for Plato himself.of an arbitrary quadratic equation or to the solution of a system of 2 equatlOns With 2 unknowns of the form X + y = a. are not similar. as given. . no other definition of this subject is given. reduces to the solution of the linear equation ax = F. Perhaps Plato got the neWs of this solution just before the pu- blication of The Republic. { xy = F. solid geometry shows how to transform any number into a cube. the app~ication and the a~dition of. Plato mentioned as the most important planimetric operations. with excess or with d~ficiency. as given. In the above passage. are not similar". Confrontation of this definition with the passage in Plato's R'epublic.140 CHAPTER V (considered as area of rectangles of sides a. changm. and to transform any rectangle into a square. xy = F. it leads to the solution. areas. For the Epinomis.y = a. two numbers are similar only if a mean propor- tional exists between them. in which the purpose of solid geometry is defined as "the enlargement of cubes and of all things which have depth". around 375. they can be made into similar numbers by means of plane geometry. Apparently the solution of Archytas had not yet been obtained. all these operations come from geometnc algebra. even if two numbers. The Epinomis proceeds now from plane numbers to spatial numbers. without excess or deficiency.

Is it he who is honored by the inscription visible at the upper edge of the bowl: "Hip(p)odamas is beautiful"? To the right ofthe pupil.C. The "musike" which the boy was taught (art of the Muses) comprised literature as well as music. we read the opemng words of an epic poem. Without doubt. I begin to sing of the full·f1owing Scamander" (one of the two rivers in plain of Troy). On the left a lesson is taking place in playing the lyre. the young man in front of the teacher presently has to recite the poem. 17. (Staatliche Museen. Berlin. Athenian school. the latter played an im porant part in the life of the Greeks. then he has to take the boy home. where the teachers of music and the grammarians gave their lessons. The other side of the bowl. about 485-480 B.) This represents the interior of a "gymnasion". represents lessons in writing and in playing the flute. PLATE 17 PL. on the right. not shown here. . Attic bowl with red ligures of the painter Duris. which the teacher in the center holds up before the pupil.. On the $Croll. beginning. as is appropriate. the "paedagogos" sits quietly waiting for the end of the lesson. with an invocation of the Muses: "Oh Muses. in reciting.

Landesbibliothek Brunswick PL. in the hair is antique and seems to have been ander the Great in honor of his great teacher (accord. Copy in marble (M.C) Beautiful Ro· man copy in marble.D. British Museum.C. after a bronze original of about 335 B. Aristotle (384-322 B. PL. 6th century A.D.). Euclid (I) Miniature from the manu- script of the Roman surveyors (agrimensores) in Wolfenbiittel.) of a famous statue. probably in bronze.C. from the early years of the Empire. 1&. It may trait. (See A. ing to K. Private collection. Plato (427-348/7 B.C.. Boehringer Platon.C. It is possible that (30-50 A. perhaps of Euclid's own time. 180. 2403).).S. This R. Schefold).. tue of Lysippus in the Pompeion in Athens. 18b. PLATE 18 PL. 1&1. PL. Blldmsse und Nachwelse.D. Schefold) the famous sta· Breslau 1935). Statuette of Socrates from the 2nd century probably by Silanion. this representation derives from a Greek por- made during the master's life (about 325 B. was perhaps (according to K. after a Greek original ofabout 370 B. The diadem be supposed that this was the statue erected by Alex. taken from an ancient original (Schefold). .

He classified all line segments. the problem of the . Obviously it is not the only pro- blem. This can be found in the dialogue Theaetetus. which produce commensurable squares. and who puts highly interesting things about Theodorus' treatment of irrational sides of rational squares into the mouth of the chief protagonist Theaetetus. into those which are commensurable and those which are incommensurable. and Plato's dialogue is dedicated to the young hero's memory. which is laid in 399. Who was this Theodorus? Diogenes Laertius calls him Plato's teacher. he must therefore have been a contemporary of Hippocrates and Democritus.e. to the men of Plato's time. thus opening the way for other problems. i. We had therefore better confine ourselves to Plato who introduces the grey Theodorus himself in his dialogue Theaetetus. The first new problem that arises here. Iambli- chus mentions him in his Catalogue of Pythagoreans. It is not surprising therefore that he looks upon solid geometry as the generalization of geometric algebra to space. It is therefore entirely logical to consider this as the central problem of solid geometry. the construction of a cube of given volume. In connection with geometric algebra. and he found a general solution of the problem which Theodorus had treated for a few special cases. Thus we see. But he is not too old to give a lecture on incommensurable line segments which arouses the deep interest of the young Athenians Theaetetus and Socrates. is the solution of the pure cubic xa = V. That is why Plato adds the words "and everything which has depth" to "the enlargement of cubes". i. by way of the Pythagoreans. his ch. Theodorus and Theaetetus. THE GOLDEN AGE 141 Finally. A very important contribution to the clarification of this problem was made by the famous geometer Theodorus of Cyrene. For us. who had the tendency to include every famous mathematician among the Pythagoreans."ational moves to a position of increasingly central importance. Theodorus of Cyrene appears as an old man in the dialogue Theaetetus. that what Plato calls plane geometry is mainly the geometric alge- bra of the Pythagoreans. the most important result of this clarification is that we recognize more and more clearly the line of development of geometric algebra from the Baby- lonians. there are also equations of the form x2(x + a) = V and other similar ones. As a result of this lecture. but this does not mean very much. at the end of the 5th century. In the poem.uacter and his merits are spoken of . because Iamblichus was a freak.e. as the geometric interpretation of the calculation with products of three factors each. the "adding" of areas or of lines is after all only the geometric equi- valent of addition. Theaetetus started thinking. Theaetetus fell on the battle field in 369.

The subject matter of Theodorus' lecture is sketched as follows by the person Theaetetus in the dialogue: "Here our Theodorus drew (or wrote) something about sides of squares (:>reel w"ap£w11 Tt lYl!aV'e) and showed (d:ot09lal11W11) that those of three or five feet are not commensurable in length with those of one foot... . in the Greek acceptation of the word. hence the word w. a mathematical discovery of his young friend. here something stopped him (or: here he stopped). Plato sketches in a brief passage. but only "sides of squares".&fUl~ can not mean "squares". force) can indeed very well mean the side which produces a square. as it does later on in Diophantus. It was the purpose of Theo- dorus' lecture to show this. 17 square feet. 5.. 17 into squares. We would represent their lengths by the numbers v3 and VS.. It matters little how Theodorus constructed the sides fII" filii' ••• 11117 of the squares whose areas are 3.. When. even though the detailed circumstances of the dialogue may be fictional. -fl." The expression "commensurable in length" (I'~"e. as Tannery does.142 CHAPTER V at great length. v3 and VSarenotnumbers and the line segments referred to have no lengths that are expressible in numbers. a little further on Theaete- tus says himself that henceforth he will designate certain sides of squares as """al'~ (in restricted sense). it is reasonable to assume that the separation of the contributions of Theaetetus. it is not necessary. perhaps he used the . is historically correct.. and describes in detail what had previously been found by Theo- dorus and what Theaetetus has added to this. . It can not have been Plato's intention to give credit to Theodorus for what is due to Theaetetus. """al'~ (impulse. Let us therefore avoid the modern notation I V3 and rather denote the side of the square by fII. "Those of three or five feet" are the sides of the squares whose areas are three and five square feet. to replace the word """al'~ by w"aJdvrJ (creating). 5. nor vice versa. but. I11'JI'IS8Tl!~) also occurs in Euclid's Book X and means that a common measure exists. and in this manner he took up one after another up to the one of 17 feet. Indeed. which gives the impression of having been dragged in. simply followed Euclid I 14 to transform rectangles of areas 3. Perhaps he Fig. they are incommensur- able. Philologically. . It is applicable only to line segments. in the dialogue itself. . from those of Theodorus.

But it can not be assumed that Theodorus followed this procedure. this can be accounted for in another way. but not for q = 3.e. on the assumption that.e.. P. that which is left never measures the one before it. . Is it not possible that he proved the incommensurability of "'a. "'1 is commensurable. the indirect proof.. if he had reasoned in this way. this process leads to the greatest common divisor.e. but a geometer. it produces the largest common measure. 1 Joco. to bring into the light. because the proof indicated above can very readily be carried through for the general case... and that this continues until the remainder be- comes so small. for the sub- tractior. but that he merely showed intuitively. to show. "'" . that and "'Q have a common measure. that the remainder can never measure the preceding one? Anderhub points out that d1C(J(paJ"B'" does not mean to prove. Anderhub. Ka..... if applied to two numbers. taken from the interesting monograph of J.. "'" . "'17 geometrically? A geometrical criterion for the incommensurability of two line segments is found in Euclid X 2: "If." The idea of subtracting in turn comes from the theory of numbers.I_"'s. H. as shown in Fig.. 47. A. Wicsbadcn 1941 . but. q = 5. one can obtain an equation of the form By means of the theory of numbers of Book VII. there is always a remainder. . 17. For. as we shall presently see.Sa... as Plato says explicitly. "'17 and "'1 are incommen- surable? One might try to extend to "'a. "17. the magnitudes will be incommensurable. when the less of two unequal magnitudes is continually subtracted in turn from the greater. that the limited accuracy of a drawing in sand compels one to stop. to make clear. there can be no common measure.. one can then show that this equation is possible only if q is itself a square. i. then it is entirely inexplicable why he treated the cases q = 3. at the start of the pro- cess. . Auasabe du ~lIe·Wukc. that we discussed earlier for "'1. each separately. Thus. Theodorus was not an arithmetician. 5. He is therefore of the opinion that Theodorus did not really prove the infinite character of the process. "'. then it becomes immediately understandable why he had to deal separately with each of the cases "'. and then stopped at q = 17. If Theodorus used this criterium."itTttI tioes "ismJt. to two commensurable magnitudes. i. More important is the question: how did Theodorus show that "'a.. if the process never comes to an end.. process runs differently in each of these cases. . "'17 the "apagoge". that.s tIt.". "'" . This is indeed possible. etc. i. and if applied in general. which probably was in aistence at the time of Theodorus. but to designate. 1 Anderhub uses this figure to explain why Theodorus stopped with "'17. How did he show that this process never ends. THE GOLDEN AGE 143 Pythagorean Theorem to pass from one to the next.

Speaking about mathematicians. Wl1 . and probably having in mind the. (Plato. then one can conclude . I am led to the following conjecture. the subtraction procedure will lead to a pair of lines whose ratio is the same as the ratio of the pair from which one started. the question as to the existence of a common measure has no sense. The question of commensurability makes sense only for line segments as objects of thought. If one can show that. But in the dialogue. after a certain number of steps. but in the case of Theodorus some further elements have to be considered. and hence nothing concerning the commensurability or incommensurability of WI' W6. W&... From the obvious fact that. pursuing their inquiry for the sake of the square as such and the diagonal as such. proof of the incommensurability of the side and the diagonal of a square. one by one. but what they really seek is to get sight of those realities which can be seen only by the mind". The Republic. Theaete- tus. Theodorus had made the incommensurability of W3 . It is therefore clear that Theodorus could not appeal to intuition to prove the incommensurability of the sides of his squares.. towards the end of the 5th century. he writes: "And do you not also know that they further make use of the visible forms and talk about them. though they are not thinking of them but of those things of which they are a likeness. the demands of rigor in proofs had already reached such a level. not about things which are seen and heard. that.144 CHAPTER V This does not appear to me as likely. and not for the sake of the image of it which they draw? . •. by means of his drawing. At the time of Theodorus. but to the fact that there are infinitely many incommensurable sides of squares. 510 DE). there always is a remainder. how could the keen critic Plato assert that Theodorus had "shown" or "made clear" this incommensurability? The most that could have been said is that. but was considered subject to proof. How then could the mathematician Theaetetus. . the simple fact that an arc is greater than or less than a semicircle. as we saw in the discussion of the fragment of Hippocrates. Book VI. ••• . How then did he discover this fact? By developing further a hypothesis of Zeuthen. was not accepted from "- figure. nothing whatever follows as to the finitude or infinitude of the process. while Theodorus had dealt with a few of them. trying to make as clear as possible a logical discrimination between the achievements of Theodorus and his own does not point to a difference in probabi- lity. a hair's breadth will measure integrally every line that is drawn. but about ideal objects which exist in thought only. It does not follow of course that every geometer necessarily was as painfully conscientious. Plato is apparently convinced that the mathematicians will agree with him. even though one may be willing to accept the conclusion without proof. in each of the first five or six steps in the subtraction process. Plato is very fond of appealing to mathematics to show that exact reasoning is possible. which he wanted to bring under one concept. when we deal with line segments which one sees and which one measures empirically. And indeed. W17 probable.

.e)(111 + e) = 1111 . It follows that. Therefore the ratio is repeated.3e) : 2e.2. But these are the seg- ments with which we started.3e : (IP .2c) = (III + 2e) : Se. each of which has to be proved by means of the calculation of areas.. Analogous proofs can be given in all cases ilia. le : (III -. e. . The next case 11118 is not interesting. ea. thus obtaining III-e.3e. III. Then we have IP' = . since the product of the means equals the product of the extremes.e W -------4 and III + e.3e) = (III + . If the upper part is removed and then attached on the right. Se : (III . one obtains a rectangle whose sides are 111.3e : (III .. In continuing the procedure we can replace e and uP .. just as _. the difference of the squares wi and el is a "gnomon".e. Theodorus was able to devise this proof or a similar one.. . we subtract 2e from uP + e.3e. and let e = "1 be the unit of length."ie) = (III + "ie) : e. This method of proof is entirely analogous to the one used by the Pythagoreans in the application of areas with deficiency or with excess. because only one proportion is required e : (IP . 48.ft. (III + e)(111 . "17' The case 11117 is especially simple. The same scheme is now applied to the segments uP + e and 2e. Now (1) e : (III .eB= 2e1.3e) : Se. If we take now III + e and I.2e) = (III + 2e) : . The proportionality (3) is proved exactly like (1). 48.e are replaced by two segments in the same ratio: (3) 2e : (111. Hence Fig. Layoff e on .e by any other two segments that have the same ratio. let" = be the side of "8 a 3 sq. Now 2e and III . and subtract the latter from the former. square. as shown in Fig.e) = (III + e) : e. Se : (. we obtain III and e.. For example. (2) (w.. We have now two segments. with the tools at his disposal. III••••• .3e1.e) _ .5. -..3e) = (II + . e and uP . THE GOLDEN AGE 145 that the process will never end. because it requires a sequence of 6 proportions e : (IP . .4e) = (II + 4e): e.• 17 square feet.el . so that the process is 0-----1 periodic and will never end. i.4e) = (III + 4e) : . Now this is indeed the case with all sides of squares whose areas are 3.e) = (III + e) : 2e. The equality (2) can be proved as follows by the use of areas: wi . because 11118 = 3111•• But 11111 is quite com plicated. the remainder is 11' . This makes it quite understandable why Theodorus closed his explanation with v' 17.e.

In the next chapter. crasis and synchysis". 414). Hippias for the quadratrix and Perseus for the spirals" (ed. First comes the systematic foundation of plane geometry: the theory of parallels. Here it is suppo!?ed that the X-axis is tangent to. Concerning this. in the Greek manner: the square on the line x equals the rectangle on the lines 2p and y) is the symptom for the parabola. . we can recognize six lines of development: 1. The famous sophist Hippias of Elis. areas of polygons. Thus. Friedlein p. for they called it the quadratix". who distinguishes the different kinds of "mixtures of the straight and the curved" in lines and surfaces as "synthesis. 356. proportions. Surveying the mathematics of the golden age. e. Hippias and his Quadratix. For Proclus. who lived around 420.g. investigated a plane curve that is called the quadratix. Nicomedes for the conchoids. we shall return to Dinostratus' quadrature of the circle. W17 • Theodorus on higher CUNles and on mixtures. Proclus writes in his commentary to Euclid: "Thus Apollonius derived the symptom for each of the conic sections. probably because Hippias did not yet use the curve for the quadrature of the circle but for another purpose. when x and y are the oblique coordi- nates of a point on the parabola. Dinostratus.g. The quadratix of Hippias is the same curve as the one described by Pappus on p.146 CHAPTER V The detailed development in the work of Theaetetus may of course have diifered from the one sketched here. The main lines of de1Ielopment. It seems that Theodorus also occupied himself with curves. and that the Y-axis is parallel to the axis of the parabola. 252 of his great compendium and of which he says: "For the quadrature of the circle. writes that "Theodorus wrongly speaks of crasis in curves". Explanation: By the symptom of a curve. the equation x2 = 2py (or. and later geometers. used a curve which derived its name from this use. we shall also reproduce there Pappus' description of the quadratrix.. the angle sum of the triangle. for the trisection of the angle. but it is highly probable that he used in- commensurability criterium X 2 to demonstrate the irrationality of ws .. the ancients meant the condition which a point has to satisfy to lie on the curve. We do not know exactly what he meant by this. German translation p. Nicomedes. Hippias is not mentioned in this passage. Ws. . e. But we do know that in this period there were others who were concerned with curves. the parabola at the origin. •.

. the geometric equivalent of the equation xl = V. to set up a theory of proportions which remains valid for incommensurable magnitudes. The development of the theory of numbers had its root in the number mysticism of Pythagoras. and. . The fourth important line of development is that of algebra. Solid geometry was developed to a point. viz. The Pythagoreans knew only three of the regular solids. 5. This problem was reduced by Hippocrates to the problem of two mean proportionals. cube and dodecahedron. are themselves commensurable" remained unsolved. there arose. The third important problem is that of the determination of the area of the circle. by use of the indefinitely continued ant- anairesis. tetrahedron. viz. an exact quadrature of various lunules. 5. the con- struction of a cube of given volume. In extending this geo- metric algebra to space. and how the problems that had remained unsolved became the starting point for further development. regular polygons. There remains as an unsolved problem. which arose naturaUy from the rules of calculation carried over from the Egyptians and the Babylonians. attained a strictly scientific character in the theory of the even and the odd. the enumeration and the con- struction of all the regular solids. The general problem "which among the sides. and led finally to the systematic establishment of the theory of the divisibility and the proportion of numbers. 17. Hippasus circumscribed a sphere about a dodecahedron. based on this proposition. as the most important problem.. . Theodorus of Cyrene demonstrated the incommensurability of the sides of squares whose areas are 3. angles and arcs in circles. that produ~e commensurable squares. at which Anaxagoras and Demo- critus could outline a theory of perspective. In the work of Hippocrates we find the careful formulation of the fundamental theorem that the area is proportional to the square of the diameter. but not solved. as found in Book VII of the Elements. 3. 4. Democritus found the formula for the volumes of the pyramid and the cone. Next we shall see how these different lines of development were continued during the fourth century. The incommensura- bility of the diagonal of a square of unit side was proved by means of the Pytha- gorean theory of the even and the odd. THE GOLDEN AGE 147 theorem of Pythagoras. An important problem remained unsolved here. Finally the problem of the irrational was formulated. 2. The Pytha- goreans carried forward the development of Babylonian algebra ("Flower of Thymaridas") and transformed it into a geometric algebra. This develop- ment is provisionally brought to a conclusion by the Elements of Hippocrates. . 6. but he did not find a rigorous proof.

his teachers in mathematics and his pupils in philosophy. not the things which are observed by the senses. the formulation is made more exact. Proceeding in this manner dialectically. spent twenty years of his life in the glorious world of the Academia. Truth can not be self-contradictory. opens with the death of Socrates by the hemlock cup (399) and closes at the moment when Alexander the Great scatters the seed of Hellenistic culture over the entire world of antiquity. the teacher of Alexander the Great. if a certain preliminary hypothesis leads to a contradiction. Socrates disputes it. The road which leads to this state is that of dialectics. It is only the ideas which have true Being.. THE CENTUR Y OF PLATO The period to be discussed in the present chapter. that means the ideas. from hypothesis to hypothesis. The interlocutor sug- gests an opinion. Now the point of view is modified. which do not instruct. again Socrates shows that the new formulation leads to a contradiction and is therefore untenable. Why did Plato attach such tremendous importance to the cultivation of Mathe- sis? Why did he require all his pupils to obtain a thorough knowledge of mathe- matics before being initiated into his philosophy? It becomes apparent from the passages concerning the square-itself and the diagonal-itself which were quoted above. The great mathematicians Theaetetus and Eudoxus and all the others enumerated in the Proclus Catalogue. In this way. then this hypothesis must be rejected. Ideas can sometimes be contem- plated. C H A PT E R VI. At the center of scientific life ." Truth. because through it alone can truth be apprehended. He guided and inspired scientific work in his Academia and outside. were his friends. but which carry on a philosophical conversation. matters continue. it was because in mathematics one can learn that it is possible to reason about things which are neither seen nor heard. hence. we can conquer the errors in which we are involved and thus free our view for an envisagement of truth.:t proof. a positive result is never attained. In The Republic he writes: "The study of mathematics develops and sets into operation a mental organism more valuable than a thousand eyes. This period is one of political decay. but this can happen only after the errors of the senses have been conquered through concentrated thought. but exist in thought only. His great pupil Aristotle. and the method of proof in dialectics is the indire.tood the personality of Plato. through reminiscence of the time when the soul lived closer to God. in moments of Grace. but for philosophy and for the exact sciences it is an era of unprecedented flourishing. This is the method applied in the strictly dialectical dialogues. but the discussion progresses to higher and higher . in the realm of Truth.

Reidemeister empha- sizes that this method of proof by a reductio ad absurdum is derived from mathematics. In the older Platonic dialogues. and sometimes. Thus. In his Mathematik und Logik bei Plato (Leipzig. etc. Archytas of Taras. he obtained a solution of the famous Delian problem. For. He also made an important contri- bution to the theory of irrationals. But this is no longer dialectics. is in the main his work. according to Plato's own words. THE CENTURY OF PLATO 149 levels.'Ry through . mathematics hardly ever occurs. in his view. in connection with his theory of music. Fragmente. which came into vogue in his time. He was on terms of intimate friendship with Plato. a wooden pigeon that could fly (Aulus Gellius. These are the reasons why Plato attributed such eminent importance to mathematical training. the duplication of the cube. The man who made Plato acquainted with the exact sciences is the Pythagorean. and mathematical understanding stands between opinion and philosophical insight. but the entire Book VIII of the Elements. Plato himself repeatedly cites the proof of the incommensurability of side and diagonal as the typical example of a mathematical argument and points out that it is by means of such an indiw!t proof that one learns something about the things themselves. and on the physical nature of sound (Diels. he also laid the number-theoretical foundation for the theory of music which is found in Euclid's "Sectio Canon is" . 1942). with its arithmetical theory of continued proportions. By means of an extremely ingenious space construction. Vitruvius reports that he wrote about machines. similar numbers. that one receives training in dialectics. but the true "ideas" which lie behind them can not have two contradictory properties. The versatility of this remarkable South-Italian Dorian is unparalleled even in that era. between the visible things and the ideas. more errors are eliminated. besides computing the nume- rical ratios for the new musical scales. Noctes Atticae X 12) and a children's rattle (Aristotle. Politics E6. by means of systematic applications of the arithmetic and the harmonic means. Mathematical objects lie. Diogenes Laertius (VIII 79-83) calls him the first to give a systematic mathe- matical treatment of mechanics. already men- tioned several times. Archytas B 1). it is in mathematics that one learns to reason con- cerning things in themselves. according to Plato. one reaches a stage in which the truth can be suggested in the form of a myth. Not only did he develop some lemmas on proportionalities of numbers and some inequalities concerning three averages. 1340b26). Observable things are variable and contradictory. he designed machines himself. dialectics is an exact method of proof and no other method of proof ever occurs in the dialogues except that of rejecting assumed hypotheses. viz. ~ite rightly does Ptolemy call him the most important Pythagorean theoretician of music. at the end. He also reflected on the connections between the sciences. who received chil.

150 CHAPTER VI him his initiation in the exact sciences and into Pythagorean philosophy. during seven successive years he was elected strategus. On the other hand. meet the line on the cylinder in a certain point. is here denoted by . when Plato was held prisoner by Dionysius. and this will. And let the semicircle I i. this problem was solved by Archytas in the following manner. and which are so readily and so naturally met. lying in the rectangle of the half cylinder. S The point. and his lack of logic.. for greater clarity.. and let the extension of this chord meet tbe tangent line to the circle at L1 in II. in the older Book VII. In his native town of Tarent he was highly respected as a statesman. can be reduced. Heiberg). during its rotation. be given. his solution of the Delian problem.e. to the construction of two mean proportionals between two given lines. which were required by the mathematicians of his day. while the extremity A of the diameter is kept fixed. This lid of dearne. 84. on the other hand. had been driven out of Italy.ignatccl in the text by . This variegated picture is concerned only with the external aspects of his life and work. while the semicircle is moved away from this plane. the tyrant of Syracuse. The duplication of the cube. if AL1 remains in its position and the triangle AIlL1 is rotated in a direction opposite to that of the semicircle. Then we have to determine two mean pro· portionals between AL1 and r. and a vertical semicircle on ALI. arises from Arcbytas' kinematical way of thinking. occupy the position L1AA. as described by Eudemus ill Eutocius (Archimedis Opera III. his inability to express himself exactly and clearly. he managed to maintain a semi-auto- cratic regime under the external forms of a democracy. the new position occupied by the moving poim. from L1 towards B. towards the plane of the rectangle. and he never lost a battle. Greater surprise will be aroused presently. more generally. but a moving point. imagine a half right cylinder constructed on the semicircle ABL1.. According to Eudemus. Let us begin with his most beautiful accomplishment. while the point of intersection is K. equal to r. Let the position of the moving semicircle be L1' KA when these lines meet 2 and let the triangle. or. It is said that. Let a circle ABL1Z be descnbed about the larger line ALI and let (a chord) AB be drawn. his creative imagination and his mastery of geometrical methods on the one hand. tben it will intersect the surface of the cylinder and mark off a line on it. which.. Let two line segments. although the law allowed only one such election. rotating in the opposite direc- tion. p. At the same time the point B will describe a semicircle on the surface of the cone. it was his letter to the tyrant which saved Plato's life. the enlargement of the cube in a given ratio. when we shall penetrate more deeply into his way of thinking and reveal the extraordinary contrast between his ingenious ideas. his errors of thought and his prolixity. A critical read- mg of Euclid's Book VIII and of the Sectio Canonis shows us Archytas struggling with the demands of rigorous proof and clear exposition. as we have already seen. 1 then the straight line AIl will descnbe the surface of a cone. so that two points in the figure arc called .g. . his spiri- tual brethren. After the aristocratic Pythagoreans.1. When this semicircle (kept in a vertical position) is now rotated. is de.1 '. ALI and r.. Let the line BEZ be drawn parallel to IIL10. e..1 is not a deftnilc point in space.1. for him .

and hence that on eA and eI. Is this not admirable? Archytas must have had a truly divine inspiration when he found this construction. SO. their common line of intersection Me is also perpen- dicular to the plane of the circle. But LI'KA Therefore is a right angle too. when A is held fixed. Since each of the two semicircles LI'KA and BMZ is perpen· dicular to the base plane. The figure which Archytas saw before his mind's eye and which he wanted to construct. Its foot Wllllie on the circum- ference of the circle. AI. 50. is evidently the right triangle A. the lines KLI' and MI are parallel. In my opinion. and we Z 0 shall have the proportionality LI'A: AK = KA: AI = IA: AM. and ALI' form a continued proportion. and let the line AA meet the semicircle BMZ in M. since it also equals AB. the choice of 1 determines both K and M (since a perpendicular to ALI' erected at 1 can be made 1 A super8uous remark.AK. r because the triangles are similar. Mland Me also be drawn. is equal to the square on Me. Hence Fig. THE CENTURY OF PLATO 151 descnbed by BbeBMZ. 49.l(A' with the two perpendicular KI and 1M. thefourlinesLl'A. Let it be dropped and let it be called KI. since the cylinder is a right cylinder. and let its line of intersection with the circle BLIZA be BZ. but to Archyw. and let the (line) drawn from I to A intersect BZ in e.l Drop a perpendicular from K to the plane of the semicircle LlBA. Let the lines KLI'. Therefore the rectangel on eB and ez. Fig. this prolixity is not to be charged to Eudernus. and A hence AMI is a right angle. while ALI' is equal to a given line segment ALI. And AM is equal to r. Thus we have found two mean proportionals AK and AI between two given lines ALI and r. Here AM. Therefore triangle AMI is simIlar to MIe and to MAe. AK. .AIandAM are in continued proportion. reproduced once again in Fig. Now Archytas observed that. so that Me is perpendicular to BZ.

not to say confusion.d Stu4~. the vividness of his spatial and kinematic imagination. On the contrary. And. In Archytas' diagram. Already in antiquity it was observed that he introduced mechanical methods into geo- metry. In one respect however. the curve of intersection will then be exactly the circle of diameter AI). 1 The style of Archytas.. in reading in Diels (Fragmentt: der Vorsokratiker) the longest connected fragment Archytas B 1. When Aristotle says that Plato derived a great deal from the Pythagoreans.152 CHAPTER VI to intersect the semicircle on ALI'). so that the line AMK will lie on a definite cone with vertex at A. There is no doubt that his ideas must have had a great effect on others. The point of intersection of the space curve with this cone is the required point K. It is seen furthermore that. particularly on Plato. If AI varies from ALI to zero. B 2 (1933).. but his reasoning is not logical. p. by using many words and by giving elaborate examples. then. everything is in motion. the variable equals this magnitude. To determine this stage. Archytas does try his best to make clear the ideas he wants to convey. as follows: If a continuous variable is first greater than a given magnitude and later less. This may be formulated. i.e. on a cylinder of which the given circle is the directrix. one is most struck. he thinks kinematically. but especially by the lack of clearness. correct in themselves. then AM will also vary from ALI to zero. 236. When Plato occasionally speaks enigmatically. The semicircle AMI will then descri- be a hemisphere whose diameter is AA (this is seen most readily by starting with the sphere of diameter ALI and determining its intersection with the rotating ver- tical plane. The quoted section reveals Archytas' wealth of ideas. we have to think in the first place of Archytas. the remark that "the same air flows more slowly through an extended 1 Compare O. for instance. He confuses the velocity of the motion which produces a sound with the velocity of propagation of the sound itself. The point K will then always lie on a vertical through I. Now he wanted to determine I in such a way that AM shall equal a given line segment r. The point K will thus describe the space curve determined on the cylinder by the rotating semicircle AKA'. Quellm . which leads him to conclude from observations. 369 and B l (1936). he can be clarity itself. Archytas hit upon the following idea: he varies AI by letting I describe a circle of diameter ALI in a horizontal plane and placing the semicircle AlG1' in a plane perpendicular to this horizontal plane. when he wants to. not only by the unbearable prolixity. Archytas is far behind Plato. he has his reasons for doing so. at some time. If now AM has a given length r. then M will fall on a definite circle on this sphere. p. he completely lacked the gift of expressing his ideas briefly and clearly. hence at some stage AM will be equal to r. as indeed most Greek mathematicians. he did not hesitate to use the principle of continuity. Becker. that higher tones are propagated more rapidly than lower ones. .

perhaps he means to say that. which we have designated before as Proposition B. But it might have been paraphrased by saying that a is produced by multiplying p a definite number of times by itself. q. discussed earlier. Archytas' logical weaknesses are found again in Book VIII of Euclid's Elements. it shows that p and tare indeed two mean proportion between 1 and p". the style of Archytas is recognized unmistakably in Book VIII . conversely. It looks as if he is afraid of stumbling on the slippery paths of logic. and VIII 7 is equivalent to the proposition from the Sectio Canonis.. We have seen earlier that the beginning of Book VIII is very closely related to his theory of music. The central problem of Book VIIlls the following: under what conditions is it possible to find one or more numbers in continued proportion between a and b? According to VIII 8 this possibility depends only on the ratio a : b. Once the necessary and sufficient condition for the existence of a fixed number of . but Archytas considers it necessary to give all the minutiae of a proof based on the epimoric ratio. In a terribly roundabout way. This may be due to the fact that the concept power was still absent. The same prolixity is found in the fragment of Archytas on the theory of num- bers. He quotes Proposition VIII 8 himself.. in his foundations of the theory of music. p. it becomes clear that a is produced in this way.this clinches the matter. A number of other propositions and of concepts from Book VIIl are quoted by Plato. It is not explicitly stated that a and b are powers with the same exponent. . pR = a. But moreover. Hence a and b are the last terms of two geometrical progressions which have the same number of terms: 1.. p2 . and. then as many numbers can be found in continued proportion between them. if the difference in pressure remains unchanged. who must have learned the theory of numbers largely from Archytas. The proof is carried through. THE CENTURY OF PLATO 153 space. Let us analyze the construction of this book. Book VIII of the Elements. so that a and b may be assumed to be relatively prime. At the very start of the proof of VIII 9. given by Boethius. Indeed it is highly probable that this book should be ascribed to Archytas. q2 . 1. An essential error occurs.. because he has stumbled more than once in other places. For this case.. at a point where he confuses a proposition and its converse (see my article in Hermes 78). . And he has good reason for such fear. as between each of them and unity. 10: If a and b are relatively prime. It is of course quite obvious that the least numbers in this ratio differ by 1. more rapidly through a short one" does not excel in clarity. the problem is solved in principle by VIII 9. the air will How more slowly when a tube is lengthened. which contains the proof that no mean proportional is possible between two numbers in the ratio (n + 1) : n.. guided by the special case n = 3. qn = b. for example.

In other respects as well. but three times. 27. The question as to the conditions under which there are possible one or two mean proportionals between two numbers is thus answered not once. and between two cubes. at the end of the book. He proves in VIII 11. It becomes clear that the whole book is put together in a confused and disorderly manner. one finds an arbitrary number of terms. ab. In VIII 18. it is formulated once more in a slightly different form. . with which the rest of Book VIII is concerned. If the given ratio. a and b are then of course squares and cubes respectively. so that the quotations become very complicated. two mean proportionals. the formulations in Book VIII are un- necessarily involved.154 CHAPTER VI mean proportionals has in this way been established. Proposition VIII 22 says: If three numbers a. bB are 3 successively proportional numbers. when c is factored into pro- portional factors. the two extremes always being powers of a and b. a might also be factorable into unequal factors. The converse is not explicitly formulated. is a : b. 12. Upon multiplying them by a and the last one by b as well. and the least ones in a given ratio. that between two squares. But the writer does not proceed in so logical and simple a manner. but proved entirely afresh. but it is hidden in a corollary to VIII 2 and. without yielding the conclusion that c can also be factored into two equal factors. But this is not the worst of the matter . one has in every case to refer back to the way in which the geometric progressions are actually formed. 12: Between two similar plane numbers ab and cd (i. in continued propor- tion. 19 we find the following generalizations of VIII 11. abB. then c is also a square. as many as may be required. then c must be one as well. These results are not obtained as special cases form VIII 10. in VIII 26. the definition of similar numbers is satisfied. Proof: From VIII 20 we know that a and c are similar numbers. bB. Every time the proof of VIII 2 has to be recalled. the cases n = 2 and n = 3 (one and two mean proportionals). but as a problem. reduced to lowest terms. A similar remark applies to VIII 23. And VIII 19. When applying it. It looks as if the author were constantly fluttering around the problem without succeeding in finding a simple formula entirely satisfactory to himself. such that a : c = b : J). The misery in Book VIII arises from the fact that this procedure is not formu- lated as a proposition. band c are in continued proportion. can of course immediately be formulated as special instances. Proceeding in this way.e. 20 state that this sufficient condition for the existence of one or two mean proportionals is also necessary. if a is a perfect square. the analogue for spatial numbers. one mean pro- portional is possible. and if a is a square. besides being factored into two equal fac- tors. because. one obtains 4 numbers in continued proportion ai. it is possible to determine one mean proportional. alb. The point of departure for the systematic development of the entire matter is VIII 2: to determine numbers. between two similar space numbers abc and de! (in which a : d = b : e = c : f) two are possible. then ai.there are extraordinary logical errors. But this conclusion is not justified.

" This sentence has been discussed earlier. In Archytas' poor logic we have therefore a very personal trait of this. But two numbers. THE CENTURY OF PLATO 155 Thus we see that in Book VIII. We shall here reproduce this remarkable fragment as interpreted by Becker.isst"It. The Mathemata in the Epinomis. Die _1rnu/iscM Epi"" . the example of the original treatises of the great mathematicians. Two plane numbers ab and cd are called similar. B. 1 It turned out that it is particularly the mathematics of Archytas which is referred to. not inversely. Archytas is con- stantly at odds with logic. according to some completed by Philippus of Mende) gives a survey of the mathematical curriculum for the future leaders of Plato's ideal state. not of the numbers which have bodies. otherwise so excellent. d. with only absolutely essential comment: "The greatest and the first is the theory of numbers. but of the generation of the even and the odd.. if their "sides" are proportional (a : b = c : d). p. When Aristotle drew up the rules of logic. which is older than Book VIII and which has an excellent logical structure. He draws most of his examples from contemporary textbooks on mathema- tics. L. and that the even and the odd are the elements of numbers (Metaphysics AS).e. ·ibid. p. v." One is naturally led to think here of Aristotle who says that for the Pythago- reans. in the realm of numbers) are not similar. which in their nature are not similar. Hippocrates adhered to the highest standards for a rigorous mathematical proof. This would be contradicted by Book VII. It would be wrong to think that such faulty logic is a general characteristic of the science of this period. but which is concerned with making similar numbers. mathematician. Philologists have vainly broken their teeth on this hard nut . which is made transparent through the fate of plane figures. HmtItS 78. in a lofty tone and in enigmatic verbiage. As we have already seen. O. which "in their nature" (i. as well as in his other works. 334. But from this it follows that the thinking of Greek mathematicians must have satisfied very strict demands of rigor long before Aristotle. in their logical structure. can still be "made 1 O. he merely codified the regularities which he found in the reasoning of mathematicians and philosophers before his time. 185. the elements of the numbers are the elements of all being. the Epinomis (a posthumous work of Plato. 191. Beck~r. there follows what is very foolishly called land measurement (geometria). and of the power which this has over the nature of being. p. B 4. "For one who has learned this. Now it is clear that such textbooks followed. Toeplitz. trying unsuccessfully to meet its strict was only in modern times that mathematicians have been able to throw some light on this matter. Qvelltn ouuI Slvdim B 2 (1933). . Waerden.

which is constantly whirling about the duplication. The repeated duplication of a ratio produces a continued proportion such as a: x = x :y = y : b. squares). "But truly. for Plato.156 CHAPTER VI similar" by means of plane geometry.') . to which the duplication of the cube can be reduced." "By means of second powers" is here the power which produces the ratio 1 : 4. and such as the proportion 1 :2 = 2 :4 = 4 : 8. is the operation of doubling numbers. "The first power of duplication progresses as numbers in the ratio 1 : 2. such as geometric. the power of duplication.g. the central problem of solid geometry. as it is ex- pressed more briefly in the next sentence. arithmetic and harmonic mean proportionals.) and thus trans- form the areas ab and cd into squares la and mi. by means of another art. volumes and ratios. For. not a human marvel. or. the pattern and type of all nature receives its mark. which are likened to hall-marks. but the progression by means of second powers ("aTeI dt)pal'w) is also a dupli. By means of these powers nature produces the pattern or idea (elOO~) and the type or genus (rbO!. after this (plane geometry). worked out more fully in the sen- tence after that. called solid geometry by the informed ones. is stated emphatically once more in the following sentence: "However what is divine and marvelous for those who understand it and reflect upon it. catio. To transform two non-similar numbers into similar numbers is the generalization of the problem of the duplication of the cube which is. to transform into similar (numbers) those which have been produced as non-similar. areas. between a and b (c and drespectiv- ely) one can always construct the mean proportionall (m resp. which is constantly whirling about the duplication and through its opposite according to the different pro- portions." The power." ." Two spatial numbers abc and de! are called similar if a : d = b : e = c : f. And again he will be able. which means that they can be represented as areas of similar rectangles (e. but a divine miracle will be revealed to him who considers.l. This becomes clear by analogy with the next sentence: "The one by means of the spatial and the tangible is once more a dupli- cation and progresses from 1 to 8. is this that through the power. to which the next two sentences refer. The opposite power is the generation of mean proportionals for the various types of proportions. That these problems constantly hover about the duplication problem. the triply-extended (numbers) similar in their spatial nature.

if prolongated. The most important receding lines. belonging to the "second style" (about 60 B. G.). Beyen (Jahr- buch deutsch. we may conclude that op is the theoretical distance between the eye and the wall. For Vitruvius' relAtions to the wall painters of his time see De Architectura VII ~14. 19. . the diagonals of the panels on the ceiling converge on one point 0 . If we suppose that the panels are constructed as squares.C. archaol. Similarly. where he gives instructions to the decorators. 137 are really pertinent to central perspective. The semi-circles in the figures have been added. These decorations show in the upper part (the double square ABeD) a perspective construction. PLATE 19 A PL. Inst. meet in one point P. because Vitruvius mentions the center of a circle to which the receding lines must converge. Since the publication of these perspective wall decorations by H. 1939) there can be no more doubt that the passages from Vitruvius guoted on p. Wall decorations from Pompeii.

This plate does not turn when the spider rotates. . one turns the spider until the point representing the star lies on the proper altitude line. Persian astrolabe. PLATE 20 PL. When in use. revolves in this ring.late on which circles are engraved. When the altitude of a star (or of the sun) is observed. the ··spider". The angle of rotation.hole sights right and left on the circumference of the spider. it is based on stereographic projection. 182. Behind the spider one sees a r. pierced in various places. The sharp pointers of the spider indicate stars. The outer ring is divided into 360 degrees. The instrument is a Greek invention. Museum of the History of Science. the instrument is suspended vertically from the ring. Oxford. the eccentric circle the zodiac. can be read oft" on the rim. See p. representing the horizon and its paral el circles at various altitudes. on the upper half there are 30 circles. The spider can rotate and thus imitate the diurnal rotation of the celestial sphere. 20. constructed in 1221/22 by Muhammed ben Abi Bakr al·Farisi. A disc. and hence the time. The observation of the altitude may be made by means of the two pin.

one does no longer distinguish between musical intervals and the corresponding numerical ratios. . In this mystical region. THE CENTURY OP PLATO 157 The last two sentences may be interpreted as follows." The arithmetic and harmonic means of 6 and 12 are 9 and 8. they corres- pond to the fifth and the fourth in the theory of music." I have clarified the meaning of this mysterious sentence in my article in Hermes. but a mystical surrender to the divine Creator of the disci- ple who has been initiated into the secrets of numbers and of harmony. the ratios 6 : 7 and 7 : 8 appear. "By starting from these (ratios) and moving from the center to both sides. nor between geometric and material bodies. it is on these four intervals that Archytas built his three scales. in which the (arithmetic) mean is as much larger than the smaller one as it is less than the greater. If. If. thinking is no longer human. The lofty conception of nature. There is in this entire world of ideas an unmistakably mystical element. which tends towards the middle. one forms again an arithmetical or a harmonic mean. formed on the example of numbers. one gets the ratio 1 : 4. The human soul communicates with the divine soul. the ratios 2 : 3 and 3 : 4 come together. it has become divine. in the language of the theory of music. rhythm and harmony of which the text speaks. The Aristotelian criticism of this world picture is based on this that the Pythagoreans do not discriminate between mathematical and material things. it (the opposite power) presented to mankind melodious consonance and measured charm of play. in accordance with the definition of the harmonic mean. so that the interval of the fourth is divided into a diminished minor third and an augmented whole tone. which has the divine spark within itself. they actually produce therefore the melodious consonance of music in play. everything Hows together. Taking a line segment and doubling it. starting from the ratio 2 : 3. "Finally the (opposite power) of duplication. that the interval of the fifth has been divided into a major and a minor third. a cube is doubled in every direction. is mystically united with the divine spirit. in accord- ance with the laws of harmony. rhythm and harmony. is thoroughly Pythagorean. in the middle of 6 to 12. finally. The ratio 6 : 9 = 8 : 12 = 2 : 3 and 6 : 8 = 9 : 12 = 3 : 4 are called in Greek Hemiolion and Epitriton. it penetrates into the marvelous plan on which nature has been constructed. Now. one gets a geometric image of the ratio 1 : 2. one obtains the ratios 4 : 5 and 5 : 6 which means. the other (harmonic mean) however by the same part of the ex- treme terms exceeding them and being exceeded by them. then the ratio 1 : 8 appears. If the ratio 3 : 4 is similarly divided by means of an arithmetic or harmonic mean. The 8 exceeds 6 by one third of 6 and is exceeded by 12 by one third of 12. which is speaking here. abandoned to the blessed dance of the Muses. It is not a sober natural science. Because the human spirit. But if a square is doubled in length and in width. All of this is characteristically Pythagorean.

at the beginning of the Sectio Canonis. which does not discriminate between related things. his geometry and his theory of music are all connected. For them the heavens "are harmony and numbers". This leads him to distinguish between similar and non-similar plane and spatial numbers. indeed. the problem of the mean proportionals can be put into geometric language. this leads to the problem of constructing one or more mean proportionals between two given line segments. this means that an epimoric interval (such as the octave. thus frequently obscuring the formulation for the reader. and he uses this method in constructing his scales. Translated into the language of the theory of music. three purely number theoretical propositions are proved. no mean proportional could exist either. This mystical way of thinking. the fifth. for Archytas. but of intervals. this problem does have a solution in geometry. one would conclude that. but also in the mathematical textbooks which develop his ideas. i.. In his theory of numbers. which is indeed the scale of the sirens. These results must hav~ been very baffling for a Pythagorean. and he proves. If one were to assume . is also directly connected with the lack of logic which we have repeatedly found in Archytas. for line segments in certain ratios. It is possible to subdivide such intervals by use of the arithmetic or the harmonic mean. such as: "What is God? Unity! or: "What is the oracle of Delphi? The tetractys. fourth or whole tone) can not be split into two or more equal intervals. the spirit with definite numbers. the Pythagoreans identify justice. that of two mean proportion- als was discovered by Archytas in his duplication of the cube. which enabled Archytas to prove that certain intervals could not be halved. And now it turns out that. It has left its traces not only in his own writings. It is this proposition which forms. Surveying now once more the mathematics of Archytas by way of the Epino- mis. it becomes clear that his theory of numbers.. But the mean proportional of two line segments can always be con- structed.g. he himself calls them "related" sciences. " According to Aristotle. E. This forces the conclusion that line segments can not always be repre- . because the Pythagoreans always assumed that everything was ordered according to numbers. he considers continued proportions and he raises the question as to the existence of one or more mean proportionals between given numbers. Finally.a. the soul.158 CHAPTER VI This mystical identification is already found in the oldest Pythagorean sayings. in contrast with the theory of numbers and the theory of music. The construction of one mean proportional had oeen known for a long time. the basis of the theory of consonant intervals. etc.a natural assumption for a Pythagorean . It was exactly the assumption that all tones could be represented by numbers and all intervals by numerical ratios. that no mean proportionals can exist between two numbers in the ratio (n + 1) : n.that all line segments can also be expressed by numbers. no mention is made of numerical ratios however. which have come down to us.

to. we shall now discuss the further deve- lopment of the famous Delian problem.. that a : b = . What we know about this problem is to be found chiefly in Eutocius. this connection was clear."-. Without further motivation.knew therefore that these proposi- tions can be applied to the theory of irrationals. The duplication of the cube.. I owe a great deal to this pamphlet and to . : 1P = ~ : tJI. two proofs occur.also a defect of course. : ell is similarly the duplicate of the ratio . the equality of the ratios is now concluded from the equality of the duplicate ratios. from whom the theories of Book VIII ana of the Sectio Canonis are derived. In the proof of X 10. which are entirely alien to the methods of proof of Book X.a". .. contains the expression ip01}op€'IJ rde ("for we have learned this") which. use is made of X 11 . Something has to be added.ubocquent corrcepondencc with Reidemeister. 1 Did not Pappus say that the theory of the irrational had its origin in the school of the Pythagoreans? Connecting with the work of Archytas. But the tradition is very confused. viz. these two proofs not only contain superfluous parts. it becomes necessary to conclude from at. For this purpose it is obsecved that the ratio of the squares a2 : 1P is the duplicate ratio of the ratio a : b. Heath and other historians of mathematics. in which a and b are line segments. the proofs of X 9. In other words: there exISt incommen- surable line segments. in accordance with proposition 20 of Book VIII. that stirred the best minds of Plato's time. whereas Book X as a whole excels in its strictly logical structure and its extremely brief and elegant proofs. It is quite probable that Archytas and his disciples. Cantor. that they have to be con- sidered as later additions. no mean proportional number is possible) can be made similar by use of the geometric construction of mean proportionals. certainly for Plato and his disciples. and d numbers. concerning mean proportionals and similar numbers. which Heiberg also considers as "spurious". the commentator of Archimedes..ik tltr a. The proof of X to. In the proof of X 9. as observed by Heath.probably the same uncritical and unoriginal author who is responsible for the Sectio Canonis . . which points out so emphatically that numbers which are not similar in tlieir nature (and for which therefore. the proof is therefore incomplete. does not occur anywhere else except in the Sectio Canonis.. Is it likely that Archytas himself clearly understood this connection between his theory of continued proportions and the existence of irrationals? I think it is. In Book X of the Elements. A.i. A study of the Epinomis. have observed various contradictions of 1 Sec Reidemeiater. so completely different in style from the rest of the book. have already made such applications themselves. Both proofs use pro- positions from Book VIII. For. : d. in a more exact formulation: ratios of line se~ments can not always be expressed as ratios of integers. THE CENTURY OP PLATO 159 sented by numbers. and that the ratio . Whoever established the text of these proofs . Leipzig 19£. or. but also gaps. : d. proves that.

"Of these we have left aside the writings of Eudoxus of Cnidos". He begins with the statement that he has found the writings of many famous men. Hippocrates of QUos observed that. in which the construction of two mean proportionals between two lines is announced. GOttingen." It is clear that Eutocius had not Eudoxus' original work before him. . says Eutocius.ei. let alone therefore in Eu- doxus.. ~. Quickly double each side of the tomb. since they started with a cube and tried to double it. and this problem was called the duplication of the cube. When he heard that the tomb was a hundred feet long in every direction. It would be foolish to suppose that something like this were possible even in an author who was but a mediocre geometer. They sent emissaries to the geometers in Plato's academy to ask them for a solution. we shall have to separate the actual history of the problem from the dramatic story found in Eratosthenes' "Platonicus" of which some fragments have been preserved. Eudoxus with so-called curved lines. fift . without spoiling the beautiful shape.. It is said that Archy.. who had a tomb bwlt for Glaucus. . After they had looked for a solution in vain for a long time. This transformed the difIiculty into another one. W . These took hold with $feat diligence of the pro- blem of constructing two mean proportionals between two given lines. and. without changing its shape. whom an oracle had given the task of doubling an altar. after some time. he treats it as if it were a continued pro- portion." He seems to have made a mistake. biot. After having discussed a number of solutions by various authors. It is further reported that. also. we shall have to consult the sources from which the story of Eutocius is derived. p. It starts as follows: It is said that one of the ancient tragic poets brought Minos on the scene. because. 15. In order to present clearly the points of view of the writers. it should be twice as great. not less great. the area is enlarged fourfold and the volume eightfold. if only one could find two mean proportiomls between two line segments. above all. .. met the same difficulty. Ga. he said: "You have made the royal residence too sma1l. we shall describe here the method of solution of each of them. "because he says in his preface that he found the solution by the use of curves. of which the luger one is double the sma1ler. The geometers then started to investigate how to double a given body. while he does not introduce curves at all into his proof. . WilalDowitt-MocUendorf. Wilamowitz 1 has shown that this letter can not possibly be genuine. . certain Delians. after having established a discrete proportion. but only a poor extract or a compilation which did not clearly reproduce the reasoning.160 CHAPTER VI which they have not found a satisfactory resolution. .. then the cube would be duplicated. since in Eudoxus' own treatise curved lines would appear. Eutocius re- produces what he calls a letter from Eratosthenes to king Ptolemy.. Phil. Eutocius himself does not say much about his sources. 1894. tas solved it with half cylinders.. but it contains very important material. but no logical errors. Nachr. v. which have been handed down to us. For when the sides are doubled. 1 U. To resolve them.

partly in artfully made verses. 2) gives the source of the second iegend: III his wrk elltitled Platollicus. They went to consult Plato. in order to be hberated from the pest. a pit. Furthermore. that is. In almost the same words. a half century after Hippocrates. who told them that the god had not given the oracle because he needed a doubled altar. is found in Plutarch (de Ei apud Delphos. the "letter" contains an extemely important document. How can this contradiction be accounted for? . this is in thy power. . a solution of the Delian problem is given by the aid of a diagram and a model. partly in prose. or the broad basin of a hollow well. 386 E). Do not thou seek to do the difficult business of Archytas' cylinders. Plato. good friend. the duplication of the cube is an old problem. and others before him. I shall reproduce the verses in Heath's excellent translation: If. Elsewhere (de genio Socratis 579 CD). Fortunately. he could not make use of Hippocrates of Chi os. Eratos- thenes condensed the entire development of the problem into a short period of time. or to cut the cone in the triads of Menaech- mUSt or to compass such a curved form of lines as is desCribed by the god-fearing Eudoxus. a quotation in Theon of Smyrna (ed. but that it had been declared to censure the Greeks for their indifference to mathe- matics and their lack of respect for geometry. the second from the Pla- tonicus. p. In fact the problem is a much older one. It is likely that the Platonicus was a dialogue in which the Delians. when God announced to the Oelians through an oracle that. by this method. 1 In the prose section. 1 For funha details we refa to the article by von Wilamowitz. According to the first version. Hippocrates of Chios. thou mindest to obtain from a small (cube) a cube double of it. Eratosthelles relates that. he adds that Plato referred the Delphians to Eudoxus and to Helicon of Cyzicus for the solution of the problem. The epigram was engraved on a marble tablet In the temple of Ptolemy in Alexandria. they would have to make an altar. Von Wilamowitz has convincin&ly demoDitrated the &enuinenea of the epipm. Thus the con- tradiction between the first and the second account in the "letter" solves itself: the first tale probably derives from historical sources. to be discussed at a later point. Hiller. the architects were much embarassed in trying to find out how a solid could be made twice as great as another one. cited above. the problem arose from a declara- tion made to the Delians by an oracle at the time of Plato. connected with a legend about Minos. twice as great as the existing one. In this dramatic story. thou canst find the measure of a fold. In the second version. apparently two different versions of the story are reported. in connection with Eratosthenes. if thou thus catch between two rulers (two) means with their extreme ends converging. THE CENTURY OF PLATO 161 We notice a certain contradiction between the two parts of the tale. Eudoxus and Menaechmus appeared. and duly to change any solid figure into another. one after the other. had occu- pied themselves with it. for it arose from the translation of the Babylonian cubic equation xs= V into spatial geometric algebra. The same story. Of course in this setting. Archytas. an epi- gram of Eratosthenes.

so that 8 lies also on a second parabola with the same vertex. It is required to construct two mean proportionals x and J between two given lines a and b: (1) a:x-x:. also receive the sceptre at thy hands. the one of Eudoxus has been lost. Ptolemy. which also refer to the . conversely. Thus may it be. and let anyone who sea this olfering say "This is the gift of Eratosthenes of Cyrene. easily find a myriad of mrans. and may he in the future. thou hast gwen him all that is dear to Muses and Kings. Suppose that the problem has been solved. of the duplication of the cube for a short distance. 0 Zeus. 51. beginning from a small hue. so that 8 lies on a parabola whose vertex is LJ." From this we see that Archytas' solution with the cylinders. Hence 8 can be con- structed as the point of intersection of the parabola and the hyperbola. Ll Z Now. that the Platonicus is his source. in that. and layoff LJZ = x and Z8 = J. Hence 8 can also be found as the point of intersection of two parabolas. on these tablets.:b. From (1) follows also ~ ~-. Happy art thou. that of Eudoxus with the curves and that of Menaechmus with the three conic sections.=. ret us return to the dramatized story in the Platonicus. as a father the equal of his SOD in youthful vigor.. The solution of Menatthmus is described by Eutocius as follows: Aaordi"8 to Menaechmus. (1) follows from (2) and (3). god of heaven. . This justifies the conjecture that two other places in Plutarch. does Plutarch say.162 CHAPTER VI Nay thou couldst. so that 8 lies on a hyperbola with asymptotes LJZ and LJK. he merely speaks of Plato. Next Eutocius proves in detail that. In neither of the two places where he cites Plato's words to the Delians. We already know the solution of Archytas. Kt----~Iri Alternately: From (1) follows (4) .xb. From (1) follows in the first place ~ ~-. after having followed the real history Fig. so thatthe point of intersection 8 furnishes indeed the solu- tion of the problem. are histo- rical.

This solution makes use of carpenter's squares with grooves and of adjustable rulers. all of them mechanical aids which Plato condemns so roundly! . but then proceeded with a mechanical solution. not even compasses and straight edge. which require extended operations with unworthy handicrafts. on which so many constructions depend. In the Platonicus. The first of these passages. It appears that Eutocius' source quoted the introduction of the genuine Eudoxus text. Then the mathematicians. in the proof of Eudoxus. We can now also understand why Eutocius could write that. The construction of Menaechmus. For example. they invented mechanical solutions. In the life of Marcellus. from the eighth book of the ~aestiones conviviales. which were then rejected by Plato. which does not involve curves. that they had theoretical solutions. by designing certain instruments which could produce mean proportionals. but they fit very well in the Platonicus. because geometry revertS to observation instead of raising itself above this and adhenng to the eternal. was solved by both with the help of mechanical methods. THE CENTURY OF PLATO 163 Delian problem and in which Plato is again the speaker. he could not detect the "curved lines" announced in his introduction. Archytas and Menaechmus. it does not involve any instrument. They introduced an elegant variation into geometry and supported with intuitive models the problems which were not fully provided with theoretical solutions. if we separate the true history of the Delian problem from the dramatic account in the Platonicus. What was then the basis for Plato's crushing criticism? This dilemma is resolved satisfactorily as well. for in this manner the good in geometry is destroyed and brought to nousht. It is possible that things happened in the reverse order. in which the immanent God is the eternal God. is purely theoretical. taken from the Platonicus. It is this order which is suggested by the words with which Plutarch starts his story in the Life of Marcellus: The highly-praised mechanics was first set into activity by the men around Eudoxus and Archytas. "Plato" expresses himself similarly. because in this way they undertook to produce two mean proportionals by a non-theoretical method. giving vent even more strongly to his horror of taking recourse to "material things. who wanted to reduce the duplication of the cube to mechanical constructions. the problem of the two mean proportionals. Plato rejects them. Plato refers the Delians to Eudoxus and his disciples. humiliated." Such words do not appear anywhere in Plato's writings. These come with mechanical constructions.0lIl """s and sections. Most astounding fdr all modem commentators is the meachanical solution which Eutocius ascribes to Plato himself. startinS /. reflect on the problem some more. but only conic sections. for the benefit of the Ildians. and finally they produce their purely theoretical solut- ion based on the intersection of curves. is the following: Plato himself censured those in the circles of Eudoxus. also come from the Pla- tonicus. but that. im- material images. which is known to us.

can do this. If one now moves the ruler in such a way that KL passes through A and GH through B. The consensus of modern opinion is that this solution can not possibly be ascribed to Plato. It only requires an outline diagram. etc. Heath adds the further argument that the Eratosthenes' epigram does not breathe a word of a solution by Plato. To accomplish this. But the Plato of the Platonicus? Can one not hear the irony in the elaborate emphasis on the craftsman- ship required for the grooves. while G remains on the line AO and K on the line OB. so that the angles remain right angles. on which given line segments OA and OB have been laid off. in such a way that AM and BN are both perpendicular to MN. then OM and ON are two mean proportionals between OA and OB. who is not a geometer. Plato said to Archytas. our attention reverts from pure geo- metry to observable things. the real Plato can not have invented this solution. With all this. is completely spoiled. Just see. not even a preliminary geometric solution of the problem.. which require extended operations with un- worthy handicrafts?" It looks as if. "Plato" imagines a carpenter's square FGH with a movable ruler KL which is kept in a position perpendicular to FG. . I am in complete agreement. even I. then the problem is solved. about those "who take recourse to material things. etc. in the Platonicus. etc.164 CHAPTER VI "Plato's solution" depends upon the outline diagram shown in Fig. and the ruler KL exactly parallel to GH. The text strongly emphasizes the craftsmanship needed to construct the grooves and the crosspieces. But in this way everything that is good in geometry. etc.? Does one not notice that these words connect closely with the contemptuous remark of "Plato" in Plutarch. etc. 52. If one succeeds in constructing points Nand M on the extensions of the sides OA and M N B ~~ ~ll OB of a right angle. Eudoxus and Menaech- mus: You have found mechanical solutions? That is nothing.

that Theaetetus would cer- tainly become famous if he reached the age of manhood. This prologue takes place in 369. since they had indeed found solutions. but never did say. an extremely attractive young man. starting from these. Euclid agrees. on the day on which Socrates is summoned before his judges. "Most excellent. he supplied various playful details and he made the dramatis personae say various things. "but he is like you in his snub nose and protruding eyes"). He is not beautiful ("now don't be angry with me" Socrates. and he hit the Platonic style very well. so that one marvels how he accomplishes all this at his age. Euclid 1 tells his friend Terpsion that he had seen Theaetetus. The friendly manner. Theaetetus must have been. In the dialogue. says 1 Not the author of the Elemenla. like a stream of oil that flows without a sound.ns and then. and that dysentery was added to his wounds. geometric solutic. who comes to see Socrates in Athens. according to Plato. And. with perfect gentleness. For Eudoxus and Menaechmus had first found purely theoretical. the logical acumen of his answers. THE CENTURY OF PLATO 165 This reconstruction of Plato's speech not only resolves all apparent contra- dictions. In the introductory conversation of the dialogue "Theaetethus". "That is not at all strange". Thus indeed. but unbelievably quick of understanding." Now Theaetetus appears himself. which plays 30 years earlier. Terpsion conti- nues. The old Theodorus of Cyrene. On the other hand. they turned their thoughts "back" from pure theory to material things. the modesty with which he conducts himself. "and indeed just now I heard some people praising him highly for his conduct in the battle". he remained true to history. "But this boy advances toward learning and in- vestigation smoothly and surely and successfully. which they might have said. relates that he has there become acquainted with a young man so marvelously gifted as he had never seen in all his long life. worked out mechanical tools for the benefit of the Delians. but it gives us moreover a good idea of the structure of this part of the Platonicus. Eratosthenes chose his words very carefully. Theodorus had never considered such a combination of excellent qualities possible. "What a man he is who you say is in dangerl" exclaims Terpsion. Theaetetus IS still very young. such as "reverting attention" to material and ob- servable things. Terpsion". completely confirm the im- pression created by the praise of Theodorus. being brought back to Athens in dying condition from the camp at Corinth. in so far as he brought Archytas. Eudoxus and Menaechmus on the stage. . my boys!". but the philosopher Euclid of Megan. Various phrases. We come now to the two greatest mathematicians of the time of Plato: Theaete- tus and Eudoxus. the year of our hero's death. in the main. "A noble man. and then Euclid recalls a prophecy of Socrates. become clear. gentle and yet brave as nane else.

From the assumed commensurability of the sides. one first deduces geo- . we represented by the shape of the oblong rectangle and called them oblong numbers. How would the proof of such a proposition be made with classical methods? Indirectly of course. the numbers which can be formed by multiplying equal factors." In the dialogue Plato gives an example of a mathematical investigation made by Theaetetus.. After having explained what Theodorus had revealed as to the irrationality of the sides of squares whose areas are 3.). and what next?" "All the lines which produce a square whose area is a square number we called lengths (. would have to be stated as follows: Line segments which produce cubes whose 1I0lume is an inteser but not a cubic number. Following Socrates. and those which form the oblong numbers we called (in a restricted sense) sides of squares (OOvd. we represented by the shape of a square and called square or equilateral numbers ." "And did you find such a name?" "I think we did. had commensurable lengths.166 CHAPTER VI Socrates. The example is rather dragged in. as being incommensurable in length with the first. Therefore. briefly indicated by Theaetetus in his last short sentence.ue~). But see if you agree . The proof necessarily breaks up into a geometrical and a number-theoretical part." "Very good. What is for us more im- portant than these trivial definitions.. to try to collect them under one name.5.. Supposing that two lines. it has to serve as an introduction to a philosophical discussion. I can find but one possible explanation for this: Plato wanted to take a subject with which the mathematician Theaetetus had actually occupied himself. The analogous proposition for spatial numbers. Theaetetuscontinues: "Now it occurred to us (Theaetetus and another yoeng man). Plato attaches the greatest value to exact definitions. we can safely take the dia- logue as a historical source for the reconstruction of the mathematics of Theaetetus. The numbers which lie between these. that is why the definitions of oblong numbers and of equilateral or square numbers are given in such detail. by which we could henceforth call all the roots. We divided all numbers into two classes. is the proposition given at the end.. are incommensurable with the unit of lensth.17 square feet. which produce squares of areas n and 1..." "Most excellent. but not a sqUQ1'e number. are incommensurable with the unit of length. "I think Theodorus will not be found liable to an action for false wit- ness. but only by multiplying a greater by a less or a less by a greater. my boys!" Here ends the mathematical part. . and are therefore always contained in unequal sides. one would try to show that n is a square number. . since the number of roots appeared to be infinite. The one.1477"0(. but it does not fit very well. such as three and five and all numbers which cannot be formed by multiplying equal factors. which produce a sqUQ1'e whose area is an integer. very briefly but nevertheless clearly: Line segments. but only (commensurable) in the areas which they produce. And in the same way for spatial bodies.

ue-r{!IX: in Euclid. and to the expression "potentially commensurable" (&Pd. and that the ratio of the areas of the squares which they produce is equal to that of pi and qt.e. Commensurable magnitudes have to one another the ratio which a number has to a numbet. so that (1) n :1= p2 : q2. If two magnitudes have to one another the ratio which a number has to a number.'6T(!1X: and &Pal'S' mS.uJ. Analogously for third powers. (2) Then it has to be shown. the same phrase "commensurable in length" (.u!J8T(!IX:) occurs in both.u-qHIX: and thnJa. or. .uJ. The corresponding proposition for cubes is not found in Euclid. And while in Plato. the magnitudes will be commensurable. 6 and 9 of this book are: X 5. The squares on maight lines commensurable in length have to one another the ratio which a square number has to a square number. Book X contains the more detailed mathematical development of matters briefly indicated in the dialogue. but it can be formulated by analogy with X 9.c assumpt- ion of the incommensurability of the sides into an arithmetic property of the num- bers which represent the areas of the squares.u.u~H8' a15. There is moreover a striking similarity between the terminology of the dialogue Theaetetus and that of the beginning of Book X. the words . Theaetetus begins by recalling the accomplishments of Theodorus. among which occurs as X 2 exactly the criterium for incommensurability which Theodorus had most probably used in his investigation. that it follows from (1) and (2) that n is a square number. 1300k X opens with three propositlons concerning the "successive subtractions" of unequal magnitudes. In sound and in meaning. X 9. Proposit- ions 5.'6T(!IX:) corresponds in Plato the circumlocution "commensurable relative to the areas whicli they produce".u" in Plato are closely related to the phrases . equivalently. on X 9. i. It becomes clear from all these things that the dialogue Theaetetus and the be- ginning of Book X belong together and supplement each other. The geometric part of the proof is found in Book X of the Elements. they enabled him to translate the geometr. THE CENTURY OF PLATO 167 metrically that the ratio cf the sides is equal to that of two numbes p and q. number-theoretically.u~H8' ot$. 6 and X 9 to prove the proposition he enunciated in the dialogue. This view receives ample confirmation from a scholium. Theaetetus needed propositions X 5. X 6. which says that this theorem had been discovered by Theaetetlls. and squares which have to one another the ratio which a square number has to a square number wilT also have their sides commensur- able in length. a marginal' ot$.

and which plays the same role as the unit of length ("that of 1 foot") in the dialogue Theaetetus. the number- theoretical part of Theaetetus' proof has no further difficulties. Accord- ing to this Judgment.. . a fixed line segment is introduced which we shall denote by the letter e. so that. Analysis of Book X of the Elements. and thus to solve in complete generality a problem which Theodorus had only been able to handle for squares of areas of 3 to 17 square feet. we shall now proceed first to a discussion of the further contents of Book X.t K.168 CHAPTER VI It is Zeuthen's judgment 1 that the greatest merit of Theaetetus is however not to be found in the geometric analysis of the irrationality problem. In my opinion the merit of Theaetetus lies therefore not in his contribution to the theory of numbers. Then it follows from VII 27 that pi and q' (and r and f as well) are relatively prime. Zeutben.. as we have seen before. he was able to extend the entire theory without difficulty ("like a stream of oil that flows without a sound") to the sides of commensurable cubes. In Euclid measurable areas are called expressible (emdc"). 5. It follows that n = pi and 1 = q'. Theaetetus had discovered and proved several propositions from the arithmetical Books VII and VIII which were needed for this proof. by VII 21. reason as follows: q and p in (1) can of course be taken to be relatively prime. One can. o. I do not share Zeuthen's view on this. but in his study of incommensurable line segments which produce commensurable squares. such as the sides of squares of areas 3. but also non- measurable lines. This enabled him furthermore to answer the question as to which sides of squares are commensurable with the unit of length. We shall call a line segment (or an area) measurable. in the proof therefore that (2) can hold only if n is a square. not only measurable lines therefore. The case of cubes proceeds analogously. investigated by I H.. once one has at his disposal the propositions of Book VII. Book VII is of older date and forms the foundation of the Pythagorean theory of numbers. But. Expressible lines on the other hand are all those lines which produce measurable squares. He introduced the exact concepts "commensurable in length" and "potentially commensurable". if it is commensurable with the fixeci line segment e (or with the square eI). but in the arithmetic part.'1. for in- stance. . For.. But in the left member of (1) occur n and 1 which are also relatively prime and hence also the least in this ratio. in Definition 3. which led to the classification of line segments into commensur- able "lengths" and incommensurables "sides of squares" .•• 1910. 395. In order to get a picture of the other accomplishments of Theaetetus. Finally. . At the very start of Book X.u Yidm. v.. so that n is a square.. and in proposition X 9 he formulated the necessary and sufficient conditions on two squares under which their sides are commensurable in length.... they are the least of the numbers in that ratio.1... •. p. G.

the binomial and the apotome 2. • It is known from a statement of Eudemus that Theaetetus had studied these three irratiolLllitics and that he had related them to the three means. Cambridge IMass.. Thomson.").. All these proofs are based on one fundamental idea which runs as a guiding thread through the entire book: to prOfle properties of any type of line. A straight line whose square is equal to such an area is called a medial line. THE CENTURY OF PLATO 169 Theodorus. All other line segments are called unreasonable 1 (cUOYO!. and an apotome in only one way as a difference a . This terminology exhibits a first consequence of the principle which classifies line segments according to the squares which they produce. this basic idea already turns up in the first part of Book X and in the dialogue Theaetetus. See G.}. It is proved in Book X that all these new types of lines are "unreasonable". i.. A medial area is the area of a rectangle whose sides a and b are expressible. The question now becomes therefore: Is this area equal to that of the square of a binomial u + 11. it is necessary first to make a + b into an area. hence the word medial. which are introduced immediately after the binomial and the apotome. by constructing a rectangle of base e and height a + b. Junge and W. In our notation {ft." Book X. In order to define a magnitude which corresponds to our v' a + b. one constructs a square on this line and one in1lestisates the properties of this square. The sum a + b of two expressible but incommensurable Jines is called bi- nomial ("one with two names"). 1930.. and "irrational" with "non·measurable".. and the apotome to the harmonic mean. I have chosen these other words here. Is (3) (u + 11)1 = e(a + b)? 1 In modcrn translations one frequently finds the words "rational" and "irrational" in place of "cxprcssible" and "unreasonable". apotome ("the one cut off"). Therefore the Greek ideas do not permit the extraction of a square root. "rational" is syno- nymous with "measurable". the binomial to the arithmetic mean. Euclid defines three funda· mental irrational lines. for Theaetetus derived the incommensurability of certain line segments from the ratio of their squares. Apparently. the medial to the geometric mean. Tire . can best be understood if one starts from the following problem: under what conditions is the square root of a binomial (or an apotome) Itself a binomial (or an apotome)? A binomial a + b is not an area but a line segment. also that a binomial can be represented i\S a sum a + b in only one way. We would say. Properly speaking.b. it is an area v'" where r isa rational number. It is this which happens every time in X 54-59 and 91-96. but incommensurable. Harvard Semitic Series VIII. The other classes of irrationals. This line satisfies the equation xl = ab and is therefore a mean pro- portional between a and b. these "unnameable" magnitudes had not been either recognized or named in the earlier stages of the theory." of PQ1'1'"'.b of ex- pressible lines.. This did take place later on. For instance.e.36 and 73. the medial.. In X 21. it is shown that the square on a binomial can not be a medial area. . viz. because in modern mathematics. to prove that a binomial can not be a medial.. and are mutually exclusive. their difference a .

.. in X 35 they are both medial. + " = ta.e. the new unknowns x and y are then the projections of the sides on the hypothenuse (see Fig.. the solution of these equations is indicated.I. In X 3.y . 54.3-35. _ c. Now. in Greek terminology. = eVa 2 _ b2 • . which "applied" to AB = c leaves a Z square as deficiency". in X 33.I = ta. In X 3.(YzJ)1 . This gives the following condi- tIOns: . Now. 54). From the fact that the binomial can be split in one way only into expressible terms. . and y when added to and subtracted from %c. by the use of geometric algebra. measurable? For ~ and f!. = ".c. the conclusion follows readily that the two terms on the left of (4) must separately be equal to the terms on the right.170 CHAPTER VI Development leads to (4) ("I + . the situation is reversed. The method remains the same in all these cases: setting ea = cI and eb = cd. which yields the required line segments x Fig.I = (x + 1)c _ c2 .2 _ ye. a being the larger ot these terms (5) "I + .. the square root x III = V(Y2')1 . for various cases. new unknowns are introduced by setting (7) "I = xc. so that x and yare the roots of a quadratic equation.. whose hypothenuse is AB = c. or. From (6) and (7) we obtain for x and y the equations (8) x + 1 . ~ + f!. = (x . we have in place of (5) (6) "I + . + 2". we have the equations "I .(Yzdl'. x and y are the sides of a rectangle equal to the square on Yzd.%Vc' . For example.JI E occurs. Therefore these two areas must be measurable. is a measurable area and 21W a medial area. and 1111 = Y2'J.1)c = 2"".1_. in X 34. . one can interpret 14 and 11 geometrically as the sides ZA and ZB of a right triangle ZAB. and 21111 . this application is made every time according to the rules of the game. What is now the condition which the given areas ea = c· and eb = cd must satisfy in order that 14 and 11 may be expressible and thus ~ and f!.3-35. which may be put in the form (9) ". eb.. _ ta +tb. x. Since u' + f!. i. ea is a measurable area and eb a medial area. In the construction. "I + .

If these conclitions are satisfied. If ~ij is not satisfied. and a is commensurable with e. are commen- surable or. they would then bow both sum and product of u· and . and replaced the areas 141 and . It is to this roundabout procedure that all the complications in the proofs are due... and u· + ..... but not b: fourth binomial... 14+". measurable. and hence both 141 and .. and a commensurable.... instead of by means of a and b.. a and b both incommensurable with e: sixth binomial. and hence both uS and . non-measurable... It is possible to express conditions (ij and ~ij in terms of 14 and ". viz.. condition ~ij means that 141 . but not 14 and " and case 1) uS + . case 5) 141 + .. b is commensurable with e. and hence both u· and .. non-measurable.. so that b is not commensurable with e: first binomial. but not ""... hence the discrimination of the 6 cases can also be so expressed. what amounts to the same. case 3) uS + . the Babylonians would simply square the second equa- tion. In Euclid these are defined as follows: If ~I) holds... medial. by the lines x and y whose sum and product are then known. The six cases can therefore be formulated as follows: Let 141 and . not measurable and not commensurable... a and b both incommensurable with e: third binomial. but UJI only medial.. be commensurable.. non-measurable.... From the final conditions (I) and ~ij. Let uS and .. case 6) u' + .. but "" measurable. be incommensurable. measurable... Conditions (I) and ~ij give rise to a division of binomials into 6 subclasses. b is commensurable with e. the square of an area does not have any meaning. the auxiliary line c has again been eliminated.. In X 29-35 it is shown that in each of these six cases. and a is commensurable with e. are commensurable... and uv both medial. line segments exist which satisfy these conditions. for them. . but "V -".. but not a: fifth binomial... so that a is not commensurable with e: second bi- nomial. (Iij ..... that u· and . the root of e(a + b) is again a binomial.. and case 4) uI + . and "" non-measur- able. The Greeks could not apply this method. THE CENTURY OF PLATO 171 (I) a and e commensurable. This is why they introduced the auxiliary line c.. To solve equations (5). In accordance with (9). which do or do not satisfy condition (Iij. because.. The line of thought is simple and at bottom purely algebraic.... case 2) 142 + . Moreover X 29-32 also serve to show that line segments a and b exist..

.va minor. The same fundamental idea prevails throughout the book. p. and all 13 types are mutually exclusive.g. 1 Up to X 28 it goes fairly well. ul + vB and 2uv. because the square (u + .. in case 6).".. the study of the 13 irrationalities is a unit. . Leidcn 1634. etc.. est a plusieurs devenue en horreur.v. 1 .s Iu Mbtooi. For example. All these segments are "unreasonable" because their squares are unreasonable.v. but when the existence proofs start with X 29 (liTo find two potentially commensurable expressible straight lines.. As early as 1585..""".)2 = (u 2 + p2) + 2up is the sum of two medial tit lJrusts••~ JOftI irum. Simon Stevin wrote: liLa difficulte du dixieme livre . In all other cases the line segments u + fI and u . The author succeeded admirably in hiding his line of thought by starting with his constructions. T"'·ill_ Pri_ MdWia tit N_. matiere trop dure a digerer et en laquelle n' aper~oivent aucune uti lite" . and by placing at a still later point the division into 6 types of binomials. because in this case conditions m and (II) are valid. or were these introduced later on? It seems to me that all of this is the work of one mathematician. but these are not used until the higher irrationalities appear on the scene. u + fI is called "producing two medial areas". lOa. In case 1). Propositions X 17 and 18 concerning the measurability of the roots of a quadratic equation precede the introduction of binomial and apotome."i9UU tit Si..IIu . In case 4)..'ul-AII T. For. such that the difference of the squares described on them is equal to the square on a straight line commensurable with the first". The theory of the binomial and the apotome is almost inextricably interwoven with that of the 10 higher irrationals. the bmomlal and the apotome.) one does not see very well what purpose all of this is to serve. even before having introduced the concept of binomial which does throw some light on the purpose of these constructions. e.v an apotome. the square on a medial segment is a medial area. This conclusion finds further confirmation in the close connection between the tenth and the thirteenth books. M.the entire book is the work of Theaetetus.. voir jusqu'a l'appeler Ie croix des mathematiciens. the same methods of proof are applied in all cases. Hence . etc. u + v is called a major and u . and an equally long one (X 73-110) concerning the properties of u .. Su. and to 13 types when the medial is counted in. Thus the 6 cases lead to 12 types of irrational segments u + fI and u . u + v is a binomial and u . Book X does not make easy reading. But who is this author? Has the same Theaetetus who studied the medial.fI are given names to indicate the type of square which they produce. also delined and investigated the ten other irrationali- ties.172 CHAPTER VI There follows a long set of propositions (X 36---72) concerning the properties of u + v. in each of the 6 cases. because in all these cases the squares have mutually exclusive properties. The latter contains a.W-ifuu Juf.-M. etc.c.

Thus we conclude that the author nf Book XIII knew the results of Book X. but now everything is taken up anew and in a different manner. 250. "which haa already been obtained explicitly in Book II. because he mentions them in the Timaeus. but he was not acquainted with its systematic development in Book II. In modern notation." The evident intention of this last sentence is to say that Euclid did not revise this book. whether they can be represented as a binomial or an apotome. J. This makes inevitable the conclusion that the two books are due to 1 See E. an equilateral pentason be inscribed. On the other hand. Indeed it is very probable that the occasion for the careful study and classification of irrationa- lities in Book X is found in the fact that these irrationalities occur as sides of regular polyhedra. the side of the pentason is the i"ational line called minor. the side of the pentagon is Y4 V 10 . Groningcn. namely the cube. he did know the contents of Book X very thoroughly. he would certainly have used the diagram of II 11 (division of a line into mean and extreme ratio).2yt5. this gave rise to the theory of Book X. Dijksterhuis 1 observes rightly. the thirteenth. which would have shortened the arguments considerably. It opens with 12 propositions on the golden section and on the regular pentagon and triangle inscribed in a circle. while the octahedron and the icosahedron are due to Theaetetus. . or if not. This can be formulated in classical form as follows: it is the side of a square whose area is equal to the difference between a measurable and a medial area. for he makes constant use of the classification of irrationals! Thus we read in XIII 11 : If in a circle which has its diameter expressible. but took it in its entirety from an older work. three of the aforesaid five figures being due to the Pythagoreans. 17 that the edge of the regular icosahedron inscribed in a sphere is a minor and the edge of the dodecahedron an apotome. THE CENTURY OF PLATO 173 The theory of the resular polyhedra. These topics have been fully dealt with in Books II and IV. are treated the five so-called Platonic figures. Dr Ele""". If Euclid had written Book XIII himself. the pyramid and the dodecahedron. This book also carries Euclid's name because he embodied it in the Elements. p. as Tannery has observed. Thus Book XIII refers to Book X. "Time and again. which however do not belong to Plato.N """ EucliJes II. A scholium in Book XIII states: "In this book. results are obtained implicitly". They are named after Plato. Dijkstuhuis. but that moreover." The author of Book XIII evidently knew the methods of "geometric algebra". but the reverse also takes place. 1930. Indeed. Book XIII strongly creates the impression of being an entirely inde- pendent treatise. complete in itself. The ques- tion was raised whether such magnitudes can be expressible. we find in XIII 16. Analogously. the theory of Book X was developed with a view to its applications in Book XIIl.

makes it equal to Ar and completes the pyramid. the author knew all of this in advance. moreover Ar is the height of the pyramid. As Book X. It starts again by the con- a. To construct this side and the other magnitudes. proportions appear very rarely in either book. etc. The author never reveals the line of thought which leads him to his constructions by the way of algebraic analysis. so does Book XIII employ chiefly the methods of geometric algebra. e. so that ALI is the required side. form a second regular pentagon AMNEO. He starts at the other end. mentions under Theaetetus that he was the first to write on the so-called 5 solids (or that he was the first to construct them). erects a perpendicular to AB Fig. meeting the semicircle in A and draws ALI. erects a perpendicular to the plane of the circle at its midpoint. Then ALl2 = %ABI. The mid- points of the arcs EZ. then he constructs from them the required figure (such as one of the regular poly- a.a = 1lI/5. required for the construction.e. of the tetrahedron. the statement is made that the square on the diameter will be one-and-one-half times as great as the square on the side of the "pyramid".g. and rA the radius of the circumcircle of the base. We already know his name: Theaetetus. with the aid of a semicircle on the given diameter of a line segment BLJ = r. first he constructs (rapidly and simply. Right at the start. but obscurely) the magnitudes to which he is led by algebraic analysis. he divides the given diameter AB in the ratio 2 : 1 by means of a point r. 55.d hedra) and he ends by showing that this figure has the desired properties. A r B draws a semicircle on AB. Very beautiful is the construction of the icosahedron.174 CHAPTER VI the same author. that it has a circumscribed sphere of unit diameter. in r. Now he constructs an equilateral triangle in a circle of radius r LJ.0 of the decagon. so that OE is the side 21. in the shortest and most elegant manner. In both books geo- metric constructions are cleverly interwoven with algebraic calculations. a regular pentagon EZH8K is inscribed in a circle of radius r. and on each of them the length r is laid off. Thus another such pentagon II PETY is obtained in a plane parallel to that of the first drawmg. Apparently. Our conclusion is confirmed by an undeniable similarity in style and in method of proof. usually called "Suidas". The 10th century compendium "Suda". Each of the vertices IIPETY is connected with the two adjacent . struction. such that . Next. Let us take as an example the construction of the regular tetrahedron inscribed in a sphere of given diameter AB (XIII 13). but he does not betray the source of his know- ledge. Then he proves that the faces are equilateral triangles and that the entire solid can be inscribed in a sphere of diameter AB. i. Now perpendiculars to the plane of the pentagon are erected at the points AMNEO.

Fig. Then the solid is completed. In Definition 5 of Book V we read: "Magnitudes are said to be in the same ratio. De Elementen van Euclides II. The lateral edges of these pyramids are V~02 + r2 = Z6' Thus all the faces of the icosahedron are equilateral triangles. The fragment of Theaetetus concludes with a proposition 'concerning the re- lative size of the edges of the 5 solids.6 and 9 of Theaetetus we could take from the Elements. Euclid's proofs contain logical errors for which we can not hold Theaetetus responsible. so that the solid can be inscribed in a sphere of diameter d. when. consisting of 10 triangles. and any equimultiples whatever of the . Indeed. if any equimultiples what- ever be taken of the first and the third. following Eudoxus. indeed the proof is decidedly inferior in rigor to the other proofs. or to Euclid himself. Thus a wreath is obtained. I see no reason to ascribe the statement to Theaetetus. by a regular five-sided pyramid of height ~o. except the five that have been mentioned. THE CENTURY OF PLATO 175 vertices of the first pentagon. e A Fig. 57. which was not yet known to Theaetetus. and the distance between two opposite vertices always turns out to be r + 2~o =V5r2 = J. for the construction of the dodecahedron. the pro- position is incorrect in this formulation. 56. which is bounded by equal. As Dijksterhuis remarks. Moreover these proofs depend on the theory of proportions developed in Book V. Euclid adds the statement that no other solid can be constructed. this is not possible for the proofs. The formulation of propositions X 5. above and below. the first to the second and the third to the fourth. equilateral and equiangular polygons. it has indeed a very loose connection with the rest. which are proved to be equilateral. The theory of proportions in Theaetetus. The interested reader is referred to Dijksterhuis.

literally "balancing against each other". 6 and 9. According to Aristotle this definition would therefore have to be the following: two areas and two lines are proportional. and na = mb implies nc = mJ. the former equimultiples alike exceed. the latter equimultiples respectively taken in corresponding order. The older definition of Book VII: "Numbers are pror0rtional when the first is the same multiple.176 CHAPTER VI second and the fourth. As we shall see presently. 0 the second that the third is of the fourth". The definitions are therefore of little value to Theaetetus for his non-measurable line segments.g. etc. deduction. is a rectangle. or the same parts. parallel. he could not yet make use of these definitions. but not to incommensurable line Sf'g- ments. If these parts are called band c. are alike equal to. or a parallelogram and the assertion is that the areas. which is used especially for sums of money. Evidently he is speaking of a definition of the concept of proportionality. e. for instance in drawing up the balance sheet. or the same part. because the definitions of the concepts rectangle. are propor- tional to the parts into which the height is divided. or alike fall short of." This says that the proportionality a : be: d means that na > mb implies nc > mJ. without a sharp definition. this ingenious definition is due to Eudoxus. if the areas and the lines have the same antanairesis. found propositions X 5. might use the ratio concept naively. strict logician as he was. in his early youth. The commentator Alexander of Aphrodisias adds here that by antanairesis. Now Aristotle calls this "immediately clear". for the areas have the same antanairesis from them as have the sides: and this is the definition of "the same ratio". But what definition of proportionality did he start from then? A remarkable passage from the Topica of Aristotle (158b) throws some light on this question: It appears also in mathematics that the difficulty in using a figure is sometimes due to a defect in definition. if the line seg- ments a and b have no common measure. The area here referred to. no matter how the integers m and n are chosen. . nor parts. nor a part. into which a line parallel to the base divides the rectangle. subtract. is applicable to numbers. in proving that the line which cuts the area parallel to one side (of a parallelogram) divides similarly both the line which it cuts and the area. then it is a question of showing that lib : IIC = b : c. because a is neither a multiple of b. He must have started from another definition of proportionality. Neither can I imagine that he. the fact asserted becomes immediately clear. and na < mb implies nc < mJ. whereas if the definition be given. from the verb dVT-av-aL(!siv. if the "definition" is stated. are of no importance in this connection. When Theaetetus. But what is the antanairesis? The lexicon derives dvravai(!sa".

1 O. This definition of proportionality makes it easy to prove a number of properties of proportions. because in Euclid VII 2 and X 2. that from a: b = e: d. p. the interchange of the means. it was indeed exactly the proof of this proposition. (a . i. when the antanairesis. one deduces first the equality of the areas ~ ~=k then interchanges b and c. And now it is curious that. . for the determination of the greatest common divisor. if a> b. for line segments. Dijksterhuis and Becker have given.e. e.b) : b = (e . "the proposition could be proved in general". On the basis of this definition. THE CENTURY OF PLATO 177 Aristotle means the same thing as by anthyphairesis. Eut/Qxus. which at first led to difficulties. for. the remainders again equally often. the concept of magnitude). This actually brings us some help. etc. solids and periods of time" (viz.d) : d. the following explanation of the words of Aristotle. Post. the verb dv&v<pawei" means "to take away in turn" the smaller of two numbers or line segments from the larger one. is the following: two magnitudes a and b are proportional to c and d. if a can be subtracted from b (or b from a) as often as c from d (or d from c). Becker has advanced an ingenious hypothesis for this. we can conclude e :d= a: b.g. But after the introduction of the general concept which includes numbers as well as lines. causes difficulty. for solids and for periods of time.l From the pro- portionality (1) a : b = c : J. 15. proceeds with a and b in the same way as with c and d. But one property. Que'''' und Studie" B 2. 311. according to Aristotle.Studi. What was the old proof for line segments? O. The defimtion of proportion which Aristotle has in mind. independently of one another. says Aristotle in Anal. "this proposition was proved separately for numbers. then the rectangle ab can be taken away from the rectangle ac equally often. and. "Formerly".n. etc. viz. Becker. if it is possible to layoff the height b a certain number of times on the height c. Zeuthen. and (a + b) : b = (e + d) : d. On the basis of this etymology. the subtracting in turn of the smaller from the larger. and finally returns to the proportionality (3) a: e = b: J. and b: a = d :e. it does indeed become "immediately clear" that the rectangles ab and ac are proportional to band c. 3. The proof for numbers can be found in Book VII (VII 13).

This is exactly the first proposition of Book X.1 how P can be derived from R. requires use of the so-called "lemma of Archimedes". . This in- dicates that we are on the right track.g. Evidently. II : b = . that (3) was indeed derived from (2) in this manner. and so forth. Follow- Ing his usual procedure. In Euclid the lemma of Archimedes is usually applied in the following form R. the work of Theaetetus. Consequently. If in a proportion the consequents are equal then the ante- cedents (ad and be) are equal as well. it is used exclusively in the proof of proposition X 2: If. then the remainders. But what purpose does X 1 then serve? It serves for the proof of proposition P. then at some time. make it very plausible. among these is R = X 1. would always be multiples of E. The proof of X 2 by means of X 1 is very elegant. why does this proposition X 1 appear here? In Book X. based on the antanairesis-definition. e. Becker has shown (loc. both line segments). since lI:c=ad:cJ=cb:cJ=b:d follows from (4) and (2).. Euclid omitted this part. then a certain multiple nA exceeds B. and if A is less than B. The fragments from Aristotle. when the less of 11110 unequal magnitudes is continually subtracted in turn from the greater. thus obtaining at the same time a criterium for the commensurability of two line segments or two areas. that Aristotle talked about: (~) II : c = ad : cd. which have been cited. but there is not the slightest diffi- culty in proving X 2 without X 1. from the remainder a piece larger than one half of it. because he had already given another . If a piece larger than one half of A is taken from A. 3 he established the theory of the infinite or finite antanairesis. In propositions X 2. based on the antanairesis-definition to which Aristotle refers. Theaetetus began his book with an ex- position of the theory of proportions. O. from (1) follows ad : bd . as follows: if the magnitudes A and B had a common measure E. he started with lemmas which would be needed later on. obtained in the alternate subtractions.g. 1 Let A and r be comparable magnitudes. that which is left "ever measures the one before it. that is formulated as follows: Q. The proof of P. For. the magnitudes will be Incommensurable. The deduction of (2) from (1) involves another proposition. cit. It is probable that the next thing was the theory of proportions. X 1 is not necessary as a preliminary for X 2. now it remains to prove: P. If A and B are comparable magnitudes (e.178 CHAPTBR VI It is this last step which involves the proposition on rectangles. and A larger than r. For. :d= be : bJ. so that the sequence of remainders would have to end after a finite number of steps. indeed constantly diminishing multiples. which in turn is needed for setting up the theory of proportions! All becomes clear now. the remainder will be a magnitude less than r.

Eudoxus was famous not only as a mathematician. he was a model of moderation. on the contrary. makes clear why this new theory (due to Eudoxus) was given pre- ference. But Plato did not agree with Eudoxus' views on the ideas. a walk of two hours each way from Plato's Academy.. joy. In jest his friends called him Endoxus. . king of Sparta. he held discussions on philosophical que- stions with Plato. Of still greater st. except that he and his followers added a number of less im- portant propositions and remarks. 5 and 9-13). The next main division of Book X. mentions that Eudoxus "flourished" about 368. just as pure white is present in visible white and thus pro- duces being-white. he went to Athens to learn philosophy and rhetoric. Theaetetus then proceeded to develop his theory of expressible magni- tudes and of their ratios. Euclid replaced the proofs by others of which the method was borrowed from Archytas and his school. He studied mathematics with Archytas in Tarentum and medicine with Phili- stium on the island of Sicily. a philo- sopher and a geographer. Some years later his friends enabled him to undertake a journey to Egypt. on the testimony of Aristotle.1ture and more famous than Theaetetus is his younger con- temporary. After that he established a school at Cyzicus on the sea of Marmora. on the basis of his theory of proportions (X 4. which attracted a large number of pupils. From Agesilaus. which would mean that he was born around 400. he received a letter of recommendation to the Pharaoh N ectanebus. but also as a medical man and especially as an astronomer. In this part. He was so poor. This does not by any means justify the conclusion that he advocated a dissolute life. The chronicle of Apollodorus. in his native town of Cnidos on the Black Sea. was left practically un- changed by Euclid. the renowned. on the ideas and on the supreme good. nor with his doctrine of pleasure as the highest good. in the Philebus he combats these tenets with a variety of argu- ments. situated between Helio- polis and Cercesura. that he had to live in the harbor- town Piraeus. a chronological classic of later date. He died at the age of 53 years. He also taught that pleasure. When he was 23 years old. cited above. goodness and strength of character. thus causing the being-thus of things. still in existence in the days of the Emperor Augustus. intended to clarify the very difficult subject. one of the most brilliant figures of his time: Eudoxus of Cnit/os. It is reported that in Egypt he learned astronomy from the priests of Heliopolis and that he made observations himself in an observatory. The Aristotle fragment from the Anal. Around 365 he came once more to Athens with his pupils. which is concerned with the 13 kinds of irrational lines. At that time he stood in high repute. or "blended with" observable things. highly honored as a lawgiver. THE CENTURY OF PLATO 179 theory of proportions in Book V. He held the opinion that the ideas are present in. he was moreover an eminent orator. "since all living beings strive after it". is the highest good. post.

that they debated and that they collaborated to the good of the sciences to which they were both devoted. as an example the four concentric spheres for Jupiter. the others serve to account for the motion of sun. . The lack of uniformity in the motion of the moon is not accounted for in this system. the ecliptic. See also Th.%tIIIriscb.. On the second sphere is a great circle. just as the opposite statement of Strabo.. Abh. Leipzig 1877. i. Math. further particulars can be found in the treatise of Schiaparelli. it has moreover a slow rotation in the same sense about the poles of the ecliptic. Heath. The exterior one of these three rotates in one day about the poles of the equator. is almost certainly exagge- rated. The following bird's-eye view of the system will reveal its extra- ordinarily ingenious construction. to explain. 1 1. slightly inclined to the ecliptic. as they become known to us through the best sources.e. Eudoxus was the first to give an answer to this question. This motion serves to account for the "recession of the nodes" of the lunar orbit. Let us consider. this sphere also shares the motions of the first two spheres. which could serve to "save". Each planet requires four spheres. 180 CHAPTER VI This shows that.. Around this center. An oblique circle lies on this first sphere. It is not difficult to describe the motion of the 3 concentric spheres which govern the motion of the moon. Eudemus relates that Plato proposed to the astroaomers the question as to the uniform circular motions.-m. 27 concentric spheres rotate. to say that they respected each other. and it carries the moon. Oxford 1913. this is the orbit of the moon. The third sphere rotates about the poles of this circle. From communications of Simplicius and Aristotle. It is best in accord with the character of the two men. Of these the exterior one carries the fixed stars. SchiApardli. in its motion it carries the two others along. Gesch. This explains why the moon shares the diurnal motion of the stars. The model for the sun is similar to that for the moon. moon and the 5 planets. Schiaparelli was able to reconstruct almost completely the extremely ingenious planetary system designed by Eudoxus. The second sphere is carried along by the motion of the first. A'. I. 1 The spherical earth is at rest at the center. that Eudoxus was beloved by Plato. The period of revolution of this third motion is the draconitic period of the moon. But the model for plane- tary motion is incredibly clever. the planetary phenomena. the sun and moon three each. that Eudoxus and Plato were enemies.s 0/ Saw•. Eudoxus was opposed to Plato on a number of points. which the reliable Simplicius takes from the astronomer Sosigenes. The story of Dio- genes Laertius. Eudoxus as astronomer. Sp""'" tIu EuJ-us. although his philosophic views were closely related to those of Plato.. just as the sphere of the fixed stars. This appears also from the following statement. IN hD.

______ result of the slowly advancing motion of the second sphere. The poles about which the third sphere rotates with respect to the second. Callippus improved the system by adding two each to the number of spheres for sun and moon. 58. motion in a loop. An important part of this is occupied by data concerning the 12 signs of the zodiac. viz_ that it did not supply an explanation for the variable luminosity of the planets. Consequently the axis of rotation of the fourth sphere is inclined to that of the third. whose transverse axis lies in the plane of the ecliptic. but had to make way a half century later for other. that of the fixed stars. of importance for the determination of time during the night. Aristotle adopted all these improve- ments and supplemented them with "retrograde" spheres. the third has three and the fourth four motions. This accounts for the fact that these two rotations in opposite senses do not completely neutralize each other. which aroused admiration in the profession. There is little doubt that Eudoxus gathered this knowledge . the period is the synodic period of Jupiter (13 months). Still greater fame than he reaped with his theoretical astronomy. it was ~=~I='-~-'-~--~-'-~"~---~-'~--~~~l called the Hippopede. becau~c in this model their distances from the earth remained constant. As a ~ --. Eudoxusgarnered from hisdescri ption of the constellations and of the rising and setting of the fixed stars.--. This curve can also be considered as the intersection of the sphere with a thin cylinder tangent to the spliere. But all these emen- dations failed to meet the fundamental objection to the entire system. but they do not coincide with the poles about which tltt fourth sphere rotates. the planet. Front and lateral views of the Hippopcdc as thus producing a to-and-fro the intersection of a sphere and a cylinder. the center of this figure-r:lght curve describes the entire ecliptic in 11 to 12 years. Their rotational velocities are equal but opposite in sense. The second sphere rotates about the poles of the ecliptic in the sense opposite to that of the diurnal motion. would describe a horizontal figure-eight curve (see Fig. 58). If there were nothing but these last two motions. Fig. lie on the ecliptic. and the stars which rise and set at the same time as the initial points of these signs. and one to each of the others. which is attached to the fourth sphere. exhibiting great similarity with the apparent motion of Jupiter with respect to the fixed stars. this serves to explain Jupiter's sidereal period of 11-12 years. The rotations of the third and the fourth spheres are used to account for the alternate direct and the retrograde motions of Jupiter. The second shares this motion and has an additional one. THE CENTURY OF PLATO 181 The exterior sphere again has only one motion. better systems.-___ .

omen. the "Arachne". For example. com"". A fine Persian specimen. after having observed the altitude of a star or of the sun.. 59. is shown on Plate 20. giving data on the first and last appearances of the most important fixed stars. which looks somewhat like a spider and which is indeed called Arachne in the Greek treatises on this instru- ment. Thus one can determine the time of day. Vitruvius. Furthermore Eudoxus designed a perma- nent calendar. All these circles are represented in stereographic projection. lies on the circle of this altitude. The instrument can be made to imitate the daily rotation of the fixed stars by turning the spider in the ring which surrounds it. :>.. Paris 1947. the Phaenomena 1. on which circles have been scratched. a central projection on the equatorial plane from the south pole. and have also ~een translated into modern languages. Michel. A papyrus has been preserved which contains an excerpt of this work. To use the instrument at different geographical latitudes. called the astrolabe. Behind the spider one can see a removable plate. others circles parallel to the horizon at different elevations. and also some pointers. . He alsQ wrote on the construction of celestial globes. adds: "Some say however that Salzburg (sec page 183 and plate 21) Apollonius (invented the spider)" What is this spider? There is an astronQmical instrument.. representing the ecliptic. On the Arachne one sees an excentric circle. Traite de l'astrolabe. from the year 1223. which were read and admired throughoutthe ancient world. this plate has to be changed. namely. the spider. one can rotate the spider until the pointer which carries the name of this star. much in use among the Arabians and the Persians.1.e. It is said that Eudoxus also invented an astronomical instrument. This instrument contains a revolving disc. from whom this statement Fig. indeed the stereographic projection of a circle is again a circle. or the point on the ecliptic which corresponds to the position of the sun.nown chieOy from the critique of Hipparchus (in A.nitius) and from the poem of cd .. ... One of his two works on the celestial sphere. A dis- 1 Ira contents arc Ir.. the reader may consult H. on the equinoxes and the solstitia. or of the night. The entire spider is a "stereographic projection" of a part of the sky. a aIPaiea..182 CHAPTER VI from a rotating celestial globe. For further particulars. Eouloxi ph. Ptolemy and Theon of Alexandria have written treatises on the astrolabe. was later put into verses by the poet Aratus. But. the Roman architect. Reconstruction of the bronze disc of is obtained. i. One of these circles represents the horizon. Several problems can be solved by means of the astrolabe. which carry the names of bright fixed stars.

Eudoxos was also famous as a medical man. exactly the reverse of the astrolabe. I belIeve that it was Apollonius. 41. p. As an example he always quotes the third of the Euclidean axioms which have just been cited. But it is high time now to turn to his most important accomplishments.. Hipparchus and Ptolemy as the inventors. driven by hydraulic means. Things which coincide with one another are equal to one another. stereographic projection. Ptolemy is out of the question. There is another instrument. Jahreshefte osterr . . Bronze""'i.. Proclus says moreover: "Eudoxus added three proportions to the three others and he continued the investigation of the section. Neugebauer. z. His work in geography includes a "journey around the earth' .. at present however there is quite general agreement as to what has to be ascribed to Eudoxus (Plate 22). viz. If equals be subtracted from equals. In this instrument. 59). Eudoxus. Exceedingly vague is the statement in the Proclus catalogue. p.ol. while the rotating disc behind it represented the sky with the zodiacal signs (see fig. Our sources mention Abraham." s. Wien 6 (1903).." Aristotle already knows "the so-called general axioms" which form the founda- tions for all demonstrative sciences and must necessarily be accepted by anyone who wants to gain knowledge. in which a "spider" is found. which represented the horizon and the hour circles. from Eudoxus. which agrees exactly with his deSCrIption 1 (Plate 21). Sources for the mathematical work of Eudoxus are rather meagre and uncertain. and a piece of a large bronze disc was found in Salzburg. Isis 40 (1949). The Roman architect Vitruvius has described it. It is therefore quite possible that Euclid took these axioms. The whole is greater than the part. Both instruments are based on the same principle. Does Proclus perhaps refer to the propositions in Book V of the Elements. Inst. Apollonius. begun by Plato. the wholes are equals. the history of the foundation of several cities. which hold equally well for line segments. to his mathematics. the remainders are equal. concerning general magmtudes and their ratios. etc? The word "general" might also lead one to think of "general understandings" (or axioms). used constantly in the theory of ratios of Book V. areas..". making use of analyses.• reh. which con· tains i. the "spider" was a network of wires. but it might equally well have been Eudoxus or Hipparchus. angles. 240. that Eudoxus "10- creased the number of so-called general theorems"..w. related to the astrolabe. The inventor of this principle must have been a mathematician of stature. but we know nothing further about this.. which are enumerated as follows at the beginning of Book I of the Elements: "Things which are equal to the same thing are also equal to one another. It is some kind of "bad-weather-clock". If equals be added to equals. because Vitruvius lived about 130 years earlier than he. THE CENTURY OF PLATO 183 cussion of these treatises and further references to the literature are found in an article by O." 1 Sec A Rehm.a.

2). Book XII starts with the proof that the ratio of the areas of two circles is equal to that of the squares on their diameters (XIII. then in Book VI (proposition 30). The proof depends on XI: "Two unequal magnitudes being set out. along with the theory of proportions. . there will be left some magnitude which will be less than the lesser magnitude set out. Inscribe then in the second circle a square EZHe. The" discoveries of Eudoxus" will there- fore have to be traced to Books V. by means of enclosing the circle between inscribed and circumscribed polygons. Suppose that the two circles were not proportional to the squares on the diameters BLJ and ze. this last set of propositions is due to Theaetetus. It will be recalled that Hippo- crates' quadrature of lunules was based on this proposition." The proof of the proposition on the ratio of the areas of two circles starts as follows. In reality however the circle is never exhausted. Can we specify them more closely? Let us first have a look at Book XII. it consists of Books X and XIII. then the ratio of BA2 to ze 2 would be equal to that of the first Circle to an area l:. but that Hippocrates was not able to prove it rigorously. is indeed abandoned after a finite number of steps in the proof of XII 2. The proof in Book XII rests on two pillars: the theory of proportions of Book V. and the concept of the indirect proof. first in Book II (proposition 11) in connection with the Pythagorean geometric algebra. whose areas differ by less than an arbitrarily given area. in his Elements. the unhappily chosen name "exhaustion method" is used. which would have to be either greater or smaller than the second circle. Weare brought a bit further along by Proclus' statement that. VI (theory of proportions) and to Book XII. There is no doubt that the material in Books I-IV (foundations of plane geometry without proportions) and in Book XI (foundations of solid geometry) is of older date. The arithmetical books VII- IX can here be left out of consideration. As we have seen.184 CHAPTER VI The "section" referred to by Proclus is probably the "golden section". the attempt at "wearing out". It is treated three times in Euclid. and finally in Book: XIII (propositions 1-6). The exhaustion method. to use the terminology of Dijksterhuis. based on the idea that the circle would finally be exhausted by inscribed polygons of a constantly increasing number of sides. and from that which is left a magnitude greater than its half. Suppose first that l: is smaller. and if this process is repeated continually. and it is shown that what remains of the circle is less than an arbitrarily given area. it is therefore possible that VI 30 comes from Eudoxus." We have already excised the work: of Theaetetus from the Elements. Euclid "collected many of the discoveries of Eudoxus and completed many of those of Theaetetus. For the method used in this kind of proof. if from the greater there be subtracted a magnitude greater than its half.

a somewhat subtler formulation of the proof could have avoided it. In spite of this imperfection. this is a contradiction. 61. in K. the proof remains a scientific accom- plishment which compels admiration. A. The ratio of the first circle to the area E is then the same as that of the polygon in the first circle to the polygon in the second. which is larger than the circular segment. for the inscribed polygons approach the circle in the strict sense. THE CENTURY OF PLATO 185 This is larger than one half of the circle. the existence of a fourth proportional is tacitly assumed. while E is less than the second polygon. Then the ratio of BJ2 and Z8 1 is equal to that of the polygon in the first circle and the similar polygon in the second circle. The in- scribed polygon that has been obtained is then larger than the area E. that their difference can be made less than an arbitrary area. then (by X 1) there must ultimately remain circular. the modern limit concept. Now bisect the arcs EZ. because double each triangle is a rectangle. with full exactness. because a circumscribed square is exactly twice as large as the inscribed square. corresponding circular segment. M and N and construct the octagon EKZAHM8N.'f------. quod erat demonstrandum. Therefore the areas of the two circles have the same ratio as BJI and ze 2 . But the first circle is larger than the first polygon. etc. by bisecting the arcs each time. is larger than one half of the Bt-------J. Proposition X 1 which is used in this proof. hence it also leads to a contradiction. At the beginning of this proof. Interchanging the means one finds the proportion: First circle: First polygon = E : Second polygon.d .8 H r Fig. Now con- struct a similar polygon AS BOrIIJ P in the first circle. If this process is repeated. 60. This assumption is not necessary. segments. etc. whose sum is less than the difference between the circle EZHe and the area E. the same thing takes place in all the analogous proofs in Book XII. The case in which 1: is larger than the second circle is reduced to the first case by an interchange of the two circles. Then each of the triangles EKZ. It contains. while the circle is less than the circumscribed square. Fig. in turn rests on a tacitly assumed .

and by means of a simI1ar lemma they proved 4° that every cone is one third of a cylinder on the same base and with the same height. of exceeding one another". 3° that every pyramid is one third of a prism on the same base and with the same height. What Archimedes attributes here to the "older geometers" is exactly the brief contents of Book XII. Def.. this follows readily from previously proved propositions. he adds that the older geometers had also used this "lemma. who invented the method used in all these proofs? No one but Eudoxus. A preparatory step is XII 3: Every triangular pyramid ABr. 2° that spheres have triple the ratio of their radii. which was formulated in the following way by Archimedes in De Sphaera et Cylindro: "The larger of unequal lines. they proved: 10 that circles have double the ratio of their radii. and if each of them is divided into two prisms and two pyramids. by use of this lemma." "for 2. then the sum of the prisms in the first pyramid is to the A:-="""---~:. But. For. and the proofs do indeed depend in each case on the lemma. Euclid assumes everywhere tacitly that two line segments. As an example of Eudoxus' ingenuity. :This word "for" creates tbe impression that tbe "Ancients" had used the lemmA.. areas or solids exceeds the smaller in such a way that the difference. in the preface to De Sphaera et Cylindro.1 can be divided into two equal and similar triangular pyramids AEH8 and 8KA. the sum of the prisms being greater than the sum of the pyramids. Archimedes ascribes the proofs of 3° and 4° to Eudoxus. can exceed any given individual of the type to which the two mutually compared magnitudes b-:long. the bases of the given pyramids. mentioned by Archimedes. It is however hidden in V. Ardoi. I shall briefly reproduce the proof for the pyramid (3°). always have a ratio. Archimedes formulates the same postulate for areas. etc." 1 In the Qgadratura parabolae. 1 Compare Dijksterbuis. Neither does Euclid formulate it explicitly. which ue apable.186 CHAPTER VI postulate. 62. when multiplied.. 4: "Magnitudes ue said to have a ratio to one another. without having formulated it explicitly. as indicated above.-----~B sum of the prisms in the second pyramid as Fig. added to itself. two areu.1 and two equal prisms BZK-EH8 and HZr-eKA. 139. The proof depends on the fact that prisms with equal heights have the ratio of their bases.Ju I. . Next comes XII 4: If two triangular pyra- mids with equal height are given. p. For XII 3-7 produce exactly the result mentioned under 3° • XII 10 says the same thing as 4° and XII 18 the same as 2 0 • The remaining pro- positions are conclusions from or preparations for these four main theorems.

Pyramids which are of the same height and hafle triangular bases are to each other as the bases. The same result follows for pyramids with an arbitrary number of sides. 63. Furthermore. Magnitudes are said to be in the same ratio. as we have seen. At every step. It is shown in XII 7. There is in Book V a scholium. pyramids is therefore one third of the prism. Proclus also attributes a number of "general theorems" to Eudoxus and because there is more- over a certain unmistakable relatedness between the methods of Book V and those of Book XII." The scholium sounds plausible. when multiplied. when. Both operate constantly with proportions. A ratio is a sort of relation in respect of size between two magnitudes of the same kind. and are therefore equal. 4. Each of these Fig. approxi- mations and inequalities which are used ultimately to prove equalities by a reduc- tio ad absurdum. cited above. THB CENTURY OF PLATO 187 Now comes the key proposition: XII 5. so that in the end there is a remainder which is less than any prescribed solid . Up to this point we could appeal to Archimedes." Nevertheless. which have. The theory of proportions. "it is right that this book carries the name of Euclid. two by two. by means of the exhaustion method. The place of the inscribed polygons is now taken by the prisms which arise from the continued division of the pyra- mids. of exceeding one another. the prisms absorb more than half of the pyramids. and both use. 3. by division into triangular pyramids. was found by Eudoxus. the first to the second and the third to the fourth. if any equimultiples whatever be taken of the first . is entirely analogous to that of XII 2. explained in this book.this is the basis of the proof. Def. Weare therefore prepared to join the scholiast and to attribute Book V to Eudoxus. Magnitudes are said to have a ratio to one another which are capable. is generally considered to be due to Euclid. The proof. particularly since. are the following: Def. 5. according to which "it is said that the general theory of proportions. in a rigorous manner. this proves the principal result for triangular pyramids. the scholia in Euclid usually square with what is known from other sources. that a triangular prism can be divided into three triangular pyramids. both are based on the postulate of Eudoxus. whose author is unknown. because the exposition in the form of the Elements and its adaptation to the system of geo- metry. The definitions on which ~his book is based. according to the same scholium. equal bases and equal heights. but for Book V this indisput- able authority leaves us in the lurch. Det.

:st is said to have a greater ratio to the second than the third has to the fourth. and the same has to the less a greater ratio than it has to the greater. and a multiple of c will exceed kb. 175-6). the greater has to the same a greater ratio than the less has. and magnitudes to which the same has the same ratio are equal. and any equimultiples whatever of the second and fourth. then a: b> c: J.b will exceed c k(a . a:c>b:c. then the fi.f the first part can be formulated as follows in modern algebraic notation: let a > b and let c be a third magnitude comparable to a and b. we obtain from V 8: V 9. 7. After these preliminary propositions.:J=e:! . As a supple- ment there appears Def. so that (m + l)c> kb ~ me. b» c. On this definition is based the following auxiliary proposition: V 8. Of magnitudes which have a ratio to the same. we conclude almost at once: V 12. of the equimultiples. we find the principal properties of pro- portions. makes it possible for Eudoxus to develop the theory of proportions in masterly fashion. but the multiple of the third does not exceed the mul- tiple of the fourth. the former equimultiples alike exceed.. viz. By a reductio ad absurdum. that which has a greater ratio is greater. Magnitudes which have the same ratio to the same are equal to one another. where m and n are natural numbers. Of unequal magnitudes. In a continued proportion a. which we have reproduced and interpreted before (see p. in complete generality for arbitrary magnitudes. the multiple of the first magnitude exce~ds the multiple of the second. Following Dijksterhuis.. The second part is proved similarly. That is: if ma> nb. or alike fall short of. This admirable definition. and that to which the same has the greater ratio is less. A multiple of a . the latter equi- multiples respectively taken in corresponding order. Suppose that this is accomplished by (m + 1)c but not by me. By addition we find ka> c + me = (m + l)c.188 CHAPTER VI and third. 7. V 10. the proof o. But kb «m + l)c. The proposition now follows from def.b=. . are alike equal to. but me> nd. When. From the definition of proportions..

The whole disc had a diameter of 1. Bronze disk. one for every 2 days of the year. 59 and p. Part of the dial of a Roman water clock. The circle. at which the disc is broken off. Rehm (see Fig. representing the sun. A knob.22 m. 21. had to be moved from one hole to the next every two days. PLATE 21 PL. Along the circle were 182 or 183 holes. represents the ecliptic. . which was placed before the rotating disc. The hour of the day was indicated by the position of the knob with respect to the hour lines on the ··spider". reconsttucted by A. 183 with footnote). It was turned by water force so as to make one rotation a day. found in Salzburg.

PL. o-j Euctemon and Callippus. 21 . In the last column.23 form the most valuable part of the teaching. II p. "Art of Eudoxus".8 we read: From Orion's to Sirius' setting days two. Mem. to Euctemon tIl days 90. 23.23 of the papyrus) contain an excerpt from the calendar of Eudoxus (see p. Thus we read in col. 22. On the back. to Callippus days 89. 12 verse lines are written. tv the rest of the text and the drawings are rather primitive. with additions taken from calendars of Democritus. in which the names of the 12 zodiacal signs are written. In coli. to Democritus days 91. Part of a Greek papyrus. 183). the initial letters forming the words "Art of Eudoxus". 49. scientif.8: From fall equinox to winter solstice according to Eudoxus days 92. The columns 21 . '0 nomica) text. znd K. A popular astro. Illustrations in Roll and Codex p. 22. we see a circle representing the sky. line 7 . 407. ~ The 3 preceding columns (col. Weitzmann.C. line 3 . now in the Louvre. . The name Leptinus is in the last line but one. tv See P.. Tannery. written in Egypt between 331 and 111 B. called "Teaching of Leptmus".

ma : !lib = 4 : b. and in an amazingly simple manner. I can not resist the temptation of comparing the modes of thinking of Theaetetus and Eudoxus. then b = d. mentioned by Aristotle: V 16. then mb = nd. then mb < nd. but this solution is not known to us. in one step for arbitrary magnitudes. the two great geometers of this period of florescence. then b > d : and if a = c. From a : b = c : d follows ma : mb = nc : nd. As a special case (equal terms). is a . They are both extremely keen. We do not know to what extent he is responsible for the creation of Book VI (geometrical applications of the theory of ratios). the extremely important proposition on the interchange of the means. THE CENTURY OF PLATO 189 the sum of any number of antecedents has the same ratio to the sum of their consequents as any antecedent has to its consequent: (4 + c + e) : (b + tl + f) . For Theaetetus. The other cases are treated similarly. It this not a masterpiece of logic? The remaining usual properties of proportions follow now without any diffi- culty. But there is a characteristic difference: Theaetetus thinks in a way which we moderns would call "algebraical". by use of V 10: tl <b. and hence c : tl> c : b. He has also given a solution of the Delian problem by means of the intersection of "curves". we conclude from V 14 that. Theaetetus and Eudoxus. they will also be proportional alternately. then b < d. Therefore a : c = b : d. rational or irrational. logical thinkers. one finds for every natural number m V 15. Proof: From a > c follows by means of V 8 4: b> c: b. then mb > nd. and if ma < nc. every line segment. and if ma = nc. If four magnitudes be proportional. if ma > nc. From a : b = c : d follows a : c = b : d. This concludes our discussion of the most important mathematical accomplish- ment of the great Eudoxus. This is preceded by an auxiliary proposition derived from V 8--10: V 14. Hence.4 : b. If a : b = c : d and a > c. And now follows. but from this we obtain. Proof. they are both very ingenious in finding geome- trical constructions. Eudoxus on the other hand is a typical "analyst". and if a < c.

which were introduced only later on by Apollonius. section of an acute-angled cone (ellipse). whether there did not exist for him a shortcut to geometry. no matter how closely they approach it in size. for travellers through the country. derived from arithmetic. His two solutions have already been discussed. as becomes clear from the epigram of Erastothenes (see p. he proceeds similarly with the pyramid. . or whatever you like. the cone and the sphere. which arise in the process of alternate subtractions. and then Menaechmus and his brother Dinostratus. Theaetetus determines the edges of the regular polyhedra from the algebraic properties which follow from their construction. The Proclus catalogue mentions in the first place Amyclas. of whom we know nothing further. 161-2). The most important accomplishment attributed to Menaechmus is the dis- covery of the conic sections. For Eudoxus on the other hand. one does not know whether they are based on actual occurences. line segments are continuously variable magnitudes. on the two meanings of the word element. and from this procedure he obtains indirectly the properties he wants to prove.190 CHAPTER VI separate entity with definite algebraic properties. they . but most probably the original version related to Menaechmus. medial or apotome." The same story is told about Euclid and king Ptolemy. The ancient names of the conic sections are: section of a rectangular cone (parabola). which he used to solve the Delian problem. Menaechmus did not yet use the words parabola and hyperbola. which can approach limits and which can be approximated arbitrarily closely by other line segments. section of an obtuse-angled cone (hyperbola). The greatest of these three is undoubtedly Menaechmus. According to Proclus he increased the number of spheres in the planetary theories of Eudoxus and Calli pus. Theaetetus determines a ratio by means of a sequence of integers. there are royal roads and roads for ordinary citizens. He wrote on the foundations of geometry. When Alexander the Great asked him. After these stars of the first magnitude. but in geometry there is but one road for all.Eudoxus encloses the circle between inscribed and circumscribed polygons. The only cones considered by the "ancients" were cones of revolution. we come to the lesser lights of Plato's circle. constructible in a definite man- ner. sharply distinguished from other line segments. a pupil of Eudoxus. As always with such anec- dotes. and on the difference between theorems and pro- blems. But he did construct his curves as plane sections of cones. But Eu- doxus determines a ratio by means of its place among the rational ratios which enclose it on both sides. who acquired fame as an astronomer and as a geometer. he answered: "0 King.

But it also gives the length of the quadrant LIB and hence the quadrature of the circle. Let the straight line rB rotate uniformly about r so that B descnbes the arc BELl. If rZE is one definite position of the rotating line and Z the point of intersection with the line which movC5 parallel to itself. Pappus describes as follows the construction of this curve: Descnbe a circular arc BELl about r in a B A square ABrLI. which he needed for his duplication of the cube. according to the definition. It is not known. y2 = xb (parabola). and let the line BA move uniformly r towards LI. . VII 30. obtuse or acute vertical angle.o. p. Brwill be to the perpendicular Z11 as the entire arc BLI is to the arc ELI. was the first who used it for squaring the circle.1 Fig. so that both rB and BA wI11 coincide with rLl at the same moment. THE CENTURY OF PLATO 191 distinguished them according to the right. in the section on Apollo- nius. and Eutocius. it is mentioned that Dinostratus and Nico- medes used for the quadrature of the circle a curve.e. Later. remaining parallel to LI. The proof is made by the same mdirect method. It is clear that. it can be used for the division of an angle in an arbitrary number of equal parts. once this curve has been drawn. 168. of Apollon. Let r both uniform motions take place in the same time. 1 Thus they obtained the parabola from the rectangular cone. . where e is the terminal pomt of the quadratrix. In the large compendium of Pappus. which for that reason was called the quadratrix (TETeaycovICovlJa). the brother of Menaechmus. but it seems that Dinostratus. their equations xy = ab (hyperbola). then. how Menaechmus obtained from this method for generating the conic sections. Dinos/fa/us. These two moving lines intersect in a point which moves along with them and which describes a curve BZe. cutting the cone by a plane per- pendicular to a generating line. II (ed. their "symptoms".. we shall return to the history of the conic sections. Heiberg). Collect. i. For Pappus proves that the arc LlEB is to the line segmentBr as Br is to re. It appears from other sources that Hippias of Elis had already found this curve.. the hyperbola from the obtuse-angled cone and the ellipse from the acute-angled one. which is characteristic of Eu- 1 See Pappus. arrd used each type to generate one kind of conic section. Commentary on the Con. which must have been written in the time of the emperor Diocletian (284-305).

To this one can reply that the two motions. Then. we have arc AEB : Br = rA : rK ~ arc AEB : arc KZH. which I made. which is impossible. that the terminal point e of the curve is not denned. extend rz to E and draw the perpendicular ZA. the velocities must be in this ratio. i. gave the result rA : I'8 = 3. both moving lines coincide with rA. appears to me as only partially justified. that the point e is completely defined theoretically by the relation arcAEB: Br = Br: I'8. defined as above. are theoretically determined. since. Sporus says: how can the uni- form motions be defined as long as the ratio of the line AB to the arc BELf is un- known? Indeed. and that practically the point e can also be determined very accurately by drawing a smooth curve through the other points. neither theoretically. But there is something valid in the remark of Sporus.14: 2. which intersects the quadratrix in Z. so that Br=arcKZH. In case the' fourth proportional is less than re. at the end of the motion. if the antecedents in this proportion are equal. then the consequents are also equal. is there anything to object to in the definition of the curve. describe an arc of circle KZH about r. so that arcZK =ZA. 192 CHAPTER VI doxus' exhaustion proofs. Having thus determined the length of the circumference. The critique on the definition of the quadratrix made. and that the curve can readily be drawn practICally by constantly bisecting line AB and also the arc BEA. and. later proved rigorously by Archimedes. Dinostratus knew the proposi- tion. But. then the fourth proportional rK is either greater than or less than I'8. If rK > I'8. the quadrature of the circle? Apparently. That this smooth curve passes exactly through the point e. it follows that Br: ZA = arcHZK: arcZK. by Sporus. so that they have no definite point of intersection. since rK was defined as the fourth pro- portional. Therefore. according to Pappus. because the ratio of the velocities has been defined theoretically as the ratio of AB to the arc BEA. nor from the point of view ofthe draughtsman. thus obtaining an arbitrary number of points on the curve. from the definition of the quadratrix. and hence the curve.&al to . On the other hand one can say.e. a contradiction is reached in similar manner. If the asserted proportion does not hold. follows rigorously from the proof given by Pappus. that the area of a circle is eql. A and K being interchanged. A careful drawing. how does one nnd the area.

. . but not by sight . according to which he was the teacher of Archesilaus. so . he said. how he himself had taunted him and how Glauco then wanted to sing the praises of astronomy after the manner of Socrates: . I would never say that he really learns . for my part.. THE CENTURY OF PLATO 193 that of a triangle. an astronomer after Plato's heart was Autolycus of Pitane. These can be apprehended only by reason and thought.. But how. did you mean that astronomy ought to be taught contrary to the present fashion if it is to be learned in a way to conduce to our purpose? . it will be well to recall what Plato says in The Republic (Book VII 529 A) about astronomy as a science.for nothing of the kind admits of true knowledge .How could it be otherwise than absurd? he said. or some equivalent proposition. that this study compels the soul to look up- ward and leads it away from things here to those higher things.. but would think it absurd to examine them seriously in the expectation of finding in them the absolute truth with regard to equals or doubles or any other ratio. for I do not think. Socrates relates how Glauco had first praised astronomy for its practical value. for apparentiyif anyone with back·thrown head should learn something by staring at decorations on a ceiling. whether gaping up or blinking down. as the fairest and most exact of material things. For anyone acquainted with geometry who saw such designs would admit the beauty of the workmanship.Thus. that one who was an astronomer in very truth would feel in the same way when he turned his eyes upon the movements of the stars? He will be willing to concede that the artlsan of heaven fashioned it and all that it contains in the best possible . whose base is equal to the circumference and whose height equals the radius. but down. then. To appreciate the significance of these works. This agrees with the tradition. am unable to suppose that any other study turns the soul's gaze upward than that which deals with being and the invisible.. For I. since they are decorations on a visible sur- face. you would regard him as contemplating them with the higher reason and not with the eyes.nor would I say that his soul looks up. your rebuke is deserved. the founder of the" middle Academy". said I: these sparks that paint the sky. Then. but we must recognize that they fall far short of the truth.For it is obvious to everybody. we must regard. namely of real speed and real slowness in true number and in all true figures both in relation to one another and as vehicles of the things they carry and contain. . Autolycus is especially important. Perhaps you are right and I am a simpleton. we must use the blazonry of the heavens as patterns to aid in the study of those realities. I think. We can therefore place Autolycus at 320 or 310. because he is the most ancient mathematician- astronomer of whom two works have been completely preserved. . the movements. But if anyone tries to learn about the high things of sense. We shall see presently that Autolycus was older than Euclid.It may be obvious to everybody except me. You seem to me in your thought to put a most liberal interpretation on the "study of higher things".A fair retort.Do you not think. A genuine academician. to be sure. even though he study floating on his back on sea or land. . just as one would do who chanced upon diagrams drawn with special care and elaboration by Daedalus or some other craftsman or painter. said I.. which succeeded Plato's old Academia.

I at least do think so. Where shall we find an ideal astronomy. all of this sounds very strange. Proposition 1 is: When a sphere rotates uniformly on its axis. idealized objects. mathe- ma. "God always geometrizes" says Plato. The works of Eudoxus and Callipus. such as material points and perfect spheres. and that of the month to the year. describe parallel-circles about the poles. ordered the visible umverse. One can not know anything with certainty about observable objects. accustomed as we are to modern empirical natural science. but. as in the study of geometry. based on exact foundations. is possible. like geometry and mechanics. Truth does occur in mathematics because it is ehct. of Heraclides of Pontus are argely lost. In proposition -4 a great circle is introduced which does not take part in the rotation.It is by means of problems. no "truth" tan exist. Nevertheless this is easily understood. said I. about which the sphere rotates. Plato is entirely right. that we will pursue astronomy too. and of their relations to the month. if we are to have a part in the true science of astronomy .tically pure motions. it is concerned with theoretical. where no judgment. Following the pattern of the true. We can discover some of the causes of change. To us.. What does not agree with our present views is this that Plato calls the idealized motions the "true" ones. such as Plato demands. but. hence our word "horizon"). in which we do indeed find an astronomy to Plato's taste: pure kinematics of points and circles on a uniformly rotating sphere. a geometry of moving points. But we do have Autolycus' book On the rotating sphere. . In proposition -4 this circle is perpendicular to the axis.. then all points of the sphere which do not lie on the axis. circles and spheres? For this science was certainly cultivated in the Academia. an of the other stars to these and one another. which move in space in accordance with mathematical laws and to which observable celestial bodies correspond only approximately. but never the totality of all causes. Hence. the Creator has. but when it comes to the proportions of day and night. he said. according to Plato. And truth is divine. in planes perpendicular to the axis. in proposition 5 it passes .194 CHAPTER VI manner for such a fabric. divine. and we will let be the things in the heavens. now that I hear it from you. Plato starts from the position that exact judgments are possible only concerning idealized objects. then.though they possess bodies and are visible objects . the "bounding" circle (oe'Cwv = bounding. at bottom. Theoretical astronomy does not cQnsider actual celestial bodies. because through their variability they constantly escape us.and that his unremitting quest is the realitIes of these things? . do you not suppose that he will regard as a very strange fdlow the man who bdieves that these things go on for ever without change or the least deviation .

must have existed before Autolycus. geometry is taught 1 Thiodo. was collected in the work of Euclid in a manner which was destined to remain exemplary for thousands of years. On the rotating sphere there is also an oblique circle which does take part in the rotation. which has been preserved and in which these propositions are found again with their proofs. an entire world has learned geometry from it. the Phaenomena. Hultsch concludes that a spherical geometry. both Autolycus and Euclid quote. One can not get very far beyond such generalities before the invention of trigo." This shows that Autolycus antedates Euclid. Bruges 1927. We are able to get some idea of what this must have been like from the much later Sphaerica of Theodosius 1. This book deals with the sun and the 12 signs of the zodiac. de Tripoli. nometry. the rotating oblique circle is openly called the "circle through the centers of the signs of the zodiac". His "Elements" constitute one of the greatest successes of world literature. vcr &cl«. because it touches larger circles than those which the horizon touches. with visible an-<l invisible rising and setting of stars. then this second circle WIll rise and set on that part of the bounding circle that lies between the parallel circles which it touches. the horizon is an arbitrary oblique circle. a Sphaerica. are entirely left out of consideration. He proves i. Euclid (Phaenomena VI~ quotes this proposition in the following form: "The circle of the zodiac rises and sets on that part of the horizon which lies between the tropical circles. From this fact. apparently the ecliptic. visible things such as the zodiac in the sky. For.. without proof. Frequently.isinS and senins of stars.a... It was Hipparchus who first determined accurately the times of rising and setting of the signs by means of trigonometric computations with the aid of his table of chords. THE CENTURY OF PLATO 195 through the poles. . propositions concerning circles on the sphere..s. Euclid treats the same topics in his Phaenomena. that some signs of the zodiac require a longer time than others for rising and setting. that shows many points of contact with Autolycus. But AutoIycus remains entirely abstract. LtJ splobi. Even at the present time. in and below the zodiac.. traduit par P. At the end of the 4th century. as cultivated in the school of Plato. but beginning with proposition 6. true to Plato's ideal. in a similar work of Euclid. Less abstract is Autolycus' second book: On the . all the mathematics. Thus propo· sition 11 is: When an oblique circle bounds the visIble part of the sphere and when another oblique grat circle touches grater parallel circles than those which the bounding circle touches.

the first two in the Elements. no matter in how modest a way. who quotes somewhere a proposition from the Elements. 'But what shall I get by learning these things?' Euclid called his slave and said 'Give him threepence. cit. The "Elements" constitute the conclusion of a sequence of similar works. 00 tht 'phtre aod . in his Catalogue. As shown in The Republic and in other works. asked Euclid. Proclus." (Compare Heath. . and school geo- metry is known in England as "Euclid". what excellent judgment he shows in postponing to Book V the difficult theory of proportions of Eudoxus and in treating the most important topi~s of school geometry in Books I-IV of the Elements without the use of proportions! This makes it possiblr. the study of the four subjects: arithmetic. 2. 2nd ed. his modesty and the kindness of his judgment on anybody who aided the development of mathematics. calls him a contemporary of king Ptolemy I (305-285) to whom he had the courage to say to his face that there exists no royal road to geometry. 2 This indicates that the date "about 300" can not be far from the truth. 1 He was older than Archimedes (t 212). Euclid was a diSCiple of the Platonic school. when he had learnt the first theorem. Plato prescribed for everyone. to get hold of the first four books without being scared off at the start by things which are beyond them. I Pappus. This is all we know about the man himself. And there are other such instances. since he must make gain out of what he learns'. geometry. And indeed. astronomy in the sense of Plato. of Theudius. Both Leon and Theudius belonged to the circle of Plato's Academia. all mentioned in the Proclus Catalogue. and. harmony and astrono- my. He is the greatest schoolmaster known in the history of mathematics. Stobaeus relates the following interesting story: "some one who had begun to read geometry with Euclid. Collecti. of Leon. It is exactly these four which are dealt with in the works of Euclid. I. the Elements of Hippo- crates. p. in the book entitled "Phaenomena". But of much greater importance are his works. 3). I. prop.) praises his scrupulous honesty. Pappus (loc. even for mediocre pupils. VII 34: "Apollonius lived for a long time in Alexandria with the pupils of Euclid". What do we know about his life? From a statement of Pappus we learn that he taught mathematics in Alexandria. According to Proclus. 'yliotirr. For example.196 CHAPTER VI in English schools from an English adaptation of the Elements.h. His works do indeed fit perfectly into the tradition of that school. I Archimedes. We may well suppose therefore that their Elements were used for the teaching of mathematics in the Academia. which bear witness even to-day to his unequaled gifts as a teacher. the theory of unIformly rotating spheres. as a preparation for philo- sophy. The thirteen books of Euclid's Elements.. his remarkable didactic qualities fuUy entitle Euclid to this fame. the third in the Sectio Canonis.

It is rather remarkable that a pedagogue of Euclid's excellent qualities is sometimes so illogical. but it seems to me that Proclus exaggerates his praise of Euclid. these parts are on a very high mathematical level. especially from Theaetetus (Books X and XIII) and from Eudoxus (Books V and XII). is due to Euclid himself. We see therefore that also in this respect. but that. This proposition says: "In a right triangle. this apparently means that it occurs in these works. We turn now to a brief discussion of Euclid's other works. Proposition VI 32 is also formulated very carelessly. It is very difficult to say which original discoveries Euclid added to the work of his predecessors. especially the middle one of the arithmetical books (Book VIIQ and the related Sectio Canonis. But this is not an isolated phenomenon. he is himself excellent. Euclid's level is apparently determined by that of the predecesspr whom he follows. since. he also explained the more general proposition by means of an irrefutable argument. THE CENTURY OF PLATO 197 Whenever Aristotle presupposes in his hearers a knowledge of a geometrical proposition. . but this conclusion goes a bit too far. Euclid is by no means a great mathematician. We have already seen that the m03t important and the most difficult parts of the Elements have been taken from other authors. while other parts. Euclid continues the tradition of the Academia. described in similar manner on the right sides". Did Proclus really know Euclid's predecessors? Could he compare their proofs with those of Euclid? We know nothing about this. In connection with the Pythagorean Theorem. found in the Elements. or the translations of Thaer and of Heath. his standard goes down. a figure of arbitrary shape described on the hypothenuse. according to VI 20. fall far below this level. I am sorry. as Dijksterhuis observes. but when he copies from a less eminent author. not a creative genius. These contain" logical errors and the formulations in them are sometimes confused. the proof is not quite complete not even for rectilinear figures. To become more intimately acquainted with the Elements. Moreover. The proof of VI 31 in the Elements is "irrefutable" only for rectilinear similar figures. It can be said that Proclus attri- buted to Euclid himself the more general proposition of the sixth book. in the sixth book. not only that he gave a very clear proof of this proposition. Like the arithmetical Books VII and IX. the interested reader has available the excellent work of Dijksterhuis. For curvilinear figures. the exhaustion method of Book XII would have to be called upon. When he is guided by a first-rate author. these have the ratio of the squares of homologous sides. Proclus says in his commentary on Euclid: "I admire the writer of the Elements. Euclid is first of all a pedagogue. such as Theaetetus or Eu- doxus. is equal to the sum of similar figures." It has been con- cluded from this statement that the proof of the Pythagorean Theorem.

If a magnitude A and the ratio A : B are determined. 198 CHAPTER VI The "Data" is a book of great importance for the history of algebra. If an area is determined in shape and in magnitude. Then analogous propositions concerning polygons and their areas. If X is a definite amount greater than in ratio to y. then each side is determined in magnitude. In geometric dress. Dijksterhuis has called my attention to the fact that "'deter· mmed"' better expresses the meaning. y2 _ ax 2 + C. gives a definite result. 11 introduces the concept" a definite amount greater than in ratio". one would say that X and Y are linearly related: X . or one angle and the ratio of the including sides. A typical example is the follow- Ing: Proposition 19.g. (3) xy = F.5: Proposition 55. after having been diminished by a definite amount C. then X is a definite amount greater than in ratio to Z. then other magnitudes are also determined 1.58 refer to the application of areas. Proposition 7. xy = F. Proposition 86 treats a more troublesome set of equations. for example: Proposition 2. xy = F and (2) x + Y = a. but Dr. and Y is a definite amount greater than in ratio to Z. when two angles are given. If A + B and A : B are determined. Def. propositions 84 and 85 deduce that two straight lines are determined. such . A magni- tude X is said to be a definite amount greater than in (definite) ratio to Y.ratio to Y: (X . In modern terminology. they state that the application of definite areas to definite line segments.aY + C. e. then B is also deter· mined. has a definite .C) : Y = determined. . or their sum and their product. 1 The text spaks everywhere of "'given"'. Propositions 10-21 correspond to certain operations to which such linear equations can be subjected. This leads quite naturally again to the domain of geometric algebra.Y = a. such as substitution. VIZ. then A and B are also determined. without defect or with defect or excess of a definite form. Propositions 57. The work consists of propositions of the following form: when certain magnitudes are given or deter- mined. Next come simple propositions about straight lines given "in position" and about triangles given "in shape". when their difference and their product. From this. We have discussed these propositions at an earlier point.). if X. are given. it is again a question of the old-Babylonian normal problems (1) x .

When two straight lines AB and ALI subtend a definite rectangle and if the square on AB is a definite amount greater than in ratio to the square on ALI. instead of y. then the rectangle AB· Ar is determined.1 Propositions 88-95 refer to circles. therefore Z2 and hence B z are known.. n:t 0re generally. of PiS"'" (1915).e. Namely. or.. zI : (2y-z}l. But the third equation determines the ratio X. I hope that what has been presented gives an idea of the character of the Data. The set of equations (3) is strongly reminiscent of Babylonian algebra problems. Euclid writes ALI. hencezl :y(y -z) is known. 6S and 92 state: If a line is drawn from a point A 'II. i.ision of ftsures. If the point lies on the perimeter. consequently alsozl : [4y(y -z)+zI). AE and EA.C.z) = axl .: y()' -z). etc. But yz is also given...g. THE CENTURY OP PLATO 199 It is formulated as follows: Proposition 86. But I have made no other changes in his reasoning. Thus zI : X. it is transformed into a rectangle C-yz. then AB and ALI are determined. 91 p.. A y!.e. On the di. thus obtaining the product and the difference of f and axI. 1 Of this opusculum only a version in Arabic has been preserved. x I z E I All this is of course formulated in geometrical language. but Euclid accomplishes his purpose by introducing a new unknown z.. the solutJ(~n is very elementary. by means of a line 10 a gI~en direction. . z and y -z.. the ratio z:x-C:F is determined. in place of (3) to a set of three equations: (4) XJ = F. Archibald. 1 Englioh cranslation by R. from which follow z : y and also zI : yz. etc. then the solution of a quadratIc equation is required. C. .oint. y(y . • outside or inside a given circle. or through a given {. E. yz . &diJ's BoM . From the first two of these. into two parts that have a given ratio. it is applied to the width y. The Babylonians would simply have squared the first of these equations. This leads.. This is not possible in geometric algebra. if it lies IOside or outside. carried out of course by means of the application of areas. TherJore the ratio z: (2y-z) is known and hence z: 2y. which cuts the circle in B and in r. Vi. in order to subtract the given area C from known. It deals with the problem of dividing a given rectilinear figure into two equal parts .

o. 8 The Optica deals with perspective. VII 13 ltd. classical and modern mathematicians. . Just as the Babylonians. PseuJaria (on errors of reasoning in mathematics). Hath. the period of Archimedes.'~ "d'lJO'I'OO) with the theory of music. Euclid finds that the square on the line of division is equal to one half of the sum of the squares on the parallel sides. Z Euclid's works on applied mathematics have been preserved. Porismata. Euclid has thus actually collected the whole of elementary mathematics as known at the time of Plato in textbooks. £1 tt seq•. 648). Eratosthenes and Apollonius.. Tonot neoa bmpa'l'elq."".'-"'ics I. • All Greek writings of Euclid art found in the supub edition of Htibug and Menge.200 CHAPTER VI The familiar Babylonian problem of bisecting a trapezoid by means of a line parallel to the base.i. period of florescence. are his pupils. o. p. he and his Alexandrian school initiate the subsequent.<li*. The Catoptrica is concerned with reflections by mirrors. d' E. p. The Sectio Canonis (KaTaToJ. and at clarifying the concept Porisma. Hultsch II. or loci on surfaces). (surfaces as loci. Conica (On conic sections). the Alexandrian era. All of us. the rising and setting of the sections of the ecliptic). Us po. Porismata are a kind of proposition which fall between theorems and constructions. What is this? According to Pappus I. His'"" of Gmk M". appears also. . Lost geometrical writings. at the same time.. and particularly Cauchy.llt. Pappus gives some examples. I have no judgment of my own concerning the value of the various attempts at reconstructing the con- tents of the Porismata. if possible still more brilliant. • Sa Th. He concludes a brilliant period of flowering of mathematics and. Phaenomena with elementary theoretical astronomy (rotation of the celestial sphere. 1 Pappus.

The first Ptolemy founded the Musaeon. philologists (going by the name of grammarians). The same men who brought about the great development of astronomy.).C. mathe- maticians. The arts and sciences of this period have a character totally different from that of the classical era. in "The Birds". philosophers and poets. The classical art of a Homer. Aristophanes could make an allusion to the quadrature of the CIrcle. In astronomy. established not only a powerful kingdom. Alexandria became a flourishing commercial metropolis and also a cultural center of the highest rank. A refined culture reigned at this Hellenistic royal court. All those who cultivated letters and science flocked to Alexandria. historians. The kings Ptolemy Soter.C. The indictments of Anaxagoras and of Socrates for atheism show to what extent the people of Athens were concerned with and ex- cited about philosophy. Hipparchus (130 B. poetry was developed and refined. It included a world-famous library. a Pindar. Aristotle could presuppose a knowledge of elementary geometry. so tradition has it. The culmination of this development was the great Syntaxis Mathematica. receiving ample salaries from the royal treasury. the young hero Alexander (Plate 26) had selected the location of the future world center Alexandria. but they promote~ the arts and the sciences in a truly regal manner. but in his lectures on logic. It goes without saying that. the "Almagest" of Ptolemy (140 A. With the shortest possible delay. were also the foremost mathematicians of their day and brought mathematics to a state of unequaled Horescence. geographers. careful observations were made and theories were established to explain the observations. CHAPTER VII THE ALEXANDRIAN ERA (330-200 B. astronomers. to which Ptolemy Euergetes added the entire collections of books of Aristotle and of Theophrastus. the citizens of Athens did not read mathematical treatises. Aristarchus. In astronomy. and. a Sophocles was written for all the people and was enjoyed by all the people. Archimedes. The works of Homer were analysed and purged of interpolations. and given the start to the construction of the city. but also with a thorough knowledge of geo- graphic conditions and of the possibilities of transportation. but the founda- tions of all these theories were laid in the Alexandrian era. which brought together poets and scholars of the very first order.) With the insight of a man of genius.) and Ptolemy completed the work of their great Alexan- drian predecessors. who succeeded each other during the years from 305 to 322. at the Olympic games as well. in general. the science of chronology was foun- ded. In the leading . Ptolemy Philadelphus (Plate 26) and Ptolemy Euergetes. The exceptionally popular Histories of Herodotus were recited publicly in Athens and.D. such as that of the epicycle and the excenter. Eratosthenes and Apollonius.

Clouds. but for the scholars and the lovers of art. the arts and sciences flourished. Seleucia. Syracuse.C. but the external circumstances are evident. relieved only now and then by a temporary revival. and Apollonius lived there as well. would fall behind and would therefore appear to fly off in the opposite direction. at the princely courts of Alexandria. Archimedes resi- ded at the court of Hiero and of his son Gelon in Syracuse. bridge construction and army organization. Aristarchus advanced the bold hypothesis that the earth rotates in a circle about the sun. there existed an unprecedented urge for know- ledge. Plato's Aca- demy and the school of Eudoxus in Cyzicus drew pupils from near and far. the arts and sciences languished. . the arts and sciences were no longer a matter for the people. in times of prosperity and in times of distress. The kings who came after Ptolemy III did not wish to devote money to science. If the earth had such an enormously rapid motion. a hunger for culture. Every one knew: knowledge is power! Indeed. for in 280 B. in Ptolemy. it may well be doubted whether Alexander could have conquered the world without his technically perfect engines of siege. as Archimedes tells us also. in Asia Minor. Indeed. Eratosthenes was the librarian in Alexandria. in Hellas and on the shores of the Black Sea. of the unsurpassed geniuses Archimedes and Apollonius. for instance. in view of the status of mechanics at that time. But during the era of Hellenism. The sophists received royal honoraria. then everything that was not clinched and riveted to the earth. for the highly cultivated circles. developed more fully. the study of books was imperative for the mastery of the ne- cessary foundations and methods. At the end of the preceding chapter we discussed Euclid. Plato and Aristotle. Wherever the Greeks lived. without a staff of geographers and of financial-economic experts. in North Africa.C. because he represents less the initiation of a new florescence than the conclusion of the previous one. As related by Archimedes in the "sand-counter". without the training in politics given by men like Thucydides. the oldest of the Alexandrian mathematicians. he observed the summer solstice. We shall have to return to the inner causes of this decay. was followed by centuries of decay. Aristarchus of Samos lived at the beginning of the 3rd century. When royal favor and royal support ceased. without the Greek naval architecture. free and untrammeled. there are weighty arguments against the motion of the earth.202 CHAPTER VII circles of the principal Greek cities. on the Ionian isles. of the versatile scholar Eratosthenes and of the ingenious geo- meter Nicomedes. in Italy. Science was largely tied to libraries. says Ptolemy. We proceed now to the period of highest flowering of ancient mathematics. The period of flowering of mathematics in the 3rd century B. Such arguments are already found in Aristotle and. to the era of the keen-witted mathematician-astronomer Aristarchus. Most astronomers rejected this hypothesis.

is the very interesting short treatise "On the distances of sun and moon". We know this also from the "sand-counter". In of s. when the plane. 66. If the earth does not drag the clouds along. is that it would require uninterrupted changes in the apparent distances between the fixed stars. i.. sin 3°). The ratio of the diameter of the sun to the diameter of the earth is greater than 19: 3 and less than 43 : 6. Heath.i. since the Greeks did not know the law of inertia and required a force to account for every motion. It would carry us too far afield to go into this more deeply.e. there is nothing to be said against this. passes through our eye. The only one of the works of Aristarchus which has been preserved. with a right angle at .. Oxford 1913. derived from observation. In reality this angle is not 87°. the centers A. because it gives for the first time a method for approximating the sine of a small angle (viz. From the point of view of Greek dynamics. We do not know how Aristarchus met these arguments.... Aristarchus deduces in this book the following propositions in a rigorously mathematical manner.. To meet this difficulty. and moon form a triangle ABr. uses eclipse observations in an ingenious way. 1. This proof is important for the history of trigonometry. which separates the light part of the moon from the dark part. and in which the angle A is equal to 1/30 of a right angle. The question is to prove that the hypothenuse I T.. earth r O~==~~==========~T A B Fig. It depends on the hypothesis. It is a great merit of Sir Thomas Heath that he called attention to the mathematical value of this treatise and that he published a translation with an excellent historical-astrono- mical commentary.. Copomicw. but the mathematical derivation does not lose any of its elegance from this discrepancy. 1 From certain "hypotheses". The "". Aristarchus assumed that the sphere of the fixed stars has a radius so large that the earth's orbit could be considered as a single point in comparison with it. 2. The distance earth-sun is more than 18 times but less than 20 times the distance earth-moon. 3. it amounts to the following: At the instant referred to in the hypotheses. but 89°50'.. THE ALEXANDRIAN ERA 203 would be overtaken by the rotation of the earth and would hence lag behind. The proof of 3. Another argument against the rotation of the earth around the sun. the angle between the lines directed to sun and to moon is exactly a risht ansle less 1/30 of a riSht angle. Band r of sun. The diameters of sun and moon have the same ratio as their distances. that at the instant at which the moon is exactly in the second quarter. but we do want to discuss the proof of proposition 1. A. they have to lag behind.

and Rj30. we find therefore BA : BK < 10 : 1.oI:4:-"===::=::IE (2) ZE :HE> 15 :2. 68. (3) LIE: ZE = (LIZ + ZE) : ZE> 12: 5 = 36 : 15. Indeed. If BLI is the diagonal of the square ABELl.. BZ is the bisector of angle EBLI and H L EBH = Rj30. r Fig.we obtain LIE: HE> 36 : 2 = 18: 1. Aristarchus calculates as follows (his procedure is somewhat abridged here): BLiI : BEt = 2 : 1> 49 : 25. then the tangent inequality gives BI. but of right sides and of chords of circular arcs. 67. but less than 20 times the side Br. a fortiori: (4) BH :HE> 18: 1. From (2) and (3) . and 1/10 of a circumference. LlZ:ZE=BLI :BE>7:5. the tangent inequality occurs akeady in Euclid's Catoptrica (proposition 8). If BK and BA are these chords. To estimate EZ.:1 but supposes them to be known. 1:=======:!. Fig. corresponding to inscribed angles of Rj3 and Rj30 respectively.) and (5). so that.:ii'-:"!E and BE is a diameter. equal to 1/. to such a degree of accuracy as to enable him to conclude that the . In the proof Aristarchus uses inequalities which can be formulated as follows in modern terminology: Let CIt and p be acute angles and CIt> p. and he does not prove the inequalities A . Related to this reasoning is the method used in Archimedes' measurement of the circle to determine the perimeters of the inscribed and circumscribed regular polygons of 96 sides. The proof of the proposition is contained in (4. He applies the sine inequality to the chords of two arcs.~F. Aristarchus applies the tangent inequality to two angles of Rj4. (5) BE : BK = 2BA : BK < 20 : 1. Then we have tan CIt CIt sin CIt (1) tan p> p> sin p' He does not speak of tangents and sines. where R is a right angle.204 CHAPTER VII AB of such a triangle is more than 18 times.

I (3rd ed. 51. Muller and O. Vogel. : xt = v5: 1> 265: 153.: Xa . Cantor.. since v349450> 59P/•. this leads to ..> 2334 1/. 1351 153 < v3 < 780' To calculate the side of the circum- scribed polygon.. hence. Archimedes' method of obtaining (8) closely resembles the derivation of (3) by Aristarchus.u. : xl = 2 : 1 = 306 : 153. : 153. p.(zt. Mlllh . and Za : Xa> 1172 1/. we have (6) . : x. The side of the circumscribed polygon of 96 sides is 2x6 the diameter is 2r..). tier Malh . + r) : XI> (306 + 265) : 153 = 571 : 153. Bisecting once again. 281 and p. Now the angle HE]' is bisected and . I shall briefly explain his procedure here. lI..mber. : 153. 1 For conjectures on this point see C. 153. + XII) : xa 2 > (571 2 + 1532) : 1532 = 349450 : 23409. : x2• (8) . hence the ratio of the diameter to the perimett!r of the 96-gon is larger than 4673Yz : 96 x 153 = 4673Yz : 14688. THE ALEXANDRIAN ERA 205 perimeter of the circle is less than 3 1 /rJ. Archimedes now continues as follows: Zit: X21 = (. Tocplitz in Que/len rnoJ SluJins. without stating where he got these1 • 265 . Fig.> 2339 1/. 69. p. Archimedes starts with a radius Er = " a tangent line rz and r E an angle r EZ equal to Ys of a right angle. : XI and Zs : Xa. in which H divides the side rz. K. 316: Th. . and finally . but more than 3 1°/n d. we find (9) Zz : xa> 591 1 /. then the segments XI and . Now. J. and Zt. 286. (7) zt. and z. Guclt... + r) : (x2 + yzl = . so that (zt. D. p. p. : xs> 4673Yz : 153. : X. Compare also M. : Xa and Za : Xa are calculated just like . : xa> 1162 1 / 8 : 153. Setting rz = Xi and EZ = Zt. Z Archimedes starts with two bounds for v3.. Mlllh. 41. : 153. Z. V. 8.. where d = 2r is the diameter. Since we shall presently be discussing Archimedes. Archimedes finds that . if EH = Zs is the bisector of angle E. are proportional to the sides. Heath Histo" 0/ Grttlr.

the side of the inscribed triangle.60° = 12°. His method 1 Following the German translation of Manitius. which subtends an arc of 720 .= 1/120 of the diamettr. when the chords subtending the arcs AB and AL-----------. the side of the inscribed square. Now he deduces the well·known "Theorem of Ptolemy": the rectangle on the diagonals of a chordal quadrilateral is equal to the sum of the rectangles on the two pairs of opposite sides. takes the diameter of the circle as 120p and calculates to begin withl. Archimedes proves in a similar man- ner that the perimeter of the circle exceeds 3 10/ 71d. subtending an arc of goo. that the side of the inscribed hexagon which subtends an arc of 60°. which subtends an arc of 36°. I' = p/ffJ and 1" = 1'/60.206 CHAPTER VII But the last of these numbers is less than 3 1 / 7 times the first.r) = 37p 4'SS". Ptolemy calculates sexagesimally. the side of the inscribed pentagon. . 36°. he can now calculate the chords of the supplemen- tary arcs. Thus.2 + t2 = 70p 32'3". so that the perimeter of the circumscribed 96-gon. It served to compute the chord Br. Starting with these chords.g. he could calculate in the same manner the chord which corresponds to the sum of two arcs. By taking two chords AB and AI' in opposite senses. subtending an arc of 120°. 70. we use the following notation: 1. is equal to V. the chord of an arc of 72° . is equal to V3. occur again in Ptolemy (ISO A. we do not learn what use Ptolemy made of it. Ptolemy found.D. and a fortiori that of the circle is less than 3 1/ 7d. for the chord of an arc of 144 the supplement of 0 . already used by Aristarchus. By consideration of the inscribed 96.. is equal to t = Yz(vS.) who used them in an essential manner in the computation for his famous Tables for the lengths of chords. because the sum of the squares of supplementary chords equals the square of the diameter. When we learn this proposition in school. is equal to 6Op .2 = 103p SS'23" . which sub- tends the difference of the two arcs AB and Ar. Thus. the side of the inscribed decagon. arcs BLJ and rLJ are known. The inequalities (I).:~'" Ar and those subtending the supplementary fig. he finds: chord 144° = 114P?'37". e.2 .gon. is equal to VOl = 84p Sl'lO" .

. he computes successively the chords of arcs 2 0 . and which enables him to determine the chord of one half of an arc whose chord is known.C. chord %0 =OP47'8". who lived three centuries before him (150 B. and on the other hand more than one same amount. chord 10 < 1P2'50". He takes therefore Chord 10 = 1 P2'50". Then. and secondly that Chord 1%0 : chord 10 < 1 % : 1.) has written a work "On the theory of the lines in the circle" which certainly contained a table of chords. By use of the half-angle formula he finds Op31 '25" for Chord %0. he derives the following proposition: If two unequal chords are constructed in a circle. We have seen that this theorem was already known to Aristarchus and that he used it in his estimate for sin 3 0 • From this proposition follows firstly that chord 10 : chord 3/40 < 1 : 3/4. which is equivalent to our modern formula 2 sin· %at = 1 . Tannery conjectures that Hipparchus and Ptolemy derived their method of calculation from Apollonius. by means of the addition formulas. chord 10 < 4/3 x OP47'8". In order to calculate the chord of arcs of 10 and %0 . thus reasons Ptolemy. we can set the chord equal to I p 2'S()"". "without appreciable error". etc.cos IX. sin p. Ptolemy obtains furthermore a formula. But this is a mere conjecture. 2Yz 0 . 1 %0 and %0 : chordlYz° = IP34'15". and thiS better approximation may indeed have been the starting point for a table of chords. For Hipparchus. "Since the chord of an arc of lOis on the one hand less. the ratio of the longer chord to the shorter is less than that of the corresponding arcs. Chord 10 > 2/3 x IP34'15" Chord 10 > lP2'50". probably increasing 10 at a time. By use of the Pythagorean Theorem. Hipparchus used the table for the computation of the instants of rising and setting of fixed stars and of the signs of the zodiac. Thus he finds successively the chords for arcs of 6 0 . We only know that Apollonius gave a better approximation of 1t than Archimedes. Y. THE ALEXANDRIAN ERA 207 amounts to our formulas sin {at± {Jl = sin at cos P± cos a. to 1800 • It is in this way that Claudius Ptolemy obtained his table of chords. But he was not the first to calculate such a table.

because. It is said that Archimedes once gave Eratosthenes a practically insoluble problem. "he would not begrudge mathematicians the pleasure of finding out for themselves". 411. Thinking about this problem.. he was the sonl of the astronomer Phaedias.25 (1918). Everyone knows the story of the golden sacrificial wreath of king Hiero (Plate 26). 104 (1883).C. .. St".. addressed to Eratosthenes.hl. without supplying proofs. Dijksterhuis. C. we can not restrict ourselves to the few established facts. . A". of which Archimedes had to determine the purity 3. 255.Je. 1938..I. which had to satisfy nine conditionsl . BIa.ies about Archimedes.. yellow and dappled cows and bulls. who had written a treatise on the diameters of sun and moon. at first usually without proofs. "so that those who pretend to have discovered every- thing themselves. But to playa trick on his conceited colleagues in Alexandria.-f9otuI . . it is true that he moved in the highest circles at the court of Syracuse and that he was on terms of friendship with king Hiero and with his successor Gelon.500 digits! • For further particulars see E. Sometimes years passed before Archimedes. Was this spiteful remark intended for Eratosthenes? One is inclined to think so.4.729. be a multiple of 9304. A_. there lived in Syracuse. far from the wo~ld center Alexandria. The equations ultimately lead to a "Pell equation": . • See R. The least solution i. which called for the numbers of white. when one reads the ironically eulogistic introduction to the "Method". for raising water. published complete proofs of his propositions.. also in use in the Spanish silver mines (Plate 23). he 1 This at least is the interpretation given by F.hi.. 80 enormous that the number of cattle would require a number of more than 206. the brilliant gemus Archimedes. but we must also take into account the legends which have gathered around him and his fabulous discoveries. 208 CHAPTER VII During the middle of the 3rd century B. His friendship with the Alexandrian astronomer Conon of Samos also points to a residence in Alexandria. TillCrt. 2488. may fall into a trap by asserting to have found something which is impossible". Groningen (Noordhof).I. he sometimes slipped in a few false propositions. as he wrote himself in the introduction to the work "On Spirals". problno. the greatest mathematician of antiquity. MIIIII. called Cochlias. black. from which I have borrowed a great deal. No. From Syracuse he used to communicate his mathematical discoveries to Conon by letter. Diodorus states that there he invented a hydraulic screw. of a puaage from Archimedes' ·'Sand·Counter". It is not known whether he was of distinguished origin. It is probable that he has been in Egypt. M"". p. urged by his friends. p. Archibald. According to his own report. with the subsidiary condition that.u. To obtain an adequate idea of his personality and of the tremendous impress of Archimedes upon his contemporaries. AsITO•• Nodtr.

Archimedes is believed to have spoken the winged words: "Give me a spot where I can stand. after a long siege. which. so-called scorpions. they took to /light exclaiming that Archimedes had once again in- vented a new machine for their destruction". . Livy and Plutarch tell elaborate tales about the machines." 1 Briu<llS \V. so he taunted his own technicians. byoccupy- ing himself in some tangible manner with the demands of reality".rmed giant in Greek mythology. invented by Archimedes. and then ran home naked. cranes at the seashore which.. under the personal direction of the 70-years old mathemati- cian. In 212. smaller catapults. a wonder of technical accomplishment. dropped big rocks or heavy blocks oflead on the Roman ships. and. and to reveal his mind to ordinary people. He designed an apparatus which could be operated by one man alone. by weighing the water that is spilled from a filled vessel when the object is immersed.ts a hundr<d . THE ALEXANDRIAN ERA 209 went into the bath. was to be launched. and then to /ling them down suddenly upon the water (Plate 23)." Polybius. Marcellus finally succeeded in taking the town_ For Archimedes himself. in general. and promptly got the idea for determining the volume (either. in their beauty and their excellence. turned outwards. repelled the attack of the Romans on the city of Syracuse. "Shall we have to continue to fight this geometrical Briareus"l. who has knocked out our bat- tering rams and who surpasses the mythical giants with c1 hundred arms in the many projectiles which he drops on us at one time?" But the Roman soldiers were scared to death: "If they only saw a rope or a piece of wood extending be- yond the walls. eureka!" When the famous ship Syracosia. Archimedes will be believed no matter what he says". every skill which is exercised for its practical uses. these mechanical inventions were but the "byproducts of a playful geometry. which he used to cultivate at an earlier time. which king Hiero had ordered built for his collegue Ptolemy. or according to others.. Marcellus. there was difficulty until Archimedes was called into consultation. remain entirely outside the realm of necessity. so writes Plutarch in the Life of Marcellus. "who empties the sea with our ships. he did not want to leave behind any writing on these sub- jects. and I shall move the earth. And again: "Although these discoveries had brought him the fame of superhuman sagacity. which threw a hail of projectiles through the embrasures. On a similar occasion. powerful catapults which dropped from afar heavy rocks on the Roman legions. equipped with the most refined luxury. by measuring the upward pressure). when king Hiero emphatically urged him to direct his art somewhat away from the abstract and towards the concrete. as lowbrow and ignoble. so writes Plutarch. shouting "Eureka. and he only gave his efforts to matters which. or which reached out with iron grip to lift the bow of the ships. he considered the construction of instruments. as Vitruvius believed. The king himself launched the ship and exclaimed: "From this day forward. who personally directed the assault of the /leet stood aghast.

Apuleius and Hieronymus. draws his lines on a board which lies on the floor. to the fact that we have here a copy of an antique prototype. which Winter has brought together. Berlin 1924) is right in thinking that the sandbox. The fact that in our mosaic. 82. in his dissertation (Zur Kunst der romischen Repu- blik. he would draw geometric figures in the ashes. and he would draw lines on his anointed body with his fingers. in the right foreground. Indeed. all the other more recent writers. in truth a prisoner of the Muses. he was forcibly driven to the bath and to chrism. A later legend has it . and of the attitude of Archimedes and of the folds in his garment that the mosaic is not antique. by an examination of the cuirass and the cloak of the soldier. F. in spite of the explicit orders of Marcellus. such a sandbox was placed on a table. Archimedes. asked him (according to Plutarch) to wait a few moments until he had finished the solution. This dramatic episode is represented in a famous mosaic. have always thought that Archimedes wrote in the sand on the floor. On other paintings. indicates an antique prototype. abacus) covered with sand. the sandbox is placed on a table nust therefore be attributed to the happy inspiration of a renaissance artist or. It is an open question whether the soldier had intended to kill him from the start or merely wanted to lead him to Marcellus. and slew him in anger. he forgot to take food and he neglected the care of his body. the scholars draw in the sand on the floor itself." At the sack of Syracuse in 212. This becomes clear from the statements in Iamblichus. But the ancients made their geo- metric drawings on a board (plinthium.if not true. Valerius Maxi- mus and. W inter also calls attention to this. Winckelmannsprogramm. it was said that this mosaic came from Herculaneum. abacium. and this is more probable. but that it comes from the school of Raphael. a mathematician. In Raphael's Athenian school. and when. reproduced in Plate 24a. don't touch my figure".210 CHAPTER VII Plutarch says that he was possessed by mathematics. so possessed was he by a great en- chantment. on his authority. that the expression of Archimedes' head much resembles that of a head from Herculaneum. when he did not obey immediately. lost in the study of a figure. W. and placed on the grave. a scholium in Persius speaks of "the table into which the geometers describe their loci and their measures. Marcellus rendered all honms to the relatives." Apparently this fact was no longer known during the renaissance. it seems to me that F. which he had drawn in his sandbox. Nevertheless. "Constantly held in thrall by an ever-present Siren. as was often the case. But. Goethert made it highly probable. For convenience in use. considered to be of Democritus (see Plate 24b and c) and that the attitude of the soldier strongly recalls that of a warrior on the sarcophagus of Alexander. which Archimedes has before him. Winter (Der Tod des Archimedes. Berlin 1931). At the auction of Jerome Bonaparte's estate. it makes a good story - that Archimedes said to the soldier "Fellow. in . a Roman soldier killed the grey-haired mathematician.

But his greatest accomplishment in the field of astronomy.. and then one saw how the sun disappeared and how the moon entered into the shadow-cone of the earth with the sun on the opposite side .. We recall that. much greater than the radius of the earth's orbit. was the construction of a planetarium. but little.2 The book which Archimedes himself wrote about this "On the making of spheres" has been lost. The principal purpose of the Sand-Counter is to explain. because anyone who wants to give himself the 1 Claudius Ptolemy. while the distance to the fixed stars is much. The works of Archimedes will not be discussed at great length. to Saturn and finally from there to the realm of the fixed stars. he calculated the number of stadia from the earth to the moon. even if one were to suppose that the universe has the size assumed by Aristarchus. ". if it were entirely filled with sand. that the volumes of these solids have the ratio 3 : 2. this would require a larg~ book by itself. About Archimedes as an astronomer we know. see the moon rise above the earth's horizon after the sun. And it is not necessary to do this. from the moon to Venus. with an inscription concerning Archimedes' greatest dis- covery. Archimedes seems to have favored the view of "most astronomers" that the earth is at the center of the universe. Syn. to the sun. to Mars. De republi. by means of an effective notation for large numbers. the moon and the 5 planets. he made an error of % day in the observation of the solstices. In the Sand-Counter. at every turn.axis III 1. that would be contained in the universe. in the system of Aristarchus. unfortunately. • CIcero. The text breaks off in the middle of a sentenco. the earth revolves about the sun. that. especially admired in antiquity. a revolving open sphere with internal mechanisms with which he could imitate the motions of the sun. thence to Mercury. just as occurs in the sky every day. "one could. I 14. THE ALEXANDRIAN ERA 211 accordance with Archimedes' own wishes. circum- scribed about a sphere.. to Jupiter. a small error indeed. According to Macrobius. a representation of a cylinder. . According to a statement of Hipparchus1 . he still found the monument. it is a simple matter to write down a number greater than the number of grains of sand. writes Cicero who had himself seen the apparatus. When Cicero was quaestor in Sicily. he describes an apparatus which he used to measure the sun's apparent diameter. on the whole a not very reliable Roman writer. "When Gallus set the sphere in motion". with sphere and cylinder over- grown with underbrush and thorns. A single turn started the entire complicated movement with the varying periods of rotation of the different celestial bodies.

. as Wallis said. namely by the use of mechanics. of which Volume I was published by P. to a survey of the contents of the principal works.. For it is clear that they were not discovered by the steps which lead up to them in the finished treatises. although not furnishing scientific proofs of them. weighing elements of a figure against elements of another simpler figure the mensuration of which was already known. In this book Archimedes tells us how he discovered certain theorems in quadra- ture and cubature. and a discussion of a few of the important methods of proof.1 I shall therefore restrict myself here to some general characterizations. the Danish philologist Heiberg. he says. so happily discovered by Heiberg. and (2) the rigorous demonstrations of them by orthodox geometrical methods which must follow before they can be finally accepted as established: "Certain things". moreover there exists in Dutch the excellent work of Dijksterhuis. the gradual revelation of the plan of attack. "It is not possible to find in geometry more difficult and troublesome questions or proofs set out in simpler and clearer propositions".a. Heath (History of Greek mathematics II. . while he wished to extort from them assent to his results". although they had to be demonstrated by geometry afterwards because their investigation by the said method did not furnish an actual demonstration. And indeed (again in the words of Wallis) "not only Archimedes but nearly all the ancients so hid from posterity their method of Anal} sis (though it is clear that they had one) that more modern mathematicians found it easier to invent a new Analysis than to seek out the old". 20) has the following to say about the general character of the works of Archimedes: The treatises are. 15-17 and 20 (1938-44). the second part wu printed only in the periodical Eutli. "as it were of set purpose to have covered up the traces of his investigation. A partial exception is now furnished by The Method of Archimedes. by the method. are so impressive in their perfection as to create a feeling akin to awe in the mind of the reader. But it is of course easier. to supply the proof than it is to find it without any previous knowledge. Czwalina or Ver Eecke. p. without exception. German and French trans- lations of Heath.. "first became clear to me by a mechanical method. of Euclid and of Archimedes). Noordhoff in 1938. monuments of mathematical exposition. As Plutarch said. A~ the same time he is careful to insist on the difference between (1) the means which may be suffic- ient to suggest the truth of theorems. If the geometrical treatises stood alone. There is at the same time a certain mystery veiling the way in which he urived at his results. when we have previously acquired. some knowledge of the questions. the masterly ordering of the propositions. as if he had grudged posterity the secret of his method of inquiry." We begin with a discussion of this work from which we can best acquire an understanding of Archimedes' line of thought_ The "Method". Archimedes. went to Con- stantinople to study a papyrus from the library of the San Sepulchri monastery in 1 It is regrettable that the book' has remained incomplete. vola. In 1906. who had already provided us with so many excellent text-editions (i. the finish of the whole. Archimedes might seem. the stern elimin- ation of everything not immediately relevant to the purpose.212 CHAPTER VII pleasure can read them himself in the excellent English.

EA through an arbitrary point 0 of the parabola. originally covered with Greek letters. within the parabola. which was first obtained by Archimedes himself in his treatise on the equilibrium of plane figures. obviously a mathematical text with diagrams. Heiberg succeeded in restoring and deciphering nearly the whole of the old text. Thus far everything is completely rigorous. 1. Now comes the crux of the method: because the triangle AT Z consists of all lines (like EM). but in addition the extremely important work "Method". parallel to the axis of the parabola. which had been believed to be lost. the midpoint of AT. parallel to AB. Draw also ME. It follows from properties of the parabola. therefore the triangle Ar Z. We shall now briefly indicate the contents of this work. and because the parabolic segment ABr consists of all lines. Draw AZ. In the positions which they occupy. like EO. so that K is their common center of gravity. Area of the parabolic segment. placed where it is. if transferred to the point e. such that KX = Y:JKr. For. placed with its centroid at e. parallel to Fig. at the other end e of the lever. Let ABr be a parabolic segment. It contained p. Since now the lever arm Ke of the parabolic seg- . it follows then that this segment TH will be in equilibrium with the segment ME. The same conclusion holds for all segments drawn in the triangle AT Z. which Archimedes assumes as known that BA = BE. afterwards scraped off by monks and written upon anew. and furthermore (1) EM:EO = Ar:AE=Kr:KN= Ke :KN. a line segment TH = EO. The first example of Archimedes is at the same time best suited to explain his mechanical method. Through A. Now consider re as a lever with ful· A crum at K. parallel toAE. Now Archimedes states that the segment ABr is 1Ya times as large as the triangle ABr. Extend rB beyond its intersec- tion K with AZ and make Ke = KI'. to its point of intersection Z with the tangent line r rEZ. Frem the law of the lever. the pro pro· tionality (1) states exactly that the weights of the two segments are inversely pro· portional to their lever arms. bounded by a straight line Z AT and a parabola ABr. they are in equilibrium with their sections within the parabola. For the centroid of the triangle ATZ is at the point X. so that NE = NM and KA = KZ. Now we suspend.uts of various known works of Archimedes. 71. It was a so-called "palimpsest". Now we are practically at home. a lineA BE is drawn. which can be drawn in the triangle. placed where it is. will be in equilibrium with the parabolic segment. THE ALEXANDRIAN ERA 213 Jerusalem.

incorrect and that the heuristic derivation should be supplemented by a rigorous proof. then EJ12 + E02 = EA2 + E02 = A02 = AE· Ar. . furthermore of a cone through this second circle with vertex atAandaxisAr. If now MEN is an arbitrary line in the plane of the circle ABrLJ.--. the other arm Ae is made equal to Ar. But triangle ArZ is twice as large as ArK. But. of the sphere. Let ABrLJ be a great circle on the sphere. A M face of the cone in II and P. (2) (EP2 + E02) : EN2 = AE. nevertheless it carries conviction. and since the triangle is in equilibrium with the segment. cutting this circle in 0 and E and the sur· Fig. placed at its own center E. To conceive of a parabolic segment or of a triangle as the sum of infinitely many line segments. Archi· medes is fully aware that this conception is. the weight of the triangle must be equal to three times that of the segment.N_ _ _ _ _---. hence the parabolic segment is equal to '/a of triangle ABr. and finally of a cylinder EZHA with axis Ar on the circle EZ as a base.. who thought of the integral Iy dx as the sum of infinitely many terms y dx. Ar : Ara = AE : Ar. 2. Archimedes remarks himself that this argument does not give a rigorous proof of the proposition.. is closely akin to the idea of Leibniz. Think of a second great circle on the diameter BLJ in a plane perpendicular to that of the first..211 CHAPTER VII ment is three times as long as the arm KX to the centroid of the triangle. Archimedes announces that the volume of the circumscribed cylinder is 1 % times as great as the volume H. Volume of the sphere. 72. in contrast with Leibniz. and hence four times as large as triangle A Br. and the circles on the diameters II P and EO are transferred to e.. It follows from (2) that in this position they will be in equilibrium with the circle on the diamet'er MN. Now Ar is thought of again as one arm of a lever with fulcrum at A. Hence the ratio of the sum of the circles on the diameters lIP and EO to the circle on the diameter MN is equal to the ratio of AI: to Ar. parallel to BLJ. as a matter of fact. whose base is a circle with diameter EZ.

will be in equilibrium with the cone. Centroid of a sesment of a paraboloid of retlolution.1 to Br. But Eudoxus has shown that the cone equals Ys of the cylinder. this result is obtained even more readily than the two preceding ones. Hence sphere plus cone equals half of the cylinder. 73. the ratio of the cylinder to the sum of cone and sphere will be equal to that of Ae to AK. In On the sphere and the cylinder he proved the correctness of all these results rigorously.1 again as a lever with fulcrum at A. of this cylinder. 5.1A to B. so that A8 = ALI. then the volume of the finite segment is 1% times as great as that of a cone of the same base and the same axis. placed at 9.. Since the weight of . we find EO' : EfJ2 = ALI : AE. be in equilibrium with the sphere and cone together. or 1/3 of the smaller cylinder tPX'PIJ. By using the same method. Archimedes constructs in the segment a cone ABr r with vertex at A. To show that this centroid K divides the axis ALI in the ratio 2 : 1.A~--¥"--+--_I. Volume of a spheroid. and considers 8. If a paraboloid of revolution is cut by a plane perpendicular to the axis. i. The result may also be formulated as follows: the sphere is four times as large as a cone whose base is a great circle of the sphere and whose height equals the radius. in its own position. This led Archimedes to the idea that the area of the sphere is equal to that of four great circles. Volume of a sesment of a paraboloid of retlolution. For. THE ALEXANDRIAN ERA 215 Because the cylinder EZHA is made up of these circles. he said. From this the conclusion is again drawn that Fig. it will probably be true that every sphere equals a cone whose base is equal to the sur- face of the sphere and whose height equals the radius. as 2 : 1. will be in equilibrium with the circle on PI1. Taking OE as an arbitrary line in the plane of the parabola BAr. placed where it is. an ellipsoid of revolution equals % of the circumscribed cylinder.e. placed where it is. therefore the sphere is 1/. both placed at B.!. hence the circle on OE. just as every circle is equal to a triangle whose base is the perimeter of the circle and whose height equals the radius. the cylinder will. cut off by a plane perpendicular to the axis.e. produces. parallel 9_ _ _ _ _--. B the segment of the paraboloid. transferred to 9. 3. By means of the same method it is found that a "spheroid". 4. Since K is the centroid of the cylinder. i.

11. The method is the same as that used in the preceding case. then the volume of the solid common to these cylinders equals 1/. 7. this proof proceeds as follows: . Finally he announces the following proposition: 13. but equal to polyhedra bounded by planes. Centroid of a hemisphere. The proof is not found on the palimpsest. Archimedes gives two strict proofs for the proposition concerning the area of the parabolic segment: one mechanical and the other geometrical. pass a plane through the center of the base and an edge of the upper base. The method is identical with that in example 2.216 CHAPTER VII the segment equals 3/1 of that of the cone. with mutually perpendicular axes. In entirely analogous manner. the following cases are treated: 8. is equal to 1/8 of the entire prism. Volume of a sesment of a spheroid. 10. Intersection of two cylinders. Centroid of a spherical sesment. Cylindrical sesment. Then the volume of the segment of the cy- linder bounded by the cylindrical surface. from which the proposition follows. but Archimedes made it into a rigorous proof by means of the exhaustion method. Next follow the cubatures of two remarkable solids. Let a cylinder be constructed in a right prism of square base. The mechanical derivation is essentially the same as that in the "Method". 6. Centroid of a sesment of a spheroid. of the volume of the cube. If. In the case in which the segment is formed by a line Br perpendicular to the axis. 9. Archimedes derives this proposition by his mechanical method and then gives a strictly geometrical proof. whose base is the inscribed circle of the base of the prism. Volume of a sesment of a sphere. We can see very clearly how Archimedes supplemented his heuristic consider- ations with rigorous proofs. Centroid of a sesment of a hyperboloid of reflolution. from the treatise on The quadrature of the parabola. two cylinders are constructed. the oblique plane and the base. in a cube. bounded by cylindrical surfaces and planes. the ratio of the lever arms A9 and AK is also equal to 3 : 2. 12.

and more than three times x the sum of the trapezoids Z~.. Z. and similarly BA : BZ = trap. Z. Divide Br into an arbitrary number of equal parts and. HN and IE parallel to the axis. etc.trap.AZ. they will be in equilibrium with P. ZM. X. To prove this.r than three times the sum of the trapezoids KE. X.:. TH. etc. etc. YI : trap. LI' which are respectively in equilibri- um with the trapezoids LIE.Ir__.l:Z : trap. etc. D. are smaller than the trapezoids KE. But the sum P + X + '1' + D + LI' is % of the trangle BrLl. NI. to its point of intersection LI with the tangent at r... The sum of all these Pig.d B E r that the triangle BrLl is less r---------=r---=r-TZ~~H~. EZ. MH. TH : trap. But if. and con- nectr with the points in which these lines intersect the parabola. THE ALEXANDRIAN ERA 217 Draw &J. AZ. EZ.Br : BE = EE : E~ . instead of from these points. AZ. MH and NI and the triangle lEr.. A Z. From the proportion (1) of the "Method". draw lines EA. A Z. they are suspended from the lines BE. if these trapezoids LIE. KE. are suspended from their "right" vertices E. they will be in equilibrium with the trapezoids KE. . The sum of the areas P + X + ':P + D + LI' is therefore less than the sum of the trapezoids KE. Now it is asserted . ZE. LIE : trap. Therefore. BA : BI = trap. and the triangle lEr. Yland the triangle Elr. AZ. HM p and NI and the triangleElr. parallel to the axis. etc. BA : BH = trap.. I. He and III and the triangle lor. 74. At the extremity A of the lever he suspends the areas P. etc. which extend beyond the parabola. we have BA : BE . so that % of this triangle is less than the sum of the trapezoids KE. Archimedes extends r B to A such that BA = Br and he considers AT as a lever with fulcrum at B. etc. H. ':P. through the points of division E. Therefore the areas P. suspended from A. etc. areas will then be in equili- iJ brium with triangle BrLI and therefore equal to one third of its area. X.

by an argument from mechanics. Now he constructs.» ~ + (~-I. @H.218 CHAPTER VII In an entirely analogous manner Archimedes proves that Ya of the triangle Br.1.. can be made less than the ratio of the larger to the smaller of two arbitrarily given magnitudes. Sometimes Archimedes uses ratios in- stead of differences and proves that the ratio SI : s. Archimedes constructs a triangle ABr.s. which we may present as follows.: (1) s]> s> s•• (2) s]> z> s•. we shall not give all the steps of the indirect proof in the sequel. by drawing through the midpoint. at least as eleg- ant.s. which contradicts (2). no and the triangle rOB.1 (in our case 1/.1 is greater than the sum of the trapezoids4)Z. But this sum is exactly equal to the triangle BrK. @n. This mechanical proof is followed by a geometrical derivation. Therefore z = s. Archimedes has enclosed s be- tween two areas St and s. then s .) > ~ + (~ . is equal to the sum of the trapezoids IN>. from which the con- clusion z = s follows. 4)@.1 of the base Ar a line L1 B parallel to the axis. The difference between the first series of trapezoids plus triangle and the second. in each of the two parabolic segments determined by the arcs AB . It is a technique which originated with Eudoxus. s is less than z. It is clear that the inequality s] . With slight variations.) -~. In contrast with Archimedes.1. Then: Z> s + (s]-I. which contradicts (1). Let s denote the area of the parabolic segment and z one third of the triangle Br. In the parabolic segment ABr. ~ ~-~-& Suppose now that s were greater than z.) -~. but we shall merely derive the inequalities (I).s•• and s> Z + (s] . take 8 less than the difference z . through the interior of which the parabola passes.Z> 8 = s] . that z is also between s] and s.s: Z-S> 8 . on the other hand.. we have to prove then that s = z. this indirect method of proof recurs repeatedly in the works of Archimedes. Ss <8 can perfectly well replace the equality (3). which is an arbitrarily small part of the triangle Br. and he has shown. Choose 8 as less than the difference s-z. whose difference is equal to an arbitrarily small part of the triangle Br. interior to the parabolic segment. Now follows the proof by the "exhaustion method". m and the triangle lI'O. This completes the mechanical part of the proof. (2) and (3).). If. ~ -s•.

triangles A ZB and BHr in the same manner. triangle ABr is one half of the cir- cumscribed parallelogram ArEe.e. and then he derives from the properties of the parabola the conclusion that the sum of these two triangles is equal to % of triangle ABr. A . E + E/3 ~ 4E/3 = A/3. and so forth. so that we obtain by addition B + r + A + E + B/3 + rl3 + A/3 + E/3 . Archimedes proves and formulates very explicitly that the parabolic segment. Next Archimedes proves the formula for the sum of such a geometric progres- sion: if an arbitrary set of numbers A. Indeed. 1 S. in accordance with a well-known proposition. the modern concept "sum of an in- finite series" means nothing but the limit of a finite sum. Archimedes does not speak about the sum of the infinite geometric progression. E form a geometric progression with ratio %. the next step will lead to 4 triangles. The proof is very simple: B + B/3 = 4B/3 ~ A13. Now. 75. i. If B/3 + r /3 + A /3 is subtracted from both sides and A added. is exactly equal to '/a of the first term. Moreover. . then there will ultiinately remain less than any arbitrarily given area. p.. he does have command of the concept.(A + B + r + A)/3.e. but. then their sum. again more than half. although he does not know the expression "sum of an infinite series". 177. differs from this sum by less than an arbitrarily given area. which we encountered in our discussion of Book X of the Elements. although greater than every partial sum A + B + r + A + E of his geometric progression. If r we remove now from the parabolic segment first the triangle ABr. and therefore greater than one half of the parabolic segment. more than one half. 1 Moreover the areas of the triangles which have been removed from a geometric progres- sion of ratio %. r. If this process is con- E tinued. we obtain the asserted statement: (4) A + B + r + A + E + E/3 = 4A/3. and A then from the remaining segments the two inscri· Fig. THE ALEXANDRIAN ERA 219 and Br. i. A + A/3 = 4A/3 = r13. bed triangles AZB and BHr. B. a magnitude which differs from a finite partial sum by less than an arbitrarily given positive e. i. whose sum is equal to % of the sum of the two triangles which have just been considered. increased by Ya of the last term. r + rl3 ~ 4r.3 ~ B13.e. and if this process be continued.

then that one is the shorter. it is not much of a trick to show. such as "The whole is greater than the part". a partial sum such as A + B + r + LI + E. 2) If two curves in one plane have the same endpoints and are concave on the same side. Then the excess of 4. as the calculation with an arbitrarily small 8 is sometimes called. areas or volumes. but not on the other side. and surfaces which span plane curves and which lie entirely on one side of the plane of the bounding curve.A13. his thinking is entirely modern. or on them. In this respect. other postulates are required.A13. that the parabolic segment can neither be greater nor less than 4. is added to . 5) If the difference between two unequal lines. at which the last term E is less than the difference 4. And.A 13 . supposing that the segment exceeds 4. For the measurement of plane areas and of volumes. On sphere and cylinder J.A13 over the segment would be greater than E.AI3. to which we shall return presently. always lie on the same side of the curve or the surface. would also exceed 4. contrary to formula (4. For. which occur in the summing of infinite series and in limit operat- ions. For how could one know otherwise that the perimeter of a circle is greater than that of an inscribed polygon and less than that of a circumscribed po- lygon? Archimedes starts therefore with some new axioms. . Such a curve or such a surface he qualifies as "concave on one side".). where A is the first inscribed triangle. which differs from the segment by an arbi- trarily small amount. were for Archimedes an open book. All this is found in Archimedes.) The analogue of 2) for surfaces.220 CHAPTER VII After this. the "epsilontics".A13. if all the line segments which connect two arbitrary points of the curve or the surface. by manipulating inequalities. and hence less than E. 3) Of all surfaces. if the segment Z were less than 4.A13 over A + B + r + LI + E is E 13. in essentially the same words. But for the measurement of lengths of arcs and of curved surfaces. He considers bounded plane curves which lie entirely on one side of the line joining their endpoints. . but this would require that the segment Z is less than A + B + r + LI + E. were adequate. And now he postulates : 1) Of all lines with the same endpoints the straight line is the shortest.Z. but the excess of 4. spanning the same plane curve. combined with the "lemma of Archimedes". 4. to a point. It shows that the estimations. which is impossible. and if one of them is entirely enclosed by the other (or coincides with it in part). the axioms laid down by Eudoxus. we would but have to continue the series A + B + . the plane has the least area.

while with his hands he holds a horizontal bar. 363. invented by Archimedes of Syracuse and called "snail" (lroXlUx. lHiO). in the British Museum. draining an irrigation by means of hydraulic in· struments were actively carried on du- ring the period of Alexander. Place of origin uncer· tain.C. pp. operated by a negro slave. Terracotta statuette from the late Ptolemaic period. that the Delta was irrigated by means of a special machine. which rests on two sup- ports. As seen on our representation. 360. Vatican Museum.) on ac· count of its form .. According to Rostovtzeff (Social and Economic History ofthe Hellenistic World. 230. 2(9). Rostovtzeff adds: This contrivance appears to have come into common use in the Delta in Helle· nistic times and is still employed In some parts of middle Egypt. 23b. Augustan relief from Palestrina. The Greek historian Diodorus states posi. PLATE 23 PL. about 100 B. made in ac· cordance with the system of Archime· des. PL. the slave moves the cylinder round the screw with his feet. Photo Alinari. Hydraulic screw. Roman warship. tively. Rome. (see p. .

Na· pIes. Various elements in the framework (the waterhens .in the four corners and the tendrils) are very successful imitations of Roman examples. Frankfurt am Main). Democritus( 1). (Stadtisches Kunstinstitut.porphyrions . probably from the school of Rafael. 2"". 24c. PL.i) PL. Warrior on one of the gables of the Alexander Sarcopbagus at Istanbul (see page 210). see p. Mosaic. . Bronze copy after a Hellenistic original (3rd century B. 24b. (Photo Ali"". originally thought to be authentic Roman work. From the" V ilia dei Papiri" in Herculaneum National Museum.). Copy (or falsification).C. PLATE 24 PL. m the same mate· rial. 210. (460-400 B.C. Death of Archimedes.

closes the sphere. w to the deter- mination of the curved areas of a risht cylinder and of a risht circular cone. of which the number of sides is divisible by 4. in the case of the cylinder. From postulate (4) we derive (1) area of I" <area of the sphere < area of Cn • Now let A be a circle whose radius is the diameter of the sphere and whose area is therefore four times as great as that of a great circle on the sphere. for the sphere as well. by use of an argument with which we have already be- come familiar. The same con- clusion holds for en and hence. This is the famous "Postulate of Archimedes". on the basis of postulate 4). THE ALEXANDRIAN ERA 221 itself a number of times. exhaustion method. in both cases by enclosing them between an inscribed and a circumscribed prism or pyramid. He constructs an inscribedand a circumscrib- ed regular polygon for a circle.:I+-. a circumscribed polygon of the same number of sides generates by rotation a solid en. already used by Eudoxus. The only difficulty lies therefore in the proof of inequality (2). From his . similar to In' which en- Fig. but probably explicitly formulated for the first time by Archimedes.about an axis through two diametrically oppo- site vertices produces a solid In' inscribed in the sphere and bounded by sections of conical surfaces. In both cases he constructs a circle whose area equals that of the cylinder or the cone. we derive from (1) and (2) the conclusion that the sphere has the same area as the circle A. In the same manner. is a sum of areas of sections of conical surfaces between parallel CIrcles. Revolution --t+--+---t-----+--. Archimedes proceeds nc. The area of I. It turns out that the volume of In is equal to that of a cone. 76. For example.. Next Archimedes takes up his famous deter- mination of the area and the volume of a sphere. Then Archi- medes proves the following inequality: (2) area of In < area of A < area of Cn. by means of tht:. Since the ratio of the areas of the similar solids I" and en can be brought arbi- trarily close to 1. After a few preliminary propositions. the radius of this circle is a mean proportional between the height and the diameter of the cylinder. it will exceed every prescribed magnitude of the same kind. whose base is a circle equal in area to the solid itself and whose height is the radius of the sphere circumscribed about In. Analogous inequalities hold for the volumes.


theorems on the area of a conical surface, Archimedes deduces that the area In
equals that of a circle whose radius R is given by
RI = o· s,

where a is the side of the revolving polygon and s
the sum of all the diagonals which are perpendicular t+-+--;-+-+.::::::~
to the axis of rotation. The product a . s is now trans·
formed into d . ~ where d is the diameter of the circum-
scribed circle and ~ is the diagonal whIch subtends one
less than half the number of sides. The ratio of d to ~
of course approaches 1. so that in the limit R2 tends Fig. 77.
to J2. These are the main features of the proof.
In an entirely analogous manner, the area of a spherical segment and also the
volume of a spherical sector are determined.

On sphere and cylinder II.

This treatise gives the solution of two problems which lead to cubic equations.
The first, to determine a sphere equal to a given cone or cylinder, leads to a pure
cubic of the form

which is solved in the familiar way, by determining two mean proportionals
between band c. Archimedes does not say how this is to be done. This fact led
his commentator Eutocius to the elaborate report on the duplication of the cube
on which we have been happy to draw earlier.
The second problem, to divide a sphere in a given ratio, by means of a plane,
leads to a cubic equation of the form

or, in Archimedes' formulation: to divide a line segment a into two parts x and
a - x, such that
(0 - x) : b = ,2 : x3,

Archimedes promises to give an analytical and a synthetic solution, but this part
of the treatise has been lost. Eutocius has however found a manuscript in Dorian
dialect, which he attributes to Archimedes and which gives first the necessary
condItion for the solvability of (4) and then a solution by means of the intersect-
ion of a parabola and a hyperbola. The condition for solvability is that bc2 be
at most equal to the maximum value of x2(a - x), which is attained when x =


On spirals.
When a straight line revolves uniformly about a point O. while a point P.
starting from 0 moves uniformly along the line. then the point P describes a spiral.
Archimedes derives the characteristic property
of a point on the spiral in polar coordinates.
then determines the tangent line at an arbitrary
point on the spiral and finally the area contained
between two arbitrary radii vectores and be-
tween two successive windings. or within the
first winding. The adjoining figure indicates
how this area is enclosed between sums of cir-
cular sectors. The only difficulty in the proof by
the exhaustion method is the summing of the
series 12 + 22 + 3 2 + ... + n2 • For this sum
Archimedes gives the following formula. In
Fig. 78.
geometric formulation (Proposition 10):
3[a 2 + (21)2 + (Ja)2 + ... + (na)2] = n(na)2 + (na)2 + a(a + 21 + 3a ... + na).
On conoids and spheroids.
By conoids Archimedes means paraboloids of revolution and hyperboloids of
revolution of one sheet. by spheroids he means ellipsoids of revolution. He
determines the volume of the spheroid and of all segments of all spheroids and
conoids. which are cut off by an arbitrary plane. as well as the area of the ellipse.
The volume of the sphere is of course contained in this as a special case. but the
method of proof is totally different. Archimedes might have transformed the
spheroids into spheres by stretching or contracting (a trick which he himself ap-
plies to the ellipse in proposition 4). but this method is not applicable to conoids.
For this reason he invented a new method. which is invariant with respect to
stretching and contracting. According to his own saying he first found the results
for the paraboloids. and only later on. and with
great difficulty. those for the ellipsoids and the
hyperboloids. The method consists in this that
the segment of a conoid or a spheroid under con-
sideration is divided into slices of equal thickness
by planes parallel to the base. as shown in the
adjoining figure for the paraboloid. and that each
of these slices is enclosed between two cylinders. Fig. 79.
one less and the other greater than the slice.
In the case of the paraboloid. the outer cylinders form an arithmetic progression
whose difference is equal to the smallest term:
~ = A + 2A + 3A .;- ... + nA.


while the inner cylinders form the same progression except for the last term:
~ = A + 2A + ... + (n - 1)A.
Now, says Archimedes, it is clear that the sum of n terms, each equal to nA, is
less than 2S1 and greater than 2S2 :
(1) ~> n. nA> 2Sz.
The geometrical meaning of n . nA is a pile of equal cylinders, all on the same
base as the segment of the paraboloid and all of the same height. These form to-
gether a cylinder C. Thus, for (1), we can write:
(2) ~>C>2S2'
Moreover, we have
(3) ~>S>~.
where S denotes the volume of the segment of the paraboloid.
Since the difference Sl - S2 = nA represents the volume of the cylinder at the
bottom of the pile, which can be made arbitrarily small, the familiar method leads
form (2) and (3) to the result
S = Y2G.
In the case of an ellipsoid or a hyperboloid, the cylinders do not form an arith-
metic progression. But their sum can be reduced to the sum of a sequence of areas
of rectangles x(a ± x), whose altitudes x form an arithmetical progression b, 2b,
3b, ... , nb. If now ~ designates again the sum of n such rectangles and S2 the
sum of the same rectangles, except for the last term B, i.e. in modern notation, if
II .. -1
~ = 1: kb(a + kb). ~ = 1: kb(a + kb). B = nb(a + nb).
1 1
then Archimedes proves the inequality
(4) nB : ~ > (a + nb) : (1/3 a + l/s nb) > nB : ~.
which replaces the inequality (1) in the proofs. In the proof of (4), Archimedes
starts from the formula for the sum of squares, which he proved in "On Spirals".

The notion of integral in Archimedes.
Riemann gave a rigorous definition of the integral by enclosing it between a
"lower sum" and an "upper sum", as indicated in the adjoining figure. The integral
f"y dx is then the area under the curve between Y
.the ordinates x = a and x = b, and the X·axis; the
"lower sum" is the sum of the areas of the rectangles
below the curve, and the "upper sum" is the sum
of rectangles. of somewhat greater height, which
cover the area. The treatise on conoids and spheroids L.L---1~'--.l....-..1....~b-~X
shows that Archimedes was familiar with this method a
of inclusion and that he used it for the determination Fig. 80.
of volumes. But, it seems to me that one ;:an not say that he was familiar with


the concept of the integral. His integrals always remain tied to a definite geo-
metric interpretatIon, as volumes or as areas of plane figures. We have no evidence
that he understood that one single concept is the foundation of all these geometric
interpretations_ For instance, in the quadrature of the parabola on the one hand,
and in the determination of the volume of the sphere and the spheroid on the
other, he bases his rigorous proofs on totally different methods, although both
lead to the same integral

.fx(a- x)dx .

so that any method effective for one of these problems could also have been used
for the other.
Nevertheless, his rigorous determination of areas and volumes make Archi-
medes the precursor of the modern integral calculus.

The book of Lemmas
(Liber Assumptorum) has been preserved only in Arabic, not in Greek. It begins
with a set of propositions concerning a figure, called the Arbelos or cobbler's
knife, bounded by three semicircles, tangent
to each other at their extremities. The Arbelos
has the same area as the circle which has BD
as a diameter (see Fig. 81). This line divides
the Arbelos into two parts, whose inscribed
circles are equal. Furthermore Archimedes
indicates how to express the diameter of the
C D A inscribed circle of the Arbelos itself in terms
Fig. 81. of AC, when the ratio of the parts into which
D divides AC is given.
The book contains furthermore a remarkable proposition: 8. Extend a chord
AB of an arbitrary circle by a segment BC
equal to the radius, and draw the diameter
FDE through C. Then the arc AE is three C
times as great as the arc BF. The proof
can be given very easily, e.g. as follows:
LADE - L DAB + L ACD = L AnD + E
L BDC = 2 . L BDC + L BDC = 3. L BDC.
By means of this proposition, the trisection
of a given arc can be accomplished as fol- .
lows: draw the diameter EF and layoff FIg. 82.
the segment BC equal to T, in such a manner that the extension of CB goes through


A, e.g. by means of a ruler on which distances equal to r have been marked off.
Then the arc BF equals one third of the arc AE.
The Greeks frequently applied such "Neusis" constructions, in which a seg-
ment of definite length is pfaced between two lines in such a manner that the ex-
tension of the segment passes through a fixed point. Hippocrates used them in the
quadratures of his lunules.

The construction of the regular heptagon
has also been preserved in Arabic only. C. Schoy discovered it in his study of a
treatise of Al-Biruni, on the construction of the regular nonagon. 1 Tabit ben
Qurra, who still had seen Archimedes' treatise in Greek, complains of the poor
condition of the manuscript, so that it was only after a great deal of effort that he
succeeded in unraveling the proofs. B K A Z
The treatise, which Tabit has brought down y x
to us, begins with a number of propositions
on right triangles. But propositions 16 and
17, which deal with the construction of the
heptagon, are independent of these.
Proposition 16. Let the diagonal BC of the
square ABDC and the transllersal DTEZ be so
drawn that the triangles DTC and ZAE are equal
in area. (One sees readily, by rotating the Fig. 82&.
transversal, that such a position always exists.) Drop the perpendicular TK from T
on AB; and let the segments ZA, AK and KB be denoted by x, y and z respectively.
Then it follows that
(1 ) AB.KB =AZ2 ,i.e. (y + z)z - xl, and
(2) ZK.AK-KBI ,i.e. (x + y)y _ Z2.

These statements follow easily from the equality in area of the two triangles.
Archimedes does not speak about the construction of the transversal and of the
segments x and y. But they can readily be c.~rried out by the use of conic sections.
For, if we set y + z = a, the equations (1) and (2) take the form
(3) a(a -y) - xl
(-4) (x + y)y - (a - y)2.
Equation (3) represents a parabola, when x and y are interpreted as rectangular
coordinates; and (4) is then the equation of a hyperbola. These two curves have
three points of intersection in the finite part of the plane; one of these lies in the
first quadrant.
1 C. Schoy, Die rriS""olfltt,i,,/ten uh,.. tks ... Al·Bi","i, Hannovu, 1927. Set also J. Tropfkt, aitsc"'. _h. U1III
••'''''. U.tmicbt, 59 (1928), p. 195, and Di, Si.b.."lsbhandl..S des Arcbilfltdu, Osiris 1, p. 636. I follow Tropfkc in
the notations.


Now Archimedes constructs a triangle AKH, of which AK = y, HK = z and
AH = x, he circumscribes a circle about triangle BHZ and asserts that BH is
equal to the side of the inscribed regular heptagon! Is this not breath-taking?
The proof can be found in Tropfkc. But one can not help asking how Archi-
medes hit upon this scheme. Tropfke's answer is as follows:
Let the regular heptagon BHLZGEF be
drawn, let the diagonal BZ meet HE in K
and HG in A, and let BG meet HE in T. We
denote the segments ZA, AK and KB by x,
y and z respectively and we designate by
B Z ex the inscribed angle which intercepts one
seventh of the circumference. Then we find
that AH = x and KH = z. The triangles
ZHK, HAK and HTA are similar, because
each of them has one angle equal to ex and
another equal to 2x. But from this fact we
obtain immediately the proportionalities
E z : (x + y) = y : z. and
Fig.82b. z : X = X : (y + z).

which are equivalent to (2) and (1) respectively.

The other works of Archimedes.
We have already briefly mentioned the Measurement of the Circle and the Sand
In spite of the importance of the treatises on mechanics, we can only barely
mention them here. To begin with, in On equilibria of plane figures, the famous law
of the lever is derived strictly logically from a set of axioms: two magnitudes are in
equilibrium if their distances are in inverse ratio to their weiShts.1 By means of it, the
centroids of parallelogram. triangle and trapezoid are determined, and, in Book
II, the centroid of a parabolic segment and of the section of a parabola between two
parallel lines.
In the treatise On floating bodies, Archimedes first derives the laws of upward
pressure for bodies, whose specific gravity is less than, equal to or greater than that
of the liquid. A real masterpiece is the treatment of the stability of the equili-
brium of a floating right segment of a paraboloid of revolution. One has to read
this splendid piece oneself in order to appreciate its value.
In the fifth book of his Collectio. Pappus relates something about Archimedes
investigation of semi-regular solids, bounded by:
1 A careful analya. of this deduction and a discUliion of E. Mach·s critique is found in W. Stein. Qwlltn gnd
SlvJj",. B I. p. 221.

a)(s . which was mentioned at an earlier point2. a historian. or 8 triangles and 18 squares. His epigram on the duplication of the cube. 8 hexagons and 6 octagons.c)· Somewhat younger than Archimedes is Eratosthenes of Cyene.b)(s . without being at the top in anyone. it was from him that man learned to determine scientifically the dates of historical events. a geographer. or 32 triangles and 6 squares. 30 squares and 12 pentagons. and he wrote a great work on ancient Greek comedy. p. taking the form of a walk through the sky of Hermes. conscientious and he hdd artistic tastes. or 12 squares. He made an estimate of the inclination of the ecliptic. Indeed he was eminent in various domains. or 12 pentagons and 20 hexagons. This is perhaps the reason why his friends called him Beta . He described the starry heavens in a poem. based on the view that the earth is round. Archimedes is also the discoverer of the well-known formula for the area of a triangle. 20 hexagons and 12 decagons. set Bibliothtca MathtJll&tica Xl 3. or 20 triangles. or 6 squares and 8 hexagons. and of the earth's perimeter. who accomplished excellent results in a great diversity of fields. industrious. or 30 squares.228 CHAPTER VII 4 regular triangles and 4 regular hexagons or 8 triangles and 6 squares. of the distances of sun and moon. He also collected myths about the constellations. He was learned. the allround sportsman. Life. but also with philosophical questions and with matters relating to the theory of music. 11-78. but h~ lacked the genius of Archimedes. in which he encounters various pretty adventures. He was the founder of critical chronology. or 20 triangles and 12 decagons. TM boo" "''''''''i''8 1M d"mb. usually attributed to Heron: Vs(s . . St:t p. a philo- logist and a poet. According to an Arabic traditionl . His dialogue Platonicus deals not only with the Delian problem. or 8 triangles and 6 octagons. young Eratosthenes travelled from his birthplace on the I Al-Biruni. or 20 triangles and 12 pentagons. 160-1. He designed a new map of the world. the five-sports-athlete. or 80 triangles and 12 pentagons. Another nickname for him was Pentathlos. Around the year 260.a "second string man". as a mathematician. is a sample of his subtle poetic art.

how he sometimes spent the night in a big urn (the famous vat). Ariston was called "the siren".saw). they were dog- matists and their views of nature especially were long antiquated.. SchwOln. Xenocrates and Polemon. Zeno and Chrysippus. but they tried to make it into a moral code for reliable functionaries.our concept of "cynicism". Plato's successors. i. On account of his magnetic oratorical talents. Following the example of Plato's Academia and Ari- stotle's Lyceum. even lasting into the Roman period. he accepts life as it comes. The wise man. we must remember that they lived in the time of the great Hellenistic monarchies with their staffs of gover. Leipzig 1943. It is probable that he had been 1 E. "The wise man is like a good actor. were indeed influenced by the kynic dogma. he said. the colonnade. who plays the parts of Agamemnon and of Thersites equally well".klJp!e aus der A. who has experienced justice as the inner truth. He did not indoctrinate his pupils. he taught that man can only attain real liberty by liberating himself from his desires. The doctrine of Diogenes has little connection with . wearing nothing but a cloak and sandals. Very popular were also the Kynics.e.?ment officials. but he trained them in the dialectical method. Eratosthenes must have been around fifty when he was called to the court of Ptolemy III to educate the crown prince. "the dog" (". But Eratosthenes admired him above everything else for his uncompromising ethics. THE ALEXANDRIAN ERA 229 coast of Africa to Athens. how he ridiculed all conventions and all imagined needs unmercifully. how he walked through the city. in order to study philosophy. the first stoics. It was the period in which the schools were constantly wrangling with each other. does not need a rule of behavior for every possible case. carrying his wooden watercup. He him- self had an excellent teacher of mathematics in Autolycus. He disputed relent- lessly the naively-empirical epistemology of the Stoics. who took this designation after Diogenes. At the same time he became the head of the world-famous library.hed the "garden" and Zeno the "stoa". OwoJu. He equally admired Archesilaus who had reorganized the Academy. There are numerous anecdotes about Diogenes. Epicurus had establi. According to Schwarzl . He also maintained uncompromisingly the re- quirement that every philosopher had to begin by learning mathematics. but under their direction. he will know for himself how to act. but stands above it inwardly. because the coryphees of philosophy were then in Athens. nevertheless Stoicism and Epicureanism continued to be the fashionable tendencies. they had to defend propositions against his attacks. the Academy had come into the blind alley of sterile speculations on numbers./ike. Neither Epicurus nor Zeno were great philosophers. were very decent and well-meaning people. But Ariston of Chi os pitched into the compromi- ses and the moral casuistry of the Stoa. Archesilaus had put a stop to this and had enthusiasti- cally taken up the cudgels against the dogmatism of the Stoa. In his own opinion he arrived there at the most favorable moment. .


in Alexandria for some time before this, because his life-work is unthinkable
without constant use of a great library. Where else would he have found the
material for his map of the world, for his chronography, his philological work and
his celestial travelogue? Neither could he learn mathematics and astronomy in
Athens, but he could in Alexandria.
Indeed, Eratosthenes was a typically Alexandrian scholar. The refined culture
of the Hellenistic royal court, where poets, philosophers and grammarians com-
peted in the purification of the language, in learned arguments and in artfully
made poetry, was his element. The subtle compliment to the king and his son,
which occurs in his epigram, betrays the well-versed courtier.
In his later years, Eratosthenes grew blind. He died the "philosopher's death"
by suicide.
Chronography and measurement of a desree.

The chronological work of Eratosthenes excels in critical precision. Is was his
principle to eliminate all unverifiable legends and to date the events exclusively on
the basis of authentic documents (such as the list of winners in the Olympic games)
and reasonable estimates. It is true that, in connection with the Trojan war, he
could not avoid deviating from this principle.
His measurement of the earth was equally conscientious. An older estimate,
mentioned by Archimedes, started from the assumption that the distance from
Lysimachia on the Hellespont to Syene in Egypt is 20,000 stadia. But this line
went over land and sea, so that it was impossible to verify the distance. For this
reason Eratosthenes preferred to take a smaller distance, which could be measured
accurately, viz. from Alexandria to Syene, situated practically due south of
Alexandria. Probably on the authority of professional counters-of-steps, he put
the distance from Syene to Alexandria at 5000 stadia. Then he established the
fact that at the time of the summer solstice, the sun is exactly in the zenith in Syene,
but at an angle of 1/50 of -4 right angles from the zenith in Alexandria. This gave
him 50 X 5000 = 250,000 stadia for the perimeter of the earth. Since we do
not know the exact length of a stadium, we can say little more than that the order
of magnitude is about right.

Duplication of the cube.

Recalling what has already been said about the problem (see p. 160), we shall
now discuss the mechanical solution given by Eratosthenes which he had en-
graved in stone in the temple of the king-god Ptolemy. I shall reproduce in trans-
lation the text, which appeared on the stone tablet, without the commentary
(superlluous, in my judgment) of the forged Eratosthenes letter. To understand it.
one has to know that at the top there was a model in bronze, consisting of 3 tri-


angular or rectangular plates, which could be pushed back and forth between a
fixed and a rotating ruler. Under this were the following figure and text.
A To determine two mean proportionals in
continued proportion to two given lines. Let

~ BrA
the lines AE and LI e be given. Then I move the
plates of the instrument towards each other
until the points ABrLl lie on one line. Now.
since AE and BZ are parallel. KA and KB
have the same ratio as KE and KZ. And since
E Z H 8 K AZ and BH are also parallel. this ratio also
equals that of KZ to KH. Hence KE is to KZ
Fig. 83. as KZ is to KH. But AE and BZ also have the
same ratio. as well as BZ and rHo In the same way we prove that BZ is to rH as rH is to
eLI. Therefore AE. BZ. rH and LIe form a (continued) proportion. Thus two mean pro-
portionals between two given (lines) have been found.
If now the given lines are not equal to AE and LI we shall obtain the mean proportionals
by making AE and LIe proportional to them; we reduce them to these and thus the task
will have been carried out.
In case more mean proportionals have to be found. we take every time one more plate in
the instrument than the number of mean proportionals to be constructed. The proof reo
mains the same.
Theory of numbers.
The best known discovery of Eratosthenes in the theory of numbers is the
sieve of Eratosthenes, a method for sorting out the prime numbers among the odd
numbers, transmitted by Nicomachus of Gerasa (Introductio arithmetica, chapter
13). The odd numbers, beginning with 3, are written down, going as far as one
wishes. Now one takes the first number 3 and crosses off all multiples of 3. There
are every time two places between two successive ones. Now one takes the second
number 5 and crosses off all its multiples, which now have 4 places between
them. Next one takes the third number 7, etc. In the end only the prime numbers
are left.
Another arithmetical theory which has to be ascribed to Eratosthenes is the
generation of all kinds of medieties (or mean proportionals) from the geometric
mean proportional, and of all proportionalities from equality, as they are explained
by Nicomachus, Theon of Smyrna, and Pappus.
The explanations of Theon and of Pappus begin with a philosophical intro·
(Theon): " Eratosthenes says that ratio is the source of proportionality. and the
origin for the generation of everything which is produced in an ordered way. For all pro·
portionalities arise from ratios and the source of all ratios is equality."
(Pappus): "Proportionality is composed from ratio and equality is the origin of all
ratios. Geometric mediety indeed has its first origin in equality; it establishes itself and also
the other medieties. It shows us. as says divine Plato. that proportionality is the source of


all harmonies and of all rational and ordered existence. For he says that the only connecting
link between all sciences. and the cause of all existence. and the tie between everything that
has become. is the divine nature of proportionality."
The words which are here attributed to Plato, do not occur in the dialogues,
and Theon ascribes almost the same words to Eratosthenes. The explanation is
very simple: it is not the writer Plato who is speaking here, but the Plato of the
dialogue Platonicus of Eratosthenes.
In the same way, Aristotle says sometimes: "Socrates says in The Republic ... "
and at other times "Plato says in The Republic". Of course, the reference is al-
ways to the person Socrates who appears in the dialogue of Plato's Republic.
In Theon, Eratosthenes discusses elaborately the elements from which various
kinds of entities can be evolved: numbers from the unit, magnitudes from the
point, ratios from equality.
But what is the meaning of the words "Geometric mean proportionality pro-
duces itself and also the other medieties"?
In Pappus, the 10 medieties (,uBaoTijTBa) are defined as follows:
1) A - B = B - r or A + r = 2B: arithmetical mediety;
2) A : B = B : r or AT = /JZ: geometrical mediety;
3) (A-B): (B-r) = A :r: harmonic mediety;
4) (A - B) : (B -r) = r : A: subcontrary of the harmonic mediety;
5) (A - B) : (B - r) = r : B: fifth mediety;
6) (A - B) : (B -r) = B : A: sixth mediety;
7) (A -r) : (A - B) = B : r, or A = B + r: seventh mediety;
8) (A -r) : (A - B) = A : B: eighth mediety;
9) (A -r) : (A - B) = A : r: ninth mediety;
10) (A -P) : (B -r) = B : r: tenth mediety.
The first three of these are old-Pythagorean, the next three were discovered by
Eudoxus, the last four by "later "writers. After having quoted the passage from
"Plato", cited above, Pappus continues as follows:
"We shall now show how to generate the 10 medieties from the geometrical one. First
now following theorem:
Proposition 17. Let A. Band r be three proportionals and set
E = B -I- r.
Now Lt. E. Z are again three proportional terms".
Pappus proves this proposition by transforming proportionalities; in pro-
position 18, he shows briefly, by three examples, how all proportionalities in
three terms (A, B, r) arise, by repeated application of proposition 17, from the
equality (I, I, 1). This is developed more fully by Theon, and by Nicomachus.
Starting from the proportionality in 3 terms (I, I, I), one obtains by application


of proposition 17, successively
2 4.
3 9. etc.;
and by inversion
4 2 1. etc.;
then, by applying proposition 17 again,
4 6 9. etc.. etc.
The inverse process makes it possible to transform every proportionality in
three integers into the equality (1, I, 1). None of this is very profound, but it is
rather nice.
How the geometric proportionality, thus having generated itself, produces also
the other medieties, is shown in the succeeding set of propositions 20-27, of
which we present one by way of example.
Proposition 20. If A. B. r are proportional. then
A = 2A + 3B + r
constitute a harmonic mediety.
These propositions mean that the quadratic equations, which are equivalent
to the various medieties, such as, e.g.,
E(A + Z) = 2AZ
for the harmonic mediety for A, E and Z, are all reducible to the normal form
B2 = Ar.
by means of linear substitutions such as
A = 2A + 3B + r. E = 2B + r.
The propositions which have been mentioned have without doubt been taken
from Eratosthenes' treatise on medieties, referred to twice by Pappus in Book
VII. In the Platonicus, Eratosthenes evidently hinted at this theory and developed
some of its philosophical implications, but in his mathematical treatise he has add-
ed the proofs.
Theon brings his explanation, taken from Adrastus, to a close in the following
"Eratosthenes proves that all figures are composed of certain proportionalities. so that
they originate from equalities and are again transmuted into equalities; but it is not necessary
to speak about this now".
If the word "figures" were here replaced by "medieties", everything would be
1 Pappus naturally do.. not consider the arithmetical and the seventh medieties which lead not to quadratic
but to linear equations. The Latin commentator Commandinus and Hultsch did not adhere to this sensible restric·
tion. and have insened two propositions. 19 and 24. which do not appear in Pappus. Hultsch has even "translated
back" th..e interpolations into Greek.


perfectly clear; but what does Eratosthenes mean by "all figures", Ta aVIIJ.aTa
It is my conjecture that the "figures" are intended to be conic sections, given by
quadratic equations in homogeneous coordinates, and that Eratosthenes wants to
say that it is possible to transform all such equations by means of linear substi-
tutions to the form HZ = Ar or A : B = B : r.
A few years ago, when I casually suggested this explanation, it appeared to me
as very risky and as having little chance of being confirmed. But recently I dis-
covered that Tannery had made a similar conjecture on the basis of entirely differ-
ent sources. His starting point is the following:
In Chapter VII, enumerating books on geometric analysis and on geometric
loci, Pappus mentions, along with books on those subjects by Euclid and by
Apollonius, a work of Eratosthenes in two volumes: lIeel p.eI10T-qT(JW, "On me-
dieties". A bit further on, he talks about certain geometrical loci (Tono£), which
Eratosthenes called Tono, neol1 p.eI10T1/Taa, i.e. "geometrical loci referring to me-
dieties". Of these loci he says that they "belong to the types mentioned before,
but differ from them by the peculiarity of the hypothesis". The types mentioned
before are straight lines, circles, conic sections, higher plane curves and surfaces.
It becomes clear from another place (p. 652), where the Tono, neoa p.8I10T1/TOO
appear along with other types of curves, that they are not surfaces, and at a still
different place (p. 672) they are placed in contrast to the straight line and the
circle. Hence these Tono, have to be conics or higher plane curves.
It appears from various sources that Aristaeus and Apollonius were much in-
terested in "geometrical loci on three or four lines", defined as the loci of
points, whose distances from 3 or 4 given lines form a proportionality; for 3
lines this would be a proportionality of 3 terms


and for 4 lines one of -4 terms


Such loci are conics, as shown partly by Aristaeus, and completely by Apollonius.
Now one can define for all the other medieties loci of points, whose distances (per-
pendicular or oblique) to 3 given lines form a mediety (harmonic, etc.). Such
curves are again conics and Tono, neol1 p.8110T1/Tal1 would be a good name for them.
For this reason Tannery surmises that these are the curves which Eratosthenes has
defined in his work on medieties. If this surmise is correct, it would at the same
time account for the passage in Theon.
Before coming to the last geometrician of genius of antiquity, to Apollonius of
Perga, we must say a word about a competitor of Eratosthenes, the geometer


Chronologically, Nicomedes comes between Eratosthenes and Apollonius,
since he criticised Eratosthenes' duplication of the cube, while Apollonius refers
somewhere to the cochloid of Nicomedes. Apparently he belonged to the same
group of Alexandrian mathematicians as Eratosthenes and Apollonius.
The cochloid or conchoid of Nicomedes is described in the following manner by
Pappus (Book IV, proposition 20): Let r,1E be perpendicular to AB. Let the line
r e


Fig. 84.

r,1E revolve about E, so that,1 always remains on AB while ,1r preserves its
length. Then r describes a curve ArM which has the following property: If a line
EHe is drawn to the curve through E, then the curve and the line AB determine on
this line a segment He which is always equal to r,1. The point E is called the pole
of the conchoid. AB its ruler and r,1 its distance.
Pappus states that Nicomedes described an instrument, with which the curve can
be drawn. He proved that on both sides of E, it approaches the line AB, and that
every line drawn from a point on AB to the side on which the curve lies, will cut
the curve somewhere. From this follows: If two lines AB and AH are given, and


Fig. 85.

if E is a point external to the angle BAH, then it will always be possible to con-
struct a line segment of given length HK in such a manner that H lies on AH, and
K on AB, while the extension of HK passes through E. This "neusis-construction"
can be performed by determining the point of intersection of the line AH with the
cochloid whose "pole" is E, whose "ruler" is AB and for which the given length
is the "distance".

86. LAHL1 =2·LE. B r Now the angle rBE is exactly Ya of Fig. p.. Furthermore. As we have seen. Proof: Bisect L1E in Hand draw AH. . Pappus shows that the neusis of Nicomedes can also be obtained by the inter- section of a circle and a hyperbola. Co ... twice as long as AB.r be perpendicular to Br. (This is the way Pappus has 1 At least according to Proclus. LrBL1 = LE.. Then the line segments HL1. the given angle ABr. and let A. He applied his neusis-construction to two famous problems. The trisection of the angle. But it looks as if Pappus claims the bonor of this can· otruction for him. The duplication of the cube in Nicomedes. every bit as simple.df... between which we wish to construct two mean proportionals. The underlying idea of the trisection will be seen to be the same as that of the trisection of Archimedes... Draw AE / / Br. in Eucl. while Archimedes calls for a neusis between a circle and a straight line. since A E / / Br. and such that its H extension passes through B... and. since a cochloid is defined and drawn more easily than a hyperbola.. so that finally L ABr = L rBL1 + L ABL1 = L E + 2· L E = 3· L E. Complete the parallelogram ABrL1. Let ABr be the given angle. But Nicomedes only uses a neusis between two straight lines.. and Br in E.. viz. Now construct between the A lines Ar and AE a line segment . the trisection of the angle1 and the duplication of the cube. ~E .236 CHAPTER VII Nicomedes took great pride in the discovery of this curve. since L AHL1 is an exterior angle of the triangle A HE which is also isosceles. HA and HE are all three equal to one half of L1 E (be- cause the right triangle AL1 E can be inscribed in a semicircle) and therefore equal to AB. Extend L1A until it meets the extension of r B.. T12. Let AB and Br be two given line segments...... In the isosceles triangle ABH we have L ABL1 = L AHL1. Nicomedes used the intersection of his cochloid with a straight line. this neusis-construction can be car- LI ried out with the aid of the cochloid.. Bisect AB in A.EL1..

p. ortorZ. At the museum he had the nickname Epsilon. AA = AM : AB = AM : AK = rB : rK. it follows at once that AM = 9 Z. and there he learned mathema.tics from the pupils of Euclid.l Br and determine Z so as to make rz = AA (Again simpler: M determine Z so that BZ and rz are both equal to BA. extend KiJ and BA to K their point of intersection M. The more he worries over it. Apollonius' period of florescence falls around 210 B. the line KiJM Fig. Finally. and then drawing AH. As a young man. Z Strictly speaking. you can look up the proof in Heath I. Draw re / / HZ. as it is said. . 87. translated by Ver Eecke (I. the construction would be much simplified by making BH = Br. p. If you do not succeed. who discovered not only the proof but the construction itself. so that AM can be replaced by e Z and we can say: rK and ez are the two required mean proportionals between AB and Br. A single hint: determine KZ from the Pytha- gorean theorem and make use of the second proportion derived above. or in the third book of Pappus. and ez : AA = ez : 9K ~ rH : rK = 2· rB : rK. Draw EZ .:"'-~----~ rK and AM are the two required mean proportionals between AB and Br.C. is genuine or not. This "neusis-construct- ion" can be carried out by intersecting the line rK with a c~hloid. but it is certain that he was a great astronomer as well as a great mathematician. under Ptolemy Philo- pator. THE ALEXANDRIAN ERA 237 it. Now draw a line zeK through Z to the extension of Br in such a way that the segment e K will be equal to AI----~ AA. he came to Alexandria.. which has Z as "pole". It is of little importance whether this stor. which will then bisect AB). re as "ruler" and rz as "distance". the more he will admire the ingenuity of Nicomedes. Then H~----=~-":.. From the propor- tions AM : 2 . We come now to the last really great geometer of antiquity: Apollonius of Persa. is superfluous.. because he established a theory of the moon and the crescent of the moon has the shape of an fl. 261. 43). Let the reader try to find the proof for himself. The proof requires only a consideration of the part of the figure below HK.

• In the theory of the epicycle. The first of these propositions deals with the epicycle hypothesis. It is not certain whether he referred here to Apollonius of Perga or to another Apollonius. the excenter. 88. or vice versa. nor the theory on which they were based. Then the linear velocity of the planet on the epicycle is v = rco. we find for the linear velocity of H. Ptolemy states that Apollonius proved two important propositions which show us how to determine the points on the planetary orbit.e. The second proposition. the roles are reversed. analogous to the first.-int of intersedion. In the same way. at which the direct motion. we should argue about as follows: Let r be the radius of the epicycle. In the theory of the excenter. called epicycle. and co the angular velocity of the planet H on the epicycle. as seen from the earth. Here it is supposed that both motions. arritled at H. the planet will appear to be standing still. who informs us that he used tables of Apollonius to determine the position of sun and moon at the time of eclipses. whose center describes a small circle about the observer. changes to a retrograde motion. Fig.. i. the planet traverses a large circle. The theory of the epicycle and of the excenter was the foundation of the theoretical astronomy of the Alexandrian mathemati- cians. If we had to prove such a proposition. fIIhlCh cuts the epicycle in such a fIIay that one half oJ the segment BH interior to the epicycle has the same ratio to the segment ZH from the eye to the proximate ". and it continued as such until Claudius Ptolemy. then the point H lies precisely at the boundary between the direct and the retrosrade motion. as the angular tlelocity of the epicycle has to the angular tlelocity of the planet in the epicycle. it is as follows: If a stra'Jht line ?-HB is dratm from our eye Z. is concerned with the excenter hypothesis.238 CHAPTER VII Our first witness to this fact is the astrologer Vettius Valens. neither do we know how these tables for the sun and moon were arranged. !J the angular velocity of the center E of the epi- cycle. each planet describes a small circle. whose center travels at the same time on a larger circle about the observer. resulting from the rotation . that of the epicycle and that of the planet in the epicycle have the same sense of rotation. who crowned the work of the astronomers of antiquity with his masterly "Almagest". Let us there- fore rather go to more reliable sources to learn something about the astronomy of Apollonius..

the parallelogram of velocities was therefore not entirely foreign to the Greeks. so that triangle AEZ : triangle AEr> sector AEH : sector AEr. THE ALEXANDRIAN ERA 239 of the entire epicycle about Z V= ZH-!J. The parallelogram of the velo- cities will then consist of two right triangles.Qr =HO ·wr. ZH . This is exactly the result of Apollonius.r=HO·II. but he wanted to show incontrovertibly that before this instant the motion is direct. and retro- grade after. Apollonius based his proof on the following lemma: If in triangle AI. If the perpendicular EO is dropped from E to ZH. Now triangle AEZ> sector AEH. As viewed from Z. This reasoning is of course entirely non-classical. the composition and resolu- tion of velocities do not occur anywhere in classical writings. According to Ptolemy. in which Br > Ar. not because he was not capable of it. and hence rLi :LlB>LB:Lr. then these right triangles are similar to the triangle EOH. Proof: Complete the parallelogram ALIrE and extend BA and rE to their point of intersection Z. This gives the proportion V:lI =HO :r. triangle AEr < sector AEr. a segment rLi is laid off on rB. He wanted not merely to determine the instant at which the apparent angular velocity is zero. greater than or equal to r A. 89. z Then. Nevertheless I con- sider Apollonius capable of having obtained his conclusion by an argument of this kind. but because he wanted more. the planet will then appear to be standing still. In the Mechanics of Heron. ZH:HO=w:Q. Consider first the case rLi = r A. a circle of radius AE about A as a center will pass through r. then rLi : LIB> L B : L r. Fig. But ZE : Er = ZA : AB = rLi : LIB. when th( resultant of these two velocities passes through Z. I think that Apollonius did not give a simple deduction of this type. or B ZE : EP> L EAZ : LEAr. the composition of uniform rectilinear motions does occur. .

90. Is not this proof a masterpiece of classical reasoning? The proof that the apparent motion of the planet along the arc HK is retrograde. which has to L HEK a ratio equal to that of the angular velocities. to that of the angular velocities of the epicycle and of the planet. Suppose now that ZHB has been determined so that YzBH : ZH = angular velocity of the epicycle : angular velocity of the planet as stated in the proposition.e. and its extension meets it in the "apogee" A. is direct. EK and EH. z On the other hand. In triangle BKZ. and hence. Suppose that the arc HK is directed towards A. the ratio of YzBH to HZ is equal Fig. BK. conic sections. Next comes the proof of the proposltion. through a distance whose angular measure is equal to L HZN. in accordance with the lemma BH:HZ> LHZK. in direct motion. instead of being equal to it. we have BH>BK. or towards the perigee r. let this angle be L HZN. A Draw ZKA.240 CHAPTER VII It is obvious that the ratio is increased further if rLJ is greater than A r. the center of the epicycle has travelled. The diameter ZE intersects the epi- cycle in the "perigee" r. when this arc is laid off towards the perigee. We have then to A show that. Conica i. : L HEK. :LB. Thus the resultant motion of the planet is a forward one. therefore the ratio of these angular velocities is greater than LHZK:LHEK. twice the inscribed angle B is equal to the corresponding central angle HEK. is greater than L HZK. therefore YzBH : HZ> L HZK.: 2L B. During the time in which the planet describes the arc KH of the epicycle. But. Therefore motion through the arc KH of the epicycle has moved the planet backwards by a smaller amount than the amount by which the motion of the epicycle through the angle HZN has moved it forward. This masterpiece has . is the title of Apollonius' most famous work. and in the second case retrograde. as seen from Z. in the first case the motion of the planet in this arc. It follows that the angle. is carried through in similar manner. YzBH: HZ> L HZK. Layoff from H an arbitrary arc HK towards the apo- gee A.

Thus Apollonius had important precursors. In Menaechmus we already met with the equations of the parabola (1) = px . The conic sections before Apollonius.. as he says himself. i. 51-100 for further details. The "symptom" of the curve. Neugd.tlation: Va Becke. THE ALEXANDRIAN ERA 241 rightly aroused the utmost admiration among all mathematicians of antiquity and of modern times. Qwlltll """ StwIieIo B 2.. he put the theory on a more complete and more general foundation than his predecessors. but occasionally also to oblique axes. ZlII" ali/uft A'sehnl II. by means of equations.e.". Archimedes I. on conics as geometrical loci... but we can get an idea of the work because Archimedes frequently quotes propositions from it. but. . A short time before Euclid. Archimedes always gives the equations of ellipse and hyperbola in the "two- abscissas form". and the distances AQ = x and BQ = Xt are called the "abscissas". supplemented in a few important points by the "Apollonius-Studien" of Neugebauer . About 300 the theory of conic sections was developed far enough to enable Euclid to write a textbook on this subject..iu Oft Cots. . us Cotsi.. Menaechmus already knew the parabola and the equilateral hyperbola. tl'A. Studi.llmoius (192<4).x. but this book is already in existence. I O. In what does this generalization consist? Archimedes and the more ancient writers already represent the conics systema- tically by means of "symptoms".. 1 A complete discussion of the Conics is out of the question here. Let AB = a be the major axis of the conic. The interested reader is referred to Dijksterhuis. StaiOfll (1896).2 and of the equilateral hyperbola (2) xy = A. which is as follows. The perpendicular 'PQ = Y from a point on the conic to AB. T. It would fill an entire book. And Archimedes proves incidentally some propositions on conic sections. In the case of the ellipse we have therefore Xt = a . which he needed in his determinations o( areas and volumes. indeed he used their intersection for the duplication of the cube. 215. p. tor the hyperbola Xt = a + x. it is the excellent work of Zeuthen.olucr. is called the "ordinate". usually referred to rectangular coordinate axes. and to sketching in broad outline what was known about conic sections before him and what he has added himself.... Aristaeus wrote a book on "spatial loci" . Die Lehre von den Kegelschnitten im Altertum.e.. French trao.• I shall therefore restrict myself to giving an impression of the structure and the style of Apollonius' work. These "elements of conic sections" have been lost. i. Apollmoius 0/ P. p. the condition to be satisfied by every point 1 English translation: Heath.

. In I 21. This is the reason why the parabola was called "section of the rectangular cone".242 CHAPTER VII P of the curve. This already comes very close to modern analytic geometry. which makes it possible to put the elaborate formulations of the ancients in a very compact form.. the hyperbola "section of an obtuse-angled cone". 91. and therefore t = ~. each of the conic sections was obtained from one type of cone of revolution.I: xxt. Fig. Xl and y are the abscissas and the ordinate of another point on the curve. before Archimedes. : d. we can write instead of (3) (5) . 92. instead of ocB (e : d)B. depending upon the pr?blem with which he is concerned. and ellipse "section of an acute-angled cone". then (3) can be reo placed by (4) .x) or f = IXX(X + a). In this very convenient notation. we were quite naturally led to introduce a new symbol «B for a magnitude which has a given ratio « to B. (4) and (5) derived from the definitions of the conics? We have already pointed out that. He chooses the second point (x. If x. : xxl = . is then in both cases: (3) . j) in various ways. : xxl = IX. When IX is given as the ratio . Apollonius gives the symptom in the same form (4) and in 120 the analogous form for the parabola: r:r=x:x In the discussion of the Data. by cutting it with a plane perpendicular to a generator.. It is in this form that Archimedes always writes the symptom. But how were the symptoms (3). For the circle« = I. were« is a given ratio. I = iXXXl and thus. for ellipse or hyperbola yI = otX(a . B • Fig. we can also write.

whose vertex may even be chosen arbitrarily in a plane of symmetry of the ellipse. on which the ellipse lies.. THE ALEXANDRIAN ERA 243 The ellipse as section of a cone accordins to Archimedes. but the proof of proposition 8 is more concise.. The method of proof is exactly the same in both cases. Construct now in the plane through En. .. But exactly the same proof can be used to derive the symptom if it has not yet been obtained.d Sp/otroia. LI the center. We have to find then a cone with vertex r. propo.ia. N one half of the other axis. 93. Make rE = rB and draw through LI a line ZH parallel to EB. Archimedes first considers the case in which the vertex of the cone lies on a per- pendicular to the plane of the ellipse and through its center (proposition 7) and then he reduces the general case to this special case (proposition 8). In Archimedes 1 we find a proof that every ellipse can be considered as a section of a circular cone. LlH. r and let the point lie in a plane through AB perpendicular to the plane of the ellipse. in case N2 is equal to LIZ. 7. Archimedes starts from tqe symptom of the ellipse in the form (5) and proves that the curve represented by this symptom does indeed lie on a right or on an oblique circular cone.ition. 1 On O"". Let us therefore look a little more closely at this proof. perpendicular to the plane ABr. It proceeds as follows: Let AB be a principal axis of the given ellipse. . 8. a circle. r Fig.

From proposition 7. The symptom of the given conic allows us to replace the left member by eK2 : AK· BK. From e drop the perpendicular eK to AB. also lies on the surface of this cone. perpendicular to the plane BrE). parallel to the first. and hence MA :rA = €JK :rK. of which AB is a principal axis. Then it follows from (6) that (7) N2 : LIZ . which leads to the conclUSion that r. From (9) and (10) we get. and therefore that e is on the surface of the cone. then the ratio of the products. e If this were not the case. which does not lie on this conical surface. e and M are collinear. Draw AM. while the square on the other. BK : rK2. AB.1M2: All· AP = eKI: AK· BK. such as II P and EB. Moreover (10) AIl· AP : rA2 = AK .of the intercepts is the same for LI as for A: (8) LIZ· LlH: LlA . which he obviously supposes to be known and which may be formulated as follows: If through a giflen point LI two lines. By composition we obtain from (7) and (8) N2 : LlA . E. LlH = AMI: EA . Now he uses a lemma from plane geometry. The lemma is readily proved by multiplication of the two proportions LIZ : LlA = AE : AIl. perpendicular to BE (the intention here is. . viz. are drawn to the sides of an angle ArB. there would be a point on the ellipse. Here Archimdes uses the symptom of the ellipse on EB. and through another point A two lines. We have then to show that the ellipse. already proved. such as AB and ZH.244 CHAPTER VII but otherwise an ellipse of which EB is one principal axis. again by composition AM2: rAI = eK2: rK2. the ratio AM2 : EA • AB is the same for every point on the ellipse and therefore equal to the ratio of the squares of the semi-axes. is determined by (6) N2 : LIZ . Finally draw lIP through A parallel to AB. so that M lies on the surface of the cone. LIB = AMI : AIl· AP. and LlH: LIB = AB : AP. AP. it follows that this circle or ellipse will lie on a right or on an oblique cone with vertex r. LIB = AE· AB : AIl . LlH = EB2 : 1:2. Extend rK to its inter- section A with EB. so that (9) .

and thus to derive the symptom of the conic section from the assumption that r. It is easy to devise analogous proofs for the parabola and the hyperbola.e. But before passing on to this matter. whose symptom is Fig. . in the fact that they were chiefly concerned with proving that a curve. Such proofs have been worked out by Dijksterhuis. For it is exactly this proof which is particularly simple when the plane of section is per- pendicular to a generator. and. are collinear. Rather do I look for the explanation. in place of (7) AM2=EA·AB. instead of (6) N2 = AZ . Menaechmus first of all. i. M p--. Why did "the ancients". but of all these cones. And. we have to answer another question. 94. p. 53-54. can always be obtained as a conic. The further development of this line of thought is found in Chapter 21 of Zeuthen's book. There is an infinite number of cones of revolution. whose equation is given in the forms (1) or (5). as a section of a cone. This would also account for the conciseness of his statement. which are cut by a given plane in a given conic. this was exactly the case which Me- naechmus and Euclid needed for the derivation of the symptom of their "section of the acute-angled cone". restrict themselves to sections made by planes perpendicular to a generator of the cone? Was the general case perhaps too difficult for them? This can hardly be supposed. THE ALEXANDRIAN ERA 245 HOflI were the symptoms deril'ed originally? It is not much of a trick to reverse the order of the steps in the above proof. The proof becomes still simpler. For in that case. The cone is then a right circular cone. AH and EB = E. with Zeuthen. A question and an answer. are by far the easiest to construct. one obtains. e. in Archimedes I. that these "ancients" had already used such a method of proof and that Archimedes only generalized and inverted a well- known proof. the two whose ver- tex lies in a perpendicular to the plane of this section at one of its vertices. The supposition that Menaechmus actually proceeded in some such manner is made still more plausible by the fact that similar proofs are found in Apollonius. when EMB is a Q B circle. One can imagine. It is therefore possible to obtain the symptom of the curve A8B from that of the conic EMB by means of simple proportions from plane y geometry.

as a result of the construction. the plane of the conic will intersect this plane in a line AB. The axis of an oblique circular cone is for Apollonius the line which joins the vertex T with the center of the base. hence the equation of the circle IS agaIn 8K2 . This line will then lie in the plane of the conic AB and also in the plane of the circle lIK. EZ. and 8P:8A-AA :AT. 8P by 8K2 and if we designate the ratio in the right member by IX. 11-13.246 CHAPTER VII The derivation of the symptoms according to Apollonius. If we replace 81I . and he multiplies the proportions 81I: 8B . This line is taken as the X·axis.81I· 8P.l is entirely analogous to the one which we dug out of Archimedes.P. Suppose now that an arbitrary plane cuts the base in the lint. through K. thus he obtains 81I· 8P : 8B· 8A . In the plane of this circle. ArchImedes. parallel to the circle of the base. If the plane TrA is taken as the plane of drawing. The ordinate K8 of an arbitrary point of the conic is drawn parallel to EZ. 8K is perpendicular to ~he di~meter PlI. .Ar· AA : Art. This is a special case of the very same lemma which we have already met in T Z Fi8·95. we find as the symptom of the ellipse 1 <Awiu I.Ar : AT. Now Apollonius draws TA / / AB. then he draws in the base a diameter perpendicular to EZ. only simpler and more general.

The geometric diagram .directions are no lonser mutually perpendicular. hyperbola and ellipse have their origin the Greek words for "application". in our earlier notation (11) y2 = OtXXl . we know. 98. but the difference lies here. The equations are exactly the same as the earlier ones. the terminology of such applications of areas comes from geo- metric algebra. consist of two unequal parts. in such a way that there is a rectangle in "excess". THE ALEXANDRIAN ERA 247 Ke l = ot· (eA.x. In both other cases. which we shall henceforth designate by PQ = y. The diagrams of Apollonius. from now on called AAl . Fig. Apollonius reduces equations (11) and (12) to a new and very useful form. 96. For the hyperbola (Fig. now no longer constant. A. eB). and therefore otXl = ota + otX. In the ellipse (Fig. "excess" and "defect".and y. Fig. whose base x and height otX have a constant ratio ot = p : a. 97) Xl = a + x. The names parabola. so that otXl = ota . i is also equal to a rectangle of base x and height otXl but this altitude is p p E o Fig. en. we have Xl = a . For the parabola. here reproduced. Apollonius finds in a similar manner (12) y2 = p. The new element in Apol1onius is therefore that the cone and the plane of the conic no lonser have a common plane of symmetry. By the use of geometric algebra. Equation (12) simply states that i is equal to a rectangle of base x and constant height p. 98). one could speak of a geometric and an algebraic diagram. is no longer perpendicular to the diameter AB. so that the x. the rectangle is therefore applied" to the line segment II p = aot in such a manner that there is a rectangle in "defect" whose base and height have again the constant ratio ot = p : a. or. the rectangle is therefore "applied" to a constant line segment p = ota.otX. But there is still more. that the ordinate K8 in the plane of the conic.

(14) ii2 = lXa2 = ap. just stated in geometric language. the word "hyperbola" means what we call a branch of a hyper- bola. The algebraic diagram is not oblique. the new FIg. p. The areas which are here added and subtracted correspond to the terms of an equation in modern analytical geometry. rather than in some other way. The proportion p : a = IX~ : ~ can also be expressed geometrically by drawing a straight line AlDE. i. 99. are sometimes called latus transversum and latus rectum. m which the + sign applies to the hyperbola and the . For Apollonius.sign to the ellipse. his reasoning is elegant and crystal clear. Conjugate diameters and conjugate hyperbolas. the oblique axes. This concept is not original with Apollonius. latus transversum ii being defined by (14) and the new latus rectum by the p proportion (15) ii : a = a : p.248 CHAPTER VII consists of the conic section. 15) that the two conjugate diameters are interchangeable.e.. that the ellipse which has DDl as the direction of the abscissas. Apollonius calls the midpoint of AAl the center of the conic section and the line through the center in the direction of the ordinate. whose ordinates are obtained o immediately from equation (11) by setting x = Xl = a Y2a. and Apollonius proves geometrically all the algebraic trans- formations performed on the equation. he calls "opposite hyperbolas". which occur in these equations. This is what makes his work hard to understand. and also a virtuoso in hiding his original line of thought. the abscissa x and the ordinate y. is given by a symptom of exactly . the diameter conjugate to AAl • In the ellipse the conjugate diameter cuts the curve in two points D and Dl . The relation between the areas of the applied rectangles and the square i. by /1 2 = ii. can be expressed as follows in algebraical formulas: (13) y2 = X axl = X· (alX± ax) = x(p± ax) = x{p± (p : a)x}. the same form as the original symptom (13). x and y.e. but rect- angular. ApoIIonius proves ~. i. it expresses the algebraic relation between a. but one has to guess at what led him to reason in this way. for . in which we are actually interested. and CAl as 01 that of the ordinates. The magnitudes a and p. Apollonius is a virtuoso in dealing with geometric algebra. The line of thought is mostly purely al- gebraic and much more "modern" than the abstract geometric formulation would lead one to think. The two branches which we consider as forming together one hyperbola. The length = DDl of the conjugate diameter is obviously the mean proportional between a and p.

These are defined as follows at the end of the first book: Apollonius starts with a pair of "opposite hyperbolas" and draws through the center C a line DDl in the direction of the ordinates. Apollo- nius succeeds in proving all the propositions concerning conjugate diameters etc. 100. is the intro- duction of the concept of pairs of conjugate hyperbolas. 7. Next pis determined by (15). . a mean proportional between a and p. By introduction of the conjugate hyperbolas. p. realized at the end of the first book. as Descartes expresses it. What probably is new. 1 Zeuthen has shown however that Apollonius did have a definite purpose in mind. diameter DDl . GioroitrU (ed. Then two new "opposite hyper- bolas" are constructed on the Fig. such that DDl = ii is. but "writes down whatever happens to come to his mind". the ordinates having the direction of CA. so that a is a mean propor- tional between ii and p. in how many points at most a circle or a conic section can meet a pair of opposite sections. A superficial reading of the first book of Apollonius' Conica may well create the impression that the author does not follow a definite method. in accordance with (14). van Schooten). which appeared naturally in the space-derivatIOn of 1 Dcscutcs. versum to these are the conjugate hyperbolas. For this purpose he has to show first that the diameter AAl with the correspond- ing direction of the ordinates. On this line equal distances CD and CDl are laid off on either side of C. in his pamphlet against Conon. THE ALEXANDRIAN ERA 249 he says in the preface to Book IV that. Tangent lines. and that he included precisely those propositions which were needed for the accomplishment of this purpose. It is proved in Book II that they have the same asymptotes. Nicoteles has already stated (without proof). can also be obtained as sections of right circular cones and were therefore identical with the types that were known of old. This aim was the following: to prove that the sections which he had initially defined by means of an arbitrary oblique circular cone and which he had characterized by a "symptom". the latus rectum being equal to ii and the latus trans- p. for hyperbolas as elegantly and as simply as for ellipses.

the ellipse and the circle. QA is extended by an amount AR equal to QA. is the only one. 36. are given: (17) CQ.250 CHAPTER VII the symptom. is a tangent line in that sense. Apollonius conceives of a tangent line as a line which has one point in common with the conic section. issuing from the center C: CQ. R is determined by use of the proportion (16) RA : RAt = QA : QAt ~ 34). it is proved that the tangent line at P. At a later time. then PR is tangent at P (Proposition 33). In I 32. Similar relations hold for the conjugate diameter (I 38). . is tangent to the curve. 102 Now he constructs the tangent line at an arbitrary point P. in the direction of the ordinates. R Fig. Fig. Then follow a number of simple propositions about lines which intersect a conic. For the hyperbola. For the parabola. can be replaced by any other diameter PP1 . the construction is very simple: The ordinate PQ is drawn through P. In I 37.CR:yl=a:p. we find a repetition of the statement that the line through A.CR = CAl. He has to begin therefore by determining the tangent line at an arbitrary point of the curve. 101. constructed in this way. but the statement is sharpened by the additional remark that between this line AC and the curve no other line is possible. and that no other line through P is possible between the conic and this tangent line. this relation among 4 points received the name "harmonic position" . it is shown to begin with. Moreover a relation is indicated by means of which the position of the point of contact P can be determined when R. and hence Q. that the line through A in the direction of the ordinates. In I 17. but which lies outside it everywhere else. He proves this by showing that every other line AD meets the conic in a second point H. the corresponding direction of the ordinates always being that ofthe tangent line at P. the proportion (16) is transformed into a relation between segments. In I 35.

The two-abscissas-form of the equation of the ellipse and the hyperbola was found to be .cl . (hyperbola). CQ as the new abscissa. c • fJc . and {Jx•• and II.I . (c . . First he replaces the squares in the left members by two similar rectangles c • {Jc and x • • {Jx. n. having equal angles. a parallelogram with sides y and "y with the same angles as II and Ill' Then equations (21) and (22) can be expressed .x. and propor- tional to these rectangles. J• or (21) x. 104. THE ALEXANDRIAN ERA 251 The equation referred to the cenler.fJc.' (ellipse and circle). = Pig.' { .)(c + x. (II : p)y' (20) ca . then we obtain for the hyper- bola XXl = (x. fJx. Then. {Jx.I .) . III a similar parallelogram with sides x.1 . in which {J is an arbitrary ratio. . .c)(x. + c) = x. if we take x.a. x. c· In the same way. and for the ellipse or the circle XXl . + y • ".c . i. (18) When the origin of coordinates is translated to the center C. 103.I : xxt = P : II. J'''' . This "central equation" is formulated in a somewhat different form by Apollo- nius in Proposition 41. or. equivalently. Next the three rectangles occurring in th~ left: and in the right members of (21) and (22) are replaced by three parallelograms. Pig. and if we set AC = %a = c. x" ~ (II : p). fJc. " re- presenting the composite ratio of a : p and {J.e. one finds for the ellipse or the circle (22) x• . so that equation (13) now becomes (19) X. Let II be a parallelogram with sides c and {Jc. . . (19) becomes x• .

into a form in which the two diameters AC and EC. The ratio f3 can still be chosen at will. already discussed. by means of area- calculation. a symp' tom is found of exactly the same form as that referred to the original diameter and the corresponding direction of the ordinates. taking as axis of abscissas the diameter EC and as direction of the ordinates that of the tangent line at E. is the following I 42. The diagrams have been drawn for this case: Fig. The two-tangents theorem and the transformation to new axes. he changes the first to parallelograms. Apollonius wishes to demonstrate that all diameters are equivalent. Through an . To accomplish this. whose sides are parallel to the two diameters and their ordinates. 105. Fig.252 CHAPTER VII as follows: (23) III . is the formulation best adapted to the transformation of coordinates. he transforms the symptom of the conic. 106. Let a parabola be given through the oris in A with diameter AD. which Neugebauer has called the "two-tangents propositions".II2 II (hyperbola) = (24) III + II2 II (ellipse or circle). at an arbitrary point E of the parabola. to which Apollonius now proceeds. = In the applications. and their corres- ponding ordinate-directions. next to a trape- zoid and a triangle. appear symmetrically. The form (25) of the central equation. the symp- tom is an equation involving rectangles and squares. In proposition 41. The first of these. the leading part is played by two propositions. the tangent line ED and the ordinate EZ. Apollonius always chooses as the angle of the parallelo- grams. In all of this. and finally (in Book III) to oblique quadrangles. then to triangles. referring to a para- bola. starting from an arbitrary point E on the curve. that. the angle between the directions of the abscissas and the ordinates. and draw.e. In its original form. i. Bisection of the parallelograms gives a relation between the areas of triangles: (25) 6 CQM ± 6 PQR = 6 CAL.

National Museum. that of Alexander. a large patrician house in Pompeii. the exact representation of the armour. Although the means of expression appear to us as very sober (the representation of the background is restricted to a single feature. The face of the Persian falling backwards under Darius' battle-wagon. Copy from about 100 B. Chapter ~ 3). is reHected in his shield. Naples. almost naturalistic. These optical theories and the tendency towards tIl careful observation of nature are reSected in this painting. In every other respect this work must bave been a "novum" in the art world. The style is very realistic. there is foreshortening. after a famous Greek painting pro- bably of Philoxenus of Eretria (end of 4th century B. probably in imitation of the "four-colors technique" of the original.C. Mechanical and acoustic problems were discussed in the school of Aristotle.PL. The mosaic is in subdued brownish tints. Everywhere. Euclid wrote an Optica and a Catoptrica which discussed perspective and the theory of optical images (see p. Strato of Lampsacus made physical experiments.). the bare tree). Timocharis and Aristyllus made observations on the moon and the stars (Ptolemy.42 m) found in the Casa del Fauno. space is represented very · strongly.C. showing the victory of Alexander the Great over Darius. The almost cruel rendering of the pitiless battle IS realistic. (Photo Alina. clothing and equipment. 2(0). is seen in its full width.C.). Almagest 7. Only the principal figure.) or of Gaugamela (333 B. es- pecially in the representation of the horses. 25.92 m by 3. which is worked into the close crowding of the conHict. Battle of Issus (334 B. A mosaic (5. As a result of this foreshortening. the tumult of battle is shown very suggestively. king of the Persians. We realize that this work of art originated in a period when applied mathematics and experimental natural science rs were being developed vigorously. of the incidence and reSection of light are naturalistic.i) ~ . as seen from the collection of problems ascribed to Aristotle.C.

(British Museum). Cameo (sardonyx) in the Kunsthistorisches Museum in Vienna (early Hellenistic). that the upper layer or layers are prfserved for the representation. A cameo achieves its peculiar effect. derive from the famous Alexander-portraits of his court-sculptor Lysippus. (283-247 B. through the stone. con- sisting of two or more layers of differ- ent colors. with name inscribed. as well as numerous other portraits of Alexander. Hiero II of Syracuse 269/8-214 B. being fashioned in such a manner.C. Paris. PLATE 26 PL. PL. Silver Syracuse coin. 26b. Portrait of Ptolemv II Phila- delphus. (Photo Alinarij . 26c. herma. 260. marble. Idealised portrait of Alexander the Great. Louvre. PL. This.C. while the rest is cut away down to the darker layer which serves as a background.) and his wife Arsinoe. from Tivoli.

THE ALEXANDRIAN ERA 253 arbitrary point P on the parabola. 106. par.. and par.l Thus we see that I 42 is nothing but a transformation of the equation of the parabola. the last equality follows from th'e symptom of the parabola. The proof is very simple. parallel to the tan$ent line L M E E D and the ordinate EZ. ALEZ. 107) L Fig. that it con- tains two diameters. AIMQ. because f:!. Since AZ = AD. is proportional to x. ALMQ : par. 1 Following Neugebauu. . ALMQ. Fig. AD and EL. By letting P winciJ. analogous to the transformation of the central equations of ellipse and hyperbola in I 41.. But we have also Fig. 108. hence ~ PQR = par. the factor of proportionality is found to be 1. so that ~ PQR : par. The two-tangents proposition for the hyperbola. Complete the parallelosram AZEL.l to AQ = x. 107. except that the parallel EL is now replaced by a line EC which E connects E with the center (see Figs. to their intersectIons with the diameter AD. and two tangent lines AL and ED. This pre- pares for interchanging the roles played by the points A and E. the line of thought of the proof can also be formuLtted as follows: As P varies. But y. ~ EZD = par. ALEZ = AQ : AZ. ALEZ = AQ : AZ.PQR is proportional to par. ALM<.PQR is proportional to PQ' = . drafll the lines PR and PQ. But the formulation is already of such a character. thue- fore f:!. the ellipse and the circle is exactly the same. Then we hcwe (26) trianglePQR = paralldogram ALMQ. EZD = par. f:!. wirh f. 109. ALEZ. The similar triangles PQR and EZD are pro· portional to the squares on their altitudes z ~ PQR : ~ EZD = PQI : EZI = AQ : QZ.

6. • The form (27) does not yet show immediately that the roles of A and E. and (29) . Fig. with their diameters and tangent lines. from which the conclusion (25) follows. CHAPTER VII The proposition Q 43) is now. PQR = . this is the reason why the .6. of which two sides are the diameters CM and CR. PQR = .e..6. CAL (ellipse) .6. that the ratio y = QR : QP is obtained by composition from the ratio p = ZE : ZC and the ratio a : p. which Apollonius had obtained earlier (I 37). It only remains to show that the hypothesis. are interchangeable.6. 110 and 111). as Zeuthen shows. 112. is fulfilled. for example. then (27) can be replaced by (28) . CAL (hyperbola). because the trapezoid ALMQ equals the difference between the two similar triangles CAL and CyM. i. Apollonius does not use improper quadrangles. But this is an im- mediate consequence of a property of the tangent line. of the diameter to the latus rectum. while the two other sides are lines through P parallel to the tangent lines at A and E. CQM . In Fig. the left member of (28) or (29) is the quadr- angle CMPR. For if the trapezoid ALMQ is represented as the differ- ence between the constant triangle CAL and the variable triangle CQM. In these cases the two-tangent proposition can therefore be formulated very simply as a propolition on areas: the area of the quadrangle CMPR remains constant when P moves along the curve (see figs. L c R Fig. 110. But a slight transformation suffices to gain this result. CQM + . 111. In the cases drawn in our figures. This equality follows at once from (25). analogous to (26): (27) triangle PQR = trapezoid ALMQ. But there are conditions under which the same formulation of the proposition would require the introduction of the concept "improper quadrangle". the "area CMPR" would have to be understood to mean the difference in area of the triangles CMS and PRS.6.

III 3: P and P' both arbitrary.g.lngcJ to M'. THE ALEXANDRIAN ERA 255 formulates his area propositions in a somewhat different manner. one divides the quadrangle 1 The second endpoint of the line P'M in Figs. . in the case of Fig. 113. III 2: Limiting case P' = A.P arbitrary. III 13-15: The same for two conjugate hyperbolas. the significance of the area-propositions lies in this that they can be considered as symptoms of conic sections referred 10 two arbitrary diameters as axes of coordinates. III 4-12: The same for two opposite branches of a hyperbola. This formulation has the further advantage of being valid also for the parabola. c = R In the third book. Fig. Fig. P'. According to Zeuthen. e. If. e. for which a center does not exist. P = E. 113-115). In the two other cases. 114 and 115 should be ch. III 3 is evidently equivalent to Zeuthen's formulation CMPR CM'P'R'. Apollonius treats the area-propositions in a systematic sequence: III 1: Limiting case P' = A. it is stated that for two arbitrary points P. 116. This is a clear indication of the great value which he attached to these propo- sitions!l c Fig. In III 3. 113...g. 111. the areas SPMM' and SP'R'R are equal (see Figs. 112.

viz. the interchange of the roles of the points A and E. Apollonius treats almost Pig. 117.50 he demonstrates that the two-tangents proposition is equally valid with respect to the new diameter as with regard to the original diameter AC. The symptom of the ellipse then takes the form nl - rzxI-+ fJxJ + constant.256 CHAPTER VII CMPR ~ means of the dotted lines in a parallelogram CQPT and the two tri- angles PyR and PTM.2ED. 116. Fig. The consideration of the conjugate hyperbolas in III 13-15 is necessary to make possible a treatment of the cases in which the axes do not intersect the conic. where I is the point of intersection of the tangents AL and ED. while D is on the diameter AC and L on the diameter EC. In the first book. a diameter and a direction of the ordinates. 115. with analogous results in all other cases. the latus rectum p and (in the case of a hyperbola or an ellipse) the latus transver- . In 149. Apollonius now proposes ~ 52. The new latus rectum p' is determined from (30) p' = (EI : EL) . and if one designates the sides of the parallelogram by x and y. Cones of rnolution through a giflen conic. after that there is little difficulty in de- riving the symptom of the curve referred to the new diameter and the corresponding direction of the ordinates. In I ~8. R Fig. then it becomes dear that the paral- lelogram is proportional to xy and the two E triangles to XI and f. exclusively those area-propositions which are needed for his immediate purpose. he proves that the new diameter bisects all the cliords parallel to the tangent line at E.53) the following problem: to "find" a conic section when a fixed point A is given.

The general case is reduced to this L one by first constructing a principal axis. then F must lie on a semicircle of diameter EC. a' = 2EC the latus transversum and p' the latus rectum. It only remains to select T on this circle in such a way that the conic will have the desired ratio p : a of the latus rectum to the latus transversum. Let E be the given fixed point. the latus rectum is determined in such a way that the point E satisfies the symptom of the curve. the latus transversum AB = a is found by doubling AC. It is required to find a point Fig. A suc h t hat t he tangent AL IS . Without preamble. Finally. from the two ratios p' : 2ED and 2ED : rI. which we shall not enter into here. which must contain the vertex T of the cone. which Apollonius applies ova and over again to proponions of the form (31). THE ALEXANDRIAN ERA 257 sum a. Draw the ordinate EFjjLA. CF. parallel to DE.GFt : GE. AC is then the required principl. a line FG between the semicircle and the diameter EC. but it is not liard. He does not even say how this is to be done. . this is indeed the characteristic property of the tangent line in E . perpen- dicular to the diameter AC.E1 : EL = GF : of re~olutio~. to construct. with a given principal axis AB.e. one forms the composite ratio p' : rI. This analysis is not found in Apollonius. EC the given diameter. that the given diameter is a principal axis. We shall begin here with the latter construction because it connects directly with the figure just drawn and with o B formula (30). Apollonius passes through AB a plane perpendicular to the plane of drawing. If now. as shown in I 34. whose symptom is give~. From the construction it becomes clear that "finding" means to obtain the curve. 1 Neugebauer has shown that a general method. because our only purpose is to indicate the main structural outline of the first book. This amounts to choosing the vertical angle of the cone arbitrarily. Draw also FGjJDE. in such a way that the proportion (31) holds. . Then we have from (30) p' : 2ED . i. 118. 1 Now he draws CFD and determines A on this line in such a way that CAl = CD . Ge. a cone of revolution whose intersection with the plane of drawing is the required conic. The constructions are first earned out for the case In which the ordmate directIOn is perpendicular to the given diameter. we obtain (31) p' : a' .l axis. luits the purpose also in this casc. and constructs in this plane an arbitrary circle through A and B. and also 2ED : a ' = ED : EC = GF : Ge. Apollonius says: draw. as the section ~f a c. This is a simple problem of plane geometry.

III 41-44 are propositions about tangents. determined on two fixed tangents.258 CHAPTBR VII The second book gives a detailed treatment of the theory of conjugate diameters and principal axes. the construction of tangents through a given point. in which the ratio is given. The power-proposition is stated as follows: If through a lIariable point Z two lines are drawn in gillen directions. It is a well-established custom to name propositions after mathemati- cians. The first problem. then I is the join of the points of contact of the tangents through D. These propo~itions can be. sometimes called the Proposition of Newton. which meets the conic in E and F and the line I in G. who did not discover them. 43: ellipse and hyper- bola) have a given value. In III 16-23. in contemporary terminology. etc. with every point D. such that either the ratio of these segments (1II 41: parabola). then the ratio ZD·ZT:ZE·ZK is constant. The special cases which arise when. segments with given initial points. determining on a conic two chords EK and DT. of the hyperbola also comes up for discussion. furthermore prol?ositions concern- ing the segments determined on an arbitrary secant by two conjugate hyperbolas and their asymptotes. Apollonius associates. This topic is closely related to modern projective geo- metry. by definition. used to construct the tangents through a given point. a line " conjugate to the diameter through D. of asymptotes and of conjugate hyperbolas. by a variable tangent. and probably have been. Propositions 30-40 refer to what is nowadays called the theory of poles and polars. When D is external to the conic. was solved . the "asymptotic equation" xy = const. on two given lines. one of the four points "is at infinity" are dealt with as separate pro- I positions. Newton himself says explicitly that he has taken the proposition from "the ancients". such that for every line through D. which intercepts. Apollonius obtains the power-proposition_directlJ from the area-propo- sitions III 1-3. Propositions 41-43 reduce these constructions to the following problem: to draw a line through a given point. relations are indicated which hold between two sets of points. the proportion DE : DF = GE : GF is valid. One or both lines may be tangents. or their product (III 42. exactly as in the well-known case of the circle. or on two asymp- totes. and then treats a few special cases and applications (III 24-29). Of greater importance is The third book We have already discussed the area-propositions III 1-15. Of course. a power-proposition is developed. internal to the conic or external.

The foci and directrices of all three conics were therefore known. Pappus already knew that these loci are conics. The conic determined by (32) obviously passes through the 4 vertices of the 1 Neugebauer has shown in his Apollonios. indeed equal to the major axis (III 51. Apollonius proves that the products of the seg- ments from A to the points determined on the rays AP. But the preface just cited shows that Apollonius could solve the problem quite easily by his method and Zeuthen has made it clear (Abschnitt 7. make equal angles with the tangent line at P (III 48). according as this ratio is equal to. Loci involvins 3 or 4 straisht lines. Apollonius says that the propositions of Book III can be used to solve completely the problem of the "loci involving 3 or 4 straight lines". u from four given lines satisfy an equation of the form (32) xz = IX' }U. This proposition is also very closely related to modern projective geometry. z. The last propositions. according to which the locus of points for which the distances to a fixed point and to a fixed line have a given ratio. The rays AP are brought to intersection with a line through B parallel to the tangent line at A. . It was shown by Zeuthen (Lehre von den Kegelschnitten im Altertum. not completely disposed of by Euclid. The tangent propositions III 41-44 are applied in the theory of foci (III 45-52). 52). mentioned also by Pappus. If the last two lines coincide. THE ALEXANDRIAN ERA 259 by Apollonius himself in his work On cuttins off a ratio. p. is a para- bola. and the rays BP with a line through A parallel to the tangent line at B. This famous problem. and that their sum or their difference is constant. is the following: to deter- mine the locus of points whose distances x. have a constant value. when IX is a given ratio. which join a point P on the curve to the foci Pl and P2 . 8) that the propositions of Book III are indeed sufficient for this purpose. the second in his work On cuttins off a product.Studien that those for the parabola follow from III 41 exactly u those for the ellipse and hyperbola follow from III 42. It would have been an inconceivable denseness not to have observed that every ellipse. y. hyperbola and parabola can be obtained in this way. and of the segments from B to the points determined on the rays BP. state the relations between the two pencils of rays. generated when a variable point P on the conic is joined to two fixed points A and B on the same conic. In the preface to his Conica. we have a "locus involving 3 straight lines". greater than or less than 1. 53-56. 1 However Apollonius treats only the focal properties of the ellipse and the hyperbola. an hyperbola or an ellipse. Descartes prides himself on the fact that his analytical geometry t:nables him to determine these conics quite simply. 213) that Euclid already knew the proposition. He proves that the lines.

260 CHAPTER VII quadrangle formed by the four given lines. or a pair of opposite hyperbolas) through five given points. He proves that these maximal or Fig. Two conics can not meet in more than four points. But he does prove that the problem can have only one solution. In his preface to Book IV. But he gives even more than he promises: he determines all the lines through 0. besides the principal axis it- self. III 53. but he did not publish the proof. situated sym- metncally. The fifth book treats. an ellipse. which intersect the conic at right angles (nowadays. a parabola. and the third proposition (III 53) has not yet been thought of by anybody". we ° call them normals). probably because the necessary preparations and the complicated distinction of cases would have carried him too far afield. "Nicoteles of Cyrene has justly attacked Conon on this point. he investigates the positions of for which there are two. 119. a circle. III 36. thus writes Apollonius. A pair of opposite hyperbolas meets another such pair in at most four points. Indeed the fourth book is chieBy occupied with the proof of the proposition that two conics can never have more than 4 points in common. three or four solutions. The solution of the problem of the "locus involving 4 lines" therefore solves at the same time the problem of passing a conic (i. Nicoteles said also that he could prove the second propositions (III 36). he says that Con on of Samos had dealt with the first case (III 25). Next he . and he studies t--::~~H-f--j~-. in the words of Apollonius in the introduction.e. A conic or a circle can not have more than four points in common with a pair of opposite hyperbolas. He starts by taking the point on ° one of the principal axes. the problem of the shortest point ° and the longest line segments from a to a conic. In Apollonius this proposition is broken down into 3 propositions: III 25.~~---I carefully the maximum and minimum properties of these hnes. Apollonius does not take up the problem of the conic through five given 'Points. but that his proofs are not correct. or two maximal or minimal lines through 0. minimal lines OM are always perpendIcular to the tangent line at M. In that case there are either no. The constant at can be chosen so as to make the conic pass also through a fifth given point.

In the ellipse. to draw a straight line through a given point. Most of the modern editions of Apollonius contain Halley's very successful reconstructions. VII 13. that has been preserved in its entirety. On cunins off a ratio. This eighth book has been lost. is equal to the rectangle on the axes. There is only one other work of Apollonius. setlenth and eighth books. which leads to a quadratic equation. the distinction of the possible cases as to the number of solutions) of problems which are taken up in the eighth book. Apollonius gives first an analysis of the problem. in two books. the seventh book contains further· more propositions which are useful for the diorism (i. The sixth is of minor importance. at which the number of normals through 0 jumps from 2 to 4. The sum of the squares on two conjugate diameters of an ellipse is equal to the sum of the squares on the principal axes. and also in two conjugate hyperbolas. This work. namely: 1. treats the following problem: given two straight lines and a point on each of them. The difference of the squares on two conjugate diameters of a hyper- bola is equal to the difference of the squares on the principal axes. or inversely.e geometrical compendium of Pappus: . starting at the given points. THE ALEXANDRIAN ERA 261 allows 0 to move along the ordinate HO. but Halley succeeded in recon- structing a considerable number of its problems on the basis of a set of lemmas of the eighth book. the parallelogram on two conjugate diameters. he restricts himself to determining the limiting positions G1 and G. According tQ Apollonius' own statement. under the angle at which they intersect at the center. such that the segments cut off on the given lines. perpendicular to the principal axis. These limiting points lie on a curve"which envelops all these normals and which we call the evolute of the conic. besides the Conica. and he determines on all these ordinates the limiting positions G1 and GI . VII 31. The sixth. fore very position of H. The contents of the following works are known to us only from excerpts in th.e. which are found in Pappus. have a given ratio. The principal propositions of Book VII are the following: VII 12. and then a synthesis consisting of the solution of the quadratic and the proof that this actually solves the problem. Further works of Apollonius. Apollonius does not introduce this curve.

CP has a given ratio to the rectangle Bp·DP. On a fixed division. from one given point. C. On plane loci. A number of propositions on loci which turn out to be circles or straight lines. Furthermore. 5. p. or (b) are parallel. see Th. which (a) are extensions of one another. Heath. and if P describes a straight Iillt:' or a circle.262 CHAPTER VII 2. and If (2) OP . On contacts. B. or (c) include a given angle and which either (1) have a given ratio. Observe that this proposition contains as a special case the "inversion With respect to a point 0": If (a) OP and OQ have the same direction. 4. as expressed by Apollonius of course. D be given on a straight line. this proposition is formulated as follows: If two lines are drawn. But still more: All transformations of the plane. This treats an entirely analogous problem in which the two segments which are cut off are required to have a given product. to determine a given rectangle. such that the rectangle AP . a line or a circle. History of Greek Mathematics II. Determine on the same line a point P. or. which carries circles and straight lines into circles and straight lines. or from two. In Pappus. and if the endpoint of one of these lines describes a circle or a straight line.e. all similanty transformatIOns are special cases. 182. These two books contain the famous tangency problem of Apollonius: Given three things. This problem also leads to a quadratic equation. For a reconstruction of the contents of this book. as does the next: 3. OQ is constant. the so-called circular trans- formations) are contained as special cases in Apollonius' general proPOSition. Let the points A. or (2) have a given product. Determine a circle which passes through each of the given points and which is tangent to the given lines and circles. then the endpoint of the other will also describe J circle or a straight lme. then Q Will also describe a straight line or a circle. each of which may be a point. which carry circles and straight lines into circles or straight lines (i. 011 cutting off areas. Was Apollonius aware of this? . The most interesting proposition deals with a type of transformation of the plane.

in this last work. Steele. According to Eutocius. It is true however that such constructions were con. THE ALEXANDRIAN ERA 263 6. On Neuses. sometimes expressed. p. than the one given by Archimedes. On unordered i"ationalities. In this book. B 3. a closer approxi- mation to:rr. . It is contradicted by the numerous constructions. The reader will recall that a "neusis construction" (neusis = inclination) was for the Greeks the construction of a line segment of given length. and somewhere Pappus says that. sidered to be more: elementary. 9. On the helix. more advanced means should not be used. 1 The following book is mentioned by Hypsicles in the introduction to the so- called 14th book of Euclid: 7. extended if necessary. Apollonius gave. Comparison of the dodecahedron and the icosahedron. 11. their areas have the same ratio as their volumes. Apollonius indicates a number of cases in which the construction can be carried out by means of compasses and straight edge. which have been handed down. which passes through a given point. See in this connection A. D. The rapid delivery. 1 The idea. for the duplication of the cube and the trisection of the angle. whenever a consuuction is possible by means of compasses and straight edge. that the Greeks only permitted constructions by means of compasses and straight edge. A general treatise.. If a dodecahedron and an icosahedron are inscribed in the same sphere. 10. is inadmissible. We can only conjecture whether this approximation has any connection with the origins of trigonometry and whe- ther Apollonius knew or made tables of chords. Entirely lost are: 8. 287. Qutllnl und Sludien. and of which the endpoints have to lie on two given straight lines or circles.

the subsidies shriveled. Indeed the works of Apollonius were but little read and were even partly lost. but not for retrogression and actual decay. of which the inner meaning was no longer understood.). did not invite mathematicians. Greek geometry had run into a blind alley. The "Method" of Archimedes was lost sight of. particularly in England. the Roman hegemony with the accompanying exploitation. We have already seen that science had become the concern of' courtiers. Political and economic conditions of course playa very important role. the algebraic tools for a systematic investigation were lacking. solved some small problem. The Roman emperors had very little use for pure science. crushing taxes and. All of this is certainly very important. who. 'the Arabs started algebra anew. which Archimedes and Apollonius had left for them. The geometry of the conics remained in the form Apol- lonius gave it. Higher plane curves were studied only sporadically. until Descartes. but it was not truly understood. For it only accounts for periods of scientific inactivity. while importing Greek sculptors. until it was attacked anew in the 17th century. and geometry was applied to practical and to astronomical problems. pedagogues and historians. crumbs from the board of the great. The theory of irrationals was interpreted by commentators. How did this come about? Sometimes it is said: "Greek civilization got old and lost the spark of life. But apart from trigonometry. Geometric algebra and the theory of pro- portions were carried over into modern times as inert traditions. It is true that there were some epigones. such as that of Pappus of Alexandria (300 A. but this is only a general summary. now and then. CHAPTER VIII THE DECAY OF GREEK MATHEMATICS External ClJuses of decay After Apollonius Greek mathematics comes to a dead stop. It is also true that compendia were written. When Caesar was besieged in Alexandria. a large part of the famous library was burned. and the problem of integration remained where it was. When it happened that the kings did not wish to spend money on science. Germs of a projective geometry were present. but it remained for Desargues and Pascal to bring these to fruition." Very true. not an explanation. brought about the end of prosperity in the Hellenistic countries. from a much more primitive point of view.D. For in astronomy the development was quite . and the wealthy Romans. such as Diodes and Zenodorus. nothing great nothing new appeared. Wars. but it does not furnish a complete ex- planation. depending upon libraries and upon royal subsidies. later on. which led to the development of plane and spherical trigonometry. BrieBy.

How totally differently things went in mathematics! Euclid was studied dili- gently. for Ptolemy himself refers to other authors. Theaetetus and Apollonius were at bottom algebraists.D. Each of these great astronomers stood on the shoulders of his precursors.) continued the work of Apollonius and also took Babylonian observations into consideration. and who needed simple little rules for setting up and solving algebraic equations. who uncovered these inner causes most clearly in his "Lehre von den Kegelschnitten im Altertum". AI-Khwarizmi. and no essential progress was ever lost. Ptolemy resumed the work of Hipparchus and carried theoretical astronomy to a truly admirable point of development. And this was unavoidable as long as the requirements of strict logic were maintained. improved the Ptolemaic system. followed by a new growth on a quite different foundation.. fractional. The great Arabic astronomers. Hipparchus (150 B. For "numbers" were integral or. What is the cause of this? Obviously. Many of the writings of the great mathematicians were lost. but in mathematics a long-continued decline. It was Zeuthen. but . general political and economic factors do not adequately explain the decline. nor Apollonius. The Indian Surya-Siddhanta is to a large extent based on Greek astronomy before Ptolemy. rather then on Greek learning. from which in turn the algebra of the Renaissance was derived. a theory of line segments and of areas. Greek algebra was a geometric algebra.g. Around 150 A. at most. And the 300 years between Hipparchus and Ptolemy had not been a period of complete idleness. they thought algebraically even though they put their reasoning in a geometric dress. the father of Arabic algebra. others were reputed to be extremely difficult and were practically not read at all. He wanted to write a book intelligible to simple folks. but after such periods the work was always continued from the point at which it had stopped. since these causes should have had the same effect in astronomy. who had attempted to explain the motion of the planets by means of epicycles and excenters. in astronomy we have a progressive development. 1. e. not of numbers.. that Arabic algebra is based. It is on his work. the father of modern astronomy. with inheritance problems. The inner causes of decay. purposely avoided "Greek erudition". and so did Kepler.C. It is true that there also shorter and longer periods of inactivity occurred. There must be inner grounds for the decay of antique mathematics. THE DECAY OF GREEK MATHEMATICS 265 different. even though not un- interrupted. Thus. such as al-Battini. but not Archimedes. The difficulty of seometric alsebra. We have already frequently pointed out the importance of the algebraic ele- ment in Greek geometry. Even Copernicus started from the Ptolemaic system. who were concerned.

thoroughly versed in transforming pro- portions with the aid of geometric figures. and only books remained. But one can not get any farther. to obtain results by this extremely cumbersome method. this set bounds for Hellenic algebra. one has to have recourse to the bothersome tool of proportions. that is 2. Something has to be added.266 CHAPTER VIII at any rate rational numbers.x) = be2 in the form (a-x):b =. It does honor to Greek mathe- matics that it adhered inexorably to such logical consistency. Anyone can use our algebraic notation. but only a gifted mathematician can deal with the Greek theory of proportions and with geometric algebra. The ancients did not have this tool. one has to be a mathematician of genius. as long as each generation could hand over its method to the next. reduced the cubic equation Xl = V to the proportion a : x = x : y = y : b. The proofs are logically sound. at the same time. examples can be found in Apollonius (e. but one fails to see the guiding line of thought. and Archimedes wrote the cubic xB(a . in Book V). instead they had the oral tradition. besides. but they are not suggestive. also those of the third degree. An oral explanation makes it possible to indicate the line segments with the fingers. But to get beyond this point. Equations of the first and second degree can be expressed clearly in the language of geometric algebra and. Hippocrates. This is seen very clearly in . it became extremely difficult to assimilate the work of the great precursors and next to impossible to pass beyond it.1 : Xl. To understand the line of thought. for instance. Reading a proof in Apollonius requires extended and concentrated study. But. Instead of a concise algebraic formula. One feels caught as in a logical mousetrap. one is compelled to transcribe these sentences in modern concise formulas.g. But as soon as some external cause brought about an interruption in the oral tradition. All of this disappears in the written formulation of the strictly classical style. while the ratio of two incommensurable line seg- ments can not be represented by rational numbers. one can emphasize essentials and point out how the proof was found. if necessary. everything went well and the science flourished. The difficulty of the fII1'itten tradition. As long as there was no interruption. In this manner one can get to equations of the fourth degree. one finds a long sentence. in which each line segment is indicated by two letters which have to be located in the figure.

e. but he had to study the written works and there he encountered the same difficulties that we meet. Some- times it happens that Apollonius uses. and it is from this source that our algebra was developed. the division of a sphere into two spherical segments whose volumes have a given ratIO. and he solved the Delian problem by means of new curve. all the works of the foremost mathematicians and astronomers. he wrote extended commentaries. This illustrates clearly that.g. but from Arabic algebra. He has written a book on burning-mirrors. Pappus (320 A. A further development of mathematics urgently required a concise algebraic notation. diligent and enthusiastic. He was gifted.C. a relation between line segments. And this is what the Arabs did. Frequently Ptolemy. Pappus would then add the analogous proofs for the other cases. a commentary on Ptolemy'S Almagest and one on the tenth book of Euclid. to treat expressions like 2 + .D. the "cissoid". had treated only one of the possible cases. To pro- gress. We come now to The epigones of the great mathematicians 1. And his great compendium. to multiply and divide magnitudes unconcernedly.y5 simply as numbers without worrying about their irrationality. to which we have re- ferred repeatedly. people found it difficult to understand matters which would become at once clear in an oral exposition.) had the magnificent library of Alexandria at his disposal.. he wrote an explanatory "lemma". from proposi- tions which occur explicitly in Euclid. or whoever it was. Diocles. it was necessary first to take a step backwards. THE DECAY OF GREEK MATHEMATICS 267 The commentaries of Pappus of Alexandria. It was necessary to return to the primitive Babylonian point of View. perhaps a proportion or a relation between products. Whenever Pappus found a proof difficult or incomplete. The notation used by the Arabs and by the Italians was still rather clumsy. e. without giving an explicit proof. little step by little step. . already in the days of Pappus. further development of the rigorous Greek methods could not lead to such a notation. By the use of conics he solved the "problem of Archimedes". Usually such relations can be taken from the figure by anyone who has had the necessary practice in the transformation of products and ratios. Diocles probably lived during the 2nd century B. it was simplified by the Frenchmen Vieta and Descartes 1. 1 A mort complttt description would demand also mtntion of the Italian Bombdli and of the Dutchman Simon Suvin.g. In such cases Pappus would deduce the relation. In order to assist in overcoming these difficulties and to simplify matters for those who were to come after him. consists largely of commentaries on the classics. The Algebra of the Italian Renaissance did not derive from Greek geometric algebra.

isoperimetric figures.. that the extremes rH and H9 have a given ratio. The "symptom" of the cissoid. CHAPTER VIII Eutocius transmitted both constructions in his commentary on Archimedes.) who mentions him. this can be replaced by HLJ : H9 = rH : HZ. We shall discuss only the second of these. In view of the excellent quality and style of his work he is probably closer to Archimedes than to ~intilian. This is r/'"=::::::::---=r----t=----!r--~.1 rLJ and then writing the proportiona- lity LJH :H9 = LJK :KE. the author of a very interesting book on. 2.1 accomplished by locating the intersect- ion of the cissoid and the line rIl. then one has obtained a so. Zenodorus. thus one obtains (1) rH :HZ = HZ :HLJ = HLJ :H9. and ~intilian (75 AD. It remains then only to multiply B the four terms of the proportion (1) by . which determines on the two mutually perpendicular radii Ar and AB. its equation in rectangular coordinates rH and H9. seg- ments Ar and All in the given ratio a : b. obtain two mean proportionals between given line segments a and b.. is found readily by drawing EK .C. a proportionality factor. K lution of the Delian problem. Let E and Z lie on the circle. 120. whom he mentions.e. A This says that HZ and HLJ are two mean proportionals between r H and H9. on either side of and at equal distances from B. The rosoid was defined as follows: Let AB and rLJ be two mutually perpendicular diameters of a circle. in order to FIg. Drop the perpendicular ZH form Z to r LJ and let LJ E meet this perpendicular in 9. a : b. Then the point 9 will describe the cissoid.). If one takes care furthermore. i. . lived between Archimedes (250 B. But HZ is a mean proportional between rH and HLJ.

The questions investigated and partly solved by Zenodorus are the following: which plane figure of given perimeter has the largest area. and which solid of given area has the largest volume? It is easy to guess at the answers to these questions. bounded by conical surfaces.. Zenodorus concludes that of all figures of equal perimeter.C. the isosceles is the largest. Among triangles of equal base and equal perimeter. Among all polygons of equal perimeter and equal numbers of sides. the circle is larger. 3. Hypsic. 10. contains the proofs of 14 propositions. of less volume than the sphere of the same area. But. the one which has the larger number of angles is the larger. This conclusion is justified only if "figures" are understood to be only circles and polygons. It was not until 1884 that Hermann Amandus Schwarz gave a perfectly rigorous proof of the isoperimetric properties of circle and sphere. Next Zenodorus states that among all solids of equal surface area. but very difficult to prove them. In this short treatise. Of two regular polygons with the same perimeter. Two of his writings have been preserved. the most important of which are the following: 1. His treatise. But he only proves the following propositions: 13. If a circle and a regular polygon have the same perimeter. 11. The sum of two similar isosceles triangles is larger than the sum of two non- similar triangles with the same bases and with the same sum of their perimeters. Zenodorus had come a long way. 3. 14. it is taken to be self-evident that a largest polygon of given perimeter exists. 7. It then follows readily from 7 and 10 that this largest polygon must have equal sides and equal angles. for his time. . Each of the five Platonic polyhedra is smaller than the sphere with equal area. In the proof of 11. this places him in the Pythagorean tradition. known to us from excerpts found in Pappus and in Theon of Alexandria. From 3 and 11. If a regular polygon (with an even number of sides) revolves about one of the longest diagonals. He wrote on polygonal numbers and on the harmony of the spheres. the circle is the largest. a solid is generated. les compares the areas and volumes of a dodecahedron and an icosahedron. which is included as such in most editions of Euclid. THE DECAY OF GREEK: MATHEMATICS 269 Isoperimetric fisures are figures of equal perimeter. not long after Apollonius. by using methods of Weierstrass. Hypsicles probably lived in Alexandria around 180 B. the sphere is the largest. the regular polygon is the largest. the first is the so-called Fourteenth Book of the Elements.

The circles which circumscribe the faces of these two solids are equal to each other. but on a rough approximation. This primitive method of calculation is really not worthy of a Greek mathema- tician. according to a proposition of Aristaeus. During the light part of each day. Hypsicles uses the same method to determine the initial moments of the separate degrees of the ecliptic. From this it follows r~adily that the volume of the dodecahedron: the volume of the icosahedron = the area of the dodecahedron: the area of the icosahedron. e. exactly 6 signs rise above the horizon. Hypsicles proves now that this ratio is also equal to edge of the cube: edge of the icosahedron and he gives a construction of this ratio by means of plane geometry. astrological. In the same way. Hypsicles also wrote a work. although it is sufficiently accurate to serve for practical.270 CHAPTER VIII inscribed in the same sphere. and those of the signs from Libra to Pisces a descending one.g. Half of the longest day is therefore A + B + C. The longest daylight period is therefore the day on which the first three signs and the last three rise. Hence A +B +C= 105. In addition. and. If the times for Aries to Virgo are ABC D E F. The remaining times can now be calculated without difficulty. Hypsicles divides the day into 360 "degrees of time". the sun rises and at the end the opposite point. be- cause at the beginning of the day. called Anaphorai (Risings) dealing with the times required for rising and setting by the various signs and degrees of the ecliptic. the longest day lasts 210 degrees of time. He simply assumes that the times of rising for the signs from Aries to Virgo form an ascending arithmetic progression. B=35. for Alexandria. one finds from D + E + F = 75 that E = 25. The sum of each of the arithmetic progressions is therefore 180 degrees: A + B + C + D + E + F = 180. . because the sum is equal to three times its middle term. Hypsicles assumes furthermore that. His method is not based on exact trigonometric calcula- tions. then those for Libra to Pisces are FED C B A. Like the Babylonians.

C. 32. p. nor whether he used only plane triangles. Neugebauer has shown that these methods of calculation are directly related to the Babylonian lunar calculations. for an arbitrarily chosen radius R.D. .D. We do not know whether Hipparchus was himself the inventor of trigonome- try. or both. After that the Arabs introduced the tange nt and the cotangent. since what we now call the sine of an angle is the quotient by the radius of one half of the chord of twice the intercepted arc 2: ~ aR . in their great standard work Surya Siddhanta and in the Aryabhatiya of Aryabhata. 121. The Greeks therefore worked with tables of chords.dies 10 (1951). who lived at a not much later period.. tables of sines appear. such as our laws of sines and cosines. which greatly simplified the computations. But Hipparchus. was able to carry through such calculations. Plane trisonometry. where further referenc . An interesting exam pie: 1 See my article Babylo. Manilius and Firmicus Matemus. Ptolemy does not know any rules for oblique triangles. To understand this. the Greek astronomers Hipparchus (150 B. the Indian astronomers changed from the chords to the sines. Similar arithmetic progressions are found in the writings of the Ro- man astrologers Vettius Valens.ia.. we have to take a look at the History of trisonometry. As early as the fifth century A.) always used chords of arcs of circles. We have discussed already how these were calculated. 1 Apparently Hypsicles was not yet up to an accurate trigonometrical calculation of the times of rising and setting of the signs of the zodiac. cosine and tangent.hd (2a:) R Fig. When he is dealing with oblique triangles. Journal Near Eastern St. In place of sine. or only spherical triangles. Astronomy lII.) and Ptolemy (150 A. 1 • a SID a: = 2R . • It is only since around 1800 that we make tbe division by the radius. In fact it makes little difference whether one operates with chords or with sines. to the literature are given. he always divides them into right triangles. Up to that time the sine was one half the chord of twice the arc. For this reason we are compelled to draw chiefly on the works of Menelaus and of Ptoiemy for the earliest history of plane and spherical trigonometry. THE DECAY OF GREEK MATHEMATICS 271 applications. for his writings are lost for the greater part.

and also.L BAA is known. hence the arc EZ = 6°48'. these arcs are seen under angles BA A = 3°24' and BAr = 0037'. B. Ptolemy obtains EH = 1p17' 30" and rE = 1"20'23" when AE = 120". In the semicircle on EA. The altitude-line re again divides this triangle into two right triangles. In triangle ArE the angle at E is known as a peripheral angle with a given arc. hence the arc EZ in the semicircle on AE as diameter is known. the ratios of the altitude-line to the two including sides are then determined. In the same way: L EAZ = L BEA . the diameter AE being taken as equal to 120p. from the table of chords. In the right triangle Eer both angles are known. From a point A. equal to 7 p7'. We see therefore that every time an oblique triangle such as ME or rAE. Ptolemy has to solve the following· problem: The points A. in order to determine the moon's epicycle. and the chord EZ is. Draw EA and Er. Again setting Er = 1p20' 23" . The arcs BA = 53°35' and BAr = 150"26' are known from observations. The result is of course equivalent to our law of sines. the chord EZ. whose angles are known. By the same method. one finds AE = 17p55'32" in the same unit in which AE = 120p. so that the ratios of Ee and of re to Er can be calculated from the table of chords. outside the circle. when the diameter EA is set equal to 120p. 122. Setting the radius of the circle equal to 60. Ptolemy reasons as follows: Let E be the second intersection of A B with the circle. and drop perpendiculars EZ and EH from E to A A and A r respectively. how far is A from the center? B Fig. the peripheral angle EAZ = 3°24'. we find re = 1"0'8" and Be = 0"53'21". If. by means of a proportionality factor. and thus the ratio of the sides follows.r lie on a circle. the line EZ is given the measure 7 p 7'. according to the table of chords. is divided into two right triangles.272 CHAPTER VIII In Chapter 6 of Book 4 of the Almagest. . Finally drop the perpendicular re from r to EA.

rE and L1E.1. being the chord of an arc of 96°51'. By a change of unit for Ar. even without trigonometry.17"3'57"'. a chord BE = 117"37'32" is extended by a segmentL1E = 631"13'48". But.1 from the center K? This is easy. The words of Ptolemy suggest that this method was taken from Hipparchus. Ee = 17"2' 11". By use of the table of chords. . L1M + KM2. according to the table of chords. Then L1A· L1M . By using the table of chords backwards. Addition of the arc BrA = lSOO26'. one finds arcrE = 6°44'1". is.36". But from geometric algebra. in the unit in which the diameter of the circle is 120". which makes Ar = 89"46'14". in terms of the unit which makes the diameter of the circle equal to 120".L1B. one finds L1E = 631"13'48" and rE = 7"2'50". equal to 89'46"14". therefore Ae = AE . when L1E = 120". Ar... we know that L1KI = L1A.(L1E + EB). makes arc :C======K:=====~:M~============-- Fig. L1E = 472·700"5'32". The only problem now remaining is the following: In a circle of diameter 120". How far is . Draw the diameter AKM through . 123. so that Ar . L1E . the corresponding chord BE is found to be 117"37'2". This completes the solution of the proposed problem. BrE = 157° 10' 1". THE DECAY OF GREEK MATHEMATICS 273 In the same units we had found AE = 17"55'32". The theorem of Pythagoras now gives AJ'2 = Ael + rei = 291"14'. so that L1K1 and hence L1K are known: L1K = 690"8'42".

that has been lost and that probably contained tables of chords. we shall also . in the previous chapter. " We now pass on to Spherical trigonometry and take up first the work of Menelaus. written a Spherics. which we already see applied by Hipparchus. stan and planets. various propositions are proved on spherical triangles (i. he has Fig. starting from three selected lunar eclipses. The third book opens with the famous . Besides a work on chords. engineer and architect. Paul Rome. he says: "We shallfollow a theoretical method. astrolabic instrument for observing the moon. In the first two books.a. Indeed. own description. Ptolemy'. which has been preserved. congruence propositions).274 CHAPTER VIII For. Menelaus made astronomical observations in Rome. 124...D. as reconstructed by Mr. from Ptolemy'. In 98 A.

1 We shall not enter into this further./Ti•• Leipzig 1900. II rLl. 125. Ptolemy wants to determine the declination ~ of the sun A Z E------ r a Fig.. meeting AB in H. dtr T. 1 Sec A. ...ansflersal proposition also repeatedly applied by Ptolemy. E on Ar.. AE = EH= LlZ'EH= LlZ'BE. Then: B r rA rLl rLl LIZ rLl ZB Fig. and let rLl and BE meet in Z. for instance. 17..i8D. it runs as follows Let LI lie on AB. p. one draws the line EHthrough E. THE DECAY OF GREEK MATHEMATICS 275 T. von Braunmuhl.. he finds the relations chd (2rB) chd (2rZ) chd (2L1B) chd(2AE) chd(2LlZ)' chd(2BA)' and chd (2rA) chd (2rLI) chd (2ZB) chd (2AEi = chd (2L1Z)' chd (lBB)' From the transversal proposition. In Book I. By projection from the center. rZLI and BZE are great circles on the sphere. relations on the sphere. 127. Fig. Menelaus derives from these ratios. but rather take up the applications of the transversal proposition in the Almagest. o. AEr. the second relation. Chapter 14. And the ratio r A : AE is obtained by composition of the ratios rLl : LIZ and ZB : BE. In modem notation: rE rz LIB rA rLl ZB AE= LlZ'AB and AE = LlZ'BE' To prove. Menelaus derives other propositions which have played a role in spherical trigonometry. compounded from the ratios rz : LIZ and LIB : AB. If ALlB. Then the ratio of r E A to AE is equal to the ratio.sdoidtt. 126.

26 no. one hardly finds a single allusion to mathematics. Even the architect Vitruvius. Selsk.D.e.e. says next to nothing about mathematics. who loves to display his learning in various fields. then He = 6 is the required declination. Let Z be the pole of the equator and let the great circle ZH meet the equator in e. Astrology was tremendously popular.. in the time of Vergil. Neugebauer. mathematics had sunk to a pitifully low level.·ftl.Jtlel. but he had no interest at all in mathematics. so that EH = 1 is the given longitude. The inclination of the ecliptic. i. s. and for these mathematics was indispensable. 1 I O. chd(lBOO) chd(lBOO) chd (21) chd (2e) = chd (26) . Kg\. Let the sun be in H. Cicero knew a good deal of philosophy and something of astronomy. must of course also be given. during the period in which the power of the Roman empire was spread. 2 (1938).276 CHAPTER VIII for a given longitude 1. the astrologers needed astronomical cal- culations. These methods enable Ptolemy to solve all right spherical triangles. The period of Menelaus and of Ptolemy was one of revival in mathematics. we have now chd (2ZA) chd (2Ze) chd (2HE) chd (2AB) = chd (28H) . Ptolemy draws the circle ZBA. chd(lBOo)' or. Danske Vid. How can we establish this date? In Heron's Dioptra a method is developed (unnecessarily complicated) to determine the difference in time between Rome ~d Alexandria by means of a iunar eclipse and Neugebauer observed that the eclipse which is used there as an toxample is probably the one of the year 62. According to the transversal proposition. M. chd (2EB)' I. There is a beginning of a new revival of interest in Heron of Alexandria around 60 A. of which two sides. the angle at E between the ecliptic and the equator. HisI. or an angle and one side are given. which passes through the poles of the equator and of the ecliptic and is therefore perpendicular to both. But during the two centuries before Menelaus. . And this appears to be sufficient for spherical astronomy. of Horace and 0 Ovid. chd (26) = chd (2~k~d (2e) . During the reriod of splendor in Roman literature. This agrees with our modern formula sin 6 = sin 1 sin s. since the chord of the arc of 180° is equal to 120 in Ptolemy's units.

he treats the con- struction and the use of various measuring instruments.13 = 8 21 -14 = 7 21 . particularly those for volumes. The proof makes use of the inscribed circle of the triangle. .II)(S ."2- Cantor ascribed this formula to Heron. This prescription is equivalent to the well-known formula lI+b+c A = Vs(s .15 = 6 21 x 8 x 7 x 6 = 7056 V7056 = 84. Several formulas occurring in Heron. and gives two methods. because it appears in Heron without indication of a source. This is as if one were to attribute to Hiitte all formulae in Hiitte's Pocket Book for Engineers. As to the others. lifting machines. It is explained as follows: 13 + 14 + 15 = 42 42/2 = 21 21 . but usually the source is anonymous. In book I. such as water clocks and diopters. book III the division of areas and volumes in a giv"n proportion. Finally a number of works on areas and volumes. applied geometry and appiied mechanics. Heron next gives a geometrical proof of this formula. etc. and of machines. 14. book II the calculation of volumes. s=-". The second method avoids the calculation of the height. 15. the best known of which is called Metrics. Book I of this treatise treats the calculation of areas. for which no source is indicated! However. war engines. THB DECAY OF GRBBI: MATHBMATICS 277 Heron's work is in the nature of an encyclopaedia of elementary geometry. In a series of works. Occasionally some- thing is added. such as pneumatic machines. automatic machines. He also wrote a commentary on Euclid and a work on definitions.:. Usually such things are transmitted practically unchanged. sometimes the name of the inventor of a formula is mentioned. Heron assumes the sides to be 13. Some of the numerical examples in Heron's treatise are already found in cuneiform texts. are taken from Archimedes.b)(s -II). The first consists in first calculating the height of the triangle. century after century. we now know from an Arabic manuscript that "Heron's formula" is due to Archimedes. the history of the rules of calculation and of the numerical examples is almost impossible to trace. most remarkable is the calculation of the area of a triangle with given sides. by a method well- known from our schoolbooks.

This may have been Saint Dionysius. 255. Let us rejoice in the masterworks of Archimedes and of Apollonius and not mourn the loss of numberless little books on calculation.It -iIi.. His principal work. Geschichte der Mathematik) : Hier dies GrabmaI deat Diopbantus. Beginners are inclined quickly to lose courage. By far the most original mathematician of late antiquity (the Roman era) is Diophantus of Alexandria. dedicated a book to Diophantus. who became bishop of Alexandria in 247. And. It is mankind's really great thoughts that are of importance.. als es dem Schicksal erlag. I have begun with an explanation of the nature and the power of numbers.).. We leave it to the reader to determine from this.. ouvrage chino. Duu em Siebentel noch. da schloss tt <las Biindnis der Ehe. who became bishop of Laodicea in 270. ~ut it Will be easy for you to understand it. die Halfte der Jahre Hatt' es des Vaters erreicht. Drauf vier Jahre hindurch durch der Grtissen Betrachtung den Kummer Von sich scheuchend. but it is next to impossible to prove their dependence or to trace the road along which they were transmitted. his life is described as follo~s ~ have taken the translation from Cantor. B 2.' du me siMe. This may appear to be more difficult than it is. auch er Iwn an <las irdische Ziel. StartJDg with the foundations on which the matters are based.D. Qwllm ""J Sttuim. It is uncertain when Diophantus lived. Wehe <las Kind. Noch ein Zwolftel wu. The oldest Chinese collection of problems on applied proportions 1 looks like an ancient Babylonian text. Knabe zu sem gewahrtc ibm Gott ein Sechstel des Lebens.278 CHAPTER VIII Heron's examples are repeated endlessly in the Roman era and during the Middle Ages. I P L. sprosst' auf der Wange der Ban. because it is not yet known. Nach fiinf Jahren entsprang aus der Verbindung ein Sohn. <las vielgeliebte. found in the Anthologia Palatina. But Psellus says that the learned Anatolius of Alexan- dria. after all. it is not very important. that you exert yourself to learn the study of num~ problems. In an algebraic puzzle rhyme. Le ''''''.' de . Schauet <las Wunder! Dutch des Entschlafenen Kunst lwet sein Alter der Stein. which age Diophantus reached. the Arithmetica is dedicated to the "highly honored Dionysius". according to Bombelli he lived during the reign of Antoninus Pius (150 A.9. not their dilution in popularizations and in collections of problems with solutions. . van Hce. The dedication starts as follows: Knowing. This leads me to think that most probably Diophantus lived about 250. thanks to your enthusiasm and to my explanation.. according to Abulfaradj during that of Julian the Apostate (350). p. highly honored Dionysius.

Diophantus knows also that a number of the form 8n + 7 can not be written as the sum of three squares. when either A or C is zero. how Diophantus proceeded in solving such equations. We also know that the Babylonians were acquainted witl1 the general solution . which have only one positive solution. r.b. which usually have infinitely many solutions. In the remaining 5 books. As far as diophantine equations of the second degree are concerned. one finds especially indeterminate equations. His method varies from case to case. He is usually satisfied when he has found one solution. Zl. which are effective. THE DECAY OF GREEK MATHEMATICS 279 The Arithmetica contains 189 problems with their solutions. with the aid of solved examples. Diophantus also has a method for "double equations" of the form (3) ax! + bx + . is first indicated by Aryabhata (500) and elaborated by Brahmagupta (625). we find in Diophantus the general solution of the equation of the "Pythagorean triangles" (1) xl + r = Zl an<:l also special methods for indeterminate equation of the form (2) M+Bx+C=r. or when A + C is a square and B is zero. it makes no difference to him whether the solution is integral or fractional. JxI + ex +f . There is not a trace in his work of a syste- matic theory of diophantine equations. This about completes the general ideas to be found in Diophantus. Nearly all the problems of the first book lead to determinate equations such as as . also called Diophantine equations. or asa = b. We shall see presently. . At an earlier point we have discussed the methods ascribed to Pythagoras and to Plato for the solu- tion of the diophantine equation (1). when A or C is a square. The general solution of diophantine equa- tions of the first degree ax± by -. Diophantus admits only pontifle rational solutions. About The precursors of Diophantus in the field of indeterminate equations. we know precious little.

Wertheim'. . as to the tradition to which Diophantus belongs. I also remind the reader of the "Flower of Thymaridas" and of Babylonian al- gebra. but his algebra is so mature that we can not assume that he discovered everything himself. In a papyrus from the second century A. It is rather a general impression of relatedness which makes itself felt when one knows the cuneiform texts and then looks through Heron or Diophantus. • Anthologia Palatina. systems of linear equations are found with two or more unknowns. p. A large number of them is found under the name of Metrodorus in the Greek Anthology. p. were very popular in Greece. I In contrast to these popular collections. Sci. e. papyrus 620. without stories. one gets more and more the impression that he has drawn on older sources which in some way or other are connected with Babylonian algebra. of which I Karpinski and Robbins. there has always existed a more popular tradition of small algebraic problems and methods of solution. Certainly problems 27-30 in the first book of the Arithmetica are closely related to Babylonian problems. On the other hand. dressed up in verse form. The algebra of Alkhwarizmi can hardly be accounted for on the basis of the Greek and Indian sources which we know. too many connecting links are missing for this. I. See. such as the epitaph cited above.. It is probable that the tradition of these algebraic methods was never interrupted so that. German translation of Diophantus. which we shall encounter again in Diophantus. Alkhwarizmi was the first writer on algebra. we do have an idea with respect to determinate equations. and which ends in Arabic algebra. The Pythagoreans also knew infinitely many solutions of the diophantine equation (see "lateral and diagonal" numbers. into China and India. or the Aryabhayta of Aryabhata or the Algebra of Alkhwarizmi. Connections with Babylonian and Arabic alsebra. 125). with radiations into Greek culture. a tradition which originates in Babylonian algebra. 311.q. along with the scholarly tradition of Greek geometry. I also refer to the "cattle problem of Archimedes". As early as the time of Archimedes.g. little algebraic problems.. or the Chinese "classic of the maritime isle". Diophantus gives nearly always a purely arithmetical formulation.D.280 CHAPTER VIII y= z. They are: To find two numbers. Not much more is known. We have no real proofs for the existence of such an uninterrupted tradition. According to all Arabic sources.ce 70 (1929). Michisa. and so are the solutions.

= AedJ". For the addition of terms. if necessary.• Cambridge 1910. Diop". 28) the sum and the sum of the squares are given. the other to a . Either he expresses the other unknowns in terms of that one. Alkhwarizmi does not even use the abbreviations of Diophantus. Diophantus sometimes calls. THE DECAY OF GREEK MATHEMATICS 281 27) the sum and product are given. (S3) K Y = K{.~ unity. In a similar way. besides the "number" '. (s') K Y K = KV{JOIro{JO. his algebra is 1 For a detail. no matter how complicated. greater than and less than.d discussion see Th. p. 1 Just as Alkwharizmi. following their example. .s and then obtains an equation in s. something like an inverted 1'. 30) the difference and the product are given.~ square. multiplication. which looks somewhat like an s and which is also found in the Michigan papyrus 620. •" of Greek Matlttm4tics. He had a special symbol for the unknown (which we shall always denote by s). Moreover he has names for reciprocals of powers such as . but 12 units. or he gives them arbitrary values to be modified afterwards. Diophantus reduces all his problems. for them he uses the following abbreviating symbols: (l)Af = Mo. to equations in one unknown. = cubocube. or the same author's Hi. just as the Arabs did later on. he sets the numbers equal to s + band s . Thus he does not simply write 12. Diophantus simply writes them in a row. LfYrMt/f = 3$· + 12. "the side". equality. Heath. Alkhwarizmi also reduces all his problems systematically to equations in one unknown. in order to satisfy the other conditions of the problem. number 2. 29) the sum and the difference of the squares are given..{Jo.g. (s) .1. The method of solution agrees completely with the Babylonian method. the square s' occurs also. Diophantus sets one of the numbers equal to a + s. e. he expresses in words. Diophantus uses special names for the powers of the unknown s. and. . When the difference 2b is given. cube. 456. so does AI· khwarizmi. or from the final a of Arithmos.2. etc. Diophantus has a definite sign for subtraction. (S2) LfY = Lfv"a". Perhaps this symbol was developed from the first letters. When the sum 2a is given. 2nd ed . Alkhwarizmi usually writes 12 dirhems. The algebraic symbolism of Diophantus is still rather primitive."tw of AIex4nJria.d. Everything else. And yet Alkhwarizmi is independent of Diophantus.b. the Italians..

it then takes e. p. or hs +c= lUI. Diophantus explains how one multiplies and divides powers of the unknown.. then multiply thia by itoelf. To divide a given number which is the sum of tllO ~uares into two other squares. together 5s. are exactly the 6 standard forms of Alkhwarizmi.3. or eight and subtract from it ~lf the number of rooto. Add thia to the nine and thirty. The squares are therefore (s + 2)1 = .IS to one ' .&. Take ~lf the number of '''''. • A typial example from the algebra of Alkhwarizmi. the sum ia equal to nine and thirty. Then one of the squares is Sl + 4s + -4 and the other 4s1 + 9 . and (2s . Their sum is indeed 13.l to nine and thirty dirhema. In the part of his work that has been preserved. and ten of ito "". how polynomials are multiplied and how terms with the same power of s can be combined. Next follows what the Arabs call aljabr and almukibala. e. or lUI = h and it is easy to solve... case five. We shall attempt to throw some light on Diophantus' procedure by means of a few typical examples From Book II. hence s = 8/5. The 6 forms of equations of the first and second degree. which gives sixty.g. this promise is not redeemed. or lUI = hs. Diophantus promises to discuss at a later point the case..g.. and there remains three. take the square root. M_. In free translation. In the simplest case..Is an eq .. namely five.. ed. t~t ia if you add ten "".N/. Problem 9. although it becomes evident that he knew how to solve quadratic equations like lUI + hs = c.12s. the sum of the squares of 2 and 3. 1831.. the form IU = h. in the traDslation of Rosen (TIre -'B''''' 0/ M""-'I .. these operations will leave only one term on each side of the equation. or lUI +c. + 13 . and the result ia five and twenty. Aljabr (hence our word algebra) is the transposition of negatives to the other side.. mentioned here. This has to equal 13. the solution is as follows: Let the given number be 13.four. 2s .282 CHAPTER VIII still entirely "rhetorical" and in this respect more primitive than that of Dio- phantus.3)1 = 1/•. It looks therefore as if the solution of the definite equations is a traditional topic for Diophantus. 1 In his introduction. and almukibala is subtracting equal terms from the left and right members. The solution ia as follows. This is tbe root.. as well as the Babylonians.. 5) is the following: "A ' . Rosen. Let the side of one square be s + 2 and that of the other 3 units less than an arbitrarily sdected multiple of s. t~t ia in thi." . hs. Much more original is his virtuoso treatment of indeterminate problems. and later writers restauratio and oppositio. London. the Indians and the Arabs. in which two terms remain on one side.

. Let the given numbers be 2 and 3. Another typical ingenious device. This is accomplished by factoring the difference and equating the factors to x + y and x .8s. then the product of the first two is 13. increased by 12. Therefore s + 2 and s + 3 must be squares. then this product itself is 4. The method is applied for the first time in Problem 11. To find three numbers such that the product 0/ tflO 0/ them. one sets it equal to (21 .I. THE DECAY OF GREEK MATHEMATICS 283 Problem 20. is derived from the rule xS. equal to s + 3 . Choose for the two number the expressions sand 21 + 1.g. In such a case their difference has to be equal to the difference of two squares. The square of the first plus the second then becomes automatically a square viz. Diophantus sets the required numbers equal to simple expressions in s. Most of the problems in Book II are solved by this method. multiplied by itself. e. we call something like this a . multiplied by itself.. Set the product of the first two plus 12 equal to 25. Set the product of the second and third numbers plus 12 equal to 16. equal to s + 2. the second to s-1. The product of the first and third is 5211. Therefore set the first number equal to 13s.. The square of the second plus the first becomes 411 + 5s + 1. The method is applied whenever each of two expressions in s has to be a square. 4 and %. difficulty to the solution s = I'll' The identical method is used in both examples. produces emy time a square. . Half the sum of these factors is x.3 in the first example. hence the third is 41.2)1. He obtains then a quadratic equation in s and he knows how to arrange matters cleverly so that either the term in Sl drops out.)(x -. To make this into a square. so that he can solve for s rationally. Take their difference and look for two numbers whose product is equal to this.y. Now the product of the first and third plus 12 must also be a square. the "method of double equality". we add s to each. and half the difference is y. hence 5211 + 12 must be a square. which leads without. To find tflO numbers such thilt the square 0/ either added to the other sifles a square. Sl + 21 + 1 = (s + 1)1."1M' The number to be added is therefore "1M' From Book HI. in each case onefinds s . double equality". Set &alf the difference of these numbers. To add the same (required) number to tflO Siflm numbers so as to ma1ce each of them a square. such as s + 2 and 2s ... But the second number is S-1. or set half their sum. Problem to. = (x + . or the constant term. Thus one finds the equation 411 + 5s + 1 = 411 + 4 .

so that one finds 3s + 18 = 5sl . Then Sl + 1 times 18 plus 2%. Let 3 and 5 be the added numbers.l' . e. and then increased by 2%.. the second to S-l and the third to %s.. s. s + 3.g . . the product of the two sums sifles a square. This is easy 1.2". each of which would produce a square if increased by 12. Addition of 3 and 5 to these parts gives s + 3 and 6 . Now 52 was found as the product of 13 and 4. he looks for the causes of the difficulty and for the changes in the assumptions which will enable him to solve the problem.. The product of the first and third is S3.g.s. one finds s = 18.)". His formula was evidently b+~ s=------ a 1 e'II" by . to 41'. we can choose 4 for one of the squares and % for the other. This has to be equal to a square. must be a square..%)(3% + Yz) = (3Y. and again 12 = 3 X .. Apparently Diophantus knew how to solve the general quadratic equation of the form asl = 2bs + c. .. then the square becomes S2 + 6s + 9 and we find s = %. He wrote the solution as a fraction with denominator a and he knew that the solution is rational if hi + ac is a square. if siflen numbers are added to them respect. or 1&1 + 20% must beasquare. The required square Sl is therefore 324.. then we find s = 78/. To diflide unity into two parts. = '/18' The first part becomes ./•• the second 18/18 . and (Yz)1 + 12 = (3 %)". Let us therefore start all over again and set the first number equal to 4s. Set it equal to (&+9)1. this coefficient multiplied by 18. he starts by assuming certain expressions for his unknown numbers.. for example. when this brings him into trouble. Thus we have to look for a square which. Their product is therefore 3s + 18 . Choose as the side of the square. Now we return to the original question: 3s + 18 . in- creased by the square of one half of 3.. the other 1 . Takethis four times: 72s' + 81 has to be a square. Take for it 324s1. What Diophantus applies here is the "method of the wrong hypothesis". increased by 1.flely. From Book IV.IY.s. hence S3 + 12 must be a square.Sl. e. if these were both squares their product would be a tile method of the "double inequality" and factoring 12 into two facton 12 _ 2 X 6 - (of ..284 CHAPTER VIII 1£ the factor 52 were a square. produces a square. We still have to make sure that that the product of the first and third plus 12 is a square.. such that.2)(4 + 2) . Let it be equal. In order that the equation may have rational solutions. This solves the problem. and. The coefficient 5 was a square plus 1. This equation can not be solved in rational numbers. and multiplied by 18. (Introduction of a new number s:) Let Sl again be the required square. it would be easy to satisfy this condition. Thus we have to find two squares. must become a square. increased by 12. is a square. = (3% . Let one part be s.. Thus we iind 2" + 12 . A very pretty example of the method of the wrong assumption is the following: Problem 31. since each of them..

this is certain. p.. From that time the Greeks also began to show on many occasions a marked love of the colossal. Pyramids).. were quite uncommon for the Greeks. representations on coins (compare Plate .. nothing un· usual for us.D.----. fig.D.C.. built by Sostra· tus of Cnidus during the time of Ptolemy Soter (3ll-283 B. Before. . Kultur.. It seems to be connected with older to- werlike graves in Asia minor.). P. PL. .. An older fonn is 'the fire·column (harbor of Piraeus). Our plate is taken from his book "Der Pharos von Alexandrien". the Mausoleum of Halicarnassus(about350B. The tower is a splendid example of . --. Reconstruc· . based chie8y on . The lighthouse stood at the entrance of the port of Alexandria.: the French name "phare'')... 1909.. R. Vb. only from about 50 A.. 19. Va. "the most famous harbor of the ancient world". prototype of all lighthouses (notice . Alexandrian coin (2nd century A. Thiersch. is perhaps. The line of development. Waner.)shows a notabk tendency to verticalism.. More than N -0 100 meters high.... Completed about 280 B. The structure survived un- til the 14th century A. It is still open to question whether this and similar towers served as lighthouses from the start. starts in the Orient (Tower of Babel and cognate monuments. PLATE 27 PL. in the Alexandrian Pharos.. Hell. There .. . High towers. tech- nology and art in the Hellenistic period. -. 36.) with a representation of the lighthouse at Alex- andria.. See B.-. Lighthouse on the island Pha· ros near Alexandria. C> Vb)....C. Lycian gra- ves..D.. tion by A. With Alexander the Great the situation changed.C.~ the collaboration of science.. also a reminiscence of Persian watch ~ towers. .. . the I ····· -.

vaults and cupolas. It is quite dilferent with the Romans. The construction of arches. but the Hagia Sophia is not conceivable without the ftourishing of the Roman construction of vaults and cupolas. PLATE 28 PL. vaults and cupolas did not playa role of importance in Greek architecture.. The "Hagia Sophia". Already in the second half of the 5th century B. concrete). Democritus studied the theory of the arch. these forms have been developed enormously. yet the Greeks applied it but little in practice and not until late in the Hellenistic period. one has but to think of the Pantheon. lime.C. . Oriental inftuences are also active. 28. helped also by the introduction of new building materials (bricks. Connecting with early Roman and Italic constructions of arches.

the second 20/41. This is not a cube.12 has to be a square.6Y2. 1 has to be made greater than 3. Y = 2pq. Their difference is 12 . Now 1 was found by dividing 9 by 5. Then the second part is 4 . Reduce these numbers to the denominator 64. Diophantus imposes turther conditions on the sides or the area. again produces a square. Book VI takes up right triangles with rational sides.6%1.e.s and the condition becomes: 9s . then one finds 1 = 9/ 5. z = p2 + ql. and the square between 2 and 1 %.' Therefore the first number will be 2l/u. from which follows 5 = 1"/41. i.e. Now let the first term of the progression be 42 %. 16/11 . If 1 lies between 3 and 4. Now it is easy: the square is 100/840 . we equate this to the smallest expression 6%s -12. so that the numerators become 80 and 128. Resolve this in two factors: sand 1 . Let the triangle be formed with the aid of two num bas. Diophantus knows that he can obtain such triangles "with the aid of" two num· bers p and q. Look for a square which. i. Returning to the original problem. Prom Book VI. In each of the problems of Book VI. The hypothenuse is then S2 + 9. in which he operated with inequalities: Differently. but less than 4.52 = (26/11}52. A simple example of the "method of the double equation": Problem 1. But we can not subtract 3 units from this. e. decreased by 12. The hypothenuse less the base is now 18. To find a (rational) right triangle such that the hypothenuse minus each of the sides gives a cube.12 and 6%1 . then the square plus 1 must lie between 3 and 2%. we find the equations 95 . sand 3. One has to take therefore (in place of 3) a number . THE DECAY OF GREEK MATHEMATICS 285 Diophantus gave a second solution for the same problem. Thus w. tine equation x2 + y2 = Z2. To find three numbers in geometrical progression such that each of them decreased by 12 gives a square. then the middle term is 6%1. An example: Problem 1. Half their difference multiplied by itself is 119/11 . This is easy: the square is 42%.12 must be a square.e. Therefore we have to look for a square that is greater than 1 % but less than 2. according to the formulas x = p2 _ q2.g. the third term S2. Where does the number 18 come from? It is twice the square of 3. the height 6s and the base S2 .~ get 1 = H1/104 • etc.9. i. Let the first part be s less the 3 units which have to be added to it. solutions of the diophan. Now each of the two expressions S2 . by a square plus 1.' Prom Book V. Set this square equal to 411 .

and he determines the area bounded by the spiral and an arc of circle. for the unknowns such expressions in s. . A . the work is done.7 AI. he gives a new construction for the five regular solids inscribed in a given sphere. is a cube. the "Mathematical Collection". etc. in French. and s becomes 2. He shows how a Neusis-construction. 1 The reader will gradually have gotten some idea of his principal work. if we set s . either in Heath's English translation. he frequently selects. Rome.320 he observed a solar eclipse and he tells about it in his commentary on the Almagest.2. In the same book Pappus defines a type of spiral on the sphere. on which we have already drawn several times. 4. Diophantus has more than one condition which the unknown numbers have to satisfy. This gives us a great deal of varied information about what was contained in the lost works of Euclid and of Apollonius. We happen to know when Pappus lived: on October 18. 4. should get hold of the work of Diophantus himself. the method of analysis. he formulated them as lemmas. then lsi has to be a cube. Therefore. In Book 3 for example. the duplication of the cube. and the base 96. Take for this cube 8. The difference of hypothenuse and base is a cube. can be reduced to the intersection of two conics. they satisfy them no matter how s is chosen. the method being taken from Archimedes. Book 4 contains an interesting generalization of the Theorem of Pythagoras and a number of nice propositions on circles inscribed in the Arbelos of Archimedes. the height 40. The hypothenuse becomes Sl + 4.. The hypothenuse is 104. or in that of Ver Eecke. Whenever. on· the quadrature of the circle. In this work. e. Anyone in whom these examples have developed a taste for it. Bruxelles. the height 4s and the base Sl . + 4 . "in general". Wherever explanations of or supplements to the works of the great geometers seemed to him necessary. Call this number again s. p. the trisection of the angle. i. Now this expression is a square of side s . The last condition then leads to an equation in s from which s can be determined.6. In one more case has the Alexandrian school produced a mathematician of calibre. s' The other condition requires that the hypothenuse less the height. Now form again a triangle with the aid of the numbers s and 2. Pappus supplemented and extended in several points the work of his predecessors. be a cube. i. Pappus brought together everything in the work of his predecessors which interested him: on higher plane curves. sc. then we get s . namely Pappus of Alexandria.2 equal to a cube. Soc. sa. The triangle is therefore formed from 10 and 2. taken twice. as he does here.e.286 CHAPTER VIII whose square.It. that they satisfy all conditions save one.10.en'. I Sec A. used somewhere by Archimedes. as in this case.4s.g.4.e. But furthermore..

In the introduction to Book 7 of the Collectio. three which lie on one line (a). because it gives a survey of a large number of works on geometric analysis and on geometric loci. which is essentially the same as what we know as the "theorem of Pappus-Guldin". stated as follows: B Pig. THB DBCAY OP GRBBK MATHBMATICS 2137 In Book 5. Book 8 is largely devoted to mechanics. J in Fig. . If. 128. we shall now discuss thirteen lemmas on the Porisms of Euclid. S two points are held fixed. pass through one point S. lines (p and q). nearly all of which are lost. To give the reader some notion of the high level of this powerful work. Pappus might have added that the three lines p. Book 7 is extremely important. and if two of P' 129 the remaining three (P and Q) lie on fixed . . supplemented by some propositions of Pappus himself. In Book 6. in case of parallelism. Pappus formulates a proposition on the vo- lume of a solid of revolution. A porum of Euclid. The book contains moreover a construction of the principal axes of an ellipse when two conjugate diameters are given. or. given as the perspective of a circle. 1213). The motivation for this is the problem of deter- mining the diameter of· a cylindrical column from an arbitrary fragment of the column. Pappus determines the center of an ellipse. . Historically. but it also contains a construction of a conic through 5 given points. Pappus says that a number of porisms of Euclid can be summarized into a single proposition. They contain ideas which were destined to become later on of great importance for projective geometry. c. As a discovery of his own.g. q. the work of Zenodorus on isoperimetric figures is reproduced. of the 6 points of intersection of 4 lines (a. then the sixth point of intersection (R) will also lie on a fixed line. b.

. Lemma II is analogous. added by Pappus.) are cut by a fixed line a. This proposition. in our modern terminology. then. then the lines joining corresponding vertices (P and P'. If one can prove that the sixth point of intersection G is uniquely determined by the remaining five points B. whose 6 sides are being met by one line. d and b'. C. in lemmas I and II. and this proves the proposition of the four lines. E and F is some- times called the Second Theorem of Desargues or the Theorem on the complete quadrangle. one obtains the diagram of the well-known Theorem of Desargues: If two triangles PQR and P'Q'R' are so placed that the points of intersection of corresponding sides (b and b'. and conversely (fig. EZ = AZ . We shall not discuss the generalization to more than 4 lines. d and d') lie on a line (a). Pappus formulates this proposition in his own manner. viz. The diagram of the proposition on the 4 lines can also be interpreted as the figure of a complete quadrangle PQRS. whose sides (PQ. fixed lines are drawn in full. Lemma I runs as follows: E Let figure ABrL1EZH be drawn and let (1) AJ:JF=AZ:ZH and let e be joined to K. Q and Q'. JE : AJ . 130. etc. then it follows imme- diately that. D. c and c'. and in lemma IV for the case that all six points B . Br : AB . the sixth point will "lie at infinity". E and F. for the case that one of the sidt>-s of the quadrangle is parallel to a. when P moves on the line ES and Q on the line FS. (1). as we shall see presently.1 H r 5 of the 6 points satisfy the relation Fig. This is indeed the case. I assert that eK is parallel to Ar. The theorem on the complete quadrangle.G are at finite distance. d'). except that Ke / / Ar is part of the hypotht>sis Lemma IV says: Let figure ABrL1EZHeKA be drawn and let (2) AZ . c. C. QR. If A Z.288 CHAPTER VIII In our diagram variable lines are dotted. the point R must move on the line GS. but we shall raise the question: how may Euclid have proved the proposition for 4 lines? It is an obvious conjecture that the proof can be given by means of the lemmas which Pappus then develops. rz. Rand R') ·pass through one point (S). that G is uniquely determined by B. c'. If one ima- gines two different positions of the dotted lines (b. One recognizes in the figure the complete quadrangle BEeK. D. 129).

P 19. A One recognizes again the complete quadrilateral (HeKA) whose 6 sides are brought to intersection with one line. r. then the points of intersection H. for this is exactly what he formu- lated as Lemma III. Lemma III.. Fig. Draw the lines eE and eLi so as to intersect the three lines AB. ELI anJ Er are drawn. coincide. Br. Lemma X is the converse of lemma III A and lemma XI is concerned with the special case in which one of the transversals is t>a- raIle! to one of the four lines. AZ. In his proofs he might have used the fact that the cross-ratio of four !. Br = 8E·Hz:eH·ZE. 133. Then (2) is replaced by the proportion (3) AB : Br = ALI : Llr. Pappus proves his lemmas every time E by drawing paralle!lines and by cleverly juggling with proportions. AT. ALI. H. If the points E and Z are taken en the lines AB and r LI. By means of these lemmas Pappus pr{)ves the famous Theorem of Pappus. B. of two points each. E. Nowadays we say that the six points A. LI. 8~----~--~------~ Z H which runs as follows: Fig. LemmaV refers to the special case in which of the six points two sets. Z are in A tJ E Z in"olution. rLi : eLi . M will lie on one line. It is however true that he only considers the special case tJ in which one of the four points is the point of intersection of the two lines. 132. In modern terminology. · 131 . . 132).oints remains un- changed under projection on another line. the re- lation (2) expresses that the cross-ratio (AZELI) is equal to the cross-ratio (ZABr). Z lie on one straight line. In this case we speak of four harmonic points (fig. I assert that eB . and if ALI. K. BZ. THE DECAY OP GREEK MATHEMATICS 289 I assert thu the points 8.

.oclw Diadochw. After Hypatia. Indeed. Theon of Alexandria (380 A.. The later bis~op Synesius calis her his "mother. According to the church history of Socrates.stl II (1936). Rume. . One day. The Athens school. Although on terms of friendship with many Christians. which had access to its own ample financial means.. He acquired his greatest fame not as a mathematician. 134. sister and reverend teacher". mortal enemy of bishop Cyrillus. In the Christian community she was accused of having incited Orestes against Cyrillus.. with freedom and yet with decency. maintained itself for a longer time. One of her friends was the Roman prefect Orestes. There is a tale that she brought a desperately amorous lover to his senses in a drastic manner (see "Suidas" under Hypatia). which says that the points of intersection of opposite sides of a hexagon inscribed in a conic lie on a line. she moved among men. This became the cause of her downfall. 5th century. she was charming. she was eloquent.) was a rather mediocre mathematician. Alexandrian mathematics came to an end. she had remained faith- ful to the traditions of pagan Hellenism. su. her corpse was torn to pieces and burned. and she became professor of Platonic philosophy in Alexandria.D. in the temple of St'rapis.0. had been destroyed by the Alexandria mob. The Academia of Plato. the only remaining library. She had connectibns in the highest circles..290 CHAPTER VIII Pappus formulates and proves the proposition first for the case that AB and rLI are parallel (lemma XI~ and then for the case A E B that they intersect (lemma Xll~.d. d·AItx. but as the father of HJPIIIia a very learned woman. but also in practical affairs her counsel was appreciated. she wrote learned commentaries on Diophantus and on Apollonius. She was handsome. Nowadays the theorem is considered as a special case of the Theorem of Pascal."""". C. heroine of romantic atrocity tales. This happened in 418. a new edition of Euclid's Elements with rather unimportant additions and emendations. she was hauled from her carriage by the mob. A womiln. not only in literary matters.toin: de Popp" " Je 1'1. I'AI. stripped and assassinated with shards.i. dragged to the church. The industrious neo-Platonist Proclus (450) 1 Sec 11. He wrote a commentary on the Almagest 1 and he prepared Fig. P. already in 392.. on her way home..

. Leipzig. a commentary on Euclid. 1 Prench translation by P. During the time of Damascius. Anthemius has written on burning-mirrors. Simplicius and Damascius emigrated to Persia.. Brup 19-t8. on the order of the Emperor Justinian. I After him. at any rate its last part. . for the writer speaks of the "famous Isidore" as his teacher.. They we-re experts not only in architecture.. The last center of Greek culture was Constantinople.a. After these last flutterings. edited and translated by Maniti. Here lived Isidore of Milete and Anthemius of Tralles. It is probable that the so-called 15th Book of the Elements. besides highly im- portant fragments of old philosophers. the history of Greek mathematics dies like a snuffed candle. H"". comes from the school of Isidore. became heads of the school.. the last rampart of the pagan world.riposis. the school of Athens. through whom have been preserved for posterity. there came to Athens also Simplicius. THE DECAY OF GREEK MATHEMATICS 291 wrote i. was closed. But in 529. • Procl. but also in mathematics. Book I 1 and a very nice opusculum on astronomical hypothesis. 1909. Va Recke.. the excellent commentator of Aristotle. Isidore of Alexandria. and Damascius of Damascus. called the Great. the architects of the noble Hagia Sophia (Plate 28). The book contains propositions on regular polyhedra. also the fragment of Eudemus on the Iunules of Hippocrates.

~. 58 2f>7.141. vtttical 87 Akkad 38 Angular velocity 239 Akousmata 107 Antanairesis 127. 267. 201 Agesilaus 179 Anaximandtt 82. 127.140. Applica:ion of areas 118-123. 128. rhetorical 282 Apuleius 210 Algebraic symbolism 280. 290 Apogee 240 Algebra 6. 265 Anthemius 291 Al-Biruni 226. 116. 277 Algebra.248.258. 183. 119. 178. 196. 58. 207. 80. 242 Altunate interior angles 117 Abulfaradj 278 Alyattes 83 Academia 148. 198. ApoUonius 6. 190 Adaptation 140 Analysis 90. 201. 35.201. 199.198. division of 191 Ahmes 15 Angle.202. 176. 106.126. 132 Aratus 182 Al~est50. 139. 198. 281. 63-73. 281 Aqueduct 102 Algorithms. trisection of 225 . Algebra. 145. 37. 123. 90. 290 Amasis 83 Academy 90.201. 265 199. 59.82. 286. Apices 52. 234. 229.290 Arbelos 225.200. 247 Algebra. 172.267. 237. 147. 282 Arachne 182 Allman 100. 265.275. 64 Alexandtt Polyhistor 101.85. 128. 128 Antiphon 130.229 Amicable numbers 98 Acropolis 106 Amyclas 90. 193.44 Aljabr 282 Arabic numerals 51-59 Alkhwarizmi 58. 45. 177. papyrus 17 Aetius 129 Anatolius of Al~dria 278 Aeschylus 106.202. 82. 276. 200. indian 279. 58 Al-Mansar 57 Abraham 183 Almuklbala 282 Abscissa 241. 286 Adelhard of Bath 58 Anaphorai 270 Adrastus 233 Anastasi I. 265. 56. 264. 179 Al-Batdni 57. ApoUodorus 179 118-125. 190. 281. 236. INDEX Abacus 47. 46. 44. 286 AI-Ma'mun 57 Arc. 290 281 Apotome 169 Algebra. 129.201. 230.247. geometric 71. 263 Akhmen 36. arabic 57. 194. 139. 208. 282 146. 132 Al~dtt62. 176 AO 8862 63. 280. Agatharchus 136 136. 140. 72. 148. 196. 247. 280 Alcaius 84 Anthyphairesis 177 Alcmaeon 102. AO 6484 77 210 AO 6770 73. 280. 131. 235. 266. italian 267. 205 Algebra. 212. 31. 179. 102 Apagoge 143 Alexandria 196. 57. 136 Anaxagoras 82. 281 Approximation 32. 128 Agni-pudna 55 Anaximenes 82 Agrimensor 277 Anderhub 143 Aha-calculations 27-29 Angle. 97 Al~dtt of Aphrodisias 131.272. trisection of 225. 237 -263. 130. 80.48 Angles.173. 66. 280.118-125. 228 Anthologia Palatina 278. 292 Applied mathematics 29. 270. 115 A-ra 42.

179. 82.129. 90. Method 137.242. 49. 284 Archimedes. 116. 101. 212-215. naval 202 104. 192 Astronomy. post. principal 257. 89. 196. 56. On spicals 208. 277 Augustus 82 Area of a sphere 215. 110-114 211. 82. 202. 85. Archyus6. propositions on 254 Ayuta 52 Aries 270 Aristaeus 234. 83-88. 85.119. 208 Aristotle. 97. 201 Babylonian influence 7. 68. 290 Area of a quadrangle 32. Asoka 53 224. 270. 148. Desphaera et cylindro 186.270 Archimedes. 228 Axis. 138. 230. 201. 221 Autolycus 82.107. Sand-counter 47. arabic 265 Archimedes. 75. 186. 280 206 Aryans 51 Archimedes. 62. 202. 223. 281 183. Architecture. 131. 165. Topica 176 200. 84 Archimedes. 57. 278. 179. 176. wrong 27. 271. 81 Aristophanes 130. greek 31. 94. 102. 29. Physica 131 Archibald 199. 91. Politics 149 Archimedes 6.227 Assyria 62. modem 5. 229 Aristotle. Aryabhatiya 271. Arithmetic progression 77 280. 57. 196. 198. 58.201. 279 Archimwes. measurement of the circle 204. 204. 82. On equihbria of plane figures Assurbanipal 62 213. Anal. 56. Astronomy. 205. 35. Arjuna 52 220-222. 125. 9. 184. 106 Astronomy. 100. 30 Area of a trapezoid 32. 202. 84. 171. 202-204. 200. Metaphysics 15. 67.267 Artemis 102 Archimedes. 157. 85-87. 152 Basket 33 . 107. 270-273 118.uadratura parabolae 186. 177. 82.96. 94..268. 265 Area of a circle 32. 227 Astronomy.277. 265 Area of a cone 221 Asymptote 235.266. 59. 83. 101. 106 Area of a rectangle 32.201. 279 Archimedes. 149-159. 264. 287 Areas. Aristotle.216.illet 48 Aristotle. 220. 280. On floating bodies 227 Astrolabe 182 Archimedes. 243 Assumption. 84.286 Al'ithmetica 61. 179 Ba. 72.294 INDEX Archesilaus 193. 96. 101. egyptian 31 Architecture 102. 208-228. 232 Baghdad 57 Aristotle. 34 Athens 106. 62. 51. Aristotle 7.241. 189.180-184. 92. 125 Aristarchus 7. 224 Astrology 276 Archimedes. 131. Q. 155. 86. 221 ~abhata 55. 281 Archimedes. 119. 139. 7. babylonian 31.264. 161. 258 Area of a cylinder 221 Athenaeus 90 Area of a hemisphere 33. Aristoxenos 107 236.263. 55. 201. 16. SO.115. As 41 264 Asia Minor 83 Archimedes. 73 Axioms 183 Area of a triangle 32. 180. 89 Atomists 82. 36. On conoids and spheroids 223. postulate of 186 Astronomy. 197. 199.270 Babylon 56. 229 Area of a spherical segment 222 Auxiliaries 26. indian 54. 219 102. 140.214. 125. 211 Babylonian calculating 37-45 Ariston 229 Babylonian geometry 72-76. book of lemmas 225 Arithmos 125.193-195. 179 Astronomy.241. 208. 108. 72. Lemma of 178. 107. 46. 193-195. Ba·si 44 109. 57. 102. 178. 47. 9. 109. 16. 57. 237-240.

85 Central equation 251-252 Bidez 103 Cercesura 177 Binomial 169-172 Chace 32 Bisector 204.271-276 BM 85200: 71 Chronological summary 15. Egyptian 17-30 Cochlias 208 Calculating. Proclus 90. 70 Chinese mathematics 278. 190. 242-245 Cantor 6. 75 Circle. 203-207 BM 85196: 76 Chords. 190. 48-51.263. 286 Calculating. Cassites 62 177. 124. von 275 Circle. 278 Cone. circumference of 81. 108. 186. 184. 153 Chrysippus 229 Bombelli 267 Cicero 211. 280 Berlin papyrus 6619 28. 85.207. 155. tables of 206. 152. Greek 45-50 Cochloid 235 Calculating. 137. 45. 276 Bonaparte. 75. 119. 287 Bhudda 52 Center of gravity 213-216 Bible 33. 58-60 Commensurable 142. 41 Cissoid 268 Byzantium 131 Clark 55 Cleostratus 84 Caesar 264 Cnidos 179 Calculating. 159. 82 BM 85210 : 76 Chronology 201.91. cylindrical 287 Calculating. 87. Babylonian 37-45 Cobbler's knife 225. 138. 192. 230 Boethius 60. area of 221 Cassina 35 Congruence 87 . 111. 89. 114. Jerome 210 Cifra 59 Brahmagupta 279 Cilicia 83 Brahmi digits 52. 62. 196 Belt of signs of zodiac 129 Cattle problem 208. 32 Circle. potentially 167 57 Commentaries 267 Calculations. squaring of 132. 194 Conchoid 235 Callisthenes 85 Cone 75. 31. 61. 8. area of 32. 205 Chaldeans 84. 228. 263 Briareus 209 Circle. 132. Sumerian 37-45 Commandinus 233 Calculating. 129 "Concave on one side" 220 callipus82. 29. 81. 195. 150. 286 Bushel 28. 29. 29 Cauchy 200 Beta 228 Cavalieri 138 Bh1skara 55 Celts 41 Bhattotpala 55 Center 248. Hindu 51-58 Column. 181. 277. quadrature of 123. 269 Braunmiihl. 76 Chords 88. 192. 168 Calculation with fractions 19-29. INDEX 295 Becker 7. 79 Chordal quadnlateral 206 BM 85194 : 68. 99. 76. Beer 29. 83. 132. 183. Cambyses 94 221. 54 Circle. 75. 190. 205. Commensurable. 192 Bread 23. 53. 194. 139. 178 Catalogue of geometers. 166. segment of 132 Burgess 54 Circle. measurement of 204-207 Bruins 74. 131-135. 286 BUr 66 Circle. Western-European 47. 280 BM 34568 : 74. 100. 199. 94 Blass 208 Charmides 48 BM 10250: 20 Chiera 38 BM 13901 : 69.31 135. 205. 192. sekem 26 Compasses 263 Calendar periods 62.

267 Continuity 152 DeIians 160. 175. 139. Cynics 229 180 Cypher 39 Dionysius ISO. 191. 90. 264-291 Division. 264. 58 Cubic numbers 166 DigIts. 71 Diels 92. 35. 286. 149. Diodes 139. 210 Courtier. 260 Delian problem 9. 128. 151. 137. Decimal fraction 39. 85 Diagonal 77 Cube 100. indian 51-58 Cubocube 281 Dijksterhuis 9. 149. 259. 102. Counting board. 147. Salamis 47. 241. 159-165. 146.228. hours of the 84 Division of figures 199 Decay 202. conjugate 248 237. 48 129.222. 138. 84. area of 221 Diodorus 208 Cylinder. 197.296 INDEX Conic sections 119. 62. 140. measurement of 230 Conjugate hyperbolas 248. 48. 264 Desargues 264. 267 Croesus 83. 229. 160. 202. 84.230. Dialectics 148 151. 265 Delivery. 200. 263. 211. rapid 263 Corinth 165 Democedes 85 Cosmic solids 90 Democracy 106. 215. . 92. 166. 247 Conjugate diameters 248. 91. 101. 240-261. 198-199 Division of angle 191 Datta 51. 132. 85 Diophantus 278-286. 240 Deficiency 123. 139-141.124. 83. 136. 136. 132. 203. 1. 85 Division 19. 208. duplication of 136. 156. 236. 109. 136.)9-141. 22-26. 53. fixed 262 . 22-26 . 193. 220. Gobar 52.266 Digits. 138. 16. 249 Degree. Egyptian 19. 1 I Counters of steps 230 Democritus 15. 281 Diagonal numbers 126-127 Cube. 173.202 Dioptra 103. 228 Cross-ratio 289 Determined 198 Croton 84. 40 177. 51 234. 156. 150. 245 Curves 146. 54 Division. 270 Conoids 223. 264. 86 Description of heavens 182. 290 Cyzicus 179. 226. 288 Crasis 146 Descartes 249. Diameters. 214. 268 Dickson 98 Cube roots 44. 161 Copernicus 6. Constantinople 212. 152 Cubic equations 44. 159. 190. 184.141. 149. segment of 216 Diogenes 229 Cylindrical segment 216 Diogenes Laertius 86. 71. 230-231. 104. 278 Cyprus 83 Dionysus 93 Cynllus 290 Diophantine equations 279 Cyrus 62. 287 Declination 275. 224 Delatte 93. Cuneiform writing 38. 149. 291 190. 141. 270. 96. 106. 286 Dinostratus 82. 267-268 220-222 Diodenan 191 Cylinder. 191-193 Cylinder 137.30 Day. 212. SO Data 121. 94 Conon 208. 276 Cornca 200. 39. 119. 276 Czwalina 212 Diorism 261 Directrix 259 Daedalus 193 Dirhem 282 Damascius 291 Distance of ships 87 Damascus 291 Distances of sun and moon 203 Dandapani 52 Divisibility 11 0 Darius 48. 202. 249 Degrees 50. 88.

150. 201. solar 86. 30 183. 268 Equations. 116.184. 189. explanation of 128 Eratosthenes. 91. normalized 124 Duplication of unit fractions 22-26 Equations. 72. 286 200. 73. 184. 238-240. Epicurus 229 197 Epicycle 56. 161. 173. 46. 198. 197 E1leipsis 123 Euclid. 184. 123. 89. Elements II 118. 129. 223. INDEX 297 Dodecahedron 100. 139. 124. 179. 87 Euclid 7.98. 197 Empedocles 95 Euclid. 222. 281 Equations. Pell 208 Dynamics 166. 276 195-200. Catoptrica 200. 118-121. 147. Phaenomena 195. 208. Elements X 110. 135.173. 162. 265.290 Egyptian calculation with fractions 19-28 Euclid. Epigones 267 197 Epimoric 111. Elements XV 291 Epirtiton 50 Euclid. 199 Easter 59 Equator 276 Eclipse.136. double 279. 110-115. observation of 203 Etruscan dodecahedron 100 Eclipses. 201. 252 Duplatio 18 Equations. 199 Egypt 15-36. 31-35 Euclid.184. 274 Eratosthenes 49. sieve of 231 Eclipses. 197 Ellipsoid 215. 202. 173. systems of 63-67. Egyptian division 19. 73. 265 Euclid. Elements VIII 140.183. 82. lunar 128. free form of 90. 168. 175. Elements XlII 100. 156. Epogdone 50 270 Epsilon 237 Dodecahedron. 104 Euclid. Elements 1100. 269. 101. 158 Euclid. 284 Earth's measurement 230 Equations. 196 Euclid. 80. 121-124. 109. 170. 159. 112.135. 186-187. Ecliptic 128. Elements Xl 139. 263. 174. Equations. 140. method of 283 Equality. 82. 197 Elements 90-92. 135. 47 136. 198. 159-165. 30.200 . 267. 176.179. 143. 204 Edfu 32 Euclid. 71.179. 90. 228-234 Eclipses. Elements IX 97. 196.117.92. 84 Euclid. 285 Equations. linear 66. 147. Elements XIV 263. 184 Eplcharmus 109 Euclid. 196. 287 Ecphantus 107 Euclid. 127. 208 Euclid. 230-231. Data 121. 228. 168. central 251. 220 Double equality. 195. Conica 200 Education. cubic 44. 124. 94. 139. 22-26. 266 Duplicate ratio 159 Equations. diophantine 279 Duplication of the cube 136. 108. 196. 276. Eclipse. Elements V 175. 107. 180. 237. Elements 91. Egyptian multiplication 18. 285 149-151. 73 143. 139-141. prediction of 86. Elements VI 184. 72. 120-124. Epaminondas 82 197 Ephesus 102 Euclid. Elimination 70.187 Element 190 -189. 223 Euclid. 57. 140. 116. 197. 70. Elements XII 132. 142. 78.190. 241-256. 197 E1eusis 93 Euclid. Equations. Divisiop of figures 199 Egyptian calculating 17-30 Euclid. 184. quadratic 69-71. 173. 140. 198 236-237.129. 184. Enlightenment 106 146 159. 91. double 283 "Double equations" 279. 88. 196. 184. 127. 270 Epinomis 9.167-172. Etruscan 100 Epsilontics 136. 59. 155. 170. 160. Elements VII 49. 291 Egyptian geometry 15. 177. 153- Ellipse 190. 83.178. 196 Ekpetasmata 137 Euclid. Elements III and IV 135. 137. 287 155. 155-159 Euclid. Earth's orbit 202 80. 269. 149. 196 Eleatics 82.


Euclid. Porismata 200 Fractions. unit 19. 20
Euclid of Megara 165 Frank 107
Eudemus 82. 87-89.91. 92. 118. 123. 125. Free form of education 90. 104
129. 131. 132. 150. 151. 180. 291 Freudenthal 9. 56. 73. 277
Eudoxus 7. 82. 83. 90.114. 136. 139. 148. Fritz. von 93
160. 161. 163. 164. 179-190. 194. 196. Frustrum of pyramid 34. 75. 76. 81
202. 218. 220. 221. 232
Eudoxus. Phaenomena 182 Gallus 211
Euergetes 201 Gandz 15. 33. 66. 80
Euler 136 Gani 55
Eupalinus 102. 103 Gaudentius 95
Eureka 209 Gautama 51
Euripides 106 Gellius 149
Eurytus 107 Gelon 202.208
Eusebius 107 Geminus 85
Eutocius 6. 47. 136. 139. ISO. 159. 163. General theorems 90. 183
191. 222. 263. 268 General understandings 183
Even and odd 96. 108-110. 116 Geographers 202
Evolute 261 Geometria 108
Excenter 201. 238-240. 265 Geometric algebra 71. 118-126. 141. 173.
Excess 123. 140. 247 198-199. 247. 265
Exhaustion method 184-187. 192. 216- Geometrical loci 90. 234. 259. 260
225 Geometrical progression 77. 111. 112. 153.
Expressible lines 168 219
Ezechias 103 Geometry. babylonian 72-76. 81
Geometry. egyptian 15. 31-35
Fabricius 103 Geometry. greek 15. 87
Fibonacci 60 Geometry. plane 140
Fifth 95. 96. 107. 157 Geometry. projective 258. 259. 287. 288
Figurate numbers 98-100 Geometry. solid 136-141. 156
Figures 234 Geometry. spherical 194. 195
Firmicus Matemus 271 Gerbert 58. 60
Flower of Thymaridas 116. 147.280 German numerals 59
Foci 259 Germanic peoples 41
Forms. standard 80 Gerstinger 49
Formula of Heron 228. 277 Gillain 31
Formulas 68. 120 Given in position 198
Fossey 38 Given in shape 198
Fotheringham 84 Glaucus 160
Fourth 95. 96. 107. 157 Gnomon 84.99. 104. 121. 145
Fourth proportional 185. 192 Gnomic numbers 99
Fractions. calculation with 19·29. 48-51. 57 Gobar digits 52. 58. 59
Fractions. decimal 39. 51 God 158. 194
Fractions. egyptian 19-29 Goethert 210
Fractions. greek 19. 48-SO. 115. 116 Golden Age 106-147
Fractions. indian 49 Golden proportionality 94
Fractions. natural 19. 20. 21 Golden section 90. 101. 121. 173. 184
Fractions. roman 41 Gonda 54
Fractions. sexagesimal 37-44. SO. 51 Gopa 52
Fractions. splitting 48 Gorgias 116
Fractions. sumerian 40 Gravity. center of 213-216


Greater ratio 188 Hieroglyphics 17, 18
Greatest common divison 49, 113, 115, 127 Hieronymus 210
Greek astronomy 31. 85-87, 101.102, 104, Hindu calculating 51-58
107, 129, 180-184, 193-195,201. 211. Hipparchus 51. 82, 85,182,183,195,201,
237-240,264,270-273 207,211, 265,271,273,274
Greek calculating 45-50 Hippasus 82, 106, 107, 147
Greek multiplication 47, 48 Hippias 90, 146, 191
Griffith 17 Hippocrates 82, 90, 92, 114, 131-136,
Grone Heinrich 59 139, 141. 144, 147, 155,160,161. 184,
Guldin 287 196, 226, 291
Gur 66 Hippolytus 128
Gyges 83 Hippopede 181
Histaeus 85
Halley 261 Homer SO, 201
Halsted 56 Hopfner 94
Halys 84, 86 Horace 276
Hammurabi 38, 62, 63 Horizon 194
Harmonic mean 94, 107, 156, 157,232,233 Horologium 104
Harmonic position 250 Horus 32
Harmonica 108 Hours of the day 84
Harmony, theory of 93, 95-96, 107, 111, Hultsch 195, 200, 233
157, 158 Hydraulic screw 208
Harpedonaptai 15 Hyksos 16
Heath88,92, 111.159, 161.164, 180, 197, Hypatia 290
200,203,205,212,237,241,262,281, Hyperbola 162, 190, 191, 222, 226, 236,
286 241-260
Heavens, description of 182, 193,202,228 Hyperbolas, conjugate 249
Hee, van 278 Hyperbolas, opposite 248
Hegira 57 Hyperbole 123
Heiberg 132, 159, 200, 212, 213 Hyperboloid 216, 223, 224
Heinrich, Grone 59 Hypomnemata, pythagorean 101
Helicon 161 "Hypotheses" 203
Heliopolis 179 Hypotirosis 291
Heliocentric system 202 Hypsicles 85, 263, 269, 270
Helix 263
Hellas 83-85, 202 lamblichus 56, 91, 94, 97, 116, 141. 210
Hellenism 148, 201. 202 Ib-si 44
Hemiolion 157 Icosahedron 173-175, 269, 270
Hemisphere, area of 33, 34 Ideas 148
Heptagon, regular 226, 127 Inclination of planes 31. 32, 68
Heradides 82, 94, 194 Incommensurable 142, 159, 166, 168
Heraclitus 82, 93, 95 Indian algebra 279, 280, 282
Herculaneum 210 Indian astronomy 54, 55, 56, 57,58, 265
Hermotimus 90 Indian digits 51-58
Herodotus 15, 16, 83, 84, 86, 102, 106, Indian fractions 49
201 Indivisibility of unity 115
Heron 45, 49, 77, 82, 103, 104, 139, 228, Induction, mathematical 126
239, 276-278 Inequalities 136, 185, 187, 204, 220, 239
Hicetas 82, 107 Initiation 93, 107
Hieratic 18 Integral 214, 224, 225
Hiero 202, 208, 209 Integration 264


Interval 158 Leon 90. 196
Inverses. table of 43. 44 Lepsius 32
Involution 289 Lesson·text 73. 74
Ionia 83. 84 Lever 213. 216. 217
lrrational90. 110. 117. 125. 141, 159. 160. Uvy93
166.168-172. 178. 179.190.264 ubrary 201. 202. 229. 264. 267. 290
Irrationalities. unordered 263 Labra 270
Isidore the great 291 Limit 184. 219. 220
Isidore of Milete 291 Lindemann 100
Isocrates 94 Line. tangent 249-256
Isoperimetric 268. 269 Linear equations 66. 73. 116. 124. 198
Italian algebra 267. 281 Lines. expressJble 168. 169
Italy 84. 107 Literature. roman 276
Livy 209
Jerusalem 103. 213 Loci. geometrical 90. 234. 259. 260
Julian the apostate 278 Logic 154. 155
Junge 169 Logistics 115. 116
Jupiter 181 "Long year" 129
Justinian 291 Lucian 95
Lunar eclipse 128. 274
Kahun papyrus 17 Lunules 131-136
KaDkara 52 Lyceum 229
Karpinski 280 Lydia 83
Kaye 54 Lysimachia 230
Keller 59
Kendra 56 Machines 149. 202. 209. 277
Kepler 5. 6. 265 Magi 94.96
Kharosti numbers 53 Magnitude 177. 187. 188
Kingdom. middle 16 Major 172
Kingdom. old 16.30 Mamercus 90
Klein. J. 116 Mana "II
"Knowledge without intellect" 93 Mani\ius 271
KOdel59 Manitius 182. 206. 291
Konig 85 Marcellus 163. 209. 210
Koti 52 Mathemata 108
Kugler 85 Mathematical induction 126
Mathematical proof 7. 88. 89. 90
Lal45 Mathematics. applied 29. 31. 80. 277
La1itavistara 51 Mathematikoi 107
Langdon 84 Maurice de Nassau 172
Larsa 72 Mean and extreme ratio 102. 121. 173
Lasos 107 Mean. harmonic 94. 107. 156. 157. 232. 233
Lateral numbers 126. 127 Mean proportionals 90. 118. 119. 134. 136.
Latus rectum 248
Latus transversum 248 268
Least common multiple 49. 112-115 Means 94. 231-234
Leather scroll 2n Measurement of circle 204-207
Leimma 96 Measuring instruments 104.277
Lemma 267 Mechanical method 212. 213
Lemma of Archimedes 178. 186. 221 Mechanics 149. 163. 212. 227. 277. 287
Leodamas 90 Medes 83. 84


Medieties 231-235
Meissner 38
Menaechnus 82. 90. 139. 161. 162. 190- Neusis 134. 226. 235. 236. 237. 263. 286
191. 241. 245 Newton 5. 6. 258
Menelaus 82. 271. 274-276 Nicomachus 97. 98.
Menge 200 Nicomedes 82. 139. 146. 191. 235-237
Menninger 48. 51. 52. 53 Nicoteles 249. 260
Mesotetes 232. 234 Nile floods 15
Method 137. 208. 212-215. 264 Nineveh 84
Meton 130 Nipsius 87
Metrics 277 Niyuta 52
Michel 183 Nodes. recession of 180
Michigan papyrus 620. 280. 281 Nonnalized equations 124
Middle kingdom 16 Nonnalized problems 122. 198
Milete 85 Normals 260
Minor 172 Notation. positional 38-41. 53-61
Minos 160. 161 Number mysticism 93-100
Mishnat ha-Middot 33 Number symbols 52. 53
Mixtures 146 Number systems 17.37
MKT63 "Numbers" lOB. 265
MohaIllllled 57 Numbers. auxiliary 26. 30
Money 41 Numbers. figurate 98-100
Monochord 95 Numbers. gnomic 99
Moon 128. 180. Numbers. lateral 126. 127
Moon in second quarter 203 Numbers. parallelopiped 98
Moscow papyrus 33. 34 Numbers. pentagonal 98. 99
Muhammed ben Musa 57 Numbers. perfect 97. 98. 109
Muller 205 Numbers. poetic 54. 55
Multiplication Numbers. prime 115. 231
Multiplication. egyptian 18. 30. 47 Numbers. rational 169
Multiplication. greek 47. 48 Numbers. similar 140. 154. 158
Multiplication tables 42. 43. 46. 59 Numbers. spatial 140. 154. 156. 158
Musaeon 201 Numbers. square 98. 99. 154. 166
Music. theory of 94.95. 107. 111. 157. 158 Numbers. triangular 95. 99
Musical scales 96. 149. 157 Numerals 51-58
Myriad 46 Numerals. arabic 51-59
Mysteries 93. 94 Numerals. babylonian 38. 39. 40. 56
Mysticism 94. 157. 158 Numerals. egyptian 17. 18. 29. 30
Mysticism. number 93-100 Numerals. german 59
Numerals. greek 45. 46. 47
Nabonassar 62 Numerals. indian 51-58
Nabopo1assar 84 Numerals. roman 17.38. 39. 59. 60
Natural science 194 Numerals. sumerian 48
Naucratis 83
Naval architecture 202 Octahedron 173
Nebukadnezar 84 Octave 95. 96. 107. 158
Nectanebus 179 Oenopides 82. 90. 129. 130
Nehemiah 33 Old kingdom 16.30
Neoclides 90 d'Ooge 97. 99
Neo-Pythagoreans 82. 91. 96. 97. 116 Opposite hyperbolas 248
Neugebauer 34. 38. 41. 42. Oppositio 282

.302 INDEX

Oracle 160, 161 Perseus 146
Ordinate 241. 242 Persia 83-85
Orestes 290 Persian wars 82, 106
Orgies 107 Persius 210
Orient 83, 84 Perspective 129, 136, 137, 287
Orphic prophets 93 Pesu 29
Ovid 276 Phaenomena (Eudoxus) 182
Phaedias 208
Palimpsest 213 Philebus 180
Pamphile 88 Phillipus 90. 91. 155
Pannekoek 85 Philolaus 107
Pappus 6, 47, 82, 110, 125, 139, 146. 159, Philon 139
231. 267. 286-290 Philoponius 131
Pappus, Collectio 1-46, 191. 196,200.232, Philosophy 106. 148, 149, 179, 193, 228,
259, 229,2.30
Pappus. commentary on Almagest 286 Pindar 201
Pappus. commentary on Book X 169 Piraeus 179
Pappus, theorem of 287. 289 Pisces 270
Papyrus, Akhmen 36,48 Pisthetaerus 1.30
Papyrus, Anastasi I 17 Plane geometry 140
Papyrus, Berlin 6619 28, 29 Planetarium 211
Papyrus, Kahun 17 Planets 128. 129. 180, 238
Papyrus, Michigan 62Q 280, 281 Plato 6.7, 92, 101. 102,
Papyrus. Moscow 33. ·34 106. 107. 109. 110, 125, 129. 138-142.
Papyrus. Rhind 16-32 144, 148, 149. 152, 161. 162, 163, 1M.
Papyrus. Vindobon 49. 56,194.202,231,232,
Parabola 1-46. 162. 190, 191,213,216-.219 279
241-259 Plato. Epinomis 9.83,112,139.140,155-
Parabola. quadrature of 216·-219 159
Parabole 123 Plato. Gorgias 116
Parabolic segment 213, 216-219 Plato, Philebus 180
Paraboloid 215. 223, 227 Plato. Republic 115. 126. 138, 144. 148,
ParaIldogram 121. 122 193
ParaIIdopiped numbers 98 Plato. Theaetetus 141. 165. 166
Paramekepipedon 45 Plato. Timaeus 112. 173
Pargiter 55 P1atonicus 160. 161. 162. 163. 164, 228,
Part 49, 114 232.233
Pascal 264, 290 Pleasure 179
Put 16.33 Plutarch 96.100, 117. 138. 161. 162, 163,
PdI equation 208 164. 209. 210. 212
Pdoponnesian war 82, 106 Poetic numbers 54. 55
Pentagon, star 100. 117 Polar 258
Pentagonal numbers 98, 99 Pole 235.258
Pentagramma 101 Polemon 229
Pentathlos 228 Polos 84
Perfect numbers 97, 98, 109 Polybius 47, 209
Pericles 82, 106, 128 Polycrates 102
Perigee 240 Polygons. regular 117. 135. 136, 146, 184,
Perimeter of circle 81. 192, 205, 263 204, 205. 206, 228, 269
Period of Saturn 129 Polyhedra. regular 90, 100, 138. 173-175,
Perpendicular 129 286,291


Polyhedra. semi-regular 228 92-102. 104. 106. 107. 117. 118. 125.
Porism 200. 287. 288 147.279
Position. given in 198 Pythagoras. theorem of 100.
Position. harmonic 250 117. 118. 124. 136. 143. 147. 197.207.
Positional notation 38-41. 53-61 237.273.286
Postulate of Archimedes 186 Pythagoras. the tradition of 116. 117. 118
Potentially commensurab