THE EXACT SCIENCES

IN ANTIQUITY
THE EXACT SCIENCES
IN ANTIQUITY

BY

O.NEUGEBAUER

Second Edition

DOVER PUBLICATIONS, INC.

NEW YORK
 Published in Canada by General Publishing Com-
pany, Ltd., 80 Lesmill Road, Don Mills, Toronto,
Ontario.
Published in the United Kingdom by Constable
and Company, Ltd.

This Dover edition, first published in 1969, is an

unabridged and slightly corrected republication of
the second edition, published in 1957 by Brown Uni-
versity Press.

Standard Book Number: 486-223J2-9

Library of Congress Catalog Card Number: 69-20421

Manufactured in the United States of America

Dover Publications, Inc.
180 Varick Street
New York, N. Y. 10014
\
\
\

i
i
I
I

I
I

To
RICHAR.D COURANT
in Friembhip emil Gratitude
 vii

FROM THE PREFACE TO THE FIRST EDITION

The first series of Cornell University's "Messenger Lectures on the

evolution of civilization" was given by James Henry Breasted,
eminent Egyptologist and founder of the Oriental Institute of the
University of Chicago. Few scholars have contributed so much
to our understanding of ancient civilizations and have attracted
the interest of scholars and laymen alike to the study of the ancient
Near East. I personally feel a great debt of gratitude towards
Breasted whose "History of Egypt" was my first stimulus towards
the study of ancient oriental civilizations, a field of research which
has occupied me ever since and about whose role in the history
of science I shall report in the following pages. That I was able
to follow this road from the early days of my graduate study in
Gottingen is due to the never failing encouragement and support
of R. Courant. But more than that lowe him the experience of
being introduced to modern mathematics and physics as a part
of intellectual endeavor, never isolated from each other nor from
any other field of our civilization. In dedicating these lectures to
him I only acknowledge publicly a debt which has profoundly
influenced my own development.
The following chapters follow closely the arrangement of six
lectures which I delivered at Cornell University in the fall of 1949.
I fully realize that this form of presentation forced me into many
statements which actually should be qualified by many conditions
and question marks. I also realize that the following pages will
give ample opportunities to quote statements and to utilize them
in a sense which I did not imply or did not foresee. And I have
no doubts that many a conclusion will have to be modified and
corrected. I am exceedingly sceptical of any attempt to reach
viii Preface

a "synthesis"-whatever this term may mean-and I am convinced

that specialization is the only basis of sound knowledge. Never-
theless I have enjoyed the possibility of being compelled for once
to abandon all learned apparatus and to pretend to know when
actually I am guessing. This does not imply that I have ignored
facts. Indeed, I have consistently tried to keep as close as possible
to the source material. Only in its selection, in its arrangement,
and in its coherent interpretation have I permitted myself much
greater freedom than is usual in technical publications. And in
order to counteract somewhat the impression of security which
easily emerges from general discussions I have often inserted
methodological remarks to remind the reader of the exceedingly
slim basis on which, of necessity, is built any discussion of
historical developments from which we are separated by many
centuries. The common belief that we gain "historical perspective t t
with increasing distance seems to me utterly to misrepresent the
actual situation. What we gain is merely confidence in generaliza-
tions which we would never dare make if we had access to the
real wealth of contemporary evidence.
The title liThe Exact Sciences in Antiquity" is not meant to
suggest an exhaustive discussion of this vast subject. What I
tried to present is a survey of the historical interrelationship
between mathematics and astronomy in ancient civilizations, not
a history of these disciplines in chronological arrangement.
Since the works of Sir Thomas L. Heath provide an excellent
guide for Greek mathematics, I see no need to summarize their
contents in a series of lectures. For Greek astronomy no similar
presentation exists. but its highly technical character makes it
impossible to discuss any details in the present book. Conse-
quently, the main emphasis is laid on mathematics and astronomy
in Babylonia and Egypt in their relationship to Hellenistic
science.
In the notes which follow the single chapters I have added
some technical details which seem to me relevant to further
study. I have also quoted several works which lead far away
from the direet path of the present approach because I feel that
this book will have best reached its goal if it convinces the reader
that he finds here only one of many ways of approach to a
subject whieh is much too rich to be exhausted in six chapters.
 Preface ix

Instead of attempting completeness I have tried to convey to the

reader some of the fascination which lies in active work on
historical problems. I wished to confront him with one of the
ever-changing pictures which one forms as a kind of guiding
principle for future research.

PREFACE TO THE SECOND EDITION

Tu devlem rapomable
pour toqJoun de ce qUB
tu as apprivoiBl
S.uNT-ExUPBRY

When preparing this second edition, I was helped again by my

friends and colleagues, particularly by A. Sachs, but I alone am
responsible for any statements which might be incorrect or might
become untenable in the light of further research.
I am very thankful to Brown University and in particular to
its librarian Mr. D. A. Jonah for having made possible the publi-
cations of this book. Mr. Torkil Olsen in Copenhagen was very
helpful in arranging for the printing which was completed with
traditional craftsmanship in Odense, Denmark.
In order to keep this book up-to-date, many additions referring
to recently obtained results have been made. Large sections on
Egyptian astronomy and on Babylonian planetary theory have
been rewritten. Two appendices are entirely new, one on Greek
Mathematics, the other on the Ptolemaic system and its Copernican
modification.
I hope I have avoided, in spite of these amplifications, con-
verting my lectures into a textbook.
o. N.
 xi

TABLE OF CONTENTS
Page
List of Plates. . • . • . • • . • . . . . . . • • . . . • . . . . • • • . . • • • • • • • • • • • • • • . • . . • • • • xiii
Introduction • . . • •. . . . . . . . . . . • • . . . . • . . . . . . . . . . • . . . . . . . . •. . • • •. • •. . 1
Chapter I. Numbers.....................•......................... 3
Bibliography to Chapter I .........•............................. 23
Notes and References to Chapter I ....•..•......•......••..•••.•.• 23
Chapter II. Babylonian Mathematics ....••.•......•....•.....•...... 29
Bibliography to Chapter II ..•.....•....•......•...••........•... 49
Notes and References to Chapter II •......•••••...••..•.....•..... 49
Chapter III. The Sources; their Deciphennent and Evaluation •.•..•..•.• 53
Noles to Chapter III. • • . • . • . . . . . • • . . . • . • . . . . . . . . . . . • • . . . . . . . . . . . 66
Chapter IV. Egyptian Mathematics and Astronomy •...............•..• 71
Bibli.aphy to Chapter IV.•...•.••....•.....•....••.......•..•• 91
Notes and References to Chapter IV •..•.•..•.••••.....•••••.••••• 92
Chapter V. Babylonian Astronomy ....•.•.•.•.....•.........••..••.• 97
BibHography to Chapter V .•.•••.......•••••.••••.•••.••••••.•.. 138
Notes and References to Chapter V ..••..•.••..•••••••••••••••.•.• 139
Chapter VI. Origin and Transmission of Hellenistic Science•••••.•••.••. 145
BibUos:raphy to Chapter VI•••••.•..•......•••..•.....•••••••.••. 177
Notes and References to Chapter VI ....••.•••••....•..••..••••••. 178
Appendix I. The Ptolemaic System •••••••••••••••••••••••••••••••••• 191
Bibliography to Appendix. I ......•.•....•....•............••.... 206
Notes and References to Appendix I •..•.......••............••.•• 206
Appendix. n. On Greek Mathematics .••...•••••••.••...•..•...••.•.• 208
Bibliography to Appendix II•..•••...••••••••.••••.•••••.••.••••• 225
Notes and References to Appendix II .•.•.•.••.•.•.•...••••••••••• 226
The Zodiacal and Planetary Sips .....•.....••.•.•....•.•••.•.•••••• 228
Chronological Table..•...........•....••.•••••..••••..•••••.•.•... 229
Index ••••••••.••••..•.•.•.••••••.•.•.•.•••••••••••••••.•..•.•..• 230
 xiii

LIST OF PLATES
(Plates 1-14 appear followiDI p. 240)

ChronolO8ical &heme (p. xvi).

1 "September", Book of Boon of the Duke of Berry.
2 Lunar Table, P. Lund,lnv. 35a.
3 Old Babylonian Mathematical Text, BM 85194 obv.
4a Multiplication Table for 10, VAT 7858 obv.
4b Economic Text from Para, VAT 12770 obv.
5 Table of Squares, P. Cairo Inv. 65445.
6 a Square and DiagoDal, YBC 7289.
6b Fragments from Warka, Excavation Photo.
7a Pyth8lorean Numbers, Plimpton 322.
7b Lunar Ephemeris, A 3412 rev.
8a Copy of YBC 4712, Neuaebauer MKT II pI. 59.
8b Copy of BM 34589, KusJer SSB I pI. XX.
9a Old Babylonian Mathematical Text, YBC 4712.
9b Ephemeris for Saturn, BM 34589 obv.
10 Prom the CelliDg in the Tomb of Senmut.
11 Prom the CeiIiDI in the Tomb of Ramses VII.
12 Geometrical Problems, P. Cornell,lnv. 69.
13 Demotic Planetary Tables, liverpooL
14 Prom Strassmaier's Notebook.
 xv

"And when he reaches early adolescence he must become

possessed with an ardent love for truth, like one inspired, neither
day nor night may he cease to urge and strain himself in order
to learn thoroughly all that has heen said by the most illustrious
of the A.ncients. And when he has learnt this, then for a prolonged
period he must test and prove it, observe what part is in agreement,
and what in disagreement with obvious facts; thus he will choose
this and tum away from that. To such a person my hope has
been that my treatise would prove of the very greatest assistance.
Still, such people may be expected to be quite few in number,
while, as for the others, this book will be as superftuous to them
as a tale told to an ass. I t

GALEN, On the natural faculties, Ill, 10.

[Translation by ARTHUR JOHN BROCK, M. D.
The Loeb Classical Library p. 279/281.]
 xvi

IneLta Iran Mesopota Syria Egypt Asia lIlinor Greece Italy Spah\
-3500

Writing
Writins
-JOOO

Elall\
Sumer ............
Pyrami. -2500
Alclcecl OIc1Ki~_

UrJl[ ...........
Middle Crete -2000
KiJ\9dolll
H-llIUI"IIpl
H kills Line.. B
V.dic
'Rtr,ocl Hittites -1500

New
ffin,cl.m
Greelc -1000
Auyru." Empir. AlplaaW
EWKaJllI
Buddha Medea
-.sao P....lan Emplre -.500

earth. .
'Ptoleml••
0 0

Roman ElIlplre

Gupta
Saanians By~.ntlum
SOO 500

Isla",
1000

~
1506
THE EXACT SCIENCES
IN ANTIQUITY
 1

INTRODUCTION
The investigation of the transmission of mathematics and as-
tronomy is one of the most powerful tools for the establishment
of relations between difTerent civilizations. Stylistic motives, religi-
ous or philosophical doctrines may be developed independently
or can travel great distances through a slow and very indirect
process of diffusion. Complicated astronomical methods, how-
ever, involving ·the use of accurate numerical constants, require
for their transmission the direct use of scientific treatises and will
often give us very accurate information about the time and
circumstances of contact. It will also give us the possibility of
exactly evaluating the contributions or modifications which must
be credited to the new user of a foreign method. In short the
inherent accuracy of the mathematical sciences will penetrate
to some extent into purely historical problems. But above and
beyond the usefulness of the history of the exact sciences for the
history of civilization in general, it is the interest in the role of
accurate knowledge in human thought that motivates the following
studies.
The center of "ancient science" lies in the "Hellenistic" period,
i. e., in the period following Alexander's conquest of the ancient
sites of oriental civilizations (Frontispiece). In this melting pot of
"Hellenism" a form of science was developed which later spread
over an area reaching from India to Western Europe and which
was dominant until the creation of modern science in the time
of Newton. On the other hand the Hellenistic civilization had its
roots in the oriental civilizations which nourished about equally
long before Hellenism as its direct influence was felt afterwards.
The origin and transmission of Hellenistic science is therefore
the central problem of our whole discussion.
2 Introduction

I restrict my subject to the exact sciences simply because I feel

totally incompetent to deal with subjects like medicine or the
natural sciences, though much important information could be
obtained for our problem from an investigation of these fields.
Medicine and astronomy, for example, are closely related in the
Greek medical schools; similarly, medieval medicine was deeply
affected by Hellenistic astrology. The sciences of drugs, plants,
stones, and even the animal kingdom show many points of
contact with astronomical or astrological doctrines; our use of
the name of Mercury for a substance and for a planet is a still
living witness of this. Nor did medieval or Renaissance artists
pride themselves on being ignorant of the sciences. The sculptures
of Gothic cathedrals and the paintings and miniatures of the Middle
Ages are full of astronomical or astrological references which
were significant to the contemporary man. Thus, it is a quite
artificial restriction which we impose upon the following dis-
cussions in limiting ourselves to exact mathematics and mathe-
I
matical astronomy.
And even within these narrow limits it was necessary to lay
undue weight on the part of mathematics as compared with
astronomy. The basic mathematical concepts are simple and
much more familiar to the modem reader than the corresponding
astronomical facts and their ancient presentation, which often
will be rather strange even to a professional modern astronomer.
All I could hope to do within the given framework was to remind
the reader on many occasions of the paramount role which
astronomy played in the history of science. I do not hesitate to
assert that I consider astronomy as the most important force in
the development of science since its origin sometime around
600 B.C. to the days of Laplace, Lagrange, and Gauss. And I
hasten to say that the history of the origin of astronomy is one of
the most fragmentary chapters in the history of science, however
great our gaps may be for other periods and other problems.
Consequently I am convinced that the history of mathematical
astronomy is one of the most promising fields of historical
research. I hope that this will become evident, at least to a
certain extent, to any reader of the following chapters.
 3

CHAPTER I

Numbers.
1. When in 1416 Jean de France, Duc de Berry, died, the work
on his "Book of the Hours" was suspended. The brothers Lim-
bourg, who were entrusted with the illuminations of this book,
left the court, never to complete what is now considered one of
the most magnificent of late medieval manuscripts which have
come down to us.
A "Book of Hours" is a prayer book which is based on the
religious calendar of saints and festivals throughout the year.
Consequently we find in the book of the Duke of Berry twelve
folios, representing each one of the months. As an example we
may consider the illustration for the month of September. As
the work of the season the vintage is shown in the foreground
(Plate 1). In the background we see the ChAteau de Saumur,
depicted with the greatest accuracy of architectural detail. For us,
however, it is the semicircular field on top of the picture, where
we find numbers and astronomical symbols, which will give us
some impression of the scientific background of this calendar.
Already a superficial discussion of these representations will
demonstrate close relations between the astronomy or the late
Middle Ages and antiquity. This is indeed only one specific
example of a much more general phenomenon. For the history
of mathematics and astronomy the traditional division of political
history into Antiquity and Middle Ages is of no significance. In
mathematical astronomy ancient methods prevailed until Newton
and his contemporaries opened a fundamentally new age by
the introduction of dynamics into the discussion of astronomical
phenomena. One can perfectly well understand the "Principia"
without much knowledge of earlier astronomy but one cannot
read a single chapter in Copernicus or Kepler without a thorough
 Chapter I

knowledge of Ptolemy's "Almagest". Up to Newton all astronomy

consists in modifications, however ingenious, of Hellenistic
astronomy. In mathematics the situation is not much different
though the line of demarcation between "ancient" and "modern"
is less sharply drawn. But also here the viewpoint could well be
defended that all "modern" mathematics begins with the crea-
tion of analysis. thus again with Newton and his contemporaries.
2. We shall not worry, however, about historical doctrines.
Merely as an illustration for the continuation of ancient traditions
we shall briefly analyze the calendars of the Book of the Hours
of the Duke of Berry. Both the outermost and the inner ring
contain numbers, the inner ring from 1 to 30, the outer ring from
17 to 30 and from 1 to 15. The appearance of these numerals
is not quite the one familiar to us today-ef., e. g., the 14 and
15 at the righ~ end of the outer circle-but everyone will easily
decipher their values and recognize that they are the familiar
"Hindu-Arabic" numerals which penetrated into Europe, begin-
ning in the 12th century, from the Islamic world. They super-
seded more and more the Roman numerals of which we find also
two representatives in our picture. The right half of the outer
rim shows the inscription initium libre gradus XV "beginning of
Libra 15 degrees" and in the inner rim we read primationes lune
memis septembris dies XXX "the primationes of the moon of the
month of September, 30 days". Implicitly we have here a represen-
tation of a third method of expression of numerals, namely, by
number words. September, and similarly October, November,
December, are denominations of the months as 7, 8, 9, 10
respectively-thus reflecting a period of the Roman calendar
when the end of the year fell two months later than in our present
calendar.
3. These three types of numerical expression can be found in
many examples all over the world. The writing of numbers by
simple words without the use of any symbols whatsoever is very
common indeed. A variant of it is the method found in Greek
inscriptions which use abbreviations like II for IIENTB "five"
or Ll for LlBKA "ten". One calls this the "acrophonic" principle,
where the first letter stands for the whole word.
The Roman system is perhaps the most widespread method,
historically speaking. The smallest numbers are simple repetitions
 Numbers 5

of 1. This holds even for our present number symbols where 2

and 3 originated from = and == by connecting lines in cursive
writing. The same system prevailed in Egypt, in Mesopotamia) or,
for the smallest units, in the Greek inscriptions just mentioned.
Roman V is probably half of the symbol X as D (= 500) is half
of ([) = 1000. This latter symbol was only later conveniently
interpreted as M for "millett thousand. Similar individual symbols
for 10, 100, 1000 are found in Egypt. Their repetition and com-
bination readily Yields the intermediate numbers. Arrangement
may playa role, as in IV = 5 - 1 in contrast to VI = 5 + 1. As
an explicit case of subtractive writing may be mentioned an Old-
Babylonian form for 19. In this period "one" would be r;
"ten," <; thus 21 = ~but ~ = 20 -1 = 19. Here the sign
r- LAL "subtracf' is written bctween 20 and 1. Later, 19 would
be written only 4f = 10 + + +3 3 3 from which a final cursivc
form ~ originated ill the Seleucid period.
Fundamentally different from all these methods is the "place
value notation tt of our present system, where neither 12 nor 21
represents 1 + 2 or 2 + 1 but 1 times ten plus 2, and 2 times
ten plus 1 respectively. Here the position of a number symbol
determines its value and consequently a limited number of
symbols suffices to express numbers, however large, without the
need for repetitions or creation of new higher symbols. The
invention of this place value notation is undoubtedly one of the
most fertile inventions of humanity. It can be properly cOinpared
with the invention of the alphabet as contrasted to the use of
thousands of picture-signs intended to convey a direct represell-
tation of the concept in question.
4.. Before returning to the history of number symbols we shall
draw some additional information from the calendar of the Book
of Hours. The wide middle zone shows the pictures of "Virgo"
and the scales of "Libra", headed by the inscriptions finis graduum
virginis "end of the degrees of Virgo" and the already quoted
"beginning of Libra 15 degrees." Virgo and Libra are signs of
the zodiac, i. e. sections of 30 degrees each in the yearly path
of the sun among the fixed stars as seen from the earth. Conse-
quently our picture indicates that the sun travels during Sep-
tember from the 17th degree of Virgo to the 15th degree of Libra,
6 Chapter I

or a total of 29 degrees, as can be counted directly by tallying

the spaces on the outer rim. Because September has 30 days the
. one d ay 29
sun covers In 30 = 58 d egrees or 58IDlnutes
60 · f
0 arc per

day. This corresponds very well to the facts. Because it takes the
sun slightly more than 365 days to travel the 360 degrees of the
whole zodiac, the average daily travel must be slightly less than
one degree per day. If we repeat our computation for all the
12 folios of our calendar we find, however, a faster movement
of 10 per day for November, December, and January. This is
counterbalanced by a slower movement of about 56 minutes in
the months from March to July. This again reflects facts correctly.
The sun moves fastest in Winter, slowest in Summer; and we
shall see that this phenomenon was accurately taken into con-
sideration both in Greek and in Babylonian astronomy of the
Hellenistic period. One calls this irregularity of the solar motion
its "anomaly". It is certainly not to be expected a priori to find
this concept carefully represented in a prayer book of the early
15th century.
6. An additional numerical notation occurs in the inner ring
of the calendar. Here we find associated with symbols of the
moon the following letters: b k 8 9 f d m a i etc. If we assign
to these letters numbers according to their position in the alphabet
we obtain:
2 10 18 7 6 4 12 1 9 17 6 14 3 11 19 8 16 5 13
These numbers are obviously connected by the following simple
law: always add 8 to the preceding number in order to get the
next number; in case the total exceeds 19, subtract 19. Thus
2 +8 = 10 10 +8 = 18 18 +8 = 26; 26 - 19 = 7.
The next number should be 7 +
8 = 15 = P followed by
15 + 8 = 23; 23 - 19 = 4 = d. The text, however, has f
followed by d. Henee we must correct f into p, and this correetion
is confirmed by the calendars for the other months where we
always find the arrangement 9 p d. The remaining part of the list
is correct. In the last plaee we have 5 +
8 = 13 to be followed by
21 - 19 = 2 whieh is the first number of our list. Thus the list
repeats itself after 19 steps.
 Numbers 7

The question as to the significance of the number 19 leads us

directly back to the 5th century B.C. when a cyclic scheme of
intercalations was actually introduced in the Babylonian calendar
and unsuccessfully proposed in Athens by Meton, who was,
however, honored by his contemporaries with a statue and by
modern scholars with the attachment of his name to the cycle.
The basis of this cycle can be explained very simply as follows.
The time between two consecutive conjunctions of sun and
moon is about 291 days. This interval is called one "lunation."
A lunar month is therefore either 29 or 30 days long. Consequently
12 lunar months amount to 354 days or about 11 days less than
one solar year. After three years a deficiency of about 33 days
has accumulated, making it necessary to add a 13th month to
one of the three lunar years in order to bring the beginning of
the lunar year roughly back to the beginning of the solar year.
More accurate recording of the beginnings of lunar months and
the beginnings of solar years shows that 19 solar years contain
235 lunar months, i. e., 12 ordinary lunar years of 12 months
each and 7 intercalary lunar years of 13 months each. This
19-year or Metonic cycle is quite accurate; only after 310 Julian
years do the cyclically computed mean new moons fall one day
earlier than they should. This simple cyclical computation not
only formed the basis of the calendar of the Seleucid empire
in antiquity but is similarly the foundation of the Jewish
and Christian religious calendar, especially so far as Easter is
concerned. The same cycle appears, though in a slight disguise,
in the luni-solar computations of two of the earliest astronomical
works of India, the Romaka- and the Paulisa-Siddhinta (about
fifth century A.D.), whose Western origin is apparent from their
names and confirmed by many details.
6. By means of this cycle the Middle Ages solved the problem
of establishing the dates of the new moons, at least for purposes
of the religious calendar, though the actual facts might differ by
several days. The "primationes lunae" or new moons in our
Book of the Hours are determined as follows: As the fIrst year,
"a," of the cycle a year is chosen when the new moon fell on
January 19 (cf. Fig. 1). From now on we operate with alternating
lunations of 30 or 29 days respectiYely, with occasional additions
of one day such that two 30-day months follow one another. In
8 Chapter I

this way one obtains February 18 for the next new moon, (30
days after January 19), then March 19 (29 days) after February
18, etc. Continuing this process1) we reach September 13 as a
new-moon date for "year aU and indeed the letter "a" is given
at this date below a little crescent in our calendar miniature for
September. Continuing with alternating
29 and 30 day lunations we reach Janu-
ary 9 of the second year, called &lb." For
September we find "b" marked at day 2;
for October one finds October 1 and Oc-
tober 31 for year "b," etc. This procedure
leads eventually to an arrangement of
It letters, representing the numbers from

~fI.
• a = 1 to t = 19, exactly in the form which
• we see in the special ease of September.
J.- "
;'

The scheme ends where it began, with

January 19, if we make the two last
lunations 29 days long. This final excep-
• tion to the rule of alternation was called
saltus lunae, the "jump of the moon".
In order to know which date is sup-
posed to be a new moon one need only
know which number the present year
has in the 19-year cycle. This number is
called the "golden number" because, as
Fig. 1. a scholar of the 13th century expressed
it, "this number excels all other lunar
ratios as gold excels all other met~ls." In the twelfth century
this very primitive method was considered by scholars in Western
Europe as a miracle of accuracy, though incomparably better
results had been reached by Babylonian and Greek methods since
the fourth century B.C. and though these methods were ably
handled by contemporary Islamic and Jewish astronomers.
7. Scientific progres, can perhaps be best measured by the
number of previously separated facts which become under-
standable under a new common viewpoint. By this standard the
recession from a lunar ephemeris of the third century B.C. to a
l)One pair of SO-day lunatioDS was inserted between March and July. In Fig. 1
SO-day intervals are indicated by solid lines. 29-day intervals by dotted lines.
 Numbers 9

lunar calendar 1700 years later is rather drastic, even disregarding

all astronomy and looking at simple arithmetic only. Instead of
one uniform numerical system, several computing devices are
used side by side. This even goes so far that for the handling of
the above-mentioned lunar letters special tricks were invented to
establish their sequence by means of the segments of the fingers
to which these letters were assigned. Thus one bas come back
to a wide-spread technique of "computing on the fingers".
These methods form a substratum of primitive mathematical
lore which has been found in the most different ancient civiliza-
tions, as well as among nations of the Near and Far East. Probably
its earliest occurrence is in Egypt in a passage of the "Book of
the Dead" which in turn is based on a spell of the "Pyramid
Texts." In the "Spell for Obtaining a Ferry-Boat" the deceased
king tries to convince the ferryman to let him cross a canal of
the nether world over to the Eastern side. To this the ferryman
objects with the words: "This august god (on the other side)
will say, 'Did you bring me a man who cannot number his
fingers?':' But the deceased king is a great "magician" and is
able to recite a rhyme which numbers his ten fingers and thus
satisfies the requirements of the ferryman. It seems obvious to
me that we are here reaching back into a level of civilization
where counting on the fingers was considered a difficult bit of
knowledge of magical significance, similar to being able to know
and to write the name of a god. This relation between numbers
(and number words) and magic remained alive throughout the
ages and is visible in Pythagorean and Platonic philosophy,
the Kabbala, and various other forms of religious mysticism.
8. We return once more to the diverse methods of writing
numbers. Four different types of writing can be illustrated on the
calendar of the Book of Hours: the place value notation still in
use today; the Roman numerals operating with individual sylnbols
for the different groups of units; complete number words; and
finally alphabetic numerals. From Greek inscriptions we have
added a fifth method, the "acrophonic" writing which consists,
however, only in an abbreviation of number words. We shall
now discuss other variants of Greek numerical notation.
The first is smclly alphabetical and is found on Athenian
coins of the second century B.C. On these coins, of the so-called
10 Chapter I

"New Style", the months of issue are denoted by the letters A

to M representing the numbers 1 to 12 for an ordinary year,
adding N = 13 for a leap year of the Athenian lunar calendar.
The same principle is followed in Ptolemaic coins from Egypt
rr
where the numbers AA BB
.... K."6",.,, ••" I" ndp I.~"...
.....,. Me...
·etc. occur, obviously indicating
25, 26, 27 etc. after the first 24
..... I~ letters from A to !J were exhaust-
I.' 0 k .. jJ" I
"
ed. It is clear how one in prin-

--, , .,,.
, .1.'• • II " o •
I o • IJ
" ciple could continue this system.
• I~ " 9. Much more important, how-
IJ • o • IJ
"
, ,
'I.' Jr 0
• I ,." ever, is another modification of
• IJ fl. the alphabetic numerationwhich
,, .." .-
,I.' ~
~ 0
0

• III. fl.
10
, ,.
I If 0
• f&C is extensively used in Greek ma-
,.. ,
II.'
, 14
, II
0
• IJ
~ thematics and astronomy and
• IJI
'",..
0
II.' • tIC 0
• I also in economic and literary
16
f

fl.'
c
I

...
I .
•• I~
0

0 ,
•
,.,~ documents, e.g. in Greek papyri.
c •• I., 0 •• IJ 14 Though this system of Greek
c --
,,
• IJ ,..
• " .,"" '"..
ct: 0
" 0 « II
numerals is often described in
»
• •" -,
'fl.'

... ... •
If
I., books on the history of mathe-
0
10 tIC 0
• IJ
.1.'
• • ,
., 00 • IJ lC matic and elsewhere, I shall
• tIC ~ 0 • IJ I.,11 sketch the way one Rlight be able
..., .-.
'1.' •
• ,,
• to decipher this system in any
..
,\ 0 « IJ IIJ
• "I.' • • 00 •
IIJ k •
.t sufficiently elaborate mathema-
" tical or astronomical text. This
Fig. 2. can at the same time serve as
an illustration of how one proceeds in similar cases with less
well known context.
I take as our example a table from Ptolemy's Almagest" U

(Fig. 2). The heading says "Table of straight lines in the circle,"
i. e. table of chords. The first column is described as "arcs."
In this column we find in every second line the familiar Greek
letters in the arrangement of the alphabet. We are obviously
dealing with numbers; thus we make the simplest assumption
tX = 1 P = 2 y = 3 d = 4 8 = 5. Thereafter one should
expect C = 6 but Cappears only one step later and we are forced
to read the intermediate symbol ~ as 6... Thereafter we obtain
again a regular sequence C= 7 fJ = 8 9 = 9 t = 10. Following
the alphabetic order one might expect ,,= 11 ,\ = 12 etc.
Actually, however, we find ux = 11 tP = 12 etc., in other words,
 Numbers 11

+ +
combined symbols 10 1, 10 2, etc. This is readily confirmed
by the continuation of our table (not reproduced here) where at
the proper place &8 = 19 is followed by x =20 ~ =21 etc. Con-
tinuing in this fashion one will meet once more a disturbance of
the standard alphabet when after n = 80 a strange sign 9 signifies
90. Then follows e = 100 a = 200 'I' = 300 until (» = 800,
followed again by a special sYmbol '1' (or h.. or ep) = 900.
Though the three symbols ( 9 and t:p are not members of
the classical Greek alphabet they are well known to the historian
as remnants of the earliest form of the Greek alphabet which still
shows these three letters in actual use. Consequently the alphabetic
numerals were invented when the Greek alphabet had not yet
eliminated these three sounds which it took over with the rest
of the alphabet from the Phoenicians. Considerations of this
type allow us to date the origin of the Greek alphabetic number
system to about the 8th century B.C. and to localize its invention
with great probability at the city of MUetus in Asia Minor.
Returning to our table we have still omitted every second line.
It is obvious, however, that one would guess that the sign l'
represents 1 because we then can read
1 1 Ii 2 2i 3 3i etc.
We can confirm this hypothesis immediately by means of the
second column. This column contains three subcolumns which
we already can transcribe by means of our previous decipherment
with the exception of the new symbol 0 in the very first place.
Calling this symbol x we read
x 31 25
1 2 50
1 34 15
2 5 40
2 37 4
3 8 28
ele. The structure of these three columns of numbers is obvious.
In the second and third column we observe alternatingly smaller
and larger numbers whereas the numbers in the first column
either remain unchanged or increase by one. This last obser-
vation compels us to assign to x the value "zero." Consequently it
12 Chapter I

is plausible to consider the numbers in the second and third

column as fractions and to assume that the numbers as a whole
increase from 0 to 1, 2, etc. Indeed, if we look at the last numbers
we find that they increase from 25 to 50, then fall down to 15 but
increase again by 25 to the next 40. We would have a constant
increase by 25 if we had the sequence 25 50 60 + 15 60 + 40
or if 60 units of the last place would amount to 1 of the preceding
place. This is easily tested in the preceding column. The numbers
31 2 34 5 37 show again almost constant increase if we take
a total of 60 as one higher unit:

31 31 + 31 = 60 + 2 62 + 32 = 60 + 34
60 + 34 + 31 = 120 + 5 120 + 5 + 32 = 120 + 37
etc.
The increase is either 31 or 32 and it is 32 when and only
when 60 units of the third column have accumulated. And when-
ever 60 units of the second column have accumulated, the number
in the first column increases by one. Thus we have a system of
numbers which behave exactly like degrees, minutes, and seconds,
or like hours, minutes, and seconds; the fractions are sixtieths
of the next higher unit. We call such fractions Usexagesimal
fractions" and write numbers of this type in the following form:
0,31,25
1,2,50
1,34,15
2,5,40
We can say that these numbers show a constant difference 0,31,25.
Later in our table the differences become smaller and smaller,
but this is exactly what one should expect. If the first column
indicates arcs increasing by i degree, as is indicated by the
numbers already known, then we must expect that the chords
do not grow simply proportionately with the arcs, though this
might hold for very small angles at the beginning of the table.
But what are the units used in our table? That the first column
indicates degree~ is obvious from the fact that the table ends with
180, i. e., with the straight angle. The chord to 180° must be the
diameter; the table gives for this entry the value ex 0 0 = 120,0,0.
Thus the radius is 60. This is confirmed by the chord 60 for 60°,
 Numbers 13

as is correct for the equilateral triangle where chord = radius.

We do not need to discuss in detail the third column, called
"sixtieths," of the table of chords. We know already that the
chord of lOis 1,2,50. Hence the chord for 0;10 (or for 1 minute
of arcl )) will be 0;1,2,50 as given in the third column. In general
the third column gives the coefficients of interpolation for single
minutes, as is easily confirmed from the section reproduced on
p. 10.
to. Our example of a Greek numerical table familiarized us
with several interesting features of the most important type of
Greek numerals. The borrowing of the Greek alphabet from the
Phoenicians explained the symbols for 6, 90, and 900. We found
a special sign for i, a phenomenon which could be amplified
from papyrus documents and other sources. We found a special
sign for zero, used exactly as our zero. And finally we have seen
the sexagesimal system in full use, both in the familiar division
of the circumference of the circle into 360 "degrees" of 60 minutes
or 3600 seconds each, and in the division of the radius into units
of consecutive sixtieths.
These features are not restricted to an isolated case like the
table of chords which we quoted. All Greek astronomical works,
containing hundreds of extensive numerical tables, are based on
exactly the same procedure. According to the prevailing doctrine
that Greek mathematics is essentially geometry, the historians of
mathematics have badly neglected the enormous amount of
numerical computations which are readily accessible in works
like Ptolemy's "Almagest" or Theon's "Handy Tables." But
long before these classics were written, Greek astronomical
papyri were covered with computations. While Ptolemy or Theon
are today preserved only in Byzantine manuscripts, we do have
papyri from the Ptolemaic periodS) onwards. In these papyri
we can find, e. g., the zero sign as it was actually written. An
example is a papyrus written in the second century A.D. (cf. Pl. 2
and the transcription on p. 163). Near the end of the last line
1) I apply here a notation which will be used throughout in the subsequent
pages. A semicolon separates integers from fractions. while all other sexagesimal
places are separated from one another by a comma. Thus 1.1 means 61 but 1;1 -
1 + /no Neither comma nor semicolon has any counterpart in the actual texts.
I) "Ptolemaic period·' refen to the dynasty of the Ptolemies who ruled over
Egypt during the last three centuries before our era. Ptolemy the astronomer. about
150 A.D•• has nothing to do with this dynasty.
14 Chapter I

preceding the empty line in the upper part of the papyrus, one
finds representing zero a sign which looks like ...... In other
astronomical papyri are found similar symbols varying from
forms like T or ~ to m. In the form is and related variants
this zero symbol is found until the latest periods in Arabic geo-
graphical and astronomical manuscripts where numbers were
written in the alphabetic notation. Only in Byzantine manu-
scripts do I know of the bare a-like shape which is usually
considered as the first letter of Greek aV()811 "nothing... The
papyri do not support this explanation (which is in itself very
implausible since omicron already represented a numerical value,
namely 70) but suggest an abitrarily invented symbol intended
to indicate an empty place. This would correspond exactly to the
Babylonian zero symbol which is also not a letter or a syllable but
a mere separation mark.
11. In order to make this remark fully understandable, I have
to explain briefly a main point in the chronology of "Babylonian"
mathematical and astronomical source material. The texts of
which I speak are clay tablets, generally about the size of a hand,
inscribed with signs which were pressed into the surface of the
once soft clay by means of a sharpened stylus. This script is
called "cuneiform," i. e. wedge-shaped, because the individual
impressions have a deeper "head" and a finer line at the end,
thus resembling a wedge. Cuneiform tablets with mathematical
contents are known to us mostly from the so-called "Old-Babyl-
onian" period, about 1600 B.C. (cf. PI. 3). No astronomical texts
of any scientific significance exist from this period. while the
mathematical texts show the highest level ever attained in Ba-
bylonia.
The second period from which we have a larger number of
texts is the latest period of Babylonian history, when Mesopotamia
had become a part of the empire of Alexander's successors,
the "Seleucids." This period, from about 300 B.C. to the beginning
of our era, has furnished us with a great number of astronomical
texts of a most remarkable mathematical character, fully com-
parable to the astronomy of the Almagest. Mathematical texts
from this period are scarce, but they suffice nevertheless to
demonstrate that the knowledge of Old-Babylonian mathematics
had not been lost during the intervening 1300 years for which
texts are lacking.
 Numbers 15

Thus it is essential to remember that we are dealing with

mathematical texts from two periods, "Old-Babylonian" from
about 1800 to 1600, and "Seleucid" from 300 to 0, whereas
astronomical texts belong only to the second period.
12. The development of the numerical notations in Mesopo-
tamia took as many centuries as the development of writing from
a crude picture script to a well defined system of complicated
signs. We shall for the moment deal only with the final product
as it appears in the mathematical texts of the Old-Babylonian
period. And we shall again use the most direct approach by
deciphering an actual text.
Plate 4,a shows a tablet whose size is about 3'1 by 2 inches
(and about ! of an inch thick). In the middle of the text is
visible a column of signs which obviously represent numbers
in ascending order. The tablet is not quite cleaned from incrusta-
tion of salt or dirt but it is clear that the signs look about as
follows:

Counting of the vertical wedges leads directly to the readings

1, 2, 3, etc. up to 9. Then follows <
which must be 10, and
consequently we can also read the remaining signs as 11, 12, 13.
Using this exceedingly plausible hypothesis, we should also be
able to read the right-hand column of signs. The first five look
as follows

Obviously we must read these signs as 10, 20, 30, 40, 50 if the
first sign represents 10 as we have established in our first list.
But what follows is

III * III

which we transcribe consistently as

1 1,10 1,20 III * III 2 2,10

each III indicating a broken line. These signs continue the previous
ones if we interpret the first "1" as 60 and then read 1,10 as
16 Chapter I

+ +
60 10 = 70 and 1,20 as 60 20 = 80. The broken lines should
contain 90, 100, and 110. The next sign "2" should be 120, in
excellent agreement with our interpretation of "I" as 60, while
the last sign 2,10 must be 120 + 10 = 130. Thus we have
obtained all multiples of 10 from 10 to 130, line by line, corres-
ponding to the numbers 1 to 13. In other words, our table is a
multiplication table for 10, which we now can transcribe as
follows:
1 10
2 20
3 30
4 40
5 50
6 1
7 1,10
8 1,20
9 1,30
10 1,40
11 1,50
12 2
13 2,10

The notation 1,10 = 70 1,20 = 80 2,10 = 130 etc. is

Usexagesimal" in the sense that 60 units of one kind are written
as 1 of the next higher order. This is exactly the same principle
we found in Ptolemy's table of chords. The only difference
consists in the fact that Old-Babylonian texts have not yet devel-
oped a special sign for "zero". This appears, however, in both
mathematical and astronomical cuneiform texts of the Seleucid
period, as we shall see in later examples. Thus we have reached
complete identity of the principle of numerical notation for
astronomical tables of the Hellenistic period, whether written
in cuneiform or in Greek alphabetic numerals. Only in one point
is the Greek notation less consistent than the Babylonian method.
In the latter all numbers were written strictly sexagesimally,
regardless of whether they were integers or fractions. In Greek
astronomy, however, only the fractions were written sexagesim-
ally, whereas for integer degrees or hours the ordinary alphabetic
notation remained in use also for numbers from 60 onwards.
 Numbers 17

Thus Ptolemy would write 130 17 20 where a cuneiform tablet

would have 2 10 17 20. In other words, the Greeks already
introduced the inconsistency which is still visible in modern
astronomy, where one also would write 130°17'20". The other
inconstistency of the modern astronomical notation, namely, to
continue beyond the seconds with decimal fractions, is a recent
invention. It is interesting to see that it took about 2000 years of
migration of astronomical knowledge from Mesopotamia via
Greeks, Hindus, and Arabs to arrive at a truly absurd numerical
system.
19. The example of our present system of numeration for
degrees, hours, measures and ordinary numbers should suffice
totally to discredit the popular idea that a number system was
"invented" at a certain moment. Yet innumerable "reasons"
have been advanced why the Babylonians used the basis 60 for
their number system. I shall not make any attempt to discuss
here the history of the sexagesimal system in any detail, but a few
points must be mentioned because they are of importance for the
historical approach to the development of number systems as
a whole.
First of all, there exists a common misconception as to the
gcmerality of the use of the sexagesimal system. The very same
tablet which contains hundreds of sexagesimal numbers, column
beside column, to compute the dates of the new moons for a
given year, might end with a "colophon" containing the name of
the owner of the tablet, the name of the scribe, and the date of
writing of the text, the year being expressed in the form 2 me
25 "2 hundred 25" where the main text would express the very
same date sexagesimally as 3,45. In other words, it is only in
strictly mathematical or astronomical contexts that the sexagesimal
system is consistently applied. In all other matters (dates, measures
of weight, areas, etc.), use was made of mixed systems which
have their exact parallel in the chaos of 60-division, 24-division,
12-division, 10-division, 2-division which characterizes the units
of our own civilization. The question of the origin of the sexagesi-
mal system is therefore inextricably related to the much more
complex problem of the history of many concurrent numerical
notations and their innumerable local and chronological varia-
tions.
18 Chapter I

But it is not enough to realize that the 60-division is only one

of several contemporary norms between higher and lower units.
The essential point lies in the use of the place value notation,
regardless of the value of the ratio between consecutive units.
No historical theory of the origin of the sexagesimal system is
acceptable if it does not account also for this extraordinary feature,
namely, the use of the same small number of symbols for different
values, depending on the arrangement. A variety of "bases" is
well known from number words and number writing all over
the world. The place value notation, however, is the most striking
feature of the Babylonian system.
A problem of this kind cannot be solved by speculation, but
only by a systematic analysis of the written documents. Fortun-
ately there is an abundance of source material available. The
early association of Assyriology with Biblical problems and the
Hellenistic and Roman concept of "Chaldaeans" as equivalent
to astrologers or magicians is today still reflected in the wide-
spread idea that the majority of Babylonian documents are con-
cerned with religion, magic or number mysticism. In fact, how-
ever, the overwhelming majority of cuneiform texts concern
economic items. Tens of thousands of such documents were
unearthed and, although only a small fraction has been made
available in modern publications, they suffice to obtain a fair
sampling of the use of numbers through all periods of Mesopota-
mian history. Especially for the earliest period of writing the
economic records are almost the only class of existing documents
and the number signs are among those signs which one can read
with certainty even for periods where the interpretation of the
other signs is very problematic.
In the earliest phase of writing the signs are still recognizable
as pictures which were scratched in the soft clay with the sharp-
ened edge of a stylus, probably made of reed. The number signs,
however, were impressed with the round end of the stylus. A
slanted position of the stylus produces a roughly ellipse-shaped
impression; a vertical position results in a circular sign. The
former represents ordinary units, the latter tens. Thus we can
read a 40 in the second section from the top in the second column
from the left in PI. 4 b. In the first column we have a 2 in the
7th section from the top. The subsequent sections are damaged
 Numbers 19

but one can still recognize the traces of 7, 8, and 9 in their

respective compartments preceding a lOin the next to the last
section at the left lower corner.
Beside these basic elements, many modifications of number
symbols were in use for different classes of objects, such as
capacity measures, weights, areas, etc. Among these a clear
decimal system has bcen recognized with signs for 1, 10, and
100. The numbers 1 and 10 we have already described. The
100 was written as a circular impression which looks like 10,
but is made much bigger. Thus 100 is simply "big 10". Another
system proceeds sexagesimally, at least partially. Distinct units
are 1 and 10 as before. A big 1 represents 60. Two big units
written in opposing directions are combined into one sign to
form 120. A 10-sign added in the middle gives 1200. A very big
10 sign stands for 3600. Variations of these systems, both decimal
and more or less sexagesimal, can be cstablished at different
localities. The main facts, however, al'C common to all of them,
namely, the existence of a dccimal substratum and the use of
bigger symbols to represent higher units. This latter fact is
obviously the root for the development of the place value nota-
tion.When the script slowly became simplified and standardizcd,
the distinction between bigger and smaller signs of thc same type
disappeared. Whereas originally one big unit, meaning 60, and
one 10 symbol were written to denote 60 + 10, later a simple
"1'" followed by a 10 was read 70, in contrast to a 10 followed by
1 meaning 11.
Combined with this, another process was taking placc. In
economic texts units of weight, measuring silver, were of primary
importance. These units seem to have been arranged from early
times in a ratio 60 to 1 for the main units •'mana"' (the Greek
IJ1vii "mina") and shekel. Though the details of this process cannot
be described accurately, it is not surprising to see this same ratio
applied to other units and then to numbers in general. In other
words, any sixtieth could have been called a shekel because of
the familiar meaning of this concept in all financial transactions.
Thus the "sexagesimar" order eventually became the main
numerical system and with it the place value writing derived
from the use of bigger and smaller signs. The decimal substratum,
howevcr, always remained visible for all numbers up to 60.
20 Chapter I

Similarly, other systems of units were never completely extin-

guished. Only the purely mathematical texts, which we find well
represented about 1500 years after the beginning of writing, have
fully utilized the great advantage of a consistent sexagesimal
place value notation. Again 1000 years later, this method became
the essential tool in the development of a mathematical astronomy,
whence it spread to the Greeks and then to the Hindus, who
contributed the final step, namely, the use of the place value
notation also for the smaller decimal units. It is this system that
we use today.
14-. The Babylonian place value notation shows in its earlier
development two disadvantages which are due to the lack of a
symbol for zero. The first difficulty consists in the possibility
of misreading a number 1 20 as 1,20 = 80 when actually
1,0,20 = 3620 was meant. Occasionally this ambiguity is over-
come by separating the two numbers very clearly if a whole
sexagesimal place is missing. But this method is by no means
strictly applied and we have many cases where numbers are
spaced widely apart without any significance. In the latest period,
however, when astronomical texts were computed, a special
symbol for "zero" was used. This symbol also occurs earlier as
a separation mark between sentences, and I therefore transcribe
it by a "period." Thus we find in Seleucid astronomical texts
many instances of numbers like 1,. ,20 or even 1,.,. ,20 which
apply exactly the same principle as, e. g., our 201 or 2001.
But even in the final phase of Babylonian writing we do not
find any examples of zero signs at the end of numbers. Though
there are many instances of cases like .,20 there is no safe ex-
ample of a writing like 20,. known to me. In other words, in all
periods the context alone decides the absolute value of a sexage-
simally written number. In Old-Babylonian mathematical texts we
find several cases where a final result was written by means of
individual symbols for the fractions, e. g., 1,30 might be called
"1 and i" which shows that we should transcribe 1;30 = 1 i and
not 1,30 = 90.
The ambiguity with respect to fractions and integers is of no
importance for the practice of computation. Exactly as we multiply
two numbers regardless of the position of the decimal point, one
can also operate with the Babylonian numbers and determine
 Numbers 21

the absolute value at the end if necessary. For the numerical

process itself it is indeed a great advantage that one does not
need to worry about special values for fractions and integers.
It is precisely this feature which gave the Babylonian system its
tremendous advantage over all other number systems in antiquity.
Though this will become more obvious in the subsequent discus-
sion of (and comparison between) Babylonian and Egyptian
mathematics, one example nlay be given now to illustrate this
point.
A multiplication by 12 would be performed by an Egyptian
scribe in two steps. First he would multiply the other factor by
10 (simply by replacing each individual symbol by the next
higher one) and then he would double the other factor. Finally
he would add the two results. Thus for the multiplication of 12
by 12 he would arrange his figures as follows:
1 12
I 10 120
I 2 24
total 144
giving him 144 as the result of the addition of the two items marked
by a stroke. Let us now assume that the other factor was a
fraction, say the "unit fraction" 1, or, as we should write in
imitation of the Egyptian notation, 5'. The scribe would again
proceed in two steps, namely, multiplication by 10 and by 2.
The first gives the result 2. The second, however, would need
a table of duplications of unit fractions where the double of '5
appears to be listed as '8 15 (indeed f = ! +
1'7;). Thus the com-
putation would be
1 5'
I 10 2
I 2 3 15
total 2 3 15
A contemporary Old-Babylonian scribe would solve the same
problems by using a multiplication table for 12 exactly of the
same type as we have described above, p. 16, for 10. In line 12
he would directly find the result 2,24. Of course, so far we have
only established the fact of a better organized procedure in
22 Chapter I

Mesopotamia but nothing inhinsically inherent in the Babylonian

notation. This is different for the second problem, however. The
Babylonian scribe would know (or take this information from a
table of reciprocals) that i corresponds to "12" (0;12 = if in
our notation when we use a zero symbol). Hence y leads again
to finding the value of 12 times 12 or again to 2,24 (we would
write 2;24). In other words the Babylonian process completely
avoids special rules for computing with fractions, whether unit
fractions or not, and requires only that one remember correctly
the place value of each contributing number, exactly as we
must do in placing the final decimal point. The historical con-
sequences of this simplification can scarcely be overestimated.
14 Q. The advantages of the Babylonian place value system over
the Egyptian additive computation with unit fractions are so
obvious that the sexagesimal system was adopted for all astron-
omical computations not only by the Greek astronomers but also
by their followers in India and by the Islamic and European
astronomers. Nevertheless the sexagesimal notation is rarely
applied with the strictness with which it appears in the cuneiform
texts of the Seleucid period in Mesopotamia. Ptolemy, for ex-
ample, uses the sexagesimal place value system exclusively for
fractions but not for integers. Thus he will write 365 as '&' E 8
(300,60, 5) but not as ~ 8(6,5). The same procedure was followed
by the Islamic astronomers and is the reason for our present
astronomical custom to write integers decimally and then use
sexagesimal minutes and seconds.
Extreme consistency in the use of the sexagesimal place value
system is found in the Latin version of the "Alfonsine Tables"
(about 1280). Here we find a date like 1477 Sept. 20 6;1,36h
expressed as 2,29,49,32;15,4,0 days. Indeed 1476 Julian years
(of 3651 days each) contain 24,36 ·6,5;15 days = 2,29,45,9 days.
To this are added the 4,23 days until Sept. 20 and the fraction
O;15,4d = 6;1,36h • This gives the above total of days, counted
from A.D. 1 Jan. O.
Also Copernicus often used consistently written sexagesimal
numbers, pal1icularly in his tables of mean motions. For example,
for the moon he gives the following mean motions in consecutive
Egyptian years (of 365 days each)
 Numbers 23

1 2,9;37,22,36°
2 4,19;14,45,12
3 0,28;52,7,49 etc.
where we (and Ptolemy) would write for the integers 129, 259,
and 28 respectively.
The perfection to which Islamic scholars developed numerical
methods has only recently become clear, especially through the
work of P. Luckey on aI-Kishi, the astronomer royal of Ulugh
Beg in Samarqand. AI-Kishi died in 1429; one of his last works
is a treatise on the circumference of the circle in which he de-
termines (correctly) 2n as 6;16,59,28,1,34,51,46,15,50. And since
he had invented, a few years earlier, the decimal analog of the
sexagesimal fractions, he also converts the above number into
decimal fractions: 6.2831853071795865.

BIBLIOGRAPHY TO CHAPTER I
The best existing book on numbers and number systems is Karl Menninger,
Zahlwort und ZifTer. Aus der Kullurgeschichte unserer Zahlsprache, unserer
Zahlschrift und des Rechenbretts. Breslau, Hirt, 1934. The sections about the
Babylonian number system are not always reliable. In general, however, this
work is far superior to the majority of books on the history of numbers and of
elementary computing.
Kurt Sethe, Von Zahlen und Zahlworten bei den alten Aegyptern und was
fur andere V6lker und Sprachen daraus zu lemen ist. Schriften d. wiss. Ges.
Strassburg 25 (1916). This work is fundamental for the understanding of the
role of fractions within number systems in general.
For calendar computation in general see F. K. Ginzel, Handbuch der
mathematischen und technischen Chronologie, 3 vols., Leipzig, Hinrichs, 1906-
1914. The lunar calendar of Western Europe during the Middle Ages is discussed
in vol. III.
An excellent introduction to Babylonian civilization is given in Edward
Chiera, They Wrote on Clay, Univ. of Chicago Press; several editions since
1988.

NOTES AND REFERENCES TO CHAPTER I

ad 1. The "Book of the Hours" of the Duke of Berry was originally published
by Paul Durrieu, Les Ires riches heures de Jean de France, Duc de Berry,
Paris 1904. The twelve calendar miniatures are reproduced in color in Verve
24 Chapter I

No. 7 (1940), unfortunately excluding the zodiacal figure (Umelothesia") which

followed the calcndar. It was discussed in detail by Harry Bober, The Zodiacal
Miniature of the Tris Riches Heures of the Duke of Berry -Its Sources and
Meaning, Journal of the Warburg and Courtauld Institutes 11 (1948) p. 1-34.
ad 2. D. E. Smith and L. C. Karpinski, The Hindu-Arabic Numerals,
Boston, 1911. Julius Ruska, Zur Iltesten arabisehen Algebra und Reehenkunst,
Sitzungsber. d. Heidelberger Akad. d. Wiss., philos.-hist. Kl. 1917, 2. G. F. Hill,
The Development of Arabic Numerals in Europe exhibited in 64 Tables. Oxford,
Clarendon Press, 1915.
It is a mistake to assume that the Islamic mathematicians and astronomers
consistently used the "Hindu-Arabic" numerals. By and large the Hindu-Arabic
numerals are restricted to mathematical context, whereas astronomical tables
use the alphabetic numerals. In Egypt the Greek or Coptic alphabetic numerals
remained in use for centuries after the Arabic conquest.
A nice detail about the transmission of the numerical notation of the Hindus
to the Islamic world is accidentally preserved in the autobiography of Ibn SInl,
the "Avicenna" of the Middle Ages. He was born in 980 near Bukhara, which
was then undcr the Iranian Dynasty of the Samanids. When he was about ten
years old, missionaries of an Islamic sect, called Ismaelites, came to Bukhara
from Egypt. Through the teaching of these missionaries Ibn Sinllearned about
the Hindu method of ~ompuling. Without this explicit bit of information nobody
would have dreamed that Indian influcnce reached southem Russia via Egyptl
(Cr. Ibn Sini's autobiography, translated in Arthur J. Arberry, Avincenna on
Theology, London 1951.)
ad 3. For the development of the Greek number system and its relation to
Phoenicia cf. Wilhelm Larfeld, Grieehische Epigraphik, 3rd ed., MUnchen,
Beck, 1914 [Handbuch der k1assisehen A1tertumswissenschaft vol. 1]. See esp.
p. 290 ft'. Furthermore M. N. Tod, The alphabetic numeral system in Attica.
The Annual of the British School at Athens, 45 (1950) p. 12&-139.
The acrophonic numerals are often called "Herodianic" because a gramma-
rian Herodianus (second century A. D.) discussed these numbers. The name
seems to have been introduced by Woisin in his thesis, De graecorum notis
numeralibus, Lipsia 1886.
Recent discussion, textual evidence and bibliography in Marcus Niebuhr Tod,
The Greek acrophonic numerals, The Annual of the British School at Athens
No. 37, Sessions 193&-37, p. 23&-258 (London 1940). For examples cf. B. D.
Meritt-H. T. Wade-Gery-M. F. McGregor, The Athenian Tribute Lists (Cam-
bridge, Harvard Univ. Press. 1939) vol. I passim; e. g. the photo p. 74 Fig. 98
and corresponding copy on PI. XXI.
ad 6. W. E. van Wijk, Le nombre d'or. :Etude de chronologie technique
suivie du texte de la Massa Compoti d'A1exandre de Villedieu. La Haye, Nijhoft',
1936. This work contains a valuable introduction to the medieval cyclic calendars
in Europe. Cf. also Nils Lithberg, Computus, Stockholm 1953 (Swedish) =
Nordiska Muscets Handlingar 29; very complete bibliography.
ad 7. Jean-Gabriel Lemoine, Les anciens procedes de calcul sur les doigts
en orient et en occident. Revue des etudes islamiques 6 (1932) p. 1-58 [with
extensive critical bibliography].
Egyptian numbering of fingers: Kurt Sethe, Ein a1tlgyplischer Fingerzlhl-
 Notes and References 25
reim. Zeitsehr. fUr Aegyptische Sprache 54 (1918) p. 16-39. Battiscombe Gunn.
UFinger-Numbering" in the Pyramid Texts. ibid. 57 (1922) p. 71 f.
For the relations between alphabetic numerals, number mysticism, astrology,
etc., see Franz Dornseiff, Das Alphabet in Mystik und Magie, Stoicheia 7,
2nd ed., Leipzig, Teubner, 1925.
ad 8. Examples for Athenian coins of the New Style: Bulletin de Corre-
spondance Hellenique 58 (1934) PI. I. These coins show the Athenian owl
standing on the Panathenaic amphora. The numerals for the months are often
inseribed on the amphora and are therefore called "amphora letters". For
Greek coins in general see. e. g.• Barkley V. Head. Historia Numorum, A
Manual of Greek Numismaties, Oxford, Clarendon 1911.
For the Ptolemaic coins with double letter numerals cf. Reginald Stuart
Poole, Catalogue of Greek Coins. The Ptolemies, Kings of Egypt, p. 44 and
Pl. VIII, 5. These numbers seem to represent the years of an era in honor of
Queen Arsinoti U (270 B.C.): cf. Head, Hist. Num. p. 850.
ad 9. The Greek symbols for 6, 90, and 900 are usually called stigma,
qoppa, and sampi respectively. The fll'St is originally F (or similar) and therefore
also called cCdigamma", that is double-gamma. Later, it assumed forms which
were similar to the ligature of c and T in Byzantine manuscripts and it was
therefore called "stigma" (since the 7th or 8th cent. A.D.). Its original name
is Waw. Qoppa is the Q of the Phoenician alphabet. The sampi is originally
written with only one middle stroke (cr. the form in PI. 5). The name usampi"
has been in use since the 17th century A.D.; the Phoenician original is an
S-sound called ~ade.
The division of the circumference of the circle into 360 parts originated in
Babylonian astronomy of the last centuries B.C. The sexagesimal number
system as such is many centuries older and has nothing to do with astronomical
concepts.
ad 10. For an example of an inscription with large number symbols in the
alphabetic notation see Inscriptiones Graecae vol. 12,1 (Insularum Maris
Aegaei) Berlin 1895, No. 913. This inscription, found in Keskinto (Rhodes) and
dating from the second century B.C., lists the basic numbers of a theory of
planetary motion; the author is unknown. For a discussion cr. P. Tannery,
Memoires scientifiques vol. 2 p. 487 fl.
In the ordinary alphabetic notation the numbers 1000, 2000 etc. are written
by means of (x, P. etc. which precede the symbols of lower order. Often accents
are added in order to avoid confusion with I, 2. etc. Several cases, both from
inscriptions and papyri, are known, where the symbol for 900, the usampi",
with (x, p, ... as superscript was used for 1000,2000. etc. (cf. Larfeld. quoted
above p. 24 in the note to Section 3, p. 294).
In papyri of the early Ptolemaic period ODe finds, in addition to sampi,
also fI'U = 500 ( +) 400 for 900, and besides Q) also fJT = 500 ( +) 300 for 800.
Cf. Mahaffy. Flinders Petrie Papyri vol. 3 p. 98 etc. An example from a school-
book of the third century B.C. is shown on PI. 5 from P. Cairo. Inv.65445
(published by O. Gueraud et P. J ouguet, Publications de la Societe Royale
EuPtienne de Papyrologie, Textes et Documents. Vol. 2, Cairo 1938). The
column on the left and the middle column constitute a table of squares of which
the fonowing part is clearly readable:
26 Chapter I

6 6 36 100 100 1 ·10000

7 7 49 200 200 4 ·10000
8 8 64 300 300 9 ·10000
9 9 81 400 400 16 ·10000
10 10 100 500 500 25 ·10000
20 20 400 600 600 36 ·10000
30 30 900 700 700 49 ·10000
40 40 1600 800 800 64 ·10000

Note in PI. 5 the signs for 6 and 900. The sign for 1000 (in 1600) is an (¥
with an attached loop. The multiples of 10000 are written as a II (first letter of
the Greek word for 10000) with the factor written over it. The last column gives
a list of the fractions of the drachma; preserved are the symbols for the following
unit fractions: 8, 12, 24,3, 6,8, 48. Note that the unit fractions are written with
the ordinary number signs plus an a£cent. The only exception is p' which does
not mean} but f denoted here by 3. Its corresP.9nding drachma sYmbol is a
combination of the symbols for 2 and 6. Indeed, 3" = 2 6. +
Ordinarily the arrangement of the alphabetic numerals is stricOy from higher
to lower numbers. In datings, however, one finds also the inverted order: ct. for
examples from Mesopotamia Yale Classical Studies 3 p. 30 fT. (clay bullae
from Uruk); Excavations in Dura-Europus, Preliminary Report IX, 1 p. 169 fT.;
Klio 9 p. 353. For Macedonian inscriptions (between 131 B.C. and 322 A.D.)
cf., e. g., Tod. The Maeedonian Era; The Annual of the British School at
Athens, No. 23 (1918-1919) p. 206-217 and No. 24 (1919-1921) p. 54-67.
That the Arabic form for the zero sYmbol (a litOe circle with a bar over it
and related forms) is simply taken from Greek astronomical manuscripts was
recognized by F. Woepcke in 1863 (Journal Asiatique, Sere 6 voL 1 p. 466 ft.).
A table showing dift'erent forms in Arabic manuscripts as well as in Greek
papyri is given by Rida A. K. Irani, Arabic numeral forms, Centaurus 4 (1955)
p. 1-12. In a Byzantine manuscript, written about 1300 A.D. a sign like II is
used for zero beside 0 (Vat. Graec. 1058 fol. 261 ft.), apparently under Islamic
influence.
ad 13. The most comprehensive collection of the evidence on early number
signs is found in the fIrSt edition of Anton Deimel, Sumerische Grammatik der
arehaistisehen Texte, Roma, Pontificium Institutum Biblicum, 1924 (Chapter IV).
More recent evidence, especially concerning the decimal system, is given in
A. Falkenstein, Archaisehe Texte aus Uruk, Leipzig 1936, (sign list at the end).
For the picture of a stylus see, e. g., S. Langdon, Excavations at Kish, vol. 1,
Paris 1924, PI. XXIX and Falkenstein, 1. c., p. 6.
The texts from Uruk also revealed the existence of a system of fractions
strictly proceeding on the principle of repeated halving. A very important
feature of cuneiform numerical notation is the existence of special signs for },
-1-, I, and t which are in very common use also in later periods, even occasionally
in mathematical texts. These "natural fractions" undoubtedly play an important
role in the arrangement of metrological units. Obviously one will group higher
units in such a form that they admit directly the forming of these most common
parts. This leads naturally to a grouping in 12 or 30 or 60. All these ratios do
occur in one or another of the parallel systems ofunits in Mesopotamian metrology.
 Notes and References 27

A strictly decimal notation occurs occasionally in mathematical texts. The

following is an example from an Old-Babylonian text (published Neugebauer-
Sachs, Math. Cuneiform Texts, p. 18). The number 1,12 which occurs in the
text is transcribed in the heading as "4 thousand 3 hundred and 20" which is
indeed the equivalent of 1,12,0. This example shows at the same time the lack
of an absolute determination of the place value in this period of number writing.
We may interpret 1,12 as 1,12,0 = 4320 or as 1,12 = 72 or as 1;12 = I} etc.
Only the context permits the determination of the absolute value of a number
written sexagesimally.
The lack of a notation which determines the absolute value of a number
made it possible to misinterpret simple tables of multiplication or reciprocals.
When Hilprecht, in 1906, published a volume of "mathematical, metrological
and chronological tablets from the Temple Library of Nippur" he was convinced
that these texts showed a relation to Plato's number mysticism. In book VIII of
the "Republic" Plato gives some cabballstic rules as to how guardians of his
dictatorially ruled community should arrange for proper marriages. By some

found on the tablets. Thus 1,10 (i. e. 70 or 1.

wild artifices, Plato's cabbala was brought into relationship with the numbers
etc.) was interpreted to mean
195,955,200,000,000 and in this fashion whole tablets were transcribed and
"explained".
As to the origin of the sexagesimal place value notation, it may be noted
that it is quite common that fractions of monetary units came to mean fractions
in general. As an example can be quoted the Roman as, which is It, of the
uncia (ounce). In the measurement of time, however, as is if of one hour
(Jahreshefte d. oesterreichischen archaeol. lost. in 'Vien 37, 1948, p. 111).
Mixed writings are also quite common. An example from an astronomical
procedure text (ACT No. 811 a, obv. 27) is 1 me 1,30 me for "190 days". Here
1 me means "1 hundred" (me being an abbreviated writing of the Babylonian
word for 100), while 1,30 is the sexagesimal writing for 90, and the final me
means "day", probably an abbreviation.
ad 14. No definite answer can be given to the question when the zero sign
was introduced in Babylonian mathematics. \Ve feel sure that it did not exist.
say, before 1500: and we find it in full use from 300 B.C. on. A table of squares,
found at Kish. tentatively dated by the excavator, S. Langdon, to the period
of Darius (500 B.C.), contains four cases of a "zero" written exactly like 30.
It is omitted in one case. Cf. Neugebauer, MKT I p. 73 and II pI. 34. For the
possibility of an earlier date (about 700) see MeT p. 34 note 95.
One might expect that the Babylonian notation should often lead to errors,
e. g., by mistaking 10,2 for 12 and vice versa. Numbers of this type are, however.
ordinarily written with very careful spacing such that <rr is hardly ever to
be taken as -«Tr. The descriptions of the Babylonian number systems in the
current textbooks arc generally quite misleading on this very point. Nevertheless,
there do exist cases where the proper combination of tens and units becomes
very doubtful. We even have examples of large numbers, written in two or more
lines, where, e. g., the 50 of a 56 was written at the end of one line and the 6
at the beginning of the next. For such "split writings" cf. Neugebauer-Sachs,
Math. Cun. Texts p. 13 note 69 and Neugebauer, ACT vol. 2, index.
28 Chapter I

In practical computations, the Babylonian scribes oceasionally committed the

same type of mistakes which arise when we are careless with the decimal point.
As an example may be quoted an astronomical text which concerns the risings
and settings ofMereury dUl'iDg the years 146 to 122 B.C. (BM 34585, Neugebauer
ACT No. 302 obv. IV. 30). The scribe had trouble with interpolations. The
table at his disposal contained the entries
15 42
45 36
The problem consists in interpolating the value of the right-hand column for
the value 31;20 of the left-hand argument. Obviously the answer should be
36 + 13;40 • 0;12 - 36 + 2;44 - 38j44. The utronomical problem required
the addition of this result to a~other number 1;20. Henee the ftnaI result should
be 40;4. In the text. however. we find 37;22,44. Obviously the scribe incorreedy
determined the place value of 13;40 • 0;12 and wrote Oj2.44 instead of 2j44.
This gave him u the result of the interpolation 36;2,44 and therefore as the
final answer Ij20 + 36;2.44 - 37;22.44.
It must be said. however, that the number of errors in the texts is compara-
tively very small. I have had the experience that I committed many more errors
in eheekiDg the ancient computations than there were in the original documents.
Often errors in a text are very helpful because they constitute one of the main
tools for establishing the details of a numerical procedure followed by the
ancient computer.
ad 14. Paul Luckey, Der Lehrbrief iiber den KreisumfaDg (ar-risila aI-
muhltlya) von c}amlld b. MU·Od ai-Kill. Abh. d. Deutsehen Akad. d. Wiss.
zu Berlin. KI. f. Math. u. aJJcem. Naturwiss. Jahrg. 1950 Nr. 6. Akad.
... Verl.•
Berlin 1953. Furthermore: Paul Luckey, Die Reehenkunst bei Gamlld. b.
Mu·Od ai-Kill mit RfickbUcken auf die Jiltere Geschiehte des Rechnens. Abh.
f. d. Kunde d. Morgenlandes 31,1 (1951).
 29

CHAPTER II

Babylonian Mathematics.
15. The following chapter does not attempt to give a history of
Babylonian mathematics or even a complete summary of its
contents. All that it is possible to do here is to mention certain
features which might be considered characteristic of our present
knowledge.
I have remarked previously that the texts on which our study
is based belong to two sharply limited and widely separated
periods. The great majority of mathematical texts are "Old-
Babylonian"; that is to say, they are contemporary with the
Hammurapi dynasty, thus roughly belonging to the period from
1800 to 1600 B.C. The second, and much smaller, group is
"Seleucid", i. e. datable to the last three centuries B.C. These
dates are arrived at on quite reliable palaeographic and linguistic
grounds. The more than one thousand intervening years influenced
the forms of signs and the language to such a degree that one is
safe in assigning a text to either one of the two periods.
So far as the contents are concerned, little change can be
observed from one group to the other. The only essential progress
which was made consists in the use of the "zero" sign in the
Seleucid texts (cr. p. 20). It is further noticeable that numerical
tables, expecially tables of reciprocals, were computed to a much
larger ex~ent than known from the earlier period, though no new
principle is involved which would not have been fully available
to the Old-Babylonian scribes. It seems plausible that the
expansion of numerical procedures is related to the development
of a mathematical astronomy in this latest phase of Mesopotamian
science.
For the Old-Babylonian texts no prehistory can be given.
We know absolutely nothing about an earlier, presumably
30 Chapter II

Sumerian, development. All that will be described in the sub-

sequent sections is fully developed in the earliest texts known.
It is customary to postulate a long development which is sup-
posedly necessary to reach a high level of mathematical insight.
I do not know on what experience this judgment is based.
All historically well known periods of great mathematical dis-
coveries have reached their climax after one or two centuries of
rapid progress following upon, and followed by, many centuries
of relative stagnation. It seems to me equally possible that Babyl-
onian mathematics was brought to its high level in similarly
rapid growth, based, of course, on the preceding development
of the sexagesimal place value system whose rudimentary forms
are already attested in countless economic texts from the earliest
phases of written documents.
16. The mathematical texts can be classified into two major
groups: "table texts" and "problem texts". A typical representative
of the first class is the multiplication table discussed above p. 16.
The second class comprises a great variety of texts which arc all
more or less directly concerned with the formulation or solution
of algebraic or geometrical problems. At present the number
of problem texts known to us amounts to about one hundred
tablets, as compared ",ith more than twice as many table texts.
The total amount of Babylonian tablets which have reached
museums might be estimated to be at least 500,000 tablets and
this is certainly only a small fraction of the texts which are still
buried in the ruins of Mesopotamian cities. Our task can there-
fore proPerly be compared with restoring the history of mathe-
matics from a few torn pages which have accidentally survived
the destruction of a great library.
t 7. The table texts allow us to reconstruct a small, however
insignificant, bit of historical information. The archives from
the city of Nippur, now dispersed over at least three museums,
Philadelphia, Jena, and Istanbul, have given us a large percentage
of table texts, many of which are clearly "school texts", i. e.,
exercises written by apprentice scribes. This is evident, e. g.,
from the repetition in a different hand of the same multiplication
table on obverse and reverse of the same tablet. Often we also
find vocabularies written on one side of a tablet which shows
mathematical tables on the other side. These vocabularies are
 Babylonian Mathematics 31

the backbone of the scribal instruction, necessary for the mastery

of the intricacies of cuneiform writing in Akkadian as well as in
Sumerian. Finally, many of our mathematical tables are combined
with tables of weights and measures which were needed in daily
economic life. There can be little doubt that the tables for multi-
plication and division were developed simultaneously with the
economic texts. Thus we find explicitly confirmed what could
have been concluded indirectly from our general knowledge of
early Mesopotamian civilization.
18. Though a single multiplication table is rather trivial in
content, the study of a larger number of these texts soon revealed
unexpected facts. Obviously a complete system of sexagesimal
multiplication tables would consist of 58 tables, each containing
all products from 1 to 59 with each of the numbers from 2 to
59. Thanks to the place value notation such a system of tables
would suffice to carry out all possible multiplications exactly
as it suffices to know our multiplication table for all decimal
products. At first this expectation seemed nicely confirmed except
for the unimportant modification that each single tablet gave
all products from 1 to 20 and then only the products for 30,
40, and 50. This is obviously nothing more than a space saving
device because all 59 products can be obtained from such a
tablet by at most one addition of two of its numbers. But a more
disturbing fact soon became evident. On the one hand the list
of preserved tables showed not only grave gaps but, more discon-
certingly, there turned up tables which seemed to extend the
expected scheme to an unreasonable size. Multiplication tables
for 1,20 1,30 1,40 3,20 3,45 etc. seemed to compel us to assume
the existence not of 59 single tables but of 3600 tables. The
absurdity of this hypothesis became evident when tables for the
multiples of 44,26,40 repeatedly appeared; obviously nobody
would operate a library of 60S = 216000 tablets as an aid for
multiplication. And it was against all laws of probability that we
should have several copies of multiplication tables for 44,26,40
but none for 11, 13, 14, 17, 19 etc.
The solution of this puzzle came precisely from the number
44,26,40 which also appears in another type of tables, namely,
tables of reciprocals. Ignoring variations in small details, these
tables of reciprocals are lists of numbers as follows
32 Chapter II

2 30 16 3,45 45 1,20
3 20 18 3,20 48 1,15
4 15 20 3 50 1,12
5 12 24 2,30 54 1,6,40
6 10 25 2,24 1 1
8 7,30 27 2,13,20 1,4 56,15
9 6,40 30 2 1,12 50
10 6 32 1,52,30 1,15 48
12 5 36 1,40 1,20 45
15 4 40 1,30 1,21 44,26,40
The last pair contains the number 44,26,40 and also all the
other two-place numbers mentioned above occur as numbers
of the second column. On the other hand, with one single excep-
tion to be mentioned presently, the gaps in our expected list of
multiplication tables correspond exactly to the missing numbers
in our above table of reciprocals. Thus our stock of multiplica-
tion tables is not a collection of tables for all products a . b, for
a and b from 1 to 59, but tables for the products a • b where b
is a number from the right-hand side of our last list. The character
of these numbers b is conspicuous enough; they are the reciprocals
of the numbers b of the left column, 'written as sexagesimal frac-
tions:
i = 0;30
1 = 0;20
1 = 0;15
etc.
1.~1 = 0;0,44,26,40.

\Ve can express the same fact more simply and historically more
correctly in the following form. The above "table of reciprocals"
is a list of numbers, band h, such that the products b • bare
1 or any other power of 60. It is indeed irrelevant whether we
write
2 • 30 = 1,0
or
2 • 0;30 = 1
or
0;2 • 30 = 1
or
0;2 . 0;80 = 0;1 etc.
 Babylonian Mathematics 33

Experience with the mathematical problem texts demonstrates

in innumerable examples that the Babylonian mathematicians
made full use of this flexibility of their system.
Thus we have seen that the tables of multiplication combined
with the tables of reciprocals form a complete system, designed
to compute all products a • b or, as we now can write, all sex-
a
agesimal divisions b within the range of the above-given table
of reciprocals. This table is not only limited but it shows gaps.
There is no reciprocal for 7, for 11, for 13 or 14, etc. The reason
is obvious. If we divide 7 into 1 we obtain the recurrent sexage-
simal fraction 8,34,17,8,34,17, ... ; similarly for n the group
5,27,16,21,49 appears in infinite repetition. We have tables which
laconically remark "7 does not divide", "11 does not divide",
etc. This holds true for all numbers which contain prime numbers
not contained in 60, i. e. prime numbers different from 2, 3, and
5. We shall call these numbers "irregular" numbers in contrast
to the remaining "regular" numbers whose reciprocals can be
expressed by a sexagesimal fraction of a finite number of places.
We have mentioned one exception to our rule that all multi-
plication tables must concern numbers b or, as we shall call them
now, regular numbers. This is the case of the first irregular
number, namely 7, for which several multiplication tables are
preserved. The purpose of this addition is clearly the completion
of all tables a • b at least for the first decade, in which 7 would
be the only gap because all the remaining numbers from 1 to 10
are regular. Thus we see that our original assumption was correct
for the modest range from 1 to 10. Instead, however, of expanding
this table up to 60, one chooses a much more useful sequence of
numbers, namely, those which are needed not only for multipli-
cation but also for division. The mere multiplications could
always be completed by one simple addition from two different
tables. This system of tables alone, as it existed in 1800 B.C.,
would put the Babylonians ahead of all numerical computers
in antiquity. Between 350 and 400 A.D., Theon Alexandrinus
wrote pages of explanations in his commentaries to Ptolemy's
sexagesimal computations in the Almagest. A scribe of the
administration of an estate of a Babylonian temple 2000 years
34 Chapter II

before Theon would have rightly wondered about so many words

for such a simple technique.
The limitations of the "standard" table of reciprocals which
we reproduced above (p. 32) did not mean that one could not
transgress them at will. We have texts from the same period
teaching how to proceed in eases not contained in the standard
table. We also have tables of reciprocals for a complete sequence
of consecutive numbers, regular and irregular alike. The reci-
procals of the irregular numbers appear abbreviated to three
or four places only. But the real expansion came in the Seleucid
period with tables of reciprocals of regular numbers up to 7
places for b and resulting reciprocals up to 17 places for b.
A table of this extent, containing the regular numbers up to
about 17 . lOll, can be readily used also for determining approxi-
mately the reciprocals of irregular numbers by interpolation.
Indeed, in working with astronomical texts I have often used this
table exactly for this purpose and I do not doubt that I was only
repeating a process familiar to the Seleucid astronomers.
19. Returning to the Old-Babylonian period we find many
more witnesses of the numerical skill of the scribes of this period.
We find tables of squares and square roots, of cubes and cube
roots, of the sums of squares and cubes needed for the numerical
solution of special types of cubic equations, of exponential
functions, which were used for the computation of compound
interest, etc.
Very recently A. Sachs found a tablet which he recognized as
having to do with the problem of evaluating the approximation of
reciprocals of irregular numbers by a finite expression in sexages-
imal fractions. The text deals with the reciprocals of 7, II, 13,
14, and 17, in the last two cases in the form that b • b = 10
instead of b • b = 1 as usual. We here mention only the two
first lines, which seem to state that

8,34,16,59 < ,.
but
8,34,18 > 7.
Indeed, the correct expansion of" would be 8,34,17 periodically
repeated. It is needless to underline the importance of a problem
 Babylonian Mathematics 35
which is the first step toward a mathematical analysis of infinite
arithmetical processes and of the concept of "number" in general.
And it is equally needless to say that the new fragment raises
many more questions than it solves. But it leaves no doubt that
we must recognize an interest in problems of approximations for
as early a period as Old-Babylonian times.
This is confirmed by a small tablet, now in the Yale Babylonian
Collection (cf. Pl. 6a). On it is drawn a square with its two diag-
onals. The side shows the number 30, the diagonal the numbers
1,24,51,10 and 42,25,35. The meaning of these numbers becomes
clear if we multiply 1,24,51,10 by 30, an operation which can
be easily performed by dividing 1,24,51,10 by 2 because 2 and
30 are reciprocals of one another. The result is 42,25,35. Thus we
have obtained from a = 30 the diagonal d = 42;25,35 by using
VI = 1;24,51,10.
The accuracy of this approximation can be checked by squaring
1;24,51,10. One finds
1;59,59,59,38,1,40
corresponding to an error of less than 22/60'. Expressed as a
decimal fraction we have here the approximation 1.414213 ..
instead of 1.414214 ... This is indeed a remarkably good approx-
imation. It was still used by Ptolemy in computing his table of
chords almost two thousand years later.
Another Old-Babylonian approximation of VI
is known to be
1;25. It is also contained in the approximation of V2
which we
find in the Hindu Sulva-Siitras whose present form might be
dated to the 3rd or 4th century B.C. There we find

whose s~xagesimal equivalent is

1;25 - 0;0,8,49,22, ... = 1;24,51,10,37, ....
The possibility seems to me not excluded that both the main term
and the subtractive correction are ultimately based on the two
Babylonian approximations.
36 Chapter II

JO. The above example of the determination of the diagonal of

the square from its side is sufficient proof that the "Pythagorean"
theorem was known more than a thousand years before Pythagoras.
This is confirmed by many other examples of the use of this
theorem in problem texts of the same age. as well as from the
Seleucid period. In other words it was known during the whole
duration of Babylonian mathematics that the sum of the squares
of the lengths of the sides of a right triangle equals the square of
the length of the hypotenuse. This geometrical fact having once
been discovered. it is quite natural to assume that all triples of
numbers 1. b. and d which satisfy the relation [I + bl = dI can
be used as sides of a right triangle. It is furthermore a normal
step to ask the question: When do numbers 1. b. d satisfy the
above relation'/ Consequently it is not too surprising that we find
the Babylonian mathematicians investigating the number-theoret-
ical problem of producing "Pythagorean numbers". It has often
been suggested that the Pythagorean theorem originated from
the discovery that 3. 4. and 5 satisfy the Pythagorean relation.
I see no motive which would lead to the idea of forming triangles
with these sides and to investigate whether they are right triangles
or not. It is only on the basis of our education in the Greek
approach to mathematics that we immediately think of the
possibility of a geometric representation of arithmetical or alge-
braic relations.
To say that the discovery of the geometrical theorem led natur-
ally to the corresponding arithmetical problem is very different
from expecting that the latter problem was actually solved. It is
therefore of great historical interest that we actually have a text
whieh clearly shows that a far reaching insight into this problem
was obtained in Old-Babylonian times. The text in question
belongs to the Plimpton Collection of Columbia University in
New York.
As is evident from the break at the left-hand side. this tablet
was originally larger; and the existence of modem glue on the
break shows that the other part was lost after the tablet was
excavated. Four columns are preserved. to be counted as usual
from left to right. Each column has a heading. The last heading
is "its name" which means only "current number". as is evident
from the fact that the column of numbers beneath it counts simply
 Babylonian Mathematics 37

the number of lines from "lst" to "15th". This last column is

therefore of no mathematical interest. Columns II and III are
headed by words which might be translated as "solving number
of the width" and "solving number of the diagonal" respectively.
"Solving number" is a rather unsatisfactory rendering for a term
which is used in connection with square roots and similar oper-
ations and has no exact equivalent in our modern terminology.
We shall replace these two headings simply by "b" and "d"
respectively. The word "diagonal" occurs also in the heading of
the first column but the exact meaning of the remaining words
escapes us.
The numbers in columns I, II and III are transcribed in the
following list. The numbers in [ ] are restored. The initial
r
numbers "[1 in lines 4 fl. are half preserved, as is clearly seen
from the photograph (PI. 7 a). A "1" is completely preserved in
line 14. In the transcription I have inserted zeros where they
are required; they are not indicated in the text itself.

I II ( .. b) III(-d) IV
[1,59,0,]15 1,59 2,49 1
[1,56,56,]58,14,50,6,15 56,7 3,12,1 2
[1,55,7,]41,15,33,45 1,16,41 1,50,49 3
[1,]5[3,1]0,29,32,52,16 3,31,49 5,9,1 4
[1,]48,54,1,40 1,5 1,37 5
[1,J47,6,41,40 5,19 8,1 6
[1,]43,11,56,28,26,40 38,11 59,1 7
[1,]41,33,59,3,45 13,19 20,49 8
[1,]38,33,36,36 9,1 12,49 9
1,35,10,2,28,27,24,26,40 1,22,41 2,16,1 10
1,33,45 45 1,15 11
1,29,21,54,2,15 27,59 48,49 12
[1,]27,0,3,45 7,12,1 4,49 13
1,25,48,51,35,6,40 29,31 53,49 14
[1,]23,13,46,40 56 53 15

This text contains a few errors. In 11,9 we find 9,1 instead of 8,1
which is a mere scribal error. In 11,13 the text has 7,12,1 instead
of 2,41. Here the scribe wrote the square of 2,41, which is 7,12,1
instead of 2,41 itself. In 111,15 we find 53 instead of 1,46 which
38 Chapter II

is twice 53. Finally there remains an unexplained error in 111,2

where 3,12,1 should be replaced by 1,20,25.
The relations which hold between these numbers are the
following ones. The numbers band d in the second and third
columns are Pythagorean numbers; this means that they are
integer solutions of

As band d are known from our list, we can compute I and find

Line I Line I Line I

1 2,0 6 6,0 11 1,0
2 57,36 7 45,0 12 40,0
3 1,20,0 8 16,0 13 4,0
4 3,45,0 9 10,0 14 45,0
5 1,12 10 1,48,0 15 1,30
dB
If we then form the values of II we obtain the numbers of

column I. Thus our text is a list of the values of ~, b, and d,

for Pythagorean numbers. It is plausible to assume that the values

of 1 were contained in the missing part. That they have been
explicitly computed is obvious.
If we take the ratio b for the first line we find 1,59 = 0;59,30
I 2,0
that is, almost 1. Hence the first right triangle is very close to
half a square. Similarly one finds that the last right triangle has
angles close to 30° and 60°. The monotonic decrease of the
numbers in column 1 suggests furthermore that the shape of the
triangles varies rather regularly between these two limits. If one
investigates this general fact more closely, one finds that the
values of ~ in column I decrease almost linearly and that this

holds still more accurately for the ratios ~ themselves (Fig. 3).
This observation suggests that the ancient mathematician who
composed this text was interested not only in determining triples
of Pythagorean numbers but also in their ratios :. Let us inves-
 Babylonian Mathematics 89
tigate the mathematical character of this problem. We know that
all Pythagorean triples are obtainable in the form
I = 2 pq b = pi - ql d = p'l. + q2
where p and q are arbitrary integers subject only to the condition
that they are relatively prime and not simultaneously odd and
p > q. Consequently we obtain for the ratio 1d the expression

d = 1(p . q + p . q)
I
d
where p and q are the reciprocals of p and q. This shows that I
are expressible as finite sexagesimal fractions, as is the case in
our text, if and only if both p and q are regular numbers.
This fact can be easily checked in our list of numbers by
computing the values of p and q which correspond to the I, b,
and d of our text. Then one finds a very remarkable fact. The
numbers p and q are not only regular numbers, as expected,
but they are regular numbers contained in the "standard table"

I"'-.. ~
~
sJ:
K
r----.
~
~

~
~
i"-.
..........
r-.... ~
"""-
-~ ..... ~

-r-.... I"- .........

......... ~ """-

,
~
I I I I I I I I I I I I I I I
'0, 2 3 4 6 7 8 9 10 " 12 /3 1+ 1$
Fig. 3.

of reciprocals (p. 82) so well known to us from many tables of

the same period. The only apparent exception is p = 2,5 but
this number is again well known as the canonical example for the
computation of reciprocals beyond the standard table. This
40 Chapter II

seems to me a strong indication that the fundamental formula

for the construction of triples of Pythagorean numbers was
known. Whatever the case may be, the text in question remains
one of the most remarkable documents of Old-Babylonian
mathematics. We shall presently (p. 42) return to the question
how a formula for Pythagorean numbers could have been found.
21. Pythagorean numbers were certainly not the only case of
problems concerning relations between numbers. The tables for
squares and cubes point clearly in the same direction. We also
have examples which deal with the sum of consecutive squares
or with arithmetic progressions. It would be rather surprising if
the accidentally preserved texts should also show us the exact
limits of knowledge which were reached in Babylonian mathe-
matics. There is no indication, however, that the important
concept of prime number was recognized.
All these problems were probably never sharply separated
from methods which we today call "algebraic". III the center of
this group lies the solution of quadratic equations for two un-
knowns. As a typical example might be quoted a problem from a
Seleucid text. This problem requires the finding of a number such
that a given number is obtained if its reciprocal is added to it.
Using modern notation we call the unknown number x, its
reciprocal x, and the given number b. Thus we have to determine
x from
xx= 1 x x = b. +
In the text b has the value 2;0,0,33,20. The details of the solution
are described step by step in the text as follows. Form

(:r = 1;0,0,33,20,4,37.46,40.

Subtract 1 and find the square root

v(it- 1 = V0;0,0.33.20.4.37,46.40 = 0;0,44.43.20.

The correctness of this result is checked by squaring. Then add

b
to and subtract from 2 the result. This answers the problem:
 Babylonian Mathematics 41

z = -b + y- = 1;0,0,16,40 + 0;0,44,43,20 = 1;0,45

2

;: = b _
2
y- = 1;0,0,16,40 - 0;0,44,43,20 = 0;59,15,33,20.
Indeed, z and x are reciprocal numbers and their sum equals the
given number b.
This problem is typical in many respects. It shows, fIrst of all,
the correct application of the "quadratic formula" for the solu-
tion of quadratic equations. It demonstrates again the unrestricted
use of large sexagesimal numbers. Finally, it concerns the main
type of quadratic problems of which we have hundreds of ex-
amples preserved, a type which I call "normal form": two num-
bers should be found if (a) their product and (b) their sum or
difference is given. It is obviously the purpose of countless ex-
amples to teach the transformation of more complicated quadratic
problems to this "normal form"
z·y=a
z±y=b
from which the solution then follows as

x=:+v(:r~a
g=±:~V(~r~a
simply by transforming the two original equations into two linear
equations
z±y=b
z =F y = y-bS-=F-
4 -a.
In other ;words, reducing a quadratic equation to its "normal form"
means finally reducing it to the simplest system of linear equations.
The same idea can be used for finding three numbers, a. b, c,
which satisfy the Pythagorean relation. Assume that one again
started from a pair of linear equations
a=z+y
b=z-y
42 Chapter II

realizing that
all = bl + ell if ell = 4xy.
Assuming that x and yare integers, then a and b will be integers;
V V
but e = 2 xy will be an integer only if xy is an integer. This
condition is satisfied if we assume that x and U are squares of
integers
x = pi y = ql
and thus we obtain the final result that a, b, and e form a Pytha-
gorean triple if p and q are arbitrary integers (p > q) and if we
make
b = pi _ qll e = 2pq.

This is indeed the formula which we needed for our explanation

of the text dealing with Pythagorean numbers.
!!. It is impossible to describe in the framework of these
lectures the details of the Babylonian theory of quadratic equa-
tions. It is not really necessary anyhow, since the whole material
is easily available in the editions quoted in the bibliography to
this chapter. A few features of this Babylonian algebra, however,
deserve special emphasis because they are essential for the
evaluation of this whole system of early mathematics.
First of all, it is easy to show that geometrical concepts play a
very secondary part in Babylonian algebra, however extensively
a geometrical terminology may be used. It suffices to quote the
existence of examples in which areas and lengths are added, or
areas multiplied, thus excluding any geometrical interpretation
in the Euclidean fashion which seems so natural to us. Indeed,
still more drastic examples can be quoted for the disregard of
reality. We have many examples concerning wages to be paid for
labor according to a given quota per man and day. Again, problems
are set up involving sums, differences, products of these numbers
and one does not hesitate to combine in this way the number of
men and the number of days. It is a lucky accident if the unknown
number of workmen, found by solving a quadratic equation, is
an integer. Obviously the algebraic relation is the only point of
interest, exactly as it is irrelevant for our algebra what the letters
may signify.
Another important observation concerns the form in which all
these algebraic problems are presented. The texts fall into two
 Babylonian Mathematics 43

major classes. One class formulates the problem and then

proceeds to the solution, step by step, using the special numbers
given at the beginning. The text often terminates with the words
"such is the procedure". The second class contains collections
of problems only, sometimes more than 200 on a single tablet
of the size of a small printed page. These collections of problems
are usually carefully arranged, beginning with very simple cases
e. g., quadratic equations in the normal form, and expanding
step by step to more complicated relations, but all eventually
reducible to the normal form. One standard form of such collec-
tions consists in keeping the condition xy = 10,0 fixed but
varying the second equation to more and more elaborate poly-
nomials, ending up, e. g., with expressions like

(3x+2y)I+-h {4[l«x + y) -(I + 1) (x- y»]2+(X+y)2}

= 4,45,0.

Investigating such series, one finds that they all have the same
pair x = 30 Y = 20 as solutions. This indicates that it was of
no concern to the teacher that the result must have been known
to the pupil. What he obviously had to learn was the method of
transforming such horrible expressions into simpler ones and to
arrive finally at the correct solutions. We have several tablets of
the first class which solve one such example after another from
corresponding collections of the second class.
From actually computed examples it becomes obvious that it
was the general procedure, not the numerical result, which was
considercd important. If accidentally a factor has the value 1
the multiplication by 1 will be explicitly performed, obviously
bccause this step is necessary in the general case. Similarly we
find regularly a gencral explanation of the procedure. Where
we would write x + y the text would say "5 and 3, the sum of
length and width". Indeed it is often possible to transform these
examples directly into our symbolism simply by replacing the
ideograms which were used for "length,., "width", "add",
"multiply" by our letters and symbols. The accompanying
numbers are hardly more than a conycnient guide to illustrate
the underlying gcneral process. Thus it is substantially incorrect
if onc dcnics the use of a "general formula" to Babylonian
44 Chapter II

algebra. The sequences of closely related problems and the

general rules running parallel with the numerical solution form
de facto an instrument closely approaching a purely algebraic
operation. Of course, the fact remains that the step to a consci-
ously algebraic notation was never made.
23. The extension of this "Babylonian algebra" is truly remark-
able. Though the quadratic equations form obviously the most
significant nucleus a great number of related problems were also
considered. Linear problems for several unknowns are common
in many forms, e. g., for "inheritance" problems where the shares
of several sons should be determined from linear conditions
which hold between these shares. Similar problems arise from
divisions of fields or from general conditions in the framework
of the above mentioned collections of algebraic examples.
On the other hand we know from these same collections
series of examples which are equivalent to special types of equa-
tions of fourth and sixth order. Usually these problems are
easily reducible to quadratic equations for Xl or Xl but we have
also examples which lead to more general relations of 5th and
+
3rd order. In the latter case the tables for n l n l seem to be useful
for the actual numerical solution of such problems, but our source
material is too fragmentary to give a consistent description of
the procedure followed in cases which are no longer reducible to
quadratic equations.
There is finally no doubt that problems were also investigated
which transcend, in the modern sense, the algebraic character.
This is not only clear from problems which have to do with
compound interest but also from numerical tables for the con-
secutive powers of given numbers. On the other hand we have
texts which concern the determination of the exponents of given
numbers. In other words one had actually experimented with
special cases of logarithms without, however, reaching any
general use of this function. In the case of numerical tables the
lack of a general notation appears to be much more detrimental
than in the handling of purely algebraic problems.
J.f.. Compared with the algebraic and numerical component
in Babylonian mathematics the role of "geometry" is rather
insignificant. This is, in itself, not at all surprising. The central
problem in the early development of mathematics lies in the
 Babylonian Mathematics 45

numerical determination of the solution which satisfies certain

conditions. At this level there is no essential difference between
the division of a sum of money according to certain rules and the
division of a field of given size into, say, parts of equal area.
In all cases exterior conditions have to be observed, in one case
the conditions of the inheritance, in another case the rules for the
determination of an area, or the relations between measures or
the customs concerning wages. The mathematical importance of
a problem lies in its arithmetical solution; "geometry" is only
one among many subjects of practical life to which the arithme-
tical procedures may be applied.
This general attitude could be easily exemplified by long lists
of examples treated in the preserved texts. Most drastically,
however, speak special texts which were composed for the use
of the scribes who were dealing with mathematical problems and
had to know all the numerical parameters which were needed in
their computations. Such lists of "coefficients" were first identified
by Professor Goetze of Yale University in two texts of the Yale
Babylonian Collection. These lists contain in apparently chaotic
order numbers and explanatory remarks for their use. One of
these lists begins with coefficients needed for "bricks" of which
there existed many types of specific dimensions, then coefficients
for "walls", for "asphalt", for a "triangle", for a "segment of a
circle", for "copper", "silver", "gold" and other metals, for a
J

"cargo boat", for "barley", etc. Then we find coefficients for

"bricks" • for the "diagonal"• for "inheritance" • • for "cut reed"
etc. Many details of these lists are still obscure to us and demon-
strate how fragmentary our knowledge of Babylonian mathematics
remains in spite of the many hundreds of examples in our texts.
But the point which interests us here at the moment becomes very
clear, namely, that "geometry" is no special mathematical disci-
pline but is treated on an equal level with any other form of
numerical relation between practical objects.
These facts must be clearly kept in mind if we nevertheless
speak about geometrical knowledge in Babylonian mathematics,
simply because these special facts were eventually destined to
play a decisive role in mathematical development. It must also
be underlined that we have not the faintest idea about anything
amounting to a "proof" concerning relations between geometrical
46 Chapter II

magnitudes. Several tablets dealing with the division of areas

show figures of trapezoids or triangles but without any attempt at
being metrically correct. The description of geometry as the
science of proving correct theorems from incorrect figures cer-
tainly fits Babylonian geometry so far as the figures are concerned
and also with regard to the algebraic relations. But the real
"geometric" part often escapes us. It is, for instance, not at all
certain whether the triangles and trapezoids are right-angle
fIgUres or not. If the texts mention the "length" and "width" of
such a figure it is only from the context that we can determine
the exact meaning of these two ternlS. If the area of a triangle is
found by computing i Q • b it is plausible to assume that Q and b
are perpendicular dimensions, but there exist similar cases where
only approximate formulae seem equally plausible.
There are nevertheless cases where no reasonable doubt can
arise as to the correct interpretation of geometrical relations.
The concept of similarity is utilized in numerous examples. The
Pythagorean theorem is equally well attested; the same holds
for its application to the determination of the height of a circular
segment. On the other hand only a very crude approximation
for the area of a circle is known so far, corresponding to the use
of 3 for 1f. Several problems concerning circular segments and
similar figures are not yet fully understood and it seems to nle
quite possible that better approximations of 1f were known and
used in cases where the rough approximation would lead to
obviously wrong results.
As in the case of elementary areas similar relations were known
for volumes. Whole sections of problem texts are concerned with
the digging of canals, with dams and similar works, revealing
to us exact or approximate formulae for the corresponding
volumes. But we have no examples which deal with these objects
from a purely geometrical point of view.
24 Q. After completion of the manuscript, new discoveries were
made which must be mentioned here because they contribute
very essentially to our knowledge of the mathematics of the Old-
Babylonian period. In 1936 a group of mathematical tablets
were excavated by French archaeologists at Susa, the capital of
ancient Elam, more than 200 miles east of Babylon. A preliminary
report was published in the Proceedings of the Amsterdam
 Babylonian Mathematics 47

Academy by E. M. Bruins in 1950 and the following remarks

are based on this preliminary publication, though I restrict myself
to the most significant results only. The texts themselves still
remain unpublished, more than 20 years after their discovery.
The main contribution lies in the direction of geometry. One
tablet computes the radius r of a circle which circumscribes
an isosceles triangle of sides 50, 50, and 60 (result r = 31;15).
Another tablet gives the regular hexagon, and from this the
approximation Vi I:d 1;45 can be deduced. The main interest,
however, lies in a tablet which gives a new list of coefficients
similar to those mentioned above, p. 45. The new list contains,
among others, coefficients concerning the equilateral triangle
(confirming the above approximation V3 ~ 1;45), the square
(¥2 I:d 1;25), and the regular pentagon, hexagon, heptagon, and
the circle. If An denotes the area, sn the side of a regular n-gon,
then one can explain the coefficients found in the list as follows:
A5 = 1;40 . s~
A6 = 2;37,30 . ':
A7 = 3;41 . s~.

If we, furthermore, call c, the circumference of the regular

hexagon, c the periphery of the circle, then the text states
c6 = 0;57,36 . c.
3
Because c, = - c, the last coefficient implies the approximation
n
n I:d 3;7,30 = 31
thus confirming finally my expectation that the comparison of the
circumference of the regular hexagon with the circumscribed
circle must have led to a better approximation of n than 3.
The relations for A5 , A6 , and A7 correspond perfectly to the
treatment of the regular polygon in Heron's Metrica XVIII to XX,
a work whose close relationship to pre-Greek mathematics has
become obvious ever since the decipherment of the Babylonian
mathematical texts.
Also in many other respects do the tablets from 8usa supple-
ment and confirm what we knew from the contemporary Old-
Babylonian sources in Mesopotamia proper. One example deals
48 Chapter II

with the division of a triangle into a similar triangle and a trapezoid

•
such that the product of the partial sides and of the partial areas
are given values, the hypotenuse of the smaller triangle being
known. This is a new variant of similar problems involving sums
of areas and lengths or the product of areas. One of the tablets
from Susa implies even a special problem of the 8th degree,
whereas until now we had only the sixth degree represented in
the Babylonian material. The new problem requires that one
find the sides x and y of a rectangle whose diagonal is d, such
that xy = 20,0 and x 8 • d = 14,48,53,20. This is equivalent to
a quadratic equation for x'
x 8 + a1x' = bl
a = 20,0 b = 14,48,53,20. The text proceeds to give the step-
by-step solution of this equation, resulting in x' = 11.51,6,40
and finally leading to x = 40 Y = 30.
25. However incomplete our present knowledge of Babylonian
mathematics may be, so much is established beyond any doubt:
we are dealing with a level of mathematical development which
can in many aspects be compared with the mathematics, say,
of the early Renaissance. Yet one must not overestimate these
achievements. In spite of the numerical and algebraic skill and
in spite of the abstract interest which is conspicuous in so many
examples, the contents of Babylonian mathematics remained
profoundly elementary. In the utterly primitive framework of
Egyptian mathematics the discovery of the irrationality of V2
would be a strange miracle. But all the foundations were laid
which could have given this result to a Babylonian mathematician,
exactly in the same arithmetical form in which it was obviously
discovered so much later by the Greeks. And even if it were only
due to our incomplete knowledge of the sources that we assume
that the Babylonians did not know that pi = 2 ql had no solution
in integer numbers p and q, even then the fact remains that the
consequences of this result were not realized. In other words
Babylonian mathematics never transgressed the threshold of pre-
scientific thought. It is only in the last three centuries of Babylonian
history and in the field of mathematical astronomy that the
Babylonian mathematicians or astronomers reached parity with
their Greek contemporaries.
 Bibliography; Notes and References 49
BIBLIOGRAPHY TO CHAPTER II
The description of Babylonian mathematics given here is far from complete,
even if measured by the fragmentary existing material. A more detailed descrip-
tion has been given in the author's "Vorgriechische Mathematik", Berlin,
Springer 1934. Any serious study must be based, however, on the texts themselves,
in order to get a proper estimate of the sometimes fiueDt boundaries between
established facts and modem interpretation. The majority of the now available
texts are published in the following works:
O. Neugebauer, Mathematische Keilschrift-Texte, 3 vols., Berlin, Springer,
1935/37 = Quellen und Studien zur Geschichte der Mathematik, Abt. A, vol. 2.
[Henceforth quoted MKT.]
O. Neugebauer-A. Sachs, Mathematical Cuneiform Texts. American
Oriental Society, New Haven, 1945 = Am. Oriental Series vol. 29. [Henceforth
quoted MCT.]
Both these works give photographs, copies, transcriptions, translations, and
commentaries. The main part of the material published in the first-mentioned
work was republished in transcription and with translation by F. Thureau-
Dangin, Textes mathematiques babyloniens, Leiden, Brill, 1938 = Ex Oriente
Lux vol. 1.
Since the publication of the first edition of this book several more mathematical
texts have come to light, supplementing but not essentially changing the general
picture given in the preceding pages. Most of these new texts are again Old-
Babylonian problem texts, published by Sayyid Taha Baqir in Sumer vol. 6
(1950) p. 39-54, p. 130-148 and vol. 7 (1951) p. 28-45 (found in Tell J:farmal,
just south of Baghdad). For a list of coefficients and related subjects cf. A.
Goetze, Sumer 7 (1951) p. 126-154. Additional texts were discussed by Bruins
in Sumer 9 and 10 (1953/54) but only in exerpts or in very unreliable transcrip-
tions. A problem text of unknown origin, concerning a circular city (similar to
MKT I p. 144) was published by W. F. Leemans in Compte Rendu de la
seconde re;ncontre assyriologique intern., Paris 1951, p. 31-35. Two fragments
of problem texts and 16 table texts from the Late-Babylonian archive in Babylon
are reproduced in Pinches-Strassmaier-Sachs, Late Babylonian Astronomical
and Related Texts, Providence, Brown University Press, 1955. Cf. Also Sachs,
J. Cuneiform Studies 6 (1952) p. 151-156.

NOTES AND REFERENCES TO CHAPTER II

ad 17. There exists a single fragment of a mathematical text written in
Sumerian (MKT I p. 234 f.). Because Sumerian was still practiced in the schools
of the Old-Babylonian period nothing can be concluded from such a text for
the Sumerian origin of Mesopotamian mathematics. The same holds for the
exeeedingly frequent use of Sumerian words and phrases throughout all periods.
That mathematics was taught in scribal schools can hardly be doubted. At
what level such instruction started and to what extent it was the common knowl-
edge of scribes it is impossible to say. There exists a text, probably itself written
50 Chapter II

for use in scribal schools, in which the trying life of a schoolboy in such a
"Tablet House" is dramatically described. Cf. S. N. Kramer, Schooldays, a
Sumerian Composition Relating to the Education of a Scribe. J. Am. Oriental
Soc. 69 (1949) p. 199-215.
ad 18. The structure of the system of tables of multiplication and reciprocals
was first described by the present author in a series of papers entitled "Sexagesi-
malsystem und babylonische Bruchreehnung" I-IV published in Quellen und
Studien zur Geschichte der Mathematik, Abt. B., vols. 1 and 2 (1930-1932).
The methods for the computation of reciprocals not contained in the standard
table were analyzed by A. Sachs, Babylonian Mathematical Texts I, Journal
of Cuneiform Studies 1 (1947) p. 219-240. The transformation of sexagesimal
fractions to unit fractions was discussed by the same author in "Notes on
Fractional Expressions in Old Babylonian Mathematical Texts", J. of Near
Eastern Studies 5 (1946) p. 203-214.
That a scribe was sometimes not quite sure when a number was regular or
irregular is shown by a statement found in a text now in the British Museum
(MKT I p. 224,12 and p. 184), to the effect that "4,3 does not divide". This is
wrong because 4,3 == 3·1,21 and both 3 and 1,21 are regular numbers whose
reciprocals can be found in the standard table.
ad 19. Approximations of V2 can be found as follows. Obviously 1 == 1;30
is a first approximation, though larger than the correct value because (ir II:::

t
== = 2;15. Consequently we obtain an approximation which is too small by
I.
dividing 2 by The result is • == 1;20. The mean value of these two opposite
approximations is 1;25 which is one of the two attested values. We can
repeat this process; 1;251 == 2;0,25 is too large. Thus 2 divided by 1;25 or
1;24,42,21, ••. is too small. The mean value of 1;25 and 1;24,42,21 is 1;24,51,10
which is the second approximation found in our texts.
We have no proof that this was the way in which these values were found
but there is also no way to disprove this possibility. Cf. also below p. 52.
ad 20. In linie 2 of column III of the Plimpton tablet we find for d the value
3,12,1 instead of 1,20,25. Following a suggestion by R. J. Gillings (The Australian
Journal of Science 16, 1953, p. 54-56) one might explain this mistake as the
result of two errors: In computing

dl == pi + qI == (p + q)1 - 2pq (p == 1,4 q = 27)

the scribe replaced - 2pq by + 2pq and then wrote down only 2·27·1,0 =
54,0 instead of 2 ·27 . 1,4 == 57,36. Thus he found not d - 2,18,1 - 57,36 ==
1,20,25 but 2,18,1 + 54,0 = 3,12,1.
A different suggestion was made by E. M. Bruins, in Sumer 11 (1956)
p. 117-121 which contains, however, incorrect and unfair statements as to the
readings of the text.
It may finally be remarked that the construction of Pythagorean triples by
means of two numbers p and q from

1=2pq b=pl_qI
 Notes and References 61

was well known in Hellenistic times. Diophantus, whose relationship to the

oriental tradition is obvious, uses it frequently (e. g. VI, 1). The same holds
for the Indian mathematicians, e. g., Mahavira (about 850 A.D.) or Bhascara
(about 1150); d. Mahavira. 00. Nangacarya p. 209 and LDavati VI, 135 and
145, Colebrooke.
ad 22. Examples for Donhomogeneous terms:
MCT p. 74: addition of areas and volumes.
MKT I p. 243; MKT II p. 63: addition of lengths and areas.
MKT II p. 64: addition of length and volume.
MKT I p. 513: addition of number of days and of men.
Thureau-Dangin. TMB p. 209 f.: addition of sheep and rams.
For a careful classification of quadratic equations see Solomon Gandz, The
Origin and Development of the Quadratic Equations in Babylonian. Greek,
and Early Arabie Algebra. Osiris 3 (1937) p. 405-557.
ad 29. For cubic equations see MKT I p. 208 tr., and vol. III p. 55. Fourth
order: MKT I p. 420; p. 456 and III. p. 62; p. 471 ft.; p. 498. Fifth order:
MKT I p. 411. Sixth order: MKT I p. 460.
Tables for an: MKT I p. 77 ft. For logarithms d. MKT I p. 362, MCT p. 35.
The tables for an were found at Kish, east of Babylon. and at Tell Harmal
near Baghdad.
An interesting table of special square roots is contained in the following
text (MKT III p. 52):
1 e 1 ~si8
1.2,1 e 1,1 ~si8
1,2,3,2,1 e 1,1,1 ~si8
1,2,3,4,3,2,1 e 1,1,1,1 ~si8

A following IIcatch line" points to a succeeding table for the cube roots of 1,
1,3,3,1 etc. The knowledge of the binomial coefficients lies, of course. fully
within the 'reach of Babylonian algebra.
ad 24. It is of interest to remark that not only were similar triangles frequently
used in the solution of problems which have geometrical backgrouDd. but that
the arithmetical concept IIratio" had a special term. MKT I p. 460 Jr. we have
series of examples where the IIratio" xll/ is to be computed from quadratic
equations. That the ratio of two numbers is treated as an entity is indeed a
very important step in the development of algebra.
The area A of a circle is usually determined from its circumference c in the
form
A = 0;5· ell
1
where 0;5 = 1/12 is an approximation of 4 . For many examples d. MCT
~«M~~ n
Problems concerning circular segments are published MCT 56, MKT I p. 188.
Cf. also MKT I p. 177, p. 230; MCT 134ft. All these problems cause trouble-
which is a certain indication that we have not yet found the proper key to this
part of Babylonian geometry. A very interesting early text concerning ornamental
patterns of clrcles and squares was published by J. C. Gadd (d. MKT I p. 137ft.).
Unfortunately no solutions are given.
52 Notes and References

For the inaccuracy of figures ct. MCT p. 46 and p. 54. For the approximate
determination of volumes see MKT I p. 165, p. 176; for areas MCT p. 46.
For a clear case of a figure which must be exactly a right triangle in order to
make the following simllarity relations correct, d. the tablet published by
Sayyid Taha Baqir, Sumer 6 (1950) p. 39-54.
ad 2.f. a. The paper referred to in the text was preceded by a preliminary
note in the August 1950 issue of the French popular journal UAtomes. Tous les
aspects scientiftques d'un nouvel age", (p. 270 f.). This article also gives a
photograph of a triangle with its circumscribed circle.
The value V3 ~ 1;45 can be obtained immediately by the process described
above p. 50 in the case of VI. StartiDg with the obvious estimate V3 ~ 1:30,
3
one obtains as the next value 1;30 = 2 and hence i (2 + 1;30) = 1;45. The
same approximation is found in Heron, Metrica XXV (who uses 1;44 also).
Metrica XXI contains VI ~ 1;25.
The value n "., 31 does not seem to be attested in the preserved literature
of antiquity. As its first appearance, Tropfke (Gesehichte d. Elementarmathe-
malik IV(8) p. 279) quotes a passage from DUrer in 1525. In Babylonian material
this value was hesitating1y mentioned as a possibility in MCT p. 59 note 152k.
This conjecture is now fully confirmed.
 58

CHAPTER III

The Sources; Their Decipherment and Evaluation.

26. There are many forces which cooperate in the destruction
of source material, none more powerful than continuous peaceful
life. The lack of interest in the far remote past will invariably
change and eventually destroy what l·emains from earlier gener-
ations.Without violent catastrophies there would hardly be any
archeology. If Mesopotamian cities had not been turned into
desert hills we would have no chance of finding the hundreds
of thousands of documents from which Babylonian history is
written.
It would be pointless to describe in any detail the unending
sequence of disaster which provided us with the material for our
studies. We shall devote our discussion exclusively to the modern
attempts at reviving the past and penetrating into the intellectual
life of previous generations.
Following the general plan of these lectures I shall not try to
describe familiar aspects of historical research at great length.
I

I will emphasize, however, some specific methodological facts

which are directly related to our main topic. the investigation of
ancient science.
27. Best known of all, of course, is Greek science. The great
classics are carefully edited and many of them are available
in excellent translations. For the historian of science this is very
pleasant indeed, but far from sufficient. Nobody would expect a
historian of English literature to remain satisfied with an edition
of Shakespeare or Chaucer. Also the history of a science can only
be written if more is available than the "classics". The predeces-
sors, the pupils. the related authors are still exceedingly difficult
to reach.
Let me mention only one very typical example. Ptolemy's
54 Chapter III

"Almagest" was edited in Greek by Heiberg in 1898-1903. An

excellent critical and annotated translation into German by
Manitius was published in 1912-1913. But the planned glossary
was never printed, making it exceedingly difficult to check
Ptolemy's terminology with other sources. This fact is also re-
sponsible for serious gaps even in the best modern Greek diction-
aries, not to mention that the history of astronomical terminology
in general is practically a terra incognita. Modern editions have
played havoc with abbreviations, symbols, drawings, etc. in the
original manuscripts. One has to do practically all the work over
again if one should try to investigate the development of symbols
like the zodiacal signs or the planetary sigla. The best list of
mathematical, astronomical and chemical sYmbols is still the
collection made by Du Cange in the appendix to his "Glossarium
... " (1688) which, in turn, is based on a sheet, now in the
Bibliotheque Nationale, written around 1480 by Angelo Poliziano,
the teacher of Piero di Medici. This is a characteristic example of
the true state of affairs in the study of the history of scientific
developments.
It is only recently that scholars following A. Rome and A. De-
latte, have begun to publish editions where the figures and their
lettering are taken as part of the text. With these recent exceptions
no edition can be trusted in the least with respect to appearance,
lettering or even existence of figures. The question, for instance,
how the ancients depicted geometrical relations on a sphere cannot
be seriously discussed on the basis of the existing printed texts.
Ptolemy's other works are slowly being published. One volume
appeared in 1907, containing, among others, important writings
on the theory of sun dials and on stereographic projection which
is the basis of the famous astrolabe, one of the most important
instruments of medieval astronomy. The "Tetrabiblos", the
"Bible of astrology", was published twice during World War II.
One edition, by E. Boer, appeared in Germany, Greek text only;
the other, by F. E. Robbins, Greek with English translation, in
the Loeb Classical Library. Thus once the experiment was un-
willingly made of testing the uniqueness of the modem text-
critical methods. It is amusing to see that the differences begin
with the title and continue in varying degree in the division of
chapters and sections. Of course in essence the results are the
same, but the details are by no means identical.
 The Sources; Their Decipherment and Evaluation 55

An enormous literature has clustered around Ptolemy's "Geo-

graphy", one of the most influential books of antiquity. Never-
theless, no reliable edition exists. The task is indeed of great
difficulty. The constant use of this work has greatly affected its
tradition and it is a major enterprise to restore the original version
of a text which to a large extent consists of geographical names
and numbers uncheckable by internal evidence which is for-
tunately available in the case of astronomical tables.
To summarize, we may say that even Ptolemy's work is only
in part available, disregarding completely lost works, fragments
of which mayor may not appear on papyri or in some obscure
oriental library.
Early Greek astronomy from its beginnings about 400 B.C. to
Ptolemy (about 150 A.D.) is almost completely destroyed, except
for a few very elementary works which survived for teaching
purposes. But the rest was obliterated by Ptolemy's outstanding
work, which relegated his predecessors to merely historically
interesting figures.
As to Ptolemy's successors we should be in a much better
position. Pappus's and Theon's commentaries, written in the
4th century, were widely used and have in part survived. They
are now in the process of publication by A. Rome. We are still
very badly off so far as the tables are concerned, though at least
a preliminary publication by Halma exists, more than 100 years
old and bristling with misprints and errors. But almost nothing
has been done with Byzantine or European m.edieval tables. Thus
all work on mathematical astronomy of the Middle Ages is most
seriously handicapped by the fact that almost no tables are
accessible, though hundreds of them can be found listed in
library catalogs. The much publicized "progress in the study of
U

the history of science is difficult to reconcile with the shocking

neglect of a great wealth of source material which is of primary
importance for our knowledge of Byzantine astronomy. The study
of the problem of the interaction between Islamic science and the
West is precluded as long as these sources remain unpublished.
What we really need is not bibliographies and summaries, but
competent publications of Islamic, Greek, and Latin treatises.
28. There is one group of sources which will become of in-
creasing importance when systematically utilized: the astrological
writings. During the last 60 years a "catalogue" in 12 volumes
56 Chapter III

of Greek astrological writings has been completed. The text is

Greek, the notes are Latin, the indices are restricted to proper
names and occasionally to selected terminology. And the content
can only be called repelling-hundreds and hundreds of pages
of the dryest astrological nonsense. Nevertheless I think that the
scholars who have undertaken this publication, foremost of all
Franz Cumont, have contributed enorlnously to the study of
ancient civilization, far beyond the narrow limits of the history
of astrology. To quote only one work, I mention Cumont's
"L'Egypte des astrologues" (1937). Here, stripped of all astrology,
the background of the daily life and of the contemporary in-
stitutions of Hellenistic Egypt is depicted as it becomes visible
from the mishaps and fortunes predicted for the men and women
who consulted the astrologer. And another fact of great historical
importance becomes increasingly clear from these researches,
namely, that the date of origin of this mighty astrological lore
must be fixed to the Ptolemaic period in Egypt and thus appears
as a truly Hellenistic creation.
We shall come back to this aspect of our problem in the last
chapter. At the moment we will only consider the astrological
literature as a source of information for the history of astronomy.
Indeed, these texts contain innumerable scattered fragments of
computations concerning the moon, the planets, positions of stars,
their risings and settings. These computations are often almost
hopelessly distorted. Many centuries of tradition through hand-
written copies have badly affected numbers which were of little
interest, if not unintelligible to the scribes. Nevertheless, we obtain
from these passages many references to methods which belong
to the period between Hipparchus and Ptolemy. One of the most
unexpected discoveries was made by W. Gundel. In an Old-
French and a related Latin astrological manuscript of the Renais-
sance he found imbedded the fragments of a star catalogue of
the time of Hipparchus. Slowly there emerges from scattered
scraps of information a whole system of astronomical methods
which are very different from the classical "Ptolemaic" system
but which are of primary importance for the study of the origin
and transmission of Hellenistic astronomy.
!9. The majority of manuscripts on which our knowledge of
Greek science is based are Byzantine codices, written between
 The Sources; Their Decipherment and Evaluation 57
500 and 1500 years after the lifetime of their authors. This
suffices for one to realize the importance of every scrap of
papyrus from a scientific or astrological treatise. Here we have
originals which were written during the Hellenistic period itself,
not yet subject to the selective editing of late centuries. It can
be said without any exaggeration that the relatively young field
of papyrology has truly revolutionized classical studies-even if
by natural inertia this has not always become evident in the
standard curricula.
The fascinating story of the recovery of papyrus treasures from
the soil of Egypt has often been told and need not be repeated
here. A masterly little book by one of the leading scholars in
the field, H. Idris Bell, "Egypt from Alexander the Great to the
Arab Conquest" (Oxford, 19(8), will not only give the reader a
summary of papyrology, its history and methods, but is itself a
brilliant study of the diffusion and decay of Hellenism, the very
problem one facet of which is the subject of the present lectures.
Papyrology is one of the best organized and most pleasantly
managed fields of the humanities. An unusual spirit of cooperation
has survived two great wars. Great series of publications of texts
have appeared regularly and the intrinsic difficulties of the field
and its enormous spread into highly specialized disciplinE'S,
especially law, agriculture, economics, etc. have created the
active cooperation of scholars of neighboring disciplines. Con-
sequently, generally usable editions were produced with trans-
lations, commentaries and excellent indices, glossaries and hand-
books-indeed a pleasant contrast to, say, the fact that the Arabic
version of Euclid is published only with a Latin translation (1897-
1932) so as to make life miserable to a mathematician who
would perhaps for once want to look into the Oriental tradition
of a classic in his field.
In spite of the very active and successful work of papyro-
logists ,their number is much too small to cope with the large
amount of material which has reached museums and smallel·
collections. Many hundreds of papyri and papyrus fragments are
rapidly disintegrating into dust after having been purchased at
high prices from antiquity dealers. From my own very limited
experience I could quote several instances where papyri got lost
in smaller collections, which have neither facilities nor competent
58 Chapterm

personnel for the proper handling of these documents. Especially

well preserved texts are valuable for the antiquities trade and
are therefore most exposed to rapid destruction. Again I might
mention one typical example. One of the most interesting astronom-
ical papyri eventually reached in part the Carlsberg collection
of the University of Copenhagen. H. O. Lange and I published
this text, which was written probably at the beginning of our era.
It contains a Demotic translation and commentary to a much
older hieratic text whose hieroglyphic replica is still preserved
in the Cenotaph of King Seti I (about 1300 B.C.). One of its
subjects is the description of the travel of the "decans u over the
body of the sky goddess, who was depicted on the ceilings of
tombs and temples as a representation of the vault of heavens
under which we live. Our papyrus was first seen in the posession
of an antiquities dealer in Cairo. At the time, the text still contained
at the beginning the picture of the sky goddess with all the
constellations and their dates of rising and setting. When the text
reached Copenhagen the picture was gone. No doubt it was sold
to some private collector and is probably lost forever. Thus a
vital part for the understanding of the text vanished almost at
the same moment its importance was recognized.
Needless to say, such happenings are contradictory to existing
antiquities laws. Also needless to say, many texts disappear from
excavations or are "found" by natives who have long learned
that papyri can be sold profitably instead of burning them at
their camp fires or using them as fertilizer. But it remains a
rather depressing fact that a large percentage of all these spoils
is destined to end unread and unpublished in climates less
favorable than the soil of Egypt.
30. But all that we have mentioned so far in source material
is child's play compared with ancient Mesopotamia. It is barely
a hundred years since cuneiform writing once more became
intelligible; and it was only shorlly before the beginning of the
present century that so fundamental a fact as Sumerian being
a language of its own though written with the same characters
as its later Semitic successors, Akkadian, became generally
recognized. But while decipherment and interpretation pro-
gressed in slow steps, texts were found in tremendous numbers
from the very beginning. The first collection of reliefs and tablets
 The Sources; Their Decipherment and Evaluation 59

arrived in France in 1846, having been excavated three years

before in the ruins of Khorsabad, near Mosul, by the French
consul Botta. In 1849/50 Layard found in the ruins of Nineveh,
then called Kuyunjik, the first palace library; in 1853 followed
the discovery of the Library of Ashurbanipal by Hormuzd Ras-
sam. Some 20,000 tablets in the British Museum now bear the
inventory letter K (for Kuyunjik) or Rm. (for Rassam). Perhaps
about a quarter of these two collections is published today.
This ratio between existing and published texts might seem
rather small. Actually it is unusually high and only due to the
fact that it is the result of a century of work on one of the most
famous discoveries in the Near East. In the meantime many
tens of thousands of tablets have found their way into museums,
providing source material which would require several centuries
for their publication even under the concentrated efforts of all
living Assyriologists.
91. At this point it may be useful to insert a few general
remarks about excavations and publication of texts because
little is known about these problems outside the small circle of
the initiated.
A modern excavation is a highly complex enterprise. A staff
of architects, draftsmen, photographers, epigraphers, and phil-
ologists have to assist the archeologist in his field work. But this
is only the fust and easier phase of an excavation. The pre-
servation of the ruin, the conservation of the objects found, and,
most of all, the publication of the results, remains the final task
I

for which the work in the field is only the preliminary step.
Here a sad story indeed must be told. While the field work
has been perfected to a very high standard during the last half
century, the second part, the publication, has been neglected
to such a degree that many excavations of Mesopotamian sites
resulted only in a scientifically executed destruction of what was
left still' undestroyed after a few thousand years. The reasons
for this fact are trivial. The time required for the publication of
results is a multiple of the requirements for the field work. The
available money is usually spent when a fraction of the original
planned excavation has been completed, benefactors are hard
to find to pay for many years of work without tangible or spec-
tacular results, and the scholars get interested in special aspects
 60 Chapter III

of the problem involved or go out for new material instead of

performing the tedious work of publishing the thousands of
details which the accidents of exavation have provided. The
final result is not much different from the one obtained by the
treasure-hunting attitude of the earliest excavators.
Many an excavation, if not all, had to be stopped before com-
pletion or had to restrict itself from the very beginning to a few
trenches crossing the ruin in the hope of getting a general insight
into the character of the stratification. Then one or another
promising building was investigated in greater detail. What
resulted is a ruin left with deep sears, an easy prey for the natives
to extract all exposed bricks, to tunnel for more without too many
difficulties, and to have access to deeper layers and thus to con-
tinue the excavation" in their own fashion and for their own
U

benefit. In this way the natives must have found thousands of

tablets which were then sold at high prices by antiquities dealers
to the very same museums which spent the initial money for
the removal of many tons of sand and debris.
Let me illustrate the effect upon the special studies under
discussion here. Until 1951 not for a single astronomical or
mathematical text was its provenance established by excavation.
. The only apparent exceptions are a number of multiplication
tables from Nippur or Sippar but nobody knows where these
texts were found in the ruins. Consequently it is, e. g., completely
impossible to find out whether these texts came from a temple,
a palace, a private house, etc. Not even the stratum is known to
give us a more accurate date of the texts. In other words, if those
texts, which were officially "excavated", would have been found
by Arabs, we would be no worse off than we are now. But while
the Arabs in their "clandestine" exploits dig only relatively small
holes, a scientific excavation has destroyed beyond any hope all
traces of the locality where the texts have been found. Thus we
are left with the texts alone and must deterntine their origin from
internal evidence, which is often very difficult to interpret.
A long story could be told about the "methods" to obtain the
needed information. Texts which for more than 50 years were
lying in the basement of a great museum could be relatively
dated from the newspaper in which they were wrapped. That
gave a plausible date for the "expedition" which found the texts
and hence the place from which they were excavated.
 The Sources; Their Decipherment and Evaluation 61

A whole class of texts was identified as follows. A German

expedition before 1914 had worked in the city of Uruk, a most
important site because it contains structures which reach from
the earliest periods down to Seleucid times. There the Germans
must have found the debris of an archive of which, however, all
the good tablets had been removed by the Arabs. These tablets
finally found their way into the collections of Berlin, Paris, and
Chicago, forming one of the most important groups of texts for
the study of Seleucid astronomy. The Arabs were not interested
in small fragments. These were left at the site and were then
carefully sifted and photographed by the expedition. By courtesy
of the Berlin Museum, I obtained prints of these photographs
(cf. PI. 6 b) showing the fragments neatly arranged on a desk of
the expedition. The records about the place where they were found
were lost in the meantime. The fragments themselves had also
been lost. By means of very extensive computation I succeeded,
however, in establishing the relationship of these splinters to
bigger pieces from the above-mention~dmuseums. Thus it became
possible to restore whole tablets, the parts of which are now on
different sides of the Atlantic. Finally, the small fragments
themselves were rediscovered in Istanbul. But the main question
of their accurate provenance remains unanswerable.
82. The Mesopotamian soil has preserved tablets for thousands
of years. This will not be the ease in our climate. Many tablets
are enc;rusted with salts (cf. PI. 9 a left and photograph which
shows fucrustations along the crack; the right-hand photograph
gives the same tablet after cleaning). A change in moisture pro-
duces crystals which break the surface of the tablets, thus rapidly
obliterating the writing. I have seen "tablets" which consisted
of dust only, carefully kept in showcases. To prevent this, tablets
must be slowly baked at high temperature and thereafter soaked
to remove salts. But only great museums possess the necessary
equipment and experienced staffs, not to mention the fact that
these methods of conservation were often kept as museum secrets.
Many thousands of tablets have been acquired at high cost by
big and small collections only to be destroyed without ever being
read or recorded in any way.
The publication of tablets is a difficult task in itself. First of
all, onf must find the texts which concern the specific field in
question. This is by no means trivial. Only minute fractions of
62 Chapter In
the holdings of collections are catalogued. And several of the
few existing rudimentary catalogues are carefully secluded from
any outside use. I would be surprised if a tenth of all tablets in
museums have ever been identified in any kind of catalogue.
The task of excavating the source material in museums is of
much greater urgency than the accumulation of new uncounted
thousands of texts on top of the never investigated previous
thousands. I have no official records of expenditures for ex-
cavations at my disposal, but figures mentioned in the press
show that a preliminary excavation in one season costs about as
much as the salary of an Assyriologist for 12 or 15 years. And
the result of every such dig is frequently many more tablets
than can be handled by one scholar in his lifetime.
There exists no simple method of publication. Photographs
alone are in the majority of cases not sufficient, even if their
cost were not prohibitive. Tablets are often inscribed not only on
both sides but also on the edges. Only multiple photographs
taken with variable directions of light would suffice. Thus cost
and actual need have resulted in the practice of hand copies.
Many different styles of copying were developed by individual
scholars, varying between an almost schematic reproduction of
the signs to a minute reproduction of details. The reader may get
an impression of this situation from PIs. 8 and 9 which show
an ephemeris for Saturn (Seleucid period) and an Old Babylo-
nian mathematical text in hand copies and photographs.
The ideal method of publication would be, of course, direct
copying from the text. In practice this is often excluded by the
scattering of directly related material all over the world. Even
with great experience a text cannot be correctly copied without
an understanding of its contents. Practically no text falls at the
first attempt. Thus repeated collation, joining with other frag-
ments, and comparison with other texts are needed. It requires
years of work before a small group of a few hundred tablets is
adequately published. And no publication is "final". Invariably
a fresh mind will find the solution of a puzzle which escaped
the editor, however obvious it might seem afterwards.
93. The process of decipherment follows no fixed rules. Every
special class of texts requires the slow construction of a technical
dictionary. The results of reading difficult signs and words must
be recorded in card files which allow repeated checking of and
 The Sources; Their Decipherment and Evaluation 63

comparison with previously obtained results. Only continued ex-

perience leads to a more rapid understanding of a certain type
of texts. Though it is impossible to describe this process ad-
equately the special situation which prevails in a specific class
of mathematical texts might illustrate the great advantages which
one has in reading mathematical texts of this type.
The text from which I shall discuss a few lines is reproduced
in PI. 8 a and PI. 9 a. At first sight it seems impossible to make
any sense out of its numbers, which show no relations which
could be explained as the result of consecutive operations. Hence
one has to abandon the idea of reading the text as a unit. This
is confirmed by the short note at the left lower corner of the
reverse. This "colophon" reads: 48 im-su dub-13-kam-ma. The
second part means "13th tablet" and characterizes the text as
a part of a series of at least 13 related tablets. The first part
must refer to the contents. Sometimes the number of lines is
indicated; for this 48 is too small a number. But one can easily
cheek that the total of small boxes of two or three lines of text
sepal ated from one another by horizontal lines amounts to about
40 or 50. This is confirmed by similar texts where the number of
im-su exactly agrees with the number of sections. Thus we know
already that each section has to be treated individually. Obviously
the shortness of these sections suggests that we are dealing with
problems only, not with their solutions worked out in detail. This
explains the lack of obvious connection between numbers.
Now we are ready to transcribe a few lines of the text, simply
following the general Assyriological rules for the transcription
and interpretation of signs.We ignore all difficulties of individual
readings but we indicate by [ ] a destroyed section. Then we
obtain for the lines from 12 to 17 in the left column of the reverse

...............
12 gar-gar us d[ ah ]5
13 a-ra 2 e-tab
-
14
-
us dah-ma 1
15
16 -
sag dah-ma 35
a-ra 2 e-tab
17 dah-ma 50
64: Chapter III

Several terms can be translated directly; gar and dah are words
known to indicate addition; a-ra is known from the m~ltiplication
tables, corresponding to our "times". The same word occurs,
e. g., in each line of the multiplication table of PI. 4 a. The words
ui and sag are known to mean length and width respectively.
Because no numbers are directly associated with them, we
transcribe them by x and y. The particle -ma represents some-
thing like "and thus"; we represent it simply by an equal sign.
The phrase e-tab seems to suggest another addition because tab
is the counterpart to lal "minus" which we know already (p. 5)
from the number sign 20 lal 1 = 19. In order not to complicate
our discussion unnecessarily we shall anticipate the result that
here the whole phrase a-ra 2 e-tab must mean "multiply by 2"
without any reference to addition. This is indeed in line with the
original meaning of the sign "tab" which consists of two parallel
wedges, thus indicating "duplication". We finally remark that we
invert for the sake of convenience the order of "x" and "add"
and write + x instead of x +. Then we obtain the following
"translation"•

12 ++ x[= )5
13 ·2
14 +x = 1
15 + y = 35
16 ·2
17 +( ) = 50
Weare now facing a new difficulty. At the beginning we tried
to read the text as a unit and found that we had to break it
up into single problems. Now we have the single problems but
they are obviously too short to make sense. Line 15, e. g., requires
that y is added to something and then gives the sum 35. And
exactly the same difficulty arises in the other examples. Thus
we are compelled to introduce an unknown quantity f, which
might depend on x and y, to which all the other quantities are
added. Thus we interpret line 15 as

f+ y = 35.
 The Sources; Their Decipherment and Evaluation 65

It is plausible to assume for the next example the interpretation

2 f+(y) = 50

because otherwise we would have two new unknowns instead of

f and y whereas our interpretation makes the second example
the direct continuation of its predecessor. But under this assump-
tion we can determine from these two equations the values of f
and y, and find f = 15 Y = 20. We can put this result to an
immediate test. Line 12 seems to indicate

f+
x[= ]5
2f+x=1

Obviously the second relation is impossible for positive numbers

because 2 f + x cannot be 1 if f + x already is at least 5. But
here the place value notation comes to our rescue. Instead of
"1" we can read 1,0 = 60. Thus we assume 2 f + x = 60 and
using f = 15 we obtain x = 30. Substituting this into the first
equation we obtain f + x = 45 in excellent agreement with the
traces in line 12. Thus we have reached as a first consequence
of our hypothetical interpretation that x = SO y = 20 f = 15.
Again we are able to test this result. Line 12 is the tail end of
a bigger section of 6 lines. Following essentially the same method
of decipherment we can translate two sentences as follows

n (2,40 - (x + y»
and
+(55 - y).

If we here substitute ~ = SO and y = 20 we find that the first

expression has the value 10, the second 5. The gar-gar "+" in
line 12 connects the two preceding expressions. Thus we find a
total 10+ 5 which is indeed the value 15 of f. Hence we have
not only confirmed our results but have also determined f as a
function of x and y:

f= n(2,40 - (x + y» + +(55 - y).

In other words we can now summarize all four problems as

followS:
66 Chapter In

(+x=45
2 (+ x = 1,0
(+ y = 85
2 (+ y = 50
where f stands for the above-mentioned expression. Obviously
these equations do not suffice, if one takes them singly, to deter-
mine x and y. On the other hand they cannot be used simulta-
neously because there are too many. Thus one has to look for
additional information higher up in the text. Applying exactly
the same procedure to preceding sections, one finds a simple
scheme. There exist several larger sections which define similar
functions g, h, etc., always followed by variations of the above
form 9 +x, 2g +x, etc. At the very beginning, however, we find
one more condition which turns out to mean
xy = 10,0.
This condition is common to the whole text as one relation
between x and y. All subsequent sections contain individual linear
relations between these unknowns, thus leading to quadratic
equations for x and y. We know already that x = 30 y = 20
are the common solutions. Thus our decipherment is completed.
What I have described here is, of course, a simplified story
of what actually happened when texts of this type became known,
but the essential s1eps were precisely the same. In this way it
was possible to establish many technical terms. The results can
be tested in other classes of texts which contain the details of
the working out of given problems. And it is clear that the deter-
mination of the meaning of a text is generally the easier the more
complicated the mathematical context is because this leaves
fewer possibilities for the interpretation of the procedure. A con-
text which contains only a few numbers combined by addition
or subtraction is of almost no help in the determination of ter-
minology. The advanced algebraic level of Babylonian mathe-
matics was of the greatest help in its being decoded.

NOTES TO CHAPTER III

Extensive bibliosraphical references for the edition of Greek mathematical
and astronomical works are given in R. C. Archibald, Outline of the History
of Mathematics, 6th edition, published as a supplement to the American Mathema-
tical Monthly 56 No.1 (1949).
 Notes 67
ad 27. The history of the zodiacal and planetary symbols is virtually unknown.
To my knowledge no study has been made which was based on the evidence
from manuscripts or epigraphic representation. No symbols appear in the
Greek papyri. It is a widespread but wrong belief that the Egyptian hieroglyph (!)
for the sun was used in ancient astronomical texts to denote the sun. The standard
symbol in the Middle Ages is d, but never (!) which is commonly used, however,
as the abbreviation for oveavo~ (heaven) or "t",lo~I) (circle). In the latter
usage it is shown in Fig. 3a (line 4) in a manuscript of the 10th or 11th century.
The solar symbol is directly above it (line 3).

The same holds, to my knowledge, for the papyri (cf. P. Warren 21 from the
early third century A.D. and Archiv f. Papyrusforschung 1, 1901, p. 501, a
papyrus of the second century A.D.; furthermore Karl Preisendanz, Die griechi-
sehen Zauberpapyri, Leipzig 1928-1931, vol. II, index p. 213). Ordinarily,
however, the names of sun, moon, and planets are written out in full. Demotic
planetary I tables of the Roman imperial period contain symbols for the planets
and for the zodiacal signs, apparently based on their Egyptian names and with
no relation to the mediaeval symbols.
That there is stm much left to be done even with the great classics might be
mustrated by the fact that Books V to VII of Apollonius's Conic Sections were
never edited because they are only preserved in Arabic. The only existing
(Latin) translation was made by Edmund Halley in 1710.
One of the many desiderata in the publication of mediaeval tables is the
"Alfonsine Tables" which were completed around 1270 under the auspices and
active support of Alfonso X (who ruled from 1252 to 1282). Of the Spanish
original only the introduction is preserved, but for the Latin versions Haskins
(Studies in the History of Medieval Science, p. 17) mentions 75 manuscripts
and 13 early editions. Cf. Alfred Wegener, Die astronomischen Werke
Alfons X. Bibliotheca Mathematica ser. 3 vol. 6 (1905) p. 129-185.
No aecurate estimate can be given about the quality and importance of
Byzantine astronomical handbooks but it seems evident from the most superficial
use of catalogues that this material must amount to many thousands of folios
of texts and tables. It is not known to what extent these treatises continue the
1) Also in composite words like en (f) for mlHtM~ (epicycle).
68 Chapter III

Ptolemaic-TheoDic tradition, are influenced by Islamic works, or have added

new ideas of their own.
The wealth of Islamic material is shown by E. S. Kennedy, "A Survey of
Islamic Astronomical Tables", Trans. Am. Philos. Soc. N.S. vol. 46,2 (1956)
p. 123-177. Of the more thaD a hundred works listed, only two had been pub-
lished (al-Battln! by NaDino 1899-1907 and al-Khwlrizmi by Bjernbo, Besthom,
and Suter 1914).
ad 28. The "CataIogus Codicum Astrologorum Graecorum" (Bruxelles,
Lamertin, 1898-1953) is not a "catalogue" in the strict sense. The first part of
each volume gives a description of the astrological manuscripts. The larger
second part contains editions of significant sections of these texts. The volumes
are arramged according to countries. Oriental libraries have not beeD utilized;
any amount of source material might be expected from the Near East.
For an important criticism of some results in Cumont's book see L. Robert.
Studes epigraphiques et philologiques, Bibliotheque de I'ecole des hautes
etudes 272 (1938) p. 72 ft.
The edition of Ptolemy's astrological writings has been completed by Lam-
mert and Boer in 1952 (Opera III, 2 Teubner).
An edition of Greek horoscopes with historical and astronomical commen-
taries by the present writer and H. B. Van Hoesen is to appear in the Memoirs
of the Am. Philos. Soc. (1957).
Gunders discoveries were published in "Neue astrologische Texte des
Hermes Trismegistos", Abh. d. Bayerischen Akad. d. Wissenschaflen, Philos.-
histor. AhL, N.F. 12 (1936). The results concerning the old star catalogue are
discussed on p. 131-134 and presented in the list of p. 148-153. Gundel de-
termined the date of the catalogue by comparing the given longitudes with the
longitudes of the star catalogue in the Almagest and assuming a change of 10
per century, acceptiDs Ptolemy's constant of precession. Though Gundel's main
result, namely, that a large number of the positions indicate the time of Hippar-
ehus, cannot be doubted, one can raise objections against details of the discussion.
With a single exception (No. 63), all longitudes in Gundel's list are integer
degrees (No. 63 gives 291) and are thus obviously rounded-oft values. Ptolemy's
catalogue, however, gives longitudes to an accuracy of 10'. Thus one cannot
compare these two sets of values without allowance for the roundiDg-oft errors.
Two possibilities must be considered; either rounding oft to the nearest integer
or simple cancellation of fractions. Experience with very large numbers of
rounded-oft values in Babylonian and Greek. astronomy suggests, contrary to
our habit, preference for the second method. Consequently I compared the
longitudes of all stars of Gundel's list with the longitudes which they had at
130 B.C. according to the catalogue of Peters and Knobel (Ptolemy's Catalogue
of Stars, Carnegie Institution ofWasbington, Publication No. 86,1915; p. 74 ft.).
A comparison is possible for 59 values from Gundel's list and the correspondiDg
values for 130 B.C. disregardiDg fractions, however close to one. For only one
star does the longitude appear to be 10 less than expected for 130 B.C.; and for
one other star it is 2 0 greater than expected. For 39 stars or 66 per cent one
obtains exactly the same numbers while 18 stars or 30.5 per cent have longitudes
10 greater than found for 130 B.C. In other words 96.5 per cent of the stars of
 Notes 69
Gundel's list have longitudes eorrect for the time from 130 B.C. to 60 B.C.
Thus they were taken either from mpparehus's catalogue itself or from the
eatalogue of an astronomer of the next generations.1) Gundel's hypothesis,
however, of a star catalogue whieh preceded Hipparchus and which gave the
positions in ecliptie coordinates is disproved.
The deeply rooted conviction that the Greeks were inclined towards philos-
ophieal speculation only, but neglected observations and experiments made it
an easily accepted theory to consider Ptolemy's star catalogue as a trivial
modifieation of Hipparchus's catalogue, simply assuming that Ptolemy added
2;40° to Hipparchus's longitudes, in spite of his explieit statement of independent
observations. In the meantime, Boll has shown (Bibliotheca Mathematica sere 3
vol 2, 1901, p. 185-195) that Hipparchus's catalogue covered only about 850
stars as compared with more than 1000 of Ptolemy·s. Finally, Vogt demonstrated
(Astron. Naehr. 224, 1925, ools. 17-54) that for 60 stars of Bipparehus's eat-
alogue, only 5 may have been utilized by Ptolemy whereas the majority un-
doubtedly show independent observations.
Gundel, who overlooked Vogt's paper, still operated under the assumption
of a purely schematic relation between the two catalogues. Of Hipparchus's
writings, only his Commentary to Aratus is preserved (edited by Manitius,
Leipzig 1894, with a German translation). This work is undoubtedly an early
work of Hipparehus, written before the discovery of precession. This follows
from the fact that the positions of stars are never given in ecliptie coordinates
(longitude and latitude) but in a mixed ecliptic-deelination system (cf. Fig. 30
p. 184). It was obviously the discovery of preeession that later led Hipparchus to
introduce real ecliptie coordinates because longitudes increase proportionally
with time whereas latitudes remain unchanged.
ad 29. H. O. Lange-O. Neugebauer, Papyrus Carlsberg I. Ein hieratisch-
demotischer kosmologiseher Text. Kgl. Danske Vidensk. Selskab, Bisl.-filol.
Skrifter 1, No.2 (1940). A new edition with many improvements in detail is
in preparation by R. A. Parker and the present author for inclusion in a larger
work on Egyptian astronomieal texts.
ad 31. Budge, The Rise and Progress of Assyriology, London 1925, writes
(p. 136 f.): "As soon as the dealers and offieials in Baghdad knew that Bassam
was out of the country they began to make excavations on their own aecount.
They employed the workmen who had been employed by Rassam .•. The
British Museum bought several collections, and as there was keen competition
in Paris and Ameriea prices began to soar, and in a short time ... tablets ...
for whieh the finders were paid five piastres each in Baghdad, were fetehing
£ 4 in London". This coneerns the period from 1882 to 1887.
The first reliable information about date and loeation of mathematieal
1) I suspect that for three consecutive stars of Gundel's list (Nos. 45 to 47) the
complete values are preserved in "Hermes, De XV stellls" published by Louis
Delatte, Textes latins et vieux fran~s relatifs aux Cyranides (Bibl. Fac. Philos.
et Lettres Univ. Li~e 93, 1942). There one finds for the first three stars the following
coordinates <yo 15;27 <yo 27;20 ij (Text <YO, out of orderl) 9;28 corresponding to Gundel's
er 15, <yo 27, ij 9. (Delatte p. 246, 249, 250 respectively.) The list "De XV stellis"
shares with Gundel's text the preference for the first half of the zodiac; t of the stars
of Delatte belong to this semicircle as compared with I in Gundel.
70 Chapter In

problem texts of the Old Babylonian period was given by Taba Baqir in Somer 7
(1951) p. 28 f. According to the field records of the recent excavations in Tell
~armal these texts come from a private house. The joint expedition of the
University Museum of the University of Pennsylvania and the Oriental Institute
of the University of Chicago to Nippur has now finally established that the
"Tablet Hm" which Hilpreeht, the original excavator, thought should represent
the "Temple Library", actually belongs to "residential quarters of varying
date" (D. E. McCown in J. Near Eastern Studies 11, 1952, p. 175).
ad 99. A more extensive description of the method of decipherment of
algebraic problem texts is given in my paper "Der VerhiUtnisbegrift' in der
babylonisehen Mathematik", Analecta Orientalia 12 (1935) p. 235-257.
 71

CHAPTER IV

Egyptian Mathematics and Astronomy.

34. Of all the civilizations of antiquity. the Egyptian seems
to me to have been the most pleasant. The excellent protection
which desert and sea provide for the Nile valley prevented the
excessive development of the spirit of heroism which must often
have made life in Greece hell on earth. There is probably no
other country in the ancient world where cultivated life could be
maintained through so many centuries in peace and security. Of
course not even Egypt was spared from severe outside and interior
struggles; but, by and large, peace in Mesopotamia or Greec~
must have been as exceptional a state as war in Egypt.
It is not surprising that the static character of Egyptian culture
has often been emphasized. Actually there was as little innate
conservatism in Egypt as in any other human society. A serious
student of Egyptian language, art, religion, administration, etc.
can clearly distinguish continuous change in all aspects of life
from the early dynastic periods until the time when Egypt lost
its independence and eventually became submerged in the
Hellenistic world.
The validity of this statement should not be contested by refer-
ence to the fact that mathematics and astronomy played a uni-
formly insignificant role in all periods of Egyptian history. Other-
wise one should deny the development of art and architecture
during the Middle Ages on the basis of the invariably low level
of the sciences in Western Europe. One must simply realize that
mathematics and astronomy had practically no effect on the
realities of life in the ancient civilizations. The mathematical
requirements for even the most developed economic structures
of antiquity can be satisfied with elementary household arithmetic
72 Chapter IV

which no mathematician would call mathematics. On the other

hand the requirements for the applicability of mathematics to
problems of engineering are such that ancient mathematics fell
far short of any practical application. Astronomy on the other
hand had a much deeper effect on the philosophical attitude of the
ancients in so far as it influenced their picture of the world in
which we live. But one should not forget that to a large extent
the development of ancient astronomy was relegated to the status
of an auxiliary tool when the theoretical aspects of astronomical
lore were eventually dominated by their astrological interpreta-
tion. The only practical applications of theoretical astronomy
may be found in the theory of sun dials and of mathematical
geography. There is no trace of any use of spherical astronomy
for a theory of navigation. It is only since the Renaissance that
the practical aspects of mathematical discoveries and the theor-
etical consequences of astronomical theory have become a vital
component in human life.
35. The fact that Egyptian mathematics did not contribute
positively to the development of mathematical knowledge does
not imply that it is of no interest to the historian. On the contrary,
the fact that Egyptian mathematics has pl·eserved a relatively
primitive level makes it possible to investigate a stage of develop-
ment which is no longer available in so simple a form, except
in the Egyptian documents.
To some extent Egyptian mathematics has had some, though
rather negative, influence on later periods. Its arithmetic was
widely based on the use of unit fractions, a practice which probably
influenced the Hellenistic and Roman administrative offices and
thus spread further into other regions of the Roman empire,
though similar methods were probably developed more or less
independently in other regions. The handling of unit fractions was
certainly taught wherever mathematics was included in a curri-
culum. The influence of this practice is visible even in works of the
stature of the Almagest, where final results are often expressed
with unit fractions in spite of the fact that the computations
themselves were carried out with sexagesimal fractions. Sometimes
the accuracy of the results is sacrificed in favor of a nicer appear-
ance in the form of unit fractions. And this old tradition doubtless
contributed much to restricting the sexagesimal place value
notation to a purcly scientific use.
 Egyptian Mathematics and Astronomy 73
36. There are two major results which we obtain from the study
of Egyptian mathematics. The first consists in the establishment
of the fact that the whole procedure of Egyptian mathematics is
essentially additive. The second result concerns a deeper insight
into the development of computation with fractions. We shall
discuss both points separately.
What we mean by the "additivity" of Egyptian mathematics
can easily be explained. For ordinary additions and subtractions
nothing needs to be said. It simply consists in the proper collec-
tion and counting of the marks for units, tens, hundreds, etc.,
of which Egyptian number signs are composed. But also multipli-
cation and division are reduced to the same process by breaking
up any higher multiple into a sum of consecutive duplications.
And each duplication is nothing but the addition of a number to
itself. Thus a multiplication by 16 is carried out by means of
four consecutive duplications, where only the last partial result
is utilized. A multiplication by 18 would add the results for 2
and for 16 as shown in the following example

1 25
I 2 50
4 100
8 200
I 16 400
total 450

In general, multiplication is performed by breaking up one factor

into a series of duplications. It certainly never entered the minds
of the Egyptians to ask whether this process will always work.
Fortunately it does; and it is amusing to see that modern com-
puting machines have again made use of this "dyadic" principle
of multiplication. Division is, of course, also reducible to the same
method because one merely asks for a factor which is needed
for one given number in order to obtain the second given number.
The division of 18 by 3 would simply mean to double 3 until
the total 18 can be reached
1 3
I 2 6
I 4 12
total 18,
74 Chapter IV

and the result is 2 +

4 = 6. Of course, this process might not
always work so simply and fractions must be introduced. To
divide 16 by 3 one would begin again with
I 1 3
2 6
I 4 12
+
and thus find 1 4 = 5 as slightly below the requested solution.
What is still missing is obviously 16 - 15 = 1, and to this end
the Egyptian computer would state
"""3 2
181

which means that f of 8 is 2, 1 of 3 is 1 and thus he would find

5 "3 as the solution of his problem.
Here we have already entered the second problem, operations
with fractions. As we have said in Chapter I, Egyptian fractions
are always "unit fractions", with the sole exception of f which
we always include under this name in order to avoid clumsiness
of expressionl ). The majority of these numbers are written by
means of the ordinary number signs below the hieroglyph c:=>
"r" meaning something like "part". We write therefore '5 for the
expression "5th part"' = i. For t we write i whereas the Egyptian
form would be "2 parts" meaning "2 parts out of 3", i. e., f.
There exist special signs for 1 and 1 which we could properly
represent by writing "half" and "quarter" but for the sake of
simplicity we use 2 and 4 as for all other unit fractions.
We shall not go into the details of the Egyptian procedures for
handling these fractions. But a few of the main features must
be described in order to characterize this peculiar level of arith-
metic. If, e. g., 3 and 15 should be added, one would simply
leave 3 15 as the result and never replace it by any symbol like
I. Again 3 forms an exception in so far as the equivalence of 2 6
and 3 is often utilized.
Every multiplication and division which involves fractions
leads to the problem of how to double unit fractions. Here we
1) We disregard here another "complementary fraction" (that is. a fraction of
the form 1 - Ii) which is written with a special sign. namely I. because it plays no
role in the Egyptian arithmetical proeedures known to us.
 Egyptian Mathematics and Astronomy 75
find that twice 2, 4, 6, 8, etc. are always directly replaced by
1, 2, 3, 4, respectively. For twice 3 one has the special symbol 3.
For the doubling of 5, 7, 9, ... however, special rules are follow-
ed which are explicitly summarized in one of our main sources,
the mathematical Papyrus Rhind. One can represent these rules
in the form of a table which gives for every odd integer n the
expression for twice 'ii.
This table has often been reproduced and we may restrict
ourselves to a few lines at the beginning:

n twice Ii
3 2 6
5 3 15
7 4 28
9 6 18
etc.

The question arises why just these combinations were chosen

among the infinitely many possibilities of representing 2/n as the
sum of unit fractions.
I think the key to the solution of this problem lies in the
separation of all unit fractions into two classes, "natural" frac-
tions and "algorithmic" fractions, combined with the previously
described technique of consecutive doubling and its counter-
part, consistent halving. As "natural" fractions I consider the
small group of fractional parts which are singled out by special
signs or special expressions from the very beginning, like 3, 3, 2'
and 4. These parts are individual units which are considered
basic concepts on an equal level with the integers. They occur
everywhere in daily life, in counting and measuring. The remain-
ing fractions, however, are the unavoidable consequence of
numerical operations, of an "algorism", but less deeply rooted in
the elementary concept of numerical entities. Nevertheless there
are "algorithmic" fractions which easily present themselves,
namely, those parts which originate from consistent halving. This
process is the simple analogue to consistent duplication upon
which all operations with integers are built. Thus we obtain two
series of fractions, both directly derived from the "natural"
fractions by consecutive halving. One sequence is 3, 3, 6, 12, etc.,
76 Chapter IV

the other 2, 4, 8, 16 etc. The importance of these two series is

apparent everywhere in Egyptian arithmetic. A drastic example
has already been quoted above on p. 74 where we found that
3 of 3 was found by stating first that 3 of 3 is 2 and only as a
second step 3 of 3 is 1. This arrangement 3 -+ 3 is standard even
if it seems perfectly absurd to us. It emphasizes the completeness
of the first sequence and its origin from the Unatural" fraction 3.
If one now wishes to express twice a unit fraction, say 5, as
a combination of other fractional parts, then it seems natural
again to have recourse to these two main sequences of fractions.
Thus one tries to represent twice 5 as the sum of a natural fraction
of '5 and some other fraction which must be found in one way or
another. At this early stage, some trials were doubtless made
until the proper solution was found. I think one may reconstruct
the essential steps as follows. We operate with the natural frac-
tion 3, after other experiments (e. g., with 2) have failed. Two
times 5 may thus be represented as 8 of 5 or 15 plus a remainder
=
which must complete the factor 2 and which is 1 3. The question
of finding 1 =3 of -5 now arises. This is done in Egyptian mathe-
matics by counting the thirds and writing their number in red
ink below the higher units, in our ease
1 3 (written in black)
3 2 (written in red).

This means that 1 contains 3 thirds and 3 two thirds. Thus the
remaining factor contains a total of 5 thirds. This is the amount
of which 5 has to be taken. But 5 fifths are one complete unit and
this was a third of the original higher unit. Thus we obtain for
the second part simply 3 and thus twice 5 is represented as 3 15.
This is exactly what we find in the table.
For the modern reader it is more convenient to repeat these
clumsy conclusions with modem symbols though we must re-
member that this form of expression is totally unhistorical. In

order to represent f in the form of .!. + !.x we

m
chose ..!:.. as a
m
natural fraction of 1, in this ease i· i = ra. For the remaining
fraction we have
 Egyptian Mathematics and Astronomy 77

Thus we have the representation

211
-=-+-
5 15 3
of the table. In general we have
21151
-=-.-+-.-
n 3 n 3 n
and the second term on the right-hand side will be a unit fraction
when and only when n is a multiple of 5. In other words a trial
with the natural fraction 1 will work only if n is a multiple of 5.
This is indeed confll"med in all eases available in the table of the
Papyrus Rhind which covers all expressions for
n
!
from n = 3
to n = 101.
We may operate similarly with the natural fraction 1. Then
we have
21131
-=---+-.-
n 2 n 2 n
which shows that we obtain a unit fraction on the right-hand
side if n is divisible by 3. For n = 3 we obtain

211
3 = 6 +2
and this is the relation 3 = 2 i which we quoted at the beginning.
All other eases in the table for ii's which are multiples of 3 show
.•
the same d ecomposltion ··t
operating WIh - 1 as one t erm.
2n
It is clear that one can proceed in the same manner by operating
with 4, 8, etc. or with 6, 12, etc. In this way, more and more
eases of the table can be reached and it seems to me there is
little doubt that we have found in essence the procedure which
has led to these rules for the replacement of 2 Ii by sums of unit
fractions.
87. For our present purposes it is not necessary to discuss in
detail all steps in the structure of Egyptian fractional arithmetic. I
hope, however, to have made clear the two leading principles, the
strict additivity and the extensive use of the "natural fractions".
78 Chapter IV

A few historical remarks must be added. The Papyrus Rhind

is not our only document for the study of Egyptian arithmetic.
The other large text, the Moscow papyrus, agrees with rules
known from the Papyrus Rhind. We have, however, an ostracon
from the early part of the New Kingdom where the duplication
of '7 is given as 6 14 21 instead of 4 28 of the standard rule.
Much more material is available from Demotic and Greek papyri
of the Hellenistic period. Here again, deviations from the earlier
rules can be observed, though the main principle remains the
same. In other words we cannot assume that once and forever
a system of fractional tables was computed and then rigidly
maintained. Obviously several equivalent forms were slowly
developed but without ever seriously transgressing the original
methods. This latter fact is of great historical importance. The
handling of fractions always remained a special art in Egyptian
arithmetic. Though experience teaches one very soon to operate
quite rapidly within this framework, one will readily agree that
the methods exclude any extensive astronomical computations
comparable to the enormous numerical work which one finds
incorporated in Greek and late Babylonian astronomy. No wonder
that Egyptian astronomy played no role whatsoever in the devel-
opment of this field.
98. It would be quite out of proportion to describe Egyptian
geometry here at length. It suffices to say that we find in Egypt
about the same elementary level we observed in contemporary
Mesopotamia. The areas of triangles, trapezoids, rectangles, etc.
are computed, and for the circle a rule is used which we can
transcribe as A = (~dr if ddenotes the diameter. Correspond-
ing formulae for the elementary volumes were known, including
a correct numerical computation for the volume of a truncated
pyramid. This, as well as the relatively accurate value 3.16 for n
resulting frOD1 the above formula, give Egyptian geometry a lead
over the corresponding arithmetical achievements. It has even
been claimed that the area of a hemisphere was correctly found
in an example of the Moscow papyrus, but the text admits also
of a much more primitive interpretation which is preferable.
A vivid description of the main topics of Egyptian mathematics
is given in a papyrus of the New Kingdom, written for school
 Egyptian Mathematics and Astronomy 79

purposes. It is a satirical letter in which an official ridicules a

colleague. The section on mathematics runs as follows l ): "Another
topic. Behold, you come and fill me with your office. I will cause
you to know how matters stand with you, when you say 'I am
the scribe who issues commands to the army'.
"You are given a lake to dig. You come to me to inquire con-
cerning the rations for the soldiers, and you say 'reckon it out'.
You are deserting your office, and the task of teaching you to
perform it falls on my shoulders.
"Come, that I may tell you more than you have said: I cause
you to be abashed (?) when I disclose to you a command of
your lord, you, who are his Royal Scribe, when you are led
beneath the window (of the palace, where the king issues orders)
in respect of any goodly (?) work, when the mountains are
disgorging great monuments for Horus (the king), the lord of the
Two Lands (Upper and Lower Egypt). For see, you are the clever
scribe who is at the head of the troops. A (building-) ramp is to
be constructed, 730 cubits long, 55 cubits wide, containing 120
compartments, and filled with reeds and beams; 60 cubits high
at its summit, 30 cubits in the middle, with a batter of twice 15
cubits and its pavement 5 cubits 2). The quantity of bricks needed
for it is asked of the generals, and the scribes are all asked to-
gether, without one of them knowing anything. They all put
their trust in you and say, 'You are the clever scribe, my friend!
Decide for us quickly! Behold your name is famous; let none
be found in this place to magnify the other thirtyS) I Do not let
it be said of you that there are things which even you do not
know. Answer us how many bricks are needed for it?
"See, its measurements (?) are before you. Each one of its
compartments is 30 cubits and is 7 cubits broad."
On the whole, one can repeat here what we have already said
for Babylonian geometry. Problems concerning areas or volumes
do not constitute an independent field of mathematical research
but are only one of many applications of numerical methods to
practical problems. There is no essential difference between the
determination of the acreage of a field in special measures and
1) From Erman. Egyptian Literature. p. 223 f.
I) These explanations are due to L. Borchardt. Cf. for a figure Quellen und
Studien zur Geschichte der Mathematik. sere B. vol. 1. p. 442.
8) Perhaps the frequently mentioned ··College or the Thirty".
 80 Chapter IV

the distribution of beer to temple personnel according to different

ratings. This is a state of affairs which holds to a large extent
even in the Hellenistic period and far beyond it. In Arabic
mathematics the "inheritance" problems play an important role,
while similar examples are found already in Old-Babylonian
texts. The geometrical writings of Heron, whether authentic or
merely ascribed to him, contain whole chapters on units, weights,
measurements, etc. Of course, since the Hellenistic period, even
the writings of Heron and related documents show the influence
of scientific Greek geometry. But, by and large, one has to
distinguish two widely separate types of "Greek" mathematics.
One is represented by the strictly logical approach of Euclid,
Archimedes, Apollonius, etc.; the other group is only a part of
general Hellenistic mathematics, the roots of which lie in the
Babylonian and Egyptian procedures. The writings of Heron and
Diophantus and works known only from fragments or from
papyrus documents form part of this oriental tradition which
can be followed into the Middle Ages both in the Arabic and in
the western world. "Geometry" in the modern sense of this word
owes very little to the modest amount of basic geometrical knowl-
edge which was needed to satisfy practical ends. Mathematical
geometry got one of its most important stimuli from the discovery
of irrational numbers in the 4th or 5th century B.C. and remained
rather stagnant from the second century B.C. onwards, except
for those additions of spherical geometry and descriptive geometry
which were introduced by their astronomical importance. On
the other hand, geometrial theory had a negative effect on the
algebraic and numerical methods which were part of the Oriental
background of Hellenistic science. A real insight into the mutual
relations between all these fields was not reached before modem
times.
39. The role of Egyptian mathematics is probably best described
as a retarding force upon numerical procedures. Egyptian astron-
omy had much less influence on the outside world for the very
simple reason that it remained through all its history on an ex-
ceedingly crude level which had practically no relations to the
rapidly growing mathematical astronomy of the Hellenistic age.
/
Only in one point does the Egyptian tradition show a very
beneficial influence, that is, in the use of the Egyptian calendar
 Egyptian Mathematics and Astronomy 81

by the Hellenistic astronomers. This calendar is, indeed, the only

intelligent calendar which ever existed in human history. A
year consists of 12 months of 30 days each and 6 additional days
at the end of each year. Though this calendar originated on purely
practical grounds, with no relation to astronomical problems, its
value for astronomical calculations was fully recognized by the
Hellenistic astronomers. Indeed a fixed time scale without any
intercalations whatsoever was exactly what was needed for
astronomical calculations. The strictly lunar calendar of the
Babylonians, with its dependence on all the complicated varia-
tions of the lunar motion, as well as the chaotic Greek calendars,
depending not only on the moon but also on local politics for its
intercalations, were obviously far inferior to the invariable
Egyptian calendar. It is a serious problem to determine the number
of days between two given Babylonian or Greek new year's days,
say 60 years apart. In Egypt this interval is simply 50 times
365. No wonder that the Egyptian calendar became the standard
astronomical system of reference which was kept alive through
the Middle Ages and was still used by Copernicus in his lunar
and planetary tables. Even in a civil calendar the Egyptian year
of 366 days was revived during the Middle Ages. The last Sasanian
king, Yazdigerd, based the reformed Persian calendar on this
year, shortly before the collapse of the Sasanian monarchy under
the impact of expanding Islam. Nevertheless the "Persian" years
of the Era Yazdigerd (beginning A.D. 632) survived and are often
referred to in Islamic and Byzantine astronomical treatises.
A second Egyptian contribution to astronomy is the division of
the day into 24 hours, though these "hours" were originally not
of even length but were dependent on the seasons. These •'seasonal
hours", twelve for daylight, twelve for night, were replaced by
"equinoctial hours" of constant length only in theoretical works
of Hellenistic astronomy. Since at this period all astronomical
computations were carried out in the sexagesimal system, at least
as far as fractions are concerned, the equinoctial hours were
divided sexagesimally. Thus our present division of the day into
24 hours of 60 minutes each is the result of a Hellenistic modifica-
tion of an Egyptian practice combined with Babylonian numerical
procedures.
Finally, we have to mention the "dccans" (to use a Greek term)
82 Chapter IV

which have left no direct traces in modern astronomy. This is

curious enough since the decans, as we shall see, are the actual
reason for the 12-division of the night and hence, in the last
analysis, of the 24-hour system. Again in Hellenistic times the
Egyptian decans were brought into a fixed relation to the Babylo-
nian zodiac which is attested in Egypt only since the reign of
Alexander's successors. In this final version the 36 udecans" are
simply the thirds of the zodiacal signs, each decan representing
10 0 of the ecliptic. Since the same period witnesses the rapid
development of astrology, the decans assumed an important
position in astrological lore and in kindred fields such as alchemy,
the magic of stones and plants and their use in medicine. In this
disguise the decans reached India, only to be returned in still
more fantastic form to the Muslims and the West. Their final
triumph lies in the frescoes of the Palazzo Schifanoria in Ferrara
under Borso d'Este (about 1460).
In tracing back the history of the Egyptian decans we discover
the interaction of the two main components of Egyptian time
reckoning: the rising of Sirius as the harbinger of the inundation,
and the simple scheme of the civil year of 12 months of three
decades each.
99 Q. Here is not the place to attempt a description of the history
of the Egyptian calendar. Its basically non-astronomical character
is underlined by the fact that the year is divided into three seasons
of four months each, of purely agricultural significance. The only
apparent astronomical concept is the heliacal rising of Sirius
which, however, obtained its importance only by its closeness
to the rising of the Nile, the main event in the life of Egypt. There
existed, finally, a lunar calendar which regulated festivals in
relation to the phases of the moon. As a matter of fact, as R. A.
Parker has observed, there are different variants of lunar calendars
to be distinguished, one of which was also eventually schematized
and brought into a fixed relation to the schematic civil calendar
with its twelve 30-day months and 5 epagomenal days.
When the decans first appear-on coffin lids of the Middle
Kingdom-the civil calendar had long been established. To it is
now set in relation a series of constellations, 36 in number, though
with small variations in arrangement and limitations. Only two
 Egyptian Mathematics and Astronomy 83

can be directly identified, namely Sirius and Orion. Some con-

stellations cover more than one decan; on the other hand there
are decans "preceding" or "following" a constellation, indicating
groups of stars of lesser significance. We shall see that all these
decans belong to a zone of the sky roughly parallel to and south
of the ecliptic (cf. Fig. 3 b p. 87).
The astronomical representations on the coffin lids are, in all
probability, poor replicas of ceilings in royal tombs or temples
which were imitated in the modest coffins of minor people. These
pictures represent the sky with the decanal constellations in-
scribed on them, arranged in their ten-day intervals throughout
the year, forming 36 columns with 12 lines each for the 12 hours
of the night. The name of a specific decan moves from column to
column, each time one line higher. Thus there originated a
diagonal pattern which is the motivation for the name "diagonal
calendars" for these texts.
In fact, we have here not a calendar but a star clock. The user
of this list would know the "hour" of night by the rising of the
decan which is listed in the proper decade of the month. We shall
now proceed to investigate the working of this type of "clock"
more closely, at first from a modern point of view, then turning
to historical considerations.
When we watch the stars rise over the eastern horizon, we see
them appear night after night at the same spot on the horizon.
But when we extend our observation into the period of twilight,
fewer and fewer stars will be recognizable when they cross the
horizon, and near sunrise all stars will have faded out altogether.
Let us suppose that a certain star S was seen just rising at the
beginning of dawn but vanished from sight within a very short
time because of the rapid approach of daylight. We call this
phenomenon the "heliacal rising" of S, using a term of Greek
astronomy. Let us assume that we use this phenomenon as the
indication of the end of "night" (meaning real darkness) and
consider S as the star of "the last hour of night". One day later
we may again say that the brief appearance of S indicates the end
of night. We may continue in the same way for several days, but
during this time a definite change takes place. The sun not only
participates in the daily rotation of the sky from East to West,
84 Chapter IV

thus producing the change of night and daylight, but it also has a
slow motion of its own relative to the stars in a direction opposite
to the daily rotation.
This eastward motion of the sun (completed once in one year)
delays the rising of the sun from day to day with respect to the
rising of S. Consequently, the rising of S will be more and more
clearly visible and it will take more and more time before S fades
away in the light of the coming day. Obviously, after some lapse
of time, it no longer makes sense to take S as the indicator of the
last hour of night. But there are new stars which can take the
place of S, and this procedure can be repeated all year long until
the sun comes back to the region of S. Thus year after year S
may serve for some days as the star of the last hour, to be replaced
in regular order by other stars T, U, V, ....
It is this sequence of phenomena which led the Egyptians to
measure the time of night by means of stars (or groups of nearby
stars) which we now call the decans. In the above description,
we left unanswered the tacit question: how long shall we wait
until we replace S by T, T by U, etc.? Obviously, one could be
very strict and choose daily a different star which is just in the
phase of "heliacal rising". But this sort of impractical pedantry
was not characteristic of those Egyptians, who intended to devise
some method of indicating the times of office for the nightly
service in the temples. They adjusted these times to their calendar.
As the months were divided into decades, so were the services of
the hour-stars. For 10 days, S indicated the last hour of night, then
T was chosen for the next 10 days, and so forth. During each
decade the end of night receded from dawn toward darkness, only
to be pushed back toward dawn by the heliacal rising of the next
"decantl as we shall now call the stars S, T, U, ....
So far we have only described the definition of the end of
"nighttl or the last "hour". We have made one choice: we applied
the decimal order of the civil calendar to these decanal hours.
What follows is a necessary consequence of this vital decision.
We go back to the time of year when S serves as the decan of
the last hour. Ten days later, T takes the place of S. By this time
the rising of S is clearly visible in full darkness. Since the last
hour is now indicated by T, we shall naturally say that the rising
of S marks the hour next to the last. After another ten days, U
 Egyptian Mathematics and Astronomy 85

represents the last hour, T the preceding hour, S the second-last

hour, etc. Writing from right to left as the Egyptian texts do, we
have thus obtained the following "diagonal" arrangement:

decade decade decade

8 2 1

S second to last hour

T S next to last hour
U T S last hour

How long can this process be continued? To simplify matters,

let us at first assume that a year had exactly 360 days or 36 decades.
Then we need 36 decans before S can serve again as decan of
the last hour. Our "star clock" will therefore be composed of
36 columns. The number of lines is a consequence of the following
consideration. The rising of stars can only be seen at night. The
maximum number of "hours" indicated by our decans is therefore
equal to the number of decans which we can see rise in succession
during one night. If we had complete darkness from the moment
of sunset to the moment of sunrise, and if night and day were
equal all year, one could always see exactly one-half of the
celestial sphere rise during one night. Since 36 decans correspond
to one complete circuit of the sky, 18 decans would be seen rising
each night and our list of stars would lead to an 18-division of
the night. In reality, however, the variation in the length of night
as well as twilight inftuences this number considerably. A closer
investigation shows that during summer, when Sirius rises hella-
cally, only 12 decans can be seen rising during darkness. Hence
the decadic succession of the decans leads to a 12-division of the
night. This, indeed, is the arrangement we find in the "diagonal
calendars" on coffin lids of the period from about 1800 to 1200
B.C.
It is essential to recall that it was the decimal arrangement of
the calendar which determined the spacing of the decans and thus
the number of hours to be indicated by their rising each night.
A finer division would have led to a greater number of hours,
while longer intervals would have given fewer hours. The 12-
division is therefore not an arbitrary choice of units, but the
86 Chapter IV

consequence of the decimal order of the civil calendar. The

decimal basis of time reckoning appears in another form in
the division of daylight. One of the inscriptions of the cenotaph
of Seti I (about 1300 B.C.) shows a simple sun-dial and gives a
description of its use. From this it follows that this instrument
indicated ten •'hours tt between sunrise and sunset. To this, two more
hours are added for morning and evening twilight respectively.
Thus we see that the Egyptian reckoning of hours was orig-
inally decimal for daylight, duodecimal for the time of darkness
because of the decimal structure of the calendar, and leaving two
more "hours" for twilight. The result is 24 "hours" of rather
uneven length and uneven distribution between daylight and
night. We do not know the details of the further development, but
it can be shown that this primitive system was already obsolete
when it was still depicted on the inscriptions of Seti I, giving way
to a more even distribution of 24 hours into 12 hours of night and
daylight each-a division which eventually led to the 24 "seasonal"
hours of the Hellenistic period.
39b. Not only must the independent division of darkness and
daylight have soon become obsolete, but also the decanal clocks
for the hours of night were bound to lose their usefulness in the
course of a century or two. In our description, we assumed for the
sake of simplicity a year of exactly 860 days' length. In this ease,
36 decans would repeat their service periodically, following the
diagonal pattern as described before. Actually, however, the
Egyptian civil year contained 365 days. Since the 36 decans
suffice only for the 360 days, an additional set of constellations
is required to indicate the hours of darkness for the epagomenal
days. All this was, in fact, taken into account by the inventors of
the decanal hours, as can be demonstrated by the terminal section
of the "diagonal calendars" on the coffin lids. What was not
taken into consideration, however, is the fact that 365 days do not
accurately measure the return of the sun to the same star, and
consequently, a slow but relentless change in the relation between
the heliacal rising of a decan and its date in the civil calendar
takes place. Our texts show that rearrangements of the decanal
order were attempted in order to counter the resulting disturbances.
By the time of the New Kingdom, the usefulness of the decans as
indicators of hours had ceased. An attempt to substitute the
 Egyptian Mathematics and Astronomy 87
culmination of stars for their rising also did not last. But by this
time the decans held a secure position as representatives of the
decades of the year in the decoration of astronomical ceilings,
as in the tomb of Senmut or in the cenotaph of Seti I. In this form
they continued to exist until their association with the zodiac of
the Hellenistic period revived them and made them powerful
elements of astrological doctrine.
--------------------------+60

---------------------------60
Fig.3b.

We still have to answer the question of the location in the sky

of the decans when they were first invented as the indicators of
the hours of night. From what has been said to this point, any
sequence of stars or constellations whose risings were spaced at
ten-day intervals could have been used. But additional information
is available. We not only know that Sirius and Orion figured
among the decans but that Sirius was, so to speak, the ideal
prototype of all the other decans. Its heliacal rising ideally begins
the year, just as the rising of the other decans are associated with
the beginning of the parts of the year, the decades. The rising of
Sirius occurs after an interval of about 70 days, in which the
star remains invisible because of its closeness to the sun. Similarly,
it was assumed that the same holds for all decans. The Demotic
commentary to the inscriptions in the cenotaph of Seti I describes
at length how one decan after another "dies", how it is "purified"
in the embalming house of the nether world, to be reborn after
70 days of invisibility.
Such a mythological description cannot, of course, be taken as
an exact astronomical condition for the duration of invisibility.
But there can be little doubt that the decans were essentially to
follow the cycle of Sirius. In other words, to serve as decan
88 Chapter IV

during the decade immediately after Sirius, a star must have

been chosen that not only rose ten days later, but that also had a
period of 70 days' invisibility. If these data were accurate and if
the brightness of the stars in question were known, their positions
could be determined accurately. This is not the case. Nevertheless,
the deviation from a 70-day invisibility as well as the variation in
brightness may be assumed to have remained within reasonable
limits. This suffices to localize at least a zone within which
constellations, which can serve as decans, must be located. The
result is shown in Fig. 3b1) which represents the belt of the decans
in its relation to the ecliptic and equator and to Sirius and Orion.
To attempt to go further in the determination of the decans is not
only of very little interest but would necessarily imply ascribing
to our texts an astronomical accuracy which they were never
intended to have. But we have reached insight into a sound,
however primitive, procedure of marking time at night by means
of stars and are able to localize them in a definite region of the
sky to which Sirius and Orion belong, not as exceptions, but as
the leading members of the decanal constellations.
40. The coffins with the "diagonal calendars" belong roughly to
the period from 2100 B.C. to 1800 B.C. From the New Kingdom,
more elaborate monuments are preserved. One is the ceiling of the
unfinished tomb of Senmut, the Vezir of Queen Hatshepsut;
another is the ceiling of the cenotaph of King Seti I, about 1300
B.C. The tomb of Senmut contains lists of the decans, the represen-
tation of the deities of the hours, etc., and pictures of constellations
of the northern hemisphere. On PI. 10 part of this section is
shown as copied by the expedition of the Metropolitan Museum
in New York. The Hippopotamus and the Crocodiles etc. appear
often again on similar pictures. Of special interest in the present
drawing is the fact that it shows two stages of work. Below the
sharply outlined drawings are visible faint lines which were
drawn in blue on the ceiling. The man whose one arm is near a
crocodile was missing on the original draft. The crocodile was
not drawn in a skew position buf horizontally. The traces are still
visible at the feet and legs of the standing man. The lion, which

1) Fig. 3b can be considered as an unrolling (or cylinder projection) of the

celestial sphere between declinations ± 60 0 (or fJ = 300) with the equator as line
of symmetry.
 Egyptian Mathematics and Astronomy 89
is now one register higher up, was lying parallel to the crocodile.
The base line is still visible crossing the front shoulder of the skew
crocodile. The head of the lion was at the man's shoulder, the
lion's front paws are to the right of the man's belt. The previous
inscription occupied the place of the final lion. These details are
of interest because they demonstrate drastically that artistic
principles determined the arrangement of astronomical ceiling
decorations. Thus it is a hopeless task to try to find, on the sky,
groups of stars whose arrangement might have been the same
as the depicted constellations seem to require. Astronomical
accuracy was nowhere seriously attempted in these documents.
In the tombs of Ramses VI, VII, and IX a new type of astronom-
ical text appears. Here we find depicted observations which were
made to determine the hours of the night throughout the year.
For the first and for the 16th day of each month we see pictured
a sitting man (PI. 11) and above (or, better, behind) him a
coordinate net with stars. The accompanying inscription men-
tions, for the beginning of the night and for each of the 12 hours
of the night, a star and where it will be seen ..over the left ear",
"over the right ear", "over the left shoulder", or the "right
shoulder", etc.
The horizontal lines in the coordinate network represent the
hours, the vertical lines the positions. The stars are entered as
named in the text-at least in principle, except for the innumer-
able errors which the craftsmen committed. Obviously we are
dealing here with a much more refined method of time measure-
ment than in the coffins of the Middle Kingdom. Nevertheless these
texts were mechanically copied over much longer periods than
they could possibly cover correctly. Much effort has been spent to
identify these new lists of stars, often without the realization that
the contents of the texts in a purely philological respect have not
been safely established, because the available copies were made
in the early days of Egyptology and often without consideration
for variants in other copies and related texts. Only a new edition
of this whole material can provide the necessary basis for such
studies.
41. With the Ptolemaic period, Egyptian astronomy changes
in aspect. A totally new element, the Greco-Babylonian zodiac,
appears on the monuments. The ceilings of the Hellenistic
90 Chapter IV

temples erected and restored by Ptolemaic kings and Roman

emperors, truly represent the chaotic mixture of astro-mythology
and astrology of the Hellenistic period.
Beginning with the second century B.C., also astronomical
(or, more accurately, calendaric) and astrological papyri appear,
written in Greek or in Demotic or both. The earliest Demotic and
Greek horoscopes were written close to the beginning of our era.
Also proper "astronomical" texts written sometimes in Demotic,
sometimes in Greek, have been found. We have planetary texts
from the time of Augustus to Hadrian. In these, the dates when a
planet enters a zodiacal sign are recorded. These texts are based
on computations, not on observations, as is evident from the fact
that entrances into a zodiacal sign are also noted when the planet
is in conjunction with the sun, thus being invisible.
Another text of the Roman period, written in Demotic, un-
doubtedly represents an older Egyptian method, probably unin-
fluenced by Hellenism. We have mentioned before that lunar
calendars played a role since early times side by side with the
schematic civil calendar of the 365-day year. An inscription of
the Middle Kingdom mentions "great" and "small" years, and
we know now that the "great" years were civil years which
contained 13 new-moon festivals in contrast to the ordinary
"small" years with only 12 new moons. The way these inter-
calations were regulated, at least in the latest period, is shown
by the Demotic text.
This Demotic text contains a simple periodic scheme which
is based on the fact that 25 Egyptian civil years (which contain
9125 days) are very nearly equal to 309 mean lunar months.
These 309 months are grouped by our text into 16 ordinary years
of 12 lunar months, and 9 "great" years of 13 months. Ordinarily,
two consecutive lunar months are given 59 days by our scheme,
obviously because of the fact that one lunar month is close to
291 days long. But every 5th year the two last months are made
60 days long. This gives for the whole 25-year cycle the correct
total of 9125 days.
In this way one had an exceedingly handy scheme of deter-
mining by means of a simple rule the dates of all lunar festivals
in such a way that no grave error could develop for many cen-
turies, though the single new moon or full moon could deviate
 Egyptian Mathematies and Astronomy 91

by ± 2 days or perhaps even more. Yet the Egyptians were

obviously satisfied with their scheme in the same way as the
Jewish and Christian calendar of the Middle Ages confidently
relied on a periodic scheme which showed equally serious
deviations from the facts.
In summary, from the almost three millenia of Egyptian writing,
the only texts which have come down to us and deal with a
numerical prediction of astronomical phenomena belong to the
Hellenistic or Roman period. None of the earlier astronomical
documents contains mathematical elements; they are crude ob-
servational schemes, partly religious, partly practical in purpose.
Ancient science was the product of a very few men; and these
few happened not to be Egyptians.

BIBUOGRAPHY TO CHAPTER IV
Our knowledge of Egyptian mathematies is primarily based on the following
texts, an of whieh were written in the Middle Kingdom or the Hyksos period.
1. Mathematical Papyrus Rhind published first by Eisenlohr in 1877.
Modem publieation by T. E. Peet, London, 1923; additional material and
photographs in Chaee-Bull-Manning-Arehibald, The Rhind Mathematical
Papyrus, Oberlin, Ohio I 1927, II 1929.
2. MOICOW Mathematical PaPlJrus published by W. Struve, Quellen und
Studien zur Gesehiehte der Mathematik, sere A. vol. 1 (1930).
3. P. Berlin 6619. Published by Sehaek-Sehaekenburg, Zeit8ehr. f.
aegyptisehe Spraehe 38 (1900) p. 135 fl. and 40 (1902) p. 65 f.
4. P. Kahun. Published by F. 11. Griffith, Hieratie Papyri from Kahun
and Gurob, London 1898 Pl. VIII and p. 15 fl.
5. Leather roll British Museum 10260 by S. R. K. Glanville in J. Egyptian
Archeology 13 (1927) p. 232 fl.
6. Wooden tablet. Cairo 26367 and 26368. ReeueD de travaux relatifs i la
philologie et i l'areheologie egyptiennes 28 (1906) p. 62 fl. and Catalogue
generale ... du Musee du Caire, Ostraea, 1901 Pl. 62-64 and p. 95 f.
For the late period, Demotie papyri should be added. One large Demotie
text was found in Tunah el Gabal aecording to m. London News 104 (1939)
p. 840 and Chronique d·~te 14 No. 28 (1939) p. 278. No information about
this text could be obtained. Fragments of geometrical Demotie texts are in the
Carlsberg Colleetion of the University of Copenhagen, to be published by
A. Volten.
Greek papyri are very closely related to the Egyptian texts. For this material
ef. Andre Deleage, Les eadastres antiques jusqu'i Diocletien, £tudes de
papyrologie 2 (1934) p. 73-228. K. Vogel, Beitrige zur grieehisehen Logistik I.
SitzuDgsber. d. Bayerisehen Akademie der Wissenseh., Math.-nat. Abt. 1936
92 Chapter IV

p. 357-472. Cf. also MitteUungen aus der Papyrussammlung der National-

bibliothek in Wien, Grieehische Hterarisehe Papyri I (1932) [Gerstinger and
Vogel].
A great number of geometrical and arithmetical problems ue found on
papyri. A systematic investigation of this scattered material would be worth-
while. I mention only especially large multiplication tables for fractions from
the fourth century published uNo. 146 in Michigan Papyri vol. DI (1936).
Many smaller tables are preserved both in Greek and in Demotic.
An excellent brief summary of Egyptian mathematics wu given by T. E. Peel.
Bun. John Rylands Library 15 (1931). A detailed study by the present author
of the arithmetical methods is to be found in QueUen und Studien zur Gesehichte
der Mathematik, sere B, vol. 1 (1930) p. 301-880, on geometry p. 413-451.
For a deeper understanding ofthe background which determined the chuaeter
of ElYPtian arithmetic, the study of the following works will be of great help:
Lucien Levy-Bruhl, Fonctions mentales dans les soeietes inferieures (1922);
Heinrich Schaefer, Von aegyptiseher Kunst (1919), and Kurt Sethe, Von
Zahlen und Zahlworten (quoted p. 23).
General works on Egypt: J. H. Breasted, A History of Egypt, New York,
Scribner. A. Erman, The Literalure of the Ancient Egyptians; Poems, Narra-
tives, and Manuals of Instruction, from the Third and Second Millenia B.C.,
New York, E. P. Dutton, 1927. Egyptian mathematics is described in O. Neu-
gebauer, Vorgrieehisehe Mathematik. Berlin, SpriDger, 1934 (quoted u
"Vorlesungen"). There exists no modem work on Egyptian utronomy. An
edition of all available ESYPtian astronomical texts by R. A. Parker and the
present author is in prepuation. For literature concerning special problems
see the notes given below to Nos. 39 fI.

NOTES AND REFERENCES TO CHAPTER IV

ad 96. A detailed analysis of the table for ~ of the Papyrus Rhind is given
n
in my book "Die Grundlagen der igyptischen Bruchreehnung", Berlin, Springer,
1926. This theory is summarized in my "Vorlesungen" p. 137 fI. Modifications
were proposed by van der Waerden "Die Enlstehungsgesehichte der aegyp-
tisehen Bruchreehnung" in "QueUen und Studien zur Gesehichte der Mathematik"
sere B vol. 4 (1937) p. 359-382.
ad 97. As stated in the text we find already in the New KiDgdom an exception
to the rules of the Papyrus Rhind for the duplication of unit fractions. William
C. Hayes, Ostraca and Name Stones from the Tomb of Sen-Mut (No. 71) at
Thebes, The Metropolitan Museum of Art, Egyptian Expedition [Publications
No. 15] New York 1942, published an ostracon (No. 153) which contains the
following computation1)
1) The restoration of the original problem given by Hayes seems to me very
doubtful. In the ftrst line one can read safely only i i4 i 2i and I see no reason
for restoring "cubit, palm(t)" at the beginning. The four fractions obviously form
two pairs but I do Dot undentand their relation to the subsequent operations.
 Notes and References 93
1 '7 (black)
3 (red)
2 6 14 21 (black)
3 2 121 (red)
4 "2 14 (black)
102 1 2 (red)

where the numbers below the main lines ue written in red. Obviously we He
deaIiDg here with a multiplication of 7. The standard procedure would be

1 '7
2 '428
4 2 14

Thus we see that the ostracon uses a ditJerent (and more complicated) expression
for twice 7. The analysis of this decomposition is useful for the understanding
of the method which we have described in the text. The standard decomposition
would operate with the natural fraction '4 and determine the fraction which
remains for 2 - i of 1. The result is i.
In the ease of the ostracon we find "auxiliary numbers" written in red below
the fractions. Under 21 we find 1. This means that 21 is introduced as a new
unit; consequently we find 3 below 7. This shows us that we are dealing, not
with the natural fraction i of the sequence 2, i, ..., but with S which belongs
to the sequence I, 3, i, ... Hence we obtain now 21 as one term an«!,.must
find the remainder_which is obtained by multiplying 7 by 2 - 3 = 1 3. We
know already that 3 = 2 i. Thus we have to multiply 7 by 1 2 6. Here again
auxiliary numbers must be introduced, counting '6 as 1, which will lead to

116
2 3
I 6 1
If we take here the first and last term we have 7 new units. Thus we see that
1 i of '7 is i. There remain "2 of 7 which is 14. Thus we have found for the
remainder '6 14 and for the whole of twice '7 the form

6 14 21.
The above computation shows how important it is to begin with the proper
natural ~ction. The use of i leads to a two-term expression, whereas the use
of 3 forced us into a three-term decomposition. I am sure that the Egyptians
never saw behind the underlying reason of divisibility but simply operated by
trial and error. The reader might find the above explanation exceedingly clumsy
and hypothetical. Only a systematic study of many available examples can give
the necessary experience so that one becomes rea1ly famlliu with this type of
uithmetieal rules. A useful illustration, however, is a group of problems from
the mathematical Papyrus Bhind which I have analyzed in detail in my"Vor-
94 Chapter IV

lesungen", p. 139ft. Fortunately, these examples also deal with '7 and its sub-
fractions, and in the majority of cases all the auxiliary numbers which help to
"multiply" fractions are preserved.
ad 39. For Egyptian time reckoning see K. Sethe, Die Zeitrechnung der
alten Aegypter im Verhlltnis zu der der andern Vtilker. Nachr. d. k. Gesellschaft
d. Wissensch. zu Gtittingen, Phil.-hist. Kl. 1919 and 1920. Also L. Borchardt,
Die altlgyptische Zeitmessung, Berlin, De Gruyter, 1920. The Egyptian lunar
calendars are discussed by R. A. Parker, The Calendan of Ancient Egypt,
University of Chicago Press, 1950.
For the later history of the decans see W. Gundel, Dekane und Dekanstem-
bilder. Studien d. Bibliothek Warburg 19 (1936).
ad 39a. For the origin of the Egyptian calendar cr. O. Neugebauer, Die
Bedeutungslosigkeit der "Sothis-periode' fiir die iilteste igyptische Chronologie,
Acta Orientalia 17 (1938) and The Origin of the Egyptian Calendar, J. Near
Eastem Studies 1 (1942). Also H. E. Winlock, The Origin of the Ancient
Egyptian Calendar, Proc. Am. Philos. Soc. 83 (1940), and R. A. Parker in
the book quoted in the preceding section.
The "diagonal calenders" were first discussed by A. Pogo, Isis 17 and 18
(1932) and Osiris 1 (1936). The location of the decans from their period of
invisibility was given by O. Neugebauer in "Vistas in Astronomy" (ed. by
Arthur Beer) vol. I p. 47-51 (London, 1955). For the tomb of Senmut see
A. Pogo, Isis 14 (1930).
The Seti Cenotaph was published by H. Frankfort, in Memoir 39 of the
Egypt Exploration Society (2 vols.) London 1933. For the discussion of the
astronomical ceiling cf. H. O. Lange-O. N eugebauer, quoted p.69 ad No. 29.
A very puzzling text from a Ramesside papyrus (concerning lucky and
unlucky days) was published by J. Cerny, Annales du Service des Antiquites
de rEgypte 43 (1943) p. 179 f. There we find a scheme for the length of daylight
and night from month to month, varyiDg linearly between a minimum of 6
hours and a maximum of 18 hours. Another scheme for the variation of
daylight is discussed by J. J. Clere. Un texte astronomique de Tanis, K6mi
10 (1949) p. 3-27.
ad 41. For the planetary tables cr. O. Neugebauer, Trans. Am. Philos.
Soc., N.S. 32 (1942) with additions in Knudtzon-Neugebauer, Zwei aslro-
nomische Texte, Bull. de la soc. royale des lettres de Lund 1946-1947 p. 77 ft.
Discussion by van der Waerden, Egyptian "Eternal Tables', Koninkl. Nederl.
Akad. van Wetensch., Proc. 50 (1947) p. 536 ft. and p. 782ft.
The dating of four of these planetary tables was related to a peculiar accident
which is worth mentioning as an example of how the most unlikely cOmbinations
may occur and mislead us in our conclusions. The four tables under discussion
are inscribed on wooden tablets which were originally bound together like the
pages of a book by meaDS of strings which were strung through holes in one side
of the wooden frame (cf. PI. 13 which shows Tablet II). These tablets were first
published in 1856 by Brugsch, one of the great pioneers of Egyptology. Each
tablet mentions regnal years and it was natural to arrange them accordingly,
because these years formed a complete sequence as fonows
 Notes and References 95

Tablet I year 9 to 15
Tablet II year 16 to 19 and 1 to 3
Tablet III year • to 10
Tablet IV year 11 to 17

Because these texts were obviously written in the Roman period, Brugsch
concluded that the first ruler must be Trajan, whose reign lasted 19 years and
whose successor was Hadrian, whose reign lasted longer than 17 years. These
conclusions proved to be correct, however, only for tablets I, II, and IV. Checking
of the astronomical data showed immediately that No. III could not be the
continuation of No. II nor could it be the predecessor of No. IVI). Indeed it
was easy to show that the years ". to 10" were not the years of Hadrian but of
Vespasian, 30 years before Trajan. Hence we know that by mere accident
Tablet III seems to fit between II and IV. Similar eases may occur, more often
than we think, in historical research but escape discovery simply because the
rigorous astronomical check is not applicable.
The 25-year cycle was discovered in the Demotic papyrus No.9 of the Carls-
berg collection, published by O. Neugebauer and A. Volten in Quellen und
Studien zur Gesehichte d. Mathematik, ser. B, vol.• (1938). This 25-year cycle
was well known and often used in Hellenistic astronomy. Ptolemy, e. g., arranges
his tables of syzygies according to it (Almagest VI, 3).
One must not misinterpret the expression "25-year cycle" as a parallel to
the previously mentioned "19-year cycle" (or "Metonic cycle"; d. p. 7). In
the first case the 25 years are Egyptian calendar years of exactly 365 days each.
In the second ease the years are tropical years, i. e., time intervals which are
astronomically defined and which involve fractions of days. The first cycle
comprises 309 mean lunar months at the end of which the same Egyptian civil
day appears again as the date of a new moon or full moon. In the second cycle
235 mean lunar months bring the same lunar phase back to the same season,
but it depends on the local calendar whether or not this restores also the calendar
date. Because the Greek astronomers operated consistently with the Egyptian
calendar in their tables, the 25 year cycle was by far the most convenient cycle
to use.
From the enormous wealth of written documents from ancient Egypt we
have only one doubtful reference to a partial solar eclipse of 610 B.C. -assuming
that this is the correct interpretation of the text (d. W. Erichsen in Akad. d.
Wiss. u. Lit. Mainz, Abh. Geistes- u. Soz. Wiss. 1956, No. 2.1) Not a single
Egyptian observation is quoted in the Almagest, although Ptolemy gives extensive
references to earlier observations on which his theory is based. There exists one
Coptic eclipse record of 601 A.D.(I), first identified by Krall and Ginzel

I) This was correctly realized by William Ell is, Memoirs Roy. Astron. Soc.
25 (1857) p. 112 but, strangely enough, Ellis did not determine the correct date
of No. III.
S) The alleged eclipse report of Osorkon (9th cent. B.C.) does not concern an
actual eclipse, as was shown by R. Caminos in Analecta OrientaUa 37 (1958)
p. 88 fr.
96 Chapter IV

(S. B. Akad. d. Wiss. Wien, math.-nat. Cl. 88, 2 [1883] p. 655) and again by
E. B. Allen, J. Am. Oriental Soc. 67 (1947) p. 267.
Ap~. The reader may have missed a reference to the astronomical or
mathematical significance of the pyramids. Indeed, a whole literature has been
built up around the "mysteries" of these structures, or at least one of them, the
pyramid of Khufu (or "Cheaps"). Important mathematical constants, e. g., an
accurate value of n, and deep astronomical knowledge are supposed to be
expressed in the dimensions and orientation of this building. These theories
contradict flatly all sound knowledge obtained by archeology and by Egypt-
ological research about the history and purpose of the pyramids. The reader
who wants to see an excellent account of these facts should consult the paper
by Noel F. Wheeler, Pyramids and their Purpose, Antiquity 9 (1935) p. 5-21,
161-189, 292-304 and L. Borchardt. Gegen die Zahlenmystik an der grossen
Pyramide bei Gise, Berlin 1922.
For the very complex historical and archaeological problems connected with
the pyramids, cr., e. g., J. P. Lauer, Le probl~me des pyramides d'~te,
Paris 1948, and I. E. S. Edwards, The Pyramids of Egypt, Penguin Books,
1952. How little one knows about the significance of the arrangement of rooms
and corridors in the interior is particularly evident in the case of the "Bent
Pyramid" at Dahshur: cr. A. Fakhry's recent excavation reports in Annales
der Service des Antiquites de l'Egypte 51 (1954) p. 509 ft and 52 (1955) p. 563ft.
 97

CHAPTER V

Babylonian Astronomy.
4-t. There is scarcely another chapter in the history of science
where an equally deep gap exists between the generally accepted
description of a period and the results which have slowly emerged
from a detailed investigation of the source material. This discre-
pancy has its roots as far back as the Hellenistic tradition about
the "Babylonians" or "Chaldeans" who are innumerably many
times mentioned in ancient writings. especially in the astrological
literature. Thus magic. number mysticism. astrology are ordinarily
considered to be the guiding forces in Babylonian science. As far
as mathematics is concerned. these ideas have had to be most
drastically revised since the decipherment of mathematical texts
in 1929. But for more than 70 years the same sort of revision
resulted from the discoveries of Epping and Kugler in Babylonian
astronomy. Thanks to the work of these scholars. it very soon
became evident that mathematical theory played the major role
in Babylonian astronomy as compared with the very modest
role of observations. whose legendary accuracy also appeared
more and more to be only a myth. Simultaneously the age of
Babylonian astronomy had to be redefined. Early Mesopotamian
astronomy appeared to be crude and merely qualitative. quite
similar to contemporary Egyptian astronomy. At best since the
Assyrian period. a turn toward mathematical description becomes
visible and only the last three centuries B.C. furnished us with
texts based on a consistent mathematical theory of lunar and
planetary motion. The latest astronomical text has been identified
recently by Sachs and Schaumherger. with the date of 75 A.D.
These late theories. on the other hand. proved to be of the highest
level. fully comparable to the corresponding Greek systems and
of truly mathematical character. Simultaneously it had to be
98 Chapter V

admitted that we know next to nothing about the details of

horoscopic astrology in Mesopotamia in sharpest contrast to the
overwhelming abundance of astrological documents from Hellen-
istic Egypt and the Roman and Byzantine period.
Finally it has been repeatedly remarked by competent observers
that the almost proverbial brilliance of the Babylonian sky is
more a literary cliche than an actual fact. The closeness of the
desert with its sand storms frequently obscures the horizon.
This is the more essential as the majority of problems in which
the Babylonian astronomers were interested are phenomena close
to the horizon. The lunar calendar requires observation of the
first visibility of the new crescent in the western horizon. The
last visibility of the moon happens at the eastern horizon. Dis-
appearance and reappearance of the planets are phenomena close
to the horizon and it seems that also ••opposition" of a planet
was defined as rising or setting at sunset and sunrise respectively.
Only eclipses and occultations will usually be observable under
favorable conditions. It is certainly the result of this situation
that Ptolemy states that practically complete lists of eclipses are
available since the reign of Nabonassar (747 B.C.) while he
complains about the lack of reliable planetary observations. He
remarks that the old observations were made with little com-
petence, because they were concerned with appearances and dis-
appearances and with stationary points, phenomena which by
their very nature are very difficult to observe. It is worth noting
that this precise description of the planetary observations of the
Babylonians by a competent astronomer had almost no effect
on the current evaluation of Babylonian astronomy while the
vague but abundant references of the Hellenistic astrologers to
Chaldean wisdom completely dominated the picture which later
centuries developed of Chaldean astronomy.
49. Our description of Babylonian astronomy will be rather
incomplete. The historical development will be given in bare
outline. As in the case of Egypt. a detailed discussion of the few
preserved early texts would require not only too much room but
would also unduly exaggerate their historical importance. For
the late period, however, the opposite situation prevails. A great
number of texts exist from the Seleucid epoch and only a very
extensive and highly technical discussion could do justice to the
 Babylonian Astronomy 99

mathematical and astronomical achievements of this period.

Obviously it is impossible to do this in the present frame.
We begin our survey with a short description of the earlier
development. Then, before entering the discussion of the Seleucid
period, a few remarks about our source material and its pro-
venance will be necessary. Our description of the mathematical
astronomy of the Hellenistic period we shall start with an outline
of the theory of the solar and lunar motion because this theory
is undoubtedly the most characteristic section of this whole
development. The planetary theory will then be summarized so
far as essentially new ideas which go beyond the methods already
known from the luni-solar problems are apparent. We shall,
finally, touch upon the few facts which are known about the
milieu in which these texts originated.
44.. To begin our historical sketch with a negative statement,
we can say that nothing is known about a Sumerian astronomy.
Mythological concepts which involve the heavens, deification of
Sun, Moon, or Venus cannot be called astronomy if one is not
willing to count as hydrodynamics the existence of belief in a
storm deity or the personification of a river. Also the denomination
of conspicuous stars or constellations does not constitute an
astronomical science.
One of the earliest documents with definitely astronomical
trend is a tablet in the Hilprecht Collection in Jena, Germany.
The text was probably written in the Cassite period but copied
from an original composition which was oldeJ". Its formulation is
quite similar to a familiar type of Old-Babylonian mathematical
texts. The document begins with a list of numbers and names
which might be interpreted as follows: "19 from the Moon to
the Pleiades; 17 from the Pleiades to Orion; 14 from Orion to
Sirius", and so on for eight stars or constellation, ending with
the statement that the total (of what?) is 120 "miles" and the
question "how much is one god (i. e., star) beyond the other
god?" Then begins the "procedure", exactly as in a mathematical
text. It consists in dividing each of the given numbers by their
sum, which is 1,21, well known to us as the last entry in the
standard table of reciprocals (cf. p. 32). Each of these results
is converted into "miles" and lower units of distance and ex-
plained as the distance from one of the previously enumerated
100 Chapter V

celestial objects to the next. The text ends with the customary
Hsuch is the procedure" and the names of the scribe who copied
the text and the one who verified the copy.
This text and a few similar fragments seem to indicate some-
thing like a universe of 8 different spheres, beginning with the
sphere of the moon. This model obviously belongs to a rather
early stage of development of which no traces have been found
preserved in the later mathematical astronomy, which seems to
operate without any underlying physical model. It must be
emphasized, however, that the interpretation of this Nippur text
and its parallels is far from secure.
Another group of probably contemporary texts represents a
division of the sky into three zones of 12 sectors each. Each
zone contains the names of constellations and planets and simple
numbers in arithmetic progression like 1 1,10 1,20 etc. up to
2 and down again 1,50 1,40 until 1. This is probably the earliest
occurrence of an arithmetical scheme which was later developed
into an important tool for the description of periodic phenomena,
the so-called zigzag functions. In the present ease, the numbers
are so simple and so obviously schematic that many different
interpretations which explain them equally well or equally badly
can be proposed.
There is another class of early documents which deserves
mention because it contains the earliest records of actual ob-
servations in Mesopotamia. For several years of the reign of
Ammisaduqa the appearances and disappearances of Venus
were recorded. Because the dates are given in the contemporary
lunar calendar, these documents have become an important
element for the determination of the chronology of the Ham-
murapi period. From the purely astronomical viewpoint these
observations are not very remarkable. They were probably made
in order to provide empirical material for omina; important
events in the life of the state were correlated with important
celestial phenomena, exactly as specific appearances on the livers
of sacrificial sheep were carefully recorded in the omen literature.
Thus we find already in this early period the fIrst signs of a
development which would lead centuries later to judicial astrology
and, finally, to the personal or horoscopic astrology of the Hel-
lenistic age.
 Babylonian Astronomy 101

It is difficult to say when and how the celestial omens developed.

The existing texts are part of large series of texts, the most
important one called "Enuma Anu Enlil" from its initial sentence,
similar to papal bullae in the Middle Ages. This series contained
at least 70 numbered tablets with a total of about 7000 omens.
The canonization of this enormous mass of omens must have
extended over several centuries and reached its final form perhaps
around 1000 B.C.
Historically much more interesting than this mass of purely
descriptive omens are two texts which were called "mul apin".
The earliest preserved copies are dated around 700 B.C., but
they are undoubtedly based on older material. They contain a
summary of the astronomical knowledge of their time. The first
tablet is mostly concerned with the fixed stars which are arranged
in three "roads", the middle one being an equatorial belt of
about 30 0 width. The second tablet concerns the planets, the
moon, the seasons, lengths of shadow, and related problems.
These texts are incompletely published and even the published
parts are full of difficulties in detail So much, however, is clear:
we find here a discussion of elementary astronomical concepts,
still quite descriptive in character but on a purely rational basis.
The data on risings and settings, though still in a rather schematic
form, are our main basis for the identification of the Babylonian
constellations.
Around 700 B.C., under the Assyrian empire, we meet with
systematic observational reports of astronomers to the court.
Obviously the celestial omens have now reached primary im-
portance. In these reports no clear distinction is yet made
between astronomical and meteorological phenomena. Clouds
and halos are on equal footing with eclipses. Nevertheless, it had
been already recognized that solar eclipses are only possible at
the end of a month (new moon), lunar eclipses at the middle.
The classical rule that lunar eclipses are separated from one
another by six months, or occasionally by five months only,
lnight well have been known in this period. We should reeall
here Ptolemy's statement that eclipse records were available to
him from the time of Nabonassar (747 B.C.) onwards.
It is very difficult to say when this phase developed into a
systematic mathematical theory. It is my guess that this happened
102 Chapter V

comparatively rapidly and not before 500 B.C. Up to about

480 B.C., the intercalations of the lunar calendar show no regul-
arity whatsoever. One century later, however, the rule of 7 inter-
calations in 19 years at fixed intervals seems to be in use, and
remains from now on the basis of all the lunar calendars which
were derived from the Babylonian scheme, including the lunar
calendar of the Middle Ages discussed in the first chapter.
A luni-solar intercalation rule presupposes the recognition of
a relation which indicates that m lunar months are equal in
length to n solar years. In the specific case of the 19-year cycle
m = 285 and n = 19. In the preceding period a "year" was an
interval of sometimes 12 or sometimes 13 months, where probably
the state of the harvest decided the need for a 13th month. The
existence of a cycle, however, proves that a more precise astron-
omical definition of "year" was adopted. We cannot give accurate
data about the mean length of such a year or how it was deter-
mined. There are good reasons, however, which point to an ob-
servation of the summer solstice as the point of comparison. At
any rate, it is the summer solstices which are systematically com-
puted, whereas the equinoxes and the winter solstices are simply
placed at equal intervals. Because much more accurate methods
were known in the Seleucid period, it is plausible to assume that
the scheme of the 19-year cycle represents a slightly earlier phase
of development.
We shall see that period relations of the above-mentioned type
form the very backbone of Babylonian mathematical astronomy;
these are relations which state that s intervals of one kind equal
t intervals of another kind. Mathematical astronomy is fully
developed at about 300 B.C. at the latest. The 19-year inter-
calation cycle is certainly one of the most important steps pre-
ceding the later astronomical methods, that is to say, later than
about 450 B.C. Roughly to the same period, probably the fourth
century, belongs also the invention of the zodiac. The constellations
which lent their names to the zodiacal signs are, of course, much
older. But it was only for mathematical reasons that a definite
great circle which measured the progress of the sun and the planets
with respect to exactly 30°.10ng sections was introduced. Indeed,
the zodiac was hardly ever more than a mathematical idealization
needed and used exclusively for computing purposes. Actual
 Babylonian Astronomy 103

positions in the sky were expressed until the end of euneiform

writing with referenee to well known bright stars. This primitive
system was still in use in Greek horoseopes of the Roman period,
exactly as in Babylonian texts, side by side with the determination
of positions by degrees and zodiaeal signs.
We may now enumerate the tools whieh were available at the
end of the "prehistory" of Babylonian astronomy whieh extends
from about 1800 B.C. to about 400 B.C. The zodiae of 12 times
30 degrees as referenee system for solar and planetary motion.
A fixed luni-solar ealendar and probably sollie of the basie
period relations for the moon and the planets. An empirieal
insight into the main sequenees of planetary and lunar phenomena
and the variation of the length of daylight and night. The use
of arithmetie progressions to deseribe periodieally variable
quantities. And, above all, a eomplete mastery of numerieal
methods whieh eould immediately be applied to astronomieal
problems. The utilization of these possibilities marks indeed the
erueial step.
45. The next seetion will bring us to the discussion of the
eompleted system of the Babylonian lunar theory. This diseussion
is based on texts whose signifieanee was first reeognized by
Fathers Epping and Strassmaier. In 1881 there appeared in
a Catholie theologieal periodical, the "Stimmen aus Maria Laaeh",
an artiele "Zur Entzifferung der astronomisehen Tafeln der Chal-
diier" by J. Epping of Quito, Equador, with an introduetion by J. N.
Strassmaier in LOlldon. This paper eoneerns a faseinating report
of the first dl' :pherment of astronomieal tablets whieh then were
arriving in London in ever increasing numbers. The two authors
were fully eonseious of the importanee of their diseoveries.
Indeed, this first paper eontained the eorreet determination of
the zero point of the Seleueid Era. and that of the Parthian Era,
thus providing for the first time a solid ehronologieal basis for
the history of Mesopotamia after Alexander the Great. But much
more was done for the understanding of Babylonian astronomy
itself. Suddenly it beeame elear that arithmetieal progressions
were skillfully utilized for the predietion of lunar phenomena,
with an aeeuraey of a few minutes. The names of the planets
and of zodiaeal eonstellations were eorreetly determined and the
road opened for the translation of astronomieal reeords. On ten
104- Chapter V

pages Epping described discoveries which were to inaugurate a

new epoch in the history of science.
Eight years later Epping published a small book, entitled
"Astronomisches aus Babylon" in the supplements to the "Stim-
men aus Maria Laach". Here one finds an account of the
leading ideas of the Babylonian theory of the moon as well as
a detailed discussion of planetary and lunar almanacs. Epping
died in 1894-. The period of initial discoveries came to a conclusion
in the monumental works by Father Kugler, published between
1900 and 1924-.
The texts on which Epping's and Kugler's work are based
come exclusively from the British Museum. For many years
Strassmaier copied there thousands of tablets, the majority of
which belong to the latest periods of Babylonian history. These
copies were collected in notebooks, of which one page is repro-
duced as a characteristic sample on PI. 14-. It was on Strass-
maier's initiative that Epping began the study of Strassmaier's
transcriptions of astronomical texts. When the decipherment
proved to be successful, Strassmaier excerpted, from his note-
books, astronomical texts on special sheets, often adding ex-
planatory remarks. These sheets were then sent to Epping for
final investigation, and after Epping's death, to Kugler. Kugler's
successor, Father Schaumberger, and I myself got the main
portion of our texts from Strassmaier's copies which he had
entered in his voluminous notebooks during the 1880's and
1890's. Not a single one of these texts was ever published in the
official publications of the British Museum; and no information
whatsoever is available concerning similar tablets which the
British Museum may have acquired after Strassmaier ceased
copying.Without Strassmaier, Epping, and Kugler, the few other
astronomical texts so far published would probably have been
laid aside in other museums too. It is very likely that no trace
of this enormously rich material would have penetrated to the
outside world, and Babylonian astronomy would still appear to
us in the light of a few texts from the earliest periods and of the
omens of Eniima-Anu-Enlil. A few numbers will illustrate this
situation. From Strassmaier's notebooks and from Kugler's pub-
lications about 24-0 astronomical texts and fragments were reco-
vered, all of which were probably found in one archive in Baby-
 Babylonian Astronomy 105

Ion. From the inventory numbers of the British Museum, one can
conclude that these texts reached the museum between November,
1876 and July, 1882. During these six years the number of tablets
increased from over 32,000 to more than 46,000 and one could
expect that many hundreds of astronomical texts would be among
these masses of texts. Indeed, in 1953 it became known that
T. G. Pinches, before 1900, had copied some 1300 pieces of
astronomical texts. This material was then put at the disposal of
A. J. Sachs who published it with the addition of many related
pieces in 1955. Thus we have now the major part of one ancient
archive at our disposal, as far as it had reached the British Museum.
46. The mathematical astronomical texts fall into two major
groups: "procedure texts" and "ephemerides". The texts of the
first class contain the rules for the computation of the "ephem-
erides", which, in turn, are similar to a modern "nautical alma-
nac", giving for a specific year (or for some specific sequence
of years) the lunar or planetary positions at regular intervals. If
the "procedure texts" were complete and if we fully understood
their technical language, they might suffice for the actual com-
putation of the ephemerides. In fact, however, none of these
assumptions is satisfied. The preserved texts are badly damaged
or totally missing for many of the steps; their terminology is far
from clear, at least to us; and it might be justly asked if even
a complete set of procedure texts would not have required sup-
plementary oral explanation before it could be used for actually
computing an ephemeris. Consequently the ephemerides them-
selves form the major basis for our researches, and the procedure
texts often play the role of very welcome testing material for the
rules which we finally abstract from the completed ephemerides.
In the subsequent discussion we shall, however, make no sharp
distinction between these two groups of sources and we shall act
as if we had explicit rules at our disposal, though they are often
actually only obtained from a very complex interplay between
related fragments of both classes of texts.
The number of available astronomical tablets from the Seleucid
period is not at all large. I know of less than 250 ephemerides,
more than half of which are lunar, the rest planetary. The number
of procedure texts is about 70, the majority of which are only
small fragments. Thus our knowledge of Babylonian mathematical
106 Chapter V

astronomy is based on about 300 tablets. To this number can

now be added the vast mass of some 1000 non-mathematical
astronomical texts from Pinches-Sachs; it will take nlany yean
of patient work before the conclusions can be drawn from thi\!l
great variety of new sources for the earlier development of
Babylonian astronomy in all its theoretical and practical aspects.
47. The fundamental problem of the Babylonian lunar theory
is determined by the calendar. So far as we know, the Babylonian
calendar was at all periods truly lunar, that is to say, the "month"
began with the evening when the new crescent was for the first
time again visible shortly after sunset. Consequently the Ba-
bylonian "day" also begins in the evening and the "first" of a
month is the day of the first visibility. In this way the beginn-
ing of a month is made dependent upon a natural phenome-
non which is amenable to direct observation. This is certainly
a very simple and natural definition, as simple as the concur-
rent definition of the "day" as the time from one sunset to
the nex.t. But as is often the case, a "natural" definition leads
to exceedingly complicated problems as soon as one wishes to
predict its consequences. This fact is drastically demonstrated
in the case of the lunar months. A very short analysis will illus-
trate the intrinsic difficulties.
A "lunar month" obviously contains an integer number of
days. How many? A rough estimate is easily obtainable. No two
consecutive reappearances of the new crescent after a short
period of invisibility of the moon are ever separated by more
than 30 days or by less than 29 days. Thus immediately the
main problem arises: when is a month 30 days long, when 29?
To answer this problem we must obtain an estimate not only of
the lunar motion, but also of the motion of the sun. In one year
of, roughly, 365 days, the sun moves once around us; that is
to say, after this time the sun again comes back to the same
star, having completed a great circle of 360°. Thus the solar
motion per day is close to 1° and therefore close to 30 ° in one
month. The time from one new crescent to the next is obviously
about equal to the time from invisibility to invisibility. But the
moon is invisible because it is close to the sun. Thus a month
is measured by the time from one "conjunction" of the moon
with the sun to the next. During this time the sun traveled about
 Babylonian Astronomy 107

30°; the moon, however, traveled not only 30°, but completed
one additional whole rotation of 360°. Hence 390° are covered
in about 30 days; this shows us that the moon must cover about
13° per day.
Now the real difficulties begin. In order to make the mst
crescent visible the sun must be sufficiently deep below the
horizon to make the moon visible shortly before it is setting
(Fig. 4). The evening before, the moon was still too close to the
sun to be seen. Hence it is necessary to determine the distance
from the sun to the moon which is required to obtain visibility.

'\U1s j
Western Horl zon

Fig. 4.
o
The distance between them depends on the relative velocity 'Of the
two bodies. We have found that the moon moves 13° per day,
the sun 1° per day; thus the distance in question, the so-called
"elongation", increases about 12° per day. But this estimate is
no longer accurate enough to answer the question as to the
moment when the proper elongation is reached. Neither the sun
nor the moon moves with constant speed. Thus the daily elon-
gation might vary between about 10° and 14° per day. This shows
that our problem involves the detailed knowledge of the variation
of both solar and lunar velocity.
But even if we had insight into the variable velocity of both
bodies the visibility problem would not be solved. For a given
place, all stars set and rise at fixed angles which are determined
by the inclination of the equator and the horizon. The relative
motion which we were discussing before is a motion in the ecliptic,
which makes an angle of about 24° with the equator. Consequently
we must know the variations of the angles between ecliptic and
horizon. For Babylon we find a variation from almost 30° to
almost 80° (Fig. 5). Thus the same elongation produces totally
different visibility conditions at different times of the year.
Let us assume that also the problem of the variation of the
108 Chapter V

Western
..... . .. Horizon
... .
. ~
~
...... ~l.
..
Ecl.\ I
quo
Fig. 5.

angles between ecliptic and horizon is satisfactorily answered.

Then we must still remember that only the sun travels in the
ecliptic whereas the moon deviates periodically from it between
the limits of about +5° and _5° in "latitude". This deviation
is measured perpendicularly to the ecliptic. If the ecliptic is
almost vertical to the horizon (as is the case in spring), then the
latitude has relatively little effect upon the visibility (cf. Fig. 6).
In the fall, however, the full effect of the latitude is felt in bringing

I +5-
\ I
-~
•• I
• I
I
••
••
w. I
I
I
Ho...
I

•I I

$
\Ecl.
Fig. 6.

the moon nearer to, or farther away from, the horizon (Fig. 7).
Thus we need also the knowledge of the variation of the latitude
of the moon.
All these effects act independently of each other and cause
quite irregular patterns in the variation of the length of lunar
months. It is one of the most brilliant achievements in the exact
sciences of antiquity to have recognized the independence of
 Babylonian Astronomy 109
these influences and to develop a theory which permits the pre-
diction of their combined effects. Epping, Kugler, and Schaum-
berger have indeed demonstrated that the lunar ephemerides of
the Seleucid period follow in all essential steps the above outlined
analysis.
Before turning to the description of these ephemerides we can
observe that the solution of the problem of first visibility readily

w. Hor.

Fig. 7.

permits the solution of some other problems which were also

of great interest. First of all, the day by day positions of sun
and moon can easily be established as soon as the laws which
determine the variation of solar and lunar velocity are known.
Thus it is not surprising to find tables which give the daily motion
of sun or moon. Secondly, one can solve the problem of last
visibility of the moon by applying essentially the same argument
to the eastern horizon and the rising of sun and moon. Finally,
both the first and last visibility require as a preliminary step the
knowledge of the moments of conjunction which fall in the middle
of the interval of invisibility. Exactly the same considerations
lead to the computation of the moments of opposition. If we
combine this knowledge with the rules which determine the
latitude of the moon, we can answer the question when the
moon will be close to the ecliptic at oppositions or conjunctions.
In the fIrst case we can expect a lunar eclipse, in the second a
solar eclipse. Thus it is only a logical step which leads from the
computation of the new moons to eclipse tables wmch we find
derived from the ephemerides.
I hoPe that this superficial summary of the main results and
problems of the lunar theory will suffice to give an impression
of the inner consistency and the truly mathematical character of
this theory. The following discussion of some details is nol only
110 Chapter V

intended to clarify the single steps but also to illustrate a pre-

viously made remark to the effect that at no point of this theory
are the traces of a specific geometrical model visible.
4.8. The main tool for the computation of the ephemerides is
arithmetic progressions, increasing and decreasing with con-
stant difference between fixed limits. As an example we offer the
following excerpt from the first three columns of an ephemeris for
the year 179 Seleucid Era, i. e. 133/132 B.C.:
XII. 28,55,57,58 22, 8,18,16 Cf
2,59 I 28,37,57,58 20,46,16,14 ~
II 28,19,57,58 19, 6,14,12 U
III 28,19,21,22 17,25,35,34 ms
IV 28,37,21,22 16, 2,56,56 cSt
V 28,55,21,22 14,58,18,18 1JR
VI 29,13,21,22 14,11,39,40::Q:
VII 29,31,21,22 13,43, 1, 2 III
VIII 29,49,21,22 13,32,22,24 t
IX 29,56,36,38 13,28,59, 2 11
X
XI
29,38,86,38
29,20,86,38
13, 7,35,40 =
12,28,12,18)(
XII 29, 2,36,38 11,80,48,56 CV'
In the first column we have dates, beginning with an inter-
calary (13th) month, called XII.; then follows the year 2,59 of
the Seleucid Era and all the months from I to XII of this year.
These dates have a more accurate meaning. The first XII. does
not mean the whole month XII. but the moment of the mean
conjunction which falls at the end of this month. Similarly, each
subsequent month signifies the moment of the mean conjunction
of this month. Consequently the time inverval from line to line
represents always the same amount of one mean synodic month.
The arithmetical structure of the second column is simple to
analyze. All the numbers of the first three lines end in 57,58.
Then follow six lines ending in 21,22 and finally we have four
lines which have 36,38 as terminal figures. Thus we need only
concentrate our attention on the first two places. The first three
lines show a fixed decrease of 18 in the second place:
28,55 28,37 28,19
 Babylonian Astronomy 111

The next group shows an increase of 18 from line to line:

28,19 28,37 28,55 29,13 29,31 29,49
Then follows again a decreasing sequence:
29,56 29,38 29,20 29,2
with difference of 18. If we plot these numbers in a graph with
equidistant points representing the consecutive lines, then we

. M
.. ..... .
'

~.
....... .. .. ..
'
' '
'"
' .
~
..... y .....
~

]1[2 I ][ ][ :mn: lX:

Fig. 8.

obtain a sequence of points which lie on straight lines of alter-

nating slope ± 18 (cf. Fig. 8). We call such sequences ulinear
zigzag functions".
The straight lines intersect in a maximum value M and a
minimum value m which can easily be computed from our table.
One finds
M = 30,1,59,0
m = 28,10,39,40

From similar tables one can demonstrate that the same extremal
values were used. Consequently our linear zigzag function is
bounded by a fixed maximum M and a fixed minimum m and
therefore forms a periodic function of amplitude
LI =M - m = 1,51,19,20
and mean value
p. = t (M + m) = 29,6,19,20.
Finally we introduce the concept of "period" P. The abscissa in
our graph is divided into equidistant steps, each of which repre-
sents a mean synodic month. We now can ask for the distance
between two consecutive points of maximum (or minimum)
112 Chapter V

Fig. 9.

measured in these units which represent mean synodic months

(cf. Fig. 9). By a simple geometrical argument one finds that
p=2L1
d
where L1 is the amplitude M - m and d the difference from one
line to the next. Substituting our present numbers LI = 1,51,19,20
and d = 18,0,0 one finds
p = 3,42,38,40.
18,0,0

We can cancel common factors in numerator and denominator

or we can express this ration as a sexagesimal fraction. Thus we
obtain finally:
2,46,59
P = 3 3 = 12;22,8,53,20.
1 , 0

In other words we have shown that two consecutive maxima or

minima in the zigzag function of the second column are separated
by 12;22,8,53,20 mean synodic months or, roughly, by slightly
more than 121 months.
The astronomical significance of the second column is revealed
by means of the third column. The first line of the third column
is obviously to be interpreted as a point in the ecliptic of longitude
~ 22;8,18,16. If we add to this the value 28;37,57,58 which we
find in the second line of the second column we obtain
~ 50;46,16,14 = ~ 20;46,16,14
and this is the longitude found in the second line of the third
column. The same rule applies for all subsequent lines and we
can therefore say that the second column contains the differences
of the third column. Combining this result with the fact that the
 Babylonian Astronomy 113

first column contains the dates of the consecutive conjunctions,

we can say that the third column gives the monthly longitudes
of the moon and also of the sun because we are dealing with
conjunctions. The second column gives the monthly progress of
the sun or the solar velocity. Thus we have reached the important
result that the ephemeris under discussion represents the yearly
variation of the solar velocity by means of a linear zigzag function.
Another important item of information is contained in the
value we have found for the period P of this zigzag function,
namely
2,46,59
P = 3 = 12;22,8,53,20 months.
13, 0

This shows not only the value which was adopted here for the
length of the year, measured in mean synodic months, but we
can read this relation also in the form
13,30 years = 2,46,59 months
or
810 years = 10019 months.
It seems as if this relation would imply the use of observational
records going back more than 800 years. This conclusion is,
however, too hastily drawn. First of all, it can be shown that
other columns of the same type of ephemerides are based on
the simpler relation
P
= 46,23 = 12;22,8 months
3,45
or
225 years = 2783 months.
But neither can this relation be taken as the direct result of
observations. The period of a zigzag function is given by the
quotient of 2Lt and d where LI is the amplitude M - m and d
the difference. The values of d in a linear zigzag function are
usually simple numbers-in our example 18,0,0 and not, per-
haps, 17,59,59-as is easy to understand in view of the practice
of computing an ephemeris, where the value of d has to be
added or subtracted in every single line. The accuracy of the
value of LI = M - m is reflected in the number of sexagesimal
114 Chapter V

places of .V and m and thus of all intermediate places. Again

it is reasonable to choose conveniently small numbers for M and
m and Lt. In other words the value of P depends on small cor-
rection in the values of Lt and d and does not depend solely
on the initial empirical relation between the number of years
and the corresponding number of months.
This situation is typical throughout Babylonian astronomy.
The ephemerides alone are never a reliable source for the in-
vestigation of the basic empirical facts. At present it is com-
pletely impossible to write a "history" of Babylonian astronomy
in its latest phase. All we do have is the ephemerides in a form
excellently adapted to practical computation and to predicting
new mooDS, eclipses, etc. We do not know, however, which
empirical elements were actually used for the determination of
the basic parameters nor are we able to retrace the steps by
means of which the theory was formed.
49. The example of an ephemeris which we have quoted in
the preceding section is based on the assumption that the variation
of the solar velocity follows the scheme of a linear zigzag function.
All texts which show this pattern will be called texts of "System B".
In contrast to this we classify ephemerides as belonging to a
"System Aft if the solar velocity is assumed to be constant on
two complementary arcs of the ecliptic in the following way.
From 11R 13 to )( 27 the sun moves 30 0 in each mean synodic
month; from )( 27 to 11R 13 with a motion of 28;7,30 0 per month

I
K
I
T
I J1Jl~

Fig. 10.
I I
JC
t
...
(cf. Fig. 10). It is easy to show that this corresponds exactly to
the relation which we mentioned at the end of the last section,
namely,
1 year = 12;22,8 months
and which also occurs otherwise in ephemerides of both systems.
Obviously it seems to be a much more natural assumption to
let the solar velocity vary continuously instead of having discon-
 Babylonian Astronomy 115

tinuous variations at lJR 13 and )( 27. Nevertheless Kugler has

seen that System A is in general more primitive and therefore
older than System B, and this has been confirmed by subsequent
study. In particular it can be shown that the assumption of a
solar movement of the type of System B leads to rather com-
plicated consequences and it is on this basis that System A
adopted its cruder pattern. Here again we meet with the fact
that purely mathematical considerations exercise an essential
influence on the details of the theory behind which the original
empirical data and general concepts are veiled from our sight.
Though the chronological priority of System A seems to be
well established we have no means to determine the date of
origin of either one. It is furthermore a curious fact that both
systems were simultaneously used during the whole period (from
about 250 B.C. to about 50 B.C.) for which ephemerides are
preserved. This coexistence of two different methods of com-
puting ephemerides is not a matter of "schools" in so far as both
"systems" are attested both at Babylon and at Uruk, the places
of origin of the only two archives to which we can safely assign
our texts. It is difficult to explain why both methods were kept
alive in spite of the fact that System B was certainly an im-
provement over System A in several respects. In the planetary
theory a still higher multipiicity of procedures exists simultane-
ously, very much contrary to our modern scientific habits.
50.We now turn to a rapid summary of the lunar ephemerides
without explicitly attempting to derive our statements from the
textual material or to analyze in detail the general theory which
formed the basis upon which the numerical procedures were
built. Our sketch includes both systems, leaving aside variants
which occur especially in System B.
The general arrangement of all ephemerides is identical. Each
line represents a month, each column a specific "function" like
solar velocity, lunar velocity, etc. We denote these functions by
capital letters, which we also use for the corresponding columns.
The majority of lunar ephemerides cover one year but we also
have texts which concern two or even three years. The general
appearance of such a text with its different columns is shown
on PI. 7b. The columns of ephemerides always proceed from
left to right.
116 Chapter V

The first column in all ephemerides is the column of dates T,

giving the year of the Seleucid era and the consecutive months
(cf. the example on p. 110). Because the edge of a tablet is par-
ticularly exposed to destruction, one often meets the problem of
restoring the date of an ephemeris. This can be done by con-
tinuing preserved columns until one reaches the corresponding
column of another dated text. In this very way it is possible to
show that all texts of System A form one consistent set of ephem-
erides throughout the whole interval (of two centuries) at our
disposal. System B, however, shows a much lower degree of
uniformity.
The next column, tP, is peculiar to System A only. Its period is
identical with the period of the variable lunar velocity and its
units are time degrees. It is used for the computation of the
variable length of the synodic month (column G) under the
preliminary assumption of constant solar velocity. The details of
the construction of this linear zigzag function tP are not yet clear
but it is now certain that it is related to the 18-year cycle, the
socalled "Saros" of 223 mean synodic months. This period is
slightly shorter than 239 anomalistic months and therefore also
the length G of the synodic month almost repeats itself after one
Saros. The slight difference of time in the length of two months
one Saros apart is the difference of tP. From this change of G
after one Saros can then be found the corresponding change of
G from month to month. This is an interesting case of an important
method of ancient astronomy: the accumulated error after the
lapse of a relatively short approximate period (here 18 years or
223 months) is used to deternune the correction from step to step
(here a single synodic month).
The next column is column A of System B and gives the solar
velocity as described in our example of p. 110, column II. From
this is derived column B, containing the longitudes of the moon
and of the sun at conjunction or, for full moons, the longitudes
of the moon, the sun being 180 0 distant. In System A, column B
is derived without explicit mention of the velocity (column A)
because in System A only two velocity values are used, and
thus there was no reason to repeat them in a special column.
The subsequent columns, C and D and variants, give the
length of daylight or night corresponding to the solar longitude
 Babylonian Astronomy 117
of column B. The functions C and D are computed according to
independent arithmetical schemes designed to represent quan-
titatively the variation of the length of daylight during the year.
The underlying problem is one of spherical trigonometry but it
was solved here by arithmetical devices similar to the approx-
imations of a sinusoidal curve by a linear zigzag function.
The two following columns, E and lJI, describe the variations
of the latitude of the moon and the magnitude of eclipses. As
we have remarked previously, the consecutive lines of an ephem-
eris refer to the consecutive conjunctions or oppositions. If the
latitude of the moon is known for these moments, one is able
to judge the possibility of an eclipse and to compute, if necessary,
its magnitude. The latitude itself is again found by means of
zigzag functions. The "eclipse magnitude" is expressed in a
slightly different way than is customary today, but it is easy to
transfer it directly into a measure for the depth of immersion of
the lunar disc into the shadow. It is interesting to see that this
quantity was computed in many ephemerides for every month
and not only for every sixth (or perhaps fifth) month when an
eclipse is possible. In other words, a method had been developed
for computing "eclipse magnitudes" as a function of the latitude
such that the numbers obtained gave the size of the eclipse
correctly for real eclipses. For non-ecliptic conjunctions, how-
ever, these values behave exactly as if the distance from the
shadow was introduced as eclipse magnitude, allowance being
made for negative distances if the shadow is not reached, positive
values giving the depth of immersion for a real eclipse. This
shows a remarkably abstract attitude in the Babylonian pro-
cedure, which unhesitatingly introduces quantities for purely
mathematical convenience, in principle very much the same as
the use of complex numbers in modern mechanics.
The next column, P, gives the variations of the lunar velocity
in a form similar to column A for the solar velocity. In column G
we find the length of the synodic months under the assumption
of a constant solar velocity but a variable lunar velocity as
indicated by column F. At the beginning of our discussion we
had to make the assumption that consecutive lines represented
mean conjunctions, separated by the mean length of a synodic
month. This mean length would be produced by the conjunctions
118 Chapter V

of a sun and a moon, each moving with its own mean velocity.
In column G this assumption is partially abolished insofar as
only the sun is moving with its mean velocity and the answer
is given to the question how much a given variation in the lunar
velocity influences the spacing between consecutive conjunctions.
Obviously G will show the same period as F; the value of G will
be small and the month will be short if the moon moves fast,
i. e., near the maximum of F. This is indeed the relation between
F and G.
The next step, J, gives the necessary corrections to G because
of the variable solar velocity. In System B, column J is a dif-
ference sequence of second order due to the fact that column A
is a linear zigzag function. Here it becomes evident why the
inventor of System A preferred to assume a simple step-function
for the solar velocity; the corrections for variable solar velocity
are much more complicated in System B than in System A. After
the correction J has been found, the algebraic sum K of G and J
gives the length of the synodic month as it results from the
variability of both sun and moon. If the moment of one con-
junction is known, one need only add to it the amount of K
found for the length. of the following SYnodic month and one
obtains the moment for the conjunction of the next month.
Actually a slight complication is introduced here by the use of
the Babylonian calendar, which requires that the beginning of
a day be counted from actual sunset and not from midnight.
Hence a correction for the transformation from midnight epoch
to evening epoch is required. This can be done easily by means
of columns C and D which give us the length of daylight or night.
After this transformation is carried out, we obtain in column M
the dates and moments of all consecutive conjunctions referred
to sunset. Thus the first goal of the lunar theory has been reached:
the moments of the actual conjunctions or oppositions are known.
51. For the computation of eclipses no more information is
needed than has been collected thus far. We have in column M
the time of the conjunction or opposition expressed in its relation
to sunset or sunrise. From column '1', we know the distance of
the moon from the shadow. The ephemerides and eclipse tables
show with full clarity that one knew that solar and lunar eclipses
were subject to the same conditions, namely, sufficiently small
 Babylonian Astronomy 119

latitude near new or full moon. The problem of determining

these moments and of describing the motion in latitude was
solved very successfully by means of arithmetical methods. Con-
sequently one obtained quite satisfactory results for the prediction
of lunar eclipses (cf. Fig. ttl». For a solar eclipse we should
know more; specifically, we should be able to judge whether the
vertex of the shadow cone touches our particular locality, assuming
all other circumstances are favorable for an eclipse. This problem
can be solved only if sufficiently accurate information about the
actual distances of sun and moon from the earth are available,
together with a correct knowledge of the relative sizes of these
bodies. There is not the slightest reference to any of these quantities
in Babylonian texts. Tables for solar eclipses are computed
exactly like the tables for lunar eclipses with no additional
columns corresponding to "parallax", i. e., quantities depending
on the above-mentioned distances and sizes. Consequently the
Babylonian texts do not suffice to say anything more than that a
solar eclipse is excluded or that a solar eclipse is possible. But
they cannot answer even approximately the question whether a
possible solar eclipse will actually be visible or not. One has to
remember that this is the state of affairs during the last period of
Mesopotamian astronomy, from about 300 B.C. to O. Before
300 B.C. the chances for the correct prediction of a solar eclipse
are still smaller. At all periods, exclusion of an eclipse of the sun
is the only safe prediction that was possible.
5J. The remaining part of the ephe~erides concerns the
fundamental problem of the lunar calendar: to determine the
evening of first visibility after conjunction when the new crescent
again becomes visible. We have already discussed (p. 107) the
three major factors which determine the visibility of the new
crescent at a sunset following conjunction, namely, elongation,
variable inclination between ecliptic and horizon, and latitude
of the moon. It is exactly these three quantities which are found
in the columns 0, Q, and R of ephemerides of System B. Column 0
for the elongation is preceded by a column N which gives the

1) Fig. 11 illustrates the results obtained for the magnitude of lunar eclipses,
expressed in digits such that 12 means totality. The omission of modern values
indicates that no eclipse would have been visible, according to modern computation.
The same is indicated by ancient values ~ o.
 120 Chapter V

I
'.
20~·,::-----t------+------+------J-----~,
.' "r
-'L -rL-'\ ~.,

I: ~~~-=- ·c~ ~'·-"""-L--=·:a-q-I'- '

i'··~~'·"., ,?--\.
__
.
10 20 25

-----+------+-----.,..~.......------+---~..,.---I-20
..;'~' ,.... \
---A--,--,~-.+---)- \. r-\~---+IO
{-t----...po...='--.,.--4-~y Col- 0

30 35
. _.•"elen! •••••-tnocl.rft
Fig. 11.

time difference between the moment of conjunction and the sub-

sequent sunset at which the fIrst crescent might be expected. For
this particular evening one computes how long the new crescent
will be above the horizon after sunset. If the resulting time
difference between sunset and the setting of the moon is long
enough to secure visibility, then the initial guess was right and
the evening which starts the new month is known. If the resulting
value seems too high, the computation has to be repeated for
one day earlier. If the first result seems too low, a new value
must be found for 24 hours later. In some cases alternative results
are recorded in the final column P, corresponding to either a
29·day month or a 30-day month.
Though the general outlines of this part of the lunar theory
are clear, many details are still obscure, chiefly because of the
difficult terminology of the procedure texts and the rounding-off
of the numbers involved in the actual ephemerides. An added
difficulty results from the fact that the ephemerides of System A
do not give any of the columns N, 0, Q, and R but list only
the final result P. Nevertheless, a few additional facts can be
 Babylonian Astronomy 121

established. First of all, it is clear that the determination of the

influence of the variable angle between ecliptic and horizon is a
problem of spherical trigonometry, and the same holds for the
influence of the lunar latitude. Exactly as in the case of the
length of daylight, this problem was solved by means of fixed
arithmetical schemes. The procedure texts give lists of coefficients
by which the elongation has to be multiplied in order to obtain
for different solar longitudes the proper amount of difference in
time for setting, and a similar device is followed for the latitude.
The main difficulty for us consists in discovering on what grounds
the decision was made as to whether a given value in the final
column P was sufficient for visibility or not. It seems as if nol
P alone had been used, but the sum of the elongation 0 and the
value of P. Indeed the brightness of the new crescent depends
essentially on the width of the illuminated sickle of the moon,
and this width is proportional to the elongation. Thus it is reason-
able to say that even a small value of P, caused by closeness of
the moon to the horizon, might be compensated for by a greater
brightness of the sickle, and, vice versa, a very small sickle
might not be visible even at a relatively great distance from the
horizon. Thus the sum of 0 and P is indeed a very reasonable
parameter to be used as a criterion of visibility.
53. Concluding our summary of the lunar theory, we must still
mention the texts which concern the daily motion of the sun and
the moon. Indeed, there exist ephemerides which give the longi-
tude of the sun from day to day, assuming a constant mean
velocity of 0;59,9° which is slightly too high a value.
Similar ephemerides also exist for the moon, though under
the assumption of a variable lunar velocity. This variation is,
as usual, expressed in the form of a linear zigzag function. The
mean velocity is assumed to be 13;10,35° per day, a value which
appears again and again in ancient and medieval astronomy.
The extremal values are m = 11;6,35° and M = 15;14,35° from
which one derives a period
4,8
P =9 = 27;33,20 days.

This indicates that one "anomalistic" month is given the length

of 27;33,20 days or, expressed as a relation between integers,
122 Chapter V

that 9 anomalistic months contain 248 days. This relation is not

quite accurate, as can be shown by comparison with the pre-
viously discussed ephemerides. In the latter we find that column F
for the lunar velocity is based on the relation
4,29 anomalistic months = 4,11 synodic months.
Substituting in this equation the value 27;33,20 days for the
length of the anomalistic month we obtain for the synodic month
a value close to 29;31,54 days. But from column G one derives
for the mean synodic month the length of 29;31,50,8,20 days,
which is again one of the classical parameters of ancient and
medieval lunar theory. Hence it is clear that 27;33,20 is slightly
too high a value, caused by the desire to obtain conveniently
short numbers for the parameters of the zigzag function for the
daily motion. We shall come back to this remark in our last
chapter (p. 162).
54.. Before describing the Babylonian planetary theory, we
shall discuss the main features of the apparent movement of
the planets from a modern point of view. We know that the
planets move on ellipses around the sun, the earth being one of
them. We shall derive from these facts the apparent motions as
seen from the earth. In order to simplify our discussion, we shall
replace all orbits by circles whose common center is the sun.
The eccenbicities of the elliptic orbits are so small that a seale
drawing that would fit this page would not show the difference
between the elliptic and the circular orbits.
We utilize furthermore the fact that the dimensions of our
planetary system are so minute in comparison with the distances
to the fixed stars which constitute the background of the celestial
sphere that we commit no observable error at all if we keep
either the sun or the earth in a fixed position with respect to the
surrounding universe. Hence we will proceed in the following
way.We shall start with the circular motion of the planets around
the sun and then keep the earth fixed and ask for the resulting
motion with respect to the earth. This will answer our question
concerning the planetary phenomena.
The first step is absolutely trivial. We know that the earth is
a satellite of the sun, moving around it once in a year. In order
to obtain the appearances seen from the earth we subtract from
 Babylonian Astronomy 123

Fig. 12a-b.

all motions the motion of the earth. Thus we see that by arresting
the motion of the earth we obtain the appearance that the sun
moves around the earth once per year. Its apparent path is called
the ecliptic (cf. Fig. 12a and b).
Secondly we consider an "inner" planet. Mercury or Venus.
which moves closer to the sun than the earth (Fig. 13a). If we
stop the earth we need only repeat Fig. 12 in order to obtain
again the motion of the sun. The orbit of the planet remains a
circle with the sun in its center. Hence the geocentric description
of the motion of an inner planet is given by a planet which moves
on a little circle whose center is carried on a larger circle whose
center is the earth. The little circle is called an "epicycle". the
large circle is the "deferent".

b.
Fig. 13a-b.

Finally we have an "outer" planet. Mars. Jupiter. or Saturn.

whose orbit encloses the orbit of the earth (Fig. 14a). From the
earth E the planet P appears to be moving on a circle whose
center S moves around E. Thus we have again an epicyclic
motion (Fig. 14 b). In order to establish a closer similarity with
the ease of the inner planets we introduce a point C such that
the four points S. E. p. and C always form a parallelogram.
SP is the radius of the planetary orbit; because EC = SP we
124 Chapter V

......-:"0

" ·'C

b.

c.
Fig.14a-e.

see that C lies on a circle with center E. Similarly ES is the

radius of the solar orbit, and, because ES = CP, we see that P
lies on a circle around C. Thus the planet P moves on an epicycle
whose center C travels on a deferent whose center is E (Fig. 14c).
Thus we have established an exact analogue to the case of the
inner planets. In both cases the planet has an epicyclic move-
ment. In the case of the inner planets the center of the epicycle
coincides with the sun. For the outer planets the center C of the
epicycle moves around E with the same angular velocity as the
planet moves around the sun, while the planet P moves on the
epicycle around C with the same angular velocity as the sun
moves around the earth.
In order to avoid misunderstandings, I shall repeat once more
the assumptions upon which our above results rest. These assump-
tions were (a) that the planetary orbits are circles with the sun
in their common center; (b) that all planetary orbits lie in the
same plane. Accepting these two assumptions we have seen that
the planetary orbits with respect to the earth consist of epicycles
whose centers move with uniform velocity on deferents having
the earth as center. In other words, if we disregard the small
eccentricities of the planetary orbits, and if we also neglect the
small inclinations of these orbits, then the epicyclic motion gives
a correct description of the planetary orbits with respect to the
 Babylonian Astronomy 125

earth. Indeed it is only a matter of mathematical convenience

whether one computes first the longitudes of the earth and the
planets heliocentrically and then transforms to geocentric co-
ordinates, or whether one carries out this transformation first
and then operates with epicycles.
For a finer theory of the planetary phenomena the above
assumptions are too crude. It is easy, however, to see in what
directions one should move in order to reach higher accuracy.
The eccentricity of the orbits can be taken into consideration by
assuming slightly eccentric positions of the earth with respect to
the centers of the deferents. The latitude can be accounted for
by giving the epicycles the proper inclination. Both devices were
followed by the Greek astronomers.
56. We have now seen that the planets move with respect to
the earth on epicycles. This makes it particularly simple to
understand the main features of the planetary motions as seen
from the earth.We begin again with an inner planet. Its angular
velocity about the center S of its epicycle (cf. Fig. 15) is greater
than the angular velocity of S about the earth E. If the planet P
is on the part of its epicycle which is removed from the earth,
the motion of P is added to the motion of S and the planetary

Fig. 15.

motion appears greater than the motion of S. We call this the

"direct" motion. Between A and B, however, the planet moves
backward faster than its epicycle is carried forwardl ), thus it
appears to be "retrograde".
The same figure allows us also to describe the visibility con-
ditions. If the planet P and the sun S are seen in the same, or
1) It is easy to see that the points A and B lie somewhat inside the two points
where the lines from B are tangential to the epicycle.
126 Chapter V

Fig. 16.

in nearly the same, direction from E, the planet is invisible

because of the brightness of the sun. Thus a certain "elongation"
of the planet from the sun is required to make the planet visible.
Fig. 16 shows that the arc of invisibility between X and S near
"superior conjunction" is much greater than between {J and r
(near "inferior" conjunction). The visible arc from r to X rises
before the sun; thus the planet is "morning star". The arc from
B to {J sets after the sun; thus the planet is "evening star". Fig. 17
describes the same phenomena once more in a graph with the
abscissa representing time whereas the ordinates represent geo-
centric longitudes. The straight line represents the motion of
the sun.
In similar fashion one obtains for an outer planet a graph as
given in Fig. 18. Now the motion of the planet is slower than
the motion of the sun. Retrogradation occurs near opposition, 9,
when the sun and planet are seen in opposite directions from the
earth. Consequently the retrogradation of an outer planet is fully
visible in contrast to that of an inner planet, where a part of the
retrograde motion becomes invisible near inferior conjunction.
An outer planet becomes invisible only once in each cycle: near
conjunction, {J to r. The points 4) and '1', where direct motion
changes to retrograde motion and vice versa, are called the "first"
and "second" stationary points respectively.
56. It is in the theory of the planets that the contrast between
the Babylonian approach and Ptolemy's theory as presented in
the Almagest becomes most visible. In the Ptolemaic theory a
definite kinematic model is assumed, based on epicyclic motion,
which closely corresponds to the description of the planetary
 Babylonian Astronomy 127

~A.1(,;
-I,
0< ..::., "
, '
~/ J'
. ~-, " 'I:.
IIll.. "

.... /"!
~ (
~6 ~Uthq,tf
360

':' J~r
,,' 0:<{".S-~
- ,• , ,
. ., "
.N
-

, ".
"
,-'II.
1:
Fig. 17. Fig. 18.

motion given in the preceding sections. Thus the geocentric

longitude of the planet can be computed for any given moment t.
It is then a secondary problem to determine those values of t
for which the planet is in one of the characteristic phenomena
which we denoted by Greek letters.
The Babylonian method follows the exactly opposite arrange-
ment. The f11"st goal consists in determining the "Greek-letter
phenomena", and thereafter the longitude of the planet for an
arbitrary moment t is found by interpolation.
This difference in approach is, of course, the result of the
historical development. The Babylonians were primarily in-
terested in the appearance and disappearance of the planets in
analogy to the first and last visibility of fixed stars-e. g. Sirius-
and of the moon. It was the periodic recurrence of these phenom-
ena and their fluctuations which they primarily attempted to
determine. When Ptolemy developed his planetary theory, he
had already at his disposal the geometrical methods by means of
which the solar and lunar anomalies were explained very satis-
factorily, and similar models had been used also for an at least
qualitative explanation of the apparent planetary orbits. Thus it
had become an obvious goal of theoretical astronomy to otter a
strictly geometrical theory of the planetary motions as a whole
and the characteristic phenomena lost much of their specific
interest, especially after the Greek astronomers had developed
enough observational experience to realize that horizon phenom-
ena were the worst possible choice to provide the necessary
empirical data.
128 Chapter V

57. Whatever phenomenon the Babylonian astronomers wanted

to predict. it had to be determined within the existing lunar
calendar. Suppose one had found that a planet would reappear
100 days from a given date. What date should be assigned to
this moment? Obviously one should know whether the three
intermediate lunar months were. perhaps. all only 29 days long.
or all three were 30 days long, etc. This question could be an-
swered perfectly well by lunar ephemerides whose goal it was to
determine whether a given month was 29 or 30 days long. But
planetary phenomena proceed very slowly. One single table for
Jupiter or Saturn could easily cover 60 years and more. To
determine calendar dates so far in advance would have meant
the computation of complete lunar ephemerides for several
decades. Furthermore. the actual computation of the planetary
motion had to be based. in any ease. on a uniform time seale.
All these difficulties were at once overcome by a very clever
device. One used as unit of time the mean synodic month and
divided it in 30 equal parts. The Babylonians seem not to have
had a special name for these units. referring to them simply as
udays". Modern scholars have used the term "lunar days"; I
shall use the corresponding term of Hindu astronomy. namely.
utithi".
The fact that the Babylonian calendar was a strictly lunar
calendar has the effect that the total duration of a number of
calendar months will not deviate more and more from the cor-
responding total of mean SYnodic months. Dates expressed in
tithis will never be far off from real calendar days. usually not
more than ±1 day. Thus the Babylonian astronomers in their
computations simply identified the results given in tithis with the
dates in the real calendar. This is the standard procedure for all
planetary texts.
The use of tithis implies that one did not try to reach. for the
planetary phenomena. the same accuracy which was obtained in
the lunar theory. While one went to great lengths to determine
all possible influences upon the first and last visibility of the
moon. we find no similar devices used for the planetary phe-
nomena. The latitude of the planets. for instance. is nowhere
taken into consideration for the planetary ephemerides. On the
other hand. several concurrent Usystems" are used simultaneously,
 Babylonian Astronomy 129

as we have seen also in the case of the two systems of the lunar
theory. The different systems of the planetary theory are obviously
modeled after the two main systems of the lunar theory. They
either operate with step functions (type "A") or with linear zigzag
functions (type UB"). The variations within ephemerides of type A
consist in the use of different numbers of steps for each period.
There exists. for instance. a theory of Jupiter with only two
ecliptical zones of different velocity. while another method
operates with four zones. two intermediate steps being inserted
between the extremal values of the previous model. The variations
of texts of type B consist in small changes (rounding-off) of
parameters; similar variations are also known in System B of
the lunar theory.
The basic idea of all planetary ephemerides is, however. the
same. It consists in the separate treatment of each characteristic
phenomenon by itself as if this phenomenon were an indepen-
dent body moving in the ecliptic.
Let us consider. as an example. the first appearance r of
Mercury as a morning star. We assume that we are given (by
observation or by previous computation) the moment to and the
longitude Ao of Mercury when it again became visible in the
morning after a period of invisibility at inferior conjunction (cf.
Fig. 16 and 17 p. 126 f.). We call this point r o in our diagram
Fig. 19. If both the sun and Mercury would move with constant
velocity and if this motion would fall in the equator. then sub-
sequent morning appearances I. r r.. ... would be spaced
equidistant in the diagram. keeping a fixed distance from the
graph of the solar motion. Actually. however. these assumptions
are not satisfied. Therefore the spacing of the points r o• r l , r ll•
. . .. shows periodic irregularities. The Babylonian theory tries
to describe these irregularities precisely in the same fashion as
the solar and lunar theory described the variable velocity of
these bodies. Hence for an ephemeris of type A the ecliptic is
divided into zones such that the progress of the phenomenon r
in each zone is given by the same amount, with discontinuous
changes at the boundary.
For Mercury and r we have three zones with discontinuities
at n 1. 1116. and n o. Suppose that r o is given to be in n 17.
The velocity in the zone which stretches from n1 to 11 16 is
130 Chapter V
,
-':'3 '
,
r:3
o ,,' 0

, ,,
-=-2. "
0,'
,
0 J;
,,
-
-,
,
"0 ~
o, " ,
,
,,
" 0 r;

Fig. 19.

1,46°. Consequently r 1 will be in eft 17 1,46 = en 2,3 = t 3. +

This point still lies within the same zone; thus we again add
1,46° and obtain t 1,49 = )( 19. In this step we have crossed,
however, the boundary l1 16 and have entered the second zone
by an amount of 1,3°. In the second zone the velocity is no
longer 1,46 0 but 2,21;20° or 1 greater than the previous velocity.
Thus we have to raise the arc of 1,3° also by 1 of its amount,
i. e., by 21°. Thusr. will not be )( 19 but )( 19 21 = C'f 10. +
In the same fashion, subsequent positions providing us with
all longitudes of r l , raJ .... can be found. The corresponding
dates, expressed in tithis, are determined by a simple rule which
makes the time differences linearly dependent upon the differences
in longitude. Thus two columns of an ephemeris can be computed,
one giving the longitudes, the other the dates for consecutive r's.
How one determined the parameters which characterize the
distribution of the r's is a difficult question which cannot be
answered completely. Only this much is easy to see: some counting
of the number of r's must have been made such that an integer
number of first appearances of Mercury as morning star cor-
responds to an integer number of years. The above mentioned
parameters are based on the relation that 848 years contain 2673
risings of Mercury. Exactly as in the lunar theory, no historical
conclusions can be based on these numbers. The size of the
zones, the particular zonal velocities and their ratios must be
comparatively handy numbers, and the period relation derived
from these numbers reflects nothing more than the final com-
promise between empirical facts and computational require-
ments.
 Babylonian Astronomy 131

The next step in the computation of the phenomena of Mercury

consists in finding longitudes and dates for all consecutive fIrst
appearances as evening star beginning with a given point So.
The principle of the procedure is perfectly analogous to the
procedure for the r's, but the zones and velocities are different.
The discontinuities are now located at ~ 26, )( 10, and sm 6
and the velocities are 1,46;40°, 1,36°, and 2,40°, as compared
with 1,46°, 2,21;20°, and 1,34;13,20° in the previous case. The
period relation is now expressed by the equivalence of 480 years
and 1513 risings of Mercury. This drastically illustrates our
previous remark that no historical conclusions can be drawn
from these relations because it is obviously absurd that the ob-
servations of the r's should extend for centuries farther back
than the observations of the S's. It is furthermore clear that the
two periods should be identical because every r must always
be followed by exactly one S, and vice versa. This fact was of
course evident to the Babylonian astronomers and the two periods
deviate from each other only as 3;9,7,38,.. from 3;9,7,30. It is
only the adjustment of the determining parameters of the zones
which causes an apparent discrepancy, the effect of which was
negligible in practice.
The preceding steps provide us with all r's and S's. We still
have to find the E's and D's, that is, the corresponding settings
of Mercury. One might expect to find two additional schemes
which yield these data in the same fashion as the r's and S's
were found before. This is, however, not the case. Kugler sus-
pected from the fragmentary ephemerides at his disposal that
the E's were computed from the r's by means of fixed additive
amounts depending only on the longitude of F; and, similarly,
the D's from the S's. Kugler's hypothesis has been fully con-
firmed by texts from Uruk. Tables were computed which give
for every degree of every zodiacal sign the amount of longitude
and the number of tithis which must be added to a given r
(or B) in order to find the subsequent E (or D). In other words
we have fixed arithmetical schemes which determine as function
of 1 the relationship between consecutive risings and settings of
Mercury.
Fig. 20 illustrates the graphical representation of one of these
curvesl ). One point needs special emphasis. The region near the
1) The longitudes of D and S.
132 Chapter V

/ ~
30
( 1\ ~

20
J \ ~
i\---- / '
ID

o
Fig. 20.

minimum indicates that in this part of the zodiac the correspond-

ing disappearance should follow very shortly after the preceding
appearance. The text declares this whole zone as cases of in-
visibility. In other words, for this region the planet is never
visible at all though one "rising" and one "setting" should be
theoretically counted in order to maintain the correct period
relation. In the representation of our graph (Fig. 17 p. 127) there
exists a zone of the zodiac where one whole bulge remains so
close to the sun that it should be dotted as invisible.
It is interesting to remark that the most up-to-date modern
tables for the computation of risings and settings of Mercury are
based on ephemerides which contain dates and longitudes for
actually invisible phenomena. In the ancient texts these eases are
denoted by an ideogram LU whose significance became clear
only when the computation of the X's and D's was fully under-
stood. Without this knowledge, however, one had to assume an
extremely high visibility for the horizon in Babylon in order to
cover all recorded eases, not realizing that these records con-
tained invisible and visible risings alike. Consequently our
modern tables assume a much too high degree of visibility of
Mercury in Mesopotamia and therefore yield results which have
only a very general resemblance to the facts and are definitely
wrong in the critical cases for which they were computed.
58. The computation of ephemerides for the other planets
follows in general the ideas which we have illustrated in the case
of Mercury.
 Babylonian Astronomy 133

The discovery, among Pinches's copies (cf. p. 105), of several

planetary procedure texts and ephemerides disclosed the existence
of a much greater variety of methods in detail than we had
assumed on the basis of the material accessible to Strassmaier
and Kugler. The following is a short summary of the main
systems according to my present knowledge.
For all planets the central problem consists in establishing the
variations in the amount of the synodic arcs between phenomena
of the same kind as function of the ecliptic. If the synodic arcs
are known, then the synodic times are easily found since they
are the time required for the sun to travel between two consecutive
phenomena of the same kind. Though some correction had to be
applied to this basic principle, we need not discuss the details
here. For the longitudes, however, the variety of methods is of
real interest since it shows that great efforts were made to describe
properly the appearances of the planetary motion.
As is to be expected, the most regular variation in the SYnodic
arcs is found in the case of Saturn. We know of two systems, one
proceeding with two zones of constant SYnodic arc ("System AU),
one with synodic arcs which vary according to a linear zigzag
function ("System B U). Both systems concern all five phenomena:
r, tP, 9, W, and D.
For Jupiter we know of two closely related procedures following
System B, for the same five phenomena as in the case of Saturn.
We also have, however, a variety of methods of type A: two
zones, four zones. three zones, and variants as to location and
amount of the discontinuities. Only a fraction of these systems is
attested in actual ephemerides, though they are relatively numerous
for Jupiter.
While there is, in principle, only little difference between the
theory of Jupiter and that of Saturn, we find a different procedure
followed in the case of Mars. We know of a six-zone System A
(all zones being two zodiacal signs long), but applied only to
first and last appearance, rand D, and to first station, tP. For
the retrograde are, however, four different methods are known
for the determination of the amount of the retrograde arc from tP
to opposition 8, and probably equally many existed for the arc
from 9 to W. In other words, we have here a situation similar to
the procedure that we described in the case of Mercury. After the
134 Chapter V

position of tP has been found by System A, the position of 8

and 1p are determined by means of pushes whose (negative)
values depend on the longitude of tP. The position of the sub-
sequent phenomenon, D, is entirely independent of the size of
the preceding retrogradation.
Venus is treated quite differently. Here one utilizes fIrSt of all
the fact that in eight years Venus completes five synodic periods
which, in the mean, are only 2l ° short of the initial position in
the zodiac. Now for each of the six characteristic phenomena
(H, 'P, D in the evening, r, tP, E in the morning), numerical rules
are given which indicate the sYnodic ares and times in a sequence
of five consecutive phenomena of the same kind. Thus we are
given a table of 5 times 6 or 30 pushes for the longitudes and
equally as many for the times. The total of 5 pushes for longitudes
is such that it results in a deficit of 2;30°, which is what it should
be for each cycle. Similarly, the dates recede by 4;10 tithis after
the 99 lunar months which correspond to 5 sYnodic periods. This
general procedure is known in at least two variants, but no process
of type A or B is attested so far. Much of the detail escapes us
since our sources concerning Venus are particularly fragmentary.
For Mercury we have described in detail a procedure following
System A for the appearances rand H while the disappearances
E and {J were found by means of pushes depending on the
longitudes of the preceding appearances. In other words, the
lengths of the ares of visibility are prescribed as function of the
longitude of the first appearances. We also know of the existence
of a complementary system: the disappearances E and Dare
computed first according to a System A (though using zones which
are different from the previous ones) and the ares of invisibility
are prescribed as pushes E -+ Hand D -+ r respectively. In a
general way, the results are equivalent to the previous ones, but
are different in detail.
For all planets we now know of a variety of methods for
computing from one given phenomenon all subsequent phenomena
of the same kind. The question arises as to how the initial values
were chosen. One could, in principle, assume that one set of
phenomena within one SYnodic period was determined by obser-
vation, e. g., longitudes and dates of r,
tP, 9, 'P, {J for Saturn or
Jupiter, r, tP; {J for Mars, etc. We know that the Babylonian
 Babylonian Astronomy 135

astronomers were keenly aware of this problem and tried to

develop rules which allow us not only to compute consecutive
phenomena of the same kind. but also to pass from one phenom-
enon to its neighbor of another kind. This problem is by no means
trivial since in an ephemeris of the type
ro 4)0 8 0 Y'o Do
r1 4)1 9 1 "PI D 1
r2 4). 8. Y'2 D.

all rows are determined as soon as the first row is given. The
rules in each later line are thus a eonsequenee of the rule which
determines the relations in the first line. In other words. one
must cheek the consistency of a rule which leads from a r to 4)
to ... D with the previously given methods of computing vertically.
This problem has been correctly solved in some eases. in other
cases the rules are clearly intended only to serve as approxima-
tions. In all eases one sees. however. the tendency to restrict the
empirical data to a minimum. Indeed for a consistent system of
rules for computing in lines as well as in columns one single
value suffices for computing all following phenomena. This is
the ideal of a mathematical astronomy of the purest kind.
For all planets there arose the final problem of describing
their daily motion. We have now in principle reached the know-
ledge of all longitudes and dates for the typical phenomena.
Using our graphs we can say that we know the positions of all
points whieh we have denoted by Greek letters. The problem
remains to determine the intermediate curves. Though only few
texts are preserved which permit an investigation of this problem.
we at least know that interpolation schemes were devised such
that one could start from one given value and reach in a number
of steps. given by the difference in date. the next characteristic
value. These interpolation schemes are built upon difference
sequences of second or even third order. Using modern termin-
ology. one may say that one determined simple polynomials
which satisfy with sufficient accuracy the conditions which are
expressed by the relative position of consecutive characteristic
points in our graph of the planetary motion. One can only admire
the elegance and skill which is reflected in all these arithmetical
136 Chapter V

methods. We are still far from a full appreciation of them since

we know so little about the underlying empirical material which
was so skilfully applied to provide the basic parameters of a
real mathematical theory.
69. It is natural to ask who were the astronomers who de-
veloped and used this theory. I see no way of answering such
a question satisfactorily. One can do no more than enumerate
the few facts that we know. The texts from which all our infor-
mation comes were parts of two archives, one in Uruk, one in
Babylon. There is no proof against the existence of other archives
and we are unable to judge the relations between two or more
centers of astronomical activity. We know very little about the
Babylon archive, because the Babylon texts rarely have colophons.
Thus we are almost completely dependent upon the colophons
of the Uruk texts. These colophons follow more or less the fol-
lowing pattern: "Tablet of A, son of B, son of C, descendant of
M; hand of ( = written by) R, son of S, son of T, descendant of N.
Uruk, month m, day d, year y (of the Seleucid era), X being
king". Many tablets contain an invocation at the beginning:
..According to the command of the deities Anu and Antu, may
it go well". Some colophons add a curse against whoever removes
the tablet which was written by the scribe, it is said, "for the
prolongation of his days and for the well-being of his posterity",
and occasionally we read that "the informed may show the
tablet to the informed but not to the uninformed".
The investigation of the kinships mentioned in the colophons
allows us to establish two scribal families which were engaged
in writing ephemerides or were their owners (if this is the meaning
of the phrase "tablet of"). One family mentions Ekur-zikir as
their "ancestor", a man who is given the title of "mashmash-
priest of Anu and Antu of the Resh sanctuary, scribe of (the
series) Enu.ma-Anu-Enlil, from Uruk". The "series" mentioned
is the famous series of astronomical omens mentioned in the initial
sections of this chapter. The second family has Sin-leqe-unnini
as its ancestor, "scribe of Enuma-Anu-Enlil, kalu-priest of Anu
and Anto, from Uruk". Both these "ancestors" are known from
colophons of other tablets of the Seleucid period and the question
to what extent these scribal families were real families or merely
scribal schools seems undecided. Also the significance of owner-
 Babylonian Astronomy 137
ship and scribe escapes us. We do not know, e. g., whether the
"scribe" of an ephemeris was its actual computer or not. All that
one can safely say is that our tablets came from "priestly"
circles, but this says little more than the trivial statement that
they were written by professional scribes. And no information
about the origin of these methods can be obtained from the
colophons of the Uruk tablets.
The Babylon texts give us still less information about their
scribes. From Pliny, Strabo, and Vettius Valens, however, are
known names of three Babylonian astronomers who seemed also
to appear in the colophons of our texts. One, Sudines, seemed
--
to be contained in the second half of Anu-ahhe-su-iddina, but the
latter turned out to be a misreading of Anu-aha-usabsi. The
second name, Naburianos, seems to be attested o~e, in doubtful
context in one of the very latest tablets, in the form Naburimannu
but the reading itself is not really certain. And there is still less
proof that Naburianos is mentioned as the inventor of the lunar
System A to which the text belongs. Finally there is the name of
Kidenas which corresponds to cuneiform Kidinnu. This name
appears in a few colophons in the connection "tersitu of Kidinnu"
which was guessed to mean "lunar tablet of Kidinnu" or "system
of Kidinnu" and thus Kidinnu is usually considered to be the
inventor of System B. This may be so, but real proof is missing.
The term tersitu is a complete puzzle in this context because it
is otherwise known to denote some tools or ingredients in con-
nection with the manufacturing of glazed bricks.
Attempts were made to give accurate dates for the invention
of the lunar theory. They were based on a comparison of modern
computation with the results of the ancient theory. The slowly
accumulating error of the ancient theory was supposed to be
zero at the beginning, and this led to the alleged date of the alleged
inventors Naburianu and Kidinnu. It suffices to say that this
method presupposes the accuracy of the initial values, a hypo-
thesis which is far from even being plausible. It is furthermore
assumed that the parameters used in the actual computation of
the ephemerides are exactly identical with the empirical values
or, at least, with the values theoretically abstracted as correct
from some observations. But we have seen that the parameters
of the ephemerides were adjusted for the purpose of convenient
138 Chapter V

computation. The errors caused by this procedure are very small:

nevertheless they influence quite essentially the results of com-
putations which are themselves based on the investigation of the
small deviation from the factual motionlJ. Hence there is no hope
of obtaining, in this way, accurate information as to the date of
invention of mathematical astronomy. For the time being, we
must be satisfied with general historical considerations, however
inconclusive they may appear. Otherwise one can only hope that
a tablet may be found (and perhaps even published) which gives
us direct information about the theoretical and empirical founda-
tions of the whole theory.

BIBLIOGRAPHY TO CHAPTER V
All texts known to me by 1955 which concern mathematical astronomy are
published in O. Neugebauer, Astronomical Cuneiform Texts, London, Lund
Humphries. 1955 (3 vols.). This edition contains complete transcriptions.
translations, and commentaries. Henceforth quoted as ACT.
Copies of all other available texts from the British Museum with much
information about unpublished material are given in T. G. Pinches-J. N.
Strassmaier-A. J. Sachs. Late Babylonian Astronomical and Related Texts.
Providence. Brown University Press, 1955.
For a modem comprehensive discussion of the role of divination and astrology
see J. C. Gadd. Ideas of Divine Rule in the Ancient East, The Schweich Lectures
of the British Academy 1945 (London 1948).
The reader should be warned against the use of Jeremias, Handbuch der
altorientalischen Geisteskultur. With the use of an enormous learned apparatus,
the author develops the 'panbabylonistic" doctrine which flourished in Germany
between 1900 and 1914. only to be given up completely after the first world
war. The main thesis of this school was buDt on wild theories about the great
age of Babylonian astronomy, combined with an alleged Babylonian "Welt-
anschauung" based on a parallelism between "macrocosm and mierocosm".
There was no phenomenon in classical cosmogony. religion, literature which
was not traced back to this hypothetical cosmic philosophy of the Babylonians.
A supreme disregard for textual evidence, wide use of secondary sources and
antiquated translations, combined with a preconceived chronology of Babylo-
nian civilization, created a fantastic picture which exercised (and still exercises)
a great influence on the literature concerning Babylonia. Kugler was one of the
few scholars in Germany who did not fall for these theories. In a little book
called "1m Bannkreis Babels" he demonstrated drastically the absurdities which
can be reached by the panbabylonistie methods. He collected 17 pages of
striking parallels between the history of Louis IX of France and Gilgamesh,
showing that Louis IX was actually a Babylonian solar hero.
 Notes and References 139

The panbabylonistic sehool no longer has any followers. But it seems to me

that Kugler's example should be studied by every historian because it demon-
strates far beyond its original purpose how easy it is to fit a large body of evidence
into whatever theory one has decided upon.

NOTES AND REFERENCES TO CHAPTER V

ad 42. The latest dated cuneiform tablet (75 A.D., thus the time of Vespasian)
is an "Almanac" in the classification of Sachs (J. Cuneiform Studies 2, 1948,
p. 280). It was found in Dropsie College in Philadelphia and in all probabUity
came from Babylon. The exact date was established by Schaumberger.
ad 44. The story of the text from the "Frau Professor Hilpreeht Collection
of Babylonian Antiquities im Eigentum der UniversiUit J ena" (this title is
accurate) is somewhat peculiar, though not unique. Six lines from the reverse
were published in 1908 in the Sunday supplement of the Miinchner Neueste
Naehrichten. Rather fantastic interpretations were made on the basis of this
excerpt until Thureau-Dangin in 1931 suggested the explanation of the numbers
as meaning distances in depth. During all these years the text could not be
eheeked because it was "lost". In 1931, however, I got permission to have
access to the closely guarded Jena collection, whose rich material, incidentally,
furnished me with the key to the understanding of the relationship between
multiplication tables and division (el. above p. 31 f.). In going through this
collection I found a tablet with the label "One of the 5 important Nippur texts
from my desk drawer"; this turned out to be the lost text. Shortly thereafter,
I was informed by the authorities in Jena that it was only by mistake that I
had been admitted to the collection and that I was forbidden to publish any
text from this collection. Neverthe1ess I reserved for myself the privilege of
remembering my newly aequired knowledge, and since then my copy of the
text has been used by other scholars. The essential passages are discussed in
a review in Quellen und Studien zur Gesehichte d. Math., Sere B, vol. 3 p. 273ft.
(1936).
For the Old-Babylonian observations of Venus see Langdon-Fotheringham-
Schoch, The Venus Tablets of Ammizaduga, Oxford, 1928. The chronological
conclusions of this work have been disproved by subsequent arehaeological
evidence.
A comprehensive study of the series "Enilma Anu Enlil" was begun by
E. F. Weidner in the Arehiv fUr Orientforsehung, vol. 14 (1942) p. 172-195,
308-318 and vol. 17 (1954) p. 71-89. The reader will find in this paper a de-
scription of the very complex structure of this "series" with its supplementary
series of excerpts, commentaries, etc. It is important to realize that we have only
very few original texts with astrological omina from a period before late Assyrian
and Neo-Babylonian times. Consequently even the history of this early stage of
astrologiealliterature must largely be reconstructed from much later documents.
For the two tablets of the series umul Apin" see Bezold-Kopff-Boll,
Zenit- und Aequatoriaigestirne am babylonisehen Fixsternhimmel, Sitzungs-
beriehte d. Heidelberger Akad. d. Wiss., phil.-hist. Kl., 1913, No. 11, and
140 Chapter V

E. F. Weidner, Ein babylonisehes Kompendium der Himmelskunde, Am. J.

of Semitic Languages and Literatures 40 (1924) p. 186-208. Cf. furthermore
B. L. van der Waerden, Babylonian astronomy II. The thirty-six stars. J. of
Near Eastern Studies 8 (1949) p. 6-26; also part III: The earliest astronomical
computations (ibid. 10, 1951, p. 26-34).
The number of attested leap years in Babylonian texts is now sufficiently
great to show that the 19-year cycle was introduced into consistent calendaric
use very close to 380 B.C. This gives Meton's announcement of the cycle in
Athens a priority of about 50 years and opens the possibility of an originally
Greek discovery. On the other hand, A. Sachs has put forward (J. Cuneiform
Studies 6, 1952, p. 113) arguments which connect the Babylonian intercalation
rules with observations of the heliacal risings of Sirius prior to 380 B.C.
It may be remarked in this connection that uyear" in astronomical context
always means sidereal years in Babylonian texts (d. Neugebauer ACT p. 70)
but that there is no reason for asSUllliDg that one realized the difference between
sidereal, tropical, and anomalistic year. Apparently it was Ptolemy who ftrst
defined "year" as meaning utropical year" (d. Almagest III, 1 p. 192 f. Heiberg).
In recent years, it has become commonplace (e. g., the first edition of this
book or van der Waerden in ArchIv fUr OrientforsehUDg 16, 1953, p. 22) to
consider 419 B.C. as the earliest aUested date for the mention of the real zodiacal
sips in Babylonia. A copy of the text in question, an astronomical diary for
the year -418/417, has been published by E. F. Weidner in Archlv fUr Orient-
forsehung 16 (pI. XVIII) and shows, on the contrary, that zodiacal signs had
not yet been introduced. Four passages occur (obv. 7, 11, rev. 8, 11 f.) where
planets are said to be "behind" or"in front of" the alleged zodiacal signs. From
this it is clear that eeliptical constellations, not zodiacal signs, are referred to
[A. Sachs].
ad 46. Budge, The Rise and Progress of Assyriology, London 1925, writes
about Strassmaier as foRows (p. 228): uHe was convinced that it was a waste
of time to complle an Assyrian Dictionary, or to write a history of the Sumerian
and Babylonian civiUzations, whllst so many tens of thousands of tablets in the
British Museum and elsewhere remained unpublished". Today one may repeat
this statement, oDly replacing "tens of thousands" by uhundreds of thousands".
ad 46. The uobservational" texts are discussed by A. Sachs, A Classification
of Babylonian Astronomical Tablets of the Seleucid Period. J. Cuneiform
Studies 2 (1950) p. 271-290.
ad 49. As an example of the solar motion according to System A, I have
computed the data for the same year which we used on p. 110 for System B.
These elements are readily obtainable from ephemerides slightly earlier or
slightly later. As a matter of fact, all texts of System A form a uniform ephemeris
with no disturbances at an from the earliest to the latest text known. This is
not the ease for System B and therefore every comparison between the two
systems must reckon with the possibility that the texts of System B show some
small individual deviations. Nevertheless it is clear from the difference in method
that deviations of about 2° in solar longitude may occur in the column B. This
does not imply, however, that the final columns show equally large deviations.
The ephemeris for System A leads to the foRowing values
 Notes and References 141

XII. 22,18,45 "Y'

2,59 I 20,26,15 tI
II 18,33,45 D
III 16,41,15 1m

IV 14,48,45 .no
V 12,56,15 III
----------
VI 12,56 ~

VII 12,56 Il
VIII 12,56 t
IX
X
XI
XII
12,56
12,56
12,56
----------
11,56,15"Y'
- r\

)(

The dotted lines indicate the discontinuities at lIl13 and )( 27 where the monthly
solar velocity changes from 28;7,30 0 to 30 0 and vice versa.
The agreement with System B is quite close for the last month of the year.
For 2,58 XII. System B gave "Y' 22;8,18,16 as compared with "Y' 22;18,45. For
2,59 XII we had "Y' 11;30,48,56 against "Y' 11;56,15 now. But for the middle of
the year the linear zigzag function leads to III 14;58,18,18 as compared with III
12;56,15 just before the discontinuity.
ad.60. O. Neugebauer, "Saras" and lunar velocity in Babylonian astronomy.
Kgl. Danske Vide Selsk., mat.-fys. medd. 31,4 (1957).
ad 61. The "Saros". It has become customary to call the relation
223 synodic months - 242 draconitic months
the "Babylonian Saras" and to assume that it was the basis for the prediction
of eclipses by the Babylonians and their successors.
Unsuccessf1J1 protests against the use of this terminology have been made by
»,
Ideler (1825 1», Tannery (1893'», SchiaparelU (19081», Bigourdan (1911 4 and
».
PaDDekoek (19176 Only Ideler, however, gave an aecount of the origin of this
term, and it therefore seems to me worthwhile to outline the main steps as a
beautiful example of the creation of generally accepted historical myths.
The Sumerian sign IAr has, among others, the meaning "universe'· or the
Ute. As a number word it represents 3600, thus being an example of the trans-
formation from a general concept of plurality to a concrete high numeral.
In the special meaning of 3600 years, "Saras" is used by Berossosl) (about
290 B.C.) and, following him, by Abydenus') (second cent. A.D.) and by
Synkellosl) (about 800 A.D.).
1) Handbuch der ••. Chronologie I p. 213•
•) Recherches sur l'histoire de l'astron. anc. p. 317
8) Seritti I p. 75.
C) L'astronomle p. 33.
I) The origin of the Saros. Koninklijke Atad. van Wetenschappen te Amsterdam,
Proceedings 20 (1917) p. 943-955. Cf. also Quellen und Studien z. Gesch. d. Mathe,
Ser. B. voL 4 (1937-1938) p. 241 ft. and p. 407 ft.
t) Fragm. 29 ft. (Schnabel, Berossos, p. 261 ft.). ,
') Schnabel, Berossos, p. 263, 30a (in line 29, correct
') Schnabel,B~08,p.261ft.
Ha"
"to read xal ,,,).
142 Chapter V

An astronomical meaning is associated with "Saras" for the first time in

the encyclopedia of Suidas (about 1000 A.D.). There "Saros" is explained as
"a measure or number with the Chaldeaos" and then the remark is added
that one Saros contains 222 months, i. e., 18 years and 6 months, while 120
Saroi correspond to 2222 years1). The first relation implies that one year contains
exactly 12 months. This excludes the Babylonian calendar. The second relation
is a consequence of the first if one considers 2222 to be a scribal error for 2220;
otherwise it is senseless. In no case is there any relation to eclipses.
Pliny, NH II, 561), discusses the recurrence of eclipses after 223 months.
The manuscripts contain di1ferent readings of this number: 213 or 293 or 222
or 2358 ). Edmund Halley had at his disposal a Pliny text which gave 222. He
realized that only 223 made sense and assumed that a similar correction should
be made in Suidas, whose source he thought to be Pliny. He overlooked,
however, that all the other figures in Suidas contradict the change of 222 into
223 and that they are only the expression of the trivial relation that one year
contains 12 months. Thus Halley assumed that Suidas intended to say that
223 months were ealled one "Saros", and he published this conjecture in the
Philosophical Transactions 1691 (p. 535-540; reprinted in the Acta Bruditorum
1692, p. 529-534).
Halley's hypothesis was severely criticized by Le Gentil in 176(4) after it
had been presented as a fact by Montucla in the first edition of his Histoire
des mathematiques (1758). A more cautious formulation in the second edition
(1802) was too late to have any effect. Since Montucla it has been an accepted
doctrine of textbooks that the Babylonians used the "saros" for the prediction
of eclipses. From Kugler (1900) we know how eclipses were actually computed
during Seleucid period, namely, by a careful investigation of the latitude of the
moon in relation to the syzygies. Nevertheless, there are certain indications that
the periodic recurrence of lunar eclipses was utilized in the preceding period
by means of a crude 18-year cycle which was also used for other lunar phenomena.
The myth of the Saras is often used. as an "explanation" of the alleged
prediction by Thales of the solar eclipse of -584 May 28. There exists no cycle
for solar eclipses visible at a given place; all modern cycles concem the earth
as a whole. No Babylonian theory for predicting a solar eclipse existed at 600
B.C., as one can see from the very unsatisfactory situation 400 years later; nor
did the Babylonians ever develop any theory which took the inftuence of geo-
graphical latitude into aecount. One can safely say that the story about Thales's
predicting a solar eclipse is no more reliable than the other story aboutAn-
axagoras predicting the fall of meteors.
Even from a purely historical viewpoint the whole story appears very doubtful.
Our earliest source, Herodotus (1,74), reports that Thales had predicted "this
loss of daylight" to the Ionians correctly "for the year" in which it actually
happened. This whole formulation is so exceedingly vague that in itself it exe-Judes
the use of any exact method. The farther we move away from the time of Thales,

1) Ed. A. Adler IV p. 329. _

I) Ed. lan-Mayhon'1 p. 144: Loeb Class. Libr. I p. 204/205.
I) Cf. the critical apparatus in lan-Mayhoft. The number 235 was obviously
suggested by the 19-year ·'Metonic" cycle which contains 235 months.
6) M&noires for 1766 of the Acad. Royale des Sci., Paris, p. 55 ft.
 Notes and References 143
the more generous do the ancient authors become in assigning to him mathe-
matical and astronomical discoveries. I see not a single reliable element in any
of these stories which have become so dear to the histories of science (cr. also
p. 148). In this connection may be quoted the summary of an article by R. M.
Cook, Ionia and Greece, 800-600 B.C., J. Hellenic Studies 66 (1946) p. 67-98:
"My tentalive conclusion is that we do not know enough to say definitely whether
in the 8th and 7th centuries the Ionians were generally the pioneers of Greek
progress, but that on the present evidence it is at least as probable that they
were not."
Such a cautious outlook is, however, far from common. Only one of the
amusing fairy tales, spun out of the story of the Thales eclipse, may be mentioned.
In a learned work on "Die griechisch-rlSmische Buchbeschreibung verglichen
mit der des vorderen Orients" (Halle 1949) C. Wendel declares (p. 20 tI.)
that the "Bahnbrecher" of Ionian science must have had a library of "amazing
richness" at their disposal and Thales must have been its "spiritual founder".
Needless to say Thales's studies in Egypt are also taken very serionsly. Un-
fortunately, we know from Diodorus I, 38 that Thales knew so little about
Egypt that he could propose the theory that the inundation of the Nile began
at "the mouths of the river when the etesian winds ... hinder the flow of the
water into the sea".
ad 52. The problem of column Q can be easily formulated as follows.
Column 0 establishes the elongation .dA of the moon from the sun for the critical
evening by multiplying the time since conjunction by the relative velocity between
sunset and moonset. This is equivalent to asking for the time it takes the are .d A
of the ecliptic to set and this in turn is equal to the time of rising of the dia-
metrically opposite section of the ecliptic. But for the rising times of ecliptic
arcs, arithmetical schemes are known (cf. below p. 159 and Fig. 26) which
solve the problem of transforming longitudes A into right ascension 0:. Exactly
the same coefficients appear in the transformation from 0 = .dA into Q = .do:
as I have shown in J. Cuneiform Studies 7 (1953) p. 100 to 102.
ad 56. For the schematic computation of the dates of the appearances (r),
oppositions (8), and disappearances (D) of Sirius cr. A. Sachs, Sirius Dates in
Babylonian Astronomical Texts of the Seleucid Period, J. Cuneiform Studies 6
(1952) p. 105-114.
ad 51. O. Neugebauer, The Babylonian method for the computation of
the last visibility of Mercury. Proc. Am. Philos. Soc. 95 (1951) No.2.
ad 58. For Babylonian planetary theory cr. vol. 2 of my Astronomical
Cuneiform Texts. How these methods originated is still largely unknown.
A. Sac h s has given a classification of the not stricOy mathematical texts (J.
Cuneiform Studies 2, 1950, p. 271-290) which also clearly distinguish between
observational elements and computed data, e. g. on the basis of periodic re-
currence.
Apart from these well defined classes, there exist procedure texts which
evidenOy belong to an intermediate state preliminary to the final mathematical
theory. Texts of this type still present serious difficulties because of unknown
terminology, composite character of historical background, etc. Several copies
of such texts have been published by Thureau-Dangin in Tablettes d'Uruk
(Paris 1922) and in Pinches-Strassmaier-Sachs (cr. p. 138). So far
144 Chapter V

Sachs and I have succeeded only in explaining the meaning of one of these texts
to a reasonable degree of certainty (J. Cuneiform Studies 10, 1956, p. 131-136).
ad 59. The secrecy of ancient oriental sciences has often been assumed
without any attempt to investigate the foundation for such a hypothesis. There
exist indeed examples of "cryptographic" writing both in Egypt and Mesopo-
tamia. The Cenotaph of Seti I, for instance, contains cryptographic passages
in the mythological inscriptions which are written around the sky goddess. Some
of the passages use rare readings of hieroglyphs. some are simply incorrectly
arranged lines of the original from which the artist copied. In a related text
concerning a sun dial the words are written backwards, as if reflected by a
mirror, but this causes no real difficulty in reading the text. On the whole,
however, all texts with mathematical or astronomical context show not the
slightest intention of concealing their meaning from the reader. I think one
can only agree with T. E. Peet, the editor of the mathematical Papyrus Rhind,
that we have no reason for assuming the existence of any secret science in Egypt.
The same holds for Babylonia. The Old Babylonian mathematical texts are
as plainly written as possible. From the latest period there exist a few texts
which give lists of numbers and signs obviously for coding and decoding
purposes. A few words and proper names are written in such a code in the
colophons of two ephemerides. The ephemerides themselves as well as the
procedure texts show no trace of an attempt to hide their contents. If many
details remain unintelligible to us, it is our ignorance and missing texts which
cause the difficulties, not an intentionally cryptic writing. I think the remark
found occasionally in colophons of Uruk texts that the text should only be
shown to .. the informed" is not to be taken too seriously. It hardly indicates
much more than professional pride and feeling of importance of members of
the scribal guild.
Cryptographic devices occur also in Greek manuscripts of the Byzantine
period, hased, e. g., on a simple substitution of letters and their numerical
values in inverse order; d. e. g., V. Gardthausen, Griechisehe Palaeographie II
(2nd ed., Leipzig 1913) p. 300 tI. Magical texts are, of course, full of secret
combinations of letters; astrological texts, however, are practically free of such
secrecy.
The sixth chapter of the SOrya-Siddhlnta deals with a graphical representa-
tion of the different phases of an eclipse. It ends with the remark, "This mystery
of the gods is not to be imparted indiscriminately: it is to be made known to
the welltried pupil, who remains a year under instruction". Burgess says rightly,
"It seems a litUe curious to find a malter of so subordinate consequence •..
guarded so cautiously ...... The same holds for the construction of a celestial
globe (S.-S. XIII, 17 and similarly Paiica-S. XIV, 28). Similarly one of the
most trivial chapters in the Paiica-Siddhintika (XV) is called the "secrets of
astronomy". Here one gets the impression that we are dealing with a very old
section which could have been omitted without any harm to the understanding
of the rest.
The name of Kidenas == Kidinnu is customarily associated with the city of
Sippar and its school of astronomers, mentioned by the Greek writers cited
above p. 137. A. Sachs has realized, however, that the passage in question was
misread by Strassmaier (d. Neugebauer, ACT I p. 22 colophon Zo). We have
not a single astronomical text which came from Sippar.
 145

CHAPTER VI

Origin and Transmission of Hellenistic Science.

60. Any attempt to reconstruct the origin of Hellenistic mathe-
matics and astronomy must face the fact that Euclid·s "Elements"
and Ptolemy·s "Almagest" reduced all their predecessors to
objects of mere "historical interest'· with little chance of survival.
As Hilbert once expressed it, the importance of a scientific work
can be measured by the number of previous publications it
makes superfluous to read.
Because Euclid's work falls not much more than a century
after the beginning of scientific mathematics, it has been easier
to restore its prehistory than is the case with astronomy. The
early date of Euclid (about 300 B.C.) leaves room for two or
more centuries of active development carried out by men like
Archimedes and Apollonius. Ptolemy, in 150 A.D., lives close
to the end of the Hellenistic age, and his work comprises practi-
cally all astronomical achievements which could be reached with
the mathematical methods of antiquity. The careful analysis,
on purely mathematical grounds, of Euclid's work has given
valuable information about the preceding major steps on which
it was built. Ptolemy·s work is exclusively concerned with the
description of one unified method for the representation of the
celestial phenomena. On the basis of the Almagest we would
have no idea about the existence of totally different methods,
Greek and Oriental, which preceded and occasionally even
survived the Almagest.
Finally one must realize that the "Elements" of Euclid concern,
with very few exceptions, a purely Greek development in a sharply
defined direction. Ptolemy's astronomy is probably built to a
large extent on results obtained 300 years earlier by Hipparchus,
who in turn was influenced both by Greek and by Babylonian
ideas. Hence the problems connected with the history of astron-
146 Chapter VI

omy are far more involved than is the case with mathematics.
Furthermore Greek mathematical procedures are directly intel-
ligible to a modern mathematician whereas ancient astronomical
treatises operate with a terminology and with problems and
empirical and numerical methods which are no longer familiar
in our time. This situation is also reflected in the modern discussion
of these problems. For the history of Greek mathematics, there
are quite competent and complete presentations but we are far
from this goal in the history of ancient astronomy.
61. To say that Greek mathematics of the Euclidean style is a
strictly Greek development does not mean to deny a general
Oriental background for Greek mathematics as a whole. Indeed,
mathematics of the Hellenistic period, and still more of the later
periods, is in part only a link in an unbroken tradition which
reaches from the earliest periods of ancient history down to the
beginning of modern times. As a particularly drastic example
might be mentioned the elementary geometry represented in the
Hellenistic period in writings which go under the name of Heron
of Alexandria (second half of first century A.D.). These treatises
on geometry were sometimes considered to be signs of the decline
of Greek mathematics, and this would indeed be the case if one
had to consider them as the descendants of the works of Archi-
medes or Apollonius. But such a comparison is unjust. In view
of our recently gained knowledge of Babylonian texts, Heron's
geometry must be considered merely a Hellenistic form of a
general oriental tradition. The fact, e. g., that Heron adds areas
and line segments can no longer be viewed as a novel sign of the
rapid degeneration of the so-called Greek spirit, but simply
reflects the algebraic or arithmetic tradition of Mesopotamia.
On this more elementary level, the axiomatic school of mathe-
matics had as little influence as it has today on surveying. Conse-
quently, parts of Heron's writings, practically unchanged, survived
the destruction of scientific mathematics in late antiquity. Whole
sections from these works are found again, centuries later, in one
of the first Arabic mathematical works, the famous "Algebra"
of al-Khwarizmi (about 800 to 850). This relationship can be
especially easily demonstrated by means of the figures. In order
to make the examples come out in nice numbers, the figures were
composed from a few standard right triangles. One of these
 Origin and Transmission of Hellenistic Science 147

standard examples is shown in Fig. 21, which appears in Heron

as well as in al-Khwirizmi. Two right triangles with sides 8, 6.
and 10 are combined into an isosceles triangle of altitude 8 and
base 12. Then a square should be inscribed. The resulting linear
equation yields for the side of the square 4 2 5 10, i. e. 4f. The
style of the formulation of these problems. the way of solving
them in special numerical examples-all this closely resembles
the Babylonian mathematical texts. A similar comparison could

42siii
12
Fig. 21.

be carried out for various parts of Hellenistic and Arabic mathe-

matics. such as the inheritance problems, the algebra of the
Diophantine type. etc. This does not mean that Hellenistic or
even Arabic authors were able to utilize Babylonian material
directly. All that we can safely say is that a continuous tradition
must have existed, connecting Mesopotamian mathematics of the
Hellenistic period with contemporary Semitic (Aramaic) and
Greek writers and finally with the Hindu and Islamic math-
ematicians.
62. The question arises whether any Oriental influence is
apparent in the scientific branch of Greek mathematics. My
answer to this question cannot be proved by documentary
evidence, but the following working hypothesis seems to me to
account for the known facts: the theory of irrational quantities
and the related theory of integration are of purely Greek origin,
but the contents of the "geometrical algebra" utilize results
known in Mesopotamia.
To substantiate these statements a few remarks must be made
about the historical development of Greek mathematics. First
of all, it seems necessary to distinguish sharply between the
axiomatic style of mathematics, which is the work of Eudoxus
and his contemporaries in the fourth century B.C., and the
148 Chapter VI

mathematics usually connected with the Ionian and South-

Italian schools. I see no reason to deny to the earlier period a
comparatively large amount of mathematical knowledge which
might comprise, or even exceed, in certain points the knowledge
attested in Mesopotamian sources. It seems to me evident, how-
ever, that the traditional stories of discoveries made by Thales
or Pythagoras must be discarded as totally unhistorical. Thales,
e. g., is credited with having discovered that the area of a circle
is divided into two equal parts by a diameter. This story clearly
reflects the attitude of a much more advanced period when it
had become clear that facts of this type require a proof before
they can be utilized for subsequent theorems. To the later mathe-
maticians it seemed natural to assume that theorems which had to
be established first on logical grounds should also come first chro-
nologically. Actually the Greek historians acted in exactly the same
way as modern historians do when no source material is avail-
able to them: they restored the sequence of events according to
the requirements of the theory of their own times.We know today
that all the factual mathematical knowledge which is ascribed
to the early Greek philosophers was known many centuries before,
though without the accompanying evidence of any formal method
which the mathematicians of the fourth century would have
called a proof. For us, there is nothing to do but to admit that
we have no idea of the role which the traditional heroes of Greek
science played. It seems to me characteristic, however, that
Archytas of Tarentum could make the statement that not geometry
but arithmetic alone could provide satisfactory proofs. If this was
the opinion of a leading mathematician of the generation just
preceding the birth of the axiomatic method, then it is rather
obvious that early Greek mathematics cannot have been very
different from the Heronic Diophantine type.
It is also generally accepted that the essential turn in the
development came about through the discussion of the conse-
quences of the arithmetical fact that no ratio of two integers could
be found such that its square had the value 2. The geometrical
corollary that the diagonal of a square could not be "measured" by
its side obviously caused serious discussion about the relation
between geometrical and arithmetical proof. The "paradoxa"
concerning continuity, both of space and time, made the relation
 Origin and Transmission of Hellenistic Science 149

to the whole problem of determination of area and volume evident.

One way out might have been the assumption of a somewhat
atomistic structure of geometrical objects by means of which the
problem of area or volume would have been reduced, though in
not too clear a fashion, to a counting of discrete elements,
"atoms".
The reaction of the mathematicians against this type of specula-
tion seems to have led to two major steps. First of all, one had
to agree exactly on a system of basic assumptions from which
alone the rest had to be deduced; this gave rise to the strictly
axiomatic procedure. Secondly, it had become clear that one
should consider the geometrical objects as the given entities such
that the case of integer ratios appeared as a special case of only
secondary interest; this led to the problem of how to formulate
classical arithmetical and algebraic knowledge in geometrical
language. The result is the familiar ·'geometrical algebra" of
Greek mathematics. It is these two essential steps which are fully
to the credit of the Greek mathematicians.
The situation changes when we ask about the origin of mathe-
matical relations which were incorporated in the systematic
building of geometrically demonstrated laws. Everything which
is directly related to the theory and classification of irrational
quantities is, of course, Greek; and the same holds for the rigid
theory of the processes of integration. The elementary theory of
numbers, however, mayor may not eventually be based on much
older oriental material. I do not doubt that any connection with
the name of Pythagoras is purely legendary and of no historical
value.
The most interesting question, however, seems to me the
problem of the origin of the "geometrical algebra". We have
seen that the Babylonian treatment of problems of second degree
consists in reducing them to the "normal form" where two quan-
tities, x and y, should be found from their given product and
their sum or difference. It seems significant that the geometric
formulation of this problem leads precisely to the central problem
of the geometrical algebra, a problem which is otherwise rather
difficult to motivate. This problem is known as the "application
of area", which consists, in its simplest form, in the following.
Given an area A and a line segment b; construct a rectangle
150 Chapter VI

W/lilffh,'~/IffllM y
~ y
Fig. 22.

of area A such that one of its sides falls on b but in such a way
(cf. Fig. 22) that the rectangle of equal height and of length b
is either larger or smaller by a square than the rectangle of area A.
The identity of this strange geometrical problem with the Babylon-
ian "normal form" is at once evident when we formulate it
algebraically. Let us call, in both cases, a: and y the sides of the
rectangle. Then we are given
xy =A.

In the first ease a square should remain free; its sides are y and
we must require
x + y = b.
In the second case, a square should exceed the rectangle of side b;
thus we should have
x - y = b.

These are indeed the two normal forms (cf. p. 41).

Attempts have been made to motivate the problem of "applica-
tion of areas" independently of this algebraic background. There
is no doubt, however, that the above assumption of a direct
geometrical interpretation of the normal form of quadratic
equations is by far the most simple and direct explanation. I
realize that simplicity is by no means equivalent with historical
proof. Nevertheless the least one must admit is the possibility of
the above explanation. That the numerical solution of quadratic
equations was taught in contemporary Babylonian scribal schools
cannot be doubted in view of the fact that this is still attested
even in later periods of cuneiform writing. The only serious
question consists in the specific way in which such knowledge
found its way to Greece. Here we are left to mere speculation.
 Origin and Transmission of Hellenistic Science 151

But it seems to me not to require too much imagination to think

of a diffusion of mathematical knowledge from the Near East
to Greece in a period close to the eve of the Macedonian offensive
against the Persian empire.
In the history of speculative thought much has been said about
the direct contact of Plato and Aristotle with Orientals. An Iranian
is reported to have informed Plato about the religion of Zarathus-
tra. Callisthenes, the nephew of Aristotle, is even supposed to have
brought Babylonian astronomical records to Athens. This latter
report is not too trustworthy, as T. H. Martin has rightly empha-
sized in 1864. No mention of it comes directly from Aristotle but
it is based only on Porphyrius (third cent. A.D.) through Simplicius
(sixth cent. A.D.); and worst of all, these Babylonian observa-
tions are supposed to have reached 31,000 years back. \Vhat-
ever the case may be, a few years later Babylonia was under
Greek domination and literary evidence is then no longer required
to prove that the Greeks had access to Babylonian science from
this time onwards. Nevertheless, the need for caution remains
when the contact under discussion especially concerns the work
of Eudoxus. I see no good reason to deny the possibility of his
travels to Egypt. It seems to me certain, however, that there was
nothing to learn from the Egyptians themselves, and the hypothesis
that Babylonian science had reached Egypt before Greece seems
to me only to substitute one name of an unknown quantity for
another. At the present state of our knowledge, none of these
stories contributes significantly to our insight into the historical
events.
69. The Greeks themselves had many theories about the origin
of mathematics. A favored one, which is still kept alive in modern
textbooks, makes the necessity of repeated land measurement
responsible for geometry. Modern authors have often referred
to the marvels of Egyptian architecture, though without ever
mentioning a concrete problem of statics solvable by the known
Egyptian arithmetical procedures. A much more sophisticated
attitude is represented by Aristotle, who considers the existence
of a "leisure class", to use a modern term, a necessary condition
for scientific work. Our factual knowledge about the development
of scientific thought and of the social position of the men who
were responsible for it is so utterly fragmentary, however, that
152 Chapter VI

it seems to me completely impossible to test any such hypothesis,

however plausible it may appear to a modern man.
It seems to me equally impossible to give anyone conclusive
Itexplanation" for the origin of higher mathematics in the fifth
and fourth century in Athens and the Italian colonies. On the
negative side, however, I think that it is evident that Plato's role
has been widely exaggerated. His own direct contributions to
mathematical knowledge were obviously nil. That, for a short
while, mathematicians of the rank of Eudoxus belonged to his
circle is no proof of Plato's influence on mathematical research.
The exceedingly elementary character of the examples of math-
ematical procedures quoted by Plato and Aristotle give no support
to the hypothesis that Theaetetus or Eudoxus had anything to
learn from Plato. The often adopted notion that Plato "directed"
research fortunately is not borne out by the facts. His advice to
the astronomers to replace observations by speculation would have
destroyed one of the most important contributions of the Greeks
to the exact sciences. Plato's doctrines undoubtedly have had
great influence upon the modern interpretation of Greek sciences.
But if modern scholars had devoted as much attention to Galen
or Ptolemy as they did to Plato and his followers, they would
have come to quite different results and they would not have
invented the myth about the remarkable quality of the so-called
Greek mind to develop scientific theories without resorting to
experiments or empirical tests.
64.. The structure of our planetary system is indeed such that
Rheticus could say "the planets show again and again all the
phenomena which God desired to be seen from the earth".
The investigations of Hill and Poincare have demonstrated that
only slightly different initial conditions would have caused the
moon to travel around the earth in a curve of the general shape
given in Fig. 23, and with a speed exceedingly low in the outer-
most quadratures 01 and O. as compared with the motion at new
and full moon. Nobody would have had the idea that the moon
could rotate on a circle around the earth and all philosophers
would have declared it as a logical necessity that a moon
shows six half moons between two full moons. And what could
have happened with our concepts of time if we were members of
a double-star system (perhaps with some uneven distribution of
 Origin and Transmission of Hellenistic Science 153

F
Fig. 23.

mass in our little satellite) is something that may be left to the

imagination.
Actually, however, the initial conditions of our planetary
system were chosen in such a way that all the satellites of the
sun-and our own satellite as well-behave with great modesty.
Their orbits can be closely approximated by circles such that
the simplest possible model of a circular motion with constant
speed leads immediately to very reasonable results for the descrip-
tion of the solar and lunar phenomena. On the other hand, the
deviations from the trivial circular orbits are just great enough
to be observed and to challenge an explanation, but small enough
such that again comparatively simple modifications of the trivial
solution give satisfactory results. The successive approximations
of the Babylonian lunar and planetary theory reflect this situation
perfectly. At the basis lies the counting of the periodically recurrent
phenomena; the properly chosen periodic functions-zigzag
functions or step functions-suffice to describe the deviation from
a trivial mean motion.
Perhaps a little before these methods were developed in Meso-
potamia, perhaps almost simultaneously, a most decisive step
in another direction was made by Eudoxus. The then recent
discovery of the sphericity of the earth must have suggested a
corresponding sphericity of the sky and a circular motion of the
celestial bodies. Eudoxus's theory may have well started from the
following consideration. The motion of sun and moon can be
described as the combination of the uniform motions of two
concentric spheres: one is the fast daily rotation about the poles
of the equator; the other is slow and proceeds in opposite direction
about an inclined axis which is perpendicular to the ecliptic.
Eudoxus saw that a similar combination is capable of explaining,
154 Chapter VI

at least qualitatively, also the most striking phenomenon of

planetary motion, the retrogradations. The motion of two homo-
centric spheres allows of two trivial limiting types. If the two axes
are made to coincide, the body which is fixed to the equator of
one of the two spheres moves simply in a circle with the difference
velocity. If, secondly, the two opposite velocities are made of
equal amount the body remains stationary as seen from the center.

Fig. 24.

But one case remains to be investigated: what motion results from

equal opposite velocities but inclined axes? Eudoxus found that
the orbit is an 8-shaped curve (Fig. 24). Now one can superimpose
a third rotation about an axis which is perpendicular to the plane
of sYmmetry which represents the plane of the ecliptic. Conse-
quently, the point P no longer follows a closed curve but proceeds
with a certain mean velocity in the ecliptic. Simultaneously,
however, there appears a periodic deviation from the ecliptic,
or a motion in latitude. Finally, one will obtain retrogradations
if the longitudinal component is less than the backward motion
in the original figure eight. Thus it is demonstrated that, at least
qualitatively, even the apparent irregularities of planetary motion
can be described by a combination of circular motions of uniform
angular velocity.
In spite of the great importance, in principle, of the discovery
of Eudoxus, it is quite obvious that a model of this type has grave
shortcomings.
For example, the observed retrogradations of the planets do
not recur in curves of identical shape as would be the case in
the Eudoxian model. Another difficulty lies in the large variation
 Origin and Transmission of Hellenistic Science 155

Fig. 25.

of the brightness of planets. that seemed to indicate corresponding

variations in their distance from the earth. We do not know who
first succeeded in explaining these and similar anomalies by
means of a much more flexible modification of the theory of
uniform circular motion. We know. however. that Apollonius
(about 200 B.C.) used the simple device of viewing uniform
circular motion. not from the center of the orbit. but from a slightly
eccentric point. This obviously has the effect that the motion
appears fastest where the circle is nearest to the observer and
slowest at the opposite point. But Apollonius proved more. He
demonstrated that an eccentric movement of this type can always
be replaced by an epicyclic motion where the center of the epicycle
moves on a circle with the observer at its center and with a radius
of the epicycle equal to the eccentricity (cf. Fig. 25). All that is
needed is to regulate the angular velocities in such a way that
the point P and the observer E remain the vertices of a parallel-
ogram SPCE. But as soon as epicycles are introduced. it is also
clear that the motion of P around S can be chosen in such a way
that P appears to have a retrograde motion if observed from E.
as we have shown in the discussion of Fig. lSb and 14c (p. 123 f.).
Hence the model of homocentric spheres could be abolished. and
uniform description of all celestial motion was obtained by means
of eccenters and epicycles. But the main principle. the fundamental
role of eircular motion. seemed to have been splendidly vindicated.
This conviction remained the cornerstone of celestial "dynamics"
of ancient astronomy comparable to a law of inertia.
In principle. however. ancient astronomers pretended only to
"describe" the appearances. not to "explain" them. All that was
actually observable was angular motions. the only exceptions
156 Chapter VI

being the distances for sun and moon obtainable by means of

parallax. For the planets, however, neither theory nor observa-
tions were accurate enough to obtain reliable information as to
their distances. In our discussion of the geocentric description
of a heliocentric motion in Figs. 13 and 14, S represented the sun.
Now we try only to describe the direction EP under which the
planet P appears from E. Hence we can no longer say that S is
the sun but only that ES is the direction to the sun. But otherwise
all our conclusions remain valid. Thus we can say that the angular
motion of an inner planet is described by an epicyclic motion
such that the direction from E to the center S of the epicycle
coincides with the direction from E to the sun. And an outer
planet P moves on its epicycle around C in such a way that CP
is always parallel to the direction from E to the sun. This is
indeed the basic formulation of the Greek planetary theory by
means of epicycles, with the obvious refinement that we should
say "mean sun" instead of simply "sun". This theory is a correct
description of the appearances so far as the angular motion is
concerned and it would be a correct heliocentric theory if the
correct scale were chosen. Second-order deviations from this first-
order approximation could be explained by added eccentricities
and similar devices which were brought to perfection by Ptolemy.
Only greatly refined observations could eventually disclose the
defects of the supposition of strictly circular motions.
65. If one looks back on this sketch of the development of the
Ptolemaic planetary system, one sees no reason for the assumption
of oriental influences. All that we know about Egyptian astronomy
rules out any possible influence from this source. The Babylonian
theory, on the other hand, is known to us to have reached about
equally excellent results-by means of methods which nowhere
point to an interpretation through a combination of circular
motions or any other mechanical model. Indeed, zigzag and
step functions practically exclude any such attempt. Nevertheless,
Babylonian influence is visible in two different ways in Greek
astronomy: first, in contributing basic empirical material for the
geometrical theories which we have outlined in the preceding
section; second, in a direct continuation of arithmetical methods
which were used simultaneously with and independently of the
geometrical methods.
 Origin and Transmission of Hellenistic Science 157

The first influence was revealed when the Babylonian lunar

theory was deciphered by Epping and Kugler. Exactly the same
constants which determined the periods of several of the most
important zigzag functions in the Babylonian theory arc attested
as the relations from which the mean motions were derived in
the Greek theories, especially by Hipparchus (as nearly as we
can teU from Ptolemy's references in the Almagest). Because
the earliest Babylonian ephemerides antedate-though only by
a narrow margin-the time of Hipparchus, it cannot be denied
that at least the empirical foundations of the Babylonian theory
must have been known to Hipparchus. How this knowledge was
transmitted to him and how much he knew about the actual
technique of computing ephemerides cannot be answered from
our sources. Ordinarily, the teaching activity of the Babylonian
Berossos (who moved to the Greek island of Cos about 270 B.C.)
is considered responsible for the transmission of much of the
astronomical knowledge to the Greeks. This may actually be the
ease, though the extant fragmentary excerpts from his writings
contain no specific reference to mathematical astronomy. What
we would really need in order to understand the details of trans-
mission is a Greek commentary to Babylonian ephemerides and
procedure texts. Somewhere the great step from year-by-year
ephemerides to tables based on mean motions, as we know them
from the Almagest, must have been made. That we cannot answer
such a question even approximately demonstrates how little we
know about the earlier period of Hellenistic astronomy outside
Mesopotamia.
66. Babylonian influence is not restricted to providing essential
constants for the determination of the parameters of the geo-
metrical models of our planetary system. A much more direct
development of the Babylonian arithmetical methods has be-
come visible from Greek papyri and from occasional references
to technical details in the astrol~cal literature. Exactly as the
"Greek" mathematical literature must be divided into two classes,
a purely scientific development and a more elementary tradition
which is closely related to the oriental tradition, so also do astron-
omical procedures faU into two groups: one leading to the Almagest,
the other probably best known among the astrological authors
for their computations of the positions of the celestial bodies for
158 Chapter VI

horoscopic purposes. I call this second class of procedure the

"arithmetical methods" or the "linear methods" because they
are essentially based on difference sequences of first order.
It must be realized that no such classification is anything more
than a convenient matter of speech and that there exist many
contacts and influences between both extremes. Most of all,
the reader should be warned not to take the expression "arith-
metical methods" as implying that the methods of the Almagest
somehow exclude numerical procedure. The opposite is true.
Not only does the Almagest contain a great number of numerical
tables, which in turn are based on an enormous amount of
numerical computation, but the final goal of the Almagest is
exactly the same as that of the "arithmetical methods", namely,
to provide numerical data for astronomical phenomena. But
the Almagest is unique in its desire to explain the empirical
foundations and the theoretical reasons for its procedures. And
the way always leads first to a definite geometrical model, from
which the resulting arithmetical consequences are then derived.
The linear methods, however, proceed on exclusively numerical
grounds precisely in the fashion which is now familiar to us from
the Babylonian texts. Though no theoretical treatise concerning
these linear methods is preserved, it is clear that they rest on
cleverly designed procedures and on empirical material which
is probably rather similar for both methods. Nevertheless, the
concept of a geometrical model seems to be completely absent
for this second type of astronomical literature. This may be
compared with the lack of a strictly axiomatic structure for the
Heron-Diophantus type of Hellenistic mathematics.
The direct survival of Babylonian methods can be recognized
most easily in an important problem of mathematical geography.
In Hellenistic and medieval geography, one frequently finds the
latitude of a locality expressed by means of the ratio of the longest
to the shortest daylight for the region in question. Alexandria,
e. g., falls in the zone for which this ratio is 7: 5; that is to say,
the longest daylight is 14 hours and the shortest night, assumed
to be equal in length to the shortest dayl), is 10 hours. Similarly,
1) This convenient assumption of exact symmetry was always made in ancient
astronomy. Actually. however. atmospheric inOuenees make the sun visible longer
than would be the ease with a mathematical horizon. Consequently the shortest
daylight is longer than the shortest nlghL
 Origin and Transmission of Hellenistic Science 159

Babylon is characterized by the ratio 3: 2. Consequently it becomes

an important problem to determine the length of daylight astro-
nomically. Daylight will be longest when the sun is at the summer
solstice, zO°. At sunrise on this day, the point z 0° rises because
the sun is located at this point. At sunset on this day, the sun and
the point m5 0 0 set in the west, and the point" 0° rises because,
at any moment, 180 0 of the ecliptic must be above the horizon.
Consequently, we can say that, during the longest day, the six
signs z, st, TJR, ~, 11\., t have risen. Similarly, we can say
that, during the shortest day, the six signs from ~ to n have risen.
.w0J---1---~oooQ,,;:~+---+--

3 6 9 12 3 6 9 12
Syst. A Syst. :B
Fig. 26.

Thus we are able to determine the variation in the length of daylight

if we know the time of rising of each individual zodiacal sign.
If we call lX I the rising time or "ascension" of <'f, lXlI of ~, etc.,
then we know that the longest daylight is given by the sum
lX,+ ... + lX9, and the shortest daylight by lXIO + ... +
lXs. In
general. for any time of the year, we know the length of daylight
if we know the rising times for the semicircle of the ecliptic which
begins at the point in which the sun is located.
The values of the ascensions depend on the variable inclina-
tion between ecliptic and horizon and vary in a rather complicated
way.l) Nevertheless, arithmetical schemes to account for their
variation were devised for the computation of column C (length
of daylight) in the Babylonian lunar ephemerides (cf. p. 116 f.),
such that the extremal values show the ratio 3: 2 mentioned above
as the characteristic ratio for the latitude of Babylon. These
arithmetical schemes are slightly different for System A or B.
In System A, the lX'S increase and decrease with constant differ-
1) For an observer on the equator the rising times are called "right ascensions",
a term whieh is still in use in modern astronomy.
160 Chapter VI

- /
/
- ~
~

r\.
, r--
~

(f
3 , 9 ,2.
Fig. 27.

ence; in System B, twice the ordinary difference is assumed in

the middle (cf. Fig. 26). The correct curve would look like Fig. 27.
Its shape depends on the geographical latitude, the indentation
at the top becoming more pronounced as we move north. Tables
for these ascensions are given in all ancient and medieval astro-
nomical works. These tables demonstrate concretely what we have
said before about the classification of astronomical literature.
They may be computed by means of spherical trigonometry;
this then leads to values as represented in Fig. 27. This is
the case, e. g., for the tables in the Almagest and in the later Greek
and Arabie works of similar character. In many smaller treatises,
however, one finds the ascensions for various geographical
locations computed with schemes which are exactly of the type
of System A or B, represented in Fig. 26. This is the ease not only
in the West but also in Indian astronomical texts (beginning
with the sixth century A.D.) and the same holds for the
majority of the astrological writings until the end of the Middle
Ages.
It may seem very strange that primitive arithmetical schemes
were used long after the correct bigonometrical solutions had
been found and utilized for the computation of tables. And
not only were trigonometric and arithmetic schemes coexistent,
but the arithmetical devices maintained the duplicity of Systems
A and B inherited from the Babylonian ephemerides. This is a
nice example of the Clconservatism" of the human race as a whole
because this parallelism of equivalent methods is attested for Ba-
bylonians, Greeks, Romans, Jews, Christians, and Moslems alike.
But even for the limited field of the history of mathematics,
the problem of the "ascensions" is of great interest. The careful
 Origin and Transmission of Hellenistic Science 161

investigation of early Greek spherical geometry, especially of

Theodosius and Menelaos, has shown that this problem is one
of the major goals of the whole theory. It is probably Menelaos
(about 100 A.D.) who first saw that spherical geometry must be
based on great circles only. In the preceding period, either only
qualitative results were reached or else graphical methods were
used. One of these seems to have been based on the discovery
that stereographic projection of the sphere maps circles into
circles. This fact was certainly known at the beginning of our era,
as is shown by the construction of mechanical clocks which
represent the celestial motions in a plane, similar to the later
astrolabes. Hipparchus, who had no spherical trigonometry at
his disposal, may have solved spherical triangles by the method
of stereographic projection.
67.While Babylonian origin is quite obvious in the arithmetical
treatment of the problem of ascensions and length of daylight,
a much more complex situation is encountered in the theory of
the lunar motion. Insight into this part of Hellenistic astronomy
is of rather recent date and far from complete. In fact, it is more
appropriate to say that a new (and very promising) chapter of
research has been no more than barely begun. I shall outline
how this came about because it is typical of the accidental
fashion in which we actually proceed, in defiance of all attempts
at planning the road of research in advance. This, of course,
is really not too surprising because only those objects can be
reached in a systematic fashion whose outlines are already
fairly well determined. And this is certainly not the case with
Hellenistic astronomy and its descendants.
In 1922 Thureau-Dangin published a copy of a tablet from
Uruk, now in the Louvre, dealing with the daily movement of
the moon. It was discussed in 1927 by Schnabel, who pointed
out that the parameters agreed with the values given by Geminus
in his "Introduction" (about 100 B.C.). Several additional tablets
of the same type have been identified since then, and it is possible
to show that they belong to a consistent ephemeris of the moon's
daily motion extending at least from the year -194/3 to -181/0.
These texts are based on a zigzag function which we have already
mentioned on p. 122. Its period is 248 days; in other words, it
is assumed that the smallest integer number of days between,
162 Chapter VI

say, two minima of the lunar velocity is 248 days. This interval
covers 9 complete oscillations of the lunar velocity, or 9 "anomal-
istic months", but the length of one of these anomalistic months
is not an integer number of days. Its length is given by our zigzag
. 248 • I .
function as the quotient 9 = 27;33,20 days. This va ue IS
slightly too large and was obviously chosen in order to obtain a
conveniently round number for the difference of this zigzag
function, namely, 0;18 degrees per day. A more accurate value
can be derived from the Babylonian lunar theory itself, namely,
from the columns Fand G(cf. p. 117 f.). One finds 27;33,16,26,54,
assuming full accuracy of the numbers used for F and G. Thus
it is clear that we are dealing again with two concurrent and
slightly different methods of Babylonian astronomy for the
description of the lunar motion. The one is based on highly
accurate values and is used implicitly in the lunar tables for full
and new moons. The other, for the day-by-day motion, is based
on conveniently rounded-off parameters. It is the history of this
second method which we will now analyze.
The first step was made by Schnabel in his paper of 1927.
In a short appendix he remarked that the relation 9 anomalistic
months = 248 days was not only known to Geminus but also
occurs in Hindu astronomy. This was fully in line with a discovery
which had been made by Kugler in 1900, namely, that the ratio
3: 2 of longest to shortest day used by both systems in columns
C and D of the Babylonian lunar ephemerides also apPears in
Hindu astronomy, though this ratio is totally incorrect for the
main parts of India.
The next step was in a new direction. E. J. Knudtzon identified
as astronomical the fragments of two Greek papYri in the Library
of the University of Lund, Sweden, and sent me photographs
shortly after the end of World War II. One of these fragments
turned out to be part of a papyrus now at the University of
California, belonging to the larger class of Demotic and Greek
papyri dealing with planetary motion (cf. above p. 95). The
other fragment, however, proved to be a new type of lunar ephem-
eris based on the Babylonian relation: 9 anomalistic months =
248 days. The calendar used is based on Egyptian years (of
365 days each) and, in the fragment at hand, on the regnal years
 Origin and Transmission of Hellenistic Science 163

of Nero and Vespasian. The papyrus gives dates, 248 days apart;
and corresponding longitudes, 27;43,24,56° apart. The following
example of 3 consecutive lines will illustrate the procedure:

year 6 month VII day 2 11R 5;3,21,31

7 III 5 ~ 2;46,46,27
7 XI 13 1l\. 0;30,11,23.
From this the mean lunar velocity can easily be found. Because
248 days contain 9 complete revolutions, the moon moved not
only 27;43, ... during this period but 9 times 360 0 more. Hence
we add 54,0° to the previous number. Dividing the total 54,27;43,
24,56° by 248 gives for the daily motion 13;10,32,16, ... which
is slightly less than the standard Babylonian mean value of
13;10,35 0 per day. Obviously the deviation noted is merely a
consequence of the fact that the period of 248 days is itself only
a rounded-oft" value. This could be confirmed from the very same
text. The process described so far is repeated only 11 times. After
every 11 of these ordinary steps, henceforth called D, one "big"
step L1 of 303 days was inserted, with a corresponding motion of
the moon of 11 times 360 plus 32;33,44,51°. The lunar velocity
resulting from one such big step is 13;10,36,23, ... , slightly larger
than expected. This shows that we are dealing with a process of
successive approximations. Indeed, if we consider the 11 ordinary
steps plus one big step as one higher unit C = 11 D +L1 of
3031 days, then we find for it the mean motion 13;10,34,51,57, ...
This value is not only very close to the value 13;10,35 but we
know that 13;10,35 itself must be the result of a small rounding
oft". Ptolemy's value, e. g., is 13;10,34,58,33,30,30. Similarly the
value for the anomalistic month represented by C is a much
better approximation than in D; one finds 27;33,16,21, ... which
is very close to the expected 27;33,16,26, .... Thus we see that
higher groups were designed in order to avoid the accumulation
of error which was allowed in the single steps.
It can be shown that the moments chosen were the moments of
minimum velocity. Dates and positions being known for the
apogees of the moon, the positions of the moon for any other
date can be found simply by operating with the well known zigzag
function for the lunar velocity, starting with the minimum and
going linearly up and down until the given date is reached. This.
164 Chapter VI

computation will never lead to large errors because one always

starts from the nearest minimum, whose position is well estab-
fished by the general process.
Thus we see before us in a Greek papyrus a method of purely
arithmetical character, based on Babylonian parameters and
Babylonian schemes but adjusted to the Egyptian calendar. It is
difficult to say whether the formation of groups of the type
+
C = 11 D LI is also of Babylonian origin or a later invention
of the Alexandrian astronomers. Against Babylonian origin might
be held the fact that the process D alone was used for more
than 13 years in the preserved texts (p. 161); but this does not
exclude the possibility of the existence of the improved procedure
in other texts. Admission of our ignorance seems the best pro-
cedure.
How little we really know about this form of Hellenistic
astronomy became evident shortly after the publication of the
Lund papyrus. An investigation of dates of solstices in the Hel-
lenistic period led me to check a Greek papyrus of the John
Rylands Library, published in 1911, because it contained such
data at the end. The solstices turned out to be uninteresting but
it now appeared at once obvious that the main text contained,
among others, the rules for the computation of the scheme of
the Lund papyrus. This discovery made at least one thing clear:
there existed "linear methods" of far wider extent than one
could possibly have deduced from the silence of Ptolemy and his
commentators.
The Ryland papyrus revealed further details of the whole
method and raised new problems. It showed that the whole
process, as preserved in the Greek papyri, is based on the Era
of Augustus and that also the cycle of 25 Egyptian years was
incorporated-the very same cycle which is known to us from
the Demotic papyrus Carlsberg No.9 (cf. p. 95). Thus the mixture
of methods became still more evident for Hellenistic astronomy.
But there appeared also difficulties in details; for instance, the
step LI should have been inserted one line later, following the
rules of the Ryland papyrus, than it actually appears in the Lund
papyrus. We can express this also in the form that somewhere
one unexpected step D was inserted.
Today I am still unable to explain every detail in these Greek
 Origin and Transmission of Hellenistic Science 165

texts but the general direction in which to go now appears clearer.

Again returning to the study of the solar theory, I intended
to investigate the transmission of Ptolemy's theory of precession
to the Arabs, and it was only natural to include here the Hindu
sources. This led me to the Paiica SiddhAntikl of Variha Mihira,
written about 550 A.D. We shall come back to Hindu astronomy
presently (p. 173); for the moment it suffices to say that the Pafica
Siddhintiki contains certain rules for the computation of the
lunar motion based on the processes now know to us from
Greek sources. Thibaut, who edited the Paiica Siddhintiki in
1889, found it very difficult to understand these passages. He
eventually found the key to the problem in the book Kila San-
kalita by J. Warren. The latter had traveled extensively in South-
ern India and had recorded the astronomical lore of the natives
in the book mentioned, published in Madras in 1825. In this
book, he describes a method followed by the Tamil inhabitants of
the Coromandel coast for the computation of the lunar motion.
His informants no longer had any idea about the reasons for the
single steps which they performed according to their rules. The
numbers themselves were not written down but were represented
by groups of shells placed on the ground. Thus

•• • • •• •
• • • •
• • ••••
•••••
•
means 7 zodiacal signs and 19;5,1°. Nevertheless they carried out
long computations for the determination of the magnitude.
duration, beginning and end of an eclipse with numbers which
run into the billions in their integral part and with several sex-
agesimal places for their fractions. Simultaneously they used
memorized tables for the daily motion of the sun and moon
involving many thousands of numbers. Certain elements can be
dated astronomically as referring to an epoch of 1200 A.D. But
the Patlca SiddhAntiki already demonstrates the existence of
186 Chapter VI

these methods seven centuries earlier. And, finally, they go back

to the Greek papyri, though the Indian sources go slightly beyond
the steps known from the Hellenistic sources. One begins with
the "Devaram" period D = 248 days. From this one forms the
Calanilam period C = 11 D + LI = 3031 days. The next step is
new; it consists of forming the Rasa Gherica period R = 4C + D
= 12372 days. The value of R is almost identical with the Ba-
bylonian value for the anomalistic month; one finds 27;33,16,26,
11, . . .. where only the last figure is different from the expected
value (cf. p. 162).
Whatever remains to be clarified, it is evident that the methods
found by Warren still in existence in the 19th century are the
last witness of procedures which go back through the medium
of Hellenistic astronomy to Babylonian methods of the Seleucid
period. I do not doubt that this specific case of the lunar theory
is only one of many similar instances where very close contact
between Hindu astronomy and originally Babylonian methods
can be established. We shall return to this question at the end of
this chapter.
When the question of contacts is raised, it might seem tempting
to assume a direct relationship between India and Mesopotamia
without the Hellenistic intermediary. At the present rudimentary
stage of our knowledge of such questions, any definite answer is
more a matter of guess and of taste than of real evidence. Never-
theless, it seems to me more plausible to assume the way through
the Greek and Persian civilization of the Sasanian period than
through a direct contact.
For this I may give three major arguments. First, the fact that
the terminology as well as the methods of Hindu astrology are
clearly of Greek origin; for example, the names of the zodiacal
signs are Greek loan-words. Similarly, the basic concepts of the
planetary theory of the Surya Sidhanta are influenced by the
Greek epicyclic models and not by the Babylonian linear methods.
This argument no longer holds for the linear methods themselves.
But here-and this is my second argument-may be mentioned
that precisely the coastal region from which our information about
Tamil astronomy comes was a center of Roman trade. We have
ample evidence for this, e. g., the anonymous "Periplus of the
Erythrean Sea", written in the first century A.D., which contains
 Origin and Transmission of Hellenistic Science 167

a detailed account of the commerce between Egypt and India,

the harbors and the kind of goods that were traded, etc. This is
fully corroborated by .' ~chaeological evidence, most drastically
in 1946 by the discovery of a large Roman emporium in Arika-
medu in the outskirts of Pondicherry, the very same place where
LeGentillearned in 1769 for the first time about the linear methods
from his Tamil informants. This contact with the West has its
climax in the time of Augustus and in the frrst century A.D., but
Roman coin hoards reach into the fourth century. All this is
confirmed by repeated references to the "Yavanas" (i. e., uIoll-
ians" for "Greeks"!» in Hindu astronomy and Tamil literature.
And finally, the chronology of Hindu astronomy: linear methods
as well as trigonometric models point to the early centuries A.D.,
not B.C.
\Vhatever later discoveries might reveal, at present it seems
reasonable to assume that Babylonian methods, parameters and
concepts reached India in two ways, either via Persia or via the
Roman sea routes, but only through the medium of Hellenistic
astronomy and astrology.
68. We have already remarked that essential parameters
ascribed by Ptolemy to Hipparchus are identical with the cor-
responding parameters of the Babylonian theory. As long as
Ptolemy was our only source, it was natural to assume that
Hipparchus's theory was strictly geometrical and exactly of the
type later followed by Ptolemy. Now, with the linear methods
coming gradually to light, the situation becomes much more
complex. There can be no doubt that Hipparchus used geo-
metrical methods; we know this, e. g., from his determination
of the eccentricity of the solar orbit, as related by Ptolemy. But
one can no longer exclude the possibility that he also worked
with linear methods, as would be no more than natural for
anyone who had access to Babylonian material. Greek astro-
logical literature contains several references to linear methods
which are associated with the name of Hipparchus. It has been
customary to discard such references as apocryphal simply
because one thought the great astronomer unworthy of such

1) This includes, of course. the Romans; also the Periplus speaks only about
"Greek" ships sailing for India. Conversely, the "Romans" of Islam are the Byzantine
Greeks.
168 Chapter VI

elementary arithmetical schemes. I do not want to say that we

should now revert to the opposite extreme and consider such
references as genuine without further investigation. But it is
clear that Ptolemy did not intend to be historically complete in
his references to methods used by his predecessors, and that
only a systematic collection of all references to Hipparchus, by
Ptolemy and by all other sources, can help us to obtain a more
complete picture of the contents of his writings.
69. One of the main reasons for the transmission of astronomical
knowledge from one nation to another was undoubtedly the
spread of the belief in astrology as the one science which gave
insight into the causes of the events on earth. It has often been
said that astronomy originated from astrology. I see no evidence
for this theory. It seems to me much more plausible to assume
that one major incentive for the development of astronomy con-
sisted in attempts to achieve regularity in the intercalations of
the lunar calendars. The best description of the true situation
might be the statement that we know equally little about the
origin of astrology or astronomy and that the relative influence
of these two disciplines upon one another is largely a matter of
conjecture. As we have explained in Chapter V, we are on
safe ground for astronomy only for the Seleucid period in
Mesopotamia. Almost all documents concerning astrology in
Mesopotamia belong to the same period, but their number is very
small compared to the astronomical texts. We have about
ten horoscopes from cuneiform tablets, and still fewer texts con-
cern astrological doctrines as they are known to us in such enor-
mous amounts from Greek sources. In Egypt the earliest horo-
scopes, Demotic or Greek, are from the reign of Augustus. To the
same period belong the earliest astrological treatises and a general
discussion about the validity of astrological doctrine. The rapid
spread and enormous development of astrology during the first
Roman imperial period is paralleled in the spread of Christianity,
Mithraism, and related creeds. In neither case can the speed and
extent of this spread be utilized as a chronological argument.
The only chronological criteria can come from the texts them-
selves. One of these criteria is the arrangement of the planets.
In the cuneiform texts of the Seleucid period the standard ar-
rangementis
 Origin and Transmission of Hellenistic Science 169

Jupiter-Venus-Mercury-8aturn-Mars.
The reason for this arrangement is unknown; the commonly
given explanation that the first two planets are beneficiary, the
last two malevolent, with Mercury doubtful, does not appear in
cuneiform sources. The ordinary arrangement in the Greek
horoscopes is
Sun-Moon-Saturn-Jupiter-Mars-Venus-Mercury,
except for eases where an arrangement is chosen which depends
on the special horoscope, that is, following the positions of the
celestial bodies in the zodiac at the given moment. I think these
two lists reflect the difference between the two astronomical
systems very clearly. The Babylonian system has nothing to do
with the arrangement in space. The Greek system, however,
obviously follows the model which arranges the planets in depth
according to their periods of sidereal rotation. This is reflected
even in the arrangement of the days of the planetary week which
we still use today. Here the Sun is placed between Mars and
Venus, and the Moon below Mercury. Every one of the 24 hours
of a day is given a "ruler" following this sequence. Beginning,
e. g., with the Sun for the first hour one obtains

hour 1 2 3 4 5 24
day 1
Sun Venus Mercury Moon Saturn... Mercury
day 2
hour 1 2 3 ...............
24
Moon Saturn Jupiter ...............
Jupiter
day 3
hour 1 ...... ...
. etc.
Mars

The "ruler" of the first hour it then considered to he the ruler

of the day and thus one obtains for seven consecutive days the
following rulers
Sun Moon Mars Mercury Jupiter Venus Saturn
which is our sequence of the days of the week and also the
arrangement of the planets in Hindu astronomy.
Here we have a system which is obviously Greek in origin
not only because it is based on the arrangement of the celestial
170 Chapter VI

bodies according to their distance from the earth but also because
it supposes a division of the day into 24 hours, a form of reckoning
which is not Babylonian but a Hellenistic product of ultimately
Egyptian origin (cf. p. 86). It is totally misleading when this
order is called "Chaldeanu in modern literature.
As we have said before, the astrology which is known to us
from the Assyrian period is quite different from the Hellenistic
personal astrology. The predictions concern the king and the
country as a whole and are based on observed astronomical
appearances, not on computation and not on the moment of
birth. In addition, the zodiac never appears. Hellenistic horo-
scopes, however, concern a specific person and depend upon
the computed position of the seven celestial bodies and of
the zodiacal signs in their relation to the given horizon, for
a given moment, the moment of birth. Around this is woven
an enormous system of doctrines concerning the evaluation of
these data and of secondary data which can be derived by
all kinds of artifices so as to obtain a greater variety of pos-
sibilities. It is interesting to observe that the actually preserved
horoscopes contain very little, if anything at all, of these theoretical
speculations. The great majority contain nothing but the bare
results of the computations for the given moment. This makes
these texts useful documents for the study of purely astronomical
and chronological questions, but they help us very little for the
history of astrology as such and of the astronomical methods
imbedded in its doctrines. Nevertheless, the patient work of
scholars like Bouche-Leclercq, Cumont, Boll, Bezold, Kroll, Rehm
and many others has shown the existence of different components
of diverse origin. There exist predictions which fit only very
specific circumstances, like the destruction of the Persian empire
by Alexander or the wars between his successors in Syria;
finally, there is a great mass of references to Egypt under the
rule of the Ptolemies. The references to constellations, especially
their simultaneous risings and settings, made it possible to dis-
tinguish between two widely different celestial maps, a "sphaera
barbaricau and a "sphaera graecanicau • Yet the fact remains that
the evidence for direct borrowings from Babylonian concepts
remains exceedingly slim. The main structure of the astrological
theory is undoubtedly Hellenistic. On p. 68 we have discussed
 Origin and Transmission of Hellenistic Science 171
the remnants of the oldest available catalogue of stars, contained
in the astrological writings which go under the name of Hermes
Trismegistos. The fact that these star coordinates correspond to
the time of Hipparchus or his direct followers (cf. p. 69) is an
added argument for the origin of this major work of astrology
in the second century B.C.
Though it is quite plausible that the original impetus for
horoscopic astrology came from Babylonia as a new develop-
ment from the old celestial omens, it seems to me that its actual
development must be considered as an important component of
Hellenistic science. To a modern scientist, an ancient astrological
treatise appears as mere nonsense. But we should not forget that
we must evaluate such doctrines against the contemporary back-
ground. To Greek philosophers and astronomers, the universe
was a well defined structure of directly related bodies. The concept
of predictable influence between these bodies is in principle not
at all different from any modern mechanistic theory And it stands
in sharpest contrast to the ideas of either arbitrary rulership of
deities or of the possibility of influencing events by magical
operations. Compared with the background of religion, magic and
mysticism, the fundamental doctrines of astrology are pure
science. Of course, the boundaries between rational science and
loose speculation were rapidly obliterated and astrological lore
did not stem-but rather promoted-superstition and magical
practices. The ease of such a transformation from science to
humbug is not difficult to exemplify in our modern world.
70. To the historian of civilization, astrology is not only one
of the significant phenomena of the Hellenistic world but an
exceedingly helpful tool for the investigation of the transmission
of Hellenistic thought. As an example may be quoted Abu
Ma'shar, who died in 886 and is an early representative of
Hellenistic astrology among the Arabs. Boll has shown that he
utilized a Persian translation, made in 542, of the "Sphaera
Barbarica" of Teukros. Thus Abu Ma'shar becomes an impor-
tant source for an early Hellenistic lore of constellations. The
famous astrological paintings in the Palazzo Schifanoja in Ferrara,
made in the second half of the 15th century, are influenced by
the doctrines of Abu Ma'shar's astrology. On the other hand,
his writings were translated into Latin, into Greek, into Hebrew,
172 Chapter VI

and from Hebrew into Latin. These "translations" are often only
freely handled versions, incorporated in other material of diverse
origin. There exist even complete cycles of translations and
borrowings from Greek back into Greek. For example, chapters
from an astrological poem in hexameters by Dorotheos of Sidon
(first century A.D.) were used by Abu Ma'shar, who in turn
provided the prototype for a Byzantine dialogue "Hermippos".
Similar cycles can be established for astronomical tables and
treatises which reached Byzantium in the 13th century.
There is found, however, in Abu Ma'shar's writings another
component which makes them of great interest for our problem
of tracing the transmission of Hellenistic science. Indian asterisms
appear in Abu Ma'shar, and their source is found in the astro-
logical writings of Varlba Mihira, the same author of the sixth
century A.D. in whose astronomical work we found the use of
the linear methods for the lunar motion, otherwise known to
us from Greek papyri and finally from cuneiform tablets. Fol-
lowing the unmistakable traces of very specific astrological doc-
trines, one can reconstruct the road which connected Hellenistic
Mesopotamia with Hellenistic Egypt, with pre-Islamic Persia,
and with India. We are obviously entitled to aSSUlne that the
same road was followed by the transmission of mathematical
astronomy even if no more is available to us than the two extremal
ends in Mesopotamia and India.
In the ease of the lunar theory, at least one missing link, the
papyri, were available. In the ease of the planetary theory,
however, not even that much is known from Greek sources.
Nevertheless, we can now understand whole sections in Variba
Mihira's Pailca Siddhlntiki by means of the Babylonian plan-
etary texts. We have seen how the planetary phenomena were
described in Babylonian texts by means of step functions which
we generally called "System AU. Precisely the same idea is
found in the Pailca Siddhintiki. The same holds for fundamental
period relations and even for special parameters. A few examples
may be quoted. For Saturn and Jupiter we know from cuneiform
sources the synodic periods

94,16 = 28;26,40
and 6,31
36
= 10;51,40
 Origin and Transmission of Hellenistic Science 173

and for Venus the synodic arc of 3.35;30°. All three values are
used by Varlha Mihira.
Very strange values seemed to be assigned by Varaha Mihira
to the duration of the synodic revolutions of the planets; for
example.
Mars 768f days
Mercury II4/, days
Jupiter 3931 days
Venus 5751 days
Saturn 3721 days.

The comparison of these numbers with Babylonian parameters

immediately gives the solution of the problem. Not "days" are
meant. but degrees. Indeed the mean synodic arc for Mercury
is 1.54;12.24•... = 5~~2 which is exactly the Hindu value; and
the same agreement is found for Venus. The mean synodic
arc for Mars. however. is. according to the Babylonian theory.
6.48;43.18.. .. which is very close to 6.48;45 = 408! whereas
the Pailca Siddhintiki gives 768!. But the difference is 360°.
or one complete rotation. This explains also the replacement of
"degrees" by "days"; the Hindu "days" are simply sexagesimal
fractions of a sidereal year. or. expressed differently. the mean
solar motion is assumed to be 1° per day. This is confirmed for
I

the case of Jupiter and Saturn. Instead of 39S; days we con-

sider only 33+ = 33;8.34•... ° and this is again in good agree-
ment with the Babylonian value 33;8.45° for the mean synodic
arc. Similarly for Saturn: 3721 - 360 = 12;40° as compared
with 12;39.22.30° in the Babylonian theory.
We stand today only at the beginning of a systematic investi-
gation of the relations between Hindu and Babylonian astronomy.
an investigation which is obviously bound to give us a greatly
deepened insight into the origin of both fields.
71. The fact that a close relationship between Babylonian
linear methods and sections of the Pailca Siddhintiki can be
established is only one facet of the general problem of the evalua-
tion of the role of Hindu astronomy in the history of science.
In it the Panca Siddhintiki of Variha Mihira is of special im-
174 Chapter VI

portance because it forms a chronological fixed point of the first

importance. Varlba Mihira's date is well established by his use
of the year 427 Saka Era = 505 A.D. as an epoch and by con-
siderations which lead to about 590 A.D. as an upper limit for
his lifetime. But it is important not only that the Paiica Siddhin-
tiki is in this way a well dated and comparatively early docu-
ment; it is also a historical source of a unique character in Hindu
astronomy. Its name indicates that it is based on five Siddhintas,
and it actually contains a summary of the contents of the five great
astronomical treatises which were in existence in Variha Mihira's
time. Thus we have here an early historical report on source
material which is no longer extant, or at least no longer extant
in exactly the same form. On the other hand Varlba Mihira is
also one of the main sources of al-Birfini's report on Hindu
astronomy and astrology, written about 1030 A.D. Consequently
Varlba Mihira occupies a central role for the study of Hindu
astronomy.
The Siirya Siddhinta has to be considered as the main canon
of Hindu astronomy. It is supposed to have been revealed by
the Sun (Siirya) at the end of the Golden Age (2163102 B.C.)
to a Maya Asura. Some manuscripts contain the additional com-
mand of the Sun to Maya: "Go therefore to Romaka-city, your
own residence; there, undergoing incarnation as a barbarian,
owing to a curse of Brahma, I will impart to you this science"l).
This is closely paralleled by a passage in Varlba Mihira: caThe
Greeks, indeed, are foreigners, but with them this science [astro-
nomy] is in a nourishing state"lI). The origin of the SOrya Sid-
dhinta is dated by modern scholars to about 400 A.D. while its
present version may be as late as about 1000 A.D. That this
work contains several much older and very primitive sections
which it combines rather startlingly with the Greek theory of
epicyclic motion has been apparent to all scholars since al-
Biriini, who characterizes Hindu mathematical and astronomical
literature somewhat drastically as caa mixture of pearl shells
and sour dates, or of pearls and dung, or of costly crystal and
common pebbles"').
1) SOrya Siddhlnta I, 8 (Burgess).
') Brhat SaAblta II, 15 (Kern). An Inslgnlftcantly dUferent translation Is pro-
posed in Isis 14 (1930) p. 391 or in the translation by V. Subrahmanya Sastri,
Bangalore 1947, p. 19 (Sloka 14).
') India I (Sachau I, p. 25).
 Origin and Transmission of Hellenistic Science 175

Though the Greek influence on the Surya Siddhanta is evident,

it is equally obvious that the Greek theory has undergone a quite
independent transformation in many details both with respect
to the values of numerical constants and to the general theory.
That modifications of this type went on continuously is explicitly
proved by the comparison of the existing Siirya Siddhanta with
the summary of VarAha Mihira in the Panca Siddhantika. On
the whole, however, it seems as if the Siirya Siddhanta were the
most consistently modified Hindu treatise. What we know about
the Romaka and the Paulisa Siddhanta from the material which
was incorporated in the Pafica Siddhantika seems to bring these
treatises closer to the Hellenistic sources. Their names support
this view; the "Romans" are, of course, the Greeks of the Roman
or Byzantine empire; and al-Biruni considers the Paulisa Sid-
dhanta the work of Paulus Alexandrinus, an astrologer of the
fourth century A.D. Thus we obtain again as the period of con-
tact roughly the time of origin of the Surya Siddhinta, i. e.,
about 400 A.D. And it is perhaps significant that the earliest
occurrence of the place value notation can be traced back again
to the Paulisa Siddhanta.
On the other hand, it has long been recognized that the bor-
rowings of Hindu astronomy from Greek astronomy show no
influence of the refinements of the Ptolemaic theory. Also the
astrological theory reflects, at least partially, the oldest strata of
Hellenistic doctrine. This would lead to a period between Ptolemy
(150 A.D.) and Hipparchus (150 B.C.) or even slightly earlier.
This seents to leave a serious gap of several centuries between
the date of the Hellenistic sources and reception by the Hindus.
There is, however, increasing evidence forthcoming to bridge
this gap. Ibn Yunus, the great Arabie astronomer (died 1009),
famous as author of the Hakemite tables, quotes Persian ob-
servations of the apogee of the solar orbit made about 470 A.D.
and 630 A.D. Nallino has shown that both Teukros and Vettius
Valens were translated into Pehlevi, the pre-Islamic or Middle-
Persian Iranian language. We have already seen that Teukros
(first century B.C.?) accounts for early elements in Greek astro-
logy; Vettius Valens is a younger contemporary of Ptolemy but
uninfluenced either by the Almagest or by the Tetrabiblos. The
otlquoted passage where Vettius Valens declares that he "did not
compute eclipse tables himself but used Hipparchus for the sun,
176 Chapter VI

Sudines and Kidenas for the moon", relates him directly with
the linear methods of the Babylonians, and with whatever
geometrical or al"ithmetical methods were used by Hipparchus.
If we assume that these sources reached Persia without being
modified by the scientific theories of Ptolemy, then we have a
satisfactory explanation for the main features both of the linear
and of the geometrical methods found in the Paiica Siddhintiki.
72. The way back leads again via Persia. It is well known
that the scientific activity of Islam originated under the Abbasid
Khalifate in Baghdad; al-Khwirizmi, Thibit ibn Qorra, Abu
Macshar belong to this period (9th century).
Al-Khwirizmi's astronomical tables have been preserved
through Latin translations; they show a curious mixture of Hindu
and Greek methods. The relation between his mathematical writ-
ings and the Hellenistic tradition has already been mentioned. A
century later appears al-Biriini, another native of Khorazm. He
not only transmitted Hindu knowledge to the West, but the tells
us that "most of their books are composed in Sloka [verses],
in which I am now exercising myself, being occupied in com-
posing for the Hindus a translation of the books of Euclid and
of the Almagest, and dictating to them a treatise on the con-
struction of the astrolabe, being simply guided herein by the
desire of spreading science". On the other hand, al-Birnni trans-
lated an astrological work of Varlha Mihira into Arabic!).
The history of the transmission of Hellenistic science throughout
the Islamic world need not be told here. The general trend is no
longer in doubt and has often been described. What is less
generally known, however, is the fact that for every specific
question of astronomical or mathematical theory we are still
groping in the dark because of a most deplorable lack of edited
source material. With the splendid exception of al-Battini's
tables, none of the great astronomical tables of the Middle Ages-
Arabic or Latin, Hebrew of Greek-is available in modern editions
for the period between Ptolemy and Copernicus. The history of
the ancient mathematical sciences is a field in which one need
not go far to find fertile soil ready to be cultivated.
73. We have come to the end of our discussion, which has
brought us back to the civilization of the Middle Ages from which
1) India XIII and XIV (Sachau I. p. 137 and p. 158).
 Origin and Transmission of Hellenistic Science 177

our journey started. By patiently following the connections of

mathematical and astronomical theory we moved from period
to period and from civilization to civilization. Our road often
went parallel to the road pointed out by historians of art, religion,
alchemy, and many other fields. This is not surprising. It only
underlines the intrinsic unity of human culture. The role of
astronomy is perhaps unique only in so far as it carried in its
slow but steady progress the roots for the most decisive develop-
ment in human history, the creation of the modern exact sciences.
To follow this specific aspect of cultural history seems to me
worthy of our efforts, however fragmentary our results may be.
In the "Cloisters" of the Metropolitan Museum in New York
there hangs a magnificent tapestry which tells the tale of the
Unicorn. At the end we see the miraculous animal captured,
gracefully resigned to his fate, standing in an enclosure sur-
rounded by a neat little fence. This picture may serve as a simile
for what we have attempted here. We have artfully erected from
small bits of evidence the fence inside which we hope to have
enclosed what may appear as a possible, living creature. Reality,
however, may be vastly different from the product of our imagi-
nation; perhaps it is vain to hope for anything more than a picture
which is pleasing to the constructive mind when we try to restore
the past.

BIBLIOGRAPHY TO CHAPTER VI
A good introduction to the political and cultural history of Hellenism is
w. W. Tarn and G. T. Griffith, Hellenistic Civilisation, London, Arnold,
1952.
The following work gives an excellent picture of the multiple paths of scientific
contacts duriDg the Middle Ages in the West: Charles Homer Haskins, Studies
in the History of Mediaeval Science. 2nd ed. Cambridge, Harvard University
Press. 1927.
One of the most remarkable books about Hindu science and its relations to
the Muslims is Alberuni's India, An Account of the Religion, PhDosophy,
I

Literature, Geography, Chronology, Astronomy, Customs, Laws, and Astrology

of India about A.D. 1030; an English edition with notes and indices by Edward
C. Sachau, London 1910. For special works on Hindu astronomy cf. the notes
ad No. 67. Cf. also M. Steinschneider, Zur Geschichte der 'Obersetzungen
aus dem Indischen ins Arabische etc., Zeitschr. d. Deutschen Morgenllndischen
Gesellschaft 24 and 25 (1870 and 1871).
178 Chapter VI

For the influence of Iranian thought during the Hellenistic period see Joseph
Bidez et Franz Cumont, Les mages hellenises; Zoroastre, Ostanes et Hystaspe
d'apres la tradition greeque; 2 vols. Paris, 1938.
For the early relations between Greeks and India see W. W. Tarn, The
Greeks in Bactria and India, 2nd. ed. Cambridge 1951 and A. K. N arain, The
Indo-Greeks, Oxford 1957.
As a summary of Hindu science should be quoted the chapter on science by
W. E. Clark in "The Legacy of India" (edited by G. T. Garratt, Oxford 1937).
A detailed summary of the literature up to 1899 is given by G. Thibaut in his
article "Astronomie, Astrologie und Mathematik" in vol. III, 9 of the uGrundriss
der Indo-Arisehen Philologie und Altertumskunde". Very useful is James
Burgess, Notes on Hindu Astronomy and the History of our Knowledge of It
(J. of the Royal Asiatic Soc. of Great Britain and Ireland, 1893, p. 717-761)
where one finds complete references to the early literature which contains much
important information which is no longer available otherwise.
The translation by E. Burgess of the SOrya Siddhinta, quoted below p. 186,
contains extensive commentaries which must be read by any serious student
of this subject. For the "linear methods" in Hindu astronomy cf. the references
to Le Gentil and Warren on p. 186. For the form which the Greek theory
of epicyclic motion of the planets took in India and then in al-KhwirizmI, see
O. Neugeba uer, The transmission of planetary theories in ancient and medieval
astronomy, Scripta Mathematica, New York, 1956.
E. S. Kennedy, A survey of Islamic astronomical tables, Trans. Amer.
Philos. Soc., N.S. 46 (1956) p. 123-177 is a publication which shows the great
wealth of material still available but barely utilized for the investigation of
medieval astronomy, its Greek, Islamic and Hindu sources and their interaction.

NOTES AND REFERENCES TO CHAPTER VI

ad 61. The date of Heron Alexandrinus is fixed in the second half of the
first century A.D. by the fact that he quotes as an example the lunar eclipse of
A.D. 62 March 13 [cf. Neugebauer. Kgl. Danske Vidensk. Selsk., hist.-fiIol.
medd. 26, 2 and 7 (1938 and 1939) and A. G. Drachmann, Centaurus I (1950)
p. 117-131].
Our only certain knowledge about the date of Diophantus rests upon the fact
that he quotes Hypsicles (opera I p. 470, 27 and 472, 20, ed. Tannery) and
that he is quoted by Theon Alexandrinus (Comm. in Almag. I, p. 453, ed. Rome).
The date of Hypsicles can be roughly estimated as about 150 B.C.; Theon's
date is fixed by the solar eclipse of 364 June 16. One now usually follows
Tannery, who argued for a date about 250 A.D. Tannery's argument is based
on an admittedly corrupt passage from Psellus (about 1050) and is in general
very hypothetical, as has been pointed out by J. Klein in Quellen und Studien
z. Gesch. d. Math.• ser. B vol. 2 (1934) p. 133 note 23. Klein himself favors an
earlier date and suggests possible contemporaneity with Heron. Klein's arguments
are largely based on the stylistic similarities between Heron and Diophantus,
an argument which has lost much of its weight since we have become aware
that both authors more or less represent common Oriental-Hellenistic tradition.
 Notes and References 179
The dedication in both eases to a certain Dionysius is not decisive, not only
because this name is very common but also because of the use of difTerent
tiUes given to Dionysius.
A medieval dating is preserved in the "History of Dynasties" of Bar Hebraeus
(1226-1286), also called Aba 'l-Faraj, a learned Jacobite bishop. His "History
of Dynasties" is written in Arabic1) and is an abridged and modified version of
his Syriac "Chronography"·). Only the Arabic version quotes Diophantus and
his "Algebra", referring it to the time of Julian the Apostate (361-363). This
date is just barely reconcilable with the upper limit for Diophantus's lifetime,
mentioned above. That Bar Hebraeus was a competent mathematician and
astronomer is weD known8), but we do not know on what authority his chronolo-
gical statement was based. I see no possibility for confirming or disproving it
and it seems to me that we must admit that Diophantus cannot be dated
with any accuracy within 500 years.

Fig. 28.

Writings of the type of Heron's "Geometry" were undoubtedly widespread

in antiquity and formed the backbone of instruction in elementary mathematics.
This explains the relatively large number of papyrus fragments containing such
texts. As an example can be shown an unpublished papyrus of the CorneD
CoDection (d. Pl. 12). The figure in the lower part of the right column is a
typical example of the building up of a more complicated example from the
simplest cases (cf. Fig. 28).
It is precisely from the construction of such examples that we can demonstrate
direct relations between geometrical treatises. For example, the concept of
isosceles triangle is illustrated by a triangle of side 10 in Heron, Metriea I,
XVII (opera III p. 48, 49); in Heron, Geometrica 10 (opera IV, p. 224, 225);
al-Khwlrizmi, Algebra (Rosen p. 80); al-Blroni, Astrology 22 (Wright). The
general triangle of sides 13, 14, and 15 is used by Heron, Metrica I, V and VIII;
Geometriea 12; Mishnat ha-Middot 9 (Gandz p. 46); Mahlvlra VII, 53 (ed.
Rangacharya p. 199); al-Khwlrizmi p. 82; Bhaseara, Ldlvati VI, 165, 168
(Colebrooke p. 71, 73). In fact this triangle is composed of two Pythagorean
triangles of sides 13, 5, 12, and 15,9, 12 respectively. They occur again in two
problems of the Pap. Ayer (Am. J. of Philology 19, 1898, p. 25 ft.).

1) Edited with Latin translation by Edward Poeoeke. Oxford 1663.

I) Edited with English translation by A. W. Budge. Oxford 1932.
e.
8) Cf•• g•• F. Nau. Le livre de l'aseension de l'esprit ••• Cours d'astronomie
redig~ en 1279 par ••• Bar Hebraeus (Biblioth&que de l'kole des hautes ~tudes
121. Paris 1899/1900).
180 Chapter VI

The fact that a Hebrew treatise is part of this tradition (d. S. Gandz, The
Mishnat ha-Middot; Quellen und Studien zur Geschichte der Math., Abt. A.
vol. 2, 1932) is perhaps not as isolated a phenomenon as it may appear. The
Heronic corpus itself contains several references to Hebrew units of measure
(Heron. Opera V p. 210-219). Similarly, many concepts of Judaism are found
in the magical papyri and in related practices involving numbers and the
alphabet, the so-called gematria. There is certainly some basis in the ancient
terminology which uses Umathematici" in the sense of magicians or astrologers.
Another example for the transmission of mathematical knowledge is found
in al-BlrOnl's India XV (Sachau I p. 168 f.), where he says uBrahmagupta
relates with regard to Aryabhata ... that he fixed the circumference as 3393, ...
the diameter as 1080". The reason for this expression. which contains a common
factor 9, becomes obvious if one recalls that 1080 is an important metrological
unit in oriental astronomy. As an example can be mentioned the division of
one hour in 1080 parts (chelakim) in Hebrew astronomy. The sexagesimal
equivalent of :~:~ is 3;8.30 which is the approximation of n used in the Almagest
(VI. 7).
In the same section al-BlrOnl states that "Pulisa employs ... in his calcula-
tions ... 3 11;:0' ... The same relation is derived from the old theory, which
Ya~~ub ibn rlri~ mentions in his book. Compositio Sphaerarum, on the
authority of his Hindu informant ..... Indeed the same value is used by al-
+
Khwlrizml (Algebra, ed. Rosen p. 72 198 f.). Its decimal equivalent is 3.1416,
its sexagesimal equivalent 3;8,29.45,36. For a discussion of these and related
values see B. Datta, Journal and Proceedings of the Asiatic Society of Bengal.
N.S. 22 (1926) p. 25-42. esp. p. 26 f.
The use of the value 3 for n is commonly found in Hellenistic texts and goes
back to Old Babylonian mathematical texts. Another example common to
Babylonian and Heronic mathematical education is the computation of the
volume of "ships" of prismatic form (d. Neugebauer, MKT II p. 52 and
Heron, Opera V pp. 56, 128, 130, 172).
Traditionally it is assumed that Hellenistic science reached the Arabs through
the intermediate stage of Syriac versions of the Greek works. Though this may
be so in many cases it is certainly an oversimplification. G. Bergstrlsser edited
and translated in 1925 a work of Hunain ibn Ishlq concerning the translations
of the writings of the physician Galen, a contemporary of Ptolemy (Abh. fiir
die Kunde des Morgenlandes 17, No.2; d. also Max Meyerhof, New Light
on ~unain ibn IshAq and his Period, Isis 8, 1926, p. 685-724). ~unain, a
Nestorian, played a central role in the early phase of Arabic translation. He
was born in 809 and died in 877; his search for Greek manuscripts led him an
over the Near East and to Alexandria. He must have seen and compared
hundreds of them and accumulated a large collection of his own. From his
report one learns how these translators worked, comparing defective manuscripts,
restoring, explaining, excerpting. There is no such simple sequence, Greek ~
Syriac ~ Arabic, visible. By far the greatest number of works is directly trans-
 Notes and Referenees 181
lated from Greek either into Arabie or into Syriac. There are many eases where
Syriac translations were the basis for Arabie versions but also the opposite
order oeeurs. ~unain's report covers about 50 years and concerns more than
130 books ascribed to Galen or his sehool. Only 10 titles were not translated
aecording to ~unain. From the rest 179 Syriae and 123 Arabie versions were
listed by lIunain, of whieh he himself contributed 96 and 46 translations
respectively, not counting revised versions. Of these translations 81 were made
for Arabie eustomers, 73 for men who read Syriae.

.x
.
~ ~,~

.. ..
~" ,
.
....
,,
..
, '

.....-=--•;•;p. "5'"
", :
", :
____-J
'R
Fig. 29.

What is said here about medical literature mayor may not hold for mathe-
matieal or astronomical works. One must not forget that praetical need and
local conditions might have been widely different for different fields of learning.
For astronomy, e. g., a transmission via India undoubtedly plays a great role
whereas we have very little evidence for a Syriae intermediary.
ad 62. The theory of "applieation of areas" attained great importanee in
ancient mathematies beeause of the diseovery that the eonie seetions could be
incorporated in this theory. Indeed, our modem names ellipse, hyperbola, and
parabola are direetly taken from this theory. The ease of the ellipse might be
quoted as an example. Assume as given two "coordinated" directions (from
which our use of the word "coordinates"), here denoted as the tE- and II-direetion
(Fig. 29). Let E and 11 be two given parameters, to be represented by line segments
in the tE-direetion and perpendieular to it, respeetively. Then a point P(tE, 11)
will be a point of an ellipse with OQ == E as diameter and with the II-direetion
as conjugate direction if the area of the square with side 11 equals the area of
the rectaDgle :a;' whieh is "applied" to OR == 11 in sueh a way that a small
reetangle (RS) remains whose sides have the given ratio of 11 to E. This is only
182 Chapter VI

a slight generalization of the "elliptic" ease of the application of areas described

in the text (p. 149 f.), where the remaining reetaDg1e was a square. In our notation
P(:.:, ,) is determined by
:.:' 'I
11:==' - - : = -.
f -:.: f

In the case of the hyperbola one requires an excess for the rectangle :a:', whereas
the parabola corresponds to exact equality of If and :I:fJ. For detaDs and figures
ef. O. Neugebauer, Apollonius-8tudien, Quellen und Studien zur Geschichte
der Mathematik, Abt. B vol. 2 (1932) p. 215-254.
The historical sequence of these discoveries seems to be as follows. Since
Old-Babylonian times the knowledge of solving the main types of quadratic
equations existed. The discovery of irrational quantities led to the geometrization
of these methods in the form of application of areas (4th century B.C.). Shortly
afterwards the conic sections were discovered, as I think, from the investigation
of sun dials (cf. p. 226). At any rate the conic sections were at first considered
as curves in space, unrelated to algebraic problems. Finally the relation to the
application of areas was established. as found in Apollonius (3rd century B.C.).
Figures which illustrate the configurations in space from which the relations
between the plane areas were derived are given in Quellen und Studien zur
Gesehiehte der Mathematik, Aht. B vol. 2 (1932) p. 220 f.
ad 63. The relationship between mathematics and Plato's theory of Ideas
has been the subject of innumerable publications. For a realistic discussion of
the whole problem cf. H. F. Cherniss, The Riddle of the Early Academy,
University of California Press, 1945.
The uneasiness which a good Platonist felt when he was dealing with astro-
nomical theories based on observations is nicely seen in the introduction of
Proclus to his "Hypotyposis" (Greek with German translation by Manitius,
Leipzig, Teubner, 1909).
ad 64. The eight-shaped curve on which a planet P moves according to the
combined motion of two inclined spheres was called "Hippopedett • Its qualities
can easily be deduced from the fonowing consideration (Fig. 29a). We project
the motion of the planet on the plane of the great circle BP0T, which corresponds
to the horizontal eirele in Fig. 24 p. 154. The planet itself moves on the inclined
plane BP' which appears in our projection as an ellipse. If the motion of the
planet started. at R, we can obtain its position after a motion of angle IX by
first moving by this amount from R to p' and then turning P' back by -IX. Since
p' as point of an ellipse is the vertex of a right triangle OPT0' we see that P is
the vertex of a congruent right triangle SPR. Thus P runs once through a eirele
of diameter RS while P' moves through the semicircle RT. And since the angle
at 0 and therefore also at S is IX, we see that the are from R to P is 2 IX. Thus P
moves on its circle with twice the angular velocity of P'. Since P in Fig. 29a
represents only the projection of the planet, the orbit on the sphere is the inter-
section of the sphere with a straight circular cylinder of diameter SR. This gives
the two loops of the Hippopede, with R as double point (Fig. 24).
A detailed comparison between the Ptolemaic theory of the motion of the
moon and the modern theory was given byA. F. Mobius, Gesamme1teWerke IV
 Notes and References 183
T
....
I
.".'
,
I
. I
I
I
I

,,• ·
I
•
I
•
I
I

·••.
I

•..
..
,
'. '.
--~~--
R

Fig. 29a.

("Meehanik des Himmels"). Cf. also Paw Kempf. Untersuchungen tiber die
ptolemiische Theorie der Mondbewegung. Thesis. Berlin 1878; C. J. Schu-
macher. Untersuchungen tiber die ptolemiiische Theorie der unteren Planeten;
MUnster, Aschendorfl. 1917; P. Boelk. Darstellung und PrUfung der Mereur-
theorie des Claudius Ptolemaeus. Thesis. Halle. 1911. The transformation from
the geocentric to the heliocentric system is often hailed as one of the greatest
discoveries of modem science. though foreshadowed by the Greek genius. In
fact. however. the equivalence of these two modes of description of the observable
phenomena had scarcely been forgotten by the astronomers of the Middle Ages.
Aryabhata (about 500 A.D.) argued for a movable and rotating earth (Aryab-
hatiya IV. 8; trsl. Clark p. 64 fl.) and al-Biriini (1030 A.D.) remarks in his
"India" (trsl. Sachau I p. 276 fl.) rather casually. "Besides, the rotation of the
earth does in no way impair the value of astronomy. as all appearances of an
astronomic character can quite as well be explained aecording to this theory
as to the other:'
ad 65. Cf. also O. Neugebauer, Notes on Hipparehus. Studies presented to
Hetty Goldman, New York 1956. p. 292-295.
ad 66. It seems to me possible that a horoscope for the year 137 A.D. (P. Paris.
19. lines 11/12) has preserved for us the ancient name for the "linear methods".
If we are correct in restoring this passage, the astrologer tells us that he computed
the position of the sun according to the method of "greatest and smallest [veloc-
ity)". This would be an appropriate description of a linear zigzag function. as
used for·the solar velocity in System B of the Babylonian theory.
The development of the ancient and mediaeval concept of "clima" can
be described quite simply. In Babylonia originated the norm that the ratio M:m
of longest to shortest daylight was 3:2 as well as the two "systems" A and B
for the rising times which determine the variation of length of daylight during
the year.
These arithmetical methods were transplanted to Alexandria. In the second
century B.C. we find Hypsicles using the System A of rising times for the computa-
tion of the length of daylight in Alexandria. the only modification being that
184 Chapter VI

Fig. 30.

M:m is now given the value 7:5, or M = 3,30° == 14b, m = 2,30° == lOb. The
same can be done, of course, for any given ratio M: m. But it is characteristic
that this expansion to arbitrary localities was again made in a strictly linear
fashion by varying M in constant steps of 4°. In this way seven zones or "cli_
mates" were distinguished, number one being Alexandria (M == 3,30), number
two M == 3,34, and so forth to 3,38 ••• up to a seventh climate with M == 3,54
(about 43° north). Unfortunately, the classical value M == 3,36 for Babylon
does not fall into this scheme. Consequently, we find a second type of division,
again in steps of 4°, but now with Babylon and M == 3,32 as the starling point
(though always ealled "second e1ima", a fact which demonstrates the Alexandrian
origin of all these devices). Since we have a choice, in all eases, between System A
and System B for the rising times, we have in principle four possibilities for the
rising times in each "clima": either M == 3,30 (Alexandria) plus a multiple
of 4°, or M == 3,36 (Babylon) plus a multiple of 4° to be used either with rising
times of System A or of System B. This explains the great variety of figures found
in the texts for the rising times for different climates.
As in so many other eases, no trace of these primitive schemes is tolerated
in the Almagest There Ptolemy gives (in Book II) a table of rising times for
ten zones (beyond the equator, which is the so-called "sphaera recta") such
that M increases from zone to zone by one-half hour. Alexandria with M == 14b
(here called "the low country of Egypt") is obviously part of this system, which
is extended until M == l7 h (" = 54;1°). In practice, however, again only seven
of these zones were accepted as "climates", this Ume beginning with the second
of Ptolemy's zones ("Meroe" with M == l3b ) and ending with M == 16b or
M:m == 2:1 (Southern Russia). In this system Alexandria lies in the third
e1ima, the next being "Rhodes", the fifth "Hellespont" (M == l5b).
In later periods new additions were introduced according to special needs.
For example, for Byzantium M == 15;15b was adopted, but sueb new zones
did not interfere with the standard system of the "seven climates" which re-
mained a basic element in mediaeval geography. Only gradually did the geo-
graphical latitude replace the greatest length of daylight as the defining parameter
of a locality. This development reaches its climax in the tables of rising times of
al-KlshP) which proceed from degree to degree of geographical latitude, from
sphaera recta to " == 60°.

1) Unpublished; cf. E. S. Kennedy, Trans. Am. PhDos. Soc. N.S. 46 p. 164.

 Notes and References 185
The inventor of the system of "seven climates" is unknown: a faet which is
demonstrated by the great number of theories forwarded to solve this problem.
All that is reasonably eertain is that it originated in Alexandria. There is no
trace in Babylonian astronomy of varying values of M according to geographieal
latitude. The linear methods are frequently used in astrologieal literature and
it is therefore not surprising that we fmd the rising times for Babylon (System A)
even in India (YarAha Mihira, Brihat Jataka I, 19). It is incorrect, however, to
distinguish "astrologieal" and "geographical" elimates.
For the history of the climates see Ernst Honigmann, Die sieben Klimata
(Heidelberg, Winter, 1929) and my paper "On Some Astronomieal Papyri and
Related Problems of Ancient Geography" in the Trans. Am. Philos. Soe., N.S.
32 (1942) p. 251-263. For the rising times in Babylonian astronomy d. my paper
in J. Cuneiform Studies 7 (1953) p. 100-103 and ACT I p. 240.
The great importanee of the problem of ascensions is also visible in the only
preserved writing of Hipparehus, his commentary to Aratus. In a truly masterful
paper, H. Vogt has investigated Hipparehus's methods ("Versueh einer Wieder-
herstellung von Hipparehs Fixstemverzeiehnis" Astron. Naehriehten 224, 1925,
cols. 17-54). As Boll demonstrated, Hipparehus's eatalogue of stars recorded
about 850 stars. From this last eatalogue about 350 stars are mentioned in the
Alatus Commentary. Vogt investigated 879 numerieal data coneeming these
stars; 471 of them are spherieal eoordinates, 408 eoneem simultaneous risings
or settings or culminations. Hipparehus does not use the ecliptie eoordinates
"longitudes" and "latitudes" which have been standard in star eatalogues ever
since the Almagest. Of the 471 preserved numbers, 64 are declinations, 67 are
right ascensions, and 340 are ecliptie arcs determined by the intersection with
the ecliptie of the eircle of declination through the star (d. Fig. 30).
The aneient astronomers rightly had greater eonfidenee in the accuraey of
their mathematieal theory than in their instruments. As an example may be
quoted the faet that Ptolemy determines the ecliptie coordinates of fixed stars
by measuring their distanee from nearby positions of the moon, whose longitude
is then l4ken from eomputation.
For a fragment of a Babylonian star eatalog d. Saehs, J. of Cuneiform
Studies 6 (1952) p. 146-150.
A model of a meehanical clock, based on stereographie projeetion. is de-
scribed by Diels, Antike Teehnik, 3rd ed. (1924) p. 217 and PI. 18. Stereographie
projection is applied by Ptolemy in his "Planisphaerium" (opera II p. 225-259
Heiberg) in order to determine the rising times by using only plane trigonometry.
The same eonstruetion is used for the "astrolabe", an instrument whieh allows
us to represent the celestial phenomena by means of the rotation of a network
of eircles over a dise representing the plane of the equator upon whieh the
celestial i sphere is projeeted from the south pole. Cf. my paper "The Early
History of the Astrolabe" in Isis 40 (1949) p. 240-256.
ad 67. The Babylonian value 27;33.20 days for the anomalistie month is
also mentioned by Geminus (about 70 B.C.) in his Introduetion to Astronomy
18. The derivation. however, given by Geminus is wrong. He says that 717
anomalistie months contain 19756 days, but from this it fonows that one month
equals 27;33,13,18, ... days and not 27;33,20. Geminus is here telescoping two
dift'eren~ methods into one.
186 Chapter VI

References for the publications quoted in the text (p. 165):

Schnabel, Zeitschr. f. Assyriologie 37 (1927) p. 35 and p. 60.
Knudtzon-Neugebauer, BuB. soc. roy. des lettres de Lund, 1946-1947,
p.77-88.
Neugebauer, Danske Videnskab. Selskab, hist.-fiIol. Meddelelser 32, No.2
(1949).
Thibaut-Dvivedi, The PaichasiddhAntikA, the Astronomical Work of
VarAha Mihira. Benares 1889.
John Warren, Kala Sankalita, a Collection of Memoirs on the Vuious
Modes According to which the Nations of the Southem Parts of India Divide
Time. Madras 1825.
Warren had a predecessor in the French astronomer LeGentil who was
sent to India to observe the Venus transits of 1761 and 1769. He missed the first
because of the French-English war, the second because of clouds. He learned,
however, a great deal about Tamil astronomy and gave an excellent description
of the methods for computing eclipses in the Memoires of 1772, II of the French
Academy. The French scholars of this period, Cassini, LeGentil, Bailly, Delambre,
had reached a clear distinction between the linear methods of the Tamil astron-
omers and the trigonometric type of the Sarya Siddhlnta. This insight has been
lost in the subsequent literature.
Translation of the Sdrya-SiddhAnta by Ebenezer Burgess, reprinted 1935,
University of Calcutta, from J. Am. Oriental Soc. 6 (1860) p. 141-498.
For a discussion of the Tamil methods for the computation of lunar eclipses
cr. O. Neugebauer, Tamil Astronomy, a Study in the History of Astronomy
in India, Osiris 10 (1952) p. 252-276 and B. L. van d~r Waerden, Die Be-
wegung der Sonne nach griechischen und indischen Tafeln, S. B. Bayer. Akad.
d. Wiss., Math.-nat. Kl. 1952 p. 219-232 and by the same author Tamil Astron-
omy, Centaurus 4 (1956) p. 221-234.
There are many evident indications of a direct contact of Hindu astronomy
with Hellenistic tradition, e. g., the use of epicycles or the use of tables of chords
which were transformed by the Hindus into tables of sines. The same mixture of
ecliptic arcs and declination circles is found with Hipparchus (cr. p. 185) and
in the early Siddhlntasl ) (called "polar longitude" and "polar latitude" by
Burgess). The extensive use of the sexagesimal system is common to both
Greek and Mesopotamian astronomy. The use.of utithis··, which are so charac-
teristic of Hindu astronomy, is not yet attested in Greek texts but we know so
little about the linear methods in Hellenistic astronomy that we may assume
that the use of "lunar days" penetrated into Hellenistic astrology from Babylonian
texts exactly in the same fashion as the planetary periods and the lunar theory.l)

I) Hipparehus divides not only the ecliptic but also the equator into 30 0 sections
and denotes them by the names of the zodiacal signs (d. Manitius's edition of
Hlpparehus's Commentary to Aratus p. 295). In the Stirya Slddhlnta. the zodiacal
signs are used in similar fashion to denote arcs on any great circle.
II) The use of ..tithls". that Is, of tbfrthieths of mean synodic months. was first
discovered in Babylonian planetary texts by Pannekoek (Koninklijke Akad.
van Wetensch. te Amsterdam. Proceedings 19. 1916. p. 684-703) and by van der
Waerden (Eudemus 1. 1941. p. 23-48). For the oceurenee of these units in Babylo-
nian lunar ephemerides d. O. Neugebauer, ACT I p. 119. In late Hindu astronomy
 Notes and References 187
The occurrence of the ratio 3:2 for the longest and shortest days might be taken
as a sign of direct Mesopotamian influence though also this element is a part
of the Hellenistic tradition of the "climates". Also the arrangement of the planets
aeeordiDg to the "rulers" of the days of the week (d. p. 169) indicates primarily
Hellenistic intluence.
For the Roman sea-routes and for Roman settlements in India, see R. E. M.
Wheeler, Arikamedu: An Indo-Roman Trading-Station on the East Coast of
India. Ancient India. No. 2 (1946) p. 17-124. For a summary see Martin P.
Charlesworth, Roman trade with India. Studies in Roman Economic and
Social History in Honor of Allan Chester Johnson. Princeton 1951, p. 131-143.
A translation. with extensive commentary. of the Periplus was published by
W. H. Schoff. The Periplusofthe Erythrean Sea. New York-Philadelphia 1912.
A relatively early date for Greek-Persian-Hindu contacts seems to be obtain-
able from a passage in the D6nkart. Book IV. according to which Hindu books
on grammar and on astronomy and horoscopy as well as the Greek Almagest
reached the court of Shapur I (about 250 A.D.); d. Menasce, Journal Asiatique
237 (1949) p. 2 f.
ad 68. Hipparchus is often quoted in the astrological literature. As an
example 'might be mentioned Vettius Valens I, 19 for the elongation of the
moon. This method uses the same epoch (Augustus - 1) as P. Ryl. 27 and is
therefore not genuinely Hippuehian. Nevertheless it may go back to Hippuchus
just as other linear methods were developed from Babylonian originals. It was
F. Boll who first emphasized that the ancient reports connecting Hippuchus
with astrology have to be taken seriously in view of the time of origin of astrol-
ogical doctrine in the second century B.C. (d. Boll's lecture "Die ErforsehUDg
der antiken Astrologie" in Neue Jahrbiicher fiir das Klassische Altertum 21.
1908, p. 103-126). F. Cumont speaks about "Hippuque. dont Ie nom doit
atre place en tAte des astrologues comme des astronomes grecs" (KHo 9. 1909.
p. 268).
ad 69. The earliest known horoscope is cast for the year 410 B.C. (A. Sachs.
Babylonian horoscopes, J. of Cuneiform Studies 6 (1952) p. 49-75). The
remaining cuneiform horoscopes belong to the Seleucid period. The earliest
Greek horoscope is the horoscope of the coronation by Pompey of Antioehus I
of COlDlllllgene in 62 B.C. on the Nimrud Dash. Horoscopes on papyrus or in
Greek literature start at the beginning of our era.
An early indication of knowledge of Babylonian astrology in Greece was
pointed out to me by Professor H. Cherniss. Proclus (who died in A.D. 485) in
his commentary to Plato's Timaeus (III, 151 Diehl) quotes Theophrastus, the
successor of AnstoUe (died 322 B.C.). as saying that the Chaldeans were able to
prediet•.in his time, not only the weather from the heavens but also life and
death of all persons.

the tithis have become of variable length. being thirtieths of the true lunar months.
Thus one finally introduced the variability of these units which had been invented
in order to avoid the fluctuations of the true lunar calendar. In the classical mndu
astronoqay. however• (e. g.. in the Sorya Siddbinta) the titbis are of fixed mean
length; cf. Olaf Schmidt. On the Computation of the Ahargana. Centaurus 2 (1952)
p.140-180.
188 Chapter VI

Still one generation earlier leads to an oft-quoted remark of Cicero (De

divinatione II, 42, 87) that Eudoxus has written that one should not believe
the Chaldean practice of predicting the fate of a person from the date of his
birth.
Though the possibility of an early spread of Babylonian astrology to the
West cannot be denied, caution seems to me to be necessary. It is by no means
certain that prediction "from the day of birth.. means astronomical prediction.
On the contrary, we have a very similar earlier reference by Herodotus, who
says~ (II, 82) that the Egyptians "assign each month and each day" to a god
and that "they can tell what fortune, what end, and what disposition a man
shall have aeeordiDg to the day of his birth". Here we see clearly that prediction
from "the day of birth.. means not at all prediction from the planetary positions
and from the position of the zodiac at the hour of birth. As Brugsch has seen
(Herodotus ed. Stein, 1881, p. 88 note) Herodotus is referring to a practice
which is directly attested from texts, e. g., the P. Sallier IV of the British Museum,
written in the New Kingdom (d. F. Chabas, Le calendrier des jours fastes et
nefasles de l'annee 6gyptienne [1870]; Oeuvres divenes, vol. IV p. 127-235;
also F. W. Read, Proe. Soc. Bibl. Arch. 38 [1916] p. 19-26, 60-69, and Abd
el-Mohsen Bakir, Annales du Service des Antiquites de l'~te 48 [1948]
p. 429). These lists of lucky and unlucky days contain also predictions as to
the future fate of a person bom on a certain day, e. g., death by a crocodile,
snake bite, etc., exactly as Herodotus indicates.
A still more rudimentary form of prediction of the future of a penon can
be found in Hittite texts of the 13th century B.C. (d. B. Meissner, Klio 19
(1925), p. 432-434), where the fate of a child is made dependent upon the
month of its birth. This is, of course, fundamentally different from the planetary
horoseopy of the Hellenistic age.
An important argument for the comparatively late introduction of astrology
seems to me the frequent use of Aries 8° as vernal point in astrological texts,
i. e. the vernal point of System B of the Babylonian lunar theory, which seems
to have reached the Greeks only at the time of Hipparehus. Eudoxus, however,
uses Aries 15° as vernal point and the same holds for several astronomical and
ealendaric papyri of the Ptolemaic period in Egypt. This earlier, Eudoxian
norm appears nowhere in astrologiealliterature.
There exists, moreover, direct evidence about the type of material that was
associated with the name of Eudoxus. Bezold and Boll, Reflexe astrologischer
Keilinsehriften bei grieehisehen Sehriftstellem (S.B. Heidelberger Akad. d.
Wiss. Philos.-hist. Xl. 1911 No.7) have shown close parallels between Mesopo-
tamian omens concerning thunder, clouds, the positions of the horns of the
moon, etc., and Greek calendars of the same type which are also related to
Eudoxus. But nowhere in these parallels does there occur any reference to
computational methods which are characteristic for horoscopic astrology.
The techniques and doctrines of Greek astrology are described in the classical
work of A. Bouche-Leclercq, L'astrologie grecque, Paris 1899. The historical
and philosophical aspects of one of the major components of astrological lore
and its branehiDg out into botany, alchemy, etc. is discussed by A. J. Festugike,
 Notes and References 189

La Revelation d'HermM Trismegiste, I, L'astrologie et les sciences occultes,

2nd ed., Paris 1950.
For the Hellenistic origin of astrology see W. Kroll, Klio 18 (1923) p. 213
and W. Capelle, Hermes 60 (1925) p, 373.
For the history of the planetary week see F. H. Colson, The Week, Cambridge
University Press 1926. Its penetration into Jewish literature has been investigated
by S. Gandz, Proc. Am. Acad. for Jewish Research 18 (1949) p. 213-254.
ad 70. For Teukros ef. Gundel in Pauly-Wissowa, Real Enc. 5 A col.
1132-1134. An upper limit for his lifetime would be given by Antiochus of
Athens, to whom Teukros is known (ef. Real Enc. 18,3 col. 122, 58 fl.). Cumont
(Annuaire de l'institut de philol. et d'hist. orientales 2, Bruxelles 1934, p. 135-
156) places Antiochus between 100 B.C. and 50 A.D., but his 8l1UDlents are of
a very indirect character. Thus we have at best these same limits for Teukros.
Teukros is called '"the Babylonian" by Porphyrius (about 270 A.D.) and by
subsequent authors1). Gundel. who realized how little is actually known about
Babylonian astrology, went to the other extreme by practically completely
denyiDg Babylonian intluenee on Hellenistic astrology and substituting Egyptian
mythology as the main agent. Consequently he adopted a rathe~ artificial
hypothesis of Eisler that uBabylon" refers to the fortress-town of this name in
Egypt (near Cairo). I see no evidence which supports this viewpoint.
The writings of Teukros are known only through excerpts preserved in later
astrological treatises. Their great historical importanee was first fully recognized
by Boll in his "Sphaera, neue griechische Texte und Untersuchungen zur
Geschichte der SternbUder" (Leipzig. Teubner, 1903). For the translation into
Pehlevi ef. Nalii no, Raceolta di Scritti VI p. 285-303. The Arabic tradition
was investigated by Steinschneider (Z. Deutsche Morgenlindische Gesell-
schaft 50,1896, p. 352-354). Cf. also Gundel, Real Ene. 5 A col. 1133.
For the dialogue uHermippus" eft Fr. Boll in S.B. Heidelberg Akad. d.
Wiss., phU. hist. Kl. 1912, Abh. 18 and V. Stegemann, Hermes 78 (1944) p.
120 note 2.
Babylonian parameters and methods found in the Pdca Siddhintika are
quoted in my paper "Babylonian Planetary Theory" Proc. of the Am. Phllos.
Soc. 98 (1954), p. 60-89.
ad 71. Datta and Singh in their "History of Hindu Mathematics", vol. I
(1935), quote (p. 59) the commentary of BhaUOtpala to the Brihat Samhita of
VarAha Mihira for an excerpt from the original Paulisa SiddhiDta in which a
huge number, ending in .. 7800, is expressed by number words in opposite
arrangement in the form. "zero, zero, eight, seven, ... .o'. It seems to me rather
plausible to explain the decimal place value notation as a modUlcation of the
sexagesimal place value notation with which the Hindus became familiar
through Hellenistic astronomy.
For the date of the Siirya-Siddhanta ef. van der Waerden, Diophantische

1) In case this name is based on ancient tradition one could not be sure that
Teukros was a citizen of Babylon, because "Babylonian" is also used for inhabitants
of Seleucia on the Tigris; cf. Tarn, The Greeks in Bacbia and India, p. 15.
190 Chapter VI

Gleichungen und Planetenperioden in der indisehen Astronomie, Vierteljahr-

sehrift der Naturforseh. Ges. in Zitrich, 100 (1955) p. 158-170.
For Ibn YQnus's reference to Persian observations d. Caussin, Notices et
Extraits de Manuserits de la Biblioth6que Nationale, tome 7, 1 and 12 [1808]
p. 2M note (1). For the date of Vettius Valens ef. O. Neugebauer, Harvard
Theological Review 4:7 (1954:) p. 65-67.
An excellent summary of Greek magic can be found in K. Preisendanz,
Zur "Oberliefel'UDlSBesehichte der spfltantiken Magie, Zentralblatt ffir BibUotheks-
wesen, Beiheft 75 (1951), p. 228-240.

In our discussions we have frequently used the word "Greek" with no further
qualiftcation. It may be useful to remark that we use this term only as a convenient
geographical or linguistic notation. A concept Uke "Greek mathematics", however,
seems to me more misleading than helpful. We are fairly well acquainted with
three mathematicians-Euclid, Archimedes, and Apollonius-who represent one
consistent tradition. We know only one astronomer, Ptolemy. And we are
familiar with about equally many minor figures who more or less follow their
great masters. Thus what is usually called "Greek" mathematics consists of the
fragments of writings of about 10 or 20 persons scattered over a period of 600
years. It seems to me a dangerous generalization to abstract from this material
a common type and then to establish some mysterious deeper principle which
supposedly CODDeets a mathematical document with some other work of art.
 191

APPENDIX I

The Ptolemaic System.

74. The following is a description of the cinematic models accord-
ing to which the tables of the Almagest were computed, allowing
us to find for any given moment the longitudes of the sun, the
moon and the five planets. This will give the reader at least some
concept of the basic methods developed in the Almagest, though I
cannot emphasize too strongly the fact that I am presenting here
only one special facet of the planetary theory of the Almagest.
Even within the restricted problem of determining the co-
ordinates of the celestial bodies, I have omitted the problem of
latitudes because an adequate discussion could not be given
without a detailed description of the empirical data which Ptolemy
had assembled and of the geometrical and numerical procedures
devised for their representation. Similarly I have almost com-
pletely ignored the very elegant methods which led from a great
variety of empirical data to the numerical determination of the
characteristic parameters of the cinematic models. Any serious
student of ancient or medieval astronomy must familiarize himself
with these details, not only in order properly to appreciate one of
the greatest masterpieces of scientific analysis ever written, but
also to be able to understand what was common knowledge for
every competent astronomer from the second to the seventeenth
century.
The planetary theory of the Almagest contains much more than
the determination of longitude and latitude. An interesting section
concerns the stationary points, another deals with first and last
visibilities. The lunar theory is also concerned with parallax,
distan~e and size of sun and moon, and the computation of
eclipses. Observational instruments are described, mathematical
tools devised. Solar apogee, equation of time, precession, fixed
192 Appendix I

star coordinates and their connection with lunar positions are

important questions which later became a center of interest for
Islamic astronomers. These processes can be adequately described
only in a work of a scale totally different from the present volume.
I hope however, that the following description will be useful as a
guide to the reading of the original sources.
75. The solar motion is represented in the Almagest by a simple
eccentric motion, or by the equivalent epicyclic motion which
depends only on one variable, the mean distance IX from the
apogee (cf. Fig. 31). The apogee is assumed to be fixed at a
distance of 65;30 0 from the vernal pointl ), the eccentricity is
e = r = 2;30 (for a deferent radius 60). The corresponding
correction 6, called "prosthaphairesis for anomaly" is negative
for values of the anomaly" = IX between 0 and 180 and positive
for the remaining semicircle; its maximum value is 2;23 0 by
which the "mean sun" at C can differ from the "true sun" at S.
The corrections ~ as function of the anomaly are given in a table
in Almagest 111,6 which is represented graphically in Fig. 33
(lowest curveII)). This correction was later called the .,equation
of center" .8)

Fig. 31.

76. The lunar theory obviously requires a model of greater

complication since it was known already to the Babylonian
astronomers of the third or fourth century B.C. that the anomalistic
month (return to the same velocity) was longer than the sidereal
1) In other words, it is assumed that tropical and anomaJistic year coincide.
S) Representing the absolute value of tJ.
8) To my knowledge, the term "equatio" appears flrst in Latin translations of
Arabie treatises.
 The Ptolemaic System 193

Fig. 32.

month (return to the same fixed star). Consequently in an epicyclic

model (cf. Fig. 32) the mean motion in longitude, expressed by
the angle A, is greater than the mean motion in anomaly, the first
being about 13;10,35°/d the latter 13;3,54°/d • The resulting motion
can also be described as an eccentric motion with a rotating
apsidal line. Measured in the units for which the radius of the
deferent is 60, the radius of the epicycle was found to be r = 5;15.
The resulting corrections ~ = 1 - 1 which leads from the "mean
longitude" 1 to the "true longitude" 1 is tabulated in Almagest
IV,10 and graphically represented as Ct in Fig. 33 in the same
scale as the equation of center of the sun.
77. The theory described so far was known to Hipparchus
though refined in several respects by Ptolemy. The determination
of the essential parameters was based on carefully selected
observations of lunar eclipses and the results obtained were very
satisfactory for the description of eclipses in general. Ptolemy,
however, realized from a masterful analysis of observations of
his own and of his predecessors that marked deviations from the
predicted longitudes of the moon, reaching a maximal amount
in the neighborhood of elongations of ± 90° from the sun, could
occur. In other words he realized that the traditional lunar theory
agreed with observations in the syzygies (conjunctions and
194 Appendix I

oppositions) but could not explain longitudes near the quadratures,

particularly for values of the anomaly" near ± 90°. In these
cases the diameter of the epicycle seemed to be enlarged over the
value found at the syzygies.
The procedure which Ptolemy followed to cope with this
situation is of interest in many respects. It provides us with a

Fig. 34.

good insight into the mathematical and astronomical methodology

of the time; the attitude toward a glaring defect of the theory is
very revealing and has repercussions in Islamic astronomy and
in the work of Copernicus; finally the method by which this
inequality of the lunar motion was accounted for influenced also
the planetary theory, both of Ptolemy and Copernicus.
As we have said above the observations suggested a dependence
of the apparent diameter of the epicycle of the moon on the
elongation from the sun. Such an effect could be produced by
bringing the epicycle closer to the observer by the following
mechanism (Fig. 34). Let Co be the position of the center of the
lunar epicycle at conjunction with the mean sun such that OFoCo
 The Ptolemaic System 195

are on a straight line and OCo = R the radius of the deferent

known from the previous model given in Fig. 32. Let 1/ be an
angle increasing proportionally with time at a rate equal to the
difference between the mean velocity of the moon and the mean
velocity of the sun; 1/ is therefore called the "elongation". Its
value is zero at conjunction; as 1/ increases, the point F is made
to move in retrograde direction on a small circle of radius sand
with center 0 such that its angular distance from the direction
from 0 to the mean sun equals the elongation 1/. At the same time
the point C, the center of the epicycle, moves forward such that
the direction OC makes an angle 1/ with the direction from 0 to
the mean sun. In this way C approaches 0 as 1/ increases from 0 °
toward 90°. At quadrature (1/ = 90) the distance C from 0 reaches
its minimum value R - 28. At opposition (1/ = 180) C is again
removed to the original distance R from O. Since eclipses can only
occur at conjunctions or oppositions, the new model agrees with
the old one for all the elements obtained from eclipses. Toward
the quadratures, however, it increases the apparent diameter of
the epicycle in accordance with the observations.
Ptolemy found one more deviation from the original theory for
positions of the epicycle at elongations nearer to the octants:
instead of counting the anomaly'Y which determines the distance
of the moon M on the epicycle from the apogee D, it had to be
measured from a variable apogee H (Fig. 35) such that the
radius HC has a "direction" (Greek: ~e6(111evt1,,~) toward a point
N which is always located diametrically opposite to the point F.
The point H is called the "mean apogee" because from it is
measured the mean anomaly 'Y. The point T, which is the apogee
of the epicycle as seen from 0, is then called the "true" apogee.
The parameters of this model are
r = 5;15 8 = 10;19
which show that the moon at quadrature could come as near as
R - 28 - r = 34;7 to the observer. This obviously means that
the apparent diameter of the moon itself should reach almost
twice its mean value which is very definitely not the case. This
discrepancy is silently ignored by Ptolemy, though he could not
1) In the "eanobie Inscription" (Opera II p. 150. 5 and 12), Ptolemy adopts
the norm R - . == 60 and consequently obtains r == 6;20 • == 12;28.
196 Appendix I

o
I
I
I
I
I
I
I
I
I

~---., ..
,, ,
, ,
,,, ' ,
\
. \
\ ",
\
\ ,,
"
\
\ '
\
\
\
\
'\ '
.. ,
\

\ ,,
\
,
\
\

Fig. 35.

have doubted that the actual geocentric distances of the moon

were very different from what his model required. Nevertheless
this model was retained by almost all his followers simply because
it proved to predict at least the longitudes correctly.
Copernicus pointed to the obvious discrepancy between
Ptolemy's lunar model and the observable parallaxes1) and
proposed another model which would keep the eenter C of the
epicycle at mean distance but would nevertheless increase the
moon's distance from C at quadratures (Fig. 36). He assumed
that the moon M was located on a secondary epicycle such that
it started its motion at conjunction at E at a distance r - 8 from C.
With increasing elongation 11. the moon would move on the small
epicycle in the direct sense by 2 '1/. thus reaching for '1/ = 90 a
distance r + s from C. Since Copernicus used r = 6;35 and
s = 1;25 for R = 60 the moon could not come closer to the
observer than R - (r +
s) = 52 which is no longer essentially
different from the distances resulting from the simple model
(Fig. 32).
1) De revolutionlbus (published 1543) IV. 2.
 The Ptolemaic System 197

o
Fig. 86.

This obvious advantage of the use of secondary epicycles

induced Copernicus to apply the same construction also to the
planetary motion and thus to initiate complications which de-
stroyed the inherent elegance and simplicity of the Ptolemaic
model.
Only recently has it been discovered1) that the same method
for the correction of Ptolemy's lunar model was used about 200
years before Copernicus by ibn ash-Shitir. Whether Copernicus
knew about his predecessor or not is impossible to decide at the
I
presentJ moment.
78. 'rhe pro~edure for computing the longitude of the moon
I

according to the Ptolemaic model is based on the tables in Al-

magest V,8. The first two columns contain the argument 8 and
360 - 8 for 8 between 0° and 180°. The 3rd column gives the
angular difference between true and mean apogee of the epicycle
(T and H in Fig. 35) as function of the double elongation 21J,
where 1J is the difference between mean lunar and mean solar
longitude at the given moment. We call this function ca(21J); its
graph is given in Fig. 37. By forming y' = y +
Cs one has found
the ariomaly which determines the angle at which M is seen
from O.I
With y' as argument one finds in the 4th column the
value ~,(y') which is the equation of center already known from
1) cr. V. Roberts, in a forthcoming article In Isis.
198 Appendix I

the simple lunar model (Fig. 32 and 33), that is, the angle under
which CM would be seen from 0 if the distance of C were R = 60.
The column c.(,,') gives the amount by which the equation of
center increase if C were at minimum distance from 0; finally
c.(2,,) indicates the fraction of c. due to the fact that for 2" < 180 0
the actual distance of C from 0 lies between the two extrema for
which the equation of center is c, and c, + c. respectively. Thus
one finds for the final equation the value

" = c, + c• . c.
180 or 1 = 1 + "
and hence for the true longitude A = A- " if y' <
if ,,'> 180. The numerical values given in Almagest V,8 are
graphically represented in Figs. 38 and 37.

1£.

Fig. 37.

79. Ptolemy's planetary theory follows the same principles as

the solar and lunar theory. As we have seen (p. 123 f.), circular
planetary orbits with the sun as center are epicyclic motions with
respect to the terrestrial observer. It is easy to see by the same
method that eccentric circular orbits lead to epicyclic motions
with eccentric deferents, the eccentricity being the resultant of the
vectors which represent the planetary and solar (or terrestrial)
eccentricity.
This model was further modified by Ptolemy on the basis of
observations mostly made by himself or his predecessor Theon;1)
he found that the center of the epicycle appears to move with its
mean angular velocity not with respect to the center of the deferent
but with respect to a point (later called "equant") located sym-
metrically to the observer. We have encountered essentially the
same idea in Ptolemy's theory of the moon where the center C
1) Perhaps Ptolemy's teacher, not to be confused with Theon of Aleundria.
the author of the "Handy Tables" who lived two centuries after PtolemY.
 The Ptolemaic System 199

of the epicycle is moving with constant angular velocity only in

so far as it is kept on the radius 0 T which moves uniformly
(Fig. 35) while the linear velocity of C is not constant. Philosophical
minds considered this departure from strictly uniform circular
motion the most serious objection against the Ptolemaic system
and invented extremely complicated combinations of circular
motions in order to rescue the axiom of the primeval simplicity
of a spherical universe.

p"'" .... . ,
".... .... y'. ,
\

E
.. '... \,
\
e
.. ...... \, G
......'\ e
. ~

o
Fig. 38.

80. Fig. 38 represents Ptolemy·s model for the outer planets.

The mean motion of the planet is represented by the motion of
the center C of the epicycle, measured by the "mean distance"
IX of C from the apogee A. The planet P moves on the epicycle
with a speed corresponding to the synodic period and measured
by the anomaly" ,... The sense of rotation of P on the epicycle is
U

now equal to the sense of mean motion, thus giving the planet
its greatest direct motion near the apogee of the epicycle and
producing retrogradation near the perigee.
In ~greement with our general analysis of heliocentric and
geocentric motion (p. 124 Fig. 14) the direction CP is parallel
to the direction from 0 to the mean sun in the ease of an outer
200 Appendix I

planet. In the case of an inner planet the angle lX increases as the

longitude of the mean sun (in fact C may he identified with the
mean sun) and the anomaly 'Y varies independent of the position
of the sun. In all cases the position of P depends on the two
independent variables lX and 'Y which can be considered to he
known for any given moment t.
For Mercury the observations could not be reconciled with so
simple a model as for the other planets-UNulle planete n'a
demande aux astronomes plus de soins et de peines que Mercure,
et ne leur a donne en recompense tant d'inquietudes, tant de
contrarietes" says Leverrier. Ptolemy's data led again to the
necessity of increasing the apparent size of the epicycle as in the
case of the moon, the only difference being that the closest ap-
proach now occurred at about ± 120 0 from the apsidal line.
Thus Ptolemy adopted a model as described in Fig. 39. The
center G of the deferent moves retrograde on a circle of radius e
and center B where e is not only the eccentricity BE but also the
distance of the observer 0 from the equant E. The center C of
the epicycle moves forward such that its progress, seen from E,
appears uniform and of the same amouut as G is removed in the

Fig. 89.
 The Ptolemaic System 201

opposite direction from the apsidal line OA. For fX = ± 120 the
radius GC of the deferent passes through E and thus brings C
nearer to 0 than for IX = 180. In other words the orbit of C with
respect to 0 has one apogee in the apsidalline but two perigees
symmetric to it at IX = ± 120.
81. The practical computation of planetary positions follows
much the same lines as for the moon. Almagest XI,11 gives
tables in 8 columns for each planet. Columns 1 and 2 contain
the common arguments 8 and 360 - 8. Columns 3 and 4 are
only used in the combination,
c' a (IX) = Ca (IX) + C.(fX)
for identical values of the mean distance IX in both columns. In
later works-Theon's tables and Islamic tables-these two
columns are always combined into one (c' a in Fig. 38), leading
from the mean apogee H to the true apogee T. Ptolemy kept the
two columns separate for didactic reasons because he wanted the
reader to see how c' a had been obtained from first locating the
center of the deferent at E and then moving it to G.
With the argument y' = " +
c' a three tables are computed:
c.(,,') gives the angle «5(,,') under which the radius r = CP of
the epicycle appears from 0 under the assumption that C is
located at mean distance from O. The column CIi("') gives the
amount of the correction which must be subtracted from c. such
that C. - Cli represents 6(,,') for the case of C at maximum distance
from O. Similarly C. + c7 gives «5(,,') for minimum distance of
C from O. Finally CS(IX) gives the coefficient by which Cli or C7.
respectively, have to be multiplied in order to give the correction
of Cli at a distance IX of C from the apogee.
Now we can formulate the whole procedure for the computation
of the true longitude A of a planet. For the given moment t the
following elements are known:
10 longitude of the apogee A
fX mean distance of C from A
" mean anomaly of the planet.
Then the true longitude A = l( t) is given by
A = Ao+ IX - C' a(fX) +6
202 Appendfx I

Fig.40a.

where c'.(tx) is found from

c'a(tx) = ca(tx) + cJ.tx)
and ~ from
~= c.(y') + cs(tx) •{ C6("') for
c.,(y') for
with
,,' = " + c'.(tx).
Fig. 40 represents these functions in the case of Mercury. The
corresponding curves for the other planets are somewhat simpler
since they have only one perigee of the center of the epicycle.
The technique of computation, however, is in all cases the same.
8!. It is illuminating to compare Ptolemy's model of the motion
of Mercury with the Copernican theory. The empirical data are,
of course, essentially the same, particularly the fact that the
smallest values for the greatest elongation of Mercury from the

Fig.40b.
 The Ptolemaic System 203

,
,,
"
Fig. 41a.

sun occur in Libra whereas the greatest values are not observed
opposite to Libra in Aries but two signs before or after, in Aquarius
and Gemini. In order to account for this observation, Mercury is
made to move on a straight line segment such that its distance
from the center of its orbit varies with the proper period. A
movement on a straight line seems not quite in conformity with
the postulate of circular motions of the celestial bodies but
fortunately Copernicus had at his disposal a device of at-TOfi,
•
who had shown that a point of the circumference of a circle of
radius i moves along the diameter of a circle of radius s inside
of which the first circle rolls (cf. Fig. 41 b).
204 Appendfx I

,"
, .. ---- ....... ,
"
,,
" ,,
\
I \ ,
P, \
, :Q
'•,, ,l '\~~~~~:.....-
,
\,
,,,
, \
\
" . . . . ."
~ _~'20.

Fig. 41b.

Fig. 41 a describes the Copernican modification of Ptolemy's

theory of Mercury. The observer 0 now moves on a circle around
the mean sun S, the physical sun S being at a distance e from S
corresponding to the eccentricity of the solar orbit1). At a distance
el from S is located the center E of a small circle of radius e.
on which rotates the point C with twice the angular velocity with
which the distance ()t, of 0 from the apsidal line SA increases. The
point C is the center of the instantaneous orbital circle of Mercury.
Its radius r = CM is made variable between the limits P and Q
by means of at-TUfrs device (Fig. 41h) such that M moves from
Q to P and back again while ()t, increases from 0° to 180°. The
anomaly" is counted from a direction CH which is parallel to SO.
811. It is, of course, of no interest whether we say that, in the
model of Fig. 41 a, 0 rotates about S or S about 0, in which case
R is called the radius of the deferent and r the radius of the
epicycle. Thus it is evident that cinematically the two models are
hardly different except for Copernicus's insistence on using circles
for every partial motion where Ptolemy had already reached
much greater freedom of approach.
The popular belief that Copernicus's heliocentric s;)·stem
constitutes a significant simplification of the Ptolemaic system is
obviously wrong. The choice of the reference system has no effect
whatever on the structure of the model, and the Copernican
models themselves require about twice as many circles as the
Ptolemaic models and are far less elegant and adaptable. In fact
1) Copernicus is not very outspoken abouCthe question whether S rotates about
S or vice versa. For his planetary theory only S is of real signifteance.
 The Ptolemaic System 205
the importance of Copernicus's work lies in a totally different
direction than generally announced. One may enumerate his
main contributions as follows:
(a) The return to a strictly Ptolemaic methodology which made
all steps from the empirical data to the parameters of the model
perfectly clear and opened the way to a refinement of the l>asic
observations which eventually led to the proper generalization of
the Ptolemaic methods. discarding the iterated epicycles of
Copernicus.
(b) The insight that we can obtain information about the
actual planetary distances if we assume that all planetary orbits
have essentially the same center. namely. the sun. Then the radii
of the epicycles of the inner planets directly give their distances
from the sun in terms of Ri for the outer planets the reciprocals
of the radii of the epicycles measure the heliocentric distances.
Again the question which body is supposedly at rest is of no
interest whatever and therefore Tycho Brahe's system was
exactly as good as the Copernican system. and equally complicated
because of the same doctrine of circularity of admissible motions.
(c) The assumption of a common center of the planetary orbits
suggested also the proper solution of the problem of latitudes.
namely. that the inclined planes of the planetary orbits pass
through that common center. Unfortunately the postulate of
circularity induced Copernicus to use the mean sun and not the
real sun as common center and thus resulted in a theory of
latitudes which labored under exactly the same complications as
Ptolemy's theory. Nevertheless. this modification of the ancient
theory of latitudes helped Kepler to find the real solution. which
then permitted the computation of heliocentric coordinates in a
uniform fashion and the finding of the geocentric coordinates
through an independent procedure.
The enormous increase of empirical data accumulated by
Tycho Brahe and his collaborators finally convinced Kepler that
even the return to the Ptolemaic model with equant could not
properly represent the observations and thus led him to abandon
the axiom of circular orbits and to the discovery of the proper
orbits.
There is no better way to convince oneself of the inner coherence
of ancient and mediaeval astronomy than to place side by side
206 Appendix I

the Almagest, al-Battini's Opus asfronomicum and Copernicus's

De revolutionibus. Chapter by chapter, theorem by theorem, table
by table, these works run parallel. With Tycho Brahe and Kepler
the spell of tradition was broken. The very style in which these
men write is totally different from the classical prototype. Never
has a more significant title been given to an astronomical work
than to Kepler's book on Mars: "Astronomia Nova".

BmLIOGRAPBY TO APPENDIX I
For the history of Greek and medieval astronomy three works must be
consulted: Delambre, Histoire de l'astronomie ancienne (2 voIs., Paris 1817)
and his Histoire de l'astronomie du moyen age (Paris 1819). Norbert Herz,
Gesehichte der Bahnbestimmung von Planeten und Kometen (2 parts, Leipzig,
Teubner, 1887 and 1894). J. L. E. Dreyer, History of the Planetary Systems
from Thales to Kepler (Cambridge 1906, reprinted New York 1958 under the
tide A History of Astronomy").
II

See furthermore Norbert H erz in the Handwtirterbuch der Astronomie

vol. I p. 1-98 (= Eneyklopaedie d. Naturwiss. In, 2 ed. W. Valentiner, Breslau
1897) and the article on Ptolemaios in Pauly-Wissowa, Realeneyclopldie
der classischen Altertumswissenschaft vol. 28, 2 cols. 1788-1859 by B. L. van
der Waerden; also Paul Tannery, Recherches sur l'histoire de l'astronomie
aneienne, Paris 1898. A good summary of the Ptolemaic planetary theory is
given in Derek J. Price, The Equatorie of the Planetis, Cambridge Univ. Press
1955 p. 98-117.

NOTES AND REFERENCES TO APPENDIX I

ad 77, simultaneously a contribution to the medievalism in modern scholar-
ship. Pierre Duhem in vol. I of his Systeme du Monde (1914) p. 494 f. has given
a description of Ptolemy's lunar theory according to which the moon would
become retrograde each month sinee he gave it the wrong sense of rotation
(1. c. Fig. 11). It is also Duhem's opinion that the retrograde motion of the
diameter FN (using my figure 85) represents u6videmment" the retrograde
motion of the nodal line (thus producing every month a total lunar and solar
eclipsel). This flagrant nonsense has now been repeated for some 40 years by
the excerptors of Duhem's work.
Duhem's total Ignorance of Ptolemy's lunar theory is a good example of the
rapid decline of the history of science. The details of the ancient theory of the
moon has been particularly well known in France since it was the object of
discussion which occupied the French academy from 1836 to 1871 and in
which Sedillot, Biot, Arago, Damoiseau, Libri, Munk, Reinaud, de Slane,
 Notes and References 207

Chasles, Leverrier, Bertrand and others participated.1) Sedillot thought he had

discovered in a passage of Abd1-Wefa (who died 998) a description of the lunar
inequality, now called "variation" and published by Tycho Brahe in 1602
(Opera II p. 100 f.) whereas Ptolemy's theory covered only the first and second
inequality (called uevection"). It is no longer possible to doubt that Sedillot
was in error and Abu'l-Wefa gave only a description of Ptolemy's construction
without any addition of his own. It is, however, of interest to remark that
P. Tannery has shown (Astronomie ancienne p. 211 tI.) that the combined
etIeets of Ptolemy's theory do not exactly correspond to the "evection" of the
modern theory of the lunar motion but also include about half of the term which
corresponds to the "variation".
ad 80. The parameters for the planetary orbits are for R = 60
for Saturn r = 6;30 e = 3;25
Jupiter 11;30 2;45
Mars 39;30 6;0
Venus 43;10 1;15
Mercury 22;30 3;0
None of my figures for the planetary models is drawn to scale since most details
would become unrecognizable in the available spaee.
In contrast to the solar theory, in which the apogee is of constant tropical
longitude, the apogees of the planetary orbits are assumed fixed with respect
to the fixed stars and thus participate in their increase in longitude by 10 per
century.
ad 82. Ni$ir ad-Din at-TOfi (born 1201, died 1274) was the director of the
famous observatory at Marlgha. He objected on philosophical grounds against
the crank-mechanism in Ptolemy's theory of the moon and of Mercury. The
relevant chapter from his "Memento on Astronomy" was translated by Carra
de Vaux in P. Tannery's Recherches sur l'histoire de l'astronomie ancienne,
Paris 1893, p. 348-359. I do not know through what medium Copernicus knew
about TUft's construction.
The motion of M on the diameter PQ (Fig. 41 b) is, of course, nothing but a
simple harmonic motion QM = s(1 - cos 2 IX).
The parameters, as determined by Copernicus (de revol. V, 27) are
R = 10.000 or, if R = 60:
e = 322 e = 1;56
el = 736 el = 4;25
e. = 212 ell = 1;16
3573 ~ r ~ 3953 21;26 ~ r ~ 23;4:3
2s = 380 2s = 2;16
Longitudes are counted by Copernicus sidereally, taking" Arietis as zero point.
This corresponds to a widespread mediaeval custom counting longitudes from
Regulus (IX Leonis), a norm which goes back at least to Ptolemy·s Canobic
Inscription (opera II p. 152. 2 f. to be corrected by p. 80. 27).
1) cr. the excellent summary by Carra de Vaox, J. Asfatique 8 s~r., 19 (1892)
p.440-471.
208

APPENDIX II

On Greek Mathematics
84.. I do not consider it as the goal of historical writing to
condense the complexity of historical processes into some kind of
"digest" or "synthesis". On the contrary, I see the main purpose
of historical studies in the unfolding of the stupendous wealth of
phenomena which are connected with any phase of human
history and thus to counteract the natural tendency toward over-
simplification and philosophical constructions which are the
faithful companions of ignorance.
To a modern mathematician who wants to get some insight
into the mechanism of Greek mathematics, the access is made
easy. The major works of Archimedes, Apollonius, Euclid, etc.,
are well edited and competently translated. The careful reading
of a treatise by Archimedes or a book of Apollonius takes time
and effort-as does the pursuit of all worthwhile knowledge-but
one is repaid with a much deeper understanding of ancient
mathematical methods than the reading of all summaries could
provide. After a solid basis has been established, works like
T. L. Heath's will provide a competent guide to related material
and give the general background.
The astronomical material is much less directly accessible and
consequently certain mathematical methods which were devel-
oped in close relationship to astronomy are much less generally
known than, e. g., Euclid's procedures. The conventional picture
of Greek mathematics-a very sophisticated branch of geometry
followed by some not quite successful attempts at algebra and
number theory during the later periods of decline-is wrong for
two reasons. First, as we have seen on many occasions in the
preceding chapters, we are not dealing with a decline from scien-
tific geometry to less exact methods of algebraic tendency, but
with two contemporary phenomena: a comparatively rapid
 On Greek Mathematics 209

development of rigid mathematical reasoning on the highest level

in contrast to a much older and little changing background of
ancient oriental and Hellenistic mathematics with a strongly
algebraic character inherited from its Babylonian origin. Secondly,
even the strictly Greek development is not adequately described
by the Euclidean-Archimedean development, which is most
familiar to the modern reader, but we must add many methods
which concern numerical and graphical problems which originated
in mathematical astronomy. In fact we see here the first instance
of the stimulus which astronomy has repeatedly given to mathem-
atics. It is the purpose of the following to illustrate some of these
less known aspects of Greek mathematics.
85. The fact that Greek plane trigonometry was based on the
tabulation of chords instead of sines produces, of course, no
essential differences between modern and ancient procedure. The
norm R = 60 introduces coefficients 1/120 but since all fractions
are written sexagesimally this only implies division by 2,0 or
numerically the same as halving. Thus we have (Fig. 42)

c
a = -2,0 crd2(X a = c sin (X

c
b = -2,0 crd(180 - 2(X) b = c cos (X

a crd 2(X a
-=
b
------~
crd (180 - 2(X)
-b = tan (X

Since the tables extend from 0 to 180 the fIrst two cases are just
as easy to handle as in our system. The only real inconvience
lies in the lack of tables for the ratios corresponding to tan (X.
As an example for the use of trigonometry in an astronomical
problem, I shall describe the method followed in Almagest IV, 6
210 Appendix II

for finding the length of the radius r of the lunar epicycle and the
position of the apogee. Weare dealing here with the simple
model of the lunar motion (as in Fig. 32 p. 193) in which the
center C of the epicycle moves on a fixed deferent of radius
R = 60. This problem is of interest for several reasons. It re-
presents a method which was certainly used and probably
invented by Hipparchus. It is the simplest case of a much more
general problem, namely, to determine the parameters of an orbit
from a set of observed positions. It has, finally, close relations to
an important problem in geodesy, namely, to find the position of
an observer with respect to three given points.
In the case of the moon the essential steps are as follows.
Observation of the motion of the moon with respect to the stars
from day to day easily reveals that its velocity is not constant.
About once every month this progress is at a minimum of about
12 0 and then again at a maximum of about 14 0 per day. Conse-
quently it is easy to count the number of periods of the lunar
velocity corresponding to a given number of lunar months.
Within a few years of observations the mean length of this
"anomalistic" period can be determined with sufficient accuracy.
If we then decide to describe this variation of velocity by means
of an epicyclic model, as in Fig. 32 (p. 193), we can consider as
known the rate of change of the "anomaly" 'Y as well as the rate of
change of the mean longitude A of the center of the epicycle. The
problem now faced consists in the determination of r and of the
moment at which the moon is exactly at its minimum speed, i. e.,
in the apogee of the epicycle.
To this end the moments and longitudes of 3 lunar positions
are determined by means of 3 lunar eclipses the longitudes of
which are accurately known through the diametrically opposite
solar positions. The reason for this procedure is very character-
istic: one reaches greater accuracy by means of a computed solar
position than from a direct comparison of the position of the
moon with respect to stars, since the coordinates of stars would
involve the measurement of angles by means of instruments the
accuracy of which is difficult to control.
The three moments for the midpoints of the eclipses and the
three corresponding longitudes furnish us with two sets of differ-
ences Llt 1, Lit., and LlA l , L1A s (cf. Fig. 43 for the first pair of posi-
 On Greek Mathematics 211

Fig. 43.

tions). Our knowledge of the mean values for the change of "
and A furnishes us. in combination with Llt 1 and Lit•• with the
values LI" l' LI" I and Lll 1• Llli by which y and l have changed
between consecutive eclipses. Fig. 43 illustrates the situation for
the first two eclipses. From the data which we have so far assem-
bled. we know that the observer at 0 sees the segment MIMI of
the epicycle under the angle Lll 1 - Lll 1 whereas this same segment
is subtended at the center C of the epicycle by the angle LlYI. A
similar situation holds for the second pair of eclipses and this
provides us with the final formulation of the problem (Fig. 44):
Three points MI' MI' M 8 on a circle of radius r subtend at its
center given angles jXl' tXl and are seen from an observer in 0
under given angles ()1' ()I. Find r and the position of 0 with
respect to MI' MI. M 8.
This problem is solved as follows (Fig. 44): consider the

Fig. 44.
212 Appendix II

point B on the straight line OMI and call 8 the distance OB. We
then have for the perpendicular B0 1 :

and also, because

and thus:
crd 2'1
M.B = s d p.
cr 2 1
• •
Similarly, since
R
1'1
R
+ 1'. = (¥1 +2 tX. - «(II + ui)
.I ..I

MaB = , crd 2('1 + 'i).

crd 2(PI + P.>
From the triangle M.MaB we can now express the chord '. in
terms of 8 and the given angles. Using the altitude MaP we have

MaP - MaB d _ 8 crd 2( 1 a + '.) d

- 120 cr tX. - 120 • crd 2(Pl + P.) • cr (¥.

and
MaB • crd2(~ + a.)
BP = - crd (180-(¥.) = - . . crd(t80 - t:¥.).
120 120 crd 2(PI fl.) +
Consequently
M.P = M.B - BP = 8 • ( •••• )
is known in terms of 8 and the given angles and therefore also

8. = VMsP· + MaPI = 8 .( ••••• )

On the other hand '. is a chord of the circle of center C, thus

r
'. = 60 crd (¥.
where t:¥. is given. Thus
60
r= ·VM~+Mapl=,·(••••. )
ad (¥.
is known if , is known.
 On Greek Mathematics 213

Fig. 45.

Obviously the absolute dimensions of Fig. 44 are not deter-

mined by the angles alone. Thus we must arbitrarily fix some
distance and this is done by defining the radius OC of the deferent
as R = 60. All other distances will then be measured in these
units.
In order to find s we have to determine the remaining angles
around C (cf. Fig. 45). Now
r
MaB = - crd lX a
60
where both MaP and r are expressed by trigonometric functions
multiplied by s. Thus s can be cancelled in the above relation
and lXa is known independently of s. The same holds for
IX, = 360 - (IX I + IXI + lXa).
Now s can be determined as follows

r
BMI = -crdlX,
60
and OMI = s + BM = S . ( •••• ).
I

Using a classical theorem which holds for a circle and a point

outside, we have
(R +r) . (R - r) = OMI • S
or
RI = r l +OMI • s = 8" • ( •••• ).

Thus s is known in units of R = 60 and therefore also r.

Finally (Fig. 45)
R 1 1
OD = -
120
crd 2fJ = -crd 21/ = 8
2
-BMI
2
+
214 Appendixll

and thus fJ is known. Thus the moon M1 at the moment of the

fIrst eclipse was on the epicycle ~' + fJ distant from the line OR.
This completes the solution of our problem.
86. We have very little knowledge about the early history of
spherical trigonometry. Beginning with quite primitive treatises
on spherical astronomy by Euclid and Autolycus (4th century
B.C.) we have several works which precede Menelaos (about
100 A.D.) in which one can recognize attempts to solve in general
problems of astronomical importance, e. g., the determination
of the rising times of given arcs of the ecliptic (cf. p. 160). We
do not, however, know how numerical problems of this type were
solved in practice; one might, e. g., assume that Hipparchus used
methods which are known to us from Hindu astronomy and, in
certain traces, even from the Almagest. These methods are
characterized by the use of the interior of the sphere for the
determination of the length of circular arcs on the sphere. For
example the Siirya Siddhinta (II, 60/63) determines the length of
daylight as follows. Assume to be given the length So of the equi-
noctial noon shadow of a gnomon of length 12 and the declina-
tion " of the sun for the day in question. Then we have (Fig. 46)
12 OB Sin 6
-=-=--
e e
where we define Sin 6 = R sin 6 according to the Hindu usage
of trigonometric functions. Furthermore, r, the radius of the par-
allel circle of declination 6, the so-called"day radius" is given by
r =R- Sin vers 6
where the versed sine corresponds to our 1 - cos 6. Finally
• e R Sin 6
Sin fX = R - = · So
r 12(R - Sin vers 6)
where fX is called the "right ascensional dift'erence".
Using the value of fX thus obtained the length of daylight is
given by 180 + 2fX degrees.
87. The above approach to the solution of problems of spherical
geometry by means of plane trigonometry applied to properly
 On Greek Mathematics 215

o
A
Cjl-------9---+------9
e B r c
··•
·•·..,
,
\
.,,
.
"

Fig. 46.

chosen planes is systematically expanded in the theory of the

"Analemma", a method which we could classify under descriptive
geometry. It existed already before Ptolemy since he criticizes the
unsystematic definitions of his predecessors. What follows is
taken from Ptolemy's very elegant treatment of the subject.
Related procedures are known from Vitruvius, the Roman
architect under Augustus, and from Heron, who wrote about 70
years before Ptolemy.
The problem itself concerns the theory of sun-dials in the
simplest form of a vertical "gnomon". Mathematically the problem
consists in defining proper spherical coordinates for the position
of the sun at a given moment for a given geographical latitude and
then to find graphically in a plane the arcs which represent these
coordinates. Ptolemy's predecessors-he calls them the "ancients"
without telling us who they were-operated with the following
system: consider, e. g., the octant of the celestial sphere which is
bounded by the planes of the horizon, of the meridian, and of
the vertical which is perpendicular to the two first mentioned
planes (cf. Fig. 47). The center of the sphere is the observer, the
vertical axis the gnomon. Let I be the position of the sun as seen
at the given moment. Then two planes are passed through I:
a vertical plane which contains the gnomon, and an inclined
plane which contains the axis OS. In this way three angles are
216 Appendix II

E
s
Fig. 47.

defined: (Xl and (XI in the coordinate planes and Pin the vertical
plane through X. Each pair (XltXl or tXIP or (XIP can be used to
define the position of X. To this "the ancients" added one more
angle. For given geographical latitude qJ the position of the equator
plane is given; its intersection with the plane SX defines a new
angle 'Y and it is clear that also the pairs tX I'Y and P'Y can be used
to define the position of X since for given 'Y also the arc tX l is
fixed and vice versa.
Ptolemy would not tolerate such inelegant definitions. Using
the same three orthogonal axes he would pass three planes
through X and one of the axes, respectively. Then six angles
«I" ..• Ps are defined, as indicated in Fig. 48, two of which
always suffice to define the position of X. The arrangement is
strictly cyclical. all angles are acute and counted from one of the
three orthogonal axes.
z

Fig. 48.
 On Greek Mathematics 217

,D

E
·•:
I
I

,·
S o-----=--OO---..:...L.-B~':----o' I'

,,
t "
" '~"
• .'

,, I
.C
I

,,
I

Fig. 49.

Ptolemy now proceeds to construct these six angles. As an

illustration of the method followed, it will suffice to take the case
of PI (called Uhectemoros"). As plane of construction we use the
plane SOZ of the meridian (Fig. 49). For given fJ we know
the angle fJ = 90 - fJ under which the equator is inclined to the
horizon. For a given moment we also know the solar longitude
and thus its declination; we also know for a given seasonal
hour (here assumed to be an hour before noon) the fraction travel-
led by the sun between sunrise and noon. Thus we construct the
trace ABC of the plane of the parallel-circle travelled at the given
day and swing this plane into the plane of the meridian. If DB
is perpendicular to ABC, then D represents the point of sunrise,
A the point of culmination for the given day!). Thus we can
find the position X of the sun for the given hour on the arc DA.
Construct XF perpendicular to AB and make FG = FX. If E is
placed such that EO is perPendicular to OF, then EG is the arc PI
which we wanted to find. Indeed, E is the east-point of the horizon
swung into our plane of construction and OF is the trace of
the plane of the hectemoros in the meridian. Thus G is the
place of the sun in the plane of the hectemoros which appears
rotated about OF. Thus EG = EX.
Similar procedures lead to the determination of the other angles.
It would lead us too far to present the details here but it is of
great principal interest to mention the mechanization of these

1) Obviously DA/DC gives the ratio of daylight to night for the date In question.
218 Appendix II

constructions such that they can rapidly be carried out for any
geographical latitude fJ and for arbitrary solar longitude 1. All
eases have some elements in common which do not depend on
fJ and A, e. g. in Fig. 49 the circle of the meridian and the parallels
to the equator. These circles are to be engraved on a plate of
metal or stone or-in a cheaper model painted in black or red
on a wooden disk. About the center is drawn the meridian and
concentric with it a circle which indicates the angles fJ correspond-
ing to the seven "climates" which have longest daylight of lSb
lS1b ••• 16b respectively. On the proper diameter are indicated
points which correspond to the hourly position of the sun at
equinox. The plate is then covered with wax so that additional
lines which depend on the special values of A and fJ can be easily
drawn. The disc can rotate about its center and a straight edge
with right angle permits one to connect points of the different
graduated circles corresponding to the swing of the projections
about the proper axes. In this way it is possible to determine the
angles in question by a procedure which is now called nomo-
graphical. In principle it is of the same character as the determina-
tion of angles on the celestial sphere by means of the circles and
disks of an astrolabe. This is a good illustration of the fact that
"Greek" mathematics was by no means rigidly restricted to some
"classical" problems, as so many modern authors seem to believe.
88. In the notes to Chapter VI No. 62 (p. 181) I have given an
example of the relationship between the theory of conic sections
and "geometric algebra" as it existed in the time of Archimedes
and Apollonius. This aspect does not exhaust by far the importance
of the ancient study of the theory of conic sections. A large part
of Apollonius's work on conic sections deals with problems which
were later, in the 19th century, classified as projective or synthetic
geometry-fields, which were developed in direct continuation
of the ancient theory. Islamic and late medieval optics (Ibn al-
Haitham, Kepler) are concerned with the focal properties. In
antiquity the conic sections are needed for the theory of sun-
dials and I have conjectured that the study of these curves
originated from this very problem.
In another ease the astronomical use of the theory of conic
sections is almost certain, that is, the proof of the fact that stereo-
graphic projection maps circles on the sphere into circles in the
 On Greek Mathematics 219

plane. This fact is a consequence of a theorem. proved by Apollo-

niusl ). that there exist on every oblique circular cone two families
of circular sections and it is easy to see that a circle on the sphere
and its image projected from one pole of the equator onto its
plane are exactly in the relation which Apollonius requires for
elements of the two families. In the existing works no proof of
this fact has come down to us but the circle-preserving quality of
stereographic projection is commonly used in the treatises on the
astrolabe and in Ptolemy's "Planisphaeriumttl).
This work of Ptolemy is another example of the combination
of descriptive geometry and trigonometric methods. and of
practical devices which. by themselves. led to the instrument later
known as the "astrolabe". The problem to be solved consists in
the determination of the centers and radii of the circles which are
the images of circles on the celestial sphere when projected from
the south pole onto the plane of the equator. In order to determine
these quantities the plane of the equator is used simultaneously
as plane of mapping and as plane of construction orthogonal to
it. For example. the radius of the circle which represents the
ecliptic is found as follows (Fig. 50). The circle abgd represents
the equator. and a plane perpendicular to it in which the diameter
bd is the axis with d as south pole. Now make az = ng = gh =

..''',
,,
.,. "
,
.. ....
,, , ... ,
,
,, I

.• •,
I

,,
I
I ·•
I
I
I

,,
I ••
,, ,,
I

• .
\
..
,
,,
......
• ,
,
...... _...... .'.-,'
Fig. 50.
1) Conics I, 5.
I) Heath is incorrect when he says (Greek Mathematics II p. 292) that Ptolemy
proves our theorem in special eases. The proofs, referred 10 by Heath, concern the
fact that also the images of great eireles intersect each other in diametrically opposite
poinls hut these proofs make use of the circularity of the images.
220 Appendix II

23;51 0 =B the obliquity of the ecliptic and project the points

n, h, and % from d onto the diameter ag which now represents the
trace of the plane of the equator. Then tm is the diameter of the
ecliptic, b the vernal equinox, t and m the solstices through which
go the solstitial circles which remain tangential to the ecliptic in
all its possible positions.
In similar fashion all the celestial coordinate systems can be
represented as families of circles in the plane. In this way it is
possible to determine by means of plane geometry the rising times
of the zodiacal signs, both for sphaera recta (tp = 0) and for
general geographical latitudes. Since the method of stereographic
projection precedes in all probability the invention of spherical
trigonometry one sees here another way of finding the answer to
problems which later were solved directly from spherical triangles.
89. The problem of mapping the sphere onto a plane arises
once more in the field of geography. Again Ptolemy is the main
source of information for us. In the first book of his "Geography"
he gives the rules for constructing a grid of curves representing
the circles of constant geographical longitude and latitude re-
spectively. The following will give a general impression of the
methods of the foundations of Greek cartography without making
any attempt to investigate the prehistory of geographical mapping.
We may only remark that Ptolemy's predecessor, Marinus of
Tyre, who wrote about 110 A.D., used for his map a rectangular
coordinate system in which the units that represented geographical
longitudes were t of the units for the latitudes tp. Consequently
the spacing of the meridians is everywhere the same as at a
latitude for which cos f{J = t. This is with sufficient accuracy
satisfied for tp = 36, the latitude of Rhodes. Thus the map of
Marinus preserves distances on all meridians and on the parallel
of Rhodes. Distances in all other directions are increasingly
distorted as one moves away from the latitude of Rhodes.
The mappings suggested by Ptolemy are much more sophisti-
cated. Two belong to the general class of conic projections, the
third is a perspective representation of the terrestrial globe. I
shall give a short description of all three methods using modern
terminology. Ptolemy assumes that the inhabited part of the
earth, the "oikoumene", lies within 63° northern latitude (Thule)
and 16;25 0 southern latitude (Hanti-Meroe", a parallel as far south
 On Greek Mathematics 221
of the equator as Meroe in Nubia lies north of it). In longitude the
oikoumene is assumed to extend 180°. and we shall count longi-
tudes L from - 90° at the western limit to +
90 at the eastern
boundary. The fIrst conic projection uses polar coordinates
(Fig. 51) which we call rand lJ. All meridians are mapped on
radii. all parallels of latitudes on circles r = const. Then three
requirements are made: (a) no distortion of lengths on meridians.
(b) nor on the parallel of Rhodes. and (c) the ratio of lengths on
the parallel of Thule and on the equator should be preserved.
The fIrst condition implies that

r=f+ c q; = 90 - f

(c a constant counted in degrees). The second condition means

for the coordinates rOt fo of Rhodes
n
-rolJ
180
= L cos fo

or
lJ = 180 . cos fo . L.
n(fo + c)

Finally the constant c can be determined from the last condition

;;1 + c
--- = COSfl
90 +c
H
222 Appendix II

where 9'1 = 63 (latitude of Thule). This leads to the value c = 25

and thus to the values of r = fJ +
25 for each latitude1). Fig. 51
is drawn to scale and shows the resulting boundaries according
to Ptolemy's construction; the arc SOlI represents 180 0 of the
parallel of Thule, MN of the parallel of - 16;25, PIT of the
equator; K lies on the parallel of Rhodes. In order to avoid the
distortions in longitude on the southern boundary, Ptolemy
arbitrarily changes the mapping south of the equator by dividing
tPZX in segments of a length as they would have had at the
latitude 16;25 north of the equator.
The discontinuity in the direction of the meridians at the
equator seemed to him less detrimental than the enlargement of
the picture beyond the length of the equator. Here mathematical
consistency was sacrificed to implausible appearance.
The second method of projection (Fig. 52) was devised to
remedy this defect and to obtain a representation which is closer
to the impression of gradually curving meridians. Again Ptolemy
requires that the radial distances correctly reflect latitudinal
differences though the radii no longer represent meridians (except
for the central meridian L = 0). Thus we have as before

(1) r = ~ + c.
For the circular arcs which now represent the meridians we
determine three points by the following conditions: preservation
of length on the parallel of Thule (9'1 = 63), on the parallel of
Syene which lies on the Tropic of Cancer (fIJI = 8 = 23;50), and
on the parallel of Anti-Meroe 9'8 = - 16;25. Consequently we
have
n
(2) - r,ll = L cos fIJ, i = 1, 2, 3.
180

The value of c in (1) determines the curvature of the limiting

parallels of the map. Ptolemy chooses c = 180 on the basis of a
simple geometrical consideration by means of which he obtains

1) ConsequenOy the north pole is mapped on the clrcle ,. == c = 25 (dotted in

Fig. 51).
 On Greek Mathematics 223

for the map of the oikoumene dimensions reasonably like the

actual ratios.
With c fixed, the map can now be constructed (Fig. 52). His
the common center of the parallels of latitude, E lies on the
parallel of Syene and therefore HE is to be made 180-23;50 =
156;10, and EO = 23;50 + 16;25 = 40;15 determines the south-
ern boundary. HN = 180 - 90 = 90 gives the image N of the
TH

-30 o
Fig. 52.

north pole. Substituting for L the same value (e. g. L = 90 for

the eastern boundary) in all three equations resulting from the
use of 9'1' 9'", 9'3 respectively gives three points through which the
meridian of longitude L must pass. In fact the curves L = const
are transcendental curves but Ptolemy replaces them by the
circle arc which is determined by the three points (L, 9',).
It may seem that this last approximation is a very crude one,
though convenient for the actual construction of the grid. In
fact, however, it is a remarkably good approximation, within the
area - 16;25 < 9' <63 and - 90 < L < 90 which contains
the oikoumene:-If (2) were requiredforall values of 9', one
would obtain the so-called "Bonne-projection" which preserves
length on all parallels of latitude. In Fig. 52. I have added in
dotted lines meridians of the Bonne projection; only for the
224 Appendix II

extreme-north-eastern area the deviation between the meridians

of Bonne and Ptolemy reach visible proportions.
The cartographic designs discussed so far are mappings in the
modern mathematical sense of this word: mathematical relations
are defined which relate a point with coordinates L, " of the
sphere to a point with coordinates r, "in the plane. This relation-
ship is not obtainable, however, as the image of the sphere seen
by an eye in suitable position. A truly perspective picture occurs,
however, in a presentation of the terrestrial globe in Book VII
Ch. 6 of Ptolemy's "Geography" though in a very inconsistent
combination with a mapping of the Bonne type. Ptolemy assumes
a terrestrial globe mounted between rings which represent the
arctic circles, the solstitial circles and equator and ecliptic. A
perspective picture of these rings is then constructed, seen from
a center of projection located in such a fashion that no part of the
area of the oikoumene is obscured by a ring. The globe within
the rings, however, is not represented in perspective but simply
as a circular frame of a map similar to the second above-described
networks. It is more a book illustration than a real map which is
described here, and is the only case in all of Ptolemy's writings
where he displays an inconsistent and totally useless construction,
thus foreshadowing the taste of the Middle Ages.
90. The history of mathematics provides good illustrations for
the fact that continuity of tradition alone is not sufficient to keep
a scientific field alive. The "Elements" of Euclid formed for
centuries the basis of mathematical instruction. Nevertheless the
significance of its axiomatic structure was not understood until
the problems connected with the foundations of analysis led the
mathematicians of the 19th century to similar methods. A particu-
lar ease in point is the theory of proportions in Book V which was
only seen in its proper role within Greek mathematics through
the theory of irrational numbers and continuity developed by
R. Dedekind since 1858. Similarly it required the recent develop-
ment of formal logic to discover the existence of similar systems
in the writings of Aristotle and the Stoic and Megaric philosophers.
I am not competent to discuss the modern aspects of Aristotle's
logic and of the later schools. One aspect, however, should be
mentioned which concerns mathematics in its narrower sense,
that is the use of letters in the formulation of syllogisms. For
 Bibliography 225
example the figure later on called "Barbara" is given as "If A is
predicated of all Band B is predicated of all C, then A must be
predicated of all Cl)". The introduction of variables, represented
by letters into logic would seem to have constituted the essential
step to a formulation of mathematical rules which we would call
..Algebra". Such a development would have been still more
natural since the axionlatic approach to mathematics originated
in the same time and among the same circle of men. Yet nothing
of this type happened and the origin of algebra is totally in-
dependent of the existence of an algebraic notation in one of the
most famous philosophical works of antiquity.
This is a good illustration for the futility of any attempt to
reconstruct "reasons" for the incidents of historical events.
Similarly the absence of algebraic notations should not have
prevented the Greek geometers from developing what was called
in the 19th century "synthetic" and "projective" geometry since
many of the basic concepts were ready at hand in the works of
Apollonius. Again such a "natural" development did not take
place and all that we may ever hope to establish in historical
research is facts and conditions but never causes.

BIBLIOGRAPHY TO APPENDIX II
There is no lack of histories of Greek mathematics and every library catalogue
will suffice to identify many of them. I therefore restrict myself here to quoting
a few comparatively recent publications which have not yet become commonly
used.
A work of Archimedes on the regular heptagon has been recovered through
an Arabic translation, a German summary of which was published by C. Schoy
in his work "Die trigonometrischen Lehren des persischen Astronomen ...
Al-BfrQnY" (Hannover 1927) p. 74-91.
The "Sphaerica" of Menelaos,likewise only preserved in Arabic translation,
or version, by Abu Na,r Man,Ul (about 1000 A.D.) is available in a critical
edition with German translation by Max Krause (Abhandlungen der Gesell-
schaft der Wissenschaften zu Gottingen, philol.-hisl. Kl., 3 Folge, Nr. 17, Berlin,
Weidmann, 1936).
A new edition of one of the earliest Greek mathematical works that have
come down to us (written perhaps about 330-300 B.C.) was given by J. Mo-
genet "Autolycus de Pilane; histoire du text, suivie de redition critique des
traites de la Sphere en Mouvement et des Levers et Couchers (Louvain 1950,
1) Translation of Analytica priora I, 4 25 b 37 by Lukasiewicz (Aristotle's Syllogistic
p. 3 and p. 10).
226 Appendix II

&eueil de Travaux d'Histoire et de Philologie, 3e ser., fasc. 37). In the meantime

O. Schmidt made the interesting discovery that the two "books" of the "On
Risings and Settings" are actually only two versions of the same work; d. the
article "Some critical remarks about Autolyeus' On Risings and Settings" in
the transactions of Den 11te skandinaviske Matematikerkongress i Trondheim
22-25 August 1949 (published Oslo 1952) p. 202-209.
A. Lejeune, Euclide et Ptolemee; deux stades de l'optique geometrique
greeque (Louvain 1948, Recueil de Travaux d'Histoire et de Phllologie, 3 e ser.,
fasc. 31) is a work of great historical and methodological interest. In a careful
analysis we see here the progress from a strictly geometrical optics to a theory
of binocular vision and physiological optics based on empirical data and
systematic experiments. Ptolemy's optical theories touch also upon the problem
of the three-dimensionality of space, a subject on which he also wrote a special
treatise, now lost.1) Ptolemy's Optics is only preserved in a Latin version of an
Arabic translation, (edited by A. Lejeune, Louvain 1956, l. c., 4e ser., fase. 8).

NOTES AND REFERENCES TO APPENDIX II

ad 85. The geodetical problem. mentioned on p. 210, is known as the
Snellius-Pothenot problem. It was solved by W. Snellius in his "Eratosthenes
Batavus" Leiden 1617 p. 203 f. J.A. C. Oudemans stated in VierteJjahrschrift
der astronomischen Gesellschaft 22 (1887) p. 345 that Ptolemy's problem was
identical with the problem of Snellius. This, however, is not the case. Ptolemy
assumes as given 8b 8., and ~, ~I and R whereas Snellius knows beyond 81
and 8. all three sides "10 SS, Sa of the triangle.
The problem has also becn discussed by Delambre, Hist. Aslron. Ancienne II
p. 164 fl. Cf. also Tropfke, Geschichte der Elementarmathematik V, 2nd ed.
1923 p. 97.
ad 87. Literature concerning the uAnalemma": Ptolemy's treatise is published
by Heiberg in Ptolemaeus, Opera II, p. 189-223 (1907). For the discussion of
the method, d. Delambre, Histoire de l'astronomie ancienne II p. 458-471
(paris 1817); J. Dreeker, Theorie der Sonnenuhren (in: Bassermann-Jordan,
Geschichte der Zeitmessung und der Uhren, Bd. I, E; Berlin 1925); and in
particular the article by P. Luckey, Das Analemma von Ptolemiius, Astron.
Nachriehten 230 No. 5498 p. 17-46 (1927), to whom we owe the understanding
of the nomographic procedures of Ptolemy's IIAnalemma".
For Vitruvius and Heron d. O. Ncugebauer, 'Ober eine Methode zur Distanz-
bestimmung Alexandria-Rom, Kgl. Danske Vidensk. Selsk., hist.-fIlol. medd.
26, 2 and 26, 7 (1938-1939).
ad 88. O. Neugebauer, On the astronomical origin of the theory of conic
sections. Proc. Amer. Philos. Soc. 92 (1948) p. 136-138. My argument is based
on the fact that the earliest form of the theory assumes that always one generating
line is perpendicular to the intersecting plane, an arrangement as in the case
of the gnomon with respect to the plane upon which the shadow is cast. The
difficulty of this conjecture lies in the ract that no sun-dials seem to be preserved
in which the gnomon is directed toward the culminating sun.
In passing, it may be remarked that the theory of sun-dials is perhaps the
1) Cf. PtoJemaeus, Opera astron. minora. ed. Helberg, p. 265 f.
 Notes and References 227
origin of one of the "classical" problems of Greek mathematics, namely, the
trisection of angles. We know from Pappus (Collection IV, 27) that he used the
conehoid of Nicomedes (2nd century B.C.) in order to trisect an angle in connec-
tion with a work of Diodorus (first century B.C.) on the theory of sun dials.
Here the problem arises of constructing the 12th part of the are which the sun
is above the horizon, because this is the equivalent of one "seasonal hour".
ad 89. The best discussion of Ptolemy's theory of map projection is given
by H. v. M!ik.-F. Hopfner, Des K1audios Ptolemaios Einftlhrung in die dar-
stellende Erdkunde, Klotho 5 (1938). Cf. Also H. Berger, Gesehichte der wissen-
sehaftlichen Erdkunde der Griechen, 2. Autl., Leipzig 1903 (p. 632 fT.).
A great variety of ancient maps is reproduced in the monumental work
Claudii Ptolemaei Geographiae Codex Urbinas Graecus 82 (= Codices e Vatieanis
selecti vol. 19), Leiden-Leipzig 1932 (4 vols.), edited and commented by J. Fi-
scher. It must be emphasized, however, that the "Geography" in eight books,
as it exists today, is in all probability not the work of Ptolemy but rather a
Byzantine compilation, as L. Bagrow has shown ("The Origin of Ptolemy's
Geographia" Geografiska Annaler 1943 p. 318-387, in particular p. 368 fl.).
The mathematical sections, ("Book I") are unquestionably genuine.
Ptolemy himself was fully aware of the fact that his second conic projection,
with circular arcs as meridians, was only an approximation, though a very good
one, to a mapping in which distances were preserved on all parallels of latitude.
It is only for the simplicity of construction that he restricted himself to three
parallels. The accurate meridian lines, for the ease of the north pole as center,
were given by Johannes Werner in connection with his translation of the first
book of Ptolemy's Geography (Niirnberg 1514 and again Ingolstadt 1533 with
an introduction by Petrus Apianus). This method of projection became very
popular after it was used in an atlas by R. Bonne (1787) and was later adopted
for the mapping of France on the recommendation of Laplace.
The fact that this "Bonne-projection" is area-preserving was, of course,
unknown in antiquity.
ad 90. For the relationship between the Greek theory of irrationals and its
modem counterpart d. R. Dedekind's monographs "Stetigkeit und irrationale
Zahien" (1872) and "Was sind und was sollen die ZAhien" (1888) as well as
his correspondence with R. Lipschitz (R. Dedekind, Gesammelte mathematische
Werke III p. 469-479). Cf. furthermore: O. Beeker, Eudoxos-Studien, Quellen
und Studien zur Geschichte der Mathematik, Abt. B vol. 2, p. 3ll-333; p. 369-
387; vol. 3, p. 236-244; p. 370-388; p. 389-410 (1932-1934).
K. v. Fritz, The discovery of incommensurability by Hippasus of Meta-
pontum, Annals of Mathematics 46 (1945) p. 242-264.
Van der Waerden, Die Arithmetik der Pythagoreer, Mathem. Annalen 120
(1947/1949) p. 127-153; p. 676-700.
In recent years a large number of books and monographs have been written
on problems of ancient formal logic. It may suffice to quote the following works:
Jan Lukasiewicz, Aristotle's Syllogistic from the standpoint of modem
formal logic, Oxford, Clarendon Press, 1951.
I. M. Bochenski, Ancient Formal Logie, North-Holland Publishing Company,
Amsterdam 1951.
Benson Mates, Stoic Logie, Univ. of California Press, 1953.
228

THE ZODIACAL SIGNS

ep Aries c Libra
as Taurus III Scorpio
u Gemini t Sagittarius
... Cancer n Caprieom
nLeo - Aquarius
111 Virgo )( Pisces

THE PLANETARY SYMBOLS

l! Saturn eSun
2l. Jupiter ~ Venus
4 Mars ~Mereury
 229

CHRONOLOGICAL TABLE
cr. also the Frontispiece.
Dates are only approximate.

-1700 Old Babylonian 360 Theon Alex.

-1300 Seti I 380 Paulus Alex.
- 650 Ashurbanipal 450 Proclus
- 430 Meton 600 Aryabhata
- 376 .Archytas 500 Bhetorios
- 370 Eudoxus 650 Varlha Mihira
- 360 Aristotle 650 Severus Sebokht
- 311 beg. of Seleueid Era 650 Brahmappta
- 300 Euclid 825 al-Khwlrizml
- 276 Aristarch 850 Abo Ma·shar
- 275 Aratus 900 al-Battlnl
- 275 Berossos 1000 ibn-Yonus
- 250 Eratosthenes 1000 ibn al-Haitham
- 240 Archimedes 1000 al-BIrOnI
- 200 Apollonius 1000 Suidas
- 150 Hipparehus 1130 Adelud of Bath
- 100 Theodosius(f) 1150 Bhaseara
- 100 Teukros(!) 1170 Malmonides
76 GemiDus 1250 Alfonso X
+ 10 Manitius 1250 Bar Hebraeus
60 Pliny 1430 U1Dgh Beg
76 Heron 1500 Copernicus
75 latest cuneif. text 1640 Bhetieus
100 Menelaos 1575 Tyeho Brahe
150 Ptolemy 1600 Kepler
160 Vettius Valens 1680 Halley
160 Galen 1686 Newton
340 Pappus 1700 CassiDi
360 Diophantas(?t) 1760 LeGentil
230

INDEX
AbO 'I-FaraJ 179 anomaly of solar motion 6. 192 <lee also
Abil'l W6fa 207 solar theory)
Abil Ma'shar 172, 176, 229 Antiochus of Athens 189
Abydenus 141 Antiochus I of Coll1JlUlgene 187
acrophonic numerals _ numbers Apononius 67. 145, 155, 182, 218. 219,
ACT 138 225.229
additivity 73 application of areas 149, 181
Adelard ot Bath 229 ap~oDmationsofNclprocals34
ahargana 187 - y2 35.47,50.52
Akkadian 31, 58 - fa 47,52
al- lee second part of name - n 23.46.47.51.52. 78. 180
Alexandria, length ot daylight 158. 184 Arabic lee Islam
Alfonsine tables 67, 229 Aratus, commentary to 69. 185. 186. 229
algebra 208, 225 Archibald 66,91
- Babylonian 40, 42, 44. 51 Archimedes 145. 229
- geometrical aJg. 147, 149, 181. 218 - Arabic trsl. 225
- al-Khwlrizmf 146 Archytas 148. 229
algorithmic fractions 75 Arlkamedu 167. 187
Allen 96 Aristarchus 229
Almagest 13, 55. 145. 214 Aristotle 151. 224. 225. 227. 229
- approximation ot n 180 arithmethic progression 40, 100 ( _ .
- al-Blriinl. translation 176 difference sequences; zigzag func-
- catal. ot stars 68. 185 tions)
- computations 33, 72. 158 arithmetical methods _ Unear methods
- edition 54 arrangement of the planets 169, 187
- hl PersIa 187 Arsaeid era 103
- planetary theory 126, 198 Aryabhala 180. 183. 229
- table of chords 10,85,209 aI for l/u 27
- table of syzypes 95 ascension _ rising time; right ase.
- transl. hlto Sanskrit 176 Ashurbanipal 59. 229
- Vi 35 Assyrian period XVI, 101. 170
(see also Ptolemy) astrolabe 54. 161, 176, 185,219
almanac, type of cuneif. records 139 astrology. eatal. of Greek 1D98. 56. 68
alphabetic numerals _ numbers - Henenistic 171, 187, 188
Ammi~duqa 100, 139 - Mesopotamian 18, 139, 168. 171. 187
amphora letters 25 - origin 100, 171, 188
Anastasi pap. 79 - time of origin 100. 187. 187
Angelo PoUziano 54 - trausmission 168, 171, 187
anomaHstic month. 121. 162. 166, 185, (_ also horoscopes)
193,210 atomistic theories 149
 Index 231

Augustus 90. 215 Breasted VII, 92

- era 164. 187 British Museum 59. 69. 104. 140
Aulolycus 214. 225 Brugsch 94, 188
auxiliary numbers with Eg. fractions 93 Bruins 47. 49, 50
Avicenna 24 Budge 69. 140. 179
Bukhara 24
Burgess. E. 144. 186
Babylon. archive of astron. texts 115.136. Burgess. J. 178
139 Byzantine period XVI
- in Egypt 189 - astron. tables 55, 172
- length of dayHght 159, 187 - era Yazdlgerd 81
- Seleucla 189 note 1 - geography 184. 227
Babylonian astrology 18, 139, 168, 171. - Hermfppus 172
187 - India 175
- calendar see lunar calendars - zero symbol 14. 26
- chronology XVI. 14, 100, 101
- intluence on Greek astron. 157
Bagrow 227 calendar see diagonal cal.• Egyptian cal.;
Bailly 186 lunar cal.
Bakir 188 Callisthenes 151
Baqlr 52, 70 Caminos 95
Bar Hebraeus 179. 229 Capelle 189
Barbara syllogism 225 Carlsberg, pap. 1 58, 69. 87
al-Battllni 176. 206, 229 - pap. 9 90. 95. 164-
Bell 57 Carra de Vaux 207
Bergstrisser 180 cartography 220. 227
Berossos 141, 157. 229 Cassini 186. 229
Bezold 139. 170, 188 Cassite XVI. 99
Bhascara 51, 179. 229 catalogue of stars
Bhattotpala 189 - ~est68, 185
bibliography 55 - Babyloni~ 185
Bidez 178 - Hipparchus 56, 69, 185
Bigourdan 141 Catal~Cod.Astrol.Graec.56.68
binominal coefficients 51 Caussin 190
al-Blrani 174. 176. 177. 179, 180. 183, celestial globe 144
229 Cerny 94
Bober 24 Chahas 188
Boelk 183 Chace 91
Boer 54.68 Chaldeans 18, 97, 170. 187
Bo1169.139,170, 171.185, 187.188.189 Charlesworth 187
Bonne projection 223, 224. 227 Cherniss 182, 187
book of hours 3. 23, Plate 1 Chiera 23
Borchardt 79, 94- chords. table of 10, 35. 186
Botta 59 chronology XVI, 14.94, 100,101,292
Bouch6-Leclercq 170. 188 Cicero 188
Brahe see Tycho Brahe circle. area 51
Brahmagupta 180, 229 - circumscribed 47, 52
232 Index

circle. segment 51 Datta 180, 189

dreular motion 153. 155, 199, 203, 205 day see length of daylight
Clark 178, 183 deeans 58, 81.82.84,87.88.94
Clere 94 decimal fractions 23
climates 183. 185. 187, 218 decimal order of hours 84. 86
clock. stars as 83 decimal system in Mesopotamia 17, 27
clocks. meehanical161, 185 declination 185, 214. 217
eoefficients. list of 45, 49 Dedekind 224. 227
coffin texts 83 deferent 123
coins 9.25 degrees 17. 25
Colebrooke 51 Deimel26
colophon on tablets 17. 63, 136, 144 Delambre 186. 206. 226
Colson 189 Delatte. A. 54
complementary fractions 74 Delatte. L. 69
compound interest 34, 44 Deleage 91
computing on fingen 9 Demotic. astron. texts 90. 91. 162. 164
- with shells 165 - horoscopes 90. 168
computus 24 - matheDl.pap.78,91,92
conchold of Nicomedes 227 - P. Carlsberg 1 58, 69. 87
conic sections 67, 181,218,226 - P. Carlsberg 9 90, 95, 164
conjunction, Inferior and superior 126 - planetary tables 90, 94, Plate 13
- moon 106, 118 Denkart 187
Cook 143 descriptive geometry 80.215.219
coordinate lines for stars In Egypt 89 Devaram 166
coordinates, the name 181 diagonal calendars 85, 88, 94
Copernicus 3.22,81, 194. 196. 197,202, Dlels 185
204, 207, 229 digamma 25
Coptic. eclipse record 95 digits, eclipse magn. 119
CorneU,pap,r1179,P~te12 dimensions of space 226
crescent see v1sibmty Diodoros 227
cryptographic writings 144 Dionysius 179
cubes, table of 34 Diophantus 51. 80. 147. 148
cubic equations 34. 44. 51 - date 178
Cumont 56, 68. 170, 178, 187, 189 - pythagorean numbers 50
cuneiform 14 distance. of fixed stan 99, 139
eyele. 18-year 116. 142 - of planets 156. 205
- 19-year see MeloDic c. DornseUf25
- 25-year 90.95, 164 Dorotheos of Sidon 172
Cyranldes 69 drachma 26
Drachmann 178
daily motion of sun and moon. Bah,yl- Drecker 226
onian 121. 163. 166 Dreyer 206
- of planets 135 Dropsie College 139
- Tamil 166 Du Cange 54
- use of titbis 186 Duc de Berry 3. 23
Darius 27 Duhem 206
Dashur, bent p,ramld 96 duplication 73.92
 Index 233
Dura 26 Erman 79, 92
DUrer 52 errors in place value 28
Durrieu 23 Euclid, Arabic version of Elements 57
Dvivedi 186 - spherical astr. 214
- Sanscrit trsl. by al-Birfini 176
Easter 7 Eudoxus 229
eceenter 125, ISS, 167, 192 - astrology 188
eclipses, Babyl. theory 109,116,117,119 - astronomy 153, 182
- records from Egypt 95 - mathematics 147, 152
- records from Mesop. 98, 101 - oriental influence 151
- Saros 116, 141 evection 207
- tables of Vettius Valens 175 evening epoch 106, 118
- Tamil methods 165, 186 evening star 126
ecliptic see zodiac excavations 59
ecliptic coordinates 69, 185, 207 - in museums 60, 62
Edwards 96 exponential function 34,44,51
Egyptian astronomy 78, 80, 82, 86, 88,
89,92, 156 Fakhry 96
Egyptian calendar 81, 90, 94, 162, 164 Falkenstein 26
(see also lunar calendars) Festugiere 188
Egyptian planetary texts 90, 94, 162 figures, in manuscripts 54
Eisenlohr 91 - in cuneif. texts 46, 52
Ekur-zlkir 136 fingers, computing on 9, 24
Elam 46 first visibility, moon 106, 119, 121
EWs~5 - planets 127, 129
elongation, moon 107, 119, 121, 187, 194 fixed stars, distances 99
- planets 126 formula, its use in Bab. math. 43
empirical basis for astron. theory 130 Fotheringham 139
(see also observations) fractions, symbols 20, 26
enfima-Anu-Enlil 101, 104, 136, 139 (see also unit fractions)
ephemerides see daily motion Frankfort 94
epicycle 123, 155, 192, 196, 205, 210
epoch see evening-ep.; midnight-ep. Gadd 51, 138
Epping 97, 103, 109, 157 Galen XV, 152, 180, 229
equant 198 Gandz 51, 180, 189
equation of center 192 Gardthausen 144
equations, degree 4 and 6 44, 51 Geminus 161, 162, 185, 229
- degree 5 44, 51 geographical latitude 158, 184
- degree 8 48 (see also climates)
(see also cubic equ.; quadratic equ.) Geography, Ptolemy's 55, 220, 224, 227
equinoctial hours 81 geometrical algebra 147, 149, 181,218
era, Augustus 164, 187 geometry, Babylonian 44, 47
- Nabonassar 98, 101 - Egyptian 78
(see also Parthian era; Salta era; - Greek 80, 149, 208
Seleucid era) Gerstinger 92
Eratosthenes 229 Gilgamesh 138
Erichsen 95 Gillings 50
234 Index

Giozel 23. 95 Hindu-Arable numerals see numbers

Glanville 91 Hindu astronomy 82, 162, 169, 172, 178,
globe 144. 224 183, 186, 214
gnomon 214 Hipparcbus 145, 157, 161, 167,193.210,
Goetze 45, 49 214.229
golden number 8, 24 - astrology 187
great yean 90 - division of equator and ecliptic 186
Greek-letter phenomena 127 note 1
Greek mathematics 80, 146, 147, 149, - eclipse tables 175
190" 208, 225 - star catalogue 56, 69, 185
Griffith 91 - stellar coordinates 185
Gu~raud-Jouguet 25 bippopede 154, 182
Gundel 56, 68, 94, 189 Hippopotamus 88. Plate 10
Gunn 25 Hittite astrology 188
homocentric spheres 153, 182
Hadrian 90, 95 Honigmann 185
Hakemite tables 175 horizon, visibility 98, 127
Halley 67, 142, 229 horoscopes 68, 90, 103, 168. 183, 187
Halma 55 hours 81. 84, 86. 169
Hammurapi XVI, 29 - division in 1080 parts 180
Haskins 67, 177 ~unain ibn Is~iq 180
Hayes 92 Hypsieles 183
Head 25
Heath 208, 219 Ibn 81nl24
heetemoros 217 ibn YQnus 175, 190, 229
Heiberg 54 Ideler 141
heliacal rising 83, 85 inaccuracy of figures in Bab. texts 52
heliocentric system 123, 183, 204, 205 India see Hindu astron.
Hellenistic period 1. 145 inheritance 44, 80
hemerology 94, 188 inhomogeneous problems 42, 48, 51, 146
hemisphere, area 78 instruments, inaccuracy 185, 210
heptagon 47, 225 intercalations 102, 140
Hermes Trismegistos 68, 171 <aee also lunar calendars)
H~pposI72, 189 interest, compound 34, 44
Herodianus 24 interpolation 28, 135
Herodotos 142, 188 invocation 136
Heron 146, 178, 179, 215,226,229 lonians 143, 167
- approx. ofV2 and Vi 52 Irani 26
- date 178 irrationality ofV2 48, 148
- Heron-Dioph. writings 80, 148, 179 irregular numbers 33, 34, 50
- regular polygons 47 Islamic period
Herz 206 - algebra 51
hexagon 47 - astron. tables 175. 178, 201
Hilbert 145 - Copernlcus197.203
Hill, G. F. 24 - decaos 82
Hill, G. W. 152 - era Yazdlgerd 81
Hilprecht 27, 70, 99, 139 - India 82, 172. 177, 180
 Index 285
- numerals 4. 14. 22. 24. 189 leisure class 151
- numerical methods 23 Lejeune 226
- rising times 160. 184 Lemoine 24
- trsl. of Greek works 57. 67. 176. 180. length of daylight. variation of. Babyl-
225.226 onian scheme 116, 159
- zero symbol 26 - Egyptian scheme 86. 94
(see allIO individual names. Battinl. - extremal values 158. 184
BIrOni. etc.) - India 162. 187
Leverrier 200, 207
Jean de France 3. 23 Uvy-Bruhl92
Jena 30. 99. 139 Lewy. J. 139
Jeremias 138 library. Assyrian 59
Jupiter 123. 129. 133. 173 LnivaU 51, 179
linear methods 158. 160, 167. 178, 183.
Kahun.pap.91 186. 187
Karpinski 24 linear zigzag function see zigzag f.
al·Klshl 23. 28. 184 Lithberg 24
Kempf 183 logarithms 44. 51
Kennedy 68. 178, 184 logic 224. 225. 227
Kepler 3. 205. 206. 218. 229 longest and shortest day see length of
Keskinto 25 daylight
Khorsabad 59 longitudes, zero point 188.207
al-Khwlrizml 146. 176. 179. 229 LU 132
Kidenas 137. 144. 176 Luckey 23, 28. 226
Kish 27. 51 lucky and unlucky days 94. 188
Klein 178 Lukasiewicz 225, 227
Knudtzon 94. 162. 186 lunar calendars 168
Kopft' 139 - Babylonian 81. 102, 106, 119, 128
Krall 95 - Egyptian 82.90.94
Kramer 50 - Greek 10. 81
Krause 225 - Middle Ages 7. 91. 102
Kroll 170. 189 lunar days Ite tithis
Kugler 97. 104. 109, 131, 138. 142. 157. lunar eclipses Ite eclipses
162 lunar month 7. 106
Kuyunjik 59 lunar (and solar) theory 152
- Babylonian 116. 143
Lammert 68 - Copernicus 196
Langdon 26. 27. 139 - Islamic 207
Lange 58. 69. 94 - Ptolemy 183. 191. 206. 207. 210
Larfeld 24. 25 lunar velocity 118, 121. 162
latitude, moon 108. 117. 119, 121 lunation 7
- planets 125. 128, 191,205
Lauer 96 Macedonian inscriptions 26
Layard 59 magic. Greek 171. 180. 190
leather roll BM 10250 91 Mahaft'y 25
Leemans 49 MahlVlra 51. 179
LeGentil142. 167. 178. 186.229 Maimonides 229
236 Index

mana 19 natural fractions 26. 75

Manillus 229 Nau 179
Manitius 54 Nero 163
mapping. geogr. see cartography Newton 3. 229
Marigha 207 Nne 82, 143
Marinus of Tyre 220 Nimrud Dagh. horoscope 187
Mars 123. 133. 173, 206 Nineveh 59
Martin 151 Nippur 27, 30, 70, 100, 139
McCown 70 nomography 218.226
MCT49 normal form or quadratic equations 41.
medicine 2, 180 150
Meissner 188 numbers and numerals
melothesia 24 - acrophonic 4. 9. 24
Menasce 187 - alphabetic 6, 10, 14. 26
Menelaos 161, 214, 225. 229 - - on Greek coins 9, 25
Menninger 23 - mndu-Arabic 4. 24, 189
Mercury 2, 123, 169 - number words 4
- Babylonian theory 28, 129, 131. 182. - Roman 4,5
134, 1'3 - subtractive 5. 64
- Copernicus 202, 207 (see also approximationsi zero)
- India 173 Nut 58
- Ptolemy 188. 200
Meritt 24 obllque ascension see rising time
Metonie cycle 7, 95. 102, 140, 142 note 8 observations, in Egypt 95
Metropolitan Museum 60. 88 -, in Mesopotamia 98, 100, 101. 127.
Meyerhof 180 137, 140
Michigan pap. 92 Old Babylonian 29
midnight epoch 118 omens 100, 101. 189, 188
Mishnat ha-Middot 179. 180 optics 218. 226
MKT49 origin of astronomy 99. 168
M6bius 183 origin of geometry 151
Mogenet 225 Orion 83, 87. 88, 99
Montucla 142 Osorkon. eclipse 95
moon see lunar Oudemans 226
morning star 126
Moscow, math. pap. 78,91 Palazzo Sehiranoja 82, 171
mul-apin 101, 139 panbabylonistic doctrine 138
multiplieation, Egyptian 21. 78. 93 Paftca Siddhlntiki 144, 165, 172. 186.
multiplication tables 16. 80. 81, 92. 189
Plate 4a Pannekoek 141. 186
Mtik 227 Pappus 55. 227. 229
papyrology 57
Nabonassar 98, 101 papyri 78, 92
Naburianos 137 - Anastasi 79
Nalllno 175, 189 - Ayer 179
Narain 178 - Berlin 661991
Nisir ad-Din at-Tusi 203.207 - Cairo 65445 25. Plate 5
 Index 237
- California 162 Pliny 137. 142, 229
- Carlsberg 1 58. 69. 87 Poeoeke 179
- Carlsberg 9 90, 95. 164 Pogo 94
- Cornell 69 179. Plate 12 Poincare 152
- Kahun 91 polar longitude and lat. 186
- Lund 35a 94. 162. Plate 2 Pondicherry 167
- Michigan 146 92 Poole 25
- Moscow 78. 91 Porphyrius 151. 189
- Paris 19 183 Pothenot 226
- Petrie 25 precession 68. 69. 191
- Rhind 75, 78. 91. 92, 93 Preisendanz 67. 190
- Rylands 27 164, 187 preservation of tablets 61
- Salner IV 188 Price 206
- Tunah el Oabal 91 primationes lunae 7
- Warren 67 prime numbers 40
paradoxa 148 problem texts 30
parallax 119. 156. 191. 196 Proclus 182. 187
paranate1lonata lee decans product of areas 42. 48
Parker 69. 82. 92. 94 projective geometry 218. 225
Parthian era 103 proofs 45. 148
Paulisa Slddhlnta 7. 175. 189 prosthaphairesis 192
Paulus Alexandrinus 175, 229 provenance of math. cun. texts 51.60. 70
Peet 91. 92. 144 Psellus 178
Pehlevi 175. 189 Ptolemaic XVI. 13 note 2
pentagon 47 Ptolemy 13. 55. 140. 145.226. 229
Periplus 166. 187 - Almagest see under A
Persian intermediary 166. 172. 175. 187. - Analemma 215. 226
190 - Canobic inscr. 195. 207
Persian years 81 - catal. of stars 68. 185
Peters and Knobel 68 - Geography 55. 220. 224. 227
Petrie papyri 25 - lunar theory 183. 191. 206. 207. 210
Phoenician alphabet 11.24 - optics 226
Pinches 49. 106. 133. 138. 143 - planetary theory 126. 156. 183. 191.
place value notation 5. 18. 22. 175. 189 198. 206. 207
planetary theory 122. 206 - Planisphaerium 85.219
- Babylonian 126. 129. 132. 143. 172. - references to earlier observations 95.
173 98. 101. 157. 167. 215
- Copernicus 202. 205. 207 - Tetrabiblos 54, 68
- Egyptian 90. 94 pyramids 96
- Greek (incl. Ptolemy) 25. 90. 126. pyramid texts 9
153,156.162.182.183.191.198.206 Pythagoras 36. 148. 149
- Hindu 172. 173. 186 Pythagorean numbers 36. 42. 50. 179
planetary week 169. 187. 189 Pythagorean theorem 36
planets. arrangement 169. 187
- symbols 54. 67.228
Plato 27. 151, 152. 182 qoppa 25
Plimpton Collectlon 36 quadratic equations 41, 51. 66. 150
238 Index

Ramses 89. 94 Schumacher 183

Rassam 59. 69 scribal families 136
ratio. Babyl. term for 51 seasonal hours 81, 86
Read 188 secrecy 144
reciprocals 32, 34. 50 SMillot 207
regular numbers 33. 39 Seleucia 189 note 1
regular polygons 47 Seleucld era 103, 110, 116.229
Regulus 207 Seleuclds XVI, 14. 29
Rehm 170 Senmut 81,88,92,94, Plate 10
retrograde motion 125. 154 series of cuneiform tablets 63, 139
Rheticus 152. 229 Sethe 23. 24. 92, 94
Rhetorius 229 Seti I. cenotaph 58, 86. 87. 88, 94. 144.
Rhind math. pap. 75. 78. 91. 92. 98 229
right ascension 143. 159 note 1 Severus Sehokht 229
rising times 143, 159, 184 sexagesimal fractions 12, 13. 16
roads of heaven 101 sexagesimal system 17, 22, 186
Robbins 54 Shapur 1187
Robert 68 ash-Shltir 197
Roberts 197 sheDs as computing aid 165
Romaka Slddhinta 7. 175 ships of prismatic volume 180
Roman calendar 4 sickle of moon 1~1
Roman numerals see numbers sidereal month 1'92
Rome. A. 54. 55 SimpJicius 151
rounding off 68 Sin IX == R· sin IX 214
Ruska 24 sin-tables 186
Rylands,pap. 164. 187 Sln-Ieq6-unninni 136
Sippar 60, 144
Saehau 177 Sirius 82. 85,87,88, 94,99, 127, 140, 143
Sachs 34. 49. 50, 97. 106, 138. 140, 143, small years 90
185.187 Smith-Karpinski 24
Bade 25 Snellius 226
Saka era 174 solar eclipse ,ee eclipses
saltus lunae 8 solar theory 6, 123. 156. 175
sampi 25 - Babylonian 110, 114, 116, 118
Saros 116. 141 - Greek 156, 186. 192, 204
Sasanlans XVI, 81 - Hindu 186
Saturn 62. 123. 133. 173 (see allO lunar theory)
Schaek-Sehackenhurg 91 Sothlc period 94
Schaefer 92 Sothls see Sirius
Schaumberger 97. 104. 109. 139 sphaera barbarica 170. 171
Sehlapare1li 141 spherical trigonometry see trigonometry
Schmidt 187,226 sphericity of earth 153
Schnabel 141. 161, 162, 186 split writings 27
Schoch 139 square roots 34,51
Schoff 187 squares. sum of 34, 40
school texts 30, 49 squares, tables of 26, 27, 34
Schoy 225 stationary points 98, 126
 Index 239

Stegemann 189 Theodosius 161.229

Steinschneider 177. 189 Theon (1st cent. A.D.) 198
stereographic projection 54. 161.185. 219 Theon (4th cent. A.D.) 13. 33. 55. 178,
stigma 25 201. 229
Strabo 137 Theopbrastus 187
Strassmaier 49. 103, 138. 140. 143. Plate theory of numben 149
14 Thibaut165. 178. 186
Struve 91 Thule 220
stylus 18,26 Thureau-Dangin 49. 139, 161
subtractive writing of numben lee tithi 128. 186
numhen Tod 24. 26
Sudines 137. 176 Trajan 95
Suidas 142. 229 triangle and circle 52
Sulva-SOtra 35 trigonometry 209, 214
sum of squares 34, 40 - plane 186.209,220
Sumerian 30, 49. 99 - spherical 80, 117. 121. 160. 161. 220
sun dial 86. 144, 182, 214, 218. 226 trisection 227
sun, symbol for 67 Tropfke 52. 226
SOrya-SiddhiDta 144,174,175,186,189. Tunah el Gabal 91
214 at-Tilst 203, 207
Susa 46 Tycho Brahe 205. 206, 207. 229
symbols lee planets, lee sun twelve-division of day 81, 86
Synke1los 141
synodic month 7, 112, 122 Uliigh Beg 23.229
synodic periods of planets 173 uncia 27
Syriac 179, 180 unit fractions 21. 50. 72, 74
Systems A and B of the lunar theory 114. Uruk. archaic texts 26
118, 140. 172. 183, 188 - archive 115. 136
- of rising times 159. 183. 185 - Seleucld astron. texts 61. 161
Systems of planetary theory 129, 133. 172
van der Waerden 92. 94, 140. 186. 189.
table texts 30 206.227
tables of multiplication 16, 30. 31, 92. Van Boesen 68
Plate 4a van Wijk 24
tables of squares etc. lee squares etc. Varlha Mihira 165, 172. 173, 176. 185.
Tablet House 50 189
Tamil 165. 167. 186 variation 207
Tanis 94 Venus 100. 123. 134. 173
Tannery 25, 141. 178, 206. 207 Venus tablets of Ammisaduqa 100. 139
Tarn 177. 178, 189 Venus transits of 1761 and 1769 186
Tell ~armal 49. 51. 70 vemal point 188, 207
temple library 70 <lee also precession)
tenilu 137 Vespasian 95. 139, 163
Tetrabiblos 54, 68 Vettius Valens 137, 175, 187. 190.229
Teukros 171. 189, 229 visibility conditions. Mercury 132, 143
Thlbit ibn Qorra 176 - moon 98. 106. 107. 119. 121
Thales 142, 143, 148 Vitruvius 215, 226
240 Index
Vogel 91, 92 year. Egyptian 81. 82. 95. 162
Vogi 69.185 - Persian 81
Volten 91. 95 - sidereal 140
volumes 46. 51. 52 - tropieal95, 140. 192
Warka Bee Uruk <Bee al80 lunar calendars)
W~.J. 165. 186.189
Warren. pap. 67 zero. in the Almagest 11
waw25 - Arabie 26
week 169. 187. 189 - Babylonian 14, 16.20,27,29
Wegener 67 - India 189
Weidner lS9 - papyri 14. 26
Wendel14S - time of origin 27
Werner 227 (see al80 place value notation)
Wheeler. N.F. 96 zigzag functions 100. 111
Wheeler, R.E.M. 187 zodiac 5. 186 note 1
Winlock 94 - deeans 81
Woepcke 26 - in Egypt 82, 89
Woisin 24 - in India 166. 186 note 1
- time of orisin 102. 140. 170. 188
Ya'qub ibn Tlriq 180 - zero point 188. 207
Yale Babylonian Collection 35, 45 - zodiacal symbols 54. 67. 228
Yavana 167
yazdigerd. era 81 Y2 35, 47. 48. 50, 52, 148
year. definition 102. 140 Vi 47.52
- anomaUstie 113. 114. 140. 192 n 23. 46, 47. 51. 52. 78, 96, 180
 PLATE 1

"Septe mber". Book of Hours of the Duke of Berry.

 PLATE ~

P. Lund. Inv. ~5 H.
 PLATE a

B~I 85194 Ohv.

l' I.ATF.: 4

o
-
l-
I-
1:"1
PLATE 5
PLATE 6
 PLATE 7

n. Plimpton 322.

b. A 3412 Rev.
 PLATE R

rr ~.. -(fr~' «
(4:V
rr~ ~ «~ <nr t «<
<' ~
~ « .!~ ~
t?t ~
fr ««, ~~ «,.
<flY 4.lr
) ~ ~,.~, {nT~" ~
t:

~ <nr ~TT ~ J r- :;:~

«' (, .Ifr-~
t«.
ff ~'J! « 'f <rrr ~W c« < ~nr ~ ~:if
) ~ ff ~ :rEf , <nr «~, ~ ~ 0/ ~r ~ ~ ~
/

\, «fF ~r ~ rp «'Y « ,
~ ~rr ~ -f-
~ ~~ 1f ~ff ~ ~(.. (fff~'" «1f ~ VI rf- ~
ff ~ fF ~rrr w:vtf ~W <t:nF ~ W' W «T ~r-- «"
<'
c(

~ lr ~., ~« <fIf"trr W ~,<W itr ~

rr .(~W Pi= «i'( <fff «.r W fff <'« Jtf<~F
IT «~"
.(~, ~
~ , <nr , W 4.1W ~, «.W ~::;:f
If _(.ff W~, ~o/ ~ ~ ~f
.(~, (rr <~W ~ «fr '- ~'t\f If ~

h. Sp. 11,62 = BM 34589.

 PLATE 9

a. YBC 4712 Rev.

b. Sp. II, 62 = BM 34589.

 PLATE 10

"S
E
c:
Cl.I
r/)
-_.----- ......- _.----~ .....o
..0
E
o
E-c
Cl.I
:S
.5

-
'cJl
.5
'Q3
u
Cl.I
&
:5
IL- <
E
~ ~Ia: '. ~
o

~~1 ::~
~~j~j
 PLATE 11

From the Tomb of Ramses VII.

 PLATE 12

P. Cornell, Inv. 69.

PLATE r:i

"0
o
...0..
Il.I

-.~
 PLATE U

From Strassmaier's Notebook.

 DOVER SCIENCE BOOKS
THEORIES OF THE WORLD FROM ANTIQUITY TO THE COPERNICAN REVOLUTION,
Michael J. Crowe. (26173-5) $6.95
A SHORT HISTORY OF TECHNOLOGY: FROM THE EARLIEST TIMES TO A.D. 1900,
T.K. Derry and Trevor I. Williams. (27472-1) $17.95
FADS AND FALLACIES IN THE NAME OF SCIENCE, Martin Gardner. (20394-8)
$7.95
lJNDERSTANDING EINSTEINS' THEORIES OF RELATIVITY: MAN'S NEW PERSPECTIVE ON
THE COSMOS, Stan Gibilisco. (26659-1) $8.95
DE MAGNETE, William Gilbert. (26761-X) $11.95
GOSSAMER ODYSSEY: THE TRIUMPH OF HUMAN-POWERED FLIGHT, Morton
Grosser. (26645-1) $9.95
ON THE SENSATIONS OF TONE, Hermann Helmholtz. (60753-4) $15.95
THE STRANGE STORY OF THE QUANTUM, Banesh Hoffmann. (20518-5) $7.95
ALCHEMY, E.J. Holmyard. (26298-7) $8.95
THE SCIENCE OF SOAP FILMS AND SOAP BUBBLES, Cyril Isenberg. (26960-4)
$9.95
MENLO PARK REMINISCENCES, VOLUME ONE, Francis JehI. (26357-6) $13.95
FROM SUNDIALS TO ATOMIC CLOCKS: THEIR THEORY AND CONSTRUCTION, James
Jespers'en and Jane Fitz-Randolph. (24265-X) $8.95
ANCIENTS AND MODERNS, Richard Foster Jones. (24414-8) $7.95
ExPERIMENTS IN TOPOLOGY, Stephen Barr. (25933-1) $6.95
THE GREAT PHYSICISTS FROM GALILEO TO EINSTEIN, George Gamow. (25767-3)
$9.95
AN ILLUSTRATED GUIDE TO LINEAR PROGRAMMING, Saul I, Gass. (26258-8) $7.95
THE SCIENCE OF MEASUREMENT, Herbert Arthur Klein. (25839-4) $16.95
ExCURSIONS IN NUMBER THEORY, C. Stanley Ogilvy and John T. Anderson.
(25778-9) $5.95 .
THE ENJOYMENT OF MATHEMATICS, Hans Rademacher and Otto Toeplitz.
(26242-1) $7.95
DISCOVERING THE NATURAL LAws, Milton Rothman. (26178-6) $6.95
RELATIVITY IN ILLUSTRATIONS, Jacob 1: Schwartz. (25965-X) $7.95
THE HISTORICAL ROOTS OF ELEMENTARY MATHEMATICS, Lucas N.H. Bunt et al.
(25563-8) $8.95
THE ExACT SCIENCES IN ANTIQUITY, O. Neugebauer. (22332-9)
ENTERTAINING SCIENCE EXPERIMENTS WITH EVERYDAY OBJECTS, Martin Gardner.
(24201-3) $9.95
GREAT ExPERIMENTS IN PHYSICS, Morris H. Shamos. (25346-5) $10.95
ALL ABOUT LIGHTNING, Martin A. Uman. (25237-X) $6.95
THE COMPLEAT STRATEGYST: BEING A PRIMER ON THE THEORY OF GAMES OF
STRATEGY, J. D. Williams. (25101-2) $7.95
FAMOUS PROBLEMS OF GEOMETRY AND How TO SOLVE THEM, Benjamin Bold.
(24297-8) $4.95
SIDELIGHTS ON RELATIVITY, Albert Einstein. (24511-X) $3.95
(continued on back flap)
 "r<;

(continued from front flap)

THE GEOMETRY OF AKr AND LiFE, Matila Ghyka. (23542-4) $6.95

MAllIEMATICS IN 1lIE TIME OF 1lIE PHARAOHS, Richard J. Gillings. (24315-X)
$8.95
EssENTIALS FOR SClENTlflC AND TECHNICAL WRITERS, Hardy Hoover. (24060-6)
$7.95
PHYsiCS AND PHILOSOPHY, Sir James Jeans. (24117-3) $7.95
SCIENCE AND MUSIC, Sir James Jeans. (61964-8) $7.95
PRACTICAL STATISTICS SIMPLY ExPLAINED, Russell Langley. (22729-4) $9.95
VISUAL ILLUSIONS: THEIR CAUSE'S, CHARACTERISTICS AND APPUCATIONS, M.
Lucklesh. (2153Q-X) $6.95
PRELUDE TO MA1lIEMATlCS, W.W. Sawyer. (24401-6) $5.95
GREAT IDEAS OF MODERN MATHEMATICS: THEIR NATURE AND USE, JagJit Singh.
(20587-8) $8.95
WONDERS OF THE SKY, Frederick Schaaf. (24402-4) $8.95
How TO CALCULATE QUICKLY, Henry Sticker. (20295-X) $3.95
1001 QUESTIONS ANSWERED ABour EARTHQUAKES, AVALANCHE'S, FLoODS AND
OTHER NATURAL DISAsTERs, Barbara Tufty. (23646-3) $7.95
BASIC MACHINES AND How THEY WORK, U.S. Navy. (21709-4) $7.95
BASIC OPTICS AND OPTICAL INSTRUMENTS, U.S. Navy. (22291-8) $16.95
LADy luCK: THE THEORY OF PROBABILITY, Warren Weaver. (24342-7) $8.95
CRYSTALS AND LIGHT: AN INTRODUCTION TO OPTICAL CRYSTALLOGRAPHY,
Elizabeth A. Wood. (23431-2) $6.95
TEMPERATURE'S VERY Low AND VERY HIGH, Mark W. Zemansky. (24072-X)
$5.95
Ji*' How TO SOLVE MA1lIEMATiCAL PROBLEMS, Wayne A. Wlekelgren. (28433-6)
$7.95
LANDMARK ExPERIMENTS IN TwENTIETH CENTURY PHYSICS, George L Trigg.
(28526-X) $8.95
INTRODUCTION TO LoGIc AND TO THE METHODOLOGY OF DEDUCTIVE SCIENCE'S,
Alfred Tarskl. (28462-X) $8.95

........

Paperbound unless otherwise indicated. Prices subject to change with-

out notice. Available at your book dealer or write for free catalogues to
Dept. 23, Dover Publications, Inc., 31 East 2nd Street, Mineola, N.Y.
11501. Please indicate field of interest. Each year Dover publishes over
~- "'--./
( '"" I

200 books on flne art, musle, crafts and needlework, antiques, lan-
guages, literature, children's books, chess, cookery, nature, anthropol-
ogy, science, mathematics, and other areas.
Manufactured in the US.A.