JHA iv (1973), 1-12

THE MESOPOTAMIAN ORIGIN OF EARLY INDIAN

MATHEMATICAL ASTRONOMY

DAVID PINGREE, Brown University

In this paper I intend to advance and offer evidence in support of an hypothesis

concerning the dependence of the mathematical astronomy of the Jyotisaveddnga
on Mesopotamian science of the Achaemenid period.' I believe that the evidence
in support of the theory that some elements of early Indian astronomy are
derived from Mesopotamia is overwhelming, and that the evidence for the rest
of my hypothetical reconstruction is persuasive. But I must enter a cautionary
note with regard to that portion which relates to the Indian intercalation-cycle:
the evidence in both the cuneiform and the Sanskrit sources is so fragmentary
that no hypothetical reconstruction of the development or of the interrelation
of their respective intercalation-cycles is more than a reasonable guess. I hope
that the reader will find my guess more plausible than those of my predecessors.

Though the Vedas and Brahmanas provide us with some crude elements of
observational astronomy, such as the standard list of 27 or 28 naksatras or
constellations associated with the Moon's course through the sky, and some
rough parameters, such as the twelve months and 360 nychthemera of a year,
mathematical astronomy begins in India with a group of related texts which I
intend to explain in this paper. The basic text of this group is the Jyotisaveddnga/'
one of the six ailgas or "limbs" studied by Vedic priests; its purpose was to
provide them with a means of computing the times for which the performances
of sacrifices are prescribed, primarily new and full moons. This brief work has
come down to us in two recensions: a shorter one of 36 verses associated with
the Rgveda, and a longer one of 43 verses associated with the Yajurveda, which
latter incorporates 29 verses of the Rk-recension. That Rk-recension was com-
posed by one Lagadha, who is otherwise unknown, or, according to another
interpretation, by Suci on the basis of Lagadha's teachings; the Yajur-recension
names no author, but has the dubious benefit of a bhasya or commentary by
one Somakara. It is the Yajur-recension that has generally been used by modern
scholars also, as it, in two of its additional verses, attempts to adjust the older
system of the Rk-recension to the familiar terms of medieval Indian astronomy.
In this paper the shorter and surely older Rk-recension will be used.
We are justified in asserting the originality of the Rk-recension not only by
its shortness, but also by its parallelism to other pre-medieval Sanskrit texts.
In particular we must discuss here the following seven works in addition to
the two recensions of the Jyotisaveddnga:
1. The Arthasdstra of Kautilya" is an ancient work on political science. Many
scholars have identified the author with the minister of Candragupta Maurya,
who established the Mauryan Empire in northern India shortly before 300 B.C.,
though it seems fairly secure that our recension of Book Two of the Arthasdstra
does not antedate the second century A.D. 4 The twentieth chapter of the second
book of the Arthasdstra prescribes the duties of the Manadhyaksa or Super-

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

2 Journal for the History of Astronomy
intendant of Measurements, among which is included the duty of supervising
the measurements of time. These time-measurements are closely related to
those of the Jyotisavedanga.
2. The SiirduiakarViivadiina 5 is now the thirty-third story in a Buddhist collec-
tion of tales about Bodhisattvas, the Divydvaddna. Originally it was an anti-
caste tract in which a king of the Matangas (that is, Candalas), Trisanku, asks a
Brahmana, Puskarasarin, to give his daughter, Prakrti, to the outcaste's son,
SardiilakarQa. Upon the Brahmana's refusal of this unorthodox request,
Trisanku proves his status as a Brahmana by displaying his knowledge of astral
divination and astronomy. Our present Sanskrit version is full of interpolations,
both in prose and in poetry; but the history of the basic core of the text can
be traced back to translations into Chinese by the Parthian prince An Shih-kao,
who settled in Loyang in A.D. 148,6 by Chu Chiang-yen and Chih Ch'ien between
ca A.D. 220 and 252,7 and again by Dharmaraksa between A.D. 266 and 317. 8
The passages which are of interest to us as they reflect the astronomy of the
Jyotisaveddnga occur in the Chinese translations of the third century.
3. The twelfth chapter of the Paiicasiddhdntiktir composed by Varahamihira
at Ujjayini towards the middle of the sixth century A.D., summarizes a Paitdma-
hasiddhdnta whose epoch is 11 January of A.D. 80. This is our most important
witness for establishing the priority of the luni-solar period-relation of the
Rk-recension over that of the Yajur-recension; for, with the exception of an
alteration in the position of the winter solstice and Varahamihira's introduction
of the concept of omitted tithis, it is in complete agreement with the former.
4. The Vasisthasiddhdnta summarized in the second chapter and in verses 1-60
of the seventeenth chapter of the Paiicasiddhdntikd of Varaharnihira represents
an Indian adaptation of a Greek version of Babylonian astronomy intermingled
with elements of the older Indian borrowings from Babylonian astronomy
which appear in the Jyotisaveddnga and in the other texts of which we have
been speaking. Though the version of the Vasisthasiddhdnta available to
Varahamihira apparently used 3 December 499 as its epoch, the reference to
Vasistha as an astronomical authority in the Yavanajdtaka'" written by
Sphujidhvaja in A.D. 269/270 indicates that the original was probably composed
in the second or early third century A.D.
5. The Yavanajdtaka itself is primarily a versification of a Sanskrit prose
translation of a lost Greek astrological text; the Greek original was composed
in Egypt, and probably in Alexandria, in the first half of the second century
A.D., while the prose Sanskrit translation was made by Yavanesvara in western
India in A.D. 149/150. In the seventy-ninth chapter of the Yavanajdtaka,
Sphujidhvaja describes a Greek adaptation of the Mesopotamian linear
astronomy of the Seleucid period, but he intersperses in this material various
elements derived from the tradition of the Jyotisaveddnga.
6. The fifth, sixth, and seventh upangas of the Jaina canon of sacred literature
are in many respects virtually identical; they are entitled respectively the
Siiriyapannatti, the Jambuddivapannatti, and the Camdapannatti.P The recension
of the Jaina canon is traditionally assigned to the Council of Valabhi in
Saurastra, which met during the reign of the Maitraka monarch Dhruvasena I

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

 The Origin of Indian Astronomy 3
from ca A.D. 519 to 549. Though these texts contain much that is certainly
far older than the early sixth century A.D., their intercalation-cycle seems to
have been adapted to that of the Yajur-recension as they interpret the 366 days
of a solar year as civil rather than sidereal.P

7. Many verses ascribed to Garga are cited by Somakara in his bhasya on

the Yajur-recension, with whose system Garga is in accord as he knows most
of that recension's additional verses. Usually this Garga has been identified
with the author of the oldest form of the Gargasamhitd, which may be as old
as the first century A.D,13; but the slokas cited by Somakara are not found
in that work. Rather one must date this Garga to a considerably later period
as he demonstrates a knowledge of the astronomy of the early medieval period
-for example, of the four categories of time-measurements (solar, lunar, civil,
and naksatra) first enunciated in the Paitdmahasiddhdnta of the Visnudharmot-
tarapurdna, a work written in all probability in the early fifth century A.D.,
though they are already hinted at by Sphujidhvaja.P
Though the Yajur-recension is thus seen most likely to be a product of the
third to fifth centuries A.D., the Jyotisaveddnga has often been dated to the
twelfth century B.C. on the basis of the position among the naksatras that it
assigns to the winter solstice. Though this argument, as we shall see, is not
worth much, the system of the Rk-recension must indeed be much earlier than
that of the Yajur-recension since we find it in one text definitely dated in
A.D. 80 and in another which in part depends on sources that may go back
to the third century B.C. I hope now to show that in fact the astronomy of
the Rk-recension was formulated in the fifth or fourth century B.C. on the
basis of information about originally-Mesopotamian methods and parameters
transmitted to India during the Achaemenid occupation of the Indus Valley
between ca 513 and 326 B.C. This date well suits the Rk-recension, whose
language is definitely post-Vedic, but not yet classical.
The time-measurements of this earliest period of Indian mathematical
astronomy were made by means of two instruments; both of them were invented
in Mesopotamia. Furthermore, the mathematical device used for interpreting
them-a linear zig-zag function-is also Mesopotamian. The first of these
two instruments is an out-flowing water-clock. Its usage is described in the
first three padas of verse 7 of the Rk-recension:
The increase in daylight and the decrease in night-time in the northern
course (of the Sun) is a prastha of water, in the southern course it is the
reverse.
This means that one employs a linear zig-zag function with a constant difference
of one prastha a day to determine the amount of water to be poured into the
water-clock for each period of daylight, with a minimum amount at the
beginning of the Sun's northern course, that is at the winter solstice, and
the maximum amount at the summer solstice.
The escape-hole (nadika) of the water-clock (ghatika or kumbha) is des-
cribed in the Siirdu!akarViivadiina1S as a round tube of gold four digits in length
with a rectangular opening at the end. A similar description is given in the
Arthasastrav:

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

4 Journal for the History of Astronomy
a nalika [i.e., 24 minutes] is measured by an adhaka of water flowing through
a hole which is four digits long [extending] from a pot [kumbha]. [It is
composed of] four rnasakas of gold.

That there is this relationship between the liquid measure, adhaka, and the
temporal measure, kumbhika, is also stated -in verse 17 of the Rk-recension,
This out-flowing water-clock is also mentioned by Sphujidhvaja'? and by
Varahamihira."
In cuneiform sources we find a reference to the use of an out-flowing water-
clock in mul Apin," a series which was probably compiled in about 700 B.C.;
the earliest copy is dated 687 B.C. A number of other contemporary or even
older texts contain similar references, in particular the fourteenth tablet of
Eniima Anu Enlil. One mathematical text indicates that the shape of the
clepsydra was a cylinder as seems to be true of the Indian equivalent. The
water-clock itself, then, appears to be a Babylonian invention of the first few
centuries of the last millennium B.C. What of its manner of use in India?
The remainder of verse 7 in the Rk-recension is also connected with Meso-
potamia. It states simply: "Six muhurtas in an ayana." A muhiirta is a
thirtieth of a nychthemeron, or two nadikas, and an ayana is the time between
two consecutive solstices. So the statement tells us that the magnitude of
difference between the longest and shortest daylights is six muhiirtas or a fifth
of a nychthemeron. This, of course, means that the ratio of the longest to
the shortest day is 3: 2-a ratio inappropriate to all parts of India save the
extreme north-west, but one that is well-attested in cuneiform texts.
In fact the earliest ratio of longest to shortest daylight in Mesopotamia is 2 : 1,
found in the so-called Astrolabe P, in the fourteenth tablet of Eniana Anu Enlil
in mul Apin, and elsewhere in tablets dated before the seventh century B.C. 20
The ratio 3: 2 used by the Indians, however, was commonly utilized in all
Babylonian astronomical texts after ca 700 B.C. This tradition must surely be
the source of the Sanskrit texts under discussion, and provides us with a
terminus post quem for those texts.
But not only do our texts use the Babylonian ratio; they also employ a
linear zig-zag function to determine the lengths of daylight in intermediate
months. Thus the Arthasastraw gives the length of daylight as 15 rnuhurtas
at the equinoxes, with an increase or a decrease of one muhiirta a month as
one proceeds from an equinox toward respectively the summer or the winter
solstice; thereby the longest day becomes 18 muhurtas, the shortest 12. The
winter solstice is placed by the Arthasdstra here in Pausya, which is the month
preceding Magha wherein the winter solstice occurs according to verse 6 of
the Rk-recension; this is probably due to an attempt already discernible in
the Paitdmahasiddhdnta of A.D. 80 to have the vernal equinox fall in Caitra,
the third month after Pausya. Blsewhere'" the Arthasdstra places the beginnings
of the two ayanas in the same months as does the Jyotisaveddnga.
A linear zig-zag function for obtaining lengths of daylight identical to the
Arthasastra's is found in the Sdrdidakamdvaddnar" in the Yavanajatakar« and
in the Vasisthasiddhdnta. 25 This methodology is, of course, undeniably Baby-
Ionian; one need only multiply the Babylonian beru or double-hours by 2;30
to produce the Indian table of muhurtas.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

 The Origin of Indian Astronomy 5
The same zig-zag function is expressed in verse 22 of the Rk-recension,
though at intervals of days rather than months.
Whatever has elapsed of the northern ayana or whatever remains of
the southern ayana, multiply it by 2, divide [the product] by 61, and add
[to the quotient] 12; [the sum] is the length of daylight [expressed in
muhiirtas].
This formula is based on three parameters-12 muhiirtas as the length of
daylight at the winter solstice, 183 "days" as the interval between two consecu-
tive solstices, and 6 muhiirtas as the difference between the longest and shortest
lengths of daylight. Then, if the time from the winter solstice in "days" is
denoted x and the length of daylight on that "day" as d(x), it is assumed that
6x 2x
d(x) = 12 + 183 = 12 + 61'
The second instrument used for measuring time in this period, though not
mentioned in the Jyotisaveddnga, is the gnomon. In later Sanskrit texts it is
called sanku, but in the earliest specimens the gnomon-shadow was named
chayapaurusa, so the gnomon itself must have been called purusa or man;
its length was normally defined as 12 angulas or digits. One of the oldest
gnomon-texts is that preserved in the Arthasdstra.r: Here it is assumed that
the noon-shadow is 0 in A~ii9ha-that is, at the summer solstice if the vernal
equinox falls in Caitra as the preceding section of the Arthasdstra prescribes.
This indicates a terrestrial latitude of about 24°-that is, approximately the
latitude of Ujjayini-whereas the use of 3 : 2 as the ratio of longest to shortest
length of daylight indicates a terrestrial latitude of about 35°-that is, somewhat
north of the latitude of Babylon. But though the length of the noon-shadow
for the summer solstice has been adjusted to Indian conditions, the Arthastistra
still employs a Babylonian linear zig-zag function to obtain the lengths of the
noon-shadows in other months of the year. It is assumed that the length of
the noon-shadow at the winter solstice is 12 digits, so that the monthly increment
or decrease in the length of the noon-shadow is 2 digits and its length at the
equinoxes is 6 digits. Precisely this same linear zig-zag function is found in the
Vasisthasiddhiinta summarized by Varahamihira," though with the months
replaced by the zodiacal signs occupied by the Sun.
In this system the noon-shadow at the winter solstice measures one gnomon-
length. Exactly the same noon-shadow at the winter solstice is given in mul
Apin,28 though that text does not yield a value for the length of the noon-
shadow at the summer solstice. Rather it forms a linear zig-zag function for
the seasonal hours after sunrise at which the length of the shadow equals one
gnomon-length; these values are 3 beru or 90 us at the winter solstice, 2 beru
or 60 us at the summer solstice, and 2;30 beru or 75 us at the equinoxes.
This implies a monthly difference of 5 us.
The remainder of the gnomon-text in the Arthasdstra gives shadow-lengths
at the passing of various fractions of the day of the summer solstice; pre-
sumably simple proportion is to be used to find the shadow-lengths on other
days. A similar, though corrupt and fragmentary, table of shadow-lengths at
the passing of each muhiirta of the day of the summer stolstice is found in
the SiirdUlakart:ziivadiina,29 and another in the Pahlavi Shdyast ne shdyast'"; and

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

6 Journal for the History of Astronomy
mul Apin also preserves such tables in which a number of seasonal-hours is
associated with each shadow-length expressed in integer gnomon-lengths.
Though the methods of computation of these several tables differ, it is note-
worthy that the shadow is assumed in mul Apin to equal 8 gnomon-lengths
after 1- seasonal hour on the day of the summer solstice or 45 minutes of an
equinoctial hour, whereas both the Arthasdstra and the Siirdulakarniivadiina state
that the shadow equals 8 gnomon-lengths after 1 muhiirta or 48 minutes of
an equinoctial hour. It seems plausible, then, that the Sanskrit texts represent
an adaptation of a lost Mesopotamian scheme reduced to expression in Indian
units of time-measurement, and with further modifications to produce a
shadow-length of 0 at noon.
The Siirdi:tlakanJiivadiina31 preserves a fragment of another early gnomon-
text in which the gnomon measures 16 digits and the longest length of daylight,
18 muhiirtas, occurs in Sravana as in the Jyotisaveddnga. This scheme may
well antedate the one discussed above. It gave the lengths of the noon-shadows
for each month in a linear zig-zag system almost identical with that of the
Arthasdstra, though the latter used a 12-digit gnomon; the one variation in
the SiirdUlakanJiivadiina is that the noon-shadow at the summer solstice is half
a digit rather than zero. This shows that in this scheme it was not assumed
that the situation at Ujjayini can be generalized for all of India. But it would
be unrealistic to think that one could compute the precise latitude for which
this shadow-table would work; it should suffice to say that it must have
originated in northern India.
From this discussion it should be clear that the mathematical astronomy of
this group of texts centering about the Jyotisaveddnga has borrowed much
from Mesopotamian astronomy of the seventh and sixth centuries B.C., and
is especially close to the compilation mul Apin, though, of course, the Indians
have altered certain elements to suit their own situation. We now must look
at the more important and troublesome part of the Jyotisaveddnga, its inter-
calation-cycle.
In the Rk-recension the elements of this yuga or cycle are not clearly stated,
but can easily be inferred. It is stated in verse 32 that the yuga consists of
five years-a choice perhaps determined by the lustrum mentioned in some
Vedic texts, though periods of two, three, four, and six years are also attested
in the literature." These five years are solar, and therefore contain 60 solar
months. That they also contain 2 intercalary months to make a total of
62 synodic months is clear from verse 4, which gives the following rule for
finding the parvan or syzygy.
Diminish [the current year-number] by 1, multiply [the remainder] by 12
[to get the solar months], multiply [the product] by 2 [to get the "syzygies"
of the solar months], add [to the product] the lapsed [syzygies of the
present year], and add [to that sum] 2 for each 60; [the result] is called
the sum of the syzygies.
The addition of 2 syzygies for each 60 "syzygies" of the solar months is, of
course, equivalent to adding 2 intercalary months for each 60 solar months. This
same intercalation-rule is found in the Arthasiistra,33 the SiirdUlakanJiivadiina,34
and the Paitdmahasiddhdnta summarized by Varahamihira."

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

 The Origin of Indian Astronomy 7
The number of "days" within this five-year yuga can be computed from
verse 18 of the Rk-recension:
The Moon travels through a naksatra in 1 day and 7 kalas, the Sun
in 13 and 5/9 days.
The Moon's total travel in 5 solar years containing 62 synodic months is
67 sidereal rotations; since in each sidereal rotation it travels by definition
through 27 naksatras, in a yuga it travels through 1809 naksatras, From
verse 16 we learn that each day contains 603 kalas, so that to find the number
of days, D, in a yuga we form the equation:
7
D = 1809
1 = 1830.X
603
Similarly, the Sun travels through 135 naksatras In 5 years at the rate of
13~ days per naksatra. And therefore
5
D = 135 X 13 = 1830.
9
A cruder parameter for solar motion-e-D] days per naksatra-s-is found in
verse 12; it yields an impossible 1845 days in a five-year yuga.
If there are 1830 "days" in 5 solar years, each one contains 366 "days" and
each ayana contains 183 "days", a parameter confirmed by the rule for the
daily increase in the length of daylight discussed above. Furthermore, in
62 months there must be 1860 tithis or thirtieths of a month; this again is
confirmed by verse 8 of the Rk-recension,
They say that the first [tithis] in [each] ayana [are] the first, seventh,
and thirteenth [in the suklapaksa and] the fourth and tenth [in the
krsnapaksa]; [the series is taken] twice.
Since 1860 tithis contain 10 ayanas, each ayana is 186 tithis or 6 synodic months
plus 6 tithis; and the constant difference in the above verse is precisely 6 tithis.
The same scheme is used for determining the tithi-numbers of the equinoxes
in verse 12.
Precisely these figures are known to lie behind the computations of the
Paiuimahasiddhiinta of A.D. 80. Neugebauer and the present author have shown
that they are to be interpreted to mean that each year contains 366 sidereal
days, which are equivalent to 365 civil days. In other words, the fundamental
relation of the five-year yuga is:
5 years = 1830 sidereal days = 62 synodic months.
This manner of expression is quite consistent with the normal practice of the
earliest texts of medieval Indian astronomy, for both the Paitdmahasiddhdnta
of the Visnudharmottarapurdna'" and the Aryabha{iya37 composed by Aryabhata
in A.D. 499 express the length of their yugas in naksatra or sidereal days, from
which the civil days can be derived by subtracting the number of revolutions
of the Sun in a yuga; and Sphujidhvaja stated the correct relation already in
the third century." One of the principal problems that modern scholars have
had with the Jyotisaveddnga is due to their understanding the text to mean that
a solar year equals 366 civil days, which is a parameter approximately three
times as inaccurate as that of the Rk-recension.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

8 Journal for the History of Astronomy
Their justification for understanding the text in this way is to be found in
verses 28-29 of the Yajur-recension-two verses not to be found in the
Rk-recension. These verses state:
A year [consists of] 366 days, 6 rtus [or seasons], 2 ayanas, and 12 solar
months; a yuga is this multiplied by 5. The sum of the days plus 5 [equals]
the risings of [the naksatra] Vasava [or Dhanistha]; [the sum of days]
diminished by 62 [equals the risings] of the Moon; [and the sum of days
diminished] by 21 [equals the Moon's transits] of the naksatras,
Since we know that the Moon travels through 1809 naksatras in a five-year
yuga, "the sum of days" in these verses is 1830, that is, the number of sidereal
days in our interpretation. But these verses also state that the number of
risings of a naksatra-s-that is.uhe number of sidereal days-equals "the sum
of days plus 5" or 1830 plus the number of revolutions of the Sun. Clearly,
then, the author of these two verses assumed that 1830 is the number of civil
days in a five-year yuga, and both medieval and modern scholars have followed
him. But I would suggest that the redactor of the Yajur-recension, in perhaps
the fourth or fifth century A.D., added these verses in order to provide a state-
ment about the fundamental period-relations of the Jyotisaveddnga including
that of sidereal to civil days; unfortunately, he did not correctly apprehend
what that relation was. As we have shown, the Rk-recension of the
Jyotisaveddnga and the Paitdmahasiddhdnta of A.D. 80 do not impose this
false interpretation on us.
Given this yuga of 5 years, of which each contains 365 civil days, and during
which period 2 intercalations are made, we must necessarily ask whether the
Indians developed it themselves or adopted it from some external source.
Though longitudes in this cycle are measured in Indian units-the 27 naksatras
now for the first time interpreted as equal arcs of 13 ;20 0 each along the ecliptic
-there is nothing in the earlier Indian texts to suggest any prior stages of
development that could have culminated in the Jyotisaveddnga. On the contrary,
not only are the time-measuring instruments, the water-clock and the gnomon,
along with the linear zig-zag functions utilized in reading them, derived from
Babylonia; some of the elements of the intercalation-cycle itself are also,
beginning with the division of the mean synodic month into thirty equal units
called in Sanskrit tithis (these units appear in a cuneiform tablet of about
600 B.C. which deals with the dates of solstices, equinoxes, and the heliacal
risings and settings of Sirius"), the thirty muhurtas of a nychthemeron being
an obvious imitation of the thirty tithis of a synodic month. Furthermore,
a year of 365 days, though Egyptian in origin, may have been adapted in Iran
as early as the second half of the fifth century B.C. 40 This possibility provides
us with a hint regarding the intermediary which transmitted so much of
Mesopotamian astronomy to the banks of the Indus and Ganges.
If we look in cuneiform texts for a five-year intercalation-cycle, we are
frustrated. But a knowledge of the basic period-relations of lunar and solar
motion are implicit in the Babylonian eighteen-year eclipse cycle, whose ante-
cedents go back to the eighth century B.C.4l And, as I have learned recently
from Dr Herman Hunger of the Oriental Institute, the second tablet of mul
Apin describes a crude intercalation-cycle in which one month is intercalated

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

 The Origin of Indian Astronomy 9
every three years. The epact in this text is 10 days; somewhat later texts of
the Achaemenid period use 10;30 days. Moreover, if legend as preserved by
Censorinus is correct, Cleostratus of Tenedos towards the end of the sixth
century B.C. had learned-inevitably from a Mesopotamian source-of an
eight-year cycle with three intercalary months.v and some claim, though on the
basis of not completely convincing evidence, that this octaeteris was used in the
Babylonian civil calendar from 527 to 503 B.C. 43 In 432 B.C. Meton in Athens
proposed the nineteen-year cycle with seven intercalations, which is based on
very fine parameters indeed; and Meton may have had Babylonian sources.
The Mesopotamian evidence consists of the actual intercalations in the civil
calendar, which are regular with respect to the year of intercalation after
483 B.C. with the exception of one anomalous intercalation in 384 B.C.,44 and
it must be remembered that the civil authorities may well have lagged far
behind the astronomers in adopting this intercalation-scheme. In Athens itself
the Metonic cycle was never employed in the civil calendar. It is not clear
what relationship the Egyptian 25-year cycle, introduced in the middle of the
fourth century B.C. during Achaemenid rule, might have to the Indian. For
the Egyptians 25 years of 365 days each contain 309 synodic months.v whereas
for the Indians 25 such years contain 310 synodic months. The contemporaneity
of these developments in areas linked by the Achaemenid empire suggests a
common origin.
It is clear, then, that Babylonian astronomers were capable of devising
intercalation-cycles in the seventh, sixth, and fifth centuries B.C., and there
is evidence both in the Greek and in the cuneiform sources that they actually
did so; and by the early fourth century B.C. they had certainly adopted the
quite-accurate nineteen-year cycle. It is my suggestion that some knowledge
of these attempts reached India, along with the specific astronomical material
we have already discussed, in the fifth or fourth century B.C. through Iranian
intermediaries, whose influence is probably discernible in the year-length
selected by Lagadha for the Jyotisaveddnga. But the actual length of the yuga,
five years, was presumably accepted by Lagadha because of its identity with a
Vedic lustrum. Not having access to a series of extensive observations such
as were available to the Babylonians, he probably was not completely aware
of the crudeness of his system. And the acceptance of this cycle by Indians
for a period of six or seven centuries or even more demonstrates among other
things that they were not interested in performing the simplest acts of observa-
tional astronomy.
This leads us directly back to the question of the date of the Jyotisaveddnga
which we briefly touched upon at the beginning of this paper. Verses 5-6 of
the Rk-recension may be translated as follows.
When the Moon and Sun travel in the heavens together with Vasava
[i.e., the naksatra Dhanistha], then is the beginning of the yuga, the
suklapaksa of Magha, and the northern ayana. The Sun and Moon
proceed northwards at the beginning of Sravi~!ha [i.e., Dhanistha], and
southwards in the middle of Sarpa [i.e., Asle~a]; [these two events occur]
always in [the month] Magha and Sravana respectively.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

10 Journal for the History of Astronomy
0
Since the tropical longitude of the beginning of Dhanistha was 293;20 in about
A.D. 500, that point in the sphere of the fixed stars coincided with the winter
solstice in about 1180 B.C. But, though this computation is correct, its relevance
to the date of the Jyotisaveddnga is not evident. We simply do not know where
Lagadha would have placed the beginning of the equal naksatra Dhanistha
with respect to the fixed stars, nor do we know the accuracy with which he
could have determined the sidereal longitude of the Sun at the winter solstice.
Since a displacement of the beginning of the equal naksatra by some 100, or
an error of 10 days in computing the date of the winter solstice, or some
combination of these two effects is all that is required to bring the date from
the twelfth century to the fifth century B.C., we should not lend much weight
to this chronological argument." Furthermore, we must consider that an
intercalation-cycle based on such a crude parameter for the length of a year
as 365 days does not inspire much confidence in the accuracy with which its
author was able or wished to endow the positions of the solstices. The scheme
could never pretend to provide more than a very rough approximation to
reality, for a few years; it would require constant readjustment if it were really
to work. The true significance lies in the fact that with it the Indians accepted
in theory the possibility of describing the periodicity of celestial motions in
mathematical terms, and that they learned of this possibility from the Baby-
lonians through their Iranian neighbours.
But there is one further question that we must raise before accepting this
hypothesis of transmission. Was this an isolated phenomena, or part of a
general Iranian influence on Indian culture in the fifth and fourth centuries B.C. ?
Unfortunately, our answer to that question is rather clouded by the scarcity
of literary or archaeological data from the period in question." We do not
know how far into India the Achaemenids penetrated, but probably their
control did not extend beyond the western parts of the Panjab as Alexander
met numerous small and apparently independent states in the Indus valley.
But Iranian influence in the early fifth century was sufficiently strong to make
possible the safe completion of Scylax's exploratory voyage down the Indus,
and Taksasila, in the region where Panini seems to have worked, was certainly
a city where cross-cultural contacts were frequent. And it is arguable that the
enormous and often-studied Iranian influence discerned in Mauryan polity,
architecture, sculpture, epigraphy, and the like in the third century B.C. was
an inheritance from the pre-Mauryan Nandas' rather than from the post-
Alexandrian Greeks' adaptations of Achaemenid forms. And parallel to the
suggested Mesopotamian-Iranian influence on Indian mathematical astronomy
is the influence of the same cultural complex on Indian omens, which first are
mentioned in the Upanisads and Buddhist canonical texts of the pre-Mauryan
period, though the oldest codifications of these omens available to us were
compiled in the early centuries A.D. It is reasonable, then, or at least so I
believe, to see the origins of mathematical astronomy in India as just one
element in a general transmission of Mesopotamian-Iranian cultural forms to
northern India during the two centuries that antedated Alexander's conquest
of the Achaemenid empire.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

 The Origin of Indian Astronomy 11
REFERENCES

1. This relationship I originally proposed in Isis, liv (1963), 231-3.

2. Both recensions along with Somakara's commentary were first edited by A. Weber, Ober
den Vedakalender, namens Jyotisham (Berlin, 1862) (the text of one recension had been
published at Bombay in 1833); the text of the Yajur-recension along with the non-Yajur
verses of the Rk-recension were reprinted by G. Thibaut, "Contributions to the Explanation
of the Jyotisha-Vedanga", Journal of the Asiatic Society of Bengal, xlvi (1877), 411-37,
esp. 413-6; the Rk-recension was reprinted by J. B. Modaka of Thana in 1885; one recension
was included in Vedasya sadanga (Bombay, 1892); both recensions were edited and com-
mented on by Bfrhaspatya (Lala Chotelal) in the Hindustan review for 1907 (reprinted
Allahabad, 1960) and again, with Sornakara's bhasya, by S. Dvivedin (Benares, 1908). The
most recent attempt at an edition and at an understanding of both texts is by R. Shamasastry
(Mysore, 1936). The more recent efforts by B. R. Kulkarni, The Lagna System ofthe Vedanga
Jyotisha (Dhulia, 1943); by G. Prasad, "The Astronomy of the Vedanga Jyotisa", Journal
of the Ganganath Jha Research Institute, iv (1946--47), 239--48; and by A. K. Chakravarty,
"The Working Principle of the Vedanga Jyotisa Calendar", Indian studies past and present,
x (1968-69), 31--42, as well as the older work by B. G. Tilak, Vedic chronology and Vedanga
Jyotisha (Poona, 1925), do not advance our understanding of this text.
3. have used the recent edition with translation and commentary by R. P. Kangle (3 vols,
Bombay, 1960-65).
4. T. R. Trautmann, Kautilya and the Arthasdstra (Leiden, 1971), 176-84.
5. I have used the edition by S. Mukhopadhyaya (Santiniketan, 1954); this is reprinted in P. L.
Vaidya's edition of the Divyiivadiina, Buddhist Sanskrit texts, xx (Darbhanga, 1959),314-425.
6. E. ZUrcher, The Buddhist conquest of China (2 vols, Leiden, 1959), i, 32--4.
7. ZUrcher, op. cit., i, 47-51.
8. Ibid., 65-71.
9. Edited, translated, and commented on by O. Neugebauer and D. Pingree (2 vols, Copenhagen,
1970-71).
10. Edition, translation, and commentary by D. Pingree to appear in the Harvard oriental series;
the reference is to 79,3.
11. These works have been printed several times by the Jainas in India; I have used the edition
of the first two by J. F. Kohl (Stuttgart, 1937). See also G. F. Thibaut, "On the Surya-
prajfiapti", Journal of the Asiatic Society of Bengal, xlix (1880), 107-27 and 181-206.
12. Suriyapannatti, I, 3, and XX, 59.
13. On the various Gargas see D. Pingree, Census of the exact sciences in Sanskrit, Series A, ii
(Philadelphia, 1971), 115b-126a.
14. For the Paitiimahaslddhiinta, see D. Pingree, "The Paltdmahasiddhiinta of the Visnudharmottara-
purdna"; Brahmavidyii, xxxi-xxxii (1967-68), 472-510. Compare Garga as cited by Somakara
on Yajur-recension II and Paitamahasiddhiinta, III, I; cf. Yavanajdtaka, 79, 6.
15. Pp. 57-8.
16. II, 20, 35.
17. Yavanajdtaka, 79, 27; this could be interpreted as either an out-flowing or an in-flowing clock.
18. Paiicasiddhiintikd, 14, 31-2.
19. See O. Neugebauer, "The Water Clock in Babylonian Astronomy", Isis, xxxvii (1947), 37-43.
20. See, e.g., B. L. van der Waerden, "The Earliest Astronomical Computations", Journal of Near
Eastern studies, x (1951), 20-34.
21. II, 20, 37-8.
22. 11,20, 61-64.
23. Pp. 53 and 100-8.
24. 79, 26 and 31.
25. Pahcasiddhiintikd, 2, 8.
26. II, 20, 39-42.
27. Paiicasiddhiintikd, 2, 9-10.
28. E. F. Weidner, "Ein babylonisches Kompendium der Himmelskunde", American journal of
Semitic languages and literatures, xl (1924), 186-208, esp, 198.
29. Pp. 54-5.
30. Translated by E. W. West, Sacred books of the East, v (Oxford, 1880),397--400.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015

12 Journalfor the History of Astronomy
31. Pp. 100-3.
32. A. A. Macdonell and A. B. Keith, Vedic index (2 vols, London, 1912), ii, 412; for a fifth-century
reference to the five-year cycle see Paitiimahasiddhanta, IV, 4.
33. II, 20, 65-6.
34. Pp. 103-4.
35. Paiicasiddhantika, 12, I.
36. III, 32.
37. Dasagitikd, I.
38. Yavanafiitaka, 79, 8.
39. Published by O. Neugebauer and A. Sachs, "Some Atypical Astronomical Cuneiform Texts. I",
Journal of cuneiform studies, xxi (1967), 183-218, esp. 183-90; for the use of tithis in this
text see 189-90.
40. So, for instance, argues S. H. Taqizadeh, Old Iranian calendars (London, 1938), on not entirely
convincing evidence.
41. A. J. Sachs, Late Babylonian astronomical and related texts (Providence, 1955), p. xxxi, A
crude relation
223 months ss 18 solar rotations + 10;30°
is found in a "saros"·text of the early fifth century B.C.; see A. Aaboe and A. Sachs, "Two
Lunar Texts of the Achaemenid Period from Babylon", Centaurus, xiv (1969), 1-22, esp, 18.
In a later "saros"-tablet there occurs a column X which registers the difference between
a year and 12 months; see A. Aaboe, Some lunar auxiliary tables and related texts from the
Late Babylonian Period (Copenhagen, 1968), 28-30.
42. De die natali 18,5; H. Diels and W. Kranz, Die Fragmenteder Vorsokratiker, I' (Berlin, 1954),42.
43. B. L. van der Waerden, Die Anfiinge der Astronomie (Groningen [NDD, 112.
44. See R. A. Parker and W. H. Dubberstein, Babylonian chronology (Providence, 1956), 6.
45. R. A. Parker, The calendars of Ancient Egypt (Chicago, 1950).
46. On the difficulties involved in accepting the precision of the statements in the Jyotisaveddnga
required for their use as chronological indicators see W. D. Whitney, "On the Jyotisha
Observation of the Place of the Colures, and the Date derivable from it", Journal of the
Royal Asiatic Society, n.s, i (1865), 316-31; cf. also D. Pingree, "Precession and Trepidation
in Indian Astronomy before A.D. 1200", Journal for the history of astronomy, iii (1972),
27-35.
47. Most of the literary evidence is discussed in S. Chattopadhyaya, The Achaemenids in India
(Calcutta, 1950); ct. also R. A. Jairazbhoy, Foreign influence in Ancient India (London,
1963), 38-47. For the archaeological evidence see R. E. M. Wheeler, "Iran and India in
Pre-Islamic Times", Ancient India, iv (1948), 85--103, esp. 92 sqq.

Downloaded from jha.sagepub.com at INDIANA UNIV on July 13, 2015