Probably the best-known story about Indian (actually Tamil21) astronomy is that told by Warren (1825), recounted in Neugebauer (1952; 1983a, p. 435) about a "kalendar maker" in Pondicherry, who demonstrated to Warren a technique to predict a lunar eclipse by means of memorized tables and the movement of shells on the ground. Neugebauer cites this as an example of a continuous tradition stemming from the 6th century (with Varahamihira) in India, back through the 3rd-century Roman empire, and ultimately to Seleucid era cuneiform tablets, no later than the 2nd century b.c. Apparently, the key to the tables lay in word association, because Warren writes about "certain artificial words and syllables" being used, and "... he did not understand a word of the theories of Hindu astronomy, but was endowed with a retentive memory, which enabled him to arrange very distinctly his operations in his mind, and on the ground." The demonstration consisted of the computation of the circumstances of the eclipse of 1825, from May 31 to June 1, with a surprising degree of accuracy: Differences from the actual moments of beginning, middle, and end of the eclipse were +4m, -23m, and -52m, respectively. The predictability of the recurrence of eclipses in this way implies a
19 Note the curious inverse relation between the date of the work and the dating of the Mahabharata. For a different view, the Tibetans held that Nyatri Tsepo, defeated in the Mahabharata war, fled to Tibet and became the first king there about 127 b.c. (Bryant 1992, p. 78).
20 The closest relevant eclipse from Oppolzer is 19 Aug. 1157 b.c. at JDN 1299060. Stephenson and Houlden (1986) show a partial eclipse track through SE Asia (eastward from southeastern Burma) for 12 Feb. 1156 b.c. (JDN 1299237).
21 A Dravidian-speaking group of Southern India and northern Sri Lanka.
knowledge of the motions of the Moon—synodic, nodal, and perhaps anomalistic—which we discussed in §2.3.4 and §2.3.5. Neugebauer documents the connections from Babylonia to India, citing (1983a, p. 435) masses of lunar and planetary theory in the Panca Siddhantika, with close parallels in Babylonian texts from the Hellenistic period. Thus, the bases are ultimately observational, but how could such material be conveniently coded to permit a rote-learned technique to predict a lunar eclipse? Both Warren and Neugebauer attempt an explanation of the particular method used by the Tamil informant.
The use of the shells was illustrated by Warren as follows. The number 248 was given the simple representation whereas the sum of five zodiacal signs (of 30° each) and 29°58'13" was denoted by
which can be understood by writing the count of these shells exactly as given:
Was this article helpful?