¹ Research in Madurai was carried out in 1976–7 for a period of 12 months, and was financed by a research grant from the Social Science Research Council, whom I would like to thank. I also thank Ms Penny Logan for supplying most of my data on domestic ritual and for her comments on this paper.

² Filliozat in Renou, L. and Filliozat, J., L'Inde classique, vol. 2, Hanoi, 1953, 720–38;Google ScholarKane, P. V., History of Dharmaśāstra, V, Part 1, Poona 1962, section 2;Google ScholarBasham, A. L., The wonder that was India, third edition, London, 1971, 491–7;Google ScholarSewell, R. and Dikshit, S. B., The Indian calendar, London, 1896;Google ScholarPillai, L. D. Swamikannu, Indian chronology, Madras, 1911Google Scholar and Panchang and horoscope, Madras, 1925.Google Scholar Pillai is particularly useful on the Tamil system, which is summarised (using Tamil terms only) by Arden, A. H., A progressive grammar of the Tamil language, fifth edition, Madras, 1969, 316–9.Google Scholar C. G. Diehl discusses the almanacs but does not explain clearly how to consult them in Instrument and purpose; studies on rites and rituals in South India, Lund, 1956, 200–11.Google Scholar Two anthropologists have discussed the calendrical system of a North Indian village: R. S., and Freed, S. A., “Calendars, ceremonies and festivals in a North Indian village: necessary calendrical information for fieldwork”, Southwestern Journal of Anthropology, XX, 1964, 67–90.Google Scholar

³ Nearly all these terms are Sanskrit words, but Tamil orthography makes them hard to recognise. To reduce this difficulty as far as possible, I have sometimes deviated from “strict” Tamil spelling and employed the grantha letters often used in modern Tamil; e.g. rāci for irāci, nakṣattiram for naṭcattiram.

⁴ The Western astronomical names of these constellations are in Renou, and Filliozat, , op. cit., 729–30Google Scholar and Basham, , op. cit., 492.Google Scholar

⁵ For detailed discussion of this problem, see Sewell and Dikshit, op. cit. and Sewell, R., Indian chronology, London, 1912.Google Scholar

⁶ The sidereal lunar month slightly exceeds 27 days and a twenty-eighth intercalary nakṣattiram has been used by the astronomers to correct the error (see Renou, and Filliozat, , op. cit., 185, 721Google Scholar and Basham, , op. cit., 492). However, modern practical astronomy does not use this extra nakṣattiram, named Skt. abhijit.Google Scholar

⁷ See Renou, and Filliozat, , op. cit., 722;Google ScholarBasham, , op. cit., 494; Government of India Rashtriya Panchang, the official government almanac published annually in English and all the major Indian languages.Google Scholar

⁸ In order, the 27 yokams corresponding to the nakṣattirams are (Sanskrit names in parentheses): 1. viṣkampam (viṣkambha), l. pirīti (prīti), 3. āyuṣmān (āyuṣmant), 4. saupākyam (saubhāgya), 5. (śobhana), 6. atikaṇṭam (atigaṇḍa), 7. cukarumam (sukarman), 8. tiruti (dhṛti), 9. cūlam (śūla), 10. kaṇṭam (gaṇḍa), 11. virutti (vṛddhi), 12. turuvam (dhruva), 13. viyākātam (vyāghāta), 14. harṣaṇam (harṣaṇa), 15. vajram (vajra), 16. citti (siddhi), 17. viyati (vyatīpāta), 18. (varīyas), 19. parīkam (parigha), 20. civam (śiva), 21. cittam (siddha), 22. cāttiyam (sādhya), 23. cupam (śubha), 24. cuppiram (śukla), 25. pirāmyam (brahman), 26. mahentiram (indra), 27. vaitiruti (vaidhṛti). The 11 karaṇams are (Sanskrit names in parentheses): 1. (kiṃstughna), 2. pavam (bava), 3. pālavam (bālava), 4. kaulavam (kaulava), 5. taitulam (taitila), 6. karacai (gara), 7. vaṇacai (vanij), 8. pattirai (viṣṭi), 9. (śakuni), 10. catuṣpātam (catuṣpada), 11. nākavam (nāga). The 60 karaṇams of the lunar month ending on New-moon day start with followed by eight cycles of the seven pavam to pattirai, and closing with the last three karaṇams. The almanac daily entry normally records only the karaṇam current at sunrise (with its ending time) and not the subsequent one, which will also end within the same day. I understand that only the highly inauspicious pattirai (viṣṭi) is of much significance. Many explanations of the karaṇams are incomprehensible; I have relied on Renou, and Filliozat, , op. cit., 722, 734.Google Scholar

⁹ The Tamil almanacs I have used are the Vākkiya Pañcāhkams by K. Muttukkiruṣṇayyankār, published in Madurai, for the years 1976–7, 1977–8 and 1978–9. Other almanacs contain almost exactly the same information. Sewell, and Dikshit, , op. cit., 14–15, reproduce a (translated) page from a Bombay almanac, which closely resembles a modern Tamil one.Google Scholar

¹⁰ Each day's equivalent date in the Gregorian (Western Christian), Indian National and Muslim calendars is also given. The Gregorian calendar is generally used in daily life in Tamilnadu; The Tamils call it the “English” (iṅkilīṣ) calendar. The Indian National calendar (Saka era), introduced by the Government of India in 1957, is only used by government offices in Tamilnadu. The Rashtriya Panchang is based on this latter calendar. Hindus, of course, never use the Muslim calendar.

¹¹ See Brunner-Lachaux, H., Somaśambhupaddhati, pt. 2, Pondicherry, 1968, 6, 8, on confusions caused by reading lunar dates as solar dates in the ancient ritual texts known as the Āgamas. These confusions also occur in some of the books published in Tamil explaining how to conduct the rituals at domestic festivals.Google Scholar

¹² For further details of the complexities of titis, see Sewell, and Dikshit, , op. cit., 17–18Google Scholar and Pillai, , Ind. chron., 46.Google Scholar If the same titi of the same fortnight occurs twice within one solar month, it is a (Skt. śūnya) titi on its first occurrence, and it is inauspicious, as is a day on which two titis end or no titi ends. Various other titis, on which Vedic study is prescribed (see Kane, P. V., History of Dharmaśāstra, II, Poona, 1941, 395–6) and which are inauspicious, are marked in the almanacs as well.Google Scholar

¹³ The lunar months are named from the nakṣattirams traditionally linked to their Full-moons, but some of these nakṣattirams differ from those specified in the Tamil tradition. Some Tamil solar months have names similar to their approximate lunar equivalents. It may also be noted here that 13 sidereal lunar months (i.e. cycles of nakṣattirams) last almost the same time as 12 synodical lunar months (i.e. cycles of titis).

¹⁴ The Rashtriya Panchang gives these data for the Central Station of India, where local mean time is Indian Standard Time; a table of correction times for all major centres in India is provided.

¹⁵ Some of this information, including the predictive data, is discussed by Diehl, , op. cit., 200–11.Google Scholar

¹⁶ The Tamil names of these 60 years are printed in some almanacs; they are also in Arden, , op. cit., 318.Google Scholar The Sanskrit names are in Renou, and Filliozat, , op. cit., 735–6.Google Scholar Fora Tamil man whose wife is alive, the 60th birthday is specially important, for then he has lived through the entire cycle and also half-way through the cumulative cycle of all the planets, believed to total 120 years and to be the allotted span of life. On the 60th birthday, the ritual includes the invocation in pots of water of all the years – regarded as divine like all units of time may be. The years are then worshipped in the water before it is used to bathe the celebrant and his wife.

¹⁷ The list of auspicious days is headed cupa muhūrtta – “days with good (auspicious) periods”. To the best of my knowledge, the auspicious days are determined simply by eliminating all inauspicious periods. I lack complete data on these periods; in Madurai, the more important ones are: āṭi, puraṭṭāci, and (for Non-Brahmans only) māci; Tuesday and Saturday; piratamai, aṣṭami, navami, amāvācai and (for Non-Brahmans only) paurṇami titis; paraṇi, kārttikai, āyilyam and keṭṭai nakṣattirams. Some of the other inauspicious periods have been referred to above.

¹⁸ Modern discussions of the Tamil ritual cycle are contained in B. Beck, E. F., Peasant society in Konku, Vancouver, 1972, 52–6, 282–3;Google ScholarClothey, F. W., The many faces of Murukan: the history and meaning of a South Indian god, The Hague, 1978, 131–48;Google ScholarDiehl, , op. cit., 158–97Google Scholar and passim; Dumont, L., Une sous-caste de l'Inde du sud, Paris, 1957, 372–403.Google Scholar But none of them deals with the problem comprehensively. There are several older ethnographic sources but few of them contain much worthwhile discussion. Many features of the Tamil cycle are, of course, also features of ritual cycles in other parts of India – a fact which considerably complicates analysis, but which I cannot discuss here. Some aspects of the cycle in the Madurai temple are discussed in my forthcoming paper, “The Divine Couple's relationship in a South Indian temple: Mīnakṣī and Sundareśvara at Madurai”, History of Religions, XIX, 05 1980.Google Scholar