¹ In its present version, this article was written by Fuller, but it has benefited from detaled discussion with I Logan, who originally developed many of the ideas in it in her Ph.D. thesis, ‘Domestic worship and the festival cycle in the south Indian city of Madurai’, University of Manchester, 1980. We would like to thank Brenda Beck, Madeleine Biardeau, Jean-Claude Galey, Tony Good, David Shulman and Burton Stein for their invaluable comments on an earlier version of this paper, some of which have been incorporated into the paper without further acknowledgement. Fuller's field-work in the Mīnākṣī Temple was carried out for 12 months in 1976–77Google Scholar (supported by the Social Science Research Council) and two months in 1980 (supported by the British Academy Small Grants Research Council) and two months in 1980 (supported by the Brithish Academy Small Grants Research Fund in the Humanities), and Logan's in Madurai homes for 15 months in 1977–78 (supported by the Social Science Research Council, Emslie Horniman Anthropological Scholarship Fund and Tweedie Exploration Fund). Names of deities, titles of texts, ritual terms, etc., other than purely Tamil ones, are transliterated from their more readily recognizable Sanskrit forms, although where relevant both Sanskrit and Tamil forms are supplied. Many sources contain brief descriptions of the festival in Tamilnadu or elsewhere, but we do not provide a comprehensive bibliography.

² Extracts from the Devīmāhātmya myth are in Dimmitt, C. and van Buitenen, J. A. B., Classical Hindu mythology (Philadelphia, 1978), 232–40Google Scholar, and from several different versions of the myth in O'Flaherty, W. D., Hindu myths(Harmondsworth, 1975), 238–49.Google Scholar

³ O'Flaherty, , Hindu myths, 242Google Scholar.

⁴ Biardeau, M., ‘L'arbre ś et le buffle sacrificiel’, Puruṣārtha, 5, 1951, 235–8; ‘Deva/Asura’ in Dictionnaire des mythologies (Paris, forthcoming), q.v.Google Scholar

⁵ Shulman, D. D., Tamil temple myths: sacrifice and divine marriage in the south Indian Śaiva tradition (Princeton, 1980), 176–192.CrossRefGoogle Scholar

⁶ Biardeau, M., ‘Le sacrifice dans I'Hindouisme’, in Biardeau and C. Malamoud, Le Sacrifice dans l'Inde ancienne (Paris, 1976), 146–51Google Scholar; ‘Devi’ in Dictionnaire des mythologies, q.v.; O'Flaherty, , Hindu myths, 238Google Scholar; Shulman, , Tamil temple myths, 91, 186.Google Scholar

⁷ The full name is Arulmiku (‘grace-bestowing’) Mīnākṣī-Sundareśvara Tirukkoyil (‘holy temple’). For description of the Temple, see C.J., Fuller, Servants of the goddess: the priests of a south Indian temple (Cambridge, 1984), ch. i.Google Scholar

⁸ Logan, ‘Domestic worship’, 215–20, 222–31, 290–308.

⁹ On the festivals, seeFuller, C. J., ‘The divine couple's relationship in a south Indian temple: Mīnākśvara at Madurai’, History of Religions, 19, 1980, 328–34; Servants of the goddess, 17–21.Google Scholar

¹⁰ On the priests and other Temple officiants, see Fuller, C. J., ‘Gods, priests and purity: on the relation between Hinduism and the caste system’, Man, N. S., 14, 1979, 462–4; Servants of the goddess, ch. ii.Google Scholar

¹¹ For brief descriptions of the worship, see Fuller, ‘Gods, priests and purity’, 460–2; ‘Divine couple’, 324–8; Servants of the goddess, 10–17.

¹² cf. Balarathnam, L. K., ‘South-Indian fasts and festivities’, Quarterly Journal of the Mythic Society, N. S., 34, 1943, 72;Google ScholarRao, B. R., ‘The Dasara celebrations in Mysore’, Quarterly Journal of the Mythic Society, 11, 1921. 303Google Scholar. (Rao's article reprinted in) Nanjundayya, H. V. and AnanthakrishnaIyer, L. K., The Mysore tribes and castes (Mysore, 1928), II, 56–71Google Scholar.

¹³ Fuller, ‘Divine couple’, 329–31.

¹⁴ ibid., 325.

¹⁵ The sari colours on the eighth and ninth nights have been verified since Fuller, ‘Divine couple’ was written;cf. p. 341, n. 41.

¹⁶ ibid. 329–31.

¹⁷ The yellow dye is painted on the sanctum walls and the śāntābhiṣeka is done solely for Minākṣī's main image, whereas her festival image is decorated and has its hair washed. It is conventional practice in the Temple to put different decorations on the festival images, and as the hair washing is followed by a procession, only festival image could be used then. It is similarly conventional practice for major bathing rituals to be performed mainly for the immovable images. The Kŏlu maṇḍapa is made of black marble, an unsuitable surface for the day, but it may be significant that the heat is concentrated in the central sanctum (see main text below). However, to ask which image actually fights the demons would be beside the point, because it is not the images but the goddess who fights; the presence of two images in the Temple no more affects the matter than do the innumerable other images also being used throughout India.

¹⁸ For details, see Logan, ‘Domestic worship’, ch. ii.

¹⁹ The popularity of Sarazvatī pūjā in Tamilnadu is specifically referred to by P., Sonnerat, Voyage aux Indes orientales et à la Chine (Paris, 1782), I, 235.Google Scholar

²⁰ cf. Diehl, C. G., Instrument and purpose: studies on rites and rituals in south India (Lund, 1956), 170–1.Google Scholar

²¹ cf. ibid., 252–3.

²² The buli during the six festivals is offered directly to the eight lords, although this is not true of the ritual held inside the Temple on ordinary days. None the less, though we cannot demonstrate it here, there is good reason to conclude that the daily sacrifice is partly intended for the eight lords as well.

²³ O'Flaherty, W. D., The origins of evil in Hindu mythology (Berkely, 1976), 58–9 and passim.Google Scholar

²⁴ Kane, P. V., History of Dharmaśāstra (Poona, 1968–1977), v, 128.Google Scholar

²⁵ cf. Diehl, , Instrument and Purpose, 252–3Google Scholar; Kane, , History of Dharmaśāstra, V, 28Google Scholar.

²⁶ The possibility that the cords are specifically those of a bride (goddess) and groom (nampiyār) is ruled out here because the cords remain on until the tenth day (i.e. after the ‘marriage’ of the eighth day: see main text below), because no ritual in the Minākṣī Temple ever suggests marriage between Minākṣī and her priests, and because the cords are also worn for festivals that certaily do not include any marriage. Cf. Beck, B. E. F., ‘The goddess and the demon: a local south Indian festival and its wider context’, Puruṣārtha, 5, 1981, 3–2.Google Scholar

²⁷ Beck, B. E. F., ‘Colour and heat in south Indian ritual’, Man, N.S., 4, 1969, 559.Google Scholar

²⁸ At Āṭi Muḷaikkŏṭṭu: Fuller, ‘Divine couple’, 338.

²⁹ For the ritual significance of heat in Tamil Hinduism, See Beck, ‘Colour and heat’, as well as discussions by Back, ‘The goddess and the demon’, 130–2 and passim; Fuller, ‘Divine couple’, 326–8; Logan, ‘Domestic worship’, 199–205; Reiniche, M. L., Les dieux et les hommes: étude des cultes d'un village du Tirunelveli, Inde du sud (Paris, 1979), 177–9.Google Scholar

³⁰ ‘Divine couple’, passim.

³¹ On this festival, see F., Clothey, ‘Skanda-ṣaṣṭi: a festival in Tamil India’, History of Religions, 8, 1969, 236–59.Google Scholar

³² Breckenridge, C. A., ‘From protector to litigant: changing relations between Hindu temples and the Raja of Ramnad’, in Stein, B. (ed.), South Indian temples (Delhi, 1978), 82–3; B. Stein, ‘Mahānavamī: medieval and modern kingly ritual in south India’ (forthcoming).Google Scholar

³³ cf. Fuller, ‘Divine couple’, 323; and Banerjea, J. N., The development of Hindu iconography (Calcutta, 1956), 497–500Google Scholar; GopinathaRao, T. A., Elements of Hindu iconography (Madras, 1916), I, 338–9, 345–6Google Scholar; Krishna Sastri, H., South-Indian images of gods and goddesses (Madras, 1916), 190, 206, 211, 220.Google Scholar

³⁴ Shulman, , Tamil temple myths, 166–76.Google Scholar

³⁵ Notably, Kolāṫṫa and Ĕṇṇai Kāppu: Fuller, ‘Divine couple’, 341–3.

³⁶ An exception is widows belonging to Vaiţṇava communities (a minority), who wear a form of the Vaiᛣṇava sect mark and not vibhūti, which is sacred to Ṥiva.

³⁷ M., Biardeau, ‘Introduction’, Puruṣārtha, 5, 1981, 15–16Google Scholar; Shulman, , tamil temple myths, 176–92.Google Scholar

³⁸ For example, the Māriyammaṉ festival described by Beck, ‘The goddess and the demon’.

³⁹ R., Dessigane, Pattabiramin, P. Z. and , J. Filliozat, Les légendes Civaītes de Kāncipuram: analyse de textes et iconographie (Pondicherry, II, 583–93.)Google Scholar

⁴⁰ Kane, ,History of Dharmaśāstra, II, 583–93.Google Scholar

⁴¹ Āṭi Muḷaikkŏṭṭu: Fuller, ‘Divine couple’, 337–8.

⁴² cf. the meaning of the white sari worn by Mīnāksī at night: ibid., 325.

⁴³ This is also shown in Minākṣā's kŏlaŭŭa and ĔṇNṇai Kāppu festivals, though Fuller (ibid., 341–3) does not suffciently emphasize the point; cf. also Shulman, , Tamil temple myths, 140–1, 166–76 (‘the lustful bridge’), 349–50.Google Scholar

⁴⁴ Beck, ‘Colour and heat’, 558.

⁴⁵ Shulman, , Tamil temple myths, 95.Google Scholar

⁴⁶ In Śaiva mythology, the god's ashes bear an erotic significance too (O'Flaherty, W. D., Asceticism and eroticism in the mythology of Śiva (Delhi, 1973), 245–7)Google Scholar and are identified in some myths as his semen (O'Flaherty, , Hindu myths, 147–8).Google Scholar

⁴⁷ Beck, ‘Colour and heat’, 558–9.

⁴⁸ Biardeau, ‘Devi’; cf. M., Biardeau, L'Hindouisme: anthropologie d'une civilisation (Paris, 1981), 147–8.Google Scholar

⁴⁹ Shulman, , Tamil temple myths, 219.Google Scholar

⁵⁰ ‘Devi’.

⁵¹ J. P. Parry, personal communication.

⁵² cf. Fuller, ‘Gods, priests and purity’, 469. The belief is, however, inconsistent with the gods' character as depicted in many myths, particularly those in which they kill a Brahman god or demon, as shown in the stories of Indra's and Śiva's Brahmanicide (O'Flaherty, , Origins of evil, 146–64Google Scholar) and Tamil myths with a similar theme (Shulman, , Tamil temple myths, 183Google Scholar).

⁵³ A., Hiltebeitel, ‘Draupadi's hair’, Puruṣārtha, 5, 1981, 179–214.Google Scholar

⁵⁴ ibid., 206.

⁵⁵ ibid., 211.

⁵⁶ Fuller, ‘Divine couple’, 334–7.

⁵⁷ Nor is there really any suggestion that it is an anointment symbolizing a regaining of sovereignty, as is so for Draupadī in the epic (HIltebeitel, ‘Draupadī's hair’, 201).

⁵⁸ Hubert, H. and M., Mauss, Sacrifice: its nature and function (London, 1964), 331Google Scholar; Fuller, , ‘Divine couple’, 331; Kane, History of Dharmaśāstra, II, 1198–201Google Scholar.

⁵⁹ Hubert, and Mauss, , Sacrifice, 48.Google Scholar

⁶⁰ ibid., 55.

⁶¹ cf. Ferro-Luzzi, G. Eichinger, ‘The logic of south Indian food offerings’, Anthropos, 72, 1977, 547.Google Scholar

⁶² cf. ibid., 547.

⁶³ Diehl, , Instrument and purpose, p. 177, n. 4Google Scholar; W., Francis, Madura (Madras District Gazeteers, Madras, 1906), 87Google Scholar. The devadāsīs gave up working in the Temple in the 1940s, and the practice of visiting the Temple for exorcism at Navarātri was banned by the Temple administration around the same time.

⁶⁴ The role of women in the domestic festival cycle in Madurai is discussed at length by Logan, ‘Domestic worship’, chs. iii and iv. Unlike Navarātri elsewhere, especially in Bengal and Nepal, the Madurai festival gives no particular prominence to virgins, although it is considered a good time for them to come of age.

⁶⁵ cf. Reiniche, , Les dieux et les hommes, 59.Google Scholar

⁶⁶ Logan, ‘Domestic worship’, 235–8.

⁶⁷ Biardeau, ‘L'arbre śamī’.

⁶⁸ B., Stein, Peasant state and society in medieval south India (Delhi, 1980), 384–92; ‘Mahānavamī’Google Scholar.

⁶⁹ See Rao, , ‘The Dasara celebrations’; Kane states that vijayadaśamī is special for Kshatriyas, nobles and kings (History of Dharmaśāstra, V, 190)Google Scholar.

⁷⁰ W., Taylor, Oriental historical manuscripts in the Tamil language (Madras, 1835), II, 103Google Scholar. According to a story related by Taylor (ibid., 5, 7, 15), Visvanatha Nāyaka was chosen as king of Madurai because he captured and then decapitated a buffalo at one stroke at the Navarātri festival in the Vijayanagara capital.

⁷¹ Taylor, ibid., 159; Rao, ‘The Dasara celebration’, 304.

⁷² Nelson, J. H., The Madura country: a manual (Madras, 1868), III, 138; cf. Breckenridge, ‘From protector to litigant’, p. 91, n. 34.Google Scholar

⁷³ cf. Stein, , Peasant state, 392Google Scholar; also Devakunjari, D., Madurai through the ages: from the earliest times to 1801 A.D. (Madras, ɛ1979]), 302, 306Google Scholar.

⁷⁴ ‘L'arbre śamī’, 224–5.

⁷⁵ D., Hudson, ‘Śiva, Mīnākṣī, Viţṇu: reflections on a popular myth in Madurai’, in B., Stein (ed.), South Indian temples (Delhi, 1978), 110–13Google Scholar.

⁷⁶ cf. Devakunjari's reference (Madurai, 302) to MInākṣī's ‘Kolu or Darbar’; Navarātri at the Vaiṣṇava temple of Varadarājasvāmi at Kāncipuram, where the god and goddess ‘grace the durbar or kolu’ of the ninth day: Raman, K. V., Sri Varadarājaswāmi temple―Kānchi: a study of its history, art and architecture (New Delhi, 1975), 105Google Scholar; the Tirunělveli temples described by Reiniche, , Les dieux et les hommes, p. 51, n. 15Google Scholar.

⁷⁷ For this myth, see Long, J. B., ‘Life out of death: a structural analysis of the myth of the churning of the ocean of milk’, in Smith, B. L. (ed.), HInduism: new essays in the history of religion (Leiden, 1976)Google Scholar; O'Flaherty, , Hindu myths, 273–80Google Scholar.

⁷⁸ For digvijaya, see Gonda, J., Ancient Indian kingship from the religious point of view (Leiden, 1966), 100–5Google Scholar; Kane, , History of Dharmaśāstra, III, 66–71Google Scholar.

⁷⁹ Fuller, ‘Divine couple’, 343–4.

⁸⁰ Biardeau, ‘Le sacrifice’, 105–6; L'HIndouisme, 107–8; ‘Viṣṇu/Śiva’ in Dictionnaire des mythologies, q.v.

⁸¹ Undoubtedly, there must be yet more aspects that could be explored if we had data on Navarātri from other temples, especially those of independent goddesses. In the study of Hindu ritual, one templeĥeven when it is as large as the Mīnākṣī Templeĥplainly does not define a ‘whole’, although the practical difficulties of collecting data from a plethora of complex temples in a large city like Madurai are not easily overcome.

⁸² ‘Le sacrifice’.

⁸³ ‘Domestic worship’, ch.ii.

⁸⁴ Sacrifice, 12.

⁸⁵ ‘l'arbre śamī’ 224–5.

⁸⁶ van den Hoek, W., ‘The goddess of the northern gate: Cěllatammaṉ as the “divine warrior” of Madurai’, in Gaborieau, M. and Thorner, A. (ed.), Asie du sud: traditions et changements (Paris, 1979)Google Scholar.

⁸⁷ Biardeau, , L'Hindouisme, 146–7Google Scholar.

⁸⁸ ‘The goddess and the demon’.

⁸⁹ cf. Biardeau, , L'Hindouisme, 153–4Google Scholar.

⁹⁰ Fuller, , Servants of the goddess, p. 191, n. 40Google Scholar.

⁹¹ Clothey, , ‘Skanda-ṣaṣṭi’, 253–4Google Scholar.

⁹² Biardeau, , ‘L'arbre śamī’, 215Google Scholar.

⁹³ cf. Reiniche, Reiniche, Les dieux et les hommes, 58–9Google Scholar.

⁹⁴ cf. Hubert, and Mauss, , Sacrifice, 85–6Google Scholar.

⁹⁵ Biardeau, , ‘L'arbre śamī’, 238Google Scholar/

⁹⁶ Shulman argues that in many Tamil myths the demon killed by the goddess is a substitute for Śiva. Although some aspects of the ritual analysed above are consistent with his interpretation, it is arguably simpler and more in a accordance with the general understanding of sacrifice to see Mahiṣāsura as a surrogate of the goddess himself; cf. Tamil temple myths, 9–11, 26–9, 90–3, 138–44, 347–42 and passim.