¹ The author takes this opportunity of thanking R. K. Swan for his kindness in making possible the visit to Tārācuram in November 1977 during which photography for this paper was carried out.

² Examples are the Rāmāyaṇa-friezes at Haḷěbīḍu, Karnataka, and at Prambanan in Java.

³ e.g. the episode of Śiva garlanding Caṇṭecura nāyaṉār, the subject of a fine sculpture at Kaṅkaikŏṇṭacoḷapuram.

⁴ Pěriya purāṇam is usually ascribed to the reign of the Coḻa Kulottuṅka II, A.D. 1133–50, and the Tārācuram temple is regarded as the work of his successor, Rājarāja II, 1150–73.

⁵ As (1) to (43) having inscribed captions. A farther eight, unnumbered, were introduced as ‘a number of other labels which are only painted with red paint, but not cut. The alphabet of these labels is nearly the same as that of the foregoing’. ARE, 1908, 81.

⁶ 2nd ed., New Delhi, 1973, Appendix, pp. 40–6, ‘Stories of Śaiva saints depicted at Dārāsuram’Google Scholar.

⁷ Sivaramamurti, op. cit., pp. 33–4. Pillai's, J. M. Nallaswami work in English, Periyapuranam (the lives of the Saiva Saints), Madras, 1955 Google Scholar, is an incomplete survey, omitting N7–N22, N54, and N55, but reproducing most of the ARE (1920) illustrations of the panels; it is referred to below as JMNP.

⁸ At Prambanan this dilemma was avoided, since the 48 Rāmāyaṇa episodes carved there are on the inside of the outer wall.

⁹ For their arrangement at Tārācuram, see list on p. 269.

¹⁰ See: Piḷḷai, Na. Ci. Kantaiyā, Tamiḻppulavar akarati, 2nd ed., Madras, 1960, 186 and 251–4Google Scholar. See also Zvelebil, K., The smile of Murugan, Leiden, 1973, 191 Google Scholar. He assumes that Cekkiḻār also made use of oral traditions, inscriptions and court-records and documents, but gives no references for the last two categories. Also see Arooran, K. Nambi, Glimpses of Tamil culture based on Periyapuranam, Madurai, 1977 Google Scholar. This is a particularly useful work.

¹¹ Sastri, K. A. Nilakanta, History of South India, 4th ed., Madras, 1976, 379 Google Scholar.

¹² Infra, p. 284.

¹³ Infra, pp. 282, 283, n. 62.

¹⁴ Tev. vii, pat. 39, vv. 1–11. This is taken as having the earliest extant form of each Nāyaṉār's name; these are in heavy type, followed by the N number in parentheses. The sequence is that adopted for the whole paper, and is based on the Tārācuram-frieze wherein Cuntarar appears as N63 and where the groups, clearly sculptured as such, are slotted into the sequence exactly as in P. They are numbered here G1 to G9.

¹⁵ See plate, I, and infra, p. 281 Google Scholar.

¹⁶ An example is N60, Tirunīlakaṇṭa ppāṇaṉār. Non-episodie, the panel yet shows the saint as playing a bow-harp; both T and NN refer to him as a bard, in v. 11 and v. 83 respectively.

¹⁷ NN, v. 20.

¹⁸ But P covers this fully, see P IV, 4, w. 1071–4.

¹⁹ Nor is there correlation between this set of 19 panels and the presence or otherwise of captions to them: 12 have captions, 7 do not.

²⁰ T, v. 10: Pattarāy ppaṇivārkaḷ (G3) and Tiruvārūr piṟantārkaḷ (G6). NN, v. 49 (G2); v. 70 (G3); v. 72 (G5); v. 75 (G8). While appearing in the correct place in the sequence, G2 has been taken as singular in the Sanskrit Śivabhaktavīlāsa (Śv.), being rendered as Satyadāsa! Infra, p. 282.

²¹ NN, Invocatory verse, unnumbered in published texts.

²² op. cit., 191.

²³ NN, vv. 8, 16, 23, 32, 40, 48, 57, 63, 69, 77 and 86.

²⁴ P vv. 550, 967, 1265, 1898, 3635, 3938, 4054, 4095, 4146 and 4170. Also in this category is the final carukkam, Veḷḷāṉaic., of 53 verses.

²⁵ Tārā. N63.1–3. The ARE (1920) author was almost certainly in error in stating that ‘probably this refers to the birth of Ālālasundara’ (p. 103). Possibly the fact that the extant P starts with the heavenly Cuntarar led the epigraphists to start at the wrong end of the Tārā. frieze in the first place. But of course the order within the set of seven Cuntarar is reversed in ARE too: the set as given there starts with the saving of the boy from the crocodile and finishes with the Saint's birth! See ARE (1908), 80–1 and ARE (1920), 102–3.

²⁶ e.g. Śv. adhyāya 26: Kalānātha for N10 Kuṅkiliyakkalaya n.; a. 29: Gonātha for N13 Āṉāya n., a probable mistranslation of Ta. -āyaṉ, see infra, p. 277; a. 30: Mūrtinātha for N14 Mūrtti n. and a. 31 Skandanātha for N15, Muruka n.

²⁷ T, vv. 2 and 6.

²⁸ See Śv. further, a. 52, 57 and 64. -bhakta also appears as a devotee-title: Śv. 53 Daṇḍabhakta for N30 Taṇṭiyaṭika' ṇ.; a. 59–62 Dabhrabhakta for N35 Ciruttŏṇṭa n.

²⁹ op. cit., 40.

³⁰ This Job-like story may be compared with that of Amarnīti n. (N6) who mislaid the god's cloth given him to look after. Śiva was irate that Amarnīti should offer another cloth to replace it, produced another of his own and said that the loss would be acceptable if Amarnīti could balance its weight in the scales. He had to seat himself, his wife and children therein before this could be achieved and be pardoned by the god for his carelessness.

³¹ Painted.

³² Śv., a. 16.

³³ Supra, n. 30.

³⁴ I follow here the text as edited by C. K. Cuppiramaṇiya Mutaliyar, II, 1042–50. Hereat are textual differences that lie outside the scope of this paper.

³⁵ op. cit., 40.

³⁶ ibid.

³⁷ See Burrow, and Emeneau, , Dravidian etymological dictionary, p. 25 Google Scholar, col. i, entry 283 (DED).

³⁸ P IV, 4, v. 1051/11, line 4.

³⁹ The distinctions between ě/e and ŏ/o were not observed in medieval Tamil script, accordingly they are not separated here.

⁴⁰ Supra, pp. 271, 273.

⁴¹ op. cit., 40, where Appar (Tiranā.) is mentioned first, and is said to be ‘graphically depicted in a number of scenes’.

⁴² op. cit., 41. Lacking a caption, Kulacciṟai was not noticed by ARE (1920), but Campantar, in the correct order, was; see p. 107.

⁴³ Canto 63; the MS floating is mentioned therein in verse 64; the story is depicted in the Tiruviḷai. paintings in the Bṛhadīśvara temple, Tanjore.

⁴⁴ T, v. 4, line 2.

⁴⁵ P V, 4, v. 1718/2.

⁴⁶ An excellent study of Kāraikkāl ammaiyār is that of ‘Kârâvêlane’ pseud., Kâreikkâlammeiyâr, auvres éditées et traduites, Pondichéry, 1956; pp. 1–4 contain a summary of her purāṇam as told by Cekkiḻār.

⁴⁷ There is a translation of the whole of Campantar's purāṇam, by Mudaliar, S. Sabaratna Life of Thiru Gnana Sambanthar, Madras, 1920 Google Scholar.

⁴⁸ op. cit., pp. 40 and 42.

⁴⁹ In fairness it should be noted that Taṇṭi deliberately sought out that part of Tiruvārūr where the Jains had their temples, and started his tank-digging there. An instance of the deep animosity between Śaivites and Jainas that informs the whole of P, and indeed other Tirumuṟai- works.

⁵⁰ Supra, p. 273.

⁵¹ ARE (1920), 106, § 43.

⁵² This point is made because of general and modern insistance that the receptacle signifies the female yoni, the influence presumably being Tantric. The author has seen two other square receptacles, one in the ‘CoḺ’ temple at Polonnaruva, Sri Lanka, another in the Kumbheśvara temple, Pāṭan, Nepal, a veritable stronghold of Tantrism! While not questioning Tantric belief, one may assume it was originally just a receptacle for liquid offerings.

⁵³ T, v. 6, line 3; P VII, 3, v. 3747/88.

⁵⁴ P VII, 4, v. 3773/26.

⁵⁵ Tiruviḷai. 41, v. 6.

⁵⁶ e.g. Cuppiramaṇiya Mutaliyar, C. K. comm. on P, vv. 3773/26 and 3777/30, VII, 30–1Google Scholar.

⁵⁷ op. cit., p. 42; ARE (1920), 106.

⁵⁸ The Śv., interestingly, calls this Coḻa ‘pretender’ Parāntaka, a name borne by two tenthcentury Coḻas, Parāntaka I (907–55) and II (957–73). See a. 65. The exiles may have formed the Nambūdiris of Kerala, as JMNP suggests, see p. 101, n. The panels show the brahmans in G1 and herein as wearing ‘buns’.

⁵⁹ Śv., a. 64.

⁶⁰ NN, v. 49; Pc., v. 45.

⁶¹ Infra, p. 284 (N51).

⁶² Śv. has Pañcapāda for Aiyaṭikaḷ, an apparent confusion between Ta. ai-, ‘five’, and ai ‘beauty’, which makes better sense!

⁶³ P IX, 2, v. 4064/1.

⁶⁴ Sivaramamurti, op. cit., 43; JMNP, op. cit., 121.

⁶⁵ VII, 398, comm.

⁶⁶ Cantos 58–60.

⁶⁷ ARE (1908), 81; ARE (1920), plate IV, fig. 30.

⁶⁸ Usually called Maṅkaiyārkk'araci and N57 in the sequence; infra, p. 29.

⁶⁹ Supra, p. 270.

⁷⁰ op. cit., 42–4 and n. 68.

⁷¹ P X, 3, v. 4123/4, line 1.

⁷² ARE (1920), 105.

⁷³ Tiruviḷai., canto 62.

⁷⁴ Lohitākṣa, Śv., a. 75; Cěṅkaṇāṉ, NN, v. 81, P XII, 4, v. 4210/14. As fast as the spider spun the web, an elephant came and tore it down. The spider entered the elephant's trunk and bit it, killing them both. When the former was reborn, his mother died after addressing him as Cěṅkaṇāṉ.

⁷⁵ See plate IX.

⁷⁶ See ibid., plate I, figs. 10, 11.

⁷⁷ T, v. 12, line 3.

⁷⁸ Supra, p. 280.

⁷⁹ Supra, p. 274.

⁸⁰ Which also include the earthly life and, specifically, Śiva's acting as go-between in the Caṅkiliyār affair, see Śv., a. 12.

⁸¹ Supra, p. 274, n. 24.

⁸² p. 103.

⁸³ The two were in fact incarnations of Kamaliṉi and Anintitai, the two maidens upon Kailāsa whom, Cuntarar, in his heavenly form Hālālasundara, had desired. Śiva drew up an agreement that he could have them, provided he consent to be born on earth.

⁸⁴ ARE (1908), 80.

⁸⁵ op. cit., 45. The author rightly notes that one of the pair is a king, but simply says that Cuntarar is directing him to a temple the other side. The king is not identified, nor is the cleaving of the waters mentioned.

⁸⁶ Plate, IV, and supra, p. 278 Google Scholar.

⁸⁷ Tev. vii, pat. 77: Paravum paric'ŏṉṟ'aṟiyeṉ. … Each verse ends with the cry for help: Aiyāṟ'uṭaiya aṭi keḷo.

⁸⁸ P VII, 4, vv. 3881/134 and 3882.

⁸⁹ See DED, p. 73, col. ii, entry 837: Ta. oil.

⁹⁰ See Tev. vii, pat. 59 and P VII, 4, v. 3903/156.

⁹¹ op. cit., p. 46.

⁹² C. K. Cuppiramaṇiya Mutaliyar states that a case of a boy being swallowed by a crocodile and disgorged was reported in the press. See P VII, comm. to v. 4240, p. 596, n.