^{page 1 note 1} Govt. Epigraphist for India (Ootacamund) Neg. No. 278, eaptioned: “Stone Image of Agni, east gopura.” This photograph is reproduced in several of the standard works on Indian iconography. So far, no one has questioned the incorrect ascription to the east gopura.

^{page 1 note 2} Banerjea, J. N., The Development of Hindu Iconography, first edition, 1941, p. 49, n. 2Google Scholar; “Hindu Iconography (Vyūhas and Vibhavas of Viṣṇu,” JIS0A., xiv (1946), p. 46.Google Scholar “Conjeeveram,” in the second reference, is doubtless a misprint for Cidambaram. In the second edition (1956), The Development of Hindu Iconography, p. 525, Banerjea merely states that the figure may represent Yajñapuruṣa.

^{page 1 note 3} Airletter dated 10th March, 1960, including permission to publish or quote.

^{page 1 note 4} Sastri, H. Krishna, South Indian Images of Gods and Goddesses, Madras, 1916, pp. 241–3, fig. 147Google Scholar; Rao, T. G. Gopinatha, Elements of Hindu Iconography, Madras, 1914, II, ii, p. 524Google Scholar; P. CLIII, fig. 2.

^{page 2 note 1} Gopinatha Rao, op. cit., II, ii, pi. CLII.

^{page 2 note 2} Ibid., Appendix B., pp. 253–5.

^{page 2 note 3} Gopinatha Rao, op. cit., I, ii, p. 55.

^{page 2 note 4} Quoted in Gopinatha Eao, op. cit., II, ii, p. 253.

^{page 2 note 5} Ibid., p. 254.

^{page 3 note 1} Monier-Williams defines aja as both “he-goat” and “ram”. Professor T. Burrow tells me that aja and meṣa stand for “he-goat” and “ram” respectively.

^{page 3 note 2} Gopinatha Rao, op. cit., II, ii, pp. 521–4—see also under Vahni in Appendix B; Krishna Sastri, op. cit., p. 243;Banerjea, J. N., Development of Hindu Iconography, 2nd ed., pp. 524–5, etcGoogle Scholar.

^{page 3 note 3} Ānandāśrama Sanskrit Series, 1926.

^{page 4 note 1} A name of Agni.

^{page 4 note 2} I have assumed the emendation dakṣiṇetare for dakṣiṇottare. The translation of this and the next two lines is uncertain because of corruptions in the text.

^{page 4 note 3} The word śāstra is used loosely here to include all works containing prescriptions for the making of images.

^{page 5 note 1} Ed. Trivandrum Sanskrit Series No. 98.

^{page 5 note 2} Ed. Trivandrum Sanskrit Series No. 121.

^{page 5 note 3} Apparently not the same text as edited in the TSS (see n. 1).

^{page 5 note 4} This is not the text edited in the Trivandrum Sanskrit Series (No. 65) which deals exclusively with architecture.

^{page 5 note 5} Banerjea, , Development of Hindu Iconography, 2nd ed., pp. 25–7.Google ScholarKramrisch, Stella, The Hindu Temple, p. 261, n. 54Google Scholar.

^{page 5 note 6} Stella Kramrisch, op. cit., p. 264, n. 60.

^{page 5 note 7} One other two-headed image also belongs to this group. It appears in a photograph in the Cohn Library collection at Oxford which bears the caption: “46— Image of Agni on the wall of the second prākāra of the Mahāliṅgasvāmin temple at Tiruviḍaimarudūr” (Tanjore District). This photograph, quite likely a print from one of the Survey negatives, shows a two-headed figure in a late style and possibly in terra-cotta or plaster instead of stone. There appears to be no third leg.

^{page 5 note 8} Horns appear on certain Pallava heads, principally those of dvārapālas, but they are of a quite different type.

^{page 5 note 9} Bhṛṅgi is depicted in South India with three legs. So is Jvarahareśvara— P. Z. Pattabiramin, “Notes d'Iconographie Dravidienne: I. Arddhanārīśvaramūrti. II. Jvarahareśvara or Jvaradeva. III. Candraśekharamūrti.”Arts Asiatiques, vi, 1 (1959).Google Scholar For a three-legged Bhairava image, seeAgrawala, R. C., “An Interesting Tripādāmūrti from Kirādu,” Journal of Indian Museums, x (1954), pp. 23–4, fig. 6.Google Scholar A three-legged, but apparently single-headed Agni has been reported from Modhera in Gujerat.Sankalia, H. D., Archaeology of Gujerat, p. 144, n. 6Google Scholar.

^{page 6 note 1} Banerjea, , Development of Hindu Iconography, 2nd ed., p. 524Google Scholar.

^{page 6 note 2} A standing figure of the Early Cōḷla period in the Nāgeśvara temple, Kumbakonam— Govindaswami, S. K., “Note on a Stone Image of Agni,” JISOA., iii, 1 (1935), pp. 45–7;Google Scholar a seated figure of roughly the same period in the Government Museum, Madras.

^{page 6 note 3} Banerjea, op. cit., pp. 524 5.

^{page 7 note 1} Zimmer, H., The Art of Indian Asia, I, pl. Bl, eGoogle Scholar.

^{page 8 note 1} J. N. Banerjea, op. eit., pp. 53–4.

^{page 8 note 2} Banerjea, J. N., “Hindu Iconography (Vyūhas and Vibhavas of Viṣṇu),” JISOA., xiv (1946), p. 46Google Scholar.

^{page 9 note 1} Gopinatha Rao, op. cit., I, i, p. 249.

^{page 10 note 1} J. C. Harle, Temple Gateways in South India: the Architecture and Iconography of the Cidambaram Gopuras (in the press).

^{page 10 note 2} i.e. Varuṇa the regent of the West, in the middle of the facade which faces west, and so on. Since the gopuras are rectangular buildings, orientated to the cardinal points, the regents of the intermediary quarters cannot actually face their directions; they are placed in niches near the appropriate corners.

^{page 11 note 1} The texture of the stone is different and the detail coarser when compared with the other statues in this gopura and I suspect that this is a modern replacement image. There are several in this gopura, faithful but coarse reproductions of late Cōḷa images.

^{page 11 note 2} This is the outermost eastern gopura of the temple, if one excepts the unfinished base of a colossal eighteenth century gopura. Jouveau-Dubreuil, by error, located this gopura in the west side of the temple.Jouveau-Dubreuil, , Archéologie du Sud de l'nde, i, p. 135Google Scholar.

^{page 11 note 3} I have described this gopura and its iconography in detail elsewhere. J. C. Harle, op. cit.

^{page 12 note 1} It is a matter for speculation whether this image emanated in some way or other from the great Vaiṣṇava temple of Śri Raṅgam only a mile or two away from Jambukeśvara.

^{page 13 note 1} Two lotus-headed female figures from the Southern Deccan also stand apart from the general Indian iconographical tradition and in a somewhat similar way. They, however, are significant expressions of Vedic and early post-Vedic religious concepts whereas the Agni/Yajñapuruṣa images are simply mechanical renderings into stone of certain features in the text which lent themselves, in a different age, to such a transposition.Kramrisch, Stella, “An Image of Aditi-Uttānapad,” Artibus Asiae, xix, 3/4, pp. 259–270.Google Scholar The two armless and legless figures of Śrī (“Śrīvatsa-Śrī”) at present in the Government Museum, Madras, are probably of a nature similar to the lotus-headed goddesses. Banerjea, op. cit., p. 376, pl. XIX, 1 and 3.

^{page 15 note 1} “Yajñapuruṣa is really Agni in its background, for fire was the main factor and ingredient in the performance of sacrifices by Vedic priests”—letter referred to above.

^{page 16 note 1} Letter dated 17th March, 1961.

^{page 16 note 2} Letter dated 15th June, 1961.

^{page 16 note 3} The author of a commentary of the Vājasaneyi Saṁhitā.