1 8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8o 



and may have been a Yaksa shrine, or the shrine of a goddess. 

 Structural temple architecture was already far advanced in and before 

 the Kusana period.' The existence of images (and Yaksa images 

 are the oldest known images in India) in every case implies the ex- 

 istence of temples and a cult. 



On the other hand it is quite certain that the word caitya some- 

 times means no more than a sacred tree, or a tree with an altar ; such 

 are designated caitya-vrksas in the Epics, and it is stated in the 

 Mahabharata, Southern Recension, 12, 69, 41 fit'., that such holy trees 

 should not be injured inasmuch as they are the resorts of Devas, 

 Yaksas, Raksasas, etc. Even when as so often happens in Buddhist 

 literature, the Buddha is represented as halting or resting at the 

 bhavanam of some Yakkha, it does not follow that a building is meant ; 

 the bhavanam may have been only a tree sacred to a Yaksa, and such 

 sacred trees are natural resting and meeting places in any village, as at 

 the present day. But in Safnyutta Nikdya, Yakkha Suttas, IV, it is 

 expressly stated that the bhavanam of the Yakkha Manibhadda was 

 called the Manimala caitya (the Jaina Silrya-praj napti says that the 

 Manibhadda ceiya lay to the northeast of the city of Mithila). As the 

 shrines of Manibhadda and Punnabhadda seem to have been the most 

 famous of all Yakkha shrines, it is most likely that the former as well 

 as the latter was a real temple, and indeed it is described as a temple 

 with doors and an inner chamber in K athdsaritsa gara, chapter XIII. 

 We know, too, that a statue of Manibhadda was set up at Pawaya,^ 

 and this must have been housed in some kind of structure. Sakyavar- 

 dhana's shrine, too, in the Tibetan text and in one of the Amaravati 

 reliefs, is a temple: so also the krtdlaya of Vajrapani in the Maha- 

 mayuri list. 



On the whole, then, we may be sure that in many cases Yaksa 

 shrines, however designated, were structural buildings. What were 

 they like? The passages cited in the present essay tell us of buildings 

 with doors, and arches {torane, which may refer either to gateways 

 like the Buddhist toranas, or, as the text has, " on its doorways," 

 probably to stone or wooden pediments, with which we are familiar 

 from the Maurya period onwards) ; "* and of images and altars within 



*Cf. HIIA, figs. 41, 43, 45, 62, 69, 70, 142: M. F. A., Bulletin, Nos. 144, 150: 

 Parmentier, L'Art khmcr primitif, p. 349, and Origine commune des architec- 

 tures dans I' hide et en Extreme-Orient, in fitudes Asiatiques. 



^ See Garde, M. B., in Arch. Surv. India, Ann. Rep. 1914-15, Pt. I, p. 21 and 

 The site of Padmavalt, ib. 1915-16, p. 105 and PI. LVIl. See also p. 28. 



* Cf. Smith, V. A., Jaina stupa of Mathura, pis. XIX, XX; Coomaraswainy, 

 in M. F. A. Bulletin, No. 150 (August, 1927). 



