Page:Archaeological Journal, Volume 5.djvu/241

This page needs to be proofread.

UPON THE ARCHITECTURE OF WESTERN INDIA. 183 The lotus and tlic elephant are the most frequent and con- spieuous subjects of Buddhist sculpture. The former is the flower and the latter the animal which Hindu legends asso- ciate especially with water and rain. The parabolic moulding therefore seems to have an inti unite connection with the mys- tic ideas of the Buddhists ; and hence it seems to have ob- tained such an extraordinary ascendancy in their architecture. On comparing the letters of the inscription (fig. 1.2) with ]Ir. Prinsep's table of " Modifications of the Sanscrit Alpha- bet," (Journal of the Asiatic Society of Bengal, vol. vii. pi. 13,) it appears that two of these belong to the fifth century, B.C. ; two to the third century, B.C. ; and six belong decidedly, and seven may be referred, to the second century, A.U. The pilaster of Buddh-Gliosh, and consequently the commencement of the balustrade order, may therefore be unhesitatingly re- ferred to the second century. On the rough wall of the unfinished hall to which fig. 1 belongs is cut an inscription in Sanscrit and cave characters. It is a mere scribble. Ball 3^^^)) Gangadhar Shastir, a learned brahmin, who had long been engaged on these antiquities, (and who furnished me with a translation of the other inscription, wdiich I have lost, but which I believe is identical with that I have given above,) could make nothing out of it. But the second letter seems decidedly to refer it to a tune not very long after Buddh-Gliosh, so that the balustrade order could not have existed long at Kenneri. Subsequently to the suppression of Kenneri the balustrade seems to have undergone the following modifications : I. The square pillar of fig. 11 was cut octagonally. See Vihara Cave Adjunta, No. 7 ; Eergusson's Rock-cut Temples of India. II. To apply another remedy to the same defect the pedes- tal was cut pyramidally, as in the Doomar Lena of Ellora. III'. A rich ornamental girdle encircled the balustrade shaft, which is thereby obscured and confounded with the pyramidal pedestal. This variety is very stum})y in its pro- portions, and this character is increased by a rich deep border which edges the pedestal. Lanka, Ellora; Eergusson, pi. 16. IV. Neither the pyramidal form nor the octagonal shaft being sufficient to reheve the extreme plainness of the im- mense pedestal, the flutes of the shaft were continued down it, which thus became nearly cylindrical. Consecpiently the