Sec. ii] FIRST EXCAVATION OF BUDDHIST SHRINES 247
but the feet and portions of the lower edge of the drapery remained, and I am hence unable to make any suggestion as to their original character. The arrangement of these four corner figures closely agrees with that observed in the Endere temple (see plan in Plate XXXVI).
That the cella had once contained other sculptured objects, but of a far more modest size, was shown by the finding of two small wooden boards lying detached on the top of the base and near the feet of the principal image. They were perfectly plain, but each bore the much decayed and almost shapeless remains of a pair of feet and legs in soft stucco reaching to a height of about 3 inches. These undoubtedly belonged to some small statue similar to the one subsequently excavated in the adjoining cella (see D. ti. 09 in Plate LIII). The exposed position of these little sculptures, built up on wooden boards for the sake of mobility, had caused even the wood of the latter to perish. The same cause had greatly damaged three small wooden panels (D. ii. 2, 4, 03), bearing paintings, which I found leaning against the lotus pedestal of the central image. There could be no doubt that these panels had been placed there by worshippers and represented objects or scenes of religious interest. But the decay of the wood and the damp which had reached the painted surfaces has made the outlines difficult to recognize even to the practised artist eye of Mr. Andrews. No reproduction of these painted panels could be attempted, and I must refer to Mr. Andrews' description in the list at the end of this chapter for any details which can still be made out. D. ii. 2 and D. H. 4 were painted on both sides ; the latter panel, measuring 12 by 4 inches, shows apparently three seated figures on either side. D. H. 03, about 82 inches wide and 58 inches high, painted only on one side, displays two seated figures with halos, each holding what looks like a swathed infant. Curiously enough both D. ii. 2 and D. ii. 03 show remains of written leaves, which had been pasted across the surface and thus had helped to efface the painting. The paper of the leaves has decayed so badly that only a few characters of the Brâhmi script in each line can now be made out. But these suffice to show that the Brâhmi writing was of the upright Gupta type of the seventh or eighth century, as it appears in several MS. finds from the neighbouring ruin D. ui. (see Plate CV I II), and that each leaf or piece of leaf originally belonged to some MS. of the usual oblong shape, one apparently showing three lines and the other five lines per page. Insignificant in themselves these almost perished scraps of writing were of value to me at the time as raising hopes of real MS. finds.
The walls of the cella, which, judging from the size of the principal image, must have been of considerable height, were found• to have been decorated inside with frescoes. The colours looked much faded and worn, especially in the extant upper portions, as if the surface had been exposed for a considerable time to climatic influences, including rain trickling down from the roof, for some time before the protecting sand invaded the building. Over a considerable portion of the extant walls the painted surface had peeled off. Nevertheless it was possible to make out that the walls west, south, and east had each been occupied in the centre by a sacred figure standing enveloped in a large vesica. As these central figures were over life-size, only the feet with a broad painted frieze below them showing lotuses and small figures of worshippers could be made out in parts. There were narrow vesicas painted around each of the four corner images, a leaf border being the most distinguishable feature in them. The western wall, the best preserved on the whole, showed between this corner decoration and the central painting remains of very graceful tracery work curiously suggestive of late Gothic, and inserted between it miniature representations of what looked like a small shrine (or prayer-flag ?) containing the figures of two seated deities. Most of the outlines seemed drawn in a kind of terra-cotta colour, while the colours between had often faded into complete indistinctness.
| || || || || |