Sec. v] DISCOVERY OF DATED DOCUMENTS 265
with Hsüan-tsang's story which represents those sacred rats as having ` hair of a gold and silver colour' 2. The worshipping attitude of the nude male figure on the right proper, carrying a leaf-shaped fan and looking towards the rat-headed king, clearly marks the sacred character of the latter. A third smaller figure on the left proper is almost completely effaced. On the reverse, too, five seated figures can be made out in faint outlines. It was only after the panel had been cleaned in the British Museum from the thin layer of adhering sand, and carefully examined by the trained eye of Mr. F. H. Andrews, that I realized the peculiar features of the rat-headed figure and its significance.
The other panel, D. iv. 4, measuring ioâ by 5, in. (see Plate LXV), shows on the obverse the red-robed figure of a standing Buddha, apparently in the attitude of teaching, and surrounded by vesica and nimbus. The standing figure on the reverse is too deleted for any certain interpretation, but attention may be called to the peculiar circular markings which appear above the knees of this figure on legs apparently bare.
Of manuscript finds the cella yielded from its north-east corner the fragments D. iv. 1-3, written in Brâhmi: characters of the upright Gupta type, recognized by Dr. Hoernle as belonging to single leaves of two separate Peithis. The language is Sanskrit, and the text apparently from some Buddhist canonical work. D. iv. 6 was subsequently discovered in the north-west corner of the cella in the form of a small ball of crumpled-up paper. This proved to consist of two pieces of paper, each measuring about II by 5 inches, found attached to one another, and according to Dr. Hoernle's examination evidently making up one document. It is written in cursive Brâhmi characters placed widely apart like the documents to be discussed presently, but on both sides of the paper, while ordinarily, owing to the thin unsized paper used for all documents whether in cursive Brâhmi or Chinese, the writing is confined to one side only. As in all cursive Brahmi documents from Dandân-U iliq, the language is the non-Sanskritic one of Eastern Îrânian type. Dr. Hoernle, without being able to indicate the purport of the document, has noted in it the mention of a definite date, the i 7th day of the month Mûnamji 3.
The remains of the ruined dwelling-house D. v, situated to the north of the cella and at a distance of only some 5o feet, had been exposed too long to retain any objects which ` treasure-seeking ' parties might consider of value. Turdi distinctly remembered having searched them years before, and having obtained there a number of khats (` writings '), though he, naturally enough, could not give any information as to their character. Nevertheless, the systematic clearing of this ruin was attended by some important results. By tracing the dividing walls some seven apartments could still be made out, besides two enclosures built with walls of rushes and plaster in which we may recognize outhouses or stables. Other apartments must have disappeared through erosion, judging from the plentiful timber debris strewing the slope adjoining. Nothing but fragments of coarse plain pottery, red and grey, and pieces of wall plaster were found on the slopes and in the rooms northward. But luckily the two southernmost rooms were still under a cover of sand from 2 to 3 feet high, which the wall portions extant there had helped to retain, and under this cover some interesting relics had escaped previous searchers' attention.
In particular, the large room forming the south-eastern end of the house, and measuring 17 ft. 6 in. by i 5 ft. inside, yielded an unexpected variety of written remains. The two fragmentary leaves (D. v. I, 2) in upright Gupta writing of the seventh or eighth century, which turned up first near the north-east corner, have proved to belong to Pôthis, probably containing