all the soil had been carried away by wind erosion. Thus on the south the original ground level had been lowered fully eighteen feet. The sand which covered the floor of the rooms still extant was only one to two feet.
But even this had sufficed to protect a number of interesting small relics. The first was a narrow slip of wood, about half an inch broad, inscribed with a column of Chinese characters, and exactly conforming in its dimensions to those ancient Chinese records which I had brought to light on my first visit to the Niya site in 1901. Similar finds of Hedin in a ruin presently to be mentioned had prepared me to find more of such Chinese ` slips ' at this site. But it was a very gratifying surprise when this was followed almost immediately by two oblong tablets in wood bearing four or five lines of faint but still perfectly legible writing in Kharoshthi. I had scarcely ventured to hope for records in ancient Indian script and language so far away to the east.
A second Chinese slip emerged from the corner of an adjoining room ; and subsequently, when we began to clear carefully the miscellaneous rubbish which had found refuge amidst the fallen pieces of timber on the eroded slopes, successive shouts of ` Khat ' soon announced the discovery of Chinese records on paper. They were fragments large and small, evidently of documents, and half-adozen in number. But better preserved and more curious were the closely packed layers of papers inscribed with a Chinese text, which seemed to have served as a backing for what had manifestly formed part of a large Buddhist painting executed on a thin ground of plaster. Little survived of the colouring, and the subject could not be made out with any certainty. The regular columnar writing on the backing leaves suggested some text or official record.
From the débris-strewn slopes came two more Kharoshthi records on wood which even on their first hurried examination offered some points of interest. One was the covering tablet of a wedge-shaped document which, though all the writing had become badly effaced through driving sand, still showed by its seal-socket, string-hole, etc., that
IIIF Color HighRes
