A Wu-shu coin, which was found lying on the surface near the north-western corner of the court, gave me the first definite indication of the early date of these ruins. But other evidence came to light when I set men to clear the refuse layers which closer inspection soon revealed in several places near the main Stûpa. They had fortunately escaped the attention of those who had previously searched the remains of the site in the manner to be presently described. The largest of these rubbish-heaps, Y. 1. a, was found to cover the slope below the eastern edge of the platform. It consisted chiefly of broken pieces of wood, reed-straw, ashes, and fragments of fabrics and worked leather, all proofs that there must also have been living quarters of some sort on the ground above. As the refuse was being carefully sifted there came to light, to my great satisfaction, below the north-eastern corner of the platform the fragment of a wooden document, about 4 inches long, bearing two short lines of Kharosthi on one side, with the rest of the surface as if scraped. Soon there followed three small ` shavings ' from wooden documents with Kharosthi writing, just like the ` shavings ' of Chinese slips which had been found by me in large numbers at the Lou-lan station and along the Tun-huang Limes.4 The type of the writing seemed to resemble closely that of the Kharosthi documents obtained from the ruins of the former, and the use of this Indian script clearly pointed in itself to approximately the same period as that to which the Lou-lan remains belong.
Among the miscellaneous objects found here, fragments of woollen fabrics, both fine and coarse, are numerous. Mr. Andrews' examination of specimens of these textiles has shown warp-rib weave in the majority of them (Ying. a. 03-4 ; I. a. 017-19, 21). This specially deserves attention, since this technique is a distinctive mark of all the Chinese figured silks recovered from the Lou-lan cemetery L.C. The total absence of silk fragments, on the other hand, might suggest that those who ordinarily attended at this sacred spot were not Chinese but of local race. Little reliance, however, can be placed on merely negative evidence of this kind. Of other relics may be mentioned a wooden pen, Ying. a. of ; fragments of lacquered wood, Ying. a. 09, I. a. 02 ; of bands, &c., of goat's hair, a. 05, 1. a. 012, 016 ; of a strong pile fabric of wool, 1. a. 015. The pieces of a vine-stem, I. a. 014, are of interest as evidence of viticulture in the neighbourhood. Two smaller accumulations of refuse, b and c, to the north and north-west of the main Stûpa yielded wheat-straw, fragments of pottery vessels, and the like (see List). Such pieces of worked timber as were found among the debris of a small completely wrecked structure to the north of the Stûpa were all of Toghrak wood. We also came upon evidence that the sacred locality had been visited down to late Buddhist times, in the shape of a K`ai yiian coin, a Tang issue, of which two fragments were picked up on the top of the refuse near the north-eastern corner of the platform.
Grouped around the central shrine were found nine smaller Stûpas, as shown in the sketch-plan (Pl. 37). They had all been dug into and otherwise had suffered much damage ; but of most the bases could still be made out, measuring between 15 and 5 feet square. No relics were found on the bare gravel around them. The position of all these marks of Buddhist worship crowded together on the little plateau left no doubt in my mind that the spot had been held sacred as a Su-bâshi or ` head of the waters '. I have had occasion before to refer to the sites thus designated, found where rivers debouch above oases of the Târim basin, and to the worship they receive from the modern occupants of the lands irrigated by those rivers. The shrines found at Kohmäri on the Kara-kâsh river of Khotan, at Toyuk, Ara-tam, and above the outfall of the two rivers of Kuchâ are all illustrations of this local worship, continued from ancient times to the present. This was obviously the place to which the inhabitants of a settlement dependent on the Shindi river for
4 Cf. Serindia, i. p. 375 ; ii• PP. 598, 646, 685, 763. 1155, 1238 ; also above, i. p. 166 sq., regarding Bâsh-koyumal
5 Cf. Ancient Khotan, i. p. 189 ; Serindia, iii. pp. 1151, above Charkhlik.
5 D 2