there to the present day," which I had so far succeeded in bringing to light. The technical details concerning the arrangement of warp, weft, and pile are explained in Mr. Andrews' description of L.A. vi. ii. 0046, a subsequently-found and better-preserved specimen. There the peculiarity of the long woollen tufts with free ends which both specimens show on the back, and which make them resemble the modern cheap Japanese rug in this respect, has also been discussed. Whether these carpets actually came from Khotan it is, of course, impossible to assert. But considering the practical monopoly which the Khotan carpet industry has enjoyed for a long time past in Chinese Turkestan and the historical evidence for its ancient date, the surmise seems to me justified.
Among the miscellaneous fragments of fabrics plentifully recovered from the refuse, L.A. i. iv. 00I1, pieces of plain woollen materials in brown, buff, and red prevail. In these and the fragments of felt, dyed yellow, red, and scarlet, it is quite safe to recognize local products. The abundance of wool in the Lop region is well attested by the early Chinese accounts and, owing to the pastoral facilities continuing in the riverine jungle of the Tarim and in the mountains above Charkhlik, still exists at the present day. On the other hand, the numerous small pieces of fine silk in a variety of rich colours, undoubtedly shreds of garments, which were found in the rubbish both at this and some other ruins of the site, were certainly derived from Chinese imports. I knew that the ancient silk trade of China with Central Asia and beyond must have moved for centuries along the very route marked by this ruined settlement, and that to it the latter owed its original raison d'être. Yet I could scarcely hope at the very outset to find so striking and instructive a relic of that early trade as came to light that day in the immediate vicinity of this ruin.
To the north-north-west, the terrace on which stood L.A. i was connected by a kind of neck with a larger piece of ground immediately south of the Sttipa base which, though attacked by wind-erosion, had yet in places retained its original level under the protection of timber débris and what appeared to be a flooring of reed fascines. It is possible that this had served as a foundation for walls, but these could no longer be traced. On clearing the ground here of a light layer of drift-sand there was found, flush with the original flooring, a small bale of yellow silk, L.A. 1. oo2 (see Plate XXXVIi), tightly rolled and evidently unused, which had become so dry and brittle that when first lifted it broke in two. Its actual width was 18; inches, its diameter 22 inches. It was useless to speculate how it had come to be left behind when the structure once standing here was abandoned, or how it had escaped those who, during the immediately succeeding period, were likely to have searched the deserted settlement for any objects of value or practical use. But what I could realize at once was that this find showed us for the first time the actual form in which that most famous product of the silk-weaving Seres used to travel from China to the classical West.
A series of interesting finds made afterwards has settled it beyond all doubt that the width shown by this little silk roll was the regular one adopted for China's most important article of export in the centuries immediately before and after the commencement of our era. Decisive evidence on this point is furnished by two strips of undyed silk, T. xv. a. i. 3 (Doc. No. 539 Plate xv), which I discovered about four months later at one of the ruined watch-stations of the ancient Chinese Limes west of Tun-huang, and the inscription on which M. Chavannes has examined and explained.12 One of them, bearing a seal imprint and complete in height, shows that the piece of silk to which it belonged had a width of 5o centimetres (19.69 inches). The other, 30.5 cm. long and incomplete, bears an inscription in Chinese, exactly indicating the origin, dimensions, weight, and price of the piece of silk : ` A roll of silk, from Rang-fu in the J ên-ch`êng Kingdom ; width 2 feet 2 inches ; length 4o feet ; weight 25 ounces ; value 618 pieces of money.'
" Cf. Ancient Khotan, i. pp. 134, 174, on the Khotan 12 See Chavannes, Documents, p. 118; for M. Chavannes'
carpet industry and early Chinese references to it. correction in his interpretation see below, chap. xix. sec. iv.