374 THE ANCIENT SITE BEYOND THE NIYA RIVER [Chap. XI
Yüeh-chin', i. e. from the old Indo-Scythian dominion 45, for whom the land route through Khotan was a more likely line of communication than the sea route.
Far more numerous, of course, were articles for which local manufacture may safely be assumed. Among the rags mixed up with the refuse were the pieces of cotton and woollen materials, showing considerable variation in texture and colour, from which representative specimens (N. xv. 012, 13, 16, 17, 18) have been reproduced in Plate LXXVI in their original colouring. The coarse cotton check (N. xv. 012), in dull yellow with blue stripes, curiously recalls the
Jharans' in favourite use through modern India. The pieces of felt, N. xv. 014, 015 (see Plate LXXVI), the one very compact and strong, the other loose but richly coloured, show that this still flourishing branch of Khotan textile industry had already attained full development. The comparison of all these fabrics with those found in the Endere temple, from which specimens are shown in Plate LXXVII, might, in view of the four or five centuries separating the two sites, possibly prove instructive to an expert student of ancient textiles. The twisted cord (N. xv. ow. k), with its hard, bead-like masses entwined at regular intervals, has the appearance of a kind of rosary. In the horn spoon (N. xv. o01. i), we have a utensil of domestic use. A large number of sheep's knuckle-bones, often painted pink like the specimen N. xv. ooi. j, showed that gambling with this simple form of dice must have had its votaries in the household. Besides these there was found also an ivory die, N. xv. oo4 (see Plate LXXIV), showing four equal rectangular sides and two square ends. By its shape and size it distinctly recalls an ancient brass die from Eastern Turkestân described by Dr. Hoernle ; but as it is not perforated, it .could not have been used for the system of divining explained in Dr. Hoernle's remarks 46. It may be noted that the arrangement of the dots distributed over the four equal sides differs from that seen on the cube-shaped ivory die found at Endere (see E. o01 in Plate LI I). A curious small relic is the oblong piece of thick and very hard leather, N. xv. 005 (see the diagram, p. 41i), rounded at one end and square at the other, pierced by three holes on each of its longer sides. Its shape and the manifest arrangement for sewing suggest the possibility of its having belonged to a piece of scale armour, where the scales of leather were fixed in the manner displayed on the Kubera statue of the Dandân-Uiliq shrine, D. 11, discussed above 47. Finally the dried date (N. xv. o01. 1), may find mention here as a curious proof that even such delicacies from the distant South were obtainable by those who once lived in this ruined dwelling.
What the remaining ruins at this site yielded in written documents must appear small in comparison with the rich finds described in the preceding sections. But their excavation was attended by some interesting results. First among these ruins was a small detached structure, N. xvi. (see plan in Plate X X X I I), situated about 70 ft. to the south of the nearest extant portion. of N. v., in what evidently was once an orchard attached to that residence. The walls of timber and plaster forming a square of 16 ft. had, owing to the insufficient protection of sand, which lay only to a height of about 4 ft., decayed badly. The interior of the room, which by its dimensions and isolated position at once reminded me of the small temple cellas of Dandân-U iliq, was found to be occupied by a platform of stamped earth, 8 ft. square and
45 See Hirth, China and the Roman Orient, pp. 23o sq. *' [For a striking confirmation of this view, see Addenda.]
46 See Report on C.A. antiquities, i. pp. 39, 44.