added finds of three more Kharosthi documents, each presenting its own special point of interest. One, L.A. r. iv. 6, was a regular wedge covering-tablet, with seal cavity exactly reproducing the shape and arrangement of the wedge-shaped double tablets in wood which my finds at the Niya Site had proved to have been in use for semi-official correspondence, as it were.' Thus the elaborate indigenous system of ancient stationery was exactly the same in this far-off corner of the Tarim Basin as in the Khotan region. Though the sand-abraded surface of the obverse no longer allowed the writing of the address which it must have once contained to be read, there could be little doubt about its having been intended for a representative of the indigenous administration and not for a Chinese official.
The other record, L.A. i. iv. 5, consisted of a rough piece of tamarisk wood, still covered with bark on the back, and inscribed on the flat inside surface with two lines in Kharosthi. Its material looked strangely uncouth by comparison with the neatly finished and smooth tablets of the Niya Site, and this observation soon drew my attention to an essential difference in the make of this ancient stationery. On examining the other Kharosthi records on wood more closely (and this applies equally to the far more numerous ones subsequently recovered at this site), I found that their rough and cracked surface was not so much due to increased exposure and corrosion as to the wood being that of the Toghrak or wild poplar, with its naturally coarser fibre, instead of the Terek or cultivated poplar invariably used in the Kharosthi tablets of Niya. I shall have occasion to refer to this significant fact further on, when I discuss the limited resources of local cultivation at these Lop sites as shown also by the almost complete absence of dead tree trunks belonging to ancient
arbours or orchards.° .
From the same refuse-strewn slope were recovered also the three fragments of a paper document, L.A. I. iv. 7, showing Kharosthi writing in a faint straggling hand. They, too, had an interest of their own as the first evidence I found of the use of paper at a period when the Kharosthi script and the . Prakrit language associated with it in the Tarim Basin were still current. That paper was, in this region and epoch, used simultaneously with wood for Chinese records also was made certain by the discovery of a number of fragmentary Chinese documents which were picked up from the rubbish of the wind-eroded slopes, L.A. i. v. i : Doc. Nos. 894, 895 ; L.A. i. iv. 2, 3 : Doc. Nos. 930-38, Plates XXIX—xXX.10 The last-named were found glued together, card-board fashion, into small roughly semicircular packets which on one side bore remains of a thin coloured
plaster coating. Whether these pieces of waste paper, proved by M. Chavannes' decipherment to contain portions of private letters, some offering a certain historical interest, had been used as a backing for a painted decoration or to fill some opening in a wall, as M. Chavannes thought, can no longer be determined. The fact of Kharosthi records at this ruin being nearly as numerous as Chinese contrasts strikingly with the great preponderance of the latter in the structure L.A. ii, which certainly marks the local Chinese Ya-mên. It lends support to the suggestion that this dwelling L.A. I may have been occupied by a petty local head-man or representative of the indigenous administration of Lou-lan.
But in addition to these written records, the search of the first ruin explored was rewarded also
by other relics of interest. In a corner of room ii there were found two fragments of a woollen pile carpet, L.A. i. ii. ooi (Plate xXXVII), much worn in most places but elsewhere still retaining its rich colours of deep claret, two browns, buff, light blue in remarkable brightness. It was the first ancient specimen of an industry attested in the Khotan region from very early times, and surviving
e Cf. Ancient Khotan, i. pp. 347 sqq., 364 sq. hereafter quoted with the serial numbers given to them by
9 Cf. below, pp. 404 sq. M. Chavannes in his Documents, as well as with their proper
10 For facility of reference Chinese records are here and site-marks.