Sec. v] CHINESE DOCUMENTS FROM N. xv. AND WRITING ON WOOD 363
Bhürja from Kashmir, the nearest area which could furnish it, very troublesome and expensive. That Bhûrja was nevertheless known in Khotan is proved by the Dutreuil de Rhins MS., and by the tiny fragment I discovered on the plaster wall of the Endere temple cella 25. Palm-leaves must have been still more difficult to obtain, though it is probable that MSS. written on this material were at times imported from India into Buddhist establishments 28. Leather was used for writing purposes, as my discoveries in N. xv. have proved ; but its preparation was certainly more troublesome than that of wood, and its cost, too, in all probability greater. Wood was thus indicated by nature as the common writing-material in the Khotan region, and probably throughout Eastern Turkestan, just as the bamboo was in China, until the introduction of the far more convenient paper rendered its use obsolete.
Why paper should not have come into use in Khotan territory even one and a half centuries after its invention in China is a question which cannot be definitely answered from our available materials. The fact itself must be considered as certain, for rich though the ruins explored by me at the Niya River Site were in rubbish remains of all kinds, not one of them yielded the smallest scrap of paper. This total absence of records on paper is all the more curious in view of the political connexion with China which did not cease, as our Chinese documents plainly prove, even after the close of the later Han dynasty. Nor can it be attributed to the possible want of the paper-mulberry tree (Broussonelia papyrifera), from the bark of which the modern paper of Khotan is exclusively manufactured, seeing that the alternative use of rags, hemp, and other substitutes was known in China from the very time of the first invention of paper (105 A.D.) 27. Whatever the cause may have been, the continued use of wood during the latter half of the third century in a distant Khotan settlement cannot be a matter of surprise when it is considered that, as M. Chavannes has shown from incontestable evidence, writing on bamboo slips was still currently practised in China itself about 200 A. D.28
23 See below, chap. xit, sec. ii.
26 Evidence in this direction is furnished by the arrangement of the Brihmi MSS. on paper from both DandanUiliq and Kuchi, which clearly shows imitation of palm-leaves, in their Pbthi shape, string-holes, &c. Bhûrja MSS.
were either rolled, as shown by the Dutreuil de Rhins MS., or else arranged in ` forms ' and bound after the fashion of our books, as is invariably the case with Kashmir Sanskrit MSS. !7 See Chavannes, Les livres chinois, p. 6.
26 See ibid., pp. 74 sq., also p. 58.