360 THE ANCIENT SITE BEYOND THE NIYA RIVER [Chap. XI
found by me before in what was undoubtedly the office-room of the ancient dwelling N. iv7; these specimens are not split at their points.
Whether these wooden pens had actually served for the writing of Chinese or of Kharosthi documents, it is impossible to decide. But as M. Chavannes has observed, the peculiar form assumed by the Chinese characters on our wooden slips distinctly points to their having been written with such styles of wood. Hence there can be little doubt that the latter in their form
and material correspond closely to the pi , which, as demonstrated by M. Chavannes, consisted
of a small stem of wood, usually bamboo, used for writing in ancient China, before the general adoption of the brush, invented about the third century B. c., caused the term to be appropriated for the designation of the latter writing implement 8. The use in Khotan of wooden pens instead of the brush for writing purposes is' specifically recorded by the Chinese Annals as late as the Liang and Tang periods 9. The quasi-archaic survival here attested is of special interest, as it furnishes an exact parallel to the long-continued use of wood as a writing material in the same region. As we shall presently see, paper had been known in China for fully one and a half centuries previous to the time when the documents of the Niya Site were still recorded on wood, and the wooden documents from Dandan-Uiliq and Rawak prove that the use of the latter had not completely ceased even as late as the eighth century 19.
Seeing how closely the Chinese documents of the Niya River Site conform in their appearance and material to what we know from literary sources of the form of written records in ancient China, it seems justifiable to inquire whether the antiquarian materials brought to light there may not throw light on other points connected with early writing arrangements in China. From a great number of passages discussed in M. Chavannes' treatise, it results that, as a single bamboo slip (chien) could afford space only for from twenty to thirty characters, a series of these slips were used for all writings of any length 11. The number of slips being necessarily great in the case of books, they used to be arranged into Lien or packets, ordinarily corresponding to chapter divisions. In order to keep together the number of bamboo slips forming such a packet, it was manifestly necessary to give them a fastening of some kind. Several texts quoted by M. Chavannes mention the silk cords or thin leather thongs which were used in antiquity for this purpose 12. But they do not furnish any indication as to the particular method of fastening by which the retention of the slips in their correct order was secured. Yet it is evident that without some such method the greatest inconvenience was bound to arise, since, owing to their narrowness and probably curved surface, the slips could not by themselves lie flat one above the other and thus retain their position, as e. g. a bundle of loose leaves of paper would.
In Indian palm-leaf MSS., where a similar need was felt, a string passed through either one or two string-holes in each leaf secured the desired order 13. The observation of two unmistakable string-holes in a blank slip of wood (N. x. 9. c) found in the office-room of N. iv, hence suggested to me whether it might not be a specimen of the kind of slip used for Chinese records where fastening in a fixed order was intended. The fact of this slip, reproduced in Plate CV, being only half an inch broad, seems a distinct indication of its having been meant to receive Chinese writing. Its length between the holes placed near the two ends is 98 in., thus agreeing exactly with the average size of the Chinese slips from N. xv. The holes are in each case
7 See above, p. 335.
8 See Chavannes, Les livres chinas, pp. 65 sqq., 7o, 72 sqq.
e Comp. Rémusat, Ville de Kholan, p. z6 ; Chavannes, Turcs occid., p. 125; above, pp. 170, 173.
10 See above, pp. 268 sq., 305.
11 See Les livres chinois, pp. 40, 43 sqq.
12 Comp. Les livres chinois, p. 43.
1S See Bidder, Indische Palaeographic, p. 89.