426 THE ENDERE RUINS [Chap. XII
Kanjur b. By a painstaking comparison with the text which the latter collection contains Dr. Barnett has succeeded in determining the correct order of all the fragments, even the smallest, and has thus been able to prepare the edition presented in Part i. of Appendix B. There, too, will be found Dr. Barnett's annotated transcriptions of the miscellaneous fragments of religious works (Part ii.), and a full translation, with exegetical notes, by the Rev. Mr. Francke, of the two religious poems which the completely preserved sheet (E. i. I I) has proved to contain (Part iii.).
Referring for details to Dr. Barnett's introductory remarks on the Salistamba-sutra MS. and to the Preliminary Notice of my Tibetan MS. finds previously published by him 6, I shall briefly indicate here the main facts as to the material, &c., of this MS., and the philological and historical interest attaching to its text. Several of the pieces consist of practically full half-leaves like E. i. 24 + 21, 10+ 13, which Plate CXV I I reproduces. In the arrangement of the lines (five per page) and the string-hole the MS. does not differ materially from the early Brahmi Pbthis previously described. But here the writing is confined to one side of the leaf only, a peculiarity which from the first attracted my attention. It has found its explanation in the interesting results of the detailed microscopical and chemical analysis to which Professor J. Wiesner kindly subjected the paper used in these leaves 7. His investigations have proved that the paper consists wholly of the well-macerated raw fibres of a Thymelaeacea, in all probability a Daphne plant, such as Daphne papyracea, which is still used in the preparation of modern Nepal paper. As this plant and kindred species of Daphne are not to be found in Eastern Turkestan, it becomes highly probable that the MS. was not written in the country, but imported from Tibet. An equally notable difference from the paper of the other ancient MSS. excavated by me at this site, as well as at Dandan-Uiliq, is presented by the very peculiar method adopted for making the body of the paper impermeable to liquid ink, and thus its surface more suitable for writing. Professor Wiesner has discovered that this object was attained here not by ` sizing' with a glue of starch, a method of which various fashions can be traced in the early Turkestan papers 8, but by heavily ` loading' the paper with unaltered starch of rice flour, and on that side only which was intended to be written upon. Professor Wiesner has discussed in full detail the interest which this method of ` loading', not previously observed by him in any ancient Asiatic papers, presents for the historical development of early paper manufacture 9. For our purposes it is also important as further evidence of the importation of this particular MS. Of the paper used for the minor Tibetan text fragments, including the sheet E. i. I I, Professor Wiesner has proved that it agrees in all main characteristics with the paper of the old Brahmi MSS. and Chinese documents contained in Dr. Hoernle's and my own collections 1Ô. These texts may, therefore, be supposed to have been actually written within the Khotan region.
The recovered fragments of the Salistamba-sutra represent about one half of the text as contained in the Kanjur, and known also from quotations of its Sanskrit original and from Chinese versions 11. The text differs but little from ` the version of the Kanjur, which indeed
s The first identification of the text was made independently by Professors C. Bendall and L. de la Vallée Poussin, the latter recognizing it in the specimens reproduced in Plate XVI of my Preliminary Report; see J.R.A.S., 1903,
6 See jR.A.S., 1903, pp. 109-14.
7 Compare Professor Wiesner's article Ein neuer Beitrag zur Geschichte des Papieres, reprinted from the Sitzungs-
berichte of the Imperial Academy, Vienna, cxlviii, 1904, pp. 14-21.
8 As, e.g., in the Chinese document D. vII. 3. See Wiesner, loc. cit., pp. ix sqq. ; also Hoernle, J.R.A.S.,1903, p. 678.
Comp. ibid., pp. 18 sqq.
10 See ibid., p. 21 note.
11 See J.R.A.S., 1903, pp. i to, 1 12.