270 THE RUINS OF DANDAN-UILIQ [Chap. IX
documents. The paper is in all cases of a thin flimsy kind, closely resembling in outward appearance that of the Chinese documents found at Dandan-Uiliq. With a single exception (D. iv. 6), only one side of the paper is inscribed, probably because the latter was not sized, and hence would allow the ink to pass through. Of D. iv. 6 it may be supposed that its paper, like that of one at least of the Chinese documents (D. vii. 3), which Prof. Wiesner has subjected to microscopical analysis, had received a loading with starch 18.
Even such cursory examination as was possible on the spot convinced me that these more or less fragmentary sheets could not have belonged to Pôthis, but had evidently served for detached records of some kind. The cursive Brahmi writing which they invariably showed was not to be found on any leaf which by its shape or material could be recognized as part of a manuscript book. The impression gained from the outward appearance of these papers has been confirmed by Dr. Hoernle's examination. This shows that they are documents closely agreeing in language, writing, and probable contents with the far more extensive collection of Brahma documents described and analysed by him in the second part of his Report". Referring to that publication for all details of Dr. Hoernle's painstaking researches, I shall restrict myself here to those antiquarian points which my own observations on the spot may help to illustrate. The manuscript materials in cursive Brahmi upon which Dr. Hoernle worked, and which corn-prised no less than thirteen sheets complete or nearly so, besides a large number of fragments, had been derived from purchases made during the years 1895-8 by Mr. Macartney and Captain (now Major) S. H. Godfrey from Badruddin Khan, the previously mentioned Ak-sakal of the Afghan merchants of Khotan 18. Internal evidence of a conclusive kind, which Dr. Hoernle has duly indicated, proves that the whole of the Brahmi documents comprised in those purchases `came from the same locality, and even belonged to the same community'. On the other hand, the results of my inquiries at Khotan and the comparison of these documents with my own finds make it appear practically certain that this locality was Dandan-Uiliq, and that the documents described by Dr. Hoernle represent chance finds made by Turdi during his earlier visits to that site. Badruddin Khan acknowledged from the first that, apart from the ample supply of forged manuscripts and block-printed ` old books ' with which Islam Akhün had furnished him, all genuine acquisitions of ancient manuscripts had reached him in the form of fragments of leaves or crumpled lumps of paper brought by Turdi and his people. My old `treasure-seeking' guide himself asserted with equal emphasis and consistency that nowhere else but at Dandan-Uiliq had he and his fraternity ever found ancient ' Khats ', and that all of such that he had sold to Badruddin Khan came from that site. In appearance, material, and, so far as he could judge, in their writing his manuscript finds had always resembled those which rewarded my search at Dandan-Uiliq.
These statements made by the two persons through whose agency the documents comprised in Dr. Hoernle's collection had reached Kashgar and Leh, is strongly supported by the evidence of the documents themselves. In paper, writing, and even the crumpled condition in which many of them reached Dr. Hoernle, they show the closest resemblance to my own finds. That their language, which Dr. Hoernle's philological acumen has successfully deciphered, is found
16 See J. Wiesner, Neuer Beitrag zur Geschichte des Papiers, pp. i i sq.
1T See Report on C.-A. ant., ii. pp. 3o sqq.
18 Captain Godfrey's contribution (G. 1) which comprised a number of the documents here under discussion, is said to have been made up of finds made near Kuchâ.