508 DEPARTURE FROM KHOTAN [Chap. XV
prints ' in a variety of ` unknown ' characters were alleged to have been found, of which Dr. Hoernle, between 1897 and 1899, published descriptions 2. In regard to these acquisitions, sold in steadily increasing numbers from Khotan, the suspicion of forgery had before presented itself to some of the Europeans resident at Kashgar, as well as to competent scholars s. But evidence was wanting to substantiate it, and in the meantime these strange texts continued to be edited and to form a subject of learned investigation. On my first arrival at Khotan I had naturally been on the look-out for such ` old books ' ; but offers in this article were then surprisingly scanty, and curiously enough the very first ` old book ' that was then shown to me supplied unmistakable proof of forgery. Hearing of my arrival, a Russian Armenian from Kokand brought to me for inspection a manuscript on birch-bark, consisting of some ten ragged leaves covered with an ` unknown ' script, and folded up into ` forms '. He had bought it for forty roubles, undoubtedly as a commercial speculation, and now wished to have his treasure properly appraised.
I saw at once that the birch-bark leaves, consisting each of a single thin layer, had never received the treatment which, as all old Bhûrja codices known to me from Kashmir show, was needed to make the material suited for permanent record. Nor had the forger attempted to reproduce the special ink required for writing on birch-bark. So when I applied the ` water-test ', in accordance with what experience in Kashmir had taught me, the touch of a wet finger sufficed to take away the queer ` unknown characters ' both written and ` block-printed '. It was significant that the ` printed matter ' of this manifest forgery, specially abundant on the margins, showed a very close resemblance to the formulas of certain ` block-prints ' published by Dr. Hoernle 4. The inquiries made at the time indicated a close connexion between Ibrahim Mulla, the person from whom the Armenian had purchased the leaves, and Islam Akhûn, the `treasure-seeker', whose alleged places of discovery I had vainly attempted to locate about Gûma. Local rumour, in fact, credited Islam Akhûn with having worked a small factory for the production of ` old books '. But at that time he was keeping away from Khotan, and there were reasons for postponing investigations about him.
The grave suspicions which these local inquiries had led me to entertain about the genuineness of all these ` old books ' in ` unknown characters ' had been strengthened almost to certainty by my explorations of the winter. Ample as were the manuscript materials which the latter had yielded, and in spite of the great variety of languages and scripts represented among them (Kharosthi, Indian Brahmi, Central-Asian Brahmi, Tibetan, Chinese, and Hebrew), I had failed to trace the smallest scrap of writing in ` unknown characters '. The actual conditions of the sites explored also differed entirely from the conditions under which those queer ` old books ' were alleged to have been discovered. There was good reason to believe that Islam Akhûn, to whom it was possible to trace most of these manuscripts and block-prints that had been purchased on behalf of the Indian Government during the years 1895-98, was directly concerned in the forgeries. He stayed away from Khotan during my first visits. He had been punished some time before on account of other impositions which Captain Deasy and Mr. Macartney had brought to the notice of the Khotan Ya-mên, and he evidently did not think it safe to attempt further deception in my case. I had no reason to regret the wide berth which Islam Akhûn had given me while I was engaged in my archaeological work about Khotan and at the ancient sites of the desert. But now, when the antiquarian evidence as to the true character of those remarkable literary relics in ` unknown characters ' was practically
2 See above, pp. ioo sqq. 4 See Report on C.-A. ant., i. Pls. V—VII, IX—XI, XIII, &c.
3 Comp. Hoernle, Report on C.-A. ant., i. pp. 57 sqq.