EXPLOEATION IN" CHINESE TURKESTAN. 771 



connection with Dandan-Uilik, Ttic Tibetan leaves, containing, as 

 Mr. Barnett of the British Museum has ascertained, portions of a 

 translation of the Salisthanil)asutra, a Buddhist canonical text, 

 undoubtedly are the oldest written remains of tliat language as yet 

 discovered. It was curious to note how the folia which originally 

 belonged to a fairl}" large manuscript had been cut up and separately 

 deposited, manifestly as votive offerings, at the pedestals of various 

 images. A pious visitor of the shrine had evidently endeavored to 

 propitiate with his text as many divinities as possible. To other curi- 

 ous discoveries made there, such as Tibetan and Chinese Sgraffitti, 

 small votive offerings of elaborately woven fabrics in silk and cotton, 

 etc., I can only allude here. But as a point of chronological impor- 

 tance it ma}^ be mentioned at least that in one of the Chinese Sgraffitti, 

 of which I brought away photographs. Professor Douglas has since 

 read a date corresponding to A. 1). 790. 



The proofs of Tibetan occui:)ation showed me that 1 had reached at 

 Endere the easternmost limits of the territor}- with the archasological 

 exploration of which I was concerned. So, on February 26 I could 

 turn back w'ith a good conscience toward the west, where several sites 

 yet remained to be examined. The journe}^ to Keriya, a distance of 

 over 180 miles, was covered in seven forced marches. The energetic 

 assistance of Khon-Daloi, the Amban, who had followed my move- 

 ments with the friendliest care and interest, allowed me to set out at 

 once with fresh laborers, transport, and supplies for Karadong, the 

 ancient site in the desert, some 150 miles north of Keriya, which Dr. 

 Sven Hedin had first visited. 



This so-called "ancient city" proved to contain little more than the 

 ruins of a roughly ])uilt (juadrangular structure, which probably had 

 served as a fortified sarai, or post, on the ancient route leading along 

 the Keriya Darya toward Kuchar in the north. My excavations at 

 this desolate spot were carried on under considerable difficulties. The 

 height of the dunes which covered the interior of the great quadrangle 

 was considerable, and daily we were visited by sand storms of varying 

 degrees of violence. The finds, which were scanty, as I had expected, 

 curiously enough included small quantities of remarkably well- 

 preserved cereals, such as wheat, rice, pulse, etc., found embedded in 

 the floor of what evidently was an ancient guard room. 



A series of Imrried marches brought me back once inoic lo the 

 vicinity of the present inhabited area. Various anti(juarian and topo- 

 graphical considei'atlons made me look out in the desert north of the 

 oasis of Gulakhma for the remains of the town of Pi-mo, which Hiuen 

 Tsiang visited on his way from Khotan to Niya, and which is probably 

 mentioned also 1)}' Marco Polo imder the name of Pein. After a search, 

 rendered difficult by the insufiicienc}' of guides and the want of water, 

 I succeeded in tracing it in an extensive debris-covered site known as 



