EXPLOKATION IN CHINESE TUEKESTAN. 7fi5 



entire!}" supported by the evidence of the numerous old Chinese coins 

 I found at the site, the latest bearing the symbols of the dynastic 

 period which corresponds to the 3'ears 713-741 A. D. 



The three weeks I spent in continuous excavations, from the earl}'^ 

 mornint^ until da3"lig'ht failed us, enabled me to explore all ruins trace- 

 able under the sand. It was a happy time for me personally, though 

 the ph3"sical conditions were trying. The severe winter of the desert 

 had already set in when I started from Khotan. During my stay at 

 Dandan-Uilik the temperature at night usually went down to a mini- 

 numi of about 10'^ F. below zero. In the daytime it never rose above 

 freezing point in the shade. The weather was cloud}^, but luckily 

 there was very little wind. Its absence is an essential condition for 

 all prolonged work in the desert. The dead trees of the little orchards 

 which once surrounded most of^ the scattered groups of shrines and 

 dwellings supplied fuel in plenty. Yet the men suffered from the 

 exposure as well as from the badness of the water, the only available 

 supply coming from a brackish well they had succeeded in digging in 

 a depression of the ground over a mile from the main ruins. My own 

 little tent, brought from India, though provided with an extra serge 

 lining, was a bitterly cold abode at night. When the temperature had 

 once gone to about 6° below freezing point, writing or reading became 

 impossible, and I had to take to my bed, however anxious I might 

 have been to study the manuscript finds of the day, etc. But, from 

 long experience, life in a tent seems the one most congenial to me, 

 and, with such fascinating work to occupy me, the four and a half 

 months spent in the desolation of the desert were indeed an enjoyable 

 time. 



During my stay at Dandan-Uilik, Ram Singh had again joined me 

 from the direction of the Keri}^ River. L had dispatched him a 

 month earlier on an independent survey of the high range which 

 extends between "" Kuen-luen No. 5" and the mountains eastward 

 where connection could 1)0 obtained with Captain Deasy's work about 

 Polu. On comparing my own plane-table fixing for Dandan-Uilik 

 with his, a gratifying surprise awaited me. NotAvithstanding that we 

 had brought our survey from entirely different directions and over 

 great distances of such deceptive ground as sandy [)lanes and dunes, I 

 found that Ram Singh's ])ositi<)n diffci'od from my own l)y only about 

 a mile in latitude and a iialf mile in longitude. 



My detaihnl survey of the Dandan-l'ilik site, together with other 

 observations of a seniitopographical, semianti(iuai"ian nature which 

 gradually ticcumulated during my (\\plorations at this and other sites, 

 make it very ])r()hal)h> tiiat tlie lands of Danchui-Tilik were irrigated 

 from an extension of the canals which had, down to an even later date, 

 brought the water of the streams of Chiraand Culakhina to the desert 

 area due south of the ruins. I nuist reserve for another occasion a 



