EXPLORATION IN CHINESE TURKESTAN. 751 



Fa-hien and other ancient Buddhist pilgrims from China have left us 

 of their experiences on the journe}" through the gorges of the Indus. 

 From Ghulmit, the second stage onward, the scanty settlers occupy- 

 ing the few patches of cultivated ground in the valley proved to be of 

 Iranian origin, speaking a Wakhi dialect closely allied to that which 

 is used b}' the Wakhan immigrants found in Sarikol. Thus, in this 

 part of the Hindukush, too, the line of contact between the great areas 

 of the Indian and Iranian families of language does not completely 

 coincide with the geographical watershed. 



After six days spent in more or less continuous climbing, Misgar, 

 the northernmost hamlet of Ilunza, was reached, where I was able to 

 discharge the hardy hill men who had carried our impedimenta without 

 the slighest damage over such trying ground. On June 28 at last I 

 crossed, by the Kilik Pass (circ. 15,800 feet above the sea), into Chinese 

 territory' on the Taghdumbash Pamir, using the 3'aks of the Sarikoli 

 herdsmen, who, by Mr. Macartney's arrangement with the Chinese 

 authorities, had awaited me at tlie southern foot of the pass. 



From Koktorok, our lirst camp on the Taghdumbash, at an elevation 

 of over 11,000 feet, we commenced our plane-table survey, on the 

 scale of 8 miles to the inch. Throughout our travels in the mountains 

 I endeavored to supplement it, as far as ni}" limited time permitted, by 

 photogrammetric work, for which I used the excellent Bridges-Lee 

 photo theodolite kindly lent tome by Mr. Eliot, the head of the Indian 

 meteorological department. Systematic triangulation ])y theodolite 

 was started at the same time with the help of the points supplied b}^ 

 the surve3's of the l)Oundary commission and Captain Deasy, while 

 regidar astronomical obser\'ations for latitude were made ])y Babu 

 Ram Singh from here onward at all camps, the exact determination of 

 which possessed topographical interest. The constant and direct super- 

 vision which I exercised over the plane-table work enabled me to pay 

 special attention to the local nomenclature. A good deal of })hilo- 

 logical and historical interest attaches to the latter in regions like the 

 Pamirs and a considera))le portion of Chinese Turkestan, over which 

 have passed the waves of great ethnic migrations. I believe, there- 

 fore, students interested in this part of central Asian geography will 

 derive some advantage from the pains I took to correctly ascertain and 

 to record with phonetic accuracy all local names throughout the terri- 

 tories covered l)v our surveys. 



From the height of the Khushbel peak, the first •'hill station"' of 

 our survey (close on 17,000 feet above the sea), I could simultaneously 

 see the ranges which form the watershed between the drainage areas 

 of the Indus, the Oxus, and the Yarkand rivers, and which politically 

 divide the territories of British India, Russia, and China, Afghanistan 

 (pi. i). Pressed for time, as I necessarily was in regard to all that 



