468 SOUTHERN DZUNGARIA 



and which stamps the entire region with an undeniable 

 character of its own. 



The change was so sudden, so complete, that we did 

 not hesitate in forming an opinion that the desiccation 

 at work on the south side of the Tian Shan does not 

 extend into Dzungaria. B}^ a curious chance our 

 route at first led us into a small, self-contained basin, a 

 kind of miniature Turkestan, with its special features and 

 peculiar conditions reproduced for us on a miniature 

 scale. 



This isolated basin, which I will call after the guard- 

 house and inn situated in its centre — Chi-ku-ching, lay 

 immediately below us as we left the gorge in the moun- 

 tains and entered the plain. It is a small basin ; its area 

 could be covered by that of Middlesex. On the north 

 it is bordered by the declivities of the Tou-shui plateau 

 — where a small amount of water drains into the basin 

 at certain seasons of the year ; on the south a semicircle 

 of low desert hills connect up with the northern wall and 

 render the basin complete. The road from Kumul to 

 Turfan crosses it from east to west, Chi-ku-ching forming 

 the junction of the Guchen road. The group of three 

 or four houses and the rather brackish water-supply 

 found at a depth of 20 ft. interested us little, the chief 

 feature of the basin being the unmistakable signs of 

 desiccation in actual progress. 



The centre of the basin, at its lowest level, was covered 

 with tamarisk mounds. These mounds stood as high 

 as 20-25 ^^■> ^^d were perfect t3^pes of aeolian action — 

 cut away underneath by persistent winds, and increased 

 above by the deposition of blown sand, with a scanty 

 growth of tamarisk on the summit. According to Pro- 

 fessor Huntington's theory as to the approximate age of 



