30 GEOLOGICAL RESEARCHES IN 



direction gradually ascending, the flat-topped hills of the table-land shutting us in 

 on both sides, till we reach a watershed from which we look down on a large, deep, 

 circular valley, covered with grazing herds, and ornamented with the gilded spires of 

 a lama-temple. This valley is shut in on the north and west by the volcanic forma- 

 tion of the plateau, but its southern wall is of granite and gametic gneiss, capped 

 here and there by thin remnants of the plateau mantle. Still farther south, after 

 passing the village of Yingmachuen the plateau formation predominates, and the 

 long descent into the valley of the Te Hai^ is entirely over its rocks. 



The great depression of the Te Hai is about twelve miles broad, and so far as the 

 plateau is concerned, appears to be open to the S. W. in the direction of its longer 

 axis. The northwestern side is formed by a serrated range of mountains, which 

 rises about 2,000 feet above the lake, between this and the plateau. The eastern 

 wall is of gneiss capped with the volcanic plateau formation, and the same would 

 seem to be the case with the southern wall, while, as we have seen, the northeastern 

 side is volcanic in its entire height. Thus the thickness of the volcanic mantle 

 varies, within a few miles, several hundred feet. 



The northeastern end of "the valley contains an extensive deposit of the terrace 

 loam. This faces the lake with a bluff that stretches N. W. S. E. across the valley. 



From this line the terrace rises toward the N. E. at first gradually, and then 

 rapidly, until in the long northeastern arm of the valley and in the side valleys, its 

 surface is several hiindred feet above the lake. 



Below this terrace a plain rises gently from the lake toward the mountains. 



The terrace deposit is a firm, stratified loam, containing, near the hills, numerous 

 fragments of the neighboring rocks and layers of gravel. It is cut into by deep 

 ravines, in the sides of one of which, about five miles east of the lake, I found 

 several species of fresh-water univalves. 



The lake is apparently about eight miles long by four or five broad. Its water 

 is salt, though far less so than seawater, and is not bitter. The flat surrounding it 

 is covered with a thin coating of soda efiiorescence.^ 



While the valley of the Kir Noor is occupied exclusively by the Mongols and their 

 herds, that of the Te Hai is cultivated by Chinese, only one or two Mongol camps 

 being seen. Ancient watch towers, that dominate these plains, and from which 

 signals could be made to the long line of similar posts on the Great Wall, are silent 

 monuments of a time when the shores of these lakes were the home of an aggres- 

 sive race, ever threatening a descent into the fertile regions of China. Rising with 

 the terrace, the road leads us to the hills that form the southeastern wall of the 

 valley, and we pass through these by a deep and rocky ravine, in which the pass is 

 situated. These hills are, as I have already said, of gneiss, characterized by an 

 abundance of garnets, and capped with the volcanic mantle. The stratification 

 trends, in the main, N. E. and dips 75° to N. W. Garnctiferous granulite, from these 



* Daikha Noor of the Mongols. 



^ For negative results of a microscopical examination of the deposits, both of the terrace and the 

 flats, see Nos. 2 and 3, in Mr. A. M. Edwards' Letter, Appendix No. 3. 



