CHINA, MONGOLIA, AND, TAPAN. 29 



growing in this gully, nor did wc meet with any others either on the plateau or 

 in its valleys. 



The lake is said to be drying up, and the Mongols say that its Avaters have flowed 

 into the Te Hai farther west, an apparently unfounded belief, as there is no surface 

 communication between the two lakes, and the natives on the shores of the Te Hai 

 were not aware of any increase in its volume. Still it is evident that the waters of 

 the Kir Noor are rapidly disappearing, and the cause, whether this be only tempo- 

 rary or a constantly operating change in the climate, has been acting for at least 

 several years. Among the lakes we have already noticed, the Chaganoussu is also 

 disappearing, and the adjoining Hoyur Noor has for several years been represented 

 only by its dry bed. 



The greater part of the plain of the Kir Noor valley is clothed with grass, and 

 supports large herds of sheep, but as we approach the recent lake-bed the surface is 

 eroded by dry, shallow water-courses, and is covered with tufts only of grass, 

 between which the groimd is bare and cracked. This was apparently a marsh sur- 

 rounding the lake of which, a little further west, the dry bed is visible covered with 

 the white soda efflorescence, and stretching several miles west, north, and south.^ 



The walls of this great valley, formed by the abrupt edge . of the plateau, are 

 marked by a series of lines at different heights, and extending apparently hori- 

 zontally, and on the same level, along the faces of both sides of the valley. They 

 are reproduced on an island-like hill that rises from the plain, and are visible at a 

 distance of from ten to twelve miles to the naked eye. They are defined, where the 

 slope is gentle, by a continuous mass of large and small fragments of rock, and on 

 the steep declivities by slight variation in the angle of slope. 



I was able to examine these lines in only one locality, and there they appeared 

 to be independent of the structure of the plateau, and I can account for them only 

 on the supposition that they mark former water levels. 



Following the road from Hoyurbaishin to the Te Hai we cross, at about the middle 

 of the valley, a small stream of fresh water flowing from the north, and which is seen 

 to empty into the. remnant of the lake a mile or two south of the road. StiU farther 

 west the road lies through a marshy tract. Two or three miles west of this we 

 reach a terrace of the lake-deposit, which descending rapidly from the western side 

 of the valley, faces the plain with a bluff. As the road ascends a ravine in this 

 terrace, the increasing proportion of fragments of granite and gneiss shows that we 

 are in the neighborhood of a rise in the granite foundation, while a few miles to the 

 north a ridge rising several hundred feet above the level of the plateau, seems to be 

 the source of the fragments in question. 



As we leave the terrace and the valley of Kir Noor, we pass a deep and gloomy 

 gorge cut through the plateau to its very foundation. Where seen it is barely 

 separated by a low ridge from a valley that leads into the Kir Noor. This chasm 

 seems to lead to the Karaoussu, a tributary of the Tourgen Gol, which is an affluent 

 of the Yellow river. The valley by which we leave the plain leads us in a S. S. W. 



' For the results of an examination of the dried mud of the recent lake-bed, see Nos. 1 and 12 m 

 Mr. A. M. Edwards' Letter, Appendix No. 3. 



