crest-line running pretty evenly. In places on the flanks were thin strips of snow. On the level ground the snow lay however in far less quantity than the day before, and in fact it soon ceased altogether. The left-hand, northerly range now bore a closer resemblance to a chain of low rounded hills, though nowhere did it exhibit any breach ; the eastern limit of the transverse breaches is Ku-schui-cha. The surface of the latitudinal valley consisted of soft earth, with little or no gravel. At first there were no eroded watercourses in this valley, and it was only in the last portion of it that we passed a series of narrow and fairly deep ravines. These gather into a main channel, which clings close to the foot of the southern range and has a west-south-west direction; nevertheless it soon disappears in the ground.
On the southern side of the northern range, which here again is somewhat more massive, there occurs a transverse glen known as Ja-ma-tschan. Through it runs a deeply excavated drainage channel, and in its bed, a short distance above its mouth, the spring of Ja-ma-tschan gushes out; it was then completely frozen over, but is said to be nevertheless salt. All the same the sheets of ice were much trampled by kulans, arkharis, and wild yaks. And curiously enough we saw some of the last-named on the northern slopes of the range, although I scarcely expected that these animals would have ventured to visit such a comparatively remote mountainous region. Still they would be less disturbed there in the winter than in the summer.
At the spring (alt., 3024 m.) the rock was the same dark schist as that which I have last described, and its dip was 78° towards the N. 12° W.
Immediately east of the spring of Ku-schui-cha is a lofty stone pyramid, with a slab bearing an inscription in Chinese script, though whether its object is to proclaim, that the country there belongs to China or to celebrate some wretched
IIIF Color HighRes
