THE SELLING-TSO ANI) THE JAGJU-RAPGA. 35
rainfall over its drainage-area is diminishing. It is however of course impossible to say how far the shrinkage is permanent, although it is probable that it is so. The farther we travelled west, during the further course of the journey, the more evident were the signs of higher lake-levels having existed at former dates, and consequently the greater was the extent of the shrinkage. In the case of the Selling-tso I estimated that the highest rampart was about 5o m. above the present level of the lake; subsequently I came across similar ramparts, at Lakor-tso, at a measured height of 133 m. above the existing lake-level. In the case of the lakes situated farther west, the old beach-lines are best developed on the sides that face west and are exposed. In the case of the Selling-tso however they appear to be developed most distinctly on the shore that faces east. This may possibly be caused by different wind relations; but I am not in a position to make any comparison in this regard, for I have not visited the eastern side of the Selling-tso, and old strand-ramparts may be developed there to an even greater extent than they are on the northern and western sides of the lake. It is most probable that the same winds prevail here that prevail in western Tibet.
September I oth. At the very beginning of the clay's march I had an opportunity to observe a peculiar form of surface, which is also undoubtedly connected with the shrinkage of the lake. On the south side of the valley of the Jagju-rapga rises a chain of craggy heights, not very lofty, but serrated and capricious. This chain grows increasingly lower towards the east and terminates in a rocky headland jutting out into the lake. Immediately south-east of Camp LXXVI there exists a gap or breach in the adjacent range. In the throat of this rocky gateway, through which we obtained an extensive view towards the south there rises a ridge, about 40 m. in height, composed exclusively of hard, compacted gravel and having a relatively steep slope towards the north. On the top it is flat as a platform; the slope that reaches towards the south is both less steep and less high. The entire ridge thus constitutes a broad rampart, stretching in a sinuous line from east to west. At its northern foot was a pool or lagoon, at that time not exhibiting any visible connection with the bay of the Selling-tso, from which it is only separated by a very low ridge. Immediately south of this and parallel with it runs a second ridge, presenting in the main the same characteristics as the first one. The only difference between them is that the second ridge turns its steeper face towards the south. Between the two lies a depression, which contained a pool; and between the southern ridge and the western wall of the rocky gateway there exists a well-marked watercourse. Now these two ridges are beyond doubt old beach-lines, although in default of an exact levelling it is difficult to make out clearly their former position in relation to the lake. I dare say that they formed together a narrow peninsula, with a lake in the middle. Possibly too the deep watercourse may have been a narrow sound, in which case the eastern part of the mountain-range would be entirely separated from the western part. Immediately south of the rocky gateway lies a third pool, bigger than the other two
The rocky walls that frame in this mountain gateway are steep and picturesque; the rocks, which consist of the same varieties as the escarpment-range in the broad peninsula, dip at an angle of 46° N. It is perfectly evident, that these
IIIF Color HighRes
