lagoons; one small portion of the range in the west appears however to plunge abruptly down into the lake. The Addan-tso forms, as we have seen, the receptacle for the surface-water that gathers off a pretty considerable area, the greater part of which lies apparently south of the lake. The reason why the Tschargut-tso does not receive any tributaries from the south is probably simply this, that the latitudinal valleys which lie on that side drain westwards into the more southerly of the two eastern bays of the Addan-tso, and eastwards into the upper course of the Alantsangpo, and more especially into the Selling-tso. The water of the Addan-tso flows out through the Tsangmo-rapga and empties into the Tschargut-tso, that stream being the only affluent of the latter lake, except for the brooks which enter it directly during the rainy season. After leaving the Tschargut-tso, which is thus nothing more than a »half-way house», a gigantic fluvial ganglion, the water continues through the Jagju-rapga eastwards into the Selling-tso. This lake occupies the very lowest part of the entire basin, and is therefore its terminal lake, so that its water is salt. In contradistinction to the Satschu-tsangpo and the Alan-tsangpo, and the other two lakes emptying themselves into the Selling-tso that we have already made acquaintance with, the Jagju-rapga maintains, we have assumed, a certain constancy of level throughout the whole of the year, the reason of this being that it receives its water from a rather large freshwater lake, the gathering-basin of a pretty extensive area; moreover the Tschargut-tso also acts as a moderator to its northern neighbour, the Selling-tso, which still further distributes and equalises the volume, tending to render it as far as possible independent of the seasons. I do not of course mean to say, that there is no fluctuation of volume in the Jagju-rapga, but it is probably very slight. To obtain absolute certainty upon this point would require a visit to that lacustrine region in winter.
It is however of great interest to turn to the very distinctly marked and beautiful old beach-lines and step-like terraces which occur on the northern shore of the northern bay of the Addan-tso and on the western shore of the northern peninsula. In the former locality I did not observe more than two terraces, but in the latter three, and of these the highest may be fully i o to 15 m. above the existing level of the lake; hence these terraces clearly correspond to the shore-ramparts which I had previously seen on the western side of the first rocky island in the Tschargut-tso. As for this peninsula, it is curious, that these ramparts are formed on its western shore only, but are totally absent on its eastern side, the explanation being of course the same as that which applies in the case of the island, namely the prevalence of westerly winds and the effects which they produce upon the western shore through the instrumentality of the violent waves and their surf-beat. The two lowest terraces on the northern shore of the bay and on the western shore of the peninsula belong of course together notwithstanding that they are in a couple of places broken, and even obliterated. There are no terraces on that part of the west side of the peninsula which lies under the shelter of the more northerly range on the southern peninsula. The existence of the two upper terraces are more difficult to explain, for both slope inwards towards each other, and thus can hardly be regarded as having anything to do with a higher water-level at some former period, but rather they would appear to represent the traces of the border-rim of
IIIF Color HighRes
