sheet of ice and its walls yellow sand-dunes. Even in the sound that connects the lake with the river several sandy islets cropped up above the ice, together with some poplars which formerly grew on firm ground at the foot of the dune-wall. The artificial canal has thus lost its importance; but that it used to be dammed is plain from its name, daschi meaning »salt-pool». At the present time the lake forms, so to speak, a lateral expansion of the river, and being at the same level with it rises and falls as the river does. When the rush of the spring freshets comes, the water pours unchecked through the broad connecting sound into the lake, until it fills it to the brim. A comparison between the breached dune-wall of Daschi-köl and the big sharp elbow north-east of Sejt-köl indicates clearly in what direction the erosive energy of the river is working in the latter locality. The shape of the loop, and the arrangement of its silt deposits, are alone sufficient to show that the river is pushing towards the south-west, and gradually washing away the sand which bars its path. There is already a distinct cavity perceptible, and it may safely be affirmed, that the portion of the sandy neck which has already disappeared is greater than the part which is left, and that it will not be long before this last is completely cut through in the same way as at Daschi-köl. If, when this happens, the Sejt-köl were to be cut off from the river for two or three years, until its level very sensibly dropped, then as soon as the river established a direct connection with the lake, it would discharge a vast volume of water into the latter, and so suffer a serious »blood-letting» before the equilibrium of level was secured between the two. It is only when this step has been accomplished that the water flows freely in and out in the way it did at Daschi-köl on 7th December. There the river did not lose a drop of its water; it was rather the lake that was the loser, in consequence of the river's slow, gradual fall. The same phenomenon is repeated lower down at each of the other lakes, but in the case of none of them is the true situation so distinctly seen as here at the Daschi-köl. The summit of every dune afforded an especially vivid and instructive view of the river's victorious struggle against the dunes. The strip of land between the river on the one side and the Sejt-köl and Daschi-köl on the other is extremely narrow, and even yet narrower the dunes which overhang it. When these last have been finally swept away, it is very probable that the river will shift its bed a good distance to the right by directing its path through the lakes. Then the lakes will be gradually filled with silt and levelled up, except for the trench which the main current carves out for itself through the alluvial deposits. In this way these lakes facilitate the transference of the river-bed towards the right, that is towards the south-west. Except when the stream thus suddenly flings itself into some lake, or series of lakes, the movement of its bed towards the right takes place very slowly, but none the less surely, along the whole of its course. When the river thus violently changes its bed, it might perhaps be supposed, that a portion of the dunes which have been breached would still continue to remain in situ on the left bank of the Tarim. In reality however that happens very seldom; at all events there are no independent, detached dunes on that bank of the river. Indeed on that side there is no sand at all, except Igis-dung and a few other fragmentary patches, all of which have been already mentioned. Nevertheless it is quite conceivable that formerly there were such dunes at several places, but they have been blown away
IIIF Color HighRes
