AN EXCURSION UP THE EASTERN WATERWAY OF THE TARIM DELTA. 463
of the lake issues a canal, which near its embouchure divides into two arms. Of these the upper one had a breadth of 2.27 m.; a mean depth of 0:54 m.; a mean velocity of 0.88 m.; and a volume of Los cub. m. in the second. The lower branch consists of a shallow part (5.3 m. broad, I dm. deep, 0.4 m. velocity, and a volume of 0.2 cub. m. in the second), and a deep part, of the following dimensions: breadth, 3.3 m.; mean depth, 0.79 m.; mean velocity, 0.83 m.; and volume, 2.164 cub. m. in the second. Thus the total volume amounted to 3.444 cub. m. in the second. Later on, I ascertained that this water came from the lake of Märdäk-köl, which I discovered in i 896. The name of the village indeed alludes to the fact. But the really interesting feature about this new river-system is that it points to the correctness of the idea I first threw out in 1896, namely that the water in this region is tending towards the east, just as the Kara-koschun is tending towards the north, in other words, is returning to the former bed of the Lop-nor. Whether the Märdäk-köl, that is the new lake situated east of Märdäktik, sends out any arm towards the east, I am unable to say; if it does, the arm in question would be analogous to the Tokus-tarim, and it is not unlikely that sooner or later it will be formed, if it is not formed already.
Eight years ago one of the older men of Märdäktik, who was by birth a native of Tschaj, carne here by chance in quest of new fishing-grounds. At that time there was only little water; but in the very next year the quantity increased rapidly, and then again four years ago there was another noteworthy increase. But within the memory of man the water had never risen so high as it had done that year 0900. Nine years and farther back there existed in this same locality nothing but a series of salt-pools (daschi). At that time, too, the river Ilek was very different from what it is now; for it then divided into two branches, of which the eastern one made its way to this point, where it formed the lakes, while the western branch again divided, sending one arm to the Tarim at Almontschuk and the other at Talkirtschin, in a district called for that reason Kujusch; here there was a lake which has now disappeared. Thus we again have proofs of how exceedingly changeable these waterways are; scarce a year passes without some radical change or other taking place, so that no detailed map holds good for longer than a short time.
My informant told me further, that when the ice broke up in the lake at Märdäktik — which the natives call simply »köl» — the level was one karetsch (0.22 m.) lower than at the time of our visit, and that it was only during the last ten days that it had risen, the cause being palpably the mus-suji. In the height of the summer the level is said to be a good deal lower than it was just then. To the south they were able to give me the names of only three lakes, namely the Kötäktik-köl, Ghilam-köl, and Nias-köl, all plainly either forming part of or connected with the Tschong-köl. The Ghilam was no doubt one of the lakes we had just traversed, this name being unknown to my canoe-men from Schirge-tschapghan. Here too the cold had been more severe than usual that year, and the precipitation more abundant. But the people had not observed any fatality amongst the fish, such as we noticed at Kara-koschun.
On the 22nd April we continued our journey northwards from Märdäktik, crossing first the lake of Sadak-köl, which is on the whole very shallow, although
IIIF Color HighRes
