352 THROUGH THE LOP DESERT [Chap. X
returned from the proposed half-way depot. As the donkeys could be given a good drink here, it became possible for us to take them and their ice loads two marches further out into the desert.
The point we had now reached by the shore of Chainut-köl had its special importance for the determination of the route which was to take us to the ` Lou-lan ' Site. From Dr. Hedin's Central Asia and Tibet and the sketch-map accompanying the first volume,5 I could see that the track we had so far followed was the same which had brought him to the lowermost Tarim after he had struck the north-western edge of the present Tarim delta or Kara-koshun on his journeys, in i 900 and 1901, from the foot of the Kuruk-tagh across the Lop desert. Tokhta Akhûn's testimony left no doubt that the place of our depot was practically the same as that at which Dr. Hedin had encamped, in March 1901, by the Chainut-köl.° Neither his big publication Central Asia nor the detailed maps accompanying it were then or before available to me. But the previously mentioned sketch-map sufficed to show that, in order to reach the ruined sites first discovered by him, I should now have to strike a route to the north-north-east, which would necessarily lead near those routes
which he had followed in the reverse direction, and actually between them if our course was kept quite straight. But in the wastes of sand and clay before us there would be nothing to guide us except the compass. Neither Mullah nor Tokhta Akhûn had ever visited the ruins from this side. However, for about a day's march further to the north the ground was still familiar to them from former hunting trips, and this allowed them to give me assurance on one point of importance. The large, shallow, newly-formed lake, which in March 1901 had obliged Dr. Hedin to make a considerable détour, had since almost completely dried up, leaving only scattered salt-water lagoons. Hence, provided that the position assigned to the ruins of ` Lou-lan ' in Dr. Hedin's sketch-map was approximately correct, I could safely steer my course by the compass without having to fear détours and consequent loss of time.
March After proceeding two miles further, we struck the southern end of a salt lagoon which lay
across dry within the area once covered by the Yangi-köl or ` New Lake ', and to which the Loplik fishermen lagoons of
Yangi-köl. now specifically applied that name. No fresh influx of water was said to have reached the narrow
curving depression for three years, and the fish in it seemed to be dying off rapidly owing to the increased saltness of the water. We found a great quantity of them stacked for drying near a reed-hut, and plenty of moribund or benumbed fish were caught by the men with their hands from under the thin ice just forming. Large stretches of boggy, salt-covered soil surrounded this lagoon and others that we passed further on, attesting the rapid shrinkage of these remnants of the ` New Lake '. Beyond a large crescent-shaped lagoon, which still retained water in its deeper northern portion, we came upon a succession of basins which showed signs of having dried up recently. We had marched across the largest of these, known to the hunters as Kurban-kullu-köl, for about a mile and a half, when the abundance of young reeds and tamarisks on its north bank induced us to halt there for the night.
The information I gathered from Mullah and Tokhta Akhttn in thé course of the day's tramp accounted for the notable change which had taken place in the physical aspects of the ground since Dr. Hedin last saw it. According to them, the basins of the Yangi-köl area were filled with fresh water by the spring floods of the Yangi-su branch of the Tarim for three years after his first visit in 1900. Since then (1903) no water had reached them, and the lagoons continued to shrink, while what water was left turned more and more salt. They had known the Yangi-köl depression since their youth, from hunting expeditions after deer and the like, and the opportunities I had, both on this and my subsequent visit to the Lop region in 1914, for testing various points of the information
6 See Central Asia and Tibet, i. pp. 891 sqq., ii. pp. a See route maps, Hedin, Central Asia, i. Pl. 21 (19oo,
173 sqq. March 22—April 7) and ii. Pl. 53 (1901, March 20—April 7).
Determination of route to Lou-lan Site.