On the 26th December the wind still continued to blow out of the south-southwest, and the atmosphere was thick with dust. The sand-dunes still continued to be high, though not so high as they were close to the Tarim. The bajir depressions now began to decrease sensibly in size, and although in the course of a march of close upon 22 km. we traversed eight of them, they profited us but little, for the greater part of the time we were travelling over heavy sand. In fact they were a hindrance rather than a help, owing, to our having continually to descend into them and then climb up out of them again at the other end. In several of the bajirs we traversed to-day the desiccation rings were plainly indicated, whether they were due to rain or to the proximity of the ground-water. But the former cause grew increasingly more likely in proportion as we advanced, for with every step we were drawing nearer to the Kwen-lun mountains; and the rains which fall there do reach a good long way out into the desert. The desiccation rings were especially plainly marked in bajirs Nos. 22, 24, and 28. The last-named was of an unusual shape, in that two »bays» were thrust out from the main body of the bajir, one to the south, the other to the south-east. Judging by the desiccation belts, the latter »bay» was the deeper. Between them lay an accumulation of sand resembling in appearance the peninsula in the Ullugh-köl. Strictly speaking, there were two depressions, parted by a broken dune-accumulation. In bajir No. 22 there was but little of the bare soil left; in fact its two thresholds appeared likely to meet ere long, when of course the bajir will totally disappear. In the northern part of bajir No. 23 there were several small level patches, entirely surrounded by low dunes. Between bajirs No. 23 and No. 24 there had been another bajir, now so filled with sand that we were unable to make out its boundaries, the sand being for a long distance uninterrupted. The northern part of bajir No. 25 was filled with small ridges of sand, stretching from north-east to southwest; it was quite patent to the eyesight, that this bajir was deepest in the south. No. 27 was a miniature bajir, or rather three small level patches which we did not think it worth while to go down into.
Thus during the course of the day the going had grown appreciably worse; the sand was obviously getting the upper hand of the level ground. Although we still had the steep wall of sand on our left, and its architecture was still unchanged, the trough or interminable sandy valley we were following was now more directly inclined to the south. At the beginning of the journey it pointed to the south-west; now however it pointed to the south-south-west and even for short distances to the south. The slopes up to the crests of the thresholds were gentle, an angle of 4° to 5°, but the individual dunes which crowned them were generally disposed from north-north-east to south-south-west. These appeared to be chiefly affected by the east wind, whereas the lowest part of the thresholds, stretching northwards, were most exposed to the effects of the north-north-east wind; hence the capping dunes of the thresholds lay crosswise, from east to west. But the descent on the other side was one unbroken slope to the south-south-west.
One great difference between this desert and the Takla-makan is, that in the latter the patches of bare ground are not only excessively rare, but occur in a most irregular way. It is true that the dune-accumulations are also dome-shaped in the Takla-makan, but they melt more into one another, and are subject to less regular
IIIF Color HighRes
