to the same long gully. In places the jardangs approached so close together that it was as much as the camels could do to get between them. Here we were still able to distinguish the three different stories, indeed we could sometimes even make out four, although the fourth, which does not exceed r o m. in height, is rudimentary and rare. Towards the close of the day's march the separate layers became more and more fused together, and it became increasingly difficult to distinguish the step-like divisions between them. At the same time the jardangs were more widely separated and grew smaller in size, the desert itself more open, and the gullies less energetically excavated.
Fig. 139. SHOWING THE HIGHEST KIND OF JARDANG.
Although the sand continued to be extremely little in quantity, nevertheless it increased as we advanced south, and in the vicinity of Camp No. XV we perceived a couple of tiny dunes to the south-east and south-south-east, though all day not one dune in the proper sense of the terni. Occasionally there was a little sand, a decimeter thick, in the gullies, and in one place a pretty large space was coated with a similar thin layer. Looking north-west we could not help noticing the slow, gentle slope of the saj or detritus-slope stretching from the foot of the mountains towards the clay desert. It is plain that Lop-nor once sent out a large bay to the north-east, namely that which we crossed over in February 1901. This I infer, not only from the direction in which the saj extends, but also from the fact that from Altmisch-bulak the clay desert can be seen continuing a long way towards the east, though swinging round to the north-east as it were along a once convex shore.
But the most striking change is that the vegetation once more puts in an appearance, and is indeed in places particularly luxuriant. First come a few very ancient tamarisks, dried up and brittle as glass, standing for the most part on high mounds, though sometimes also on the level ground. They appeared in greatest quantity on the left of our route, that is to the south-east. The stratification of the clay is just as distinct in these mounds as elsewhere in the desert, a proof that they were carved out of the general mass by the wind, and did not grow on what was originally the level ground. In these mounds there would appear to be several varieties of structure; but one law is common to them all, namely that their skeleton or framework is always composed of the roots of the bushes, and it is these roots which hold them together. In certain directions, and over pretty extensive areas, dead reeds and sedge are very plentiful. Next we threaded a belt of toghrak forest, the trees standing upright and being often pretty coarse-stemmed ; they stretch in two directions, S. 62° E. and S. 6o° W., meeting at a common point, and undoubtedly mark a portion of the former lake-shore. Everything points to the fact that the vegetation which formerly existed here was especially thick and vigorous. By far the
IIIF Color HighRes
