228 THE DESERT OF LOP.
of these groups of trees was just as low-lying as the bottom of the trench we were travelling in, the terraces may equally well have been islands in the former lake, and the gully between them, which ran north-east to south-west, may equally well have been hollowed out by the wind.
//
Fig. 141. LOW JARDANGS.
During this day's march the appearance of the desert changed again. There was toghrak forest everywhere, the trees occurring always in small scattered groups, or a few grey trunks together, all without branches. The tamarisks and kamischfields varied in both thickness and extent of distribution. The mounds were generally 3 to 4 m. high, seldom 5 m. The drift-sand gradually increased in quantity, forming small rudimentary dunes, though seldom more than one meter in height. But this, the first belt of sand we encountered, soon came to an end. After that, when it began again, it continually increased in quantity, and as it did so the jardangs decreased. Indeed in the vicinity of the first belt of sand they were already sparse, and only occurred at intervals amongst the dunes; and they were seldom more than one meter high. At the beginning of the second belt of sand the stratum of the clay terrace which I have called storey no. 2 was rare, and stunted, and consisted of extremely small, short, and ill formed ridges. In consequence the desert grew more open, and there was nothing to hinder the view. Storey no i, or the clay base upon which the jardangs stand, is itself seriously attacked by the wind, being already grooved from north-east to south-west, but the jardangs so formed are only one foot high. All the same they form very serious obstacles in the path of the traveller, and would make a journey to the south-east extremely toilsome. In one depression or hollow, 6 m, deep, and stretching east and west, we found in our path an especially large accumulation of sand. But the wind-erosion had nothing to do with the hollow: it had manifestly been scooped out by running water, and was the relict of some river-arm, a memorial of the Tarim's restless wanderings between the northern and the southern depressions of the Desert of Lop. We had already discovered similar river-beds on the north-west shore of the Kara-koschun; and the Tokus-tarim too belongs to the same chain of parallel river-beds, which one after the other have carried either the entire river or at any rate a part of it.
About one-half of the area to the south-west of this river-bed was occupied by dune-sand; the other half consisted of bare wind-furrowed clay, with the thinnest scattering of vegetation. The first living vegetation we came across in this direction consisted of a number of tamarisks, rather small and languishing, growing amongst low dunes. These again appeared to form a strip stretching from west to east, or rather perhaps from west-south-west to east-north-east, and consequently parallel to some
IIIF Color HighRes
