348 THE LOP-NOR PROBLEM.
of the lower Ilek; for instance, we have found that the Sadak-köl had become in great part filled in the course of only four years. The Tarim lakes in particular are certainly condemned to early destruction; in fact they are ambulatory basins, and when they get full of sediment, and the river has built up for itself fresh containing ramparts, it will soon quit its bed and form new lakes by the side of the old ones. Higher up, where the direction of the river and the long axis of the lakes no longer coincide with the orientation of the bajir depressions, the lakes are probably often framed about by old marginal ramparts built up by the river at still older stages of its existence. Accordingly these lakes belong to a more transient and more undecided type of lacustrine formation than do the Avullu-Tajek-Arka-köl series.
If now the lakes last named are situated, as I have proved, in former bajir depressions, the question arises, why is it that these depressions are deepest just in this part of the desert? In other words, why is it that a trench or hollow runs right across the desert all the way from Lop-nor to the vicinity of these lakes, a trench that is moreover parallel to the line of the Kara-buran, the lowermost Tarim, and the Kara-koschun? Presumably the reason is, that during the time the Tarim flowed along the bed of the Kuruk-darja, it was just the bajir depressions in the very middle of the desert that were most developed, possibly because the wind
' there, midway between the Kuruk-tagh and the Astin-tagh, put forth its greatest energy.
The existence of these typical bajir depressions affords a very strong, although I admit an indirect, proof, that, during a pretty long time, this part of the Desert of Lop at any rate was entirely free from water, and the modulation of the surface is due to no other factor but the predominant wind. In the Desert of Tschertschen we were able to convince ourselves that bajir depressions can only originate in a sandy desert, and that consequently each individual bajir is fenced in on west and east by a long ridge of sand. Accordingly it is quite natural that, if we proceed east from the vicinity of Avullu-köl, we find the bajirs decrease in size proportionally with the diminution in the volume of the sand, until finally they disappear, even as the sand does. The parts of the Desert of Lop which I crossed exhibit not the slightest trace of meridional bajir grooving, except that which runs from north-east to south-west, and is a direct consequence of the erosive action of the wind.
I have also assumed, that the prevailing east-north-east wind, which is so typical of the Lop country, grows increasingly less constant towards the west. All the same it is impossible to overlook the situation and direction of the Chotan-darja and the Kerija-darja. Both these rivers flow parallel to the chain of bajirs through which I marched in the Desert of Tschertschen; and although their courses are determined by the trough-like basin of East Turkestan, it is not impossible that they may also have been to some extent influenced by the relief and configuration of the desert itself. The lines of fall, which they suggest, taken together with those of the Jarkent-darja and the Kaschgar-darja, form, like the similar lines of the bajir depressions in the Desert of Lop, the radii of a circle, although in this instance the orientation is more excentric. The north-north-west direction of the Nija-darja, Tolanchodscha, Bostan-toghrak, Möldscha, and Kara-muran agrees perfectly with the position of the depressions in the southern part of the desert. For instance, if the Bostantoghrak carried a sufficient quantity of water to be able to flow all the way to the
IIIFカラー高解像度
