found in the fact that, whereas those sides which are exposed to the wind are greatly undermined and almost vertical, the leeward faces are intact, and present a more gradual slope down to those parts of the surface which are not protected by either houses or vegetation.
In an earlier chapter I have enlarged upon my conviction, that it is precisely the uneven distribution of the vegetation — trees, bushes, and reeds — which originally gave rise to these jardangs and to the gullies between them. After the country dried up in consequence of the Tarim having changed its course, the places that first fell a victim to the wind's excavating power were those that were distinguished by an absence of vegetation, or where the vegetation was thin and scanty. Other places were protected by the vegetation, and consequently formed elevations, that is jardangs. I look upon the 21/2 m. high platforms or pediments upon which these houses stand as an eloquent proof of the correctness of my theory. These powerful beams naturally form a much more efficient protection for them than does a solitary tree-trunk and its roots, or than a small patch of kamisch stubble. In consequence the platforms upon which the houses stand exhibit precisely the same shape as their foundations, and the position of the beams shows that nothing has been blown away from the top. In the jardangs that are held together by vegetation it is difficult, if not impossible, to determine how far and to how great a degree their upper surface has been filed down by the wind. It was for this reason that I was unable to regard the depth of the eroded gullies as a sure indication and measure of the effects of the wind-erosion. But this difficulty no longer exists in the case of the platforms we are considering; for they do show, and show unmistakably, that a layer 21/a m. deep has been planed off the circumjacent surface. Nevertheless one uncertainty still remains, namely this: we do not know at what epoch the Lop-nor shifted its position, nor how long after it did so the moisture surviving in its basin was able to nourish the vegetation sufficiently for it to offer successful resistance to the wind. But it is extremely probable that the basin of the lake dried up rapidly once the water had left it for another quarter.
I will now proceed to describe the large house-complex. At its south-southwest corner we found a quadrilateral apartment or house, measuring i 5.6 m. in breadth by 15 m. in length. Its walls were formed of basket-work made of tamarisk-branches, and were pretty substantial; but they only projected one foot above the slight layer of sand which had there heaped itself up under the shelter of the mound or platform. The fragments that projected above the sand appeared to suggest, that another similar apartment had originally stood immediately S.S.W. of the first structure. There was a mat, woven of reeds, almost entirely buried under the sand; it was precisely like the mats which are in use all over East Turkestan at the present day, partly to put under the ordinary felt-carpets (kigis), or under Chotan carpets indoors, on terraces, or in bazaars, partly also to hang up as awnings against the sun in the narrow streets of the bazaars in towns and villages, the reed-mats being supported by long thin rods of wood. What object the mat in question served it would be difficult to say; but, judging from the materials of which the walls were constructed, I should infer that the apartment was covered with kamisch mats, and probably served as an outhouse, stable, or so forth.
Hedin, yourney in Central Asia. II. 79
IIIFカラー高解像度
