112 THE ROUTE FROM KARGHALIK TO KHOTAN [Chap. V
graves are Muhammadan b. On examining one large but low mound, from the sides of which skeletons were protruding, I found that it consisted of closely packed rows of graves. These were lined with mud-bricks and covered on the top with small planks of wood. On having one of the graves opened, I found in it the remains of a child wrapped in the rough cotton material common all over Turkestan, and known as ` Kham '. The head was turned towards the west in conformity with Muhammadan custom, and the feet tied with a bandage just as is the fashion nowadays according to the assertion of my Turki followers. There could be no doubt that this cemetery had served for the burial of orthodox Muslims. Local tradition, in fact, maintains that it is the resting-place of ` Shahids ' who fell fighting the infidels. But I could not trace any definite legend, and my informants, who had shown no particular scruples about opening a grave for my inspection, readily expressed their disbelief when they saw the body of a small child in the supposed grave of a holy warrior.
Two observations make it appear very probable that this cemetery is old, and approximately of the same period as the remains of Togujai. From one of the partially exposed graves, Niaz Akhûn, my Tungani interpreter, picked up a small brass ornament representing a flying bird, exactly similar to one which had been brought to me while searching the débris at Togujai (marked M. oor. d and shown in Plate LI). I found on inquiry that it is still customary among the people to deposit small articles of this kind on the graves of their relatives.
The other observation, bearing on the configuration of the ground containing these old graves, is also of interest, but did not strike me until I had gained further experience elsewhere. The graves, as far as I could observe, invariably occupied low mounds not unlike the loess-banks or `witnesses' described above in connexion with the Tati of Kakshal, but usually larger and more rounded in shape. The fact that the contents of the graves nearest the edge of these mounds were generally exposed to view proves that the latter were subject to erosion. The ground between the mounds showed a surface of soft disintegrated loess, strewn here and there with fragments of old pottery, but not thickly. It thus resembled exactly the depressions covered with loess dust which at Tatis mark the maximum erosion of the ground. The elevation of the mounds containing graves above the rest of the ground was not great, perhaps nowhere more than 8 to ro feet.
Comparing my notes and recollections of this old burial-place, I am led to the conclusion that we have in the mounds really portions of the original ground level which the compactness of the graves has preserved from the erosion proceeding all round. This conclusion is confirmed by what I saw at Kepek-gholuk Mazar, a little shrine about half a mile to the east of the most conspicuous mound of Hasa. It consists of a collection of staffs and flags, erected on a small well-defined loess-bank, and is believed to mark the resting-place of a saint. A photograph I retain of this ` Mazar ' (Fig. i8) shows very clearly the effects of wind-erosion on the sides of the loess-bank. The latter rises about ten feet above the surrounding flat ground, and manifestly owes its preservation to tombs placed on its top. Pieces of much-decayed timber sticking out on the upper slope probably belong to tombs that have been cut away through erosion.
Drift-sand in low dunes now encircles Hasa from the north and north-east, forming a fit setting to the dreary scene ; yet there is no proof that the site of Togujai, from which this burial-ground probably received its occupants, was abandoned owing to an advance of the desert. Even at the present day there are plots of ground under cultivation north of Togujai, and the inroads regularly made there by flood-water show that irrigation would not be impossible.