Sec. iv] ANCIENT REMAINS AT MOJI 111
to issues of the Chien-yüan period (758-76o A.D.) of the Emperor Su-tsung, and a small number of uncertain Muhammadan copper coins. It is highly probable that these coins, too, come from Togujai. [Nine coins have since been identified by Prof. Rapson as probably belonging to Muhammad Arslan Khan, an early successor of Satok Boghra Khan.]
In view of the evidence of these coin-finds, it may be considered as certain that the remains of Togujai belong to a settlement which flourished well into the period following the Muhammadan conquest. The layer from which the pottery fragments and other débris are washed out bears a close resemblance in character to the culture-strata of Yôtkan, the site of the ancient capital of Khotan, to be discussed hereafter ; like them it is undoubtedly due to the gradual accumulation of rubbish. But the slight depth of the Togujai layer and the absence of earlier coins suggest that the occupation of the site, or at least that which is traceable in remains, does not date back to more than a few centuries before the advent of Islam. On the other hand, we have an indication of the lower chronological limit in the deposit of earth which covers this layer. The observations subsequently made at Yôtkan prove that this deposit must be attributed to prolonged irrigation of the ground after it had ceased to be used for dwellings 2. Within living memory of the villagers the area had been nothing but a waste, and there seemed to be evidence that during this later period its surface had in fact undergone some erosion. Combining thus the evidence afforded by the present condition of the site with that of the deposit of silt overlying the débris layer, we are led to conclude that even the latest portion of the latter must have rested in the ground for a considerable period, probably not less than four or five centuries.
It is of interest to arrive even at an approximate determination of the highest and lowest limits of time, as this helps us to classify the remarkably varied and curious specimens of pottery which the site yielded, besides other small objects in glass and metal. The glazed pottery fragments described in the list below, and partly reproduced in Plate XLII, claim particular attention owing to their relatively good technique and artistic colouring. None of the earlier sites explored by me yielded glazed ware ; and hence it may be questioned whether the art of glazing was known, or to any extent practised, in Eastern Turkestan until towards the commencement of the Muhammadan period 3. On the other hand, some of the unglazed pieces show in their decoration the survival of motives which are already met with in the wood-carving of the Niya River Site (third century A. D.), and can be traced back to Graeco-Buddhist art 4.
As at Yôtkan, I was much struck by the complete absence of any traces of structural remains below the ground. But this feature is easily explained by what we shall have to show hereafter as to the effect of irrigation and concomitant percolation upon sun-dried bricks, plaster, and timber—the sole building materials of Turkestan. There is another point of contact between the conditions presented by the remains of Yôtkan and those of Togujai. Just as the discovery of the site of the ancient capital of Khotan was brought about by the accidental formation of a ravine which laid bare its deeply buried ` culture-strata ', so Togujai, too, would in all probability have kept its old and more interesting débris hidden but for the channels scooped out by the flood-water.
From Togujai I proceeded over waste ground, covered in parts by light drift-sand (or loess dust), to an old burial-ground known as Hasa, about z s miles to the north-east of Moji Bazar. Dr. Hedin, whose special attention had been directed to this site, rightly observes that the
2 See below, chap. VIII. sec. ii. which, as we shall see, was inhabited up to the end of the
3 The only other piece of glazed terra-cotta in my thirteenth century of our era; see chap. xui.
collection comes from Uzun-Tati, the ancient P1-mo, a site 4 See, e. g., description of T. M. 004. a, below.
| || || || || || || || |