274 ANCIENT SITE OF ENDERE [Chap. VII
and of great age, and these, characteristically enough, were inside the wall and were found mostly in open spaces left between the quarters. Evidently they had grown up there before the latter were erected and had been spared for the sake of their shade when clearing and building began. The way in which these patriarchs of the desert jungle had preserved most of their minor branches (see Fig. 76), was a sign that their death had taken place in a comparatively recent period. Everything supported the impression that the abandoned settlement belonged to the Muhammadan epoch and to a not very distant part of it.10
If we take into account what can safely be inferred from the uniform roughness of the dwellings and the total absence of refuse heaps, the following suggests itself as the most likely explanation. At a time when the Endere River was following a terminal course lying west of its present bed (or beds) and perhaps connected with the dry river-bed traceable higher up near Korgach and Tokuz-köl (see Map-sheet No. 40), it must have been easy to bring water to the wide open plain now covered with tamarisk and dead or dying Toghraks. Thus a colony had been planted here in the hope of utilizing the chance offered for an agricultural settlement. But the provision of a circumvallation and the crowding of its interior with numerous dwellings, all of a uniform type and manifestly provisional, seem to point clearly to a scheme of colonization very different from the haphazard growth of scattered holdings usual in the case of such ` new lands'.
Geographical and antiquarian observations combine to make it easy to account for such a colonizing effort in this position. In discussing below the far older remains to the east of the Endere River, I shall have occasion to indicate the special importance which the area of vegetation along the terminal course of the Endere River must always have claimed in historical times as the only possible position for a half-way station on the desert route, some 200 miles long, between Niya and the oasis of Charchan. The ruins of older fortified stations found near the east bank of the river undoubtedly date from successive endeavours to establish here a settlement which would help to facilitate and protect traffic on the route leading by the Taklamakân edge from Khotan to Lop-nor and China. It thus seems reasonable to connect the later ruins of this fortified village with a systematic endeavour made in Muhammadan times for the same purpose.
The change in the site chosen for the new settlement was, no doubt, dictated by a temporary shifting of the Endere River course. From the absence of all traces of agricultural development near the site and from other indications previously mentioned, it is equally certain that the attempt must have soon failed. But of the direct cause of this abandonment it is impossible to make sure in the absence of any definite evidence. A number of causes can be thought of, which might uniformly have produced the early failure of this colonizing venture. Under the special physical conditions prevailing, another shift of the river to where it now flows, fully five miles to the east, would alone have sufficed to make irrigation impossible. Attention must also again be called to the curious observation that the rampart which had borne a superstructure of rush bundles fixed on rough beams, showed throughout its exposed portion marks of having been subjected to fire. Considering that several of the huts cleared had also their timber partially burned, it seems difficult to resist the conclusion that a conflagration had played its part in the early end of the deserted settlement, even though the particular circumstances escape us. From the wide range of conjectural explanations it will suffice to mention one only, which has been suggested to me by the structural peculiarities of the ruins. I mean the possibility of their marking one of those short-lived attempts at forcible colonization in which Central-Asian rulers down to quite recent times have so often indulged in dealing with