SECTION IV.—GENERAL OBSERVATIONS ON ENDERE AND THE `OLD
Leaving aside these insignificant remains of undetermined age found west of the river, I may briefly sum up here the main facts brought to light by my renewed survey and excavations as to the history of the ruined area eastwards, which we may comprehensively designate as the Endere Site. We have seen that, as far as extant structural remains are concerned, occupation is definitely proved for the site during two distinct periods. The earlier one is represented for certain by the ruins excavated close to the south of the circular fort first explored in 1901, and by the débris discovered below the rampart of the fort itself. To the same period may also be ascribed with much probability the great Stûpa northward and the circumvallation near it, as well as the fortified post and the traces of ancient dwellings at the southern end of the site. This earlier period may be safely assumed to have extended down to about the time when the Niya Site was abandoned, i. e. down to the end of the third or beginning of the fourth century A. D. To the later period of occupation of the site belongs the circular fort with its small temple E. I containing a graffito of A. D. 718. This period must fall within the Tang dynasty's effective domination over the Tarim Basin (c. A. D. 66o790), and in all probability nearer to its beginning than its end. It seems in any case likely that this renewed occupation was much shorter and far more restricted also in local extent.
I do not propose to discuss here the questions concerning the physical changes which may have caused, or been connected with, the successive abandonments of the site. These questions are sub_ stantially the same as have already been considered above in regard to the Niya Site, and, notwithstanding Professor Huntington's ingenious efforts, the available archaeological evidence does not appear to me to furnish critically safe answers with regard to the special circumstances under which those abandonments took place. Only two facts may be considered certain. On the physical side we must accept progressive desiccation as the explanation why during the earlier period a large settlement could exist here such as the present water-supply would no longer be sufficient to maintain. On the historical side we have the clear testimony of Hsüan-tsang, who, when passing the site about A. D. 645, found it completely deserted and its towns ruined wastes.
The archaeological evidence now brought to light of a settlement abandoned centuries before Hsiian-tsang's passage has once more confirmed our reliance on the great pilgrim's often-proved accuracy in topographical matters, and supplied us with an incontrovertible historical instance of a site in the desert reoccupied after long abandonment. But the definite identification of the Endere ruins with the deserted towns ' or ` walls'' which Hsüan-tsang saw in ` the old Tu-huo-lo country', has a wider historical interest, though for the present only in a negative sense. The fact that the pilgrim mentions this deserted settlement by the same name Tu-huo-le 45 A if, or Tukheira, which was borne by early conquerors of the Graeco-Bactrian kingdom, and which survived down to mediaeval times as a designation of Badakhshân and adjoining tracts in the Tokhâristân of Muhammadan geographers and historians, has given rise to a great deal of learned speculation and discussion.
It is not my task here to consider the many difficult questions as to the ethnic origin of the Tochari who, as we find in a well-known passage of Strabo (xi. 8. 2), were among the nomadic tribes which wrested Bactria from the Greeks in the second century B. C. ; as to the relation between them and the Yüeh-chih ; or as to their suggested identity with the Ta-hsia whom the Yüeh-chih