778 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XXI
A. D. 94, who thereby completed his pacification of the Tarim basin.5 The Wei /it), composed between A. D. 239-65, similarly associates Wei-li, Wei-hsü, and Shan as territories all dependent on Yen-ch`i or Kara-shahr, where it describes the continuation of the ` route of the centre ' to the west of Lou-lan.6 Finally the Tang Annals, also, place Wei-li correctly to the south of Yen-ch`i, without apparently furnishing any details about it.' I am unable to trace any later notices, whether Chinese or other, which can be directly connected with the tracts now comprised in the district of Kara-kum. But it deserves to be mentioned that the present official attribution to it of the Chinese designation Wei-li proves that the correct location of the ancient territory is known to the modern administration of Hsin-chiang.
Geographical conditions make it clear that the territory of ancient Wei-li derived importance mainly from the position it occupied as a kind of ` corridor ' along the Konche-daryâ. All through historical times it must have served, just as it does at present, to link the cultivable portion of the Lop region, corresponding to Shan-shan or the modern Charkhlik tract and commanding the routes which lead thence towards Tun-huang, Tsaidam, Tibet, and Khotan, with the north-eastern corner of the Târim basin and the high roads along the Tien-shan. No doubt the importance of Wei-li was necessarily even greater at a time when the ancient ` route of the centre ' connecting the Tarim basin directly with China via Lou-lan passed through it. But even after this route was abandoned, any traffic connected with trade, administration, or military movements that passed between the oases in the north and the cross-roads of Shan-shan was obliged to traverse Wei-li.
The geographical factors determining this traffic and the administrative, commercial, and strategic needs served by it have remained the same to the present day. This explains the persistent endeavours made by the Chinese administration of the ` New Dominion ', ever since its reconquest, to facilitate the use of this route from Korla to Charkhlik by the creation of agricultural settlements ; for without the local supplies that these alone can furnish, the value of this line of communication, some 35o miles in length, must necessarily remain precarious. Considerable difficulties have been encountered, with the curious result that the head-quarters of the district officer entrusted with the organization of such colonies has been moved in the course of about twenty-five years to four different localities in succession. These frequent shifts, which seem to invest the head-quarters of modern Wei-li with a quasi-peripatetic character, did not fail to attract the attention of Professor Elsworth Huntington. When visiting Tikenlik in 1906 from the side of Lou-lan, he was led to recognize in them an illustration of the physical difficulties that once also beset that ancient settlement, though to a smaller extent, and at the same time to treat the difference between the conditions indicated by these modern colonizing attempts and the conditions assumed to have prevailed at Lou-lan as definite evidence supporting his theory of a great secular change of climate.$ He attributed those shifts of the colony intended for the Wei-li head-quarters from Jan-kul to Dorai, below Tikenlik (Map No. 25. c. 3), and thence again in 1901 to Kara-kum, ` entirely to the extreme salinity of the rivers ', which would render fields irrigated from them for two or three years in succession worthless for production.9
In view of the interpretation proposed by the eminent geographer, the question of these curious shifts assumed a direct quasi-antiquarian interest. My examination of old sites to the south of the Taklamakân, the abandonment of which had also been ascribed to the increased salinity of the water-supply, had not yielded any definite evidence on the point. Nor had Professor Huntington been able himself to visit any of those former sites of the Wei-li head-quarters within a short time
5 Cf. Chavannes, T`oung-pao, 1907, pp. 208 sq. ; 1906,
pp. 234, 236.
6 See Chavannes, T oung-pao, 1905, pp. 552 sq.
7 See Chavannes, Turcs occid., p. xio.
8 See Huntington, Pulse of Asia, pp. 271 sq.
9 Cf. ibid., pp. 266 sqq.