SECTION I.—KORLA AND ITS OLD SITES '
MY stay at Korla during the last few days of December, short as it was, allowed me to appreciate fully the exceptional physical advantages which this bright oasis in the extreme north-east corner of the Tarim Basin proper enjoys. Situatefl as it is at the very debouchure of the big river which drains the Baghrash Lake, the Korla oasis has the great boon of possessing a water-supply for irrigation which is not merely abundant but of unfailing regularity. Owing to the fact that the Baghrash Lake acts as a huge natural reservoir storing the spring and summer floods from the Tien-span ranges to discharge them again gradually, the Konche-ctarya, as the river is called from Korla onwards, preserves throughout the year a remarkable uniformity of volume, a feature wholly unknown elsewhere in the river system of Eastern Turkestan. This volume is greatly in excess of the needs of the lands cultivated by the actual population.' In 1907 this was locally estimated, with what seemed a fair approach to accuracy, at 1,700 households, while the water-supply was acknowledged to be sufficient for irrigating an area which might easily support six times as many. Enough suitable land for such extended cultivation is certainly available south of the river (Map No. 49. A. 2).
It seems difficult to account for this discrepancy between conditions so exceptionally favourable and the comparative smallness of the settlement. But there is reason to assume that the same difficulties, due to the vicinity of nomadic Mongol neighbours, which at present affect colonization in the Kara-shahr territory, as explained above, have much to do also with retarding the development of Korla, now included in the Kara-shahr district. The Mongols, who, with their herds of horses, etc., frequent the scrub- and reed-covered waste lands south and south-west of Korla for winter grazing, are by no means welcome visitors for the peaceable Turki settlers. However this may be, it is certainly curious to note the same relative unimportance of Korla in ancient times as well. I believe that we have to locate there the small kingdom of Wei-ksü fg , which figures in the Han Annals in connexion with Yen-ch`i or Kara-shahr, but subsequently disappears from Chinese historical records as a separate territory.
The identification here proposed of Wei-hsü with Korla appears to me proved by the Former Han Annals' notices which place this `kingdom' with its city of Wei-hsü a hundred li from Yen-ch`i and 260 li to the west of the kingdom of Shan 111.2 That the latter must be identical with the
' Dr. Hedin calculated in March, x896, the volume in the river where it passes the bridge leading to Korla at circ. 72 cubic metres, or about 2,442 cubic feet, per second. The position of this bridge proves that the water-level here remains practically the same throughout the year; see Hedin, Reisen in Z.-A., pp. 68 sq.
To this volume must be added the by no means unimportant amount of water taken off by canals like that of Bashengiz and Shinalga which have their heads above the town.
4 Cf. Wylie, Notes on the Western Regions, J. Anthrop. Inst., xi. pp. lox, 105. In the latter passage we are told that the people of Shan, ` living among the mountains, depend on Yen-ch'i and Wei-hsü for their grain and field produce'. This is still true at the present day of the Mongols and others who frequent the westernmost Kuruk-tagh. Their supplies are drawn exclusively from Korla and, during the winter when the Baghrash Lake gets frozen, from Kara-shahr.
Mr. Wylie provisionally adopted the conjectural identifi-