Sec. ii] THE ANCIENT COURSE OF THE KONCHE-DARYA 765
` kingdom of Mo-shan', i. e. the western Kuruk-tàgh, and before continuing its eastward course to the ` town of Lou-lan '.
I have already, in Serindia, briefly expressed the belief that by the ` town of Chu-pin ' is probably meant the locality marked by the ruins of the Ying-p`an site. In support of this identification, I would call attention in the first place to the natural advantages offered by the site for a settlement of some importance, especially during the period when the ancient high road from Lou-lan towards Korla and the string of northern oases passed along the Kuruk-darya. We shall see that the line followed by this high road is marked beyond all doubt by the line of massive watchtowers which extends from Ying-p`an north-westwards to the vicinity of Korla. Archaeological evidence enables us to assign these to Former Han times, the very period when the road passing through Lou-lan possessed its greatest importance. The route, which can still be traced by the line of these ruined watch-stations for a distance of about ninety miles, led along the extreme foot of the Kuruk-tagh glacis. It was thus protected from those difficulties of communication to which periodical inundations and changing river-courses inevitably give rise in the riverine belt below the glacis edge, difficulties such as the present track from Kara-kum to Charkhlik is constantly exposed to during the flood season and the early autumn. At the same time the old course of the Konche-darya lay sufficiently near to this route to ensure easy access to water.
The traffic passing north-westwards from the oases of ancient Shan-shan, i. e. from the side of the present Charkhlik and from the adjoining riverine tract of Lop, found obvious advantage in gaining this safe ground, which was practicable at all seasons. Reference to the map will show that Ying-p`an was the nearest point where it could be reached by those coming from the south and south-east. The Ying-p`an site may thus with great probability be considered as the point of junction of the ancient ` route of the centre ' coming from Lou-Ian and a transverse road which connected it with the oases along the ancient ` route of the south ', Shan-shan (Lop), Chü-mo (Charchan), &c., and which probably passed close to the ruined fort of Merdek.15 The importance of the Ying-p`an site as a road junction must have been further enhanced by the fact that the most direct route connecting the whole of the Lop area with Turfan also led through it, then as now.ls It is easy, moreover, to realize the benefit that all traffic passing these cross-roads at the foot of the barren Kuruk-tagh derived from the fact that just here the physical conditions were such as to permit of cultivation, restricted to narrow limits, perhaps, but yet permanent.
I have had frequent occasion to point out how difficult it is throughout the Tarim basin to maintain irrigation along the river-courses, except below the point where they debouch from the mountains or else in ` terminal oases ' before they lose themselves in the sands, owing to the vagaries of the beds, the damage caused by floods to canal heads, &c. Now Ying-p`an is the only point where the Konche-darya and its continuation, the Kuruk-daryâ, are joined by a stream large enough even now at times to carry an appreciable volume of water across the thirsty Sai down to the riverine belt. That this is the case with the river of Shindi is sufficiently proved by the lagoons and the freshwater springs along their sides which are to be found here in the bed of the Kurukdarya ; for these are fed solely by the floods periodically descending from the Shindi valley. That this water-supply could even in comparatively modern times be utilized for cultivation is proved by the remains of a late Muhammadan settlement described above. The fact that this amount of water is available in the river of Shindi, whereas it cannot be found in any of the other beds
15 Regarding the Merdek site, cf. Serindia, i. pp. 452 sqq. ; for the ` route of the south ' leading to Shan-shan and thence along the oases at the foot of the Kun-lun, cf. ibid., i. p. 418.
16 We have probably an indirect reference to the use of this transverse route, from Shan-shan passing through Ying-p`an towards Turfàn, in a notice concerning a campaign of Pan Yung in A. D. 124 ; see ibid., i. p. 333.