THE REGIONS SURVEYED [Chap. II
basin from the one in the south it is mainly because it is thus easier to indicate certain topographical distinctions deserving of notice. t But both share the essential feature of containing whatever ground within the Tarim basin is capable of permanent cultivation under existing conditions.
The cultivable area is considerably greater within the western and northern belts.
This may be taken as extending in a big arc from Yarkand and Kash-gar to Korla and the tract immediately south of it. It contains the far larger number of important oases, such as Yarkand, Kashgar, Ak-su,
Kucha, and owing to them and to the short distances at which smaller ones are strung out between them, the route passing through this belt of cultivable ground has from the earliest historical times to the present day been the chief line of communication and trade within the Tarim basin. These advantages for permanent occupation and traffic result mainly from the fact that irrigation—that indispensable condition of all cultivation in this basin—is here greatly facilitated by the volume and number of the rivers as well as by physical features of the ground which favour full use of their irrigation resources.
Among such features may be mentioned the absence or comparative narrowness of
barren gravel stretches at the foot of the outer hills, which allows cultivation to be started from the very debouehure of the rivers; also the protection which the Yarkand river or Tarim with its . broad riverine
jungle belt affords from the moving sands of the Taklamakan. Hence we find in the case of the larger oases cultivation extending for considerable distances along the beds of the rivers. 3 Since all of these reach the Tarim and on their lower courses command extensive areas inundated during flood times, grazing grounds adequate for the needs of the oases are also available. On the other hand the map shows a complete absence of cultivated areas of any size along the Tarim itself after the river has passed the northern edge of the Yarkand district. This striking fact is explained by the great difficulties which the very slight fall in the river's bed and the consequent constant shifts of its course here oppose to the construction and maintenance of irrigation canals of any size.
With the northern belt of oases may conveniently be mentioned also the subsidiary
basin of Kara-shahr which adjoins the extreme north-eastern corner of the Tarim basin proper. 3 Though draining into the latter by the Konchedarya, it shows characteristic geographical features of its own. It is
divided from the plains of the Tarim by the westernmost hill-range of the Kuruk-tagh which encircles it on the south, and a considerable portion of its area is occupied by the large freshwater lake of the Baghrash-köl. This gathers the waters brought down by the Kara-shahr river from high plateaus of the Central Tien-shan and, acting as a big natural reservoir, discharges them with an almost • constant volume through the defile above Korla. An abundant water supply and an apparently less arid climate assure to Kai a-shahr chances of extensive cultivation, and the historical importance of the territory shows that in antiquity these were adequately utilized. Their present neglect seems to be largely due to the population containing a considerable element of semi-nomadic Mongol herdsmen. The latter's presence alone serves to illustrate the difference in climatic conditions between the Karashahr district and the rest of the Tarim basin.
Western and northern belt of oases.
Position of cultivated areas.
1 This distinction is supported also by evidence derived from the historical geography of the Tarim basin. The detailed descriptions given of it by the early Chinese historical record, always separate the
territories situated along the ' iorthern Rend e.
these at the south foot of the Tien-shah, from those on the 'Southern Road' which passed through Charkhlik and Khotan. In this as in many other notices of those records relating to Central Asia the keen topographical sense of the Chinese clearly reflects itself.
It is significant that though So-chic, i.e. Yiirkand,
is duly indicated as the point through which passed the `Southern Road' to the l'iimirs and beyond, yet this territory is always described in its proper geographical nexus, along with Ktishgar; see Wylie, 'Notes on the Western Regions', Journal of the Anthropol. Institute, x. p. 21, 47 sq.; Chavannes, 2'oungpao, 1907, pp. 170, 196 sqq.
2 See Sheets No. 2. D.2; 6.A, B.2; 7. D.2,3; 12.A.3; 17. A. 1. Is, C.1,2.
3 See Sheets ? o. 20. D. 4; 24. A-D. 4; 25. A-C.1.
4 Cf. Seriadia, ill. pp. 1179 sq., reta°ding the limited cultivation of gala-shahr and its causes.