Sec. iv.] THE LOP DEPRESSION AND TURFAN BASIN 47
texts has become familiar to geographers—contain moderately fresh. water, at least in the parts near the entrance of the Tarim, has given rise to the much-discussed `Lop-nor question', complicated, perhaps quite as much as elucidated by the controversial literature which since the days of Prejevalsky and Richthofen has accumulated over it in the absence of adequate surveys. 4
Beyond the northern edge of those marshes, with their fringe of dried-up salt lagoons filled at times by exceptional floods of the Tarim, there extends an area
Ancient delta of bare clay overrun by light sand and undergoing excessive wind-
erosion. It is crossed by a series of well-marked dry river-beds, and of these our surveys have proved that they represent the southern portion of an ancient delta formed by the dried-up 'Kuruk-darya', which during the first centuries before and after Christ carried the waters of the Konche-darya (together, perhaps, with some addition from northern tributaries of the Tarim, like the Inchike-daryà) to the then-occupied territory of Lou-Ian. The erosive action of wind-driven sand has covered the dried-up delta and the whole area of that ancient territory as far as the foot of the Kuruk-tagh with Wind-eroded desert of an endless succession of `yârdangs', steep ridges carved out of the
alluvial clay and separated by parallel trenches. All lie in the direction of the prevailing north-east winds which 'aspiration' draws down from the plateaus of the Kuruktagh and southern Mongolia into the Lop depression during great portions of the year. 6 Abundant archa ological evidence, brought to light at various ancient sites of Lou-lan, makes it certain that the waters carried by the Kuruk-darya reached this once habitable territory and the delta to the south and east until the first half of the fourth century A.D. At that time the early Chinese route leading from Tun-huang through the Lou-lan area and thence along the Kurukdarya to the northern belt of oases in the Tarim basin was finally abandoned, and the territory itself became an arid and lifeless wilderness. 7
Eastwards the ancient delta ends on the shores of the prehistoric sea bed. The ancient
Chinese trade and military route which had crossed this, as we know
Ancient route across from descriptions in Chinese historical records, was traced by me to dried-up sea bed.
where the difficult salt-encrusted expanse is narrowest. S Brief reference must suffice here to the curious topographical features which the ground near these shores presents in the shape of high `Mesas' of earlier lacustrine formation and .of strings of salt- coated `Yardangs'. The ancient Chinese accounts, ever exact in topographical details, duly refer to them. 9
The opposite shore of the sea-bed lies along the foot of the Kuruk-tagh. Further south we find a long bay extending to the north-east. It occupies a
Bay joining Su-lo-ho broad valley-like depression which lies between the southernmost hill basin.
range of the Kuruk-tagh and a line of high sand-ridges lining the foot
4 There is definite historical evidence in early Chinese texts that the Kara-koshun marshes and the termination of the Tirim already occupied in the. early centuries after Christ approximately the same position as at present, and that at the same time while the Kuruk-daryii still carried its waters past the Lou-lan area into the northwestern portion of the great Lop basin ; see Serindia, i. pp. 326 sqq. 419 sqq.
This evidence, fully in accord with the results of our surveys, is important, as it disposes of a recent theory on the ' Lop-nôr question' which assumes that the Ti rim took its present course to the south, with its termination in the Kara-koshun marshes, only in comparatively recent times after abandoning an older course represented by the Kuruk-daryi. A detailed analysis of that evidence with special references to the surveys of my third journey must be reserved for the final report on the latter.
6 See. Sheet No. 29. C, D. 3, 4, where the position and bearing of dry river=beds are indicated by the
rows of dead tree (wild poplar) trunks which invariably mark their banks; the direction of these rows bas been shown in the map as recorded on the plane- table. Cf. also Serindia, i. pp. 355 sqq.
6 The extent of wind-eroded ground has been marked in Sheet No. 29 and elsewhere by the use of the special ' Yardang' symbol. Regarding the exact direction of ' Yiirdangs' in this area, their formation, etc., cf. Serindia, i. pp. 353 sqq., 360, 369, etc.; also Figs. 92, 93, 105, 108.
7 Cf. Serindia, i. pp. 426 sq., for the data concern. ing this abandonment; for the position of ancient sites, see Sheet No. 29. C. 4, D. 3, 4; 32. A. 3. See Sheet No. 32. A-C.3. The actual ancient crossing of the dried-up sea-bed lay probably near a line between C. ccxixviii. a and the find-spot of Han coins as marked on the map.
9 Cf. Serindia, is pp. 340 sqq., 423 sq., regarding the location of the White Dragon Mounds' (salt-coated Yiirdaugs) and the Mesas of the 'Dragon Town'.