Sec. vi ACROSS THE ANCIENT DELTA OF THE KURUK-DARYA 209
grown up during some period after the river-bed had been left dry but when it still received subsoil water. The general appearance of this bed closely resembled that of stretches of the main Kurukdarya higher up at the numerous points where I saw it in March, 1915, south and south-east of Yardang-bulak.18 A look at the map shows that this well-marked dry river-course lies exactly in the direction of the line of dead tamarisk-cones, obviously riverine in character, which was passed in 1906 about a mile to the north of Camp 123. My plane-table traverse of the route followed on the journey to the Tarim in the last days of December, 1906, records a ` dry lagoon ' with banks zo feet high amidst strips of dead jungle, about three miles to the NE. of Camp 126a.19 This in all probability represents a broad reach of the same ancient water channel higher up.
The light was fading by the time the camel train had caught us up at this bed. Yet though I knew beforehand the difficult character of the much-eroded ground which still separated us from our goal at the ruined station, I decided to reach it that night. It meant a day saved for the work before us, and for the camels the prospect of reaching a little sooner the rest and grazing they badly needed. So we pushed on ahead of them. For about two miles we traversed a much-furrowed waste of bare clay, with scarcely any remnants of dead vegetation ; but worked stones in plenty continued to be picked up until the light failed. After a weary tramp of an hour and three-quarters across terribly eroded ground, we reached another dry river-bed, this time a very winding one and about 200 yards wide. Among the fallen trees lining its banks the men with me thought they could recognize two Jigda (Eleagnus) trees. Darkness had fallen by the time we had crossed this bed, but it was soon relieved by a glorious full moon. I had a bonfire lit on an isolated tamarisk-cone to guide the camels far behind, while we struggled on across the close-set Yardang ridges and trenches to where the ruined Stûpa of L.A. loomed ever bigger before us in the moonlight. At last, after a total tramp of twelve hours, during which we had covered some nineteen miles of marching distance, I found myself once again amidst the silent ruins to which my thoughts had turned so often since the happy days I spent at work there in December, 1906. A big fire of ancient timber, lit on the terrace bearing the Stûpa, threw the familiar features of the old deserted Chinese station into weird relief, and eased the long wait till the hard-tried camels came in, which they did safely hours later.
Before I proceed to give an account of the fresh explorations for which the Lou-lan station served as base, it will be convenient to sum up briefly the results, as regards the physical aspects of this area and their relation to periods of human occupation, yielded by the surveys made, both on this and the previous journeys across the Lop desert, between the Tarim and the Lou-lan Site. Proceeding northward from the terminal course of the Tarim two main zones are easily distinguished. In the southern zone, which extends from the Kara-koshun marshes to the north-east as far as the approximate latitude of 40° 41, we find at first lagoons, like the Chainut-köl, that receive water more or less regularly from the annual floods of the Tarim. Beyond these there stretches a series of small depressions into which water is carried only by exceptional floods at long intervals, the lakes thus formed being gradually reduced by evaporation to mere salt pools and finally left dry. The ground extending along these terminal lagoons over a distance of about lo miles or less receives sufficient moisture to support desert vegetation such as reeds and tamarisks. But nowhere would the configuration and character of the ground permit of cultivation and permanent settlement. To the east spread the marshes of the Kara-koshun, and beyond them the salt-encrusted wastes of the ancient dried-up Lop sea-bed, still likely to be reached on its margin by inundation from the Tarim when the spring floods are exceptionally high. To the west, high dunes cover the whole ground right away to the Tarim, whose course here lies from north to south.
18 See Map No. 29. A, B. 3 ; below, Chap. xx. sec. iii. 18 Cf. Serindia, i. p. 45o ; v. Map No. 6o. c. 3.