210 REMAINS OF AN ANCIENT DELTA
Zone of Near the northern limit of this zone where flood water penetrates but rarely, we meet with
eroded dead the first patches of ground left bare of living vegetation and thus exposed to wind-erosion. Beyond
delta. the line of the last dry lakelets begins the second zone, comprising a now utterly waterless desert.
Here wind-erosion, with the drift-sand that serves as its powerful instrument, is the predominant factor in determining the present surface conditions of the ground. Its excessive action in this zone is primarily due to the strength and great frequency of the ENE. winds drawn into this, the lowest, portion of the Tarim basin by ` aspiration ' from the barren plateaus of the Pei-shan and southern Mongolia. It is further facilitated by the nature of the soil, which here as throughout the bottom of the Tarim basin is formed of the clay sediments left behind, perhaps since tertiary times, by a huge inland sea.
Physical Apart from the wind-sculptured soil, the dry beds of old river-courses, lined on their banks
character- by strips of dead forest, are the most characteristic feature of this zone. Their direction unmistakistics of
dead delta. ably proves them to have formed part of the delta in which the Kuruk-daryà once carried its waters
into the great salt-encrusted waste eastwards, occupying the terminal basin last filled by the much more ancient Lop Sea. The dried-up marginal marshes of this basin, which Afràz-gul traversed in February, 1915, form the eastern limit of the zone, while westwards it is bordered by the great dune-covered area that divides the old Kuruk-daryà delta from the riverine belt of the Tarim. To the north the zone extends right up to the foot of the Kuruk-tâgh glacis, and thus includes also that portion of the ancient delta which lying north of the Lou-lan sites remained outside the surveys discussed in the present chapter.
Belt With the zone just described remains of human occupation permit us to distinguish three belts.
occupied in In the southernmost, which extends along the line clearly marked from north-west to south-east historical
period. by the ruins of L.R., L.M., L.L., L.K., we have definite archaeological evidence that the pre-
historic occupation during successive periods of the stone age was followed by settlements which were contemporary with those of the Lou-lan sites, and probably survived like these till the beginning of the fourth century A. D. Their existence indicates that the southernmost branch or branches of the Kuruk-daryà traced in this belt carried water during the first centuries of our era. The riverine tract thus watered in historical times appears, however, to have been a narrow one.
Belt not The second belt to the north comprises ground where, though ancient beds are here also
occupied in traceable at intervals, evidence of human occupation is confined to remains of the stone age, and historical
times. these, moreover, are less frequent than in the adjoining belts north and south. The width of this
belt, which may be described as extending from about 40° 12' lat. to 40° 22' lat., is roughly 16 to 17 miles. The conclusion seems justified that already in the earliest historical period to which we can go back on this ground this portion of the ancient delta lacked sufficient water for cultivation and permanent settlement. The earlier drying up of this portion and the consequent longer exposure of its ground to wind-erosion may account for the depression shown here by Dr. Hedin's line of levels, as well as for the greater effacement of the old river-beds.
Belt The third belt extends from about 40° 22' lat. northward to the gravel glacis of the outermost
watered hill range of the Kuruk-tagh. Within it are to be found the Lou-lan sites explored in 1906 as well
darya in as the remains of the same period traced in 1914. They all afford clear proof that one or two at
Han times. least of the northern branches of the Kuruk-daryà carried sufficient water during the early centuries of our era to permit of irrigation. But here also abundant relics of the stone age are to be found, showing that the whole of this belt was sufficiently watered in prehistoric times to render nomadic occupation possible. From coins and other approximately datable relics discovered in localities situated at an appreciable distance from those sites, it seems safe to conclude that physical conditions, similar to those still observed now on the Tarim, permitting of nomadic life based on fishing, hunting,