of the ` Yârdangs '—here as everywhere else in the Lou-lan area, WSW. to ENE.—had allowed us on this day to cover a total marching distance of about eighteen miles. Camp xciii was pitched in the evening close to a well-marked line of tamarisk-cones, recalling those which are often found near river-courses in the Tarim basin. Dead reed-beds of the kind above mentioned were found occupying the tops of ` Yardang ' trenches near the camp. Their height above uniformly eroded trough-like ground varied as much as eight feet. This clearly showed that water, such as alone could account for such reed growth, had returned here in comparatively recent times to ground where wind-erosion had previously been long at work, and that the level reached by the water during its temporary return had varied.
Another interesting fact was that on resuming our march on the morning of February loth we came, within less than a mile, to rows of dead Toghraks along a shallow bed winding with an approximate bearing from north-west to south-east. As appears from the map, this direction exactly coincides with that in which our survey of 1906 shows a strip of dead trees and reed-beds about four miles to the north of Camp 122.6 It is significant that if we continue the same line farther to the north-west it takes us to the ` belt of Toghraks and tamarisks ' that my plane-table record of 1906 shows about four miles to the south-west of Camp 126 a on the route from the Lou-lan site L.B. to the Tarim.' Here, too, as in the previous instances noted along our survey from L.K. and L.M. onwards, an ancient branch of the Kuruk-daryâ is still clearly traceable on the map.
The ground that we traversed for a distance of about six miles beyond this river-bed was a bare flat of clay, comparatively little furrowed by wind-erosion and scattered over with scanty dead tamarisk-cones and low dunes. Remains of the stone age, such as worked stones and coarse neolithic pottery, here made their appearance again, and were found with much frequency from about the fourth mile right through to the vicinity of L.A. As appears from the Descriptive List below, by far the most numerous among the worked stones were the narrow sharp-edged flakes, probably used as knife-blades, which I have already frequently referred to among the finds from the eroded area south of the Lou-lan sites.8 Besides miscellaneous worked stones of indeterminate character, mention may be made of the scraper, C. xciii. 022 ; the well-made stone arrow-heads, C. xciii. 016-17, 099 (Pl. XXII) ; and the jade celts, C. xciii. 0141-2, 0146 (Pl. XXII).
For reasons already indicated, it is impossible to draw any definite conclusion as to the chronology of prehistoric occupation from stone age relics found in the Lop Desert on wind-eroded soil.9 But it should be noted that such remains, whether of stone or coarse pottery, practically ceased to be found on the route we followed in 1914 from a short distance beyond L.M. until we reached the vicinity of Camp 122, and that over a corresponding direct distance of close on ten miles on the route we followed in 1906 finds of the same kind were either totally absent or very rare.10 It appears to me probable that for some reason no longer clear this belt was but little frequented during prehistoric times and during the Han period which succeeded them, except along the route leading from the station L.A. to L.K. and thus on to the old Lop capital at Miran.
The first find of a datable relic and one therefore of distinct antiquarian interest occurred at a
Dead reeds on top of Yârdangs.
Ancient river branch marked by dead Toghraks.
Worked stones and neolithic pottery frs.
Distribution of stone age relics.
See Serindia, v. Map No. 6o. c. 3.
7 See ibid. In Map No. 29. c. 4 the corresponding entry of dead tree symbols has been omitted by an oversight. Their line ought to have been shown with the bearing from NW. to SE. also extending to the entry ` Rare pottery '.
8 See below in Descriptive List, C. xciii. 013-15, 018-21, 024-6, 035-63, 082-98, 0100-40, 0158-62 (Pl. XXII). A number of these show worn edges, a sign of long use. Cf. also p. 184 above ; Serindia, i. p. 357.
9 See above, p. 197 ; Serindia, i. p. 357.
10 I regret that I did not bring this out clearly in Serindia, so far as the line followed by our route of 1906 is concerned. But my recollection on this point is supported by the plane-table record of the route in Serindia, Map No. 6o. c. 3, 4, which shows no entry of finds between latitudes 40° II' and 400 19', and also by the ` List of Objects ', Serindia, i. pp. 363 sqq., marking no definite find-place farther north than 6 miles from Camp 121.