288 THE SEARCH FOR THE ANCIENT CHINESE HIGH ROAD [Chap. VIII
have appeared insuperable to traffic. The same physical difficulties still faced us, in fact had become even greater, since Lou-lan had ceased to hold water and therewith to support life. Accordingly it did not take me long, once I was satisfied as to the character of the remains on the Mesa; to decide that it was my best course to follow the guidance afforded by our antiquarian discoveries and to continue the line to the north-east.
Old river- I had instructed Afraz-gul to keep an eye on my movements ; so it was not difficult for us to
bed traced. get into touch from the top of our respective Mesas. As I turned back from L.J. in his direction
I crossed once more the previously mentioned depression, but at a point higher up, where its character as an old river-bed was quite unmistakable. Its well-marked banks were from about twenty to twenty-five feet in height, and its width about 16o yards. It was interesting to note along its centre a line of dead tamarisk-cones thickly covered with dead roots and branches. Elsewhere the bottom of the wide bed was eroded and showed Yardang ridges and trenches running in their regular direction from ENE. to WSW.
Growth of From the fact that the top of the tamarisk-cones still remained some ten feet below the level
ibed. banks it seemed safe that the growth n of the bks id fe to conclude thhe rowth of the tamarisk-cones had started at a time
when the bed had been deserted by flowing water and become dry, but when subsoil water was still near to its bottom. In the ancient river-bed near the southern end of the Niya Site I had observed a similar chain of tamarisk-cones ; but there they were still alive on their tops and rose considerably above the level of the ancient banks, as seen in my photographic panorama.'8 The probable explanation seems to be that while both river-beds may have lost their flow of water about the same period, subsoil water at the Niya Site has remained sufficiently near to the surface to permit the continued growth of the tamarisk-cones down to the present day, when their tops rise twenty feet or more above the banks. Here, on the other hand, at the extreme north-eastern edge of the ancient Kuruk-tagh delta, subsoil water must have disappeared completely, say, some five or six centuries after water had ceased to flow in the bed. The estimate of time allowed for the increase of height in the cones here must, of course, remain purely conjectural; for the rate of growth per century may well be different on this ground from that approximately determined at sites south of the Taklamakan.19
Possible Judging from the position of this river-bed to the west of L.J., one might be inclined to connect
of river-bed. it with the northernmost terminal branches of the Kuruk-darya, which, as Map No. 29. D. 3 shows,
were traced by us to near the latitude of 40° 4o'. But no recognizable bed was crossed by us on our march from Altmish-bulak to L.I. On the other hand, the direction of the bed where we did see it closely coincides with that of the big flood-bed coming from Altmish-bulak as last sighted at its issue from the ` Sai ', about ten miles to the north-west (Map No. 29. D. 3). Whether there is a connexion between the two, and whether the bed surveyed near L.J. has its continuation in the wide salt-encrusted depression which Afraz-gul crossed in February, 1915, on his march along the western edge of the ancient Lop sea-bed about eleven miles to the south-east, could only be settled by further detailed survey. I may here record also that on my way from L.J., at a point close to the right bank of the old river-bed, I came upon pottery debris of the same coarse kind as picked up at L.I.
Grave on Rejoined by Afraz-gul, I was guided by him to a big Mesa, situated about one and a half miles
Mesa . to the NNE. of L.I. and rising to a height of over seventy feet. On a terrace of its north-western
slope he had come upon a grave exposed by erosion. The badly decayed body had rested in a coffin made of rough Toghrak trunks, and the head still retained tufts of fair hair. Though little
18 See Serindia, i. Fig. 75 ; cf. also above, p. 44 ; Desert 19 See Serindia, i. pp. 199, 284 ; iii. p. 1251.
Cathay, i. Pl. III.