Sec. i] BY THE EASTERN COAST OF THE DRIED-UP SEA 315
off to the north-east marked a bight of the coast-line, at a point seventeen miles from camp. This strange well-trodden track cropped up again towards the end of the march, and beyond it also. It puzzled us greatly at first until the increasing number of camels' footprints running along it farther on convinced both Tokhta Akhün and myself that it must have been trodden by wild camels moving along this line for a long time past. It pointed to visits paid by them to the eastern Kuruk-tagh from their present haunts along the terminal . Su-lo-ho and in the Besh-toghrak valley.
Tokhta Akhün however declared that he had never seen such a regular track used by wild Possible
camels except where it leads close to water, and from that we were still far away. I well remember explanation
how the ancient track left in the gravel by the movements of Chinese patrols along the wall of the Tun-huang Limes has remained traceable to the present day.2 I have accordingly wondered at times whether the wild camels' use of a regular track here, on ground where the nearest open water is fully sixty miles away, may not have been induced in the first instance by the convenience which a path made by man afforded, and been subsequently continued through the ages. Such an explanation must, of course, remain purely conjectural. I may, however, mention that Abdurrahim, who had also noticed the old well-marked track where he and Lai Singh's party reached the northern edge of the great bay, put upon it the same interpretation quite independently, when I questioned him on the subject after our reunion at Kum-kuduk.
But a still more curious observation awaited us. I realized, on sighting a promontory far ahead Trace of
to the east, that following the shore of the bight above mentioned would involve a considerable ancient
detour. So I decided to steer straight for a hillock rising within the bight of hard shôr half a mile recognized.
farther on and in the direct line of that promontory. My hope of finding a better surface beyond it was disappointed. But when I had ascended with Afraz-gul and Tokhta Akhün the salt-encrusted hillock, about twenty feet high, my eye was caught at once by a broad and absolutely straight line running across the hard salt surface from the western end of the bay towards the previously sighted headland. My companions, too, clearly recognized the line which passed close to the south of the hillock. It was obviously the line of the ancient Han road cutting off the detour round the bay, and its trace was as clear as only this peculiar ground could preserve it.
Tokhta Akhün was sent back to take the camels round by the shore, and then, having fixed Ancient our position on the plane-table, I followed the ancient track with ease as the depression of the surface track
marked it clearly. It at once brought back to my mind the appearance of the present caravan
track towards Tun-huang, where it cuts across the big bight on the southern shore of the Lop Sea
beyond Chindailik.3 Together with Afraz-gul I repeatedly measured the track and found that
it showed a fairly uniform width of twenty or twenty-one feet. Its surface was sunk about a foot
below the average level of the adjoining crumpled-up salt-cakes and offered tolerably good going ;
for within the track the salt-cakes were either much worn down or were covered with a layer of
soft shôr. This smoother state of the surface must have resulted in the main from the grinding
effect of heavy traffic, much of it probably in carts. But comparison with the surface noticed in
shallow drainage channels passing into the shôr from the hill-side at other points of this coast-line
suggested another explanation : an occasional accumulation of flood water in the worn-down
track, rare as it must be, may have contributed to produce its present appearance.
We were able to follow the straight track of the ancient route, thus fortunately traced here, Straightness
without a break for two miles to where it met the clay promontory already referred to, at the eastern track of ancient
end of the bay. This headland, on close approach, proved to be broken up into a series of wind-
eroded terraces, much after the fashion observed at the end of the Sai tongues projecting into the
2 Cf. Serindia, ii. pp. 656 sq., 682. 3 See ibid., ii. pp. 549 sq. ; Desert Cathay, i, p. 507.