Find of Fortunately, the first day's march on this ground brought a find which, small as it was, had the
bronze merit of being approximately datable, and important discoveries which I made in 1914, in an exactly
corresponding position further west, have left no doubt about its interpretation. We had moved about four and a half miles from Camp I2I when, near a belt of dead forest showing besides Toghraks also trunks of dead Jigda or Elaeagnus trees, Tokhta Akhûn's sharp eyes discovered lying on the ground a carefully finished bronze arrow-head, C. I 2 I. 0050. Its shape, triangular in section to the point and hexagonal where bevelled down below for the shaft, accurately corresponds to the type represented by numerous bronze arrow-heads which were picked up near to or at the ancient station of Lou-lan (cf. C. 123. 001-003 ; L.A. 0017, 0069, 0082 ; III. 00I ; VIII–IX. 005), and of which a series has been illustrated in Plate XXIX. The same type is proved, by my very numerous finds along the ancient Chinese Limes, to have been in regular use by Chinese troops during the first centuries before and after Christ, as will be seen on reference to the specimens reproduced in Plate LIII. The necessary conclusion that the type was introduced into the Lou-lan region during Han times, and was probably manufactured in China, has since received very striking illustration from a discovery made in 1914 as I was tracing the ancient Chinese route through the salt desert east of Lou-lan, when I found a considerable collection of such arrow-heads scattered along the track, just as if they had dropped out of a convoy.'
Belt occu- The archaeological evidence furnished by that single arrow-head had been sufficient to warrant
pied in Lou-
Ian belief expressed in my personal narrative that the ground north of Camp 121 was covered with
lan period. P Y P g P
riverine jungle and saw at least occasional visits of man in the early centuries of our era, when that arrow-head was likely to have been dropped by some hunter or soldier.' But it was left for my expedition of February, 1914, to reveal that, within a distance of probably not more than four miles west of Camp I2I, the wind-eroded desert was hiding a ruined fort, L. K., which must have been occupied down to about the same time as the site of Lou-lan, i.e. the beginning of the fourth century A. n., and that, some six miles further to the north-west, there survived ruined dwellings of a small settlement, L.M., undoubtedly belonging to the same period.' _ This settlement had stood near the banks of a southerly branch of the Kuruk-daryd., the well-defined bed coming from the north-west and being still clearly traceable to the east-south-east, and this is the very direction which would take it to the strip of dead forest close to which the bronze arrow-head was found in 1906. This seems to me to establish the fact that the relic came from the lower portion of the same river-bed which passes the ruined settlement of L.M., and which down to the beginning of the fourth century A. n. must have carried water. The chain of topographical evidence is completed by the plane-table
survey carried out by Surveyor Afrdzgul in February, 1915, along the north-western shore of the salt-encrusted ancient Lop sea-bed. This shows, in an exactly corresponding position, the winding terminal course of the same dried-up river branch, lined by dead Toghraks and tamarisk-cones, before it is finally lost- in the great depression, covered with hard salt, which represents the true
Lop-nor of the earliest historical period.
No datable No other datable relics from the hand of man, in fact none but objects of the Stone Age, were
northward, found on the remainder of the march to Camp 122, nor for some eight miles of direct distance
beyond it. Almost the same observation applies to the line of traverse which I followed in February, 1914, from the ruins of L. M. to the ` Lou-lan Site ', L.A. On this route the first bronze object, the fragment of an ornament, was found at a point approximately six miles north of Camp 122 of my previous journey. It is hence impossible to assert whether any of the ancient beds, marked by belts of dead jungle, which our routes crossed within this area—the close correspondence
' Cf. Third Journey, in Geogr. J., x 916, xlviii. pp. t 27 sq. ° See my Third Journey, in Geogr. J., T 916, xlviii. p. 121.
s Cf. Desert Cathay, i. p. 366.