302 FROM CHARCHAN TO CHARKHLIK [Chap. VIII
course, of which actual survey has failed to reveal any evidence,' while all topographical considerations point to the conclusion that in historical times the area best adapted for cultivation must have
lain in the present oasis and its immediate neighbourhood. The absence of datable relics precludes any definite judgement as to the period when the ` Kone-shahr' area was last inhabited, just as it is impossible to guess how much of the remains of ancient Chü-mo and of Marco Polo's Charchan lies buried in the soil now again irrigated.
Here it may be noticed that according to local information, confirmed by the appearance of the trees and other indications, the oldest portion of the present oasis is represented by the Aralclii
` Mahalla ', which lies on the right bank of the main river bed.5 It takes its name, meaning ` the island one', from the fact that on its east side there is another smaller bed known as Kone-daryâ,
which is still filled by the summer floods. If we assume the main settlement of ancient Chü-mo to
have occupied the position of Aralchi, we can account for the fact that the passage already quoted from Li Tao-yitan's commentary on the Shui thing speaks of the Charchan River as flowing ` to the
west of the walled town of Chü-mo'.6 Otherwise it might be assumed that the old flood-bed, which passes close to the ` Kone-shahr ' and is now utilized to irrigate the westernmost part of the existing oasis, represents an earlier main bed. Owing to the very uniform slope of the detritus fan, at the northern foot of which Charchan is situated, the river bed is so broad and shallow that such a change could have taken place during the last fifteen hundred years without much affecting the position of the irrigated area.
Owing to the close vicinity of the ` Kone-shahr ' to the inhabited area and the constant search for ` treasure ' proceeding among the small relics brought to light by wind erosion, there was little chance of picking up there in situ any fragments of archaeological interest. But ancient beads of
stone and glass, as well as fragments of bronze ornaments, &c., are frequently found by the villagers, and of these a small representative set, as described in the List below, was acquired without difficulty
during my brief halt. In general character these small relics resemble those obtained from the
` Tatis ' around Khotan. None are of a type which at present admits of exact dating ; but I may at least mention a bead of cornelian, Char. 001 I, showing a peculiar inlay which is a characteristic
feature in some acquisitions from Yotkan. I greatly regretted the total absence of recognizable
Chinese coins among the objects brought to me by the villagers ; but my informants declared that such finds were rare now, since all structural remains had been destroyed by burrowing. My own
search at the more distant ` Tatis ' described below yielded no intact coins, only tiny fragments which, retaining the characteristic square rim, attest their former existence. I was inclined to attribute this complete destruction of copper coins, elsewhere so common at such sites, to the much-increased force of wind erosion on ground which is practically clear of drift-sand and is scoured in most places down to the bare gravel surface.
On November 22 I paid a visit to the more distant ` Tatis ' which Saif-ullah, a local ` treasure-seeker ', had to show me to the south-west of the present oasis. Proceeding from the Yalghuz-tugh
Mazar, we first followed for about a mile and three-quarters the line of an ancient canal, still clearly
recognizable, to a point where it was found to diverge from the ` Yangi-üstang' restored by Mûsa Beg. For about one-half of that distance, or a little less, the pottery débris of the ` Kone-shahr '
Site extended. The embankment of the ` Yangi-üstang' showed sharply above the absolutely bare gravel plain. From a line about three miles south-west of Yalghuz-tugh there commenced an extensive ` Tati ' area which was found to reach, with scarcely any interruption, for about two and a half miles further to the bank of a shallow depression marking a flood bed of the Ayak-tar stream.
' Cf. Hedin, Central Asia and Mel, i. pp. 306 sq. 5 Cf. also Hedin, Reisen in Z.-A., p. 179.
See above, p. 297.