Sec. iiJ ANCIENT REMAINS AROUND CHARCHAN 30!
embankment of a canal running northward. The fact that its bottom level was several feet higher than that of the deflated ground near by was a sufficient proof of its antiquity in spite of the name Yaugi-iistaug, ' the new canal ', by which it is known now, and of the tamarisk scrub growing along it. I soon learned from my local informants that the name and the signs of returning vegetation were due to the attempt which Mûsa Beg, an enterprising local headman, had made here, some twenty years before, to utilize an ancient irrigation channel for founding a new colony below the present oasis. The water had flowed freely for some years, during which young tamarisk scrub found time to effect a footing. But soon the endeavour to assure cultivation had failed from want of adequate labour, even though the ground which irrigation from this restored canal commanded was delared to be very fertile. It was an apt illustration of the main difficulty which seems ever to have dogged the fortunes of the most isolated of Turkestan oases.'
Going to the north-east for about a mile from the point where the canal line was crossed, I was shown a ruined wall of stamped clay about fifty feet long and still rising about eleven feet above the eroded ground. This was strewn with plentiful pottery débris of ancient appearance for a short distance all round. The ruin is known only as Tam,' the wall', and its original character could not be determined. Proceeding thence across the bare Sai towards the south-western extremity of the oasis, I was shown a low and much-burrowed mound about forty feet in diameter, known as Kalaghak-doug, which from the human bones scattered around seemed to have served as a burial-place. By its layers of brushwood inserted between gravel it recalled the mounds of Tiiga-dong examined in 190! near Gulakhma.2
Within half a mile to the east of this, and quite close to the edge of cultivation, there was reached the northernmost part of the extensive ' Tati ' generally known to the people of Charchan as the ` Kane-shahr'. Some nameless saint's Ziarat, called Yalghuz-lugh Nazar (` the shrine of the lonely yak-tail '), marks its present limit northward. About a quarter of a mile south of this I could trace scanty remains of foundations of walls built of clay and bricks, all dug into by ` treasure-seekers '. The bricks measured about twenty by ten inches, with a thickness of about three and a half inches. Otherwise the site wore the look of a regular ` Tati ', the potsherds mostly of coarse material resting on eroded riverine loess. The extent to which wind erosion has here lowered the ground-level could be gauged by far-stretched loess terraces extending on the east of the ` Tati' along a flood channel which carried water to a newly irrigated part of the oasis. These terraces, which moisture had evidently helped to protect, also bore pottery débris on their top.
Recently levelled fields fringed this ` Tati ' on the north and north-east, and confirmed the
statements of my local informants that much of the ' Kbne-shahr' area had been brought under cultivation again by the gradual southward extension of the oasis. M. Grenard, who visited
Charchan in 1893 and first noticed its ' Kône-shahr', evidently found the structural remains more extensive, though not differing in character from those I could still trace.2 He was inclined to attribute them to the Charchan of Marco Polo's time, and to look for the ruins of the older Chii-mo further north about Tatran. But he based this belief on a supposed change in the Charchan River's
' The line of this old canal was crossed by me again at two points to the south-west of the present oasis. Owing to an error of the Surveyor, who passed here after dusk at the end of a long march from the foot of the mountains, Map No. 47 shows the head of the Yangi-iistang as taking off from the Ayak-tar stream instead of the Charchan River. In Dr. Hassenstein's map, based on Dr. Hedin's survey, the uppermost line of the canal is correctly delineated ; but by a reverse error it is treated there as a side branch of the
Charchan River and conjecturally shown as uniting itself lower down with the bed of the Ayak-tar, while in reality its course northward remains quite distinct. The point where the route southward crosses the canal is correctly marked in both maps.
2 Cf. Ancient Khotan, i. pp. 465 sq.
3 Cf. Grenard, Mission D. de Rht'ns, i. pp. 183 sq.; iii. P. 146.