Sec. iii] THE CHARCHAN RIVER ROUTE AND VASH-SHAHRI 305
The remains to which Ismail guided me on the following day, though manifestly of late date, were yet curious and at first sight rather puzzling. Leaving our camping place at Chong-kul (see Map No. 5o. A. 2), we made our way through a belt of spring-fed, reed-covered marshes, and then over ground where salt-impregnated patches alternated with thickets of luxuriant riverine jungle, to a ridge of tamarisk-covered sand cones known as Yalçhiiz-dong, ' The Lonely Hillock'. I found there the remains of three small isolated structures, with oblong walls of very soft brickwork, occupying the tops of separate knolls within thirty to forty yards of each other. The walls showed a uniform thickness of two feet. Their bricks were mixed with plentiful wheat straw, and measured about sixteen by eight inches on the average, with a thickness of four to five inches. The best preserved and apparently largest structure measured twenty by thirteen feet. The walls stood only a few feet above the ground, and though partially protected by drift sand which the tamarisk scrub had detained, they nowhere showed remains of any superstructures. On the slopes of the knolls, rising to about thirty feet above the plain, there lay some large pieces of poorly worked timber, apparently all Toghrak, which Ismail thought might have belonged to coffins ; but there was nothing to show their original position or use. It was not easy to account for the poor construction of these walls and their strange situation. Their survival notwithstanding the softness of the bricks suggested no great age.
The puzzle was not yet solved when, after going for about one and a half miles to the northeast across dried-up marsh, I was shown by Ismail a second group of small rectangular structures closely resembling the first in material and appearance. They occupied a low ridge by the side of what was unmistakably an old irrigation canal. Here, too, the walls built of soft bricks stood only two to four feet above ground, and showed no sign of having ever borne superstructures. A depression a short distance to the north of the ruined walls was lined by a thicket of dead Toghraks, and distinctly recalled an old river bed. After following the traces of the canal, which dead tree-stumps standing in line helped to mark, for about three-quarters of a mile eastwards, Ismail brought me to the principal group of the ' old walls ', Kône-tamlik, as he called the whole site. Here I found a row of over a dozen rectangular enclosures, built of the same brickwork but more completely preserved, stretching without any distinctive plan from east to west along the top of a low ridge. They were all detached and their size varied greatly, the largest being in the centre and measuring close on fifty by forty-two and a half feet. The walls were only some four feet in height. But on one side each enclosure showed a narrow arched gateway standing to a somewhat greater height than the rest of the walls, a clear proof that the latter were never meant to bear a superstructure.
Thus the true explanation of these strange ruins very soon revealed itself. Everything recalled the walled enclosure so often seen in Muhammadan cemeteries further west in the Tarim Basin, and this conjecture was promptly confirmed by the discovery of a grave just outside one of the enclosures near the north-east edge of the ridge. A little clearing revealed the end of a coffin formed of a hollow tree-trunk and covered on its top by a row of rough Toghrak branches laid across much after the fashion I first observed in the old Muhammadan graveyard of Hasa near Moji.2 When the few men with me had scraped away enough of the soil to display the feet of a woman or child turned due south, it became quite certain that the remains were those of a Muhammadan burial-place.3
2 See Ancient Khotau, i. p. I r 2.
' As Ismail was also Dr. Hedin's guide it is very probable that the old Mussulman burial-place with several gumbez (tombs) ' to which passing reference is made by him in Central Asia and Tibet, i. p. 308, at a point south of the
Charchan River, is identical with the remains just described. The mention made also of houses' is not at variance with this, since the dimensions recorded of the largest among them clearly point to identity with the enclosure above described.