At the time of Prejevalsky's visit in 1876, when Charkhlik was still mainly a small convict settlement, the circumvallation is likely to have been in somewhat better preservation. ' On the site of the present village of Chargalyk,' he states, ' remains of mud walls of an ancient city, called Ottogush-shari, may be seen. These ruins are reported to be two miles in circumference, and watchtowers stand in front of the principal wall.' But far more striking in appearance than the parts of the rampart still extant was the ancient mound known as the Tora, or tower, to which my attention was called, about one mile away to the north-west. There, some 300 yards from the left bank of the river in what is called ' Kurbân Beg's Mahalla ', rose a large and remarkably steep mound to a height of over fifty feet above the irrigated level. The layers of rubbish, with plentiful large stones exposed on its slopes, left no doubt about its artificial origin. On the top I could trace the much-dilapidated remains of a massive brick structure, about twelve yards in diameter. Its masonry on the north side was exposed to a height of twelve feet. The bricks had weathered badly, and only their thickness and width, four and twelve inches respectively, could be made out with certainty.
There could be little doubt that these were the remains of a Stûpa dating from Buddhist times. But the mound below was far too large to be formed by the débris of any single structure, and the strata of rubbish composing it, where laid bare by cuttings, distinctly suggested an origin from prolonged occupation, reaching back probably to prehistoric periods. Examining the layers cropping out on the steep slope above an irrigation channel which had been cut into the south foot of the mound, I was vividly reminded of the observations that I made in 1904 on the large mound of Shorgalai, near Thal, in Balüchistân.8 The ancient refuse mixed with the stones contained much dung and reed straw, and was being regularly dug away for manuring, just like the khaura earth of ancient mounds along the Indian North-West Frontier. It is likely that this practice may have led to the disappearance of other and smaller mounds within the cultivated area.
Nowhere around Charkhlik did I hear of the existence of such débris-covered Tatis ' as would seem to belong to areas of ancient occupation that had been subjected to far-advanced wind-erosion.' Nor were any of the small antiques so commonly found on such ground forthcoming for sale here. We may, I believe, safely conclude from these negative facts, and from the position of the old town walls in the middle of the present oasis, that the new settlement approximately occupies the same ground as the precursors of the modern Charkhlik. At the same time this absence of visible proofs of erosion and denudation makes it difficult for me to believe that the period of complete abandonment, which certainly preceded the creation of the existing oasis, could have been a very protracted one. Had the site been abandoned before mediaeval times, the erosive forces ever at work, as we shall see, in this wind-swept Lop-nor region would scarcely have failed to denude the soil, at least in parts, right down to the Piedmont gravel, and to efface completely those far from massive walls. These observations will explain why, in order to trace the history of Charkhlik, it will be safest to start from the later, and deal afterwards with the earlier, notices.
distance from the southern edge of the present oasis. They are not of the ' Tati ' type, and the second, near the left river bank, occupies ground which could never have been under cultivation. Both sides go back to Tang times, and I much regret not to have learned of their remains on my first visit to Charkhlik in 1906-7.