Sec. ii] KHARA-KHOTO AND ITS REMAINS 439
the abstract of his preliminary report reproduces at length.' But since this story points very significantly to some traditional folk-lore connexion between the cutting and great hidden treasures, it seems quite possible that the former owes its origin to the operation of early treasure-seekers, more persistent than others of whose handiwork plain signs were to be seen in abundance elsewhere. In support of this conjecture it deserves to be mentioned that at several points in the north wall, close to the passage through it, a number of tunnels and cavities had been burrowed into the clay rampart from the inside, while a neighbouring attempt, at the point marked 4 in Pl. 18, had effected a breach not merely deep down into the wall but also penetrating into the bastion in front of it.
In addition to the already mentioned outworks, at the gates, the arrangement of which is clearly shown by the sketch-plan in Pl. 18, the walls were defended by big circular bastions, slightly varying in dimensions, at the four corners and by rectangular bastions along the sides. There are of the latter four each on the western and eastern faces, and six and five respectively on the north and south. These rectangular bastions also vary in size, the largest, defending the outworks of the gates, having a frontage of 47 feet. Outside the western and southern faces there were found in places scanty remains of a kind of covered way, formed by a wall of stamped clay, 10 feet thick, carried in front of both curtains and bastions. Of a fosse I was unable to trace any indications.
The way in which the drift-sand has been slowly but steadily attacking the town walls of Kharakhoto offered occasion for observations of special interest to me in the light of those made at sites like those of Lou-lan, So-yang-ch`êng, An-hsi, where wind-erosion has found full play for its force. As the photographs in Figs. 242, 243, 246 clearly show, the prevailing north-western winds have heaped up big dunes against the western and northern walls. Where the corners of the projecting bastions caught them, the dunes have reached the summit of the wall, and here the corrosive action of the wind-driven sand has not only completely carried off the parapet, but cut trough-shaped breaches into the body of the wall, to a depth of 6 feet and more. The sand which has thus entered the circumvallated area has been partly left on the lee side under the shelter of the western and northern walls (Fig. 245), and partly driven across the interior until stopped by the inside of the eastern and southern walls. Here the same breaching action is repeated (Fig. 242) ; but owing to the absence inside of corners formed by bastions the position of breaches on the top of the wall is less regular. Yet the force of wind-erosion was well marked here also, as shown by the eastern wall having been carried off for a length of 32 feet to a depth of 6 feet.
With the massive solidity of the circumvallation and its comparatively good preservation the utter decay and consequent emptiness of the interior of the town presented a striking contrast. As seen from the sketch-plan in Pl. 18 and the photographs Figs. 244, 245, the greater part of the area appeared as a desolate waste with only a few ruined structures rising above the level expanse of decomposed clay and small debris, while the outlines of other buildings could be traced only with difficulty by low remnants of walls here and there, stumps left of wooden posts, or by foundations and floors of hard bricks. Standing on the height of the walls it was possible to make out certain
4 See Geographical Journal, October /909, pp. 387 sq. This story ascribes the breach to a legendary ruler of Kharakhoto, ` Khara-tsian-tsiun ', who, when besieged by an imperial Chinese army and deprived of water by a diversion of the Etsin-gol, attempted to break out here with his troops, after having buried his countless treasures in a well that had been dug near by in the vain hope of reaching water.
A wide and deep hole- undoubtedly marking a well is found close to the cutting (see 3 in Pl. 18) and may have
formed the starting-point for the particular form of the folk-lore account heard by Colonel Kozlov. That traditions of great buried treasure have gathered round the site is certain, and there is abundant evidence that treasure-seekers have made prolonged search for it by burrowing at the different ruins. All I could learn, through the defective channel of our interpreter Malum, was some vague connexion between the breach in the north wall and the siege that Chinghiz Khzn was believed to have laid to the town. About the cutting in the south wall I was able to learn nothing.