国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

640   THE END OF THE CHINESE LIMES   [Chap. XVII

The tower T. H was found to occupy the western edge of the same gravel-covered plateau, that here rose about fifty feet above a broad depression with reed-covered patches. It was in fairly good preservation, built with layers of stamped clay about three inches thick and standing to a height of fully twenty-two feet. Here large stones, stored with the purpose just explained, were still seen actually on the top of the tower. Others had fallen and got jammed in a fissure running down the upper part of it. Here, too, wind-erosion had laid bare the ground along it on the north and south to a depth of two or three feet. No débris from any adjoining structure, nor refuse layers such as would point to continuous occupation, could be traced near. The absence of potsherds was also significant. But, searching the neighbouring ground, we picked up some metal fragments, T. II. ool. a—d, including two arrow-heads of the Han type in bronze and iron.

From T. II I continued my reconnaissance along the Lop route for another four miles or so north-westwards without being able to find the ruined ` Pao-t`ai' which Rai Ram Singh, when previously surveying the ground between the several terminal river-beds, had marked on the map, and which on the strength of this has been shown as T. 11, a in No. 74. B. 3. He had sighted it from the west, as his route shows, but had not been able to visit it. My failure on May 2, 1907, to discover it might be attributed to the dust haze which a strong north-west gale had raised in the afternoon of that day. But as the same experience was repeated when I passed here again in 1914, under atmospheric conditions somewhat better, I am now inclined to believe that the surveyor's entry may have been due to some visual deception, easily accounted for on such ground where refraction and glare are apt to give the appearance of a tower to a mere natural knoll of clay. In any case it is clear that the absence of water would have made the maintenance of an advanced line of posts here very difficult in ancient times.

Tower T. n.

No more towers along Lop route.

SECTION II.—THE SOUTH-WESTERN FLANK OF THE LIMES

Watch-stations guarding flank of Limes.

Distances between watch-stations on flank.

Before we follow the line of the Limes eastwards, it is desirable to describe the observations and finds attending the exploration of the detached watch-stations, five in all, which guarded its south-western flank. The broad topographical facts, and the military considerations based upon them, which induced the constructors of the Limes to use the marsh-filled terminal basin of the Su-lo Ho as a natural flanking defence, have already been fully set forth in the preceding section. What I have to describe here will help to illustrate them in detail.

Moving from Camp 171. a to the south-south-west on the morning of May 3, I first reached, across boggy ground, a low broad terrace less than two miles distant, where Surveyor Ram Singh had previously reported some ruined dwelling. It proved to be of very recent origin and to mark what evidently was a regular grazing station, visited by Chinese Muhammadan herdsmen down to the times of the last Tungan rebellion, and perhaps later, too. Under the abundant refuse-layers that covered the clay terrace I failed to trace any ancient débris. Yet there is a topographical consideration which suggests the possibility that this point was already occupied in the times of the Limes. An examination of the map, Plate 33, shows that the distance between the tower T. iv. c and the next one to the south-west, T. v, over seven and a half miles, is far greater than that between any two of the other watch-stations, T. v, v1. a—d, which belong to the flanking line along the terminal marsh basin. There the distance is uniformly about five miles. From the map it is also seen that the position of the terrace and hut above mentioned falls very close to the almost straight line which connects those flanking posts with each other and with T. Iv. c. Both T. Iv. c and T. v are visible from the terrace, which thus could conveniently have served for an intermediate signal station. But in the absence of direct evidence this must remain mere conjecture.