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

raise their top almost to the level of the gravel plateaus to the east and west. In consequence it was possible, as I ascertained by actual observation, from the parapet of the ruined fort not merely to scan the whole of the depression for a considerable distance, but also to keep a look-out over wide stretches of the level ` Sai ' on either side and northward. This is best illustrated by the fact that I could sight from that point of vantage all the watch-towers of the Limes from T. xi to T. xix, and that in spite of their present ruined condition. I have little doubt, from what practical experience on such desert ground has taught me, that, given reasonable atmospheric conditions, it would be possible to observe from there signal-fires lit at night on the ancient towers over even greater distances, probably from T. Ix to T. xxii close to the shore of Khara-nbr, more than 3o miles in a straight line. Nor should other practical advantages be ignored which make this position obviously suited for a military and roadside station of importance. The depression both to the north and to the south affords ample grazing on reeds and scrub, and springs on the edge of the marsh, quite close to the north-west of T. xiv, provide drinkable water.

Massive   These considerations had from the first impressed me, and the fact of the position being

construction

of fort.   occupied by a ruined fort of such massive construction as T. xiv seemed naturally to raise the

presumption that it might mark the site of some sectional headquarters for this part of the Limes. But, imposing as the ruin of the fort looked with its thick walls of stamped clay, there was nothing

in its structural features to serve as a definite indication of its age. I have already briefly described

them in connexion with my first passage along this route, and Figs. 183, 184, which show the small stronghold as seen from the north-east and south-west respectively, together with Plate 4o will help

to recall the essential details. Antiquity was suggested by the fact that, in spite of the thickness

of the walls, fully 15 feet at the base, and the remarkable solidity of the clay, considerable portions of the outer faces to the north and east had fallen, obviously through wind-erosion. Its

effect was clearly marked also by the under-cutting which the north-west corner had suffered, as seen in the photographs. More difficult to explain was the big opening, resembling a pointed arch and 13 feet wide below, which appeared near the middle of the wall of the fort facing north (Fig. 183). It seemed to have been caused accidentally by the mass of clay falling outward. The gap had been closed by rough brickwork which seemed late. The true entrance to the interior of the little stronghold led through the west wall, as seen in Fig. 184, and was only 8 feet wide. I noticed repairs of distinctly old appearance in the north-west corner, where a fissure in the wall had been filled up with fascines of reeds secured by Toghrak branches.

Within the fort I searched in vain for ruins of quarters or other approximately datable remains. It is true that the interior, about 54 feet square, was covered with a crust of refuse 4 to 5 feet high. But wherever I had this cleared to the natural soil only ashes, bones of animals, and stable refuse could be found. The complete removal of this unpromising stuff could not be attempted with the limited labour and time at my disposal. At the outside of the north wall I had come upon a thin layer of refuse with pieces of the same hard grey ` mat-marked ' pottery which was common near the watch-towers. Similar finds near the surface induced me to continue the ` prospecting' towards the foot of a bare, gravel-covered hillock (Fig. 179), the western end of which lay about 70 yards to the north of the fort. It was on April 20, at which time the clearing of T. x11. a still kept almost the whole of my small working-party busy. But, as I made the man with me scrape the slope of the mound at different points, layers of straw and other stable refuse were disclosed in more than one place below the cover of gravel.

Finds such as a blank slip of wood, a tiny piece of paper with a couple of half-effaced Chinese characters, and rags of silk and hemp had first encouraged hope when the man's spade, at a point about 15 feet above the west base of the hillock and about 6 feet below its top, laid bare

No datable remains within fort.

Discoveries on hillock.