National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

644

THE END OF THE CHINESE LIMES   [Chap. XVI

Position of watch-tower T. VI. b.

Advantages for point

appui.

Quarters of watch- tower T. vi. b.

horizontally, completing what served as a kind of internal framework for additional strength. By the east face of the tower was a flight of stairs about one foot wide, of which the four lowest steps survived. Close to them were the foundations of single-brick walls enclosing five small recesses, only two to three feet wide and three to four feet long, obviously intended to hold stores. On the steeply-eroded slopes to the north and east other structural remains may have disappeared altogether. A large dung-heap on the west had fared better ; but neither there nor at the foot of the tower were any finds made. About 200 yards off to the south low bundles of reeds and Toghrak branches were discovered on the top of the gravel plateau, embedded in coarse sand and lying in a broken line about forty feet long, as if marking some enclosure.

SECTION III.—THE ANCIENT STATION T. vi. B AND ITS RECORDS

Immediately to the south of the ridge bearing T. vi. a, a deep-cut dry river-bed, about twenty yards broad, was crossed, and beyond it an area was entered where low scrub-covered ground was broken up by small plateaus and isolated clay terraces. Further on a narrow gravel-covered ridge rose to about ioo feet above the depression and, as usual, at its western end tapered away into a short line of detached terraces. Both the centre of the ridge and the last outlying clay terrace, or Mesa, were occupied by towers, a fact which, being unusual on this flank of the Limes, was bound to attract my attention at once. As the distance between them was less than three miles, and as the tower on the ridge, T. vi. b (Fig. 171), lay well behind the line, this could not have been intended for a mere signalling post. The débris adjoining the tower on the east seemed to indicate quarters somewhat larger than usual. So the thought soon suggested itself that the position marked by the ruin might have been that of some main station that controlled this flanking section of the Limes.

The numerous layers of refuse which were scattered over the gravel slopes near the ruined station seemed to agree with such an assumption and, in any case, furnished proof of prolonged occupation. There could be no doubt either that the position was one which offered special advantages for a sort of point d'appui or main guard-post. Though only about a hundred feet or so above the low ground at the foot of the ridge, it commanded a complete and open view of the southern and south-eastern shore of the marshy basin. This extended to the foot of a distant chain of low hills, entirely covered with big dunes, which was seen to come from the direction of Nan-hu and stretch far away westwards. The belt of high sands along it seemed to bend round the south-western end of the wide marsh-filled basin and to continue towards the great sand ridges flanking the Bash-toghrak valley (Maps Nos. 71, 75, 79). It was clear that any inroads attempted here from the north or west of the terminal depression of the Su-lo Ho would have to keep between the shore of the marshes and the impassable high sands. The ground left there is a narrow belt of gently sloping and absolutely bare gravel ` Sai '. Consequently, the line of watch-posts that extended across the south-eastern corner of the basin (Map No. 74. A. 4) to within sight of those high sands would, when controlled and supported from a main post at T. v1. b, effectively protect the Limes proper and the important route it was meant to safeguard from being outflanked by raiders.

The tower T. v1. b in its ruined state rose to a height of about sixteen feet and, as the plan (Plate 37) shows, had a base of some twenty-one feet square. It was built of sun-dried bricks, measuring about 142" x 7"x 5" with the usual layers of reeds at intervals. Excavation of the