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

Sec. vii]   THE GREAT MAGAZINE OF THE LIMES   711

T. xv proved to be a badly ruined tower occupying the top of a small but conspicuous clay terrace about 20 feet high, which rises near the eastern end -of the dry wind-eroded basin previously mentioned. The brick masonry, though much decayed, showed the usual intervening layers of tamarisks and stood to a height of about 13 or 14 feet. Débris covered the remains of a small room adjoining the east foot of the tower ; of neither could the exact dimensions be determined. Among the miscellaneous relics, found under this débris and described in the List below, I may mention two cast-iron hoe-blades of excellent make, T. xv. 004, 009 (Plate LIV), and a spade of welded iron, ooio (Plate LIV) ; a woven hemp string shoe, oo6 (Plate LIV), and fragments of a woollen pile carpet, 007 (Plate XLIX). A curious find was made on the steps cut into the clay which led up to the tower, in the shape of some fragments of iron, 005, which evidently belonged to old matchlocks. They were found tied up into a little bundle under about a foot of débris which had fallen from the tower above. Their iron was far less corroded than that of the hoe-blades unearthed in the ruined quarters. It seems to me very probable that the bundle of fragments had been lost here by some hunter who had used the tower, which commands an extensive view, as a look-out for wild camels, still frequent on this desolate ground. I subsequently found the top of a similar terrace north of T. xvi crowned with a sangar of clay lumps, evidently built as a screen by a still later hunter.

From this point the wall could be traced, unbroken though low, across a gravel-covered plateau to T. xvi, about 1 ? miles distant. This tower rose on the southern edge of a dry scrub-filled basin and commanded a distant view both to the north and east. Built with sun-dried bricks, 14" x 7"x 42" in size, and having layers of reeds after every three courses, it still stood in fair preservation to a height of over 13 feet. As the plan (Plate 39) shows, it measured 24 feet square at the base and contained on its top a small guard-room, i, 8 . feet square. Built against the west foot of the tower were the débris-covered remains of another small room, ii. At a distance of about 5o yards off to the south-east there rose the half-petrified remains of three stacks of reed fascines of the usual size. From the guard-room i there were recovered ten inscribed slips, among which those still decipherable are given by M. Chavannes in Doc. Nos. 579-83.2 The room ii also yielded a few fragmentary records. Among the former are two fully-dated documents, Doc. 579, 580 (Plate XVII), containing receipts of A. D. 68 and 77 for rations issued to two soldiers from

the Fu-kuei   canton of the command of Tun-huang. To the Later Han period belongs also the
fragment of an inscribed wooden stick, T. xvi. 2 (misread b), Doc., No. 581, on which is read the name of the Chin-clang #4 command instituted in 86-74 B. c.

From T. xvi the line of the wall, in places still in fair preservation, was traceable across gravel terraces and intervening shallow depressions to the tower T. xvii. This rose on the western edge of a wide lacustrine basin crossed by the Su-lo Ho and filled to the north and east with lakelets and extensive marshes. The tower, about 22 feet square at its base, was built of brickwork like that of T. xvi, but was badly broken. Its actual height was about 10 feet. Close to it, on the south side, were the scanty remnants of a wall which seems to have belonged to some small quarters. No finds attended the clearing here. About 120 yards off to the south-east there were remains of a square stack of reed fascines. About 20 feet from the tower in the

Remains of watch-tower T. xv.

Finds at watch-tower T. xvi.

Remains at tower

T. xvIi.

To these must be added probably Doc. Nos. 545, 561, the ' site-marks' of which slips, having become partially effaced, have been misread as T. xv.

I may note here that these and a few other occasional misreadings of ' site-marks' on Limes records were due to the minute markings in ink or pencil, which were all that I could make on the very limited space available, having become

difficult to read before M. Chavannes examined them in 1909-10. Their rectification could be effected only through close comparison with my original notes on the ' finds' from different stations, etc., and it was not possible for me to make this comparison when M. Chavannes' Documents were passing through the press in r g r r-12.