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. v]   THE HAN LIMES FROM YÜ-MEN-HSIEN TO AN-HSI   1137

the relation of the Han Limes remains discovered near Shih-êrh-tun to the rest of the ancient border-

line, both to the east and the west of this point, was fully cleared up only by the surveys of 1914,

and must be left for discussion elsewhere.

Close above the left bank of the stream where it irrigates the fields of Shih-êrh-tun there rises Watch-

the outermost of a succession of narrow low ridges covered with fragments of decomposed rock, as towers of

Han Limes.

seen in many places along and within the utterly barren Pei-shan ranges. It is crowned by a line of watch-towers, some badly decayed, some repaired, but all showing an ancient core. The eastern-

most of those found above Shih-êrh-tun had been refaced with rough brickwork, hiding in most

places the ancient layers of stamped clay. But a small fort below it, at the foot of thé ridge, looked decidedly ancient. Its remarkably massive construction with solid clay walls 10 feet thick, enclosing

a space 52 by 42 feet within, distinctly recalled the small fort at the ancient Yii-mên, T. xmv;

significantly enough it bore the same local designation of Hsiao fanKp'an.3 On following the ridge to the west for about half a mile I came upon a second tower, badly decayed but furnishing conclusive

proof of its antiquity. The familiar reed layers of the Han Limes separated its strata of stamped

clay at intervals of i 2-14 inches. It was about 19-20 feet square at the base, just like so many of the ancient watch-towers I had examined on the Tun-huang Limes, but was broken at a height of

about 14 feet. A third tower found over a mile further west crowning the same ridge had suffered even more, rising only to 8 feet or so, and even this remnant fissured by wind-erosion. But here, too, the thin layers of reeds dividing the strata of stamped clay were distinguishable in perfect clearness.

It was first at this tower that I noticed the slopes of the ridge, only 15-25 feet wide at the top, Remains of and about 3o feet above the adjoining ground, to be thickly strewn with half-petrified twigs of Limest tamarisks and wild poplars. Their abundance on this ridge, where trees could never have grown, agger,

at once suggested use in a rampart constructed in the fashion of the Limes wall north and west of Tun-huang. Closer search soon revealed, about 4o yards west of this tower, a portion of the

rampart or agger still in situ on the top of the ridge. The twigs and branches were embedded

between layers of detritus, used in the sanie way as the clayey gravel had been along the Limes line beyond Tun-huang. Owing to the coarseness of the material here available, the agger had been

built with greater thickness at its base, the foundation layers in situ measuring about 15 feet across. Its remains could be traced with more or less clearness for a distance of about 30o yards along the detritus-covered ridge before it finally sinks away to the level of the salt-encrusted plain at its foot. Here the line of the wall completely disappears on ground manifestly liable to be flooded at times either from the Su-lo Ho or by such drainage as may at rare intervals descend from the gravel plateaus to the north that mark the edge of the Pei-shan.

Some three miles away to the north-west a conspicuous tower rises at the foot of the nearest Eastward plateau, and towards this, as I found later, the line of the wall continued. But its further explora- continuation tion became possible only in 1914, when I revisited this ground after having traced the Limes right of Limes. through from the direction of An-hsi. On the same occasion I was able to follow up its line from Shih-êrh-tun eastwards to the little oasis of Hu-hai-tzû, also known by the significant name of Ying

fan   (military encampment), and thence across a great sandy waste to the lower course of
the Su-chou River near Mao-mei.4 The description of the line thus followed by the Han Limes and the discussion of the reasons which are likely to account for its choice must be left for a future publication. Here it will suffice to point out that the section of the Limes line stretching between

The name AN KS   means literally ' small protective

camp' ; cf. above, p. 683, note 2 a.

The fort symbol in Map No. 85. A. 2 near Shih-êrh-tun has been erroneously placed south instead of north of the

Hsi-wan Ho stream. It ought also to have been printed in red.

4 Cf. Third Journey of Exploration, Geogr. Journal, xlviii. P. 195..

1814   7 F