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

Sec. ii]   THE POSITION OF THE YU-MÊN HEADQUARTERS   693

a distance of about 9 yards to the east of the wall, except where it was seen to turn off towards the ruined tower as it neared it, thus furnishing definite evidence of contemporary origin.'a

SECTION III.—THE WALL BETWEEN YÜ-MÊN AND YANG KUAN

Before we consider the important bearing which the discovery of this transverse wall has upon the location of the Jade Gate, we must endeavour to arrive at a clear perception of its original purpose and its chronological relation to the main line of the Limes. Taking the latter question first, it appears to me certain that the construction of the transverse wall can neither have preceded, nor have been contemporary with, the extension of the main line of the Limes to this ground. In the first place, it is highly improbable that those who under the Emperor Wu-ti, in order to assure safe access to the Tarim Basin, pushed forward the fortified line in a few years over a vast distance and in the face of formidable natural difficulties, should have been satisfied here with a wall of far less strength than that which, as my explorations further east both in 1907 and 1914 have proved, was maintained by them continuously across hundreds of miles of desert. Nor could this wall have been built as an independent and earlier barrier to watch the route from the west—even if such could have been a main object of the Chinese authorities in those times of a vigorous ` forward policy ' ; for in this position at T. )(Iv it would have been quite easy to turn it from the north flank as long as the main wall did not exist to afford protection on that side. If such had been the object of the transverse wall, we ought to expect its northern end to abut upon the marshes near T. xvii or T. xvII1, which, together with those towards the Khara-nbr, would have made its flank quite safe, while the length of the line would have remained practically the same. Thus we are obliged to recognize in it a work of later date than the main line of the Limes.

But what was the true object of this later cross-wall ? Two explanations seem possible. At Purpose of

first sight the idea suggests itself of an inner line built not primarily for increased military defence, transverse

wall line.

but chiefly for purposes of police control, in order to watch effectively the coming and going of individual travellers and caravans, to levy duties, and in particular to keep the administrative area of the empire clear of the incursions of troublesome ` barbarian ' visitors from the protected territories beyond. It would be easy to find analogies for such an inner cordon line both in the organization of the lines of the Roman Limes and on modern frontiers in the East.' It is obvious that such a cordon might have been organized side by side with the maintenance beyond it of a strong line for military defence such as the main Limes wall, well secured at its extreme western flank, was undoubtedly intended to be. But in the light of a chronological fact revealed by the documents another explanation claims attention, and there are topographical observations which

Transverse wall of later origin.

15 Another curious observation, and a more puzzling one, may also find record here. From about half a mile to the south-south-east of the fort of T. xiv to beyond the newly discovered tower I noticed a low earth embankment, about 3 feet wide at the bottom and about r feet high, running parallel to the wall, outside, i.e. to the west of it, and at the regular distance of 5o yards. Its purpose remains obscure for the present, and I did not observe a corresponding feature along the Limes wall proper. Could it possibly have been intended to mark the exact limit of the administrative border, as has been assumed, I believe, of the small ditch traced outside the wall on certain lines of the Limes of the Roman Empire ?

I For an interesting parallel furnished by the lines of Hadrian's Limes in Pannonia, Dacia, Dobrucha, etc., cf. Prof. E. Kornemann's illuminating survey, Die neueste Limesforschung, Kilo, 1907, pp. 97 sq., 104 sqq., and passim. Many problems like the one discussed above, but often far more complicated, must engage the attention of scholars concerned with the study of Roman Limites.

Among modern parallels it may suffice to mention the line held on the Indian North-West Frontier by the Military Border Police posts. It runs well behind the true Limes represented by the advanced routes and passes which are guarded by the various Frontier Militia Corps of the Khyber, the Kurram Valley, Waziristan, etc.