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

T

Suggestion of defensive line.

Arguments against defensive character.

6i8   THE OASIS OF NAN-HU AND THE YANG BARRIER [Chap. XVI

to serve as a guiding line for travellers when crossing the bare gravel plateau, to and from the banks of the Tang Ho, at the time of violent dust-storms such as sweep across the desert with great frequency in the spring and summer. Incidentally, the embankment was supposed to afford some protection from the force of the specially dreaded gales that blow from the north-east or north. This may probably account for the name AT ch`iang, ` wind wall ', by which the embankment is now known to the Nan-hu people. Whether this designation is old, and whether the interpretation I heard of it rests on some kind of tradition, I have no means to decide. If correct, it would furnish for that strange dyke an explanation. which archaeologically seems a ftriori admissible. The need of guarding travellers from the serious risk of straying off the track and losing themselves in waterless desert during violent sand-storms is proved by the measures that the present Chinese administration has taken to mark the ` high road ' across desert stretches, both east and west of Khotan, with lines of closely-set poles.6 It was forcibly brought home to me more than once by personal experiences of travel on desert routes of the Târim Basin.

But there is another possible explanation of this curious embankment running across the desert which deserves attention here. The site of Nan-hu must, as we shall presently see, be identified with the ancient Yang kuan, or ` Yang barrier ', of Han times, and the connexion of the westernmost Limes with Nan-hu is proved by a secondary line of wall traceable south-east from T. xiv, the ancient station of the ` Jade Gate ', or Yii-mên, towards the end of the formerly cultivated area of Nan-hu.ba The question, therefore, necessarily suggests itself whether, in the dyke stretching across the gravel plateau to the natural fosse of the Tang Ho, we ought not to recognize the remains of a defensive line intended to protect the ` Yang barrier ', and with it the westernmost Limes, from attack in the rear. The utility of such protection could not be denied a lit-nine; for, when discussing above the ` southern route ' leading in Han times from Tun-huang to Shan-shan or Lop, I have had occasion to point out that this route, corresponding to the present itighyol, or ` mountain route', between Tunhuang and Lop, passed near to the territory in the K`un-lun held by the Jo Chiang tribe.e Together with other nomads on the high plateaus to the south of Tun-huang, such as the Little Yüeh-chih, they must have remained a potential source of trouble.' Nor does the assumption appear altogether impossible that the embankment may, as Mr. Littledale thought, represent the remains of a completely decayed ` Chinese Wall'.

Yet there are weighty arguments to be urged against such a view. Most important is, I think, the fact that the remains in question differ wholly in construction from those of any other wall on this westernmost Limes. On repeatedly examining the embankment I failed to trace in it any of those layers of fascines which elsewhere are the most characteristic feature of the wall in the Tunhuang region. Yet the materials for such fascines in the form of tamarisk brushwood and reeds could be secured in abundance both from the Nan-hu basin and from the gorge of the Tang Ho. Then again the total absence of remains of watch-towers along the embankment, apart from the one at its western end, is a very significant indication. It must further be noted that, if the line was intended to defend the road to Tun-huang from attacks on the south, we might reasonably expect it to start from the ruined circumvallation, which clearly goes back to Han times, and not from a point considerably to the north of it. The same observation also applies, and with increased force, to the position of the line relative to the area of graves which, as already related, I found extending on its south where the edge of the Nan-hu basin is approached. Had the line been laid

6 Cf. Ruins of Kholan, p. 180 ; Ancient Kholan, i. p. 468 ; Hedin, Reisen in Z.-A., p. zoz.

6a See chap. xix. sec. iii.

6 Cf. above, p. 418 ; Chavannes, T `oring pao, 1905,

pp. 526 sq., note 8.

7 For relations attested by the Han Annals between the Jo Chiang and the Huns down to A.D. 2, cf. Chavannes, T`oung pao, 1905, p. 527, note 8.