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

Sec. ii]   THE POSITION OF THE YO-MEN HEADQUARTERS   691

(Plates XI, XII). The first was intended to introduce a certain local magistrate to the Governor, while the second gives instructions as to the particular kind of official seal to be used by the latter's deputy. The title Chien jên, ` chief of a thousand men ', which is found in several documents, Nos. 305, 375, 376 (Plates Ix, xI, xII), may also be noted here, as its use for superior members of the military hierarchy established in the frontier commands of Han times is attested.12

In connexion with the documentary evidence from the site of T. )(Iv it only remains for me to point out that its identification with the ` Jade Gate ' headquarters is consistent with our knowledge of the other localities that we find mentioned besides Yii-men in the records of this site. As regards Ta-chien-tu (or Chien-tu), named in Doc. Nos. 304, 307, 356, I have already had occasion to show that it must in all probability be identified with the westernmost section of the Limes, and that it was controlled from Yü-mên.13 As the first-named document, No. 304, dates from 96 B. c. and the last from A. D. 15, we see that this administrative connexion was maintained for more than a century. The watch-station Ta fu, named in T. xiv. iii. 64, Doc., No. 309, is not mentioned elsewhere and cannot be located. We are in a better position as regards the local name Ping-wang 21t, Doc., Nos. 313, 314, 377, repeatedly met with also in records from other sites as the designation of a watch-station, a company, or a ` barrier '. From an examination further on of these records, more definite than those from T. xiv, we shall see that the name was probably borne by the section of the wall extending from T. xxii. c, on the Khara-nor, to T. )(Iv. a, where it joined the ` Yü-mên barrier '•13a

If we review the conclusions arrived at by the detailed scrutiny of the records from T. xiv, and join with them what our preceding survey of the natural advantages of the site and its topographical setting has shown us, it is easy to realize that, for the period covered by the extant dated remains of the Limes, the position of the ` Jade Gate ', the chief frontier station through which all traffic westwards by the ancient Lou-lan route had to pass, may now safely be fixed at T. xiv. Well withdrawn behind the defensive line of wall and watch-towers, and protected besides from direct attack by impassable marshes to the north-west and south-east, the position was admirably adapted to serve as a point d'apßui for the posts along the most advanced section of the Limes. It was equally well chosen as a head station for controlling traffic by the route which, from the reign of the Emperor Wu-ti down to the close of the Former Han dynasty, was certainly the main link between the Chinese empire and its Central-Asian ` sphere of influence '. Fortunately an archaeological discovery made on this ground permits us to clinch the argument, while at the same time illustrating once again the accuracy of Chinese historical records.

I refer to the discovery of a secondary line of wall which, on my first reconnaissance from the site, I found to extend from the north towards the ruined fort of T. my and, passing it quite close on the west, to continue to the south-south-east approximately in the direction of Nan-hu. Closer examination of the portion lying north of T. my showed that this wall was only about 5 feet wide at its foot, but otherwise in construction closely resembled the Limes wall, showing the identical alternation of layers of fascines and stamped clay. Badly decayed as it was in most places, the marked difference in thickness could yet be observed with certainty throughout its traced extent, and this su ffices to distinguish it clearly from the Limes wall proper. The line of this secondary wall starts from the south-western edge of a deep marshy depression which lies to the south of the Limes stations T. xv and T. xv. a, and contains several spring-fed lakelets. For about half a mile it runs due south across the bare gravel-covered top of a tongue-like plateau, and though its remains are nowhere more than 4 or 5 feet in height, the reed fascines used in it are clearly traceable

Other Limes sections mentioned.

Advantages of ' Jade Gate' placed at T. xiv.

Secondary wall line passing T. xlv.

12 Cf. Chavannes, Documents, p. 72.   13 Cf. above, pp. 636, 648.   13a Cf. below, pp. 698 sq., 717.

4T2