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

Notice of ancient

Yi1-min in Tun-huang

lu.

Reference to `Great Wall' in Tunhuang lu.

Description of Limes wall in

Sha chou chip.

Wall measures in SJra chou chah.

734   HISTORY AND RECORDS OF THE TUN-HUANG LIMES [Chap. XX

less clear towards the end of the Tang period or soon after is shown by the short text on the Mirabilia of Tun-huang, the Tun-huang lu, which Dr. Giles has edited and translated from one of our Chien-fo-tung manuscripts. In this text, which probably belongs to the tenth century and cannot be older than the ninth, we read : ` West of the city [of Sha-chou or Tun-huang] is the Yang Barrier, which is the same as the ancient Yti-mên (Jade Gate) Barrier.... It connects China with the capital of Shan-shan, but the natural obstacles of the route and its deficiency in water and vegetation make it difficult to traverse. The frontier-gate was afterwards shifted to the east of Sha-chou.' 25 Evidently local popular tradition still vaguely remembered that the Jade Gate was once situated west of Tun-huang, though it erroneously identified it with the Yang Barrier. The notice, at any rate, is of some interest as definitely mentioning the shift of the ` frontier-gate ' to the east which Hstian-tsang's Life presupposes to have already taken place, and also as correctly describing the mountain route towards Charkhlik which passed through the Yang Barrier.25a

Of much greater antiquarian interest for us is another passage which closes the Tun-huang lu, and which a notice of the Sha chou chili fortunately amplifies and corrects. ` The Great Wall, built under the Former Han dynasty, passes 63 li to the north of the city and runs due west out into the desert.' 26 We see here clearly that the remains of the Limes wall and its origin were still known to the people of Tun-huang about the tenth century n. D. The nearest point of the wall where it passed north of the Sha-chou town of Tang times may be placed, according to my surveys of 1914, at a distance of about i6 miles, which agrees very closely with the 63 li of the text.

The fragment of the Sha chou chip which Mr. Lo Chên-yü has published from a Chien-fo-tung manuscript in M. Pelliot's collection, and of which Dr. Giles has translated an extract in his note, adds to the above several very interesting details about the remains of the Limes which were known when this text was composed, apparently towards the close of the Tang period or not very long after. They deserve to be quoted in full here : ` The ancient wall is 8 feet high, io feet wide at the base, 4 feet wide at the top. It passes 63 li north of Tun-huang and extends eastwards for

18o li to the Chieh-t`ing Signal-station RI   4,4, where it enters the territory of Ch`ang-lo hsien in
Kua-chou ; towards the west it reaches as far as the Chü-tsê (Winding Lake) Signal-station

, a distance of 212 li, running out into the desert due west in the direction of the territory

of Shih-ch`êng   ttA (Charkhlik).'

Both the measurements and the distances given appear to me to be based upon carefully collected local information. Taking the ruined wall first, we find a remarkably close agreement between the width indicated for it at the base and the actual measurements which I secured from different sections of the Limes. If we assume that the record of the Sha chou chi h's informant was taken with a foot measure such as I excavated at T. dnt and T. xi, the io Chinese inches of which were equal to 9 British inches,27 we get as the result of the equation io : 9::I00 : x, a width in British measure of 90 inches, or 7 feet 6 inches, for the base of the wall. This shows a remarkably close agreement with the average of base measurements which I obtained at numerous points of the wall,

,1,

1,

26 Cf. Giles, J.R.A.S., 1914, pp. 715 sqq. ; also his re-translation, J.R.A.S., 19x5, p. 45. Our detailed examination above of the archaeological and topographical facts bearing on the true location of the Yü-mên and Yang barriers in Han times makes it unnecessary to discuss here the conclusion which Dr. Giles thought it possible to base upon the Tun-huang lu passage reproduced above ; cf. above, pp. 623 sq.

I doubt whether, in view of the archaeological evidence

now available, the late and vague statement in that passage can be used in the way suggested to explain how Li Kuang-li in 103 2.c. reached Tun-huang, while the Jade Gate stood at Nan-hu, etc.

26a See above, pp. 622 sq.

26 See Giles, J.R.A.S., 1915, p. 47, for the rectified translation.

27 Cf. above, pp. 66o, 668.