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

Sec. iv]   MILITARY AGRICULTURAL COLONIES   741

mention of the command of Tun-huang ' proves that the edict must be later than I I I B.c., while the reference made to the ` command of Chiu-ch`üan ', i.e. Su-chou, suggests that it dates from

a time when the latter still retained its original character as the true advanced base for Chinese expansion into Central Asia.1 Considering that the documents found at T. vi. b go back as far as 68 B. c.,2 there is no chronological difficulty about assuming that the imperial edict quoted in our document belongs to the years after IO2-IOI B.C., when, as we have seen above, the westernmost extension of the Limes beyond Tun-huang took place.3 I have shown also that this extension is likely to have been planned from the first to reach the natural line of defence on the south-west, including the station of T. vi. b where the record was found.

The question how a document that reproduces the imperial edict connected with that extension actually found its way to this station obviously need not trouble us. The big refuse-heap of

T. vI. b. i had undoubtedly received the contents of an old office archive. The only problem is how to explain that part of the edict which relates to the foundation of a military agricultural colony, or rather where to locate it. That the establishment of such colonies in the protected belt along

the great highway westwards, was an integral part of the Emperor Wu-ti's ` forward policy ', from

the time when it first pushed the continuation of the Great Wall ' beyond the Yellow River, is clearly proved by the Chinese historical notices discussed at the beginning of this chapter .4 Without Chinese colonies organized on a military basis it would have been impossible to assure either the maintenance of the troops needed for guarding the Limes or the provisions needed for the trade

caravans, expeditions, and political missions, the secure movement of which it was the main object

of that policy to render possible ; for any scanty settled population that the cultivable tracts along the foot of the Nan-shan may have contained while held by the Hsiung-nu and their equally

nomadic predecessors, the Yiieh-chih, was removed on the Chinese conquest, as the Former Han Annals distinctly record .3 Yet I feel equally certain on geographical grounds and from the actual survey of the ground that such an agricultural colony could not have been placed anywhere along the actual line of the wall beyond the Khara-nôr, and still less by the edge of the terminal marsh basin of the Su-lo Ho where the watch-station T. vI. b stood.

The full account I have given above of the physical features of the desert area traversed by the Limes west of Tun-huang 6 makes it unnecessary to support this statement by detailed explanations

of a topographical nature. Cultivation was never possible within historical times either in the marshy

depressions which skirt the terminal course of the Su-lo Ho or on the gravel plateaus which overlook and divide them. The plateaus could never have been irrigated from the Su-lo Ho owing to their

elevation and the general configuration of the ground. Nor could there have been, during historical times, a sufficient permanent supply of water at the foot of the hills west of Nan-hu to cross the huge gravel glacis, partly covered with high dunes, and assure irrigation south or south-west of the Limes.

That agriculture without irrigation would have been as impossible here in ancient times as it is nowadays scarcely needs emphasizing. Nothing could demonstrate better the exceptionally arid

climate of this region since early times than the bare fact that wooden documents thrown out

on a refuse-heap in the first century B. c. had survived, often in an almost perfect state, until I came to recover them from under a few inches of gravel or rubbish. These observations are supported by

the fact that nowhere along the Limes west of Tun-huang were even the slightest indications of permanent ancient occupation traceable, apart from the remains of the watch-towers and stations described.

' Cf. above, p. 724 ; Chavannes, Documents, p. v, note 5.   ' Cf. above, pp. 724 sqq.

' Cf. above, p. 647 ; Doc. No. 255.   ° Cf. Wylie, J. Anthrop. Inst., X. p. 22 ; above, p. 728.

3 See above, p. 728.   6 Cf. above, chap. xvii—xix, passim.

Location of agricultural colony.

Cultivation impossible along Limes west of Tun-huang.