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

294   THE SEARCH FOR THE ANCIENT CHINESE HIGH ROAD [Chap. VIII

mark the position of those terminal marshes, which, at the time when the Lou-lan area was under occupation, formed a fringe along the western shores of the ancient sea-bed. The same process of salt-impregnation can still be seen to-day in progress in the marginal marsh-beds of the Karakoshun which the traveller passes on the first two marches by the caravan track from Abdal towards Tun-huang. Owing to their position remote from the termination of the Tarim these beds have for long periods past received a supply of water increasingly saline, and that only at rare intervals ; ultimately, owing to shrinkage or diversion of the inflow from the Tarim, they have been left for the most part quite dry.9

Now, of those Yu marshes once fed by the Kuruk-darya Li Tao-yiian's above-quoted passage tells us that their water ` accumulates in the north-east of Shan-shan and in the south-west of the Town of the Dragon '. If we compare the position of Miran and Charkhlik—the only cultivable areas of any importance in ancient Shan-shan,10 as seen in Map No. 3o—with that indicated by the foregoing observations for the terminal marshes of the several Kuruk-darya beds (Maps No. 29. D. 4 ; 32. A. 3, 4), it is clear at the first glance that the north-eastern bearing here stated for the Yu marshes in relation to Shan-shan is perfectly accurate. Since they are said also to lie south-west of the Town of the Dragon it is obvious that the latter has to be looked for in the continuation of the same bearing from Shan-shan.

On the ground to which this direction takes us it may be asserted with absolute confidence that no real town could ever have existed either in historical times or before then. We have seen that the area to the north-east of the Lou-lan Site, L.A., during the period when the route through it was frequented, could have afforded only the scantiest subsistence to a scattered population of indigenous herdsmen and hunters. The last dead tree marking the former existence of a riverine forest belt was left behind near the burial-ground of L.C., and beyond a point close to L.I. the traces of ancient vegetation completely disappeared. The same observation, excluding the possibility of any larger settlement, applies equally to the utter wastes of bare clay, salt and gravel through which lay the routes of the surveyors farther north, as well as to the adjacent barren area of the Kuruk-tagh. We are thus necessarily led to assume that the site described in Li Tao-yüan's account as once occupied by the Town of the Dragon was in reality not a ruined site, but a locality where popular imagination, stimulated by natural features, placed the remains of a town.

The vague reference of the text to ` a great kingdom of Hu' or barbarians, the submergence of its capital by ` an overflow of the Fu-ch`ang lake ', the huge extent ascribed to the ` foundations' still preserved of this town,—all these suggest that we are dealing here with a creation of folk-lore. But the essential fact confirming this impression is that on the very ground to which we are taken by Li Tao-yüan's clear and reliable topographical indications, there is found a striking natural formation accounting for the rise of that folk-lore tale.

It is that great belt of high Mesas through which we first passed to the north of the castrum L.E. and along which our march of February 27th on the track of the ancient Chinese route had led us. With the wall-like steepness of their slopes and the fantastically eroded forms of their tops they must have suggested, to the imagination of ancient wayfarers, the walls, towers, and mansions of some vast ruined city. They called up the same ideas to us when we passed, the first travellers for many centuries, through this silent, utterly desolate scenery. That such was the impression on my Turki followers was shown by the frequency with which my attention was called by them to dome-shaped Mesa tops which they took for ` P`ao-t`ais ', i. e. Stûpas, and

9 I had a good opportunity of observing this process on my marches of February 21-2, 1907, eastwards from Abdal along the southern shores of the Kara-koshun marshes ; cf.

Desert Cathay, i. pp. 504 sq.

10 See Serindia, i. pp. 311 sqq., 335 ; above, pp. 170 sq.

Position of Yu marshes.

Town of Dragon ' not a ruined site.

Folk-lore origin of story.

Mesas suggest town walls, mansions, &c.