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

554   THROUGH THE LOP DESERT TO TUN-HUANG   [Chap. XIV

It is on this northern branch of the Lop—Tun-huang route, which for brevity's sake we may call the ` Lou-tan route ', that I was able correctly to locate in 1914 the dreaded ` White Dragon Mounds ', which the Former Han Annals describe as opposite to ` the extreme eastern border of the kingdom of Lou-lan where it approached nearest to China ',4 and to which references are made also in later records.5 There were also important antiquarian ` finds', as recorded in the preliminary account of my explorations of 1914, which helped me to follow the Lou-lan route over its desolate westernmost portion on either side of the ` White Dragon Mounds '. It is traceable even more accurately over the eastern portion. Geographical and archaeological evidence makes it quite certain that it must have followed the line marked by the ancient Chinese Limes along what I have above described as the third section of the present desert route from Lop to Tun-huang. It is there that I discovered the site of the famous Vii-min or ' Jade Gate ', from which the route to Lou-lan is declared by the Annals to start.°

As regards the remaining intermediate, or central, portion of the ancient Lou-lan route we also possess safe guidance. My explorations of 1914 have proved that the easternmost inlet of the ancient lake-bed, with its difficult and in places treacherous surface of dried-up salt bog, extends as far as the neighbourhood of Yantak-kuduk (Map No. 70. B. 3). They have further shown that, east of this point, water and camel-grazing are also obtainable along the northern edge of what for brevity's sake I may call the Bash-toghrak valley. This northern edge is certainly the shortest line towards Lou-lan, and I therefore conclude that the ancient Han route to Lou-lan followed it, branching off from the present desert route somewhere about Bash-toghrak. There are some archaeological indications supporting this conclusion, but their record must be reserved for a future publication dealing with the results of my third expedition.' From Bash-toghrak eastwards we may safely assume that the Han route skirted the southern edge of, or passed through, the ancient terminal basin I have described above, more or less along the present line of route ; for, as Map No. 74. B. 3 shows, the antiquity of this is distinctly attested by the bearing of the line on which the advanced towers of the Limes, T. I, II, are erected, since it approaches the south-eastern edge of the basin within about eight miles and forms a direct continuation of the former line.

The account given elsewhere of the trying journey by which, in the winter of 1914, I succeeded in tracing the western portion of the ancient Lou-lan route through that absolutely barren waste of salt, clay, and gravel, will suffice to show how formidable the difficulties encountered here must have been.8 Over a distance which in ancient times, too, must have been fully 120 miles if not more, there was neither water nor grazing nor fuel. How those old Chinese organizers of transport under the Emperor Wu-ti succeeded in overcoming these obstacles and opening up the route as a great line of traffic for trade and military expeditions is a problem which I cannot discuss here. But one important observation, and that of a topographical nature, may be recorded. It is, beyond question, possible to make the whole of the Lou-lan route practicable for cart traffic, a mode of transport for which the Chinese still retain a special preference and aptitude in Central-A sian regions and those adjoining them. It is referred to in some of the Chinese documents found at the Lou-lan Site,° and practical experience elsewhere makes me inclined to think that in its use may be found the chief clue for the solution of the problem.

Lou-lan route' traced.

Central portion of Lou-lan route.

Difficulties of western portion.

4 Cf. Wylie, Journal of Anlhrop. Inst., x. p. 26, and for the position of the ' White Dragon Mounds', as ascertained by me in 1914, Geogr.Journal, xlviii. p. 128; above, pp. 340 sq. The suggestion made in Desert Cathay, i. p. 525, must be corrected accordingly.

e See above, pp. 418 sq. ; below, pp. 555 sq.

6 See below, p. 555 ; chap. xix. sec. i, ii.

Cf. Geogr.,Journal, xlviii. p. 129, for a reference to the traces of an ancient canal which I found along the northern edge of the Bésh-toghrak valley and to the west of Koshkuduk (Map No. 70. B. 3).

° See Geogr. Journal, xlviii. pp. x 27 sqq.

See above, p. 412 ; Hedin, Central Asia and Tibet, ii. pp. 143 sq.