310 THE SEARCH FOR THE ANCIENT CHINESE HIGH ROAD [Chap. VIII
That the Chinese designation of ` White Dragon Mounds ' was directly derived from those salt-coated Yardang terraces through which the Lou-lan route had to pass on either side of the level plain of hard salt crust is conclusively proved by a passage which M. Chavannes in his notes on the Wei lio's account of the ` route of the centre ' has quoted from Mêng K`ang's X gt commentary on the Chien Han shu composed in the third century A. D. : 20 ` The dragon-shaped mounds have the appearance of the body of a dragon in earth which was without a head but had a tail. The highest rise to two or three chang (twenty or. thirty feet) ; the lowest to over one chang (ten feet). All of them are turned towards the north-east and resemble each other.' We have here as accurate and graphic a description of these strange white ridges as an intelligent Chinese traveller might give to-day. After what I have stated above as to their fantastically twisted shapes, and yet uniform direction, we can well understand that Chinese eyes, ever keenly observant of peculiar topographical features, should find in them a resemblance to white dragons. The commentator's statements as to the north-easterly bearing and the average height of the ` mounds ' are perfectly correct and obviously derived from an authority conversant with the ground.
The passage of the Former Han Annals is moreover of special interest because it reflects in a striking fashion the exceptional difficulties which must have attended the movement of large Chinese missions, convoys and bodies of troops, by a route leading across so great a stretch of utterly barren ground and presenting formidable natural obstacles. How traffic of considerable magnitude was maintained for centuries over such a route in the face of all its hardships is a problem to which I shall have occasion to return farther on.20a But here already we may note that another significant reference to the ` White Dragon Mounds ', to be found in the Chien Han shu, is directly due to the seriousness with which those difficulties made themselves felt long after the first opening of the Lou-lan route.
In a subsequent section of the ` Notes on the Western Regions ' the Former Han Annals tell us of a new route made during the Yüan-shih period (n. D. r-5) from the side of Posterior
Chii-shih, , offi, i, e. from the territory of present Guchen (Ku-ch`êng-tzû), north of Turfân,
` which, passing north of Wu-ch`uan W. 44, penetrated as far as the Yü-mên barrier ;
the journey was thereby reduced. The Wu-chi-hsiao-wei Hsü P`u-yü had opened [this route] in order to shorten the length of the road by one-half and to avoid the dangers of the Po-lung-tui
} fk (White Dragon Mounds).'21 We find an exactly corresponding reference to the avoidance of the Lung-tui as well as of the San-lung sha, the ` Three Ridges Sands ', in the itinerary which the Wei lio gives of this ` new route of the north ', established as an alternative to the ` route of the centre ' or the Lou-lan route.22
The topography of this ` new route ' to the Western regions which led from the Jade Gate north-westwards across the desert ranges of the Western Pei-shan has been fully discussed in Serindia.23 I have also called attention there to the very serious natural difficulties which, owing to the absence, or at least extreme rarity, of water and grazing, must have likewise beset this route even in ancient times. These have made it, probably owing to the increase of aridity, altogether impossible for regular traffic at the present day. That such a route had to be opened up and followed in preference to that through Lou-lan, mainly in order to avoid the obstacles presented by the ` White Dragon Mounds ', is perhaps the best illustration of the dread in which the ancient
20 See Chavannes, Toung-pao, 1905, p. 529, note 7. 20a Cf. below, pp. 337 sqq.
21 See M. Chavannes' rendering, ibid., p. 533, note I ; also Wylie, J. Anthrop. Inst., xi. p. 209.
22 See Chavannes, T`oung-pao, 19o5, p. 533. Incidentally