418 THE LOU-LAN SITE [Chap. XI
on the ancient route through Lou-Ian. ' As regards the roads which starting from Tun-huang and Yü-mên kuan (' Jade Gate barrier ') pass into the Western Countries, there were two before, but now there are three. The southern route is the one which starting from Yü-mên kuan leaves from the west, passes the Jo Chiang, turns westwards, crosses the Ts`ung-ling, traverses " the suspended passages ", and passes among the Ta Yüeh-chih.
` The central route is the one which, starting from Yü-mên kuan, sets out on the west, leaves
the well of the Protector-General tis jt,.a turns back at the northern extremity of the San-lung
[desert of] sand . i, passes the Chü-lu granary E. a ; then on leaving from the Sha-hsi
well fLti 4, turns to the north-west, passes through the Lung-tui 4., arrives at the ancient
Lou-lan # ` and turning westwards arrives at Ch`iu-tzü (Kucha), then attains the
Ts`ung-ling , , p`.
` The new route [of the north] is the one which, starting from Yü-mên kuan, sets out on the north-west, passes through Hêng-k`êng ;j a', avoids the San-lung [desert of] sand as well as the
Lung-tui, leads north of Wu-chuan W. , a and arrives, in the territory of Chü-shin m, at
Kao-ch`ang (Kara-khaja) ; then it turns westwards and rejoins the central route at
My observations here as regards the southern and northern of the Wei do's routes may be brief, as also those concerning the Yü-mên kuan, the ` Jade Gate barrier ', from which they, as well as the central route, are described as starting. My subsequent explorations have proved that this famous frontier station, so frequently mentioned in the Chinese Annals in connexion with events affecting the Western Countries, was situated during Han times at a point of the ancient Chinese Limes in the desert west of Tun-huang marked by the ruined watch-station T. xiv and identified by the evidence of the records I discovered there.'° That ` the southern route ' of the Wei do is identical with the one which still leads from Tun-huang along the northernmost main range of the K`un-lun, here known as Altin-tagh, to Charkhlik and thence through the string of oases in the south of the Tarim Basin is made certain by the mention of the 7o Cleiang, a nomadic tribe whose' position in the mountains between Tun-huang and Chü-mo or Charchan is quite correctly described by the Former Han. Annals." We have already seen that the list which a subsequent passage of the Wei lio gives Of the territories passed through by ' the southern route ' begins with Shan-shan and the small kingdoms' of Chü-mo, Hsiao-yüan, Ching-chueh (Niya), and Lou-Ian, described as its dependencies."L It is not necessary for us to examine this list again, mainly based as it is on that which the Former Han Annals give for their ' Southern Road', nor to trace the distant
9a For ` Puits du Protecteur ' read ' Puits du Protecteur général', a correction pointed out to me verbally by M. Chavannes; cf. Chavannes, T'oung pao, 1907, pp. 153, 154, note r.
10 See below, chap. xix. sec. i, ii.
" See M. Chavannes' translation of this passage, with full notes on other references to the Jo Chiang, in T'oungpao, 1905, p. 526, note 8; cf. also for the context of the Ch'ien Han shu passage, Wylie, J. Anlhrop. Inst., x. p. 23.
It deserves to be noted here that the Ch'ien Han shu is more accurate in describing the exact starting-point of ' the southern route ' when it states : ` On leaving the Yang barrier the first people met with on advancing are the Jo Chiang.' We shall see below (chap. xvr. sec. iv) that the ' Yang barrier' was situated at the present Nan-hu, south-west of Tun-huang,
and was connected with Yü-mên kuan by a southern flanking line of the Limes (chap. mix. sec. iii).
'2 See above, p. 328.. The obvious inversion of the order of these petty states in the Wei llo s list has been clearly pointed out, together with some minor inaccuracies and discrepancies, by Chavannes in T'oungpao, 1905, pp. 535-6, note 3.
It is difficult to decide whether the separate mention of Lou-Ian among the tracts dependent on Shan-shan is due merely to a kind of diplography, the earlier name Lou-lan for Shan-shan, as recorded by the Ch'ien Han shu, being meant, or whether we ought to recognize in it a reference to the northern portion of Shan-shan territory which had retained the old designation of Lou-lan, as attested by the documents of the Lou-lan Site, and which may, perhaps, also have enjoyed a kind of local autonomy.