i8 ACROSS SWAT AND DIR [Chap. I
Hu-mi or Wakhân to the kingdom of Chü-wei in which, as has been shown elsewhere, we must recognize Mastûj.24 The route so far indicated, devious as it may seem, corresponds exactly to the one taken by the remarkable expedition which the Chinese general Kao Hsien-chih in A. D. 747 led from Kâshgar to the successful invasion of Yasin, and which there will be occasion later to consider in detail.25 Hence it is a priori probable that Wu-k`ung's party, for its progress to Mastûj, used the Barôghil saddle which offers far the easiest passage from Wakhân across the Hindukush, and which only a few years earlier had seen the triumphant advance of Kao Hsien-chih's force.
From the Barôghil, at the head of the Yârkhun River, the routes of Kao Hsien-chih and Wu-k`ung diverged. Whereas the former led his troops straight across the high Darkôt Pass into the Yasin Valley, Wu-k`ung's party evidently descended the Yârkhun for some distance before gaining Yasin territory, and thence Udyana ; thus only is it possible to account for the itinerary which Wu-k`ung's biography indicates. After Chü-wei there follow in it ` the kingdom of Ho-lan, then the kingdom of Lan-so, then the kingdom of Yeh-ho, then the kingdom of Wu-chang-na (also
called Wu-ch`ang or Wu-ch`an)'. Now as regards Yeh-ho M. M. Chavannes was undoubtedly
right in recognizing in it a variant of the name Yeh-to , by which the notice of the Tang
Annals designates the capital of Little P`o-lü or Yasin.26
To understand the interposition of the two territories which precede Yeh-ho in Wu-k`ung's list is not difficult, if the actual topography of this mountain region is consulted. After a descent of the Yârkhun river to Chü-wei, i. e. to the cultivated part of the valley extending above Mastûj proper, the easiest route towards Yasin lies from Mastûj up the Lâspur Valley, and thence across the low Shandur Pass eastwards into the valley of Ghizar, which is joined by that of Yasin from the north at Gitpis.27 It is this route, still the main line of communication between the Yârkhun Valley and the drainage area of the Gilgit-Yasin River, that Wu-k`ung and his companions may safely be assumed to have followed. After leaving the chief place of Chii-wei which the notice of the Tang Annals calls ` the town of A-shê-yü-shih-to', and which I have been able to identify with the present village group of Shuyist, Wu-k`ung must have passed down to Mastûj, to-day the administrative centre of the upper Yârkhun Valley, and thence through Lâspur. It is evidently this narrow but fertile valley which Wu-k`ung means by the territory of Lan-so a.y `, for the characters are such as would ordinarily be used for the abbreviated transcription of a local name like *Lasapura.28
I am unable at present to suggest an equally convincing identification for Ho-lan ; but I have
no doubt that it must represent an older name of either Mastûj itself, or of one of the more important village tracts higher up the Yârkhun like those of Brep or Miragrâm.
It is, of course, impossible to determine with absolute certainty the reason which induced Wu-k`ung's party to choose the route indicated instead of the far more direct one across the Darkot Pass which, as already seen, Fa-hsien must have followed on his descent from the Pamirs to Dare' and Udyâna.22 The same applies to the question, why, having followed the Yârkhun River, they did
44 See Ancient Khotan, i. p. 15, note 31, and below, pp. 42 sq.
26 See Ancient Khotan, i. pp. 8 sqq. ; below, pp. 52 sqq.
26 Cf. Chavannes, Turcs occfd., pp. 129, note 2 ; 150 (where Sie-to is a mistake for Ye-to) ; Ancient Khotan, p. 16, note 31.
27 The only alternative route leads across the high Tui Pass, 14,700, with a difficult piece of glacier to be crossed, and is practicable only for some months in the summer ; cf. Biddulph, Hindoo Koosh, p. 56.
28 If we substitute with M. Chavannes the character po
011 by correction for the so g. of Wu-k`ung's biography it
is equally easy to recognize a form of the name of Lâspur in the transcript. The suggested identification, however, of this *Lan po with Lampâka or Lamghan is one which M. Chavannes no longer now maintains, since adopting the identification of Chü-wei with Mastûj he rightly makes Wu-k'ung proceed down the Yârkhun Valley and across the Shandur Pass to Yasin ; see Chavannes, Notes Addil., p. 43, note 4.
29 See above, p. 7.