Sec. iii] HISTORICAL ACCOUNTS OF MASTÛJ 45
time being one chief ruler as they have done more than once at intervals during the modern period of their history.21 This might easily have induced the pilgrim, who never saw the territory himself, either to use the name Shang-mi in an extended sense or else to introduce into his notice the mention of a product belonging to the neighbouring chiefship. In all other respects his account appears appropriate to Mastûj or Kashkar-Bala, and his distinct mention of the predominance of Buddhism supplies a point of interest.
A far more ancient mention of this district would exist if only Shang-mi could be definitely
identified with the Shuang-mi jip, which the Annals of the Earlier and Han dynasties name as
one of the five territories held by Yüeh-chih chiefs (hsi-hou) dependent upon Tokharistan.22 But the Han Annals do not furnish any definite indication as to the position of Shuang-mi, and the Pei shih which supplements these notices by information dating from the beginning of the seventh century A. D. contents itself with giving Che^-hsieh-mo-sun ffrgfas the later equivalent of the name Shuang-mi and placing the territory to the west of Ch`ieh-pei which corresponds to Hsiu-mi or Wakhan.23 This location, however, can certainly not be reconciled with the position of Mastûj, nor yet with that of Chitral, to the south of Wakhan. As, moreover, we have no means of identifying the hsi-hou territory of Kuei-shuang, with its capital Hu-tsao, which the Han Annals mention after Hsiu-mi and Shuang-mi and which the Pei shih places to the west of the latter,24 the question as to the position of Shuang-mi must, I think, at present be left open. However, that the power of the Yüeh-chih, established for centuries in Badakhshan, should have made itself felt also in Mastùj and Chitral is • in itself highly probable, especially in view of the historical analogies which reach down to quite recent times.25
SECTION IV.—OLD SITES IN MASTÜJ
Already the march of the first day, May i I, which I passed on the soil of Mastûj, showed that the Upper Yarkhûn Valley, notwithstanding its confined area and limited resources, was not without remains of the past. In the pretty villages of Awi, Mem, Miragram, through which the road led me about half-way, I had to be content, it is true, with such proofs of an old-established civilization as were afforded by luxuriant orchards and noble Chinar groves,—sights doubly impressive in view of the background of mighty glaciers which overhang the side-valleys and feed their life-giving rivulets. But when some four miles further up I reached the large village of Sanôghar, ensconced among magnificent Chinars and fruit-trees, I found my attention claimed not merely by the training-camp of that fine corps of cragsmen, the Chitral Scouts, but also by a veritable `old site'.
The narrow rugged ridge which rises in complete isolation between the village and the steep-banked river to a height of about 200 feet above the nearest fields, and forms a position of great natural strength, was pointed out to me as the site of Sanôgharo-noghor, the ancient fort of Sanôghar. There was nothing to be seen on its level summit, some 35o feet by 5o feet or so across, save old tombs and much rubble which probably marked walls completely decayed. But the slopes all round were thickly covered with red pottery débris of remarkable hardness, far superior to any earthenware
21 See above, pp. 33, 42.
R2 This identification was first proposed by Cunningham, J.A.S.I3., xiv. 433. Yule, J.R.A.S., N.S., vi. p. 114, is content to quote it without expressing an opinion. It is accepted by Professor Marquart, Erân.fahr, p. 243.
23 For a lucid and complete synopsis of the notices concerning the hsi-hou territories attached to the Yüeh-chih dominion in Tokhâristân which the Ch`ien Han shu and Hou
Han shu furnish, see Chavannes, Pays d'Occident, p. 43 (T'oung pao, 1907, p. 189), note 3.
As regards the identity of Ch`iehpei-Hsiu-mi with Wakhân, see below, p. 6o.
2' For the vagueness of the conjectural locations so far attempted, see Marquart, Érânfahr, pp. 245 sq.
25 Cf. above, pp. 28 sqq., and p. 33.