870 IN THE REGION OF THE UPPER OXUS [Chap. XXVI
have changed materially within historical times, nor can the volume of water available for irrigation, given the vicinity of two high snowy ranges. But flocks and herds play a great role in the economic conditions of Wakhan, and during periods when a strong rule assured safety from raiding neighbours on the side of Shughnan and the Pamirs, these must have represented a very considerable addition to local resources. Nor ought those sources of profit to be ignored which might have accrued from a brisk trade between the Tarim basin and Badakhshan by this route during periods of assured security. The vicinity of Khandüt, the old capital of Wakhàn and probably its naturally most favoured site, suggests that the spur above Yamchin, very strong by nature, was fortified, perhaps, for the purpose of serving as a safe refuge for the rulers. They alone could have carried out so extensive a scheme of defences and provided for adequately manning them.
In character the defences correspond in a very striking fashion to the mountain fastness of Adh-i-samûdh in the vicinity of Kohat, which I surveyed in 1904, and to that of Kiz-kurghan in Sariko1.18 In each case a position naturally very strong was fortified to serve as a temporary refuge in case of serious danger, not for the purpose of permanent occupation. This view is supported by the great scarcity of potsherds at the site above Yamchin—I found only a few behind the main circumvallation and within the citadel—and also by the absence of any remains of habitations except within the latter 19 The parallel offered by the remains of Kiz-kurghan is particularly instructive also in another respect. I have been able to prove in Serindia that these are identical with the mountain stronghold which is mentioned by Hsüan-tsang as the site of a legendary event ascribed by local tradition to Han times and which had become ruined long before his own passage in A. D. 642.20 If so much of the walls of Kiz-kurghan, built with sun-dried bricks and rough stonework above slopes if anything even more precipitous than at Zamr-i-atish-parast, could survive to the present day, it does not seem impossible that the fortifications of the latter site—no doubt, on the whole somewhat better preserved—were already in being when Hsüan-tsang passed through Wakhan, or were erected not very much later. For it should be remembered that the climate of Wakhan is probably quite as dry as that of Sarikol, and the snowfall on Kiz-kurghan, at a height of about 13,000 feet, if anything heavier than at the Wakhan site.
On September 5th our march down the valley past the pretty hamlets collectively known as Putup brought us, after we had proceeded some 7 miles, to a portion of the ` thalweg ' which drift-sand, carried up from the wide river-bed by the prevailing western winds, has converted into a sandy steppe, with tamarisk-cones and desert scrub curiously reminiscent of the Tarim basin. At the small village of Shitkharw, reached after a march of another 7 miles, on a fertile alluvial fan, I was joined by Qazi Qadam Shah, whose intelligent help enabled me during the next few days to secure specimens of Ishkashmi, a Galcha language not previously recorded.2' Farther down, where an extremely steep cliff (Fig. 448) rising above the river used, before the making of the Russian bridle-path, to be passable only by sure-footed men with the use of ` toe-holes ', he showed me a curious recess in the rock known as Liw-bar (Persian diw-dara). Here a demon, who was wont to kill people passing, is believed to have retired into the mountain on being vanquished by a saint.
Before reaching the village of Darshai, where we halted, a remarkably narrow canon had to be
tract, was believed to be somewhat greater, notwithstanding the drain due to emigration caused by the prevailing exactions, &c.
18 See Stein, Archaeological Survey Work in NW. Frontier Province, 1905, pp. 2 sqq. ; Serindia, i. pp. 73 sqq.
19 I may note here that I could find no traces of terraced fields on the slope within the protected area, as mentioned
in Olufsen, Unknown Pamirs, p. 187, nor of irrigation channels. To bring water to the slope below the citadel would scarcely have been possible without extensive blasting.
2° Cf. Serindia, i. p. 75.
21 These materials have been published by Sir George Grierson in Ishkdshmi, Zêbaki and Yazglzuldnzi, R.A.S. Prize Publication Fund, 1920.