864 IN THE REGION OF THE UPPER OXUS [Chap. XXVI
vaulting. The enclosing walls are of very solid construction and even near their top fully 6 feet thick. At some points the circumvallation has been strengthened by oblong bastions. Apart from small rooms built in places against the enclosing wall, the interior contains a number of detached structures, all badly decayed. Most of these are of oblong shape and divided into rooms of small size ; but a few at the north-eastern end are irregular ovals. Fig. 388 shows the rough masonry of these structures, which, no doubt, served as habitations when regard for safety compelled occupation of the fort. That the period of occupation was not continuous may be concluded from indications of repeated repairs. Local tradition ascribes the fort of Hissâr, like the other defences to be noticed farther on, to ` Kafir' occupants of Wakhân, i. e. to pre-Muhammadan times.
In spite of the hardness of the mud plaster in which the rough slabs of stone are set, it would be difficult at first sight to credit this statement of great age, were it not for the extreme aridity of the Wakhàn climate, which equals that of Sarikol.4 It must also be borne in mind that the construction of the walls here, as at the other old Wakhân strongholds examined, rough as it looks, is yet distinctly superior to that found in the houses or rather hovels occupied by the present population. Here, as at the other sites, I was strongly reminded of structural features made familiar to me by the many ruined dwellings scattered on the hill-sides in Swat and in neighbouring parts of ancient Gandhâra. These undoubtedly belong to Buddhist times, and for the most part show masonry quite as rough. Yet climatic influences on the Indian North-west frontier are undoubtedly far more destructive than those to which the ruins of Wakhân are exposed. Unfortunately at none of these sites was it possible to secure coins or other datable relics.' What scanty pottery debris could be picked up at Hissâr did not include any decorated pieces. But some of the plain potsherds showed a fine brownish surface and well-levigated clay, such as are unknown to modern local manufacture. To the question who are meant by the ` Kâfirs' to whom local belief invariably attributes these ruined fastnesses, I shall have to return farther on.6
About a mile to the west of Hissâr lies Zang (about 9,70o feet above sea-level), which comprises some forty households and is the largest village on the Russian side of Wakhân. Its terraced fields and tree-hidden homesteads stretch along a broad glen well watered by springs. To the west of Zang a steep spur rises to close on a thousand feet above the village, and at its southern extremity bears the ruins of a small fort forming an irregular oblong of about 6o yards by 25 and known as Zangibar (Fig. 395). The interior, as the sketch-plan in Pl. 46 shows, is filled with the remains of dwellings built, like the enclosing walls, of unhewn stone slabs. Up to a height of about 6 or 7 feet these are set in fairly uniform courses with hard mud plaster, while above them much rougher stonework indicates later repairs. A small square bastion projecting on the northern face and provided with loopholes appears to have been originally an isolated tower to which the rest was subsequently added. The villagers attributed the construction of the fort to ` Kâfirs ', but acknowledged that during their fathers' times it was occasionally sought as a place of refuge when raids of Kirghiz or Shughnanis threatened. The absence of pottery debris suggested temporary occupation for short periods only.
The scarcity of land capable of irrigation under existing conditions, together with the oppression prevailing on the Afghan side of the valley, accounts for the numerous Wakhi emigrants to be found nowadays to the south of the Hindukush.7 Most, if not all, Wakhis belong to the Ismailia
4 See my remarks on the old walls of sun-dried bricks at the forts of Kansir and Kiz-kurghân, Serindia, i. pp. 69, 75.
5 It ought to be borne in mind that the factor which in India, as probably elsewhere also, is responsible for most of the finds of coins and similar relics at old sites, viz. occa
sional heavy rainfall, is practically unknown in Wakhân and in Sarikol also.
6 See below, ii. p. 869.
7 I had found Wakhi colonies in Guhyal (cf. Ruins of Khotan, pp. 45 sqq.), in Mastûj (cf. Desert Cathay, i. p. 41