OLD REMAINS IN WAKHAN 871
crossed by a bridge. The cleft in which the glacier stream of Darshai has here cut its passage through the foot of a rocky spur is only a few feet wide at the top. Some circular holes, about 3 inches in diameter and cut with much skill, are seen in a rock above the eastern edge of the cleft. They are ascribed to ` Kâfirs ' of old times and undoubtedly once served for rafters carrying an earlier bridge. A large boulder to the west of the canon bears numerous coarse sgraffiti of ibex, markhar, and other game, as well as of rude figures of hunters ; their date is quite uncertain.
On the top of an isolated rocky ridge reached over bare cliffs to the north of the bridge and at a height of 200 feet above it, there rise the ruins of dwellings, as shown in the sketch-plan, Pl. 48. They are known as the ` Kafir fort ' of Darshai and cover whatever level space is afforded by small terraces. The position is by nature an extremely strong one ; for unscalable cliffs protect it from the side of the canon on the west and from the north, while the bare rock slopes towards the south and south-east are easily defended from above. The walls of the dwellings, standing in places to a height of 6 or 7 feet, are built of water-worn stones set in mud plaster. Those of a larger structure occupying the highest terrace were of a more solid construction, being built of flat unhewn stones roughly set in courses, as seen in the walls of Zamr-i-âtish-parast. The walls of this highest structure still retain in parts a facing of hard plaster. This top terrace appears to have been connected by a wall with a rocky offshoot to the east, which on a level about loo feet lower bears the much-decayed remains of a few small structures. No circumvallation was traceable elsewhere, nor was one needed to make the agglomeration of dwellings a safe temporary refuge. A narrow steep gully appears to have been used for access to water in the cleft below. All that can be safely asserted as regards the date of this ` Kâfir fort ' is that its occupation probably goes back to pre-Muhammadan times, though it may possibly have continued on occasion into a later period.
SECTION IL—THROUGH ISHKASHM AND GHARAN
Our march of September 6th was of interest since it carried us into that portion of the Ab-iPanja valley which lies at the great northern bend of the river. Though, judged from a small-scale map, it might seem but the natural continuation of Wakhân, yet we shall presently see that in linguistic respect and also politically it may claim to be considered part of a small but distinct geographical division. Less than a mile below Darshai the river becomes confined to a narrow canon-like bed, and the route above this on the right bank skirts the barren rocky slopes of a succession of defiles. Cultivation restricted to small detached patches comes to an end at the isolated homesteads of Ramanit and Udit, and the steep rocky spur of Sang, which we passed 3 miles beyond the latter, has been recognized since old times as the traditional boundary between Wakhân and Ishkâshm on the right bank. After crossing this spur, the road lay over a wide stony fan and then brought us to the village of Namadgut, situated amidst orchards and a fine expanse of cornfields. Though inhabited by Wakhis, this pleasant place is reckoned as belonging to Ishkâshm.
In a contribution to Sir George Grierson's analysis of the linguistic materials brought back by me from this portion of the Oxus valley I have pointed out that a close ethnic and political connexion has existed since early times between the Ishkâshm tract on the Ab-i-Panja and the valleys of Zébak and Sanglich on the upper Wardôj river. The same connexion is also reflected in a striking manner by the practical identity of the Galcha language spoken in all three mountain tracts. It results from well-defined geographical facts, and as these have a distinct interest of their own my remarks on them may conveniently be repeated here.1 ` We have here an interesting
I Cf. Grierson, Ishkâshmi, Zèbaki and Yâzghulàmi, pp. 4 sq. I quote my remarks with a few slight alterations in spelling
Canon of Darshai stream.
Ruined dwellings above Darshai stream.
Past border of Ishkâshm tract.
Connexion of Ishkashm with Zébak and Sanglich.