Sec. i] REMAINS OF GALEHDAR 225
the top of the outer wall. Here, too, doorways, of which three still retain large unhewn slabs as lintels, lead through the inner wall. The fact that this inner wall with its doorways faces north seemed at first very puzzling. But an explanation was afforded by the fact that about 80 feet higher up the precipitous cliffs of the two deep-cut ravines on either side converge until a space of only about 8 feet wide is left for approach to the top of the spur. No remains of a tower or wall were traced here, but the natural defensive strength of the spot is obvious and so also the likelihood of rough masonry in such a position having given way in the course of centuries. The western end of the upper wall line runs close to a narrow ravine and descends steeply for about 35 feet to where a very massively constructed barrage, 46 feet long, closes this ravine. Filled by drainage from the top of the spur the reservoir thus formed would assure a considerable supply of water for those seeking protection on it.
The position chosen for the Takht-i-pirist reminded me curiously of that occupied by the ancient hill stronghold of the Zamr-i-atash-parast surveyed by me in Wakhan on the uppermost Oxus in 1915.3 There, too, a steep spur flanked by precipitous ravines on either side and connected with the range behind only by a very narrow neck had been fortified in order to secure a safe temporary place of refuge for a large number of people ordinarily living in the valley below. There, too, the fortified line consisted of a double wall. A variety of indications had led me to assume pre-Muhammadan origin for the defences of the Zamr-iatash-parast. At the Takht-i-pirist I felt inclined towards the same conclusion, although no even approximately datable remains could be found on the surface. An early date seemed to be indicated for these defences by the negative fact that the comparatively solid construction of the walls completely lacked that use of mortar or cement which all medieval structures of this region invariably show, and for which the material is just as readily available here as it was at Sirdf or on the old routes leading across the coastal range.
SECTION II—TO WARAWI AND UP THE GULF COAST
The occurrence, on January 28th, of the Id festival closing the Ramadan precluded either a move from Nauba or more work at Haraj. I used the enforced halt for a visit to a ruined stronghold known as Kôghan. It was found to be situated about 2 miles to the south-south-west of Nauba on the top of a precipitous ridge marking the extremity of the hill spur to the east of the Tawakali stream, which we had followed on our descent into the Galehdar valley. The ridge rises with wall-like cliffs of conglomerate to a height of about 300 feet
3 See Innermost Asia, ii. pp. 866-70; iii. Pis. XLVII, XLVIII.