162 FROM KERMAN TO BANDAR ABBAS [Chap. VI
ment for me to diverge from the main programme planned for my resumed explorations. These were intended to take me first through the south-eastern parts of the Kerman province to the entrance of the Persian Gulf, and sub-
sequently along the shore of the latter into the southernmost portion of Fars, the ancient Persis. Consideration for the trying climatic conditions to be
experienced along the coast, except during the short winter of that region, made me anxious to avoid any delay in my start southward beyond that which the difficulty of securing transport imposed.
Hence no time could be spared for visiting any of the ruined sites the Surveyor had traced in the Zarand and Kôhistan tracts to the north and north-
west of Kerman. But at a number of localities in that direction he had noticed
cairns of the type so frequent in Baluchistan and Makran, marking deposits from pre-Muhammadan burials, and the convenience of access offered by the
track towards Zarand being just practicable for cars allowed me at two points to verify the correctness of the Surveyor's diagnosis by a day's excursion from Kerman.
After driving about 10 miles we passed the isolated rocky hill known as Qal`a-i-dukhtar, `the Princess's Castle', rising like an island to a height of
about 350 feet above the bare alluvial glacis. Its extremely precipitous sides are crowned on the top with decayed walls of sun-dried bricks. Other
walls of very massive construction protect approach from the flat ground on
the north to the very narrow ravine which, flanked by almost vertical cliffs, appears to have provided the only means of access to the top of the hill. Ascent
in this ravine or rift proved impracticable for us, as the masses of masonry which once carried a footpath had fallen away, and the rock faces on either side afforded no foothold.
The walls flanking the mouth of the ravine are extended for some distance over the level ground in front of it, and being joined by a cross wall containing
a gate they form a kind of outwork. These walls are broken at regular intervals by arched niches with large loopholes opening towards the outside. The very ruinous condition of this outer enclosure and the care with which the great natural strength of the fastness has been increased by defences clinging to precipitous cliffs point to early construction, probably in Muhammadan times.
But the few pieces of coarse pottery and plain glazed ware which were picked up within the outwork near the hollow marking a former well, furnished no
chronological clue. The name Qal`a-i-dukhtar is to be found elsewhere also in Persia applied to hill strongholds of evident antiquity.2 But nowhere did I learn
2 Ruined forts and natural fastnesses known by Jahrilm, north of Fasâ, Istahbânât, &c. The same this name were noted by me in Fârs at Firûzâbâd, designation is applied also to the hill fortress rising