Sec. iii] THE RUINED FORT AND STLJPA OF THE ENDERE SITE 433
The only other structure of any size of which remains could be traced within the ruined fort was the dwelling (E. v) to the south-east of the temple. The few broken posts protruding from the sand, which first attracted my attention to it (partly visible in the foreground of Fig. 5 t) proved to have belonged to an upper story which otherwise had almost completely disappeared through erosion. Its floor seemed to have been approximately on a level with that of the temple and of E. iii. There were distinct indications that the walls of this upper story had, apart from the timber framework, consisted of rows of cakes of hard-beaten sheep-dung placed vertically and joined by layers of clay. From my labourers I learned that the use of this queer building material is not altogether unknown at the present day. Below these thin walls the excavation brought to light a lower story or basement, which, as the plan shows, had main walls of stamped clay over 4 feet in thickness. The interior walls dividing some small apartments were of the usual timber and plaster type, and only 6 to 8 in. thick. It is probable that the great thickness of the main walls was intended to meet the pressure of the surrounding soil. There was no entrance leading through any of the outer walls of this basement story, which thus could have been entered only by stairs from above. Probably one or the other of the small apartments separated by timber and plaster walls had contained the ladders that served for stairs. Set into the clay walls were massive posts, which continued into the superstructure. They are seen in the photograph reproduced in Plate X, which shows the south-western corner room after partial clearing. The large and well-preserved fireplace found in this room (also seen in the photograph) showed a projecting top decorated with elaborate mouldings in stucco. In construction it closely resembled some of the fireplaces found in the Dandan-Uiliq dwellings. Its presence proves that the basement rooms, all of which appear to have had a height of 9 ft., must have been partly at least used for living purposes at certain seasons. Such wholly or partially underground apartments were, no doubt, easier to keep warm in the severe winter. They may have thus corresponded in a reverse way to the Tai-khanas of Peshawar and the northern Punjab, constructed as a cool refuge against the torrid heat of the Indian summer months, which are alluded to, I think, in Philostratus' account of Apollonius of Tyana's Indian journey. The laborious clearing of these basement rooms was not rewarded by any finds of movable objects.
No structural remains, besides those already described and the much-decayed walls of a room built against the northernmost part of the ramparts, could be traced within the circumvallation. Nor, having regard to the conditions of the sand and to the available space, did it appear likely that other ruins of any size could lie hidden beneath the drift-sand. In the northern portion of the enclosed area, where a good deal of ground lay almost bare of sand (see plan), compact layers of rubbish, consisting chiefly of horse-dung, stable refuse, and straw, were found in one place to cover the soil to a height of 3 to 4 ft. In order to assure myself of the character of these rubbish accumulations I had a broad trench dug across the whole of them down to the natural ground. But the only finds, besides the refuse described, were broken pieces of coarse pottery and occasional bits of rough cotton fabrics and felt, as well as some cotton seed. It was evident that this portion of the ground had been used for stabling during a prolonged period. Further to the north the bare ground was covered with coarse potsherds, and the same also strewed in profusion the open ground to the south of E. i and E. iii, which erosion, still actually proceeding, had lowered from 3 to 5 ft. below the floor-level of the latter buildings. Just outside the gate of the fort pottery débris was also plentiful.
Not the least interesting feature of the ruins were the remains of the ramparts which had once defended this small agglomeration of buildings. There was nothing to tell of the attacks
STEIN 3 K
Excavation of dwelling E. v.
Underground rooms of E. v.
Refuse accumulations on ground.
Construction of ramparts.