22 ACROSS SWAT AND DM [Chap. I
of the Sat-ghara temples at Ketas in the Salt Range.3 A small stream fed by springs passes close by the east face of the ruin, and waters some fine Chinars growing near it. The temple had already suffered badly when I first saw it in 1897, most of the finely cut sandstone facing having been removed, as Colonel Deane heard it, by a former Khan of Dir. It was sad now to find this stripping nearly completed by the villagers, a strange writing on the wall, as it were, by the hand of advancing ' civilization '. What carved stone still remained in 1897 on the east and north wall faces had been all carried off since. Thus I found in Gumbat itself several fine slabs of yellowish sandstone, including a large block, once belonging to a corniced frieze or a pilaster capital, walled into a blacksmith's furnace. The interior of the trefoil-arched porch on the east and the inner wall faces of the cella had suffered less damage, and the careful stone lining of the narrow passage encircling the vaulted chamber of the upper story was still intact in great part, its less accessible position affording protection.
In spite of all this damage the measurements obtained on the south and the east faces where the large corner slabs of the base were still in position, together with the indications afforded by similar structures in the Salt Range and on the Indus, made it possible rapidly to prepare the ground plan shown in Plate I. This shows that the main feature of the shrine consisted of a single cella, 9 feet 8 inches square, approached on the east by a porch 9 feet broad. The original depth of this porch was doubtful, as its outer masonry had completely disappeared. That it was once surmounted by a trefoil arch was shown by what remained of the upper portion of its side walls. The dimensions of the entrance leading from porch to cella remained also uncertain ; but traces of mouldings surviving above it showed that its shape was rectangular. Flat pilasters had flanked it. The cella was surmounted by a dome of horizontal construction, resting on seven successively projecting courses which cut off the corners and converted the square into an octagon. The dome rose to a height of 13 feet 8 inches from the floor of the cella. The porch was flanked on either side by a small recess about 2 feet 6 inches deep and 1 foot 6 inches wide. The one on the south, at a height of about 6 feet, gave access to a narrow flight of stairs built in the thickness of the cella wall, and leading to a corridor about 1 foot 3 inches wide. This formed a passage on three sides round a vaulted chamber, surmounting the cella dome at a height of 17 feet 6 inches above the base and forming an upper story. The passage, about 5 feet 6 inches high, seems to have led to another flight of stairs, which probably served for the approach of a third story. But, as the highest part of the existing structure does not rise above 27 feet from the top of the base, the height of the third story, and of the stone roof which must once have crowned the whole temple, could not be determined.
Owing to the outer walls having been completely stripped of their sandstone facing, as seen in Fig. 3, the external measurements of the structure could not be secured with absolute accuracy. On the east face 21 feet 3 inches were measured between the existing corners of the bold mouldings surmounting the base (seen on right of Fig. 3). The south face showed a central projection to feet 6 inches long, between flanking portions each of 5 feet 2 inches. A niche, 2 feet 6 inches square, occupied the centre of each wall face, a little above the level of the cella, except on the east, and was, no doubt, meant to shelter some divine image. The interior masonry consisted throughout of roughly hewn sandstone set in very hard mortar. Like the inner walls of the cella, which still showed carefully smoothed sandstone slabs, 15 to 19 inches long and 4 to 6 inches high, laid in regular courses, the exterior walls had also once possessed their proper facing of carved stone. Though this had weathered badly, I was still able in 1897 to distinguish remains of