Sec. iii] THE SITE OF STRAF 205
town area lay in plenty at the foot of the `sea wall', in places being embedded in the foundations.
Towards the western extremity of the still-surviving portion of this sea wall there are to be seen the remains of what evidently had been a large building, marked B in Plan 17. They consist of huge blocks of stone and masses of concrete, which look as if an earthquake had thrown them down in utter confusion. Between these remnants of massive masonry there are to be seen what obviously are bases for pillars. They consist of large oblong stones set in mortar and measure uniformly about 2 by 3 feet. The building appears to have measured about 55 yards in length from north to south and some 25 yards across. Is it possible that this large ruin is that of the `fine mosque with columns of teak wood' of which the remains were noted by Yâqût? 5
Before describing the observations made on the higher portion of the site, reference may be made to the ruin of a small outlying walled enclosure situated where the alluvium brought down in the torrent-bed of the Kunârak has caused the shore to jut out into the sea. It lies near the date-palm groves of Bâghi-Shaikh and about one-third of a mile to the west of the limit of the debris area of the town. Here, close by the beach of the little bay, are seen remains of walls enclosing a court about 42 yards square. It is known by the name of Bang. The walls, best preserved on the east and north faces, are 3 feet 9 inches thick and, like all those of Sirâf, built with unhewn stones set in mortar. The east face towards the sea shows four fairly well-preserved triangular projections, the sides of which are 19 feet 6 inches long (Fig. 70) . As these are not found on the other faces of the small fort, it may be concluded that the projections were intended to strengthen the face exposed to the waves. A smaller enclosure with a wall similarly protected on the east side and traceable for about 100 feet appears to have been added on the south. The exact purpose of the structure could not be determined. Judging from the masonry used and the potsherds picked up around, it may be assumed to be coeval with the town.
Proceeding from here north-eastwards we reach the extremity of the low ridge which separates the town area from the valley of Shilau containing its necropolis. Before, however, describing the remains to be found along the crest of this ridge, I may briefly notice those on a rocky knoll which faces it from the west and occupies most of the space between the mouth of the Kunârak gorge and the dry bed of the Shilau nullah that descends from east to west. On the southern slope of that knoll there rises the conspicuous ruin of a structure rightly known to the people of Tâhiri as a mosque. As seen in the photograph, Fig. 68, it is built partly on a massive walled-up terrace. Though the walls have suffered
5 See above, p. 203.