National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

much damage, the structure as a whole is in so much better condition than the
rest of the buildings of the town that its being of later date can scarcely be
doubted. The local tradition, which ascribes to it an approximate age of six
centuries, supports this conclusion.
The ruined mosque, with its entrance facing north-east, shown in Fig. 68,
contains a central hall 27 feet 6 inches wide; if its badly broken arcaded
south-western end, orientated towards Mecca, be included, it was probably of
the same length. On either side are two shallow niches divided by semi-detached
columns and having an oblong narrow window high up under their pointed
arches. The entrance, which is 7 feet wide and 8 feet deep, is adjoined on either
side by a small room measuring 8 by 12½ feet, and is approached through a porch
also 7 feet wide. The porch is provided with seats on both sides under small
niches, and its ceiling still retains stucco decoration with geometrical tracery.
Judging from the style of this decoration and the survival of stucco elsewhere on
the walls, it seems difficult to believe that the structure as it now stands can be
earlier than the fifteenth to sixteenth century. But the massive substructure of
the terrace which bears the mosque looks distinctly old. An outer room adjoin-
ing on the north-west holds a large tombstone, now upturned, bearing an
inscription in Kufic characters.
Along precipitous ledges of the same knoll, both to the west and south of
the mosque, there are to be seen rows of graves cut into the rock. Many of these
are partly destroyed owing to the rock having crumbled away, a sign of the long
period that must have elapsed since they were used for burial. Higher up on the
knoll a steep ridge to the north contains more rock-cut graves, among them three
that manifestly had remained undisturbed. On examining one of them, 1 foot
6 inches wide on the surface, we found at a depth of 3 feet stone slabs, 9
inches wide, covering the opening of the grave proper, and to the right a shelf
of the same width cut in the solid rock. Of the body interred only a few decayed
bones survived. The construction of the grave corresponded exactly with that
of the graves found at Tiz.⁶ The graves were orientated from north-west to
south-east, though, being undoubtedly Muhammadan, they ought to have lain
in the orthodox north and south direction. I shall have occasion later to refer to
similar irregular bearings observed at the great cemeteries of Shilau. On the
same ridge lay scattered half a dozen coffin-like tombstones, most of them
broken or upturned, all bearing Kufic inscriptions. Fig. 81 shows the best
preserved of these.
Crossing the depression in which the dry torrent-bed of Shilau bends south-
wards to the sea, we ascended the crest of the ridge which overlooks the main