Sec. iv] PAST THE MAZAR-TAGH OF KHOTAN 93
Extensive burrowing in another place had yielded no ` finds ' whatever, but was nevertheless Charred
of some interest. I had ascertained in 1908 that the large outer court of the fort must have been masses of
used for stabling. The obvious evidence of this, thick layers of manure, had not prevented Tokhta outer court.
Akhûn, the chief `Taklamakânchi' of Islâmâ.bâd and a queer person of the true ` treasure-seeking ' type (Fig. 86), from digging a trench 6 feet wide and fully Io feet deep along the inner side of the NE. wall enclosing this court. This labour, performed with all the perseverance of an âme damnée, had resulted only in laying bare the depth of the accumulations of horse-dung, but not without some profit in the way of archaeological evidence. It showed me that the outer enclosure wall, standing to a present height of 6 feet and about 4 feet thick, had been built over a solid layer of stable refuse already over 4 feet deep at the time. This clearly points to prolonged occupation of the hill crest at this place before the extant Tibetan fort was built.8 At the same time Tokhta Akhûn's burrowing furnished me with an explanation of certain signs that had previously puzzled me. In the court I had noticed marks of what seemed to have been a big conflagration. Yet of destruction by fire I had failed to trace any evidence within the fort. Kâsim Akhûn now explained that the charred condition of the lower masses of manure was due to a slow process of burning, such as he had often observed at the manure heaps of ` Oghils ' or herdsmen's stations. It was obviously due to this process that the walls of the court owed .the burnt appearance of their brickwork.
A more interesting discovery, however, rewarded my examination of the steep and very narrow Remains of
rocky spur with which the hill falls off eastwards to the river bank. In 1908 I had failed to visit ruined
the stacks of beflagged poles that here mark the ` Mazâr' of some alleged Muhammadan saints,
about 5o yards from the fort and about zoo feet below its level (Pl. 3). It was only now that I noticed some old pieces of timber used for the rough enclosures of two of these stacks. As I looked for the spot from which they might have been taken, my attention was caught by a small and very low heap of debris on a little shelf of the slope below and about 68 feet from the Mazâr (Fig. 91). The uniform surface of decayed sandstone debris with which it was covered like the hill slope around, had here hidden the badly eroded wall foundations of a small temple cella and its enclosing passage. The surviving remains of beams and of tamarisk sticks set in rows, which a careful clearing revealed, left no doubt that the small structure had been built in timber and wattle exactly after the manner of the shrines first excavated by me at Dandân-oilik and later at Khâdalik and elsewhere in the Khotan region. That the scanty remains were those of a Buddhist shrine became quite certain when the clearing brought to light well-carved wooden finials (M. Tâgh. c. o6-8, Pl. VI) of the same type as those found at Khâdalik,9 and pieces of hard-faced fibrous plaster, originally laid over a canvas backing, which manifestly had belonged to stucco relievos (M. Tâgh. c. 01-4). Tiny fragments of painted stucco, evidently from frescoed walls, could also be picked up in the debris. A badly weathered piece of wood, probably from a piece of panelling (M. Tagh. c. i. o1, Pl. VI), showed elaborate ornamental carving of a style found elsewhere at Khotan sites.
The dimensions of the cella, 13' 8" by II' 2", could be traced by means of the foundation beams Relics from still in situ, and, outside it on the east, the wall of an outer passage 5' 6" wide (Pl. 3). Here remains ruined Cella. of stuccoed image bases, one of semi-circular shape between two rectangular ones, each about two feet across, were still distinguishable. The structural remains must have suffered badly from wind-erosion and been reduced to insignificant dimensions long before my visit in 1908. But there is good reason to believe that their final destruction was the result of subsequent treasure-seeking operations.' From these came in all probability the above-mentioned votive plaques showing a seated Buddha (M. Tâgh. 07-10, 028) and other small relics (M. Tâgh. 020, 050) that Kâsim
8 Cf. Serindia, iii. p. 1291, for a suggestion concerning earlier date.
the watch-tower above the fort, which may possibly be of 9 See Serindia, i. p. 194, Kha. viii. 002 (Pl. XVII).