Sec. ii] THE REMAINS OF AK-SIPIL AND KIGHILLIK 479
arranged on the outside of the aureole border. The employment of the stucco fragment showing a portion of a jewelled chain (A. 043) remains doubtful. Finally, we may mention some pieces (A. 044) still showing the matrix of a wood core, used here, as in all the larger reliefs of Dandan-Uiliq and Rawak, for strengthening the stucco work.
Close to the original ground and near the centre of the mound there were found some pieces. of perished timber which retained traces of red colouring in spite of their decayed surface. Here, too, was found a tiny piece of leaf-gold which may have peeled off from some gilt statue, supporting the view expressed above as to the origin of the leaf-gold washed from the débris layers of Yotkan 7. What purpose a small piece of talc found here may have served remains uncertain. The ground round the destroyed shrine and all over the flat bottom of the Nullah was strewn with ancient pottery débris, much of it being of fine texture and superior in make to the potsherds ordinarily found at the Tatis north of the Khotan oasis. Some of the fragments of terra-cotta vessels described under A. 007 show incised work, others a carefully-smoothed surface or red glaze, while in two pieces (A. o07. g, h) there appear moulded details, probably of grotesque masks such as figure so plentifully among the appliqué ornaments of Yôtkan pottery.
About 7o ft. north of the remains of the shrine there rises from the foot of the dune eastwards a large mound composed mainly of dry dung (kighik in Turki), apparently horse-dung, but containing besides an admixture of bones, charcoal, and chopped bits of fuel. This huge refuse-heap, which measures as far as exposed over 70 ft. from north to south, with a breadth of about 5o ft. and a depth of over I r ft., has not escaped the attention of ` treasure-seekers'. The regular galleries they have tunnelled into it, two of these reaching down to the bottom and being about 18 to 20 ft. long, enabled me to ascertain its contents with relative ease. This huge accumulation of dung proves that the site, whatever its character, must have been greatly frequented 8. The dunes close by may cover other remains, but their height precluded any thought of trial excavations on either side of the little depression.
See above, p. 194.
$ Prof. L6czy, in his instructive description of the parts of Kan-su visited by the Széchenyi expedition, mentions the large accumulations of old manure visible outside the road-
side stations, half-buried by sand, on the old desert route leading from An-hsi-fan towards Khâmil; see Khinai birodaZont, p. 494.