I190 KARA-SHAHR AND ITS RUINED SITES [Chap. XXIX
a well-modelled Bodhisattva head in high relief treated in almost pure Gandhâra style and set within a decorative square frame. Its material is a fine evenly-fired clay. Fragments from the same or a similar mould were found in numbers at other ruins also. The fragment of a pottery vessel, Mi. xxiii. 0026, is of interest, as it bears a short Tibetan inscription incised before burning.
Mould for From two small detached cellas further north, xxiv, xxvii, six fragmentary Pôthi leaves in
stucco Brâhmi script were recovered, besides the fine relievo plaque Mi. xxiv. ooi (Plate CxxxvIi) and the
relievos. badly injured fragment of a painted wooden panel. Beyond these cellas again stretches a series of
shrines which retain their walls to a fair height, but have suffered badly from moisture in their interior. The only noteworthy find made here was that of a plaster of Paris mould, Mi. iv. ooi (Plate CXxXVI1), which has forms for casting a small seated Buddha and curls and zigzag locks of hair for larger relievo figures. The find was of special interest to me at the time as definitely proving the use of moulds for the production of those stucco relievos which came to light in such abundance and variety from other ruined shrines of this site. Since then information has become available about the important discovery of some thirty such moulds which was made by Professor Grünwedel's party in two cellas of group II.14 Finds of exactly corresponding moulds that I made at Khadalik have been recorded above.16
Group HI, Among the shrines cleared in group III, along the easternmost ridge, no finds were made
shrine xxvi. except in the cella xxi, where an octagonal post with a line in Brahmi on each side came to light, and in the large temple xxvi. It rises as an imposing pile on a high walled-up terrace, seen in Fig. 288 on the left. The top of this terrace measures approximately 8o feet by 68 ; it must have been once approached by stairs leading up the north-east face, but these are completely covered by heavy masses of débris and could not be cleared within the available time. The cella, 222 feet square, is enclosed by walls 3 feet thick, now reddened and hardened by fire. The interior was filled to a height of more than 8 feet by similar hard débris. Passages, about 6 feet wide and each lit by a window, led on the north-west and south-east sides to the chamber at the back, about I I feet wide. The approach to the cella leads through an antechapel or front hall of unusually large dimensions, about 37 feet deep, and this is flanked on either side by a small subsidiary chapel, about 8 feet square, another unusual feature.
Fragments The clearing of the front hall was rewarded only by a few small fragments of stucco relievos,
relievos. including the two small heads Mi. xxvi. oo1, 004. The hard-burned débris of the cella was partially
cleared, but yielded only a few small relievo pieces like Mi. xxvi. 002, 0010, still recognizable. Here, too, the outer passage walls had escaped the worst effects of the conflagration, and it was along the north-west wall and in the west corner that the remaining stucco relief fragments, including the decorative bands Mi. xxvi. 008-9 (Plate cxxxvII) with appliqué flowers, were found. They must have fallen early from stucco friezes fixed to the outer wall, as will be described presently in shrine x-xii. In some places this wall still retained the wooden pegs which helped to carry these friezes. At the south-east foot of the terrace supporting xxvi there was found a cinerary urn of rough pottery containing completely decayed human bones.
Sepulchral To the east and south-east of group III of shrines the lower ground is occupied by the
monu- sepulchral monuments of the two types above described. All of them appear to have been
enclosed by rectangular walls of no great height. Those of the ` Stûpa' type in their domed portion show a curious resemblance to the felt tents, or ` Ak-ois', of Kirghiz and Mongols. In none of these monuments were there signs of burning, but there is reason to assume that their interiors, easily accessible as they were, had been searched again and again for `treasure'. The one nearest to xxvi and best preserved (Fig. 288), with a domed chamber about 13 feet in diameter, was cleared
" Cf. Grünwedel, Allbuddh. Kullslälten, p. 192. 16 See above, pp. 158, 187.