specimens reproduced in Plate LXXX, with the corresponding fragments from the Dandân-Uiliq shrines D. i and D. II (see Plates LIV, LV), will fully bear out this observation. That this aureole in flat relief fixed on the temple wall had served to encircle a large image may be considered as certain. But of this image no recognizable trace has survived ; it was in all probability sculptured in soft stucco, just as in the Dandân-Uiliq and Endere shrines.
By the side of this general agreement in decorative design and arrangement the Kighillik fragments display unmistakable differences in style and execution, which seem distinctly to indicate an earlier date, and a close connexion also with the relief decoration of the Rawak Stûpa court, to be described hereafter. The small Buddha figures, of which portions were recovered in numerous replicas, undoubtedly represented standing Buddhas, with the right hand raised in the ` Abhayamudrâ attitude, and the left hanging down and holding the robe, as illustrated by the pieces A. 01, A.o21 in Plate LXXX. The attitude is the same as in the Dandân-Uiliq aureole reliefs D. II. 34+74, shown in Plate LIV, but the size is far larger, and approaches that of the corresponding reliefs from the Rawak aureoles (see Plate LXXXVI and Figs. 63, 64). Another important point of contact with the latter is the absence among the Kighillik remains of any fragments representing the feet or lower parts of the drapery of these Buddha figures. We shall see that at Rawak the corresponding relief-images were ranged in overlapping rows within the haloes without their lower parts being shown anywhere. The same arrangement may be assumed in the Kighillik wall decoration. In the Dandan-Uiliq aureoles the corresponding figures stood apart, and were hence fully shown.
Unfortunately the Kighillik pieces are so fragmentary that we can judge only from small details as to the general characteristics of their style and modelling. But these all indicate unmistakable superiority to the Dandan-U iliq work, and, as far as care in execution is concerned, also, I think, to the Rawak sculptures. But in comparison with the latter the superiority of the material, hard plaster of Paris as against very friable clayey loess, must not be lost sight of. Hence the chronological relation between the Kighillik and Rawak Stûpa reliefs scarcely permits of definite settlement at present. But that both of them must be assigned to a considerably earlier period than that ascertained for the Dandân-Uiliq shrines is proved by the style of the Kighillik remains quite as clearly as by the more specific evidence which, as we shall see, is available in the case of the Rawak Stûpa sculptures.
The great majority of the Kighillik fragments belong to the bodies of small standing Buddhas. One of the largest is the draped torso (A. oi) with the raised right hand (see Plate LXXX). Still better preserved drapery is shown by the replica A. 04 (see ibid.), which retains the original red colour of the robe, now burnt to deep rich brown. Traces of the original colouring remain also on other fragments (A. 05, 06, 09, 010). Portions of the stucco, by which the figures were attached to the wall at the back, are still preserved in A. 012, 017. Numerous fragments prove that they had been broken before they were blackened by fire (see e.g. A. oI I, 012). Of the well-modelled left hand grasping the end of drapery, with more of the latter loosely hanging over the forearm (see A. 02 I, ibid.), we have numerous replicas (A. 016-022). Of the head of the Buddha only two specimens, unfortunately badly injured, survive (see A. 041, ibid.). Of the small nimbus, too, with a raised border, which surrounded the head, also originally coloured, we have but few fragments (A. 05o-052, see ibid.). Relatively numerous fragments (A. 035-038, 054-057, see ibid.) belong to the border of a large aureole, formed by overlapping lotus-petals and corresponding in arrangement to the similar relief decorations of the Dandan-Uiliq shrine D. II (see Plate LIV, LV). The latter also show how the overlapping flame-tongues (A. 039, 040, see Plate LXXX) are likely to have been