138 REMAINS OF THE KHOTAN OASIS [Chap. IV
brought relics of widely different periods to rest on the same level, the chronological evidence of such
finds can never be equally certain.
The relievo fragments recovered from the débris layers of the shrine were so small that, had it not been for the evidence supplied by the stucco decoration which the walls of the Rawak Vihâra had retained still in position, it would have been difficult to form any adequate notion of their original character and arrangement. The relative ease with which the evidence of the Rawak relievos enabled us to interpret and classify these small disiecta membra of what must have been once an elaborate scheme of sculptural decoration, is itself, I think, the best proof of that close connexion in style I have referred to above. In spite of the selection already made on the spot, the number of relievo fragments to be examined was so great that it required much care and effort to group them systematically and to establish all the decorative motifs, etc., represented among them. For the results of this detailed analysis which I owe mainly to the painstaking and very experienced help of Mr. C. L. Woolley, I must refer to the Descriptive List below. In order to save repetition and waste of space stucco specimens of identical character, even if not cast from the same mould, have been listed as far as possible under one heading. For the purpose of facilitating a rapid survey of the chief decorative motifs those headings which contain such descriptions have been distinguished by asterisks.
I must content myself here with brief references to the chief types of relievos which help us to form some idea of the general scheme of decoration adopted at this shrine. In the first place it is safe to conclude from the numerous fragments of ears, fingers, toes, noses, and other parts of the body more easily preserved through complete burning,s that the walls of the cella and passage must have been lined with relievo images of Buddha and Bodhisattvas just as the photographs of the Rawak Vihâra court show them. We have remnants of them also in the numerous pieces of drapery (A.T. i. 0028, 0029, oo63 ; iv. 0049 ; v 0044, etc.). In the passage none of these images appear to have exceeded life-size, a limitation for which the relatively narrow space there available fully accounts. It may have been different within the cella, and I am inclined to think that the scantiness of sculptural finds in the corners we cleared there, may be due to the prevalence of colossal statues upon which the conflagration would have had far less effect, and of which the clay masses would in consequence decay more completely.
That the images in the passage must have been surrounded by elaborate vesicas or haloes in relievo is proved by the abundance of decorative details which their shapes and the example of Rawak show to have undoubtedly belonged to such features. A selection from appliqué ornaments of this kind is illustrated by specimens reproduced in Plate VIII. The numerous pieces of a lotus-petal border (A.T. 0051, i. 0017), the cloud scrolls (A.T. 0030), and the still more frequent rows of flame-tongues (A.T. 0016, i. 0044, 0075) can all with certainty be ascribed to aureoles of varying sizes.' The examples of the Rawak Vihâra throw light also on the way in which the fields of these small Buddha aureoles were filled with plaques of varying design, showing floral patterns, conventional ornaments, or small figures of Buddhas. Of all these the Ak-terek fragments present a plethora. Of flower ornaments more or less elaborate we have often repeated specimens in A.T. 0020, oo60 ; i. 0012 ; of conventional patterns in A.T. 0033, 0019, 0087 ; v. 0050, etc. The Vajra with its fleur-de-lis shape often figures in them just as at Rawak.10 Quite as frequent are small plaques of Buddhas seated within lotus vesicas (A.T. 0025. a, 0027, 0028 ; i. oo88 ; iii. 0089 ;
8 See A.T.0024, 0041 ; 1. 006, oolo, 0052, 0053, 0058, 0084 (Pl. IX), 0087, 0098-00102 ; ii. 0.048, 005r, 0058 ; 001, 0023, 0030, 0038, 0047, 0062, 0063, 0066, 0070; iv. 0012, 0051, 0064, 0077, 00128, 00132 ; v. 001, 002,
004, 0072 (P1. VIII).
° For the appearance of the large aureoles of Rawak, see Ancient Khotan, i. Figs. 63-5 ; ü. P1. XVIII.
10 Cf. Ancient Khotan, ii. Pl. XVIII. c, LXXXIII.