36o TO TUN-HUANG AND AN-HSI [Chap. X
As these observations have since been recorded in Serindia,23 it remains for me only to add here some brief remarks on the changes which a rapid inspection enabled me to notice in the general condition of the cave-temples. Near the northern end of their main group, where Wang Tao-shih in his capacity as restorer had throughout been particularly active, there were many walls covered with fresh plaster, hiding remains of old mural paintings, as well as abundance of new and hideous statuary in stucco, to betoken the progress of his pious efforts. Less depressing evidence of his zeal was to be seen towards the middle of the same group of caves. Here much additional work had been done in clearing the approaches to the cellas of the lowest row, which had previously become partially blocked by the accumulation of drift-sand and by the gradual rise of the ground level outside. Here, too, some fifty shrines in the upper rows, which formerly could be reached only by means of rickety wooden ladders or still more rickety galleries, had been rendered easily accessible by the simple but destructive expedient of cutting passages from one cave-shrine to another right through the rock-wall separating them. Mural paintings in the way of the openings on either side had been ruthlessly destroyed in the process.24
But there were signs in a few places of a danger of another kind to the pictorial relics in the caves ; for attempts had been made here and there to cut out particularly striking details of fresco compositions, e. g. the fine head of one of the attendants flying by the side of Buddha's car in the noble wall-painting seen in Serindia, Figs. 215, 226. Fortunately these attempts, evidently prompted by some visitor's ` collecting ' zeal, had had to be abandoned before they could proceed far, or induce emulation by local hands bent on obtaining antiques for sale ; for the very hard and uneven surface presented by the conglomerate of the rock-wall must render it a far more difficult task to remove intact the mud plaster which here bears the tempera painting than in the case of the mural paintings found at sites of the Tarim basin or Turfân. There the thick plaster backing of the paintings rests against the smooth uniform surface of a wall of brickwork, wattle and plaster, or natural clay, and it can with due care and some skill be separated from the wall without too great risk of serious damage. It may accordingly be hoped that the mural paintings of the ` Thousand Buddhas ' will fare better than those of the cave-shrines in the Turfân and Kuchâ regions, and escape exploitation by inexperienced hands, whether of amateur collectors or profit-seeking natives.
Of the special difficulties presented by the conglomerate surface at the back of painted plaster I had occasion to acquire personal experience. I have already described in Serindia the great artistic interest presented by the beautiful mural paintings which cover the walls of the small cella, less than nine feet square, marked by me Ch. II. a and situated at the northern extremity of the main group.25 These paintings, apparently executed in true fresco, differ strikingly both in style and technique from those noticed by me elsewhere at the Thousand Buddhas '. For various reasons it appeared desirable to secure specimens of the work for expert examination, and the small detached panels in the same technique decorating either side of the narrow entrance to the cella, having already suffered damage through their exposed position, offered themselves as suitable for the purpose. The experiment of their removal proved difficult. The plaster, apparently containing much lime, was thin, very hard, and firmly adhered to the gravelly surface of the rock. In spite of all our care the removal could not be carried out without numerous breaks in the panels. As, however, these left sharp edges in the plaster pieces, there is some hope left that, when reset by Mr. Andrews' skilful hands, the little panels will serve to convey some idea of the remarkable