496 THE ANCIENT ßUDDHIST SHRINES OF MIRAN [Chap. XIII
greater or lesser weight we may be inclined to attach to the spelling. The seventh name K'ibhila remains uncertain at present.
Miscel- In the eastern part of the passage I was obliged to stop digging at the approach of twilight, when
laneous pieces of frescoed plaster, large and small, had begun to emerge from the débris. They had
dubs i.nevidently peeled off from higher parts of the wall of the little rotunda, and lay closely packed
débris. Y P g P
against the painted dado, as seen in Fig. 127, which shows a subsequent stage of the clearing. It was obvious that the greatest care and much time would be needed for the recovery of these fallen fragments of fresco. But in the west and south-west segments of the passage the treasure-seekers' destructive operations had allowed nothing of the fresco decoration of the walls to survive, and here clearing could be continued without risk. By nightfall it had yielded a number of interesting relics. There were long strips of finely-woven dark-red silk, M. III. 0064-65, which, judging from the stucco found adhering to them, had evidently been pasted along the mouldings of the bases of images such as might have been placed at the entrance of the shrine. The small capital of carved wood, M. III. 0021 (Plate xLVIi), which with its leaf ornament recalls a ` Coptic ' example, may also come from the same place. The cylindrical block of wood, M. III. 0024 (Plate XLVII), resembling the double hub of a wheel, is an object of uncertain use, which may possibly have belonged to the tee on the top of the Stûpa.
Artificial Very curious relics were found of what may have been votive offerings of the last worshippers
flowers. at the shrine. We came upon a number of artificial flowers, M. III. 001 3, 0027, skilfully cut out of
strong cotton and silk fabrics in a variety of colours, and cleverly made up with wooden pegs and tufts of thread to represent stalks and stamens. The way in which these flowers had been used was made clear by the discovery of several decorated pieces of stout cotton material, which had served as a background upon which to fix them. The largest of them, M. III. 0026 (Plate XLVIII), was covered with a thin coat of plaster painted dark blue, and still retained artificial flowers stuck into it with small pegs, as well as leaves and sprays of similar fabrics fixed direct to the plaster surface. The whole was apparently meant to represent a sacred lake with floating lotuses and water-plants. In other pieces, M. in. 0025. a-c, 0028, the background is painted direct on the fabric, stout cotton or silk, and still retains the holes in which the flowers had been fixed.
Material of The first day's successful work in this unpretentious ruin had revealed unexpected, and
frescoes. therefore all the more fascinating, glimpses of the influence of classical pictorial art reaching even
the shores of Lop-nor. It meant an illuminating discovery, but also the rise of new problems. Among them there was one which had to be faced at once and which was of a practical nature. For the fine wall-paintings now about to rise from their grave there was no chance of thorough study and protection but by removal. To effect this and the distant transport in safety was bound to prove a task of very serious technical difficulty. The remains of wall-paintings, whether still in their original position or lying as detached fragments of varying size among the débris, were all executed in tempera on a stucco backing which consisted of nothing but a layer of friable clay mixed with short straw of cut reeds.4b It showed ominous cracks even where it still adhered to the wall, and it had become very brittle. This was particularly the case with the fallen pieces, though the thickness of the plaster was here often greater. Both the detaching and the handling of these fragile panels of stucco demanded the utmost care, and the method and means for safely dealing with them had still to be improvised.
Clearing, The whole of February s passed in securing the requisite timber from the jungle of Mirân and
etc.,ofmural in turning it into the materials, such as planks and boards, immediately wanted for beginning paintings.
4b For Sir Arthur Church's chemical analysis of painted Paris coating with colours applied in tempera, see Appen-
stucco specimens from M. iii and M. y, showing a plaster of dix D.