Sec. ii] THE STUPA CELLA M. tit AND ITS WALL-PAINTINGS 497
the difficult work of removal. But it was necessary also to complete the clearing of the circular passage and to make a careful record of the condition and place in which the various remains of the wall-paintings were found, before attempting their rescue. In my Personal Narrative I have described the trying physical.conditions in which this preliminary work of note-taking and photographing, as well as most of my subsequent tasks here, had to be effected, under icy gales ever blowing from the north-east, and with constant crouching in cramped positions imposed by the want of space.5 Here I may at once record what the detailed examination, made both during and after the clearing,-showed as to the disposition of the surviving wall-paintings and the probable character of the decorative scheme as a whole.
As already mentioned above, the circular passage within the rotunda (Plate 32) had been lighted by three windows, besides the entrance passing through the west wall now completely destroyed. The windows, which were found to be almost exactly orientated towards the north, east, and south, reached down to 2 feet 8 inches from the floor and had a width of 2 feet 3 inches. The four segments into which the enclosing wall was divided by windows and entrance were decorated below with a painted frieze or dado. This, on the north-east and south-east segments, where alone the walls still throughout retained their plastered surface high enough, consisted of six lunettes ; from the hollow of each there rose the head and shoulders of a winged male figure of nearly life size (Plates XL, XLI). Nothing of the dado had survived on the south-western segment ; but on the north-west, near the north window, were the badly effaced remains of two lunettes containing angels' heads. It may thus be considered as certain that the same decorative scheme was used for the dado all round the interior walls of the rotunda. It is in accordance with this that the surviving sections of the dado have been marked in the plan (Plate 32) with numbers running from i to xxiv on the right and left respectively of the north window.
The rows of lunettes were united by a black band, nearly an inch wide, which ran the whole length of the dado, separating it from the paintings on the main wall above. The top of the dado as thus marked reached a height of 3 feet to inches from the floor. The length of each lunette, as measured along the top or chord line, was about 2 feet 2 inches, and necessarily presented a slightly curved surface. Including the series of bands bordering the lower curved edge, each lunette was of an average height of about t foot 6 inches. Below the row of lunettes there stretched a horizontal band about 9 inches wide, filled with eight or nine wave lines in black, and once, so far as could be judged from the damaged surface of the plaster, extending also upwards into the spandrels left between the lunettes. This band, poorly preserved in most places, was curiously suggestive of the sea. Below it the surface of the plaster was too much perished to show any traces of painting, if there ever was any so low down.
Leaving the figures represented in this dado of ` angels ' and all questions connected with the style and technique of their pictorial treatment to be discussed further on, I may next turn to what had survived of the paintings once adorning the upper faces of the wall. Only the scantiest remains of such frescoes—to use a conveniently short, though here technically inaccurate term—could still be traced in their original position. What little there was left of painted plaster surface above some of the lunettes of the dado showed only the feet of a few small figures with the lower parts of their garments. Such remnants can be seen in Fig. 127 above the fallen pieces of plaster hiding lunette v, and also above the angel ii in Plate XL. Other traces of this kind are recorded in my notes above lunettes iv, ix, and x ; in no case, however, did the perished plaster permit of removal.
It was not from the badly broken walls but from the débris accumulated at their foot, within the north-east and south-east segments of the passage, that_I was able to recover the fragments of
S See Desert Cathay, i. pp. 462 sqq.