Sec. ii] THE STUPA CELLA M. in AND ITS WALL-PAINTINGS 499
It is this relation to Graeco-Buddhist art which invests the fresco remains of the Miran shrines with exceptional interest, however fragmentary their condition. No remains of pictorial art corresponding in date and origin to the Graeco-Buddhist sculptures of Gandhâra have as yet come to light on Indian soil, and unless the cave-temples of Bâmian, or ruins as yet unrevealed below the soil of Bactria, may prove to have preserved them, the earlier stages of an art development which was destined to exercise a far-reaching influence throughout the history of Central-Asian and Far-Eastern painting are likely to remain for ever lost to us. Whatever archaeological evidence we possess at present seems to justify the belief that in the Mix-an frescoes we have the nearest approach to that conjectured Gandhâra prototype of Central-Asian Buddhist painting which can no longer be traced in its original home. It is this which makes it worth while to examine with care even what mere fragments can teach us. They deserve full attention all the more since, I think, we may also trace in them links with Orientalized forms of Hellenistic art further west.
Some general observations as to the technique of these wall paintings may conveniently find their place here. They are all painted in tempera on what the chemical analysis made by Sir Arthur Church has proved to be a thin coating of plaster of Paris, skilfully spread over a backing of loess.' A pale pinkish pigment, derived from ferric oxide, appears to have been distributed over the white plaster of Paris surface, while it was still wet, in order to serve as a ground colour. The presence of size in the various pigments used above this could not be definitely determined, but it seems likely. It should be noted that the method here used of preparing the ground by a thin layer of plaster of Paris tinted with a ferruginous pigment appears to have continued in the Khotan region down to Tang times, as Sir Arthur Church's analysis has proved it also for the ` frescoes ' of Khâdalik.$
I have already given the reasons which prove that the wall decoration must have included at least two friezes encircling the rotunda above the height of the dado I still found in situ. It is to the lower one of these two friezes that we can with considerable probability ascribe the panel M. ni. 003, reproduced with its well-preserved colours in Plate XLII. On account of its somewhat large size and the interest presented by its subject and details, it provides a very convenient introduction to our review of the rescued remains of these paintings. The two pieces now united in the panel, measuring over 3 feet 3 inches in length and over i foot ro inches in height, were discovered in a detached condition at the foot of the dado, below the lunettes iv and v. Their position there, nearest to the wall and behind two other layers of painted plaster fragments (including M. in. 002, 004), as seen in Fig. 127, makes it probable that they had fallen from the part of the wall immediately above the angels. The broad black band which runs along the top of the extant portion of the background, and near which are the remains of a grey one, belonged, as already mentioned, to the triplicate border dividing the two friezes. The method by which the joining of the badly broken pieces of this panel was effected, and which was also employed for the preservation of the other fresco remains from Miran, is fully explained in the note below.9 This method owed
' Cf. below, Appendix D.
8 See below, Appendix D.
9 In order to protect the brittle pieces of painted plaster from further injury and to render their safe fixing and handling possible; it was necessary to replace the friable clay and straw at the back of the smooth surface layer which bears the tempera pigments by a fresh backing of plaster of Paris. This removal of the coarse original backing was absolutely necessary as a preliminary safeguard. Otherwise the salts with which, like all the clay remains of these ruins, it had
become permeated in abundance would have exuded on the coloured surface under the influence of the moisture absorbed from the plaster of Paris that was needed to hold together and strengthen the whole. But without this preparatory measure it would also have been impossible to make the separated surface pieces join accurately or the many serious cracks, etc., close up again. This very delicate operation of first removing the clay backing from the several detached fresco pieces which formed part of the same panel, and then bringing their surface layers into close contact in the correct position, was effected
3 S 2