EASTLAKE’S HISTORY OF OIL-PAINTING 281

unless exceedingly thin; and even on panel is liable to crack and detach itself, unless it be carefully guarded against damp. The precautions of Van Eyck against this danger, as well as against the warping of his panel, are remarkable instances of his regard to points apparently trivial:-

“In large altar-pieces, necessarily composed of many pieces, it may be often remarked that each separate plank has become slightly convex in front: this is particularly observable in the picture of the Transfiguration by Raphael. The heat of candles on altars is supposed to have been the cause of this not uncommon defect; but heat, if considerable, would rather produce the contrary appearance. It would seem that the layer of paint, with its substratum, slightly operates to prevent the wood from contracting or becoming concave on that side; it might therefore be concluded that a similar protection at the back, by equalizing the conditions, would tend to keep the wood flat. The oak panel on which the picture by Van Eyck in the National Gallery is painted is protected at the back by a composition of gesso, size, and tow, over which a coat of black oil-paint was passed. This, whether added when the picture was executed or subsequently, has tended to preserve the wood (which is not at all worm-eaten), and perhaps to prevent its warping.”-Ib., pp. 373, 374.

On the white ground, scraped, when it was perfectly dry, till it was “as white as milk and as smooth as ivory” (Cennini), the outline of the picture was drawn, and its light and shade expressed, usually with the pen, with all possible care; and over this outline a coating of size was applied in order to render the gesso ground non-absorbent. The establishment of this fact is of the greatest importance, for the whole question of the true function and use of the gesso ground hangs upon it. That use has been supposed by all previous writers on the technical processes of painting to be, by absorbing the oil, to remove in some degree the cause of yellowness in the colours. Had this been so, the ground itself would have lost its brilliancy, and it would have followed that a dark ground, equally absorbent, would have answered the purpose as well. But the evidence adduced by Mr. Eastlake on this subject is conclusive:-

“Pictures are sometimes transferred from panel to cloth. The front being secured by smooth paper or linen, the picture is laid on its face, and the wood is gradually planed and scraped away. At last the ground

[Version 0.04: March 2008]