vals of the windows of Merton College, Oxford, where the glass is advanced about two inches from the centre of the tracery bar (that in the larger spaces being in the middle, as usual,) in order to prevent the depth of shadow from farther diminishing the apparent interval. Much of the lightness of the effect of the traceries is owing to this seemingly unimportant arrangement. But, generally speaking, glass spoils all traceries; and it is much to be wished that it should be kept well within them, when it cannot be dispensed with, and that the most careful and beautiful designs should be reserved for situations where no glass would be needed.
XX. The method of decoration by shadow was, as far as we have hitherto traced it, common to the northern and southern Gothic. But in the carrying out of the system they instantly diverged. Having marble at his command, and classical decoration in his sight, the southern architect was able to carve the intermediate spaces with exquisite leafage, or to vary his wall surface with inlaid stones. The northern architect, neither knew the ancient work, nor possessed the delicate material; and he had no resource but to cover his walls with holes, cut into foiled shapes like those of the windows. This he did, often with great clumsiness, but always, observe, depending on the shadows for effect. Where the wall was thick and could not be cut through, and the foilings were large, those; shadows did not fill the entire space; but the form was, nevertheless, drawn on the eye by means of them, and when it was possible, they were cut clear through, as in raised screens of pediment, like those of the west front of Bayeux; cut so deep in every case, as to secure, in all but a direct low front light, great breadth of shadow.
The spandril, given at the top of Plate VII, is from the south-western entrance of the Cathedral of Lisieux; one of the most quaint and interesting doors in Normandy, probably soon to be lost for ever, by the con-
