from feeling, and that it was because they did so, that there is this marvellous life, changefulness, and subtlety running through their every arrangement; and that we reason upon the lovely building as we should upon some fair growth of the trees of the earth, that know not their own beauty.
XVII. Perhaps, however, a stranger instance than any I have yet given, of the daring variation of pretended symmetry, is found in the front of the Cathedral of Bayeux. It consists of five arches with steep pediments, the outermost filled, the three central with doors; and they appear, at first, to diminish in regular proportion from the principal one in the centre. The two lateral doors are very curiously managed. The tympana of their arches are filled with bas-reliefs in four tiers; in the lowest tier there is in each a little temple or gate containing the principal figure (in that on the right, it is the gate of Hades with Lucifer). This little temple is carried, like a capital, by an isolated shaft which divides the whole arch at about ⅔ of its breadth, the larger portion outmost; and in that larger portion is the inner entrance door. This exact correspondence, in the treatment of both gates, might lead us to expect a correspondence in dimension. Not at all. The small inner northern entrance measures, in English feet and inches, 4 ft. 7 in. from jamb to jamb, and the southern, 5 ft. exactly. Five inches in five feet is a considerable variation. The outer northern porch measures, from face shaft to face shaft, 13 ft. 11 in., and the southern, 14 ft. 6 in.; giving a difference of 7 in. on 14½ ft. There are also variations in the pediment decorations not less extraordinary.
XVIII. I imagine I have given instances enough, though I could multiply them indefinitely, to prove that these variations are not mere blunders, nor carelessnesses, but the result of a fixed scorn, if not dislike, of accuracy in measurements; and, in most cases, I believe, of a determined resolution to work out an effective sym-
