architecture is a frame for the sculpture, or the sculpture an ornament of the architecture. If the latter, then the first office of that sculpture is not to represent the things it imitates, but to gather out of them those arrangements of form which shall be pleasing to the eye in their intended places. So soon as agreeable lines and points of shade have been added to the mouldings which were meagre, or to the lights which were unrelieved, the architectural work of the imitation is accomplished; and how far it shall be wrought towards completeness or not, will depend upon its place, and upon other various circumstances. If, in its particular use or position, it is symmetrically arranged, there is, of course, an instant indication of architectural subjection. But symmetry is not abstraction. Leaves may be carved in the most regular order, and yet be meanly imitative; or, on the other hand, they may be thrown wild and loose, and yet be highly architectural in their separate treatment. Nothing can be less symmetrical than the group of leaves which join the two columns in Plate XIII.; yet, since nothing of the leaf character is given but what is necessary for the bare suggestion of its image and the attainment of the lines desired, their treatment is highly abstract. It shows that the workman only wanted so much of the leaf as he supposed good for his architecture, and would allow no more; and how much is to be supposed good, depends, as I have said, much more on place and circumstance than on general laws. I know that this is not usually thought, and that many good architects would insist on abstraction in all cases: the question is so wide and so difficult that I express my opinion upon it most diffidently; but my own feeling is, that a purely abstract manner, like that of our earliest English work, does not afford room for the perfection of beautiful form, and that its severity is wearisome after the eye has been long accustomed to it. I have not done justice to the Salisbury dog-tooth moulding, of which the effect is sketched in fig. 5, Plate X., but I have done more justice to it nevertheless than to the beautiful French one above it; and I do not think that any candid reader would deny that, piquant and spirited as is that from Salisbury, the Rouen moulding is, in every respect, nobler. It will be observed that its symmetry is more complicated, the leafage being divided into double groups of two lobes each, each lobe of different structure. With exquisite feeling, one of these double groups is alternately omitted on the other side of the moulding (not seen in the Plate, but occupying the cavetto of the section), thus giving a playful lightness to the whole; and if the reader will allow for a beauty in the flow of the curved outlines (especially on the angle), of which he cannot in the least judge from my rude drawing, he will not, I think, expect easily to find a nobler instance of decoration adapted to the severest mouldings.