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 noble instance of decoration adapted to the severest mouldings.
Now it will be observed, that there is in its treatment a high degree of abstraction, though not so conventional as that of Salisbury: that is to say, the leaves have little more than their flow and outline represented; they are hardly undercut, but their edges are connected by a gentle and most studied curve with the stone behind; they have no serrations, no veinings, no rib or stalk on the angle, only an incision gracefully made towards their extremities, indi-
