XXVII. The very awkwardness with which such occurrences of difficulty are met by the earlier builder, marks his dislike of the system, and unwillingness to attract the eye to such arrangements. There is another very clumsy one, in the junction of the upper and sub-arches of the triforium of Salisbury; but it is kept in the shade, and all the prominent junctions are of mouldings like each other, and managed with perfect simplicity. But so soon as the attention of the builders became, as we have just seen, fixed upon the lines of mouldings instead of the enclosed spaces, those lines began to preserve an independent existence wherever they met; and different mouldings were studiously associated, in order to obtain variety of intersectional line. We must, however, do the late builders the justice to note that, in one case, the habit grew out of a feeling of proportion, more refined than that of earlier workmen. It shows itself first in the bases of divided pillars, or arch mouldings, whose smaller shafts had originally bases formed by the continued base of the central, or other larger, columns with which they were grouped; but it being felt, when the eye of the architect became fastidious, that the dimension of moulding which was right for the base of a large shaft, was wrong for that of a small one, each shaft had an independent base; at first, those of the smaller died simply down on that of the larger; but when the vertical sections of both became complicated, the bases of the smaller shafts were considered to exist within those of the larger, and the places of their emergence, on this supposition, were calculated with the utmost nicety, and cut with singular precision; so that an elaborate late base of a divided column, as, for instance, of those in the nave of Abbeville, looks exactly as if its smaller shafts had all been finished to the ground first, each with its complete and intricate base, and then the comprehending base of the central pier had been moulded over them in clay, leaving their points and angles sticking out here and there, like the edges of sharp crystals out of a nodule of earth. The exhibition of technical dexterity in work of this kind is often marvellous, the strangest possible shapes of sections being calculated to a hair's-breadth, and the occurrence of the under and emergent forms being rendered, even in places where they are so slight that they can hardly be detected but by the touch. It is impossible to render a very elaborate example of this kind intelligible, without some fifty measured sections; but fig. 6, Plate IV. is a very interesting and simple one, from the west gate of Rouen. It is part of the base of one of the narrow piers between its principal niches. The square column _k_, having a base with the profile _p r_, is supposed to contain within itself another similar one, set diagonally, and lifted so far above the inclosing one, as that the recessed part of its profile