§ XXVII. The transition from the actual to the apparent cluster, in the Gothic, is a question of the most curious interest; I have thrown together the shaft sections in Plate II. to illustrate it, and exemplify what has been generally stated above.[43] [Illustration: Plate II. PLANS OF PIERS.] 1. The earliest, the most frequent, perhaps the most beautiful of all the groups, is also the simplest; the two shafts arranged as at _b_ or _c_, (Fig. XIV.) above, bearing an oblong mass, and substituted for the still earlier structure _a_, Fig. XIV. In Plate XVII. (Chap. XXVII.) are three examples of the transition: the one on the left, at the top, is the earliest single-shafted arrangement, constant in the rough Romanesque windows; a huge hammer-shaped capital being employed to sustain the thickness of the wall. It was rapidly superseded by the double shaft, as on the right of it; a very early example from the cloisters of the Duomo, Verona. Beneath, is a most elaborate and perfect one from St. Zeno of Verona, where the group is twice complicated, two shafts being used, both with quatrefoil sections. The plain double shaft, however, is by far the most frequent, both in the Northern and Southern Gothic, but for the most part early; it is very frequent in cloisters, and in the singular one of St. Michael's Mount, Normandy, a small pseudo-arcade runs along between the pairs of shafts, a miniature aisle. The group is employed on a magnificent scale, but ill proportioned, for the main piers of the apse of the cathedral of Coutances, its purpose being to conceal one shaft behind the other, and make it appear to the spectator from the nave as if the apse were sustained by single shafts, of inordinate slenderness. The attempt is ill-judged, and the result unsatisfactory. [Illustration: Fig. XVII.] § XXVIII. 2. When these pairs of shafts come near each other, as frequently at the turnings of angles (Fig. XVII.), the quadruple group results, _b_ 2, Fig. XIV., of which the Lombardic sculptors were excessively fond, usually tying the shafts together in their centre, in a lover's knot. They thus occur in Plate V., from the Broletto of Como; at the angle of St. Michele of Lucca, Plate XXI.; and in the balustrade of St. Mark's. This is a group, however, which I have never seen used on a large scale.[44]