off from view by the salience of the cornice; and it was apparently in part for the same reason that the attic was interposed in the nave. The motive is commendable. The effect of vaulting rising directly from a salient cornice Wren may justly have felt to be a bad one, but to avoid it while using classic details necessitates these strange inconsistencies.
Fig. 138.—Crossing pier and impost, St. Paul's.
Among numerous other aberrations of this pseudo-classic scheme is the treatment of the segmental archivolts of the small half domes that open out of the oblique sides of the great octagon at the crossing. The orders of the crossing piers have complete entablatures (Fig. 138), and the archivolts in question are in two parts answering to the frieze and cornice of these entablatures, which they intersect in the awkward manner shown in the figure. To have mitred the cornice of the archivolt to that of the order would have left the pilaster beneath with an incomplete entablature, and the architect preferred to run the cornice through the archivolt in this unsightly way. Such were some of the further makeshifts to which the designers of the Renaissance had to resort in their efforts to apply the classic orders to uses for which they were not adapted. But all such aberrations in the use of classic elements are superficial and open. A more radical violation of architectural veracity is found in the manner in which the buttress system is concealed. The thrusts of the nave vaulting are met by a series of flying buttresses carried over the aisle roof in Gothic fashion (Fig. 139). But it would not do to have flying buttresses appear in an ostensibly classic system, and Wren accordingly hid them from sight by a screen wall made to look like an upper story in the general view of the exterior. It is not until one mounts to the terrace of
