§ XXV. These, then, being the points most noticeable in the spirit both of the designs and the chiselling, we have now to return to the question proposed in § XIII., and observe the modifications of form of profile which resulted from the changing contours of the leafage; for up to § XIII., we had, as usual, considered the possible conditions of form in the abstract;--the modes in which they have been derived from each other in actual practice require to be followed in their turn. How the Greek Doric or Greek ogee cornices were invented is not easy to determine, and, fortunately, is little to our present purpose; for the mediæval ogee cornices have an independent development of their own, from the first type of the concave cornice _a_ in Plate XV. [Illustration: Fig. LXIII.] § XXVI. That cornice occurs, in the simplest work, perfectly pure, but in finished work it was quickly felt that there was a meagreness in its junction with the wall beneath it, where it was set as here at _a_, Fig. LXIII., which could only be conquered by concealing such junction in a bar of shadow. There were two ways of getting this bar: one by a projecting roll at the foot of the cornice (_b_, Fig. LXIII.), the other by slipping the whole cornice a little forward (_c_. Fig. LXIII.). From these two methods arise two groups of cornices and capitals, which we shall pursue in succession. § XXVII. First group. With the roll at the base (_b_, Fig. LXIII.). The chain of its succession is represented from 1 to 6, in Plate XV.: 1 and 2 are the steps already gained, as in Fig. LXIII.; and in them the profile of cornice used is _a_ of Plate XV., or a refined condition of _b_ of Fig. V., p. 69, above. Now, keeping the same refined profile, substitute the condition of it, _f_ of Fig. V. (and there accounted for), above the roll here, and you have 3, Plate XV. This superadded abacus was instantly felt to be harsh in its projecting angle; but you know what to do with an angle when it is harsh. Use your simplest chamfer on it (_a_ or _b_, Fig. LIII., page 287, above), but on the visible side only, and you have fig. 4, Plate XV. (the top stone being made deeper that you may have room to chamfer it). Now this fig. 4 is the profile of Lombardic and Venetian early capitals and cornices, by tens of thousands; and it continues into the late Venetian Gothic, with this only difference, that as times advances, the vertical line at the top of the original cornice begins to slope outwards, and through a series of years rises like the hazel wand in the hand of a diviner:--but how slowly! a stone dial which marches but 45 degrees in three centuries, and through the intermediate condition 5 arrives at 6, and so stays.