§ III. We must therefore have our flat stone in either case; and let _b_, Fig. IV., be the section or side of it, as it is set across the wall. Now, evidently, if by any chance this weight happen to be thrown more on the edges of this stone than the centre, there will be a chance of these edges breaking off. Had we not better, therefore, put another stone, sloped off to the wall, beneath the projecting one, as at _c_. But now our cornice looks somewhat too heavy for the wall; and as the upper stone is evidently of needless thickness, we will thin it somewhat, and we have the form _d_. Now observe: the lower or bevelled stone here at _d_ corresponds to _d_ in the base (Fig. II., page 59). That was the foot of the wall; this is its hand. And the top stone here, which is a constant member of cornices, corresponds to the under stone _c_, in Fig. II., which is a constant member of bases. The reader has no idea at present of the enormous importance of these members; but as we shall have to refer to them perpetually, I must ask him to compare them, and fix their relations well in his mind: and, for convenience, I shall call the bevelled or sloping stone, X, and the upright edged stone, Y. The reader may remember easily which is which; for X is an intersection of two slopes, and may therefore properly mean either of the two sloping stones; and Y is a figure with a perpendicular line and two slopes, and may therefore fitly stand for the upright stone in relation to each of the sloping ones; and as we shall have to say much more about cornices than about bases, let X and Y stand for the stones of the cornice, and Xb and Yb for those of the base, when distinction is needed. [Illustration: Fig. V.] § IV. Now the form at _d_, Fig. IV., is the great root and primal type of all cornices whatsoever. In order to see what forms may be developed from it, let us take its profile a little larger--_a_, Fig. V., with X and Y duly marked. Now this form, being the root of all cornices, may either have to finish the wall and so keep off rain; or, as so often stated, to carry weight. If the former, it is evident that, in its present profile, the rain will run back down the slope of X; and if the latter, that the sharp angle or edge of X, at _k_, may be a little too weak for its work, and run a chance of giving way. To avoid the evil in the first case, suppose we hollow the slope of X inwards, as at _b_; and to avoid it in the second case, suppose we strengthen X by letting it bulge outwards, as at c.