§ 4. Now, _a_, Fig. 17, being the general type of a leaf-stalk, Fig. 18 is the general type of the way it expands into and carries its leaf;[1] this figure being the enlargement of a typical section right across any leaf, the dotted lines show the under surface foreshortened. You see I have made one side broader than the other. I mean that. It is typically so. Nature cannot endure two sides of a leaf to be alike. By encouraging one side more than the other, either by giving it more air or light, or perhaps in a chief degree by the mere fact of the moisture necessarily accumulating on the lower edge when it rains, and the other always drying first, she contrives it so, that if the essential form or idea of the leaf be _a_, Fig. 19, the actual form will always be _c_, or an approximate to it; one half being pushed in advance of the other, as at _b_, and all reconciled by soft curvature, _c_. The effort of the leaf to keep itself symmetrical rights it, however, often at the point, so that the insertion of the stalk only makes the inequality manifest. But it follows that the sides of a straight section across the leaf are unequal all the way up, as in my drawing, except at one point. [Illustration: FIG. 20.] § 5. I have represented the two wings of the leaf as slightly convex on the upper surface. This is also on the whole a typical character. I use the expression "wings of the leaf," because supposing we exaggerate the main rib a little, the section will generally resemble a bad painter's type of a bird (_a_, Fig. 20). Sometimes the outer edges curl up, _b_, but an entirely concave form, _c_, is rare. When _b_ is strongly developed, closing well in, the leaf gets a good deal the look of a boat with a keel. [Illustration: FIG. 21.]