While considering leguminous plants, a few words concerning stipels, which are so characteristic of the family, would be in place. They have been denominated as " the stipules of leaflets," but I am convinced that they have no connection with stipules whatever, but that they represent rudimentary leaflets which have their origin in a tendency to increased compounding. The habit has become so fixed in the Leguminosie that evidence of its ori gin is seldom met with. I have however seen, in Lespedeza capi tata Michx., one of the earliest leaves with the terminal leaflet only developed and the two lateral ones represented by stipels.
I have found more light on the question in other families where the same tendency to increased compounding often occurs. In Sanguisorba Canadensis L. (fig. 47) for example, very vigorous plants sometimes show rudimentary leaflets, more developed in deed than typical stipels, but in the same position. Their char acter as leaflets of secondary rank is evinced by their occasional removal to a little distance from the primary petiole. A more striking case is that of Sumbucus Canadensis L. In this species the leaves of young shoots springing up where the bushes have been cleared away are frequently partially bicompound and there are all gradations between the ordinary pinnate form and the bipinnate condition (figs. 48-50). In this case it is remarkable that the first appearance of the secondary leaflet is in the shape of a small body with both the form and position of a stipel, with the same small supporting vein and differing only in greater thick ness. These facts seem to give evidence sufficiently conclusive that stipels are in reality rudimentary leaflets. That their de velopment is not confined to the Leguminosm is farther shown by their characteristic occurrence in Staphylea trifolia L.
Another frequent foliar variation among the Leguminoste is the development of the phyllodium, which might be thought to have some connection with stipules, but the presence of both together in some genera disproves the idea.* The stipules in the Legu minosm often take the form of spines which serve for the general protection of the plant. We have an example in the well known Robinia Pseudacacia L. (fig. 51). In some of the tropical Aca cias, as for example A. spadicigera C. 47 S. (fig. 52), they take the form of enormous hollow horns which are appropriated as homes by some species of ants.t Sambucus Canadensis L. presents another remarkable char acter. The leaves of the vernal shoots from subterranean buds are furnished with stipules of the same form and in the same po sition as those of Sambucus Ebulus L., but smaller. There are
four of them at each node, they are ovate or nearly orbicular in form, small, rather fleshy and persist but a short time. Each is supplied with a small vascular bundle, originating as a branch of the nodal girdle which connects the leaf-traces. These facts give evidence of the close relationship of these two species of Sambucus, and of the characteristic presence of stipules in the ancestral form. In Sambucus Ebulus L., they are still typically developed, but in our species have become so far vestigial as to appear only in connection with the early leaves of shoots from subterranean buds, an additional evidence of the importance of the leaf-forms successively developed from such buds, in their bearing on the evolutionary development of modern adult forms.
If now we turn to the family of the Rosacem we shall find many illustrative examples of the same facts as those born out in the case of Baptisia tinctoria (L.) R. Br. But it frequently happens that basal degeneration does not take place or is only partial, re suiting in the adnate stipules characteristic of so many genera and species of the family. Agrimonia striate Michx., in the develop ment of its subterranean buds in the spring, presents an excellent series of embryonic leaf-forms. The lower ones are all simple sheathing scales completely surrounding the stein at their inser tion. Not until the eleventh leaf (fig. 53), which is three-toothed at the apex, does the differentiation of parts begin. The central tooth is the beginning of the blade with its petiole ; the lateral portions with their tips now free are the stipules. To say that they are "adnate" indicates only that they retain their primitive connection with the central-basal portion. In the twelfth leaf (fig. 54), there has been some basal degeneration, as shown by the lower point at which the three main bundles of the leaf converge and the lower position of the zigzag plexus of the stipular veins. The free tips, on the other hand, have increased in size and a small blade supported by a petiole is present in consequence of the de velopment of the central tooth. The fifteenth leaf (fig. 55) shows a stronger development of all the parts, and a branch of the main stipular bundle is seen to pass up the petiole. The adult form is attained in the seventeenth leaf (fig. 56). In it some further basal degeneration has taken place, but the adnation of the sti pules is still very prominent.