As just indicated, the only places where we may look for evidence with any rational hope of finding it, are places where some local requirement for vessels has arisen in consequence of some local development which the type does not involve. In these cases we find such evidence. Good illustrations occur in those genera of the _Cactaceæ_, which simulate leaves, like _Epiphyllum_ and _Phyllocactus_. A branch of one of these is outlined in Fig. 256. As before explained this is a flattened axis; and the notches along its edges are the seats of the axillary buds. Most of these axillary buds are arrested; but occasionally one of them grows. Now if, taking an _Epiphyllum_-shoot which bears a lateral shoot, we compare the parts of it that are near the aborted axillary buds with the part that is near the developed axillary bud, we find a conspicuous difference. In the neighbourhood of an aborted axillary bud there is no external sign of any internal differentiation; and on holding up the branch against the light, the uniform translucency shows that there is no greater amount of dense tissue near it than in other parts of the succulent mass. But where an axillary bud has developed, a prominent rounded ridge joins the mid-rib of the lateral branch with the mid-rib of the parent branch. In the midst of this rounded ridge an opaque core may be seen. And on cutting through it, this opaque core proves to be full of vascular bundles imbedded in woody deposits. Clearly, these clusters of vessels imply transformations of the tissues, caused by the passage of increased currents of sap. The vessels were not there when the axillary bud was formed; they would not have developed had the axillary bud proved abortive; but they arise as fast as growth of the axillary bud draws the sap along the lines in which they lie. Verification is obtained by examining the internal structures. If longitudinal sections be made through a growing bud of _Opuntia_ or _Cereus_, it will be found that the vessels in course of formation converge towards the point of growth, as they would do if the sap-currents determined their formation; that they are most developed near their place of convergence, which they would be if so produced; and that their terminations in the tissue of the parent shoot are partially-formed lines of irregular elongated cells, like those out of which the vessels of a leaf or bud are developed.