explains much that hitherto has found no explanation, and so shall have, in a fashion, proved it retrospectively by showing how much it would explain if it were true.
A most remarkable confirmation of my law may be found in the vegetable kingdom, in a group of facts hitherto regarded as isolated and to be so strange as to have no parallel. Every botanist must have guessed already that I have in mind the phenomena of heterostylism, first discovered by Persoon, then described by Darwin and named by Hilde brand. Many Dicotyledons, and a few Monocotyledons, for instance, species of Primulaceæ and Geraneaceæ and many Rubiaceæ, phanerogams in the flowers of which both the pollen and the stigma are functional, although only in cross-fertilisation, so that the flowers are hermaphrodite in structure ture but unisexual physiologically, display the peculiarity that in different individuals the stamens and the stigma have different lengths. The individuals, all the flowers of which have long styles and therefore high stigmas and short anthers, are, in my judgment, the more female, whilst the individuals with short styles and long anthers are more male. In addition to such dimorphic plants, there are also trimorphic plants, such as Lythrum salicaria, in which the sexual organs display three forms differing in length. There are not only long-styled and short-styled forms, but flowers with styles of a medium length.
Although only dimorphism and trimorphism have been recognised in the books, these conditions do not exhaust the actual complexities of structure. Darwin himself pointed out that if small differences were taken into account, no less than five different situations of the anthers could be distinguished. Alongside such plain cases of discontinuity, of the separation of the different degrees of maleness and femaleness in plainly distinct individuals, there are also cases in which the different degrees grade into each other without breaks in the series. There are analogous cases of discontinuity in the animal kingdom, although they have always been thought of as unique and isolated phenomena, as the parallel with heterostylism had not been suggested. In
