first sight opposed to all analogy. The sexual elements
of the same flower have become, as already remarked, differentiated in
relation to each other, almost like those of two distinct species.

With respect to the species which, whilst living under their natural
conditions, have their reproductive organs in this peculiar state, we may
conclude that it has been naturally acquired for the sake of effectually
preventing self-fertilisation. The case is closely analous with dimorphic
and trimorphic plants, which can be fully fertilised only by plants belong
to the opposite form, and not, as in the foregoing cases, in differently by
any other plant. Some of these dimorphic plants are completely sterile with
pollen taken from the same plant or from the same {141} form. It is
interesting to observe the graduated series from plants which, when
fertilised by their own pollen, yield the full number of seed, but with the
seedlings a little dwarfed in stature--to plants which when self-fertilised
yield few seeds--to those with yield none--and, lastly, to those in which
the plant's own pollen and stigma act on each other like poison. This
peculiar state of the reproductive organs, when occurring in certain
individuals alone, is evidently abnormal; and as it chiefly affects exotic
plants, or indigenous plants cultivated in pots, we may attribute it to
some change in the conditions of life, acting on the plants themselves or
on their parents. The self-impotent _Passiflora alata_, which recovered its
self-fertility after having been grafted on a distinct stock, shows how
small a change is sufficient to act powerfully on the reproductive system.
The possibility of a plant becoming under culture self-impotent is
interesting as throwing light on the occurrence of this same condition in
natural species. A cultivated plant in this state generally remains so
during its whole life; and from this fact we may infer that the state is
probably congenital.

Koelreuter, however, has described some plants of Verbascum which varied in
this respect even during the same season. As in all the normal cases, and
in many, probably in most, of the abnormal cases, any two self-impotent
plants can reciprocally fertilize each other, we may infer that a very
slight difference in the nature of their sexual elements suffices to give
fertility; but in other instances, as with some Passifloras and the hybrid
Gladioli, a greater degree of differentiation appears to b