a similar manner many plants, when
placed under unnatural conditions, fail to produce seed, but can readily be
propagated by buds. We shall presently see that pangenesis agrees well with
the strong tendency to reversion exhibited by all crossed animals and
plants.

It was shown in the discussion on graft-hybrids that there is some reason
to believe that portions of cellular tissue taken from distinct plants
become so intimately united, as afterwards occasionally to produce crossed
or hybridised buds. If this fact were fully established, it would, by the
aid of our hypothesis, connect gemmation and sexual reproduction in the
closest manner.

Abundant evidence has been advanced proving that pollen taken from one
species or variety and applied to the stigma of another sometimes directly
affects the tissues of the mother-plant. It is probable that this occurs
with many plants during fertilisation, but can only be detected when
distinct forms are crossed. On any ordinary theory of reproduction this is
a most anomalous circumstance, for the pollen-grains are manifestly adapted
to act on the ovule, but in these cases they act on the colour, texture,
and form of the coats of the seeds, on the ovarium itself, which is a
modified leaf, and even on the calyx and upper part of the flower-peduncle.
In accordance with the hypothesis of pangenesis pollen includes gemmules,
derived from every part of the organisation, which diffuse themselves and
multiply by self-division; hence it is not surprising that gemmules within
the pollen, which are derived from the parts near the reproductive organs,
should sometimes be able to affect the same parts, whilst still undergoing
development, in the mother-plant. {388}

As, during all the stages of development, the tissues of plants consist of
cells, and as new cells are not known to be formed between, or
independently of, pre-existing cells, we must conclude that the gemmules
derived from the foreign pollen do not become developed merely in contact
with pre-existing cells, but actually penetrate the nascent cells of the
mother-plant. This process may be compared with the ordinary act of
fertilisation, during which the contents of the pollen-tubes penetrate the
closed embryonic sack within the ovule, and determine the development of
the embryo. According to this view, the cells of the mother-plant may
almost literally be said to be fertilised by the gemmules derived from the
foreign pollen. With