tally from mine
in the belief that the primary germ-cell was formed within the ovarium of
the female and was fertilised by the male. My gemmules are supposed to be
formed, quite independently of sexual concourse, by each separate cell or
unit throughout the body, and to be merely aggregated within the
reproductive organs.

Lastly, Mr. Herbert Spencer ('Principles of Biology,' vol. i., 1863-4,
chaps. iv. and viii.) has discussed at considerable length what he
designates as physiological units. These agree with my gemmules in being
supposed to multiply and to be transmitted from parent to child; the sexual
elements are supposed to serve merely as their vehicles; they are the
efficient agents in all the forms of reproduction and in the repairs of
injuries; they account for inheritance, but they are not brought to bear on
reversion or atavism, and this is unintelligible to me; they are supposed
to possess polarity, or, as I call it, affinity; and apparently they are
believed to be derived from each separate part of the whole body. But
gemmules differ from Mr. Spencer's physiological units, inasmuch as a
certain number, or mass of them, are, as we shall see, requisite for the
development of each cell or part. Nevertheless I should have concluded that
Mr. Spencer's views were fundamentally the same with mine, had it not been
for several passages which, as far as I understand them, indicate something
quite different. I will quote some of these passages from pp. 254-256. "In
the fertilised germ we have two groups of physiological units, slightly
different in their structures."... "It is not obvious that change in the
form of the part, caused by changed action, involves such change in the
physiological units throughout the organism, that these, when groups of
them are thrown off in the shape of reproductive centres, will unfold into
organisms that have this part similarly changed in form. Indeed, when
treating of Adaptation, we saw that an organ modified by increase or
decrease of function can but slowly so react on the system at large as to
bring about those correlative changes required to produce a new
equilibrium; and yet only when such new equilibrium has been established,
can we expect it to be _fully_ expressed in the modified physiological
units of which the organism is built--only then can we count on a complete
transfer of the modification to descendants."... "That the change in the
offspring must, other things equal, be