as shown in the First Part, at least in regard to animals, that the new
beings which are thus at any period asexually generated do not retrograde
in development--that is, they do not pass through those earlier stages,
through which the fertilised germ of the same animal has to pass; and an
explanation of this fact was attempted as far as the final or teleological
cause is concerned. We can likewise understand the proximate cause, if we
assume, and the assumption is far from improbable, that buds, like
chopped-up fragments of a hydra, are formed of tissue which has already
passed through several of the earlier stages of development; for in this
case their component cells or units would not unite with the gemmules
derived from the earlier-formed cells, but only with those which came next
in the order of development. On the other hand, we must believe that, in
the sexual elements, or probably in the female alone, gemmules of certain
primordial cells are present; and these, as soon as their development
commences, unite in due succession with the gemmules of every part of the
body, from the first to the last period of life.

The principle of the independent formation of each part, in {391} so far as
its development depends on the union of the proper gemmules with certain
nascent cells, together with the superabundance of the gemmules derived
from both parents and self-multiplied, throws light on a widely different
group of facts, which on any ordinary view of development appears very
strange. I allude to organs which are abnormally multiplied or transposed.
Thus gold-fish often have supernumerary fins placed on various parts of
their bodies. We have seen that, when the tail of a lizard is broken off, a
double tail is sometimes reproduced, and when the foot of the salamander is
divided longitudinally, additional digits are occasionally formed. When
frogs, toads, &c., are born with their limbs doubled, as sometimes occurs,
the doubling, as Gervais remarks,[917] cannot be due to the complete fusion
of two embryos, with the exception of the limbs, for the larvae are
limbless. The same argument is applicable[918] to certain insects produced
with multiple legs or antennae, for these are metamorphosed from apodal or
antennaeless larvae. Alphonse Milne-Edwards[919] has described the curious
case of a crustacean in which one eye-peduncle supported, instead of a
complete eye, only an imperfect cornea, out of the centre of which a
portio