nic world, impelling the simpler forms to more and more complex
developments. How this law operates, what influences determine the
development of the eggs and germs, and impel them to assume
constantly new forms, I naturally cannot pretend to say; but I can
at least adduce the great analogy of the alternation of
generations. If a _Bipinnaria_, a _Brachialaria_, a _Pluteus_, is
competent to produce the Echinoderm, which is so widely different
from it; if a hydroid polype can produce the higher Medusa; if the
vermiform Trematode 'nurse' can develop within itself the very
unlike _Cercaria_, it will not appear impossible that the egg, or
ciliated embryo, of a sponge, for once, under special conditions,
might become a hydroid polype, or the embryo of a Medusa, an
Echinoderm."

It is obvious, from these extracts, that Professor Koelliker's hypothesis
is based upon the supposed existence of a close analogy between the
phaenomena of Agamogenesis and the production of new species from
pre-existing ones. But is the analogy a real one? We think that it is
not, and, by the hypothesis, cannot be.

For what are the phaenomena of Agamogenesis, stated generally? An
impregnated egg develops into an asexual form, A; this gives rise,
asexually, to a second form or forms, B, more or less different from A.
B may multiply asexually again; in the simpler cases, however, it does
not, but, acquiring sexual characters, produces impregnated eggs from
whence A once more arises.

No case of Agamogenesis is known in which, _when A differs widely from
B_, it is itself capable of sexual propagation. No case whatever is
known in which the progeny of B, by sexual generation, is other than a
reproduction of A.

But if this be a true statement of the nature of the process of
Agamogenesis, how can it enable us to comprehend the production of new
species from already existing ones? Let us suppose Hyaenas to have
preceded Dogs, and to have produced the latter in this way. Then the
Hyaena will represent A, and the Dog, B. The first difficulty that
presents itself is that the Hyaena must be asexual, or the process will
be wholly without analogy in the world of Agamogenesis. But passing over
this difficulty, and supposing a male and female Dog to be produced at
the same time from the Hyaena stock, the progeny of the pair, if the
analogy of the simpler kinds of Agamogenesis[67] is to be followed,
s