and to some
extent outside the scope of the theory.

But they would have to face a second and more important deduction from
their views, namely, that the higher animals should repeat at every
stage of their development the whole organisation of some lower animal,
and not merely agree with them in isolated details of structure. The
deduction is, however, not borne out by the facts. The embryo of a
mammal resembles in many points, at different stages of its development,
the adult state of a fish; it has gill-slits and complete aortic arches,
a two-chambered heart, and so on. But at no time does it combine all the
essential characters of a fish; nor has it ever the tail of a fish, nor
the fins, nor the shape. Any recapitulation there may be is a
recapitulation of single organs, there is never a repetition of the
complete organisation of a fish. This is indeed the fundamental
criticism of the theory of parallelism; and if it applies even within
the limits of the vertebrate phylum, so much the more does it apply to
comparisons between embryonic Vertebrates and adult Invertebrates.

There are also some lesser arguments which might be urged against the
theory of parallelism. If the theory were strictly true, no state which
is permanent in a higher animal could be passed through by an animal
lower in the scale. But birds, which are lower in the scale than
mammals, pass through a stage in which they resemble mammals in certain
respects much more than they do when adult, for in an embryonic
condition they agree with mammals in having no feathers, no air sacs, no
pneumatic sacs in the bones, no beak. Their brain also resembles that of
mammals more in an earlier stage than it does later. So, too, myriapods
and hydrachnids have at birth three pairs of feet, and resemble at this
stage adult insects, which form a higher class.

Again, were the analogy between the development of the individual and
the evolution of the _Echelle des etres_ complete, organs and
organ-systems ought to develop in the individual in the order in which
they appear in the scale of beings. But this is not always the case. In
fish the hinder extremity develops only its terminal joint, while in the
embryos of higher animals the basal joint is the first to appear.

Another consequence one would expect to find realised, were the theory
of parallelism correct, is the late appearance in development of parts
which are confined to the higher animals. In the developme