Let us experiment a little in forming our own classification of a
few vertebrates. We see a bat flying through the air. We mistake it
for a bird. But a glance at it shows that it is a mammal. It is
covered with hair. It has fore and hind legs. Its wings are
membranes stretched between the fingers and along the sides of the
body. It has teeth. It suckles its young. In all these respects it
differs from birds. It differs from mammals only in its wings. But
we remember that flying squirrels have a membrane stretching along
the sides of the body and serving as a parachute, though not as
wings. We naturally consider the wings as a sort of after-thought
superinduced on the mammalian structure. We do not hesitate to call
it a mammal.

The whale makes us more trouble; it certainly looks remarkably like
a fish. But the fin of its tail is horizontal, not vertical. Its
front flippers differ altogether from the corresponding fins of
fish; their bones are the same as those occurring in the forelegs of
mammals, only shorter and more crowded together. Later we find that
it has lungs, and a heart with four chambers instead of only two, as
in fish. The vertebræ of its backbone are not biconcave, but flat in
front and behind. And, finally, we discover that it suckles its
young. It, too, is in all its deep-seated characteristics a mammal.
It is fish-like only in characteristics which it might easily have
acquired in adaptation to its aquatic life. And there are other
aquatic mammals, like the seals, in which these characteristics are
much less marked. Their adaptation has evidently not gone so far.

Now the first attempts resulted in artificial classifications, much
like our grouping of bats with birds and whales with fish. All