implies also limbs fit for seizing and holding prey; therefore implies
claws, a certain structure of the leg-bones, a certain form of
shoulder-blade. Summing up he says, that "the claw, the scapula, the
condyle, the femur, and all the other bones, taken separately, will give
the tooth or one another; and by commencing with any one, he who had a
rational conception of the laws of the organic economy, could
reconstruct the whole animal."

It will be seen that the method of restoration here contended for, is
based on the alleged physiological necessity of the connexion between
these several peculiarities. The argument used is, not that a scapula of
a certain shape may be recognized as having belonged to a carnivorous
mammal because we always find that carnivorous mammals _do_ possess such
scapulas; but the argument is that they _must_ possess them, because
carnivorous habits would be impossible without them. And in the above
quotation Cuvier asserts that the necessary correlation which he
considers so obvious in these cases, exists throughout the system:
admitting, however, that in consequence of our limited knowledge of
physiology we are unable in many cases to trace this necessary
correlation, and are obliged to base our conclusions upon observed
coexistences, of which we do not understand the reason, but which we
find invariable.

Now Professor Huxley has recently shown that, in the first place, this
empirical method, which Cuvier introduces as quite subordinate, and to
be used only in aid of the rational method, is really the method which
Cuvier habitually employed--the so-called rational method remaining
practically a dead letter; and, in the second place, he has shown that
Cuvier himself has in several places so far admitted the inapplicability
of the rational method, as virtually to surrender it as a method. But
more than this, Professor Huxley contends that the alleged necessary
correlation is not true. Quite admitting the physiological dependence of
parts on each other, he denies that it is a dependence of a kind which
could not be otherwise. "Thus the teeth of a lion and the stomach of
the animal are in such relation that the one is fitted to digest the
food which the other can tear, they are physiologically correlated; but
we have no reason for affirming this to be a necessary physiological
correlation, in the sense that no other could equally fit its possessor
for living on recent flesh. The number and form