same in profile, with its (hypothetical)
scapulo-coracoid attached.]

As an illustration both of this and of other points which have been
mentioned, I may draw attention to what seems to me a particularly
suggestive case. So-called soldier-or hermit-crabs, are crabs which have
adopted the habit of appropriating the empty shells of mollusks. In
association with this peculiar habit, the structure of these animals
differs very greatly from that of all other crabs. In particular, the
hinder part of the body, which occupies the mollusk-shell, and which
therefore has ceased to require any hard covering of its own, has been
suffered to lose its calcareous integument, and presents a soft fleshy
character, quite unlike that of the more exposed parts of the animal.
Moreover, this soft fleshy part of the creature is specially adapted to
the particular requirements of the creature by having its lateral
appendages--i. e. appendages which in other crustacea perform the
function of legs--modified so as to act as claspers to the inside of the
mollusk-shell; while the tail-end of the part in question is twisted
into the form of a spiral, which fits into the spiral of the
mollusk-shell. Now, in Keeling Island there is a large kind of crab
called _Birgus latro_, which lives upon land and there feeds upon
cocoa-nuts. The whole structure of this crab, it seems to me,
unmistakeably resembles the structure of a hermit-crab (see drawings on
the next page, Fig. 7). Yet this crab neither lives in the shell of a
mollusk, nor is the hinder part of its body in the soft and fleshy
condition just described: on the contrary, it is covered with a hard
integument like all the other parts of the animal. Consequently, I think
we may infer that the ancestors of _Birgus_ were hermit-crabs living in
mollusk-shells; but that their descendants gradually relinquished this
habit as they gradually became more and more terrestrial, while,
concurrently with these changes in habit, the originally soft posterior
parts acquired a hard protective covering to take the place of that
which was formerly supplied by the mollusk-shell. So that, if so, we now
have, within the limits of a single organism, evidence of a whole series
of morphological changes in the past history of its species. First,
there must have been the great change from an ordinary crab to a
hermit-crab in all the respects previously pointed out. Next, there must
have been the change back again from a