ammalia
and amphibia, while their reptiles, when they possess any, do not exhibit
indications of extreme isolation and antiquity. Their birds and insects
present just that amount of specialisation and diversity from continental
forms which may be well explained by the known means of dispersal acting
through long periods; their land shells indicate greater isolation, owing
to their admittedly less effective means of conveyance across the ocean;
while their plants show most clearly the effects of those changes of
conditions which we have reason to believe have occurred during the
Tertiary epoch, and preserve to us in highly specialised and archaic forms
some record of the primeval immigration by which the islands were
originally {330} clothed with vegetation. But in every case the series of
forms of life in these islands is scanty and imperfect as compared with far
less favourable continental areas, and no one of them presents such an
assemblage of animals or plants as we always find in an island which we
know has once formed part of a continent.

It is still more important to note that none of these oceanic archipelagoes
present us with a single type which we may suppose to have been preserved
from Mesozoic times; and this fact, taken in connection with the volcanic
or coralline origin of all of them, powerfully enforces the conclusion at
which we have arrived in the earlier portion of this volume, that during
the whole period of geologic time as indicated by the fossiliferous rocks,
our continents and oceans have, speaking broadly, been permanent features
of our earth's surface. For had it been otherwise--had sea and land changed
place repeatedly as was once supposed--had our deepest oceans been the seat
of great continents while the site of our present continents was occupied
by an oceanic abyss--is it possible to imagine that no fragments of such
continents would remain in the present oceans, bringing down to us some of
their ancient forms of life preserved with but little change? The
correlative facts, that the islands of our great oceans are all volcanic
(or coralline built probably upon degraded volcanic islands or extinct
submarine volcanoes), and that their productions are all more or less
clearly related to the existing inhabitants of the nearest continents, are
hardly consistent with any other theory than the permanence of our oceanic
and continental areas.

We may here refer to the one apparent exception, whic