erminated when
man introduced dogs, pigs, and cats into the island, and himself sought
them for food. The fact that such perfectly unprotected creatures survived
in great abundance to a quite recent period in these three islands only,
while there is no evidence of their ever having inhabited any other
countries whatever, is itself almost demonstrative that Mauritius, Bourbon,
and Rodriguez are very ancient but truly oceanic islands. From what we know
of the general similarity of Miocene birds to living genera and families,
it seems clear that the origin of so remarkable a type as the dodos must
date back to early Tertiary times. If we suppose some ancestral
ground-feeding pigeon of large size to have reached the group by means of
intervening islands afterwards submerged, and to have thenceforth remained
to increase and multiply unchecked by the attacks of any more {437}
powerful animals, we can well understand that the wings, being useless,
would in time become almost aborted.[110] It is also not improbable that
this process would be aided by natural selection, because the use of wings
might be absolutely prejudicial to the birds in their new home. Those that
flew up into trees to roost, or tried to cross over the mouths of rivers,
might be blown out to sea and destroyed, especially during the hurricanes
which have probably always more or less devastated the islands; while on
the other hand the more bulky and short-winged individuals, who took to
sleeping on the ground in the forest, would be preserved from such dangers,
and perhaps also from the attacks of birds of prey which may always have
visited the islands. But whether or no this was the mode by which these
singular birds acquired their actual form and structure, it is perfectly
certain that their existence and development depended on complete isolation
and on freedom from the attacks of enemies. We have no single example of
such defenceless birds having ever existed on a continent at any geological
period, whereas analogous though totally distinct forms do exist in New
Zealand, where enemies are equally wanting. On the other hand, every
continent has always produced abundance of carnivora adapted to prey upon
the herbivorous animals inhabiting it at the same period; and we may
therefore be sure that {438} these islands have never formed part of a
continent during any portion of the time when the dodos inhabited them.

It is a remarkable thing that an ornithologist