. American coast, may in one sense be considered as
intimately connected with it; for it is certain that formerly many
icebergs loaded with boulders were stranded on its southern coast, and
the old canoes which are occasionally now stranded, show that the
currents still set from Tierra del Fuego. This fact, however, does not
explain the presence of the _Canis antarcticus_ on the Falkland Islands,
unless we suppose that it formerly lived on the mainland and became
extinct there, whilst it survived on these islands, to which it was
borne (as happens with its northern congener, the common wolf) on an
iceberg, but this fact removes the anomaly of an island, in appearance
effectually separated from other land, having its own species of
quadruped, and makes the case like that of Java and Sumatra, each having
their own rhinoceros.

{385} The comparison between New Zealand and the Cape is given in
the _Origin_, Ed. i. p. 389, vi. p. 542.

{386} In a corresponding discussion in the _Origin_, Ed. i. p. 393,
vi. p. 546, stress is laid on the distribution of Batrachians not
of reptiles.

{387} The whole argument is given--more briefly than here--in the
_Origin_, Ed. i. p. 394, vi. p. 547.

{388} See _Origin_, Ed i. p. 393, vi. p. 547. The discussion is
much fuller in the present Essay.

Before summing up all the facts given in this section on the present
condition of organic beings, and endeavouring to see how far they admit
of explanation, it will be convenient to state all such facts in the
past geographical distribution of extinct beings as seem anyway to
concern the theory of descent.

SECTION SECOND.

_Geographical distribution of extinct organisms._

I have stated that if the land of the entire world be divided into (we
will say) three sections, according to the amount of difference of the
terrestrial mammifers inhabiting them, we shall have three unequal
divisions of (1st) Australia and its dependent islands, (2nd) South
America, (3rd) Europe, Asia and Africa. If we now look to the mammifers
which inhabited these three divisions during the later Tertiary periods,
we shall find them almost as distinct as at the present day, and
intimately related in each division to the existing forms in that
division{389}. This is wonderfully the case with the several fossil
Marsupial genera in the caverns of New South Wales and even more
wonderfully so in South America, where we hav