direct consequences of more general laws.

SECTION THIRD.

_An attempt to explain the foregoing laws of geographical distribution,
on the theory of allied species having a common descent._

First let us recall the circumstances most favourable for variation
under domestication, as given in the first chapter--viz. 1st, a change,
or repeated changes, in the conditions to which the organism has been
exposed, continued through several seminal (_i.e._ not by buds or
divisions) generations: 2nd, steady selection of the slight varieties
thus generated with a fixed end in view: 3rd, isolation as perfect as
possible of such selected varieties; that is, the preventing their
crossing with other forms; this latter condition applies to all
terrestrial animals, to most if not all plants and perhaps even to most
(or all) aquatic organisms. It will be convenient here to show the
advantage of isolation in the formation of a new breed, by comparing the
progress of two persons (to neither of whom let time be of any
consequence) endeavouring to select and form some very peculiar new
breed. Let one of these persons work on the vast herds of cattle in the
plains of La Plata{399}, and the other on a small stock of 20 or 30
animals in an island. The latter might have to wait centuries (by the
hypothesis of no importance){400} before he obtained a "sport"
approaching to what he wanted; but when he did and saved the greater
number of its offspring and their offspring again, he might hope that
his whole little stock would be in some degree affected, so that by
continued selection he might gain his end. But on the Pampas, though
the man might get his first approach to his desired form sooner, how
hopeless would it be to attempt, by saving its offspring amongst so many
of the common kind, to affect the whole herd: the effect of this one
peculiar "sport{401}" would be quite lost before he could obtain a
second original sport of the same kind. If, however, he could separate a
small number of cattle, including the offspring of the desirable
"sport," he might hope, like the man on the island, to effect his end.
If there be organic beings of which two individuals _never_ unite, then
simple selection whether on a continent or island would be equally
serviceable to make a new and desirable breed; and this new breed might
be made in surprisingly few years from the great and geometrical powers
of propagation to beat out the old breed; as has happened