have seen, the egg gets rid of half its chromosomes in order to
receive an equal number from a male parent; and thus the fertilized egg
contains chromosomes, and hence hereditary material, from two different
individuals. Now, this sexual reproduction occurs very widely in the
organic world. Among some of the lowest forms of unicellular organisms
it is not known, but in most others some form of such union is
universal. Now, here is plainly an abundant opportunity for congenital
variations; for it is seen that each individual does not come from germ
material _identical with that from which either parent came, but from
some of this material mixed with a similar amount from a different
parent_. Now, the two parents are never exactly alike, and hence the
germ plasm which each contributes to the offspring will not be exactly
alike. The offspring will thus be the result of the unfolding of a bit
of germ plasm which will be different from that from which either of its
parents developed, and these differences will result in _congenital
variations_. Sexual reproduction thus results in congenital variations;
and if congenital variations are necessary for the evolution of the
living machine--and we shall soon see reason for believing that they
are--we find that sexual reproduction is a device adopted for bringing
out such congenital variations.

==Inheritance of Variations.==--The reason why congenital variations are
needed for the evolution of the living machine is clear enough.
Evanescent variations can have no effect upon this machine, for they
would disappear with the individual in which they appeared. In order
that they should have any influence in the process of machine building
they must be permanent ones; or, in other words, they must be inherited
from generation to generation. Only as such variations are transmitted
by heredity can they be added to the structure of the developing
machine. Therefore we must ask whether the variations are inherited.

In regard to the congenital variations there can be no difficulty. The
very fact that they are congenital shows us that they have been produced
by variations in the germ plasm, and as such they must be transmitted,
not only to the next generation, but to all following generations, until
the germ plasm becomes again modified. This germ plasm is handed on from
generation to generation with all its variations, and hence the
variations will be added permanently to the machine. Conge