arrow builds very
differently when its nest is in a tree or in a hole, and the
golden-crested wren sometimes suspends its nest below and sometimes
places it _on_ the branches of trees.

{284} A short discussion of a similar kind occurs in the _Origin_,
Ed. i. p. 211, vi. p. 324.

{285} This sentence agrees with the MS., but is clearly in need of
correction.

_Principles of Selection applicable to instincts._

As the instincts of a species are fully as important to its preservation
and multiplication as its corporeal structure, it is evident that if
there be the slightest congenital differences in the instincts and
habits, or if certain individuals during their lives are induced or
compelled to vary their habits, and if such differences are in the
smallest degree more favourable, under slightly modified external
conditions, to their preservation, such individuals must in the long run
have a better _chance_ of being preserved and of multiplying{286}. If
this be admitted, a series of small changes may, as in the case of
corporeal structure, work great changes in the mental powers, habits and
instincts of any species.

{286} This corresponds to _Origin_, Ed. i. p. 212, vi. p. 325.

_Difficulties in the acquirement of complex instincts by Selection._

Every one will at first be inclined to explain (as I did for a long
time) that many of the more complicated and wonderful instincts could
not be acquired in the manner here supposed{287}. The Second Part of
this work is devoted to the general consideration of how far the general
economy of nature justifies or opposes the belief that related species
and genera are descended from common stocks; but we may here consider
whether the instincts of animals offer such a _prima facie_ case of
impossibility of gradual acquirement, as to justify the rejection of any
such theory, however strongly it may be supported by other facts. I beg
to repeat that I wish here to consider not the _probability_ but the
_possibility_ of complicated instincts having been acquired by the slow
and long-continued selection of very slight (either congenital or
produced by habit) modifications of foregoing simpler instincts; each
modification being as useful and necessary, to the species practising
it, as the most complicated kind.

{287} This discussion is interesting in differing from the
corresponding section of the _Origin_, Ed. i. p. 216, vi. p. 330,
t