ence of
considerable numbers of birds in which both sexes are similarly and
brilliantly coloured,--in some cases as brilliantly as the males of many
of the groups above referred to. Such are the extensive families of the
kingfishers, the woodpeckers, the toucans, the parrots, the turacos, the
hangnests, the starlings, and many other smaller groups, all the species
of which are conspicuously or brilliantly coloured, while in all of them
the females are either coloured exactly like the males, or, when
differently coloured, are equally conspicuous. When searching for some
cause for this singular apparent exception to the rule of female
protective colouring, I came upon a fact which beautifully explains it;
for in all these cases, without exception, the species either nests in
holes in the ground or in trees, or builds a domed or covered nest, so
as completely to conceal the sitting-bird. We have here a case exactly
parallel to that of the butterflies protected by distastefulness, whose
females are either exactly like the males, or, if different, are equally
conspicuous. We can hardly believe that so exact a parallel should exist
between such remote classes of animals, except under the influence of a
general law; and, in the need of protection by all defenceless animals,
and especially by most female insects and birds, we have such a law,
which has been proved to have influenced the colours of a considerable
proportion of the animal kingdom.[122]

The general relation which exists between the mode of nesting and the
coloration of the sexes in those groups of birds which need protection
from enemies, may be thus expressed: When both sexes are brilliant or
conspicuous, the nest is such as to conceal the sitting-bird; but when
the male is brightly coloured and the female sits exposed on the nest,
she is always less brilliant and generally of quite sober and protective
hues.

It must be understood that the mode of nesting has influenced the
colour, not that the colour has determined the mode of nesting; and
this, I believe, has been generally, though not perhaps universally, the
case. For we know that colour varies more rapidly, and can be more
easily modified and fixed by selection, than any other character;
whereas habits, especially when connected with structure, and when they
pervade a whole group, are much more persistent and more difficult to
change, as shown by the habit of the dog turning round two or three
times befo