y, it would perhaps be more
in accordance with cause and effect to express the law of correlation as
a _decrease_ of intensity of colour with a _decrease_ of humidity, the
paleness evidently resulting from exposure and the blanching effect of
intense sunlight, and a dry, often intensely heated atmosphere. With the
decrease of the aqueous precipitation the forest growth and the
protection afforded by arborescent vegetation gradually also decreases,
as of course does also the protection afforded by clouds, the
excessively humid regions being also regions of extreme cloudiness,
while the dry regions are comparatively cloudless districts."[91] Almost
identical changes occur in birds, and are imputed by Mr. Allen to
similar causes.

It will be seen that Mr. Gould and Mr. Allen impute opposite effects to
the same cause, brilliancy or intensity of colour being due to a
brilliant atmosphere according to the former, while paleness of colour
is imputed by the latter to a too brilliant sun. According to the
principles which have been established by the consideration of arctic,
desert, and forest animals respectively, we shall be led to conclude
that there has been no direct action in this case, but that the effects
observed are due to the greater or less need of protection. The pale
colour that is prevalent in arid districts is in harmony with the
general tints of the surface; while the brighter tints or more intense
coloration, both southward and in humid districts, are sufficiently
explained by the greater shelter due to a more luxuriant vegetation and
a shorter winter. The advocates of the theory that intensity of light
directly affects the colours of organisms, are led into perpetual
inconsistencies. At one time the brilliant colours of tropical birds and
insects are imputed to the intensity of a tropical sun, while the same
intensity of sunlight is now said to have a "bleaching" effect. The
comparatively dull and sober hues of our northern fauna were once
supposed to be the result of our cloudy skies; but now we are told that
cloudy skies and a humid atmosphere intensify colour.

In my _Tropical Nature_ (pp. 257-264) I have called attention to what is
perhaps the most curious and decided relation of colour to locality
which has yet been observed--the prevalence of white markings in the
butterflies and birds of islands.

So many cases are adduced from so many different islands, both in the
eastern and western hemisphere, th