o
settle on a flower or a green leaf, but were many times lost sight of in a
bush or tree of dead leaves. On such occasions they were generally searched
for in vain, for while gazing intently at the very spot where one had
disappeared, it would often suddenly dart out, and again vanish twenty or
fifty yards further on. On one or two occasions the insect was detected{33}
reposing, and it could then be seen how completely it assimilates itself to
the surrounding leaves. It sits on a nearly upright twig, the wings fitting
closely back to back, concealing the antennae and head, which are drawn up
between their bases. The little tails of the hind wing touch the branch,
and form a perfect stalk to the leaf, which is supported in its place by
the claws of the middle pair of feet, which are slender and inconspicuous.
The irregular outline of the wings gives exactly the perspective effect of
a shrivelled leaf. We thus have size, colour, form, markings, and habits,
all combining together to produce a disguise which may be said to be
absolutely perfect; and the protection which it affords is sufficiently
indicated by the abundance of the individuals that possess it."

Beetles also imitate bees and wasps, as do some Lepidoptera; and objects
the most bizarre and unexpected are simulated, such as dung and drops of
dew. Some insects, called bamboo and walking-stick insects, have a most
remarkable resemblance to pieces of bamboo, to twigs and branches. Of these
latter insects Mr. Wallace says:[27] "Some of these are a foot long and as
thick as one's finger, and their whole colouring, form, rugosity, and the
arrangement of the head, legs, and antennae, are such as to render them
absolutely identical in appearance with dry sticks. They hang loosely about
shrubs in the forest, and have the extraordinary habit of stretching out
their legs unsymmetrically, so as to render the deception more complete."
Now let us suppose that the ancestors of these various animals were all
destitute of the very special protections they at present possess, as on
the Darwinian hypothesis we must do. Let it also be conceded that small
deviations from the antecedent colouring or form would tend to make some of
their ancestors escape destruction by causing them more or less frequently
to be passed over, or mistaken by their persecutors. Yet the deviation {34}
must, as the event has shown, in each case be in some definite direction,
whether it be towards some other