ed, a complete recasting of the whole. How could mere chance work a
recasting of the kind? I agree that an accidental modification of the
germ may be passed on hereditarily, and may somehow wait for fresh
accidental modifications to come and complicate it. I agree also that
natural selection may eliminate all those of the more complicated forms
of instinct that are not fit to survive. Still, in order that the life
of the instinct may evolve, complications fit to survive have to be
produced. Now they will be produced only if, in certain cases, the
addition of a new element brings about the correlative change of all the
old elements. No one will maintain that chance could perform such a
miracle: in one form or another we shall appeal to intelligence. We
shall suppose that it is by an effort, more or less conscious, that the
living being develops a higher instinct. But then we shall have to admit
that an acquired habit can become hereditary, and that it does so
regularly enough to ensure an evolution. The thing is doubtful, to put
it mildly. Even if we could refer the instincts of animals to habits
intelligently acquired and hereditarily transmitted, it is not clear how
this sort of explanation could be extended to the vegetable world, where
effort is never intelligent, even supposing it is sometimes conscious.
And yet, when we see with what sureness and precision climbing plants
use their tendrils, what marvelously combined manoeuvres the orchids
perform to procure their fertilization by means of insects,[68] how can
we help thinking that these are so many instincts?

This is not saying that the theory of the neo-Darwinians must be
altogether rejected, any more than that of the neo-Lamarckians. The
first are probably right in holding that evolution takes place from germ
to germ rather than from individual to individual; the second are right
in saying that at the origin of instinct there is an effort (although it
is something quite different, we believe, from an _intelligent_ effort).
But the former are probably wrong when they make the evolution of
instinct an _accidental_ evolution, and the latter when they regard the
effort from which instinct proceeds as an _individual_ effort. The
effort by which a species modifies its instinct, and modifies itself as
well, must be a much deeper thing, dependent solely neither on
circumstances nor on individuals. It is not purely accidental, although
accident has a large place in it; a