c elimination of the unadapted is a simple and
clear idea. But, just because it attributes to the outer cause which
controls evolution a merely negative influence, it has great difficulty
in accounting for the progressive and, so to say, rectilinear
development of complex apparatus such as we are about to examine. How
much greater will this difficulty be in the case of the similar
structure of two extremely complex organs on two entirely different
lines of evolution! An accidental variation, however minute, implies the
working of a great number of small physical and chemical causes. An
accumulation of accidental variations, such as would be necessary to
produce a complex structure, requires therefore the concurrence of an
almost infinite number of infinitesimal causes. Why should these causes,
entirely accidental, recur the same, and in the same order, at different
points of space and time? No one will hold that this is the case, and
the Darwinian himself will probably merely maintain that identical
effects may arise from different causes, that more than one road leads
to the same spot. But let us not be fooled by a metaphor. The place
reached does not give the form of the road that leads there; while an
organic structure is just the accumulation of those small differences
which evolution has had to go through in order to achieve it. The
struggle for life and natural selection can be of no use to us in
solving this part of the problem, for we are not concerned here with
what has perished, we have to do only with what has survived. Now, we
see that identical structures have been formed on independent lines of
evolution by a gradual accumulation of effects. How can accidental
causes, occurring in an accidental order, be supposed to have repeatedly
come to the same result, the causes being infinitely numerous and the
effect infinitely complicated?

The principle of mechanism is that "the same causes produce the same
effects." This principle, of course, does not always imply that the same
effects must have the same causes; but it does involve this consequence
in the particular case in which the causes remain visible in the effect
that they produce and are indeed its constitutive elements. That two
walkers starting from different points and wandering at random should
finally meet, is no great wonder. But that, throughout their walk, they
should describe two identical curves exactly superposable on each other,
is altogether u