169 
presumed, that they are not due to the selection of those 
favourable variations which occur in connection with 
sexual reproduction. He says :— 
“The meaning of these researches is self-evident. No naturalist 
will probably attempt to explain the animal colours and markings 
without the aid of Natural Selection. But it becomes less and less 
probable to admit that animal colours are a result of a selection of 
accidental variations only.”—(.Nineteenth Century, vol. xxxiii.,p.689.) 
But he goes on to say :— 
“As to the relative parts which must be attributed in the origin 
of each separate variation to Natural Selection on the one side, 
and to the direct action of environment on the other side, it would 
simply be unscientific to trench upon such questions in a broadcast 
way, so long as we are only making our first steps in discriminating 
the action of the latter agency.” 
But if the direct action of environment is already proved 
to be very great, that in itself is enough to show that the 
theory of defensive colouring is not an especially favour 
able illustration of the action of Natural Selection. If 
already it can be shown that the variations are not 
accidental, that, according to our view, is enough to prove 
that the Natural Selection of birth variations does not 
come into play. It is surely not unscientific to note how 
facts either lessen the possible action of a law or even 
call in question the truth of a theory. We may not yet 
have solved the problem : but we know enough to declare 
that the requirements of the theory are not realised when 
the phenomena are closely observed. The change is pro 
duced without Natural Selection : it is not produced by the 
selection of favourable birth variations. If this be so, we 
shall not be justified in holding the judgment any longer in 
suspense. The logical demands of the theory must remain 
the same, unless the theory itself is to be largely modified. 
If the facts already contradict the theory, what is there