But even if the difficulty of the appropriate enemy be passed over, and it be granted that a mimetic resemblance can be built up through a number of small separate steps, which have become separately established through the agency of separate species of birds with various but distinct discriminating powers, we are left face to face with an even more serious physiological difficulty. For why is it that when the end form which is supposed to have resulted from this process is crossed back with the original form all {141} the intermediate steps do not reappear? Why is it that when the altered germplasm is mingled with the original germplasm the various postulated stages between them are not reformed? For in various cases where we know the course of evolution this does occur. The pale pink sweet-pea has come from the wild purple by a series of definite steps, and when it is crossed back with the wild form the resulting plants give the series of stages that have occurred in the evolution of the pink. So also when the orange rabbit is crossed with the wild grey form and the offspring are inbred there are reproduced the black, the tortoiseshell, and the chocolate, forms which are stages in the evolution of the orange from the wild grey. If then, to take an example, the "aristolochiae" form of _Papilio polytes_ has been derived from the male-like form by a series of steps, why do we not get these steps reproduced after the germplasms of the two forms have been mingled? From the standpoint of modern genetic work the inference is that these postulated intermediate steps have never existed--that the one form of _polytes_ female came directly from the other, and was not built up gradually through a series of stages by the selective agency of birds or any other discriminating enemy.