Well, in the first place, this objection can have reference only to the first two of the five principles above stated. It can have no reference to the last three, because of these heredity constitutes the very foundation. This consideration ought to be borne in mind throughout. But now, in the second place, even as regards changes produced by climate and food, the reply is nugatory. And this for three reasons, as follows. (_a_) No one is thus far entitled to conclude against the possible transmission of acquired characters; and, so long as there is even so much as a possibility of climatic (or any other admittedly non-utilitarian) variations becoming in this way hereditary, the reply before us merely begs the question. (_b_) Even supposing, for the sake of argument, that acquired characters can never in any case become congenital, there remains the strong probability--sanctioned as such even by Weismann--that changed conditions of life may not unfrequently act upon the material of heredity itself, thus giving rise to specific changes which are from the first congenital, though not utilitarian. Indeed, there are not a few facts (Hoffmann's plants, Weismann's butterflies, &c.), which can only be explained either in this way, or as above (_a_). And in the present connexion it is immaterial which of these alternative explanations we choose to adopt, seeing that they equally refute our opponents' objection. And not only do these considerations--(_a_) and (_b_)--refute this particular objection; they overturn on new and independent grounds the whole of our opponents' generalization. For the generalization is, that the principle of utility, acting through natural selection, is "necessarily" the sole principle which can be concerned in hereditary changes of specific type. But here we perceive both a possibility (_a_) and a probability (_b_), if not indeed a certainty, that quite other principles have been largely concerned in the production of such changes.