Mr. Wallace rejects the theory of sexual selection _in toto_, and therefore nothing that can be said under this head would be held by him to be relevant. Many naturalists, however, believe that Darwin was right in the large generalization which he published under this title; and in so far as any one holds that sexual selection is a true cause of specific modification, he is obliged to believe that innumerable specific characters--especially in birds and mammals--have been produced without reference to utility (other, of course, than utility for sexual purposes), and therefore without reference to natural selection. This is so obvious that I need not pause to dilate upon it. One remark, however, may be useful. Mr. Wallace is able to make a much more effective use of his argument from "necessary instability" when he brings it against the Darwinian doctrine of sexual selection, than he does when he brings it against the equally Darwinian doctrine of specific characters in general not being all necessarily due to natural selection. In the latter case, it will be remembered, he is easily met by showing that the causes of specific change other than natural selection, such as food, climate, &c., may be quite as general, persistent, and uniform, as natural selection itself; and therefore in this connexion Mr. Wallace's argument falls to the ground. But the argument is much more formidable as he brings it to bear against the theory of sexual selection. Here he asks, What is there to guarantee the uniformity and the constancy of feminine taste with regard to small matters of embellishment through thousands of generations, and among animals living on extensive areas? And, as we have seen in Part 1, it is not easy to supply an answer. Therefore this argument from the "necessary instability of character" is of immeasurably greater force as thus applied against Darwin's doctrine of sexual selection, than it is when brought against his doctrine that all specific characters need not necessarily be due to natural selection. Therefore, also, if any one feels disposed to attach the smallest degree of value to this argument in the latter case, consistency will require him to allow that in the former case it is simply overwhelming, or in itself destructive of the whole theory of sexual selection. And, conversely, if his belief in the theory of sexual selection can survive collision with this objection from instability, he ought not to feel any tremor of contact when the objection is brought to bear against his scepticism regarding the alleged utility of all specific characters. For assuredly no specific character which is apparent to our eyes can be supposed to be so refined and complex (and therefore so presumably inconstant and unstable), as are those minute changes of cerebral structure on which a psychological preference for all the refined shadings and many pigments of a complicated pattern must be held ultimately to depend.