other parts of the general area, increase till they supplant these original forms. Then, in some limited portion of the general area, there arise two still more divergent forms, with greater mutual infertility, and with increased adaptation to the environment, enabling them to prevail throughout the whole area. The process here described, if it takes place, is not modification by natural selection. On the other hand, it _is_ modification by physiological selection. For, among the several other forms of isolation which are called into requisition, the physiological (i.e. ever accumulating cross-infertility) is supposed to play an important part. That the modification is not modification by natural selection may perhaps be rendered more apparent by observing, that in as far as _any_ other mode of isolation is involved or supposed, so far is the _possible_ agency of natural selection eliminated _as between the two or more otherwise isolated sections of a species_; and yet it is modes of isolation other than that furnished by natural selection (i.e. perishing of the less fit), that Mr. Wallace here supposes to have been concerned--including, as I have before shown, the physiological form, to which, indeed, he really assigns most importance of all. Or, as Mr. Gulick states the matter in his independent criticism:-- In the supposed case pictured by Mr. Wallace, the principle by which the two segregating forms are kept from crossing, and so are eventually preserved as permanently distinct forms, is no other than that which Mr. Romanes and myself have discussed under the terms Physiological Selection and Segregate Fecundity. Not only is Mr. Wallace's exposition of the divergence and the continuance of the same in accord with these principles which he has elsewhere rejected, but his whole exposition is at variance with his own principle, which, in the previous chapter, he vigorously maintains in opposition to my statement that many varieties and species of Sandwich Island land molluscs have arisen, while exposed to the same environment, in the isolated groves of the successive valleys of the same mountain range. If he adhered to his own theory, "the greater infertility between the two forms in one portion of the area" would be attributed to a difference between the _environment_ presented in that portion and that presented in the other portions; and the difficulty would be to consistently show how this greater infertility could continue unabated when the varieties thus characterized spread beyond the environment on which the character depends. But, without power to continue, the process which he describes would not take place. Therefore, in order to solve the problem of the _origin_ and _increase_ of infertility between species, he tacitly gives up his own theory, and adopts not only