As regards sexual selection I allow, of course, that the 'law of battle' is a form of natural selection. But where the matter is merely a pleasing of æsthetic taste, and the resulting structures therefore only ornamental, I can see nothing 'advantageous' in the sense of life-preserving. On the contrary, in most cases such structures entail considerable expenditure of physiological energy in their production. On this account Darwin says that nat. sel. must impose a check on sexual selection running beyond a certain point of injuriousness ('D. of M.,' p. 227). Now, physiological selection is never thus injurious; and although it is a 'form of isolation,' the isolation is neither so extreme nor of such long continuance as the ones you compare it with. Moreover, the environment (therefore all other or external conditions of life) remain the same, which is not the case under the other forms of isolation. Provided that the physiological change is not _in itself_ injurious, I do not see why _physiologically_ isolated forms should be less fit than those from which they have been separated, though I can very well see why this should be the case with such _geographically_ isolated forms as you mention, for there the _schooling_ is different. Lastly, physiological selection, if not in itself injurious, does not require that its children should be 'protected against the struggle for existence.' On the contrary, as I say in my paper, it is calculated to give this struggle a better chance than ever to develope adaptive character in the sexually isolated forms, because the swamping effects of intercrossing are diminished. But I really did not intend to afflict you with another jaw of this kind. I am, however, very glad that we now understand each other better than we did. At all events on my side I think I now know exactly the points which I have to make good if Nature is so constituted as to admit of my theory. One thing only I have forgotten to say, viz. that nothing can be argued against the theory from the fact of hybridisation occurring in cases where, according to the theory, it ought not to occur. This argument only becomes valid where it is found that the resulting hybrids are _fertile_. In relation to the theory, a sterile hybrid is all the same as a failure to cross. Yours very sincerely, G. J. ROMANES.