Does not this grant all that Darwin contends for? In one sense, yes; in another, no! The results appear to show that by artificial selection of some kind a group of new forms may be produced that in many respects resemble a natural family, or a genus; but if this is to be interpreted to mean that the result is the same as that by which natural groups have arisen, then I think that there are good reasons for dissenting from such a conclusion. Moreover, we must not grant too readily that the different races of pigeons have arisen by the selection of _fluctuating variations_ alone, for this is not established with any great degree of probability by the evidence. In regard to the first point we find that one of the most striking differences between species in nature is their infertility, and the infertility of their offspring when intercrossed. This is a very general rule, so far as we know. In regard to the different races of domesticated forms, the most significant fact is that, no matter how different they may be, they are perfectly fertile _inter se_. In this respect, as well as in others, there are important differences between domesticated races and wild species. The further difference, that has been pointed out by a number of writers, should also not pass unnoticed, namely, that the domestic forms differ from each other in the extreme development of some one character, and not in a large number of less conspicuous characters, as is the case in wild species. These considerations show that, interesting and suggestive as are the facts of artificial selection, they fail to demonstrate the main point for which they are used by Darwin. With the most rigorous attention to the process of artificial selection, new species comparable in all respects to wild ones have not been formed, even in those cases in which the variation has been carried farthest (where the history of the forms is most completely known). There is another point on which emphasis should be laid. If by selecting the most extreme forms in each generation and breeding from them the standard can be raised, it might appear that we could go on indefinitely in the same direction, and produce, for instance, pigeons with legs five metres long, and with necks of corresponding length. But experience has shown that this cannot be done. As Darwin frequently remarks, the breeder is entirely helpless until the desired variation appears. It seems possible, by selecting the more extreme of the fluctuating variations in each generation, that a higher plane of variation is established, and even that more extreme forms are likely to arise for a few generations; but, even if this is the case, a limit is soon reached beyond which it is impossible to go.