In considering whether the domestic breeds are descended from one species, namely, G. bankiva, or from several, we must not quite overlook, though we must not exaggerate, the importance of the test of fertility. Most of our domestic breeds have been so often crossed, and their mongrels so largely kept, that it is almost certain, if any degree of infertility had existed between them, it would have been detected. On the other hand, the four known species of Gallus when crossed with each other, or when crossed, with the exception of G. bankiva, with the domestic fowl, produce infertile hybrids. Finally, we have not such good evidence with fowls as with pigeons, of all the breeds having descended from a single primitive stock. In both cases the argument of fertility must go for something; in both we have the improbability of man having succeeded in ancient times in thoroughly domesticating several supposed species,--most of these supposed species being extremely abnormal as compared with their natural allies,--all being now either unknown or extinct, though the parent-form of no other domesticated bird has been lost. But in searching for the supposed parent- stocks of the various breeds of the pigeon, we were enabled to confine our search to species having peculiar habits of life; whilst with fowls there is nothing in their habits in any marked manner distinct from those of other gallinaceous birds. In the case of pigeons, I have shown that purely- bred birds of every race and the crossed offspring of distinct races frequently resemble, or revert to, the wild rock-pigeon in general colour and in each characteristic mark. With fowls we have facts of a similar nature, but less strongly pronounced, which we will now discuss. REVERSION AND ANALOGOUS VARIATION.