The chief means which man uses for the attainment of these results, which at times seem to border on the marvellous, is _selection_. Ever since he has possessed domestic animals he has marked out among them individuals which are better adapted than the rest to his intentions. By some kind of instinct, or _unconsciously_, as Darwin says, he has chosen them to breed from. By rejecting the types which he considers inferior, and only employing the higher types wherewith to propagate the species, he has directed the action of heredity in a definite direction, and has readily created races. Now, man has acted in this manner since the times spoken of in Genesis and by Chou-King, that is, for thousands of years. Is it then surprising that he should have multiplied around him hereditary forms which are more or less distinct from the primitive types? _Progressive selection_ would doubtless lead to numerous and varied results. Would it allow of the creation of races whose characters almost reach hemitery? The answer to this question is at least doubtful. But we have not to ask it. When by one of the actions of the conditions of life, whose origin remains obscure, an almost teratological animal form is produced, it soon disappears by the mixture of blood from different sources, if unions are left to chance. This is the reason why analogous facts are not observed in feral races. But if this form appears in a domestic animal, if it answers to any want or caprice, selection intervenes, preserves it, and multiplies it. This explains the origin of the Ancon sheep, which were all descended from a single ram of which we have spoken above; also the means by which M. Graux de Mauchamp has raised his race of sheep with silky fleeces from a single ram. These two examples show how all those peculiar races have been obtained, which in some of their characters seem to clash with the very type from which they were derived. In the canine species the beagle corresponds to the Ancon sheep; the _niata_ cattle, which have appeared in South America since its conquest, correspond to the bull-dog, etc. VIII. Races, when once formed under man’s influence, are fixed by the same causes which produced them. Their characters, which at first were entirely artificial, become more and more fixed, so much so, that even a very considerable change in the conditions of existence, never effaces them entirely. The _acquired nature_ is, so to speak, welded to the original nature of the being.