In the origin of new species by natural selection, the stress of competition determines the survival of favourable individual variations, and these, when by the continued operation of the process they have become constant, are added to those pre-existing characters of the species which are inherited from a remote past, and are witnesses of the operation of natural selection from age to age under ever-changing conditions of competition and variation. It follows, therefore, that the origin of a species can only take place once; for it is infinitely improbable that the same variation would be independently submitted under the same conditions of competition, and added to the mass of inherited characters independently gained in two distinct lines by natural selection acting in the same manner upon the same variations in the same order through all ages. Not only is it inconceivable that the same species could arise by natural selection from distinct lines of ancestry, but it is extremely improbable that the same species could arise independently in more than one centre among the individuals of a changing species; for in this case, too, it is most unlikely that the same conditions of competition would co-exist with the same favourable variations in the areas inhabited by independent colonies of the same species. [Sidenote: EARLY CRITICISM.] Under other theories of evolution--direct action of environment, supposed inherited effects of use and disuse, etc.--an independent origin, even from quite distinct lines, would be probable; and we find, accordingly, that those who would advance such theories believe in what is called the “polyphyletic” origin of species (_e.g._ the horse), and in the principle of “convergence” carried far enough to produce the same complex character (_e.g._ vertebrate teeth) twice over without any genetic connection between the forms in which the character appears. Under natural selection, however, such a result would be infinitely improbable, and hence this theory strongly supports, and indeed explains, the theory of “specific centres,” viz. that each species has arisen in one area only, and has spread from that into the other areas over which it now occurs. This view was strongly held by Lyell and Hooker after an exhaustive study of the facts then known as to the geographical distribution of plants and animals; and yet both of these distinguished naturalists seem to have feared that Darwin, in advancing a theory which was entirely consistent with their convictions and utterly inconsistent with any other views upon the same subject, was in some way undermining the conclusions at which they had arrived. Thus Lyell wrote (July 25th, 1856) to Hooker:--