We are thus led to the hypothesis, which was first introduced into the Evolution Theory by Darwin, that the prevention of constant crossing of an isolated colony with the others of the same species from the original habitat favours the origin of new endemic species, and his conclusion is confirmed when we learn that islands like the Galapagos group possess twenty-one endemic land-birds, but only two endemic sea-birds out of eleven, for the latter traverse great stretches of sea, and crossings with others of the same species on the neighbouring continental coasts will often take place. The Bermuda Islands also afford a proof that the development of endemic species is prevented by regular crossing with other members of the species from the original habitat, for although they are 1,200 kilometres distant from the continent of North America—that is, further than the Galapagos Islands from South America—they possess no endemic species of bird, and we may undoubtedly associate this with the fact that the migratory birds from the continent visit the Bermudas every year.
Madeira also confirms our conclusion, for only one of the ninety-nine species of bird occurring there can be regarded as endemic, and it has often been observed that birds from the neighbouring African mainland (only 240 kilometres distant) are driven across to Madeira. Terrestrial snails, on the other hand, will seldom be carried to Madeira by birds, and accordingly we find there an extraordinary number of endemic terrestrial snails, namely, 109 species.
Although these and similar facts indicate strongly that isolation favours the evolution of new species, it would be erroneous to imagine that every isolation of a species-colony conditions its transmutation to a new species, or, as has been maintained, first by Moritz Wagner, and later by Gulick and by Dixon, that isolation is a necessary preliminary to the variation of species—that not selection but isolation alone renders the transmutation of a species possible, and thus admits of its segregation into several different groups of forms. Romanes went so far as to regard the natural selection of Darwin and Wallace as a sub-species of isolation, and isolation in its diverse forms he regarded as the sole factor in the formation of species. He assumed that it was only by the segregation of individuals which did not vary that the constant reversion to the ancestral species could be prevented, and he regarded the process of selection as essentially resulting in the 'isolation' of the fittest through the elimination of the less-fit. The idea is correct in so far, that selection undoubtedly aids the favourable variation to conquest over the old forms, precisely because the latter, being less favourably placed in the struggle for existence, are gradually more completely overcome and weeded out, so that a constant mingling of the new forms with the old is prevented, just as it is by isolation of locality. Obviously the new and fitter forms could not become dominant, could not even become permanent, if they were always being mingled again with the old. But whether it serves any useful purpose to bring this under the category of 'isolation,' and to say that mingling with the ancestral form during transmutation is prevented by natural selection, in that favourably varying individuals are isolated by their superiority from the inferior ones, that is, the non-varying individuals which are doomed to elimination, is somewhat doubtful. For my part, I should prefer to retain the original meaning of the word, and to call 'isolation' the separation of a species-colony by spatial barriers.
Whether this factor by itself prevents the mingling with the ancestral form as effectually as selection does, and whether isolation alone and by itself can lead to the evolution of new forms, or perhaps must lead to them, must now be investigated.
I look at this question from exactly the same point of view as I did nearly thirty years ago, when in a short paper[27] I endeavoured to show that, under favourable circumstances, an individual variation of a species may become the origin of a local variety if it finds itself in an isolated region. Suppose an island had no diurnal butterflies, until one day a fertilized female of a species from the continent was driven thither, found suitable conditions of life, laid its eggs, and became the founder of a colony; the prevention of constant crossing between this colony and the ancestral continental species would not in itself be any reason why the colony should develop into a variety. But suppose that the foundress of the colony diverged in some unimportant detail of colouring, such as may at any time arise through germinal selection, from the ancestral species; then this variation would be transmitted to a portion of her progeny, and there would thus be a possibility that a variety should establish itself upon the island which would be the mean of the characters of the surviving progeny. The greater the divergence was in the first progeny of the mother-colonist, and the stronger this variational tendency was, the greater also would be the chance that it would be transmitted further and become a characteristic aberration from the marking of the original species. I then designated this effect of isolation as due to amixia, that is, to the mere prevention of crossing with the members of the same species in the original habitat.
[27] Ueber den Einfluss der Isolirung auf die Artbildung, Leipzig, 1872.
We have examples of this from the Mediterranean islands, Sardinia and Corsica, which possess in common nine endemic varieties of butterflies, most of which diverge from the species of the continent in a quite inconsiderable degree, though quite definitely and constantly. Thus there flies in these islands a variety (Vanessa ichnusa) of our common little Vanessa urticæ in which the two black spots on the anterior wing exhibited by the original species are wanting. The large tortoise-shell (Vanessa polychloros) also occurs there, but it has not varied and still exhibits the black spots. Our little indigenous butterfly (Pararga megæra), which is abundant on warm, stony slopes, quarries, and roads, flies about in Sardinia, but as a variety (tigelius), which is distinguished from the original species by the absence of a black curved line on the posterior wings.
That of two nearly related and similarly marked species, like the large and small tortoise-shell, one should remain unvaried, while the other has become a variety, shows us that amixia alone does not necessarily lead to the evolution of varieties in every case. It might of course be objected that one species may have migrated to the islands at a much earlier period than the other, and that it might be a direct effect of the climate which found expression in this way. But we have other similar cases in which one of two species has varied in an isolated region, while the other has not, and in regard to which we can prove definitely that both were isolated at the same time.
An instance of this kind is to be found in Arctic and Alpine Lepidoptera, which inhabited the plains of Europe during the Glacial period, and subsequently, when the climate became milder again, migrated some to the north into countries within the Arctic zone, and some to the south to the Alps to escape in their heights from the increasing warmth. There are many diurnal Lepidoptera which now belong to both regions, and of these some have remained exactly alike, so that the Arctic form cannot be distinguished from the Alpine form; others show slight differences, so that we can distinguish an Arctic and an Alpine variety. To the former category belong, for instance, Lycæna donzelii and Lycæna pheretes, Argynnis pales, Erebia manto, and others; to the second category belong, for instance, Lycæna orbitulus, Prun., Lycæna optilete, Argynnis thore, and some species of the genus Erebia.
This cannot be an instance of the direct effect of general climatic influences, for in that case all the nearly related species of a genus would have varied or not varied; nor have we to do with adaptations, for the differences in marking are seen on the upper surfaces of the wing, which do not exhibit protective colouring, at least in these Lepidoptera. It can only have been the prevention of crossing that has fixed the existing variational tendencies in the isolated colonies—variations which would have been swamped and obliterated if there had been constant crossing with all the rest of the members of the species.