in the majority—­not in the growth of new weapons for snatching the food from the mouth of its congeners—­food is only one out of a hundred of various conditions of existence—­but, as Wallace himself shows in a charming paragraph on the “divergence of characters” (Darwinism, p. 107), in forming new habits, moving to new abodes, and taking to new sorts of food.  In all such cases there will be no extermination, even no competition—­the new adaptation being a relief from competition, if it ever existed; and yet there will be, after a time, an absence of intermediate links, in consequence of a mere survival of those which are best fitted for the new conditions—­as surely as under the hypothesis of extermination of the parental form.  It hardly need be added that if we admit, with Spencer, all the Lamarckians, and Darwin himself, the modifying influence of the surroundings upon the species, there remains still less necessity for the extermination of the intermediate forms.

The importance of migration and of the consequent isolation of groups of animals, for the origin of new varieties and ultimately of new species, which was indicated by Moritz Wagner, was fully recognized by Darwin himself.  Consequent researches have only accentuated the importance of this factor, and they have shown how the largeness of the area occupied by a given species—­which Darwin considered with full reason so important for the appearance of new varieties—­can be combined with the isolation of parts of the species, in consequence of local geological changes, or of local barriers.  It would be impossible to enter here into the discussion of this wide question, but a few remarks will do to illustrate the combined action of these agencies.  It is known that portions of a given species will often take to a new sort of food.  The squirrels, for instance, when there is a scarcity of cones in the larch forests, remove to the fir-tree forests, and this change of food has certain well-known physiological effects on the squirrels.  If this change of habits does not last—­if next year the cones are again plentiful in the dark larch woods—­no new variety of squirrels will evidently arise from this cause.  But if part of the wide area occupied by the squirrels begins to have its physical characters altered—­in consequence of, let us say, a milder climate or desiccation, which both bring about an increase of the pine forests in proportion to the larch woods—­and if some other conditions concur to induce the squirrels to dwell on the outskirts of the desiccating region—­we shall have then a new variety, i.e. an incipient new species of squirrels, without there having been anything that would deserve the name of extermination among the squirrels.  A larger proportion of squirrels of the new, better adapted variety would survive every year, and the intermediate links would die in the course of time, without having been starved out by Malthusian competitors.  This is exactly what we see going on during the great physical changes which are accomplished over large areas in Central Asia, owing to the desiccation which is going on there since the glacial period.