A given animal or plant, [says the Reviewer] appears to be contained, as it were, within a sphere of variation; one individual lies near one portion of the surface, another individual near another part of the surface; the average animal at the centre. Any individual may produce descendants varying in any direction, but is more likely to produce descendants varying towards the centre of the sphere, and the variations in that direction will be greater in amount than the variations towards the surface. Thus a set of racers of equal merit indiscriminately breeding will produce more colts and foals of inferior than of superior breed, and the falling off of the degenerate will be greater than the improvement of the select (p. 282). Similarly M. Blanchard declares:[193] All investigation and observation make it clear that, while the variability of creatures in a state of nature displays itself in very different degrees, yet in its most astonishing manifestations it remains confined within a circle beyond which it cannot pass. And the facts of nature, as we know them, far from favouring the instability of species, exhibit a tenacity of form compelling us to treat them as practically immutable. Thus, as Mr. Carruthers points out,[194] in the notoriously variable genus _Salix_, or willow-tribe, which seems to be actively advancing towards a multiplication of its subdivisions, sub-genera, species, varieties, and hybrid forms,--one species is found, _S. polaris_, dating from before the Glacial Epoch, which has been driven from England and other lands, by climatic changes, to within the Arctic circle of both Hemispheres,--yet amid this stress of circumstances has preserved its specific identity, down even to the casual variations, which might be supposed to furnish the starting-points for new developments. Yet in this tribe, if anywhere, evidence of specific evolution might be looked for.[195] Other instances seem to show that even under new and trying conditions those creatures survive best which keep closest to the central family type, not those which diverge in any direction. Thus, of European sparrows introduced in America, Mr. Bumpus writes:[196] Natural Selection is most destructive of those birds which have departed most from the ideal type, and its activity raises the general standard by favouring those birds which approach the structural ideal. Variation supplies the raw material upon which Natural Selection is supposed to work. When we turn to examine the process by which its results should be produced, we find, quite apart from the above difficulties, a crop of others still more formidable.