It is needless to say that he stoutly resists the view of Weismann, Delbœuf, Gulick, and myself, that specific divergence can ever be due--or, as I understand him, even so much as assisted--by this principle of indiscriminate isolation (apogamy). It will be remembered, however, that Mr. Gulick has adduced certain general principles and certain special facts of geographical distribution, in order to prove that apogamy eventually leads to divergence of character, provided that the isolated section of the species does not contain any very large number of individuals. Now, Mr. Wallace, without making any reference to this argument of Mr. Gulick, simply states the reverse--namely, that, as a matter of fact, indiscriminate isolation is not found to be associated with divergence of character. For, he says, "there is an entire absence of change, where, if this were a _vera causa_, we should expect to find it[53]." But the only case which he gives is that of Ireland. [53] _Darwinism_, p. 151. This, he says, furnishes "an excellent test case, for we know that it [Ireland] has been separated from Britain since the end of the glacial epoch: ... yet hardly one of its mammals, reptiles, or land molluscs has undergone the slightest change[54]." Here, however, Mr. Wallace shows that he has failed to understand "the views of those who, like Mr. Gulick, believe isolation itself to be a cause of modification of species"; for it belongs to the very essence of these views that the efficiency of indiscriminate isolation as a "_vera causa_" of organic evolution varies inversely with the number of individuals (i. e. the size of the species-section) exposed to its influence. Therefore, far from being "an excellent test case," the case of Ireland is unsatisfactory. If we are in search of excellent test cases, in the sense intended by Mr. Wallace, we ought not to choose a large island, which from the time of its isolation must have contained large bulks of each of the geographically separated species concerned: we ought to choose cases where as small a number as possible of the representatives of each species were in the first instance concerned. And, when we do this, the answer yielded by any really "excellent test case" is unequivocal. [54] _Ibid._