theory. What we require in proof of the distinguishing feature of that theory—i.e., the discontinuity of the hypothetical somato-plasm—is not merely the obvious fact that some characters are inherited while others are not, but independent proof that inherited and non-inherited characters correspond to a continuity of germ-plasm on the one hand, and a discontinuity of somato-plasm on the other. He shows us, indeed, what was well known before, that characters developed during the lifetime of the individual are seldom (if ever) inherited, while characters developed during the lifetime of the species are always inherited. Obviously, however, this fact is no proof of the assumed correlation just mentioned, because, as Darwin has clearly pointed out, it may very well be due to the much shorter time which has been allowed for what may be termed the impress of heredity. Therefore, supposing (with Darwin and others) that living material is all of one kind, and continuous, the fact on which Weismann relies admits of being explained without resorting to his more complex supposition of living material in two kinds, the one perpetually continuous, and the other interrupted at each ontogeny. For these reasons it appears to me that, so far as the argument from “inconceivability” is concerned, it makes at least as much against the theory of germ-plasm as it does against the theory of pangenesis; and, therefore, that no argumentative advantage is gained from its use by Weismann. The truth probably is that, _whatever_ the mechanism of heredity may actually be, it is at once so minute and so complex that its action is “inconceivable,” or, more correctly, unimaginable. Be it again understood, therefore, that I am not arguing in favour of pangenesis. I am merely criticising what appears to me an unsound argument in favour of germ-plasm. All this general or merely _a priori_ reasoning with regard to inconceivability is, as I have attempted to show, as available on the one side as on the other, and so fails to yield any observable advantage to either. * * * * *