Now, it does not appear that Mr. Wallace has clearly perceived this important distinction, because he throughout speaks of "this inferiority of the hybrid offspring as the essential point." Obviously, however, the essential point is the physiological variation in the parent forms, i. e., the original occurrence and subsequent development of infertility between the first crosses. Granting to Mr. Wallace, for the sake of argument, that this development could not proceed at all, were it not for the inferiority of the mongrels; still the inferiority of the mongrels need not be the cause of this development. Therefore it is most incorrect to say, "it must be particularly noted that this effect (i. e., increase of infertility between the parent forms) would result, not by the preservation of the infertile variations on account of their infertility, but by the inferiority of the hybrid offspring." "This effect" must be due to causes which act upon the generative systems of the parent forms, even though such causes might be _counteracted_ by the _withdrawal_ of the negative condition in question.
I trust, then, it has now been rendered sufficiently clear that, no matter how infertile the hybrid progeny may become, and no matter at how great a disadvantage they may thus (or otherwise) be placed in their struggle for existence with the parent varieties, it is not apparent that their infertility (or their extinction) can ever become the cause of a further increase of infertility arising between their parent forms. Consequently, although this is the cause assigned by Mr. Wallace, when he comes to "the essential point" of showing how it is to act so as to increase cross-sterility between the parent forms, he naïvely substitutes the sentence which I have printed in italics—which assumes a "greater infertility between the two forms" as arising through any other causes that we may choose to suppose. The very thing that his entire argument professes to explain (i. e., the rise and development of cross-sterility between the parent varieties) is slipped in as granted, or given by other causes than those which are said to explain it.[6]
[6] The only conceivable way in which infertility (or other inferiority) of hybrids could react on the sexual system of their parent forms, is one which Mr. Wallace appears to have missed: at all events he has nowhere stated it. This way is as follows. Suppose A and B to be two varieties which produce comparatively infertile hybrids. In the proportion that the hybrids are infertile, or otherwise inferior, it must be a disadvantage to both varieties for individuals belonging to one to cross with individuals belonging to the other, because by so doing they are wasting their time and their energy in propagating comparatively poor offspring—thereby failing to impress their characters on the next generation as effectually as they might have done by pairing homogamously. Hence, those individuals which do pair homogamously will leave a larger number—or better quality—of offspring to the next generation, than is left by those which fail to pair homogamously. Hence, also, in the course of many generations a selective premium will be set on the homogamous pairing, A plus A, B plus B, whether such pairing be due to a sexual instinct or to a sexual incompatibility. For example, if horses and asses were to occupy the same area for a sufficient length of time, it is conceivable that the instinct which many horses now present of preferring asses to their own kind would become obsolete; because the horses or mares which have such an instinct would always fail to leave progeny that could transmit it, while such would not be the case with the horses and mares which preferred to pair homogamously, and so it might be if a physiological instead of a psychological character were concerned. But now observe, if this consideration were adduced, I should not be concerned to dispute it. For, even if such a principle of segregation does obtain, to what category does the principle belong? Clearly it does not belong to natural selection, inasmuch as a mere failure to impress individual characters on the next generation is not a matter of life and death in the struggle for existence. But, no less clearly, it does belong to physiological selection; and therefore, if it be an active principle in nature, it is an additional cause of segregate fecundity in first crosses. Moreover, such a principle, if it ever acts, presupposes some considerable degree of sexual differentiation as already given by some other cause.