tion tends to form varieties by peculiarities in the reproductive system of individuals, which render them unfit for perfect union, or cause them to remain more or less sterile, with other individuals which have not the same peculiarities.
The exact reasons are recondite, and the whole subject difficult of demonstration except from the results, since changes in the reproductive organs are not easily observable. Romanes believes this sterility to be incidental to variation, and hence one of the chief causes of the accumulation of such variation. Wherever there has been modification of the reproductive organs introducing incompatibility between two individuals, even where there has been no other change or variation, we have a valid cause of differentiation which in its consequences must be important. Compatibility or fertility between individuals is of the very essence of selection. Natural selection implies that this sexual divergence is subsequent to or coincident with divergences in other directions; physiological selection, that it antecedes them. To put the case of Romanes more fully, we will suppose that among the natural variations there occasionally occurs something to affect the reproductive organs in such wise as to produce incompatibility—i.e., incapacity of one individual with another of the parent type to unite, or sterility of such union, while it remains fertile with the variation of its own kind. This theory, of course, implies variation in the reproductive organs, or departure from the parental type, in at least two individuals of opposite sex simultaneously, and with this admission, for which we are justified in facts, physiological selection will preserve many peculiarities which need have no necessary connection with the exigencies of life.
The change may be in the organs of reproduction, introducing sexual incompatibility, or it may be due to other causes, as, for instance, the time of flowering in plants, or the season of heat in animals. Even the element of scent becomes important here, as my friend J. Jenner Weir has suggested, since it may influence sexual relationship, so that the very excretions of the body, which vary with individuals, must be allowed their part. Francis Galton has indicated a modification of Romanes's views, viz., that the primary characteristic of a variety resides in the fact that the individuals who compose it do not care to mate with those outside their pale. Incipient varieties are thus thrown off from the parent stock by means of peculiarities of sexual instinct which prompt what anthropologists call endogamy, and check exogamy or marriage without the tribe or caste. This is a very good anthropological illustration of how physiological selection may begin.
Natural selection preserves the individuals best adapted to life-conditions by destroying the less fit. Physiological selection may be said to preserve differences which have no necessary con-
