This statement must not be misunderstood. Both formulas refer to one case, to a single sexual relation, the second following directly from the first and adding nothing to it, as I set out from the point of view of an individual possessing just as much femaleness as he lacks of maleness. Were he completely male, his requisite complement would be a complete female, and _vice versâ_. If, however, he is composed of a definite inheritance of maleness, and also an inheritance of femaleness (which must not be neglected), then, to complete the individual, his maleness must be completed to make a unit; but so also must his femaleness be completed. If, for instance, an individual be composed thus: { ³⁄₄ M μ { and { ¹⁄₄ W, then the best sexual complement of that individual will be another compound as follows: { ¹⁄₄ M ω { and { ³⁄₄ W. It can be seen at once that this view is wider in its reach than the common statement of the case. That male and female, as sexual types, attract each other is only one instance of my general law, an instance in which an imaginary individual, { 1 M χ { { 0 W, finds its complement in an equally imaginary individual, { 0 M γ { { 1 W. There can be no hesitation in admitting the existence of definite, individual sexual preferences, and such an admission carries with it approval of the necessity of investigating the laws of the preference, and its relation to the rest of the bodily and mental characters of an individual. The law, as I have stated it, can encounter no initial sense of impossibility, and is contrary neither to scientific nor common experience. But it is not self-evident. It might be that the law, which cannot yet be regarded as fully worked out, might run as follows: Mμ - Mω = a constant; that is to say, it may be the difference between the degrees of masculinity and not the sum of the degrees of masculinity that is a constant quality, so that the most masculine man would stand just as far removed from his complement (who in this case would lie nearly midway between masculinity and femininity) as the most feminine man would be removed from his complement who would be near the extreme of femininity. Although, as I have said, this is conceivable, it is not borne out by experience. Recognising that we have to do here with an empirical law, and trying to observe a wise scientific restraint, we shall do well to avoid speaking as if there were any “force” pulling the two individuals together as if they were puppets; the law is no more than the statement that an identical relation can be made out in each case of maximum sexual attraction. We are dealing, in fact, with what Ostwald termed an “invariant” and Avenarius a “multiponible”; and this is the constant sum formed by the total masculinity and the total femininity in all cases where a pair of living beings come together with the maximum sexual attraction.