This supposition will be recognized by many of you as a simplified case of Mendelian inheritance of a unit character due to the presence or absence of a single determiner which can either be or not be in a germ or ovum, and which "segregates."
No case quite so simple as this can be true of human intellect, but something approximating it has been suggested as perhaps true.
Suppose, on the other hand, that the germinal basis for intellect consists in the presence, in the germ or ovum, of one or more of four determiners—, and —contributing amounts 1, 2, 3 and 4 of intellectual capacity. The fertilized ovum could then have any one of 256 different constitutions ranging from the entire absence of all these determiners to the presence of each one "duplex"—i. e., in both germ and ovum. If such duplex presence meant that the two contributions combined additively, the original intellect of the individual could range from to 20. Individuals, all of one same original intellect—10—might be of very different germinal constitutions, and so of very different possibilities in breeding. If two individuals, each of original intellect 10, were mated, it might be the case that their possible offspring would range in intellect from to 20, or it might be that they could not go below 8 or above 12.
If the number of germinal determiners of intellect is increased to five or six, the task of telling the constitution of the germs produced by any individual of known original intellectual capacity is enormously increased; and the research needed to guide the best possible breeding of man is very, very much more laborious. Moreover, instead of hoping to bring man to the best possible status (subject to the appearance of new desirable mutations) by a few brilliant rules for marriage, we must then select indirectly and gradually by parental achievement rather than directly by known germinal constitution, just as animal and plant breeders had to do in all cases until recently, and just as they still have to do in many cases. Only after an elaborate system of information concerning family histories for many generations is at hand, can we prophesy surely and control with perfect economy the breeding for a characteristic which depends on the joint contributions of five or six determiners. For it is just as hard to "breed in" a determiner that raises intellect or morality only one per cent, as it is to "breed in" one which raises it a hundred per cent.—provided, of course, the latter determiner exists. And it is thousands of times harder to discover the distribution of a determiner in the human race's germs when it is one of ten that determine the amount of a trait, than when it is one of two.
The germinal determination of intellect, morality, sanity, energy or skill is, so far as I can judge, much more like the second complex state of affairs than the first simple one. Important observations of the inheritance of feeble-mindedness and insanity have been made by Daven-