heritable. We name the former stable characters, the latter unstable, or fluctuating characters. New qualities are arising from generation to generation through variation. These variations may similarly be classed as stable variations, or mutations, and fluctuating, or unstable, variations. No result can be reached by breeding with reference to unstable variations or qualities, for they are not inherited. Qualities belonging to the unstable type can not be fixed by breeding. They are, therefore, without significance in the problems of eugenics and evolution. It is impossible, however, to discern whether an observed quality is of the stable or unstable type until one follows its behavior in inheritance.
Another fact of the greatest importance to remember is that there is probably no such thing as inheritance of vague general resemblances, but that inheritance is apparently always particular, definite, so-called unit qualities being the things inherited. The character of any individual is built up of a complex multitude of such unit qualities, each heritable separately, and the character of an individual depends upon the combination and interaction of the unit qualities that have been passed down to him from his parents, grandparents and other progenitors.
In the light of these facts, what is the essential problem, first in eugenics, then in evolution? The eugenics problem is accurately to determine the desirable unit qualities, which must be of the stable type, and to combine and fix them in the race by breeding, eliminating at the same time the undesirable unit qualities. It is the problem of finding the exact units of inheritance, and of so fixing and combining, by breeding, these valuable units in the individuals of the coming generations that we shall have a more wholesome innate character in mankind. The evolution problem is to find among the multitude of diverse human traits new desirable unit qualities of the stable type, often only in their beginnings, and to perpetuate these by breeding.
The Galton-Pearson school of English students are willing to waive accurate analysis of inheritance units, but the real problem will not be solved until we know whether the human qualities with which we wish to deal, the intellectual and moral as well as the physical, do follow the Mendelian principles in inheritance, and until we have analyzed the Mendelian qualities to their units. We have a notable example of failure to secure permanent valuable results in attempting to breed from individuals whose valued character had not been analyzed to its unit qualities. At the Agricultural Experiment station in Orono, Maine, many years of effort were given to securing a strain of fowl which would lay an unusually large number of eggs. Mere breeding from hens which laid many eggs was not found to be enough. The quality of high fecundity could not be fixed in the strain. Selection had to be continued in each generation or reversion to the general average would