There is another hypothesis which might account for the observed changes of type. If it were assumed that among the descendants of immigrants born in America there are an appreciable number who are in reality children of American fathers, not of their reputed fathers, a general assimilation by the American type would occur. Socially this condition is not at all plausible; but, on account of the importance of the phenomenon that we are discussing, it should be considered. I do not think that any of the observations that have been made are in favor of this theory. The changes that occur in the Bohemians who arrive here as young children, the different directions of the changes in distinct types, particularly the shortening of the head of Bohemians and of Italians, do not favor the assumption. Furthermore, if the modifications were due to race-mixture, the similarity between fathers and American-born children should be less than the similarity between fathers and foreign-born children, but there is no indication that this is the case. This hypothesis is also shown to be untenable by the comparisons of fathers and mothers with their own foreign-born children. These comparisons show that the differences are the same in the case of fathers and children, and of mothers and children; so that obviously the same conditions must control the relations between fathers and their children, and mothers and their children. In other words, the fathers must be considered as the true fathers of their children. Earnest advocates of the theory of selection might claim that all these changes are due to the effects of changes in death-rate among foreign-born and American-born; that either abroad or here individuals of certain types are more liable to die, and that thus these changes are gradually brought about. On the whole, it seems to my mind, the burden of proof would be entirely on those who claim such a correlation between head-index, width of face, etc., and death-rate,—a correlation which I think is highly improbable, and which could be proposed only to sustain the theory of selection, not on account of any available facts. I grant the desirability of settling the question by actual observations; but, until these are available, we may point out that the very suddenness of the changes after immigration, and the absence of changes due to selection by mortality among the adult foreign-born, would require such a complicated adjustment of cause and effect in regard to the correlation of mortality and bodily form, that the theory would become improbable on account of its complexity.