data while in other cases families having some years of possible productiveness ahead are included with others and the relative proportion of the types is not indicated. The error in these cases is therefore important and the reader is warned to accept them only with a mental allowance for this factor.
The best students make an even worse showing in this respect. The Wellesley alumnæ who are members of Phi Beta Kappa,—that is, the superior scholars—have not .86 of a child each, but only .65 of a child; while the holders of the Durant and Wellesley scholarships, awarded for intellectual superiority,5