averaging not far from two among the highly skilled. In each group of occupations I have added the Word “young” to the upper line because the men in that particular occupation average younger than the others, and therefore their families are not so near completion as are those of older men, such as foremen, bankers, officials, and the other higher categories. It seems safe to say that at least six children, and perhaps more, are born to the average mine operative, and that not less than five of these live well beyond infancy. On the other hand, it is very doubtful Whether the chemists, assayers, and metallurgists in our table will have an average of as many as three children, so that not much over two and a half Will survive far into childhood. After infancy is past the chemists’ children may have nearly as high a death-rate as the children of the miners. Snow, in a paper on The Intensity 0f Natural Selection in M an, has confirmed Pear- son’s conclusion that if the death-rate is high in the first year of life, as it is among the children of mine operatives, it is relatively low in succeeding years. On the contrary, a low death-rate in infancy means that the later death-rate is higher than it would have been under the same conditions of environment if the weaker infants had perished.

The case of the chemists appears still Worse when we remem- ber that even though the proportion of married men (91 per cent) among a picked body of leading scientists has been found by Cattell to be as large as in the population as a whole, the proportion among Harvard graduates over fifty years of age is only 72 per cent. Moreover, in both groups the number of child- less marriages is large because of the relatively advanced age of the wives. For example, among leading men of science whose families Were presumably complete, 22 per cent of those who Were married Were childless. Again among these scientific men the average completed family amounts to only 2.2 3 children. As the death-rate among the children up to the usual age of marriage is about 12o per thousand, the average number of children reaching that age is 1.96. This, however, takes no ac- count of the scientists who are not married. If they are included,


l l l l l l l l l