But the sign changes and becomes positive when we come to other comparisons. When we count the relative number of pawnbrokers and general dealers, of "general laborers" (that is, men without a trade and without regularity of occupation and employment), of employed children between the ages of ten and fourteen, of persons living more than two in a room, when we consider the infant death rate, the death rate from pulmonary tuberculosis, and the relative number of paupers,--then we find the signs of the coefficients are all positive, and on the average the coefficients are more than 0.50--a moderate to high degree of regularity of the relation. The districts characterized by the larger numbers of such individuals or by higher death rates of these kinds, are at the same time the districts where the birth rates are the higher. In a word, then, Heron found that the greater the number of professional men, or of servants employed in a community, the lower the birth rate--a very high degree of negative correlation. On the other hand, the more pawnbrokers, child laborers, pauper lunatics, the more overcrowding and tuberculosis, the higher the birth rate--a high degree of positive correlation. Little doubt here as to which elements of the city are making the greater contributions to the next generation. There may be some doubt, however, so let us consider two possible qualifications of these results. First, is not the death rate also higher among these least desirable classes? Yes, it is. Is it not enough higher to compensate for the difference in the birth rates, so that after all the least desirable classes are not more than replacing themselves? No, it is not. After calculating the effect of the differential death rate among these different social groups it still remains true that the _net_ fertility of the undesirables is greater than the _net_ fertility of the desirables: the worst classes are in reality more than replacing themselves numerically in such communities; the most valuable classes are not even replacing themselves. Second, is not this the same condition that has always existed in these districts? Why any cause for supposing that this is going to bring new results to this society? Has not such a condition always been present and always been compensated for somehow? Fortunately, Heron is able to compare with these data of 1901 similar data for 1851, and is able to show that every one of these relations has changed in sign since that date--in fifty years. The significance of this change in sign is probably clear. It means here that in London sixty years ago there was a high degree of regularity in the relation such that the more professional men and well-to-do families the community contained, the higher the birth rate; that ten years ago this had all become changed so that the more of these desirable families found in a district the lower is the birth rate. It means that sixty years ago the relation was such that the more undesirables