Whether a cost of reproduction exists among humans is still questionable. A major study of aristocratic British families finds a significant positive correlation between parity and late–life mortality, which indicates a trade–off between reproduction and longevity. This result is supported by four other studies, while earlier studies have not found a relationship or came to the opposite conclusion. We show that in natural fertility populations the relationship between fertility and late–life mortality cannot be studied correctly without considering the effects of differences in health and of mortality selection during childbearing ages because these two effects lead to a dampening of the true relationship. If these effects are controlled in Hollingsworth's genealogy of the British peerage a significant trade–off between reproduction and longevity exists for females but not for males.