A regression analysis was made of age at first reproduction in female mammals, as a function of body weight, using the data of Wootton. Data on maximal life span, also expressed as a function of body weight, were used to calculate "adult" life span, wherever possible, by subtracting the cognate value for age at first reproduction. Then a regression analysis of adult life span as a function of age at first reproduction was made. In both cases global regression lines (i.e., for whole data sets) were computed by standard least squares and by a robust method, as well as local regression lines for subgroups classified by taxonomic and ecological criteria. The slopes of the various regression lines were found to vary widely as a function of the method of classification. This result argues against the notion that the ratio of life history variables is a constant, or that one life history variable is likely to be a simple function of another. The results for bats are anomalous, in that age at first reproduction appears to be independent of body weight (over about two orders of magnitude). It is concluded that a full understanding of life history variables, such as maximal life span and age at maturity, is likely to depend on combined physiological, ecological, and evolutionary insights.