As part of a longitudinal study of the respiratory health effects of air pollution, we measured the lung function of 2,454 white adults 25 to 74 yr of age who had never smoked and who reported no respiratory symptoms. These measurements were analyzed to develop a simple model for the cross-sectional dependence of pulmonary function on height, sex, and age. Both forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) can be effectively standardized for body size by dividing each pulmonary function measurement by the square of the standing height (HT2). The age-specific distribution of these standardized measurements is approximately Gaussian, with variance that is independent of age. Plots of FEV1/HT2 and FVC/HT2 against age showed a nonlinear relationship consistent with an increase in the rate of pulmonary function loss with age. On the basis of these graphic analyses, both pulmonary function measurements were fitted to a four-parameter normative model including sex and linear and quadratic terms in age as dependent variables. This model gave predictions that were very close to those from more complicated models currently in use. Predicted percentile levels were calculated for each sex and age, and shown to describe the observations well. The estimated annual change in height-standardized lung function based on the cross-sectional model was compared with the observed change between the first and second examinations of these adults 3 yr later. The observed changes were close to predicted values, except for subjects younger than 35 yr of age at their first examination. The observed change was larger for men than for women. Such simple longitudinal comparisons are subject to selection bias. In this study, subjects in the lowest quartile of FEV1/HT2 for their age and sex at the first examination had a lower probability of providing a lung function measurement 3 yr later.