On the basis of their answers to a self-administered questionnaire, 697 nonsmoking healthy subjects were chosen from a randomly selected sample representative of the white non-Mexican-American population of Tucson, Arizona, enrolled in a longitudinal study of respiratory health. For each subject, the first satisfactory set of flow-volume data obtained during the first 3 consecutive surveys was selected for analysis. For forced vital capacity (FVC) and forced expiratory volume in one second (FEV1), the single best value for each subject was selected. Other flow-volume measurements were derived from the single test with the best sum FEV, plus FVC. These data were used to derive improved prediction equations for each sex by age group for 5 spirometric and flow-volume variables. The resulting predicted values demonstrate the effects of development, maturation, and senescence on ventilatory function. "Normal" limits are proposed that take into consideration the between-subject variability and non-Gaussian distribution of the various measurements.