We sought to characterize the physical performance of older subjects in a community-based study of aging and fitness and to establish norms of oxygen consumption during exercise in this group. A community-based census identified 3057 age-eligible individuals of whom 2092 individuals (68.4%) agreed to participate in the overall study. Of these, 1101 subjects between 55 and 94 years of age who were free of known heart disease and able to perform treadmill exercise were tested using the Cornell protocol. Of these 1101 subjects, 73.8% of women and 89.5% of men (p < .001) exceeded what was considered a minimal threshold workload [i.e., > or =2 min of exercise and a respiratory exchange ratio (RER) > or = 1.00]. Moreover, 32.9% of women and 52.7% of men achieved a RER > or = 1.10 and were considered to have achieved maximal exercise. For all subjects who exceeded the minimal exercise workload, the mean exercise duration was 10.0 +/- 4.0 min for women and 13.2 +/- 4.6 min for men. Peak VO2/kg x min(-1) decreased linearly with age for both women (y = 38.33 - 0.27 x age + 1.12 H) and men (y = 41.78 - 0.26 x age + 12.65 H - 0.15 H x age), (p < .001 for sex difference), where H is defined as presence (=1) or absence (=0) of a disease condition. However, this sex-related difference disappeared when VO2/kg x min(-1) was adjusted for lean body mass. Thus, from a community-based sample of older persons, we have provided data (by 5-year age groups) for rates of participation in treadmill exercise testing and the success rate for achieving maximal exercise. Oxygen consumption and other exercise data are presented for older, healthy subjects as well as for those with various chronic diseases associated with aging (excluding those with cardiac or cerebrovascular disease). Exercise duration, peak VO2/kg x min(-1) (normalized for total and lean body mass), and peak exercise heart rate declined with age. Most of the sex difference in peak VO2/kg x min(-1) could be explained by the greater percent of body fat in women.