We studied physical fitness and risk of all-cause and cause-specific mortality in 10,224 men and 3120 women who were given a preventive medical examination. Physical fitness was measured by a maximal treadmill exercise test. Average follow-up was slightly more than 8 years, for a total of 110,482 person-years of observation. There were 240 deaths in men and 43 deaths in women. Age-adjusted all-cause mortality rates declined across physical fitness quintiles from 64.0 per 10,000 person-years in the least-fit men to 18.6 per 10,000 person-years in the most-fit men (slope, -4.5). Corresponding values for women were 39.5 per 10,000 person-years to 8.5 per 10,000 person-years (slope, -5.5). These trends remained after statistical adjustment for age, smoking habit, cholesterol level, systolic blood pressure, fasting blood glucose level, parental history of coronary heart disease, and follow-up interval. Lower mortality rates in higher fitness categories also were seen for cardiovascular disease and cancer of combined sites. Attributable risk estimates for all-cause mortality indicated that low physical fitness was an important risk factor in both men and women. Higher levels of physical fitness appear to delay all-cause mortality primarily due to lowered rates of cardiovascular disease and cancer.