One-thousand-seven-hundred-twelve men aged 40-59, representing two demographic samples of rural areas in Northern and Central Italy, were examined in 1960 and then followed-up for 25 years. Total mortality and coronary mortality were related to the entry measurements of physical activity at work, estimated in three levels by a simple questionnaire and to four possible indicators of physical fitness, ie resting heart rate, the circumference of the right arm cleaned from the contribution of skin and subcutaneous fat, vital capacity and forced expiratory volume in 3/4 sec, the latter two adjusted by height. The four indicators of fitness and the score of physical activity were somewhat related each other in the expected direction. Both all causes and coronary mortality were indirectly related to the score of physical activity. The relative risk for age adjusted rates of sedentary vs heavy workers was 1.23 for all causes of death, and 1.72 for coronary deaths. Direct univariate relationships were found between resting heart rate and mortality (for both all causes and coronary deaths) whereas indirect relationships were found for arm circumference and the two indicators of respiratory function. The age adjusted relative risks between the upper and the bottom quartile of the distribution were, for all causes of death, 0.75 for vital capacity, 0.68 for forced expiratory volume 1.32 for heart rate, and 0.84 for arm circumference. The correspondent relative risks for coronary deaths were 0.71, 0.65, 1.23 and 1.02 respectively.(ABSTRACT TRUNCATED AT 250 WORDS)