The authors followed 775 men (aged 18-98 years) participating in the Baltimore Longitudinal Study in Aging for an average of ten years. Resting metabolic rate and fasting respiratory exchange ratio (RER) were measured by indirect calorimetry on their first visit and related to subsequent weight change. Deviations from the predicted value of resting metabolic rates (predicted from their estimated fat-free mass) were calculated. Average weight change was 0.07 kg (s.d. 6.4 kg); 122 men (15.3%) gained more than 5 kg and 40 (5.2%) more than 10 kg during the follow-up. After adjustment for initial age, body mass index, fat-free mass, and duration of follow-up, RER, but not RMR or deviations from predicted RMR, was positively related to weight change (P less than 0.001). Major weight gain (from at least 5 kg to at least 15 kg) was related to initial RER in non-obese men only (initial body mass index less than 25 kg/m2). From Cox proportional hazard regression analyses the adjusted relative risk of gaining 5 kg or more in initially non-obese men with a fasting RER of 0.85 or more was calculated to be 2.42 (95% confidence interval: 1.10-5.32) compared to men with a fasting RER less than 0.76. It was concluded that a relatively high fasting RER is a weak but significant predictor of substantial weight gain in non-obese white men.