Current recommendations for energy requirements in the elderly are based on assumed levels of physical activity relative to BMR (1.5 x BMR). The main aim of the present study was to establish whether these recommendations might be applicable to a randomly-selected group of free-living elderly men (all over 75 years of age). BMR was measured by indirect calorimetry and total energy expenditure (TEE) by the doubly-labelled-water technique. Further aims included evaluating the applicability of a variety of BMR prediction equations and whether assessed quality of life reflected any measured indices of energy expenditure. The mean value for daily energy requirement was found to be 1.5 x BMR (89 J/kg per min) but with substantial inter-individual variation (SD 0-2 x BMR; 14 J/kg per min). The bias between measured TEE and TEE estimated (1.5 x BMR) from the various BMR predictions varied according to which equation was used (-10-+ 8% of the mean) with substantial 95 YO limits of agreement (28-30%of the mean). TEE and physical activity plus thermogenesis (TEE -BMR) were positively related to activities of daily living, but no relationships were apparent between these and perceived quality of life. It is concluded that, despite considerable inter-individual variability, national recommendations for energy requirements of elderly people are applicable to this randomly-selected group of free-living men over 75 years of age but that substantial variation exists when attempts are made to estimate TEE from measurements or predictions of BMR.