In the factorial estimation of total energy expenditure it is assumed that the intra-individual variation in RMR is small. Little is known about the intra-individual variation in RMR in older subjects. The present study investigated the intra-individual variation in RMR in older people. Measurements of RMR were made in twenty-seven older subjects, mean age 71·6 (sd 6·1) years, on two separate occasions (T1 and T2) and on a third occasion (T3) in nineteen of the subjects. Measurements of height and weight were taken in all subjects. RMR measurements were made in the laboratory using a Deltatrac™ (ventilated-hood indirect calorimeter; Datex, Helsinki, Finland). All subjects had fasted overnight for 12h and refrained from strenuous exercise before measurements. The intra-individual CV in RMR (kJ/d) after T1 and T2 was 2·5% in women and 3·6% in men and was 2·6% in women and 3·4% in men after all three sets of measurements. Although mean RMR did not vary across T1, T2 and T3, there was significant ‘crossing tracks’ across the three measurement occasions in some individuals, reflecting a high degree of within-subject variability. The methods used had a significant measurement error associated with them (high R value; significant F ratio in three-way ANOVA). In conclusion, the results from the present study indicate that intra-individual variation in RMR was low in older people. The intra-individual variation in the elderly is similar to that seen in younger age groups.