In this paper, we presented a new procedure to determine daily minimum energy expenditure (MEE) in free-moving rats and mice kept in a chamber. Energy expenditure was measured for 23 h period and averaged every 10 min. Data were sorted in ascending order. MEE was estimated from the regression line with the smallest slope and the biggest intercept among the regression lines calculated between the sorted energy expenditure and the data ranking. Among three duplicate measurements in individual animals, MEE gave the smallest coefficient of variation (2.2%) as compared with actual measured-values: either the single lowest value (4.0%) or the average of the 6 lowest values (2.5%). Judging from diurnal patterns of energy expenditure and locomotor activity and from video tape observation of the rat's performance, it was confirmed that MEE represented an energy expenditure at rest. MEE decreased with fasting from days 1 to 5. MEE per body weight also declined with age, but stayed around 71-72 kcal/day/kg3/4 at 18 and 34 weeks of age in male Wistar rats. MEE in mice increased at lower ambient temperatures between 16 and 32 degrees C, but stayed fairly constant at the same temperature in repeated experiments. Thus, MEE estimated by the present regression procedure was highly reproducible and valid to determine the fundamental value of daily energy expenditure under free-moving conditions.