1. The metabolic effects of increasing or decreasing the usual energy intake for only 1 d were assessed in eight adult volunteers. Each subject lived for 28 h in a whole-body calorimeter at 26 degrees on three separate occasions of high, medium or low energy intake. Intakes (mean +/- SEM) of 13830 +/- 475 (high), 8400 +/- 510 (medium) and 3700 +/- 359 (low) kj/24 h were eaten in three meals of identical nutrient composition. 2. Energy expenditure was measured continuously by two methods: direct calorimetry, as total heat loss partitioned into its evaporative and sensible components: and indirect calorimetry, as heat production calculated from oxygen consumption and carbon dioxide production. For the twenty-four sessions there was a mean difference of only 1.2 +/- 0.14 (SEM) % between the two estimates of 24 h energy expenditure, with heat loss being less than heat production. Since experimental error was involved in both estimates it would be wrong to ascribe greater accuracy to either one of the measures of energy expenditure. 3. Despite the wide variation in the metabolic responses of the subjects to over-eating and under-eating, in comparison with the medium intake the 24 h heat production increased significantly by 10% on the high intake and decreased by 6% on the low intake. Mean (+/- SEM) values for 24 h heat production were 8770 +/- 288, 7896 +/- 297 and 7495 +/- 253 kJ on the high, medium and low intakes respectively. The effects of over-eating were greatest at night and the resting metabolic rate remained elevated by 12% 14 h after the last meal. By contrast, during under-eating the metabolic rate at night decreased by only 1%. 4. Evaporative heat loss accounted for an average of 25% of the total heat loss at each level of intake. Changes in evaporative heat loss were +14% on the high intake and -10% on the low intake. Sensible heat loss altered by +9 and -5% on the high and low intakes respectively. 5. It is concluded that (a) the effects on 24 h energy expenditure of over-feeding for only 1 d do not differ markedly from those estimated by some other workers after several weeks of increasing the energy intake: (b) the resting metabolic rate, measured at least 14 h after the last meal, can be affected by the previous day's energy intake; (c) the zone of ambient temperature within which metabolism is minimal is probably altered by the level of energy intake.