Effects of nalmefene on eating were investigated in two groups of ten male volunteers, in a double-blind placebo-controlled study. The nalmefene treated group ate 22% less, both in terms of absolute weight and caloric intake, of a standardised buffet-meal than did the placebo group. No differences in subjective ratings of hunger or satiety were found between the groups, suggesting that the reduced feeding was not a consequence of any change in motivation to eat. When analysed by nutrient content, nalmefene was found to reduce fat and protein, but not carbohydrate, intakes. Analyses of intakes of individual foods showed a differential effect of nalmefene on foods rated as highly palatable. Thus the apparent nutrient specificity of nalmefene appeared to be an indirect consequence of its effect on palatability. Nalmefene also caused slight increases in self-rated alertness, and decreases in ratings of tiredness and elation, although it was thought unlikely that these accounted for observed changes in eating behaviour. No other side-effects were detected, and performance on a choice reaction time task was unaffected. These results add weight to suggestions that endogenous opioids are involved in reward-related aspects of feeding associated with food palatability.