Abstract Naltrexone (0.05–5.0 mg/kg, SC) was administered to food-deprived rats prior to a 15-min food-preference test. Total food intake and feeding duration was reduced following administration of the opiate antagonist. However, while naltrexone reduced the consumption of the initially-preferred chocolate-coated cookies, the ingestion of the nonpreferred standard laboratory chow pellets was significantly enhanced. These data cannot be explained in terms of a general anorexic effect and nonspecific suppression of feeding responses. Instead, they indicate that naltrexone reduced preference for the highly palatable cookies, so that a feeding response to the chow pellets emerged. Under the conditions of test-familiarity, naltrexone did not reduce grooming, locomotion or rearing duration. An increase in locomotion may have been secondary to the reduction in feeding. The results agree with previous data from animal and human studies in suggesting that endogenous opioid peptide activity is involved in the palatability of preferred foods.