The purpose of this study was to determine if the beneficial effect of naloxone on formation of acute gastric mucosal lesions was brought about via central or peripheral mechanisms by measuring blood concentrations of naloxone in rats during a 4-hr period of restraint stress. The study involved administration of naloxone to rats at doses of 5, 20 and 40 mg.kg-1.hr by either the intravenous or enteral routes. Blood samples were collected throughout the period of restraint and gastric stress-lesions were counted at the end of the experiments. Both routes of administration were equally effective in preventing stress-ulceration, with only rats receiving drug intravenously showing the presence of naloxone in blood samples. Inverse linear relationships existed between mean trough blood concentrations and lesions (P = .0003), as well as a linear correlation between area under the time-concentration curve and mean trough concentrations (P = .0001). Although our results show tight correlation between blood levels and effect on lesions in the group given drug intravenously, the effect must be on peripheral rather than central opiate receptors as no detectable blood levels were found when naloxone was given enterally.