Serotonin can induce analgesia when injected directly into the brain, but analgesia after peripheral administration has been more difficult to show. The pentobarbital anesthetized mouse (PAM) model, developed to alleviate some of the problems involved in the measurement of tail flick latency, was used to assess the action of peripherally administered serotonin. Mice were anesthetized with about 65 mg/kg of sodium pentobarbital IP and their tail flick latencies measured while they were in stage III anesthesia. In these anesthetized mice, IP serotonin induced a significant analgesia that was much more robust than that found in awake mice. The analgesic effect was dose-dependent from 0.25 mg/kg to 10 mg/kg but was not blocked by the antiopiate naltrexone. Of several psychotropic agents tested, only amitriptyline, mianserin, and trazodone had significant effects on analgesia in the PAM model. The analgesic effect of serotonin was reproduced by the 5HT2 agonist DOI and totally blocked by the 5HT2 antagonist NPP. These results show the utility of the PAM model in studying nonopiate analgesia and suggest that the analgesic action of serotonin is mediated primarily through the 5HT2 receptor.