Some of the psychopharmacological characteristics of melatonin have been defined using a number of established tests of sedative/hypnotic, anticonvulsant and analgesic activity in mice and rats. The effect of melatonin in these tests has been compared to its neurotoxicity and acute toxicity. In the mouse, a low dose of melatonin (20 mg/kg i.p.) potentiated pentobarbitone- and barbitone-induced sleep. Melatonin also potentiated pentobarbitone-induced sleep in the rat. Higher doses (greater than or equal to 200 mg/kg i.p.) antagonized pentylenetetrazole, 3-mercaptopropionic acid and electroshock-induced convulsions in mice and had analgesic activity in both hot-plate and writhing tests. The presence of motor incoordination, indicated by the rotorod test, after administration of these large doses (greater than or equal to 200 mg/kg i.p.) suggests that the anticonvulsant and analgesic activities of melatonin may not represent specific neuropharmacological actions. LD50 values for melatonin in the mouse and rat were determined for different routes of administration. A sedative dose of melatonin (20 mg/kg i.p.) did not alter whole brain 5-hydroxytryptamine or 5-hydroxyindole acetic acid concentrations, suggesting that the hypothesis that the sedative action of melatonin is due to an interaction with serotoninergic neurons may need to be re-examined.