1. The present study has investigated whether the neurotoxic effects of the relatively selective 5-hydroxytryptamine (5-HT) neurotoxins, 3,4-methylenedioxymethamphetamine (MDMA or 'Ecstasy'), p-chloroamphetamine (PCA) and fenfluramine on hippocampal and cortical 5-HT terminals in rat brain could be prevented by administration of either chlormethiazole or dizocilpine. 2. Administration of MDMA (20 mg kg-1, i.p.) resulted in an approximate 30% loss of cortical and hippocampal 5-HT and 5-hydroxyindoleacetic acid (5-HIAA) content 4 days later. Injection of chlormethiazole (50 mg kg-1) 5 min before and 55 min after the MDMA provided complete protection in both regions, while dizocilpine (1 mg kg-1, i.p.) protected only the hippocampus. 3. Administration of a single dose of chlormethiazole (100 mg kg-1) 20 min after the MDMA also provided complete protection to the hippocampus but not the cortex. This regime also attenuated the sustained hyperthermia (approx +2.5 degrees C) induced by the MDMA injection. 4. Injection of PCA (5 mg kg-1, i.p.) resulted in a 70% loss of 5-HT and 5-HIAA content in hippocampus and cortex 4 days later. Injection of chlormethiazole (100 mg kg-1, i.p.) or dizocilpine (1 mg kg-1, i.p.) 5 min before and 55 min after the PCA failed to protect against the neurotoxicity, nor was protection afforded by chlormethiazole when a lower dose of PCA (2.5 mg kg-1, i.p.) was given which produced only a 30% loss of 5-HT content. Chlormethiazole did prevent the hyperthermia induced by PCA (5 mg kg-1), while the lower dose of PCA (2.5 mg kg-1) did not produce a change in body temperature.5. Neither chlormethiazole nor dizocilpine prevented the neurotoxic loss of hippocampal or cortical 5-HT neurones measured 4 days following administration of fenfluramine (25 mg kg-1, i.p.).6. In general, chlormethiazole and dizocilpine were effective antagonists of the 5-HT-mediated behaviours of head weaving and forepaw treading which appeared following injection of all three neurotoxins.7. Both chlormethiazole and dizocilpine have previously been shown to prevent the neurotoxic effects ofa high dose of methamphetamine on cerebral 5-HT and dopamine pathways. These drugs also prevent MDMA-induced neurotoxicity of 5-HT pathways, but not that induced by injection of PCA or fenfluramine. This suggests that the mechanisms of neurotoxic damage to 5-HT pathways produced by substituted amphetamines cannot be identical. The monoamine loss does not appear to result from the hyperthermia produced by the neurotoxic compounds.