Introduction: Against a background of variable results achieved by the treatment of childhood autism with dopaminergic agonists or antagonists, and opiate antagonists and mild behavioural improvements achieved following treatment with fenfluramine (Geller et al., 1982; Ritvo et al., 1984 among others) a 12 month double-blind crossover trial of fenfluramine was undertaken. [the d-isomer releases serotonin, the l-isomer modestly blocks dopamine receptors.] Methods: Six female and 14 male children and adolescents with autism were enrolled (median 9y; range 5-17 years of age) in a trial run over a 1 month baseline, 5 months of treatment, 2 months placebo, and 5 months of treatment. Parent diaries were kept. Various tests and rating scales for development, verbal and noverbal abilities along with measures of blood serotonin, urinary catecholamines and auditory event-related potentials (ERPs) were measured at 6-month intervals. Results: a) With fenfluramine treatment children lost weight, blood serotonin decreased by 60% and urinary catecholamine levels fell (25-45%, but dopamine utilization (HVA/DA) increased 2-4-fold. b) Under fenfluramine some autistic subjects improved on measures of cognitive and language function. Two mute subjects began to speak, but for the group overall the improvements did not prove significant. [British Ability Scale, Reynell developmental language and Vineland social maturity scale improvements were significant.] c) Treatment improved early stages of information processing, but impaired later stages as measured in the ERP (i.e. early negative going excitatory vs. later positive going inhibitory response, respectively in 7 patients). The early negativity changes (especially over the right hemisphere) correlated inversely with HVA/DA actvivity. d) Side effects reported included lethargy and irritability. Conclusions: While individuals showed marked reductions of stereotypies, hyperactivity, and improvements on develepmental and cognitive measures, the absence of clear group-wide effects suggests that fenfluramine would have but a limited place in the management of some patients with autism.