Abstract Owing to the long-lasting depletion in brain serotonin (5-HT) produced by high doses of the amphetamine derivative dl-fenfluramine, this drug has been suspected of neurotoxicity; the serotonergic innervation of the hippocampal formation being one of its most vulnerable targets. A first series of experiments has been carried out to determine, through a 5-HT immunohistochemical study in the rat, the validity of this claim, as well as the rate and the degree of 5-HT recovery. Five to 12 days after a high intraperitoneal dose of dl-fenfluramine (26.8 mg/kg) there is a profound, increasing reduction in the density of serotonin-immunoreactive (5-HT-IR) axons in all portions of the hippocampal formation. The fine-caliber fibers being more affected than the thicker axons. In parallel with this decrease, numerous 5-HT-IR fibers in the cingulum bundle, fornix-fimbria and the hippocampal parenchyma exhibit structural changes, similar to those produced by 5,7-dihydroxytryptamine, a specific neurotoxin of the 5-HT system. By 20 days after the treatment, abnormal axons have disappeared and the density of 5-HT-IR normal axons is markedly increased. By 40 days, the pattern of 5-HT hippocampal innervation is restored but its density is still lower than in controls. These findings, which corroborate the neurotoxic action of fenfluramine, provide anatomical evidence that the chemical ablation is followed by a slow process of axonal regeneration, started between 12 and 20 days and still uncompleted at 40 days. A second series of experiments, with chronic oral administration of dl-fenfluramine, has been conducted to determine presumptive potentiation through a cumulative effect of the drug, and to analyze the recovery process. One day after the end of the treatment with high dose regimens (210 mg/kg, distributed into 5 mg/kg twice daily for 21 days), together with a marked decrease in the density of 5-HT-IR fibers there are some rare forms of axonal pathology. 15 days later, altered axonal forms have disappeared and the density of 5-HT-IR axons has considerably increased, reaching normal levels by 30 days (56 days after the first exposure to the drug). Hence, despite the ability of the fenfluramine to exert its action on 5-HT axons all over the duration of the treatment, it does not exert a noxious cumulative effect as revealed by the low level of axonal pathology noticeable 22 days after the initiation of the treatment, and the relatively fast and complete recovery. Finally, chronic oral treatment with low dose regimens of dl-fenfluramine (42 mg/kg distributed into 1 mg/kg, twice daily for 21 days) produces depletion of hippocampal 5-HT-IR fibers without any apparent axonal damage. Moreover, under these conditions, the recovery is completed by 15 days after the end of the treatment. The absence of axonal degeneration in this last group of rats, and the fast recovery to normal levels of serotonergic hippocampal innervation indicate that the mechanisms of action of these low oral dosis of dl-fenfluramine do not involve degeneration of 5-HT fibers.