Previous anatomical investigations have reported a direct projection from substantia nigra pars lateralis to the dorsal midbrain anticonvulsant zone. The present study tested the hypothesis that the anticonvulsant properties of nigral inhibition previously attributed to substantia nigra pars reticulata were, in fact, due to the suppression of neural activity in the adjacent pars lateralis. Using the electroshock model of epilepsy, a systematic map of the anticonvulsant effects of bilateral injections of muscimol (60 ng/0.5 mu l per side) into different parts of substantia nigra was constructed. Electroshock (1 s of 40 mA 50 Hz AC) was administered via ear-clip electrodes 5 or 60 min following injections of muscimol, or 60 min after control injections of saline. To provide insight into the functional mechanisms whereby nigral inhibition might suppress tonic seizures the behavioural effects elicited by muscimol were also noted. No evidence supporting the experimental prediction was found. The most sensitive region of substantia nigra for suppressing tonic hindlimb extension was caudal pars reticulata. These data indicate a serious mismatch between the results of microinjection mapping studies and underlying patterns of anatomical connectivity. The behavioural reaction most closely associated with tonic seizure suppression was stereotyped locomotion; both were obtained maximally from caudal pars reticulata. Rostral substantia nigra was associated more with oral stereotypy, while a raised head position was observed at lateral injection sites and a lowered positioning of the head at medial locations. These data suggest that the rat substantia nigra may contain a functional organization based on a form of somatomotor topography. This organization may influence which part of the substantia nigra is most effective in suppressing seizures expressed by different muscle groups of the body.