A variety of the anticonvulsant drugs, including carbamazepine, phenytoin, primidone, phenobarbital, clonazepam, valproic acid, and ethosuximide, are available for use in the treatment of patients with seizure disorders. These anticonvulsants vary in their efficacy against experimental seizures in animals and against seizures in humans. The mechanistic basis for this variability in anticonvulsant drug action remains uncertain, but numerous mechanisms of action have been proposed. We have used mouse neurons in primary dissociated cell culture to study the action of these anticonvulsant drugs on several aspects of membrane excitability and synaptic transmission. We have proposed that the anticonvulsant drugs can be classified according to their actions on sustained high frequency repetitive firing (SRF) of action potentials and on postsynaptic gamma-aminobutyric acid (GABA) responses. Phenytoin and carbamazepine were both effective against SRF but did not modify postsynaptic GABA responses at therapeutically relevant concentrations. Phenobarbital, benzodiazepines, and valproic acid modified both SRF and postsynaptic GABA responses. Ethosuximide had no effect on SRF or GABAergic mechanisms. Based on these results, we have proposed that blockade of SRF may underlie the action of phenytoin, carbamazepine, phenobarbital, valproic acid, and benzodiazepines against generalized tonic-clonic seizures in humans and maximal electroshock seizures in animals. Enhancement of GABAergic synaptic transmission may underlie efficacy of benzodiazepines and valproic-acid drugs against generalized absence seizures in humans and pentylenetetrazol-induced seizures in experimental animals. The mechanism of action of ethosuximide against generalized absence seizures in humans and pentylenetetrazol-induced seizures in experimental animals may be by a third, as yet unknown, mechanism.