In continuation of an earlier study of our group (Neurology, 43 (1) (1993) 41-51), we present the results of an investigation of the adverse effects of carbamazepine versus phenytoin on cognitive function. Two groups of twenty-five patients are compared in an open, parallel group and non-randomized clinical investigation: a group of patients on carbamazepine (CBZ) monotherapy versus a group of similar size on phenytoin (PHT) monotherapy. The two groups do not show significant differences on variables that could confound the comparison of drug-specific adverse effects: age, gender, intelligence, type of epilepsy, seizure type, seizure frequency, EEG focus and age at onset of the epilepsy. All patients were investigated with a comprehensive neuropsychological test battery ('FePsy'), assessing the cognitive domains of 'speed factors', memory and attention. The results show lower performance in the PHT group compared to the CBZ group on all tests measuring motor speed. Additionally, there is evidence that PHT also affects the speed of central 'higher cortical' processing systems. Our investigation also shows slower performance in tasks that measure speed of information processing ('mental speed'). The other investigated areas, i.e. short-term memory, long-term (verbal and non-verbal) memory and selective attention, do not reveal statistically significant differences between the two groups. These results reconfirm that patients on PHT may suffer from cognitive side-effects even when the medication is sufficiently controlled and the drugs are given within the assumed therapeutical interval.