Twelve healthy male volunteers received phenytoin 0.5 and 1 g, lamotrigine (a new anticonvulsant) 120 and 240 mg, diazepam 10 mg and placebo orally in a double-blind, cross-over, randomized trial. Maximum drug concentrations at 4 h, measured in plasma were 11.5 +/- 2.2 micrograms ml-1 for phenytoin and 2.7 +/- 0.4 micrograms ml-1 for lamotrigine. These levels were in the therapeutic range for phenytoin and the putative therapeutic range for lamotrigine. Side effects after diazepam (mainly sedation) and phenytoin (mainly unsteadiness) differed markedly from lamotrigine which produced no important side effects. Subjective effects as measured by visual analogue scales were caused by phenytoin and diazepam but not by lamotrigine. Diazepam impaired eye movements, adaptive tracking and body sway. Phenytoin impaired adaptive tracking, increased body sway and impaired smooth pursuit eye movement. Lamotrigine produced only a possible slight increase in body sway. There were significant correlations between performance and saliva levels of phenytoin and diazepam. It was concluded that the tests used were suitable for monitoring CNS effects of anticonvulsants and that lamotrigine possibly could have a more favourable CNS side effect profile than phenytoin.