Effective prescribing of anticonvulsants requires foreknowledge of baseline pharmacokinetic data. Little such information is available about the hydantoins other than phenytoin, although one of them, mephenytoin, is widely used. Useful pharmacokinetic data should be derived from patients already exposed to anticonvulsants to reflect the induction of hepatic oxidative enzymes. Single-dose studies of mephenytoin (Mesantoin) and ethotoin (Peganone) were performed in adult inpatients on stable regimens of other anticonvulsants. Five patients received mephenytoin, 7 mg per kilogram of body weight. Serial blood sampling was performed rigorously. The time to peak concentration (Tmax) for mephenytoin was 1 hour, with a half-life (T 1/2) of 7 hours; the T 1/2 of its metabolite, 5-ethyl-5-phenylhydantion, was 96 hours. Ethotoin administration was 25 mg per kilogram in 5 patients. Ethotoin Tmax was 2 hours, with a T 1/2 of 5 hours. Saliva accurately represented the unbound fraction for all three agents. Mean salivary levels (as percentage of total levels) were 61% for mephenytoin, 73% for its metabolite, and 54% for ethotoin. The implications for therapy are that following mephenytoin administration, the metabolite 5-ethyl-5-phenylhydantoin will provide anticonvulsant effectiveness, with its long half-life producing stable blood levels on simple dose schedules. Ethotoin, in contrast, has a short half-life and would require divided daily doses to achieve a steady state. This, rather than pharmacological ineffectiveness, limits its usefulness.