The effects of benzodiazepines, barbiturates, a series of novel putative anxiolytic compounds and anxiogenic compounds are reviewed in animal tests of anxiety and on experimentally-induced seizures. It is clear from the data that drug effects on anxiety and convulsions are not always in the same direction; certain compounds are apparently both anxiolytic and proconvulsant, others are anxiogenic and anticonvulsant, others have varied effects depending on the test situation. It is suggested that this work necessitates considerable revision of our traditional concepts of an "anticonvulsant." The extent to which drug-induced anxiety is correlated with weak epileptiform activity in the brain is discussed. Finally, the Discussion considers a number of possible mechanisms that could underlie the separation of drug effects on anxiety and convulsions that is observed.