The effects on memory and psychomotor performance and the subjective effects of three anxiolytic benzodiazepines (lorazepam 2 mg, diazepam 10 mg and clobazam 20 mg p.o.) have been evaluated in a double-blind, placebo-controlled, cross-over study in 10 healthy volunteers. At each session, measurements were made prior to and +3.5 h after drug administration, except in the case of REY's test, which was presented at H + 1 h (learning) and was evaluated at H + 8 h and at H + 24 h (delayed recall). Single clinical doses of diazepam and lorazepam caused anterograde amnesia by disturbing acquisition, consolidation and retrieval. Clobazam did not impair memory. Lorazepam impaired performances in all the tests used to evaluate perception, immediate memory, reaction time, psychomotor skill and intellectual capacity. Diazepam caused a decrease in cortical arousal and the speed of perception of visual stimuli, whereas clobazam increased reaction time and reduced cortical arousal. Lorazepam caused a significant degradation of performance relative to the other two treatments.