1 Psychomotor skills and visual functions related to driving were measured double-blind cross-over in ten healthy volunteers before, and 1, 3, 5 and 7 h after a single oral administration of diazepam (10 mg), medazepam (15 mg) or lorazepam (2.5 mg). The late effects of lorazepam were tested in seven other subjects 12 and 24 h after the administration. 2 Lorazepam impaired almost all the measured skills more (P < 0.05 to 0.001) than diazepam, medazepam or the placebo. The lorazepam impairment of reactive skills and flicker fusion discrimination remained statistically significant (P < 0.05) for as long as 12 h. Medazepam impaired only reactive skills and flicker fusion, the latter remaining impaired (P < 0.05) for as long as 5 h after the administration. The magnitude and duration of the effects of diazepam were intermediate between those of lorazepam and medazepam. Diazepam impaired perceptual speed and reactive and co-ordinative skills as well as flicker fusion discrimination and visual parameters related to driving. Slight impairments in performance were measurable for up to 5 h after administration but at 7 h the results resembled those measured after the placebo. 3 The lack of alterations in adaptation to darkness, sensitivity to brightness or visual discrimination ability in bright counterlight at a time when flicker fusion discrimination was severely depressed suggests that an impaired ability to discriminate flickering light is of no or little clinical significance to driving ability. 4 It is concluded that patients receiving a 2.5 mg dose of lorazepam should not drive or operate machinery for 24 h after the administration. After diazepam (10 mg) or medazepam (15 mg) patients should refrain from driving or participating in skilled performances for only 5 to 7 hours.