1 The residual effects of two benzodiazepines, nitrazepam (10 mg) and flurazepam hydrochloride (30 mg), and pentobarbitone sodium (200 mg) were studied by adaptive tracking and by reaction time. Performance was measured at 10 h, 13 h, 16 h, 19 h and 34 h after ingestion of each drug. Impaired performance on adaptive tracking was observed at 10 h, 13 h, 16 h and 19 h after nitrazepam and pentobarbitone sodium and at 10 h, 13 h and 16 h after flurazepam hydrochloride. Enhanced performance was observed at 34 h after nitrazepam and pentobarbitone sodium. 2 Increased reaction time persisted to 16 h after nitrazepam, flurazepam hydrochloride and pentobarbitone sodium and reaction time was also increased at 34 h after nitrazepam and pentobarbitone sodium. 3 During the morning immediately after ingestion, the subjects as a group were able to differentiate correctly between placebo and drugs, but they were not able to assess accurately the persistence of the residual effects of nitrazepam and pentobarbitone sodium. 4 Flurazepam hydrochloride would appear to be a more promising benzodiazepine than nitrazepam for use as a hypnotic by persons involved in skilled activity. There was a rapid recovery of performance during the afternoon and, unlike pentobarbitone sodium and nitrazepam, subjects retained the ability to recognize impaired skill.