A placebo-controlled, double-blind, crossover trial in 11 healthy male volunteers compared clinical sedation and psychomotor function after intravenous injection of midazolam (0.05, 0.1, or 0.15 mg/kg), diazepam (0.15 or 0.3 mg/kg), or placebo (saline). The depth of sedation was estimated at 5-10-min intervals during the first hour after injection. A comprehensive battery of psychomotor tests was used to collect objective data of psychomotor performance before drug injection and 1, 3, 5, and 7 h after injection. Midazolam (0.15 mg/kg) produced the highest scores of sedation and most impairment of psychomotor performance. In most tests, the maximal psychomotor effects seen after 0.3 mg/kg of diazepam did not reach those of 0.1 mg/kg of midazolam. Although the strongest psychomotor effects were induced by midazolam, these effects disappeared sooner than those of diazepam. By 5 h after injection, 0.3 mg/kg of diazepam showed the highest scores of psychomotor impairment. The authors conclude that at least four times as much diazepam as midazolam is needed to produce equally severe psychomotor impairment. That the residual effects of midazolam terminate sooner than those of diazepam probably accounts for the occasional underestimation of the potency of midazolam in clinical practice.