Six healthy volunteers received a single i.v. dose of 'low dose' lorazepam (0.0225 mg/kg), 'high dose' lorazepam (0.045 mg/kg) and placebo by 1-min infusion in a double-blind three-way crossover study. Plasma concentrations were measured 24 hr after dosage, and the EEG power spectrum was simultaneously computed by fast-Fourier transform to determine the percentage of total EEG amplitude occurring in the 13-30-Hz range. Low and high dose lorazepam did not differ significantly in distribution volume (1.89 versus 1.81 l/kg) or elimination half-life (11.5 versus 12.2 hr); clearance was slightly although significantly reduced at the higher dose (2.08 versus 1.88 ml/min/kg, P less than .005). EEG effects were of relatively slow onset, reaching their maximum change over baseline 30 min after infusion. The duration of action was prolonged, with the fraction of EEG activity in the 13-30-Hz range still significantly above baseline 8 hr after the 0.045 mg/kg dose. Five of these subjects received 0.15 mg/kg of i.v. diazepam in a companion study of identical design. EEG effects of diazepam were shorter than those of lorazepam, probably because of the more rapid and extensive decline in plasma diazepam concentrations in the postinfusion distribution phase. In addition, the onset of diazepam's effect was immediate. In male CD-1 mice that received i.v. diazepam (8.3 mg/kg) or lorazepam (3.3 mg/kg), the brain:plasma concentration ratio was maximal 2.5 min after dosage for diazepam, but equilibration was delayed at least 30 min after dosage for lorazepam. Thus the slow onset of action of lorazepam is probably attributable to slow entry into brain.