A double-blind randomized parallel group design was used to compare the tolerance and efficacy of zopiclone 7.5 mg, flurazepam 30 mg and placebo taken each night for 3 weeks in a group of 24 out-patients complaining of sleep disturbance. Analogue rating scales were used to assess the efficacy of the treatments while tolerance and residual effects were measured weekly using a battery of performance tests (Critical Flicker Fusion Threshold, Choice Reaction Time, Letter Cancellation and Digit Span). Physical and clinical examinations, including ECG and EEG, were conducted before and after the experimental period and spontaneously reported side-effects were recorded. Analysis of variance on 24 completed cases (8 patients in each treatment group) showed both active treatments to be significantly better than placebo in shortening sleep onset latency. Flurazepam increased the duration of sleep and produced a "hangover" of impaired psychomotor function. Zopiclone had no observable effect on early morning performance and was free from residual sedative activity. The lack of residual effects with zopiclone 7.5 mg, at a dose shown to be clinically effective, suggests its use in ambulant or out-patient populations where daytime sedation could interfere with every day activities.