Six normal volunteers, aged 20 to 39 years, underwent 2 adaptation nights and three sessions of 2 consecutive experimental nights and days at 1-week intervals, according to a latin-square design. In the three sessions, subjects received either zopiclone, 3.75 mg or 7.5 mg, or placebo at 2215 h in a double-blind protocol. On nights 1 and 2 of each session, subjects were continuously monitored polygraphically, except for a 45-min provoked wake episode 135 min after sleep onset on night 2. Degree of daytime somnolence was assessed during day 1 by means of a multiple sleep latency test (MSLT) and performance evaluation was carried out during night 2 (0000 h) and day 2 (800 h and 1200 h) by means of a battery of four tests. NREM sleep stages 3 and 4 increased significantly after 3.75 mg and 7.5 mg zopiclone (p less than 0.05). No significant differences between placebo and 3.75 mg and 7.5 mg zopiclone were found at any time in the MSLT. Two performance tests (eye-hand coordination test and choice reaction time test) showed a highly significant impairment (p less than 0.01) at 0000 h with 7.5 mg zopiclone; one test (eye-hand coordination test) showed a significant impairment (p less than 0.05) at 0800 h also with 7.5 mg zopiclone and none at 1200 h. From a subjective point of view, depth and quality of sleep were improved, whereas number of awakenings and feeling on awakening were not modified. Side effects (bitter taste, jitteriness, difficulty to concentrate) were reported only with 7.5 mg zopiclone.