Adult female Sprague-Dawley rats were prepared with chronic intravenous cannulas and cortical and muscle electrodes for recording electroencephalograms and electromyograms, respectively. They were made physically dependent on morphine by automatic intravenous injections and then trained to lever press in order to self-administer morphine on a FR-20 schedule of reinforcement. Upon stabilization of morphine self-administration, one group continued to self-administer morphine, while two other groups were switched to methadone or 1-alpha-acetylmethadol (LAAM) self-administration for an additional five to ten days. Continuous EEG and EMG recordings were collected. Initially, automatic injections of morphine suppressed rapid eye movement (REM) sleep time, then tolerance developed to this effect. REM sleep time in rats self-administering LAAM, methadone or morphine was within the lower limit of the normal range. Following withdrawal, REM sleep was severely suppressed during the first 24 h with morphine and methadone, but only moderately suppressed with LAAM. Increases in lever pressing during withdrawal from morphine and methadone occurred earlier and were more intense and prolonged than for LAAM. The incidence of head shakes peaked earlier and was higher for morphine and methadone during withdrawal than for LAAM. Irritability scores increased for morphine and methadone during the first day of withdrawal, but did not show any increase until the third day for LAAM. These findings suggest that in dependent rats withdrawal from LAAM is less severe than withdrawal from morphine or methadone.