Morphine post-dependent rats were prepared with chronic cortical EEG and temporalis muscle EMG electrodes and i.v. jugular cannulae and permitted to self-administer morphine (10 mg/kg/injection) on a fixed-ratio schedule of reinforcement to reestablish and maintain dependence. Morphine was then replaced by methadone or 1-alpha-acetyl-methadol (LAAM) at a dose of 2 and 1 mg/kg, respectively. The rats developed stable patterns of self-administration of each narcotic and maintained the dependent state. The mean (+/- S.E.) interinjection interval was 2.5 +/- 0.1 h for morphine, 1.4 +/- 0.1 h for methadone and 8.8 +- 0.8 H FOR LAAM. The distribution of sleep, REM sleep, and awake within the interinjection interval for the three narcotics was similar. A short phase of behavioral stupor concomitant with EEG slow bursts emerged following the self-injections. This phase was followed by arousal with EEG activation. Sleep and REM sleep then reappeared and predominated before the next injection. LAAM manifested a relatively slow onset of action in comparison with morphine and methadone.