Aims. To investigate whether the neonatal abstinence syndrome (NAS) is different in children born to women maintained on slow-release morphine, compared with those maintained on methadone, and to compare additional drug consumption in these groups of women . Design, Setting and Participants. An open, randomized trial was conducted in an established clinic. Forty-eight pregnant women who presented to the clinic as opiate or polysubstance abusers were enrolled and maintained on either methadone (24 women) or slow-release morphine (24 women) up to and following delivery. The programme included psychosocial therapy and support for their opiate-addicted partners . Measurements. Standard urinalysis methods were used to measure consumption of cocaine and benzodiazepines during pregnancy. Injection sites were monitored to indicate additional opiate use. NAS was measured according to Finnegan score and the amount of phenobarbiturates prescribed to alleviate the symptoms . Findings. No difference was found in the number of days that NAS was experienced by neonates born to methadone or morphine maintained mothers (mean = 16 and 21 days, respectively). All children were born healthy and no serious complications arose. Fewer benzodiazepines ( p 0.05) and fewer additional opiates ( p 0.05) were consumed by the morphinemaintained women compared with those who took methadone, but no difference was seen in cocaine consumption. Nicotine consumption was reduced significantly in both groups during pregnancy ( p 0.02) . Conclusions. Both methadone and morphine are suitable maintenance agents for pregnant opiate addicts. Maintenance agents that result in a less prolonged NAS should be studied in further trials.