Evidence from both animal, and human, studies suggests that repeated administration of 3,4-methylenedioxymethamphetamine (MDMA: "ecstasy") produces lasting decreases in serotonergic activity. Serotonin is believed to play a modulatory role in a variety of psychological processes, including learning and memory. There are recent reports that polydrug users, who have used ecstasy recreationally, exhibit selective impairments in memory. However, these studies did not compare ecstasy users with polydrug users who had not taken ecstasy, leaving open the possibility that the memory deficits may be associated with a history of use of other illicit drugs. The present study used the Rivermead Behavioural Memory test to investigate immediate and delayed recall in: 25 polydrug-users who had taken more than 20 tablets of ecstasy (MDMA group), 22 participants (polydrug controls) who had never taken ecstasy, but, otherwise has personal characteristics (e.g. age, gender, education, height, weight), and illicit drug use histories, that were generally not significantly different from those of the MDMA group, and 19 participants who had not used illicit drugs but who also had similar personal characteristics (non-drug controls). Participants in the MDMA group recalled significantly fewer ideas (approximately 75% of the number of ideas recalled by participants in either of the other two groups), in both immediate and delayed recall conditions. The two illicit drug-using groups did differ in their estimated IQ scores and their duration of use of LSD, but only the latter proved to be a statistically significant covariate, and the difference in recall performance between the MDMA and polydrug controls groups remained statistically significant when this variable was treated as a covariate. The present findings provide the first evidence that deficits in memory performance in recreational ecstasy users are primarily associated with past exposure to ecstasy, rather than with the other legal and illicit drugs consumed by these individuals, and are consistent with reduced serotonergic modulation of mnemonic function as a result of long-term neurotoxic effects of MDMA in humans.