The mechanism by which buprenorphine reduces opioid self-administration in humans and animals is generally thought to be through its opioid agonist effects. Buprenorphine, given acutely i.v. to three rhesus monkeys 30min prior to a session in which a range of doses of either alfentanil or heroin was available, produced dose-related decreases in the potency of both opioid agonists. The effects of buprenorphine were generally surmounted by increasing the dose/injection of alfentanil or heroin available for self-administration, indicating that buprenorphine was acting as an opioid antagonist in this situation. These data suggest that at least part of the effectiveness of buprenorphine in reducing opioid administration by human opioid users may be via its opioid antagonist properties.