Eight heroin-dependent individuals, maintained on divided daily doses of oral morphine, participated in a 2.5-week inpatient study comparing the effects of intranasal (IN) (placebo, 12.5, 25, 50, 100 mg) and intravenous (IV) (placebo, 6.25, 12.5, 25, 50 mg) heroin. Each morning, participants received $20 and a sample dose of heroin, and each afternoon they had the opportunity to self-administer all or part of the morning heroin dose or money amount. Participants responded under a modified progressive-ratio schedule (PR 50, 100, 200, 400, 800, 1200, 1600, 2000, 2400, 2800) during a ten-trial self-administration task. During each trial, participants could respond for 1/10th of the heroin dose or 1/10th of the money amount. The total amount of heroin and/or money chosen during the self-administration task was given at the end of the task. Thus, participants received drug and/or money twice each day: once during the morning sample session and once during the afternoon self-administration session. Participants received IV solution and IN powder simultaneously during each dosing; only one route contained active drug. Heroin produced dose-related increases in break point values by both routes of administration. Although IV heroin was approximately four-fold more potent than IN heroin, the maximal break point values for both routes were not significantly different. A similar difference in potency between the IV and IN routes was found for several ratings of subjective effects (e.g., "I feel a good drug effect," "I feel high"), but maximal subjective ratings were lower for IN compared to IV heroin. These results suggest that the reinforcing efficacy of heroin is similar by the two routes of administration, but that IN heroin is less potent than IV heroin. The results also underscore the importance of evaluating drug self-administration in the evaluation of the abuse liability of drugs.