The purpose of the present study was to evaluate the reinforcing effects of smoked heroin in nonopioid-dependent nonhuman primates when an alternative reinforcer, sweetened fluid, was made available. Four adult male rhesus monkeys lived in three chambers, with heroin self-administration (0, 0.3, and 0.6 mg/kg) specific to one end of the chamber, oral sweetened fluid self-administration specific to the other end chamber, and no commodity available in the middle chamber. The length of time monkeys spent in the drug-associated chamber provided one measure of drug seeking (i.e., location preference). During self-administration sessions, a second-order schedule of reinforcement was used, with responding during the first component maintained by a brief presentation of the stimuli associated with reinforcement. Responding during the second component was maintained by a delivery of the reinforcer, and the associated stimuli. Responding during the first component provided a second measure of drug seeking. Monkeys also had choice trials each day, when they could choose to work for either commodity. Choice behavior provided a third measure of drug seeking. Each experimental day consisted of a smoking session (four smoking trials), a sweetened fluid session (four fluid trials), and a choice session (four choice trials). Monkeys typically completed all four smoking trials each day when either of the active heroin doses was available. They chose both heroin doses over fluid on 3.5 of the four choice trials, and they had a location preference for the heroin chamber. Under baseline conditions, the number of acquisition responses and the number of consumption responses (inhalations) were greater for the high dose of heroin compared to the low dose of heroin. Further, it took longer to extinguish the responding for the high dose of heroin compared to the low dose of heroin when a vehicle was substituted. During heroin extinction, acquisition responding for fluid increased, the number of fluid choices increased, and location preference shifted to the fluid chamber. These data suggest that in nondependent rhesus monkeys, measures of heroin seeking decreased when heroin was not available and seeking behavior shifted to the available alternative commodity.