Receptor theory of opioid action has provided an extremely useful interpretive framework for the discriminative stimulus effects of opioids. By and large, receptor theory has been applied to opioid actions as they are measured in in vitro and reflex systems. It is clear, however, that it can also assist in interpreting data from experiments addressing operantly conditioned behavior, and provide a link between these data and those obtained using other procedures. The current paper describes the criteria that can be used to determine whether a drug effect is receptor mediated and applies these criteria to the effects of mu and kappa opioids in drug-discrimination studies. Criteria for distinguishing between drug effects occurring through one, as opposed to two, receptor systems are described and again applied to the discriminative stimulus effects of mu and kappa opioids. The potential difficulties that can be caused by postreceptor variability and the presence of multiple receptor systems are noted, since they can modify the effects predicted from simple receptor theory, and are likely to play an important role when studies of opioid action are made in the whole animal. In discrimination studies, complicating variables include dose of the training drug, subject species, nature of the training drug, and context of the discrimination. Finally, the ability of receptor theory to guide future investigation of the phenomenon of partial generalization is explored.