Several authors have criticized the traditional resistance-to-extinction tests of conditioned reinforcement (Kelleher & Gollub, 1962;Kelleher, 1966;Zimmerman & Hanford, 1966) because they test for conditioned reinforcing effects after primary' reinforcement has been eliminated from the experimental situation. Further, Zimmerman and Hanford suggested that the extinction tests do not model nonexperimental situations in which the concept of conditioned reinforcement is often invoked. Another obvious critieism of those techniques is that they confound the extinction of a response and the extinction of the conditioned reinforcer. Consequentl~, such studies must compare the extinction responding of two groups of Ss, one that receives the supposed conditioned reinforcer for responding and one that does not.As an alternative to the extinction tests of conditioned reinforcement, Kelleher & Gollub (1962) recommended the study of stimulus functions in chained schedules, Kelleher (1966) suggested the study of brief stimuli in second-order schedules, and Zimmerman & Hanford (1966) presented a third alternative, the study of response-produced stimuli occasionally paired with response-independent food. The present paper examined a fourth technique, the study of conditioned rein force me nt in an extinction component of a multiple schedule. 1Psychon. Sei., 1970, Vol. 18 (1) Skinner (I938) demonstrated a conditioned-reinforcement test procedure that involved fust partially extinguishlng responding without the supposed conditioned reinforcing stimulus and then making the stimulus available for responding, thus producing a second extinction curve more clearly reflecting the strength of the conditioned reinforcer independently of the strength of the conditioned response. The procedure does, however, eliminate primary reinforcement from the testing situation.Multiple schedules of reinforcement {Ferster & Skinner, 1957} are those involving two or more independent reinforcement schedules, only one of which is in effect at a given time, and each of which has its own correlated stimulus condition. A multiple schedule, in wbich a neutral stimulus and primary reinforcement remain available in one stimulus condition and the procedure developed by Skinner is used in the other stimulus condition, would seem to satisfy the objections to extinction tests discussed above. Primary reinforcement would be maintained in the experimental situation, the conditioned reinforcer would be tested after the conditioned response is extinguished in the second stimulus condition, and the procedure seems analogous to situations in which the concept of conditioned reinforcement is often used as an explanatory device.The current study, then, tested for conditioned reinforeing effects in a multiple schedule in which, during the fust stimulus condition, responding produced a white noise and primary reinforcement, and, during the second stimulus condition, responding was first extinguished to a low level and then produced only the noise. METHOD The Ss were six male...