The acquisition of anxiety disorders (e.g., phobias) is often thought to be mediated by classical conditioning processes (e.g., Wolpe, 1958, Psychotherapy by reciprocal inhibition Wolpe and Rowan, 1989, Behaviour Research and Therapy, 27, 583-585). Thus, the success of exposure therapy is possibly a consequence of extinction, and factors affecting extinction in Pavlovian conditioning are potentially relevant to clinicians who administer exposure therapy. The present experiments investigated the effects of conducting extinction in multiple contexts using rats as subjects in a conditioned suppression paradigm. In Experiment 1, subjects received conditioned stimulus (CS) and unconditioned stimulus (US) pairings in one context followed by extinction of that CS in one or three other contexts. When tested in an associatively neutral context (i.e., different from those of conditioning or extinction), rats that had received extinction in three contexts exhibited less responding to the CS than rats that had received extinction in one context. In Experiment 2, CS-US training occurred in either one or three contexts, followed by extinction of that CS in three other contexts. Testing in a neutral context revealed that rats conditioned in multiple contexts showed greater responding to the CS than rats trained in a single context. The results are discussed in the framework of memory retrieval, and the clinical implications are explored.