In four conditioned suppression experiments with rats (Rattus norvegicus), backward pairings of a shock unconditioned stimulus (US) and a tone conditioned stimulus (CS) eliminated an already established conditioned response (CR), but there was recovery of the CRifthe shock was later withheld. In Experiment 1, there was recovery after backward pairings, regardless of whether the period after the US was normally shock free or not In Experiment 2, the occurrence of recovery depended on the CS's being presented closely after the US in response elimination. Levels of recovery were positively correlated with the resistance ofthe response to elimination during backward pairings (Experiments 3 and 4). Taken together, these data support the notion that recovery after backward pairings is a form of renewal (see, e.g., Bouton, 1991) and is not due to protectionfrom extinction.The usual method for eliminating an already established conditioned response (CR) is to present the conditioned stimulus (CS) in the absence of the unconditioned stimulus (US). With this method, extinction, the probability of the CR decreases over trials, although there is some tendency for responding to spontaneously recover if the animal is returned to the conditioning context after a rest period (see, e.g., Pavlov, 1927;Robbins, 1990). Other methods for response elimination have received less attention.