In three experiments with rat subjects, we examined the effects of trial spacing in appetitive conditioning. Previous research in this preparation suggests that self-generated priming of the conditional stimulus (CS) and/or unconditional stimulus (US) in short-term memory is a cause of the trial-spacing effect that occurs with intertrial intervals (ITIs) of less than 240 sec. Experiment 1 nonetheless showed that a trial-spacing effect still occurs when ITIs are increased beyond 240 sec, and that the effect of ITI over 60-1,920 sec on conditioned responding is best described as a linear function. In Experiment 2, some subjects were removed from the context during the ITIs, preventing extinction of the context. Removal abolished the advantage of the long ITI, suggesting the importance of exposure to the context during the long ITI. Experiment 3 still produced a trial-spacing effect in a within-subjects design that controlled for the level of context conditioning and reinforcement rate in the absence of the CS. Overall, the results are most consistent with the idea that adding time to the ITI above 240 sec facilitates conditioning by extinguishing context-CS associations--and possibly context-US associations--that otherwise interfere with CS-US learning through retrieval-generated priming (see, e.g., Wagner, 1981).