In order to determine the stability of the context shift effect over time, rats were trained in a passive avoidance task and tested at four different intervals afterward. Half of the subjects at each interval were tested for retention of the avoidance response in the context in which they were trained; the other half were tested in a different one. Rats tested in the different context at 1 and 21 days following learning showed the context shift effect, while those tested at 7 and 14 days did not. These results indicate that the context shift effect may be a phenomenon that changes over time.