A blocking paradigm with rats was used to evaluate whether different temporal information is encoded in simultaneous rather than forward associations. During Phase 1, the blocking conditioned stimulus (CS) was simultaneously or forward paired with an unconditioned stimulus (US). During Phase 2, the pretrained CS occurred in compound with a novel target CS that was paired in a simultaneous or forward manner with the US. Forward pretraining resulted in more blocking of a forward than a simultaneously trained target CS, and simultaneous pretraining resulted in more blocking of a simultaneously than a forward trained target CS. Thus, greater blocking occurred when the blocking and blocked CSs had the same temporal relation to the US. The results support the temporal coding hypothesis and question the necessity of predictive information in blocking.