Previous experiments in our laboratory have shown that lesions of the mammillary bodies (MB) in mice induce memory deficits in spontaneous spatial alternation (SA) tasks. However, it has also been shown that performance on memory tasks by both humans and animals depends on the degree of effort involved in a given task. Thus, we hypothesized that the involvement of MB-damaged subjects in tasks that require more effortful processing than does SA might reduce their memory deficits. Accordingly, in the first experiment, the effects of MB lesions were studied in both an SA and a rewarded alternation (RA) task. Daily test sessions consisted of six successive trials separated by a short intertrial interval (ITI) of 30 sec, which allowed us to study interference (i.e., the deleterious effect of the first trials of a session on the performance on the last trials of a session). Results of the first experiment showed that the MB-lesioned subjects were impaired on their performance on the SA (they were more sensitive to interference than were controls) but not RA task. This finding is consistent with our hypothesis. In the second experiment, the ITI in the RA task was increased from 50 sec to 3 min. At the longer ITI (3 min), we found that the MB-lesioned subjects had the same deficit in performance that the MB-lesioned subjects had on the SA task at an ITI of 30 sec. Control mice also had a similar deficit in performance at the ITI of 5 min. We suggest that even in the RA task, the performance of MB-damaged subjects remains more sensitive to time-dependent interference than does the performance of controls. The significance of our study to the understanding of amnesia (especially Korsakoff's syndrome and frontal lobe pathology) is discussed.