The ability of the rat brain to acquire or to retain specific learning tasks was tested under conditions of multiple lesions and widely different amounts of practice. Lesion targets were (a) the medial prefrontal and cingulate cortex, (b) the anterior and mediodorsal thalamus, and (c) the dorsal and ventral hippocampus. Rats were divided into seven groups. The first group received lesions of all three structural complexes prior to training in a delayed alternation and an active avoidance task. Groups 2-4 received lesions in different combinations of two of the three structural complexes prior to task acquisition. Group 5 first learned both tasks and then received the medial cortical lesion; thereafter it was retrained to criterion. Then, the thalamic lesion was made, and relearning was tested a second time. Finally, the hippocampal region was damaged, and a last relearning test was given. Groups 6 and 7 also first acquired both tasks; however, after that, they received 240 (Group 6) or 1,280 (Group 7) trials of overtraining. Following this, all three structural complexes were given lesions serially before relearning of the two tasks was tested. Nine of the ten animals of Group 1 failed to acquire the alternation task, but all learned the avoidance task. In Groups 2-4, all rats acquired both tasks. Postoperatively, rats of Group 5 were inferior to those of Group 6 in both tasks, and rats of Group 7 were the most successful animals of the last three groups. These results question the assumption that serial lesions with intermittent training between lesions have beneficial effects, and they also stress the importance of task practice, that is, of behavioral experience. It is argued that prolonged training will lead to a widely distributed storage of information within the brain. The process of wide diffusion of information will, however, be disturbed (or at least retarded) by lesions made shortly after task acquisition or task reacquisition (as was the case for animals of Group 5).