Brown, Rosvold, & Goldman, 1979; Gaffan, 1974; Mahut, Moss, & Zola- Morgan , 1981; Moss, Mahut, & Zola-Morgan, 1981; Rehbeinet al., 1980).
That is, in tasks using similar stimulus materials and similar motor responses, far more pronounced deficits appear when the reward contingency association of a stimulus varies from trial to trial than when that association is constant from trial to trial, or if the previously stable association is suddenly altered, as in reversal. Although recent work with combined hippocampal-amygdala lesions or hippocampal-cortical lesions suggests some important qualifications to the preceding statements ( Mahut et al., 1981; Mishkin, 1978; Mishkin & Saunders, 1979), the general phenomena seem to hold, and I will indicate later how the Mahut and Mishkin results are actually supportive of a distinction based on stability versus variability of memory.
A pattern of results similar to that just described for primates also seems to hold true for studies using rats as subjects, typically in maze tasks. That is, in tasks where the memory demand is discrimination ( Winocur & Olds, 1978) or a consistent location or direction ( Olton & Papas, 1979; Walker & Olton, in press; Walker, Skinner, Kosobud, Hennessy, & Black, in preparation; Winocur & Breckenridge, 1973), the brain-damaged animal seems to have a mild or transient impairment, at