The impact of past history on behavior across the life span is largely unknown. This is why the role of previous experience in subsequent memory performances has been studied in a combined longitudinal (animals repeatedly tested) and cross-sectional (animals tested once at various ages) study, in male Sprague-Dawley rats. Different types of memory (reference and working memory) and/or information processing (route or place learning) were assessed in three different tasks (T-maze, Morris water maze, and eight-arm radial maze). Our results indicate that experience prevents age-related impairments in the learning phase of the T-maze and the Morris water maze, in both middle-aged and old rats. Nonexperienced animals of the same age were found to present age-related memory deficits. However, previous experience did not have any effect on the learning of the radial maze or on the reversal performance. It is suggested that controlled processes (intentional and attentional) are impaired by aging and cannot be improved by training, whereas automated processes appear to benefit from it. These data underline the heterogeneity of cognitive aging and indicate that aging is not inevitably accompanied by a decline in performance.