Spatial learning in old mice (19 or 24 months old), some of which had been calorically restricted beginning at 14 weeks of age, was compared to that of young mice, in two separate experiments using a Morris water maze. In the first experiment, only old mice reaching criterion performance on a cued learning task were tested in a subsequent spatial task. Thus, all old mice tested for spatial learning had achieved escape latencies equivalent to those of young controls. Despite equivalent swimming speeds, only about half the old mice in each diet group achieved criterion performance in the spatial task. In the second experiment, old and young mice all received the same number of training trials in a cued task and then in a spatial task. Immediately following spatial training, they were given a 60-s probe trial, with no platform in the pool. Both groups of old mice spent significantly less time in the quadrant where the platform had been and made significantly fewer direct crosses over the previous platform location than did the young control group. As in Experiment 1, calorie restriction failed to provide protection against aging-related deficits. However, in both experiments, some individual old mice evidenced performance in spatial learning indistinguishable from that of young controls. Separate comparisons of "age-impaired" and "age-unimpaired" old mice with young controls may facilitate the identification of neurobiological mechanisms underlying age-related cognitive decline.