Little is known about the relative benefits of cognitively stimulating activities at different points in the lifespan. In a cohort of 576 older persons without dementia, we assessed current and past (childhood, young adulthood, middle age) frequency of cognitive activity; availability of cognitively stimulating resources in the home in childhood and middle age; and 5 domains of cognitive function. Past cognitive activity and cognitive resources were positively correlated with both current cognitive activity and current cognitive function. The association with cognitive function was reduced after controlling for current cognitive activity, however. Current cognitive activity was associated with better cognitive function, especially semantic memory and perceptual speed, even after controlling for past activity. The results suggest that past cognitive activity contributes to current cognition principally through its association with cognitive activity in old age. (JINS, 2005,11, 400–407.)