Three groups of older adults (mean age 72.1 years) were compared on a free recall task with categorizable lists. The nine females and two males in each group were instructed to rehearse overtly while studying. A group instructed to rehearse by category showed higher levels of free recall and category organization than either a control group instructed only to remember or a group instructed to rehearse actively at study. Strategy instructed subjects' rehearsal was organized serially early in a list and then categorically organized for the remainder of a list. Activity instructed subjects showed a high number of same-item repetitions but adopted no clear pattern of strategic category rehearsal. Control subjects' rehearsal was essentially inactive and nonstrategic, mainly consisting of single mentions of each list item and an associate. These data show that older adults' memory performance is modifiable and that efficient performance is obtained when instructional training is aimed at the processes that are crucial to task performance. Direct strategy measures, such as those used as here, yield important information about the processes underlying age differences in memory and can aid greatly in the design of training aimed at exploring older adult potential.