The present study examined individual differences in recognition memory function in a group of Old World monkeys (Macaca mulatta). Four young (9-11 years) and 10 aged (22-33 years) monkeys were tested in the same delayed-nonmatching-to-sample (DNMS) recognition memory procedure that has been widely used to study the effects of experimental hippocampal lesions in young subjects. Animals were first trained to a 90% correct learning criterion in the DNMS task using a 10-second delay between the sample and recognition phase of each trial. The memory demands of the task were then increased by gradually extending the retention interval from 15 seconds to 10 minutes. Three of the aged monkeys performed as accurately as young subjects at all delays. The remaining aged monkeys performed well at the shortest delays (15 and 30 seconds), but progressively greater impairments emerged across delays of 60 seconds, 2 minutes, and 10 minutes. These results suggest that recognition memory is only compromised in a subpopulation of aged monkeys. Moreover, aged monkeys that are impaired in the DNMS task exhibit the same delay-dependent pattern of deficits that is the hallmark of memory dysfunction resulting from medial temporal lobe damage.