Aging in men is associated with decreases in serum testosterone (T) and a decline in cognitive abilities. We sought to clarify the relationship between T, aging and cognition using the common marmoset (callithrix jacchus), which has been shown to be a useful model of human aging. Ten castrated male marmosets (ages 3-8) were given weekly injections of either T cypionate dissolved in cottonseed oil (T, n = 5) or cottonseed oil alone (controls, n = 5). Cognitive function was assessed with two tasks, the object reversals (OR) and the delayed response (DR). Marmoset behavior was recorded twice daily using a modified frequency scoring system, measuring 20 target behaviors in 15-s intervals over 5-min. For the OR task, no effect of group (F (1,8) = .51, p = .50), reversal (F (3,24) = .527, p = .67), or group x reversal interactions (F (3,24) = .640, p = .60) was found in the number of correct trials. There was a significant effect of outcome on response latency, with all monkeys having longer latencies on incorrect trials (F (1,2451) = 45.36, p