Intact monkeys and monkeys with deafferented forelimbs were seated in a restraining apparatus preventing view of the limbs and were required to perform a forelimb flexion response to obtain food reward. Initially, a food pellet was presented after completion of each response (continuous reinforcement schedule). After performance had stabilized, the rate of responding was approximately the same in the deafferented and intact animals. However, the deafferented animals reached their asymptotic performance approximately three times more rapidly than did the normal controls. In addition, the steady-state rate of responding of the deafferented monkeys was much more stable than that of the intact animals, being half as variable from day to day as well as from minute to minute within a session. One reason for the "superior" learning and performance of the deafferented monkeys might be their greater tendency toward stereotyped, nonexploratory movement. After conclusion of testing, reinforcement was delivered on fixed ratio schedules of progressively increasing magnitude. All the deafferented animals were able to achieve stable responding on a fixed ratio 50 schedule, but had more difficulty doing so than did the intact animals. This deficit illustrates the value of somatosensory feedback for maintaining behavior as performance requirements become increasingly complex and frequency of reinforcement decreases. © 1978.