Neuronal relations to active movements of individual body parts and neuronal responses to somatosensory stimulation were studied in the external (GPe) and internal (GPi) segments of the globus pallidus (GP) and the subthalamic nucleus (STN) of awake monkeys. In GPe (n = 249), GPi (n = 151), and STN (n = 153), 47, 29, and 28% of the cells, respectively, discharged in relation to active arm movements, 10, 11, and 15% to leg movements, and 22, 22, and 18% to orofacial movements. Of the neurons whose activity was related to arm movements, 26, 16, and 21% in GPe, GPi, and STN, respectively, discharged in relation to movements of distal parts of the limb. Of cells whose discharge was related to active limb movements, 37, 22, and 20% in GPe, GPi, and STN, respectively, also responded to passive joint rotation, which was usually specific in terms of joint and direction of movement. Only a small percentage of cells responded to muscle or joint palpation, tendon taps, or cutaneous stimulation. Short-latency, direction-specific neuronal responses to load perturbations confirmed the existence of proprioceptive driving. In both GPe and GPi, leg movement-related neurons were centrally located in the rostrocaudal and dorsoventral dimensions. In contrast, arm movement-related cells were found throughout the entire rostrocaudal extent of both segments, although in greater numbers caudally. In the central portions they were situated largely inferior and lateral to leg movement-related neurons. Neurons related to orofacial movements were largely confined to the caudal halves of both segments, where they were located largely ventral to arm movement-related cells. The STN cells whose activity was related to leg movements were observed largely in the central portions of the nucleus in the rostrocaudal and mediolateral dimensions. Cells whose activity was related to arm movements were found throughout the rostrocaudal extent of the nucleus, but were most numerous at the rostral and caudal poles. Neurons related to movements of the facial musculature and to licking and chewing movements were distributed over the entire rostrocaudal extent of the nucleus, where they generally occupied the ventrolateral regions. In all three nuclei, neurons with similar functional properties were sometimes clustered together. Within the arm and leg areas, however, there was no clear evidence for a simple organization of clusters related to different parts of the limb. These studies provide further evidence for a role of the basal ganglia in the control of limb movements.