CEREBELLAR LESIONS cause errors of direc-tion, range, rate, and force in voluntary movements (30). Holmes (30, 35) suggested that the primary effect of cerebellar lesions was on the speed of movements: when there was no opposing load, excessive peak velocities were reached and were inadequately braked.

The experiments described in this paper examine these aspects of motor control dur-ing experimental cerebellar dysfunction. Cebus monkeys were trained to execute a tracking task under no-load conditions requiring arm movements of accurate displacement, voluntarily determined speeds, and maintenance of target positions for a fixed period of time (39). Cerebellar control was studied by assessing the performance deficit: during and after brief reversible block of the lateral part of the dentate nucleus ipsilateral to the operant hand, using a stereotaxically implanted cooling probe (8). The dentate nucleus was chosen on anatomical grounds as a critical output link from cerebellar cortex to the contralateral ventrolateral thalamic relav to the cerebral cortex (13). It was also known that in humans clinical symptoms after damage to the cerebellar cortex are more severe when the dentate nucleus is damaged as well (30; also see ref 36), and that in monkeys dentatectomy can produce some cerebellar symptoms(26), particularly in goaldirected movements(41). The method of brief reversible coolink df local sites in the CNS under stereotaxic guidance was introduced a little over 10 years ago (2, 10, 44,