Four rhesus monkeys were trained to perform visually guided wrist tracking movements (50). While they performed tasks by wrist flexion or extension from a neutral position, simple-spike (SS) and complex-spike (CS) discharges of a single Purkinje cell (P-cell) were recorded from intermediate and lateral parts of cerebellar hemispheres (lobules IV to VI) ipsilateral to the task-performing wrist. Of approximately 400 P-cells observed, 215 (54%) significantly increased or decreased their SS discharge rate during task performance (task-related P-cells). Of these, 161 were selected for analysis of CS activity; in these P-cells, we could reliably discriminate between CS and background SS by a spike discriminator. The 161 P-cells were further classified into response locked (n = 65) and poorly locked (n = 96) cells according to temporal coupling of the SS frequency modulation to the onset of wrist movements. About 60% of the response-locked P-cells showed a phasic increase (statistical significance level: P less than 0.01) of CS firing rate at the onset of wrist tracking movement. In a few P-cells, a phasic decrease (statistically insignificant) of CS firing rate was observed with the wrist movement. In most P-cells, an increase of CS firing rate was observed with both rapid- and slow-tracking wrist movements. The increase was larger with faster step-tracking movement than with slower ramp-tracking movement. In most P-cells, the CS activity increased with both wrist flexion and extension; in some cells, however, it increased only with either flexion or extension. In most of the response-locked P-cells, the increase of CS firing rate occurred during motor time, i.e., after the onset of the EMG change in prime movers and before the beginning of wrist tracking movement. The increase occurred phasically at the onset and/or at the recovery phase of SS frequency modulation. At neutral wrist position, the maintained frequency of the CS was 0.72 +/- 0.29 CS/s (mean and SD for 161 task-related P-cells). Compared with the frequency at neutral position, the CS frequency did not change tonically during maintained flexed or extended wrist position in any response-locked P-cells. There was no increase of CS firing rate when the monkey returned the handle to center position after completing the tracking task, even in P-cells that had shown a significant increase of CS activity during tracking.(ABSTRACT TRUNCATED AT 400 WORDS)