1. We electrically stimulated the macaque monkey's frontal eye field (FEF) region to localize and to analyze the smooth pursuit eye movement representation. Rhesus monkeys were trained to fixate stationary spots of light, and trains of stimulation (usually 250–500 ms at 10–100 microA) were applied while the fixation targets remained lit and stationary. This paradigm was used in a total of 485 electrode penetrations through the arcuate sulcus region of six hemispheres in three adult monkeys. Smooth eye movements (SEMs), clearly distinct from saccades, were elicited at 86 sites in 53 of these penetrations. These SEMs had an average peak velocity of 11 degrees/s and an average latency of 39 ms. 2. The initial acceleration and peak velocity of elicited SEMs increased with stimulation intensity at any given site. On the other hand, SEM direction was characteristic of a given stimulation site and did not vary with stimulation intensity. These findings indicate that SEM amplitude is coded by the intensity of neural activity, and SEM direction is coded by the location of this activity within the cortex (“rate” vs. “place” codes). 3. SEMs elicited in the presence of a stationary fixation target (closed-loop conditions) typically reached a plateau velocity early in the stimulation and maintained that velocity throughout most of the stimulation train. However, when retinal slip was eliminated by artificially stabilizing the fixation target on the fovea (open-loop conditions), the electrical stimulation caused the eye to accelerate for longer periods and to attain higher velocities than in the closed-loop condition. Eye velocities obtained at the same site in open- and closed-loop conditions diverged from one another approximately 100 ms after SEM onset, consistent with the visual latency of the pursuit system. These findings suggest that the FEF primarily conveys an eye acceleration signal, rather than an eye velocity goal, to the pursuit system, and that this signal can be affected by visual retinal errors before effecting the smooth eye movements. 4. SEMs were elicited from a small portion of the arcuate fundus and neighboring posterior bank lying directly posterior to the principal sulcus. Functionally, this SEM region was surrounded by the saccadic FEF and by somatic premotor cortex. 5. Even though ipsilateral, contralateral, and vertical SEMs were elicited, the distribution of SEM directions was skewed toward ipsilateral movements. This tendency was more pronounced for sites in the arcuate fundus, whereas SEMs elicited from the posterior arcuate bank were often directed contralaterally and vertically.(ABSTRACT TRUNCATED AT 400 WORDS)