1. We electrically stimulated the intermediate and deep layers of the superior colliculus (SC) in two rhesus macaques free to move their heads both vertically and horizontally (head unrestrained). Stimulation of the primate SC can elicit high-velocity, combined, eye-head gaze shifts that are similar to visually guided gaze shifts of comparable amplitude and direction. The amplitude of gaze shifts produced by collicular stimulation depends on the site of stimulation and on the parameters of stimulation (frequency, current, and duration of the stimulation train). 2. The maximal amplitude gaze shifts, produced by electrical stimulation at 56 sites in the SC of two rhesus monkeys, ranged in amplitude from approximately 7 to approximately 80 deg. Because the head was unrestrained, stimulation-induced gaze shifts often included movements of the head. Head movements produced at the 56 stimulation sites ranged in amplitude from 0 to approximately 70 deg. 3. The relationships between peak velocity and amplitude and between duration and amplitude of stimulation-induced head movements and gaze shifts were comparable with the relationships observed during visually guided gaze shifts. The relative contributions of the eyes and head to visually guided and stimulation-induced gaze shifts were also similar. 4. As was true for visually guided gaze shifts, the head contribution to stimulation-induced gaze shifts depended on the position of the eyes relative to the head at the onset of stimulation. When the eyes were deviated in the direction of the ensuing gaze shift, the head contribution increased and the latency to head movement onset was decreased. 5. We systematically altered the duration of stimulation trains (10-400 ms) while stimulation frequency and current remained constant. Increases in stimulation duration systematically increased the amplitude of the evoked gaze shift until a site specific maximal amplitude was reached. Further increases in stimulation duration did not increase gaze amplitude. There was a high correlation between the end of the stimulation train and the end of the evoked gaze shift for movements smaller than the site-specific maximal amplitude. 6. Unlike the effects of stimulation duration on gaze amplitude, the amplitude and duration of evoked head movements did not saturate for the range of durations tested (10-400 ms), but continued to increase linearly with increases in stimulation duration. 7. The frequency of stimulation was systematically varied (range: 63-1,000 Hz) while other stimulation parameters remained constant. The velocity of evoked gaze shifts was related to the frequency of stimulation; higher stimulation frequencies resulted in higher peak velocities. The maximal, site-specific amplitude was independent of stimulation frequency. 8. When stimulating a single collicular site using identical stimulation parameters, the amplitude and direction of stimulation-induced gaze shifts, initiated from different initial positions, were relatively constant. In contrast, the amplitude and direction of the eye component of these fixed vector gaze shifts depended upon the initial position of the eyes in the orbits; the endpoints of the eye movements converged on an orbital region, or "goal," that depended on the site of collicular stimulation. 9. When identical stimulation parameters were used and when the eyes were centered initially in the orbits, the gaze shifts produced by caudal collicular stimulation when the head was restrained were typically smaller than those evoked from the same site when the head was unrestrained. This attenuation occurred because stimulation drove the eyes to approximately the same orbital position when the head was restrained or unrestrained. Thus movements produced when the head was restrained were reduced in amplitude by approximately the amount that the head would have contributed if free to move. 10. When the head was restrained, only the eye component of the intended gaze shift