1. We recorded eye movements in six rhesus monkeys before and after unilateral labyrinthectomy and quantified the compensation for both the static and the dynamic disturbances of the vestibuloocular reflex (VOR). 2. When first recorded after labyrinthectomy (18-20 h postlesion), all animals had a spontaneous nystagmus with mean slow-phase velocities ranging from 24 to 54 degrees/s measured in darkness and 0-4 degrees/s measured in the light. The level of nystagmus diminished quickly, and by postoperative day 25 mean values ranged from 4 to 22 degrees/s, measured in darkness. The waveform of individual slow phases was variable, but in the first postoperative week its trajectory usually showed an increasing, or an increasing then decreasing, velocity. This finding indicates that peripheral vestibular lesions can alter the function of the ocular motor eye-position integrator. 3. The VOR gain (eye velocity/head velocity, corrected for spontaneous nystagmus) during rotations (30-300 degrees/s) in the dark was diminished from nearly 1.0 preoperatively to approximately 0.5 when first measured after labyrinthectomy, except for rotations toward the lesioned side at high speeds for which the gain was even lower. Within the first few postoperative days, for rotations toward the intact side, the VOR gain increased rapidly, to approximately 0.8. For rotations toward the lesioned side similar behavior was noted for stimuli of 30-60 degrees/s, but at higher velocities compensation proceeded more slowly. By 3 mo postoperatively gains had reached values ranging from 0.77 to 1.03 for rotations toward the intact side and from 0.61 to 0.98 for rotations toward the lesioned side. Values were higher for lower-velocity stimuli. 4. Caloric testing with ice water in the unoperated ear elicited nystagmus with a mean value of maximum slow-phase velocity of 129 degrees/s preoperatively and 195 degrees/s 3 mo postoperatively. There was no caloric response on the lesioned side. From the increase in caloric responses from the intact ear we infer considerable restoration of spontaneous activity of vestibular neurons on the deafferented side. 5. The time constant of the VOR was a function of stimulus speed preoperatively with a maximum mean value of 35 s for a 60 degrees/s stimulus. After labyrinthectomy the VOR time constant was low (6.0-9.1 s) at all speeds. Subsequently, in three animals only, there was a small increase (2-3 s) in VOR time constant during the 3-mo period following labyrinthectomy. These results indicate that labyrinthectomy profoundly and persistently impairs the action of the vestibular velocity-storage mechanism.(ABSTRACT TRUNCATED AT 400 WORDS)