1.) Eye movements were recorded in four Java monkeys (M. fascicularis) after unilateral microinjections (1 microliter, concentration 1 micrograms/microliter) of the GABA antagonist, bicuculline, and the GABA agonist, muscimol, into oculomotor related regions of the vestibular nuclei. Eye movements were investigated in the dark and light during spontaneous eye movements, vestibular stimulation (sinusoidal: 0.2 Hz, +/- 40 deg/s, and velocity trapezoid: 40 deg/s2 acceleration, 120 deg/s constant velocity), and visual-vestibular conflict stimulation. 2.) Bicuculline and muscimol injections consistently led to specific eye movement changes, which were maximal 5-10 min after bicuculline injection (muscimol 10-30 min), and lasted 90-120 min (muscimol 2-4 h). Control injections with NaCl (0.9%) into the responsive area and with bicuculline 2-3 mm more lateral showed no effect. 3.) Bicuculline induced a spontaneous nystagmus of 40.9 deg/s (average, range 10.5-93 deg/s), beating in 60% of the cases to the contralateral and in 40% to the ipsilateral side. The analysis of the slope of the slow phase gave no evidence for an additional gaze holding deficit. The VOR gain in the dark showed a slight decrease (pre: 0.96; post: 0.86) on average. The time constant of decay for slow phase nystagmus velocity after vestibular ramp stimulation was reduced, reflecting a 'velocity storage' deficit. After bicuculline injections nystagmus suppression in the light and during visual-vestibular conflict stimulation was generally well preserved. 4.) After muscimol injections horizontal gaze holding was severely affected. Each saccade was followed by an exponentially decreasing postsaccadic drift with a time constant as short as 250 ms (average 414 ms). The eyes always drifted towards a null-position, which generally did not coincide with the midposition of the eye. The null-position could move up to 35 deg to the contralateral or ipsilateral side. The highly distorted eye movements after muscimol injections prevented VOR-measurements based on eye velocity. Instead vestibular stimulation led to a shift of the null-position with an amplitude corresponding to a gain (eye position/stimulus position) of 0.17 (average) at 0.2 Hz (+/- 40 deg/s). Vertical eye movements did not show a major gaze holding deficit. 5.) From the experiments it can be concluded that the inhibitory transmitter GABA plays an important role for eye movement generation within the vestibular nuclei. Bicuculline induces mainly a vestibular imbalance with little evidence for a neural integrator deficit. In contrast unilateral muscimol injections lead to a complete, reversible loss of function for the common horizontal neural integrator, which converts eye velocity into eye position signals. The accompanying shift of the null-position reflects an additional vestibular imbalance.