1. One hundred and fifty-two vestibularly activated neurones were recorded in the parieto-insular vestibular cortex (PIVC) of four awake Java monkeys (Macaca fascicularis): sixty-two were tested systematically with visual stimulation and seventy-nine were tested with various somatosensory stimuli. With very few exceptions all vestibular neurones tested responded to visual and somatosensory stimulation, therefore being classified as polymodal vestibular units. 2. A most effective stimulus for all fifty-eight visually activated PIVC units was movement of a large structured visual pattern in an optimal direction. From forty-four units responsive to a horizontally moving optokinetic striped drum, twenty-nine were activated with optokinetic movement in the opposite direction to the activating vestibular stimulus ('synergistic' response), thirteen were activated optokinetically and vestibularly in the same direction ('antagonistic' responses) and two were biphasic. The gain of the optokinetic response to sinusoidal stimulation (average 0.28 (impulses s-1) (deg s-1)-1 at 0.2 Hz, 56 deg amplitude) was in a range similar to that of the vestibular gain at low frequencies. At 1 Hz some units only showed weak optokinetic responses or none at all, but the vestibular response was still strong. 3. With different 'conflicting' or 'enhancing' combinations of optokinetic and vestibular stimulation no generalized type of interaction was observed, but the responses varied from nearly 'algebraic' summation to no discernible changes in the vestibular responses by additional optokinetic stimuli. With all visual-vestibular stimulus combinations the responses to the vestibular stimulus remained dominant. 4. The optokinetic preferred direction was not related to gravitational coordinates since the optokinetic responses were related to the head co-ordinates and remained constant with respect to the head co-ordinates at different angles of steady tilt. 5. Almost all PIVC units were activated by somatosensory stimulation, whereby mainly pressure and/or movement of neck and shoulders (bilateral) and movement of the arm joints elicited vigorous responses. Fewer neurones were activated by lightly touching shoulders/arms or neck, by vibration and/or pressure to the vertebrae, pelvis and legs. 6. A most effective somatosensory stimulus was sinewave rotation of the body with head stationary. The gain of this directionally selective neck receptor response was in the range of vestibular stimulation. Interaction of vestibular and neck receptor stimulation was either of a cancellation or facilitation type.(ABSTRACT TRUNCATED AT 400 WORDS)