1. In four Java monkeys (Macaca fascicularis) 152 vestibular neurones were recorded in the parietal cortex located in the upper bank of the lateral sulcus near the posterior end of the insula. We called this region parieto-insular vestibular cortex (PIVC). PIVC extends about 6-8 mm in the anterior-posterior direction from the posterior part of the insula into the retroinsular region (stereotaxic co-ordinates: anterior 4-12 mm, lateral 16-19 mm and vertical 3-6 mm). 2. About two-thirds of the neurones recorded from this region responded to vestibular stimuli; the non-vestibular neurones responded predominantly to somatosensory stimulation of the neck and shoulder region. The PIVC area is a polysensory field, since almost all vestibular neurones were also activated by somatosensory and visual stimuli. Large-field optokinetic stimulation was the most effective visual stimulus. 3. With vestibular stimuli, responses to angular acceleration were dominant; steady tilt in darkness rarely led to any change in neuronal spontaneous activity. Of sixty-four neurones tested by rotation in more than one plane, fifty-four responded to excitation of semicircular canal receptors aroused by rotation in more than one of the three experimental planes (roll, yaw, pitch). Compared with vestibular brain stem units PIVC neurones discharged with a higher variability. 4. In the responses to horizontal rotation of the animal 38% type I, 53% type II and 9% type III units were recorded (classification according to Duensing & Schaefer, 1958). The gain measured with horizontal sinewave rotation was lower by a factor of about 4 in PIVC neurones as compared with the responses of vestibular neurones in the brain stem or thalamus (VPL). The phase response characteristics and the gain increase with increasing sinewave stimulus frequency, however, were in the same range as observed in neurones of the afferent vestibular system. 5. When the vestibular responses to sinusoidal rotation were tested in all three experimental planes (yaw, roll, pitch), the response strength as expressed by the amplitude of the peristimulus time histograms (PSTHs) differed for the three rotational planes. For different units the relative sensitivity to rotation in each of the three planes also differed. We concluded from this observation that different PIVC units had different optimum sensitivity planes for rotation with respect to the head co-ordinates, whereby all possible planes are represented.