The responses of single units in the cerebellum, the vestibular nuclear complex and adjacent regions of the brainstem and in the oculomotor nucleus were studied in decerebrate, paralysed rainbow trout (Salmo gairdneri). Natural vestibular stimulation was provided by horizontal, sinusoidal oscillation of the fish and extraocular muscle afferents of the eye ipsilateral to the recording were activated either by passive eye-movement or by electrical stimulation of the trochlear (IV) nerve in the orbit. Unit responses to vestibular and/or orbital stimuli were examined in peristimulus-time histograms interleaved in time. In the cerebellum and brainstem, of 124 units exposed to both types of stimulus, 26 (21%) responded only to vestibular input, 26 (21%) were affected only by the orbital signal and 23 (18%) received both signals. The remaining 49 units (39%) responded to mechanical stimulation of the head or body or to vibration; they were labelled "polymodal" and discarded. The recording sites of 56 units were verified by histology; 30 were in the cerebellum and 26 in the brainstem. Input from the eye muscles had excitatory or inhibitory effects upon the vestibular responses. The effects of the orbital signal were usually phasic but rare tonic responses also occurred. About half (15 of 34) of the units which responded to passive eye-movement showed statistically significant differences in the magnitude of their responses to horizontal and to vertical eye-movement. More units preferred horizontal movement (11) than preferred vertical passive eye-movement (four). Note that the plane of vestibular stimulation was always horizontal. In the region of the oculomotor nucleus, of 19 units, five (26%) gave vestibular responses only and three (16%) were affected only by the orbital signal; three units (16%) with polymodal responses were discarded. Of the eight units carrying both signals, histological confirmation that the recording site lay in the column of cells forming the oculomotor/trochlear nuclei was obtained in four. The responses and interactions were similar to those found in the brainstem. The results present two principal points of interest. 1. They reinforce the accumulating body of evidence that, in species with widely different oculomotor and visual behaviour, signals from extraocular muscle proprioceptors reach the vestibulo-ocular system; this, in turn, suggests that these signals may play some rather fundamental role in the oculomotor system.(ABSTRACT TRUNCATED AT 400 WORDS)