We measured torsional vestibular and optokinetic eye movements in human subjects with the head and trunk erect, with the head supine and the trunk erect, and with the head and trunk supine, in order to quantify the effects of otolithic and proprioceptive modulation. During active head movements, the torsional vestibulo-ocular reflex (VOR) had significantly higher gain with the head upright than with the head supine, indicating that dynamic otolithic inputs can supplement the semicircular canal-ocular reflex. During passive earth-vertical axis rotation, torsional VOR gain was similar with the head and trunk supine and with the head supine and the trunk erect. This finding implies that static proprioceptive information from the neck and trunk has little effect upon the torsional VOR. VOR gain with the head supine was not increased by active, self-generated head movement compared with passive, whole body rotation, indicating that the torsional VOR is not augmented by dynamic proprioceptive inputs or by an efference copy of a command for head movement. Viewing earth-fixed surroundings enhanced the torsional VOR, while fixating a chair-fixed target suppressed the VOR, especially at low frequencies. Torsional optokinetic nystagmus (OKN) evoked by a full-field stimulus had a mean slow-phase gain of 0.22 for 10 degrees/s drum rotation, but gain fell to 0.06 for 80 degrees/s stimuli. Despite this fall in gain, mean OKN slow-phase velocities increased with drum speed, reaching maxima of 2.5 degrees/s-8.0 degrees/s in our subjects. Optokinetic after nystagmus (OKAN) was typically absent.(ABSTRACT TRUNCATED AT 250 WORDS)