1. The gain and phase of the horizontal (H.v.o.r.) and vertical (V.v.o.r.) vestibulo‐ocular reflexes were measured in rabbits. The V.v.o.r. was evoked by sinusoidal rolls about the longitudinal axis of the rabbit. This axis was maintained at different orientations with respect to the earth horizontal axis: V.v.o.r. 0°, prone; V.v.o.r. 90°, ‘nose‐up’; V.v.o.r. 180°, supine; V.v.o.r. 0°L, left side down. 2. In contrast to the H.v.o.r., the V.v.o.r. 0° had a higher gain (eye velocity/head velocity) and a smaller phase lead (eye position + 180° with respect to head position) at low frequencies of sinusoidal vestibular stimulation (± 10 deg, 0·005‐0·05 Hz). At higher frequencies (0·05‐0·8 Hz), the H.v.o.r. and V.v.o.r. 0° were equivalent in both gain and phase. 3. The low‐frequency gain of the V.v.o.r. was smallest in the ‘nose‐up’ orientation. The V.v.o.r. 90° was equivalent in both gain and phase to the H.v.o.r. over the entire range of frequencies tested (0·005‐0·8 Hz). Threshold angular accelerations for the H.v.o.r. and V.v.o.r. 90° were below 0·04 deg/sec2. 4. The compensatory eye movements of the H.v.o.r. were frequently interrupted by anticompensatory re‐setting eye movements. These anticompensatory re‐setting eye movements were present in the V.v.o.r. 90°, but not in the V.v.o.r. 0°. 5. An estimate gain and phase of the otolithic component of the V.v.o.r. 0° was derived by subtraction of the V.v.o.r. 90° (semicircular canal signal) from the V.v.o.r. 0° (semicircular canal signal+otolith signal). This procedure was based on the assumption that the signals from the otolith organs and vertical semicircular canals combine linearly. 6. The V.v.o.r. 180° provided an interesting test of the assumption of linear combination of otolithic and semicircular canal signals by reversing the phase of the modulated otolithic signal. The data indicated that the V.v.o.r. 180° is non‐linear.
ASJC Scopus subject areas