The horizontal and vertical vestibulo ocular reflex (VOR) of head tilted (het) mutant mice was compared to C57BL/6 controls. Eye movements were recorded in darkness using a temporarily attached search coil. Contributions of semicircular canal versus otolith organ signals were investigated by providing a canal only (vertical axis) or canal plus otolith organ (horizontal axis) stimulus. In controls, rotations that stimulated only the canals (upright yaw and on tail roll) produced accurate VOR timing during middle frequency rotations at 0.5 Hz (gain 0.27, phase error 6°), and a phase advanced VOR during low frequency rotations at 0.05 Hz (0.05, 115°). In het mutant mice, these rotations produced a highly attenuated VOR response and phase errors at both 0.5 Hz (0.11, 42°) and 0.05 Hz (0.01, 36°). In controls, rotations that stimulated both the otolith organs and semicircular canals (upright roll and on tail yaw) produced higher VOR gains overall than were elicited during vertical axis rotations, with phase accurate VOR at both 0.5 Hz (0.52, 4°) and 0.05 Hz (0.34, 9°). In het mutant mice, these rotations produced a highly attenuated VOR response and phase errors at both 0.5 Hz (0.14, 51°) and 0.05 Hz (0.01, 43°). During constant velocity rotations about an earth horizontal axis, eye velocity bias and modulation were virtually absent in het mutant mice, while robust in controls. Control mice produced compensatory ocular deviations in response to static head tilt, but responses in het mice were weak and inconsistent. These results show that het mice not only lack all aspects of otolith mediated VOR, but also are deficient in canal mediated VOR. Because semicircular canals are normal in het mice, we conclude that central neurons of the canal VOR are dependent on otolith organ signals for normal performance. [Copyright &y& Elsevier]