Abstract
1. Flatfish metamorphose from a larval form that swims upright like a standard fish to an adult that lies on one side, with both eyes on the upper side, having rotated posture 90° relative to gravity and the former normal posture. Adult Citharichthys stigmaeus and Hypsopsetta guttulata were used in behavioural and physiological experiments to determine whether the postural change is a peripheral or central phenomenon. 2. Cleared and sectioned specimens verify that the otolith organs, unlike the eyes, do not rotate within the skull, and so do not maintain the normal vertebrate orientation with respect to gravity. 3. Ocular compensation to lateral tilt shows that tactile cues, vision, and the semicircular canals are inadequate to produce tilt responses, but elimination of otolith function abolishes tilt responses. The major postural role of the otolith organs is not lost. 4. Selective removal of otoliths demonstrates that the flatfish utriculus has only a minor role in tilt responses, and that the sacculus-lagena is required, unlike the situation in other vertebrates. The details of the ocular compensation responses are similar to those of standard fishes. Each sacculus lies at an angle of up to 45° when in the normal position, but unilateral loss does not change the phase of the response curve, indicating that the null response is set for a non-zero value of gravitational shear, unlike the null at zero shear to the utriculus in other vertebrates. 5. Hysteresis effects suggest a differential sensitivity between tilts near the normal and the upside-down null positions. The narrowness of the effect argues against mechanical restrictions. Possibly the vertical utriculus is useful only near the normal, as an accessory organ, like the vertical lagena in other vertebrates. 6. Neural units recorded from both eighth nerve and medulla show the expected activity properties of regular and irregular rate, tonic and phasic responses to tilt, directional dependence and ‘multi-valuedness’, as in other vertebrates. No novel response types are found, nor any distinctive ‘into-level‘ types described for some vertebrates. Vibration sensitivity is associated with irregular rate, and exclusively vibration-sensitive units are apparent only in the utriculus. The shift in functional relations of the otolith organs relative to gravity is not apparently compensated for by any major change in peripheral afferent gravistatic unit properties. 7. An increasing distinction between the null at normal and the null upside-down shown by limited data on ocular compensation in three year-classes of flatfish. A central change in vestibular function is suggested that might be dependent on experience, as is gradual compensation to a vestibular lesion. 8. Since peripheral changes are not responsible for the postural change, alternative central mechanisms are proposed, including central weighting of input, recognition of a complex input pattern, and plasticity of connexions, all of which have received some supporting evidence from these results.