As an individual moves, the eyeballs must rotate at the same speed, but in the opposite direction, as head rotation for images to be stabilized on each retina. Such movemenis are made automatically by the vestibulo-ocular reflex (VOR). Its effectiveness may be measured by the gain of the reflex: eye velocity divided by head velocity. Normally the gain is close to 1.0, so that images do not slip on the retina when the head moves. The VOR must have some type of maintenance system that monitors its gain (by vision] and corrects it when it falls out of calibration. In a young animal this system is responsible for maintaining calibration as the organism grows. Growth of the skull-which alters the angular relations between the orbits' and the semicircular canals-and changes in the mechanical parameters of the eyeball and eye muscles must be compensated if the system is to continue working. In the mature animal the task becomes one of preserving the gain of the vestibulo-ocular system in spite of the continual disturbing effects of cell death during aging. The enormous anatomical variations between members of a species suggest that genetic programming alone cannot maintain the VOR.'The mechanism responsible for optimizing the VOR has been described as a motor-learning system capable of making parametric changes by modifying synaptic~ trengths.~,~ The ability of this repair mechanism to change the VOR has been demonstrated by adapting animals and humans to optical devices that change the relationship between the apparent motion of visual objects and head rotation. Melvill Jones and his colleagues have studied the effect of chronic vision through Dove prism^.^-^ These prisms reverse the seen world from left to right and make it appear to move in the same direction as the head when the latter moves. After an adaptation period, the gain was found to be reduced. Miles and his colleagues found that a telescope lens system that magnifies the visual world can be used to raise the VOR gain.','In each of these and similar situations, the gain always rose or fell in such a way as to lessen or eliminate retinal image slip during head movements. This form of adaptive motor plasticity now has been well established in a variety of species.

Not only must the speed of a compensatory eye movement be correct, its direction must be just opposite to that of the head to keep the line of sight stationary in the visual environment. We can think of vestibular eye compensatory movements as occurring in a plane parallel to the plane of head rotation. The wrong orientation of the plane of the eye movement would cause images to slip on the retina. If the planes of action of pairs of extraocular muscles were exactly parallel to those of synergistic pairs of semicircular canals, the arrangement