The disability resulting from uni-or bilateral loss of vestibular function was assessed by exposing subjects, in darkness, to random rotational displacements on a motorized chair away from a "center" position and requiring them to rotate themselves back to center using a "joystick" control. Normal subjects were accurate within 10 to 15 percent over 180-degree rotations. Alabyrinthine subjects failed to sense their rotation to any useful extent. Subjects with unilateral loss of function as a result of surgery produced hypometric responses to ipsilateral rotation and large amplitude, but inaccurate responses to contralateral rotation. This pattern became more symmetric and normometric with adaptation to loss of function. The procedure promises to be a more useful measure of adaptation than assessing vestibulo-ocular reflexes.