A series of experiments were made on human performance in controlling optical relative movement. The aim was to test the influence of different kinds of relative movement on visually controlled steering tasks. Within adjacent displays on a computer screen random dot patterns moved in a fixed direction at continually changing speeds (Exp. 1) or at constant speed and in continually changing directions (Exp. 2). The subject was required to compensate for the unpredictable modulations of the pattern movement by means of an isometric joystick. The task was to adjust relative movements involving pure translation, symmetric convergence, divergence, or shear. Analysis indicated that the task performance was not dependent on the special kind of relative movement. However, performance was significantly higher in tasks where directionally disturbed relative movement had to be controlled compared to those situations in which relative movement varied with respect to speed.