Gibson emphasized that the texture of a plane that yields motion perspective provides richer, more naturalistic information than do laboratory environments that consist of such artificial elements as one or more spots floating in space. Wallach noted that in kinetic depth displays, in contrast to simple examples of motion parallax, not only do points within the object shift in direction as the object moves but the contours of the object simultaneously change in orientation and length. He believes it is this coupling of length change and orientation change that is crucial to the perception of depth generated by motion. Other factors may influence the effectiveness of motion as a depth cue. For example, in the kinetic depth effect the stimulus is a connected object. Stanley Anton demonstrated in our laboratory that when only the ends of the rotating rod are visible, as separate spots, the depth effect does not occur. The amount of depth separation may also influence the effectiveness of motion as a depth cue. In the case of stereopsis, when the disparity is too great, one typically sees double images of the non- fixated contour rather than one contour farther or nearer than the fixated one. If this rule holds for motion parallax, it may explain the impressive sense of depth achieved by motion of a band within a random dot pattern in the experiment Rogers and Graham conducted. The actual depth simulated in that experiment was far less than that employed in the experiment on luminous circles described earlier. Still another factor may be whether it is the display or the observer that is moving. For ideal conditions such as in the Rogers and Graham experiment or in a kinetic depth display, object movement suffices, but for poorer conditions such as in the luminous circles experiment, only movement of the observer may yield a depth effect. When the observer moves, information is available that the changing retinal image is not necessarily the result of object motion.