Using a stabilization device capable of covering 30° of the visual field, targets of various shapes were studied extensively. Very large and consistent differences were found between the percentage disappearances of most targets and this was found not to be a function of length of line or boundary. The presence of corners and intersections in a target increased percentage disappearance. Jagged, angular figures disappeared more than rounded, topologically similar ones. When acute angles were present, the disappearance rate was radically raised.

About 10% of the time when a target disappeared, it disappeared as a complete unit, though this figure varied somewhat from target to target and was very high in the case of the circle. A mean of about 15% of all disappearances were ‘patterned’ or ‘structured’, and these proportions held true for all subjects tested.