The role of knowledge of the reversibility of reversible figures was tested in four experiments. Two ambiguous figures, the vase-face figure and a depth-reversing pyramid-hallway figure were shown to high school students. In the Uninformed condition, subjects were not told that the figures were reversible. A sampling procedure was used in which subjects reported what they perceived at 5-sec intervals. Viewing durations of up to 3 min were used, and approximately half of all subjects did not reverse at all during the uninformed condition, whereas virtually all subjects reversed quickly and frequently once they knew that the figures were reversible. These results are not consistent with neural fatigue models of perceptual reversal.