An occluded contour becomes visible with reversal of disparity.

Suppose that there is a circle in front of a cross, the central part of which is occluded by the circle. One usually perceives these figures, as such, although the objective description of the stimulus pattern might be that there is a circle with four limbs attached to it. Here the occluded contours of the cross might be 'inferred.' Until now it has not been demonstrated clearly whether this 'inference' is based on higher cognitive processes or on perceptual processes. However, by manipulating binocular disparity so that the four limbs are seen in front of the circle, the occluded contours become visible, the color of the limbs spreads into the inside of the contours, and the color spreading surface appears transparent, suggesting that the 'inference' is on a perceptual basis. Interestingly, cornered subjective contours of the cross are observed.