Purpose: This study examined the dissociation of perception and action when participants judged the orientation of a frontal trapezoid window or pointed to its edges. Viewed monocularly the perspective cues in the display create the illusion of a window strongly slanted in depth. When viewed binocularly, the illusion is reduced. Method: Participants (N=14, mean age 24 yrs.) wore LCD goggles allowing random presentation of monocular and binocular vision over trials. For the Perceptual Task participants aligned the orientation of a rod that was slowly rotated by an experimenter to match the orientation of the display (total viewing time: 7.5s). For the Motor Task participants pointed with their index fingers to the side of the display without touching it (total viewing time: 2.5s). Finger movements were recorded with an optoelectronic camera system. The perceived orientation angle (perceptual error), the transverse plane angle between the two finger tips (motor error) and reaction time (RT) were analyzed. Results: The average perceptual error with respect to the true orientation of the display was 13.9° for binocular vision and 24.9° for monocular vision. The average motor error was 1.3° for binocular vision and 19.6° for monocular vision. Preliminary analysis revealed an average RT of 214ms for binocular vision and 227ms for monocular vision. Conclusions: Perspective cues for slant induced errors in both perception and action. However, during pointing towards the trapezoid window only the monocular presentation caused significant motor error. These results demonstrate that binocular information has a differential influence on perception and action, which may reflect processing differences in the dorsal and ventral visual stream.