Is a perceived shape based on its retinal image?

Although the processing of phenomenal shape might be supposed to begin at an early stage, with the shape of the retinal image of an object, it is possible that it does not begin until a later stage at which the locations of the parts of the object have been perceived. Such perceived locations are based on a compensation or constancy mechanism that takes account of eye position. Ordinarily these two possible bases of shape perception--retinal image and perceived location--are confounded. To separate them the parts of a shape were presented sequentially, during which time the eyes were in motion. The eye movement did not alter phenomenal locations of the parts vis-à-vis one another but did yield an entirely different composite retinal image of the parts. Another method employed was to change the location of each sequentially presented part with respect to a displacing frame of reference. By and large the results indicate that the composite shape perceived is based on the perceived location of the parts of the object with respect to one another, rather than on the composite retinal image.