Abstract
Purpose. To examine changes in magnitude of the Brentano version of the Mueller-Lyer illusion by varying the chromaticities of the shaft and arrows along the cone-excitation axes and the occlusion of the shaft. Method. The Brentano version of the Mueller-Lyer figure was displayed on center of a CRT monitor subtending a visual angle of 48° x 37° at a viewing distance of 45 cm. The stimulus consisted of a shaft of 15.2° and 3 arrows of 2.3° per side forming 45° angles to the shaft. All stimuli were equiluminant to the gray surround (11.6 cd/m2) except for the black stimuli. The combinations of shaft/arrow chromaticities were black/black, L−/L−, L−/L+ and L−/S− as expressed in the DKL cone-excitation space. When the chromaticities of the shaft and arrows were different, the arrows occluded the shaft or were occluded by it producing a total of 6 stimuli. The magnitude of the illusion was measured by having observers adjust the position of the middle arrow using keys on a keyboard until judged that the 2 segments of the shaft to either side of the arrow appeared equal in length. When satisfied, the observer pressed a key to indicated it. The screen turned homogeneous gray for 3 s and the next stimulus appeared. Observers made these adjustments 6 times per randomly presented stimulus. Observers were 45 naïve, normal trichromats. The magnitude of the illusion was obtained by dividing the longer by the shorter segment of the shaft. Results and Discussion. Consistent with the past research, stimuli with identical shaft/arrow chromaticity produced more illusion than stimuli with different chromaticities. L−/S− stimuli tended to produce more illusion than L−/L+ stimuli suggesting that the chromaticity similarity of the shaft/arrow components, not the chromaticity change along one axis, increases the illusion. When the middle arrow occluded the shaft rather than the opposite, the illusion was smaller. This is possibly because the observer can see the clear segmentation of the shaft by the middle arrow.
This study was supported by National Institute of Health National Eye Institute research grant EY13936.