Effect of stimulus size on chromatic discrimination

In psychophysical experiments, discrimination thresholds were measured in different directions around points in the MacLeod-Boynton chromaticity diagram, while the eye was maintained in a state of constant adaptation to a metamer of D65. A spatial forced-choice procedure was used: a brief (150 ms) disk divided into four sectors was presented, and the observers' task was to detect the sector that differed from the other three. The diameter of the test disk varied from 32 min to 2.4^∘ of visual angle. Sensitivity was probed at several different referent positions in the chromaticity diagram, including the adapting chromaticity. The data for each referent were fitted with ellipses. In the case of the largest test size (2.4^∘ diameter), ellipses were predominantly oriented so that their longer axis was aligned with the line connecting the center of the ellipse to the chromaticity of D65 (the adaptation point). Along such radial lines, colorimetric purity varies, and the orientation of the ellipses reflects reduced sensitivity to saturation differences compared to hue differences. With decreasing test size, the ellipses change their orientation so that their longer axis is rotated toward a tritan direction, and the detection of changes in chromaticity depends primarily on the activity of long- and middle-wave cones. However, these general principles must be modified in several ways according to the region of the chromaticity diagram that is being probed.