Distinction between two classes of color constancy

Purpose. If a color perception as an object attribute (surface-color) was equal to mere appearance(apparent-color), e.g. white paper appeared white, this condition could be called "perfect color constancy." If not, it should he called "partial color constancy." Different mechanisms are possible to participate in the two classes of color constancy. So, we tried to find a border between two classes of color constancy, based on the sensitivity change of the visual system. Methods. We built a room with gray walls inside, with an illuminant on the ceiling. The illuminant was designed to change its chromaticity continuously from D₆₅ to blue, yellow, green or purple, at a constant illuminance. After 5 minutes of adaptation to an illuminant, the observer was asked to make either "unique-white (apparent-color)" or "white paper (surface-color)" setting with a stimulus on a CRT monitor. This CRT was carefully placed behind an aperture on the front wall, so that there were virtually no illuminant reflection on the surface of the CRT, when the observer viewed it through the aperture. Result. The surface-color settings exactly followed the illuminant chromaticities. Under illuminants near D₆₅ chromaticity, the observer did not noticed that he/she was in a chromatic illuminant, and the apparent-colors were almost the same as those in the surface-color matches. But a difference between surface- and apparent-color was clearer, under more chromatic illuminants. Conclusions. We looked into the change in visual sensitivity, by using cone-responses. According to this analysis, we suggest a border between partial color constancy and perfect color constancy, as a function of illuminant chromaticity.