Color-naming responses to red light of varying luminance and purity.

Abstract

A 10×10 Greco-Latin square composed of two 10-value sets of neutral density filters projected with red and tungsten lights provided a combination image in which the arrangement of hues appeared unsystematic. Nine subjects gave a total of 1326 color-naming responses from which were derived estimates of hue, lightness, and saturation for each cell of the square. Saturation varied as a function of the purity of the red light; lightness was found to be associated mainly with relative luminance, and hue was related both to proportion of red light and to relative luminance. Hue ranged over nearly the entire circuit, while lightness and saturation varied over considerable ranges. Analysis by concentric layers or rings of the sampling space described the response-events clearly, while analysis by constant chromaticity lines through the sampling space showed that single hues did not lie along such lines.