During the past 150 years, the discrimination of colors has been studied within three experimental traditions: the König tradition of measuring wavelength discrimination; the MacAdam tradition of measuring confusion ellipses within the interior of the chromaticity diagram; and the Stiles tradition of measuring increment thresholds for monochromatic flashes. By the second half of the twentieth century, it was clear that color discrimination could not be explained in terms of the retinal cone signals alone, as Helmholtz had hoped. Instead, a successful model would require chromatically opponent neural channels at a post-receptoral level. A particularly insightful paper was that of Le Grand (1949), who re-analyzed MacAdam's ellipses in terms of two channels, one that represented the signal of the short-wave cones scaled by the current luminance and a second that differenced the signals of the long- and middle-wave cones. This attractively simple (and long dominant) account, in terms of two 'cardinal' channels, has been increasingly challenged in the present century; and very recent retinal research suggests several different pathways by which the short-wave cones may contribute to color discrimination. Any satisfactory model of discrimination must certainly incorporate one fundamental finding: discrimination is optimal at the chromaticity to which the observer is adapted. Since we do not know the degree to which MacAdam's single observer was adapted to the chromaticity currently being matched, MacAdam's ellipses are not a secure basis for deriving a uniform color space (a diagram in which equal distances would represent equally noticeable color differences) or for estimating the number of distinct color sensations that a normal trichromat can potentially experience. A related consideration is that a sensation of difference is logically not the same as a difference of sensations: some color discriminations may depend on edge-detecting channels that extract differences between local regions, and in principle it is possible for different pairs of chromaticities to give rise to the same sensation of difference.
Mollon et al. (Tue,) studied this question.