Abstract
Some properties of color vision are hardly consistent with the opponent interaction of cones and suggest a significant rod contribution. An example is the long-wave-threshold difference between the two basic blue-yellow and red-green sensations whose spectral intensities were calculated by E.Schrödinger from the color-matching data. The third basic function of achromatic sensation has never been calculated for lacking linear-transform criteria and the general intensity of color sensation was conventionally considered instead. The 1955 Stiles-Burch 2° RGB-functions have been analyzed with a modification of the method of detecting some criterion mathematical properties of unknown basis functions from their experimental linear combinations (3rd Intl.Conf. “Independent Component Analysis and Blind Signal Separation”, UCSD, pp.200–205, 2001). The third basic function has turned out to be zero at the wavelengths under 417 nm and over 660 nm as well as proportional to the blue-yellow component over 540 nm. The function calculated (max. ∼540 nm) was independent of wavelength choice within the spectral ranges. It closely corresponded to the Helmholtz's ‘white content of the pure spectrum colors’ and further evidence for their perceptual saturation differences. The achromatic component inconsistent with a sum of cone signals is discussed with involving the latest rod contribution concepts.