Aug. 1922]
Colorimetry Report 1920-21
577
important, therefore, that the observers should be carefully selected so as to have standard visibility and color excitation curves. The direct application of this method is limited to colors possessing spectral hues. In order to specify purples, it is necessary to determine the wave-length of homogeneous radiation which when mixed with them matches the heterogeneous standard, i.e., which yields a gray of the given brilliance. It is clear, however, that the method is capable of dealing with colors of all degrees of saturation.
C. Trichromatic Analysis. (29; 30.) In this method the variable stimulus is composed of three constituents having appropriate spectral constitutions, which yield, respectively, colors corresponding in hue to the two end, and the middle regions of the spectrum. The colors ordinarily chosen are red, green and blue, and to render the field of applicability of the system as wide as possible the three stimuli should each be maximally homogeneous. Relative spectral distributions within the components remaining constant, their proportions and total luminosity are adjusted until a match is obtained. The measured color can then be designated by three intensity or luminosity values, one for each of the three components. If the spectral distributions of the components are determined by fairly narrow-banded filters, this method is capable of dealing with practically all reflection colors, which are relatively unsaturated, but is not satisfactorily applicable in its simplest form to many saturated filter or spectral colors. In order to extend the method to deal with the latter it is necessary to add a variable quantity of “white” to the sample which is being measured. If this is done the colorimetric significance of the added white can be recorded in the resultant measurements by subtracting its amount from each of the three-color readings, yielding a specification in terms of red, green and blue, with one minus coefficient, a mode of expression which is in no way incompatible with the trichromatic principle. The trichromatic analyses are of fundamental interest on account of their maximally direct relation to the triadic response mechanism which apparently underlies all color vision,