Page:JOSA-Vol 06-06.djvu/24

From Wikisource
Jump to navigation Jump to search
This page has been proofread, but needs to be validated.
546
L. T. Troland
[J.O.S.A. & R.S.I, VI

number of hue steps in cycles of color with uniform saturation,and it is quite uncertain whether the change in number of steps follows the simple geometrical analogy: H=2πs, where s is the saturation measure in threshold steps from the equivalent gray. (This clearly raises the question as to whether the space of the psychological color solid is Euclidean.) The angular magnitudes in this system would preferably be measured from an axis through the normal hue of extreme spectral red, on account of the stability of this color in relation to its stimuli, and all functions of these magnitudes will be periodic.

The positions of the psychologically primary hues in the “hue scale” are matters of considerable interest. On Jones’ spectral chroma scale, taking the zero in the violet and unity at extreme spectral red, the primary blue, green and yellow lie at .24, .41, and .68, respectively, the red having a value slightly in excess of 1.00, on account of the necessity of including a slight amount of blue in the stimulus for red. (95). It will be seen that the separations of the primaries on this scale are by no means equal.

3. The Saturation Scale

Careful determinations of the number of just noticeable saturation steps between each maximally saturated color and white have not yet been made. Nutting and Jones (67; 68) find about 20 such steps for red, green and blue, the thresholds for change in the per cent. white varying with the given per cent. white as shown in Table 4

Table 4
Saturation Scale Data

Per cent. white 0 10 20 30 40 50 60 70 80 90 100
Threshold white 4.7  4.6  4.5  4.4  4.2  4.0  3.7  3.4  3.0  2.5   2.1

These values, like those for wave-length sensibility, are independent of brilliance over a wide range of intensities.

The Chairman has attempted to determine the relative saturations of the spectral colors by an application of the flicker photometer. It is assumed that when any color is alternated with