Page:JOSA-Vol 06-06.djvu/15

 (a) When plotted in the form of a curve, the ordinates of a spectral distribution represent “intensity” per abscissa unit (frequency or wave-length, as the case may be); and the intensity concept, for the essential case of the incidence of the radiant energy upon the retina, will be: energy per second per unit area. To be completely specific, the function must express absolute values, but this is often difficult in practice.

(b) The wave-length unit which is ordinarily employed in colorimetrics is the millimicron which is correctly symbolized by m&mu; (not &mu;&mu;).'

(c) It is to be noted that wave-length, strictly interpreted, does not furnish a reliable specification of the color-stimulating capacities of radiant energy, as the response of the visual system depends upon frequency, while wave-length may vary independently of frequency. Since wave-lengths can only be interpreted in colorimetrics as reciprocal representations of frequency, it would be desirable theoretically to employ frequency directly in formulating spectral distributions. A suggested unit of frequency is the fresnel, defined as one vibration per trillionth (10$-12$) of a second. provides means for interconverting between millimicrons and fresnels.

(d) Spectral distributions of transmission, reflection, luminosity, etc., which are often employed to specify “color,” may be regarded as constituents or as developments of the essential distribution function (vide infra).

D. —for the purposes of colorimetrics—is radiant energy, sensibly all of the intensity of which lies within a single spectral region so small as to exhibit— under the conditions most favorable for discrimination—no perceptible hue difference within the region.

E. —The purity of any sample of radiant energy, with respect to any one of its constituents, may be defined in general as the ratio of the intensity of this constituent to the total intensity of the sample. By physical purity we may mean such a