Color Artists speak of cold, warm, advancing, and retiring colors. By warm are meant red, orange and yellow, and any tendency toward them ; it has no connection with heat, but is suggested rather by the lumi nous glow of sunlight. An advancing color seems to be nearer the eye than one which is called retiring; the effect is evident in only a few simple cases: orange seems to be nearer than green when both occur in juxtaposition on the same surface, green nearer than violet, yel low than.red, red than blue, black than yellow. The reader may easily experiment with the excellent colored coated-papers suitable for the purpose, various sizes of which can be obtained in black, white and in 18 spectral hues, with two tints and two shades of each. Color has three characteristics: hue, purity or saturation, and luminosity or brightness. Hue is color per se, such as blue, pink, black, etc. • purity depends on freedom from admixture witii white; only mon ochromatic colors are saturated. Luminosity de pends on the amount of light given by the color. A tint is a departure from a normal or dominant spectral hue in regard to purity, i.e., tints are whiter than the arbitrarily selected normal. A shade is darker than the normal and may be produced by decreased illumination or by the addition of black. It must be clearly under stood that luminosity does not affect purity: a tint cannot be made to look as saturated as its normal merely by decreasing the illumination, thus pink never looks red in a subdued light. A compound color can be defined accurately only in terms of the characteristics of its monochro matic components. But every color, simple or compound, except purple, can be matched so far as the eye is concerned by a tint or a shade of some simple spectrum hue.
Color constants are measured with a spec trophotometer, which is a combination of a spectroscope with a photometer. The spectro scope produces an ordinary white-light spec trum, one narrow transverse band of which can be passed through the photometer; at the same time white light of any intensity is superposed on this colored band. The colored light which is to be examined—in the case of a colored surface the light reflected by it — also passes through the photometer and forms a field next to the combination of spectral ray and white light. By selecting the proper hued spectrum ray, controlling the added white and adjusting the resultant until it matches in color and brightness the light to be examined, we can determine hue, saturation and luminosity. De scriptions of instruments and full references are given by Luciciesh. The following are typical results; they will vary with the reflection coefficient of the surface, i.e., with the relative amount of white light it reflects. The wave lengths of the hues are expressed in microns: 1 micron =0.001 millimeter. The columns headed aper cent white* give the amount of white light which had to be added to the normal spectrum hue to reduce its saturation.
These results are color matches rather than color analyses. They show, for example, that highly colored butter, manila paper, chrome yel low paint, the nitrogen tungsten lamp and the carbon arc light are all of about the same hue arid differ only in purity and brightness, i.e., they
are tints and shades of one another. As the three observers did not use identical light sources in their experiments, the purities in the tables are relative. The following values from Hurst ( (Color,' p. 17) give a rough idea of the lumi nosity of, or the amount of white light reflected by, various pigments in terms of white paper as a standard.
White paper 100 Emerald green 51 Vermilion 20 Ultramanne 50 Orange red 40 Umber 22 Yellow ochre 56 Chrome yellow 61 On comparison with the first table it is seen that although vermilion is very pure, very nearly saturated, it reflects comparatively little lir. The two sets of numbers are quite indepen ent and express two different characteristics of color.
Color Analysis.— Analysis of a compound color is used here to mean separation into spec tral components. For example, French ultra marine is a fair match to what is regarded as the typical spectral blue of wave length 0.472 microns; both give rise to the same sensation and are therefore physiologically identical, yet the spectrum ray is monochromatic and the pig ment decidedly complex. If the light reflected from a colored surface is analyzed spectro scopically it is found to contain all the spec trum hues in varying intensity.
The following diagrams, adopted from Luc kiesh, give spectroscopic analyses of certain pig ments. Vertical distances indicate the relative intensities with which the wave lengths along the base are reflected and correspond to lumi nosities in the absorption spectra. The area under a curve measures the luminosity of the pigment itself. White is a horizontal line at greatest height because it reflects all colors equally; black reflects none and therefore is chosen as the base. Gray would be some inter mediate horizontal line. No analyses of mixed pigments are as yet available and it is difficult to predict the exact form of the resultant curve. However, we can arrive at approximate results. According to the diagrams a surface painted with cadmium yellow will look intensely yellow when illuminated with monochromatic yellow light, orange in orange light, red in red and green in green, but in blue or violet light it will be grayish. French ultramarine will be violet, blue, green in corresponding illuminations but grayish in yellow, orange, red. The only color which cadmium yellow and ultramarine both reflect is green, the others being largely but not entirely absorbed. The green produced by a mixture of these two pigments although far from being monochromatic is much more nearly so than the pigments themselves. But its luminosity has been considerably reduced by absorption; in fact a mixture of monochromatic blue and yellow pigments would be black be cause no light would be reflected at all. Mixed pigments tend toward blackness; painters there fore select pigments which possess inherent color and do not mix unless a subdued, murky or grayed effect is desired. Unfortunately there is as yet no accepted theory of inherent color; it is not known why chromium oxide is green, hydrated ferric oxide red or ferric hy droxide yellow. Consult (Berichte d. deut. chem. Gesell) (1906, 39 p. 1959), and Journal of the Chemical Society (1906, 89 p. 1787) ; see also DYEING.