fact, seven fringes on the side of the film which is covered by the green glass will be seen to cover about the same distance as six fringes on the red side.*

Since it was shown in Fig. 390 that the distance between two dark bands corresponds to an increase of one-half wave length in the thickness of the film, we conclude, from the fact that the dark bands on the red side are farther apart than those on the green side, that red light must have a longer wave length than green light. The wave length of the central portion of each colored region of the spectrum is roughly as follows:

.000042 cm.

.000052 cm.

.000046 cm.

Violet

Let the red and green glasses be removed from the path of the beam. The red and green fringes will be seen to be replaced by a series of bands brilliantly colored in different hues. These are due to the fact that the lights of different wave length form interference bands at different places on the screen. Notice that the upper edges of the bands (lower edges in the inverted image) are reddish, while the lower edges are bluish. We shall find the explanation of this fact in § 473.

465. Composite nature of white light. Let a beam of sunlight pass through a narrow slit and fall on a prism, as in Fig. 440. The beam which enters the prism as white light is dispersed into red, yellow, green, blue, and violet lights, although each color merges, by insensible gradations, into the next. This band of color is called a spectrum.

We conclude from this experiment that white light is a mixture of all the colors of the spectrum, from red to violet inclusive.

* The experiment may be performed at home by simply looking through red and green glasses at a soap film so placed as to reflect white light to the eye.

466. Color of bodies in white light. Let a piece of red glass be held in the path of the colored beam of light in the experiment of the preceding section. All the spectrum except the red will disappear, thus showing that all the wave lengths except red have been absorbed by the glass. Let a green glass be inserted in the same way. The green portion of the spectrum will remain strong, while the other portions will be greatly enfeebled. Hence green glass must have a much less absorbing effect upon wave lengths which correspond to green than upon those which correspond to red and blue. Let the green and red glasses be held one behind the other in the path of the beam. The spectrum will almost completely vanish, for the red glass has absorbed all except the red rays, and the green glass has absorbed these.

We conclude, therefore, that the color which a body has in ordinary daylight is determined by the wave lengths which the body has not the power of absorbing. Thus, if a body appears white in daylight, it is because it diffuses or reflects all wave lengths equally to the eye, and does not absorb one set more than another. For this reason the light which comes from it to the eye is of the same composition as daylight or sunlight. If, however, a body appears red in daylight, it is because it absorbs the red rays of the white light which falls upon it less than it absorbs the others, so that the light which is diffusely reflected contains a larger proportion of red wave lengths than is contained in ordinary light. Similarly, a body appears yellow, green, or blue when it absorbs less of one of these colors than of the rest of the colors contained in white. light, and therefore sends a preponderance of some particular wave length to the eye.

467. Color of bodies placed in colored lights. Let a body which appears white in sunlight be placed in the red end of the spectrum. It will appear to be red. In the blue end of the spectrum it will appear to be blue, etc. This confirms the conclusion of the last paragraph, that a white body has the power of diffusely reflecting all the colors of the spectrum equally.

Next let a skein of red yarn be held in the blue end of the spectrum. It will appear nearly black. In the red end of the spectrum

it will appear strongly red. Similarly, a skein of blue yarn will appear nearly black in all the colors of the spectrum except blue, where it will have its proper color.

These effects are evidently due to the fact that the red yarn, for example, has the power of diffusely reflecting red wave lengths copiously, but of absorbing, to a large extent, the others. Hence, when held in the blue end of the spectrum, it sends but little color to the eye, since no red light is falling upon it.

Soak a handful of asbestos or cotton batting in a saturated salt solution; squeeze out most of the brine; pour over the material a quantity of strong alcohol. When ignited, this will produce a large flame of almost pure-yellow light. In a darkened room allow the yellow light to fall strongly upon a spectrum chart of six colors. The only color on the chart that appears natural is the yellow.

468. Compound colors. It must not be inferred from the preceding paragraphs that every color except white has one definite wave length, for the same effect may be produced on the eye by a mixture of several different wave lengths as is produced by a single wave length. This statement may be proved by the use of an apparatus known as Newton's color disk (Fig. 441). The arrangement makes it possible to rotate differently colored sectors so rapidly before the eye that the effect is precisely the same as though the colors came to the eye simultaneously. If one half of the disk is red and the other half green, the rotating disk will appear yellow, the color being very similar to the yellow of the spectrum. If green and violet are mixed in the same way, the result will be light blue. Although the colors produced in this way are not distinguishable by the eye

from spectral colors, it is obvious that their physical constitution is wholly different; for while a spectral color consists of waves of a single wave length, the colors produced by mixture are compounds of several wave lengths. For this reason the spectral colors are called pure and the others compound. In order to tell whether the color of an object is pure or compound, it is only necessary to observe it through a prism. If it is compound, the colors will be separated, giving an image of the object for each color. If it is pure, the object will appear through the prism exactly as it does without the prism.

By compounding colors in the way described above we can produce many which are not found in the spectrum. Thus, mixtures of red and blue give purple or crimson; mixtures of black with red, orange, or yellow give rise to the various shades of brown. Lavender may be formed by adding three parts of white to one of blue; lilac, by adding to fifteen parts of white four parts of red and one of blue; olive, by adding one part of black to two parts of green and one of red.

469. Complementary colors. Since white light is a combination of all the colors from red to violet inclusive, it might be expected that if one or several of these colors were subtracted from white light, the residue would be colored light.

P

L

R

S

To test this experimentally let a beam of sunlight be passed through a slit s, a prism P, and a lens L, to a screen S, arranged as in Fig. 442. A spectrum will be formed at RV, the position conjugate to the slit s, and a pure white spot will appear on the screen when it is at the position which is conjugate to the prism face ab. Let a card be slipped into the path of the

FIG. 442. Recombination of spectral colors into white light

beam at R, so as to cut off the red portion of the light. The spot on S will appear a brilliant shade of greenish blue. This is the compound color left after red is taken from the white light. This shade of blue is therefore called the complementary color of the red which has been subtracted. Two complementary colors are therefore defined as any two colors which produce white when added to each other.

Let the card be slipped in from the side of the blue rays at V. The spot will first take on a yellowish tint when the violet alone is cut out; and as the card is slipped farther in, the image will become a deep shade of red when violet, blue, and part of the green are cut out.

Next let a lead pencil be held vertically between R and V so as to cut off the middle part of the spectrum; that is, the yellow and green rays. The remaining red, blue, and violet will unite to form a brilliant purple. In each case the color on the screen is the complement of that which is cut out.

470. Retinal fatigue. Let the gaze be fixed intently for not less than twenty or thirty seconds on a point at the center of a block of any brilliant color - for example, red. Then look off at a dot on a white wall or a piece of white paper, and hold the gaze fixed there for a few seconds. The brilliantly colored block will appear on the white wall, but its color will be the complement of that first looked at.

The explanation of this phenomenon, due to so-called "retinal fatigue," is found in the fact that although the white surface is sending waves of all colors to the eye, the nerves which responded to the color first looked at have become fatigued, and hence fail to respond to this color when it comes from the white surface. Therefore the sensation produced is that due to white light minus this color; that is, to the complement of the original color. A study of the spectral colors by this method shows that the following colors are complementary.

Red
Bluish green

Orange
Yellow Violet
Greenish blue Blue

Greenish yellow

Green
Crimson

471. Color of pigments. When yellow light is added to the proper shade of blue, white light is produced, since these colors are complementary. But if a yellow pigment is added to a blue one, the color of the mixture will be green. This is