red in perfect contact. For the same reason, a rainbow formed in Venus and Mercury will be destitute of green rays, and have a brilliant bow of white light separating two coloured arches; while in Mars, Jupiter, Saturn, and the Georgian planet, the bow will exhibit only four homogeneous colours.

From his analysis of the solar spectrum, Newton concluded, "that to the same degree of refrangibility ever belonged the same colour, and to the same colour ever belonged the same degree of refrangibility;" and hence he inferred, that red, orange, yellow, green, blue, indigo, and violet were primary and simple colours. He admitted, indeed, that "the same colours in specie with these primary ones may be also produced by composition. For a mixture of yellow and blue makes green, and of red and yellow makes orange;" but such compound colours were easily distinguished from the simple colours of the spectrum by the circumstance, that they are always capable of being resolved by the action of the prism into the two colours which compose them.

This view of the composition of the spectrum might have long remained unchallenged, had we not been able to apply to it a new mode of analysis. Though we cannot separate the green rays of the spectrum into yellow and blue by the refraction of prisms, yet if we possessed any substance which had a specific attraction for blue rays, and which stopped them in their course, and allowed the yellow rays to pass, we should thus analyze the green as effectually as if they were separated by refraction. The substance which possesses this property is a purplish blue glass, similar to that of which fingerglasses are made. When we view through a piece of this glass, about the twentieth of an inch thick, a brilliant prismatic spectrum, we find that it has exercised a most extraordinary absorptive action on the different colours which compose it. The red part of the spectrum is divided into two red spaces

separated by an interval entirely devoid of light. Next to the inner red space comes a space of bright yellow, separated from the red by a visible interval. After the yellow comes the green, with an obscure space between them, then follows the blue and the violet, the last of which has suffered little or no diminution. Now it is very obvious, that in this experiment, the blue glass has actually absorbed the red rays, which, when mixed with the yellow on one side, constituted orange, and the blue rays, which, when mixed with the yellow on the other side, constituted green, so that the insulation of the yellow rays thus effected, and the disappearance of the orange, and of the greater part of the green light, proves beyond a doubt that the orange and green colours in the spectrum are compound colours, the former consisting of red and yellow rays, and the latter of yellow and blue rays of the very same refrangibility. If we compare the two red spaces of the spectrum seen through the blue glass with the red space seen without the blue glass, it will be obvious that the red has experienced such an alteration in its tint by the action of the blue glass, as would be effected by the absorption of a small portion of yellow rays; and hence we conclude, that the red of the spectrum contains a slight tinge of yellow, and that the yellow space extends over more than one-half of the spectrum, including the red, orange, yellow, green, and blue spaces.

I have found also that red light exists in the yellow space, and it is certain that in the violet space red light exists in a state of combination with the blue rays. From these and other facts which it would be out of place here to explain, I conclude that the prismatic spectrum consists of three different spectra, viz. red, yellow, and blue, all having the same length, and all overlapping each other. Hence red, yellow, and blue rays of the very same refrangibility coexist at every point of the spec

trum; but the colour at any one point will be that of the predominant ray, and will depend upon the relative distance of the point from the maximum ordinate of the curve which represents the intensity of the light of each of the three spectra.

This structure of the spectrum, which harmonizes with the old hypothesis of three simple colours, will be understood from the annexed diagram, where MN is the spectrum of seven colours, all compounded of the three simple ones, red, yellow,

Y

Fig. 7.

B

N

and blue. The ordinates of the curves R, Y, and B will express the intensities of each colour at different points of the spectrum. At the red extremity M of the spectrum, the pure red is scarcely altered by the very slight intermixture of yellow and blue. Farther on in the red space, the yellow begins to make the red incline to scarlet. It then exists in sufficient quantity to form orange, and, as the red declines, the yellow predominates over the feeble portion of red and blue which are mixed with it. As the yellow decreases in intensity, the increasing blue forms with it a good green, and the blue rising to its maximum speedily overpowers the small portion of yellow and red. When the blue becomes very faint, the red exhibits its influence in converting it into violet, and the yellow ceases

to exercise a marked influence on the tint. The influence of the red over the blue space is scarcely perceptible, on account of the great intensity of the blue light; but we may easily conceive it to reappear and form the violet light, not only from the rapid decline of the blue light, but from the greater influence of the red rays upon the retina.

These views may, perhaps, be more clearly understood by supposing that a certain portion of white light is actually formed at every point of the spectrum by the union of the requisite number of the three coloured rays that exist at any point. The white light thus formed will add to the brilliancy without affecting the tint of the predominant colour.

In the violet space we may conceive the small portion of yellow which exists there to form white light with a part of the blue and a part of the red, so that the resulting tint will be violet, composed of the blue and the small remaining portion of red, mixed with the white light. This white light will possess the remarkable property of not being susceptible of decomposition by the analysis of the prism, as it is composed of red, yellow, and blue rays of the very same refrangibility. The insulation of this white light by the absorption of the predominant colours I have effected in the green, yellow, and red spaces, and by the use of new absorbing media we may yet hope to exhibit it in some of the other colours, particularly in the brightest part of the blue space, where an obvious approximation to it takes place.

Among the most important modern discoveries respecting the spectrum we must enumerate that of fixed dark and coloured lines, which we owe to the sagacity of Dr. Wollaston and M. Fraunhofer. Two or three of these lines were discovered by Dr. Wollaston, but nearly 600 have been detected by means of the fine prisms and the magnificent apparatus of the Bavarian optician. These lines are G

parallel to one another, and perpendicular to the length of the spectrum. The largest occupy a space from 5" to 10" in breadth. Sometimes they occur in well-defined lines, and at other times in groups; and in all spectra formed from solar light, they preserve the same order and intensity, and the same relative position to the coloured spaces, whatever be the nature of the prism by which they are produced. Hence these lines are fixed points, by which the relative dispersive powers of different media may be ascertained with a degree of accuracy hitherto unknown in this branch of science. In the light of the fixed stars, and in that of artificial flames, a different system of lines is produced, and this system remains unaltered, whatever be the nature of the prism by which the spectrum is formed.

The most important fixed lines in the spectrum formed by light emitted from the sun, whether it is reflected from the sky, the clouds, or the moon, may be easily seen by looking at a narrow slit in the window-shutter of a dark room, through a hollow prism formed of plates of parallel glass, and filled with any fluid of a considerable dispersive power. The slit should not greatly exceed the twentieth of an inch, and the eye should look through the thinnest edge of the prism where there is the least thickness of fluid. These lines I have found to be the boundaries of spaces within which the rays have particular affinities for particular bodies.