in the heavens or on the earth, will give a spectrum. If it be a solid or liquid body, we shall have what is called a continuous spectrum—that is, a colored band bright from end to end, with no sign of any dark or bright lines about it. By continuing to heat this body until it becomes a mass of incandescent gas, the spectrum will become entirely changed, and will consist of a series of bright lines on a dark background, the number and position of the lines depending on the substance heated. But suppose, now, that the light from an incandescent solid or liquid body passes through a gas, what kind of a spectrum should we have? Experiment shows that in this case we get a continuous spectrum crossed by dark lines, these dark lines being produced by the peculiar power that a gas possesses of absorbing those particular rays of light which it emits. Thus we see that if we are dealing with incandescent solid or liquid bodies, we obtain continuous spectra; if with incandescent gases, bright-line spectra; and if with absorption, dark-line spectra; the position of the lines in all cases revealing the chemical nature of the substances.
So much, then, for the general idea of the nature of a spectrum. There are some additional points to be considered when we are dealing with stars. If we observe the spectrum of a star at rest, we shall obtain lines, whether bright or dark, in their normal place in the spectrum. These lines will be peculiar to certain substances, and, in fact, their presence in the star is determined simply by them. If we deal with the light from a body which is not an apparent point, the lines will still keep the same positions, for the same reason, but each one of them will be broadened equally.
Let us now suppose the star no longer stationary, but moving with a considerable velocity. In this case the wave-length of each line will be no longer the same; but the line will have altered its position in the spectrum to an extent depending on the movement of the star toward or from the earth. The result produced in the spectrum will be the same with regard to the number of lines as was the case when the star was assumed to be motionless, but the lines will all have received a slight shift, either to one side or the other of their initial positions, according as the star is approaching or receding. If instead of one we now deal with two stars of the same chemical and physical structure, traveling with different velocities, either toward or away from us, the spectrum would show each line doubled, and the more rapid the relative motion the coarser will be the doubling. If the stars were so physically constituted that the same chemical substances were present in both, but giving bright lines in one and dark lines in the other, the spectrum would present a series of bright lines, each accompanied by a dark one, on one side or