prominences, we find that in general, their spectra consist simply of bright lines. Such then is the spectrum of part at least of the gaseous matter which surrounds the sun, and it is the upper portion of the absorbing medium which cuts out these black lines from what would otherwise be a continuous spectrum, and you easily trace what lines it does cut out. For instance, here is a dark line (C) in the red [see diagram, p. 192, which shows, as through the same slit, the spectra of

the sun and of a prominence], which is due to hydrogen gas. Well, we find these red flames owe their redness to the particular colour of this line of hydrogen. So this bright red line is one of the main features of the prominences. Then we find a yellow line very nearly coincident, as you see, with the lines of sodium. Nobody as yet knows what is the chemical substance which produces this particular line. It corresponds to no absorption line usually found in the sun's spectrum

(though you observe a trace of it in the spot spectrum which I last showed you), and therefore it must be due to a substance in a peculiar condition capable of radiating, but of having its absorption made up for,—some substance which possibly we may not yet know. Possibly it may not be a terrestrial substance at all. But it occurs here, very nearly giving a coincidence with sodium; but its light is not only more refrangible, but it wants the distinctive property which sodium has of giving a double line. Then we find several other lines, including two-or I may say three-more, due to hydrogen; so that the spectrum of these flames consists mainly of the spectrum of incandescent hydrogen gas. Here is another drawing of a small portion of the spectra of the sun and a prominence, which shows the exact coincidence of the bright and dark lines.

Suppose now we had a telescope to which the spectroscope could be adjusted on looking at a red prominence without the spectroscope we should see one image, but it would be an image which consisted partly of the red, partly of the green, partly of the blue, partly of the violet rays of hydrogen; but if we combine telescope and spectroscope, the combination would enable us to separate from each other, along the line of dispersion, the various colours; and the edge of the sun would be treated in the same way. All its colours would be spread out from one another, but they would be spread out at a disadvantage compared with the colour of a monochromatic line. Because however far you separate one such line from another, you do not weaken either. They remain, except in so far as reflection from the surfaces of the prisms, and absorption within the prisms, weaken them, as strong

as ever. But if you take a corresponding portion of sunlight, then, since it gives practically a continuous spectrum, you spread it uniformly over as long a space as you choose. So by the aid of this property, as the solar spectrum is practically continuous, except where there are interceptions of light, you can spread it out, and thus weaken it throughout as much as you please; whereas the other spectrum consists of perfectly definite bright lines, which you may spread as far apart from one another as you please, but which you cannot individually weaken. Hence, however strong be the glare of sunlight, sufficient dispersive power will enable us in fine weather to examine the spectrum of the red. flames.

This is perfectly analogous to the observing stars by daylight, which, you are aware, is done in every fixed observatory by means of a good telescope. It is simply because the diffused light of the sky allows itself to be weakened farther and farther as we spread it over a larger and larger image, while the light of the star always comes from the same definite point; because no one has yet made a telescope showing a star's disc (except as a delusive appearance due to diffraction), so that, magnify it as you please, its light comes from the same definite point. So it remains of the same brightness, while the background may be made as dark as you please by spreading it out. In that way, by combining the spectroscope with the telescope, and widening, or altogether dispensing with the slit, it is possible. to study the phenomena of these red flames, and, in fact, the whole behaviour of gaseous matters round the edge of the sun's disc, without waiting for a total eclipse. This is an extremely beautiful adaptation of

means first made theoretically by Lockyer, and afterwards by Janssen, but brought into practice nearly simultaneously by the two astronomers.

Here is the result as applied to a particular portion of the sun's circumference. The body of the sun we

will suppose to be under that picture. These are simply eruptions of glowing gas from the sun's apparent surface. On the same scale there would be another image, a green image, situated almost at the end of the room; then a long way beyond, an indigo, and finally a violet

one. But we have by means of the prisms separated that particular image from the others, and thus we have here a monochromatic representation of what is above the surface of the sun, in so far at least as incandescent hydrogen gas is involved. When I point out

that the change from the first figure to the second took place in the course of a few minutes, you will see what exceedingly rapid changes are going on in these selfluminous clouds; and when I further tell you that the height of this prominence, which is a stream of hydrogen rushing violently up from a rent in the surface of the sun, is something like 70,000 miles, you will see on what a stupendous scale, and with what tremendous velocities, these phenomena are constantly taking place.

So far then for the sun. When we compare the spectra of different stars with that of the sun, we come to some very curious conclusions. We find four classes of spectra, as a rule, among the different fixed stars. which have seemed of importance enough to be separately examined. The first class of spectra are those of white or blue stars. You see an admirable example in Vega, and another in Sirius, or the dog-star. All these white stars have this characteristic, that they have an almost continuous spectrum with few and broad dark lines crossing it, and these few for the most part lines of hydrogen. These stars are in all probability at a considerably higher temperature than the sun; and their atmospheres are in even more violent agitation than is that of the sun. Then you come to the class of yellow stars, of which our sun is an example. In their spectra you have many more dark lines than in those of the white stars, but you have nothing of the nature of nebulous bands crossing the spectrum, such as you find in the third class; still less have you certain curious zones of shaded lines which you have in the fourth class of stars. This classification seems to point out the period of life, or phase of life, of each particular star or sun. When it is first formed, by the impact of