We learn but little, however, from these experiments, of the nature of the substances of which the sun is composed; for the rays which we have produced are all brilliant, while those of the solar appearance are black. Let us see, then, in pursuing this subject, if it would not be possible for us to obtain these black lines with our artificial lights. Let us produce, in analyzing the Drummond light, a perfectly continuous appearance. Now, let us make this appearance, before reaching the screen, pass through a deep layer of hypoazotic acid. Immediately you see it discontinued. It is like the solar appearance, crossed over by a multitude of black lines. The hypoazotic acid is not the only gas that produces this result. The vapor from brome, that of iodine, will give equally the black lines in the same circumstances, only these lines are different from those we have just seen in the experiment made with the hypoazotic acid. Thus, the gases, the vapors, possessing the property of absorbing certain luminous rays, certain colors, these rays, found no longer in the appearance, are necessarily replaced by the black lines we have just observed. All the gases, all the vapors, could not, I am convinced, produce this result; for it is clear that their power of absorption, being less considerable, could not make itself felt, unless by means of a stratum the thickness of which should be greater than that which we are able to use in our experiments. We find a proof of this in what passes in the atmospheric air. Under a feeble thickness no sensible absorption is produced; but it is certain that the atmospheric mass absorbs a great number of rays, and consequently gives birth to many black lines; for in the solar appearance we observe new and very marked lines, when the sun being near the horizon, his rays pass through a bed of air of very considerable thickness. These rays are principally owing to the vapor of water. We can equally affirm the absorbent power of the atmosphere which surrounds the planets Saturn, Jupiter, and Mars. Their appearances contain lines very different from the solar appearance. Yet, as the light which they transmit to us comes to them from the sun, we are forced to conclude that that light undergoes some modification in travelling over its transparent path. It is the atmosphere of the planets which produces this result.

The sun also possesses an atmosphere, as we have seen, and this atmosphere ought necessarily to exercise an influence on the rays which traverse it. Such is, in fact, the origin of the rays which we notice in the solar appearance. They are owing to the atmospheric absorption, and the bed of transparent but absorbent vapor which surrounds the atmosphere, and which the rays pass through before they spread themselves in space.

But how are we to ascertain the nature of the vapors which produce the black lines we observe? Here physical science comes again to our aid, and the question we have just put finds its answer in a recent discovery. We have seen that a certain substance in burning gives birth to certain luminous rays which characterize it. We have also seen that this same substance, in a state of vapor, absorbs, on the contrary, certain rays, and produces in consequence certain black lines which are equally characteristic. Now, by a singular coincidence, these two powers, emissive and absorbent, are identically the same. Each substance, in a state of vapor, absorbs precisely the rays which it is capable of producing in combustion, so that the black streaks produced in the first case occupy identically the same place as the brilliant lines observed in the second. We may demonstrate this interesting theory by the following experiment, due to M. Toucault. We know that sodium produces in burning a beautiful yellow light. Well, let us burn some sodium in the coals, and between these two substances the electric light is produced. The metal while it is burning volatilizes largely; the vapors which are produced absorb precisely the rays which they should have emitted in their combustion; and you see that in the yellow, instead of a brilliant line, we have a very dark line. What we have just seen take place with the sodium has been equally proved by experiments on a great number of metals, and, by induction, we may extend the application to all those on which it has been impossible to make experiments.

Let us apply this principle to what concerns the light of the sun. The photosphere is composed of condensed substances, precipitated in a solid or a liquid state, floating in a transparent and absorbent atmosphere. This matter, being simply incandescent, ought to present to us a continuous appearance, and this continuity can be disturbed only by the absorption of the solar atmosphere. From this it follows, that to ascertain the chemical nature of the substances which compose this atmosphere, it will be sufficient to compare the black lines of the sun with the bright lines of our artificial lights. This has been done. M. Kirchoff first discovered that the sun contains sodium; for the line D of Fraunhofer coincides perfectly with the brilliant lines of this metal. It is equally well known that iron, copper, and twenty other substances which exist upon the earth in a solid state, would, at a temperature of five millions of degrees, be necessarily in a state of vapor.

After having thus made a chemical analysis of the sun, astronomers wish to go further; they have sought to know equally the composition of the stars. We have been led by this to some very remarkable consequences; we have been able to make a kind of classification of these stars, and to determine the group to which our sun belongs. It remains, then, for us now to apply the spectral analysis to the myriads of stars which stud the heavens, to those far distant suns, the greater part of which, perhaps, surpass in grandeur and brightness that which is the centre of our planetary system. It remains for us to interrogate these scarcely perceptible bodies, sparkling at such an incalculable distance, and to demand and draw from them the secret of their chemical composition. This enterprise is daring, but it is not rash. The difficulties are alarming; yet learned men are not discouraged, for they are accustomed to see difficulties disappear before strenuous and persevering labor.

We commenced our study of the stars with the complicated instruments which we employ for the sun; but we soon found out that this complication was useless. We have been able to reduce our instruments to the number of two, a cylindrical glass and a prism. And M. Wolff, of the Paris Observatory, has succeeded recently in suppressing the cylinder, keeping only the essential element, that is, the prism intended to produce the appearance.

We have examined a great number of stars, and I am going to submit to you some of the results at which we have arrived. You see at this moment the appearance which the star Orion presents. This star is of a yellow color; the appearance which it produces is deeply streaked; and it is one of the most beautiful in the heavens. You will find there the line D of sodium, and the line b of magnesium. These are two fundamental lines which have served as marked points to compare this appearance with that of the sun. Besides sodium and magnesium, a of Orion contains iron, copper, and several other known metals; but it is singular that hydrogen is not found there in the free state, as in the sun. There is, then, some essential difference between the stars, of which you will be more convinced as we go further into the subject. Here is the appearance of Sirius. You see it is not nearly so fine. You will find two large bands in blue, in the place of the streak F of the sun; two others in violet; and one, very faint, in yellow. The two first are attributable to hydrogen, and the last to sodium; but we know not to what substance the violet is owing. In the green there are also some very fine lines, but very difficult to seize.

What is most remarkable is, that all the white stars present the same appearances, and half the stars that are visible belong to this type. Thus the fine stars of the Lyre, of the Eagle, of the Bear, Castor, etc., ought to be ranged by the side of Sirius. There is, however, an exception in [zeta] of the Bear, which is a yellow star. The magnificent stars of Arcturus, of the Goat, of Procyon, belong, on the contrary, to the class of which our sun is a type, except that the iron line E is much more marked. Their color, of light yellow, led to the inference that they were analogous to the sun, and the supposition has been confirmed by spectral analysis. All know substances have an appearance which is peculiar to them, and which characterizes them. Can we say as much of the stars? Do they also present marked differences in their appearance? This has been the subject of very interesting researches. The task has been undertaken at the observatory of the Roman College, and it has led to a result altogether unforeseen, namely, that the stellar appearances appertain to only a very limited number of types. We may classify them in three groups. The first group is that of the white stars like Sirius; the second, that of the yellow stars, of which Arcturus and the second are members; and Orion may be regarded as a type of the third, in which we ought to place a of Hercules, and [Beta] of Pegasus. These two last-named stars have very remarkable appearances. They seem formed of a multitude of channels, which are divided by large black bands. This form of appearance shows us that the stars which belong to this type are surrounded with atmospheres heavily charged with vapor. In this group enters all the red stars, and in particular Omicron of the Whale, that celebrated star which has been called The Wonderful. Several small stars of a blood-red color have appearances resembling each other. It is remarkable that in all the appearances belonging to stars of this type, the black lines occupy the same place, which proves that in general they are all made alike.