and in consequence we have the peculiar phenomenon of a constantly burning black sodium flame (Fig. 64).

I can also show you in a third way the fact that sodium vapour opaque to the light which it gives off.

is

I have prepared a tube containing some sodium which I can convert into vapour. By heating the tube as I am doing, it will become filled with sodium vapour, and you will see that it is perfectly colourless and trans

parent when we look at it with the white sunlight; but when we look at it with the yellow sodium light it will appear to be opaque. We shall then see that the tube containing the sodium vapour throws a dark shadow on the screen. [The lights were turned down, and the screen was illuminated with a yellow sodium flame.] Now the tube looks black; we cannot see through it; it throws a dark shadow. [Light was again admitted.] Now, by the daylight, it is colourless. This shows us, then, very distinctly, that the sodium vapour is opaque for the rays which itself can emit.

Thoroughly understanding, then, the nature of the phenomena with which we have to deal, let us follow Kirchhoff to the interesting conclusions which he draws from this experiment. He states that from this fact it appears likely that glowing gases have the power of especially absorbing rays of the same degree of refrangibility as those they emit; and that therefore the spectrum of such a glowing gas can be reversed, or the bright lines turned into dark ones, when light of sufficient degree of intensity, giving a continuous spectrum, is passed through it. This idea was further confirmed. by substituting for the sodium flame the flame coloured by potassium, when dark lines appeared in the exact position of the characteristic bright lines of this metal. Bunsen and Kirchhoff have likewise succeeded in reversing the flames of lithium, calcium, strontium, and barium ; whilst Dr. Miller and M. Cornu have also reversed some of the lines in the spectra of copper and other metals. I can here show you the reversal of the red lithium line on the screen. For this purpose I bring on to the carbon pole of the lamp some salt of lithium, together with a piece of metallic sodium. The sodium will reduce the

lithium salt to the metallic state, and I can then show you that we have got not only a dark sodium band, but a dark lithium band in the red part of the spectrum. Now the reversed lines of both these metals are clearly seen.

In speaking of this subject I may mention that Professor Cornu has recently made some singular observations respecting the reversal of the lines of certain metals. When endeavouring to obtain photographs of the three magnesium lines situated in the ultra-violet, he found that impressions of five instead of three lines were left on the plate. This he showed was due to the reversal of two of the lines which appeared dark on a light ground, whilst the third line could not be reversed. By using a powerful electric arc, the three bright wellknown magnesium lines (6) can be reversed, one coming out after the other. In a similar way Cornu has reversed certain of the lines of the following metals: sodium, thallium, lead, silver, aluminium, cadmium, zinc, and copper. Why some only of the lines of these metals can be reversed, or why none of the lines of such metals as antimony, gold, and bismuth can be reversed, are questions which we are at present unable to answer.

Generalizing from these facts, Kirchhoff has arrived, by the help of theoretical considerations which I am unable now to lay before you, at a law, previously partially enunciated by Prevost of Geneva and by Prevostaye and Dessains in France, and extended by Dr. Balfour Stewart in this country, which expresses the relation between the amount of heat of any wave-length which a body receives and that which it emits. This law has been called the law of exchanges. It asserts that the relation between the amount of heat emitted and that which is absorbed at

1 Phil. Mag., Sept. 1871.

any given temperature remains constant for all bodies; and that the greater the amount of heat emitted, the greater must be the amount of heat absorbed. Kirchhoff has proved that the same law holds good for light as well as for heat; that it is as true of the luminous as of the heat-giving rays; and for rays of different kinds, if we compare the same kind of rays: for instance, if we compare red rays emitted with red rays absorbed, or yellow rays emitted with yellow rays absorbed. From this we see that an incandescent gas which is giving off only certain kinds of light-that is, whose power of emission is limited to light of certain definite degrees of refrangibility—must have the power of absorbing those kinds of light, and those kinds only. This is what we find to be the case with the luminous sodium vapour : it has a very high power of emission for the "D" rays, and it has a proportionately high power of absorption for that kind of light; but for it alone. And we see that every substance which emits at a given temperature certain kinds of light must possess the power, at that same temperature, of absorbing the same kinds of light.'

We must remember, however, that the emissive and absorptive powers of substances can only be compared at the same or nearly the same temperature. This is of very great importance, for it has been supposed that in some cases the law of exchanges does not hold good, the comparison between absorption and radiation not having been made at the same temperature. It must not be assumed that because the bright lines of the incandescent iodine spectrum, for instance, do not corre

1 Report on the Theory of Exchanges, by B. Stewart (Brit. Assoc. 1861); Kirchhoff on the History of the Analysis of the Solar Atmosphere (Phil. Mag., Fourth Series, vol. xxv. p. 256).

spond to the dark absorption bands of the gas at a much lower temperature, therefore the law is faulty or incorrect. We must compare the lines at the same temperature.

Now we know that the same kind of law holds good with the other vibrations known to us-the vibrations of the air which we call sound. We are all acquainted with what is called resonance. When we sing a particular

note in the neighbourhood of a piano, that same note is returned to us. The particular vibrating string which can emit that note has the power also of absorbing vibrations of that particular kind, when proceeding in a straight line, and emitting them again in all directions. We are not, therefore, without analogy, in the case of sound, for the absorption and emission of the same kind of undulation by the same substance.

We will now pass to the application of this principle of the reversibility of the spectra of luminous gases to the foundation of a solar and stellar chemistry. How does this principle assist us in our knowledge of the constitution of the solar atmosphere?

In order to map and determine the positions of the bright lines found in the electric spectra of the various metals, Kirchhoff, as I have already stated, employed the dark lines in the solar spectrum as his guides. Judge of his astonishment, when he observed that dark solar lines occur in positions coincident with those of all the bright iron lines! Exactly as the sodium lines were identical with Fraunhofer's lines D, so for each of the iron lines, of which Kirchhoff and Ångström have mapped no less than 460, a dark solar line was seen to correspond. Not only had each iron line its dark representative in the solar spectrum, but the breadth and degree of shade of the two sets of lines were seen to agree in the most perfect