violet rays, and all the others will be less and less refracted as their wave-lengths grow longer and longer till we reach the lowest red in the spectrum; and therefore instead of having a set of coloured discs, as by the first method, succeeding one another with their centres along a line, and overlapping, you will have a set of parallel coloured images, each no broader than the slit itself, and you can make the slit as narrow as you please. In every part the consecutive images lie side by side, contiguous to one another; but if there be light of any wave-length or any particular refrangibility which is wanting, then the space corresponding to that will be left as a dark line (an unilluminated image of the slit) across the otherwise continuous coloured band. You see hanging on the wall a coloured plate representing the solar spectrum, formed in the way I have just pointed out, and you can see those dark lines across it. Only a few of the chief ones are figured. The number of those whose position is already carefully measured, or photographically registered, amounts to many thousands. [See diagram, p. 190.] They were first noticed by Dr. Wollaston, about the beginning of this century, but he paid very little attention to them; and they were re-discovered a considerable time afterwards by the great optician Fraunhofer, whence they have been called Fraunhofer's lines. Of those lines one of the most remarkable is that to which Fraunhofer gave the name of D, which you will see upon the picture near the boundary between orange and yellow. When, however, a very perfect prism is used, and a telescope is employed instead of the screen. to receive the spectrum, then we are enabled to see that this line is double. This gives it a very remarkable characteristic,-two almost Now, Fraunhofer observed No further action seems to have been taken with regard to this very remarkable coincidence, until Professor Miller of Cambridge, in 1849 or 1850, made a more exact experiment, with the view of comparing these yellow lines in the flame of a spirit-lamp with the dark lines of the solar spectrum, so as to test whether they are exactly coincident with one another or not. The result of his measurements was that the closeness of coincidence was so great that it was impossible, with his finest instruments, to find any divergence between them. The two bright lines exactly corresponded with the two dark ones as to refrangibility, and therefore also wave-length. It had been conclusively shown by Swan, that the two bright lines in the light of a candle are due to common salt, which pervades the air everywhere, and of which the very minutest trace is capable of producing this yellow light. was then that Stokes at once took the additional step required, and explained that the glowing vapour, which is capable, when it is the source of light, of giving these definite bright lines, is itself, when used as an absorbing medium, capable of absorbing these and these only; and therefore that Miller's test of the exact coincidence of the bright and dark lines was a complete proof that there exists in the sun's atmosphere this vapour of sodium. It That occurred about 1850, and ever since that time the fact that sodium exists in an incandescent state in the sun's atmosphere, has been taught (as an experimentally ascertained truth) by Sir William Thomson and others. This was the birth of Spectrum Analysis, as applied to celestial objects. It is curious to find that the deservedly celebrated Foucault had some years before made the same experiment in an even more convincing form than that which Miller had adopted. He found that the light, of what is called the electric arc, has in its spectrum these two bright lines; but that when he looked at sunlight through the electric arc and allowed the sunlight to come in so strongly as to overpower the electric light, then the electric light actually cut out the D lines from the solar spectrum more powerfully than if it had not been present. Although it was there giving out these lines strongly, it was not competent to fill up the wants in the solar spectrum, but actually made the deficiency more glaring than before. Then, to test whether it was really the case that this electric arc was absorbing these particular kinds of light, Foucault very ingeniously took advantage of the fact that the carbon points between which the electric arc is usually formed become incandescent, and reaching a higher temperature, are very much more brilliant than the arc itself. By means of a small mirror he reflected the white light from these carbon points through the electric arc, and found that whenever it passed through the arc, instead of getting brighter at those places, it had those very lines cut out of it; but whenever it passed beside the electric arc it had no deficiencies. Curiously enough, he seems to have derived no definitive conclusion from this. Then, again, exactly the same statement was made in Sweden by Ångström in 1853. He says that an incandescent gas gives out luminous rays of the same refrangibility as those which it absorbs. Each one of these three thus completely made and recorded the discovery of the physical basis of spectrum analysis before 1854; but, of the three, Stokes alone made the application which really constitutes celestial chemistry. Fox Talbot had, long before, distinctly pointed out the use of the prismatic method for distinguishing terrestrial substances in a flame. As I have already told you, Sir W. Thomson has, certainly ever since 1852 (probably a year or two sooner), regularly given in his public lectures in Glasgow University the statement that there is sodium vapour in the sun's atmosphere; and that, to find other constituents of solar and stellar atmospheres, all that is wanted is a comparison of the dark lines in their spectra with the bright lines in the incandescent vapours of various terrestrial substances.1 But it was not till 1859 or 1860 that this was known generally, or was applied to any purpose further than to the mere recording of the existence of sodium in the vapours around the sun. I should like to read a quotation from some remarks I made a year or two ago to the Royal Society of Edinburgh upon this curious subject:2 It is difficult now-a-days, when so many philosophers are engaged almost simultaneously at the same problem, to decide which of their successive steps in advance is that to which should really be attached the title of discovery (in its highest sense) as distinguished from mere improvement or generalisation. You have only to look at the recent voluminous discussions as to the discoverer of the Conservation of Energy, to see that critics may substantially agree as to facts and dates, while differing in the most extraor : 1 President's Address, Brit. Ass. 1871. See Stokes, Nature, January 6, 1876. Thomson writes to me with reference to this (January 23, 1876) 'I never imagined that Stokes thought I was generalising too fast, or that I was generalising at all. I felt that I had learned the whole thing from him on a foundation of absolute certainty. . . . All I said in my Edinburgh Address on this matter is, I believe, irrefragable.' 2 Proceedings R.S.E., May 15, 1871. N |