Amongst the pioneers of knowledge in this direction were Sir John Herschel[373]--who, however, applied himself to the subject in the interests of optics, not of chemistry--W. A. Miller,[374] and Wheatstone. The last especially made a notable advance when, in the course of his studies on the "prismatic decomposition" of the electric light, he reached the significant conclusion that the rays visible in its spectrum were different for each kind of metal employed as "electrodes."[375] Thus indications of a wider principle were to be found in several quarters, but no positive certainty on any single point was obtained, until, in 1859, Gustav Kirchhoff, professor of physics in the University of Heidelberg, and his colleague, the eminent chemist Robert Bunsen, took the matter in hand. By them the general question as to the necessary and invariable connection of certain rays in the spectrum with certain kinds of matter, was first resolutely confronted, and first definitely answered. It was answered affirmatively--else there could have been no science of spectrum analysis--as the result of experiments more numerous, more stringent, and more precise than had previously been undertaken.[376] And the assurance of their conclusion was rendered doubly sure by the discovery, through the peculiarities of their light alone, of two new metals, named from the blue and red rays by which they were respectively distinguished, "cæsium," and "rubidium."[377] Both were immediately afterwards actually obtained in small quantities by evaporation of the Durckheim mineral waters.
The link connecting this important result with astronomy may now be indicated. In the year 1802 it occurred to William Hyde Wollaston to substitute for the round hole used by Newton and his successors for the admittance of light to be examined with the prism, an elongated "crevice" 1/20th of an inch in width. He thereupon perceived that the spectrum, thus formed of light, as it were, _purified_ by the abolition of overlapping images, was traversed by seven dark lines. These he took to be natural boundaries of the various colours,[378] and satisfied with this quasi-explanation, allowed the subject to drop. It was independently taken up after twelve years by a man of higher genius. In the course of experiments on light, directed towards the perfecting of his achromatic lenses, Fraunhofer, by means of a slit and a telescope, made the surprising discovery that the solar spectrum is crossed, not by seven, but by thousands of obscure transverse streaks.[379] Of these he counted some 600, and carefully mapped 324, while a few of the most conspicuous he set up (if we may be permitted the expression) as landmarks, measuring their distances apart with a theodolite, and affixing to them the letters of the alphabet, by which they are still universally known. Nor did he stop here. The same system of examination applied to the rest of the heavenly bodies showed the mild effulgence of the moon and planets to be deficient in precisely the same rays as sunlight; while in the stars it disclosed the differences in likeness which are always an earnest of increased knowledge. The spectra of Sirius and Castor, instead of being delicately ruled crosswise throughout, like that of the sun, were seen to be interrupted by three massive bars of darkness--two in the blue and one in the green;[380] the light of Pollux, on the other hand, seemed precisely similar to sunlight attenuated by distance or reflection, and that of Capella, Betelgeux, and Procyon to share some of its peculiarities. One solar line especially--that marked in his map with the letter D--proved common to all the four last-mentioned stars; and it was remarkable that it exactly coincided in position with the conspicuous yellow beam (afterwards, as we have said, identified with the light of glowing sodium) which he had already found to accompany most kinds of combustion. Moreover, both the _dark_ solar and the _bright_ terrestrial "D lines" were displayed by the refined Munich appliances as double.