If we cut a narrow slit in the screen on which the spectrum falls, in a direction per pendicular to its length, the light which passes through has a definite refrangibility, and can no longer be drawn out by a prism into a spectrum. This experiment also is due to Newton.
If the slit in the shutter be very narrow, and the prism be adjusted to the most favor able position (so that the incident and refracted rays make equal angles with the sur faces on which they impinge, and from which they escape, respectively), we see that after all the solar spectrum is not continuous. It is found to he crossed at intervals by dark hands, showing the absence of rays of certain definite refrangibilities. The phe nomenon is found to be the same whatever be the substance of the Prism; so that these rays arc really wanting in sunlight.
This important discovery was made by Wollaston, hut the hands were first carefully observed and measured by Frannhofer. from whom they are commonly called leraun hofer's lines. We owe to Fraunhofer the invaluable snggestion of employing a telescope to examine the spectrum. The refracted rays are received directly on the object-glass of the telescope, which forms en image of the spectrum to be examined with the aid of the eye-piece, the screen being dispensed with. Wollaston had seen only five lines; Frann• haus at once discovered four hundred; Brewster, with more perfect apparatus, counted two thousand; and now, with a train of prisms. and powerful telescopes, their number seems beyond computation. They show every variety of breadth and distinctness, and are grouped in the most manner. For referenee, Freunhofer selected some of the more prominent, to which he attached the earlier letters of the alphabet. By their help he was enabled to measure refractory indices (see REFRACTION) with a precision completely unlooked for. If the light of a candle, a bright gas-flame, a white hot wire, or a lime-ball in the oxhydrogen flame, he examined in the same way, no such lines are seen. But sonic of them, and others not apparently belonging to sunlight, were found by Fraunliofer in the spectra of various fixed stars—while the light of the moon and planets seemed to give spectra similar to that of sunlight.
The first to throw any light on'this subject was Brewster. He showed that when light passed through nitrous acid gas it's spectrum was interrupted by countless lines; and that they increased in number and breadth by the application of heat to the gas, so that at a high temperature a thin layer of this gas is opaque to direct sunlight. Hence it was natural to conclude that the dark bands in the solar spectrum are caused by absorp tion in some medium lying between us and the sun. It is to be observed, however, that
this is ou the supposition that light m ht as it comes immediately from the sun would give, like that of the lime ball, a continuous spectrum. But Brewster went further. He showed that some of Fraunhofer's lines depend on the altitude of the sun, that is,, on the greater or less space of air, fog, and vapor through which his rays must pass before reaching the earth. Some of them, then, are caused by absorption in the earth's atmos phere.
But we must now look to another class of phenomena. A spirit lamp flame gives a very feeble spectrum; and, if a little common salt be put on the wick, although the flame becomes instantly very much brighter, no alteration is produced on the spectrum, save the appearance of a bright yellow line, crossing it at the place where the dark line called by Fraunhofer D appears in sunlight. On examining this line carefully Frani'. borer found that it, like D, is double—and he verified that these two rays were exactly (so far as refractive index goes) two of those wanting iu sunlight, and in the light of some of the stars.
About the same time Talbot and Herschel (q.v.) showed that the colors given by Lithia, Stroutia, etc., in a spirit flame were, like that produced by common salt, due to the production of light of several perfectly definite refrangibilities; so that the spectrum of the lamp-flame was crossed in each case by a series of bright lines always the same when the same body was placed in the flame; and they suggested (in 1825) the applica tion of this method to the qualitative analysis of minerals, etc., when the presence of extremely minute quantities of different bodies has to be ascertained. This was, in reality, the foundation of SPECTRUM ANALYSIS; and the method was, we may say, almost complete so far as practice is concerned. The theory, however. was left incom plete, so far as regards the cause of dark lines in the solar spectrum. Foucault (in 1849) seems to have been the first approach the true explanation. Au experiment of his, from which, however, he drew no inferences, contains the complete theory. When salt is placed in the voltaic arc (Electric Light, q.v.) the spectrum gives the double bright line (coinciding with the double dark line b) already referred to. When sunlight passes through this arc its dark line D is strengthened, instead of being filled up by the yellow light from the arc as we might have expected; and when one of the white-hot ca•bon points (which gives a continuous spectrum) is looked at through the yellow arc, the double dark line D appears in its spectrum.