PHYSICS NINETEENTH CENT. SPECTROSCOPY. 485 



limelight spectrum. Further, it is obvious that, in the light from the 

 incandescent lime, those rays that would have otherwise have made the 

 D spaces as luminous as the rest of the spectrum, were stopped or ab- 

 sorbed by the sodium-coloured flame. In other words, from the lime- 

 light there passed through the flame rays of every degree of refrangi- 

 bility, except only rays corresponding with the two D lines, to which 

 rays the sodium vapour was completely opaque. Kirchhoff likewise 

 ascertained that when the flame was coloured with potassium com- 

 pounds instead of sodium compounds, there appeared, in the spectrum 

 of the limelight behind the flame, dark lines corresponding exactly 

 with the bright lines which were seen when the flame alone was 

 viewed by the spectroscope. 



These and other facts of the like kind were generalized by Kirchhoff 

 into a conclusion which may be thus stated : When any substance ren- 

 dered luminous by heat emits rays of a certain definite refrangibility , the 

 substance has the power of absorbing at the same temperature rays of that 

 identical refrangibility. The like theoretical law applying to rays of 

 heat had before this been proposed by Prevost, and by Provostaye 

 and Desains, and this law was fully elucidated by Balfour Stewart 

 'very shortly before Kirchhoff announced that it holds good for light 

 also. The importance of Kirchhoff's generalization will appear 

 more clearly when we have shown the remarkable explanation of the 

 Fraunhofer lines which he deduced from it. As the researches con- 

 nected with spectroscopy became, from about the period we have now 

 reached, spread over several ever-widening fields of inquiry, it will 

 be more convenient to indicate the chief landmarks of discovery in 

 each province of spectroscopic investigation, rather than to follow the 

 chronological order and be passing continually from one branch of 

 the subject to another. We have brought the reader to the period at 

 which Kirchhoff clearly announced the relations that exist between 

 the dark and the bright lines of spectra. But for the present we shall 

 consider the progress of spectroscopy as a means of chemical analysis. 

 The application of the spectroscope was suggested, as we have already 

 seen, by Fox-Talbot and by Sir John Herschel, and some spectra of 

 coloured flames were described by Professor W. A. Miller in 1845. 



In 1857 Professor Swan, published an elaborate research "On 

 the Spectra of the Flames of the Hydro-carbons." The first recorded 

 observation of line spectra includes, as we have seen, a reference 

 to a hydro-carbon spectrum (page 479). Swan had the advantage 

 of using in his experiments the now well-known Bunsen gas-lamp, 

 which provides the best means for the examination of flame spectra 

 generally. He observed the constant occurrence of the yellow (D) 

 line in the spectra, and he was the first who proved that this line 

 is due solely to sodium compounds. He proved the extreme deli- 

 cacy of the spectroscopic indication of the presence of sodium, by 

 showing that quantities of sodium compounds, wholly imperceptible 



