On the Spectra of Chemical Compounds. 445 



those obtained by bringing the metals themselves into the 

 electric arc. His observations have been confirmed and ex- 

 tended by Diacon, by Lecoq de Boisbaudran, and by Lockyer. 

 Pliicker and Hittorf obtained different spectra from such 

 non-metals as nitrogen, sulphm-, selenium, and carbon, by 

 varying the conditions under which the spectra were pro- 

 duced. Nitrogen, for instance, showed two band-spectra, and 

 one line-spectrum ; compounds of carbon with oxygen, nitro- 

 gen, and hydrogen yielded five different spectra, which were 

 all regarded as spectra of the element carbon, the differences 



being attributable to the influence of temperature. Angstrom 

 and Thalen regarded the different spectra obtained by Plucker 

 and Hittorf as caused by the different carbon compounds exa- 

 mined by them : the line-spectrum of nitrogen they looked on 

 as the spectrum of the element itself ; one of the band-spectra 

 they regarded as that of nitric oxide ; the second band-spectrum 

 they could not explain*. 



Kirchhoff obtained different spectra from calcium chloride 

 by varying the conditions under which the salt was volatilized. 

 The spectrum produced by using an electric current of very 

 weak intensity was the same as that noticed when the salt was 

 brought into the gas-flame. By employing a current of 

 greater intensity, or a flame of gas and air, a different spectrum 

 was produced. Mitscherlich^s experiments, in which he ob- 

 tained the spectrum of calcium oxide by bringing a mixture 

 of calcium and ammonium nitrates into a flame of gas and 

 air, and the spectrum of calcium chloride by volatilizing the 

 chloride in a flame of hydrogen and chlorine, seemed to point 

 to an explanation of the phenomena noticed by Kirchhoff. 

 The spectrum obtained when a current of low intensity was 

 employed was probably due to undecomposed calcium chloride, 

 mixed, especially towards the close of the experiment, with 

 calcium oxide ; when a stronger current was employed, the 

 chloride and oxide were probably dissociated, and lines due to 

 the presence of the elements began to appear : this dissociation 

 was rendered complete by using a current of great intensity ; 

 the lines which then appeared were entirely caused by the 

 elements in the free state. 



But the study of emission-spectra alone cannot enable us to 

 answer the question. Has each chemical compound a distinct 

 spectrum of its own ? Emission-spectra tell us that the ele- 



* Helmholtz supposes that the chemical atoms yield line-spectra, the 

 molecules (or complexes of atoms) band-spectra. In accordance with this 

 view, Pliicker's nitrogen-spectrum of the second order would be caused by 

 nitrogen atoms, and that of the first order, which was not explained by 



Angstrom and Thalen, by nitrogen molecules. 



