172 Prof. Mitscherlich on the Spectra of Compounds 



therefore the temperature of the discharge is so lowered by the 

 liquid electrodes that the gases are not ignited sufficiently to 

 produce luminous phenomena. The latter mode of experiment 

 is important, because spectra of the metals and of their com- 

 pounds can thus be produced free from any admixture with 

 spectra of the gases. 



The spectra produced by the methods described are mostly 

 depicted in Plates I. and II.; the luminous parts are expressed 

 in as accurate shading as is possible by printing-ink. The 

 lines produced by method 8 (liquid electrodes) are designated 

 by e in the Plates. As regards the investigations themselves I 

 remark as follows. 



The spectra of the compounds of copper and of bismuth can 

 be very conveniently obtained pure. Chloride, bromide, and 

 iodide of copper, and salts of oxide of copper, such as the acetate 

 and nitrate, were investigated by method 6 (volatilization in the 

 glass tube). The spectra of chloride^ bromide, or iodide of cop- 

 per are obtained perfectly pure and very distinct by methods 3 

 (CI and H) and 4 (Br, or I, and H), using even any salts of cop- 

 per. The spectrum of fluoride of copper is best obtained by 

 method 2 (H, or coal-gas, and 0), using fluoride of copper and 

 fluoride of ammonium. If in the latter method but little oxygen 

 is admitted to the burning gases, a new line is formed in the same 

 place in all spectra of the copper compounds. If the tempe- 

 rature is increased by copious addition of oxygen, the spectra 

 disappear almost entirely, and instead of it the above line 

 appears very bright with a few other feebler ones. This line 

 might perhaps be due to a degree of oxidation of the copper, 

 were it not that it is formed, as I shall subsequently state, when 

 oxygen is completely excluded ; hence it must belong to copper 

 as metal. 



It follows therefrom that the spectrum which is only obtained 

 from the oxides of copper, and which I formerly considered to 

 be that of the metal, belongs to an oxide of copper. There 

 occur, therefore, in this, as in other copper compounds, at high 

 temperatures the lines of metallic, copper. Hence at a very 

 high temperature of ignition all copper compounds are decom- 

 posed into their constituents, whereas these compounds remain 

 undecomposed at a lower temperature — that, for example, of the 

 hydrogen flame. 



The spectrum of metallic copper produced by electricity con- 

 tains, besides the lines obtained by the flames, other lines of dif- 

 ferent luminous intensity. This is also the case with other 

 metals, with the sole difference that the additional lines in these 

 are of less luminosity. I cannot decide how these new lines are 

 formed, since the greater intensity of the light of the electrical 



