178 Prof. Mitseherlich on the Spectra of Compounds 



held good, inasmuch as each of the spectra is composed of the 

 spectrum of the metal and that of the oxide. 



In the same paper, by the experiment in which when chloride of 

 potassium is brought into a flame with much chloride of ammo- 

 nium no spectrum is perceived, I have shown that under certain 

 circumstances metallic compounds of the first order, even when 

 volatile, may give no spectra. This is confirmed by the fact that 

 even the yellow colour of the sodium flame almost disappears if a 

 flame containing sodium is brought over strongly volatilizing sal- 

 ammoniac, — and further by the fact that if any sodium, lithium, 

 or potassium compound is investigated by method 3 (H and 

 CI), not even a coloured flame, and still less a spectrum, is 

 formed. The result is similar if compounds of the alkalies are 

 volatilized in burning sulphuretted hydrogen; it can be dis- 

 tinctly observed that the interior core retains its colour per- 

 fectly if these compounds are volatilized in it, but that in the 

 outer core, where the sulphur is already burned, a feeble colour 

 is formed. Under certain circumstances, metals, even when 

 volatilizing in the flame, may show no spectrum — thus, for in- 

 stance, in the case of any of the compounds of mercury, ex- 

 cepting the cyanide, and excepting mercury itself, by method 1 

 (wick of platinum wire), 2 (H, or coal-gas, and 0), or by method 6 

 (volatilization in the glass tube). If the mercury compounds are 

 heated higher, for instance by the electrical spark, or if the cy- 

 anide is investigated by method 2 (H, or coal-gas, and 0), the 

 spectrum is observed very distinctly. 



Nickel, cobalt, and aluminium, investigated in the most 

 varied compounds by method 2 (H and 0) and method 6 

 (evaporation in the glass tube), gave no perceptible spectrum; 

 this was also the case with the compounds of titanium, tungsten, 

 vanadium, molybdenum, uranium, platinum, and palladium in- 

 vestigated by method 6 (volatilization in the glass tube) and other 

 methods. When aluminium was burned in bromine, a conti- 

 nuous brightness only could be produced, as was the case with 

 the combustion of zinc in iodine. Arsenic and antimony, exa- 

 mined as chlorides by method 3 (H and CI) and by method 6 

 (volatilization in the glass tube), in the form of other compounds 

 and as metal, showed a luminosity without brighter or darker 

 lines in the spectrum. Only by method 7 (solid electrodes) 

 could lines be found in the methods adduced, which I thus 

 investigated. I have not investigated the still rarer metals and 

 their compounds. 



From the decompositions which, according to these investiga- 

 tions, take place in the flame, it would follow that if, for instance, 

 metallic copper is heated with common salt, the chlorine which 

 is liberated at a high temperature by the decomposition of the 



