M. Mitscherlich on the Detection of Chlorine by the Spectrum. 449 



directly over the ammonia and quite near the spiral, the oxygen 

 burns, producing the long-drawn sound of the chemical har- 

 monicon. The oxygen-ammonia flame hangs then as a greenish- 

 yellow bubble at the aperture of the glass tube, and, without 

 being extinguished, moves up and down. The cork which sup- 

 ports the platinum spiral can only be laid crosswise over the aper- 

 ture of the boiling flask, as it is often projected by the explosions. 



Mitscherlich* describes a method according to which, as he 

 states, chlorine, iodine, and bromine compounds can be discovered 

 by the spectrum-apparatus, and even the smallest quantities of 

 these halogens detected. 



The method depends upon the property which copper has of 

 forming compounds with the halogens, which are very hard to 

 decompose by the influence of other substances, and are only 

 resolved into their constituents at the highest attainable tempe- 

 rature. The spectra of these compounds are admirably adapted 

 for recognizing the smallest quantities of the halogens. In this 

 respect copper has a material advantage over other metals, inas- 

 much as their haloid compounds are decomposed into their con- 

 stituents at lower temperatures, and the spectra either of the 

 oxide or of the metal are obtained. 



The application of the new method is as follows. The dried 

 solid substance is mixed with half its weight of sulphate of ammo- 

 nia, and a tenth its weight of oxide of copper, and heated accord- 

 ing to method 3 of Mitscherlich's paperf — that is, is placed in the 

 bulb of a glass tube and slowly heated in a current of hydrogen 

 gas. The hydrogen being inflamed, is at first coloured by some 

 lower degree of oxidation of copper until the oxide is reduced : 

 this is too undefined to give a definite spectrum ; subsequently 

 the spectra of the haloid salts of copper occur. With very feeble 

 reactions the chlorine compound is best recognized by the lines 

 at 105 and 109, and by the brightnesses 85 to 87 of the scale; 

 the bromine compound by the brightnesses 85, 88^, and 92; 

 and the iodide of copper by the brightnesses 96, 99, and 102^. 

 These numbers refer to spectra of the salts which appeared in 

 PoggendorfFs Annalen of last year, and of which improved copies 

 in colour are promised by the author. 



By the above method, J per cent, of chlorine, J per cent, of 

 bromine, and 1 per cent, of iodine can be detected without much 

 practice. A skilful observer can detect far smaller quantities. A 

 disadvantage of the method is that the sulphate of ammonia, 

 which decomposes at a high temperature, gives the ammonium 

 spectrum and obscures the reaction. In order, then, to detect the 



* Poggendorff's Annalen, August 1865. 

 t Phil. Mag. S. 4. vol. xxviii. p. 169. 



