1871.] Spectra of Metallic Compounds. 61 



a way as to prevent the vapourised substance from coming 

 into contact with the air. He thus obtained the spectra of 

 the chlorides unmixed with the spectra of the corresponding 

 oxides. 



The feeble illumination in the green part of the spectrum, 

 when hydrogen is burned in chlorine, may fairly be 

 attributed, according to the same interpretation which we 

 apply to the spectra of the metals when burned in chlorine, 

 to the chloride of hydrogen or hydrochloric acid ; and the 

 broad bluish-green nebulous band when hydrogen is burned 

 in air or oxygen, to the spectrum of aqueous vapour. The 

 spectra of metals as obtained by Mitscherlich with the 

 electric spark have been re-determined by Huggins and 

 Miller with great care, and the lines obtained have been 

 referred to a scale in which the atmospheric lines form 

 fiducial points. 



The maps given by Mitscherlich and Diacon, being 

 referred to an arbitrary scale, are intelligible with difficulty. 

 This difficulty applies not only to the position of the lines, 

 but in a still greater degree to their relative intensity and 

 brightness. It was very much to be desired that the benefits 

 of their labours should be made available to us in the present 

 advanced state of spectroscopy, and an attempt was made 

 to reduce their measurements to normal wave-lengths 

 according to Angstrom's tables. But this was altogether 

 impossible with Diacon's map, since the intervals had been 

 micrometrically determined, and no comparison has been 

 made with the solar or other standard lines. In 

 Mitscherlich's maps, the lines a, a/D, e, b, and f are marked, 

 but when a graphical construction was attempted, by 

 making the values of these points, as given in the maps, 

 the ordinates, and their corresponding wave-lengths the 

 abscissas, of a curve, the curve was so irregular that the 

 attempt had to be abandoned. 



Professor A. R. Leeds has instituted some interesting 

 experiments on the spectra of the metallic compounds. He 

 employed the flame of a Bunsen burner, since it is to this 

 source of heat that the spectra in ordinary laboratory work 

 is referred. The resulting spectra are of two different kinds. 

 In the case of oxygen salts, the spectra of the oxide of 

 the metallic radical is obtained, the lines and bands being 

 more or less broadened and brightened according to the 

 degree of volatility of the salt. With haloid salts, the 

 spectrum proper to the compound and also- the spectrum of 

 the oxide of the metallic radical is obtained. Two instru- 

 ments were employed in the observations ; one, a single prism 



