ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 



133 



' It follows that in the haloid compounds of barium (excepting the 

 fluoride) the distance between corresponding spectral lines is directly- 

 proportional to the atomic weight, and that in the haloid compounds of 

 calcium and strontium (excepting again the fluoride) these distances are 

 inversely proportional to the atomic weights. Further, that there is such 

 a point in the spectrum of each metal that the bands keep their relative 

 distances from it in the diflerent compounds. Here also we must except 

 the fluorides.' 



Mitscherlich's measurements were not sufficiently accurate to prove 

 his statement satisfactorily, but we can easily test it by means of the 

 more exact measurements of Lecoq de Boisbaudran, who has carefully 

 mapped the spectra of the chloride, bromide, and iodide of barium. A 

 glance at Fig. 2, which represents these spectra according to Lecoq's 



Fis:. 2. 



Barium 

 chloride 



Barium 

 bromide 



Barium 

 iodide 



-r 1 1 1 f I 1 1 1 1 1 1 1 1 I 



■»7 4a 43 so SI S2 5J 5* SS 56 57 58 S9 SO 6/ 



measurements will show their similarity. In Table VII. the first column 

 gives the wave-length of the bands seen in the spectrum of the barium 

 chloride. The strongest band is denoted by o, and the order of the 

 Greek letters gives the order of the intensity. In the second column a 

 comparison is given between the observed wave-lengths of the spectrum 

 of barium bromide, and the same wave-lengths calculated on Mitscher- 

 lich's supposition that the distance between the lines is proportional to 

 the atomic weight of the compound. The band a is taken as starting 

 point. The third column gives the calculated value for the band jo of 

 barium iodide, which compound only shows a and /3. 



Table YII. 



It will be seen that for the four central bands of barium bromide the 

 agreement seems good, but the two outer bands do not follow the rule, 



