Mi; i HAIJLKS |K WATTEVILLK ON FLA.MK SPECTRA. 155 



The flame spectra of the six metals studied which l>elong to MENDELEEFF'S second 

 group are character! xt-d l>y the presence in each of them of one extremely strong line, 

 wlio.se brilliancy far surpasses that of any other line. 



This line exists in the spectrum of every portion of the flame and has the wave- 

 length : 



2852*22 in the spectrum of magnesium, 



3075-99 x.inc, 



3261-17 cadmium, 



4226-91 calcium, 



4607*52 strontium, 



5535-69 barium. 



If we classify these six elements into two sub-groups, the one containing 

 magnesium, zinc, and cadmium, whose atomic weights are respectively 24 '36, 65*4, 

 and 112'4, the other containing calcium, strontium, and barium, whose atomic 

 weights are 40, 87*6, and 137 '4, it will be noticed that the wave-length of the line 

 diminishes with the atomic weight of the element. It seems possible, therefore-, that 

 all these six lines correspond to each other. There seems no doubt on this jM>int as 

 n-ganls the lines of magnesium, calcium, strontium, and barium. RYDREKU already 

 unites these rays, and there is strong evidence that they should IK- grouped together. 

 The correspondence of the other lines is more doubtful. Taking account of the dis- 

 placements due to differences of atomic weight, it seems possible that the mercury 

 line at 3650*31, which is strong in the arc, is the line corresponding to the above 

 in the mercury spectrum. 



MERCURY. 



The attempts which have been made to obtain the flame spectrum of mercury have 

 U't'ii unsuccessful. No lines of mercury have been seen in the flame into which has 

 been blown the spray from solutions of metallic mercury in nitric acid, and from 

 aqueous solutions of mercuric chloride to which sodium chloride has been added. 

 M. GOUY has elsewhere stated the same fact. 



It may be of interest to note that no mercury lines are present iu the solar 

 spectrum. 



x _ 



