1901.] and Melting Points to Atomic Mass. 9 



Magnesium. 



Zinc. 



Cadmium. 



Mercury. 



51841 



5173 \ triplet 

 5168 J 

 4571 

 2852 



4812"] 



4722 \ triplet 

 4680 J 

 3076 



5086 ] 



4800 \ triplet 

 4678 J 

 3261 



547 

 435 



Mercury appears to give only a continuous spectrum in tlie oxyhy- 

 drogen flame. Both zinc and cadmium give weak spectra, proving 

 that the emissive power of their molecules in the oxyhydrogen flame 

 is low compared with many other metals. Mitscherlich* records two 

 lines in the spectrum of cyanide of mercury, heated in the oxyhy- 

 drogen or oxycoal-gas flame, which correspond to two lines of the 

 strongest triplet in the arc and spark spectra of mercury. The wave- 

 lengths of the lines of the triplet are 5460, 4359, and 4078. This 

 triplet is given by the merest trace of mercury in a vacuum tube. 



There is every reason for believing that the four triplets correspond 

 to one another. The lines are all given by the same value of m in 

 Kydberg's formula. The work of Preston,! of Lord Blyths wood and 

 Dr. Marchant, ! and of Keese§ on the Zeeman effect give additional 

 support to this view. 



Observer. 



Magnesium, zinc, 

 and cadmium. 



Mercury. 



Mercury. 



Preston. 



Blytliswood and. 

 Marchant. 



Reese. 



Strength of 

 magnetic field. 

 C.G.S. units. 



20,000 



40,000 



6,000 



24,000 



24,500 



Least refrangible 

 line 



Middle line 



Most refrangible 

 line 



j 



Nebulous 

 triplet 

 Quartet 

 Triplet 



Triple 

 triplet 



Sextet 

 Triplet 



Triplet 

 Quartet 



Doublet, trip- 

 let, doublet 

 Sextet 



i Triplet, "prob- 

 ably more." 

 Triplet. 

 Triplet, 



There remain the isolated lines : two of magnesium and one each of 

 zinc and cadmium. 



* ' Phil. Mag.,' Sept. 1864, plate. 



f ' Trans. Eoy. Dublin Soc.,' vol. 7, series ii, ~No. 2. 



t ' Phil. Mag.,' vol. 49, p. 384, 1900. 



§ ' Astrophys. Jour.,' Sept. 1900, p. 120. 



I, Gray has now resolved this into nine lines, 'Brit. Assoc.,' 1901. 



