MODELS OF POSITIVE ION 309 



About three years ago some experiments were 

 carried out by tbe author on the fluorescence of 

 the metallic vapours of sodium and potassium. 

 They were undertaken with the view of determining 

 whether the Zeeman efi"ect exists in the case of 

 gases in a state of fluorescence or to determine 

 whether the fluorescent light is polarised in a 

 magnetic field, even in the continuous part of the 

 spectrum. Sodium gives the Dj and Da lines 

 when fluorescent, as discovered by Wiedemann and 

 Schmidt, as also the continuous or fluted bands 

 corresponding to the absorption spectrum of the 

 gas, when cold. Similarly with the fluorescence 

 spectrum of potassium vapour it was again the 

 absorption spectrum that was visible whilst the 

 bright K lines were invisible. The experiments 

 were entirely negative as regards the efi'ect of a 

 magnetic field on the polarisation, and, moreover, 

 the effects which it was hoped would be detected 

 at the borders of the bands could not be found. 



The condensation of the metallic vapour on the 

 walls of the tube was a source of very great 

 trouble in the above experiments ; but the 

 condensed vapour on the glass showed also a 

 distinct change of refrangibility, and the fluores- 

 cence was much greater. 



It is interesting to note that the Hg vapour also 

 shows a very marked tendency, under very slight 

 provocation, such as the falling of mercury drops 

 in a mercury still, to luminesce when it condenses 

 on the glass. The light is green and characteristic 



