46 Colorado College Studies. 



In the last column is added the corresponding grouping 

 of the lines as determined by Professor INIichelson.^ It is 

 thus seen that the groupings made by ]\Iichelson according to 

 the structure of the spectral line is the same as that given 

 when the lines are grouped according to the value of the ratio 

 e/m. 



A study of Tables I. and II. and Fig. 2 leads to the follow- 

 ing conclusions : 



(1) A classification of lines according to the amount of 

 magnetic shift is of little value; but a classification 

 according to the value of e/m is significant. 



(2) The smaller the ratio e/m the less the broadening of 

 the component lines and the simpler the structure; 

 vice versa, the larger the ratio e/m the more the broad- 

 ening and the more complex the structure. 



Professor J. J. Thomson has shown that there is reason to 

 believe that the ionic charge is always a constant. If this 

 be applicable to the present case, we may conclude from Table 

 II. that the larger the ionic mass the simpler is the structure 

 of the line ; and the smaller this ionic mass the more complex 

 the structure. We also conclude that the difi'erent spectral 

 lines of a substance are due to distinct ions. 



^ Astrophys. Jr., 7, p. 13G, 1898; Nature, March 9, 1899. 



