44 



Colorado College Studies. 



showing that for a given value of A the change of wave-length 

 should be proportional to the value of H, provided no con- 

 straint is present. It is found that a constraint is present 

 at low temperatures, but disappears as the temperature rises. 



a:::-"i- 



Fig. 1. 



The conditions of high temperature and low pressure are 

 best fulfilled by the vacuum tube and hence this was used as 

 a source of illumination. The arrangement of apparatus is 

 shown in Fig. 1. 



The difficulties encountered were such as to reduce the 

 satisfactory observations to the lines of sodium, mercury, 

 cadium and zinc. The results are summarised in Table I. 

 and also shown graphically in Fig. 2. 



Table I. 



Substance. Line. 



Sodium^ Yellow line. 



Mercury Yellow line. 



Mercury Green line. 



Mercury Violet line. 



Cadmium Red line. 



Cadmium Green line. 



Cadmium Blue line. 



Zinc Blue line. 



Magnetic shift 

 for H = 10,000. 



0.414 

 0.256 

 0.310 

 0.240 

 0.262 

 0.240 

 0.274 

 0.288 



A. U. 



'The echelon spectroscope shows the separation of the com- 

 ponents of D, to be about two thirds of that of D^. The value here 

 given belongs to D,, the line having the greatest magnetic shift. 



