( li ) 



the lines of the triplet are broadened. If the supposition of § J con- 

 cerning the relative position of A, B, and holds, the width of the 



s 

 middle component becomes — , but that of the outer ones .v. The 



components are most intense in the centre. 



s 



12. The ratio of the axes of the ellipses in § 10 is 1 : 1 -j — resp. 



7' 



s 

 1:1 if the electric effect is small i-elatively to the magnetic one. 



r 

 Hence the intensities corresponding to vibrations parallel to OX 



and to OV, differ bj an amount proportional to — . 



r 



If the light is examined parallel to the magnetic lines of force 

 and the separation of the magnetic doublet is 100 times the distance 

 ,9, then the difference of the intensities of the vibrations parallel to 

 the axes of the ellipses w^ould be V50 P^i't of the intensity corres- 

 ponding to the vertical or horizontal vibrations. Moreover the difference 

 of intensity would have opposite sign with both components. 



This method seems capable of rendering important services in 

 searching for an electi-ic effect. F'or differences of intensity of 2 7o 

 can be ascertained ^) with certainty by using photographic-photometric 

 measurements. We, therefore, must be able to discover an electric 

 effect one hunderd times smaller than the magnetic effect traceable 

 by means of a spectroscope with the maximum resolving power 

 serviceable under the conditions of the experiments. It does not 

 look now entirely impossible to ascertain under favourable circumstances 

 an electric effect of the order of magnitude estimated by theory. 



13. If the electric field is non-uniform, but gradually increasing 



upwards, then the components bend more 

 and more towards the red (Fig. 3). 



The middle component A bends more than 

 the other ones, if the case considered in § 5 

 is under consideration (parallel electric and 

 magnetic fields). 



\ 



;volch 



Fig. 3. 



If, however, the electric force is vertical 

 and the magnetic one horizontal, then the 

 position of the components must become 

 that sketched in the next figure; it is now 

 the middle component which gets the smaller 

 curvature. 



\ 



— ► vwveh- 



Fig. 4. 



1) Gomp. P. P. Koch. Ann. d. Phys. Bd. 30. S. 841. 1909. 



