(10) 



Taking now the lower signs in (14; and (15). 



1 ,/ 



We have ?^ := a -\- — (s — v r^ -{- s") =: a — ö, and 



s — Vr'-^-s"^ . r 



^ 5 , 



The coefficient of i is positive and <^1. The electron performs an 

 elHptic vibration, left-handed and in the phine of A, Y; the axes 

 are again parallel to OX and OY, but the major axis pai-allel to OA". 

 We now get a dissymmetrical triplet. 



Let A and B be the two lines of the electric doublet, the electric 

 field alone being present. 



3. If now the magnetic field is set up Jj 



remains with vibrations parallel to the mag- 



— vv^t iietic force. Tvvo new components are added 



— s -^ \^<s'~A which one may consider as originating from 



A and B by a displacement equal to a 



distance 0. 



If the sign of r is negative, that is if the. 

 magnetic force is reversed, then our consi- 

 derations still apply with but a suiall change. 

 To the frequency n=:b -\- n then corresponds 



q — x . 



r 



The coefficient of i is in absolute measure > 1, but negative. 



The value of the coefficient of i with the lower sign is the same 

 as above, only the sign of the expression is reversed. 



The figure still covers the case, but the motion in the ellipses 

 takes place in the opposite sense. 



The product of the two values of q determined by (15) is equal 

 to unity. Hence the product of the horizontal, as well as of the 

 vertical axes of both the vibration ellipses is always equal to unity. 



It may be not inappropriate to make here the remark that the 

 lines of the triplet considered in this § exert a kind of "repulsion" 

 upon each other; Lorentz ') proved that such mubt be generally the 

 case for two spectral lines. 



11. If one has to do with an oscillating electric tield (cf. § 7 



1) LoRENTz. Encyclopadie d. math. W. V. 3. Heft 2. Magneto-optische Phanomene 

 No. 36 u. No. 53. 



