( 346 ) 
lines can be drawn from general principles, independently of any 
particular theory. 
One of these results is that light of a definite frequency radiated 
along the lines of force, can never show a trace of linear or elliptic 
polarization; it must either be unpolarized, or have a circular pola¬ 
rization, partial or perfect. 
Larmor 1 ) in 1900 concluded that perfect circular polarization of 
the components of the doublet, would prove that the corresponding 
permanent types of vibration in the molecules are exactly circular. 
In order to explain the more complicated magnetic effects Voigt 1 3 * ) 
specialised the general theory of vibrating eleciric systems by supposing 
particular magnetic links between electrons and by introducing the 
hypothesis that the luminous particles take a definite orientation 
under the action of the field. A particular direction, the “axis” of 
each particle, becomes under the action of the field parallel to the 
lines of force. 
A rotation of the particles around this specified axis undoubtedly 
is possible, at least it is not excluded. This rotation has no influence 
upon the frequency. The orbits of the electrons however are rotated 
as also are the orbits of the “equivalent” electrons recently introduced 
by Lorentz 8 ) in order to simplify the theory of systems containing 
a number of electrons. 
The linear polarization of the light emitted normally to the field, 
proves that the orbits of the electrons either are straight lines parallel 
to the lines of force or ellipses in planes perpendicular to the field. 
The completeness of the circular polarization parallel to the magnetic 
force would prove that the ellipses are circles. Partial circular pola¬ 
rization however would prove the existence of ellipses with all 
possible fortuitous orientations in planes normal to the field. 
In general one would expect, that the components of a magnetic¬ 
ally subdivided line emit partially polarized light parallel to the 
field. Without hypothesis or further measurements this more general 
statement must even be regarded as the most probable. A quantitative 
examination of the ratio of circularly polarized light contained in 
the total light emitted by the components, has not been made 
till now. 
2. My observations with Rowland’s grating are consistent with 
b Larmor. Aether and Matter, p. 845. 
2 ) Voigt. Magneto- und Elektrooptik, Teubner, Leipzig, p. 98. 1908. 
3 ) Lorentz. Theorie der magneto-optischen Phanomene. Encyclopadie der matbe* 
matischen Wissenschaften. V. 3. Heft 2. p. 217. 1909. 
