294 



SCIENCE. 



[N. S. Vol. XII. No. 295. 



mon salt between the poles of the electro- 

 magnet and focused the light on the slit of 

 his spectrometer, arranging the flame so 

 that the D-lines were sharply defined. As 

 soon as the magnet was excited both lines 

 widened out very much. By a careful 

 series of subsidiary experiments he showed 

 that the widening was due directly to the 

 action of the field and was not a secondary 

 effect such as might be caused by changes 

 of density in the flame. 



These results were communicated before 

 publication to Professor Lorentz, who 

 showed Dr. Zeeman that the widening could 

 be predicted from Lorentz's theory that 

 light is generated by the vibrations of elec- 

 trically charged particles or ions ; and that 

 the same theory indicated that the edges of 

 the widened lines should be plane-polarized 

 or circularly-polarized according as the 

 light falling upon the slit came from the 

 source in a direction perpendicular or par- 

 allel to the lines of magnetic force, and that 

 the amount of the widening would give the 

 ratio of the charge to the mass of the lumi- 

 nous particles. Zeeman was able to verify- 

 fully the predictions as to polarization, and 

 deduced from Lorentz's equations, as a 

 rough value for the ratio e/ni, the value 

 10'. 



The substance of the reasoning which led 

 Lorentz to his conclusions is this : the mo- 

 tion of any ion can be resolved into a 

 rectilinear component along the lines of 

 force and two circular components in oppo- 

 site senses in a plane normal to the same 

 lines. These moving charges constitute 

 currents, and consequently there are elec- 

 tromagnetic forces acting on the particles 

 carrying the charges, owing to the presence 

 of the magnetic field. The component 

 along the lines of force is unaffected, while 

 one of the circular components is accel- 

 erated and the other retarded by an equal 

 amount. Hence we have present three dis- 

 tinct vibrations : one linear, along the lines 



of force, with the same period as the undis- 

 turbed motion ; and two circular, in oppo- 

 site senses, having periods, one a little 

 longer and the other a little shorter than 

 the original. Hence if the difference of 

 periods introduced is sufficient to resolve 

 the three lines in the spectrometer, and the 

 light falling upon the slit comes from the 

 source in a direction parallel to the lines of 

 force, two lines instead of one will appear 

 in the spectrum, one composed of right- 

 handed and the other of left-handed polar- 

 ized light ; for the linear vibration can send 

 no wave in its own direction. On the 

 other hand, if the light proceeds directly 

 across the lines of force to the slit, a triplet 

 will be seen, consisting of a central com- 

 ponent polarized in a plane normal to the 

 lines of force and two lateral components 

 polarized in a plane through the lines of 

 force (bearing in mind that on the electro- 

 magnetic theory of light the direction of 

 electric force is normal to the plane of 

 polarization). But, if the difference of 

 period is smaller, the lines of the doublet 

 in the one case and of the triplet in the 

 other, being streaks of finite width and not 

 lines in the mathematical sense, will over- 

 lap, forming a single widened line whose 

 edges alone show any definite polarization. 

 Lorentz's mathematical treatment, which 

 we will omit here, leads to the expression 



J A 



IF' 





where A is the wave-length of the original 

 line, J/ is the difference in wave-length in- 

 troduced between the extreme components 

 of the triplet by the external magnetic 

 force H, e is the charge on the moving ion, 

 m its mass, and V the velocity of light. 



It follows from this that if e/m is the 

 same for all the luminous ions which give 

 the spectrum of any one substance, the 

 separation J A is proportional to the square 

 of the wave-length ; but it became evident 



