34 BELL SYSTEM TECHNICAL JOURNAL 



but the components parallel to the applied field are jA/2t and 

 gjeh I {4:Trmc) , respectively.) For some values of s, I and j, e.g. 4, 2 

 and 2, ^ is greater than 2, and for some values it is less than 1. 



The sign as well as the magnitude of the rotation is of importance. 

 All experiments are consistent with the idea that the magnetic moment 

 is due to the spinning or circulation of negative electrons rather than 

 of positive charges. 



The results to be described below show that in ferromagnetic 

 materials generally the value of g has nearly the value two and not at 

 all the value one, so we conclude that ferromagnetic processes are 

 concerned primarily with the spins of the electrons and not their 

 orbital motions. When a change in magnetization takes place we 

 therefore attribute it to a change in the direction of spin of some of 

 the electrons, and believe that the orientations of the orbits are 

 disturbed but slightly. This change is illustrated in Fig. 19. In some 



/ \ / \ 



/ \ / \ 



fa) I + . 11 f ,' + , \ 



)\ i; 



\ y \ ' 





.b. ( .. \\ [[ .. j 



H»0 



Fig. 19 — In the common ferromagnetic materials a change in magnetization is 

 effected by a change in the direction of electron spin, not in the direction of motion 

 of the electron in its orbit. 



paramagnetic materials, on the other hand, the reorientation of orbits 

 plays an important part. 



Gyromagnetic Experiments. The first gyromagnetic experiment to 

 be performed successfully was magnetization by rotation. After an 

 unsuccessful trial by Perry ^^ in 1890, the experiment was considered 

 independently in 1909 by Barnett ^^ who in 1914 obtained the result, 

 then inexplicable, that g was approximately twice the classical value 

 one. Richardson, 2* in 1907, was the first to propose rotation by 



^1 J. Perry, as quoted by Barnett, ref. 27. 



22 S. J. Barnett, Science, 30, 413 (1909); Phys. Rev., 6, 239-270 (1915). An 

 accidental error in the calculation of the results was corrected in Joiir. Wash. Acad. 

 Sci., 11, 162 (1921). Magnetization bv rotation. 



23 0. W. Richardson, Phys. Rev., 26, 248-253 (1908). 



