698 



2m 

 Sm^ — T (4) 



§ 2. Consequence of the existence of a nimjnetic moment 

 of momentum. 



Any change of the moment of momentum ^'3)i of a magnetized 



body gives rise to a couple 6* determined by the vector equation 



dm dl 



d~ — 2 — = 1,13.10-7- (5) 



dt dt ' 



where the numerical coefficient has been deduced from the known 



e 

 value of ~ for negative electrons. 

 m 



It has been our aim to verify the relation exjiressed l)y (5). We 

 shall show in the first place that the calculated effect is not too 

 small to be observed. Let the body be an iron cylinder with radius 

 R, which can rotate about its vertical axis. We shall deduce from 

 (5) the angular velocity to the cylinder acquires by the leversal of 

 a longitudinal magnetisation, which we suppose to have tiie satura- 

 tion value Is- Denoting by Q the moment of inertia of the cylinder, 

 and writing A for the above coefficient 1,13.10—", we find 



Qw = \ddt = 2;i/,. . 



Now, if the saturation value of the magnetisation per cmMs 1000, 



M 



which is not a high estimate, we have 1,^=^^. 1000. The moment 



of inertia is (2 =: è ^^H', and we find for H = 0,1 cm 



to = 0,6 . 10-2, 

 an angular velocity that can easily be observed. 



§ 3. Description of the method. 



At first sight it seems that equation (5j may be tested in the 

 following way. A soft iron cylinder C is suspended by a thin wire 

 D coinciding with the axis of the cylinder prolonged, the period of 

 the torsional oscillations being a few seconds. Let the cylinder C 

 be surrounded by a coil K whose axis coincides with that of C. 

 Then, on reversing a current in K, a rotation of (' ought to be 

 observed. In reality, however, this simple method cannot be thought 

 of. As the field of the coil will not be uniform the cylinder 

 would probably show highly irregular motions completely masking 

 the effect that is sought for. 



