314 * Prof. A. Schuster on Magnetic Precession. 



atomic weights tend to approximate to whole numbers far 

 more closely than can reasonably be accounted for by any 

 accidental coincidence. The chance of any such coincidence 

 being the explanation is not more than 1 in 1000, so that, to 

 use Laplace's mode of expression, we have stronger reasons 

 for believing in the truth of some modification of Prout's 

 law, than in that of many historical events which are uni- 

 versally accepted as unquestionable. 



XXVIL On Magnetic Precession. By Arthur Schuster, 

 F.R.S., Professor of Physics at the Owens College, Man- 

 chester*. 



1. TN a previous paper f I discussed the possible effects of 

 -L electric inertia, confining myself to the case of con- 

 ductors at rest. But Hertz, in* his interesting experiments 

 on the subject, showed that the most delicate method of 

 investigating the influence of inertia is based on the apparent 

 electromotive forces which are introduced by the motion of 

 conductors. If electricity possesses inertia, the rotation of a 

 body through which currents pass, affects the flow of these 

 currents in the same manner as the earth's rotation affects 

 the direction of currents of air on its surface. 



Hertz obtained only negative results, but could fix an upper 

 limit to the possible inertia of electric currents. It occurred 

 to me that this inertia, even if below the limits given by 

 Hertz, might show accumulated effects, when the currents 

 last for a sufficient time. If the earth's magnetism be 

 •due to electric currents, general considerations suggested 

 to me that the effects of inertia might explain the secular 

 variation. The following investigation shows that indeed 

 inertia would cause a " magnetic precession" precisely of the 

 character of the secular variation, but that this precession 

 would be very much slower than the variations which are 

 actually observed. 



2. If m is the mass of positive electricity in unit volume, 

 and u the velocity, the energy of an electric current, so far as 

 it depends on this mass, would be mu 2 per unit volume, 

 assuming for the sake of simplicity that positive and negative 

 electricity move with the same velocity. If i is the current- 

 density, we may substitute J//i 2 for this energy, where //. has 

 the same meaning as in my previous paper. If q is the 



* Communicated by the Physical Society : read December 14, 1900. 

 f P. 227 of this volume. 



