800 
MR. .T. LARMOR ON A DYNAMICAL THEORY OF 
by the distribution of magnetic intensity over the outer boundary of the magnets. 
And the principle of energy by itself shows that if we bound the two aggregates by 
moving surfaces which always pass through the same particles of the medium, the 
Increment of the kinetic energy outside is equal to the work done in the actual motion 
by the pressures transmitted across the surfaces of the two aggregates; though we 
ai’e unable to extend this result to arbitrary virtual displacements of the surfaces. 
Nor is the method of § 58 now applicable to complete the proof, because it is 
impossible to have an equipotential surface surrounding a magnetic system. 
108. In all theories which ascribe the induction of electric currents to elastic 
action across the interveiiing medium, a discrepancy arises when the induction is 
ju’oduced by movement through a steady magnetic field : for in such cases there is no 
apparent play of electric force across the field. This difficulty may perhaps disappear, 
on the present view, when we regard such a field, not as an absolutely steady motion 
like fluid circulation round fixed cores, but as the statistically steady residue of 
elementary elastic disturbances sent out through the medium by the molecular 
discharges which maintain the inducting currents, or by changes of orientation and 
other disturbances of the molecules of the permanent magnets, such as are involved 
in any kinetic theory of matter. These elastic disturbances do not spread out 
indefinitely as waves, but come to an end when the medium has attained a new 
steady state which they have been instrumental in forming. The progress and decay 
of each small disturbance generates a current on the secondary system, whose 
integrated amount would be null if that system were at rest : but in the actual 
circumstances of movement during the progress of the induction there will be a 
residual value. The aggregate of such differences between elementary direct and 
reverse induced currents would constitute the observed total current. Thus as 
regards induction, change of the magnetic field of a permanent magnet would act in 
the same way as that of an ordinary current, notwithstanding that if each molecule 
of the magnet were held fixed there might (§ lOG) be no induction. 
On these grounds, the field of a permanent magnet would be regarded, not as a 
steady circulation of the eether, absolutely devoid of elastic reaction, but as the 
statistically steady resultant of the changing fields of the incessantly moving 
molecules which make up the magnet. The steady field of motion associated with a 
fixed magnetic molecnle would be maintained by fluid pressure alone ; but when the 
molecide is rotated, some agency is required to prevent slip during the establishment 
of the new steady motion ; and in this way the elasticity may come into play. 
In ordinary hydrodynamics, the process of the establishment of a fluid motion is kept 
out of sight: it is simply assumed that the motion can be set up without slip, and 
that it is set up practically instantaneously throughout the field. In the present 
problem on the other hand, something formally equivalent to slip does occur across 
the dielectric gaps in each electi ic circuit; and this circumstance modifies the process 
of establishment of the )notion. 
