440 Prof. Fitzgerald on the Structure of Mechanical 
right angles to the direction of discharge. Their velocity of 
rotation evidently represents the magnetic force accompanying 
the discharge, and the momentum of the wheels represents the 
electrokinetic momentum of the current, 7. e. its self-induction. 
This is further illustrated by this, that if the frictional re- 
sistance be small enough, this momentum will carry the wheels 
beyond their positions of equilibrium, and there will result an 
oscillating discharge, such as occurs when an electric condenser 
is discharged through a sufficiently small resistance. If we 
suppose a certain amount of frictional resistance at any point 
along the line of discharge, we may see that the energy ex- 
pended on friction is conveyed to the place by the bands in 
the surrounding nonconductor and comes in at the side of the 
conductor, in accordance with Prof. Poynting’s theorem as 
to the direction of the flow of energy in an electrodynamic 
system. 
The mutual induction of two circuits may also be illus- 
trated by the model. Sufficient has been explained, how- 
ever, to show how electrostatic and electrokinetic phenomena 
are represented on the model. If a sudden movement of 
rotation be communicated to any set of wheels, it is evident 
that inertia will prevent their neighbours being instantaneously 
turned, while the connecting bands will be strained. Rotation 
will, however, be communicated to the neighbouring wheels, 
and from them to their neighbours, by a process which is a 
species of wave-propagation. If we consider the nature of 
the disturbance which is thus propagated, we see that it 
consists in a rotation whose axis is at right angles to the 
direction of propagation, and of a polarization of the bands 
which is at right angles both to the axis of rotation and to 
the direction of wave-propagation. ‘These are respectively a 
magnetic and an electric displacement, which are at right 
angles to one another and to the direction of wave-propagation. 
This is exactly in accordance with Maxwell’s electromagnetic 
theory of light-propagation. It is thus seen how the same 
model that can represent electrostatic and electromagnetic 
phenomena also illustrates luminiferous phenomena by its 
small oscillations. 
If we try to produce a tridimensional model by means of 
wheels geared together by bands, we are met by the following 
difficulty. The energy of the mode! we have been considering - 
may be represented in the following way:—Let ¢ represent 
the angular rotation of any wheel from a given position ; 
then the kinetic energy of an element will evidently be pro- 
portional to ¢’, while the potential energy of an element, 
depending as it does on the difference of rotation of neigh- 
