80 Professor G. F. Fitz Gerald [March 21, 



plate, though the rapid ones are quite stopped by either — thus showing 

 that wave-propagation in a conductor is of the nature of a diffusion. 



In all cases of diffusion where we consider the limits of the 

 problem, terms involving the momentum of the parts of the body- 

 must be introduced. It appears from elementary theories of diffusion 

 as if it were propagated instantaneously, but no action can be propa- 

 gated from molecule to molecule, in air, for instance, faster than the 

 molecules move, i. e. at a rate comparable with that of sound. In 

 electromagnetic theory corresponding terms come in by introducing 

 displacement currents in conductors, and it seems impossible but that 

 some such terms should be introduced, as otherwise electromagnetic 

 action would be propagated instantaneously in conductors. The 

 propagation of light through electrolytes, and the too great trans- 

 parency of gold leaf, point in the same direction. 



The constitution of these waves was then considered, and it was 

 explained that if magnetic forces are analogous to the rotation of the 

 elements of a wave, then an ordinary solid cannot be analogous to the 

 ether because the latter may have a constant magnetic force existing 

 in it for any length of time, while an elastic solid cannot have con- 

 tinuous rotation of its elements in one direction existing within it. 

 The most satisfactory model, with proj)erties quite analogous to those 

 of the ether is one consisting of wheels geared with elastic bands. 

 The wheels can rotate continuously in one direction, and their rota- 

 tion is the analogue of magnetic force. The elastic bands are 

 stretched by a difference of rotation of the wheels, and intro- 

 duce stresses quite analogous to electric forces. By making the 

 elastic bands of lines of governor balls, the whole model may 

 have only kinetic energy, and so represent a fundamental theory. 

 Such a model can rej)resent media differing in electric and magnetic 

 inductive capacity. If the elasticity of the bands be less in one 

 region than another, such a region represents a body of higher electric 

 inductive capacity, and waves would be propagated more slowly in it. 

 A region in which the masses of the wheels was large would be one 

 of high magnetic inductive capacity. A region where the bands 

 slipped would be a conducting region. Such a model, unlike most 

 others proposed, illustrates both electric and magnetic forces and their 

 inter-relations, and consequently light propagation. 



In the neighbourhood of an electric generator the general distri- 

 bution of the electric and magnetic forces is easily seen. The electric 

 lines of force must lie in planes passing through the axis of the 

 generator, while the lines of magnetic force lie in circles round this 

 axis and perpendicular to the lines of electric force. It is thus 

 evident that the wave is, at least originally, polarised. To show this, 

 the small-sized oscillators with parabolic mirrors were used, and a 

 light square frame, on which wires parallel to one direction were 

 strung, was interposed between the mirrors. It was shown that such 

 a system of wires was opaque to the radiation when the wires were 

 parallel to the electric force, but was quite transparent when the 



