400 



SCIENCE. 



[N. S. Vol. IV. No. 90. 



on a gas is not destroyed by a considerable 

 rise in temperature ; it is, for example, not 

 destroyed if it be sucked through metal 

 tubing raised to a red heat. The conduc- 

 tivity is, however, destroyed if the gas is 

 made to bubble through water, it is also de- 

 stroyed if the gas is forced through a plug 

 of glass wool. This last effect seems to in- 

 dicate that the structure which confers con- 

 ductivity on the gas is of a very coarse 

 kind, and we get confirmation of this from 

 the fact that a very thin layer of gas ex- 

 posed to the Eontgen rays does not con- 

 duct nearly so well as a thicker one. I 

 think we have evidence from other sources 

 that electrical conduction is a process that 

 requires a considerable space — a space large 

 enough to enclose a very large number of 

 molecules. 



Thus Koller found that the specific re- 

 sistances of petroleum, turpentine and dis- 

 tilled water, when determined from experi- 

 ments made with very thin layers of these 

 substances, was very much larger than that 

 determined from experiments with thicker 

 layers. Even in the case of metals there 

 is evidence that the metal has to be of ap- 

 preciable size if it is to conduct electricity. 

 The theory of the scattering of light by 

 small particles shows that, if we, assume the 

 truth of the electro-magnetic theory of light, 

 the effects should be different according as 

 the small particles are insulators or con- 

 ductors. When the small particles are non- 

 conductors, theory and experiment concur 

 in showing that the direction of complete 

 polarization for the scattered light is at 

 right angles to the direction of the incident 

 light, while if the small particles are con- 

 ductors, theory indicates that the direction 

 of complete polarization makes an angle of 

 60° with the incident light. This result is 

 not, however, confirmed by the experiments 

 made by Prof Threlfall on the scattering 

 of light by very small particles of gold. 

 He found that the gold scattered the 



light in just the same way as a non-con- 

 ductor, giving complete polarization at 

 right angles to the incident light. 

 This would seem to indicate that those 

 very finely divided metallic particles no 

 longer acted as conductors. Thus there 

 seems evidence that in the case of conduc- 

 tion through gases, through badly conduct- 

 ing liquids and through metals, electric 

 conduction is a process which requires a 

 very considerable space and aggregations 

 of large numbers of molecules. I have not 

 been able to find any direct experimental 

 evidence as to whether the same is true for 

 electrolytes. Experiments on the resist- 

 ance of thin layers of electrolytes would be 

 of considerable interest, as according to one 

 widely accepted view of electrolysis con- 

 duction through electrolytes, so far from 

 being effected by aggregations of molecules, 

 takes places by means of the ion, a structure 

 simpler than that of the molecule, so that 

 if this represents the process of conduction, 

 there would not seem room for the occur- 

 rence of an effect which occurs with every 

 other kind of conduction. 



In this building it is only fitting that 

 some reference should be made to the ques- 

 tion of the movement of the ether. You 

 are all doubtless acquainted with the 

 heroic attempts made by Prof. Lodge to 

 set the ether in motion, and how suc- 

 cessfully the ether resisted them. It seems 

 to be conclusively proved that a solid body 

 in motion does not set in motion the ether 

 at an appreciable distance outside it; so 

 that if the ether is disturbed at all in such 

 a case, the disturbance is not comparable 

 with that produced by a solid moving 

 through an incompressible fluid, but must 

 be more analogous to that which would be 

 produced by the motion through the liquid 

 of a body of very open structure, such as 

 a piece of wire netting, where the motion of 

 the fluid only extends to a distance compar- 

 able with the diameter of the wire, and not 



