228 CONSTITUTION OF ELECTRONS [APR M. 



uniform convection of a system at any speed however 

 high should show no internal influence, provided the con- 

 stitution of matter is wholly electric, i.e. provided atoms 

 are active only through the aether. The very important 

 (because purely electrical) experiment of Trouton and 

 Noble is a case in confirmation. So is the absence 

 of influence on either of the two phenomena, magnetic 

 rotation and double refraction (Rayleigh and Brace). 



It is true that this way of explaining the absence 

 of any second order aberration-effect was received with 

 scepticism by Poincare' (Paris Congress, 1900), whose criti- 

 cism was that if in future the third order effect also 

 needed annulment some new artificial arrangement could 

 still be added on, to do even that. The result of the 

 argument has now, however, gained the independent 

 support of Lorentz (1904), whose investigation has sug- 

 gested, what in fact is easily verified, that Larmor's 

 work proves to be exact: his restriction to the second 

 order being unnecessary. He was hardly concerned, how- 

 ever, to go further, since the hypothesis of the infinitesimal 

 effective size of the electron already limited the investi- 

 gation, nor does it seem that Lorentz's work really carries 

 the matter further. Larmor's attitude has been all along 

 (Phil. Mag., June, 1904), that this provisional hypothesis 

 holds the eld until some effect of uniform convection 

 presents itself : every new negative result is a steady 

 corroboration of it: but like any other physical repre- 

 sentation it must ultimately reach the limits of its appli- 

 cation. 



Only negative electrons are known in the free state. 

 It remains unsettled whether this is due to some one- 

 sidedness in the experimental means hitherto employed, 

 or whether physical nature is in reality intrinsically un- 

 symmetrical. If the positive electron were a region of 



