378 On the Discharge of Electricity through Gases. [June 16, 



perly insulated, we easily get a current through the galvanometer 

 which is nothing but a branch current from, the main discharge. Such 

 a leakage current, even when it is weak, considerably increases the 

 effects described in this paper. 



These experiments show conclusively that there is nothing peculiar 

 in the gaseous state of a body to prevent any electromotive force how- 

 ever small from producing a current. If a finite electromotive force is 

 required under ordinary circumstances the fact cannot be accounted 

 for, as Edlund and others have attempted to do, by a special surface 

 resistance which has to be overcome by a finite difference of potential 

 at the surface. 



I think the facts are very well accounted for by the theory which I 

 have proposed in my last paper. If the two atoms of a gas making 

 up the molecule are charged by opposite electricities, but are held 

 together in addition by molecular forces, a finite force is required to 

 overcome the latter. But as soon as that force is overcome and the 

 atoms themselves are set free to diffuse and constitute a current, these 

 atoms will be able to follow any electromotive force which we may 

 apply. If, then, we have auxiliary electrodes, these electrodes will 

 establish their electric field which we can never screen off completely 

 from any other part of the vessel except by closed surfaces. The 

 atoms, with their positive and negative charges, will diffuse across to 

 the auxiliary electrodes and give off their electricity to them. No 

 finite difference of potential is required in the auxiliary electrodes, 

 because even if there is work done in making an atom interchange its 

 positive for negative electricity, that work is undone again at the 

 other pole, where atoms of a similar kind interchange negative for 

 positive electricity. 



This I believe to be the general explanation of the phenomena de- 

 scribed in this paper. In order to account for the peculiar difference 

 between positive and negative electricity which appears in the experi- 

 ments done at atmospheric pressure, also those mentioned under 5 

 (p. 377), we must make some further supposition. I have already 

 mentioned in my last paper that, according to the theory I have pro- 

 posed, we must imagine the molecules to be broken up at the negative 

 pole, and I believe that this fact will ultimately be found to account 

 for this apparently unsymmetrical property of the two electricities ; 

 but I should like to strengthen my case by further experiment before 

 going into details on this point. 



I should like, in conclusion, to point out an important application of 

 these results. I have last year obtained by calculation results which 

 seem to show that the principal cause of the diurnal variation of ter- 

 restrial magnetism is to be looked for in the upper regions of the 

 atmosphere. Professor Balfour Stewart at various times suggested 

 that the air currents in these regions may, owing to the lines of 



