ON CHEMICAL COMHIXATIOX AXD ELECTRIC DISCHARGE. 187 



or would form excellent nuclei for the formation of loosely attached systems 

 of molecules suitable for conveying the electric discharge. A quantity of 

 water vapour alternately condensing and evaporating would throw the 

 surrounding gas into a state in which it could transmit the electric dis- 

 charge. This would be the case even though the existence of the drops 

 was so transitory that the water was to all appearance continually in the 

 form of vapour. That some drops should be formed, having a very brief 

 existence it is true, even when the pressure of the water vapour is far 

 below that of saturation, is what we should expect on almost any dynamical 

 theory of evaporation. The effect of the drops in promoting the existence 

 of free atoms would of itself tend to increase the likelihood of their 

 formation, as we know from the experiments of R. v. Helmholtz and 

 E-icharz that condensation is much facilitated by the presence of free ions. 



Thus, a small quantity of aqueous vapour, portions of which, occasionally 

 condensed for an instant into drops, might put some of the gas into a con- 

 dition approaching dissociation, inasmuch as in this condition the atoms 

 could easily be detached from each other. We might expect that when in 

 this state the gas might enter into chemical combinations much more freely 

 than before this dissociation took place. The effect of a drop of water on 

 a molecule consisting of a positively electrified atom combined with a nega- 

 tively electrified one near the surface of the drop is worthy of considera- 

 tion. Pure water, though not a conductor, has yet such a high specific 

 inductive capacity that its electrostatic effect is much the same as if it 

 were a conductor. Now if we have a molecule consisting of a positively 

 electrified atom. A, with a charge -f e and a negatively electrified atom, B, 

 with a charge — e, on or near the surface of a conducting sphere, the effect 

 of the electrification induced on the sphere is the same as if we had at 

 A' a point on the same radius as A inside the sphere, and close to the sur- 

 face a negative charge equal to — e, and at B' a point inside the sphere 

 close to the surface and on the same radius as B a positive charge equal 

 to +e- Now, the negative charge at A' will almost neutralise at B the 

 force arising from the positive charge at A, while the effect of the charge 

 at B' on B is along the radius at B, and thus has no effect parallel to the 

 surface of the sphere ; there would thus be nothing to oppose the motion 

 of B away from A, and similarly of A away from B, as long as this motion 

 took place parallel to the surface of the sphere. Thus, for movements of 

 this kind the molecule would be dissociated by the presence of the sphere ; 

 this dissociation of the gas would facilitate the production of the elec- 

 trification over its surface, also its tendency to enter into chemical com- 

 bination. 



We have hitherto been considering the passage of electricity through 

 the gas. If we proceed to study the passage of electricity from a gas to a 

 metal, we shall find that it is also facilitated by the presence of a third 

 substance. This effect was very well shown in an experiment which I had 

 occasion to make on the discharge of electricity through mercury vapour. 

 A bulb for the discharge was prepared by taking a large closed vessel filled 

 with freshly distilled mercury ; this was connected by a glass tube to the 

 discharge bulb ; one end of a capillary tube was fused to this bu'b, tne other 

 end dipped under mercury. When the mercury in the large vessel was 

 heated, the arrangement acted like a mercury distiller and pumped itself : 

 to hasten matters, however, as much air as possible was taken out of the 

 arrangement by an air-pump before beginning the distillation. The ap- 

 paratus was kept distilling for a whole day, the discharge tube being sur- 



