430 Mr. J. J. Thomson on a Theory of 



lating the electromotive force necessary to produce discharge 

 from the pressure when this is less than that giving the mini, 

 mum electrical strength, viz. Y = '67/\/p (where V is the dif- 

 ference of potential per centimetre, and p the pressure in 

 millimetres of mercury), I find that at a pressure of about 

 •0001 of a millimetre of mercury the electric tension just 

 before discharge would equal the pressure of the gas; so that 

 it is at pressures comparable with this that the experiment 

 ought to be tried. 



Let us now pass on to the case where the intensity of the 

 electric field is so great that the dielectric can no longer insu- 

 late, and the electricity is discharged. 



It will be instructive to consider for a moment what hap- 

 pens when a compound gas is raised to such a temperature 

 that it is dissociated, or an elementary one until its molecules 

 are split up into atoms. If the gas is at a low temperature, 

 say 0° C, when heat is first applied, so far as we can tell the 

 whole of the heat is employed in raising the temperature and 

 increasing the radiation, and no heat is rendered latent; in 

 other words, the alteration in the molecular structure of the 

 gas absorbs no work. This state of things continues until we 

 approach the temperature at which the gas begins to be dis- 

 sociated ; then a large fraction of the heat supplied to the gas 

 is used up in altering the molecular structure, and only a part 

 of it is spent in raising the temperature and increasing the 

 radiation. 



If we look on this from the point of view of chemical com- 

 bination which we took before, we may regard it as showing 

 that, if any energy be supplied to the gas when the ratio of the 

 paired to the free time is so large that the gas exhibits none 

 of the phenomena of dissociation, the consequent diminu- 

 tion in the ratio of the paired to the free time does not 

 absorb any of the energy; but if the ratio of the paired to the 

 free time be so small that the gas exhibits some of the phe- 

 nomena of dissociation, then a diminution in the ratio of 

 the paired to the free time will absorb a considerable amount 

 of energy. The same statements will apply to an elementary 

 gas, except that in this case the change of structure consists 

 in splitting the molecules up into atoms of the same kind, 

 while in the compound gas they were split up into atoms of 

 different kinds. 



Let us now apply these considerations to the case of the 

 electric discharge. The disturbance to which the gas in an 

 electric field is subjected makes the molecules break up sooner 

 into atoms than they otherwise would do, and thus diminishes 

 the ratio of the paired to the free times of the atoms of the 

 gas; as the intensity of the electric field increases, the dis- 



