402 Prof. Thomson, On some experiments on the [May 10, 



narrowed so that its dimensions in one direction are comparable 

 with the space it takes to reverse the direction of motion of a 

 molecule, then the gas will be electrically stronger in that direction 

 than in one at right angles to it. This will have the effect of 

 confining the discharge to one direction, and so making it more 

 intense in that direction than it would have been if it had been 

 free to spread out equally in all directions : this perhaps will 

 explain Goldstein's result that the narrowing of the tube produces 

 much the same effect as would be produced if there were a 

 secondary negative electrode at the contraction. 



If this theory of the effect of the limitation of the space 

 taken to reverse the direction of motion of a molecule be true, 

 we should expect the following results, which as far as I know 

 have not been investigated. 



The ratio of the electric strength of a thin layer to a thick one 

 should be greater at low pressures than at high ones, because the 

 ratio of the average distance the molecules travel along the lines 

 of force before their direction of motion is reversed to the dis- 

 tance between the plates is greater at low pressures than at high 

 ones. 



The difference between the electric strength of a thin and a 

 thick layer will depend upon the intensity of the force between 

 the molecules, because the greater the intensity of the force the 

 less the space required by a molecule to reverse its direction ; 

 but the smaller the force between the molecules, the greater the 

 coefficient of viscosity, so that the difference between a thick and 

 a thin layer ought to depend upon the coefficient of viscosity of 

 the gas. 



The motion of the gas itself produced by the electric discharge 

 must produce important effects on the discharge. When the 

 electric discharge passes through a gas, the pressure of the gas 

 increases, this increase however only lasts as long as the spark 

 passes. This was originally observed by De la Rive, it was after- 

 wards re-discovered by De la Rue and Miiller, when investigating 

 the discharge produced by their large chloride of silver battery ; 

 by keeping the current on they were able to get an increase of 

 pressure in a large bell-jar of about 30 per cent, of the original 

 pressure (Phil. Trans. 1880, p. 86). This increase has been at- 

 tributed by Dr Schuster (Proceedings of the Royal Society, 1884), 

 to the decomposition of the molecules of the gas by the discharge, 

 but though I believe some of the increase to have been due to 

 this cause, yet the effect seems to be too large to be altogether 

 explained in this way, for if the molecules were split in two it 

 would require about 50 per cent, of the molecules to be dissociated, 



