242 Professor J. J. Thomson [April 13, 



electrode connected to an electrometer, passed into the bulb, and was 

 arranged so that it could be moved about from one part of it to the 

 other. When the electrodes were metal and the bulb was filled with 

 the electro-negative gas oxygen, the electrode received a positive 

 charge in whatever part of the bulb it was situated ; if now the bulb 

 was filled with hydrogen at atmospheric pressure, then in the regions 

 remote from the arc the electrode received a positive charge, but in 

 the immediate neighbourhood of the arc itself it received a negative 

 charge. When the pressure was reduced the region in which the 

 charge was negative contracted, and finally at pressures about one- 

 third of an atmosphere, seemed to disappear, and the electrode got a 

 slight positive charge in whatever position it was placed. If now, 

 instead of using metallic electrodes we use well-oxidised copper ones, 

 and repeat the experiment in hydrogen, working at a pressure when 

 there was only positive electricity when the electrodes were bright 

 and polished, we find that with the oxidised electrodes every particle 

 of positive electricity is taken out of the tube, and a negative charge 

 is left. This negative charge remains until the copper oxide is 

 completely reduced ; when this occurs the negative charge disap- 

 pears, and is replaced by positive. Thus, under the same conditions 

 as to the nature of the gas and the pressure, the bright copper elec- 

 trodes leave a positive charge in the gas, while the oxidised ones 

 leave a negative charge. 



The most probable explanation of these results seems to me to 

 be the view that the communication of electricity from gas to the 

 electrode, or from the electrode to gas, is facilitated by the tem- 

 porary formation of something of the nature of a chemical compound 

 between the gas and the metal. In all such compounds the metal is 

 the electro-positive element, and has the positive charge, the gas 

 being the electro-negative and carrying the negative charge. Now 

 consider the case when the negative charge is on the gas, and the 

 positive charge on the metal ; then the gas and metal have got the 

 charges proper to them in any compound they may form, and are 

 thus in a fit state to combine, or, according to this view, allow the 

 negative electricity to pass from the gas to the copper. But, now, 

 suppose the gas was positively electrified, the gas and the metal have 

 now opposite charges to those proper to them in a compound, and 

 before the union of gas and metal in this state could result in 

 anything but a most unstable compound, an additional process must 

 be gone through — i. e. the charges on the gas and metal must be 

 interchanged. Thus the conditions for the combination of the gas 

 and metal are more complex when the gas is positively electrified 

 than when it is negatively electrified, and thus, on the view that the 

 communication of electricity between the gas and the metal involves 

 a sort of chemical combination, we see that the negative electricity 

 will escape more easily from the gas to the metal than the positive. 

 Now consider the case when the gas was hydrogen, the electrodes 

 oxidised copper ; the hydrogen combines now not with the metal, but 



