882 Prof. J. S. Townsend and Mr. V. A. Bailey on 



a final state o£ equilibrium is obtained in which the pro- 

 portion of ions to electrons is constant. In this case W 

 and k would be functions of Z'j) alone. A stream of this 

 kind may be distinguished from a stream comprising only 

 free electrons, by the effects obtained with magnetic forces*. 



In all cases the proportion of ions to electrons in a gas at 

 constant pressure diminishes as the electric force increases. 

 This holds even in the case where water vapour is present. 

 With the forces and pressures used in these experiments 

 there does not appear to be any indication of the presence 

 of ions in nitrogen or hydrogen. 



The following experiments indicated in a direct manner 

 the presence of ions with electrons in a stream in oxygen. 

 The ratio of the charges received by the electrodes E 2 and 

 E 3 to that received by E x in the apparatus fig. 1, was deter- 

 mined while large magnetic forces were acting which tended 

 to deflect the stream towards the electrode E x . The greatest 

 force which it was possible to use ^ould have been too small 

 to produce an appreciable deflexion of particles having- 

 masses of the same order as the mass of a molecule. 



"When the magnetic force was increased to the maximum 

 available, the charges n 2 -f n 5 were about one per cent, of the 

 charge n^ in the experiments with nitrogen and air, but with 

 oxygen a much higher proportion remained on the electrodes 

 E 2 and E 3 under similar conditions. 



It had previously been observed that electrons tended to 

 form ions in moving through imperfectly dried gases, the 

 effect being particularly marked with the smaller values of 

 Z/p. This, however, does not seem to be the cause of the 

 presence of ions in the oxygen used in these experiments. 



The oxygen was obtained by the electrolysis of barium 

 hydrate, and was dried in a flask with phosphorus pentoxide. 

 Practically the same results were obtained with two speci- 

 mens of the gas, one of which had been in the drying flask 

 for seven days and the other for fourteen days. It is 

 therefore improbable that the effect was due to water vapour. 

 Further experiments are being made on this point and will 

 be reserved for a future communication. 



The velocity W" and the quantity k obtained in these 

 experiments with oxygen at the lower pressures and the 

 higher forces are probably not much different from the exact 

 values for streams of electrons, and these will be used for 

 the present in order to compare oxygen with nitrogen and 

 hydrogen. 



* J. S. Townsend andH. T. Tizard, Proc. Roy. Soc. A. lxxxvii. (1912). 



