

[ 198 ] 



XV. b Rays. 

 To the Editors of the Philosophical Magazine. 

 Gentlemen, — 



IN a paper read before the Royal Society *, I described 

 some experiments whose object was to test whether an 

 electron ejected b}^ an a particle has a component of motion in 

 the direction in which the a particle is moving. The experiment 

 consisted in measuring the ionization produced by collision 

 between an aluminium plate and a parallel sheet of aluminium 

 foil through which a particles were projected on to the 

 aluminium plate. It was found that the "collision" leak 

 was always greater when the foil was at a negative potential 

 in comparison with the plate than when it was at a positive 

 potential. From these experiments I concluded that the 

 ejected electrons have a component of motion of the nature 

 aforesaid. 



Dr. Campbell has recently described some experiments in 

 several papers to the Philosophical Magazine t, in which. 

 using a similar arrangement, he measured the number 

 of electrons given off from the metal surfaces where the 

 a particles emerge, and where they enter (no ions being 

 produced by collision), and found no difference in the 

 numbers in the two cases. He therefore concluded that the 

 ejected electron has on the average no component of motion 

 in the direction of that of the ionizing a particle. 



In my paper I stated that probably the gas itself plays 

 directly only a small part in the production of this asymmetry. 

 This is, however, not the case, as I will now show. In order 

 to explain what happens in the gas I must refer to a previous 

 paper J. It is there pointed out that the nature of an ion 

 cluster in a gas at constant temperature is not constant, as is 

 usually supposed, but is constantly going through a cycle of 

 changes. These changes are similar in character to the 

 chemical changes going on in a partly dissociated gas, and 

 are regulated by the same laws, namely, the laws of thermo- 

 dynamics and the law of mass action. The free ion and 

 every ion cluster has, therefore, an average period of life 

 before becoming a cluster of a different nature. This pro- 

 perty of the ion must play an important part in the production 

 of ions by collision. Thus, consider a number of ions in a 

 gas not under an electric field. At any instant there are a 



* Proc. Roy. £oc. A. vol. lxxxiii. p. 195 (1910). 



t Aug. 1911, p. 276 ; Jan. 1912, p. 46 ; April 1912, p. 463. 



X Proc. Camb. Phil. Soc. xvi. pt. iv. p. 285. 



