Cliarghnj thvoiigh Ion Ahsorpllon. 505 



throiioh a nuinhcr of narroNV tub(\-i, so that if any ions pass 

 out, as Townseiurs experiments have shown, there is bound 

 to be an excess of positive ones. The material particles are 

 thus left behind with a nep^ative charge. That this is the 

 })rocess by which the charge is obtained, and not that the sand 

 and smoke leave the place of ionization ne^ratively charged, 

 was proved by the fact that it \vas possible to collect the 

 positive ions which, streaming away from the sand, left it 

 negatively charged. To do this a cylinder of metal was so 

 placed that the sand passed through it immediately on leaving 

 the riame. The positive ions diffusing out of the stream of 

 sand were caught by the cylinder, which at once showed a 

 positive charge. No charging of the cylinder could be 

 detected when the sand passed through it without first passing 

 through the flame. 



In all the experiments so far described, we have obtained a 

 charge as the result of ion absorption ; but it has in every 

 case been accompanied by the removal of air from the action 

 of a strong ionizer to a place where diffusion could come into 

 play. In the atmosi)here such a removal does not take place, 

 for we have strong reasons to believe that the natural ionization 

 in the atmosphere is mainly due to a radioactive gas which 

 the air carries with it ; therefore none of the cases described 

 above can be directly applied to the problems of atmospheric 

 electricity. Even the charging of the metals in Zeleny^s 

 experiments, which has been used as the foimdation of Elster 

 and GeiteFs theory, has no parallel in the atmosphere, for in 

 Zeleny's experiments the air was first removed from the 

 ionizer, then, as a result of diffusion, the charge appeared. 

 Zelenv's experiments do not prove that an insulated con- 

 ductor in air which is constanth/ and uniformli/ ionized will 

 become negatively charged. Experiments which will now 

 be described show that such an insulated conductor does not 

 become negatively charged. 



When working with a quadrant electrometer it was 

 necessary to lead a long copper w4re to it from apparatus a 

 considerable distance away. In order to protect this wire 

 from electrostatic induction it was led through a long wooden 

 box covered on the outside with earth-connected tinfoil. It 

 was then found that the wire although insulated became 

 positively charged when left for any considerable time to itself. 

 In order to see whether this charge had anything to do with 

 the material of the wire, the copper one was replaced by one 

 of zinced iron; this latter became negatively charged. To 

 investigate the effect more thoroughly, a large cage, one 

 and a half metres long and one metre high and ^\ide, was 



