152 



Mr. C, T. E. Wilson. 



always present requiring, in order that water should condense upon 

 them, exactly the same degree of supersaturation as the nuclei pro- 

 duced in enormously greater numbers by Rontgen rays; and I con- 

 cluded that they are identical with them in nature and that they are 

 probably ions.* While, however, later experiments proved that the 

 nuclei formed by Rontgen or uranium rays can be removed by an 

 electric field and are therefore ions, similar experiments made with the 

 nuclei which occur in the absence of ionising radiation led to negative 

 results.! In the light of facts brought forward in the present paper I 

 should now feel disposed to attribute the negative character of the 

 results in the latter case to the small number of nuclei present.! 



Subsequently to the publication of the work on the behaviour of 

 ions as condensation nuclei, Elster and Geitel showed that an electri- 

 fied conductor exposed in the open air or in a room lost its charge by 

 leakage through the air ; and that the facts concerning this conduction 

 of electricity through the air are most readily explained on the suppo- 

 sition that positively and negatively charged ions are present in the 

 atmosphere. The question where and how these ions are produced 

 remained, however, undetermined ; it would therefore be incorrect to 

 assume their properties, and in particular their behaviour as condensa- 

 tion nuclei, to be necessarily identical with those of freshly produced 

 ions ; the carriers of the charge might consist of much more consider- 

 able aggregates of matter than those attached to the ions with which 

 the condensation expeiiments had been concerned. Moreover, so long 

 as the source and conditions of production of these ions remained 

 undetermined, one could not assume their presence in the regions of 

 the atmosphere where supersaturation might be expected to occur. 



Before going further afield in search of possible sources of ionisation 

 of the atmospheric air, it seemed advisable to make further attempts to 

 determine whether a certain degree of ionisation might not be a 

 normal property of air, in spite of the somewhat ambiguous results 

 given by the condensation experiments to which I have referred. 



After much time had been spent in attempts to devise some satis- 

 factory method of obtaining a continuous production of drops from the 

 supersaturated condition, I abandoned the condensation method, and 

 resolved to try the purely electrical method of detecting ionisation. 

 Attacked from this side the problem resolves itself into the question, 

 Does an insulated-charged conductor suspended within a closed vessel 

 containing dust-free air lose its charge otherwise than through its 

 supports, when its potential is well below that required to cause 

 luminous discharges 1 



* ' Camb. Phil. Soc. Proc.,' vol. 9, p. 337. 

 f 'Phil. Trans.,' A, vol. 193, pp. 259-308. 



X The similar results obtained with nuclei produced in air exposed to ultra- 

 violet light require, however, some other explanation. 



