CORRELATION OF PHENOMENA 69 



from these would be sufficient to enable them to 

 preserve their identity for a considerable time.^ 



Granting, for the present, that they do so with- 

 stand the presence of surrounding molecules, how 

 then is their formation to be accounted for ? We 

 may, I think, attribute it to the radiation which is 

 set up when the electrons or corpuscles are stopped 

 by collision with molecules. These " pulses " would 

 behave like ultra-violet light, which, as we know 

 from the experiments of C. T. E. Wilson, can 

 produce such condensations ; whilst the intensity 

 of the radiation within molecular distances should 

 be sufficient to bring these aggregations we are in 

 search of into being. What is more, the influence 

 of a spark on particles of dust is well known 

 as the readiest method of bringing down such 

 particles. And what is this due to, but to the 

 formation of great aggregates of dust, by electrical 

 cohesion, much in the same fashion as we have 

 supposed the molecules themselves to do ? 



The experiments of Hertz on the pressure due to 

 the discharge itself, indicate a great increase of that 

 pressure in the luminous column close to the positive 

 electrode. And is not this what we should expect 

 from the condensations in this region ? 



It seems, therefore, that the phenomena of light, 

 produced directly by the discharge of electricity 

 through gases, can be brought into harmony with the 



1 This may be inferred from Poynting's calculation of the 

 repulsion between radiating spheres the size of molecules. 

 Phil. Trans., 1904. 



