THE SPECTRUM OF HYDROGEN 251 



behave in the opposite fashion in a hehum tube. After the 

 condenser is removed they show only in the centre of the 

 capillary, but gradually spread outwards. Oxygen, mercury, 

 and sulphur lines behave in this way. These experiments point 

 to an actual separation of the elements, caused by the condensed 

 discharge, hydrogen being " blown out " of the capillary, and 

 the other elements collected into it. These phenomena are 

 very probably connected with those observed by Wood. The 

 current densities he employed in his long tubes were very high, 

 and one effect of introducing a condenser into a circuit is to 

 increase the current density while the discharge is actually 

 passing. If the densities he used were sufficient to drive the 

 heavier atoms of the impurities to the centre of the capillary, 

 they would there " catalyse " the production of the Balmer 

 series, in virtue of the effect described earlier in the article, and 

 the ends of the capillary would contain hydrogen sufficiently 

 pure to exhibit a strong secondary spectrum. As the tube was 

 run, the impurities would gradually disappear, and the Balmer 

 lines would be weakened throughout the tube. No explanation 

 of these observations has yet been given, and there is a clear 

 field for further experimental work. 



