﻿and the Balmer Series Spectrum. 541 



In all of the recent work, tubes of pyrex glass, carefully 

 cleaned with hot chromic acid, have been used. The tubes 

 usually get into good condition after a few minutes' operation, 

 and show less luminosity through a green ray-filter (which 

 is opaque to the Balmer lines) than the tubes of soft glass 

 used in the earlier work. 



One of the most discussed problems in spectroscopy is 

 why the presence of water vapour in the hydrogen enhances 

 the Balmer series and suppresses to a great degree the 

 secondary spectrum. As I showed in the earlier papers, 

 if dry hydrogen is employed in the long tube, fed in at 

 intervals through a palladium tube and pumped out with the 

 tube in operation, the discharge eventually becomes white, 

 and all of the Balmer lines disappear except H a , which is 

 so faint that it appears of a dull brick-red colour, in contrast 

 with the secondary spectrum. In view of what we now 

 know, it appears as if the glass wall of the tube, when 

 thoroughly freed from adsorbed water vapour or oxygen, 

 acts as a powerful catalyzer of the atomic hydrogen, which 

 never reaches a sufficient concentration to cause the Balmer 

 spectrum to appear. 



Dr. Irving Langmuir, with whom I discussed these 

 results, made the very valuable suggestion that the glass 

 surface might be " poisoned " by the oxygen. He has 

 found, in the course of an extended study of the production 

 of atomic hydrogen by an incandescent tungsten wire, that 

 the presence of small traces of oxygen prevented the 

 formation of atomic hydrogen by " poisoning " the cata- 

 lyzing surface of the metal. This being the case, the 

 oxygen must also render the surface of the (comparatively) 

 cold tungsten incapable of bringing about the recombination 

 of the atomic hydrogen, as had been found to be the case. 



This makes it appear extremely probable that the part 

 played by water vapour in bringing out the Balmer series 

 is merely that of supplying a " poison " (oxygen) for the 

 catalyzing wall of tube, thus permitting a high concentration 

 of atomic hydrogen in the tube under the action of the 

 heavy discharge. With a feeble discharge, as I showed in 

 the earlier paper, the secondary spectrum predominates and 

 the Balmer lines are weak. This is probably due to the fact 

 that the atomic hydrogen is not formed fast enough to get 

 ahead of the catalyzing power of the tube wall. 



If dry hydrogen is admitted through palladium, and the 

 tube brought to the white stage by long operation, it is 

 found that if a condenser is put in parallel with the tube, 

 the discharge becomes red and the Balmer lines appear. 



