﻿540 Prof. R. W. Wood on Atomic Hydrogen 



attack appeared to be a study of the " Infected spots " 

 which I spoke of in the earlier papers. These are portions 

 of the tube in which the discharge appears white or pink, in 

 contrast to the fiery purple exhibited by the remainder of 

 the tube. The spectrum of these spots shows the Balmer 

 series only to the 12th or 14th member, and a fairly strong- 

 secondary spectrum. I suspected that they were due to a 

 contamination of the wall of the tube ; and if the cause 

 could be found, it might be possible to improve in some way 

 the condition of the rest of the tube, and so obtain a more 

 complete series of Balmer lines. 



The clue was obtained, as I have shown in a recent paper 

 in the Proc. Roy. Soc, by the accidental entrance of a 

 speck of sealing-wax into the discharge tube. This was 

 speedily changed to a spot of stannous oxide by the heat 

 of the discharge, and examination of the spot with a lens 

 showed that it was covered with minute globules of metallic 

 tin. 



A section of tube was now fine-ground on the inside with 

 carborundum, and this ground portion made a part of the 

 long hydrogen tube • the portions to the right and left of 

 the ground portion were purple, and showed the pure Balmer 

 spectrum, while in the ground portion the discharge was 

 white, and showed a very strong secondary spectrum. 



The ground-glass tube was tried as a result of an experi- 

 ment with a tube of unglazed porcelain, inserted at the 

 middle of the tube, which was then bent at two right-angles, 

 so that the discharge in the porcelain tube could be viewed 

 or photographed " end-on/' The porcelain tube gave only 

 secondary spectrum with the first four or five Balmer lines. 



A tungsten wire was then inserted in the tube at a spot 

 which showed the pure Balmer spectrum. The wire was 

 raised to a white heat by the discharge, though a fine thread 

 of soft glass inserted in the same way was not even softened. 

 In the vicinity of the wire the secondary spectrum came out 

 strong, and further experimenting showed that the secondary 

 spectrum appeared a second or so before the wire became 

 incandescent. Addition of oxygen to the hydrogen suppressed 

 the heating of the wire. 



As I have shown in the paper previously alluded to, the 

 action of the wall of the tube at an infected spot, the speck 

 of stannous oxide, the tungsten wire, and the ground-glass 

 surface appears to be a catalytic one, these surfaces causing 

 a recombination of the atomic hydrogen, thus furnishing 

 molecular hydrogen at a rapid rate, the breakdown of which 

 bv the current causing the secondary spectrum. 



