﻿and the Balmer Series Spectrum. 543 



tube near the electrode bulbs. In one tube of pyrex glass 

 which had been very carefully cleaned with chromic acid, 

 with especially clean bright electrodes, it was found that the 

 white discharge extended to a distance of only 3 or 4 cm. 

 from the bulbs, while in the earlier tubes it often reached 

 to a distance of 20 or 30 cm. 



In the earlier papers I have drawn attention to the circum- 

 stance that if the hydrogen tube is brought to the white 

 stage, and then highly exhausted, with the current shut off, 

 if a small amount of air or nitrogen is admitted, the discharge 

 is of a most beautiful golden-yellow colour, resembling the 

 discharge in pure helium. Photographs of the spectrum 

 showed that the second positive spectrum (violet and ultra- 

 violet bands) was nearly absent, the yellow colour being due 

 to the first positive spectrum, consisting of red, yellow, and 

 green bands. 



Applying the catalysis theory to this result, the indications 

 are that the first positive spectrum is due to the nitrogen 

 molecule, the second to the atom. It was found that the 

 yellow discharge was obtained only if a very small amount 

 of air was admitted, doubtless due to the fact that if too much 

 air was introduced there was enough oxygen to poison the 

 avails of the tube. I have not yet been able to get the second 

 positive spectrum free from the first, but no very great amount 

 of work has been done in this direction. Possibly by intro- 

 ducing an excess of oxygen it can be accomplished. 



It was found that the tungsten wire was not heated to 

 visible luminosity in the discharge in air, while a platinum 

 wire of the same size was raised to a white heat. Platinum 

 therefore seems to be a catalytic agent for atomic nitrogen, 

 while tungsten is inoperative. I have not yet tried platinum 

 in pure nitrogen, and it may be that the oxygen plays a part 

 in the surface reaction which heats the platinum. 



It seems to be now clear why a more complete series of 

 Balmer lines is obtained in the solar corona and probably 

 in nebulae (provided sufficient exposure were given) than in 

 vacuum tubes. The luminous gases are in these cases not 

 in proximity to catalyzing surfaces, and consequently atomic 

 hydrogen of 100 per cent, concentration can exist. Of 

 course the possibility of a different type of excitation still 

 remains. 



To further extend the series in the laboratory, it will 

 probably be necessary to devise a method of more com- 

 pletely poisoning the walls of the tube, or abolish the wall 

 entirely, as can be done perhaps by means of a very powerful 

 discharge of the ring type excited by high-frequency in- 

 ductive effects, in an electrodeless tube. 



