574 Prof. J. J. Thomson on 



phosphorescence are visible. This is, 1 think, an important 

 matter in connexion with the interpretation of these results ; 

 for at these low pressures we have to deal not only with the 

 gas with which the tube was originally filled, but also with 

 the gas which is given off by the electrodes and the walls of 

 the tube during the discharge : and it might be urged that at 

 these low pressures the tube contained nothing but hydrogen 

 given out by the electrodes. I do not think this explanation 

 is feasible, for the following reasons : — 



(1) The gas developed during the discharge is not wholly 

 hydrogen : if the discharge is kept passing long enough to 

 develop so much gas that the discharge through the gas 

 is sufficiently luminous to be observed by a spectroscope, the 

 spectrum always showed, in addition to the hydrogen lines, 

 the nitrogen bands ; indeed, the latter were generally the 

 most conspicuous part of the spectrum. If the phosphorescent 

 screen on w 7 hich the positive rays impinge is observed during 

 the time this gas is being given off, the changes whieh take 

 place in the appearance of the screen are as follows : — If, to 

 begin with, the pressure is so slow that the phosphorescent 

 patches are reduced to two bright spots, then, as the pressure 

 begins to go up owing to the evolution of the gas, the deflexion 

 of the spots increases. This is owing to the reduction in the 

 velocity of the rays consequent upon the reduction of the 

 potential difference between the terminals of the tube, as 

 at this stage an increase in the pressure facilitates the passage 

 of the discharge. In addition to the increase in the displace- 

 ment, there is an increase in the area of the spots giving a 

 greater range of values of e/m ; this is owing to the increase 

 in the number of collisions made by the particles in the rays 

 on their way to the screen. As more and more gas is 

 evolved, the patches get larger and finally overlap ; the 

 existence of the second patch being indicated by a diminution 

 in the brightness of the phosphorescence at places outside its 

 boundary. As the pressure increases the luminosity gets 

 more and more continuous, and we finally get to the con- 

 tinuous band as shown in fig. 6. At this stage it is 

 probable that there may be enough luminosity to give a 

 spectrum showing the nitrogen lines, indicating that a con- 

 siderable part of the gas in the tube is air. It is especially 

 to be noted that during this process, when gas was coming 

 into the tube, there has been no development of patches 

 in the phosphorescence indicating the presence of new rays ; 

 on the contrary, one type of carrier — that corresponding to 

 e/m = 5x 10 3 — has disappeared. The presence of the nitrogen 

 bands in the spectrum shows that nitrogen is carrying part 

 of the discharge, and yet there are no rays characteristic of 



