240 Sir J. J. Thomson : Further 



parent, and the process would go on evidently leading to 

 instability. It* even an appreciable fraction of the cathode 

 particles were ejected from the cathode by this method, we 

 should expect to find in the cathode stream sets of cathode 

 particles, a u a 2 , a z — moving more rapidly than the main body 

 of these particles, and such that the energy possessed by an 

 individual particle in the set a 3 exceeded that possessed by 

 one in a { by the energy acquired by a cathode particle in 

 falling through the dark space, while the energy of one in 

 set a 3 exceeded that of one in a 3 by the same amount, though 

 the number of particles in « 2 would be much smaller than 

 that in <x u and the number in a s still smaller than in a 2 , and 

 so on. As we have no evidence of the existence of a spectrum 

 of this kind in the cathode rays, I think we must conclude 

 that practically the whole of these rays are produced by the 

 positive ions. 



The smallness of the electric force at the end of the dark 

 space is responsible, I think, for the luminosity in the negative 

 glow, for when the field is strong, as in the dark space, the 

 corpuscles and positively charged atoms and molecules acquire 

 such great velocities that they rush past each other without 

 combining. When, however, this field vanishes, positive and 

 negative particles may only have small relative velocities and 

 combine and give out light in the process. It is in the 

 negative glow that we must look for new compounds formed 

 by the union of atoms dissociated by the discharge. 



Chemical Compounds produced by the Electric Discharge in 

 Gases at very low Pressures. — The study of the positive rays 

 has revealed the existence of some compounds which have 

 not hitherto been detected by chemical methods. The mea- 

 surement of mje for the lines on the plate gives us the 

 molecular weight of the carrier producing the line, or of some 

 submultiple of it if the carrier is charged with more than one 

 unit of electricity. In some cases this leads to the identifi- 

 cation of the line without difficulty ; in other cases more than 

 one kind of carrier might give the right molecular weight : 

 for example, if we found a carrier having a molecular weight 

 28 it might be N 2 , CO, or C 2 H 4 : in such cases further 

 experiments are necessary. If it was found, for example, 

 that the line was absent when nitrogen had been very care- 

 fully eliminated from the gas, there would be a strong pre- 

 sumption that the line was due to the nitrogen molecule. By 

 observing the behaviour of the line when the gas in the 

 discharge-tube is varied, we can often determine without 

 ambiguity the gas to which it is due. 



