194 Professor Sir J. J. Tliomson [April 10, 



puscle, M the mass of the ion being supposed very large compared 

 with m. Thus if the relative velocity falls below a certain value the 

 ion and the corpuscle will form a neutral doublet and will cease to be 

 a possible constituent of the Canalstrahlen. If the ion is moving 

 much more rapidly than the corpuscle, then V will be the velocity of 

 the ion, and we see that the smaller this velocity the more likely is it 

 to have its charge neutralized. M being the mass of the ion 

 I M V^ = P e, where P is the potential-difference moved through by 



Me" 



the ion, thus the ion will be neutralized unless P > — . --^=r. Thus 



m AB 



to protect a heavy ion, for which M^ is large, from being neutralized 



it must be subject to a much stronger electric field than would be 



necessary for a light ion ; thus, if tliere were a mixture of different 



gases in the discharge-tube, the ions formed from the lighter gases 



would persist longer than those formed from the heavier ones. 



An illustration of this result is furnished by the fact that, as I 

 showed in the paper,* the only ions besides those of hydrogen which 

 can be observed in the Canalstrahlen are those of the next highest 

 gas helium, which, when the discharge passes through helium, can be 

 observed in the Canalstrahlen without difficulty. 



The places where the neutralization of the positive ions by the 

 corpuscles takes place will be either quite close to the cathode or when 

 the cathode is perforated in the region behind the cathode ; for in 

 front of the cathode where the positive ions are produced, though the 

 velocity of these ions will be small, since they are in a feeble electric 

 field, yet the corpuscles which have come from the cathode will have 

 passed through a great potential-difference and will have a very high 

 velocity ; thus the relative velocity of the positive ions and the cor- 

 puscles will be very large. Quite close to the cathode the velocity of 

 the corpuscles will be very small, and though the velocity of the ion 

 will be much greater than in the former case, yet since the mass of 

 the ion is so much greater than that of the corpuscle, the velocity 

 acquired by the ion under the same potential-difference will be small 

 compared with that acquired by the corpuscle, so that the relative 

 velocity of the two close to the cathode will be much less than at a 

 greater distance in front of it, so that combination is much more likely 

 to occur near to the cathode, or if the cathode is perforated behind it. 



If the forces between a small positive ion and an uncharged mole- 

 cule are independent of the atomic weight of the molecule, or only 

 increase slowly as the atomic weight increases, then such an ion is 

 more likely to attach itself to a light molecule than to a heavy one ; 

 for we can show that the condition that the ion and the molecule 

 should separate is that 



M + M' 



* Phil. Mag. ser. 5, xiii. p. 661. 



