192 Professor Sir /. J. Thomson [April 10, 



which e/m = JlO"* would reach it when V = 450, those for which 

 elm = ^iglO^ when V = 56, those for which V = gVlO-^when V =28. 

 Thus for ions of different atomic weights the stages are well separated, 

 and the relative numbers of the ions of the different kinds could be 

 determined. With the comparatively large hole used in the experi- 

 ments described above it was quite easy to observe the gradual 

 diminution in the number of the lighter ions, as each dose of oxygen 

 was supplied to the tube and then pumped out. This method of 

 analysis is applicable at pressures far below those at which even 

 spectrum analysis is available. 



By reversing the potentials in the ionization chamber we can 

 collect and send through the opening in the plate negative ions 

 and corpuscles which are present in large numbers in the gas. The 

 corpuscles, on account of their small mass, are prevented from 

 reaching either the Faraday cylinder or the disk by a comparatively 

 small magnetic force, and then only negative ions get through to the 

 conductor. The proportion of these reaching the disk and cylinder 

 with changes in the electric and magnetic fields show variations of a 

 similar character to those observed for the positive ions. 



The relative rates at which the cylinder or disk charged up ac- 

 cording as positive or negative ions were supplied to them from the 

 ionization chamber was determined. When the magnetic field was 

 weak the rate of charging was much more rapid with negative than 

 with positive ions ; this was due to the excess of corpuscles in the 

 ionization chamber ; when, however, the magnetic field was strong 

 enough to stop the corpuscles, the rate of charging under potential- 

 differences of the order of about 200 volts was about the same for the 

 negative as for the positive ions, while with smaller differences of 

 potential, say 25 volts, the rate of charging with negative ions was 

 only about ]. of that with positive. The positive ions seem, in the 

 ionization chamber, to be moving more rapidly than the negative, for 

 with no electric field in that chamber both cylinder and disk acquire 

 a positive charge when the magnetic field is strong. 



With the apparatus we have been describing we can measure 

 simultaneously the values of ejm for the ions and the Canalstrahlen 

 in the same vessel, and the experiments we have described show that 

 we can get a complete change in the character of the ions without 

 any change in the nature of the Canalstrahlen ; this is, I think, 

 strong evidence that the particles composing the Canalstrahlen are 

 the same from whatever source they may be derived. 



It might, however, be urged that although the tube might be 

 cleared of hydrogen to begin with, this gas might be driven by the 

 discharge out of the cathode and that this might be the source of 

 the Canalstrahlen, and I have noticed a phenomenon which at first 

 sight suggests this view. I have observed that under some condi- 

 tions there is a lag amounting in some cases to half a minute or so 

 between starting the discharge and the appearance of the phosphor- 



