10U8] 



on the Carriers of Positive Electricity. 



191 



small to be detected, for such ions under a field of 100 volts would 

 have been able to make their way to the disk against a magnetic 

 field due to 1 ampere, but not that due to 2 amperes or more ; thus 

 if these had been present in any considerable number, the number 

 reaching the disk when i - 1 would have been appreciably greater 

 than when t = 2. 



The fact that 20 per cent, under these voltages reach the Faraday 

 cylinder is due, I think, to the obliquity of the ions as they come 

 through the hole, and to the diffusion they suffer in passing tiu'ough 

 the gas. Under the smaller voltages the effect of the magnitude of 

 the magnetic field is very apparent ; thus until the voltage is above 

 20, the majority of them are stopped l)y a field of 4 amperes, indi- 

 cating that the mass of the majority of the ions is not greater than 

 480/20 or 24 times the mass of an atom of hydrogen. In fact that 

 the majority of the ions have masses comparable with that of the 

 molecule of oxygen, and are not aggregates of several molecules. 



Though the preceding list shows that the number of ordinary 

 hydrogen ions in this gas was too small to be detected, yet when the 

 Canalstrahlen produced in a tube in direct connexion with the one 

 in Avhich the ionization occurred were investigated they were found 

 to be well developed, and to give exactly the same value for e/m as 

 when the apparatus had been filled with hydrogen, as in the experi- 

 ments already discussed. 



By measuring the relative numbers of ions carried to the cylinder 

 and the disk by different voltages against a constant magnetic field 

 we can readily estimate the relative amounts of heavy and light ions 

 in the gas. Indeed I think that by using a very small hole in the 

 plate N a very fak analysis of the gas in the ionizing chamber might 

 be made. Thus suppose the magnetic field were that due to 8 am- 

 peres through the coils of the electromagnet ; then with apparatus 

 of the dimensions used in one of the experiments the ions for which 

 ejm = 10^ would not reach the disk until V = 900, while those for 



