190 f Professor Sir J. J. Thomson [April 10, 



it would require about 700 volts to drive them across when the 

 cuiTent was 2 amperes, thus the difference Avhich sets in between the 

 results when t = 1 and t = 2 at 160 volts indicates the presence of a 

 considerable number of ions for which c/m = lOi 



For voltages between oO and 160 there is no appreciable difference 

 with currents ranging from 1 to 4 amperes, while for voltages less 

 than 30 there is an appreciable diminution in the numl^er which get 

 to the disk when the current through the electromagnet is raised 

 from 1 to 4 amperes. This indicates that there are some ions which, 

 under a voltage of say 25 volts, are stopped when the magnetic held 

 is that due to 4 amperes, but can get across when it is due to 

 2 amperes. Since 1100 volts would just drive particles for which 

 e/m = 10^ across the field due to 4 amperes, 25 volts will drive 

 particles for which e/m = 107(1100/25) = lOV-U: iicross this field. 

 For the field due to 2 amperes 620 volts are required to drive 

 particles for which e/m = 10^ 25 volts will drive particles for which 

 e/in = 10^/(620/25) = 10-^/25. From the preceding results we 

 infer the presence of ions for which ^/wi is between 10^/44 and 107^5. 

 Such ions might V)e molecules of nitrogen or oxygen due to traces of 

 air in the discharge-tube ; these, however, are only a small fraction 

 of the whole number of ions. The pressure of the hydrogen in this 

 case was about * 003 mm. It is necessary to work at pressures so 

 low that the mean free path of the ion is large compared with the 

 distance d. 



Let us now compare the results obtained when the apparatus had 

 been repeatedly filled with oxygen obtained by heating permanganate 

 of potash in a tube fused on to the discharge-tube, running the coil 

 with the gas at the pressure when the discharge passes most easily, 

 and then "fining and repumping ; tbe oxygen on its way from the 

 permanganate to the discharge-tube went through a worm immersed 

 in liquid air to free it from any traces of water vapour given off from 

 the permanganate. No hydrogen lines cjould be detected in the 

 spectrum. The Faraday cylinder had been taken down between this 

 experiment and the preceding one, and slightly altered so that the 

 radius of the critical circle in this case when ^^ = 4, « = 2 • 5, is • 47 

 cm., hence the potentials required to force ions for which e/in =10^ 

 across to the disk when currents 1, 2, 3, 4 amperes flow through the 

 coils of the electromagnet are respectively 72, 270, 400, and 4<S0 

 volts. The pressure of the oxygen was O'OOO mm. 



The following (see Table on next page) are the results of the 

 experiments, n as before being the percentage of ions which reach the 

 disk. 



The figures in this case are quite different from those for hydrogen. 

 We see that for voltages over 100 the charge on the disk is not 

 appreciably diminished when the current through the electromagnet 

 is raised from 1 up to 4 amperes ; this shows that the number of 

 ions with masses comparable with those of a hydrogen atom is too 



