278 Prot. J. S. Townsend on 
the mean value of K for positive and negative ions being 
used. 
From the values of « and K for ions in air produced by 
ultra-violet light 
Nea 
The following table of values of N x ex 10” may 
be deduced from Prof. Zeleny’s determinations of the 
velocities :— 














Positive ions | Negative ions | Positive ions | Negative ions 
in moist gas.| in moist gas. | in dry gas. | in dry gas. 
TAGE cavenen Ae al apie a elie [1-46 131 
“Onygen..ceeeeenes 1S Alig ak | 1-63 1:36 
| Hydrogen ...... | 1-24 1:18. | 8268 oe ae 
_ Carbonic ada 1-01 ‘87 | ‘99 | “93 | 
| | | 

The mean values of N xe for the different gases are 
ARUP ih Bob) cal Bde poe, gale em D pea 
OXVGeN. yi neha ote Oar. 
Capoonc Acid.) lalla 
Hiydrogeny ge yeue. pL 20 pte 
In addition it has been shown that the values of w and K 
for ions in air are both inversely proportional to the pressure 
for pressures between 760 and 200 millimetres. 
The discrepancy between the above numbers is not greater 
than the probable experimental errors, and they afford 
evidence of the equality of the charges. There is also 
evidence from other investigations which leads us to believe 
that the above values of N x e should all be equal. This may 
be deduced from experiments on the ionization of molecules 
produced by collision, which are of a much simpler kind 
than the experiments which are necessary for the determi- 
nation either of the velocities or the rates of diffusion. 
It has been shown* that the negative ions produced in 
gases by the action of Roéntgen rays or by collisions are all 
exactly the same as the ions set free from a zine plate by 
the action of ultra-violet light. 
In order, therefore, to obtain the most probable value of 
N xe we are justified in taking the mean of the above 
* J.S. Townsend, Phil. Mag. June 1902. 
