262 
MR. J. S. TOWNSEND ON THE DIFFUSION OF IONS PRODUCED IN AIR 
If t — qtj, p 0 = 10 4 electrostatic unit, and u — 450 centims. per second, then 
- - P becomes '006. This fraction is bigger than the correction to be applied to y, 
P o 
as can be seen by taking a simple example. Consider a case where 50 per cent, of 
ions are lost in - 9 - 0 -th of a second when diffusion is taking place in a gas having a 
mean electric density 10 -4 electrostatic unit. There are less ions per cub. centim. 
in the gas near the surface than in the gas near the centre, so that the effect of 
mutual repulsion must be much less than the above estimate, which is made on the 
supposition that the distribution is uniform. We may safely consider that when p 0 t 
is less than 10~ G , the effect of the charge does not introduce an error of 1 per cent. 
Section II .—Ions produced by a Radio-active Substance. 
The apparatus which was used for experiments with a radio-active substance is 
shown in fig. I. It consists of a large brass tube, A (60 centims. long, 3'5 centims. 
in diameter), and two smaller tubes, B 1 and If, (16 centims. long), which fitted lightly 
into A. The tubes B had brass electrodes, E, supported by brass rods, F, which 
passed through ebonite plugs, C4, in the tubes B. 
The fine tubes in which the diffusion takes place consisted of two sets of twenty-four 
each. The longer set I\ (4 centims. long and 1 millim. internal diameter) passed 
through holes in two brass discs, a x and a\, which fitted exactly into the large tube A. 
The tubes r i\, two of which are represented in the figure, projected 2 millims. from 
the discs at either end. The set of short tubes T, were passed through holes in a 
disc a.,, and projected 2 millims. on each side of the disc, so that the conducting air 
should enter each set under exactly the same conditions. The holes in the discs 
tnrough which the tubes T passed were arranged on a circle of 2 centims. diameter, 
