
of Demonstrating the Ionisation of Air. 689 
The value of g, the rate of production of ions, used above 
was deduced from measurements of the leakage of electricity 
through air contained in a small vessel of silvered glass. As 
a part of the effect is almost certainly due to a somewhat 
easily absorbable radiation from the walls, a smaller value of 
g is to be expected in a larger vessel such as that used in the 
experiment now described ; g has, moreover, been shown by 
several observers to depend on the material of which the walls 
are composed. The experiments of H. L. Cooke*, who used 
a brass vessel of 1100 c.c. capacity, are more nearly com- 
parable with those of the present investigation. The value of g 
found by him (without special shielding of the apparatus) was 
about one-third of the value used in the above calculations. 
If we use this value the effect of recombination becomes still 
less important in comparison with that of diffusion ; and if we 
ignore recombination, the maximum number of negative ions 
in the absence of an electrical field, being proportional to g, 
is reduced to less than 900 per c.c. 
I have not yet succeeded in making any direct determina- 
tion of the number of drops actually { produced on expansion. 
A superior limit to the number was, however, obtained 
indirectly by observing the rate of fall of the drops—the 
method adopted by J. J. Thomson in his determination of the 
charge carried by an ion. In this method of finding the 
number of the drops, the total quantity of water which 
separates out from each c.c. as a result of a given adiabatic 
expansion is calculated and assumed to be equally distributed 
among the drops; while the radius of the drops is obtained 
from the rate of fall by the use of Stokes’ formula. When 
the drops are so few and the fall so rapid as in the present 
experiments, one cannot assume that they attain their maxi- 
mum size; in other words, that sensibly all the available 
water is condensed upon the drops. The value found for 
the number of drops by dividing the total available water 
by the volume of each drop (as “obtained from the rate of 
fall) will therefore be too high, but may be considered as a 
superior limit below w hich the actual number of dr ops 
really lies. 
A series of observations, in which comparatively rough 
measurements of the time taken by the drops to fall were 
attempted, gave the following resu ults. Expansions capable 
of catching all the negative ions produced drops all of 
which had fallen to the lower plate in less than three seconds ; 
the temperature being 14°C. The distance between the 
? H. L. Cooke, Phil. Mag. vol. vi. p. 403 (1903). 
Phil. Mag. 8. 6. Vol. 7. No. 42. June 1904. 3A 
