270 MR. J. S. TOWNSEND ON THE DIFFUSION OF IONS PRODUCED IN AIR 
was found that the electrometer gave no deflection when the coil was turned on for 
several minutes. 
Method of Conducting Experiments with Ultra- Violet Light. 
The small battery of Clark cells (H, fig. 3) was insulated, and its terminals were 
connected to It and A so as to make the potential of It about 6 volts less than A. 
The tube A was connected to the negative terminal of a battery of 40 lead cells, 
the other terminal being to earth. The tubes B and the electrometer connections 
were arranged in the same way as when experimenting with the radio-active sub¬ 
stance. The observations were made in a slightly different manner. 
The quadrants connected to F were insulated and the stream of air allowed to 
run through the apparatus. About fifteen seconds were allowed for the stream of 
air to become constant; the coil working the spark gap was then turned on for a 
fixed time (twenty seconds generally). The electrometer deflection can be read when 
the spot of light becomes steady, which is an advantage of this method. When the 
coil was not working the electrometer reading did not vary more than '5 division per 
minute, and was not affected by letting the stream of air pass through the apparatus. 
The deflections n 2 (obtained in a similar manner when the ionized air passes 
through the short tubes T 2 ) were about twice as big as the deflections n ,, so that the 
latter observations were made twice in order to cover the same part of the scale. 
The deflections obtained when the rod B was connected directly to A were about 
Yoth of the deflection obtained when the zinc plate was 6 volts negative compared 
with A. 
The potential of A was changed to 80 volts positive, and the zinc plate made 
positive with respect to A, and no deflection was obtained on allowing the coil to run 
for two minutes. 
The following tables give the deflections obtained with different velocities Y 
through the tubes T. The numbers n 1 and n 2 are deflections per minute, and the 
coefficients of diffusion K were deduced from the curve, Section II. The air in these 
experiments was at atmospheric pressure H. The temperature only varied from 
16° to 18° during the experiments, so that the rates of diffusion may be taken as 
corresponding to a temperature of 17° centigrade :— 
Table IX.—Dry Air. Table X.—Moist Air. 
H. 
n 2 . 
Y. 
K. 
761 
32-8 
654 
356 
•0427 
748 
44-0 
76-2 
475 
•0438 
761 
55-7 
109-3 
377 
•0440 
H. 
«i. 
no. 
Y. 
K. 
762 
45-0 
814 
368 
•036S 
772 
47-0 
90-0 
337 
•0380 
