Genesis of Ions by the Motion of Positive Ions. 745 
for the large differences between the two values of @ for the 
two gases by this effect alone. If we suppose that the 
positive ion in a gas has approximately the same mass as a 
molecule of the gas, then the positive ions in hydrogen would 
acquire much larger velocities along the same paths than the 
positive ions in air, and large differences in the effects arising 
from collision might be expected. 
Some confirmation of this theory is afforded by the results 
of experiments made with air containing a small percentage 
of hydrogen. In this case, although only a small number of 
the positive ions in the mixed gas are derived from the 
hydrogen molecules, still, owing to the high velocities they 
acquire along the free paths, they would be more effective in 
ionizing other molecules than the positive ions derived from 
the air. Consequently we should expect that the values of 
8 for the mixture would be considerably greater than the 
values of @ for air at the same pressure. , 
The following is the result of an experiment made with a 
mixture of air and hydrogen at a pressure of 4 mms., the 
partial pressure of the air being 95 per cent. and that of the 
hydrogen 5 per cent. of the total :— | 
Pee cies "2. "os: ‘4, 5 6 Ti 8 
ee FO), .xn».-c-- 4°85 | 10°8 | 24°5 56 | 1387 .| 410 | 2170 
fn | 
la=7°91 | | ‘ ae eg | | 
B= 0110 MA eso o! | 4°88 | 103°) 244 | 55 132 | 390 | 2400 |. 
| | 
011, and the value of 2 undergoes only a small change when 
the gas contains 5 per cent. of hydrogen. 
In determining the currents in a gas at low pressure by means 
of an electrometer, we found it necessary to adopt certain pre- 
cautions in the arrangement of the apparatus when the differ- 
ence of potential between the electrodes exceeded the minimum 
sparking potential. It was found that the electrometer was 
sometimes very unsteady when there were potential-differences 
of 400 or 500 volts between the electrodes. The gas would 
insulate for some minutes and then apparently conduct, and 
a large electrometer deflexion would take place, although no 
glow appeared in the gas and the potential was obviously 
iess than the sparking potential. The cause of these irregu- 
larities was found to be due to an effect taking place along 
