228 ~=Rev. P. J. Kirkby on the Passage of Hlectricity 
Jess exhausted. In this way seven distinct series of 
experiments were carried out, for each of which fresh 
hydrogen and oxygen were prepared. 
Owing to the method of preparing the gas, certain small 
inequalities were inevitable. It is clear that as the gas 
becomes exhausted these departures from the proper chemical 
proportion of two volumes to one will become emphasized. 
For example, if originally there is a deficiency of 3 per cent. 
in oxygen, when half the gas is gone there will bea deficiency 
of 6 per cent. Therefore, after a time, experiments on this 
plan may cease to give accurate results, and only the first 
few of each series can be relied on. 
Of course by must be a function of Ap so long as Ap is 
large ; and it only ceases to involve Ap when Ap is small, 
for then 
Ap A leap 
NO via go: 
For this reason it seemed necessary to have Ap as small 
as was consistent with its sufficiently accurate determination. — 
The values of the mean pressures p in millimetres were taken 
: A : 
as abscissas, and AO were taken as ordinates, and the results 
were plotted on squared paper. It was found, however, that 
though the points thus determined suggested a mean curve 
and proved that ie depends on p, they failed to give any 
conclusive result since the variation from the mean curve 
was sometimes about 10 per cent. 
But when the potential-difference of the plates (which may 
be called X) during the passage of the current is divided 
by p and taken as abscissa and = a6 as ordinate, these 
discordances disappear, and the points so determined ap- 
proximate to a curve within the limits of expected errors of 
observation. 
The following seven tables give the results of the experi- 
ments. 
AQ=the quantity of electricity passed, in coulombs. 
p=the mean pressure in mm. during the passage of AQ. 
Ap=the fall of pressure in mm. 
X=the potential-difference between the parallel plates, 
whose distance apart was 0°97 cm., as calculated 
by Ohm’s law. 
