UNDER THE INFLUENCE OF CHEMICAL ACTION. 
7 
standard voltage. The results were then expressed as fractions of the standard satura¬ 
tion current ruling at the time the particular measurements were taken. The instanta¬ 
neous value of the standard current was ascertained in various ways, depending on the 
rapidity with which it was varying. When the variation was small it was sufficient to 
take the arithmetic mean of the determinations of the standard current immediately 
preceding and following the measurement of the current under the given conditions. 
For somewhat larger degrees of variation the geometric means of the corresponding 
current values were found to give reliable results. Where the rate of variation was 
still more rapid, it was necessary to note the times of the various measurements, to plot 
a curve like fig. 2 showing the value of the standard current as a function of the time, 
and to ascertain from this curve the instantaneous value of the standard current at the 
time of the particular experiment. This method is, of course, one which is reliable 
under any circumstances; but, generally speaking, the rate of change of the standard 
current was so slow that it was not necessary to record the times at which the different 
measurements were made. This was an advantage, as each experiment involved a 
large number of settings and readings. The experimental manipulation was not par¬ 
ticularly easy in any event, and each additional item which had to be recorded made 
a series of operations increasingly tiring and correspondingly liable to involve erroneous 
records. 
This preliminary discussion will, I hope, give a general idea of the method of experi¬ 
menting adopted. It will probably be more profitable to leave further details until 
the results of the particular experiments are dealt with. 
§ 2 .—The Characteristic (Current — E.M.F.) Curve for Carbonyl Chloride. 
These data all refer to the system in which the source of emission is a spherical drop 
of the alloy NaK 2 of radius about 1 mm. surrounded by a concentric spherical electrode 
of copper of radius 2 -5 cm. The pressure of the gas in the apparatus increased fairly 
unifor ml y with the duration of each experiment, the extremes of pressure recorded lying 
between the limits 0 -001 mm. and 0 -087 mm. With the exception of one experiment, 
the final pressure did not exceed 0 -020 mm. It is probable that, in the exceptional 
case in which the final pressure of 0 *087 mm. was recorded, there was a small leakage 
of air from outside the apparatus. The data afford no evidence that the maximum 
pressure of the C0C1 2 ever exceeded 0 -001 mm., and it may have been much less. The 
pressures recorded are those of the unabsorbed products of the reaction, and are 
therefore probably due to carbon monoxide. The sensitiveness of the electrometer 
was 570 divisions per volt, and the capacity of the electrometer and its connections 
0-00012 mfd. 
In most of the series of measurements a determination of the photo-electric current 
under the same voltage, due to the blue light from the mercury vapour lamp transmitted 
through a Wratten filter No, 50, was made immediately after each determination of 
