UNDER THE INFLUENCE OF CHEMICAL ACTION. 
9 
of the whole of the completed experiments were plotted and compared together, it 
was found that there was a definite method about these variations. This will be recog¬ 
nised from an inspection of fig. 3, which shows three of the characteristics as actually 
determined. The numbers alongside the crosses on curve 1 express the order in which 
-— --->- 
the different points were determined. It will be seen that each characteristic is 
consistent enough in itself. The variability, normally about 5 per cent., which seems 
rather pronounced on the flat part of the curves, is due to the cumulative effect of the 
errors inherent in the determination of the currents. The percentage error should be 
independent of the magnitude of the currents, and consequently is not so noticeable 
on the lower parts of the diagram where the relative currents are smaller. It is 
probable that the main source of error lies in the variation of the rate of emission 
with the growth of the drops, and it will be noticed that it is most pronounced in 
curve III., where the drops were formed with exceptional slowness (at the rate of 4 in 
90 seconds). However, I believe that the accuracy of the measurements is sufficient 
to establish the conclusions which it is intended to draw from them. 
It will be noticed that the shapes of the curves are all much the same, the most notice¬ 
able difference being that they are spaced widely apart on the voltage axis. In fact, 
it is clear from an inspection of fig. 3, that if any two of the curves are given suitable 
horizontal displacements, they can be made nearly to coincide with the remaining one. 
The explanation of such a result seemed obvious. The shapes of the characteristic 
curves are the same in the different experiments, but they occur at different places 
along the voltage axis. Now the voltage Y plotted along this axis is that given by a 
VOL. CCXXII.—A, 
c 
