UNDER THE INFLUENCE OF CHEMICAL ACTION. 
25 
side. These small currents are important, and the collective data bearing on them will 
be considered later. An examination of the experimental conditions on which the 
results plotted in fig. 12 are based shows that the values denoted by © are here liable 
to much more serious errors than those denoted by x , which therefore are alone relied 
on in obtaining this part of the curve. The numbers written alongside the points 
marked 0 indicate the order in which these points were determined. It should be 
mentioned that point 1 is unreliable owing to a very rapid change in the standard current 
when it was determined. 
The second series of points indicate a small but definite falling off in the saturation 
currents with rising potentials. This may be due to some bias of chance errors, but 
it does not seem likely to be, as the data for the first series also support it. There is no 
evidence of such an effect in the COCl 2 data for the spherical electrode plotted in fig. 4, 
but it may be masked by the errors of observation. In fact, the experimental data 
which look most reliable for this part of the curve point to a rising current with rising 
negative voltages in those experiments. On the other hand, although the matter was 
not explicitly tested at the time the experiments were made, the chlorine data which 
form the basis of fig. 7 show, on re-examination, a distinct drooping tendency towards 
higher voltages. The evidence is thus not clear that this tendency is peculiar to the 
cylindrical anode. It may be illusory, but that is unlikely. It may be a general effect 
present with the spherical electrodes, but masked by errors. There are several physical 
effects which the electric field might exert on the drops which might give rise to such 
a phenomenon. 
VOL. ccxxn.—A. 
E 
