UNDER THE INFLUENCE OF CHEMICAL ACTION. 
29 
to a higher degree of accuracy than one-tenth of a volt would obviously be a matter of 
some difficulty. 
The data now under consideration clearly determine the contact differences of poten¬ 
tial ruling between the electrode surfaces during the experiments. If v is the velocity 
with which the electrons reach the receiving surface, V the retarding potential recorded 
by the voltmeter, K the contact potential, then for illumination by light of frequency 
v of a source whose threshold frequency is v 0 , in general, 
i m ,v 2 = h ( v - v 0 ) — e(V + K). 
At the intersection of a curve such as that in fig. 14 with the volt axis, Y is just suffi¬ 
ciently great to reduce v to zero. Calling this value of Y, V 0 , clearly 
K = h/e (v - „ 0 ) - V 0 . 
For the experiment to which fig. 14 refers, V 0 = — 1 -07, v = 7 -40 X 10 u , and j/ 0 is 
between 6-02 X 10 14 and 6-25 X 10 u . Consequently the contact potential K was 
between the two limits 1 -544 and 1 -639 volts. For the slightly different conditions 
ruling when the curves in fig. 12 were taken, the limits for the contact potential would 
be:—For the full curve with points thus, ©, 1 *62 and I -72 volts; for the series with 
points thus, x , l -78 and 1 -88 volts. 
§ 7.— The Approach to the Voltage Axis. 
It is well known that photo-electric characteristic curves for monochromatic illu¬ 
mination approach the voltage axis at finite angles, indicating a finite limit to the 
maximum kinetic energy, whereas the characteristic curves for thermionic electron 
currents approach this axis asymptotically. An examination of figs. 4, 6, 7 and 12 
suggests that in this respect the chemical electron curves differentiate themselves from 
the photo-electric and resemble the thermionic ones. The point is an important one 
for the interpretation of the results. It is, perhaps, not so easy to be sure about it as 
might appear from an inspection of the diagrams, inasmuch as in many cases this part 
of the curves depends on the measurement of small deflections liable to considerable 
errors. Probably the best way of testing this question is by photo-electric and chemical 
curves taken simultaneously, since any errors will then be liable to affect both curves 
in a similar manner. Fig. 15 shows the results of such an experiment with chlorine 
using the spherical electrode. The points marked x represent the relative chemical 
currents, and those marked © the relative photo-electric currents taken simultaneously. 
The values are relative to the — 3 volt values, and the currents were not saturated at 
this voltage. Some trouble arose during the measurements on the saturation part of 
the curve, and the value of the saturation current could not be got accurately, so that 
the vertical scale may be a little different from that of most of the other figures in this 
