14 PROF. 0. W. RICHARDSON ON THE EMISSION OF ELECTRONS 
3 X 10 -13 amperes. The pressure of gas in the apparatus covered the wide range from 
0 -001 mm. to 0 -087 mm. In the different series the chemical saturation currents rose 
or fell with time or did each of these alternately. In one series it was almost constant. 
In one case the standard photo-electric current Avas 10 times as great as the standard 
chemical current, in another it was less than 1 per cent, of it, whilst in another case these 
currents were about equal. The number of drops per second varied between 1 in 
10 seconds and 1 in 22| seconds, possibly more. The size of the drops probably varied 
also. 
In spite of these wide changes in the conditions, all the points in fig. 4 fall close to 
the continuous curve drawn. It is not certain that they are not all coincident with 
this curve within the limits of likely experimental error, with the single exception of 
the observation marked a. An examination of the data makes it practically certain 
that the electrometer was not functioning when this observation was taken, as the 
photo-electric observation immediately following it also gave a zero deflection when 
it should have given a measurable deflection according to the run of the rest of the 
photo-electric curve. 
The coincidence of the points with the smooth curve in fig. 4 proves that this curve 
gives, approximately at any rate, the true shape of the C0C1 2 characteristic, and co nfi rms 
the suspicion that the moving about, along the volt axis, as in fig. 3 of the experimental 
curves, is due to changes in the contact potential between the surfaces. 
The plot of VI. • did not agree with series I. to V., and is not included in fig. 4, 
but is shown separately in fig. 5, together with a copy DEF of the composite curve in 
fig. 4 shifted 0 -25 volt to the right of its position in that diagram. In this experiment 
the points to the right of + 1 volts were taken first and the others later. It will be seen 
that the first five points taken agree with fig. 4, but the later points would have to be 
given a different displacement to bring them on to the common curve. This seems a 
