8 
PROF. 0. W. RICHARDSON ON THE EMISSION OF ELECTRONS 
the chemical current. The actual measurement made was that of the combined photo¬ 
electric and chemical emission, but the instantaneous value of the chemical current 
could be obtained from the determination immediately preceding after allowing for 
its variation with the time, and thus the value of the photo-electric current obtained 
by subtraction. The light transmitted by the filter is not strictly monochromatic, but 
for most practical purposes it can be regarded as consisting of the violet group hung 
between 4347 and 4358. Whilst I have not made a spectrographic examination of the 
light transmitted by this filter, it probably lets through a certain amount of the group 
4046-4077 and a smaller amount of 4916. Consequently no particular quantitative 
reliance is placed on these photo-electric data, but they afford a valuable indication 
as to the difference between the chemical and photo-electric characteristic curves and 
also supply a useful guide as to the state of the surface of the alloy. Later on, when 
accurate photo-electric information became imperative, a monochromatic illuminator 
and a set of light filters were obtained. The reliability of these later photo-electric 
measurements, in so far as it is dependent on the monochromatism of the light used, is 
to be regarded as of a distinctly higher order than those dealt with in this part of the 
paper (see p. 23 et seq.). 
Preliminary tests with this apparatus showed that the chemical-effect currents were 
not increasing appreciably when the negative (accelerating) potential on the drops was 
increased from 2 to 3 volts ; so that — 3 volts was adopted as the potential for measuring 
the standard current. The procedure adopted, when photo-electric measurements 
were included, was first to measure the current with the light cut off at — 3 volts, then 
with the light still cut off at the voltage under test, say, x volts, then at x volts with 
the drop illuminated, then at — 3 volts again with the light off, then with the light off 
at a new voltage, say, y volts ; then at y volts with the light on, then at — 3 volts with 
the light off again, and so on, until a complete set of data had been obtained. From 
these results the values of the relative chemical currents, i.e., the values of the fraction 
obtained by dividing the current at the voltage under test by the instantaneous standard 
current and also the values of the photo-electric currents at the different voltages, 
were calculated in the manner already explained. When no photo-electric measure¬ 
ments were made the procedure was the same, except that the measurements with 
the drop illuminated were omitted. In all cases the successive settings and readings 
were made as rapidly as possible. 
At first sight the results got in this way seemed very inconsistent. For example, in 
different experiments made at intervals perhaps of some weeks, but under conditions 
which were identical so far as I could ascertain, the relative currents at zero volts (as 
compared with — 3 volts as standard) would differ by 100 per cent, or more. Later 
on it appeared that the current at zero volts had vanished altogether, and that the 
current, instead of being approximately saturated, was increasing rapidly between 
— 2 and — 3 volts. However, when the precaution was taken of ensuring that the 
standard current was really on the saturation part of the curve, and when the results 
