34 
PROF. 0. W. RICHARDSON ON THE EMISSION OF ELECTRONS 
It is not so reliable a guide to the distribution among the electrons as emitted at the 
o o 
source, on account of the difficulty in estimating the effects of the factors enumerated 
on p. .11 ante, which tend to prevent the attainment of saturation at zero field, in modi¬ 
fying that distribution before the electrons reach the receiving electrode. The picture 
may also be distorted owing to the inexact location of the true zero of potential differ¬ 
ence. A study of similar phenomena, as displayed bv tliermionically emitted electrons, 
supplies a useful guide towards the interpretation of the present chemical data. In the 
thermionic case it is found that such factors exert their major effect in distorting the 
original distribution in the neighbourhood of zero field, and the shapes of the curves in 
the larger retarding fields are little affected. Thus this method of analysis leads to 
much more reliable results for the faster than for the slower electrons. 
Such an analysis, when applied to fig. 4, is shown in fig. 17, curve 1. ; when applied 
to fig. 7, in fig. 17, curve II., and when applied to fig. 12, in fig. IS. In each case an 
assumed zero has been taken which mav be considerably wrong in the two curves in 
fig. 17. but which cannot be out by more than 0-10 volt in fig. 18, if, as I believe, the 
experimental determinations are reliable. Curves I. and II. of fig. 17 are seen to be 
verv much alike and considerably different from fig. IS. The data in fig. IS which are 
here referred to are the experimental points marked thus 0. The meaning of the 
crosses and the full curve on this diagram will be explained below. However, all three 
have certain important points in common. They show a distribution, such that fol¬ 
low energies the number within a given range dE — edX varies only slowly with E 
