o 
PROF. 0. W. RICHARDSON ON THE EMISSION OF ELECTRONS 
that the electrons are liberated by the chemical action in such a manner that for practical 
purposes they can be considered to be clear of the fields of force of the atoms or mole¬ 
cules of origin. As to the validity or otherwise of this last assumption there was no 
prior evidence, but the experiments show that it is at least approximately satisfied and 
that small accelerating electric fields have little or no effect in increasing the electron 
emission. Subject to the validity of the assumptions referred to, the currents in any 
accelerating field should be constant, whilst their values in any retarding field will be 
a measure of the number of electrons whose kinetic energy when emitted exceeds the 
equivalent applied retarding potential difference. It should, in fact, be equal to the 
number of such electrons divided by the electronic charge. A characteristic curve 
satisfying such conditions for the case of the concentric spherical electrodes will there¬ 
fore solve the problem so far as the total kinetic energy is concerned. Stated in this 
way the matter appears very simple, but it has been found in practice to be fraught 
with very serious difficulties which have taken a long time to overcome. 
In all the experiments the metal acted on was a liquid alloy of sodium and potassium. 
In the majority of them this had an initial composition corresponding to the formula 
NaKo. The alloy of this formula appears^ to have the lowest melting point, and it is 
a quite mobile liquid at room temperatures. In some of the early experiments an alloy 
of higher melting point having the initial composition NaK was used. This fact will 
be mentioned when it is necessary to refer to those experi¬ 
ments. Where no specific reference to the composition of the 
alloy is made, it can be assumed that it was close to that 
given by the formula NaK 2 . There is reason to believe that as 
the alloy is used the potassium is consumed faster than the 
sodium, with a consequent increase in the viscosity of the 
alloy. This effect is not so noticeable with the alloy NaK 2 as 
with NaK, which after a time becomes almost solid. 
After a certain amount of preliminary skirmishing with 
apparatus of different designs, that finally used in the measure¬ 
ment of the characteristic curves for concentric spheres is 
shown in a working drawing in fig. 1. The supply of the 
liquid alloy was kept in a glass reservoir (a sejDarating funnel) 
vertically above the tube A to which it was sealed. Between 
A and the reservoir was a glass stop-cock for regulating the 
flow of the alloy. A fine platinum wire trailed down the 
extension BC and was sealed through the glass between A and 
the stop-cock, and so brought to the outside of the apparatus 
for the purpose of applying any desired potential to the alloy 
in the tube BC. The dilated portion B of the tube AC was ground to fit air-tight into 
the neck of the main bulb I), and the joint could be sealed with mercury or wax at the 
* H. Le Chatelier, ‘ Recueil de Constantes Physiques,’ p. 352, fig. 3 (Paris, 1913). 
