130 DE. W. G. DUFFIELI), MESSES. THOS. H. BUENHAM AND A. H. DAVIS ON 
PAET IL—THE OEIGIN OF THE PEESSUEE. 
By Prof. W. G. Duffield. 
In the foregoino’ pages experimental evidence has been adduced demonstrating 
(l) the existence of a pressure upon the poles, (2) its variation with arc length, 
(3) its variation with current strength. 
There is reasonably good agreement between the results obtained by different 
methods in the various series of observations both for direct and alternatinsf current. 
In view of the delicate nature of the observations, it is satisfactory to find that the 
rigorous examinatio 2 i given in the section upon the separation of the electromagnetic 
and pressure effects confirms both qualitatively and quantitatively the existence of 
the pressure. 
Assuming that the corrections for convexion currents have been made with approxi¬ 
mate accuracy it remains to attempt to account for the reaction upon the poles. 
Padiometer Action and Evaporation. 
Before giving the evidence in favour of the reaction being occasioned by electronic 
projection from the poles, it is necessary to consider whether it is explicable either by 
radiometer action, produced by the departure with increased velocity of air molecules 
after Impact upon the hot poles, or by the evaporation of carbon atoms. 
Against the effect being due to either of these causes we liave the experimental 
evidence that, whereas for a constant current the reaction upon the cathode remains 
constant or diminishes with increasing arc-length (figs. 10 and 11), the amount of 
carbon lost from the poles increases rapidly with the arc-length^ over the same range. 
As the amount of carbon consumed depends upon the rate of evaporation or (and) 
upon the access of air molecules to the poles, it would be expected that, if the 
reaction depended upon either of these factors, it would increase with the consumption 
of carbon, which is not the case. 
furthermore, radiometer action is not usually appreciable at atmospheric pressure, 
though, when the object has been of very small dimensions, it has been observed at 
about ^ atmosphere. It may be argued that radiometer action is to be expected 
because the pole face is curved and the intensity of the reaction against the air 
molecules is not necessarily equal to the reaction upon the poles ; against this we 
have experimental evidence that, starting with a flat pole face, the reaction upon the 
pole became perceptibly less as it burnt to the usual curved form. 
Experiment 3, p. 125, shows that the hot pole is subject to a negative pressure when 
the current is off, indicating that, if any radiometer effect exists, it is very small and 
is masked by convexion effects. 
Again, if we assume that a pressure can arise from the expulsion of carbon atoms 
* “ Consumption of Carbon in the Electric Arc,” Duffield, ‘ Eoy. Soc. Proc.,’ A, vol. 92, p. 122, 1915. 
See Diagrams 1, 2, 4, 5. 
