40 ProceedinffS of the Ro//al Irish Academy. 



charged to various potentials, aud the corresponding currents to the upper 

 plate are read off by means of the electrometer. 



If all the ionization is produced at X, and if only one kind of ion is 

 present, then, as the voltage is increased from zero, the electrometer will 

 receive no charge until the field applied is sufficient to carry the ions across 

 the vertical space between the plates, while the air-stream carries them 

 through tiie horizontal distance of 25 cm. But, since all the ions are produced 

 very nearly at the. same place, they will have almost identical paths, and they 

 will all arrive at the upper plate for a certain small increase in voltage. If 

 the voltage is still further increased, there will be no increase in the current, 

 as all the ions are already captured. If the current-voltage graph is plotted, 

 it should show a zero value for current up to a certain point, then an abrupt 

 upward bend at a certain voltage, and then a horizontal part. Knowing the 

 dimensions of the apparatus and the quantity of air passing through it per 

 second, we are enabled by an observation of this "critical voltage " to calculate 

 the mobility of the ions. 



But the form of curve described cannot be realized in practice. We have 

 assumed that all the ionization is produced in a certain restricted region ; 

 but thu more penetititing ladiations will cause a weak general ionization 

 throughout the whole air-space. '1 he current due to this ionization will 

 increase smoothly as the voltage increases, so that our experimental curve 

 will really be iluo to the superposition of the step-like curve on this smooth 

 curve. When these experiments were initialed, it was e.xpected, therefore, 

 that the curves obtained would show a gradual upward slope, then an abrupt 

 step correspiindiiig to the imlinary small ion, and then a resumption of the 

 gradual rise. The actual curves obtained, however, were not of that simple 

 nature. 



Befoi-e discussing the graphs obtained, some further remarks are necessary 

 as to the conditions of working, in the beginning, the ordinary air of the 

 laboratory was drawn tiirough the apparatus. It was found, however, that 

 the electrical readings, while frequently quite good, were occasionally 

 unsteady, and the evidence seemed to point to variations in the humidity of 

 the air as the source of the unsteadiness. It was decided, tlierefore, to use 

 air of some definite degree of humidity; and, as the ditliculties involved in 

 drying large volumes of air are veiy considerable, it was arranged to work 

 with saturated air. The an-angement adopted then was to pass the air from 

 one water-sealed gasometer through the apparatus into another identical 

 gasometer. The gasometers were coupled together so as to move at the 

 same rate. The results given in this paper, therefore, refer altogether to 

 saturated air. 



