170 Scientific Proceedings, Royal Dublin Society. 



ACCUKACY OF THE METHOD. 



The accuracy of the method obviously[depends simply on the 

 constancy of the ionisation current, and on the accuracy with which 

 the time-intervals are measured. Numbers are given below to 

 show how very constant the ionisation current is. In practice it 

 is well to screen the space between the plates A and B from air 

 currents, as such currents, if strong, may blow away the ionised 

 air, and diminish the ionisation current. As regards the radio- 

 active substance used, uranium is preferable to thorium or radium, 

 as it gives off no emanation. If radium or thorium is used, it 

 should be in a closed vessel to prevent the emanation from escap- 

 ing, otherwise the ionisation current will not be steady. 



The method involves the use of a quadrant electrometer, which 

 to some may appear an objection to the method. The writer's 

 experience, however, is that no sensitive scientific instrument gives 

 less trouble in working than a quadrant electrometer, when it has 

 once been put in good order. When the capacities being compared 

 are small, no great sensitiveness of the electrometer will be required 

 — say 60 millimetre scale-divisions for one volt with scale one 

 metre from electrometer. When large capacities are being 

 compared, greater sensitiveness will be necessary, unless a very 

 large quantity of uranium is used ; but there is no trouble 

 in having an instrument one hundred times as sensitive as 

 above. 



Some Experiments with this Method. 



(a) We shall first give some numbers to show the constancy of 

 the ionisation current in the above arrangement, and the accuracy 

 with which the time required to charge any system through a 

 given range of potential can be measured. The system charged 

 consisted of the electrometer, a capacity marked *001 microfarads, 

 and the condenser formed of the plates between which the uranium 

 is placed. 



The time taken for the spot of light to move over fifty scale- 

 divisions was taken with a stop-watch reading to fifths of a 

 second. A series of seven observations was made, giving the 

 following numbers, no observation being rejected. 



