156 



Mr. C. T. E. Wilson. 



the gold leaf through one scale division, a series of Clark cells was 

 inserted between the condenser and its earth connection, and the 

 number of scale divisions through which the gold leaf moved on 

 reversing the Clark cells was determined ; contact between the leaking 

 system and its supporting rod being of course made before and after 

 the reversal. The scale values of the Exner electrometer were deter- 

 mined similarly. 



In the apparatus now described, a movement of the gold leaf of the 

 leaking system through one scale division corresponded to a fall of 

 potential ranging from 0*56 volt at the top of the micrometer scale 

 to 0*60 volt at the bottom of the scale. 



Any imperfection in the insulating power of the sulphur bead will, 

 as we have seen, tend to give too low a value for the leakage. The 

 error thus introduced was, however, found to be negligible ; for the 

 rate of fall of potential of the leaking system was sensibly the same 

 when its potential was equal to that of the supporting rod as towards 

 the close of an experiment when this difference was greatest. 



The apparatus used in the earlier experiments differed in some 

 respects from that which has just been described. The vessel was of 

 Ibrass in the form of a short cylinder, 6 cm. long and 5 cm. in radius, 

 the flat ends being vertical, each being provided with a rectangular 

 window closed by a glass plate, so that the position of the gold leaf 

 might be read. A purely mechanical contact-maker was used instead 

 of the magnetic one. With the voltage usually employed, a move- 

 ment of the gold leaf over one scale division corresponded to a change 

 of potential of 0'36 volt. 



With this apparatus, filled with air at atmospheric pressure (whether 

 this had been filtered or had merely been allowed to stand for some 

 hours in the apparatus), a continuous fall of potential of about 4*0 

 volts per hour occurred, showing no tendency to diminish even after 

 many weeks. Contact had to be made with the supporting rod (kept 

 as described at constant potential by means of the condenser) about 

 once in twelve hours to prevent the image of the gold leaf from going 

 off the scale of the microscope. 



Although care had been taken to avoid bringing the apparatus, during 

 or after its construction, into any room where radio-active substances 

 had been used, it was considered desirable to repeat the experiments 

 elsewhere than in the Cavendish Laboratory (where contamination by 

 such substances might be feared), and with pure country air in the 

 apparatus. Experiments were therefore carried out at Peebles during 

 the month of September, but with the same results as before obtained. 



The rate of leakage was the same during the night as during the 

 day, and was not diminished by completely darkening the room in 

 which the experiments were carried out. It is plainly, therefore, not 

 due to the action of light. 



