116 ROYAL SOCIETY OF CANADA 



The electrometer used was of the Dolezaleck type, the quadrants be- 

 ing insulated with ebonite supports, and it had a sensibility such that for 

 a potential of one volt, it gave a deflection of about GOO mm., on a 

 scale about one metre from the needle. A few readings were taken 

 with another instrument of the same type, with amber supports. No 

 difl:erence was observed, however, between the results obtained with the 

 two instruments. 



The chief difference between this arrangement and that of the 

 previous investigators lay in the fact that they used air-tight receivers, 

 while those of the writer were not hermetically sealed. This is of 

 some importance, as it has been shewn ^ that the conductivity of air 

 enclosed in a metallic receiver gradually increases with the time it is 

 confined. Meteorological conditions might also have a disturbing 

 influence. 



III. Discussion of Results. 



(a) Change of sign of charge acquired by the insulated cylinder. 

 A glance at Table I will show the great variation observed in the sign 

 of the charge acquired, by the respective insulated cylinders. Even 

 with a selected combination of cylinders, the sign acquired was not 

 always found to be the same, for, in some cases, it was possiible to 

 reverse the sign by simply cleaning the metals. Lead No. 1 in lead 

 No. 2 (vide Table I, Nos. 3 and 4) changed from positive to negative 

 after the cylinders had been thoroughly cleaned by being abraded with 

 emery paper, and then washed with distilled water, dilute hydrochloric 

 acid, water, ammonia, water and alcohol. On the other hand, tin No. 2 

 in tin No. 2 was reversed from negative to positive, after being washed 

 in the same manner (vide Table I, Nos. 26, 28 and 29). 



1 McLennan and Burton, Phys. Rev., Vol. XVI, No. 3, 1902 



