48 Proceedings of the Royal Irish Academy. 



production when the air is very dry is accounted for by the high insulation of 

 the linen which thus gets charged oppositely to the metal, and this opposite 

 charge leaks back to the metal on successive contacts. We tried the effect of 

 placing a large number of sharp earth-connected points close to the surface of 

 the linen, and found that when the metal was positive little effect was pro- 

 duced on the form of the curve, but when the metal was charging negatively 

 the presence of the points increased its rate of charging in thy air. This is in 

 nt with the fact that the point discharge takes place more easily when 

 the point is negative. The small apparent rate of production of charge in 

 a very dry atmosphere may therefore be explained by the high insulation of 

 the linen or silk. 



The varying conductivity of the linen as the amount of moisture alters 



es iK>t. however, explain the essential difference in the two types of curves. 



It doea not explain why the apparent rate of production as we approach 



saturation falls off rapidly when the metal is positive and not when the metal 



i? negati 



educed pressures. 



We i. '. pi led t-> carry out experiments in air and other gases at 



red ucf 1 \ 



A i. irate ap| is constructed, retaining, however, the same 



method ing the friction and measuring the rate of production of 



cbarg was formed bj a bell-jar resting on a metal plate. 



An iron through a hole in the metal plate and projected below it 



to a leng . - ber than the barometric height. The disc on which the silk 



or linen rubbi tied, was attached to an iron rod which passed down 



through etei tul>e and was rigidly fixed to a suitably shaped mercury 



rvoir, which coul ued by a motor. It was easy to arrange the 



iiy was ejected even at high s[>eed8. The necessary 



conn made through insulating pings passing through the metal 



• this ap| :he necessary adjustments could be made, the 



bell-jar put in position, made air-tight with wax, and a pressure of -001 mm. 



■ating mercury pump in a short time. Liquid air 

 .1 with an apparatus of this type it was not possible to 

 reach mu - without too much waste of time. 



The usual method of working was to reduce the pressure as low as 



the motor working and take observations admitting gas in 



II quantities at intervals. A two W ly Btop-cock allowed small quantities 



to be admitted when desired. The pressure was measured with a M'Leod 



gai;. 



