OP ARTS AND SCIENCES. 



265 



The time of charging was about 0.2 sec. The time was varied, 

 but the results were not materially changed. I could not find a 

 time of charging that would make the residue disappear. Of course 

 the residue, as well as the charge itself, varies with the time of 

 charging. 



The following tables show the variations, with time of charging 

 varying from O.UOl sec. to 30 sec, of the capacity of each condenser. 

 The charging in short times was accomplished by means of the pendu- 

 lum apparatus described in the Proceedings of the Academy, Vol. 

 XXIV. p. 148. For longer times, a key was used. The measure- 

 ment was made on discharging immediately after charging. All these 

 measurements have been plotted, using the time in seconds for abscissas 

 and the galvanometer deflections as ordinates. The curves are nearly 

 parallel, and close to the axis of Y at first, but change rapidly, and iu 

 a good condenser become nearly parallel to the axis of X in a few 

 seconds. The results give the means of four to ten observations. 

 Where there is as much as two per cent disagreement in the observa- 

 tions, the mean is marked with a question mark. The time curve of 

 a condenser is one of the best tests of its goodness. The capacities 

 of leaky condensers vary greatly with the time. 



Condenser No. 1. 



CONDEKSEB No. 2. 



