SHUDDEMAGEN. — RESIDUAL CHARGES IN DIELECTRICS. 



515 



capacity approximately equal to that of the air condenser. These 

 sheets were placed on a wooden base of the same size, but nothing was 

 put on top, and no pressure was applied. The tinfoil ends were, as 

 usual, soldered together with low melting point solder and were furnished 

 with copper wire terminals. Finally the edges of the pile of paraffin 

 sheets were melted together, and melted paraffin was run all over the 

 tinfoil ends so as to insulate the whole from the air. 



Three such condensers were built up. The first one was made with 

 great care, only the most perfect sheets being used for it. The other 

 two were not so carefully prepared and their sheets were considerably 

 thinner. The first one, "Pure Par. P," showed no leakage current 

 whatever under 520 volts on the d'Arsonval galvanometer, while the 

 other two leaked very slightly. But the most pleasing observation was 

 that each of these three condensers showed almost no residual charge 

 formation. In fact, in the region of small charging intervals, where 

 the mica condensers still show a considerable residual forming current, 

 none whatever could be observed in the three pure paraffin condensers. 

 Nor do the throws obtained for the shortest charging time bear any 

 evidence of a probable increase of residual forming current for still 

 shorter charging times. 



The observations taken are shown in Table XLV (520 volts across 

 charged condenser and d'Arsonval give no deflection). 



It will be seen from these figures of Table XLV that no certain evi- 

 dence for a measurable residual charge exists, in the region of charging 

 times up to .0017 second. With the combination of condensers here 

 used the ballistic throws for the shortest charging intervals should, if 



Observations on March 27, 1908. 



TABLE XLVI. 

 " Pure Par. Q " vs. Air. 



