Vol. X, No. 6.] Measurements with Quadrant Electrometers. 237 



INS.] 



Table VI, 



Drying agent, sulphuric acid. Sealed zinc case. October bth 



to 22nd. 



Hours after 

 closing. 



Leak after insulating (volts per minute). 



Quadrants only. 



4 



0015 



12 



0-0013 



24 



o-oois 



36 



-0028 



48 



-0035 



60 



0-0038 



Quadrants and 

 sulphur plug. 



Quadrants and 

 condenser. 



0*0043 



*000< f 



0-0039 



0-0004 



1 ) -0035 



-0006 



1 1 -0045 



0008 



0-0057 



0010 



0-0062 



00012 



In all the above measurements the rate of leakage observed 

 was slightly greater than the figures given in the Tables, on 

 account of the diminution of deflection due to the gradual loss 

 of charge on the needle. This was found to be practically 

 independent of the drying agent used. The apparent leak did 

 not exceed the true leak by more than 0*001 volt per minute in 

 any case, and was usually about 0*0008 volt per minute in 



excess. 



(3) Increase of accuracy obtained. 



A few examples are added to show that with a constant 

 zero and the high degree of insulation already secured, the 

 accuracy of measurements of capacity by the ordinary method 

 of mixture is much increased. 



(a) Effective capacity of the electrometer. 



Consecutive tests with a condenser of known capacity gave 

 the following results for capacity of the quadrants and attached 

 key, when the needle was charged to 16 volts : 



February 17th, 1910 



526 

 565 



60-9 



57-6 

 53-8 



55*6 

 547 



545 



January 8th, 1914 



2625 



26-20 

 26-24 

 26-19 



26-22 

 26-42 

 26-32 

 26-29 



Mean 



Probable error of a 

 single observation 



55-8 



26-27 



1-75 



004 



