Vol. X, No. 6.] Measurements with Quadrant Electrometers. 231 

 [N.S.] 



rants, except perhaps in very dry weather, is usually two or 

 three times as large as the maximum defined above, and in wet 



weather becomes still greater. 



The tubes of the air-condenser 



were insulated from one another by ebonite. This substance is 

 extremely sensitive to moisture, and the leak without the use of 

 drying agents was from ten to more than a hundred times its 

 proper value. 



Table I. 



Electrometer and condenser enclosed in teak case. Quadrants 



charged to 1*018 volt before insulating Front of case 



opened only for the measurements. No drier used. 



Date 



1913. 



April 3rd 

 „ 4th 

 July 31st 

 August 7th 



8th 

 13th 



Leak after in 



VTING ( VOLTS PER MINUTE ) 



Quadrants only 



Quadrants + Condenser. 



0*032 

 0-023 

 039 



-090 



• • 



025 



0-025 

 032 



Calcium chloride was introduced into the case on August 9th. 

 Table 11 shows that this had no effect four aays later. It 

 also shows tnat when the case was kept open for some hours 

 ou a very damp day the increase of the leak, especially on the 

 ebonite of the condenser, was very iapid. 



Table II. 



August 13th, 1913. Raining. Front of case kept open 



from 1 1 a.m. 



Time. 



Leak afteb insulating (volts per minutj-. 



Quadrants only. 



11-0 a.m. 

 11-20 a. m. 

 11-45 a.m. 

 11-50 a.m. 



2-30 p.m. 



'2-40 p.m. 



Quadrants -+- Condenser. 



070 



0-160 



0300 



