230 Journal of the Asiatic Society of Bengal. [June, 1914. 



If an ordinary cylindrical air-condenser is used as the 

 standard capacity, it is convenient to enclose it in the elec- 

 trometer case. The standard cell may also be shut up in this 

 case, but if the silk were to break it would be short-circuited, 

 so that it is preferable to take the wire which supports the 

 mercury cup connected to it through the wall of the case by 

 means of a sulphur plug. A similar plug is necessary in any 

 case in order to establish connection between the quadrants 

 and the testing vessel outside. 



On the table which carries the lamp and scale, arrange- 

 ments are made so that the observer can easily operate the 

 threads to charge, insulate, or discharge the quadra. its and 

 any other apparatus connected to them on the key. It is thus 

 possible either to check the capacities at any time or to keep 

 the condenser connected in parallel with the quadrants when it 

 is desirable to increase the capacity of the latter. 



Provided that the points of the cylinders which make 

 contact with the mercury are well amalgamated, the trouble 

 caused by zero-shift on insulating the quadrants disappears. 



The capacity of a key of this kind is small, between five 

 and ten electrostatic units. 



(2) Improvements in insulation. 



Using a key of the type just described, a series of observa- 

 tions on the insulation of the various parts of the apparatus 

 was made from July to October, 1913, i.e. at the most un- 

 favourable season. 



In electrometer work, it is usually considered that the 

 insulation is satisfactory when the rate of leak from the 

 quadrants which after being charged to about one volt are 

 then insulated does not exceed -01 volt per minute l This is 

 a reasonable standard to accept, and it will be seen that with 

 the arrangements adopted by us there is no difficulty, even 

 during the monsoon, in Keeping the leakage at less than half 

 this amount for many days 



The capacity of the air-condenser used in our measure- 

 ments was about 80 E.S.U., or at least three times as large as 

 that of tue quadrants and attached key. Hence when the 

 quadrants and condenser are connected in parallel, the rate 



of leak 

 iasulation 



00 



— — — — — — — w -^ ***** ^-^ *w^ « 



The surface of the ambroid insulators of the quadrants 

 must of course be carefully cleaned. A point often overlooked 

 is that these insulators are hollow, and the wails oi tne cavity 

 are usually more in need of cleaning than the exterior. 



When no drying agent is used, the leak from tue quad- 



* e.g. Makower and Gei 



