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



pair of quadrants is insulated after being earthed. As this 

 varies from time to time, and may amount to two or three 

 scale divisions or even more, it is impossible to obtain any- 

 thing like accuracy in measurements, such as those on capacity, 

 in which a knowledge of the true deflection is required. Of 

 course the zero-shift hardly matters when ionisation currents 

 are being observed. 



This defect can usually be traced to the fact that the wire 

 which makes connection with the quadrants is supported on 

 some insulator, and that this has become electrified by some 

 accidental disturbance. When this insulator consists of paraf- 

 fin with the usual mercury cups let into it, the shift of zero is 

 sometimes very marked, and on several occasions we have found 

 that after a key of this type has been merely moved from one 

 place to another it has had to be kept with all its cups earthed 

 for more than twenty-four hours before the effect disappears. 



For measurements such as the ionisation of gases due to 

 radioactivity, the only essential key is a simple one to connect 

 the quadrants to earth. For general work, however, it is con- 

 venient to use a key which will give all the connections required. 

 If this key is made up as a separate apparatus it requires its 

 own and a more complicated type of insulation, thus increas- 

 ing the liability to accidental electrification. 



In all measurements it is desirable to have some means 

 for ascertaining the capacity of the electrometer and other 

 apparatus used in connection with it, at any time, since the 

 capacity of the quadrants depends to a large extent on the 

 potential of the need.e. The number of connections which 

 should therefore be arranged on the key is five, namely : — 



(1) To one pair of quadrants (the other pair always earthed). 

 l2) To one pole of a standard cell, the other being earthed. 

 (H) To earth. 



(4) To a condenser of known, preferably variable, capacity. 



(5) To the ionisation or other apparatus used in the 



measurements. 



The only special device necessary is one to prevent a short- 

 circuit of the standard cell by accidental connection between 

 (2) and (3). 



The following arrangement which we have adopted secures 

 these requirements, and at the same time reduces the amount 

 of insulation required to a minimum : — 



A stout brass wire is screwed underneath the electrometer 

 to the terminal of one pair of quadrants. This carries at its 

 other end four small brass rings, through each of which a brass 

 cylinder, kept in metallic connection with the wire by a sol- 

 dered spiral, can be pulled in the usual way by means of silk 

 loops, the insulation of the latter being further improved by sul- 

 phur on the hooks. The pointed and amalgamated ends of the 



