1920.] Measurements with Quadrant Electrometers. 17 
of the quadrants through its amber support, carrying four 
hinged soft iro n armatures about 3°5 cms. long by 0°5 cms. 
wide. Each sekhibin has a well- ‘amalgamated copper point 
rivetted at its end which makes connexion when doshas oS with 
the corresponding mercury cup placed underneath it. The 
armatures are suspended by short brass spirals from ft hori- 
en necessary, one electrometer together with its key 
can be lifted from the serrate and replaced by another, 
without any further adjustment 
One of the two mercury se not shown in the diagram is 
connected to a cylindrical air-condenser, with amber insulation, 
supported vertically inside thé case, the outer cylinder being 
earthed. is air-condenser is used for determining the capa- 
city of the senpion eee at any time. The fourth cup is 
supported by a wire through a sulphur or amber plug on the 
side of the lower part of the case (C) and is used to charge the 
quadrants to the potential of one or more cadmium cells placed 
outside the case. 
t is necessary to have one more sulphur or amber plug 
passing through the side of the base (C) in order to lead in the 
potential for charging the electrometer needle toa suitable 
voltage. The wire passing Saba this plug is connected 
either to the binding screw (S) when a conducting suspension 
is employed, or to the binding screw (7’) on the charging 
arrangement supplied by the makers for use with quartz fibre. 
In the latter case the additional arrangement for charging the 
needle from outside the hermetically sealed case, mentioned in 
our previous paper, is used. Alignment marks are put on both 
rti out 
lifted up to examine the internal fittings = replaced in a 
few seconds with hermetical sealing as befor 
Tn order that the arrangements deascibed — work satis- 
erga a few important points require attentio 
(1) Should the material of the oaoreraret needle be 
slightly magnetic, the comparatively strong magnetic fields 
when one or other of the electromagnets is excited 
would alter the zero. We have found, however, that when 
and a fair degree of sensitiveness (about 300 mm. at on 
metre for one volt on the quadrants) there is no sncuaclabie 
disturbance of zero due to this cause. 
(2) Another possible source of zero shift is the nara 
which takes place in this arrangement when contact betwee 
