FORCE OF THE WESTON NORMAL CELL IN SEMI-ABSOLUTE VOLTS. 181 
VII. The Deflection Observations. 
(a) Method of Observation. 
The duplex system which had been installed by Prof. Callendar was used for the 
observation of the deflections.* The suspended coils were fitted with a plane parallel 
mirror 0'3584 cm. thick, 5 cm. in diameter, and silvered on both sides. Two metre 
scales, accurately divided on plane milk-glass, were suitably mounted on a rigid frame 
of copper pipe at distances of approximately a metre and a half east and west of the 
mirrors respectively. The parts of the scales where the readings occurred were 
illuminated by strong reflectors, and the rest was wrapped in asbestos to preserve an even 
temperature. A pair of two-foot telescopes with two-inch apertures were mounted on 
the same frame immediately below and at right angles to the centres of the scales. 
Both the mirrors and the telescopes had been specially constructed by Brashear for 
this work. The deflections were taken simultaneously through each of the telescopes 
which were provided with filar micrometers. It was found, however, that after some 
practice, readings could be estimated through the telescope directly to within three 
hundredths of a millimetre. (As the time for taking a reading was limited, this plan 
of estimation was adopted and proved to be of quite sufficient accuracy for the mean 
results.) A constant circulation of water was maintained in the supporting pipes, and 
the distances between the scales could be measured with great accuracy. The 
dynamometer was protected by non-magnetic covers which were provided with mica 
windows. 
In these observations the value of tan 0 was deduced from readings taken first in 
one direction and then in the other. It is not necessary to introduce the zero positions ; 
for if we let S equal the mean of the values obtained for the distances between the 
readings for opposite deflections, and if /3 is the average distance between the scales 
and m the thickness of the mirror, then 
tan 2 9 = —-—> 
p—m 
and tan 0 can at once be deduced. This duplex method of reading eliminates small 
errors in the setting of the telescopes and scales. 
Separate series of observations were taken with the fixed coils both reversed and 
interchanged in position. This eliminated errors introduced by any asymmetry in the 
position of the copper tape within its channel. 
(6) The Correction for Torsional Variations in the Controlling Couple. 
Preliminary observations showed slight variations which depended on the time 
elapsing between the reversing of the deflections and the taking of a reading. There 
* Callendar, loc. tit. 
