462 
Proceedings of the Royal Society of Edinburgh. [Sess. 
being insulated, the charge on the compensator rod is proportional to the 
potential of the compensator tube. This charge is equal and opposite to 
that induced on the gold-leaf system, and is therefore proportional to the 
gold-leaf deflection, for the capacity of the leaf system in the final position 
is constant. Hence the ratio of the new potential to the old potential is 
the same as the ratio of the new deflection of the gold leaf to the old 
deflection.) This method is slightly different from that adopted by Wilson, 
but gives identical results, as is shown by the following data, the reduction 
factor being the ratio of the new deflection of the gold leaf to the deflection 
at standard potential. By altering the potential of the compensator the 
following series of factors were got : — 
Wilson. 
Authors. Wilson. 
Authors. 
J 1-15 
M3 ) < 
D12 | ; 
1 -99 
•991 
\ 114 
: -99 
•98 J 
t -93 
•931 
i *86 
•88 1 
\ -93 
■94/ i 
[•86 
•87/ 
/ 107 
\ 1*07 
105 \ 1 
1-05 J 1 
j^-93 
•93 | 
/ -87 
•891 j 
[ 'SI 
•831 
l -87 
•89 j , 
1 -81 
•82 / 
In general, the sensitiveness of the gold leaf was from 2*5 to 30 micro- 
scope scale divisions per volt, and at the calibration one division of the 
compensator was equivalent to approximately 0T3 x 10 -2 E.S.U. of charge, 
so that, if readings be made to a tenth of a scale division, charges could be 
estimated to within 0*2 x 10~ 3 E.S.U., i.e. 0'5 x 10~ 5 E.S.U. per sq. cm. of 
test-plate. 
When the electrometer was used for open-air observations it was placed 
on the wooden box in which the instrument was kept when not in use, 
and thus the test-plate was at a standard height (63 cm.) above the ground 
during the observation. To a terminal attached to the base of the instru- 
ment an earth- wire was connected. The method of observation and order 
of operations were the same as adopted by Mr Wilson ; a full description is 
given in his paper (p. 538). 
Reduction of Charge and Current through the Test-Plate to 
NEIGHBOURING GROUND VALUES. 
To determine the factor by which the charge and current per sq. cm. of 
the test-plate had to be multiplied to deduce the corresponding values for 
the neighbouring ground a large wooden test-plate was used. This test- 
plate was square, had an area of 966 sq. cm., and was surrounded by a 
