464 
PROFESSORS A. W. REINOLD AND A. W. RUCKER ON 
the instrument has been left to itself four or five days, and no re-charging has been 
found to be necessary during the last eighteen months. We have found it convenient 
to connect the outer case of the electrometer to earth, and one of the electrodes to 
the induction plate. This, though not the most sensitive arrangement, has been 
found to give the best results. The two electrodes are also connected by fine silk- 
covered wires to short copper rods which dip into mercury cups placed on a sheet of 
glass immediately in front of the electrometer. Two other mercury cups, connected 
with the points whose difference of potential is required, stand close by, and gutta¬ 
percha-covered copper wire bridges, which can be arranged parallel or cross-wise, 
enable the electrification of the quadrants to be reversed, and readings to be obtained 
right and left. This arrangement was found to work quite satisfactorily, and was 
more to be depended on than Sir William Thomson’s reversing key, -which was tried 
for a time and abandoned. 
The deflection of the needle of the electrometer, as measured on the scale, not being 
proportional to the difference of potentials of the quadrants, it was necessary to 
correct it. It was thought that, A —B being the difference of potential of the 
quadrants and cl the deflection, a value might be assigned to /x which would make 
A — B approximately proportional to d (1 —-); and it was found on trial that the 
A 4, 
value /x = 10,000 satisfied the requirements sufficiently well. 
A circuit was formed containing a battery of six Daniell’s cells and a constant 
external resistance of 12,000 ohms. Two points in the circuit separated by a 
resistance varying from 1,000 to 6,000 ohms were connected with the electrometer, 
and the deflections read off. The following table shows the proportionality between 
the deflections corrected according to the above rule, and the difference of potential 
obtained by multiplying the smallest corrected deflection by 2, 3, 4, 5, 6 in succession. 
The agreement may be regarded as perfectly satisfactory, when it is remembered that 
the electrometer readings cannot be relied on to much less than 0’5 of a division. 
Table III.—Calibration of electrometer scale. 
Resistance 
= 1000 n. 
] 
Right. 
Deflection ( = cf 
Left. 
)• 
Mean. 
d 2 
10 4 
7 d2 
a 10 4 
52 n. 
6000 
328-5 
320 
324-2 
10-5 
313-7 
312 
5000 
266-5 
267 
266-7 
7-1 
259-6 
260 
4000 
214-5 
210 
212-2 
4-5 
207-7 
208 
3000 
158-5 
159 
158-7 
2-5 
156-2 
156 
2000 
104-5 
106-5 
105-5 
11 
104-4 
104 
1000 
52-5 
52 
52-2 
0-2 
52-0 
52 
The distance of the electrometer from the scale and lamp was about a metre, and the 
