of the Movable-coil Galvanometer. 85 
upon the constancy of permanent magnets, which, to say the 
least, is rather precarious*. 
Numerical Results of Measurements. 
I shall now communicate some measurements and tests to 
which the latest forms of instruments have been subjected, in 
order to illustrate the degree of accuracy obtainable. The 
first set was undertaken to ascertain the relation actually 
existing between the solid ring for currents and the coil of 
wire for E.M.F., which, it will be remembered, was intended 
to be such that amperes with the solid ring should accurately 
correspond to volts with the wire coil. The experiment was 
conducted as follows: — A current, from my constant battery 
with acid flowf, was sent through the solid ring and a copper 
voltameter in circuit for a certain time, the deflections to the 
right and' left being observed every five minutes. The mean 
of these deflections was taken as corresponding to the amount 
of copper deposited. The copper-sulphate solution and the 
electrodes consisted of pure materials. The mean of the gain 
of the kathode and the loss of the anode was taken. The 
amount corresponding to one ampere per hour is 1*164 
gramme of copper or 3*96 grammes of silver, the latter figure 
being that adopted by Messrs. Siemens and Halske, of Berlin^. 
The calibration of the fine-wire coil for volts was performed 
by means of a number of Daniells, each compared with a 
Raouh?s standard cell filled with pure sulphate solutions 
and having the E.M.F. 1*115 volt, according to Dr. Alder 
Wright's experiments § . The figures obtained were as fol- 
lows: — 
a. With the solid Ring. 
Copper obtained =11*435 grammes. 
Time of electrolysis = 60 minutes. 
Mean deflection =47°*3. 
Position of ring P = 2. 
Hence the current corresponding to the unit deflection of 45° 
with the ring vertical =4*531 amperes. 
b. With the fine-wire Coil. 
Number of Daniells = 4. 
E.M.F. thereof = 4 x 1*109. 
Deflection obtained =44°* 6. 
Position of coil P =1. 
* Another advantage undoubtedly is that the galvanometer requires 
no variable shunt, by which errors mav very easily be introduced, 
t Rep.f. exp. Phys. xviii. p. 633 (1882). 
j See latest instructions for use of their torsion galvanometers. 
§ Proc. Phys. Soc. v. p. 80 (1882). 
