86 Construction of the Movable-coil Galvanometer. 
Hence the E.M.F. corresponding to the deflection of 45° with 
the ring vertical = 4525 volts, the correction for the resistance 
of the cells being applied. 
These two results agree very closely indeed, showing only a 
difference of - 13 per cent. This is the more remarkable as 
the two kinds of measurements have nothing in common, 
being in fact based upon data quite independent of each other, 
thus proving that it is admissible to substitute for such instru- 
ments the calibration in volts for that in amperes. 
A further series given in Tables III. and IV., and carried 
out with great care, clearly show that, for any given current- 
strength or E.M.F., the result of the measurement is almost 
identical in whatever region the readings are taken. 
Table III. is obtained with currents from my constant bat- 
tery passed through the solid ring. The six different current- 
strengths were obtained by the insertion of suitable resist- 
ances, and were as nearly as possible in the proportion of the 
whole figures 1 to 6. 
Table IV. contains measurements with the fine-wire coil, 
thereby using the E.M.F. of ordinary Bunsen cells connected 
in series, and varying in number from 2 to 12. 
From these Tables it will be seen that, on the one hand, the 
deflections extend over the greater part of the tangent-scale, 
i. e. from 3 0- 4 till 78°'2, and, on the other, the position of the 
coil varies from the multiplying power 1 to 10 — the quantities 
measurable being therefore in the proportion of 1 to 500, yet 
the accuracy arrived at may be pronounced fully satisfactory. 
Combining the results of all these measurements with the 
solid ring and with the wire coil, the mean error of a single 
reading becomes 0'35 per cent., and the probable error 0*24 
per cent. 
The last set of measurements had for its object to show that 
a compensating-magnet of the description proposed does not 
affect the readings. The results are embodied in Table V.; 
they were obtained with one of the older forms of solid-ring 
galvanometers provided with a sine-scale. The curved con- 
trolling magnet of a mirror-galvanometer, 20 centim. long 
and 2 centim. broad, was strongly magnetized and placed at 
a distance of 24 centim. in the manner formerly specified, 
and so arranged that it could be turned on a horizontal axis. 
Three different positions were given to the magnet — viz. one, 
in which it assisted the earth's magnetism, another, in which 
it did not act upon the needle, and a third, in which the earth's 
magnetism was partly neutralized. By altering the resistance 
in circuit, the deflections with the vertical ring were made 
equal in all three cases, viz. 63°" 5. 
