MEASUREMENT OF DISCHARGES. 535 



An equally simple result is obtained if the closed solenoid is 

 replaced by a cylindrical coil so long that the correction for the 

 bases is very small. If n^ is the number of turns of the inducing 

 wire for unit length, and ri the number of turns of the induced wire, 

 we have (551) 



M = 



or, introducing the length / of the wire, the total number n of 

 windings, the length L of the coil, the mean radius a and the 

 radius a l of the mean circle, 



L n 



The effect of the bases will be calculated as in 769. 



1117. The simplest method of measuring the auxiliary quantities 

 contained in the expression for the resistance R, is to eliminate the 

 constants of the galvanometers, and the ratio of the terrestrial 

 components, by a common current, giving a deflection A in the 

 galvanometer, and a deflection 8 in the ballistic galvanometer, 

 provided with a shunt /*. We have then 



M tan ft fj. tan 5 



The distance of the scales need not be very exactly measured, 

 since only the ratios of the angles ft and A on the one hand, S and a 

 on the other, appear in the formulae; and the corrections for the 

 length of the needles remain very feeble. 



A shunt is then necessary for comparing galvanometers. The 

 experiments of Mascart and of Glazebrook* made by this method 

 show that, with suitable precautions, it may give very exact results. 



In order to avoid the use of a shunt, Professor Rowlandf and 

 F. Weber J have adopted a more complicated arrangement. The 

 ballistic galvanometer is formed of two symmetrical coils, and of 

 such dimensions as to enable the constant g to be calculated. An 

 independent frame, consisting of a single wire and placed in the 



* GLAZEBROOK. Phil. Trans. 1883, p. 223. 



t ROWLAND. American Journal of Science and Art, Vol. XV., pp. 281, 325, 

 429. 1878. 



F. WEBER. Vierteljahrsschrift der Naturforsch. Gesellschaft in Zurich, 

 p. 273. 1877. Phil. Mag. [6], Vol. V., pp. 30, 127, 189. 1877. 



