214 MEASUREMENT OF CURRENTS. 



The length 2/ is rather difficult to determine for a magnet of any 

 given form, but for a short cylindrical bar it may be assumed, in 

 conformity with the experiments of Coulomb, that the pole is at a 

 third of the same length from each end, and that consequently the 

 value of / is equal to - of half the length of the magnet. 



*) 



When the angle B varies from zero to 45, the factor (1-5 sin 2 5) 



/2 



remains less than 1*5. If, then, the ratio is of the order of experi- 

 mental errors or, at any rate, of the approximation which we aim 

 at the law of tangents will be applicable within the same interval. 

 Beyond this limit the sensitiveness rapidly diminishes. 



In order to extend these limits, M. Bertin* has proposed to 

 place the coil at an angle of 45 with the magnetic meridian, and 

 to pass the current so that the needle is brought towards the coil. 

 Equation (7) becomes then 



H sin 5 



G cos (a - 8) ' 

 or, if 6 is the angle a-8 of the needle with the coil, and a = 45, 



The angle 6 may thus vary from +45 to -45, so that the total 

 deflection may amount to 90. The strength of the current which 

 can be measured by the galvanometer without introducing a cor- 

 rection is thus doubled. It is not, however, possible by this method 

 to reverse the current and get rid of the defects of adjustment. 



837. In the original instrument of Pouillet the coil consisted 

 of a single circle 25 to 30 cm. in diameter, with a needle 7 to 8 cm. 

 in length playing on a pivot and provided with an index movable on 



/ 2 i 



a graduated circle. The ratio was then about , and the term 



a* 16 



of correction of the same order. 



* BERTIN. Ann. de Chitn. et de Phys. [4], Vol. xvi., p. 25. 1869. 



