OF THE MAGNETIC PERMEABILITY OF MAGNETITE. 569 



First. The kick due to the secondary coil S, fig. I, when no specimen of rock was in 

 the tube. 



This kick was balanced at the outset by the other secondary S'. The truth of this 

 balance was afterwards tested throughout the range of temperatures used for the mag- 

 netite. A correction curve was then plotted with temperatures for abscissae, the 

 ordinates representing the residual kicks due to the imperfect balancing of the two 

 secondaries. 



7. Resistance of$ varies. — Second. The variation of the resistance in the secondary 

 circuit due to heating P and S. 



The kick of the galvanometer needle varies [ceteris pai"ibus) inversely as the total 

 resistance, R, of the secondary circuit. Hence, in order to calculate the variation of 

 kick which would have been obtained had R remained constant throughout, it is neces- 

 sary in each case to multiply the observed kick by the corresponding value of R. This 

 was accordingly done (see Art. 16, vi.). The values of R were determined by a post- 

 office box, the adjustable resistance of which was shunted to give extreme sensitiveness. 



8. Resistance o/*P varies. — Third. The change of the resistance of the primary coil, 

 due to heating, changed the current passing through it. 



This alteration of resistance was compensated by an adjustment of r, fig. 1. Before 

 taking a kick, this resistance was adjusted until the current regained its standard value. 



9. Field at Galvanometer changes on reversal. — Fourth. The change of the magnetic 

 field at the galvanometer consequent upon the reversal of the current in the primary 

 coils. 



This was neutralised by the compensating coil C, fig. 1. Previous to taking a set of 

 readings for the kick, the galvanometer was disconnected from the secondary circuit, and 

 then this coil was adjusted until no deflection of the galvanometer needle was produced 

 by the reversal of the current in the primary coils. 



10. Leakage. — Fifth. There might be a leakage between either primary coil and its 

 secondary. 



This was, therefore, investigated as follows : — The galvanometer being in series with 

 the secondary coils, it was observed what permanent deflection occurred upon reversal 

 of the primary current. The deflection thus obtained at the beginning of the determina- 

 tion was less than two small scale divisions, and at the close less than one such division, 

 the current in the secondary due to leakage in the two cases being respectively of the 



order 3q qqq ooo ^h and gn qqq q qq^ 1 oi * the primary current in use. A small correction 



was applied to the kicks for this leakage (see Art. 16, iii. and iv.). 



11. Sensibility might vary. — Sixth. Any appreciable variation of the sensibility of 

 the galvanometer would invalidate a comparison of the kicks obtained at different 

 times. 



To prevent, as far as possible, this irregularity, the whole determination was con- 

 ducted in a separate room, specially free from magnetic disturbances. Further, the 



