OF THE MAGNETIC PERMEABILITY OF MAGNETITE. 571 



IV. Method of Experimenting and Eesults. 



15. Practice of the Method. — The modus operandi of the investigation is therefore 

 as follows : — 



Suppose the kick at a temperature of about 300° C. is desired. The gas stove is 

 arranged to give the flame judged most suitable. The temperature is read at intervals. 

 When the required temperature is approached, if the rise of temperature is still rapid, 

 the stove is lowered until the temperature very slowly rises towards the 300° C. After 

 the temperature has been fairly steady for some time, the reading of the potentiometer 

 and the temperature of the cold junction are recorded. The compensating coil C, fig. 1, 

 is set. The primary current is read and adjusted to its standard value. The primary 

 current is reversed, by pulling the commutator bar, and the kick observed and regis- 

 tered. The current is again read and adjusted, if necessary, to its standard value. The 

 primary current is now reversed by pushing the commutator bar, and the kick again 

 noted. The mean of these two kicks is taken. The two opposite reversals of the 

 primary current are repeated several times, and the mean of the various means is used 

 as the true value for the kick at the temperature in question. At the close of the 

 observations for the kick the potentiometer and cold junction are read a second time, 

 and the adjustment of the compensating coil C, fig. 1, again verified. 



16. Results. — The results obtained are exhibited graphically in the annexed curves, 

 namely : — 



(i. ) Calibration of thermo-electric couple. 



(ii.) Kicks obtained at the various temperatures. 



The apparently anomalous character of this curve in the neighbourhood of 

 325° C. was confirmed by a second investigation of this particular region. It 

 is therefore allowed to stand as originally plotted. 



(iii.) Corrections to kicks, owing to 



(a) Imperfect balancing of coils themselves, and to 

 (/3) Leakage between primaries and their secondaries. 



(iv.) Corrected kicks, being the resultant of (ii.) and (iii.). The ordinates of (ii.) are, 

 of course, the algebraic sums of the corresponding ordinates of (iii.) and (iv.). 



(v.) Resistances of secondary circuit for various temperatures of hot junction HH, 

 fig. 2. 



(vi.) Curve whose ordinates are proportional to the apparent susceptibility, k', of the 

 magnetite specimen throughout the observed range of temperatures. The 

 ordinates of this curve are the products of the corresponding ordinates of 

 curves (iv.) and (v.). 



