the Galvanic Current through Iron. 



15 



of the diminution of resistance observed by Schuster and 

 Stewart* on a magnetized copper wire. 



31. The following series of experiments serves to show that 

 the influence of the magnetizing is demonstrable even when 

 accompanied by considerable influences of temperature. On 

 making certain simplifying assumptions, it is easy to ascer- 

 tain the law according to which the resistance in the circuit of 

 the principal current changes with the time through lateral 

 radiation of the spiral. If now, in addition to this, we carry 

 out a longer series of alternate determinations of iv and iv, and 

 note in each the time t, we obtain by the numbers iv Q the con- 

 stants of that law. If we then construct the curve represent- 

 ing the law (it is in general a transcendental) and compare it 

 with the curve deduced from observations, we find different 

 ordinates for the abscissae corresponding to the times of the 

 determination of the quantities w. These differences refer to 

 the magnetization. I give a series of experiments of this kind 

 with the Gramme machine and the wire F 3 ; w Q , ivi, w 2 are the 

 resistances for p = 0, p — 400, and p = 800. 



t. 



w o 



w l 



w 



w 2 



w 





observed. 



observed. 



calculated. 



observed. 



calculated. 



Wj — w . 



h m 















2 50 



03285 













3 7 





03294 



0-3299 







-0-0005 



13 



0-3315 













17 



0-3326 













20 





0-3332 



0-3339 







-0-0007 



22 



27 





0-3338 



0-3348 



0-3346 



0-3381 



-0-0010 



w 2 —w 



34 0-3416 



l 











-0-0035 



Foi 



u\ 



- =S 1 we find, according to the above, three values 



of ascending magnitude, for B 2 one only, namely : — 

 S 1= -0-0015, -0-0021, -0-0030; 



therefore, mean, 



S 1= -0-0022; and S 2 = -0-0106. 



It is of course presupposed here that the discontinuities pro- 

 duced by the variation of p are masked by the residual thermal 

 action. 



The value of S 2 here should agree with the value of S in 29; 

 for H, M, n have the same values. But it is more than double 



* Phil. Mag. [4] xlyiii. p. £35 (1874) j Pogg. Ann. cliii. p. 205. 



