142 Dr. F. Auerbach on the Passage of 



current the worse the more intense the circular magnetization 

 is ; but by the added longitudinal magnetization a portion of 

 the circular is annulled. In fact my experiments show, in 

 iron as well as in steel, a diminution of the resistance in con- 

 sequence of feeble magnetizings. 



The behaviour is different when the longitudinally magne- 

 tizing force is great ; a longitudinal magnetizing will then 

 result from its cooperation with the circular magnetizing force 

 of the principal current. And here the theory leaves a blank. 

 The resistance of longitudinally magnetized iron can be less 

 or greater than that of the unmagnetic. In the former case 

 the resistance-function has nowhere a minimum or a maximum ; 

 but rather the resistance constantly rises from the state of satu- 

 rated longitudinal magnetism to the state of saturated, circular 

 magnetism : this case is realized, as the experiments show, in 

 hard steel. In the other case the resistance-function has a mi- 

 nimum value for the unmagnetic state ; my experiments have 

 in fact yielded this result in iron and soft steel : here, to one 

 and the same ordinate of the resistance-curve two abscissae 

 correspond; that is, two magnetic states (namely, longitudi- 

 nally and circularly magnetic) correspond to the same resis- 

 tance. The conjectural form of the resistance-curves is repre- 

 sented in PL I. fig. 2. This gives the explanatison of experi- 

 ments 7, 21, and 25, both so far as their result were positive, 

 and also so far as they were negative. By making use of the 

 laws laid down respecting magnetization by a circular current 

 or by a spiral*, and respecting circular magnetization by the 

 current flowing through the wire itself f , we can calculate what 

 must be the ratio of the electromotive forces of the magneti- 

 zing and of the principal current (that is, the ratio M : H) at 

 given values of n, w , I, and d, in order that 8 (for iron and 

 soft steel) may vanish. But as the smallness of the values of 

 8 hardly permits this calculation to be tested, I forbear to 

 carry it out. I will only mention that, according to the ex- 

 periments, that ratio appears to depend not only on n, ic , I, 

 and d, but also on the nature of the wire. That the depen- 

 dence on the thickness is considerable can be made evident by 

 the following consideration. The longitudinally magnetic mo- 

 ment taken by the wire is, with equal magnetizing forces, 

 nearly proportional to the square root of the thickness J. The 

 case is different with the circularly magnetic moment ; indeed, 

 with equal magnetizing force the circularly magnetic moment 

 appears not to depend essentially on the thickness: but the 



* Conf. Wiedemann, Galv. ii. 1, pp. 180, 329. 



t Streintz, I. c, 



\ Dub, Blectromagnetismus (1861), p. 197. 



