42 Dr. A. Stoletovv on the Magnetizing -Function of 



values of the function k for certain values of its argument, in 

 which for the cylindrical form of the iron an ellipsoidal one ap- 

 proximating to it as nearly as possible was substituted. The 

 following numbers resulted : — 



R. 



| 



*• 1 



R. 



, 



296 



250 



1512 



8-4 



301 



235 



1583 



81 



612 



16-9 



1773 



74 



823 



13-5 



1975 



67 



967 



120 



2080 



6-4 



1184 



10-2 



2397 



57 



1297 



9-5 



2484 



5-6 



As unit for R, 



mgr- 



mm. J . sec. 



was taken, after Gauss. 



From this we see that, with rising values of the argument, 

 the function k diminishes, at first rapidly, then more slowly, and 

 approaches asymptotically either zero or infinity — a fact already 

 indicated previously by the observations of Joule and Miiller. 



The same method of calculation can with greater right be ap- 

 plied to the more recent experiments of Von Quintus Icilius*, 

 because the form of the iron was really ellipsoidal. Some of 

 these experiments were also performed with feebler magnetizing 

 forces, as, instead of the direct magnetic action of the iron, the 

 induction currents excited in a spiral wrapped round the ellip- 

 soid, on the reversal of the magnetizing current, were measured. 



For the calculation of k from these experiments, however, we 

 must use only the most elongated ellipsoids, because with others 

 the influence of k on the quantity of the magnetic moment is 

 inconsiderable, and almost vanishes in comparison with the in- 

 fluence of the form of the ellipsoid. We will therefore calculate 

 the experiments with both ellipsoids (/=199, ^=1*97, a*nd 

 /=350, d=2'12; I is the polar axis, d the equatorial axis, both 

 expressed in millimetres) . If m is the magnetic moment of an 

 extended ellipsoid of rotation which is magnetized by a constant 

 force X acting parallel to the polar axis, we have 



where S represents a number to be calculated from the ratio of 

 the axes of the ellipsoid, viz. 



S=4^<r(^-l) g log nat^±| - i) , 

 * Pogg. Ann. vol. cxxi. pp. 134 & 137. 



