Soft Iron, especially with weaker decomposing -powers. 41 



netic force is directed along the axis of the cylinder, and its mag- 

 nitude is R. The iron is then magnetized uniformly throughout 

 its whole length ; that is, the magnetic moment m, referred to 

 the unit of volume, is the same in every point of the cylinder. 



The ratio ^ we designate as the value of the magnetizing-func- 



tion for the argument R. If M, L, and T represent the units 

 of mass, length, and time, R as a magnetic force is a quantity of 

 the dimensions M*L~**T _1 ; the quantity k(R) is a pure number. 



If, now, k is known for every value of R, we have all that is 

 necessary to enable us to determine theoretically the magneti- 

 zation of any mass we please of isotropic iron of which the form 

 and dimensions are known and which is in a given magnetic 

 field, so far as the coercive force of the iron can be neglected. 

 It is true that only a few special cases can be considered from 

 this point of view ; but this is, perhaps, not owing to the inde- 

 terminateness of the question, but solely to the analytical diffi- 

 culties of the solution. 



On the course of k when R becomes greater or less, as well as 

 on the question how far it turns out different for different 

 sorts of iron, the information at present existing is still rather 

 unsatisfactory. Most observers have experimented with cylin- 

 drical rods — a case in which a strict theory can only be carried 

 out on the assumption that the rod is of infinite length and 

 thinness. On the other hand, the magnetizing-force made use 

 of, and the magnetic moment it produces in the iron, have for 

 the most part not been given in absolute measure, which makes 

 the calculation of k impossible. So far as I know, such abso- 

 lute measurements have only been made by Weber and Von 

 Quintus Icilius. 



The latter had to do with iron ellipsoids (in a homogeneous 

 magnetic field) ; while Weber used cylindrical rods, which could 

 only be considered by way of approximation, as very extended 

 ellipsoids, and to that extent admit a theoretical treatment. 



Neither of the two physicists mentioned has calculated the 

 values of k from his experiments; they contented themselves 

 with the consideration of the magnetic moment of the mass of 

 iron. This, however, is not adapted to show clearly the universal 

 dependence of the magnetization upon the magnetizing-force, 

 since the magnetic moment of a cylinder of unlimited length, or 

 an ellipsoid, is conditioned not only by this force, but also by the 

 form of the iron. 



Kirchhoff* first, from Weber's measurements f, calculated the 



* Crelle's Journal, vol. xlviii. p. 374. 

 t Electrodyn. Maassbest. iii. art. 26. 



