656 CONSTANTS OF MAGNETISATION. 



in the case of long magnets ; it also represents satisfactorily the 

 experiments of M. Bouty on the magnetic moment of a needle 

 magnetised to saturation at any rate so long as the length exceeds 

 10 times the diameter, but it does not suit shorter bars. 



If, finally, a bar already magnetised be temporarily magnetised, 

 the total magnetisation* will not be represented by Green's formula, 

 but by the sum of two similar formulae with different constants. 

 The rigid and the temporary magnetism would be almost inde- 

 pendent of each other, the poles of the latter being much nearer 

 the ends, and its distribution comparable with that of long magnets. 



Comparison of the magnetic moments of bars of the same 

 sections and of different lengths gives thus a closer value for the 

 position of the poles, although the hypothesis of constant masses 

 which it implies is less exact the shorter are the bars. 



1218. Kohlrauschf endeavoured to determine the magnetic 

 length 2L of a magnet by the second term of the series which 

 expresses the action of the magnet as a function of increasing 

 powers of the inverse distance (1153). Thus, for points in the axis 



2Mf 2L 2 



and by experiments made at two different distances, we may deduce 

 the constant 2L. 



Working with very various magnetising forces on solid or hollow 

 cylindrical bars, tempered or annealed, the length of which was from 

 10 to 30 times the diameter, or even with rectangular bars, the 

 dimensions of which were 44 x 2-3 x i, the ratio of the constant 2L 

 to the length of the magnet varies from o'8i to o'86 ; we may then 

 consider this fraction as being in all cases sensibly equal to 0*83 

 or five-sixths. 



Nevertheless, apart from the fact that the experiment is of no 

 great accuracy, it does not give us the real distance of the poles. 

 Suppose, for instance, the case of a linear and symmetrical magnet, 

 whose linear density is A at a distance x from the centre. The 

 mass for length dx being Xdx, the magnetic moment of the two 



* BOUTY. Journal de Physique, Vol. V., p. 346. 1876. 

 t KOHLRAUSCH. Wied. Ann. Vol. xxn., p. 41 1. 1884. 



