Jan. 1 6, 1890J 



NATURE 



-DO 



result as depending upon ihe construction of the material. 

 Let us, however, consider the simplest isotropic arrange- 

 ment ; let us suppose that one material is in the form of 

 spheres bedded in a matrix of the other : if the spheres 

 are placed at random this is clearly an isotropic arrange- 

 ment. ,The result is very different according as the 



matrix or the spheres are of the magnetic material. 

 Suppose that the volume of the spheres is one-half of the 

 whole volume. In Fig. 7 we have approximately the 

 curve for iron, for a mixture of equal quantities of iron 

 and a non- magnetic material ; the spheres being non- 

 magnetic and the matrix iron, and for a mixture, the 



Fig. 4. 



spheres being iron and the matrix non-magnetic. 

 Observe the great difference. When the spheres are 

 iron, the induction is near four times the force for all 

 values of the force. When the matrix is iron, the induc- 

 tion is near two-fifths of the induction when the material is 

 iron only. 



Fig. 5. 



In speaking of the properties of bodies which, like 

 manganese steel, are slightly magnetic, it may be well 

 here to enter a caution. But little that is instructive is 

 to be learned by testing filings, or the like, with magnets, 

 as these show but little difference between bodies which 

 are slightly magnetic and those which are strongly 



magnetic. Suppose the fihngs to be spheres ; m the 

 following table are given comparative values of the forces 

 they would experience in terms of /t, if placed in a 

 magnetic field of given value, /x having its ordinary 

 signification — that is, being the ratio of the kick on the 

 galvanometer when a ring is tried made of the material 

 of the filing to the kick if the ring is made of a perfectly 

 non-magnetic material : — 



Non-magnetic body. 



Manganese steel with 12 per cent. 



Manganese!steel with 9 per cent. 



