250 



NATURE 



[yan. 1 6, 1890 



great as for non-magnetic substances. This extraordinary- 

 property is possessed by only two other substances 

 besides iron — cobalt and nickel. On the same figure are 

 curves showing on the same scale what would be the 

 deflections for cobalt and nickel, taken from Prof. 

 Rowlands's paper. You observe that they show the same 

 general characteristics as iron, but in a rather less degree. 

 Still, it is obvious that these substances may be broadly 

 classed with iron in contradistinction to the great mass of 

 other bodies. On the other hand, diamagnetic bodies 

 belong distinctly to the other class. If the deflection with 

 a non-magnetic ring be unity, that with iron, as already 

 stated, may be as much as 2000 ; that with bismuth, the 

 most powerful diamagnetic known, is 0*999825 — a quantity 

 differing very little from unity. Note, then, the first fact 

 which any theory of magnetism has to explain is : Iron, 

 nickel, and cobalt, all enormously magnetic ; other sub- 

 stances practically non-magnetic. A second fact is : 

 With most bodies the action of the primary current on 

 the secondary circuit is strictly proportional to the 

 primary current ; with magnetic bodies it is by no 

 means so. 



You will observe that the ordinates in these curves, 

 which are proportional to the kicks or elongations of the 



galvanometer, are called induction, and that the abscissas 

 are called magnetizing force. Let us see a little more 

 precisely what we mean by the terms, and what are the 

 units of measurement taken. The elongation of the 

 galvanometer measures an impulsive electromotive force 

 — an electromotive force acting for a very short time. 

 Charge a condenser to a known potential, and discharge 

 it through the galvanometer : the needle of the galvano- 

 meter will swing aside through a number of divisions 

 proportional to the quantity of electricity in the condenser 

 — that is, to the capacity and the potential. From this 

 we may calculate the quantity of electricity required to 

 give a unit elongation. Multiply this by the actual re- 

 sistance of the secondary circuit and we have the impulsive 

 electromotive force in volts and seconds, which will, in 

 the particular secondary circuit, give a unit elongation. 

 We must multiply this by 10^ to have it in absolute C.G.S. 

 units. Now the induction is the impulsive electromotive 

 force in absolute C.G.S. units divided by the number of 

 secondary coils and by the area of section of the ring in 

 square centimetres. The line integral of magnetizing 

 force is the current in the primary in absolute C.G.S. units 

 — that is, one-tenth of the current in amperes — multiplied 

 by 47T. The magnetizing force is the line integral divided 



Fig. I. 



by the length of the line over which that line integral is 

 distributed. This is, in truth, not exactly the same for 

 all points of the section of the ring — an imperfection so 

 far as it goes in the ring method of experiment. The 

 absolute electro-magnetic C.G.S. units have been so 

 chosen that if the ring be perfectly non-magnetic the 

 induction is equal to the magnetizing force. We may 

 refer later to the permeability, as Sir W. Thomson calls 

 it ; it is the ratio of the induction to the magnetizing 

 force causing it, and is usually denoted by fi. 



There is a further difference between the limited class 

 of magnetic bodies and the great class which are non- 

 magnetic. To show this, we may suppose our experiment 

 with the ring to be varied in one or other of two or three 

 different ways. To fix our ideas, let us suppose that the 

 secondary coil is collected in one part of the ring, which, 

 provided that the number of turns in the secondary is 

 maintained the same, will make no difference in the 

 result in the galvanometer. Let us suppose, further, 

 that the ring is divided so that its parts may be plucked 

 from together, and the secondary coil entirely withdrawn 

 from the ring. If now the primary current have a 

 certain value, and if the ring be plucked apart and the 

 secondary coil withdrawn, we shall find that, whatever 



be the substance of which the ring is composed, the 

 galvanometer deflection is one-half of what it would have 

 been if the primary current had been reversed. I should 

 perhaps say approximately one-half, as it is not quite 

 strictly the case in some samples of steel, although, 

 broadly speaking, it is one-half. This is natural enough, 

 for the exciting cause is reduced from — let us call it a 

 positive value, to nothing when the secondary coil is 

 withdrawn ; it is changed from a positive value to an 

 equal and opposite negative value when the primary 

 current is reversed. Now comes the third characteristic 

 difference between the magnetic bodies and the non- 

 magnetic. Suppose that, instead of plucking the ring 

 apart when the current had a certain value, the current 

 was raised to this value and then gradually diminished to 

 nothing, and that then the ring was plucked apart and 

 the secondary coil withdrawn. If the ring be non- 

 magnetic, we find that there is no deflection of the 

 galvanometer ; but, on the other hand, if the ring be 

 of iron, we find a very large deflection, amounting, it may 

 be, to 80 or 90 per cent, of the deflection caused by the 

 withdrawal of the coil when the current had its full value. 

 Whatever be the property that the passing of the primary 

 current has imparted to the iron, it is clear that the iron 



