1300 
Let now the iron be magnetized originally in such a way that the 
magnetization is due to the magnetic moments lying within a 
definite cone. To demagnetise it, we shall then begin with a field 
in the direction of the magnetisation and of such an intensity which 
we shall denote with H, that the magnetization caused by this field 
is greater than the original. All magnetic moments within the 
h 
cone bg Or will then form a sharp angle with the direction of H,. 
1 
The number of magnetic moments the directions of which are lying 
between the conical surfaces with the half top-angles «a and a + da 
may be represented by 21m sin eg. de, where n is a constant. If the 
magnetic moment of one crystal is 7’, each of these magnetic moments 
gives a component of the magnetization mm’ cosa in the direction H/,. 
Therefore the considered magnetic moments contribute to the magneti- 
zation in the direction H, 22nm’ sin « cos a. da. Integrating this between 
O0 and a, we obtain for the magnetization in the direction H, 
h 
M = 2xnm’ (1 — cos? a). Now cos a == En and let us put 2 anm' = m, 
1 
which thus represents the magnetization when all moments have been 
. 2 
tipped over viz. at saturation, so that M, =m (: — a If after 
Hy?) 
the reversal of the current which produces the external field, the 
magnetic force is H, (we shall suppose H, << H.), part of the magnetic 
moments changes sign, the magnetization in positive sense diminishes 
h? 
thereby by m (: = ) while that in negative sense becomes 
2 
h? ee ; 
M, = —m{|{1—-—~). After the next change of direction of the 
lab 
current again part of the moments changes sign in positive sense, 
the magnetization in negative sense diminishes (in absolute sense) by 
pie . . i . . . 
m(t — a) while that in positive sense increases by this amount. 
8 
In this way we may proceed always decreasing the external field. 
On a continual decrease of the field we shall obtain at the end 
H<h, so that no moments can change sign any longer. If now 
HA, is still just >h, while Hz, <h, the erystals that bave been 
tipped over at the last moment will keep their negative direction. 
The sum of all magnetizations in positive sense is now given by 
M, — M, + M,...— Ms,, that of the magnetizations in negative 
sense by — M, + M,...— 2 Mz,, so that the resulting magnetization is 
M=M,—2M,+2M,...—2My. 
