534 
PROFESSOR J. A. EWING ON EXPERIMENTAL 
is withdrawn. This entirely erroneous opinion has probably been established by 
experiments on rods whose length was insufficiently great to prevent them from 
demagnetising themselves more or less completely. In fig. 2 we see how great is the 
effect of a small amount of reversed force in reducing the residual magnetism. To 
illustrate the extent of the self-demagnetising action which occurs in even a moderately 
long rod, we may take the following ideal case. Suppose that an infinitely long, 
straight, round rod of radius r has been uniformly magnetised in the direction of its 
length, and the magnetising force withdrawn, leaving it with a uniform residual 
magnetisation whose intensity is 3- Then imagine a length 2 a to be suddenly cut 
out of it, and the ends to be instantaneously removed. At the instant of separation 
the magnetisation may still be supposed to be uniform, and the self-demagnetising 
force which then acts at any point in the axis of the rod at a distance x from the 
centre is 
a — x a-\-x 1 
\/r~ + (a—xY \/r 2 + (a + x) 2 \‘ 
Such a distribution is of course self-destructive; but it is interesting to examine 
the numerical values of the demagnetising force it would cause, with a given value for 
3 and for the ratio of length 2a to radius r. Let the rod’s length be fifty times its 
diameter, and suppose that it consists of the same soft iron as the ring of fig. 2, and 
has been similarly treated before cutting, so that the residual value of at the 
instant of cutting is 9000. The following numbers show the resulting values of the 
demagnetising force due to the ends at various points along the axis of the bar 
Distance from centre expressed as Self-demagnetising force in 
a fraction of the length. c.g.s. units. 
0-5 
4500 
0-495 
2488 
0-49 
1318 
0-48 
475 
0-45 
87 
0-4 
22-7 
0-35 
10*3 
0-3 
6-0 
0-25 
4-6 
0-2 
2-9 
0 
1-8 
Comparing these with fig. 2 we see that throughout very nearly its whole length 
the rod would be subjected in this ideal case, by its own magnetism, to a force more 
than sufficient to remove that magnetism altogether. It need not therefore cause 
surprise that there is little residual magnetism in the equilibrium distribution which 
is arrived at after any applied magnetising force is withdrawn from a rod whose 
length is even as much as fifty times its diameter. 
