78 
SIR WILLIAM THOMSON ON THE 
applications and removals of a pulling stress of 174 lbs. to about 1470 scale divisions, 
when the bar was found to have been brought to a nearly permanent condition ; 
and the average effect of applying this stress was to increase the magnetism by 
16 divisions, and of removing the stress to diminish the magnetism by the same 
amount. The magnetism of the bar further increased during 11 successive appli¬ 
cations and removals of a pull of 259 lbs. to 1568 scale divisions, and the average 
effect of applying and removing the pull was to increase and diminish the magnetism 
by 28 scale divisions. After 11 applications and removals of a stress of 325 lbs. the 
magnetism was found to have increased to 1664 divisions ; and the average effect of 
“on ” and “ off” was found to be an increase and diminution of 45 divisions. 
233. Passing now to the nickel bar of Table II., we see that when the bar was 
placed in position it showed 10 divisions of positive magnetism (or magnetism of the 
same polarity as that induced by the earth). The application of 146 lbs. of pull gave 
it 29 divisions, and the removal of this pull 10 divisions additional magnetism, that is, 
both “ on " and “ off ” increased the magnetism. A second application of the pull gave 
a diminution of 5, and removal an increase of 6 divisions. The remainder of the 
procedure was similar to that followed in the case of the soft iron, and with the 
exception of the first result after the bar was placed in the clamps, the effect of “ on ” 
was always to diminish the magnetism of the bar, and of “ off” to increase it. 
After the bar was placed in the clamps the effect of the successive operations was 
on the whole to gradually augment the total magnetization of the bar from the value 
10 to the value 155, at which it stood when the experiment was concluded. 
234. From Table III., we see that the bar of cast cobalt, when placed in position, 
had its true north pole down, and gave a deflection of 392 divisions. Ten “ ons ” and 
“offs” with 136 lbs. diminished this deflection by 85 divisions. The effect of “ on” 
was then to increase the magnetism by 15 divisions, and of “ off ” to diminish it by 
the same amount; that is to say, “ on ” increased the magnetism of the bar, “off” 
diminished it. A few blows of a mallet reversed this magnetism, and caused the bar 
to give a deflection of 90 divisions in the opposite direction. The effect of “ on ” was, 
as with the nickel bar, to diminish the magnetism, and of “off” to increase it, 
the effect of 146 lbs. being 25 divisions of the magnetometer scale, of 174 lbs. 29 
divisions. The bar broke before the effect of the application of 249 lbs. could be 
observed. After the reversal of the magnetism by tapping the bar while under the 
influence of the vertical component of the earth’s magnetic force, there was very little 
gradual change in the magnetization of the bar. 
The seemingly anomalous effect obtained with the cobalt bar, when placed with its 
true north pole down, according to which the effect of the application of stress 
was to increase the magnetism of the bar, and of the removal of stress to diminish it, 
the magnet’s length, instead of the actual half length, a somewhat nearer approximation to the value 
of A might have been obtained, but, as the approximation was at best a rough one on account of the size 
of the angle, it was not thought necessary to make this refinement. 
