IRON AND STEEL IN A ROTATING MAGNETIC FIELD. 
737 
The effect of passing - from a high induction to a low one, allowing a short time 
between each reading, w r as examined, the result being shown in Diagram 8. There 
was no definite change in the hysteresis, the curves of rising and falling currents 
being practically identical. This point will be referred to below, when more rapid 
changes are taken. 
The effect of speed on hysteresis will be examined later, when experiments will be 
described in which the induction is kept constant and the speed varied. 
Finally, a test was made at an indefinitely slow speed by slowly turning the 
armature through a short distance and then stopping it. The deflexion was then 
read, while it was standing still. Since the smallest error of centring produced an 
additional pull either forwards or backwards, there are some irregularities in the 
readings, and a high induction could not be reached. The result is in complete 
accordance with the test at higher speeds, but it will be noted that the maximum 
value, though occurring at the same induction, viz., between 16,000 and 17,000, is 
not so high as before. This is due to a small backward movement of the armature 
immediately the magnet stopped, perhaps a kind of readjustment, which does not 
have time to take place when running. It is only a small effect, about 10 per cent, 
of the value of the hysteresis at that point, and is less marked at points beyond the 
maximum. The smallest actual rotation was sufficient to prevent its occurrence, the 
values then being closely in accordance with those at higher speeds. The curve is 
shown in Diagram 9. 
[As it was found that the speed of rotation was without influence on the hysteresis, 
and that the eddy currents were almost negligible at moderate speeds, the greater 
number of readings were taken at speeds of 15 to 30 revolutions per second, since the 
deflexions were very much steadier than at low speeds. With the latter there was 
tendency for the spot of light to oscillate, especially at high inductions, owing to small 
errors in centring, which it was almost impossible to eliminate altogether. The results 
obtained at low speeds, down to a speed of one revolution per second, were identical 
with those at high speeds.—2nd September, 1896.] 
The above experiments prove very clearly the nature of hysteresis in a rotating 
field, and bear out most completely Mr. Swinburne’s deduction from the molecular 
theory of magnetism. The agreement of the phenomena with the previously 
suggested deduction forms a strong verification of the truth of this theory while it 
also easily explains various other points which have been noted. The three stages of 
the curve correspond precisely with the three stages of the molecular movement, and 
the difference between the curves for iron and steel readily follow from the difference 
in their structure. In the soft iron, as has been shown by Ewing, the molecules 
move in larger combinations than in steel, and hence it is not until a higher induction 
is reached that uniformity of movement is produced. When once past the maximum 
the decrease is very rapid, and both metals appear to give curves reaching approxi¬ 
mately a zero value at an induction between 20,000 and 21,000. 
MDCCCXC VI.—A. 5 B 
