IRON AND STEEL IN A ROTATING MAGNETIC FIELD. 
.731 
It may be noted that, while the first and last parts of the curve are in good 
agreement, there is considerable divergence near the maximum, and repeated 
experiment failed to remove this. In fact, when kept running at the maximum 
value, the deflection was by no means steady, showing that it is a genuine phe¬ 
nomenon, and not due to errors in the apparatus. Occasionally, in the latter half of 
the curve, the same phenomenon occurred, the value altering slightly for a few 
readings, and then coming back on to the original curve, without any alteration or 
stoppage of the apparatus. 
In these experiments the induction was pushed as high as possible, and a 
consistent continuous curve was obtained down to a very low value of the hysteresis. 
The minimum value obtained at an induction of 20,200 is only one-thirteenth of the 
maximum value, and the curve shows no signs of turning off again. It is therefore 
highly probable that the hysteresis vanishes altogether at a slightly higher induction, 
although the saturation point has hardly been reached. These last readings are very 
difficult to obtain, as the smallest irregularity or incidental error entirely vitiates 
them, and for some time the results were not good. However, the curve given was 
obtained several times, and the points marked are the mean of several readings, all in 
good agreement, so that it may be taken as correct. 
The comparison with the ballistic curve exhibits the same features as in the case of 
hard steel, the relation between the two curves being singularly alike. It may 
therefore be concluded that the positions will be the same for other samples of iron 
and steel, since these two occupy extreme positions among the various types. 
That the hysteresis in a rotating armature at low induction should be greater than 
in an alternating field is quite intelligible, since the movement is more gradual, and is 
free from sudden shocks. There is also not so much choice in the direction of 
movement, and hence some of the molecular combinations will offer more resistance 
to dissociation. The point is of considerable importance in the design of large 
dynamo armatures, which are usually worked at an induction of about 10,000 or 
8,000 C.G.S. At this part of the curve the value of the hysteresis is some 50 per 
cent, higher than that given by a ballistic test, and allowance must be made accord¬ 
ingly for the larger amount of heating. On the other hand, in small ring armatures, 
which are worked at a high induction, the hysteresis will be considerably lower than 
the value given by the ballistic method. 
5 A 2 
