IKON, STEEL, AND NICKEL TUBES IN THE MAGNETIC FIELD. 545 



of the magnetising current are quite in accord with the manner in which the volume 

 change goes through its successive values in increasing and decreasing fields. But the 

 rapid excursions are much greater than the corresponding volume changes in the slowly 

 described cycle, a fact which is simply explained as being due to inertia. The tube, in 

 responding to the first effects of the reversed field, is thrown into an oscillation, which 

 momentarily carries it beyond the limit of statical strain. 



The arrow heads on the graphs (Plate VI.) indicate the direction in which the cycle 

 is described. The fields are measured horizontally, the changes of volume vertically. 

 Each complete cycle is taken after the cyclic state has been established by repeated 

 applications of the limiting field in alternating directions. The condition at the begin- 

 ning of the cycle is taken as the zero, although there is no doubt that the tube in this 

 condition is very far from being in a neutral condition. The effect of the recently 

 applied field must still persist in some residual form. The want of perfect symmetry in 

 all the cases shown is due to some magnetic bias existing in the tube — a bias which it is 

 almost impossible to remove, except by heating to a very high temperature. And even 

 when this has been done, the very first magnetising force applied must at once establish 

 a bias in the direction in which it is applied. 



It will be seen that the ascending branches of the cyclic graph of Iron Tube III. 

 reproduce the characteristics of the curve in Plate III. — first an increase of volume, then 

 a decrease, which reaches a minimum about Field 150. The descending branches do not 

 show the positive maximum. 



In the two cyclic graphs for Iron Tube IV. the same features are evident. The one 

 cycle is for the range of Field ±71, and the other for the range±252. Both were taken 

 on the same day with the tube in exactly the same condition. The graph of the smaller 

 cycle bears an obvious resemblance to the central part of the graph of the larger. But 

 there are very instructive differences. The field for which the volume change is zero is 

 higher in the larger cycle, as is also the field corresponding to the positive maximum. 

 Moreover — and this point I wish particularly to emphasise — the maximum volume 

 change is nearly twice as great in the higher range cycle as in the lower. Here, then, 

 we have reproduced in the slowly described cycles the same peculiarity already noted in 

 the rapid excursions when the cycle is gone through as quickly as possible. Just as 

 the rapid excursion is greater for the higher magnetising force, so the slowly described 

 cycle of volume changes touches a higher maximum when the range of the cycle is 

 greater. The limits of the cycle have a profound influence upon the detailed form of 

 the graph. 



The cycle for Steel Tube IV. presents features very similar to those of the Iron Tube 

 cycles — the differences being differences of degree. The first effect as the field is being 

 applied is an increase of volume ; and yet if we look at the Table on page 537, or to the 

 curve on Plate IV. we see that in the lowest fields there is decrease of volume. In short, 

 the undulation in the ascending branch of the cyclic graph has no existence in the 

 curve of the same steel tube on Plate IV. In this case, then, the positive maximum is 



