IRON, STEEL, AND NICKEL TUBES IN THE MAGNETIC FIELD. 543 



be expected that the initial effect upon the tube will be a sudden increase of volume 

 followed very quickly by the final decrease. Then at break of the current there is the 

 further complication of the self-induced current, which in its short-lived intensity may 

 well be supposed to have a peculiar effect on the volume. 



But it will be noticed that the magnitude of the initial excursion is in all cases 

 greater than the greatest final increase of volume obtained in the low fields, and becomes 

 greater and greater the higher the limiting fields chosen. It thus appears that this 

 effect depends on the immediately preceding magnetic condition of the tube. The 

 cycle, it must be remembered, has been established, and it is on the reversal of the current, 

 either from positive to negative or from negative to positive, that the phenomenon is 

 observed. If the current, after being removed, is put on again in the same direction, 

 the effect is considerably modified, and in certain cases is quite non-existent. It is 

 evident, in fact, that we are face to face with another of those complicated phenomena 

 which depend on the previous magnetic history of the material, and especially on the 

 immediately preceding magnetic condition. Similar effects in the relation of magnetisa- 

 tion to varying loads and of magnetisation to the magnetising forces have been studied by 

 Kelvin, Villari, Feomme, Auerbach, Warburg, Ewing, and others. In a former paper* 

 I have examined the same phenomena in the manner in which iron and nickel wires twist 

 when subjected to longitudinally and circularly magnetising forces ; and NagaokaI has 

 successfully carried out a like investigation for the elongations and contractions of iron 

 and nickel wires when magnetised longitudinally. 



What was demanded in the present case was a study of the cycle of volume changes 

 associated with an imposed cycle of magnetising force. This was done in a considerable 

 number of instances, the field being increased by convenient stages from zero to the 

 chosen positive maximum, then diminished to zero, reversed, increased to the greatest 

 negative value, and finally reduced again to zero. The change of volume produced at 

 each alteration of field was measured, so that the total change of volume corresponding 

 to any stage could be obtained by simply adding together algebraically the successive 

 changes up to that stage. Five examples have been chosen, almost at random, three from 

 the Iron Tubes, one from the Steel Tubes, and one from the Nickel Tubes, all with the 

 brass cap. The cycles are given in full in the table following, and the corresponding 

 graphs are shown in Plate VI. 



The first is for Tube No. III. ; and it is instructive to compare the limiting ranges of 

 volume change in the slowly described cycle with the abrupt excursions indicated in the 

 table on page 542. The cycle is for the range of field + 220, for which the excursion 

 at make would be about + 8. But the positive maximum in the slowly described 

 cycle is only +1'5. How shall we account for the extra 6'5 ? Again, as the field is 

 being slowly diminished the volume change passes through a negative value —5*28, 

 which is distinctly less numerically than the excursion at abrupt break, namely — 8. 



Thus we see that the existence and direction of the excursions at make and break 



* Trans. R.S.E., vol. xxxvi. (1891). Phil. Mag., January 1894. 



VOL. XXXVIII. PART III. (NO. 13). 4 D 



