MAGNETIC SHIELDING IN HOLLOW IRON CYLINDERS. 649 



values of e£B 2 /<ifI 2 . as given by the tangent of the curves at the points referred to, were 

 reduced from the normal of 6000, as given in Table II. for shield B, to a value of about 

 3500. The B-H curve for shield B given in fig. vn., if slightly sheared over, suffi- 

 ciently well represents the cyclic changes of the B 2 induction component due to the 

 superposed alternating force H 2 . (If, however, the values of H 1 are unduly increased, 

 the B 2 -H 2 cycle becomes lop-sided. This was not further investigated. 



If comparison be now made between this curve and the c} T clic changes of the B 1 

 component, it is at once seen that the lowest values of B x occur at the positive and 

 negative extremes of the superposed force H 2 ; and the highest at those points of the 

 B 2 cycle, where B 2 is changing most rapidly with respect to H 2 The maximum and 

 minimum values of B x thus each occur twice during one complete reversal of the 

 superposed force, and the similarity of the B x cycle to the effects of repeated positive 

 and negative twisting of an iron rod or wire upon its longitudinal magnetisation may 

 also be noted. 



We may therefore conclude that the final cyclic changes of the B x induction com- 

 ponent in the direction of the magnetising force Hj kept at a constant value, follow at 

 least approximately the changing permeability (cZB 2 /V7H 2 ) impressed upon the iron by 

 the superposed alternating force H 2 carrying the B 2 round its complete cycle. The 

 effects of the previous history of the iron may be wiped out, but hysteresis appears 

 in another form. 



III. Magnetic yEolotrojiy of Demagnetised Iron. 



§ 30. It is well recognised that, however completely iron once magnetised may be 

 apparently demagnetised by magnetic means, it is not in the same molecular condi- 

 tion as it was previous to its having been magnetised, or after it has been raised to a 

 high temperature. Ewing gives a very marked example of this (Magnetic Induction 

 in Iron, 3rd ed., p. 98), where, by subjecting iron to a certain sequence of magnetis- 

 ing forces, it is possible to leave the iron at zero magnetisation and in zero field. 

 Re-application of the magnetising force, however, shows " a striking want of directional 

 symmetry when it is subsequently magnetised " in a positive or negative direction. It 

 is pointed out by way of contrast that iron which has been made neutral by the process 

 of demagnetising by decreasing reversals, does not show this directional difference as to 

 whether the subsequent magnetising force be positive or negative. 



But here ag;ain we know that the molecular condition in which the iron is left even 

 by this process must be very different from that of originally unmagnetised iron, or 

 iron demagnetised by the action of heat. The following example may be given, also 

 from Ewing (loc. cit., pp. 340-341) : When a magnetising force is applied and then 

 repeatedly reversed, " successive repetitions of the process give a gradually diminish- 

 ing range of magnetic change," — " but when demagnetised by reversals after the 



