650 MR JAMES RUSSELL ON 



magnetising force has been raised to a high value " this property is lost. " In this 

 respect, then, a wire demagnetised by reversals differs from the same wire in its. 

 'primitive annealed state." 



It has already been stated (§ 14) that the shielding to an external transverse field 

 is higher when the hollow iron cylinders or shields have been previously demagnetised 

 by decreasing reversals of the circular as compared with the longitudinal field 

 (see § 54). This is the case immediately in point, as it raised the question whether 

 iron, after having been demagnetised by decreasing reversals, shows directional 

 permeability (not as to subsequent forces being positive or negative, but) to subsequent 

 forces at right angles to that used in the demagnetising process, or whether the effects 

 were spurious, depending upon other causes entirely. 



§ 31. In the preliminary experiments, the shields A and B were dispensed with. A 

 sheet of transformer iron was taken, and a piece cut from it in the form of a cross. The 

 two arms were equal and measured about six inches in length by three-quarters of an 

 inch in breadth. The ends of one of the arms were bevelled, bent round so as to meet, 

 overlapped and secured in position, forming a sheet iron ring. The ends of the other 

 arm were bent in the opposite direction and formed into a second ring similar to the first. 

 These rings, after having been heated to redness in a bunsen flame, were wound with wire 

 through which independent currents could be sent, thus forming two magnetic circuits, 

 M and N. The piece of sheet iron, three-quarters of an inch square and not wound with 

 wire, was necessarily common to both circuits. An exploring coil of fine wire was also 

 wound round one of the circuits, M, close up to this mutual square of iron. It is evident 

 that this mutual square could be demagnetised by decreasing reversals of either circuit, 

 and that these are at right angles to one another. 



The magnetic circuit M was now demagnetised by decreasing reversals. The 

 magnetising force was then re-applied by increments and the induction measured at 

 each step by means of the exploring coil and ballistic galvanometer. The circuit M was 

 again demagnetised, followed in this instance by decreasing reversals of the magnetising 

 force in the circuit N. This left the square common to both circuits demagnetised by 

 means of a force at right angles to M. The induction in the M circuit was again 

 measured as before. It was found that for low values of the magnetising force the 

 induction was lower when the iron square common to both circuits had been demagnetised 

 by the force at right angles. (The same result was obtained when the iron circuit M 

 was opened out and did not form a closed magnetic circuit.) These experiments were 

 repeated many times, always with consistent results. The difference, which was of the 

 order of 5% (the iron circuits being closed), entirely disappeared as the induction increased. 

 It ought to be borne in mind that the mutual square formed only a small part of the 

 magnetic reluctance of the circuit. 



These preliminary experiments, made under different conditions, confirmed the 

 supposition that the observed differences in the shielding ratio were essential and not 

 spurious effects, and were due to a difference in the permeability of the iron at the early 



