UPON MAGNETISATION, AND CONVERSELY, IN IRON, STEEL, AND NICKEL. 495 



been tabulated. On the other hand, the effect of the load in decreasing magnetisation 

 in nickel is very much greater in the annealed than in the quenched condition. 



Note on the Effect of Torsion. — When the annealed nickel wire was linked 

 directly to the lever, and not by means of a thread, as is here the case, the effect of 

 the load to reduce the magnetisation was less. The load curve might even exceed 

 the normal curve without load for a short distance where clB/dH. was greatest. I 

 could trace this rise to no other source than a minute torsional effect which disappeared 

 entirely when the method of connecting the wire eliminated such a possibility. The 

 effect could be reproduced in an exaggerated form by twisting the wire a very few 

 degrees per 100 cms. of length. 



Strictly speaking, therefore, the experimental results are based upon the effects of 

 vibrations upon that particular condition which has been reached by subjecting the 

 magnetic metals first to annealing and then to demagnetisation by decreasing reversals 

 with a load not exceeding 0'5 kilo, per sq. mm. of sectional area, the load remaining on 

 throughout the experiments. It may be observed that, while load may either increase 

 or decrease permeability, the effect of vibrations is always to increase permeability. 



A few experiments were made in which the wires were soldered to the gong and no 

 load used. These will be referred to later. 



Superposition of Vibrations and Field. 



The importance of the order and manner in which mechanical vibrations and field 

 (magnetisation) may be superposed the one upon the other has already been mentioned. 

 The order of superposition is distinguished in the same way in which the superposition of 

 electric oscillations and field was distinguished in the paper already referred to. 



A. Mechanical vibrations are superposed upon constant field. 



B. A change of field is superposed upon mechanical vibrations permanently acting. 



Experimental Methods under A Conditions. 



After demagnetisation by decreasing reversals (the revolving commutator being in 

 all cases used), the field is put on by steps of increasing reversals,* followed by thirty 

 to thirty-five reversals of the pre-arranged field maximum. One-half the average of two 

 consecutive galvanometer readings determines the value of the induction at this field 

 maximum. Single steps (the first being zero) are then taken from the fixed maximum 

 to a sufficient number of points all round the normal loop. At each point, mechanical 

 vibrations are superposed by simply ringing the electric bell. This necessitates 

 demagnetisation, followed by thirty reversals of the maximum field value after each 

 step taken and before the next observation is made. 



Two curves result from the galvanometer readings taken during this process. The 



* This process tends to preserve the symmetry of the loops with reference to the origin. 



