of the Coefficients of Induction. 465 



will suddenly diminish as the change of current alters from an 

 increase to a decrease, or vice versa. 



The general shape of the curve connecting L with ^ will 

 be the same whether the value of «£) has been diminished to 

 zero from positive or from negative values before the experi- 

 ments are made. This follows from the fact that the slope at 

 a point on a Ewing's cycle increases from each cusp to the 

 point of inflexion. The coefficient L will be a maximum for 

 those values of the magnetizing current C at which the points 

 of inflexion occur. As the cycle only has one point of in- 

 flexion on each branch, the two parts of the curve connecting 

 L with will each have one maximum point, and one only. 

 The two maximum values of L will not, however, occur at the 

 same value of the current. They will occur at a small positive 

 value of C for currents changing in the positive direction, and 

 at a small negative value of for currents changing in the 

 negative direction. From this it follows that for increasing 

 currents, L will at first increase and then diminish; while for 

 currents decreasing to zero, L will continually increase. This 

 statement must be modified for small cycles, which do not 

 generally have points of inflexion. Here the coefficient will 

 increase continually from one cusp to the other, and decrease 

 discontinuously on rounding each cusp. 



Experiments were also made on the field-magnets of a 

 Gramme dynamo, the currents used varying in this case up to 

 30 amperes. The same kind of results were obtained. 



12. Several remarkable effects were obtained, due to pre- 

 vious history of magnetization, which are easily explicable 

 on reference to the researches of Professor Ewing*. 



Suppose the current has been altered until a point P (see 

 fig. 1) on the curve of magnetization is reached. The effect 

 of a small change of current will now depend upon whether 

 the current be increased or diminished. If the current be 

 changed in the same way as it was altered last, a point A on 

 the curve of magnetization will be reached ; while if it be 

 changed in the opposite direction, a point B will be reached. 

 The slope of the line PA will be quite different from that of 

 PB, and as these slopes represent the values of the coefficient 

 of self-induction at the point P, it follows that this coefficient 

 has always two distinct values whatever the state of mag- 

 netization and however that state has been attained. The 

 curves obtained by Professor Ewing imply that the coefficient 

 is greater for a change of current in the same direction as the 

 last than for one in the opposite direction ; for if P be any 



* "Experimental Researches in Magnetism," Phil. Trans, part ii. 



1885. 



