yo =-_a. 



considered in connexion with the Dynamical Theory. 9 



Now, according to the connexions, N = L or 5L; and so in the 

 first case the spark disappears, while in the second it falls short 

 of the full spark by only one-fifth. 



While considering the dynamics of the field of two currents, 

 I noticed that the initial induced current due to a sudden fall of 

 a given current in the primary wire is theoretically greater the 

 smaller the number of terms of which the secondary consists ; 

 for in calculating the energy of the field, it makes no difference 

 whether we have a current of any magnitude in a doubled 

 circuit, or twice that current in a single circuit. The same 

 conclusion may be arrived at by the consideration of the ana- 

 lytical expression for the initial induced current 



M 



N 



for if the secondary circuit consists essentially of a single coil of 

 n terms, we have, cceteris paribus, M oc n, while N oc n 2 , so that 



Vo oc - The whole induced cu.vreutj'ydt gcMqcw. Interme- 



M 2 

 diate to these is the heating-effect fy^dt, which oc -jt , and is 



therefore independent of n. Thus it was evident that neither the 

 galvanometer nor electrodynamometer were available for the veri- 

 fication of this rather paradoxical deduction from theory, at 

 least without commutators capable of separating one part of the 

 induced current from the rest. On the other hand, it appeared 

 probable that the smaller total current, in virtue of its greater 

 maximum, might be the most powerful in its magnetizing action 

 on steel. 



With the view of putting this idea to the test of experiment, 

 I bound three wires of *001 inch diameter, and about 20 feet 

 long, together into a coil whose opening was sufficient to allow 

 it to pass over the coil A. The ends of the wires were free, so 

 that they could be joined up in any order into one circuit, which 

 was also to contain the magnetizing spiral. It is evident that 

 if the currents are parallel in the three wires (an arrangement 

 which I will call a) 3 then 



M = 3M , N = 9N , 



M N being the values of the induction-coefficients for one wire ; 

 while if in the two wires the current flows one way round and in 

 the third the opposite (b), we shall haveM = M , N = N . Inas- 

 much as the self-induction of the magnetizing spiral was relatively 

 very small, these may be regarded as the induction-coefficients 

 for the secondary circuit as a whole. This arrangement was 

 adopted in order that there might be no change in the resistance 



