56 Mr. J. J. Thomson on some Electromagnetic 



do not know enough about the action on magnets of currents 

 first in one direction and then the opposite to say how much. 

 The results obtained when the condenser was put in the 

 primary appear to have important bearings on the theory 

 of Kuhmkorff's coil. The secondary of a RuhmkorfFs coil 

 will no doubt possess very considerable electrostatic capacity; 

 this will, as we have seen, give rise to local currents, which 

 exist although the secondary be closed. Now these local cur- 

 rents will interfere with the main current, and will fritter 

 away energy. The experiments just described show that the 

 introduction of a condenser into the primary circuit will dimi- 

 nish very materially the intensity of these local currents, and 

 so cause much less energy to be frittered away. I therefore 

 think it probable that part of the cause of the superior effici- 

 ency of a coil provided with a condenser in the primary is due 

 to this effect of the condenser. 



Some points of interest arise when we keep EF out of the 

 secondary circuit, still keeping the magnetizing-spiral in the 

 secondary. Thus, the intensity of magnetization of a needle 

 placed in the magnetizing-spiral is found to be different, accord^ 

 ing as the coil EF is closed, has its ends free, or has them con- 

 nected with the plates of a condenser. 



We shall begin with the case when the secondary ABCD 

 is closed. Then with EF closed, a needle vibrated 24 times 

 a minute ; with EF open, 44 times; and with the ends E and 

 F connected with the plates of a condenser, 43 times. In 

 another experiment a needle vibrated 23 times with EF closed, 

 33 times when EF was open, and 36 times Avhen E and F 

 were connected with the plates of a condenser. It is easy 

 to see that the magnetization should be greater when the 

 coil EF is open than when it is closed; for when it is closed, 

 an induced current is formed in it, which, since the electro- 

 kinetic momentum of the coils is the same after the break as 

 it was before, must be at the expense of the current in the 

 secondary, and thus the current in the secondary is diminished. 

 But it is not so evident, a 'priori, why joining the ends E and 

 F to the plates of a condenser does not make the coil EF 

 behave almost as if it were closed ; we can, however, I think, 

 find an explanation by going to the dynamical equations. 



Let L denote the coefficient of mutual induction between 

 the primary and secondary circuits, M the coefficient of self- 

 induction of the secondary, and N the coefficient of mutual 

 induction between the secondary and the coil EF; then if 

 i, j, k denote the currents through the primary, secondary, and 

 the coil EF respectively, the dynamical equation is 



