Induction- Coil Potentials. 569 



frequencies used in this experiment the induction coefficients 

 of the coils are not very different from their values for 

 steady currents. 



The Coefficient of Coupling. 



The coupling coefficient, k, was determined by finding the 

 frequency of the oscillations of the secondary coil (connected 

 only to the oscillograph and ball electrodes), with the primary 

 coil open and with the primary terminals connected by a 

 short piece of wire. The ratio of the squares of these fre- 

 quencies is 1 — P. These oscillations were started by sparking 

 to the ball electrodes with two pointed conductors connected 

 to a second induction-coil, and the curve of potential was 

 photographed in the usual way. With the primary closed 

 the oscillation curve only appeared with great difficulty, and 

 many exposures were taken before a photograph sufficiently 

 good for measurement was obtained. The amplitude of the 

 curve was small, but was sufficient to allow the period to be 

 determined with fair accuracy. The frequency was found 

 to be 1240. With the primary open the frequency was 

 599*4. This was measured near the end of the curve where 

 the amplitude, and consequently the magnetizing current, is 

 small, in order to minimize as far as possible any error 

 arising from the difference in the degree of magnetization 

 of the core in the two experiments due to the shieldin en- 

 action of the induced current in the primary when on closed 

 circuit. 



These frequencies give the value *768 for k 2 . The value 

 given in the former paper * for the same two coils was 

 •8142. This was determined by a similar method, in which, 

 however, a condenser was connected to the terminals of the 

 secondary coil. It appears, therefore, that the coefficient of 

 coupling of the coils (for natural oscillations) is less when the 

 terminals of the secondary coil are free or connected to 

 bodies of very small capacity, than when they are connected 

 to a condenser of considerable capacity, in which case the 

 oscillating current in the secondary is more nearly uniformly 

 distributed. 



The matter was further tested by determining the coupling 

 coefficient with the secondary coil connected to various 

 capacities, the largest being a leyden-jar of about *001 

 microfarad, and the smallest consisting of two tinfoil sheets, 

 19*94 square centimetres in area, separated by a glass plate 



* X. c. (2), p. 716. 



