392 Dr. Norman Campbell : Experiments on 



It appears that the ratio is more nearly constant for the value 

 o£ Lj determined from the period of the oscillations, but an 

 intermediate value would be better than either. The primary 

 self-inductance appears to vary with every method used for 

 measuring it. 



General Conclusion* 



20. It seems, then, that the theory of the magneto which 

 neglects the effects of interaction between the circuits and 

 the iron core is not applicable in any important respect to 

 machines of ordinary construction. It will predict neither 

 the absolute values of the peak potential nor the variation of 

 that potential with the primary capacity and the coupling. 



An important improvement in the agreement between 

 theory and experiment can be produced by inserting in 

 series with the primary circuit a choking coil with an air- 

 core ; even with this coil in place the quantitative predic- 

 tions of the theory are not fulfilled, but there is a qualitative 

 agreement between theory and experiment concerning the 

 relation between the peak potential and the constants of 

 the circuits. 



These results were very disappointing in view of the pur- 

 pose for which the work was undertaken, namely the practical 

 improvement of the machine. It seems that the only con- 

 siderable improvement which could be obtained would be 

 produced by decreasing greatly the losses in the iron core ; 

 no alteration of the coupling, the inductances, and the 

 capacities which is practically possible seems likely to lead 

 to any useful result. The inclusion in the primary of the 

 air- core choking coil increases the peak potential obtainable 

 with a given primary current, but in practice the decrease 

 which it would produce in the primary current would more 

 than counterbalance the increase in the ratio of peak potential 

 to primary current. In order to avoid the decrease in the 

 primary current it would be necessary that the choking coil 

 should have a resistance small compared with that of the 

 primary in the armature ; to attain this condition and at the 

 same time make its inductance comparable with that of 

 the armature would involve the use of a mass and volume 

 of copper which is quite outside the bounds of practical 

 possibilities. 



How far a reduction of iron losses is possible is a matter 

 to be determined by further investigation. If hysteresis is 

 the main source of loss, little improvement can be hoped for; 

 but this alternative is not very probable. If hysteresis were 

 the main agent, the loss and the damping should vary rapidly 



