300 Experiments on the High- Tension Magneto. 



quantity c. For it is the peak potential which alone is of 

 much practical importance, while Oj and c are the quantities 

 which are most completely at the disposal of the designer. 

 According to the theory there is for each value of c an 

 optimum value of C 1? which gives the greatest peak potential. 

 The value of this peak potential with optimum Ci does 

 not vary very greatly with c ; so long as c lies between 

 and 0*5, the greatest peak potential obtainable does not vary 

 by more than 5 per cent.; but the optimum value of Ci varies 

 greatly. If c = 0, the value is 0; but it can be increased 

 considerably by increasing c. Now in practice a lower limit 

 is set to Ci by the necessity of avoiding sparking at the 

 break in the primary circuit ; indeed, until an adequate 

 theory of the machine was produced, it was thought that the 

 only purpose of inserting a condenser in parallel with the 

 primary terminals was to suppress such sparking. It is 

 therefore clear that, since this lower limit is set to Ci, the 

 optimum value of c will not be 0, but will be some higher 

 value such that the optimum value of Ci is not less than the 

 lower limit set by the necessity of avoiding sparking ; and 

 the theory indicates that the attempt, which seems to have 

 been made in the design of most actual instruments, to make 

 the coupling as close as possible and c as small as possible is 

 mistaken, and that better results would be obtained by de- 

 creasing the coupling and increasing c. The second part of 

 this paper will be devoted to examining how far this pre- 

 diction is correct. But it should be noted that evidence is 

 already available that the theory may fail in this respect al^o. 

 For the relation predicted by it between the peak potential, 

 d, and c is intimately connected with the occurrence of 

 " arches " (such as are shown in the full-line curves of figs. 4 

 and 6) in the curve relating the peak potential and Ci. Now 

 the experimental curves do not show these arches ; there are 

 indications of very shallow maxima and minima which scarcely 

 exceed the possibilities of experimental error, but there is 

 noihing whateA^er corresponding to the sharp cusps which are 

 so notable a feature of the theoretical curves. It must, 

 therefore, be doubtful from the outset whether the theoretical 

 relation is fulfilled. [To be continued.'] 



Summary. 



1. The object of the experiments is to investigate more 

 nearly the applicability of the theory of the high-tension 

 magneto which has been given by Prof. E. Taylor Jones in a 

 recent paper. 



