484 Dr. Norman Campbell on the "Break" 



made very readily on any magneto. If we remove the 

 magnets and bring the primary terminals outside the 

 machine, we can supply from a battery a primary current 

 which is independent of the speed*. With any given 

 primary condenser and any given speed, it will be found 

 that, if the current is below a certain limit, no sparking 

 occurs, bat that if it is increased above this limit, sparking 

 appears. The limiting value can be fixed within com- 

 paratively narrow limits ; in one experiment, for instance, 

 it was found to lie certainly between 22 and 2*5 amp. 

 Suppose now that we supply a current double the limiting- 

 value ; then we shall very nearly (but not quite) double the 

 rate of rise of potential across the terminals. Suppose now 

 we double the speed of the magneto shaft and so double the 

 speed of separation of the terminals ; then at any moment 

 during the rise of potential the relation between the spark 

 potential and the potential difference between the terminals 

 should be exactly the same as it was at some corresponding- 

 moment with the original values of the current and the 

 speed. This conclusion involves only the assumption that 

 the rate of rise of potential is proportional to the current 

 broken; it does not involve any assumption concerning the 

 relation between the spark potential and the distance. 

 Accordingly the current which is now broken in the primary 

 should again be the limiting current ; in other words, by 

 doubling the speed of break we should expect very nearly 

 to double the limiting current. But as a matter of fact we 

 find that the limiting current is not doubled or nearly 

 doubled by doubling the speed ; no certain evidence could 

 be produced that increasing the speed of the break 20-fold 

 produces a change in the limiting current which is greater 

 than the error (10 or 20 per cent.) in the determination of 

 that current. This conclusion appears absolutely inconsistent 

 with the theory which has been given of the circumstances 

 which determine the occurrence of sparking. 



(3) Effect of the electrodes. — The clue to this apparent 

 failure of a theory, so obvious that it has not generally been 



* The same experiment can be tried without disconnecting' the 

 primary circuit by making use of the fact that the primary current in a 

 magneto is almost independent ot the speed over a considerable range ; 

 it is usually possible within this range to change the speed 100 per 

 cent, without changing the primary current more than 10 per cent. ; 

 and if we start on one side of the maximum and decrease or increase the 

 speed through the maximum to the other side, we can vary the speed 

 more than 100 per cent, and regain the original value of the primary 

 current. The value of the maximum current can be readily changed by 

 changing the strength of the magnets. 



