PIERCE. THK COOPER HEWITT MERCURY INTERRUPTKR. 407 



Having in tliis way obtained the amount of heat required to give one 

 centimeter of expansion on the scale of eacli of the calorimeters, the 

 resistance of the mercury interrupter for the oscillatory discharge was 

 measured for various inductances and capacities of the discharge circuit. 

 The resistance of the interrupter was found to vary with the capacity 

 and inductance. When this resistance was large, the calorimeter B was 

 used in the comparison; when it was small, C was used. To make sure 

 that the two comparison calorimeters were consistent, the resistance of 

 the interrupter was occasionally measured by both B and C and found 

 to give concordant results. 



In making the final measurements with the oscillatory current the 

 mercury interrupter in the calorimeter A was put in the discharge circuit 

 in series with the known resistance of one of the comparison calorimeters. 

 The secondary of the transformer was connected to the condenser, a 

 switch in the primary of the transformer was closed, and the condenser 

 was allowed to charge and discharge for a time varying from twenty to 

 sixty seconds. The expansion of the oil of the two calorimeters was 

 read, and then observations on the cooling of the calorimeters were taken 

 for four minutes, so that the correction for cooling could be estimated. 



At the end of the series of resistance measurements, the inductances 

 and capacities of the various circuits were measured with an accuracy of 

 about one per cent by the photographic measurement of the time of the 

 condenser discharge. A sketch of the method of the computations for 

 this purpose is given in section IV, p. 399 of this paper. The advan- 

 tage of this method of determining the constants of the circuit is that it 

 gives these constants for the required frequency. 



The following tables (Tables IV, V, and VI) give a series of results 

 for the resistance of the interrupter for various capacities and inductances. 



From these tables it is seen that the resistance of the mercury inter- 

 rupter decreases with increasing capacity of the condensers, and decreases 

 with decreasing inductance of the discharge circuit. 



These facts might perhaps be anticipated from the relation between 

 the voltage and the current in the mercury arc with direct current, and 

 the relation between the capacity, inductance, and current in the con- 

 denser discharge. 



For direct currents greater than 3 amperes the voltage about the 

 mercury arc is practically constant, and equal to 16 volts for the partic- 

 ular bulb here employed ; therefore, 



I R = constant = 16. (1) 



