684 Dr. J. A. Fleming on Magnetic Oscillators 



through the constantan wire and observing the corre- 

 sponding scale-readings of the Paul galvanometer. The 

 curve plotted out proved to be very approximately a parabola, 



Fig. 1. 



the galvanometer deflexion being very nearly proportional 

 to the square of the heating current sent through the con- 

 stantan wire. When electric oscillations were set up in the 

 circuit they heated the constantan wire and caused a 

 deflexion o£ the Paul galvanometer from which by the 

 previous calibration their mean-squre value could be deter- 

 mined. In addition the oscillation valve V joined in as shown 

 in fig. 1 passed a continuous current through the galvano- 

 meter G in its circuit, and readings were taken of this current 

 in microamperes corresponding to various measured values 

 of the oscillatory current in the square circuit which passed 

 through the constantan wire. 



The object of this calibration was to be able to use the 

 valve subsequently as a means of measuring the oscillatory 

 current induced in the circuit when it had a value less than 

 that which could be easily measured by the therm ojunction, 

 for when associated with a suitable mirror-galvanometer it 

 provides the means of measuring a high frequency current 

 of a few microamperes. Moreover it does not introduce 

 any extra resistance into the circuit, as would have been the 

 case if the constantan wire and themojunction had been 

 permanently inserted into the circuit. Observations showed 

 that the galvanometer current did not increase proportionately 

 to the condenser plate potential difference, but tended to a 

 " saturation " value as shown by the curve in fig. 2. 



The large (8-feet) square coils were then set up in the 

 quadrangle of University College, one being used as a trans- 

 mitter and the other as a receiver circuit. 



