684 



Professor J. A. Flemiyig 



[May 24, 



two fine resistance wires, against one of which a sensitive thermo- 

 j unction of iron and bismuth is attached. This enables me to 

 measure the value of the current in the cymometer bar. The process 

 of measurement is then as follows : We place the cymometer alongside 

 the antenna and slide along the handle slowly, thus altering its time 

 period or natural frequency. We observe the current and frequency, 

 and plot a curve called a resonance curve showing the secondary 

 or cymometer current in terms of the frequency (see Fig. 4). This 

 curve rises to a maximum value, sometimes very sharply, the maxi- 

 mum corresponding to the condition of exact syntonism between the 

 antenna and cymometer circuits.* From the curve we can determine 

 the sum of the decrements of the cymometer and antenna. A second 

 experiment made with a known additional resistance inserted in the 

 cymometer bar enables us to eliminate the decrement (D) of the 

 cymometer itself, and thus find that of the antenna alone. When 

 this is done we know what percentage each oscillation in the antenna 

 is of the previous one. Suppose we agree that when the oscillations 

 have decayed away to 1 • per cent, of their initial value, the train 



Strongly Damped Oscillations 



Feebly Damped Oscillations 



Undamped 

 >Pscillations 



Fig. 5. 



shall be considered to be finished, then another simple formula 

 M = (4'6 05 4- D) / 2 D enables us to find the number of the com- 

 plete oscillations M in a train when we know the decrement D.j 



* If the damping of the secondary circuit is small, as it is in the case of the 

 cymometer circuit, then the resonance curve is very sharply peaked or rises 

 quickly to a maximum when the primary oscillations are feebly damped, pro- 

 vided always that the " coupling" or mutual inductance of the two connected 

 circuits is small. 



t See " The Principles of Electric- Wave Telegraphy," Fleming, p. 167. 



