Air-Blast upon tlie Spark Discharge of a Condenser. 679 



inserted as an added resistance in the circuit o£ the cymo- 

 meter. The cymometer consists of a circuit including a spiral 

 wire having inductance L and a condenser of capacity C. 

 which can be continuously varied in the same proportion, so 

 that the oscillation constant (\/CL) o£ the circuit can be given 

 any value between certain limits. The cymometer was em- 

 ployed to take a resonance curve of the spark circuit by the 

 usual Bjerknes-Drude method, (i.) when the spark-balls were 

 not subjected to the air-blast, and (ii.) when the jet of air was 

 thrown between them. 



An extremely steady jet of air for this purpose can be 

 obtained by the use of a small Lennox blower, which is a fan 

 driven by an electric motor taking a, current of about 1 ampere 

 from any electric-lamp supply circuit. The resonance curve 

 is obtained by plotting the values of the root mean-square 

 current (a) expressed as a fraction of the maximum current 

 (A) induced in the cymometer circuit by the damped oscil- 

 lations in the spark- ball circuit corresponding to various 

 values of the natural frequency n or oscillation constant for 

 any setting of the variable inductance and capacity of the 

 cymometer circuit. Then, if A is the maximum value of this 

 secondary current in the cymometer corresponding to a 

 certain natural frequency N or oscillation constant when 

 resonance between the circuits exists, and if 8 1 and 8 2 are the 

 logarithmic decrements per semiperiod of the oscillations in 

 the spark and cymometer circuits respectively, we have by 

 Bjerknes' formula 



*+M— (i-5)>/j£p 



provided that such values of n are selected that n does not 

 differ from N by more than say 5 per cent. If then a 

 known resistance R is added to the cymometer circuit, tnus 

 increasing its decrement by a known amount 8 2 ', an ^ a fresh 

 set of observations taken, another resonance curve can be 

 plotted which should lie wholly outside of the first when 

 their maximum ordinates of both curves are taken as unity. 

 From the formula given by Bjerknes we know then that 



A 2 (Si-t-^)^ = A 1 2 (S l + S 2 + ^)(S 2 + ^'), 



where A and A x are the mean-square values of the maximum 

 or resonance currents in the two cases. 



The actual observed quantity which gives us a is the 

 deflexion of the needle of the galvanometer in connexion 

 with the thermocouple pressed against the resistance wire 

 inserted in the cymometer circuit. If there is no air-blast 



