W. G. Cady — Electric Arc between Metallic Electrodes. 245 



self-inductance greatly increased it, since the penetration of 

 the oscillations was stopped by reflection from the coil. 



§ 23. Varying the supply e.m.f . from 600 to 1000 volts had no 

 effect upon the intensity or frequency, as long as the current 

 remained unchanged. Reversing the direction of the current 

 also had no effect. 



§ 24. Increasing the current increased the frequency, as 

 illustrated in the following table, in which the last column 





Table III. 





Generator 









current 



Frequency 



Deflection 



0-09 amp. 



8-6 X 



10 5 



49mm 



o-io " 



8-1 



u 



104 



0*11 " 



9*1 



a 



109 



0-12 " 



9-3 



a 



56 



0'13 " 



9-45 



a 



39 



gives the hot-wire deflections, serving as a measure of the 

 intensity of the oscillations. The frequency observed is that 

 of the third harmonic. If the intensity of the oscillations 

 diminished with increasing current, it would be possible to 

 account for the increase in frequency on the supposition that 

 with decreasing intensity the oscillations penetrated less far 

 into the circuit, i. e. that the effective length of the " oscilla- 

 tor" was less. But the third column shows that the intensity 

 was greatest for a current of 0*11 amp., and diminished as the 

 current increased or decreased (§ 8). The change in frequency 

 must be due to the changing resistance of the heated vapor, 

 which affects the frequency according to the formula 



_ 1 /~T~ ~W~ 



f ~ 2i,y LC UD' 

 An increase in the discharge current, decreasing the resistance 

 R between the electrodes, must clearly raise the^ frequency. 

 The apparent mean resistance, derived from the measurement 

 of current and e.m.f., varies under different conditions between 

 2000 and 5000 ohms. It is hardly conceivable that the resist- 

 ance for oscillations can be as high as this, but it is evident, at 

 least, that the discharge possesses a high resistance, which 

 changes markedly as the current is varied. 



As no means suggested itself for measuring directly the 

 resistance of the discharge for oscillating currents, I tried the 

 effect of inserting a large known non-inductive resistance 

 (graphite rod) in the circuit close to the discharge. This 

 experiment failed to throw light on the problem, for when the 

 resistance was large enough to modify the frequency at all, it 

 absorbed the oscillations and reduced their penetrating power 

 to such an extent that the frequency was increased instead of 

 decreased. The resistance of the discharge would not be 



Am. Jour. Sci. — Fourth Series, Vol. XXVIII, No. 165.— Septembeb, 1909. 

 17 



