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



oscillations are integral multiples of the fundamental. More- 

 over in the present case it is not impossible that some one por- 

 tion of the circuit, whose natural period is not far from that 

 of one of the harmonics, may introduce a disturbing element. 



§ 28. By using in the resonating circuit a coil of very small 

 self-inductance, frequencies have been observed as high as nine 

 million. This must be a harmonic of the order of the thirty- 

 sixth, unless it is due simply to a small portion of the discharge 

 circuit in which feeble oscillations of very high frequency 

 were excited. 



§ 29. Are the oscillations undamped? — A satisfactory answer 

 to this question cannot be obtained from most of the resonance 

 curves. Some of the curves examined, including that in fig. 2, 

 gave as the sum of the decrements in the primary and second- 

 ary circuits values ranging as high as 0*124. As the secondary 

 decrement could be computed from the constants of the con- 

 denser circuit, the apparent decrement of the primary oscilla- 

 tions was found by subtraction. This gave, in the case of the 

 curves referred to, primary decrements of from 0*07 to 0'10. 



But on other occasions, when the discharge consented to take 

 }3lace steadily for a sufficiently long time, sharply-pointed reso- 

 nance curves were obtained, showing a primary decrement no 

 larger than O005. 



It looks as if the glow-arc oscillations were of themselves 

 undamped, the flatness of the curves being caused by the con- 

 tinually fluctuating frequency and intensity. If, on the other 

 hand, the discharge were a succession of true sparks, one would 

 exj;>ect the damping to be greater and the higher harmonics 

 less pronounced. 



There is much similarity between the production of these 

 oscillations and those studied by M. Wien for the generation 

 of undamped oscillations by " Stosserregung."* The essential 

 differences are made clear in §§ 9, 10, 13, and 14. Moreover, 

 in Wien's method the damping of the secondary circuit must 

 be small in comparison with that of the primary. Although 

 we do not know the resistance of the discharge-gap to the 

 oscillations described in this paper, still the' evidence of the 

 resonance curves points to a larger damping in the secondary 

 than in the primary circuit. 



§ 30. Quantity of matter liberated from the cathode per 

 oscillation. — During oscillations, the anode is not perceptibly 

 attacked. The cathode wastes away so slowly that it seemed 

 worth while to ascertain how small a fraction of a gram is 

 volatilized each time the discharge is on the arc phase. The 

 oscillations were accordingly allowed to pass as continuously 

 as possible for one hour, the current being 0*12 amp., funda- 

 *M. Wien, Ann. Phys., xxv, p. 625, 1908, 



