98 Cacly and Vinal — Electric Arc. 



circuit. The effect was decidedly better at atmospheric pres- 

 sure than at any reduced pressure. No oscillations, however, 

 could be detected when the arc was surrounded by the flame 

 of a Bun sen burner. 



hi H and acetone vapor. The tests with benzine vapor and 

 with illuminating gas indicated that the presence of hydro- 

 carbons around the arc greatly facilitates the production of 

 oscillations. In the attempt to find a carbon compound that 

 should give rise to less carbon dust, we tried a mixture of 

 hydrogen and acetone, as described by Fischer.* This was 

 produced by passing the hydrogen from a Kipp generator 

 through the upper part of a bottle containing acetone. It 

 was found better not to let the hydrogen bubble through the 

 liquid. In this manner enough acetone vapor was present to 

 permit of strong oscillations without causing a very rapid accu- 

 mulation of carbon dust. Nevertheless it was necessary to clean 

 the electrodes frequently, and at no time w r ere the oscillations 

 comparable in constancy with those from the spark or Dud dell 

 arc. Using the most careful precautions, we could not keep 

 the oscillations from being fickle and irregular. Hardly ever 

 could a resonance curve be obtained without the necessity of 

 bringing the electrodes together, rotating one of them, or other- 

 wise changing the character of the oscillations. Hence the 

 values of frequency given below can be considered only as 

 rough approximations. Nearly all the observations to be 

 described were obtained with the discharge in hydrogen and j 

 acetone. 



The frequency in illuminating gas or acetone is much greater 

 than in air. The carbon compounds around the discharge seem 

 to make the cathode base, on the glow phase, more concen- 

 trated, and thus to accelerate its rise in temperature. The 

 discharge is steadier and more quiet than in air, and the 

 frequency is comparatively constant, being determined by the 

 electrical constants of the circuit (cf. §4). 



§11. Tests with Different Metals. Any metal, or carbon, 

 may serve as anode, though when carbon is used the produc- 

 tion of soot is annoying. As cathode we have tried Pt, Fe, 

 Ag, Cn, Al, Pb, and soft solder. All give initial oscillations 

 of apparently the same intensity and frequency, though altera- 

 tion of the surface of some of these soon destroys the effect. 

 No difference in frequency could with certainty be detected 

 whether the cathode was of Ag, Cu or Al. A fine copper wire 

 produced the same effect as a large disc. With carbon as 

 cathode only the feeblest oscillations could be observed — 

 though it was with this aiTangement that the oscillations were 



* Fischer, Ann. Phys., xxviii, p. 57, 1909. 



