474 M. I. Pupin — Electrical discharges, etc. 



b c of narrow bore. An L-shaped tube a with a small orifice 

 was fitted by means of a rubber stopper into the neck of the 

 reservoir. By means of this tube and the tube d of the prin- 

 cipal reservoir communicated with two other reservoirs, which 

 I call the externa] reservoirs. 



The external reservoir connected with d communicated 

 with a mercury pump. "When the exhaustion had reached 

 the point at which a steady rectilinear discharge could be 

 forced from b to c, a stopcock connecting d to its external res- 

 ervoir was shut off and the exhaustion continued until a good 

 vacuum was obtained in the external reservoir d. The dis- 

 charge was then started. It was a perfectly steady, narrow, 

 rectilinear column of crimson luminosity, surrounded by a 

 phosphorescent ellipsoidal column. But as soon as the above 

 mentioned stopcock was turned on, the blast coming from the 

 orifice a played up the column and the rectilinear path became 

 curved at the point where the blast was acting. The dis- 

 charge acted as if it bent around to get out of the way of the 

 blast. The observation that the effect of the blast upon the 

 phosphorescent column was incomparably stronger than upon 

 the crimson column needs no comment. The weaker the dis- 

 charge the stronger is the effect of the blast, and vice versa. 

 The effect of a blast upon the oscillatory spark discharge of a 

 powerful Ley den jar battery is not perceptible. 



In discharges through very poor vacua the heating effect is 

 very unequally distributed throughout the vacuum jar. The 

 temperature at certain points is enormously higher than at 

 others. The result is that a very violent motion of the gas is 

 set up, which motion may sometimes, on account of the 

 effects pointed out in the last experiment, produce streamers 

 of double curvature. 



During an experiment with the apparatus given in fig. 10, 

 which I performed on February 8th, 1892, before the astro- 

 nomical section of the JSTew York Academy of Sciences, a leak 

 occurred so that the vacuum was exceedingly poor by the time 

 I was ready to start the discharge. Finding that the discharge 

 gave no sign of starting, I risked turning the whole power of 

 the 10 h. p. alternator on the induction coil. Long sparks shot 

 immediately from almost every point of the edge of the tin- 

 foil. The discharge between the brass sphere and the tinfoil 

 looked like a black faced Medusa with fiery serpents dancing- 

 all around her head. On repeating this experiment I found 

 that in very poor vacua the discharge streamers very often 

 assume the form of spirals of very long pitch. 



All these phenomena suggested to my mind a very strong 

 similarity between the streamers of an electrical discharge 



