470 M. I. Pupin — Electrical discharges through poor 



latter part of this paper may perhaps furnish a clue in tracing 

 the connection between sunspots and auroral discharges). 



III. PHENOMENA INDICATING A TRANSLATIONAL MOTION OF 



THE GAS. 



An interesting phenomenon was observed in the experiments 

 with bulbs A, B, fig. 8, when the vacuum was diminished by 

 turning a stop-cock C several times around. The vacuum pres- 

 sure was about 20 mm . The induction coil had to be strained 

 considerably to force a discharge through the long glass tube. 

 The discharge looked like a luminous jet shooting from the 

 tube into the bulbs, and in its path around the corners it seemed 

 to strike against the necks of the bulbs at a and l> from which 

 points it was reflected and glided along the surface towards 

 the points e and f. Inside of the bulbs the jet oscillated 

 rapidly ; it was also split np in several parts, each part consist- 

 ing of numerous more or less intense streamers. A slight 

 modification in the curvature of the necks modified the general 

 outline of the luminous jet without changing its general charac- 

 ter. With the increase of the gas pressure the phosphorescence 

 appeared and seemed to be strongest at a and b. It was very 

 strong in the tube C, leading to the stop-cock, although this 

 tube was entirely free from the discharge proper. The height 

 to which the phosphorescence rose in this tube increased with 

 the current. Every slight variation in the current strength 

 caused a simultaneous variation in the height of the phosphor- 

 escent column in C. (When the discharge ceased there was. a 

 strong phosphorescent after-glow all along the long tube). A 

 similar behavior on the part of the phosphorescent gas, which 

 I observed in the experiment described in this Journal, April, 

 1892, leads to the conclusion that the phosphorescent gas must 

 have a translational motion due to its being ejected from the 

 path of the discharge proper. This strengthened my belief in 

 a transitional motion of the gas along the path of the discharge 

 proper, which belief was due to the phenomena in the experi- 

 ment just described. It was also strengthened by the phe- 

 nomenon observed in the experiment described by me in the 

 paper cited above, the phenomenon namely that a discharge 

 streamer made to curve out (by the repulsive action of another 

 parallel streamer) so as to strike against the walls of the vacuum 

 jar will rebound from this wall just as a curved jet of water 

 would if it struck against a rigid surface. In another experi- 

 ment with an apparatus similar to that described in the above 

 paper a thin rectangular sheet of mica was suspended between 

 and parallel to two discharge streamers and it was found that 

 it prevented their action which I described in that paper, but 



