Vacua, and on Coronoidal discharges. 469 



But if this increase in the conductivity is due to a rise in 

 the temperature of the gas along the path of the first dis- 

 charge and to nothing else, how can the fact be explained that 

 a long, thin, discharge streamer when forced through a poor 

 vacuum can be maintained steady, and permanent in form, 

 even if the discharge continues for several minutes? It 

 should broaden out continually and become more and more 

 diffused as the adjacent particles of the air get heated. In my 

 experiments on solitary discharge streamers in poor vacua (see 

 this Journal, April, 1892), I did not observe any appreciable 

 widening out, but I did observe a phosphorescent halo around 

 the streamer which, as Prof. J. J. Thomson assumes, (1. c.) 

 was very probably due to dissociated oxygen molecules that 

 were ejected from the path of the discharge. (See further 

 below the effect of a blast on a discharge streamer). 



Still another experiment which shows that something of the 

 nature of a dissociation of the gas molecules is going on along 

 the path of the discharge. A thick German silver wire, 60 cm 

 long, was bene zig-zag fashion into 12 zig-zag parts and placed 

 in horizontal position at the bottom of a bottle like the one in 

 fig. 9. A wire passing through a rubber stopper in the neck 

 of the bottle connected this zig-zag electrode to one of the 

 poles of the induction coil. The other electrode, a small 

 brass sphere, was vertically above the zig-zag electrode, imme- 

 diately under the rubber stopper. The shortest distance be- 

 tween the two was about 30 om . The vacuum was about 3 mm . 

 The discharge started between the nearest points of the elec- 

 trodes, that is between the lowest point of the sphere and one 

 extremity of the zig-zag electrode. It had the form of a band 

 about 3 cm wide, intensely luminous at each end, but only very 

 faintly luminous along the intervening three-fourths of its 

 length. The length of the less luminous interval increased 

 with the decrease of the e. m. f., but diminished with the 

 increase of the gas pressure ; it also seemed to have a different 

 color, but I did not care to examine this point more closely. 

 The phenomenon that interested me more was the gradual 

 creeping of the discharge along the zig-zag electrode from one 

 of its extremities towards the other. It did not increase in 

 breadth but left its trail along the zig-zag electrode in form of 

 a faintly luminous halo which surrounded this electrode just 

 like a narrow luminous tube. Both the color and the gradual 

 lateral motion of the discharge reminded me very much of the 

 aurora borealis of Feb. 13th, 1892. (In this connection it is 

 well to remark that when the e. m. f. is below the critical 

 point this auroral discharge can be started by powerful disrup- 

 tive discharges of a Leyden jar in its vicinity. This in con- 

 nection with observations on coronoidal discharges given in the 



