Electric Discharge on Gases and Vapours. 525 



dissociation by the discharge and the consequent diminution 

 of pressure. I do not think, with Thomson, that the molecules 

 cannot, while the atoms can conduct. 



I think I have shown that while pure hydriodic-acid 

 vapour does not conduct to any appreciable extent at ordi- 

 nary temperatures when pure and dry, it does conduct when 

 diluted with any diluent. Further, that electrolysis of vapours 

 is only possible to any appreciable extent when the mechanical 

 action of the spark has caused such dilution of the pure 

 substance beforehand by dissociating it in part. 



Further, I made an extended series of experiments on the 

 action of the electric discharge on the following gases and 

 vapours : viz. carbon tetrachloride, chloroform, carbon disul- 

 phide, silicon tetrafluoride, marsh-gas, olefiant gas, and coal- 

 gas. 



Where one of the products of decomposition is a solid, it is 

 for obvious reasons not possible to use the apparatus above 

 described to show a polar distribution of the products of 

 decomposition. Now all these substances furnish solid de- 

 posits excepting silicon tetrafluoride, which was not decom- 

 posed. 



The hydrocarbons yield deposits of carbon, and therefore it 

 was impossible to find a difference in the gases furnished by 

 our apparatus. However, a remarkable fact was observed, 

 which I believe speaks strongly in favour of actual electro- 

 lysis, in the case of hydrocarbons especially. It is that the 

 deposit of carbon for these latter always grew out from the 

 positive pole in a way that would indicate its having been 

 separated out at that pole and not mechanically projected 

 against it. The experiment is most conveniently made with 

 coal-gas, and the sparking done in a cylinder over mercury. 

 The carbon grows from, the positive pole in form of a very 

 fine filament, very hard and very brittle, and possessing most 

 excellent conducting power. The film is very thin, indeed 

 thinner than a horse-hair. On examining it under the 

 microscope it is found to be toothed in a very regular way, 

 the teeth being sharp cones projecting towards the negative 

 pole. The deposit could not possibly have been formed by 

 mechanical projection of the carbon. From its character the 

 carbon has not been carried to the pole and deposited there, 

 but is, it seems to me, a true electrolytic separation. In any 

 other way than this it would be difficult to explain the den- 

 dritic forms. Another thing that speaks against the mecha- 

 nical projection supposition is that the filament is a mere 

 slender thread. These forms have the greatest likeness to, 



PhiL Mag. S. 5. Vol. 33. No. 205. June 1892. 2 N 



