Water- Vapour in contact ivitli Hot Platinum Wire. 631 



Although it was recognized that a gas heated by contact 

 with a fine platinum wire must be very rapidly cooled, yet it 

 was feared that as the cooling took place there might be an 

 appreciable amount of recombination in the immediate 

 neighbourhood of the hot wire, owing to the catalytic action 

 of the platinum vapour thrown off from the incandescent 

 metal. 



The amount of the dissociation would obviously bo reduced 

 by this source of error, and the magnitude of the discrepancy 

 between it and the observed values would become greater as 

 the temperature of the wire, and the consequent amount of 

 metallic vapour thrown off it, increased. 



It was not evident at the time when these determinations 

 were made, that it could be independently ascertained whether 

 the error in question influenced the results to an appreciable 

 •extent. More recently, however, while performing a series 

 of experiments for a different purpose, the extreme rapidity 

 with which the separation of oxygen and hydrogen took place 

 in pure water-vapour at about 15 mm. pressure was noticed ; 

 and the idea occurred that, by taking cognizance of this fact, 

 an experiment could be devised to test whether any combi- 

 nation of the electrolytic gas took place in the immediate 

 neighbourhood of the wire. 



The principle of the proposed experiment was that the 

 water-vapour and electrolytic gas should be withdrawn both 

 from the surface of the wire and at some distance away from it. 

 If no recombination occurred during cooling, then the oxygen 

 and hydrogen should be present in both samples in combining 

 proportions; but in the event of a partial reunion, the sample 

 of gas taken from the immediate neighbourhood of the wire 

 should contain an excess of oxygen, whereas that taken 

 simultaneously from another portion of the water-vapour 

 should be correspondingly rich in hydrogen. For, when a 

 "wire is heated in water-vapour, electrolytic gas is produced 

 at its surface. Diffusion would at once take place, and it' 

 no recombination occurred, the oxygen and hydrogen would 

 become uniformly distributed in combining proportions 

 throughout the vapour, since any diffusion back of the 

 hydrogen from the cooler towards the hotter layers of 

 vapour, would be exactly compensated by the direct outward 

 diffusion from the hot wire. 



If, however, recombination took place in a zone at some 

 little distance from the surface of the wire, its effect would 

 be to diminish the amount of electrolytic gas in the cooler 

 and more distant portions of the water vapour ; so that 



