6o"2 Mr. A. Holt on the Decomposition of 



outward diffusion would be continually going on, with the 

 result that the gas which had the greater rate of diffusion 

 would be found in excess outside the zone of recombination, 

 while that with the slower rate would tend to concentrate 

 round the wire. 



Since these experiments on the dissociation of water-vapour 

 were performed, Langmuir (Journ. Amer. Chem. Soc. 1906,. 

 p. 1357) has carried out a very similar investigation, and has 

 obtained results iu close agreement with those of Nernst and 

 Wartenberg. He does not seem to have considered that any 

 appreciable source of error arose from this recombination in 

 the immediate neighbourhood of the wire, and this view is 

 confirmed by the experiments described in this paper: hut he 

 is of opinion that the highest temperature is reached at which 

 it is feasible to determine the amount of dissociation by this 

 method when the wire begins to disintegrate, and deposits 

 finely divided metal as a film on the inner surface of the glass 

 vessel employed. Some amount of recombination of the 

 electrolytic gas may take place in contact with this film, and 

 so cause considerable errors in the determinations. 



In his experiments, however, the apparatus employed was 

 only 0*9 cm. diameter, and was heated during each expe- 

 riment to at least 100° C. by the steam passing through it, 

 whilst 1 used a vessel 20 cms. diameter, the walls of which 

 hardly became heated during the short time occupied by each 

 determination. Under these conditions it is less probable that 

 there would be any large error from this cause, though it 

 is possible that it may be an explanation of the lowness of 

 the numbers at high temperatures. The catalytic action of 

 this spray and deposit is being further examined, and it is 

 impossible at present to say how far it really influences the 

 results. 



The values obtained for the amount of dissociation of water- 

 vapour are slightly higher than those of Langmuir and Nernst 

 and Wartenberg at low temperatures, but they steadily 

 become smaller as the temperature rises, till at 2000° absolute 

 they are only about \ of the quantities obtained by the other 

 authors. 



The cause of this marked discrepancy of results is being 

 further investigated. 



Experimental. 

 The apparatus employed for the determination of the 

 amount of dissociation consisted of a large glass bulb of 

 rather more than 4 litres capacity, in the centre of which a 

 pure platinum wire about 5 cms. long was hung. 



