Dissociation of Water Vapour. 719 



The resistance of the wire at 0° C. and 100° C. was 

 determined from time to time during the course of the 

 research, but it remained practically constant unless the wire 

 was heated above 1670°. The irregularities in the resistance 

 determination which begin about this temperature are due 

 either to the volatilization of platinum, or to changes in the 

 wire itself. On one occasion a wire was heated to 1760° 

 for a short time and then examined, when it was found to 

 have become brittle, and apparently crystalline. 



It was noticed by Langmuir that when a wire was heated 

 to a very high temperature in water vapour, the amount of 

 apparent decomposition (i. e. volume of electrolytic gas col- 

 lected) rapidly diminished, but that if the glass surfaces 

 surrounding the wire were cleaned, the decomposition at 

 once increased to the normal amount. He explained this on 

 the supposition that platinum was. sprayed on to the glass 

 from the wire, and that it caused recombustion of the elec- 

 trolytic gas. Platinum was undoubtedly deposited on the 

 walls of the large bulb used by the author, but a number of 

 experiments on this point showed that the amount of electro- 

 lytic gas which was caused to recombine was negligible, as, 

 owing to the size of the bulb, the platinum deposit did not 

 get heated sufficiently to exert a strong catalytic effect. 

 Practically the same values were obtained before and after 

 cleaning the globe. 



The electrolytic gas resulting from the decomposition of 

 the water vapour was collected and analysed in a vessel of 

 the shape shown in fig. 1. 



This gas-collector is a modified form of that previously 

 used by Chapman and Lidbury (J. C. 8. Trans. 1902, p. 1301), 

 but it is specially suitable for the estimation of small quantities 

 of gas, since the volume is determined by the weight of 

 mercury displaced. 



The collected gas always contained an excess of hydrogen. 

 This varied considerably in amount, and did not seem to bear 

 any direct relationship to the volume of electrolytic gas ; 

 sometimes it would be almost 30 per cent, of the total 

 amount of gas, and at other times not more than 1 per cent, 

 or 2 per cent. 



A number of experiments undertaken to find the cause of 

 the variations in the amount of this excess gave the following- 

 results : — 



(1) Moist oxygen is absorbed to a far greater extent than 

 was imagined by the mercury in the fall tube of 

 the pump. When the pump was working slowly, 



