2fi MR O. W. BICHAIv'DSoN ON THK IOXISATION 



conditions are apparently kept constant. The following numbers, which refer to a 

 temperature of about 900 C., illustrate the kind of thing that occurs. The wire, 

 which was 0'2 millim. in diameter, was maintained at a potential of -4-40 volts, the 

 pressure of the oxygen being atmospheric. All the conditions were kept constant ; 

 the unavoidable variation in the temperature of the wire was continuously registered 

 by the galvanometer spot, and was not sufficient to cause a variation of 5 per cent. 

 in the value of the leak. Nevertheless, readings at 2-minute intervals gave the 

 following values for the current : 146, 180, 178, 228, 158, 170, 150, 246, 166, 324, 

 198, 174, 198. 



Naturally the existence of a variability of this kind makes it very difficult to find 

 out what is the real effect of changing the controllable conditions, and the author has 

 spent a great deal of time in trying to get rid of it. In this he has been unsuccessful. 

 It looks, in fact, as though this variability is in some way or another an inherent 

 part of the phenomena. In some respects it seems to follow definite laws. It is 

 more marked at high than at low pressures, and at low than at high temperatures. 

 The positive leak at temperatures above 1200 C. seemed to show very little variability. 

 It is not due to trivial variations in the state of the gas in the tube, as the negative 

 leak, measured under the same conditions as the positive, did not show it. For the 

 same reason it cannot be due to discharges from points which might form on the 

 surface of the platinum. 



It was found to be present whatever the voltage on the filament. Boiling out the 

 tube with nitric acid and distilled water left it unaffected. Slightly heating the walls 

 of the tube with a Bunsen burner did not affect it, although heating the walls more 

 strongly was found to give the big leak mentioned on p. 19. The big leak thus 

 produced, on the contrary, was steady and did not vary capriciously with the time. 

 The variability was not due to vapour given off intermittently from the heated walls 

 of the tube, as it occurred in undiminished intensity in the tube whose walls, &c., 

 were cooled with water. It might have been ascribed to the intermittent escape of 

 occluded hydrogen, were it not that a wire which had recently been heated in 

 hydrogen did not show the effect to a greater extent than a wire which had been 

 heated in an atmosphere of oxygen over a period of about three months. 



There seems to be no escape from the conclusion that this effect is caused by some 

 periodic change in the state of the platinum surface. If the metal was continually 

 undergoing recrystallisation accompanied by the emission of absorbed gas, equilibrium 

 might possibly be incapable of ever being attained, and the results might simulate 

 those which have been observed. It does not seem advisable to speculate further on 

 this point. The real question from the point of view of this investigation is whether 

 the selection of the minimum values in the case of the measurements at high 

 pressures is legitimate or not. They certainly seemed far more definite than the 

 maximum or the average values, but apart from this and the apparently remote 

 possibility that the "jumps" are really due to the escape of some substance foreign 



