702 , Prof. 0. W. Richardson on the 



present in aluminium phosphate, and at the same time behave 

 quite differently for all salts and gases in the tube experiment 

 from what they do in the strip experiment. 



This view would also account for the sudden apparent drop 

 in the emission when the pressure was increased in the tube 

 experiment. The vapour in the tube would be compressed 

 into the bottom and so would have less platinum surface to 

 act on. The converse phenomenon on lowering the pressure 

 is also readily explained in this way. The instantaneous 

 effect on lowering the pressure is smaller than the opposite 

 effect when the pressure is raised, as it should be. 



The greater effect, even when reckoned per unit area of 

 platinum, as obtained by the tube method is readily explained 

 in this way, and is quite incomprehensible on the opposite 

 view The smallness of the electrode compared with its 

 surroundings enables the vapour to diffuse away very readily 

 in the strip experiment, whereas convection currents carry 

 it up against the hot platinum walls in the tube experiment. 



The very varied curves obtained by changing the pressure 

 at constant temperature with different salts and with the 

 same salt after it has been treated in different ways seem to 

 be symptomatic of chemical changes. By that, I mean that 

 they are probably accompanied by changes in the composition 

 of the material which correspond to the curves in some way. 

 This gives a very natural explanation of the variability and 

 complexity of the curves, as the chemical actions are probably 

 quite different in the different cases and extremely complex 

 in every case. One of the difficulties of the subject is that 

 practically nothing is known about the mechanism of chemical 

 reactions at high temperatures. All the information available 

 relates to those of the end products which happen to be 

 formed in comparatively large quantity. 



The statement that the pressure changes are in all prob- 

 ability symptomatic of chemical action is quite apart from 

 the important question as to whether chemical actions of the 

 ordinarily recognized types at high temperatures are accom- 

 panied by the emission of positive ions. The very vigorous 

 emission observed when barite was being attacked by 

 hydrogen tends to support this view, but it may have been 

 that the hydrogen was facilitating the formation of some 

 relatively volatile compound of barium and thus enabling it 

 to reach the platinum surface. The action of water on 

 sodium sulphate is another case of the same kind. There is, 

 of course, no doubt that ionization is an important feature of 

 many chemical actions at high temperatures, the cases of 

 flames, the oxidization of phosphorus, and the action of 

 phosphorus on hot platinum furnishing well known examples. 



