698 Prof. 0. W. Richardson on th 



The effects which seem to depend rather on the mode of 

 ijj testing than the nature of the material employed are : — 



1. By the strip method a very small emission occurs at 

 very low pressures which increases rapidly with increasing 

 pressure. The tube method makes the emission much less 

 sensitive to variations in the pressure at low pressures, the 

 changes being sometimes in one direction and sometimes in 

 another. 



2. In the tube experiments the immediate response to 

 increasing the pressure is a diminution in the emission, 

 followed later by the change which is characteristic of! the 

 steady end condition. The converse effect occurs on 

 reducing the pressure. 



3. The emission per unit area of the hot platinum seems 

 on the average to be greater with the tube method than with 

 the strip method of testing. 



The chemical effects are evidently very involved. On the 

 whole there is a greater tendency for the emission to be 

 diminished with increasing pressure in the case of the phos- 

 phates than in the case of the sulphates. In the case of at 

 i| least one of these substances, however, we shall see that the 



emission is not due to the substance itself but to something- 

 mixed with it. The emission from the reduction products of 

 barium sulphate is greater than that from the sulphate itself. 

 The emission from the latter probably varies very little with 

 the pressure of the surrounding air. 



Observations on Aluminium Pliosphate. 



It will be remembered that Sir J. J. Thomson found that 

 this substance gave rise to a much greater emission of positive 

 ions than a great number of representative salts which he 

 tested under similar conditions. My experiments, however, 

 do not indicate that the specimen of aluminium phosphate 

 which I used was remarkable in this respect. This difference 

 coupled with the fact, established by the experiments of 

 Davisson mentioned above, that the carriers of the ionization 

 from our specimen of aluminium phosphate had a value of 

 m/H equal to 22, led to the view that the emission of positive 

 ions ordinarily observed with aluminium phosphate is due, 

 not to the salt itself, but to something mixed with it. 



To test this view it seemed desirable to prepare aluminium 

 phosphate, if possible, in a way which would ensure freedom 

 from contamination with other inorganic impurities and 

 especially with the salts of the alkali metals. There seemed 

 a reasonable chance of this if the salt were prepared from 



