434 Dr. F. Gr. Donnan on the Relative Bates of 



on the following day showed that some disturbance had 

 occurred : — m. s. 



2 50-8 



2 51*5 



2 52-7 



It appears very probable that the platinum had occluded some 

 hydrogen, and that the resulting volume-changes were the 

 source of the variations. The apparatus was now allowed to 

 remain exhausted overnight. Experiments with oxygen next 

 morning gave the following results : — 



m. s. 

 Half-an-hour after introduction of gas. . 10 57*8 

 1J hours „ „ „ . . 10 57-3 



Apparatus left exhausted for an hour and 



fresh oxygen introduced 10 56*5 



The apparatus was left filled with oxygen overnight, pumped 

 out the next morning, allowed to remain in connexion with 

 the pump for an hour or so, and then refilled with oxygen : — 



Oxygen . . . 10 m 56 s -2, 10 m 56 s «6. 



It was accordingly concluded that the aperture had been 

 again brought to a constant state and any occluded hydrogen 

 removed. Argon was now measured : — 



Oxygen . . . 10 m 56 s '4. 



Argon . . . . ll m 47 s , 11™ 47 s . 



According to the densities alone the time for argon should be 



656 s -4x A / 19 ' 957 = 12 m 13M. 

 V 16-0 



That is to say, argon effuses too rapidly by about 3*5 per cent, 

 when compared with oxygen by the inverse square-root law. 



In order to confirm this result a second determination was 

 made. During the interval (week-end) the apparatus had 

 remained exhausted and, as will be seen, the time of effusion 

 for oxygen had considerably increased : 



El. S. 



Oxygen ... 11 13'3 



Argon .... 12 3-9 



Oxygen ... 11 12 



Mean oxygen . . 11 12*6 



The time for argon, calculated from that for oxygen by 

 means of the densities, is 12 m 31 s *2, so that the observed 

 time is 3*7 per cent, less, which agrees fairly well with the 

 previous result. 



