442 Dr. F. Gr. Donnan on the Relative Rates of 



Carbon monoxide. 



Oxygen . 

 CO . . 

 Oxygen . 



8 23-6 



7 49-8, 



8 22-7 



7 49-2 



Mean time for CO - . 7 49"5 



. 7 50-7 

 , 7 49-9 



Time calculated from densities . 

 Time calculated with viscosity . 



The results obtained for argon and carbon monoxide are 

 in agreement with former ones. If we apply no viscosity- 

 correction, then the helium effuses slower than as calculated. 

 The want of agreement between this and former results may 

 be partly due to variable admixture with argon. 



The fact remains that all the results obtained for helium 

 are affected by a more or less uncertain correction for 

 viscosity, and do not agree among themselves. If any one 

 were to be selected as better than the rest, it would be the 

 last one, as the hole appeared to be practically free from 

 viscosity-effects in this case, and the specimen of helium was 

 pretty pure. 



Saint- Venant and WantzeVs Phenomenon. 

 It was stated at the beginning that the dimensions of the 

 apparatus were chosen so that the measurements lay in the 

 region where the rate of effusion was independent of the 

 back-pressure. This was frequently confirmed by experi- 

 ment. The following results, obtained in the last series 

 (where the back-pressure rose from to 47*5 millim. during 

 the course of a measurement), will serve to confirm the state- 

 ment made above. 



Oxygen. 



Time of Effusion. 



Back-pressure. 



m. 3. 



8 23-3 



-47-5 



8 23-5 



47-5-93-9 



8 25-3 



93-9 



In this series of experiments the final back-pressure was 

 47*5 millim. as already stated, and the initial and final pres- 

 sures in the gas-reservoir 525 millim. and 322 millim. re- 

 spectively. In the other series the final back- pressure was 



