234 PROCEEDINGS OF THE AMERICAN ACADEMY 



periments were made by Drs. Mitchell and Faust, and others, in which 

 gases passed through a film of india-rubber, entering into a partial 

 combination with the material on one surface, and escaping from it on 

 the other. 



Graham not only considerably extended our knowledge of this class 

 of phenomena, but also gave us a satisfactory explanation of the mode 

 in which these remarkable results are produced. He recognized in 

 these cases the action of a feeble chemical force, insufficient to pro- 

 duce a definite compound, but still capable of determining a more or 

 less perfect union, as in the case of simple solution. He also dis- 

 tinguished the influence of mass in causing the formation or decomposi- 

 tion of such weak chemical compounds. The conditions of the phenom- 

 ena under consideration are simply these : — 



First. A material for the septum capable of forming a feeble chem- 

 ical union with the gas to be transferred. 



Secondly. An excess of the gas on one side of the film and a 

 deficiency on the other. 



Thirdly. Such a temperature that the unstable compound may 

 orm at the surface, where the aeriform constituent is present in large 

 mass, while it decomposes at the opposite surface, where the quantity 

 is less abundant. 



One of the most remai-kable results of Graham's study of this pecu- 

 liar mode of transfer of aeriform matter through the very substance of 

 solid bodies was an ingenious method of separating the oxygen from 

 the atmosphere. The apparatus consisted simply of a bag of india- 

 rubber kept distended by an interior framework, while it was exhausted 

 by a Sprengel pump. Under these circumstances the selective affinity 

 of the caoutchouc determines such a difference in the rate of transfer 

 of the two constituents of the atmosphere that the amount of oxygen 

 in the transpired air rises to forty per cent, and by repeating the 

 process nearly pure oxygen may be obtained. It was at first hoped 

 that this method might find a valuable application in the arts, but in 

 this Graham was disappointed ; for the same result has since been 

 effected by purely chemical methods, which are both cheaper and more 

 rapid. 



These experiments on india-rubber naturally led to the study of 

 similar effects produced with metallic septa, which, although to some 

 extent previously observed in passing gases through heated metallic 

 tubes, had been only imperfectly understood. Thus, when a stream ot 



