macallum] cellular OSMOSIS AND HEREDITY 149 



*' setting," he poured into one limb a solution of poftassic ferrocyanide 

 and into the other a solution of either the chloride or the nitrate of cop- 

 per. The two solutions diffused into the gelatine and towards each other, 

 farming where they met ;a regular continuous well-supported membrane. 

 From his experiments -vvith this membrane, he concluded that its i>er- 

 meability does not depend on its chemical nature, but on the conditions 

 of its fonnation. Nevertheless, it is not the size of the molecules which 

 determine wliether they shall go through tbe membrane; it is rather mor*^ 

 or less the affinity of the dissolved substances for the substance constitut- 

 ing the membrane that is the deciding factor. 



It musit not be overlooked tliat the term semi-penneable has been 

 applied to other membranes tlian the precipitated ones. An instance is 

 that of Nernst.^ He used pig's bladder to separate a saturated solution 

 of water in ether from a quantity of the same solution, in which, however, 

 benzene was dissolved. The membrane allows the ether water to pass 

 through it, but keeps back the benzene. Another example of a so-called 

 semi-permeable membrane differing from the precipitates is that o^' 

 Eaoult,- who employed a rubber membrane to separate ether and a solu- 

 tion of ether in methyl alcohol. In this experiment the ether, but not 

 "the alcohol, passed through the membrane. A like septum was employed 

 by Flusin, who found that it was " permeable " to carbon disulphide, 

 chlorofonn, toluol, ether and every organic fluid which dissolves 

 caoutchouc. The velocity, however, of the diffusion through the mem- 

 brane varied almost directly according to the solubility of the diffusing 

 compound in the rubber. 



All these results distinctly point to an action on the part of tlie 

 membrane which is not postulated in the semi-permeable membrane of 

 Pfeffer and van't Hoff. The rubber dissolves th^^ " permeating" sub- 

 stance and, in the case of the pig's bladder used by Xerast, the substance 

 of the bladder membrane dissolves the water-ether solution, but does not 

 dissolve the benzene, and hence the latter is held back. 



It is obvious that! if th.is view is correct the currently received ex- 

 planation of osmosis fails. This postulates that the solute is in a staAe 

 of gas, the molecules and the dissociated ions of which bombard the mem- 

 brane and the free surface of the solution, and this bombardment tends 

 to extend the limits of the solution, wliich can only happen if the mem- 

 brane permits the molecules of the solvent in the other compartment to 

 pass through it. TMs theory absolutely ignores any aotivit}^ on the part 



^ Theoretische Chem., 2nd éd., p. 167. 



''Zeit. fur Physik. Chem., Vol. 17, p. 737. 1895; Comptes Rendus. Vol. 121, 

 p. 187, 1896. 



