1 48 PHYSIOLOGY 



to A and from A to B. The result of this diffusion must be that there 

 is no longer a sudden drop of osmotic pressure from B to A, and the 

 result of the primary osmotic difference on the movement of water 

 will be minimised in proportion to the freedom of diffusion which 

 takes place through, the membrane. Now let us take a case in which 

 A and B represent equimolecular and isotonic solutions of a and /3. 

 It is evident that the movement of water into A will vary as A.p Bp* 

 = 0. But diffusion also occurs of a into B and of ft into A. Now the 

 amount of substance diffusing from a solution is proportional to the 

 concentration, and therefore to its osmotic pressure, as well as to its 

 diffusion coefficient. 



Hence the amount of a diffusing into B will vary as A p.ak 

 (when k is the diffusion coefficient). 



In the same way the amount of /3 diffusing into A will vary as 

 By. /3k'. 



Hence, if ok is greater than /3k', i.e. if a is more diffusible than ft, 

 the initial result must be that a greater number of molecules of a 

 will pass into B than of ft into A. The solutions on the two sides of 

 the membrane will thus be no longer equimolecular, but the total 

 -number of molecules of a -f- ft in B will be greater than the number 

 of molecules of a + ft in A, and this difference will be most marked 

 in the layers of fluid nearest the membrane. The result, therefore, 

 of the unequal diffusion of the two substances is to upset the previous 

 equality of osmotic pressures. The layer of fluid on the B side of the 

 membrane will have an osmotic pressure greater than the layer of 

 fluid in immediate contact with the A side of the membrane, and there 

 will thus be a movement of water from A to B. Hence if we have 

 two equimolecular and isotonic solutions of different substances 

 separated by a membrane permeable to the solutes, there will be 

 an initial movement of fluid towards the side of the less diffusible 

 substance. 



We have an exact parallel to this in Graham's familiar experiment, 

 in which a porous pot filled with hydrogen is connected by a vertical 

 tube with a vessel of mercury. In consequence of the more rapid 

 diffusion outwards of the hydrogen than of atmospheric air inwards, 

 the pressure within the pot sinks below that of the surrounding atmo- 

 sphere and the mercury rises several inches in the tube. 



We must therefore conclude that, even when the two solutions on 

 either side of the membrane are isotonic, there may be a movement of 

 fluid from one side to the other with a performance of work in the 

 process. In fact, osmosis may occur from a fluid having a higher 

 towards a fluid having a lower osmotic pressure. If, for example, 

 equimolecular solutions of sodium chloride and glucose be separated 

 * Ap = osmotic pressure of A, &c. 



