CH. XXII.] OSMOSIS 323 



teid, etc.) that they will not pass through such membranes are called colloids. 

 Diffusion of substances in solution which have to deal with an intervening membrane 

 is usually called dialysis. The process of filtration (i.e., the passage of materials 

 through the pores of a membrane under the influence of mechanical pressure) may 

 be excluded in such experiments by placing the membrane (M) vertically as shown 

 in the diagram (fig. 307), and the two fluids A and B on each side of it. Diffusion 

 through a membrane is not limited to the molecules of wate'r, but it may occur also 

 in the molecules of certain substances dissolved in the water. But very few or no 

 membranes are equally permeable to water and to molecules of the substances dis- 

 solved in the water. If in the accompanying diagram the compartment A is filled 

 with pure water, and B with a sodium chloride solution, the liquids in the two com- 

 partments will ultimately be found td be equal in bulk as they were at the start, and 

 each will be a solution of salt of half the original strength of that in the compart- 

 ment B. But at first the volume of the liquid in compartment B increases, because 

 more water molecules pass into it from A than salt molecules pass from B to A. The 

 term osmosis is generally limited to the stream of water molecules passing through a 

 membrane, while the term dialysis is applied to the passage of the molecules in solu- 

 tion in the water. The osmotic stream of water is especially important, and in con- 

 nection with this it is necessary to explain the term osmotic pressure. At first, then, 

 osmosis (the diffusion of water) is more rapid than the dialysis (the diffusion of the 

 salt molecules or ions). The older explanation of this was that salt attracted the 

 water, but we now express the fact differently by saying 

 that the salt in solution exerts a certain osmotic pressure : 

 the result of the osmotic pressure is that more water flows 

 from the water side to the side of the solution than in the 

 contrary direction. The osmotic pressure varies with the 

 amount of substance in solution, and is also altered by 

 variations of temperature occurring more rapidly at high 

 than at low temperatures. 



If we imagine two masses of water separated by a 

 permeable membrane, as many water molecules will pass 

 through from one side as from the other, and so the 

 volumes of the two masses of water will remain un- 

 changed. If now we imagine the membrane M is not per- 

 meable except to water, and the compartment A contains Fio. 307. 

 water, and the compartment B contains a solution of salt 



or sugar ; under these circumstances water will pass through into B, and the 

 volume of B will increase in proportion to the osmotic pressure of the sugar or 

 salt in solution in B, but no molecules of sugar or salt can get through into A from B, 

 so the volume of fluid in A will continue to decrease, until at last a limit is reached. 

 The determination of this limit, as measured by the height of a column of fluid 

 or mercury which it will support, will give us a measurement of the osmotic 

 pressure. 



If a bladder containing strong salt solution is placed in a vessel of distilled 

 water, water passes into the bladder by osmosis, so that the bladder is swollen, 

 and a manometer connected with its interior will show a rise of pressure (osmotic 

 pressure). But the total rise of pressure cannot be measured in this way for two 

 reasons : (1) because the salt diffuses out as the water diffuses in ; and (2) because 

 the membrane of the bladder leaks ; that is, permits of filtration when the pressure 

 within it has attained a certain height. 



It is therefore necessary to use a membrane which will not allow salt to pass 

 out either by dialysis or nitration, though it will let the water pass in. Such 

 membranes are called semipermeable membranes, and one of the best of these is 

 ferrocyanide of copper. This may be made by taking a cell of porous earthenware 

 and washing it out first wibh copper sulphate and then with potassium ferrocyanide. 

 An insoluble precipitate of copper ferrocyanide is thus deposited in the pores of the 

 earthenware. 



If such a cell is arranged as in fig. 308, and filled with a 1 per cent, solution of 

 sodium chloride, water diffuses in, till the pressure registered by the manometer 

 reaches the enormous height of 5000 mm. of mercury. If the pressure in the cell 



