3 6 INTE OD UCTION 



which are composed of colloids. Crystalloids are generally 

 soluble in colloids and hence pass through such diffusion mem- 

 branes ; colloids dissolve but slightly in colloids, and hence 

 they do not pass through a diffusion membrane readily, and are, 

 therefore, but very slightly diffusible. 



The process of diffusion, if uninterrupted, always continues 

 until the solution is of exactly the same composition through- 

 out. If on one side of the diffusion membrane there is a sub- 

 stance that passes through the membrane rapidly, and on the 

 other a substance that passes through slowly or not at all, there 

 will soon be an unequal condition on the two sides of the 

 membrane, for the diffusible substance would accumulate in 

 equal amounts on each side, while the non-diffusible would 

 remain where it was. On one side there would then be more 

 material exerting osmotic pressure than on the other, and if 

 the membrane were flexible, it would bulge toward the opposite 

 side. The pressure is due to the bombardment of the contain- 

 ing walls by molecules or ions of the substances in solution, 

 and hence the more molecules and ions in a solution, the more 

 pressure. When equal numbers of particles are on each side of 

 the partition, the pressure is equalized. It is quite possible to 

 have membranes permeable to one substance and not to another ; 

 such membranes are called semipermeable. Experimentally 

 they are usually produced as follows : A cup or cylinder of 

 porous clay, such as the cylinder of a Pasteur-Chamberland 

 filter, is filled with a solution of some substance and placed in 

 a solution of another substance, which, by reacting with the 

 first where they meet in the wall of the cylinder, forms the 

 proper sort of a precipitate most frequently copper sulphate 

 and potassium ferrocyanide are used, or gelatin and tannic acid. 

 A thin film or membrane of the precipitate is formed in the 

 wall, which is supported firmly by the clay, so that large pres- 

 sures can be developed without destroying the membrane. If 

 we now fill the cup with a solution of sugar or some other solu- 

 ble crystalloid, its particles will bombard the walls of the cylin- 

 der in vain ; they cannot pass through the semipermeable mem- 

 brane. On the other hand, the water can pass through, and does 

 so in an attempt to equalize the concentration on both sides of 

 the membrane, and hence the volume of fluid in the cylinder 

 increases. "This it will do until the weight of the column of liquid 

 in the cylinder balances the osmotic pressure, and in this way 

 we can measure just how great the pressure is. The amount of 

 pressure exerted by a substance in solution is thus learned to be 

 very great; a 6 per cent, solution of cane-sugar produces a 



