CHEMISTRY OF DIGESTION AND NUTRITION. 251 



other, giving finally a solution of uniform composition. The process of mixing 

 that is, of the passage of the molecules of one liquid into the intermolecular 

 spaces of the other is called " diffusion." Some liquids water and oil, for 

 example will not diffuse with each other, or, as ordinarily stated, they are not 

 miscible. When two miscible liquids are separated by a membrane, diffusion 

 still takes place through the substance of the membrane; the process under 

 these conditions is called "osmosis" or "dialysis," and it occurs independently 

 of any difference of pressure on the two sides. It is well to bear in mind 

 that, in order that osmosis may occur, it is not necessary that there should be 

 actual capillary pores in the membrane. We may suppose such pores to be 

 entirely absent, and yet osmosis be possible, since the liquids in this case, or 

 one of them at least, may be imbibed into the substance of the membrane and 

 thus be brought into contact. Imbibition, or the swelling of a membrane with 

 water, is, in fact, always preliminary to the process of osmosis. When two 

 liquids containing soluble constituents in different proportions are separated by 

 a membrane, the tendency is for osmosis to occur until an equable composition 

 is found on the two sides, diffusion equilibrium being established. This pos- 

 sibility cannot always be fulfilled, for the reason that some soluble substances 

 do not undergo osmosis, or, as we usually say, are not dialyzable. As is well 

 known, Graham separated soluble substances into two great classes the crys- 

 talloids, comprising most of the crystalline bodies, which are dialyzable ; and the 

 colloids, such as gelatin, which are not dialyzable. The rapidity of osmosis of 

 a crystalloid is measured by some form of osmometer. The simplest form con- 

 sists of a glass tube the end of which is closed by a membrane for example, a 

 piece of parchment. If we place a strong solution of sodium chloride in such 

 a tube and then bring the bottom of the membrane into contact with distilled 

 water, diffusion will take place, sodium chloride passing through the parchment 

 into the distilled water outside (exosmosis), and water passing back into the tube 

 (endosmosis). The weight of water which passes into the salt solution is much 

 greater than the weight of salt which passes into the distilled water. If the 

 process is allowed to go on long enough, the proportion of sodium chloride 

 outside and inside will be the same, but the volume of liquid inside the osmom- 

 eter will be increased greatly. In an experiment of this character it is not 

 difficult to determine what weight of water passes one way through the mem- 

 brane for a given unit (1 gram) of the crystalloid passing the other way. On the 

 supposition that this ratio is constant, it was determined for a number of crys- 

 talloids, and represents what is known as the " endosmotic equivalent," ^j^. 

 As a matter of fact, the ratio is not constant : it varies among other things 

 with the strength of solutions used. Still the term is often used ; and it is a 

 convenient one, as it expresses the approximate rate of dialysis of different 

 substances. Colloidal substances, such as albumin solutions, which dialyze 

 very slightly, have been supposed to have a high osmotic equivalent, but so 

 far at least as the proteids are concerned this seems to be an error. Recent 

 work has shown that these bodies exert only a slight attraction for water. 1 

 1 See Heidenhain : Pfliiger's Archiv filr die gesammte Physiologic, 1894, Bd. Ivi. S. 637. 



