284 THE COLLOIDAL STATE 



The former class, including substances such as salts, acids, 

 bases, cane sugar, urea, etc., which, for the most part crystal- 

 lized readily, he called "crystalloids," while for the latter class, 

 which comprise such substances as starch, albumen, and gum, 

 he devised the term " colloid ". Although there was this marked 

 difference between these two classes of substance in the rates 

 of free diffusion into pure water, it was found that the presence 

 of a colloid, in relatively low concentration, had but little 

 effect in retarding the rate of diffusion of a crystalloid, which 

 accounts for the fact that diffusion experiments can be carried 

 out in gelatine solutions (see page 300), and also that crystal- 

 loids will diffuse quite readily through colloidal membranes, 

 such as parchment, etc. 



On the other hand, it was found that such membranes 

 offered a very strong opposition to the passage of other col- 

 loids ; this observation was turned to account in the dialyser, 

 by means of which apparatus it was found possible to separate 

 crystalloids from colloids contained in the same solution. 

 Numerous modifications of Graham's original apparatus have 

 been devised, but they are all ultimately based on the same 

 principle that if a mixed solution of a colloid and a crystalloid 

 are separated from pure distilled water by a colloidal parch- 

 ment or other membrane, the crystalloid alone will diffuse out 

 at a measurable rate, whilst the colloid will remain behind. 

 The method is, indeed, to this day the only one known for 

 purifying a colloid from a crystalloid since the ordinary 

 methods applicable for the purification of crystalloids do not 

 hold for colloids. 



The origin of the terms crystalloid and colloid was, how- 

 ever, based on a misconception. The rate of diffusion of any 

 substance is in no way connected with its ability to crystallize, 

 or the reverse since, as was subsequently shown, almost all 

 crystalloids can be made under suitable conditions to give 

 solutions in which they have lost their ability for rapid diffusion, 

 and have acquired many of the characteristics of the .class of 

 substance known to Graham as colloids ; similarly, many of 

 Graham's colloids, such as egg albumen and haemoglobin, 

 have been obtained in crystalline form. The properties of the 

 colloidal solutions are, therefore, no longer regarded as being 

 due to the intrinsic properties of the substances dissolved, but 



