68 SCIENCE PROGRESS 



Osmotic pressure varies in the opposite sense. They are 

 generally not very stable bodies, but the behaviour of gelatin 

 is particularly remarkable. By continued heating at ioo° 

 the submicrons grow smaller and fewer, until, after thirty-six 

 hours [Levites, 1908] only amicrons are present and the 

 viscosity has changed from 2*29 to 1*39 and become constant. 

 The sol now refuses to gelate and is known as /3-gelatin. This 

 is very soluble but, on evaporating nearly to dryness and 

 cooling, it passes through a gel stage which by drying forms 

 a mass resembling sheet gelatin. 



The heat coagulation of albumin must be supposed to be due 

 to the chemical production of an insoluble compound. In this 

 case gelation is irreversible. 



If it be conceded that the solubility of silicic acid depends 

 on the hydration and aggregation of the molecule its gelation 

 follows at once from v. Weimarn's theory. The gel is formed 

 immediately (in the same way as that of aluminium hydroxide, 

 for example), when a moderately concentrated solution of 

 sodium silicate is neutralised with mineral acid. This is due 

 to the sudden production of a highly aggregated insoluble form 

 of silicic acid, owing to the concentration, viscosity and aggrega- 

 tion of the reaction medium. On the contrary, when the 

 silicate is slowly stirred into the acid medium, which has the 

 ordinary properties of a liquid, probably a true solution of 

 silicic acid is formed at first. This is indicated by its high 

 diffusivity. Changes then take place, possibly in the hydration 

 of the molecule, which reduce the solubility and cause gradual 

 aggregation with increase of viscosity [Garrett, 1903] and sol 

 formation. 



If it is preferred to regard the immediate precipitation of 

 silicic acid as due to electrolytes, its gelation is no less in con- 

 formity with v. Weimarn's theory. 



Returning to the reversible sol-gel transformation, it has 

 been seen that the low diffusion constant of the natural emul- 

 soids, by retarding crystallisation, allows the accumulation of 

 a sufficiently large excess concentration which, together with 

 their large value of K, will cause gelation on cooling. This 

 low diffusion is also responsible for the permanency of the gels 

 and hysteris of sols, since it prevents the subsequent develop- 

 ment of larger crystals in gels and retards changes in the 

 aggregation of sol particles. 



