1862.] 289 



Amongst other important points determined, it was found that if 

 two substances which do net combine chemically and which possess 

 different rates of diffusibility, be mixed, and be then placed in a 

 diffusion cell, they may be partially separated by the process of diffu- 

 sion; andinsome cases even chemical decomposition maybe effected by 

 this means. Thus if ordinary potash alum be subjected to diffusion, 

 the sulphate of potash will diffuse out rapidly, and leave the sulphate 

 of alumina in large proportion. 



In 1854 Mr. Graham communicated to the Royal Society a paper 

 on Osmotic Force, in which he examined the influence of a septum 

 upon the process of diffusion ; investigating particularly the class of 

 phenomena specially studied by Dutrochet. By a simple modifica- 

 tion of the apparatus employed, he was enabled to make his experi- 

 ments quantitative, and thus to give them numerical precision. He 

 found that the nature of the solution and its chemical operation on 

 the material of the septum were very important. When animal 

 membranes were employed, dilute alkaline solutions absorbed water 

 rapidly, while weak acid solutions gave out water instead of absorbing 

 it ; neutral salts having little effect in promoting these osmotic 

 movements. 



But these investigations, important as they are, constitute but the 

 introduction to the inquiry published in 1861, upon the application 

 of Liquid Diffusion to Analysis. In this remarkable memoir Mr. 

 Graham has shown that chemical compounds generally may be sub- 

 divided into two great classes, which are characterized by their rela- 

 tion to the process of liquid diffusion : one of these classes he terms 

 Crystalloids, and the other Colloids as being typified by animal 

 gelatine. The crystalloids form a solution generally free from vis- 

 cosity, and always sapid. They are especially endowed with a ten- 

 dency to diffusion through a membranous septum. The colloids, on 

 the contrary, such as gum, starch, dextrine, caramel, tannin, gelatine, 

 and albumen, are characterized by a remarkable sluggishness, and 

 indisposition to diffusion or crystallization. When pure, they are 

 nearly tasteless. Colloid bodies do not necessarily belong to the 

 organic kingdom, though they are frequently met with among its con- 

 stituents ; and owing to their tendency to undergo slow but perpetual 

 molecular change, together with their peculiar relations to water, 

 they seem to be especially suited to form the plastic materials re- 



Y 2 



