424 Mr. W, Odling [Feb. 14, 



and aluminium, is allowed to diffuse from its aqueous solution, the 

 diffusive tendency of potassium compounds being: much greater than 

 that of aluminium compounds, a portion of sulphate of potassium 

 actually breaks away from the sulphate of aluminium with which it was 

 in combination, in order to diffuse into the external or superincumbent 

 water more freely than the sulphate of aluminium can diffuse. 



Crystalloids and Colloids. — Diffusibility does not seem to be asso- 

 ciated in any definite way with chemical composition. Thus, complex 

 organic bodies, like picric acid and sugar, have much the same diffusion7 

 rates as common salt and Epsom salt respectively. Isomorphous com- 

 pounds, however, are for the most part equi-diffusive, but the groups of 

 equi-diffusive are larger than those of isomorphous bodies. The com- 

 mon salt group, for instance, includes not only chloride, bromide, and 

 iodide of sodium, which are similar in composition and isomorphous in 

 form, but also nitrate of sodium, which is dissimilar in composition and 

 heteromorphous in form. But in comparing highly diffusive sub- 

 stances on the one hand, with feebly diffusive substances on the other, 

 one broad dissimilarity becomes apparent, namely, that highly diffu- 

 sive substances affect the crystalline state, while feebly diffusive sub- 

 stances are amorphous, and, in particular, are characterized by a capa- 

 bility of forming gelatinous hydrates. Hence the distinction established 

 by Mr. Graham between highly diffusive bodies, or crystalloids, and 

 feebly diffusive bodies, or colloids. There are very many compounds 

 which can exist, both in the crystalline and gelatinous states, and 

 which present two distinct diffusion-rates corresponding respectively 

 thereto. 



Nature of Diffusion. — Liquid diffusion may be attributed to a 

 self-repulsion of the particles of the salt or other body which diffuses 

 on the one hand, and of the particles of water or other liquid into 

 which it diffuses on the other ; or it may be attributed to a particular 

 kind of mutual attraction existing between the particles of the salt and 

 of the water. Assuming the truth of this latter hypothesis, it is clear 

 that the diffusive attraction of crystalloid particles for water is greater 

 than that of colloid particles ; and, recognizing this superior diffusive 

 attraction of saline particles, it is conceivable that a salt should be 

 able to unite diffusively, not only with free water, but also with water 

 that is already in a low form of combination, as it exists in a soft solid, 

 for instance, such as jelly : and this is found to be the case. Common 

 salt, for example, diffuses into a mass of solid jelly almost as easily and 

 extensively as into a similar bulk of free water. But although the intro- 

 duction of a gelatinous substance does not interfere in any appreciable 

 way with the diffusion of a crystalloid, it arrests almost completely the 

 diffusion of a colloid body. The colloid has but very little tendency 

 to unite diffusively, even with free water, and is quite incapable of 

 abstracting and uniting diffusively with water that is in any state of 

 combination, however feeble. 



Dialysis.— AXihoxx^h. a simple diffusion into water can effect the 

 partial separation of a highly diffusible from a feebly diffusible 



