192 PROFESSOR THOMAS GRAHAM's SCIENTIFIC WORK. 



snlpliate of potassium being 3, 0, and 12, the densities of tlieir diffusion 

 molecules would be as the reciprocals of these numbers, or as 4, 2, and 1. 

 Lastly, in comparing highly diffusive substances on the one hand, 

 with feebly diffusive substances on the other, one broad dissimilarity 

 became apparent, namely, that highly diffusible substances affected the 

 crystalline state, while feebly diffusive substances were amorphous, and 

 characterized, in particular, by a capability of forming gelatinous 

 hydrates. Hence the distinction established by ^Ir. Graham between 

 highly diffusive bodies, or crystalloids, and feebly' diffusive bodies, or 

 colloids. Compounds capable of existing both in the crystalline and 

 gelatinous states he found to be possessed of two distinct diffusive rates 

 corresponding respectively each to each. 



V. 



Dialysis and osmose. — Tlie subject of dialysis was included in the paper 

 "On liipiid diffusion applied to analysis," referred to in the preceding 

 section; and some further results were communicated in ISGi to the 

 chemical society, in a paper "On the properties of silicic acid and other 

 analogous colloidal substances." * 



In the course of his experiments on diffusion, Mr. Graham made the 

 curious discovery that highly diffusible crystalloid bodies were able to 

 diffuse readily, not only into free water, but also into water that was 

 already in a low form of combination, as in the substance of a soft solid, 

 such as jelly or membrane. Common salt, for instance, was found to 

 diffuse into a semi-solid mass of jelly almost as easily and as extensively 

 as into a similar bulk of free water ; but the introduction of a gelatinous 

 substance, though not interfering in any appreciable degree with the 

 diffusion of a crystalloid, was found to arrest almost entirely the diffu- 

 sion of a colloid. The colloid, of but little tendency to diffuse into free 

 water, proved <piite in(;apable of diifusing into water that was already 

 in a state of combination, however tWh\c. Hence, although the partial 

 separation- of a highly diffusible from a feebly diffusible substance might 

 be elfected by the process of free diffusion into water, a much better 

 result was obtained by allowing the diffusion to take place into, or 

 through, the combined water of a soft solid such as a piece of membrane 

 or i)archment-paper. In the process of dialysis, then, crystalloid and 

 colloid bodies, existing in solution together, are separated from one 

 another by pouring the mixed solution into a shallow tray of membrane 

 or parchment-paper, and letting the tray rest on the surface of a con- 

 siderable excess of water, once or twice renewed. By this means the 

 crystalloid, in process of time, diffuses completely away through the 

 membranous septum into the free water; but the colloid, being quite 

 in<;apable of permeating the membrane, however thin, is retained com- 

 pletely on the tray, unable to reach the free water on the other side. 



* Chemical Society Joiunal, xvii, p. 316. 



