METHODS OF COLLOID RESEARCH 



103 



Diffusion. 



Coefficients of diffusion give information concerning the molecular 

 weight and also the size of the particles of a substance in solution. 

 Diffusion in aqueous solution is the simplest method for such investi- 

 gations. The length of time necessary for such experiments intro- 

 duces so many disturbing factors that, where possible, diffusion in 

 a jelly is to be preferred. A jelly offers a means of separating sub- 

 stances having different rates of diffusion. If a mixture of two sub- 

 stances remain for a time in a tube partly filled with a jelly, the more 

 difficultly diffusible substance will, for the most 

 part, remain in solution and can be poured off, 

 whereas the substance easily diffusible will to a 

 greater extent enter the deeper layers of the jelly. 

 Diffusion experiments in jellies teach us the prop- 

 erties of jellies swollen to various degrees, both in 

 the presence of crystalloids and in their absence. 



Diffusion in Aqueous Solution. The greatest 

 difficulty lies in avoiding agitation not only when 

 samples are being taken but also during the course 

 of the experiments. The most suitable apparatus 

 is that of L. W. OHOLM* (see Fig. 20), with which 

 HERZOG made his experiments, and that of 

 DABROWSKI. The latter (see Fig. 2 la) consists of 

 two glass vessels A and B (a siphon bottle which 

 has been divided in the middle) which are separated 

 by a diaphragm C. This diaphragm is a glass ring 

 filled with glass capillary tubes of 1 mm. bore. 

 The interspaces are filled with celluloid. 



By this arrangement currents are avoided and 

 a very considerable diffusion surface is obtained. 

 The solution is placed in A and diffuses through C and reaches 

 B from which samples for analysis are taken from time to time 

 through the tube F. The fluid in A as well as in B is slowly stirred 

 by the stirrer abd. We shall return to the consideration of 

 DABROWSKI'S experiments on the diffusion of albumin with this 

 apparatus on p. 72. 



In the extremely slow diffusion of colloids, which in the case of 

 the experiments of R. 0. HERZOG extended over more than two 

 months, absolute sterility is essential. Besides having sterile vessels, 

 the solutions are also sterilized by saturating them with toluol and 

 layering it over them. The addition of 1/2 per cent sodium fluorid 

 solution is useful also. As previously mentioned, the vessels must 



FIG. 20. Diffu- 

 sion apparatus. 

 (L. W. Oholm.) 



