HY DROP HI Lie SOLS AND GELS 149 



stone or the clay of a flower pot but with pores infinitely finer. 

 The silica gel is a mass of sand shot through with myriads of 

 ultramicroscopic capillaries (Fig. 90). The tremendous amount 

 of surface that these pores display is the physical basis of the 

 excellent adsorbing qualities of silica gel. This gel is an ideal 

 example of matter in the colloidal state. 



The structure of elastic gels or jellies presents a much more 

 difficult problem. Exceedingly little is known about their 

 internal configuration, with the possible exception of cellulose. 



It does not seem likely that the structure of jellies is a porous 

 one in the same sense as is that of coagula. Hatschek states 

 that these two kinds of gels are to be distinguished not primarily 

 by such qualities as elasticity but by the fact that coagula, of the 

 type of the silica gel, are porous, while the jellies (gelatin) are not. 

 Some conception of the conditions that any theory on gel 

 structure must satisfy is to be had from the remarkable fact that 

 certain solutions set to a rigid gel even though they possess but 

 0.2 per cent of solid matter. The most extreme case of this is 

 the germanate gel prepared by J. H. Mtiller. Germanic acid 

 and calcium hydroxide are mixed, and a firm gel results. The 

 gel appears dry, and the beaker containing it may be inverted 

 without loss, yet the gel contains but 0.1 per cent of solid matter 

 and 99.9 per cent of water. That a solution containing but 0.1 

 per cent of solid matter will form and maintain a rigid gel is 

 intelligible on the assumption that the structural units are long 

 fibrous ones, linear, crystalline rods of molecular or colloidal 

 (micellar) size. The classical micellar theory of gel structure, pos- 

 tulated by Nageli, is discussed elsewhere (see pages 118 and 252). 

 Ultrafiltration. — A turbid colloidal solution will pass through 

 filter paper. In order to filter out colloidal particles, it is neces- 

 sary to have filters with much finer pores than those of filter 

 paper, the best of which have pores of 1 /j., with 2 to 5 ^ for the 

 average. Even porcelain Berkefeld and Chamberlain filters, 

 used to rid drinking water of bacteria, have pores of 0.2 to 0.6 fj, 

 and therefore above colloidal dimensions. Special colloidal 

 filters, generally known as uUrafilters, have consequently been 

 devised. Those made by Bechhold are of filter paper soaked in 

 gelatin hardened with formaldehyde. Collodion sacs, the per- 

 meability of which can be varied over a considerable range by 

 the addition of oil, gelatin, and albumin, make excellent ultra- 



