Deposits in Metas table Inorganic Hydrosols. 



29 



The difference is similar to the well-known difference in crystallography 

 of the slowly crystallising and the rapidly crystallising solution. 



It was pointed out by Graham, as quoted in the previous paper, that time 

 is an essential factor in all colloidal solutions, and that in this respect the 

 colloids approach living organisms. 



The first method attempted was that of using stronger solutions of the two 

 reagents of the previous paper, and bringing these together in proportions 

 as close to the point of actual precipitation as possible, so as to hasten 

 operations of growth in order that contamination by organisms might be 

 safely excluded. 



For this purpose a 1-per-cent. solution of ferric nitrate and a 1-per-cent. 

 solution of sodium silicate, each in distilled water, were prepared, and from 

 the two a metastable colloidal system was constructed. 



In mixing the two solutions the ferric nitrate solution must be taken as 

 the basis and the silicate solution be slowly added to this, for if the 

 reverse procedure be adopted the silicic acid is at once thrown completely out 

 of solution. 



If, however, a volume of the ferric nitrate solution be taken, and the sodium 

 silicate solution be added to it drop, by drop with constant shaking, it is found 

 that, although every drop produces a precipitate, this, on shaking, redissolves, 

 and no permanent precipitate occurs up to a certain point. Instead, a common 

 solution is formed of colloidal ferric hydrate and colloidal silicic acid. 



This point is reached when about 4 c.c. of the silicate solution has been 

 added to 10 c.c. of the ferric nitrate solution. If this point be somewhat 

 exceeded, a brownish-coloured precipitate appears permanently. On examina- 

 tion under the microscope this precipitate is seen to consist of fragments of 

 membrane shrivelled and corrugated and showing thickened tortuous lines 

 like embedded fibres in the substance of the membrane. Short pieces of 

 fibrils project at places, at the borders of the shreds of membrane. The 

 appearance suggests that fibrils are first thrown out of solution, and that 

 subsequently in the meshes of the fibrils more connecting colloid in the 

 gel form is deposited in a thin layer, so forming the membranous shreds of 

 the precipitate. 



This view is supported by the appearances shown when the metastable 

 point is not passed. If 4 c.c. of the silicate solution be slowly added with 

 continued agitation to 10 c.c. of the ferric nitrate solution, a slightly 

 opalescent solution is obtained. This shows, on examination microscopically, 

 very few membranes, but a large number of rather coarse fibres, branching 

 and tortuous. 



These are not so well developed as those more slowly grown in the 



