-CHEMICAL RESEARCHES ON SEDIMENTS 



grain; on the other hand, they do not greatly impede the migration of 

 ions. Material which is thus and otherwise forced into solution by un- 

 equal stresses can be readily deposited on parts of the grains which are 

 less stressed or in open pores. Cementation, accompanied by diminution 

 of pore space, results, and smaller grains tend to disappear. 



Diffusion of the constituents of one crystal into another crystal which 

 is in intimate contact with it has been proposed from time to time as a 

 factor in petrogenesis, but I have seen no adequate discussion of the 

 mechanism. Diffusion could hardly be expected unless the diffusing ma- 

 terial can have a definite place in the crystal structure into which or 

 through which it diffuses. Thus, dolomite and siderite might inter- 

 diffuse to form mix-crystals of intermediate composition. Diffusion 

 should be slower in hard than in soft minerals and should increase with 

 rising temperature. 



A crystal which is forming or growing in a complex solution concen- 

 trates and holds at its surface certain of the constituents of the solution. 

 The smaller the crystal the greater this tendency. Commonly in cases 

 of very slightly soluble substances the nuclei of incipient crystals become 

 covered to such an extent with this (adsorbed) material that definite 

 crystals do not grow, but the nuclei and adsorbed material become aggre- 

 gated into "amorphous" or fibrous minerals. More or less of the sur- 

 rounding solution is usually enmeshed in the aggregates and is later 

 partly replaced by other matter. The elucidation of the genetic relation- 

 ships of materials of this character, such as chert, bauxite, glauconite, 

 hydrated oxides of iron and manganese, garnierite, phosphate rocks, et- 

 cetera, will require much cooperative study. 



In some cases slightly soluble substances may form nuclei which do not 

 at once become sufficiently aggregated as to be held by such filters as are 

 ordinarily used in water analysis. Such material may be transported to 

 considerable distances and finally be precipitated quite otherwise than 

 if it had been in true solution. Silica especially should be considered 

 from this viewpoint. 



The formation of nuclei does not take place until a certain supersatu- 

 ration, varying greatly under different conditions, has been reached. 

 Such supersaturation may be so great that the solution becomes saturated 

 with respect to a second crystalline substance, and the second substance 

 may appear and continue to crystallize to the exclusion of the first, which 

 is the more stable. The crystal structure which is to result from the 

 growth of the nuclei as compared with the structure of the ions and the 

 complex atomic groups of the solution must help to determine whether 



