THE CONSTITUTION OF THE SOIL 



57 



those shown by colloids ; the only hypothesis capable of explaining 

 the observed facts is that some of the constituents of the soil, and espe- 

 cially of the clay, are colloidal. 



This view is fully borne out by van Bemmelen's fractionations of 

 soil and of clay. By successive extraction with acids of increasing 

 concentration van Bemmelen found (22) two distinct groups of silicates 

 in the Dutch alluvial soils, one soluble in dilute hydrochloric acid in 



molecules of SiO 2 

 which the ratio mo ] ecu | es o f ^1 Q := 3 to 5> t otner soluble only in 



hot, strong sulphuric acid in which the ratio is approximately equal 

 to 2. Other soils of volcanic origin from Java gave up larger amounts 

 of base relative to the silica, but in no case were the ratios constant 

 whole numbers ; the alkaline bases showed the same lack of constancy 

 in their ratios to A1 9 O Q . 



TABLE XXVII. RATIO MOLECULES OF 



MOLECULES OF A1 2 O 



VAN BEMMELBN (22). 



ExTRACTED FROM VARIOUS SOILS. 



Alkaline bases extracted from a heavy clay, Surinam. 



Different soils gave up different proportions of alkaline bases, but 

 again without showing definite simple ratios one to another. Detailed 

 studies of clay revealed the presence of chemically unchanged crystals 



1 The higher numbers were obtained from sandy clays and the lower from heavy clays. 

 As the silica was insoluble in the acid it was extracted by digesting the residue for a few 

 minutes at 55 with dilute alkali of sp. gr. 1-04. 



5 



