1916] Sharxj : Soluble Salts and Soil Colloids 339 



10. The portion of the organic matter of the soil known as 

 humus has little or no connection with the appearance of dif- 

 fusion in salt-treated, water- washed soils. 



11. The diffusion in soils treated as described above seems to 

 be closely associated with the direct addition of sodium to, or 

 with the absorption of sodium by the soil, thereby producing a 

 new silicate complex of a colloidal character in the soil. 



12. This silicate complex is formed simultaneously with the 

 interchange of ions occurring between the salt and the soil. 



13. The washing process serves, in the case of neutral salts, 

 to remove flocculating agents. 



14. The loss of calcium and magnesium from the soil bears 

 little or no relation to the flocculation appearing in salt-treated, 

 water-washed soils except in so far as it may be a measure of the 

 absorbed sodium. 



15. The presence of the OH-ion does not seem to be an essen- 

 tial factor in the diffusion of salt-treated, water-washed soils. 



16. NaoCOg and NaOH produce markedly different effects on 

 suspensions of the Davis soil. 



17. The acid ion of the salt is not an important factor in the 

 defloeculation phenomena following the washing out of salts 

 from soils. 



18. Sodium, potassium, and ammonium seem to produce the 

 colloidal silicate complex when salts of these metals are applied to 

 soils, while calcium does not. 



19. Dilute solutions of acids and salts possess flocculating 

 powers on .suspensions of the Davis soil. 



20. It is not essential in every case to wash all of the salt 

 out in order to bring about diffusion. 



21. The facts presented are discussed in their relation to the 

 reclamation and management of alkali lands, and in their appli- 

 cation to the use of soluble salts as fertilizers. 



The writer desires to express his thanks to Professor Charles 

 B. Lipman for many valuable suggestions and for his critical 

 reading of the manuscript, and to L. E. Bailey for his timely 

 assistance in the analvtical work. 



Transmitted November 16, 1915. 



