390 Journal of Agricultural Research voi. xn. No. 6 



obtained he constructed graphs and proved that with increasing quan- 

 tities of water more material was dissolved. He therefore divided the 

 total dissolved solids into two fractions : (a) That part soluble even with 

 low moisture content, and (b) the additional material dissolved by the 

 excess of solvent. 



A number of difficulties arise in connection with this method of arriving 

 at the concentration of soil solution. In the first place, it is not safe to 

 assume that a curve based on one range of extractions can accurately 

 be extended to cover another range of extractions. Indeed, the experi- 

 mental data indicate that with the smaller proportions of water the 

 curves may change their direction very appreciably and it is unfor- 

 tunately impracticable to obtain extracts for analysis in those concen- 

 trations which correspond to optimum moisture conditions. Another 

 limiting factor, previously neglected, has been described by Stewart. 

 This concerns the differential effect of the solvent. The actual solvent 

 in any case is not pure water, but pure water plus the solids already 

 dissolved in the soil solution, and these vary enormously with changing 

 conditions, even in the same soil. It is quite obvious that this factor 

 would modify any calculations of the concentration of the soil solution 

 based on water extracts. Moreover, a question may arise with regard 

 to the relation of the soil extract to the residual solution. It is possible 

 that the equilibrium may not be the same for the low and high propor- 

 tions of water, especially in soils of colloidal character. 



Notwithstanding these difficulties met with in attempting to predict 

 accurately the concentration of the soil solution from soil extracts, it is 

 still possible to make certain valuable comparisons of the two methods. 

 If we should contrast the concentration of the soil solution, calculated 

 from the extracts to the total moisture of the soil, with the concentra- 

 tions shown by the freezing-point method, there would be a general 

 similarity of magnitude. Logically, however, a comparison is much 

 more justifiable when the extracts are calculated not to the total soil 

 water, but to the free water, in the sense meant by Bouyoucos. It is 

 then apparent that the concentration of the soil solution calculated by 

 the extraction method is from two to five times that indicated by the 

 depressions of the freezing point (Table VII). In other words, there is 

 a considerably greater quantity of total solids dissolved in the i-to-5 

 extract than is actually present in the soil solution at any given moment. 

 The general order of magnitude of the two quantities is, however, evi- 

 dently not disproportionate, the material extracted from the sandy 

 loams averaging only about twice the quantity actually present in the 

 soil solution, as shown by the freezing-point method. 



The total amounts of dissolved material in 100 gm. of soil have been 

 calculated by multiplying the percentages of free water by the concen- 

 trations in terms of parts per million of total solids. If we use the 



