842 Journal of Agricultural Research vol. vi. no. 22 



Excepting the two sands, 1 and 3, the average ratio of moisture 

 equivalent to hygroscopic coefficient is 2.33, with a maximum of 2.73 

 and a minimum of 1.92. The lowest ratios are shown by the arid or 

 semiarid soils, 1, 3, 11, and 15. The exceptional behavior of 11 and 15 

 is not to be attributed to error of determination, as, after finding these 

 exceptional ratios, we made repeated determinations of both values. 

 The ratios found for both subsoils and surface soils from Nebraska are 

 quite similar to those reported in Table III, the average ratio, 2.38, 

 being identical with that obtained for the 36 loess samples. 



If the two sands, 1 and 3, in Table VI, be omitted, the variation of 

 our ratios in Tables III, V, and VI are of much the same order as those 

 of Briggs and Shantz, shown in Table I. Thus, the divergence in our 

 conclusions as to the availability to plants of the portion of the soil 

 moisture lying between the hygroscopic coefficient and the wilting co- 

 efficient is not to be explained by any differences in our respective 

 methods of arriving at the value of the hygroscopic coefficient. Neither 

 are th£re sufficient reasons to attribute it to the particular range of 

 soils with which we have worked, for the data above show that our soils 

 range a*s widely as those which they have employed. 



Their data, as well as our own work, make it evident that in any accu- 

 rate experiments to determine the relation of the nonavailable water of 

 the soil to the hygroscopic coefficient it is not permissible to calculate 

 the value of the latter from the moisture equivalent, unless a previous 

 thorough investigation has been made to determine just what formula 

 is applicable to the soil type in question. From the data of Lipman 

 and Waynick it would appear that in the case of certain soils this indi- 

 rect method would be scarcely allowable for even the crudest studies on 

 soil moisture. However, in the case of any extensive study, involving 

 many soil types, the same general conclusions as to the relation of the 

 nonavailable moisture to the hygroscopic coefficient are to be expected, 

 no matter whether the latter value be directly determined or be calcu- 

 lated from the moisture equivalent by the Briggs-Shantz or by some 

 more satisfactory formula. 



COMPUTATION OF THE MOISTURE EQUIVALENT FROM THE MECHANI- 

 CAL ANALYSIS 



Table VII shows the concordance of the moisture equivalents directly 

 determined with the values computed from the mechanical analyses in 

 the cases of the loess samples reported in Table III, using the formula 

 proposed by Briggs and Shantz: 



Moisture equivalent = 0.02 sands + 0.2 2 silt +1.05 clay, and also a 

 modified form of this formula : 



Moisture equivalent = 0.1 4 sands+0.27 silt + 0.53 clay. In these 

 formulas "sands" include particles ranging from 2 to 0.05 mm. in 

 diameter. The separates referred to in the table as "coarser fractions" 

 include the particles ranging from 2 to o. 10 mm. It will be seen that the 



