834 Journal of Agricultural Research vol. vi, no. 22 



defined by Hilgard (10, p. 256) , and from the mechanical analysis. Subsid- 

 iary formulas for the indirect determination of the hygroscopic coefficient, 

 following as a result of the interrelationships thus established, they 

 report (7, p. 73) as follows: 



Hygroscopic coefficient = -< 



Wilting coefficient X 0.68. 

 Moisture equivalent X 0.37. 

 (Maximum water capacity— 2 1) X 0.234. 

 (0.007 sand + 0.082 silt + 0.39 clay). 



As the mechanical analysis of a soil is a far more difficult and time- 

 consuming operation than the determination of the hygroscopic coeffi- 

 cient, the latter could advantageously be calculated from the former only 

 where this is already available, as, for example, in the reports of soil 

 surveys. Even then there is a probability of introducting serious errors. 

 ThUs, with a series of loess soils (3, p. 411) it has recently been shown 

 that the values for the hygroscopic coefficient calculated by the Briggs- 

 Shantz formula agree satisfactorily with those obtained by direct deter- 

 mination only in the case of those samples which carry the smallest pro- 

 portion of very fine sand. However, by altering the values assigned the 

 sands there was obtained the following modified formula, which was found 

 applicable to all the loess soils investigated. 

 Hygroscopic coefficient = 0.005 coarser fractions + 0.07 very fine sand + 



0.82 silt + 0.39 clay. 

 "Coarser fractions" is here used to designate all soil particles having a 

 diameter greater than 0.10 mm. 



The wilting coefficient also is so inconvenient of determination that, 

 unless it has to be determined for some other purpose, it will not be used 

 to calculate the hygroscopic coefficient. 



In connection with field studies of available soil moisture on the 

 Nebraska loess, of which only a few data (1, p. 118) have as yet been pub- 

 lished, one of us had arrived at conclusions so widely at variance with 

 those of Briggs and Shantz, who, in somewhat similar studies, had em- 

 ployed the wilting coefficient, either determined directly or calculated 

 from the moisture equivalent, that we suspected the explanation might 

 lie in the differences in the values of the hygroscopic coefficient obtained 

 for similar soils by our respective methods. However, as no moisture- 

 equivalent apparatus was at that time available for our use we were unable 

 then to decide the question. Now, using 135 samples of which hygro- 

 scopic coefficient determinations had been made at the Nebraska Experi- 

 ment Station, we have determined the moisture equivalents, thus obtain- 

 ing a definite answer to the question. 



Lipman and Waynick (12) have recently reported both the moisture 

 equivalents and the hygroscopic coefficients of 27 soils, and from these 

 the ratios may be calculated. In so far as we are aware, there are no 

 published data except those in the two articles mentioned from which the 



