Relation of Xonauailable Water to Hygroscopic Coefficient .7 
water, it being assumed that the loss in the field was at the same 
rate per unit area of surface. The transpired water, thus calcu- 
lated, exceeded the precipitation and to account for the difference 
it seemed necessary to assume that the soil could have absorbed 
and rendered available to vegetation the ever-present water vapor 
of the atmosphere. Later it was found that the loss of water by 
transpiration depended upon the degree of moistness of the soil 
and that, hence, there was no evidence of any such deficit. Ex- 
periments by Mayer and by Heinrich indicated that plants were 
unable to make any use of the hygroscopic moisture and that 
none of the moisture of the air was rendered available to the plant 
thru the absorbent power of the soil, and Mayer discouraged 
even indicating the hygroscopicity of soils on the ground that it 
would cause confusion if very hygroscopic soils were marked with 
a defect in regard to the water supply when this was not of much 
practical significance for the reason that these soils were the very 
ones that had the highest water capacity. 
Following this proof of the falsity of the old conception of 
the significance of the hygroscopicity of soils came the view, 
carried to the opposite extreme, that a knowledge of the hygro- 
scopic power of different soils is practically valueless, the ques- 
tion of the usefulness of a knowledge of the relative hygrosco- 
picity of soils being apparently confused with the distinct ques- 
tion of the usefulness to plants of the hygroscopic moisture of the 
soil. 
Loughridge was the first in connection with field studies of 
soil moisture to determine the hygroscopic coefficients of all the 
samples dried and to tabulate the free water as well as the total 
water. Loughridge's work first suggested to the writer the de 
termination of the hygroscopic coefficients of the different soil 
samples, taken in connection with a field study of soil moisture 
in 1904-1.905, after he had found that there was little agreement 
between the data on the total soil moisture and his field notes. 
The data on the free water, on the contrary, agreed with the field 
observations. The results of this field study suggested the pot 
experiments reported in the following pages. These experiments 
were carried out in connection with and subordinate to various 
soil investigations, the object being to determine the relation of 
the final water content of the soil to the hygroscopic coefficient 
and to determine to what extent the moisture was removed from 
the different portions of the subsoil not penetrated by plant roots. 
For assistance in the experiments the writer is indebted to 
Dr. W. D. Bonner, Dr. E. A. Gortner, Mr. G. R. McDole. Prof. 
R. S. Trumbull and the late Mr. W. A. Wnllschleger. 
