Relation of Nonavailable Water to Hygroscopic Coefficient 17 
ter does Dot need to be elevated to the surface foot of soil in 
order to become available, g. In dry-land experiments tillage 
operations should be governed by actual determinations of the 
moisture conditions of the subsoil, but a fair idea of these con- 
ditions is revealed to the experienced eye and hand without the 
weighing or drying of samples. 
To account for the failure of very heavy crops, maturing 
during almost rainless weather, to reduce the- moisture as low 
as the hygroscopic coefficient it was suggested that the lower 
limit of water available for the normal growth of plants was 
considerably higher than the hygroscopic coefficient— 4.5 to 6.0 
per cent — depending upon the value of the hygroscopic coeffi- 
cient — for the two soil types in question in that particular study. 
For the portion of the soil moisture above this lower limit the 
designation "water probably available for the support of normal 
plant growth" or ,; X water" was used. 
Several years after the publication of the results of this in- 
vestigation the writer found that the slight modification which 
he had introduced into Hilgard's method — the employment of 
shallow pasteboard trays to hold the glazed paper instead of 
placing the latter directly on the wooden shelf — while giving con- 
cordant results with duplicate determinations made at the same 
time or in the same manner at different times, invariably gave 
values much below the true coefficients. Thus the results while 
showing the relative hygroscopicity of different samples gave too 
high a free water content. Unfortunately, on removing to the 
University of Nebraska the writer had discarded all the samples 
taken at Indian Head for moisture determinations. However, 
part of another series 1 on which determinations of the hygroscopic- 
coefficient had been made in the same way and only a few weeks 
later had been saved. New determinations on the latter gave 
values 1.5 times as high as the earlier ones. It is evident that 
had the original determinations of the hygroscopic coefficient of 
the different samples been correct the importance of the de- 
termination of the hygroscopic coefficient would have been much 
more distinctly shown and there would have been a still closer 
connection between the data on the free water content of that 
portion of the soil penetrated by plant roots and both the field 
notes at the time of sampling and the previous history of the 
fields. 
Later field studies 2 at widely separated points in the semi- 
1 Alway, F. J., and McDole, G. R. Studies on the Soils of the Northern 
Portion of the Great Plains Region — The Distribution of Carbonates on 
the Second Steope. American Chemical Journal, 1907, vol. 32. o. 275. 
2 Alway, F. J. Moisture Studies of Semiarid Soils. Report of Winnipeg 
meeting of the British Association for the Advancement of Science, 1909,. 
