30 Nebraska Agricultural Exp. Station, Research Bui. 6. 
Otherwise, the difference in water requirements may be partly 
due to the difference in soil fertility. Thus, Briggs and Shantz 
(1913, Bui. 285, p. 61) compare the water requirements of both 
wheat and sorghum in Colorado and Texas during 1910 and 1911. 
Native soil was used in the tests, which may account for part of 
the difference assigned to climate. Briggs and Shantz recognize 
this factor in 1913 (Bui. 284, p. 15). 
King (1905) compared the water requirements of corn in four 
States, using native soil in each case. No data are given by which 
we can estimate what portion of the difference in the several 
States is due to difference in soil fertility. Such comparisons 
give the combined effect of difference in soil and climate. 
The need for similar precautions regarding uniformity of soil 
fertility applies to determining the variation in water require- 
ment from year to year as well. Variation in soil fertility may 
influence the water requirement fully as much as climate. 
IMPROPER DISTRIBUTION OF SOIL MOISTURE. 
In experiments to determine the effect of variation in soil- 
moisture content, some investigators have added all the water 
either from above or from below. In those potometers having a 
low relative soil saturation, water added to a single portion may 
not become distributed thruout the soil mass, which will unin- 
tentionally affect the degree of saturation in the different parts. 
EVAPORATION FROM SURFACE OF SOIL. 
Of the forty-odd experimenters who have studied the water 
requirements for the production of dry matter with plants beyond 
the seedling stage, only eight have taken the necessary precaution 
to reduce the evaporation from the soil surface to a negligible 
quantity. Many have attempted to correct such loss by sub- 
tracting the water loss of similar uncropped pots from the com- 
bined evaporation and transpiration of cropped pots. The rate of 
surface evaporation is, however, affected by so many factors that 
not much faith can be placed in the accuracy of the method. 
The evaporation rate varies with the relative soil saturation, the 
soil texture, and the shading, which are certain to be variable 
factors between cropped and uncropped pots as well as between 
pots intended to be similar. In some experiments conducted in 
this manner, the amount of water transpired by the crop is so 
much larger than the amount evaporated from the soil that the 
experiment may indicate the character of the effect altho not in 
definite degree. We cannot assume that such experiments furnish 
dependable quantitative measurements. 
Some investigators have made the error of assuming that the 
