EFFECT OF SOIL FACTORS ON WATER REQUIREMENT. 23 
out at other times. The experiments extended from April 29 to July 
16 and July 30. The results of these investigations are given in Table 
XIII. The plants were grown in open pots, without bare check pots, 
but a correction for loss due to evaporation from the soil surface was 
made in the following way: The mean value of the water consump- 
tion and of dry matter, respectively, for four pots from each moisture- 
content series, grown without nitrogen, was subtracted from the water 
consumption and dry matter of each individual fertilized pot of that 
series. The ratio of the remainders was taken to represent the water 
requirement of the plants grown in that pot. This correction ranges 
from nearly one-half of the total water consumption in the case of the 
low-nitrogen pots to about one-seventh of the water consumption in 
the high-nitrogen pots. 
An analytical examination of this method of correcting for evapo- 
ration shows that it can not be justified. 
Let f =total water transpired and evaporated from pot without nijtrogen. 
t = total water transpired and evaporated from pot with nitrogen. 
e Q = water evaporated from soil of pot without nitrogen. 
e = water evaporated from soil of pot with nitrogen. 
w =dry weight of unfertilized plants. 
w =dry weight of fertilized plants. 
r =water requirement of unfertilized plants. 
r = water requirement of fertilized plants. 
Then, by definition, * -g 0== 
w 
and, similarly, t—e , x 
— =r- (2) 
Subtracting equation 1 from equation 2, we have — 
t-t +e -e=nv-r iv . (3) 
Pfeiffer, Blanck, and Flugel, in calculating the water requirement, 
took the difference between the water loss from the fertilized and un- 
fertilized pots and divided this by the corresponding difference in dry 
weight. The quotient was considered to be the water requirement 
of the plants in the fertilized pot. Expressed as an equation, this 
becomes — 
-f—^=r (4) 
w—w v ' 
Or, t-t =nv-rw . (5) 
Subtracting equation 5 from equation 3, we have— 
e -e=( r -r o )w , (6) 
which is the assumption that PfeifFer, Blanck, and Flugel really make 
in following their method of reducing the results. Writing equation 
6 in the form — 
e —e _ 
w Q ~ r r °' 
it will be seen, first, that their assumption is that the difference in the 
transpiration ratios of two series is proportional to the difference hi 
285 
