Transpiration as a Factor in Crop Production. 
35 
Widtsoe (1909) has been inconsistent in making a similar error 
of averages in tests where the effects of different soils and treatments 
have been averaged for four years. Data secured during one, two, 
three, or four years are given equal rating in comparing one with 
the other, and this error occurs in his Tables 6, 7, 8, 15, 20, 21, 27, 
28, and 30, while it is partially or wholly recognized and over- 
come in Tables 5 and 24. 
An error of averages may be made in averaging the water 
requirements in the production of grain and of total dry matter. 
Some potometers may fail to produce grain. For the purpose of 
comparing the relative water requirements for grain and for total 
dry matter in a certain crop, it is important that the same po- 
tometers be averaged in both cases. If certain potometers fail to 
produce grain, this should be charged against the average. If 
only four or five potometers out of six duplicates produce grain, 
the total of these four or five should be divided by the full number 
of potometers, or six. Briggs and Shantz (1914, Tables 6, 7, 13, 
14, and 24) may be cited as examples of this error. An illustration 
of the point in question may be found on page 60, Table 4, 
column 3 of this bulletin. Three out of four duplicate small 
potometers failed to produce an ear. This lack of ear develop- 
ment due to shortage of available nutrients must be charged against 
the small potometers, and the weight of the one small ear produced 
must be divided by four to secure the average performance. 
In the table which follows is given a list of the investigators to 
date who have based the transpiration requirement upon crop 
yield and used plants grown in potometers beyond the seedling 
stage. Data concerning a number of factors are also tabulated to 
show the lack of uniformity, and also the opportunity for error 
from a number of sources. A few of the investigators did not 
conform entirely in all of their investigations to methods credited 
to them in the outline. The publications of some also fail to givo 
full information concerning part of the factors. 
