131 
In the following tabic I have given the average annual increase found <>n these 
plats in .section C by the different methods of computation: 
Increase (on ml on section Cby different methods of computation. 
Plat. 
By ilea r- 
" est 
checks. 
By aver- 
age oi all 
checks. 
By ex- 
treme 
checks. 
17 
1,024 
1,071 
1,127 
1,132 
1,334 
1,166 
1,323 
1,292 
831 
912 
1,003 
1,000 
1,187 
1,463 
1,915 
2, 198 
377 
21 
458 
23 
249 
24 ? 
612 
11 
732 
26 
1,009 
1,471 
1,744 
27 
29 
A glance at this table and a comparison of the two diagrams given show that 
whereas our method of calculation yields practically consistent results, the outcome of 
either of the other methods is altogether unreliable. 
In view of this comparison I find it easy to understand why some experimenters 
who have contented themselves with the use of but one or two check plats in a series 
should have come to the conclusion that there is no possibility of obtaining reliable 
results in field experiment. 
I do not claim that it is possible to obtain the exact results in this form of experi- 
ment that may be reached in the chemist's laboratory, where all the conditions are 
under absolute control. The field experiment deals with general problems, the lab- 
oratory with special ones, and the field experiment must be interpreted with respect 
to its limitations. The field experiment bears somewhat the same relation to labo- 
ratory investigation that the oil painting does to the photograph. The painter's 
brush can never approach the camera in exactitude and minuteness of detail, yet the 
goal of photography is the blending into one harmonious and true picture of land 
and sea, mountain, forest, and stream, sky and cloud, which the master painter 
attains. The experiments I have been describing are made on a soil of very indefinite 
history; they have been in progress not quite long enough to complete two rotations 
on the separate sections. During the ten years covered by this investigation have 
occurred five seasons of greater or less injury to the wheat crops from unfavorable 
seasons and Hessian fly, in one of which the average wheat yield for the State was 
reduced to the lowest point in half a century. The corn crop has also suffered from 
drought and insects, the one just harvested especially being severely injured by white 
grub. Both this insect and the Hessian fly, however, are apt to distribute their 
operations over entire fields, so that the comparative results are not so much dis- 
turbed as they would be in case of chinch-bug attack, which we have fortunately 
thus far escaped. 
But these insect attacks have added to the practical value to the farmer of the 
experiments, in demonstrating that manure and the fertilizing materials in common 
use do not prevent such attacks, but nevertheless accomplish a useful service in 
enabling the plant to overcome them. 
Investigations of this kind may include the growing of crops in continuous culture, 
in order to study the feeding habits of different plants, but rotative cropping is nec- 
essary when the object in view is the study of fertility maintenance. But no satis- 
factory results can be secured from an experiment in rotative cropping unless each 
crop of the rotation is grown every year. An experiment in which the different 
crops follow each other without any fixed plan, either of cropping or of fertilizing 
can not be called a scientific investigation, and if the experiment is to be made this 
season here and next season yonder it might as well not be made at all. 
In illustration of the importance of continuity in this work, I give a table showing, 
in two five-year periods, the total value of the increase on a few of the plats in the 
five-year rotation under consideration, estimating corn at one-third of a dollar per 
bushel, oats at 25 cents, wheat at two-thirds of a dollar, stover at $3 per ton, straw 
at $2, and hay at $6.66f, thus reducing the crops to a common denominator based 
upon approximate market values, which has this advantage over giving merely the 
total weight, that it brings out more distinctly the increase or decrease of the more 
valuable constituents of the crop, and is therefore the more truly scientific method. 
