t<> be found on any soil that has been long in cultivation. The soil is a Bandy clay, 
quite uniform in texture, easily worked, and very responsive to treatment, whether by 
tillage. or by fertilizing. 
The land was laid off in plats 16 feet wide by 16V rods long, this width of plat 
being well adapted to the various kinds of agricultural machinery. The plats were 
slightly ridged at the outset of the work, thus providing independent surface drainage 
from each plat, and this ridging is repeated at ten-year intervals. At other times 
the land is plowed across the plats. The plats are separated by paths 2 feel wide, 
and under every alternate path a tile drain is laid, thus providing a drain on one 
side or the other of every plat, and locating the drains 36 feet apart. In a few cases 
the drains have been put under every path, and it would be better for each plat to have 
its drain either under the path or under the middle of the plat. On clay soils 18 
feet would be none too close for drains for this work. 
We have not yet discovered any reason for making the dividing spaces between 
the plats wider than 2 feet. On the contrary, with care in application of fer- 
tilizers and with our method of ridging, by which the plats are separated by dead 
furrows which the plant roots are reluctant to cross, we believe that this distance is 
better than a wider one would be. In the case of oats or wheat, sown with the 
ordinary grain drill, a 2-foot space between plats gives 2 feet 8 inches between rows 
of grain, which is sufficiently wide for cultivation when that is necessary to keep 
the weeds down. 
< 
<* 
■< 
i 
1 
3500 
3 000 
2500 
2000 
PLAT NO. 
bH im ■ 
1 t rr ■ ■ " 
1 
2 
3 
4 
5 
6 
7 
8 
9 
: 
M 
12 
!3 
14 
: 
6 
17 
18 
\ 
:: 
1 
22 
23 
24 
25 
26 
n 
28 
29)30 
Diag. II.— Ten-year average yield and increase on Section C, Ohio Experiment Station. 
It was fortunately possible to lay off our plats generally across the previous plow- 
ings, thus eliminating the discrepancies caused by old back furrows and dead furrows. 
This is an important matter, as an old ridge or furrow may completely reverse the 
results of a plat experiment. 
At an early date in the history of our work in field experimentation we became 
conviuced of the necessity for frequent repetition of check plats, and, therefore, in 
the experiments with fertilizers now in progress every third plat has been left con- 
tinuously unfertilized. This gives an unfertilized plat adjoining every fertilized one, 
and is a compromise between the ordinary method and the ideal method of leaving 
alternate plats as checks. Our only reason for not employing this last method was 
the immense amount of land required. We hoped, at the outset, that after some 
years' work the check plats might develop sufficient uniformity to permit the use of 
a part of them for other purposes; but at the end of ten years we find ourselves 
apparently no nearer ready to dispense with any of the checks than at the outset. 
In calculating the increase from the fertilizers we assume that variations between 
the check plats are progressive, and on this assumption we compare each pair of fer- 
tilized plats with the two check plats between which they stand, by treating the four 
plats as the four terms of an arithmetical series. We never make our comparisons 
on the basis of simple averages, either of all the plats or of the two nearest checks. 
To illustrate the results of this frequent repetition of checks and of our method of 
calculation, I have prepared the accompanying diagrams, which show the average 
increase found on the different plats in one of our rotation experiments, the five-year 
rotation at Wooster. 
21736— No. 142—04 9 
