400 
Journal of Agricultural Research 
Vol. XI, No. 9 
PREVIOUS STUDIES IN FIELD METHODS 
Before the development of the row method, a series of attempts were 
made to develop a standard practice for field plot variety tests. Regard¬ 
ing many of the points mentioned in the earlier papers, there is as yet 
insufficient experimental evidence. 
Smith (j 3 ) 1 in discussing com variety tests mentioned the disad¬ 
vantages of growing a short variety beside a taller one. The probability 
that some varieties have stronger foraging powers than others is given 
as a further reason for discarding border rows. Plots five rows wide 
are recommended. 
Cory (2) presents the plan of using row plantings to check the results 
of field plot experiments. The plan outlined was the use of two rows 
12 inches apart and 50 links long. From result of these studies the 
author concluded that the row results were more reliable as an indication 
of yielding ability than the results obtained from plot tests. 
The work of Marek is cited by Wheeler (16) as showing the possible 
errors due to greater growth on the edges of plots. The plan of removing 
a certain border part of each plot before obtaining yields is recommended. 
Piper and Stevenson (10) recommend certain practices for variety 
testing with the hope of standardizing methods. These recommenda¬ 
tions are based on experimental studies of earlier investigators. Papers 
by Thome (13) and Carleton ( 1 ) are especially mentioned as being 
valuable contributions to the idea of standardization. For corn variety 
work they adopted the method as outlined by Smith ( 13) with the addi¬ 
tion of two to five replications for each variety and check plots* every 
third or fourth plot. For small-grain testing the use of one-fortieth to 
one-tenth-acre plots in size with two to five replications and checks 
every third plot is recommended if one decides to use the plot method. 
For row tests a length of at least a rod with checks every fifth row and 
ten replications for each test is offered as a standard practice. 
Mercer and Hall (5) reported a careful series of experimental tests in 
which they compared the relative reliability of results obtained from 
different-sized plots and gave results showing the reduction in error 
due to plot replication. The authors decided that one-fortieth acre 
plots were large enough for all practical purposes and recommended 
five replications for each variety under test. The conclusions of Lyon 
(4) in reference to plot size and replications are in substantial agreement 
with those of Mercer and Hall. Conclusive evidence is given to show 
that it is not possible to establish a schedule of relative yields for a series 
of plots even after several years' comparison. 
Montgomery (7) presented an extensive study of the experimental 
error in nursery work with wheat for the characters nitrogen content 
1 Reference is made by number (italic) to “literature cited,” pp. 41&-419. 
