Nov. 26, 1917 
Field Technic in Rod Row Tests 
401 
and yield. He shows the value of a determination of the experimental 
error by the use of check plots. This error may then be used as a means 
of deciding which strains to eliminate. After determining the experi¬ 
mental error the rule given is 
to double the error and subtract this sum from the highest variant. The remainder 
after the subtraction represents the figure below which all strains could be discarded 
without danger of discarding a high strain 
for the character studied. This conclusion refers only to the seasonal 
conditions under which the test is made. 
From a standpoint of a study of nitrogen content Montgomery (7) states 
that single plants should be replicated 40 or more times, 16-foot rows 
5 to 10 times, and blocks 5.5 feet square 8 to 16 times. From a yield 
standpoint the author finds blocks 5.5 feet square repeated 8 to 10 times 
to be as accurate as 16-foot rows repeated 15 to 20 times. This he 
believes to be due to the competition between adjacent rows, although 
no accurate data on this question are presented. 
In a previous paper Montgomery (6) shows the effects of competition 
on the yield of different varieties when grown in the same row and 
planted alternately in the row'. Under such conditions some varieties 
seem to be much more vigorous than others. 
Increasing the length of the row or size of block is said by the author 
(7) to be a means of securing accuracy. An excellent size for rows was 
found to be 2 to 4 rods in length and blocks 5 by 16 feet in area. 
Montgomery believes the bulk method of planting in nursery trials 
to be more closely correlated with field results than when the nursery 
plots are planted in such a manner that each plant has individual room 
for development. Data given for n varieties of oats tend to substan¬ 
tiate this belief. 
VonRiimker, Teidner, and Alexandrowitsch ( 12 ) followed the practice 
in trials of pure lines of wheat for yield of making the distance between 
rows of the same pure line 13 cm. (5.1187 inches) and between the plots 
of any two pure lines 24 cm. (9.448 inches) to obviate any influence one 
pure line might exert on another growing in an adjacent row. No data 
are presented of the reasons which led them to adopt the method. 
A recent paper of Pritchard ( 11 ) gives a report of a study made in 
1912 in sugar-beet breeding investigations. Two characters, total sugar 
production per row and percentage of sugar, were used in the study. 
These were the basis for a discussion of the value of check plots and the 
comparative effect of repeated plantings upon the experimental error. 
The author concludes that check rows as a standard for the comparison 
of individual progeny rows are more efficient than the use of the. means 
of the progeny rows as a standard for comparison. Check rows 32 inches 
from the progeny rows varied less than checks at a distance of 64, 96, 
128, and 160 inches. Check rows at a distance of 64 to 160 inches seemed 
