SPRING CEREALS AT MORO, OREG. 11 
Table VIII shows that June is the windiest month, though there is 
little difference in the monthly averages for any month from April to 
August, inclusive. The highest average wind velocity for any year 
in the 5-year period was 7.5 miles per hour, in 1911. The year 1911 
was also the one with the least precipitation and the greatest evap- 
oration. 
EXPERIMENTAL METHODS. 
The experimental work with cereals is conducted for the following 
purposes: (1) To ascertain what grain varieties are best adapted to 1 
the drier districts of the Columbia Basin, (2) to improve the best 
adapted varieties by selective breeding, (3) to determine what crop- 
rotation systems will prove most profitable in connection with grain 
raising, and (4) to find out the most economical and profitable methods 
of soil tillage for cereal production in this section. 
The work with spring grains during the first few years was along 
two lines, the testing of varieties and the improvement of varieties 
by pure-line selections. No work in hybridization has been attempted. 
A large number of pure lines have been developed, some of which 
give promise of being of considerable value. In this bulletin only the 
results of the varietal experiments will be given. 
In the testing of varieties, field plats and nursery rows are used, 
the unit of comparison being a tenth-acre or twentieth-acre in the 
plat experiments and 1-rod or 2-rod replicated rows in the nursery 
experiments. Eight-rod rows and eightieth-acre plats also have been 
used in the nursery. While a rather careful study is made of the 
varieties in the nursery and in the plat experiments and numerous 
notes taken on the habits of the plants, the value of a variety has been 
judged largely by its yield of grain. 
DIMENSIONS OF PLATS. 
As shown in figure 3, the substation farm is laid out in series of 
tenth-acre plats 8 rods long and 2 rods wide. The series are sepa- 
rated by roads 16^ feet wide and are divided into blocks of 10 plats 
each by roads of the same width. Alleys 4J feet wide separate the 
individual plats within thqse blocks. 
When twentieth-acre plats were used, each tenth-acre plat was 
divided in half by an alley 22 inches wide. The twentieth-acre plats 
discussed in this bulletin actually contain only 2,062.5 square feet, or 
115.5 square feet less than one-twentieth of an acre. In computing 
yields per acre, however, no account has been taken of this shortage. 
Acre yields computed from the yields of small plats are likely to be 
slightly exaggerated on account of the increased proportion of margin. 
The plan first adopted at the substation for the varietal experi- 
ments was to grow a single tenth of an acre of each variety, with 
