10 BULLETIN 297, U. S. DEPARTMENT OF AGRICULTURE. 
was May 13. The earliest frost in the autumn during this period was 
on August 27, while the average date of the first frost was September 
14. The average frost-free period for the six years is 123 days. 
TaBLE VIIIl.—Dates of killing frosts, the last in spring and the first in autumn, with 
temperatures recorded and length of the frost-free period for each year from 1908 to 1913, 
inclusive, at the Belle Fourche Experiment Farm. 
[Data from the records of the Biophysical Laboratory of the Bureau of Plant Industry.] 
Last frost in First frost in Last frost in First frost in 
spring. fall. spring. fall. 
Frost- Frost- 
Year. free Year. free 
Tem- Tem-| period. Tem- Tem- |period. 
Date. |pera-| Date. | pera- ; Date. |pera-| Date. | pera- 
ture. ture. ture. ture. 
salle cE. | Days. sete °F..| Days. 
NGOS. ase May 20 29 | Sept 26 22 WE) MN) MOM = 5345 May 4 32 | Sept. 23 32 141 
1909..... May 17 26 | Sept. 23 3l WAS |W WSS = May 6 32 | Sept. 24 29 140 
IOs s5e2 May 23 31} Aug. 25 32 93 | 
LOSES eee May 11 30 | Aug. 27 32 107 || Average .| May 13 |.----- Sept. 145 -assee 123 
EXPERIMENTAL METHODS. 
The tests with dry-land cereals on the Belle Fourche Experiment 
Farm have been conducted in field plats and in the nursery. In the 
field plats varietal tests and tests of rates and dates of seeding have 
been included. The plats have ranged from one-fiftieth to one-tenth 
of an acre in size. In the cereal nursery the varieties have been 
grown in short rows. The use of the nursery has made it possible 
to test economically a much larger number of varieties than could 
have been grown in the field plats. Careful records have been kept 
of the behavior of the varieties included in both the plat and nursery 
experiments. 
PLAT EXPERIMENTS. 
The field tests have included varietal tests of winter wheat, rye, 
and emmer and of spring wheat, oats, barley, and flax. There have 
also been rate-of-seeding tests with spring wheat and oats and date- 
of-seeding tests with flax and winter wheat. — 
SIZE OF PLATS. 
All of the plat experiments in 1908 and 1909 and nearly all in 
1910 were conducted on tenth-acre plats. These plats were 2 rods 
wide by 8 rods long. They were arranged side by side in series, the 
plats in the series being separated by 5-foot alleys. The series were 
separated by 16.5-foot or 20-foot roads. Each plat thus had a 
5-foot alley along each side and a 16.5-foot or 20-foot road along 
each end. 
Most of the tests in 1911 and all of those in 1912 and 1913 were 
in plats made by sowing a single drill width across an 8-rod series. 
As the drill was 6 feet wide, this gave a plat of one-fifty-fifth of an 
