26 BULLETIN 297, U. S. DEPARTMENT OF AGRICULTURE. 
the following spring. The highest yield was obtained from the plats 
sown on September 16. The poor yield from the plat sown sa 
ber 1 was due to low germination. 
In the fall of 1910 the two sowings Soman at about the same 
time and the plants made a vigorous growth. The plat sown Sep- 
tember 16 showed good germination but less growth than the two 
earlier seedings. The plants on the plats sown October 1 emerged 
but made little growth before winter, while in the plats sown Octo- 
ber 16 and November 1 the seed germinated but the plants did not 
emerge. As there was little precipitation during the fall, winter, 
and spring, practically all of the plants were dead by May 1. There 
was a higher moisture content and better survival in the late-sown 
Fig. 7.—Plats of winter wheat in the date-of-seeding test on the Belle Fourche Experiment Farm. (Photo- 
graphed October 24, 1912.) 
than in the early-sown plats. The fall of 1911 was so dry that the 
date-of-seeding as well as other tests with winter wheat were a com- 
plete failure. 
In the fall of 1912 the test was replicated three times on fiftieth- 
acre plats instead of being sown on a single tenth-acre plat. An 
earlier and a later date of seeding, August 6 and November 16, 
respectively, were added to the test. The seeding of September 16 
was not made, because the soil was so dry that germination could not 
be obtained. Good stands were secured from the plats sown on each 
date.. A view of the plats on October 24 is shown in figure 7. The 
earliest date of seeding produced the highest yield, 29.2 bushels. 
Because of the variations from year to year, the difference in 
average yields obtained from the various dates of seeding was not 
great. If the plats sown October 1, 1909, had not failed to germinate, 
that date probably would have given the highest average yield for 
the four years when crops were produced. The next highest yield 
