CEREAL INVESTIGATIONS ON THE BELLE FOURCHE FARM. ork 
In 1909 and 1913, as shown in Table XI, fairly good yields of both 
grain and straw were obtained. In 1910 the yields were very small, 
and in 1912 there was not sufficient growth on any of the plats to 
make them worth harvesting. In the particularly favorable year 
1909, the higher rates of seeding gave the best yields. There was 
little difference in the yields from the different rates in the somewhat 
less favorable year 1913. In 1910, a very unfavorable year, the 
plats from the lower rates produced the best yields and in 1912 they 
made the best growth early in the season. None of the plats, how- 
ever, produced grain in the latter year. : 
The average yields of grain for the four years from the 4, 5, 6, and 
8 peck rates show very slight variation. The highest yield, 8 bushels 
to the acre, was produced from the 8-peck rate, while the lowest 
yield, 7.2 bushels, was produced from the 6-peck rate. The 4-peck 
rate produced an average yield of 7.4 bushels, only 0.6 bushel less 
than was produced from sowing 8 pecks. In other words, the extra 
bushel of seed required at the 8-peck rate produced a gain of only 
0.6 bushel in the resulting crop, so that the higher rate really entailed 
a slight loss. The yield of straw is considerably higher from the 4 
and 5 peck rates than from the 6 and 8 peck rates. 
While the data here presented are far from conclusive, they indi- 
cate that it is safer to sow 4 or 5 pecks of durum spring wheat in 
western South Dakota than to sow a larger quantity. Because of 
the smaller size of the kernels of common wheat and the consequent 
greater number in a peck, the proper rate of seeding for spring com- 
mon wheat is 3 or 4 pecks to the acre. 
NURSERY EXPERIMENTS WITH SPRING WHEAT. 
The nursery experiments with spring wheat on the Belle Fourche 
Experiment Farm have been confined to the testing of pure-line 
selections of the more important varieties. In 1908 about 100 heads 
each of durum wheats and of spring common wheats were selected 
from the varieties in plat tests. These were grown in head rows the 
following year. In 1910 these selections were grown in 60-foot rows, 
and in 1911 in 60-foot rows replicated four or five times. The durum- 
wheat selections were complete failures in 1911 and produced little 
grain in 1912, so that only a very small quantity of seed was avail- 
able for testing in 1913. Only slightly better results were obtained 
from the selections of common wheat. 
None of the selections has as yet shown any marked superiority 
over the parent varieties, though there has been considerable varia- 
tion among the different selections from the same variety. These 
variations consist principally of differences in earliness, height, yield, 
and quality of grain. In many cases a selection which has appeared 
to be particularly good one year has given poor results the next, so 
