212 NORTHEAST EXPERIMENT FARM. 
varieties, it will not usually do its best unless sown near 
June 1st. The best yields of millet were obtained in 1902 on 
bottom-land brokenin’97, on whichsod had been fall plowed 
in’01. Table CIV gives the yield in tons of seven varieties 
and the result of late sowing with two kinds. The yields 
must be discounted slightly for dampness when weighed. 
The plots were cut Sept. 16th. Siberian and Hungarian are 
of similar earliness with common millet and were fully head- 
ed. The season being cold and late, the German millet even 
in plot 7 was not headed out yet when cut. Early Fortune 
is a hog millet but has twice proven very poor, perhaps due 
to poorseed, and the plot was nearly all weeds. The Japanese 
millet is quite coarse and tall which accounts for the large 
yields but does not improve its feeding value. Plots 10 and 
11, sown June 30th are seen to be much lighter in yield than 
those sown on the 1st or 15th. Plot 9, Common, was head- 
ed and fit to cut, and plot 11 had begun to head, though the 
stalks were dwarfed. Butthe German millet on plots 8 and 
10 was quite immature. Millet must be cured in the fall, 
and the task isnot easy. This fact must be considered be- 
fore sowing too much of it. 
Peas are a success only upon rather heavy soil. In’96, 
on light sandy soil 9 varieties gave yield of from 8 to 13.5 
bushels per acre. In ’98, these varieties, grown on bottom- 
land underlaid with clay, yielded from 15 to 28.8 bushels per 
acre. They are valuable for hog feed, and where soil is favor- 
able, may be grown for such purpose. The harvesting is 
difficult. Peas should be sown as early as possible. They 
cannot be sown too early, and if late sown they suffer from 
mould. 
Corn.—The experiment farm has each year tested a num- 
ber of kinds of flint and dent corn, amounting in all to 25 
varieties of flint, and 30 of dent. No attempt was made to 
keep these strains separate as it would have been impossible, 
and they must be grown side by side for proper comparison. 
The effort has been to determine the exact status of Grand 
Rapids as a corn producing locality by finding out what 
kinds would ripen there year after year. The data obtained 
applies to other localities directly, only when the conditions 
are nearly the same, but can be used as a basis of compari- 
