I \)TAT() 1 N \ ESI'KJATIOXS 
17 
lively long and slender, so it is probable that many tubers weiV 
thru the screen endwise which would have gone over if they 
had been round tubers of the same weight. In commercial grading 
it is customary to use a 1%-inch screen for this variety. Comparing 
the four varieties which are most extensively grown in the Greeley 
district we find that the two late varieties, the Rural New Yorker 
and Pearl, rank considerably higher than the two early varieties, 
the Charles Downing and Early Ohio, in the percentage of market- 
able tubers with the first named variety exceeding the Pearl by 
6.4 per cent. 
RUNNING OUT OF VARIETIES 
This experiment was conducted in connection with the variety 
test to determine if possible whether the productivity of a variety 
can be maintained for a period of years without change of seed. 
This phase of the work was started in 1915 with seed from 10 
sources, representing seven varieties. Seed was saved from the crop 
produced by each lot and grown the following year in comparison 
with new seed which came from the same source from which the 
1915 seed was obtained. In 1917 the new seed was compared with 
stock from the same original source which had been grown at the 
station one and two years, respectively, while in 1918 comparisons 
were obtained between new seed and 1-, 2- and 3-year-old seed, all 
from the same original source. Because of the increase in the 
number of plats from year to year it was necessary to gradually 
reduce the number of varieties because of lack of space until in 
1918 there remained only two varieties, the Pearl from Wisconsin 
and the Rural New Yorker from Colorado. The results are shown 
in Table 8. 
In 1916 larger yields were obtained from the station seed in 
eight of the ten trials, the average increase in yield over that pro- 
duced by the new seed being 13.3 bushels, In the two instances 
where new seed produced larger yields the seed came from Wiscon- 
sin. In 1917 in four out of five trials the largest yields were pro- 
duced by the new seed. The yields from the two-year-old seed were 
in general slightly lower than those obtained from seed which had 
been grown at the station but one year. 
In 1918, the number of sources from which new seed was 
obtained was reduced to two, the Pearl from Wisconsin and the 
Rural New Yorker from Colorado. In the case of the Pearl there 
was a gradual decrease in yield corresponding with the length of 
