282 TWENTY-FIFTH ANNIVERSARY REPORT. 
plat tests on larger areas,than were used in testing the varieties 
showed the advantage to be with seed from productive hills not 
only when equal numbers of tubers were compared, but when the 
weights of seed of the two kinds were equal. In 1903, ’04 and ’o5, 
these comparisons were again taken up, testing seed from heavy 
and light hills from the same parent crop in rows side by side and 
repeating the series several times to overcome inequalities in soil 
conditions. Of the six comparisons, four gave substantial differ- 
ences in favor of using the seed from heavy hills, one a slight 
difference in the same direction, and one a slight difference in 
favor of seed from light hills. In the second year’s test with 
two varieties and in the third year’s test with one of them, light 
yielding hills were selected from crop grown from seed previously 
selected from similar hills, and heavy hills from crop grown from 
seed produced in heavy hills in order to ascertain if the differences 
in favor of seed from heavy hills were cumulative. No such cumu- 
lative effect was shown conclusively, but there does appear to be a 
profitable margin of gain in selecting seed for each crop from the 
best hills of the preceding crop. 
This presupposes that the grower raises his own seed; but some 
have strongly advocated a “ change of seed.” Extensive tests along 
this line have not been made at the Station, but one test,**® with the 
same variety grown from Station seed and from seed of the same 
variety from two other localities, did not favor the practice. The 
yields were somewhat better from foreign seed grown on similar 
soil, but much poorer from seed grown on soil of different character. 
Prelinunary seed treatment.— In some foreign countries, particu- 
larly France, the growers of very early potatoes plant only tubers 
that have been previously started into growth by exposing them to 
warmth and light. By this means short, thick sprouts are formed, 
which grow quickly and vigorously when the tubers are carefully 
planted with the sprouts up. A test** of this method was made by 
the Station in 1888 on four twentieth-acre plats; but no advantage 
in earliness resulted from the preliminary treatment and difference 
in yield was exceedingly small. In another test*® made this same 
year a disadvantage resulted from the use for seed of potatoes 
with long sprouts, though this loss was less than was expected. It 
appeared to make no difference whether the sprouts were broken off 
*“ Rpt. 9:383-386 (1890). 
“Rpt. 7:167, 168 (1888). 
* Rpt. 7:165-167 (1888). 
