350 REporRT OF THE HorvIcULTURAL DEPARTMENT OF THE 
DIFFERENCES IN DISTRIBUTION OF DIFFERENT VARIETIES OF GRAPES. 
How greatly varieties may vary in average specific gravity, 
that is in distribution of the seeds according to specific gravity, 
is abundantly brought out in Table I, in which are shown records 
of separations of a number of varieties of grape seeds grown 
under normal conditions in a variety vineyard. For instance, 
much the larger part of the seeds of Agawam are found in the 
range from 1.07 upward, which is the range within which alone 
germination takes place, that is the range within which good 
seeds are comprised. But in the case of Canada most of the 
seeds are comprised in the range from 1.083 downward, in which 
range germination does not take place in this variety. In the 
case of the particular observations shown in this table, the very 
great differences in distribution of seeds is probably due in large 
part to unequal protencies of the pollen which chance to alight on 
the pistil. That there is ground for this assumption is abund- 
antly proven by examinations of seeds, both cross and self-ferti- 
lized, made by the writer for Prof. Beach, and representing some 
twenty different varieties. In the case of such strongly self-fer- 
tile varieties as Concord and Worden it was found that the dis- 
tribution of the seeds as regards specific gravity was about the 
same whether the flowers were self-pollinated or cross-pollinated. 
But -in the case of varieties which Prof. Beach has heretofore 
examined and classed as imperfectly self-fertile, it was found that 
very striking differences in the distribution of the seeds as re- 
gards specific gravity—that is as regards quality—are correlated 
with the potency of the pollenizing parent. For instance, in the 
case of Little Blue, an imperfectly self-fertile hybrid, only 30 per 
ct. of the seeds from self-pollinated bunches were found in the 
range from 1.07 up; but other bunches from the same vine open 
to cross-pollination showed 60 per ct. of seeds within the same 
range. See Table II and Chart I. Even in this latter case 
there was a very large percentage of poor seed. It is a fair 
question whether a considerable part of these may not have been 
due to imperfect pollination from some source. It would be in- 
teresting to compare the specific gravity of the seeds of this 
variety self-pollinated with the seeds of the same variety cross- 
pollinated with some strongly self-fertile varieties as Concord. 
