310 TWENTY-FIFTH ANNIVERSARY REPORT. 
The first phase'* of this work started by this Station was the 
sacking of clusters on grapevines before blooming time to de- 
termine if the variety whose cluster was sacked was capable of bear- 
ing fruit without foreign pollen. It was soon found that varieties 
of grapes could not be divided arbitrarily into two classes, self- 
sterile and self-fertile, but that there were various degrees of self- 
sterility and self-fertility. One variety would set a perfect bunch | 
of fruit in the sack, another would set a bunch with many of the 
berries lacking and the bunch consequently looser than those on the 
same vine which set outside the sacks. Others, as has already been 
mentioned, would set no fruit at all if the blossoms were covered. 
From the data gathered lists were published in which the grapes 
are divided into four classes: The first, all those varieties which set 
perfect bunches in sacks; the second, those which set bunches some- 
what loose but not sufficiently so as to be unmarketable; third, those 
whose bunches were so loose as to be unmarketable; and fourth, 
those producing no berries or practically no berries from the covered 
blossoms. “ With many varieties the degree of self-fertility is not 
an unchangeable characteristic even when the vines appear to be in 
a normally productive condition, but varies under differences of 
environment. Many other varieties which have been under obserya- 
tion showed practically no variation in this respet. Usually where 
no variation in self-fertility is observed with a variety it is confined 
within rather narrow limits.” 
One hundred and sixty-nine varieties were included in the lists 
given in Bulletin 157, in which the varieties were divided as above 
indicated. 
It was found in later investigations! in which varieties were 
hand pollinated with the pollen of various other sorts, that varie- 
ties which are themselves self-sterile do not make good pollen- 
izers for other self-sterile sorts, and that the value of a variety 
as a furnisher of pollen for fertilizing is almost identical with 
its self-fertilizing “capacity. The rule was consequently deduced 
that a strongly self-fertile variety should always be taken as a pol- 
lenizer for self-sterile sorts. Thus if one wish to plant Brighton, 
a quite popular variety in some neighborhoods, although it is 
practically self-sterile, Wyoming Red would not do for a fertil- 
izer because it is also self-sterile and these two varieties planted 
“Bul. 157; also in Rpt. 17:518 (1898). 
* Bul. 169; also in Rpt. 18:361 (1899). 
