36 INTERNATIONAL CONGRESS OF VITICULTURE 



The North Carolina Station reports a start along this line with the 

 Rotundifolia grape, and it is the purpose of this paper to discuss briefly 

 certain results which have been secured at the New York Agricultural Ex- 

 periment Station during some twenty-five years of grape breeding. 



During this quarter century some 200 varieties have been studied more 

 or less extensively, and about 10,000 seedlings have been grown. One of the 

 unexpected results was the failure of many of our commercial sorts to trans- 

 mit desirable qualities, some of the best results being secured from little 

 known varieties. This is why it has seemed best to test such a large number 

 of kinds. 



As a means of analyzing the genetic composition of the varieties more 

 than 3,000 selfed, or pure, seedlings have been grown. While these have 

 given much information about the varieties used as parents, they have been 

 so uniformly lacking in vigor as to lead to the conclusion that improved sorts 

 should not be sought through this method. 



The large number of cases in which Vinifera blood is found in the grapes 

 which are ranked as high in quality, points to the desirability of adding 

 some of this blood to our native species. About 100 varieties of this species 

 have been grown on the station grounds and several of the best have been 

 used in the breeding work. 



Of the factors which are being studied, only a few have sufficient data 

 available as yet to show results which are at all trustworthy. These are 

 discussed below. 



Self Sterility. 



Grape flowers may be divided into three classes: true heraphrodites, 

 hermaphrodites functioning as females because of abortive pollen, and pure 

 males with the pistil absent or rudimentary. Among these classes there are 

 two types of stamens: those with upright filaments and those in which the 

 filaments bend backward and downward soon after the calyx cap falls off. 



Results so far secured would seem to indicate that all pure males have 

 upright stamens and that among the two classes which produce fruit, the 

 true hermaphrodites which are self fertile possess upright stamens, while 

 the self sterile sorts have reflexed stamens. As apparent exceptions to this 

 statement there are a few varieties with reflexed stamens which can set a 

 small amount of fruit when self sterilized. 



From a practical standpoint it is undesirable to grow self steriie sorts 

 since they require interplanting with other varieties in order to secure 

 pollination. From a breeding standpoint, then, one of our problems is to 

 eliminate the reflexed stamens. In studying this problem the following 

 results have been secured: 



The Cross. Ratio of Seedlings. 



Upright stamens X upright stamens = 4.3 upright: 1 reflexed. 

 Reflexed stamens X reflexed stamens = 1.2 upright: 1 reflexed. 

 Reflexed stamens X upright stamens = 1 upright: 1 reflexed. 



The cross upright stamens X reflexed stamens has failed to give a suffi- 

 cient number of seedlings to show the ratio. 



While these results do not show us any way in which to eliminate those 

 varieties with reflexed stamens they do show that we may decrease the 

 proportion by the use of varieties for parents having upright stamens. 



