328 Journal of Agricultural Research voi. 1V.N04 



SUMMARY 



In the last 25 years, during which time nearly 10,000 seedUngs have 

 been grown, various changes in the methods used at the Geneva Experi- 

 ment Station have been made as the knowledge of breeding laws has 

 been extended. 



Results have compelled the belief that improved varieties of grapes 

 will not be produced to any extent until the fundamental laws of heredity 

 are understood. The present aim is to discover these laws. 



The work is now progressing mainly along two lines: (i) The deter- 

 mination of the breeding possibilities of varieties of grapes and (2) the 

 study and interpretation of breeding phenomena. 



Nearly 200 varieties of grapes have been used in the breeding work. 



Much of the value of the early work was lost by growing too few seed- 

 lings of each cross. 



Recently Vitis vinifera has been used to a considerable extent in 

 hybridization. 



The usual method of emasculation has been ineffective in a few cases 

 and may be open to criticism. 



One of the surprises in the study of grape varieties was the failure of 

 many commercial sorts to transmit desirable qualities. 



In order to study grape varieties, nearly 3,000 selfed, or pure, seedlings 

 have been grown. These are uniformly lacking in vigor. 



The inheritance of those grape characters which have sufficient data 

 available has been discussed in this paper. 



Of the two types of stamens, refiexed and upright, the first is correlated 

 with complete, or neariy complete, self -sterility, the second with self- 

 fertility. 



Self-sterility is probably caused by impotent pollen. It exists in vary- 

 ing degrees and depends to some extent upon the condition of the vine 

 and environmental factors. 



Self-sterile varieties of grapes being undesirable from a horticultural 

 standpoint, can we ehminate those with reflexed stamens? The follow- 

 ing crosses have given upright and reflexed stamens in the indicated 



ratios : 



U XU = 4.3U:iR 



R X R = i.2 U:iR 



R X U = i U:iR 



U X R = ? 



Breeding from the upright varieties only will decrease but not elimi- 

 nate the seedlings with reflexed stamens. 



The following seem to be the results secured in the study of sex 

 inheritance : 



Hermaphrodite female X hermaphrodite male = all hermaphrodites. 



Hermaphrodite female X pure male = 3^ hermaphrodites + }4 males. 



