418 



The Journal of Heredity 



have been found by test to set in bags 

 over 33% of the bloom buds as berries, 

 while Scuppernong, the most extensively 

 planted commercial variety, sets natur- 

 ally less than W'^ of its buds as 

 berries. 



Third, by breeding for increased size 

 of bloom cluster it has been possible not 

 only to secure larger fruit clusters 

 (a breeding object), but also increased 

 productiveness. Large bloom clusters 

 result in more perfect cross pollination 

 and the size of bloom cluster is not 

 necessarily in inverse ratio to profuse- 

 ness of bloom cluster. The number of 

 bloom clusters produced by a vine 

 is ordinarily determined by its general 

 vegetative condition, and little in- 

 fluenced by the size of the individual 

 bloom cluster. Varieties with large 

 bloom clusters secure better cross pol- 

 lination and therefore are more fruitful. 

 As the ]3ollen-bearing insect is attracted 

 to the flower by its odor,^ a large bloom 

 clu.^ter has greater attracting power 

 than a small one, and when the cluster 

 is once found, all open blossoms are 

 pollinated, whereas if this number of 

 open blossoms is found on. two clusters 

 the insect might fail to visit one of 

 them. Again, the large bloom cluster, 

 in addition to having more buds open 

 at a time, blooms over a period of two 

 to three days and is therefore much 

 more likely to have sufficient buds to 

 produce a normal cluster open under 

 favorable weather conditions. One il- 

 lustration of the way these facts have 

 been used in the De])artmcnt's work- 

 seems warranted. 



Among the fruiting varieties of Mus- 

 cadine grapes Scuppernong is the oldest 

 and the leading commercial sort. Its 

 commercial value lies in its good light- 

 colored fruit, and Scupi)ernong grapes 

 have sold reiieatedh' to winemakers at 

 good i)nces when dark-fniited \'arieties 

 were not in demand. Still it can be 

 emi^hatically stated that Scui^pernong 

 is one of the most unproductive varie- 

 ties grown. An eight-year record ( 1904 - 

 1011) in a well-kept eastern North 

 Carolina vinewird consisting of a])i)roxi- 



mately 5 acres each of James and 

 Scuppernong revealed that James 

 N'ielded each year approximately twice 

 as much fruit per acre as Scuppernong. 

 Similar results have been repeatedly 

 noted since. A prime reason for the 

 relative lack of productiveness in Scup- 

 pernong has been shown by our pollina- 

 tion ex])eriments to be its ver}^ small 

 bloom cluster, which reduces the chances 

 for cross-]Jollination by insects. The 

 following data collected in connection 

 with ]:)ollination studies will emphasize 

 the bearing of size of bloom cluster 

 on productiveness of Scuppernong and 

 other commercial varieties. These data 

 represent the averages of a very large 

 number of individual obser\'ations dur- 

 ing four seasons. 



The correlation shown in this table 

 between size of flower cluster and 

 natural productiveness is most striking. 

 Working on the hypotheses suggested 

 by such data as these, the Department 

 has produced by intercrossing and 

 selection many large-clustered seedlings 

 of the Scup]oernong type. A nvmiber of 

 these ha\'ing bloom clusters of 35 to 50 

 buds, with ai)])arently correlated pro- 

 ductiveness, are now being further 

 tested with a view to determining 

 which combine the greatest ])roductive- 

 ness with high fruit quality. Thus we 

 ha\'e succeeded in improving the ]3ro- 

 ducti\'eness of this most important 

 commercial type while maintaining its 

 distinctive characters. 



• Experiments show that it is the odor of pollen in rudimentary stamens of the female blossom 

 that attracts the insects. 



