404 



G. F. SPRAGUE 



phenotypic desirability. The wide range in topcross yields obtained is evi- 

 dence of the poor relation between phenotype and performance in hybrid 

 combinations. 



The frequency distribution of topcross performance was subjected to two 

 types of samplings. In one sample the Si lines representing the best 10 per 

 cent of the population were grown, and individual plants again self pollinated 

 and outcrossed to the tester parent, Iowa 13. The distribution of the So and 

 Si topcrosses are illustrated in Figure 26.1. (The So topcross yields have been 

 adjusted to the So topcross level on the basis of the performance of the tester 



50 



40 



o 



o 



lit 



30 



20 



10 



62.5 67.5 72.5 775 825 87.5 92.5 



YIELD IN BUSHELS PER ACRE 



97.5 



102.5 



107.5 



Fig. 26.1 — A comparison of the frequency distributions of 167 topcrosses of So plants {solid 

 line) with a series of topcrosses of Si plants {doited line), representing the highest j'ieiding 



10% of the original So population. 



parent, Iowa 13.) The distribution of the Si topcross yields clearly indicate 

 that the So plants exhibiting high combining ability transmitted this char- 

 acteristic to their Si progeny. Segregation within progenies was quite ap- 

 parent, indicating that opportunities for additional selection existed. 



A group of twelve lines was chosen which provided a seriated sampling of 

 the frequency distribution of So topcross yields. These were grown in 100 

 plant progenies, and an attempt was made to self pollinate 25 of the better 

 plants in each progeny and to outcross these to the tester parent. Because of 

 differences in time of pollen shedding only 6 of the 12 lines chosen were 

 finally used (Table 26.1). 



Significant differences in yielding ability were obtained within each of the 

 six Si families. The range in yield was of about the same magnitude in each 

 family, suggesting that the So plants having the highest test cross perform- 



