480 



MUTATION AND PLANT BREEDING 



in crosses C and D, s G was larger in the high selections in all treat- 

 ments. 



Table 8. — Comparison of the Means and Genotypic Standard Deviations in the 17 



Out of 23 High and Low F 2 Families which Maintained Their Relative 



High-low Performance when Tested as F 3 Progenies. 



Treatment 



Discussion 



The mutation breeding program with peanuts has been con- 

 ducted under the hypothesis that coincident with visible muta- 

 tional change after irradiation numerous nonobservable polygenet- 

 ic changes occur at loci scattered over the entire genome. This 

 hypothesis grew out of the consideration of the within-X 2 - 

 family variability in the expressivity of ordinary morphological 

 mutants. (Robbelen (44) has reported a similar observation for a 

 chlorophyll character in Arabidopsis.) Among the several hypothe- 

 ses advanced to explain the graded expressivity in morphological 

 mutants of peanuts, the one of genetic variation in the background 

 genotype has proved to be the most plausible. The first test of 

 this hypothesis (19) came with the discovery of a large genotypic 

 variance induced in control-type X 2 sibs of morphological mutants 

 of a pure line. 



The next test of the hypothesis was obtained when Loesch 

 (unpublished), in a study of the breeding value of simply inherited 

 deleterious mutants of the same pure line, discovered a remark- 

 able variation in expressivity between and within F 2 progenies 

 of mutant plants derived from intercrosses of different morph- 



