GREGORY: EFFICACY OF MUTATION BREEDING 469 



in 1955. At the end of the F 2 and F 2 X L . year, 12 of the 15 original 

 crosses were available for experimentation. Nine of these were 

 placed in a replicated experiment on the peanut testing station 

 located in the central North Carolina coastal plain. This experi- 

 ment was lost due to fall storms. Three of the crosses were placed on 

 another station on the western edge of the coastal plain. These were 

 harvested without mishap. They were Cross I (C12xA18), Cross II 

 (C12XYT24), and Cross III (C12XYT13). C12 and A18 were F 4 

 selections from two different hybrids in F n generation. YT24 and 

 YT13 were X 3 selections in X-, generation from the same pure line. 



The F 3 and F 3 X 3 experiment involving these three crosses Avas 

 designed in the following manner. Ten F 2 generation plants were har- 

 vested individually from each of five F, generation families of the fol- 

 lowing treatments: PI, P2, P1X 2 , and P2X 2 . Ten plants were harvest- 

 ed individually from two sets of five F] generation families in the F 2 

 and F 2 X 2 . This provided an equal number of progenies of 

 P, PX, F, and FX. The individual plants and the families from 

 which they were chosen were taken at random except for the specifica- 

 tion that enough seed be produced to conduct the experiment. 

 The individual plant progenies of these selections were planted in 

 the F 3 generation and arranged in the field according to the experi- 

 mental design presented in Table 1. Dry weights of fruits were 

 obtained and appropriate analyses conducted including analyses 

 of variance of the individual F, and FiXi families. 



The overall effect of the irradiation by generation is shown in 

 Table 2 where all values are presented in percentage of the mean 

 of the entire experiment. Differences among treatments in blocks 

 were highly significant in four of the six blocks and significant in 

 the fifth. Thus the reduction in mean performance occasioned by 

 X-ray treatment is a substantial factor in the breeding expectations 

 from this material. The cross means are almost identical. 



The genotypic standard deviation (s G ) of each F : family was 

 determined as 



V p - V e 



V" r 



where V,, is the progeny mean square, 

 V e the individual F : family error mean square, and r the number 

 of replications. The treatment means and the average s G among 



