370 Inbreeding, Selection, and Heterosis 



40, 50, 60, and 70 eg respectively. The seed-weight average of 

 the offspring from these four plants was, respectively, 57.2, 54.9, 

 56.6, and 55.5 eg, a series of values which show clearly that selec- 

 tion of the larger or smaller seeds of this inbred line has abso- 

 lutely no effect on the average weight of the progeny. The av- 

 erage weight varies little in all the progenies, and what little 

 variation there is in the average bears no relation to the weight 

 of the mother, for the heaviest average of these four progenies 

 came from the lightest seed. 



Within inbred lines, the variation that is present is not in- 

 herited; it must be the result of environment. Johannsen tested 

 this out further by continuous breeding for half a dozen years 

 from both the largest and smallest types within an inbred line. 

 Six years of selection in both directions failed to produce any 

 significant difference in seed weight. 



The largest seeds of the original lot produced progenies with 

 a much higher seed-weight average than progenies of the small- 

 est seeds but both the large and small seeds produced progenies 

 of the same average weight when they were from the same inbred 

 line. The reason is clear, for the original lot of seeds was a 

 heterogeneous mixture which had undoubtedly come from a 

 number of different parents with very different genotypes and 

 which showed variation that was both genie and environmental. 

 Within any given inbred line, however, all the plants had the 

 same genotypes, and therefore any inbred progeny showed only 

 environmental variation. The bean normally reproduces by self- 

 fertilization and most plants are homozygous. A homozygote 

 will produce a progeny that consists of only homozygous plants 

 (barring infrequent mutations) which are genetically all alike 

 and are like their parent. 



If two homozygous plants which differ by a number of genes 

 are crossed, the Fi plants will be heterozygous for many genes 

 and will show only environmental variation. Various Fi plants 

 upon self-fertilization will produce Fo families which will be 

 much more variable than the Fi, but will have about the same 

 mean as the Fi. If various F2 plants are self-fertilized, and if 

 subsequent generations are also produced by self-fertilization, 

 different strains will separate out which may differ from one an- 

 other in their means, will show much less variability than the 

 Fo, will after a few generations establish a fixed value for each 



