NO. 561] GENETIC AL STUDIES ON JEN OTHER A 563 



thera hybrids following the F x for which no adequate ex- 

 planation is known. 



In one F 2 generation I have noted a distinct retrogres- 

 sion in the size of the flowers. This was the F 2 from the 

 plant 11.35La briefly described in this paper. It con- 

 tained no plants with flowers larger than those of grandi- 

 flora and a large proportion of the culture bore flowers 

 as small as or smaller than the flowers of the biennis pa- 

 rent. The biennis parent in this cross was a rather large- 

 flowered type (biennis D, petals about 2 cm. long) which 

 made the retrogression appear the more marked. 



A striking feature of the F 2 generations here consid- 

 ered has been the diverse progeny from F x sister plants 

 of the same culture. Thus the Fj hybrids 10.30La and 

 10.30L& were sisters of the cross grandiflora B X biennis 

 A and the hybrids 11.35m, 11.35a and 11.35La were 

 sister plants of the cross grandiflora B X biennis D. 

 Each plant gave its own peculiar set of types in the F 2 

 with peculiarities so pronounced that the blood rela- 

 tionship was much obscured. This is difficult to under- 

 stand except on the theory that the parent stock was 

 heterozygous ; yet there has appeared no evidence of this 

 in the cultures of the pure species. It is, however, clear 

 that I have been working with complex material and it 

 is not certain that the species of (Enothera employed in 

 my crosses have been homozygous to the degree de- 

 manded for experimentation on the behavior of unit fac- 

 tors. For this reason I have endeavored to discuss the 

 problems with full caution and I hold my point of view 

 tentatively. 



5. The Habit of "Mutation" in (Enothera Lamarck- 



Perhaps the most important observations on these hy- 

 brids of grandiflora and biennis in the second and third 

 generations have been those showing a close parallelism 



