OENOTHERA GRANDIFLORA AIT. 



203 



First generation Second generation Third generation Fourth generation 



1912 



1913 



1914 1915 1916 



/ ochracea ochracea ochracea 



lorea- 



. Alabama— grandiflora - 



grandiflora 



- lorea 



( lorea 

 -^ grandiflora 

 f ochracea 



( gigas 



j gigas-lorea 



f gigas-ochracea 



The numbers of specimens and the percentages of the splitting 

 in this pedigree are shown in table I. 



Table I 



The control lines were derived from different specimens, grown 

 in 1913 from the seed of Castleberry. They yielded the same two 

 main mutants as given in table I. In the spring of 1914, however, 

 before I discovered the presence of mutants, I had observed that a 

 large number of the seedlings were very weak, dying off during the 

 first few weeks after being planted out in the boxes. Most, if not 

 all, of these must have been ochracea, and the percentage of 20 for 

 this mutant, found in August during the period of flowering, must 

 have been far too small. For this reason it is put in parentheses, 

 and the next year I tiied to get a more reliable counting. 



The seedlings of five self-fertilized plants of 1914 were planted 

 out in boxes as carefully as possible, and before any essential loss 

 was noted. They were kept in the greenhouse and counted out at 

 the end of April. At that time some few were dead and decayed; 

 others were dead but could still be judged. Since I had observed 



