116 



White. 



elusions, but owing to complications caused by the segregation of 

 numerous other characters, the classification was extremely difficult. 



Table H. (396X402 Fo generation.) (Calycanthemy X normal.) 



Some individuals could be assigned to their particular category 

 with accuracy and speed, W'hile others were so hopelessly modified that 



Fig. 26. Extremely abnormal flowers from Fg AABB segregates 

 of 396 X 402. 



even after repeated trials and extensive study, they could only be 

 placed by guess. This state of affairs was not a particular class 

 characteristic, but true of all. In order to be as accurate as possible, 

 three separate classifications were made; one early one involving only 

 a part of the population, and two later ones in w^hich nearly all of 

 the plants were considered. Those not included, were either destroyed 

 accidentally or w^ere not yet in bloom at the time of the last inspection, 

 and as they were few in number, the absence of data from them is of 

 no significance. Table 23 gives the results of the three classifications. 

 I consider that of the second the most accurate, as most of the plants 

 were in full bloom w^hen this was made. Count 3 was taken later in 

 the season, when many of the segregates possessed few^ flowers, but 



