No. 639] L\IIl-:iUTA.\CE IX NICOTIAN A TABACUM 3oS 



The further question now arises as to whetlier there 

 are any phenotypic clitTerences between carmine and red. 

 There is a detectable difference between the fiower color 

 of macrophylla and that of purpurea, for the former has 

 distinctly more of a purplish tinge and is not quite as 

 intense in coloration as the latter. But these two varie- 

 ties differ genetically in a large number of other charac- 

 ters. It is not possible, therefore, to decide the question 

 by direct examination, because any distinctions which are 

 found to exist may depend upon differences in the 

 residual genotype rather than upon the specific factor 

 diffei-ences which we have studied. Obviously the most 

 satisfactory material for determining the differences 

 between the two colors would be two varieties which had 

 the same residual genotype, but the establishment of 

 such varieties would entail the expenditure of a consider- 

 able amount of time and labor. We can, however, obtain 

 some evidence on this problem by comparing the red F. 

 segregants of the Cuba-Macrophylla series with the 

 carmine ones from the Cuba-Purpukea series. In both 

 cases there was a certain amount of variation in intensity 

 of coloration in the Fo classes, but it was found that, if 

 they were mixed together, it was impossible to separate 

 them again into red and carmine. In cases involving 

 both classes in the same experiment, they would doubtless 

 have to be considered as making up a siiiulc plicnot x jtc. 



We have been interested in determiniim- cxpci'iinciilally 

 whether the morphological similiarities of existing 

 Tnharnrn v;iri,'ti(-s mi-lit safely be taken as an index of 

 ])livl..uviictic atliuitics. Thus Setchell,' commenting on 



