368 



HYBRID OF DIGITALIS. 



Here there is not the least tendency for any dimorphism in 

 the purple colouration of the two backcrosses (3rd and 5th 

 columns), and the variability is remarkably small. 



Such a simple character as the greater or lesser intensity of 

 the same colour is likely to depend on some one factor, and, if 

 (he purity of the two homologous factors is to be supposed to 

 remain unaffected by their contact in the hybrid, the backcrosses 

 should exhibit a strongly marked dimorphic condition, but the- 

 experimental results show that the offspring are strikingly uniform. 



If, on the other hand, colour intensity depends on several 

 independent factors, the uniformity of the backcrosses militates 

 against the supposition that the factors remain pure in the hybrid. 



Leaf: Pilosity. — The number of hairs that could be counted 

 projecting over the edge of the leaf along a definite length (2 mm.) 

 was taken as a general measure of the pilosity. 



The leaves of lutea are markedly smooth, and bear but few 

 hairs, while gloxiniceflora leaves are usually richly pilose. The 

 mean of the pilosity (20) of the hybrids was slightly closer to the 

 lutea mean (6) than to the gloxiniceflora mean (37). The mean 

 of the gloxiniceflora backcross (19) is much nearer to the hybrid 

 mean (20) than to the glox. mean (37), and the mean of the lutea 

 backcross (12) is slightly nearer to the lutea mean (6) than to the- 

 hybrid mean (20). 



Pilosity of Leaf. 



It will be seen that in the backcross (3rd column) of the hybrid 

 with gloxiniceflora none of the offspring approach the mean (grade 

 35-39) of gloxiniceflora. In the backcross (5th column) with lutea 

 the condition cannot be so clearly observed on account of the fact 



