370 



hybrid of digitalis 

 Breadth 



Leaf-ratio 



Length 



x 1000. 



If the leaf-ratio depends on a single factor, or a group of con- 

 nected factors, and these factors remain unchanged in the hybrid, 

 then in both backcrosses (3rd and 5th columns) a dimorphic con- 

 dition of the offspring should be observable, but such is not the- 

 case. If the leaf-ratio depends on two or more independent 

 factors which remain unchanged in the hybrid then much varia- 

 bility in the backcrosses would be expected, and the two series 

 would have a heterogeneous aspect. An inspection of the table 

 shows that the variability is not very large, and there is no 

 evidence of heterogeneity. 



In all the other characters studied (namely, in the leaf; size 

 of lower epidermal cells, number of stomates, rate of desiccation 

 of plucked leaves, and depth of indentations : and in the flower ; 

 breadth of corolla, ratio of breadth to length, and degree of spot- 

 ting), there is likewise no indication of the offspring of the back- 

 crosses being divisible into two sets, one resembling the hybrid 

 parent, and the other, one of the parent species. 



On the view that all these characters depend on a number 

 of independent factors which remain pure in the hybrid, there 

 should be, at least, a clearly marked exceptional variability of the 

 backcrosses, but there is no suggestion of such in any of the 

 characters dealt with. 



9. Conclusions. 



It is evident that the potentialities of an organism must 

 depend on some qualities or powers of the fertilised egg. These 



