Breeding experiments which show that hybridisation and mutation etc. 275 



is inexplicable on a Mendelian basis, and can only be accounted 

 for by different degrees of prepotency in different Fj individuals, 

 or in other words by varying adjustments between the two 

 characters. 



11. In rubricalyx x grandiflora the grandiflora race carried the 

 capacity for producing tails and dwarfs in the ratio of about 

 7:1. In accordance with expectation from this condition, 

 only two in nine Fg families produced dwarfs, and these in 

 the ratios respectively (tall : dwarf) of 5.67 : i and 7 : i. The 

 capacity for producing dwarfs in a percentage of about i in 8 

 is thus inherited from generation to generation. 



12. From these and other results it is necessary to conclude that 

 the Mendelian conception of fixed and unmodifiable unit- 

 characters which can be redistributed and shuffled regardless 

 o^ the organisms themselves which exhibit these characters, 

 is unsound, for the individual organism is the real unit. The 

 "presence-absence hypothesis implies a misconception of the 

 nature of the differences between alternative characters, and 

 of their interactions with each other. 



13. All the other numerous differences between grandiflora and 

 rubricalyx are non-Mendclian or blending characters. These 

 include the differences in foliage, buds, pubescence, and in 

 physiological development. In all these cases there is neither 

 dominance nor segregation, but every conceivable degree of 

 intermediacy is represented. In foliage, for example, rubri- 

 calyx has stem-leaves which are broad at base and with crinkled 

 surface, while in grandiflora they are pointed at base and with 

 smooth surface, though these are by no means the only diffe- 

 rences. But in the F2 every conceivable degree of these 

 characters is represented, with frequently wide variation, even 

 on the same plant. The futility of trying to explain such a 

 condition as the result of the fortuitous distribution of innu- 

 merable "unit-factors" is very evident. 



14. The physiological features, such as time of stem-formation 

 and time of blooming, are also inherited in a strictlv blended 

 and intermediate fashion. Thus the time of blooming of the 

 Fl is, in both the reciprocal crosses, between that of the two 

 parents, though in both cases nearer that of the male parent 

 in the cross. When the F^ is crossed back with the parents, 

 the hybrid is again, at least in most cases, intermediate between 



18* 



