Breeding experiments which show that hybridisation and mutation etc. 271 



The contrast which has been drawn between the morphological 

 and physiological aspects of the pigmentation-character, R, probably 

 applies to many other Mendelian characters, and may furnish an 

 inkling as to why certain characters exhibit th.e phenomenon of domi- 

 nance while others show only intermediacy or blending. For doininatice 

 in the case of R is merely the morphological aspect of a fundamental 

 difference in cell metabolism which is quantitatively inherited. 



The Oenothera hybrids described in this paper serve particularly 

 to emphasize the contrast in behaviour between simple unit-characters, 

 such as R and t, and the numerous interspecific differences between 

 0. grandiflora and 0. riihricalyx which give a blended condition in 

 Fl and a continuous series in F». The latter, or non-unit-characters, 

 include the numerous foliage and bud differences, and the conspi- 

 cuous difference in pubescence, as well as the physiological differences 

 in reaction to environment and in rate of development. In all these 

 characters there is an absence of dominance, and the characters show 

 not the slightest tendency to segregate into sharp categories, but a 

 blended and variable condition occurs in F2, with a complete series of 

 intermediates between the parental forms. Furthermore, when the 

 Fx is crossed back with either parent, the offspring do not segregate 

 but form new though variable blends, in which the original differen- 

 tiating characters are still further fractionated and lost. This statement 

 applies to the foliage, the buds, the pubescence, and the physiological 

 differences. After every effort to harmonize these carefully analyzed 

 facts with Mendelian categories, it has become obvious that these 

 characters do not obey Mendelian laws, and that any distribution of 

 hypothetical units, however numerous, only falsifies instead of clari- 

 fying the facts. 



In the introduction to this paper I pointed out that the com- 

 bined cytological and breeding data in Oenothera effectually dispose 

 of the claim that the mutation phenomena are merely due to Mendelian 

 behaviour. The data in section K show that, altogether, 10 mutants 

 or mutant-like individuals appeared in six of the Fo cultures and 

 back-crosses of grandiflora and ruhricalyx hybrids, i. e., 10 plants in 

 a total of 2794, a rather low percentage. These mutants included 

 two which were semilata grandiflora, combining the foliage characters 

 of the de Vriesian mutant semilata with certain bud and leaf cha- 

 racters derived by inheritance from grandiflora. Two other mutants- 



