84 Negative Correlation in Oenothera Hybrids 



The rosettes had the general morphological features of my 0. rubri- 

 nervis forms, but the leaves were darker green and of more shining 

 surface. The buds were more slender and more nearly terete, but the 

 capsules were long and decorated, when young, with the 2-4 deep-red 

 longitudinal bands usually seen on immature rubrinervis-c&psules. The 

 buds were brilliantly red-pigmented throughout, except on the ovary 

 and the distal portion of the free tips, which were green. The stems 

 were also brilliantly red-pigmented, as described by Gates. All three of 

 the plants were of the same type, and it seems probable that they were 

 true 0. rubricalyx Gates, despite the fact that they came from unguarded 

 seeds under circumstances doubtless favorable to crossing with many 

 other Oenotheras 1 . As my results in breeding these plants differ 

 strikingly from those reported by Gates, the possibility of the hybrid 

 nature of my original plants of this form must be kept in mind ; but it 

 must also be kept in mind that even if these three plants were hybrids, 

 they must each have been produced by the union of only two germ-cells, 

 and as I have used in the self-fertilization and in all the crosses described 

 below only one of the three original plants, any complication from their 

 possibly hybrid nature is reduced to the lowest possible minimum. Such 

 a possible hybrid origin will not detract in the least from the interesting 

 correlations with which this paper deals. It should be added that in 

 making reciprocal crosses generally, I have used in both crosses the same 

 pair of individuals, so that no unsuspected complications can have been 

 introduced by genotypic differences in individuals wrongly assumed to 

 be identical because they were phenotypically alike. This is a detail 

 of technique of the utmost importance; it will be found that the 

 genetic problems in Oenothera are sufficiently complex when the 

 elements which enter into the related crosses are reduced to the lowest 

 possible terms. 



Oenothera rubricalyx appeared in 1907 in a mixed progeny consisting 

 of 112 offspring of four self-fertilized rubrinervis-pl&nts grown at Woods 

 Hole the preceding year, from unguarded 2 seeds received from de Vries 



1 Several features of my plants suggested a relationship to 0. grandiflora, particularly 

 the rather lax rosettes, strong red spotting of the young rosette-leaves, and the develop- 

 ment of buds more slender and rounder than in my 0. rubrinervis cultures, but as none of 

 the characters presented by the offspring showed accentuated resemblance to 0. grandijlora, 

 I doubt the reality of this suggested relationship. 



2 Gates does not state that the seeds received from de Vries were from open pollinations, 

 but the experience of Oenothera-students with rubrinervis -cultures, makes probable no 

 other interpretation of the occurrence of 2 0. LamarcJciana and 1 0. oblonga among the 

 45 plants which included the rubrinervis-p&rent of the original 0. rubricalyx. 



