358 Origin and Behaviour of Oenothera rubricalyx 



descended from rahricnlyx individuals in the cultuu' having a ratio 

 of 33 : 11. They were all observed only as rosettes, and as they were 

 not observed with the same care as the family of 147 above-montionod 

 it is obviously unsafe to attempt to draw any conclusions from them. 

 It is certain, however, that the number of rubrinervis rosettes was over- 

 estimated, for the reasons given earlier in this paper. As recorded, the 

 seven cultures total 91 R : 78?-. In any case the numbers do not support 

 Shull's hypothesis of duplicate factors. 



Turning now to certain other statements in Shull's paper, we find 

 (p. 8.5) the statement "Gates (1914) wavers between the treatment 

 of the ruhricahjx type of pigmentation as a Mendelian and as a non- 

 Mendelian character. His most positive declarations on the subject 

 are that it is non-Mendelian ; but if he sincerely {sic) holds to this 

 conviction it is strange that he should continue to treat the genetic 

 behavior of this character as if it 'threw valuable sidelights on Men- 

 delian phenomena." It is difficult to realize the frame of mind in 

 which such a statement could be written, showing as it does entire 

 failure to grasp the points of view developed by me in the paper in 

 (juestion. One can only conclude, as other statements also indicate, 

 that it was not carefully read or digested. 



Among the "facts" presented in Shull's paper we find the statement 

 that the offspring of his " rubricalyx " (which we now know to be 

 rubricalyx x grandifloru) yielded (in F., or a later generation) 117 

 individuals, of which 107 were red-budded and 10 green-budded. 

 This ratio, 10'7 : 1, is rather close to the ratio 9'.5 : 1 obtained by me 

 (1914, p. 235) in an F., family of grandifiora x rubncalyx numbering 

 157 plants, and confirms my results in obtaining various aberrant ratios 

 from these ci'osses. It is astounding to find that although Shull is 

 unwilling to recognize 33 : 11 as a 3 : 1 ratio, yet he is capable of sug- 

 gesting that 107 : 10 represents a 15 : 1 ratio ! 



Without entering further into the extensive data in my paper 

 (1914), it may be remarked that two things appear to be clearly 

 proven; (1) that Oe. mut. rubricalyx originated through a single 

 unit-change producing a new character-difference which is dominant ; 

 (2) that when the new form is crossed with a distinct species — Oe. 

 grandiflora — having a different physiology, the red-budded feature 

 comes out in F„ in varj'ing ratios in different families, 3:1, 4:1, 5:1, 

 and about 10 : 1. Not only is this the case, but occasionally inter- 

 mediate individuals as regards bud-pigmentation occur, and it has since 

 been found that these when selfed breed true to their intermediate 



