G. H. SHULL Oi) 



impurity, on the ground that "this made no difference in the 

 experiments [he] then had in view " (Gates, 1913 b, p. 143). It is a 

 little difficult at the present time to imagine the nature of those genetic 

 experiments which would not be injuriously affected by such an origin 

 of the foundation stock. It can only be a matter for regret that Gates 

 (1914) has also used for his so-called "reciprocal" crosses between 

 0. rubricalyx and 0. grandiftora, a strain of the latter species from 

 Alabama, for the one cross, and one from Birkenhead, England, for the 

 "reciprocal" cross, especially in view of the fact that he has found 

 much evidence of complex hybridizations in the latter locality (Gates, 

 19 13 a). Finally, his "back-crosses" of the F l plants from this second 

 hybrid family have been made not upon plants of the parental strain, 

 but upon those of the Alabama strain. It is hardly to be hoped that 

 the elements of genetic behavior in Oenothera will be discovered by 

 these methods. 



The conclusion reached by Gates (1910) that the method of 

 inheritance of a character is determined by the nature of the character 

 itself, is also brought sharply into question by the peculiar situation 

 in my 0. rubricalyx-crosses, for I am clearly dealing with the same 

 character which Gates has repeatedly called a monohybrid Mendelian 

 dominant ; but in my cultures it is either not Mendelian at all, or, if 

 Mendelian, is affected in a complex way by several different determiners. 

 The remarkable diversity in the nature of the characters which have 

 been proved to be typically Mendelian in inheritance in various plants 

 and animals, should have made Gates's conclusion impossible. It is not 

 the externally visible, physical or chemical nature of a character which 

 determines the method of its inheritance, but the nature of the 

 inheriting-" mechanism " to which it is related, and the manner in 

 which its determiner or determiners are related to that mechanism. 

 Baur (1910) has discovered a case which clearly illustrates this point, 

 and I also have been able to confirm his results (Shull, 1914). We 

 have shown that " chloralbinism " in plants may result from several 

 different causes. In some cases the absence of chlorophyll is due to 

 the absence of a definite Mendelian gene, inherited equally well through 

 both the male and female gametes; in other cases it appears to be a 

 purely cytoplasmic defect, inherited in characteristically non-Mendelian 

 ways, sometimes only through the mother, sometimes through the 

 father and the mother, but with such irregularity that its inheritance 

 cannot be related to Mendelian phenomena. Whether Mendelian or 

 non-Mendelian, the character itself appears to be the same, namely, the 



