98 Negative Correlation in Oenothera Hybrids 



rubricalyx-determmers, but that it was also heterozygous for a third 

 gene which inhibits the action of one of the genes for the rubricalyx- 

 type of pigmentation, but not the other ; and (b) that my Lamarckiana- 

 plant, No. 118(10), contained a factor whose union with one of the 

 rubricalyx-determineYS produces, in the absence of the other, a non-viable 

 zygote. While such assumptions are perfectly proper as working 

 hypotheses, they have no other value. Assumptions of a still more 

 radical kind would have to be made to account for the peculiar 

 alternative relation between the pigmentation of leaves and stems on 

 the one hand and that of the buds on the other hand. 



In the enormous mass of genetic data already recorded for the 

 Oenotheras, there is but here and there a situation which bears more 

 than a remote resemblance to a Mendelian behavior, and in these 

 cases the observed phenomena usually present only a more or less 

 misshapen caricature of the beautiful regularity of procedure which has 

 such far-reaching applicability among many other groups of organisms. 

 It appears to me undesirable therefore t> speak of the rubricalyx- 

 character as a Mendelian unit-character because it happened to 

 constitute 75 per cent of one family of 44 plants. I believe that the 

 only other character in Oenotheras which has been accepted as Mendelian 

 in inheritance, namely, the brevistylis-chsira,cter, may well be put to the 

 test of a fuller genetic analysis. 



In view of these facts, one can only view with astonishment the 

 performance of Heribert-Nilsson (1912) in maintaining that the remark- 

 able series of genetic puzzles presented by the Oenotheras can find an 

 explanation through the recombinations of plural Mendelian determiners. 

 His entire thesis falls to the ground the instant we begin to figure out 

 some of the very simplest and most obvious consequences of such an 

 explanation. His abandon in the application of this hypothesis was 

 made possible only by his belief that students of the Oenotheras have 

 generally failed to use strictly individual analysis in their investigations. 

 Having myself never mixed the seeds from two different mothers or 

 from two different crosses, I am unwilling to believe that Heribert- 

 Nilsson has not greatly overestimated this source of difficulty in 

 interpreting the genetic phenomena in Oenothera. although it is a 

 valuable service to have pointed out so strongly as he has done the 

 importance of the strictest possible adherence to the "isolation principle." 

 The importance of this emphasis may be seen, when so careful a worker 

 as Dr. Gates grows a culture from the mixed seed of four different self- 

 fertilized parents, belonging to a group in which there was obvious 



