440 BROOKLYN BOTANIC GARDEN MEMOIRS 



notorious inharmonies between the inheritance ratios in the Oenotheras 

 and the expectation based on the usual MendeHan methods of segre- 

 gation and recombination, one may well suspend judgment regarding 

 this case as an instance of duplication, until it has been shown by 

 further analysis of one of these 15 : i ratios, that the rubricalyx 

 individuals will yield three kinds of families, characterized respec- 

 tively by the ratios i : o, 15 : i, and 3:1, and that these three kinds 

 of families are produced in approximately the ratio 7:4:4. Unless 

 this should be the result of the further study, the 15 : i ratio noted in 

 several of the F2 and F3 families must have been brought about by 

 some combination of circumstances, other than the typical Mendelian 

 distribution of two duplicate factors for the rubricalyx pigmentation. 



Gates discusses at some length two of the several methods by 

 which one may reasonably suppose duplication of factors to come 

 about. He seems to imply (Gates, 1915, p. 204) that my discussion 

 of this subject does not adequately cover the several possibilities. He 

 then proceeds to present two of the same possibilities as if they were 

 original propositions of his own. These several possibilities are (a) 

 the occurrence of independent mutations affecting in the same or 

 closely similar manner non-homologous chromosomes; (b) the mating 

 of non-homologous chromosomes; and (c) the transposition of parts 

 of chromosomes by what I have called a "sort of longitudinal crossing- 

 over" (Shull, 1914, p. 139). Only the first two of these propositions 

 are considered by Gates and he agrees with me that both of the proc- 

 esses (a) and (b) have probably actually resulted in the duplication 

 of factors. He thinks that repeated mutations were responsible for 

 the duplication of red pericarp color in Nilsson-Ehle's wheats, and 

 that mismating of chromosomes will explain the duplication which he 

 believes to have taken place in his Oenothera rubricalyx crosses. 



Upon unpublished evidence Bridges (1917, p. 454) refers to two 

 cases of duplication in Drosophila which seem to result from essentially 

 the longitudinal rearrangement of genotypic materials that I had in 

 mind when suggesting the possibility of "longitudinal crossing-over," 

 though the details of the process as understood by him are somewhat 

 different. He states that a section from the mid-region of one X 

 chromosome appears to have been removed from its accustomed place 

 or locus in that chromosome, and to have become attached to the 

 end of the other X chromosome, its mate. The full account of this 

 case will be awaited with interest. 



Accepting the validity of these several methods of duplication, 

 one may well ask in each specific case whether circumstances make 

 possible a judgment as to which method was probably responsible 

 for the duplication in question. I have assumed that the complexity 



