222 



Gates. 



the current Mendelian theories. If, after having obtained a perfect 

 3 : I ratio with the unit-character in question, one obtains other 

 ratios in crosses in later generations, e. g., 5 : i as I have done, it can 

 not be argued that this shows that the original supposed simple unit- 

 character was really compound, there being two or more independent 

 "genes" or "factors" for the same character. 



I propose to show that changed ratios of a single unit-character can 

 be obtained under conditions where such an interpretation is excluded. 



The results of Nilsson-Ehle (1909, iqii) in obtaining two ratios- 

 approximately 3 : i and 15 : i — for red and white glumes in diffe- 

 rent families of wheat derived from crosses between two independent 

 races with red glumes, seems to justify his conclusion that the red colour 

 is represented by two independent R factors. Others have, however, 

 applied this hypothesis of a multiplicity of independent represen- 

 tatives for one character whenever their ratios depart widely from 

 3 : I. The behaviour of ruhricalyx here described is incompatible with 

 this procedure for the reasons above stated and more fully explained 

 below, and should serve as a warning against the loose application 

 of the hypothesis. In the present instance some other explanation of the 

 phenomenon must be sought. 



Incidentally, it may be mentioned that the origin of ruhricalyx 

 furnishes one more proof, if such were necessary, that a sudden germinal 

 change or mutation may occur wholly independently of the shuffling 

 of hybrid characters. Evidently, also, only one of the parental germ 

 cells, and not both, underwent the original change. 



The full pedigree of the ruhricalyx used in the cress grandiflora x 

 ruhricalyx will be found in pedigree i. A ruhricalyx plant, No. IV. 6, 

 descended in the second generation from the original mutant, and which 

 was heterozygous for the red (i) was pollinated from nanella, the 

 dwarf mutant, in 1909. The Fj from this cross was grown in 1910, and 

 yielded, in accordance with expectation, seven Lamarckiana rosettes 

 and thirty-five plants which were either ruhricalyx or rubrinervis. 

 Of the latter, twelve which came into bloom were all typical ruhricalyx. 

 One of these twelve plants (which proved to be heterozygous) from 



(1) This is shown by the fact that when self-pollinated it yielded in 19 10 seven 

 rosettes, three of which were ruhricalyx and four rubrinervis. Three collateral families 

 given in pedigree i show a total of about 31 rubricalyi to 41 rubrinervis rosettes. 

 These were all grown in 191 1 and unfortunately they remained rosettes, so there is- 

 some uncertainty regarding the exact ratios, but the tendency in such circumstances- 

 is for the rubrinervis determinations to be too large owing to a temporary lack of 

 pigment development in some plants which later produce tj-pical red buds. 



