Mr Bateson, On Mendelian Heredity, etc. 153 



On Mendelian Heredity of three characters allelomorphic to 

 each other. By W. Bateson, St John's College. 



[Bead 4 May 1903.] 



The object of this note was to call attention to various possi- 

 bilities attainable by a modification of the Mendelian method. In 

 the ordinary method the constitution of the gametes in the first 

 cross (i^) is tested by breeding such individuals inter se or with a 

 pure recessive. The ensuing generation (F 2 ) will consist of a 

 mixture of dominant and recessive individuals ; but if the propor- 

 tions depart from the expected 8:1 or 1 : 1, it is not possible to 

 tell whether such departure is due to change in relative numbers 

 of dominant and recessive gametes, to imperfect segregation of 

 characters, or to change in dominance. This question can in part 

 be answered by a method which consists in crossing F x produced 

 from a parent having one dominant character, with another 

 heterozygous individual having a different dominant character 

 (the same recessive being used in both cases). In the poultry 

 experiments described, rose comb (R) and pea comb (P) were used 

 as the two dominants, single comb (S) being the recessive. When 

 RS was crossed with PS, F 2 showed the four expected forms RS, 

 PS, SS, and RP, each well characterized. The RP combs have a 

 highly peculiar structure, resembling the " walnut " comb of some 

 Malay fowls, the comb almost always being feathered to some 

 extent, and generally crossed by a curious band of small feathers. 



In F 3 , RS gave rise to only R and S birds. No doubt the PS 

 would give only P and S, and the S birds, S only. But if R and 

 P had been completely segregated in F x the RP birds should be 

 incapable of producing any singles in F 3 . In the experiments 

 RP x RP did give some singles, and therefore, assuming that S is 

 not a hypallelomorph of R or P (which is being tested), a proof is 

 provided that the segregation of R (or perhaps P) from S may be 

 imperfect. That it may also be perfect for other R and P gametes 

 has of course been previously established in several cases by 

 breeding from dominants in F 2 , some proving homozygous domin- 

 ants, while others were heterozygotes. The ratios in which the 

 several sorts of gametes are produced by the " walnut " birds have 



