154 Mr Bateson, On Mendelian Heredity, etc. 



not been clearly made out. Complications arise from the fact 

 that the segregation of R from P may be imperfect, for some 

 " walnuts " occur in F 3 from walnut x single. At present it seems 

 likely that R, P, RP, and S are produced in equal numbers, but 

 various experiments are needed to settle the point. A note of 

 these qualitative results seemed of value as indicating possible 

 lines of work, which if pursued on several subjects may consider- 

 ably extend our knowledge of gametic differentiation. 



Possibly the F and D forms raised by de Vries from Antir- 

 rhinum (Mutationsth. II. 197 and 353) are somewhat analogous 

 examples of imperfect segregation. An indication might be got 

 by breeding from the reds formed by extracted F x extracted D. 

 If whites then occur we should have proof that F was one 

 hypallelomorph + white, and D another hypallelomorph + white, 

 and the problem would be greatly simplified. 



