xni] Shirley Poppies 241 



case segregation does not occur. Even a sensible con- 

 tinuity between varieties is no proof that there is not segre- 

 gation which with careful work could be demonstrated. To 

 take for instance the case of piebald mice, it would be 

 possible with industry to collect from the results of miscel- 

 laneous breeding specimens ranging from a self-colour to a 

 black-eyed white. From such series any number of false 

 conclusions could be drawn respecting the continuous 

 variation of mice. By appealing to the " ruby " eyes of 

 certain chocolates an uncritical observer might even argue 

 that a series of intermediates connected the black and the 

 pink eyes. 



Analytical breeding immediately shows the fallacy of 

 such deductions. For it would be found that the so-called 

 intermediates consisted of numbers of genetically distinct 

 types with distinct genetic properties depending on the 

 factors which constituted them. Some would carry the 

 colour-factors for the Dutch pattern, others those for more 

 complete or less complete pigmentation, while others would 

 owe their partial whiteness to the presence of the dominant 

 factor which can suppress pigmentation in several stages of 

 completeness. 



In general it is to be concluded that when the inter- 

 breeding of recognizably distinct types produces only those 

 types again without intermediates or with only few indi- 

 viduals that can be so regarded, the fact points strongly to 

 the existence of gametic segregation ; but on the other hand 

 the appearance of even a complete series of intergrading 

 forms is not to be accepted as proof that gametic segregation 

 does not occur. 



Misconception of the nature and significance of inter- 

 mediates has deprived the work of the biometrical school 

 of scientific value as a contribution to the study of heredity. 

 This is well seen in the case of the colours of Shirley 

 Poppies, one of the subjects with reference to which copious 

 statistics have been amassed and published*. The colours 

 are red, pink, and white, of many shades, and mixtures of 

 these tints variously disposed and graded. The problem 

 was to discover the laws of inheritance of the colours. The 



* Biometrika, u. i, 1902, and iv. 4, 1906. 

 B. H. 16 



