xr COUPLING AND REPULSION 127 



be of the nature / BE \ 1 2J Be: 12^ bE : i de, it is 

 clear that we can look for the appearance of the 

 red-hooded class only as a great rarity. To be exact 

 we should expect a plant of this class to appear 

 once among 65,536 (=256^^) individuals in the Fg 

 generation. Unless impracticably large numbers 

 are bred we should expect the F^ generation to 

 consist of the three forms — purple erect, purple 

 hooded, and red erect in the proportion 2 : i : i. 

 And this is what has actually happened. Among 

 5789 plants bred in this way 



Purple hooded x Red erect 



I 

 Purple erect 



Purple erect Purple hooded Red erect 



2969 1379 1441 



no red hood has appeared, while the three classes 

 found are roughly in the proportion 2:1:1. More- 

 over the hypothesis demands that almost all of the 

 Fg purple erects should behave as the F^ plants, 

 that almost all the purple hooded should be homo- 

 zygous for B, and that almost all the red erects 

 should be homozygous for E. In other words, 

 the great majority of the purple hooded and of 

 the red erect should breed true, while the purple 

 erects should almost invariably throw all three forms. 

 This natural consequence of the high intensity of 

 the repulsion suggests a comparison with such cases 

 as that of the Andalusian fowl (p. 64). Here also 

 we have three forms of which two breed true while the 

 other throws birds like itself, and birds like each of 



