i66 MENDELISM chap. 



containing all four factors, i.e. the birds larger than 

 the Hamburgh parent ; together with others contain- 

 ing none of the four factors, i.e. the birds smaller 

 than the Sebright bantams. It may be added that 

 these very small birds were subsequently mated 

 together and gave only very small offspring like 

 themselves, as indeed they would be expected to 

 do on this hypothesis. 



The appearance in F^ of birds both smaller and 

 larger than the original small and large parents 

 suggests a point of interest. If we denote our four 

 size factors as Ay B^ (7, and D^ it is clear that 

 among the many intermediates some, e.g. AABBccdd 

 and aabbCCDDy should breed true, and at the 

 same time should give, when crossed together, F^ 

 birds larger than either of the parents. Among 

 the recombinations in Fg would be some homo- 

 zygous for all four factors. In other words, from 

 birds of not dissimilar size we should be able to 

 obtain a strain which would breed true to larger 

 size. Such a phenomenon would probably be 

 ascribed by most people to the beneficial effects 

 of a cross, inducing greater vigour in the offspring. 

 But whether this be true or not, we ought not to 

 lose sight of the fact that it may be due, either 

 wholly or in part, to a recombination of multiple 

 factors on the lines suggested by the Sebright- 

 Hamburgh experiments. 



Some such process may at times be responsible 

 for the srreat increase in flower size of which the 

 literature of modern horticulture affords so many 

 examples. Such increase is doubtless often the 

 result of a sudden sport or mutation. There is 



