742 



THE AMERICAN NATURALIST 



[Vol. XLV 



The behavior of the crest when a crested fowl is mated 

 with a non-crested variety is typical of the behavior of 

 somatic blends tending to obscure gametic segregation. 

 Davenport 15 has shown that there are two genetically 

 independent factors united in the crest : One, erectness of 

 feather growth over a certain area is dominant over the 

 normal condition; the other a continued growth of 

 feathers of this area is recessive to the normal growth. 

 Hence, in the I\ generation there is an apparent " blend. " 

 the feathers being short but erect. From Fj, however, 

 in subsequent generations, Davenport has extracted a 

 beautiful complete crest. There are many other striking 

 somatic blends — among them the case of the Andalusian 

 fowl, of the human mulatto, and of the human herma- 

 phrodite. Blends are essentially the somatic aspects of 

 the fortuitous combinations of the patent and latent 

 phases of two or more genetically independent units. In 

 this sense Galton's law may justly stand for the general 

 measure of ancestral influence — a measure of the opera- 

 tion of the laws of chance. The existence of somatic 

 blends can not be denied, for they are among the most defi- 

 nite things commonly observed in inheritance. The more 

 cursory the examination and the more general the view 

 of such cases, the more seeming the blend; however, a 

 more minute inspection often reveals the segregation of 

 the parental factors, all of which points towards the 

 minuteness of the unit character and the purity of the 

 gamete. Were blending, in the commonly understood 

 sense a fact, all individuals of a race or a strain would in 

 a few generations become identical with each other. It is 

 the creation of new units by intra-zygotic reactions and 

 intra-gametic intrusions, together with the segregation 

 and recombination of the unaltered ultimate units of 

 inheritance that have given selection such an opportunity 

 for developing so many strains and species. 



""Inheritance in Poultry," p. 69. 



(To be concluded) 



