132 



THE AMERICAN NATURALIST [Vol. XLIV 



females. In the same way the simplex determiner induces 

 color blindness in males but nothing less than the duplex 

 determiner suffices to stimulate color blindness in females. 

 One may say that, under the influence of testicular internal 

 secretions the simplex determiner suffices for develop- 

 ment, but not otherwise. No doubt the same is true for 

 many other cases of "sex-limited heredity." 



Imperfection of dominance goes even further. It is 

 evinced not only in the heterozygote but even in pure-bred 

 stock. This is the same as saying that, in breeding pure 

 stock, a character may fail to develop even when a double 

 determiner is present. Thus, in pure bred strains of 

 Houdans or Dorkings having 5 toes on each foot, from 3 

 to 4 per cent, of the offspring fail to develop the extra toe 

 on either foot. In certain races, at least, such as my 

 Tosa fowl, two parents, both with complete inhibition of 

 foot feathering, may have offspring in which this inhibitor 

 fails to activate, so that their feet are slightly feathered. 

 These cases of failure of a duplex determiner to work 

 itself out in ontogeny are of the nature of sports— at least 

 many sports are of the nature of defects arising in a pure- 

 bred strain. In poultry one frequently gets such defects 

 as failure of the neural tube to close (spina bifida), failure 

 of the lower jaw to develop, absence of appendages and 

 so on. In man, imbecility seems sometimes to occur as a 

 sport; and it is doubtless a typical defect, transmitted as 

 a recessive quality. It is probable that the same is true 

 of hairlessness in mammals and man (Waldeyer, 1884, 

 page 114, and Davenport, Science, November 20, 1908, 

 p. 729). Since characteristics may occasionally fail of 

 development in homozvgotes their regular imperfection 

 in heterozygotes is easier to understand. 



Important consequences flow from the complete failure 

 of a heterozygous characteristic. The first is, as pointed 

 out by Shull, the difficulty of determining in the first 

 hybrid generation what is recessive— since impotent 

 dominance and recessiveness yield the same result. In- 

 deed, one might be tempted to conclude with Shull that 



