474 THE AMERICAN NATURALIST [Vol. XLIV 



individual totally lacks all power, the first hybrids will 

 also fail to produce the character. Their chromosomes 

 combined and divided at random in the germ cell will 

 produce a much larger number of gametes that fall be- 

 low the standard than of those that rise to a point suffi- 

 cient to give the character when combined — in conse- 

 quence the recessives will be greatly in excess. 



These considerations may seem to throw light on the 

 question of potencies of different individuals — a question 

 that is coming more into the foreground. We can see 

 from the point of view here suggested how individuals 

 alike externally may differ very greatly in their power 

 to transmit their peculiarities to hybrid offspring. 



This conclusion is especially applicable to cases where 

 the full development of a character can only appear 

 when two groups of chromosomes (to take the simplest 

 case) are necessary to produce a character ; or, to take the 

 more extended view, when excessive amounts of chro- 

 matin must be present. It is now well established that 

 certain races lacking a character nevertheless dominate 

 in the first generation when crossed with a race possess- 

 ing a character. In such cases the failure of dominance 

 may be due to insufficient chromatin of the positive kind 

 rather" than due to an inhibiting factor as sometimes 

 assumed. 



As regards blending, it is evident that this rela- 

 tion must result from the combined action of the two 

 parental contributions to the hybrid ; the blending is the 

 sum of both effects. Such cases differ from Mendelian 

 cases in the first generation only in that one influence 

 does not exclude the other. In the second generation 

 separation into two classes of individuals does not occur, 

 but a great variety of forms appear: Nevertheless the 

 individuals may show a tendency to group around the 

 two parental and the hybrid classes, as Castle has shown 

 for long and short hair. In this sense blended inheri- 

 tance shows gradations into alternate inheritance. The 

 chief difference between the two, I repeat, is found in the 

 compatibility of the contrasted characters. So far as 



