THE GENETICS OF HABROBRACON JUGLANDIS ASHMEAD 



relative number of AB and of aB phenotypes at 

 the expense of Ab and ab respectively, or the 

 reverse may occur. It may also happen that AB 

 will be increased at the expense of aB, while 

 AB is not increased at the expense of ab. This 

 may be the result of the influence of factor B 

 versus b or to a modifier of trait A versus a 

 linked with factor B or b. A factor preventing 

 overlapping acts as a "differentiator" (Bridges, 



1919). 



AB may exceed aB as the result of differen- 

 tial viability in which case we might expect a 

 comparable excess of Ab over ab. However, the 

 ratio ab/AB cannot be predicted from (aB/AB) X 

 (Ab/AB) since in some cases the double mutant 

 type fails to be reduced proportionally to the 

 single while in others the double may be highly 

 lethal although one or both of the singles may 

 show viability equal to that of wild-type under 

 the conditions of culturing. 



Linkage of one of the genes (A or a for ex- 

 ample) with a lethal or semilethal may cause an 

 excess or a deficiency of a mutant type of nor- 

 mal viability. Thus aB may surpass AB or the 

 reverse. In this case v;e may expect a corre- 

 sponding shift between ab and Ab, but such may 

 not be obvious because of viability differences 

 or somatic overlaps between the various combi- 

 nations of A and a with B and b, or because of 

 linkage of a second lethal or semilethal with B 

 or b. V/hile lethals are invisible in the mate- 

 rial thus far presented on linkage, semilethals 

 may be either visible or invisible. Complete 

 lethals, dying as pupae, which may be identi- 

 fied as to a second trait (eye color for exam- 

 ple), have been shown in Habrobracon, Degree 



114 



