OF MUTANT CHARACTERS. 



181 



The second point attacked was the amount of crossing-over in the 

 female between the loci purple and vestigial. The same two comple- 

 mentary crosses that had furnished the material for the male tests 

 just given were used as the source of the females to be tested. Fi 

 daughters from these two matings were back-crossed singly to purple- 

 vestigial males, with the results given below. 



BALANCED INVIABILITY-COMPLEMENTARY CROSSES. 



The reason why both "coupHng" and ''repulsion" experiments 

 were made is that by combining the two sets of data one can calculate 

 a linkage value more nearly free from the errors due to dispropor- 

 tionate inviability of any class (Bridges, 1915, Muller,1916). Within 

 each back-cross the inviability effects due to a given mutant form are 

 largely neutralized. Since the inviable form occurs both as a cross- 

 over and as a non-cross-over, both of these classes are lowered, but 

 lowered proportionately, so that the linkage ratio remains practically 



Table 33. — B. C. offspring given by Fi mild-type daughters, from out-cross of 

 purple vestigial male to wild female, when back-crossed to purple vestigial males. 



undisturbed. This internal balancing holds less well for combinations 

 of characters; for any given combination occurs in an experiment 

 either as a cross-over or as a non-cross-over, but not as both, and 

 should any combination have an inviability disproportionate to tliat 

 of the component mutant forms, then the cross-over value would be 

 disturbed. The remedy for this condition is to balance the experi- 

 ments in which a relatively inviable class occurs as a cross-over by an 

 equal amount of data in which this same class is a non-cross-over. It 



