OF MUTANT CHARACTERS. 



287 



curved were classified and no attention was paid to black. The result 

 was rather unexpected, for while crimson and curved proved to be 

 independent (56 per cent recombinations), the curved flies constituted 

 only 31.4 per cent of the flies instead of 50.0 per cent (table 131). This 

 difference was not clearly recognized until the counts were totaled, 



TABLE 131. PI, crimson black curved cf X wild 9 ; B. C., F\ wild- 

 type 9 + crimson black curved father. 



and there were then no more flies hatching. However, one of these F 2 

 cultures was recovered from the discards and a count was taken of it. 

 This count showed no black curved flies whatever, and far too few blacks 

 and curved. The case of dachs-lethal was then being followed, so that 

 a hypothesis was known that covered this situation. It was concluded 

 that an autosomal lethal had arisen by mutation in this second chro- 

 mosome at a locus between black and curved. No black curved flies 



TABLE 132. Offspring given by pairs of wild-type flies from an F 2 

 to the back-cross of table 131. 



parallel 



appeared in the F 2 , because every black curved zygote (except the 

 rare double cross-over) was at the same time homozygous for the lethal. 



By crossing-over between black and the lethal ( - > , ) 



\+ + + +l c ' 



/fe++\ ' 



few black flies would be produced I - ) ; likewise the few curved 



\6 I c / 



6 I c / 



flies corresponded to crossing-over between the lethal and curved. 

 If this were the explanation, then most of the wild-type flies should 



be of the same constitution as the FI flies ( - - ] 

 together should repeat the F 2 result. \+ + +/ 



and when mated 



