OF MUTANT CHARACTERS. 



2G9 



of streak was then in disfavor (through its failure in the streak cLichs 

 back-cross), so that dachs seemed more practical. 



A star dachs back-cross furnished 1,()17 flies, of which 425, or 20.3 

 per cent, were cross-overs (table 120). Several other experiments have 

 been made which indirectly gave data on this interval. The total 

 available data (table) furnish 3,472 flies, of which 949 or 27.3 per cent 

 were cross-overs. When allowance for double crossing-over corre- 



Table 121.— Pi, star purple 9 X streak dachs d"; B.C., F, .star streak 9 X 



dach spurple cT. 



sponding to the probable coincidence of 30 is made, the locus of star 

 is indicated as 28.8 units to the left of dachs or 4(3.5 units to the left 

 of black. 



As the usefulness of star developed, there was greater necessity for 

 a still more accurate mapping of the star locus. More recent work 

 with streak had shown that tolerably accurate results even under 



