OF MUTANT CHARACTERS. 271 



TABLE 122. Summary of all cross-over data involving star continued. 



necting-link between that body and the other relatively poorly estab- 

 lished loci of the experiment. 



There was on hand a complex stock containing the two genes 

 streak and dachs (provided by Muller) and this was used in the PI 

 mating to star purple, which was made for that purpose. The FI 

 star streak females were back-crossed to males of a stock of dachs 

 purple which was brought to production simultaneously with the FI 

 cultures (table 121). 



In making the classification of the flies produced by the quadruple 



S' v 



back-cross ^-^ the first separation performed was that of 



streak, disregarding the other characters, and including as streak only 

 those quite obviously streak. These precautions prevented the classi- 

 fication of any not-streak flies among the streak, and established a 

 uniform standard for streak among all the various classes. 



The separations were of varying degrees of difficulty in the different 

 cultures. The hardest were 4999 and 5040, which were fully as diffi- 

 cult as in the abandoned streak dachs back-cross. On the other hand, 

 the separations in culture 5110 were perfectly sharp and complete. 



The streak flies totaled 396, and the calculation of the cross-over 

 values on the basis of these flies gave a star streak value of 15.9 units, 

 a streak dachs value of 16.2, and a streak purple value of 28.8 (table 

 121). 



