GENES MODIFYING NOTCH. 351 



lines is not clearly separable now from those recorded in the last sec- 

 tion. There is, moreover, the possibility that during these early expe- 

 riments, stock males of eosin ruby may have been introduced at one 

 stage. That these conditions have not affected seriously the condition 

 of the selected stock as a whole is shown by table 4, where the number 

 of normal females belonging to the potential Notch class is as high in 

 most cases as in the middle and latter parts of the preceding table. 



By introducing into the experiments the two genes eosin and ruby, 

 it is a very simple matter to identify potentially Notch females from 

 the other females with normal wings. Selecting the former makes it 

 possible to carry on the experiment by breeding in every generation 

 from those females that carry the factor for Notch, but do not show a 

 notch in the wing. In other words, if the expression of a character (its 

 phenotype) is a measure of the major factor that produces it, we should 

 expect that an extreme selection of this kind would be an excellent way 

 of fixing the factor altered by selection. 



The location of the Notch factor had shown that it lies in a region 

 of the X chromosome (fig. 92), 2.8 units from the arbitrary zero-point 

 yellow. Eosin lies 1.5 units and ruby lies 7.3 units from yellow. The 

 distance between eosin and ruby is therefore a distance so short that 

 double crossing-over never takes place within it. If, then, we use a 

 male whose sex-chromosome contains the factors for eosin and for ruby, 

 and a Notch female having red eyes (i. e., the normal allelomorphs of 

 eosin and of ruby) the gene for Notch in one X of the daughters will 

 be located in a position between the eosin ruby genes present in the 

 opposite chromosome of the same daughter, as seen in figure 92. 



Now, as said above, it would necessitate double crossing-over to get 

 the Notch gene in between the eosin and ruby genes, or, in other words, 

 double crossing-over must take place within these limits to produce a 

 Notch female with eosin ruby eyes. Of the many thousands of females 

 obtained in the course of the experiment, not a single double cross- 

 over of this kind has been observed. 



Single cross-overs have, however, been recorded in the expected 

 numbers. Thus eosin and Notch females, and eosin as well as ruby 

 males have appeared. It would of course be possible to obtain a Notch 

 fly with eosin ruby eyes by first getting a single cross-over of eosin 

 Notch and then after mating such a female to an eosin ruby male some 

 daughters, in which a cross-over between Notch and ruby would result, 

 having eosin Notch ruby in the chromosome. Such a female bred to 

 an eosin ruby male would give some daughters of the desired class. 

 As there was no need in my work for such females, I have not taken 

 the trouble to make them. 



Turning to table 4, we see that nothing further resulted from select- 

 ing the potentially normal females through about 5 more generations. 

 By potentially normal females I mean that females with red eyes and 



