270 CALVIN B. BRIDGES 



purple vestigial, and the other was purple vestigial pure for 

 vermilion. The special advantage of this latter stock lay in the 

 fact that the presence of vermilion accentuates the difference in 

 eye color between the flies that are purple and those that are 

 not, that is, vermilion purple is easier to separate from ver- 

 milion than is the case in the equivalent separation of purple 

 from red. 



This latter stock was accordingly used in the Pi mating for 

 the first back-cross test. Vermilion purple vestigial males were 

 outcrossed to females of vermilion stock (May 25, 1912). Both 

 parents were homozygous for vermilion, and the Fi flies were all 

 vermilion as expected. Both purple and vestigial are recessive. 

 When the back-cross matings came to be made, the culture 

 bottle happened to contain no virgin Fi females, since the Pi 

 mating had been made at Columbia and the Fi progeny used 

 had hatched en route to Wood's Hole. The back cross was 

 therefore made in only one way — by mating the Fi males to 

 virgin vermilion purple vestigial females of the stock kept for 

 that purpose. Five back crosses were started by mating in 

 each case a single Fi vermilion male by two or three stock ver- 

 milion purple vestigial females. At the end of ten days the 

 parents were removed from the culture bottles and were put in 

 fresh bottles in which second broods were raised. In one case a 

 third brood was raised (table 1). 



The linkage results of these back crosses were somewhat 

 unexpected, for in four of the lines no crossovers at all were 

 obtained, and in a fifth only a few. In the original F2 culture 

 several crossovers had been noted, and five F2 cultures raised 

 from the brothers and sisters of these back-crossed males were 

 giving in the neighborhood of 15 per cent of crossovers (table 2). 

 The apparent crossovers had all appeared in one culture of the 

 first and of the second broods, and for this reason a third culture 

 was raised from that particular set of parents and it also gave 

 apparent crossovers. 



A second back-cross experiment, using the simple purple 

 vestigial stock instead of the vermilion purple vestigial, was 

 started (June 25, 1912) a month later than the first and before 



