OF MUTANT CHARACTERS. 193 



TABLE 45. Summary of purple cross-over data continued. 



SPECIAL PROBLEMS INVOLVING PURPLE-AGE-VARIATIONS, COINCI- 

 DENCE, TEMPERATURE-VARIATIONS, CROSS-OVER MUTATIONS, 

 PROGENY TEST FOR CROSSING-OVER. 



We have already seen how the study of the age- variation in crossing- 

 over for the second chromosome began with the purple vestigial back- 

 cross (p. 181) and was continued and confirmed by the black purple 

 curved triple back-cross (p. 185). Some of the early data suggested 

 that the drop in the second broods was followed by a recovery and 

 perhaps even by a rise in later broods (Bridges, 1915). 



To gain further light on the course of the variation throughout the 

 life of the fly a special and extensive experiment was continued through 

 four broods. The entire length of the chromosome was covered by the 



loci chosen ( - 



\ p, c s t 



This experiment showed the normal cross-over values for the first 

 broods, the usual drop for the second broods, and a slight continued 

 drop for the third and fourth broods. However, the experiment 

 proved inconclusive because of two ill-adaptations: the chromosome 

 distances involved were so long (e. g., S' pr = 52.7) that real changes 

 could be concealed by a concomitant change in double crossing-over, 

 and further because the 10-day broods gave only four points on the 

 curve of age- variation, each point representing only the net change 

 for a 10-day period, while the real underlying curve may have changed 

 its course so that an unknown part of each 10-day period may have 

 been a fall and the rest a rise. 



The original black purple curved experiment had avoided one of 

 these difficulties in that the black-purple distance is so short that there 



