BISEXUAL REPRODUCTION 225 



merit a number of genetic factors modifying bristle number 

 were present in the stock in heterozygous condition. Selec- 

 tion attending the inbreeding served to eliminate certain of 

 these from one race and to establish them in homozygous 

 condition in the other race, after which no genetic changes 

 would be observed unless they arose de novo. As MacDowell 

 was unable to detect any such changes, he concludes that 

 none were occurring. Payne has carried out a similar selec- 

 tion experiment for changed number of bristles, in another 

 race of Drosophila, starting with the descendants of a 

 single "mutant" individual with "reduced" bristle number, 

 which appeared in an "extra-bristle" strain. Selection was 

 made among the descendants of a single pair of flies and was 

 carried in brother-sister matings in a minus selected line for 

 64 generations, and in a plus selected line for 60 generations. 

 Toward the end of the experiment, the flies of the minus line, 

 in from 96 to 100 per cent of all cases, were without bristles. 

 This degree of purity was attained gradually during the first 

 seventeen generations of minus selection, after which no 

 further genetic change was observed. But in the plus selec- 

 tion, toward the normal number of bristles, four, in other 

 races of Drosophila, progress continued longer, reaching its 

 maximum in the 55th generation when 64 per cent of the 

 individuals possessed four bristles. The two selected lines, 

 plus and minus, had thus become very different as a result of 

 selection. Payne finds evidence that two or possibly three 

 genetic factors affecting bristle number were present in the 

 plus line, two of them being sex-linked, but that in the 

 minus line only a single factor was present. (Figs. 141, 142.) 

 Zeleny inbred a race of Drosophila possessing a dominant 

 character, bar eye, meanwhile selecting for high and for low 

 grades of the character (number of ommatidia). During the 

 course of the selection two striking mutations were observed 

 of the gene under study, one a reverse mutation to "full" 

 (normal) eye, the other a mutation in the direction of selec- 

 tion toward a more reduced condition of the eye, and called 

 "ultrabar." The average size of the eye in these three alle- 



