Payne: Variations arid Selection 



7 



Thruout the paper I have adopted the terminology used by 

 Morgan and his students. I have adopted it because it is the 

 easiest way of expressing myself, and because I believe the 

 theories, which have made necessary this new terminology, 

 are the most acceptable working hypotheses we have. If 

 some readers object to locating factors in the chromosomes, 

 they have the privilege of translating my expressions into 

 any terminology they like. 



The Experiment 



An examination of 613 flies from a mass culture which 

 had been bred in the laboratory for about three months gave 

 one female with one extra bristle and 612 normal. This 

 female mated to a normal male from the same mass culture 

 was the starting-point of the experiment. For comparison, 

 counts were also made in two other lines. One line gave 

 710 flies with the normal bristle number (four), one fly with 

 three bristles, and four flies with five bristles. The second 

 line gave 2,514 with the normal number, 14 with three, and 

 15 with five bristles. All three lines, then, showed some 

 variations in bristle number. The experiment was started 

 from the line which showed the least variation. 



The pair which was mated (female five with normal male) 

 gave 228 offspring, 226 of which were normal and two females 

 with one extra bristle each. Twelve pairs of the normal 

 flies were mated. They gave 2,647 ofl'spring, 2,588 of which 

 were normal and 59 with extra bristles, a ratio of normals 

 to extras of 1 :43.8. Of the extras, 51 had one extra bristle 

 and eight had two extras. The two F^ females with one extra 

 bristle were mated to normal brothers. They gave F. off- 

 spring as follows : 935 normal, 39 with one extra bristle, and 

 four with two extra bristles, a ratio of 1 : 21.7. Neither of 

 these Fo ratios would be called Mendelian, at least, not without 

 further analysis. I mention this because extra bristle num- 

 ber in MacDowell's experiment behaved as a recessive unit 

 character, giving in F. a ratio of approximately three normals 

 to one extra. The extra bristled F^ flies from the F^ females 

 with one extra bristle were selected as parents for the next 

 generation. From their offspring the flies with the highest 

 number of extra bristles were again selected as parents. This 

 method of selecting the highest grade flies as parents has been 



