Payne: Variations and Selection 



5 



interpretative value is sure to receive scant attention." Yet 

 Pearl in this same paper accepts MacDowell's conclusions. I 

 agree with Pearl that what we need most of all is demonstra- 

 tive evidence and more of it. MacDowell used bristle number 

 on the thorax. Altho some of my conclusions may agree with 

 MacDowell's, the work is in no way a repetition of his. 



The usual method in selection experiments is to start with 

 a single pair of parents and from the offspring isolate two 

 or more strains which are carried on by mating brother and 



Figure 1 



(a), an outline of the head and thorax with the scutellum of a nor- 

 mal wild type fly, showing the position of the bristles. Note the four 

 on the scutellum. (b), a similar outline showing 15 bristles on the 

 scutellum, the maximum number found in any one fly in the strain se- 

 lected for extra bristles, (c), an outline similar to the preceding show- 

 ing no bristles on the scutellum. This is taken from the minus selected 

 strain of the mutant race "reduced". In this strain nearly all the flies 

 have extra bristles (as high as 99.52 per cent in the seventeenth genera- 

 tion) . 



sister. In the present experiment, this method has not been 

 strictly followed. A single pair of parents, female with one 

 extra bristle and normal male, was the starting-point. From 

 among their offspring, the two females with one extra 

 bristle were mated to their normal brothers. In the second 

 generation, several flies with one extra and a few with two 

 extras were found. These extra bristled flies were used as 

 parents for the next generation. In all the following genera- 

 tions, flies with the highest bristle number were used as 

 parents regardless of whether they were brother and sister. 

 Most matings, however, were brother and sister matings. In 



