A RECESSIVE CHARACTER AND SELECTION 163 



From table 1, it is to be noted that some generations are much 

 more variable than others when judged by the standard devia- 

 tions which range from 2.11 ± 0.010 to 11.30 ± 0.60 in the 

 males and 1.69 rt 0.17 to 7.68 ± 0.49 in the females. The 

 coefficients of variability are 10.85 dr 1.58 to 42.50 ± 2.61 for 

 the males and 8.86 ± 0.89 to 31.66 ± 2.19 for the females. 

 Tlie means of the wing lengths of the different generations vary 

 from 16.23 to 26.88. 



h. Control series B. Control series B had its origin in the 

 random mating of the brothers and sisters of the two individuals 

 used for selected series A, and was carried for sixteen genera- 

 tions. Table 2 gives the means, standard deviations, and co- 

 efficients of variability of the males and females separately, while 

 in table 4 are found the averages of the wing lengths of males 

 and females in the different generations. 



This series shows in general the same things as selected series 

 A. It is to be noted from text figure 1, which is constructed 

 from the means of the generations given in table 4, that the 

 wing length in the control series is smaller in every generation, 

 except the F^, than in the selected series. It is thought that this 

 difference is due to the method of handling rather than to any 

 inherent difference. In the selected series only a few parents 

 were in the breeding bottle and the offspring were removed every 

 eighteen hours or more often, while in the control series, the flies 

 were not removed so frequently and a large number of parents 

 was used. The food conditions and space in the breeding bottles 

 were not so favorable in the control as in the selected series. In 

 working with the flies, one received the impression that those of 

 the control series were smaller than those of the selected series. 

 Measurements of the body lengths of males in the sixth generation 

 of each series were made, giving 62.9 for the selected and 60.4 for 

 the control. Males were measured in preference to the females 

 in this comparison because the body size in the males is more 

 constant than in females due to the presence or absence of eggs 

 in the latter. 



c. Effect of temperature. The high means in the Fio generation 

 of control series B and in the Fio, F12, and F29 generations of se- 



