Payne: Variations and Selection 



15 



generations there is practically no change, at least not in the 

 direction of selection. From the fourteenth to the eighteenth 

 generations a second upward trend is shown, the mean reach- 

 ing 7.988 in the eighteenth generation. From the eighteenth 

 to the twenty-fifth generations there is again a standstill. 

 Most of the time the mean is below eight, but in the twenty- 

 second and twenty-third generations it is slightly above. In 

 the twenty-fourth and twenty-fifth it is approximately eight 

 again. From the twenty-fifth to the twenty-ninth generations 

 the third rise occurs until the mean reaches 9.084 in the 

 twenty-ninth generation. From the twenty-ninth to the 

 thirty-eighth generation, or to the end of the experiment, 

 the curve is again practically a straight line, with the excep- 

 tion of the thirty-third generation when the mean drops con- 

 siderably below nine. There has been, at least, no decided 

 increase in this period. In other words, the curve forms a 

 series of steps, a thing rather difficult to explain on the basis 

 of selection acting on mere fluctuating variations. It looks 

 more as tho a series of mutations had occurred during the 

 course of the experiment. When a new mutation occurs the 

 the curve goes steadily up until the race is homozygous for this 

 factor. Then it remains on the same level until something 

 else happens. As will be shown later, this is the conclusion 

 I am forced to reach in this experiment. 



A comparison of the curves for the parents and offspring 

 shows a close parallel. 



In order to see whether a return to the normal could be 

 produced by mating low grade parents, a back selection line 

 was started from the eleventh generation and carried for 25 

 generations. In all matings except the first, the parents used 

 had the normal bristle number (four). Table III summarizes 

 these data by generations. The curve plotted from the mean 

 of each generation is shown in Figure 2. With the exception 

 of the first generation no return was produced. From the 

 eighth to the twenty-fifth generation the curve is practically 

 a straight line. It shows clearly, I think, that what has been 

 done by selection cannot be undone. A mass culture bred in 

 the laboratory for more than a year has not returned to the 

 normal. 



The data given in Tables I and II and in Figures 2 and 3 

 show conclusively, I think, that selection has been effective, 

 at least as far as the twenty-ninth generation. No one, it 



