NOTES ON REGRESSION IN A PURE LINE OF PLANT 



LICE. 



H E. EWING. 



In a previous number of the BIOLOGICAL BULLETIN 1 the 

 writer published the results of some selections made within a 

 pure line of plant lice. These w r ere made in an attempt to in- 

 crease and to decrease the length of the third segments of the 

 antenna? in comparison with that of the fourth segments. The 

 species used was our common European grain aphis, Aphis aven<z 

 Fab. ; and the results obtained up to that time, which included the 

 first ten generations, appeared to show that selections from ex- 

 treme variants did not alter the mean as obtained for the line 

 without selection. These results were in accord with the results 

 obtained by other workers in pure lines. 



But it also appeared that the mean of the offspring of the 

 variants selected reverted not to the mean of the line, or strain, 

 but that it would swing pendulum-like much beyond this mean 

 only to be brought back to the former side of the mean-of-the- 

 strain base line in the next generation. 



Since the publication of this former paper I have tested further 

 this aspect of regression by selecting from opposite extremes in 

 alternating generations. These selections were begun in the 

 ninth filial generation, and continued into the fifteenth, including 

 in all seven generations. I will now give briefly the results of 

 these selections. 



An individual with the remarkably high index of 2.82 : I 

 (i. e., the third segments of the antennae were two and eighty-two 

 hunclredths times as long as the fourth segments) was selected 

 as the stem progenitor from among the individuals of the F s 

 fraternity, which fraternity has been mentioned in my previous 

 paper. From this individual there were reared five wingless 

 adults which gave a fraternal mean of 1.75 : I, i.e., .05 below the 

 mean of the strain which w r as 1.80 : i. From the F u fraternity, 



1 Vol. XXVI., No. i, January, 1914. 



164 



