REGRESSION IX A PURE LINE OF PLANT LICE. 165 



Fgi6 was selected for carrying on the strain. It had an index of 

 1.58 : I, and gave a fraternity with a mean of 1.82 : I. From 

 this FIO fraternity, Fio4 with an index of 2.11 : I was chosen. 

 T-he mean for its offspring was 1.89 : I. From this F u fraternity, 

 Fn4, formula 1.78 : I, was isolated. It gave the F 12 fraternity, 

 with a mean of 1.84 : I. Fi 2 6, formula 1.98 : I, gave the F 13 

 fraternity, with a mean of 1.85 : I. Fi 3 4, the next parent se- 

 lected, had an index of 1.67 : I, and gave the F J4 fraternity with 

 a mean of 1.69 : I. From the F J4 fraternity r -he last selection 

 was made. Fi 4 T, formula 1.70 : I, was isolated, and gav-e the 

 F ]5 fraternity with 1.81 : I as a mean. 



In each of these selections the individual with either the highest 

 or the lowest antennal index was isolaced in order to obtain the 

 next generation of descendants. As has been stated, the index 

 as obtained without selection for the mean of the pure line was 

 i. 80 : i. Thus it will be noticed that in four of the seven selec- 

 tions made, the offspring of the extreme variant gave a iraternal 

 mean which showed a regression beyond the mean of the strain. 

 In the case of the other three selections the regression of the 

 fraternal mean did not reach the mean of the strain. 



In order to observe more fully the effects of selection and the 

 action of regression, I will give here in the order of their ratios 

 the antennal formulae of the eighteen parents thus far obtained 

 in the first fifteen generations of the pure line 1 together with the 

 indices representing the fraternal means of their offspring. 

 They are as follows: 



Indices for Fraternal 

 Indices for Parents. Mean of Offspring. 



1.66 1.88 



1.67. .1.77 



1.67 1.83 



1.67 1.69 



1.70 i. 8 1 



1.77.... .1.77 



1-77 .1-71 



1.78 1.84 



1.79... .1.95 



1. 80 2.01 



1.85 1. 80 



1.86 1.77 



1 Two of the parents which had antennal indices such as to suggest mutations or 

 abnormalities are omitted. 



