PAPERS ON ZOOLOGY AND PHYSIOLOGY 317 
we obtain is 1.0465, which means that the regression was a 
little more than complete according to this way of measuring it. 
We find then that this statistical method of measuring the 
amount of regression confirms my earlier findings in regard 
to the effectiveness of long continued selection in a partheno- 
genetic pure line. 
Variation, Selection and Regression under Similar Sub- 
optimum Food and Temperature Conditions. 
It has previously been mentioned that environmental con- 
ditions in two instances caused such a marked change in the 
size of both parents and offspring as to give a considerable 
amount of correlation between the two in size variation. For 
this reason the correlation noticed in Table I is misleading 
without the use of the annoted pedigree. During the second 
score of generations in isolation 11, selections were made under 
similar suboptimum food and temperature conditions. For 
this reason we should expect but little influence in size fluctu- 
ations coming from variations of these two factors, which 
were determined as being the two chiefly affecting size. We 
would expect, therefore, that the common statistical methods 
in analysis should be more applicable for selections made un- 
TABLE VI. 
TABLE SHOWING GROUPS TO CONTAIN PARENTS OF SECOND SCORE OF GENERA- 
TIONS OF ISOLATION 11, THE NUMBER OF PARENTS TO EACH GROUP, 
THEIR LENGTHS AND THE MEAN LENGTHS OF PROGENY OF 
PARENTS BELONGING TO EACH GROUP. 
Group No. of parents Length Mean length 
(length in mm.) in group of parent of offspring 
1.50-1.54 | 1 | 1.53 mm. | 1.518 mm. 
1.55-1.59 1 | 1.59 mm. 1.549 mm. 
1.60-1.64 1 1.61 mm. 1.407 mm. 
1.65-1.69 2 | 1.69 mm. 
= | 1.69 mm. 1.552 mm. 
1.70-1.74 4 1.70 mm. 
- 1.73 mm. 
- 1.74 mm. 
= 1.71 mm. 1.515 mm. 
1.75-1.79 3 | 1.79 mm. | 
- | 1.76 mm. | 
- | | 1.79 mm. 1.574 mm. 
1.80-1.84 5 1.80 mm. 
Sy i 1.81 mm. 
- | 1.80 mm. 
- 1.81 mm. 
- | 1.84 mm. | 1.495 mm. 
1.85-1.89 3 1.89 mm. 
~ | 1.86 mm. 
- | 1.86 mm. } 1.638 mm. 
