314 ILLINOIS ACADEMY OF SCIENCE 
Probably a more accurate way of detecting inherited ef- 
fects is by a graphical and statistical study of regression, pro- 
vided of course that the numbers of individuals obtained are 
sufficient for such a test, which happily is here the case. The 
means obtained for the offspring of the parents going in the 
different classes are as follows: 
i TABLE V. 
THE MEANS OF ALL THE OFFSPRING OF PARENTS BELONGING TO THE 
DIFFERENT SIZE CLASSES 
Classes Means 
(Length indicated of 
in mm.) Progeny 
1.25-1.29 1.600 
1.30-1.34 — 
1,35-1.39 —_—_— 
1.40-1.44 1.430 
1.45-1.49 — 
1.50-1.54 1.357 
1.55-1.59 1.486 
1.60-1.64 1.497 
1.65-1.69 1.538 
1.70-1.74 1.556 
1.75-1.79 1.508 
1.80-1.84 1.554 
1.85-1.89 1.614 
1.90-1.94 1.742 
1.95-1.99 1.585 
2.00-2.04 — 
2.05-2.09 — 
2.10-2.14 — 
2.15-2.19 1.890 
We can make a plot showing the position of the different 
parent classes, the means of the offspring of the parents go- 
ing into these various classes in regard to the mean of the line. 
This is done in Figure 1. 
We find by observing Figure 1 that nearly all of the parents 
are above the mean of the pure line, which is what we should 
expect since in each case of selection the longest individual 
was chosen as the parent for the succeeding generation. Of 
the thirteen parental groups or classes we find in 12 instances 
that regression is shown by the offspring. In seven of these 
the regression is not complete, in five it is more than complete. 
or beyond the mean of the line. If we compute the mean for 
all of the parents above the mean of the line, we find it is 
1.768 mm., (it is indicated by the heavy circle in figure), and 
if we compute the mean for all the offspring of these parents 
we get 1.555 mm., (indicated by heavy arrowhead in figure. ) 
