122 Experimental and Statistical Studies upon Lepidoptera 
Figures 2 and 3 illustrate the general nature of the selection in two 
by graphical representation of the original material and the selected survivors 
Fig. 3. Selection in Type and Variability. Pupal Elimination, Males. 
Left Antenna Length (mm.). 
95 
98 
10 1 
10 4 10 7 
110 
113 116 
119 
12,2 125 
Total 
m. 
(7 
All Pupae 4 
3 
9 
10 18 
57 
51 67 
32 
11 2 264 
11-2705 
0-5549 
Survivors 
3 
6 7 
27 
29 37 
19 
5 1 134 
11-3514 
0-4686 
IV. P'upal-Imaginal Elimination. 
We may now consider the nature of that reduction which takes place at the 
time of metamorphosis, and which, although dependent upon pupal abilities, is 
manifested only when the great changes at this time prove too severe a tax upon 
the strength of a large number of individuals which were able to survive throughout 
the earlier period of pupal existence. 
It has been stated previously that only 181 out of the 310 survivors produce 
perfect moths. The rest were imperfect to a greater or less degree, or failed 
entirely to metamorphose, as follows : 
Perfect Slightly Imperfect Very Imperfect Non-metamorphosed 
Males 97 (72-3%) 14(10-4%) 18(13-4%) 5(3-7%) 
Females 84 (47-7%) 61(34%) 20(11-3%) 11 (6-27J 
It therefore appears that the females are the greater sufferers, although if the 
slightly imperfect individuals be counted with the perfect ones the males and 
females occur in about the same percentage*. 
* In studying the phenomena of sexual selection in Samia cecropia it has been found that a male 
which is imperfect to any degree will not copulate, while females which are slightly imperfect may 
sometimes copulate. These facts will be considered at length in a later communication. 
