118 Experimental and Statistical Studies upon Lepidoptera 
lies between the values 2e and 3e it will be considered that selection is 'probable, 
while if it exceeds the value 3e selection will be regarded as certain. As the same 
relation obtains between the differences and their probable errors in the case of 
comparable indices of variation, selection on the basis of variability may be 
denoted possible, probable, or certain. 
III. Pupal Elimination. 
A. Males. 
The question at issue may be stated in the following form, Are the surviving 
pupae the same as those which succumb after normal pupation, or are they a 
selected group ? The fact of limitation as regards the number of survivors stands 
unquestioned. Is elimination selective ? Table I. gives the biometric constants 
for the total or unselected material. 
TABLE I. 
All Pupae. 
Males (204) 
Females (356) 
Character 
Mean 
Standard 
Deviation 
Mean 
Standard 
Deviation 
Bust, Length 
„ Width 
„ Depth 
„ Frontal Proportions 
„ Sagittal Proportions 
18-1987 + -0318 
11 - 0466+ -023.5 
12- 0031 + -0222 
60-8220+ -1041 
66-0606 ±-0868 
0-7673 ±-0225 
0-5675 + -0166 
0-5357 + -01 57 
2-5072 + -0736 
2-0916 + -061 4 
19-3460 + -0292 
12- 7182 + -0223 
13- 5117 + -0220 
65-8455 + -08.50 
69-9438 ±-0716 
0-8192 + -0207 
0-6262 + -0158 
0-6178 + -0158 
2-3776 ±-0601 
2-0041 ± -0506 
Left Antenna, Length ... 
Width ... 
„ „ Proportions 
11 -2705 + -02.30 
4-5829 + -01 13 
40 7727 ± -0890 
0-5549+ -0162 
0-2730 + -0099 
2-1447+ -0629 
11-5124 + -0255 
4-2221 + -0082 
36-6629 + -0682 
0-5648 + -0180 
0- 2293 + -0057 
1- 9084+ -0482 
In the following Tables II. and III. are given the numerical determinations 
for answering our first question. The surviving pupae number 134 as against 130 
of the dead pupae. Although the facts appear with clearness in the numerical 
form, a brief verbal statement may not be superfluous. And in such a com- 
mentary we may examine separately two questions as to the nature of selection, 
whether namely it takes place if at all with reference to a certain type, or 
whether it expresses itself in the figures relating to variability. 
In bust measurements, the survivors are found to be longer and narrower than 
the dead individuals. In both of these characters selection in type is certain 
according to the criterion established in earlier treatment. The survivors are also 
deeper (dorso-ventrally) than the eliminated pupae, and are selectively different 
