114 Experimental and Statistical Studies }ij)on Lepidoptera 
species has likewise been under investigation, and attempts have been made to 
cross various species in order to ascertain whether heredity in such cases follows 
the principles of Mendel. 
The lepidoptera afford extremely favourable material for investigations of the 
above nature, for they can be obtained in considerable numbers, and they may 
be bred with not too great difficulty. And a qualification of special importance 
is to be found in the sharp division of their life-history into three well-demarcated 
periods, each characterized by peculiar external and internal relations. It is of 
course clear that the structural relations of the imago must be totally different 
from those either of the larva or of the pupa, and it is equally evident that these 
forms have entirely diverse relations to the " environment " at large. Therefore 
the problem of selection derives an added interest from the comparison of 
elimination as it appears at one stage with that of another period. 
Furthermore the imago and the pupa possess attributes which give to each a 
peculiar value. The former enters upon its final period of life with a rigid and 
unchangeable organization, incapable for the most part of structural alterations 
as the result of " functional adaptation." The pupa excites still greater interest 
on account of its inability to " use " its various organs in any way. Life and 
adaptive response are here minimized to their lowest values ; and helplessly 
subject to environmental influences and constitutional weakness the pupa awaits 
the period of renewed activity which culminates in metamorphosis. 
It is from the very nature of the case impossible to present the results of the 
several studies carried forward during the past few years in a simple and 
condensed form. Merely to enumerate the statistical data which have been 
accumulated would trespass beyond even liberal bounds. Therefore it has been 
deemed best to offer the statement of problems and results in a series of papers, 
each dealing with a well-restricted division of the whole field. The present long- 
deferred account, which initiates this series, indicates as well the point of 
departure for the later investigations. 
I. Introduction. 
The following account is a statement of the results of an examination into the 
occurrence of " natural selection " in the case of over-wintering pupae of Philo- 
samia cynthia, as this species occurred in New York City in the year 1899. It 
is concerned, therefore, with the existence of a definite relation between elimina- 
tion on the one hand, and the extent and character of variation on the other ; 
and particular attention is directed towards ascertaining the real basis for the 
selective process. 
In previous years, when large numbers of the cocoons and pupae of this species 
had been collected for other purposes, it had been noticed that a great many cocoons 
contained individuals which had died, although they were apparently perfectly 
