

TOWER, DARWINISM. 427 
a butterfly-net, that it was quite as apt to rest on the white-washed 
wall of my house as upon any other place, and in only two instances 
did I ever see one of these butterflies, when pursued, come to 
rest on a twig where his form and coloration might be considered 
as acting as the Hypothesis postulates. 
_On several occasions insectivorous birds were observed to attack 
these butterflies, but I was not able to discover that form, coloration 
or habits served to protect them in the slightest. One of their 
favorite haunts for spending the night was on the white-washed 
wall under the eaves of my house, where they might be considered 
to resemble dead leaves, but as a matter of fact, their resemblance 
did not protect them from certain nightflying species of goat-suckers 
that | repeatedly observed picking them off. Other night-birds were 
undoubtedly active, because a large magnolia tree which was also 
a resting place at night, showed nearly every morning one or more 
sets of wings that had been nipped off from the bodies during the 
night by some predacious vertebrate. 
It is not known, of course, either in these instances of protective 
resemblance or in the mimetic groups, what proportion of the adult 
members of any generation was eliminated, and I know of no way 
of making even an approximately reliable determination. Of those 
that were eliminated, all evidence was overwhelmingly against the 
predacious agencies that are assumed to operate, while + 91,5°, 
of the eliminated met their death through the instrumentality of 
spiders, ants, dragon-flies and other predacious insects that have 
never been alleged to exercise any discrimination. I made effort 
to discover whether there were evidences of their possessing this 
discrimination, but never discovered any. The fact that the total 
eliminated in both groups showed no aiscoverable differences 
between the eliminated and the survivors, permits of but one con- 
clusion, namely, that elimination depended on chance position 
relative to the accidents of life rather than upon protecting qualities. 
When it is recalled that all of these activities refer to the two or 
three per cent of each generation which are assumed to reach 
‘maturity, it is obvious that the selective effect of the elimination 
is too trivial to have the modifying effect necessary to produce 
