No. 418.] NESTS OF AMERICAN ANTS. 80I 
F. sanguinea, however, there is already given the generalized 
condition required as a starting point for the action of natural 
selection and the development under its guidance of cases like 
Polyergus, as Darwin suggests. In Polyergus the predatory 
instincts have been developed to a highly specialized condition, 
while the domestic instincts have retrograded pari passu as a 
natural result of the survival of the prey, till the presence of 
slaves in the nest has become a conditio sine qud non of exist- 
ence. This correlation of instincts has involved the corre- 
sponding correlation of structure which we find so beautifully 
exhibited in Polyergus.! In the predatory instincts every 
slight variation in advance would be beneficial to the species, 
while slight retrogressions would not under the circumstances 
be disadvantageous. I cannot, therefore, agree with Wasmann 
when he says ('91, p. 247) : “ Natural selection could only main- 
tain and augment useful instinct variations: but the develop- 
ment of slavery up to the Polyergus-, Strongylognathus-, and 
Anergates-stage is beneficial neither to the masters nor the 
slaves — ergo natural selection cannot have produced the 
instincts of the slave-holding ants." As good an answer as I 
can conceive to an argument of this nature is a reference to 
the cases of extremely specialized parasitism like the Cestode 
and Sacculina, both of which are connected by tolerably com- 
plete series of intermediate forms with the more generalized, 
non-parasitic members of their respective phyla. 
The symbiotic sequence suggested by Forel and Lubbock 
is objected to by Wasmann on fairly good grounds. It is by 
no means clear that the development has passed successively 
through the stages represented by these forms. Indeed, as 
Wasmann shows, the problem of symbiogenesis is much more 
complicated than it appeared when the above sequence was 
Suggested. It now seems evident that several lines of devel- 
opment have proceeded independently from cases of plesio- 
biosis (and possibly also parabiosis), which constituted the 
necessary initial stages of symbiogenesis. Thus it is probable 
that cleptobiosis, xenobiosis, and dulosis represent at least three 
"In accounting for this development of instincts and structures, it is, of course, 
necessary to regard the whole mixed colony as a single evolutionary unit. 
