aspects of the other two hypotheses and, to some extent, might be con-
sidered a synthesis of the various ideas that have been presented. The
predation hypothesis is unlikely to be a complete explanation, since the
ants in the various phylogenetic lines in which slave makers occur are
not strongly predatory and do not attack other ant colonies as a major
component of their foraging strategies (but see Topoff, 1991, for a possible
exception). None the less, opportunistic brood predation as a result of
territorial interactions is relatively common in these groups (Wilson,
1975b; Alloway, 1980; Stuart and Alloway, 1982, 1983). The transport
hypothesis provides a ready explanation for the transport of pilfered
brood and the recruitment of nest mates at the conclusion of a slave raid
but does so in a non-aggressive intracolonial context, which is quite
different from and not easily translated into the aggressive intercolonial
context of a slave raid. Indeed, if fragments of a polydomous colony
became unfamiliar and aggressive towards one another (as they might
well do (see Stuart, 1987)), then the behaviours expressed in this context
would probably be aggressive territorial behaviours, and the hypotheses
merge. Moreover, point-by-point comparisons of the slave raids of slave
makers and the territorial interactions of related free-living species
indicate that territorial behaviour includes all of the behavioural elements
seen during slave raids: scouting, fighting, recruitment of nest mates prior
to the assault on the opposing nest and thereafter, the invasion of the
target nest and the transport of the remaining brood back to the victorious
colony’s nest (Stuart and Alloway, 1982, 1983). The refining of these
behaviours by increasing the distance and frequency of raids and
increasing the acceptance of captured brood, coupled with an increased
tendency for parasitic colony foundation, would put such species well
on the way to being obligatory slave makers. Moreover, the combination
of secondary polygyny and territorial behaviour is common in the various
ancestral groups that gave rise to the slave makers; and facultative
interspecific slavery as a consequence of territorial interactions has been
reported in these groups (Wilson, 1975b; Alloway, 1980, 1997; Hölldobler
and Wilson, 1990). Thus, secondary polygyny and territorial behaviour
appear to have been key elements in the evolution of slavery in ants.
The evolution of socially parasitic relationships between closely
related sympatric species is the typical model presented for the evolution
of the relationships that we see in mixed colonies, with two, formerly
independent, free-living species evolving to form a host–parasite dyad
(Hölldobler and Wilson, 1990). However, some authors have also
suggested that sympatric speciation, in which a species essentially
generates its own parasitic sibling species, is also possible under certain
circumstances and might better explain the evolution of many inquiline
species (see Buschinger, 1986, 1990; Bourke and Franks, 1991, 1995; and
references therein).
Perhaps the best example of apparent evolution from one form of
social parasitism to another is seen in the myrmicine genusEpimyrma,
where slavery might have evolved into a form of inquilinism (Buschinger,Social Parasitism in Ants 331