theoretical explorations remain mostly untested to this date. At present,
our knowledge of the phenomenon of host manipulation by parasites
remains superficial, consisting mainly of a check-list of known systems in
which manipulation has been documented. We now have the tools to
generate testable, quantitative predictions that would shed light on how
the phenomenon has evolved; the time has come to use them.Phylogenetic Evolution of Manipulation
The most informative way to examine the evolutionary history of host
manipulation by parasites, like any other trait, is to view it in a phylo-
genetic context (Moore and Gotelli, 1990). This is the only way that allows
one to determine how often and in what circumstances the trait has
evolved. In certain taxa, the ability to manipulate hosts may be an
ancestral trait inherited by all living species and would thus have evolved
only once. This appears to be the case for acanthocephalans (Moore,
1984). In other cases, it may have evolved repeatedly in different lineages
exposed to similar conditions. Only by mapping the ability to manipulate
on to a parasite phylogeny can we begin to address these issues.
There has been only one rigorous attempt to achieve this. Moore and
Gotelli (1996) studied the effects of the acanthocephalanMoniliformis
moniliformis on the behaviour of seven species of cockroaches, all
suitable intermediate hosts for this parasite. In laboratory experiments,
the behaviour of infected cockroaches of all seven species was compared
with that of uninfected conspecifics. The behavioural measures scored
were all likely to influence a cockroach’s susceptibility to predation by
rats, the acanthocephalan’s definitive host. There was no strong concor-
dance between the distribution of observed alterations in particular
behaviours among the seven host species and a cockroach phylogeny
based on morphological features (Moore and Gotelli, 1996). This finding
suggests that the effect of the parasite is independent of phylogenetic con-
straints and has evolved separately in each host–parasite combination.
Unfortunately, the weakness of this study is that certain of these seven
host–parasite associations do not occur in nature, even if all cockroach
species proved to be suitable as hosts in the laboratory. Nevertheless, this
study shows the power of phylogenetic approaches to evaluate whether
particular feats of manipulation are unique or inherited.
Another approach is to examine how phylogenetically unrelated
parasite taxa have solved similar or even identical transmission prob-
lems (Poulin, 1995). Convergent evolution is likely to have favoured
similar types of manipulation in different taxa that have intermediate
and definitive hosts involved in similar predator–prey relationships. For
instance, two completely unrelated parasite species, the acanthocephalan
Polymorphus paradoxus(Holmes and Bethel, 1972) and the trematodeM.
papillorobustus(Helluy, 1984), both use littoral amphipods (GammarusParasite Manipulation of Host Behaviour 251