about the combinations likely to maximize the net reproductive output.
This approach has been particularly fruitful in the area of foraging ecology
(Stephens and Krebs, 1986). In the context of host manipulation by para-
sites, and assuming that the currency to be maximized is the transmission
success of the parasite to its definitive host, optimality theory allows us to
predict the optimal investment into manipulation of the intermediate host
that a parasite should make (Poulin, 1994; Brown, 1999). From both the
benefits and the costs of manipulation, the optimal level of manipulation
can be derived as the one that achieves the greatest net benefits (benefits
minus costs).
For instance, let us assume that the energy invested into manipulation
by a parasite is variable within a population and under genetic control.
This is most probably true of most cases. We might expect that the more a
parasite invests into manipulation of its intermediate host, the more it
increases its probability of transmission to the definitive host beyond
what it was initially. Without manipulation, the parasite might still be
transmitted, but with a lower probability,P, corresponding to passive
transmission. Manipulation enhances this probability, but with diminish-
ing returns, i.e. small investments in manipulation yield greater returns
per unit investment than larger investments (Fig. 12.2). Beyond a certain
level of investment into manipulation, the probability of transmission
approaches 1. At the same time, the costs of manipulation are also likely
to increase in proportion to the investment in manipulation. Little is
known about how costly it is for parasites to control the behaviour of their
hosts. Two examples, however, suggest that the cost may sometimes be of
an all-or-nothing nature. Metacercariae of the trematodeDicrocoelium
dendriticumalter the behaviour of their ant intermediate host to increase
their transmission to their sheep definitive host. Typically, of the many
metacercariae inside one ant, only one migrates to the ant’s brain to
induce the behavioural change: this metacercaria normally does not infect
the mammalian host and dies (Wickler, 1976). Similarly, metacercariae
ofMicrophallus papillorobustusalter the behaviour of their amphipod
host to facilitate their transmission to bird hosts. Those that induce the
manipulation are the ones that encyst in the brain of the amphipod; many
others encyst in the abdomen and cause no altered behaviour. Frequently
metacercariae in the brain are encapsulated and melanized by the
immune system of the amphipod host, whereas abdominal parasites are
very rarely attacked (Thomaset al., 2000). Both these examples suggest
that the probability of dying as a result of attempting to manipulate the
intermediate host can increase drastically. For this reason, the shape of
the cost function may be close to a sigmoidal curve, with costs rising
sharply with only modest investments in manipulation (Fig. 12.2). Other
functions and curve shapes are also possible, of course.
The optimal investment in manipulation, then, is the investment level
that achieves the greatest positive difference between the benefits and
the costs in terms of transmission success (Fig. 12.2). A more rigorousParasite Manipulation of Host Behaviour 247