mathematical treatment is given in Brown (1999) and more detailed
disussion is presented elsewhere (Poulin, 1994, 1998b), but the super-
ficial and qualitative treatment given above is sufficient to illustrate
the optimality approach. Many other ecological parameters are likely
to influence the optimal level of investment in manipulation, such
as the likelihood that a parasite will share an intermediate host with
conspecifics and the longevity of both the parasite and the intermediate
host following infection (Poulin, 1994). These can also be incorporated in
a quantitative mathematical model to obtain more refined predictions.
There are then two ways of testing these predictions. First, as suggested
by Poulin (1994) and Brown (1999), comparisons can be made across
species that differ in respect of passive transmission rate, mean number
of parasites per intermediate host or longevity of the intermediate host.
Secondly, using model systems where some of these variables can
be manipulated, certain predictions can also be tested experimentally.
The advantage of the optimality approach is that it generates specific248 R. Poulin
Investment in manipulation1p0
*Probability of transmission or dying earlyFig. 12.2. Probability of parasite transmission and probability of the parasite dying
early as a function of investment in the manipulation of the intermediate host.
Without any investment in manipulation, the parasite has a passive transmission
rate (p) that is greater than zero; increasing investment in manipulation yields higher
transmission probabilities but with diminishing returns (top curve). At the same time,
the cost of manipulation, or the probability of dying early, increases with the level of
investment, following a sigmoidal function in this hypothetical example (bottom
curve). The optimal investment in manipulation (*) is the level at which the net gain
(benefits minus costs) in transmission probability is maximized.