Testing alternative hypotheses for variation in amphipod behaviour and life history in relation to parasitism.

We tested two hypotheses concerning changes in investment in current reproduction for parasitised hosts, using amphipods (Corophium volutator) which act as second intermediate hosts for trematodes (Gynaecoyla aduncta). One hypothesis requires that parasites exert control over their hosts (parasite manipulation), whereas the other predicts that hosts control decisions over investment (adaptive host response). Although these hypotheses are viewed as mutually exclusive, our various results support both hypotheses. For example, female amphipods infected by late-stage larvae were often found crawling at times when predation by sandpipers (Calidris pusilla), which are the final hosts for trematodes, was likely, while uninfected females typically remained in their burrows. Furthermore, old females that were newly-infected by trematodes often aborted and ate their young. Both of these responses seem inconsistent with female investment in current reproduction, but can be interpreted as adaptive parasite manipulation. In contrast, young non-ovigerous females that were newly-infected hastened the onset of their parturial moult and thus, time to becoming receptive. This response can be explained as a host adaptation to minimise the cost of parasites. We contend that differences between parasitised and unparasitised hosts in behaviour or investment can be explained as both parasite and host adaptations, expressed at different times in the host's life history. Such compromise will help explain the persistence of parasite-host associations in nature.

An empirical model of the optimal timing of reproduction for female amphipods infected by trematodes.

Life-history theory predicts that hosts should reproduce when first infected by parasites if hosts are capable and if parasites have a lower cost on current than on future reproduction of hosts. We constructed an empirical model to explore fitness of females of the intertidal amphipod Corophium volutator that reproduced soon versus long after infection by the trematode Gynaecotyla adunca. For uninfected females, the optimal time to reproduce was at their maximum body length. However, for females infected by low or high intensities of trematode metacercariae, reproductive potential (realized fecundity) was highest for females that mated immediately after becoming infected. Even after removing a high cost of delaying reproduction for infected amphipods (high likelihood of depredation by sandpipers, which are final hosts of G. adunca), realized fecundity remained highest if reproduction occurred immediately following infection by trematodes. Results from our model support the view that early reproduction of female amphipods following infection by G. adunca is an adaptive life-history response to parasitism.