Field evidence for non-host predator avoidance in a manipulated amphipod.

Manipulative parasites are known to alter the spatial distribution of their intermediate hosts in a way that enables trophic transmission to definitive hosts. However, field data on the ecological implications of such changes are lacking. In particular, little is known about the spatial coexistence between infected prey and dead-end predators after a parasite-induced habitat shift. Here, we used an Amphipoda (Gammarus roeseli)-Acanthocephala (Polymorphus minutus) association to investigate how infection with a manipulative parasite affects the predation risk by non-hosts within the invertebrate community. First, we collected invertebrates by sampling various natural habitats and calculated the distribution amplitude of amphipods according to their infection status. Infection with P. minutus significantly reduced the habitat breadth in G. roeseli, parasitised individuals being mainly found in floating materials whereas uninfected ones were widespread throughout the sampled habitats. Second, to test if these changes also affect the risk for P. minutus to be ingested by non-hosts, we estimated the predation risk experienced by G. roeseli within the macro-invertebrate community. The habitat overlap between potential invertebrate predators and G. roeseli showed that the spatial probability of encounter was lower for P. minutus-infected amphipods than for uninfected conspecifics. For the first time, to our knowledge, a study used ecological tools to bring field evidence for the spatial avoidance of dead-end predators in a manipulated amphipod.