Additive effects of temperature and infection with an acanthocephalan parasite on the shredding activity of Gammarus fossarum (Crustacea: Amphipoda): the importance of aggregative behavior.

Climate change can have critical impacts on the ecological role of keystone species, leading to subsequent alterations within ecosystems. The consequences of climate change may be best predicted by understanding its interaction with the cumulative effects of other stressors, although this approach is rarely adopted. However, whether this interaction is additive or interactive can hardly be predicted from studies examining a single factor at a time. In particular, biotic interactions are known to induce modifications in the functional role of many species. Here, we explored the effect of temperature on leaf consumption by a keystone freshwater shredder, the amphipod Gammarus fossarum. This species is found at high densities in the wild, and relies on aggregation as an anti-predator behavior. In addition, gammarids regularly harbor acanthocephalan parasites that are known to induce multiple effects on their hosts, including modifications on their functional role. We thus assessed the cumulative effect of both intra-specific interactions and parasitism. Consumption tests were conducted on gammarids, either naturally infected with Pomphorhynchus tereticollis or uninfected, feeding alone or in groups. Our results show that increased temperatures induced a significant increase in consumption, but only to a certain extent. Interestingly, consumption at the highest temperature depended on amphipod density: whereas a decrease was observed for single individuals, no such effect on feeding was observed for individuals in groups. In addition, infection by acanthocephalan parasites per se significantly negatively impacted the shredding role of gammarids. Overall, the combined effects of parasitism and temperature appeared to be additive. Thus, future studies focusing on the impact of climate change on the functional role of keystone species may benefit from a multimodal approach under realistic conditions to derive accurate predictions. This article is protected by copyright. All rights reserved.