Trait-mediated effects of ants on arthropods: From individuals to communities
Final Report Abstract
To minimize predation risk, prey show a suite of phenotypical and behavioural responses (e.g. avoidance, changes in feeding rate) upon detecting a predator, known as traitmediated effects (TME) or non-consumptive effects (NCE). In this project, we investigated the significance of TME in intraguild interactions in terrestrial systems. We used ant semiochemicals to test the factors influencing TME of ants on spiders, since ants are intraguild predators and competitors of spiders and have strong negative effects on spider populations. Our overarching goal was to improve our understanding of context dependence in food webs and predictability in community ecology. We first tested 17 spider species to determine how their traits (foraging mode) affected their avoidance behaviour towards ants (Lasius niger). We let ants walk on filter papers before removing them to place one spider on a closed arena with either ant cues or no cues and record its behaviour. Our initial prediction was that TME would be larger in sedentary spiders, because these face stronger selection pressures in choosing a suitable site than species that forage actively throughout the habitat. Instead, we found that cursorial but not sedentary spiders species avoided ant cues, which is in accordance with the higher risk of cursorial spiders of a lethal encounter with ants. Thanks to this multi-species approach we could show that TME are widespread among intraguild predators but that they highly depend on species traits. In a second phase, we extracted cuticular hydrocarbons and trail pheromones from L. niger in order to disentangle the effects of different chemotactile cues on spider avoidance behaviour and to establish the chemical basis of TME of ants on spiders. We conducted choice experiments on filter papers where one half contained chemical cues in realistic doses (either trail pheromone or hydrocarbons) or a solvent as control, and recorded which side the spiders selected at different times up to 24 hours. Because pheromones indicate ant presence to colony members much more precisely than hydrocarbons, we expected to find a stronger effect of pheromones than of hydrocarbons. However, we found that spiders did not react to ant pheromones, but consistently avoided cuticular hydrocarbons over time. This result suggests that the longer persistence, higher concentrations and broad generality among ants of CHC compared to the species-specific and highly volatile trail pheromone of L. niger makes cuticular hydrocarbons more reliable indicators of a risky encounter. We thus worked exclusively with cuticular hydrocarbons in the subsequent experiments, where we investigated how body condition and previous encounters with ants influence avoidance behaviour and feeding rate in spiders when exposed to ant cues. Interestingly, neither factor influenced spider avoidance of cuticular hydrocarbons, whereas both factors interacted in shaping the rates of consumed collembolans in spiders in contact with ant cues. Our results indicate that the presence of ants induces strong behavioural changes in spiders that can cascade down to lower trophic levels, which in turn depend on an interplay between species traits and environmental factors. Therefore, ant chemical cues have the potential to shape community composition through changes in behaviour of other predators and should thus be taken into account to understand food-web dynamics.
Publications
- (2020) Non-consumptive effects between predators depend on the foraging mode of intraguild prey. Journal of Animal Ecology, 89, 1690-1700
Mestre, L., Narimanov, N., Menzel, F. & Entling, M.H.
(See online at https://doi.org/10.1111/1365-2656.13224)
