Disentangling factors that assemble New Zealand's ant communities

<p>Several biotic and abiotic stressors can influence community assembly. The negative co-occurrence patterns observed within many communities, for example, may derive either from behavioural similarities (e.g. species displaying high aggression levels towards each other) or habitat preference. I evaluated the role of several stressors that may shape New Zealand’s ant communities. First, I investigated (in chapter 2) the co-occurrence patterns of two native ant communities located within transitional grassland-forest habitats. I also monitored the temperature variation in these habitats over a one-year period. I found that grasslands are exposed to higher temperature variation than forest habitats. I also found that some ants are mostly associated with forest habitats and others with grasslands. Using null models to examine these communities, I found evidence that two ant species (Monomorium antarcticum and Prolasius advenus) exhibit negative co-occurrence patterns. In the reminder of my thesis I developed a series of laboratory-based experiments to examine the processes that could explain the co-occurrence patterns that I observed in these ant communities. In chapter 3, I subjected heterospecific groups of ants to interactions in controlled conditions. I asked if interspecific aggression predict the survival probability and co-occurrence patterns described in chapter 2. My results demonstrated that aggression predicted the survival probability of interacting ant species and their co-occurrence patterns. I argued that aggressive behaviour might reflect the risks imposed by competitors. Differences in aggression may thus be a key factor influencing sympatric and allopatric co-occurrence patterns of these ant communities. In chapter 4, I tested the hypotheses that arrival sequence and diet influence the strength of interactions between colonies of two species that exhibited negative co-occurrence patterns (P. advenus and M. antarcticum). When arriving first, P. advenus displayed increased aggression and M. antarcticum a defensive reaction. The adoption of a defensive reaction by M. antarcticum increased their colony survival probability. Changes in carbohydrate and protein availability modulated colony activity rates of both species. These results indicate that arrival sequence can modulate the territorial behaviour displayed by interacting species in situations of conflict. Also, I showed that these ant species adjust their foraging activity rates in according to their diet, but different species do so differently. In chapter 5, I expanded the scope of chapter 4 and asked if aggression and foraging behaviour of P. advenus and M. antarcticum change in different conditions of temperature, diet and group size. For both ant species, changes in temperature had stronger effects on small than large colonies. Small groups of M. antarcticum displayed higher foraging activity at lower temperatures. Conversely, small groups of P. advenus displayed higher foraging activity at high temperatures. Also, small M. antarcticum colonies displayed increased aggression and significantly reduced the size of large P. advenus colonies, regardless of temperature and diet. These results suggest that P. advenus and M. antarcticum perform differently at different temperatures. Furthermore, I demonstrated that the persistence of these small colonies might be related to their ability to modulate foraging activities and interspecific aggression according to the environment. I also investigated (in chapter 6) the effects of a neurotoxic pesticide (neonicotinoid) on a native (M. antarcticum) and an invasive ant (Linepithema humile). I tested whether sublethal contamination with a neonicotinoid affects foraging, fitness and the outcome of interspecific interactions between these ants. Overall, pesticide exposure increased aggression of the invasive ant and reduced the aggression of the native species. Importantly, non-exposed individuals of the invasive species subjected to interactions against exposed natives were less aggressive, but more likely to survive. These results suggest that the modification of the physicochemical environment by pesticide contamination could change the dynamics of communities and influence invasion success. Overall, this thesis highlights that synergistic effects between several biotic and abiotic factors influence community assembly. My results suggest that non-random allopatric patterns of niche occupancy observed in these ant communities are better explained by high levels of aggression displayed between pairs of species that seldom co-occur, though I was unable to falsify the hypothesis that habitat preference also plays a role in determining their distribution and co-occurrence patterns. The modification of behaviour by external factors – either natural (e.g. temperature) or human mediated (e.g. pesticide exposure) – likely has broad effects on population and community dynamics and on patterns of species co-existence.</p>