In the present study, we assessed different factors associated with the host trees (i.e., size, isolation, exotic species, and presence of EFNs) that would influence the community of arboreal ant foraging in scattered trees in an urban landscape. We found a significant shift in the ant community composition between day and night and, accordingly, a different core of interacting species between these periods. The tree size was the only tree factor that influenced the ant diversity, and importantly, this was true only for the nocturnal ant community. In relation to ant species associations with their hosting trees, we observed high levels of network specialization and modular networks. Only the nocturnal network of ant-plant interactions followed a nested pattern, and there was no relationship of tree individual specialization to any predictive variable, i.e., size, isolation, exotic species, and presence of EFNs. This study is the first to assess the effects of different factors on communities of arboreal ants in urban ecosystems, focusing on diversity patterns as well as in the species interaction networks. Our work provides valuable insights into the conservation value of scattered trees to urban biodiversity, specifically ants.
Of all factors tested, namely tree size, isolation, native/exotic tree species, and presence of EFNs, only tree size has substantial empirical evidence related to ant diversity patterns. This result is in concordance with prior studies in preserved forests, showing an increase in arthropod diversity with tree size in the Brazilian Atlantic forest (Campos et al. 2006) and Papua New Guinea (Klimes et al. 2012). Indeed, larger trees possess a higher number of essential resources, mainly shelter (Klimes et al. 2012; Klimes 2017). Arboreal ants rely heavily on existing cavities produced by wood-boring beetles as shelter resources (Blüthgen and Feldhaar 2010) and presumably, larger trees would have a higher availability of such resources (Koch et al. 2016). Additionally, larger trees often have a large number of structures and microhabitat availability in their crowns, which might allow a higher amplitude of available nesting sites, from small cavities in twigs to large ones in the main branches. The strong influence of tree size we found in our study highlights that resource use patterns in an urban area are similar to previous findings for arboreal ants from natural ecosystems. Finally, larger trees might also support a richer community of herbivores and other insects, which are important food items for arboreal ants, either as prey (Floren et al. 2002) or by producing sugar-liquid resources, i.e., honeydew (Styrsky and Eubanks 2007). Overall, in urban ecosystems, resource partitioning and limitation might also play an essential role in defining species coexistence, especially in areas with a higher concentration of scattered trees.
Tree isolation and the presence of EFNs did not have a detectable effect on our focal ant communities. While the lack of effect of tree isolation on arboreal ant species diversity seems to contradict prior studies (e.g., Gove and Majer 2006; Powell et al. 2011), it is important to stress that in our study, all trees were relatively isolated (about 6 meters apart). Thus, it was much more a situation of comparing the distance between trees than connectivity per se. Most ant species that forage on trees also nest on the same tree, or at least in adjacent ones, rarely using the ground layer for foraging between trees (Camarota et al. 2016). Instead of increasing or decreasing the diversity, the distance might effects the species pool that use the cluster of closer than distant trees. In our study, all trees were located within a range of 35 ha, and thus, the arboreal ant species pool is likely to be the same for all sampled trees. Extrafloral nectaries are important resources for arboreal ants, and many ant species are involved in protective interactions with plant-bearing EFNs (Rico-Gray and Oliveira 2007). However, the role of EFNs in structuring ant composition at the community level is not clear and is probably much more limited than initially proposed (Camarota et al. 2015; Antoniazzi et al. 2019). Additionally, the spatial scale of the study might influence the detection of any effect of EFNs on ant communities (Ribeiro et al. 2018). Potentially, EFNs are important for ant community composition but in a much more supplementary way than as a primary food resource (Camarota et al. 2015; Pacelhe et al. 2019). Moreover, these resources are seasonal, and their importance may be much higher in the peak of EFN productivity (Lange et al. 2013, but see Camarota et al. 2015).
Interestingly, the ant species foraging on exotic trees were similar to the community foraging on native trees. Notably, urban trees support a significant number of herbivorous arthropods, which in turn influence their importance for higher trophic levels, such as predators (Tallamy 2004; Urghardt et al. 2010). Ants comprise a significant part of this higher trophic level, and presumably, the community of herbivores supported by a given tree would influence the arboreal ant community. An excess of herbivores can damage scattered trees’ important properties for urban well-being by reducing photosynthesis, tree growth, and the structural condition of the tree (Dreistadt et al. 1990; Christie and Hochuli 2005). Thus, our results on arboreal ants reassure the need for balanced management of urban trees (Frank et al. 2019), focusing on the beneficial effects of native trees in hosting native arthropods and the potential of exotic trees to perform essential ecosystem services.
We observed a significant shift between the diurnal and nocturnal ant communities and detected higher than average values of specialization at the network level for a given period, as well as a different core of interacting species. While one ant species, Cephalotes pusillus, was frequent at both periods of the day, the ant species that were part of the generalist core were completely different between day and night, with Camponotus blandus and Cephalotes pusillus during the day and Camponotus melanoticus and Camponotus atriceps at night. Interestingly, in a study in a natural habitat, at the Brazilian Cerrado, Camarota et al. (2016) found the same core of ant species. We also found that both diurnal and nocturnal ant networks were modular, indicating some specialization of the ants to a given set of trees. Similar to our results, a previous study showed that disturbance did not change the topology of plant-herbivorous networks, being modular in forest interior and anthropogenic edges (Carvalho et al. 2021). Perhaps the high values of modularity are more related to stochastic factors, as the distance between the sampled trees, than any other factor related to tree and ant specializations, such as a given food reward or nesting site. We also found that only the nocturnal ant network was significantly nested, with a cohesive subset of each tree's overall interacting species. A candidate explanation for this finding might be that nocturnal ant communities have more dominant ant species than diurnal ones, competing by valuable resources under less harsh abiotic conditions (Bestelmeyer 2008). Therefore, there would be some foraging species hierarchy, ultimately leading to a more predictable distribution of foraging ant species (Dáttilo et al. 2014b).
While the shift in ant communities between day/night communities has been extensively documented in the literature, most studies have focused on natural areas (e.g., Anjos et al. 2017; Dáttilo et al. 2014a, 2015). Urban habitats can provide exciting insights into the temporal partition of food resources. Such temporal partition might reflect either the ant species’ physiological constraints or competitive abilities (Cerda et al. 1997, 1998a; Wittman et al. 2010). Thus, different levels of urbanization can result in different temporal patterns of food resource use. For example, in trees located in more exposed environments, like less wooded areas inside the cities, desiccation risk is much higher (Menke et al. 2011). Thus, in such drier areas, eco-physiological constraints may be the main factor between temporal variation in resource use, if any. However, in scattered trees located in less exposed areas, such as our focal wooded university campus, competition between ant species might be stronger than a physiological limitation in defining the species foraging times. Further studies would be essential to disentangle the main factor associated with ant species temporal food partitioning in urban ecosystems, and scattered trees provide a useful template for such studies.