We chose the best model as described above. To determine if the addition of nectar resources, nest entrance size and VCI influence ant parasitism, we used the proportion of parasitized pupaenand chose only nests from Solenopsis picea for the analysis, as this was the only parasitized species. In this case, generalized linear models were performed. We constructed 25 models to test for effects on the proportion of parasitized pupae as a function of nectar treatment, nest-entrance size treatment, VCI, number of workers, and number of pupae. Models included interactions between variables. We chose the best model as described above. We examined the role of nectar and nesting resources, as well as habitat complexity on ant colonization, reproduction, and parasitism. Nectar availability did not influence twig-nesting ant colonization, possibly indicating that carbonrich substances in the form of nectar are not a limiting resource for the overall community of twig-nesting ants in our study system. These results were unexpected, considering that most canopy ants feed on sugary resources available in the canopy in the form of plant exudates and honeydews . This could be explained by the availability of extrafloral nectar provided by trees in the genus Inga, as well as honeydew from homopteran insects, making sugary resources exceptionally available in this system. Even though this explanation might seem straightforward, the effects of sugary resources on arboreal ant communities can be variable. On the one hand, studies have suggested a strong relationship between abundance and cooccurrence of ants and sugary resources in the canopy, blueberries in containers growing specifically hemipteran honeydew and extrafloral nectar, even in systems were sugar resources in the form of honeydew are highly available , or when supplying substantial amounts of nectar in the canopy .

Similarly, sugar resources can regulate competitive interactions by allowing dominant ants to specialize on honeydew, while letting less dominant ant species feed on extrafloral nectar . On the other hand, sugar resources may minimally affect ant communities, possibly because of the lack of specificity of the arboreal ant community towards the use of extrafloral nectar, and due to the temporal variability of this resource . The fact that we did not observe an effect of sugar resources on ant colonization could mean that the ant species found in our study system do not maintain any strong relationship with sugary resources or do not experience intense competition over this resource, possibly suggesting that twig-nesting ants are mostly opportunistic when it comes to sugar resources . Alternatively, this may mean that the community of TNAs in this system has a generalized diet . Further, the high availability of nectar driven by the abundance of EFN-bearing trees in coffee systems creates a particularly carbon-rich environment where sugar is not limiting. It may also be possible that the nectar provided did not contain other important compounds such as amino acids, lipids, proteins and fatty acids, which are important for foraging decisions in addition to ant nutrition . If so, ants may not have been drawn to colonize new twigs because the nectar provided did not contain the essential compounds that are important for certain consumers . Nectar availability positively influenced ant colony size, particularly of brood, suggesting that the availability of nectar has the potential to increase TNAs colony fitness after establishment, and therefore contribute to colony success beyond the colonization stage. Only a few studies have provided robust evidence that nectar resources improve ant colony survivorship , reproduction, and growth .

For instance, colonies that feed on sugar resources in the form of EF nectar have greater body weight and produce significantly more adults and progeny . Our study found that in the case of Crematogaster spp. and S. picea, nectar availability accounts for an increase in brood number. This result provides some evidence that sugar availability becomes important once colonies have established, specifically increasing the number of workers and brood in nests. Others have found that sugar resources increase about five times the number of individuals in colonies of Cephalotes pusillus, an arboreal ant that feeds on EFNs . However, we should note that the responses to sugar resources is species-specific and could depend on specific ants’ life styles. Some species might be more responsive to amino acids, rather than to carbohydrates . For example, when offering sugary substances with different amino acid contents, the fire ant Solenopsis geminata feeds more often on the amino acid-rich substances, although S. invicta does not discriminate between substances . Even though our study only increased carbohydrates and did not account for amino acids, it is likely that our experiment accurately represented the effect of nectar resources in the community of twig-nesting ants in our system because EF nectar resources from some Inga species are hexose-dominant, meaning EF nectar is primarily composed by carbohydrates, such as sucrose, fructose and glucose . However, future studies should test for species-specific preferences to artificial and naturally-occurring nectars, and account for more complex sugars and amino-acids available in EF nectar provided by Inga species.

Our study confirms that twig-entrance size is a limiting factor for the twig-nesting ant community during colonization , and that certain ant species more frequently colonize nests with a specific entrance size. Both Crematogaster spp. and S. picea were more frequently found in twigs with the smallest entrance. We believe that the clear effect of entrance-size on overall colonization regardless of the nectar treatment is driven mostly by the fact that we provided only two distinct entrance-sizes, as opposed to a diversity of sizes, that could have allowed for a more even occupation between the smallest and the largest size and increase cavity use . Interestingly, the overall positive effect of small nest-entrance size on all life stages of Crematogaster spp. and S. picea in control sites changed in sites with nectar addition, as shown by significant interactions between nectar and size , except for workers of Crematogaster spp. . This trend showed overall more individuals in nests with a small entrance in the absence of nectar, but in the presence of nectar resources, nests with the small entrance had significantly fewer individuals than nests with a larger entrance. This effect of nectar treatment on nest-entrance size was even more dramatic for Solenopsis picea. Here, the addition of nectar completely inverted the positive effect of small size entrances observed in control sites. An explanation for this pattern is not straight forward. It is possible that when food resources are abundant , there are higher recruitment and foraging rates, which could be facilitated by using nests with largest twig-entrance size. A recent study showed that in social organisms, collective clog control is vital for colony function, thus developing the right strategy to improve traffic — for example, using nests with a large entrance size and thus increasing colony fitness — could be important for the colony . Furthermore, in sites with abundant resources, the production of more individuals to gather resources might be important, however we did not evaluate caste structure in relation to colony fitness, which could be useful to confirm whether the observed colony structure responds to requirements set by the environment . Beyond the influence of food and nesting resources on colonization and reproduction, nectar resources may also influence the interaction between ants and their parasitoids , an interaction that has been described as the “missing link” in understanding ant communities . Our study aimed at understanding the significance of nectar availability on parasitism rates of TNAs by their parasitic wasps, planting blueberries in containers by adding nectar on trees with EFNs. We expected that sites with nectar addition would have higher parasitism rates than control sites, because of increased foraging to the nectar source, increased probability of ants encountering their parasitoids, and higher parasitoid visitation and oviposition near sugar resources. In this system, it appears that the availability of both nectar and nesting resources benefits arboreal ants, but not ant parasitoids. It is possible that placing a larger number of nests could yield higher parasitism to further explore the effect of resources on parasitoid host interactions. In our study, parasitism was only explained by the number of workers in the colony, this positive association between colony density and parasitism has been shown for the Azteca sericeasur–Pseudacteon sp. system, in which attack rates of phorid flies increase at higher ant densities . Other studies have found similar density dependency with the presence of species parasitized and the number of cocoons in the nest , however in our study, the number of pupae did not improve the fit of the model.

These results potentially suggest density dependence in our system of study, although it is not clear whether the mechanism is behavioral or demographic . Furthermore, our results suggest that parasitism is not a strong force shaping the community of twig nesting ants in this system– at least for parasitoid wasps–, contrary to the strong effect of parasitoid flies on other arboreal ants . Contrary to expected, we did not find any relationship between habitat complexity and parasitism rates, in contrast to other studies that did find a positive influence of VCI on parasitism Gnamptogenys spp. and Pachycondyla spp. , as well as higher parasitism in more complex coffee agroecosystems . These results, however, do not rule out the possibility that the species of parasitoid wasps found in our study use EFNs to access their host. Further observations and experiments should be done on this regard, considering that females of Orasema have developed a highly specialized behavior that involves the use of EFNs to access their ant hosts . First, female wasps infiltrate the ant-plant interaction without being noticed, and lay single stalked eggs almost exclusively inside plant tissue nearby EFNs; after emergence, planidias become into contact with ants that use EFNs as feeding sites, from which they are vectored to the nest ; once in the nest, ants transfer planidias to their immature ant brood through feeding, and the wasp’s life cycle continues . This behavior has been described for the Orasema simulatrix and O. wayquecha species groups , Kapala , and Chalcura . However, the specific mechanism of transfer of planidia to Solenopsis picea nests still needs to be identified. Collectively, our study indicates that nectar, nest-entrance size, and vegetation complexity play an important role at different stages of the ant colony life. The benefits associated with the use of a diversity of resources, could explain why ants are exceptional predators of pests in agricultural systems. We also document that it is likely that parasitoid wasps are not a strong force shaping the community of TNAs in this system. We encourage future studies to investigate potential species-specific relationships between arboreal ants and EFN-bearing trees in carbon-rich systems, as well as the specific mechanisms and factors involved in shaping the ecology of ant parasitoids, as well as parasitoid-host interactions. Finally, we encourage future studies to continue investigating the mechanisms involved in the fascinating parasitoid-EFN-ant interaction. Coffee is cultivated in a variety of ways, with varying management intensity, both socially and ecologically . In terms of its ecological complexity, coffee cultivation ranges from the “rustic” coffee where coffee plants grow under the shade of trees in mature forests, to un-shaded “sun coffee” . This gradient could be also understood as a socioecological continuum in which multiple layers of complexity interact. A vast diversity of social groups , labor intensities , farm sizes , and extent of capital accumulation, are represented and overlap within this spectrum of ecological complexity . In this socio-ecologically diverse canvas, small shade-grown plots owned by peasant families may coexist along with the agroindustrial sector, primarily represented by large coffee plantations that also practice shade-grown agriculture for export1 . Shade-grown coffee often resembles the appearance of native forests , offering the possibility to conserve biodiversity, ecosystem services, and to improve households’ food security . Indeed, several studies since the conservation boom of the 1980’s in Latin America have addressed the ecological benefits associated with biodiversity conservation within shade-grown coffee agroecosystems. For example, pollination, pest control, reduced soil erosion, water conservation and improved scenery . In terms of coffee as a global commodity, the possibility for biodiversity conservation in shade-grown systems and the participation in alternative market niches have been part of a successful marketing strategy appealing to green and socially conscious consumers in the global north.