The human touch: a meta-analysis of anthropogenic effects on plant-pollinator interaction networks.

Background: Anthropogenic activities significantly impact natural ecosystems, leading to alterations in plant and pollinator diversity and abundance. These changes often result in shifts within interacting communities, potentially reshaping the structure of plant-pollinator interaction networks. Given the escalating human footprint on habitats, evaluating the response of these networks to anthropization is critical for devising effective conservation and management strategies. Methods: We conducted a comprehensive review of the plant-pollinator network literature to assess the impact of anthropization on network structure. We assessed network metrics such as nestedness measure based on overlap and decreasing fills (NODF), network specialization (H2'), connectance (C), and modularity (Q) to understand structural changes. Employing a meta-analytical approach, we examined how anthropization activities, such as deforestation, urbanization, habitat fragmentation, agriculture, intentional fires and livestock farming, affect both plant and pollinator richness. Results: We generated a dataset for various metrics of network structure and 36 effect sizes for the meta-analysis, from 38 articles published between 2010 and 2023. Studies assessing the impact of agriculture and fragmentation were well-represented, comprising 68.4% of all studies, with networks involving interacting insects being the most studied taxa. Agriculture and fragmentation reduce nestedness and increase specialization in plant-pollinator networks, while modularity and connectance are mostly not affected. Although our meta-analysis suggests that anthropization decreases richness for both plants and pollinators, there was substantial heterogeneity in this regard among the evaluated studies. The meta-regression analyses helped us determine that the habitat fragment size where the studies were conducted was the primary variable contributing to such heterogeneity. Conclusions: The analysis of human impacts on plant-pollinator networks showed varied effects worldwide. Responses differed among network metrics, signaling nuanced impacts on structure. Activities like agriculture and fragmentation significantly changed ecosystems, reducing species richness in both pollinators and plants, highlighting network vulnerability. Regional differences stressed the need for tailored conservation. Despite insights, more research is crucial for a complete understanding of these ecological relationships.

Impacts of land use change on native plant-butterfly interaction networks from central Mexico.

Background: Land use change is a key catalyst of global biodiversity loss and ecosystem degradation. Deforestation and conversion of natural habitats to agricultural or urban areas can profoundly disrupt plant-flower visitor interactions by altering their abundances and distribution. Yet, specific studies analyzing the effects of land use change on the structure of networks of the interactions between particular groups of flower visitors and their plants are still scarce. Here, we aimed to analyze how converting native habitats affects the species composition of butterfly communities and their plants, and whether this, in turn, leads to changes in the structure of interaction networks in the modified habitats. Methods: We performed bi-monthly censuses for a year to record plant-butterfly interactions and assess species diversity across three habitat types, reflecting a land-use change gradient. From original native juniper forest to urban and agricultural zones in central Mexico, one site per land use type was surveyed. Interactions were summarized in matrices on which we calculated network descriptors: connectance, nestedness and modularity. Results: We found highest butterfly diversity in native forest, with the most unique species (i.e., species not shared with the other two sites). Agricultural and urban sites had similar diversity, yet the urban site featured more unique species. The plant species richness was highest in the urban site, and the native forest site had the lowest plant species richness, with most of the plants being unique to this site. Butterfly and plant compositions contrasted most between native forest and modified sites. Network analysis showed differences between sites in the mean number of links and interactions. The urban network surpassed agriculture and native forest networks in links, while the native forest network had more interactions than the agriculture and urban networks. Native plants had more interactions than alien species. All networks exhibited low connectance and significant nestedness and modularity, with the urban network featuring the most modules (i.e., 10 modules). Conclusions: Converting native habitats to urban or agricultural areas reshapes species composition, diversity and interaction network structure for butterfly communities and plants. The urban network showed more links and modules, suggesting intricate urban ecosystems due to diverse species, enhanced resources, and ecological niches encouraging interactions and coexistence. These findings emphasize the impacts of land use change on plant-butterfly interactions and the structure of their interaction networks.

Changes in the core species of the ant-plant network of oak forest converted to grassland: replacement of its ant functional groups.

Land-use change in terrestrial environments is one of the main threats to biodiversity. The study of ant-plant networks has increased our knowledge of the diversity of interactions and structure of these communities; however, little is known about how land-use change affects ant-plant networks. Here we determine whether the change in land use, from native oak forest to induced grassland, affected the network properties of ant-plant networks in a temperate forest in Mexico. We hypothesize that the disturbed vegetation will be more nested and generalized due to the addition of generalist species to the network. The oak forest network comprises 47 plant species and 11 ant species, while the induced grassland network has 35 and 13, respectively. Floral nectar was the resource used most intensely by the ants in both vegetation types. The ant-plant network of the induced grassland was significantly more nested and generalist than that of the oak forest; however, none of the networks were nested when considering the frequency of interaction. In both vegetation types, the ants were more specialized than the plants, and niche overlap was low. This could be related to the dominant species present in each type of vegetation: Prenolepis imparis in the oak forest and Camponotus rubrithorax in the grassland. The central core of cold climate ant species in the oak forest was replaced by a central core of subordinate Camponotini and tropical specialists in the induced grassland. These results suggest that the increase in nestedness and generalization in the grassland may be related to the loss of the cold climate specialists from the core of the oak forest network. Our findings provide evidence that land-use change increases the level of generalization in the ant-plant interaction networks of temperate forests.

Neither ant dominance nor abundance explain ant-plant network structure in Mexican temperate forests.

BACKGROUND: Ant-plant mutualistic networks tend to have a nested structure that contributes to their stability, but the ecological factors that give rise to this structure are not fully understood. Here, we evaluate whether ant abundance and dominance hierarchy determine the structure of the ant-plant networks in two types of vegetation: oak and grassland, in two temperate environments of Mexico: Flor del Bosque State Park (FBSP) and La Malinche National Park (MNP). We predicted that dominant and abundant ant species make up the core, and submissives, the periphery of the network. We also expected a higher specialization level in the ant trophic level than in plant trophic level due to competition among the ant species for the plant-derived resources. METHODS: The ant-plant interaction network was obtained from the frequency of ant-plant interactions. We calculated a dominance hierarchy index for the ants using sampling with baits and evaluated their abundance using pitfall traps. RESULTS: In MNP, the Formica spp. species complex formed the core of the network (in both the oak forest and the grassland), while in FBSP, the core species were Prenolepis imparis (oak forest) and Camponotus rubrithorax (grassland). Although these core species were dominant in their respective sites, they were not necessarily the most dominant ant species. Three of the four networks (oak forest and grassland in FBSP, and oak forest in MNP) were nested and had a higher number of plant species than ant species. Although greater specialization was observed in the ant trophic level in the two sites and vegetations, possibly due to competition with the more dominant ant species, this was not statistically significant. In three of these networks (grassland and oak forest of MNP and oak forest of FBSP), we found no correlation between the dominance hierarchy and abundance of the ant species and their position within the network. However, a positive correlation was found between the nestedness contribution value and ant dominance hierarchy in the grassland of the site FBSP, which could be due to the richer ant-plant network and higher dominance index of this community. CONCLUSIONS: Our evidence suggests that ant abundance and dominance hierarchy have little influence on network structure in temperate ecosystems, probably due to the species-poor ant-plant network and a dominance hierarchy formed only by the presence of dominant and submissive species with no intermediate dominant species between them (absence of gradient in hierarchy) in these ecosystems.

Effect of agricultural land-use change on ant dominance hierarchy and food preferences in a temperate oak forest.

BACKGROUND: The discovery-dominance trade-off is the inverse relationship between the ability of a species to discover resources and the species' dominance of those resources; a paradigm used to explain species coexistence in ant communities dependent on similar resources. However, factors such as stress (e.g., temperature) or disturbance (e.g., removal of biomass) associated with the change in land use, can modify this trade-off. Here, we aimed to determine the potential effects of land use change on dominance hierarchy, food preferences and on the discovery-dominance trade-off. METHODS: An experiment with baits was used to investigate the dominance hierarchies of ant communities in a temperate mountain habitat in central Mexico. We evaluated the dominance index (DI), food preferences and discovery-dominance trade-offs of ants inhabiting two types of vegetation: a native oak forest and agricultural land resulting from agricultural land use and grazing. RESULTS: The ant communities in both environments were comprised of three species of ants (Monomorium minimum, Myrmica mexicana, and Camponotus picipes pilosulus), four morphospecies (Pheidole sp.1 and Pheidole sp.2, Temnothorax sp. and Lasius sp.) and one genus (Formica spp.). All Formicidae showed values of intermediate to low DI, and this factor did not seem to be influenced by the change in land use. Ants in the modified vegetation (i.e., agricultural land) were found to be numerically greater. Overall, a higher number of visits were registered to the tuna bait, although the duration of foraging events to the honey baits was longer. However, foraging times were dependent on the species considered: the generalized Myrmicinae, M. minimum, the ant species with highest DI, foraged for longer periods of time in the agricultural land and on the tuna bait. Meanwhile, the cold-climate specialist Formica spp., with a lower DI, foraged for longer periods of time in the oak (although not significant) and on the honey bait. We found little evidence of the discovery-dominance trade-off; instead, we found considerable diversity in the strategies used by the different species to access resources. This range of strategies is well represented by the generalized Myrmicinae M. minimum, the cold-climate specialists Formica spp. and Temnothorax sp., and the rare species, as the cold climate specialist Lasius sp. (insinuators). CONCLUSIONS: Our evaluation shows that transformation of the original habitat does not appear to affect the hierarchical dominance of the ant communities, but it does affect their food preferences. Species with higher DI values such as the generalized Myrmicinae are more skilled at resource acquisition in modified habitats. Our results suggest that change in land use promotes an increase in the diversity of foraging strategies used by different ant species. This diversity may contribute to resource partitioning which favors coexistence.

"Pro-bird" floral traits discourage bumblebee visits to Penstemon gentianoides (Plantaginaceae), a mixed-pollinated herb.

The pollination syndrome concept implies that flowers evolved with particular sets of characteristics, such as colors, shapes, orientations, and rewards, as a means of attracting pollinators. However, these traits may have also evolved to deter unwanted visitors. The North American genus Penstemon exhibits a great floral diversity that is mainly associated with bumblebee and hummingbird pollination. Evolutionary shifts from insect pollination to hummingbird pollination have occurred in Penstemon repeatedly, but some species maintain mixed-pollination systems and intermediate floral traits between bee- and hummingbird-pollination modes. The apparently intermediate floral traits of species with mixed-pollination systems might be potentially acting to deter bumblebee foragers. Then, bird-flower traits might be selected with increased hummingbird visitation over evolutionary time might, resulting in specialization to and the evolution of floral traits present in hummingbird-pollinated species. Here, we modified bee-pollination floral traits in Penstemon gentianoides with a mixed pollination system, to resemble hummingbird-pollination traits, and measured the effects of trait modification on bumblebee foraging behavior and plant female reproductive fitness. Our results showed that reduction in the width of the corolla tube and the absence of the corolla lip negatively affects bumblebee visitation and their efficiency as pollinators, and that the synergistic interaction of both traits enhanced the "anti-bee" effect. We conclude that acquisition of floral traits that resemble those of hummingbird-pollination enables Penstemon plant species to deter bumblebee visits.

Diversity of floral visitors to sympatric Lithophragma species differing in floral morphology.

Most coevolving relationships between pairs of species are embedded in a broader multispecific interaction network. The mutualistic interaction between Lithophragma parviflorum (Saxifragaceae) and its pollinating floral parasite Greya politella (Lepidoptera, Prodoxidae) occurs in some communities as a pairwise set apart from most other interactions in those communities. In other communities, however, this pair of species occurs with congeners and with other floral visitors to Lithophragma. We analyzed local and geographic differences in the network formed by interactions between Lithophragma plants and Greya moths in communities containing two Lithophragma species, two Greya species, and floral visitors other than Greya that visit Lithophragma flowers. Our goal was to evaluate if non-Greya visitors were common, if visitor assembly differs between Lithophragma species and populations and if these visitors act as effective pollinators. Sympatric populations of L. heterophyllum and L. parviflorum differ in floral traits that may affect assemblies of floral visitors. Visitation rates by non-Greya floral visitors were low, and the asymptotic number of visitor species was less than 20 species in all populations. Lithophragma species shared some of the visitors, with visitor assemblages differing between sites more for L. heterophyllum than for L. parviflorum. Pollination efficacy experiments showed that most visitors were poor pollinators. Single visits to flowers by this assemblage of species resulted in significantly higher seed set in Lithophragma heterophyllum (30.6 +/- 3.9 SE) than in L. parviflorum (4.7 +/- 3.4 SE). This difference was consistent between sites, suggesting that these visitors provide a better fit to the floral morphology of L. heterophyllum. Overall, none of the non-Greya visitors appears to be either sufficiently common or efficient as a pollinator to impose strong selection on any of these four Lithophragma populations in comparison with Greya, which occurs within almost all populations of these species throughout their geographic ranges.