Forbidden links, trait matching and modularity in plant-hummingbird networks: Are specialized modules characterized by higher phenotypic floral integration?

BACKGROUND: Plant-pollinator mutualistic networks show non-random structural properties that promote species coexistence. However, these networks show high variability in the interacting species and their connections. Mismatch between plant and pollinator attributes can prevent interactions, while trait matching can enable exclusive access, promoting pollinators' niche partitioning and, ultimately, modularity. Thus, plants belonging to specialized modules should integrate their floral traits to optimize the pollination function. Herein, we aimed to analyze the biological processes involved in the structuring of plant-hummingbird networks by linking network morphological constraints, specialization, modularity and phenotypic floral integration. METHODS: We investigated the understory plant-hummingbird network of two adjacent habitats in the Lacandona rainforest of Mexico, one characterized by lowland rainforest and the other by savanna-like vegetation. We performed monthly censuses to record plant-hummingbird interactions for 2 years (2018-2020). We also took hummingbird bill measurements and floral and nectar measurements. We summarized the interactions in a bipartite matrix and estimated three network descriptors: connectance, complementary specialization (H2'), and nestedness. We also analyzed the modularity and average phenotypic floral integration index of each module. RESULTS: Both habitats showed strong differences in the plant assemblage and network dynamics but were interconnected by the same four hummingbird species, two Hermits and two Emeralds, forming a single network of interaction. The whole network showed low levels of connectance (0.35) and high specialization (H2' = 0.87). Flower morphologies ranged from generalized to specialized, but trait matching was an important network structurer. Modularity was associated with morphological specialization. The Hermits Phaethornis longirostris and P. striigularis each formed a module by themselves, and a third module was formed by the less-specialized Emeralds: Chlorestes candida and Amazilia tzacatl. The floral integration values were higher in specialized modules but not significantly higher than that formed by generalist species. CONCLUSIONS: Our findings suggest that biological processes derived from both trait matching and "forbidden" links, or nonmatched morphological attributes, might be important network drivers in tropical plant-hummingbird systems while morphological specialization plays a minor role in the phenotypic floral integration. The broad variety of corolla and bill shapes promoted niche partitioning, resulting in the modular organization of the assemblage according to morphological specialization. However, more research adding larger datasets of both the number of modules and pollination networks for a wider region is needed to conclude whether phenotypic floral integration increases with morphological specialization in plant-hummingbird systems.

Local drivers of the structure of a tropical bird-seed dispersal network.

One of the major challenges in ecology is to understand the relative importance of neutral- and niche-based processes structuring species interactions within communities. The concept of neutral-based processes posits that network structure is a result of interactions between species based on their abundance. On the other hand, niche-based processes presume that network structure is shaped by constraints to interactions. Here, we evaluated the relative importance of neutral-based process, represented by species' abundance (A) and fruit production (F) models, and niche-based process, represented by spatial overlap (S), temporal overlap (T) and morphological barrier (M) models, in shaping the structure of a bird-seed dispersal network from the Brazilian Atlantic Forest. We evaluated the ability of each model, singly or in combination, to predict the general structure [represented by connectance, nestedness (NODF), weight nestedness (WNODF), interaction evenness and complementary specialization] and microstructure of the network (i.e., the frequency of pairwise interactions). Only nestedness (both NODF and WNODF) was predicted by at least one model. NODF and WNODF were predicted by a neutral-based process (A), by a combination of niche-based processes (ST and STM) and by both neutral- and niche-based processes (AM). NODF was also predicted by F and FM model. Regarding microstructure, temporal overlap (T) was the most parsimonious model able to predict it. Our findings reveal that a combination of neutral- and niche-based processes is a good predictor of the general structure (NODF and WNODF) of the bird-seed dispersal network and a niche-based process is the best predictor of the network's microstructure.

Influences of sampling effort on detected patterns and structuring processes of a Neotropical plant-hummingbird network.

Virtually all empirical ecological interaction networks to some extent suffer from undersampling. However, how limitations imposed by sampling incompleteness affect our understanding of ecological networks is still poorly explored, which may hinder further advances in the field. Here, we use a plant-hummingbird network with unprecedented sampling effort (2716 h of focal observations) from the Atlantic Rainforest in Brazil, to investigate how sampling effort affects the description of network structure (i.e. widely used network metrics) and the relative importance of distinct processes (i.e. species abundances vs. traits) in determining the frequency of pairwise interactions. By dividing the network into time slices representing a gradient of sampling effort, we show that quantitative metrics, such as interaction evenness, specialization (H2 '), weighted nestedness (wNODF) and modularity (Q; QuanBiMo algorithm) were less biased by sampling incompleteness than binary metrics. Furthermore, the significance of some network metrics changed along the sampling effort gradient. Nevertheless, the higher importance of traits in structuring the network was apparent even with small sampling effort. Our results (i) warn against using very poorly sampled networks as this may bias our understanding of networks, both their patterns and structuring processes, (ii) encourage the use of quantitative metrics little influenced by sampling when performing spatio-temporal comparisons and (iii) indicate that in networks strongly constrained by species traits, such as plant-hummingbird networks, even small sampling is sufficient to detect their relative importance for the frequencies of interactions. Finally, we argue that similar effects of sampling are expected for other highly specialized subnetworks.