Understanding the relative importance of multiple processes on structuring species interactions within communities is one of the major challenges in ecology. Here, we evaluated the relative importance of species abundance and forbidden links in structuring a hummingbird-plant interaction network from the Atlantic rainforest in Brazil. Our results show that models incorporating phenological overlapping and morphological matches were more accurate in predicting the observed interactions than models considering species abundance. This means that forbidden links, by imposing constraints on species interactions, play a greater role than species abundance in structuring the ecological network. We also show that using the frequency of interaction as a proxy for species abundance and network metrics to describe the detailed network structure might lead to biased conclusions regarding mechanisms generating network structure. Together, our findings suggest that species abundance can be a less important driver of species interactions in communities than previously thought.

We studied 86 species of hummingbird‐pollinated flowers and their pollinators at a coastal lowland site and two highland rain forest sites in southeastern Brazil. The Jaccard index for hummingbird‐pollinated floras showed greater floristic similarity between the two highland communities than between floras of high altitudes and that of near sea level. The lowland site had slightly greater richness than either of the two highland sites. Bromeliaceae was the most important family, accounting for ca. 36 percent of the bird‐pollinated species and comprising 33 percent of the richest flowers. The plant populations bloomed annually for the most part. Both the wet lowland and the dry highland floras showed flowering seasonality, whereas the cloud highland flora had aseasonal flowering. A hermit hummingbird was the major pollinator in the lowland community. A hermit and a non‐hermit hummingbird shared most of the floral resources in the two highland communities. Hermit hummingbirds are the major pollinators of hummingbird‐pollinated floras in the Atlantic rain forests of southeastern Brazil. These areas may be as rich as Central and other South American areas in hummingbird‐pollinated flowers, and altogether present an older development history when compared to the North American flora.

Large-scale geographical patterns of biotic specialization and the underlying drivers are poorly understood, but it is widely believed that climate plays an important role in determining specialization. As climate-driven range dynamics should diminish local adaptations and favor generalization, one hypothesis is that contemporary biotic specialization is determined by the degree of past climatic instability, primarily Quaternary climate-change velocity. Other prominent hypotheses predict that either contemporary climate or species richness affect biotic specialization. To gain insight into geographical patterns of contemporary biotic specialization and its drivers, we use network analysis to determine the degree of specialization in plant-hummingbird mutualistic networks sampled at 31 localities, spanning a wide range of climate regimes across the Americas. We found greater biotic specialization at lower latitudes, with latitude explaining 20–22% of the spatial variation in plant-hummingbird specialization. Potential drivers of specialization - contemporary climate, Quaternary climate-change velocity, and species richness - had superior explanatory power, together explaining 53–64% of the variation in specialization. Notably, our data provides empirical evidence for the hypothesized roles of species richness, contemporary precipitation and Quaternary climate-change velocity as key predictors of biotic specialization, whereas contemporary temperature and seasonality seem unimportant in determining specialization. These results suggest that both ecological and evolutionary processes at Quaternary time scales can be important in driving large-scale geographical patterns of contemporary biotic specialization, at least for co-evolved systems such as plant-hummingbird networks.

SUMMARYThat hummingbird-pollinated plants predominantly have red flowers has been known for decades, but well-investigated research studies are still rare. Preference tests have shown that hummingbirds do not have an innate preference for red colours. In addition, hummingbirds do not depend solely upon red flowers, because white-flowered hummingbird-pollinated plants are also common and temporarily abundant. Here we show that both white and red hummingbird-pollinated flowers differ from beepollinated flowers in their reflection properties for ultraviolet (UV) light. Hummingbird-pollinated red flowers are on average less UV reflective, and white hummingbird-pollinated flowers are more UV reflective than the same coloured bee-pollinated ones. In preference tests with artificial flowers, neotropical orchid bees prefer red UV-reflecting artificial flowers and white UVnonreflecting flowers over red and white flowers with the opposite UV properties. By contrast, hummingbirds showed no preference for any colour in the same tests. Plotting floral colours and test stimuli into the honeybees' perceptual colour space suggests that the less attractive colours are achromatic for bees and therefore more difficult to detect against the background. This underlying colour preference in bees might provide hummingbirds with a private niche that is not attractive to bees. Supplementary material available online at

Relationships between ornithophilous flowers and hummingbirds have been little studied in southern South America, where hummingbird species richness is low. We studied an ornithophilous flower assemblage and the hummingbird pollinators in a montane forest in southeastern Brazil. Twenty‐three native hummingbird‐pollinated plant species in 21 genera and 14 families were observed. Bromeliaceae, Fabaceae, Gesneriaceae, and Lobeliaceae are represented by more than one species within the assemblage. Flower shapes vary from narrow tube to bowl‐shape, but tubular flowers prevail. The variety of flower shapes and sizes results in diverse pollen placement on the body parts of hummingbird visitors, although pollen is deposited mostly on the bill. Sugar concentration in nectar averages 22.1%, and nectar volume per flower averages 16.9 μl. The plant populations bloom for one month to year‐round, and their flowering approaches the steady‐state pattern. Four flower subsets may be defined within the assemblage, each subset related to the bill size and foraging habits of the most frequent bird visitor. Of the six species of hummingbirds recorded at the study site, four are common and largely resident. The four hummingbirds differ in bill size, body mass, and favoured foraging sites, attributes which reflect their favoured flower subsets. One hermit and one trochiline hummingbird share most of the flower species they use, these two birds being the major pollinators within the flower assemblage. This montane forest community may be viewed as medium‐rich in ornithophilous flower species and poor in hummingbird species.

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.

Aim To investigate the association between hummingbird–plant network structure and species richness, phylogenetic signal on species' interaction pattern, insularity and historical and current climate. Location Fifty‐four communities along a c. 10,000 km latitudinal gradient across the Americas (39° N–32° S), ranging from sea level to c. 3700 m a.s.l., located on the mainland and on islands and covering a wide range of climate regimes. Methods We measured the level of specialization and modularity in mutualistic plant–hummingbird interaction networks. Using an ordinary least squares multimodel approach, we examined the influence of species richness, phylogenetic signal, insularity and current and historical climate conditions on network structure (null‐model‐corrected specialization and modularity). Results Phylogenetically related species, especially plants, showed a tendency to interact with a similar array of mutualistic partners. The spatial variation in network structure exhibited a constant association with species phylogeny (R2 = 0.18–0.19); however, network structure showed the strongest association with species richness and environmental factors (R2 = 0.20–0.44 and R2 = 0.32–0.45, respectively). Specifically, higher levels of specialization and modularity were associated with species‐rich communities and communities in which closely related hummingbirds visited distinct sets of flowering species. On the mainland, specialization was also associated with warmer temperatures and greater historical temperature stability. Main conclusions Our results confirm the results of previous macroecological studies of interaction networks which have highlighted the importance of species richness and the environment in determining network structure. Additionally, for the first time, we report an association between network structure and species phylogenetic signal at a macroecological scale, indicating that high specialization and modularity are associated with high interspecific competition among closely related hummingbirds, subdividing the floral niche. This suggests a tighter co‐evolutionary association between hummingbirds and their plants than in previously studied plant–bird mutualistic systems.

No abstract