Morphological Traits and Specialization of Neotropical Flower-hummingbird Networks

© 2025 Academia Sinica, Taiwan Open Access Morphological Traits and Specialization of Neotropical Flower-hummingbird Networks Kelly Christie dos Santos Costa1, Érica Vanessa Durães de Freitas1, and Walter Santos de Araújo2,* 1Programa de Pós-Graduação em Biodiversidade Animal, Universidade Federal de Goiás, Goiânia, Goiás, Brazil. *Correspondence: E-mail: [email protected] (Costa) E-mail: [email protected] (de Freitas) 2Departamento de Biologia Geral, Centro de Ciências Biológicas e da Saúde, Universidade Estadual de Montes Claros, Montes Claros, Minas Gerais, Brazil. E-mail: [email protected] (Walter Santos de Araújo) Received 30 May 2024 / Accepted 12 January 2025 / Published 22 April 2025 Communicated by Chih-Ming Hung Biological specialization plays a central role in species coexistence. While many studies focus on hummingbird pollination, research on the effects of morphological traits of both hummingbirds and plants on the specialization of interaction networks remains scarce. In this study, we aim to address the following questions: i) does the dominance of ornithophilous plant species increase the specialization of hummingbird-plant interaction networks?; ii) do ornithophilous plants exhibit a greater diversity of interactions with hummingbirds compared to non-ornithophilous plants?; iii) do the beak size and body weight of hummingbirds influence the diversity of their interactions? Research was conducted on hummingbird-plant interactions in the Neotropical region. We investigated hummingbird-plant interactions in the Neotropical region by compiling 24 networks from the literature, comprising 1,182 interactions between 34 hummingbird species and 326 plant species. We found no effect of ornithophilous plant dominance on the structure (connectance and modularity) of the networks. However, species-level interactions were influenced by morphological attributes of both plants and hummingbirds. Interaction similarity among plant species was greater for ornithophilous plants than for nonornithophilous plants. Additionally, beak size positively influenced the degree and specialization of hummingbird interactions. Our findings demonstrate that the morphological characteristics of hummingbird and plant species directly influence the diversity of interactions in Neotropical hummingbird-plant networks and species specialization. Key words: Morphological correspondence, Ornithophily, Plant-animal interactions, Specialization, Trochilidae Citation: dos Santos Costa KC, de Freitas ÉVD , de Araújo WS. 2025. Morphological traits and specialization of neotropical flower-hummingbird networks. Zool Stud 64:02. doi:10.6620/ZS.2025.64-02. BACKGROUND Several ecological processes involve interactions among species, which can be studied through the approach of ecological complex networks (Delmas et al. 2009; Landi et al. 2018; Poisot et al. 2016). Studies involving ecological networks can be characterized in various ways; however, they are typically grouped into antagonistic and mutualistic networks (Ings et al. 2009; Landi et al. 2018). Mutualistic associations between animal pollinators and flowering plants are widely studied from a network perspective (e.g., Bascompte and Jordano 2013; Olesen et al. 2007), due to their wellrecognized ecological, evolutionary, and economic importance for the functioning and maintenance of ecological communities (Ollerton 2017; Rech et al. 2016; Ratto et al. 2018). Generally, plant-pollinator networks exhibit a nested structure where specialist species interact primarily with generalists (Bascompte et al. 2006). Despite significant advances in studies Zoological Studies 64: 2 (2025) doi:10.6620/ZS.2025.64-02 1 © 2025 Academia Sinica, Taiwan aimed at describing temporal (Dalsgaard et al. 2011) and geographical patterns (Moreira et al. 2020) of plant-pollinator networks, few studies have addressed how morphological traits of interacting species affect the structural characteristics of these networks in Neotropical ecosystems. The plant-pollinator interaction occurs through the provision of floral resources by plants to visitors that, during foraging, subsequently pollinate them (Agostini et al. 2014; Willmer 2011). Many plant species depend on animals as pollination agents and it is estimated that at least 87.5% of the world’s angiosperms are pollinated by animals (Ollerton et al. 2011). Among vertebrates that act as pollinators, birds represent one of the most diverse groups (Regan et al. 2015), with hummingbirds (Aves: Trochilidae) being the primary pollinators of approximately 15% of the plant species in the Neotropical region (Bawa 1990; Las-Casas et al. 2012). However, this process often involves the participation of multiple species with different degrees of specialization (Waser and Ollerton 2006; Bender et al. 2017; Rodríguez-Flores et al. 2019). In this sense, hummingbird species tend to visit plants with specific floral morphological characteristics, commonly referred to as ornithophilous or trochilophilous (see Fenster et al. 2004; Maglianesi et al. 2015). Angiosperms exhibit extreme diversity in their floral traits (Dafni et al. 2005), and certain characteristics may favor the attractiveness to different types of animals. Plants adapted to hummingbird pollination display morphological, structural, and phenotypic traits associated with the ornithophilous syndrome (Faegri and Pijl 1979). Among these, flowers with tubular corollas and reduced diameter, contrasting colors (e.g., orange, red, and violet), lack of scent, diurnal anthesis, high nectar production, and spatial separation of the nectar chamber from the stigmas and anthers are notable (Castellanos et al. 2004). Additionally, trochilophilous plants feature pendant flowers favoring the hovering flight of hummingbirds (Faegri and Pijl 1979). However, plant communities in the Neotropical region exhibit flowers with a wide morphological variation (Dafni et al. 2005). These flowers may present varying degrees of morphological specialization concerning corolla length and shape (Waser and Ollerton 2006), as well as compatibility with the beak morphology of hummingbirds (Maruyama et al. 2014). Some flower plants are visited by multiple animal species, while others have morphology that restricts their use solely by hummingbirds (Maruyama et al. 2014; Strauss and Irwin 2004). Morphological characteristics of hummingbirds, such as beak length and body mass, can directly reflect on their success in resource acquisition (Rico-Guevara et al. 2019), as well as on the foraging strategies they employ (Mendonça and Anjos 2005). However, while body mass is of great importance in hummingbird-plant interactions (Araya-Salas et al. 2018), it is more closely associated with behavioral dominance systems among hummingbirds (e.g., Marquez-Luna et al. 2019). Thus, larger-sized hummingbird species with greater body mass tend to be dominant over smaller hummingbirds, restricting their access to defended flowers (Claudino et al. 2021). On the other hand, some studies indicate that beak length is the most important variable in explaining interaction frequency and specialization in hummingbird-plant networks (e.g., Maglianesi et al. 2014; Claudino et al. 2021). This is because beak size is directly related to morphological fit with the corolla of flowers, causing hummingbirds with different beak sizes to also use distinct floral resources (Brown and Bowers 1985; Machado 2009). Based on ecological and behavioral observations, a certain level of morphological fit between hummingbirds and plants is expected (Castellanos et al. 2004). However, morphological and phenotypic incongruities can restrict the type, number, and strength of interactions exerted by a particular species (Junker et al. 2013; Stang et al. 2009). Additionally, factors such as seasonal resource availability can make hummingbirds versatile in their foraging, also exploiting non-ornithophilous plant species during times of food scarcity (Machado 2009). The inclusion of non-ornithophilous plants in their diet can directly influence the specialization of pollinator communities and, consequently, the formation of modules (i.e., subsets of species) in the interaction networks. Since species with more specialized connections tend to form groups that interact with each other (Olesen et al. 2007). Although many studies address pollination performed by hummingbirds, including from the perspective of hummingbird-plant interaction networks (e.g., Vizentin-Bugoni et al. 2014), studies focusing on the effects of morphological traits of plants and hummingbirds on network topology are scarce. We characterized the hummingbird-plant networks using the topological descriptors at the network level and at the species level (Dormann et al. 2009). At the network level, we used network connectance that is a descriptor of the level of connectivity (i.e., specialization) among these species (Antoniazzi et al. 2018), and the network modularity which is a measure of the modular arrangement of interactions between species within the network (i.e., occurrence of specialized subsets of interacting animals and plants) (Olesen et al. 2007). At the species level, we used the descriptors degree, specialization, and interaction similarity, which measure the diversity, specificity, and sharing of interactions page 2 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan between species, respectively. In this context, this study aims to answer the following questions: i) does the dominance of ornithophilous plant species increase the specialization of interactions in hummingbird-plant networks?; ii) do ornithophilous plants exhibit greater diversity of interactions with hummingbirds than nonornithophilous flowers?; iii) does hummingbird beak length and body mass influence the diversity of their interactions? We expect that: i) high proportions of ornithophilous species positively influence the structure of interaction networks by increasing their connectivity and decreasing their modularity; ii) ornithophilous plant species have greater diversity and specialization of interactions within networks compared to nonornithophilous species; iii) hummingbirds with larger beaks are more specialized compared to hummingbirds with smaller or intermediate beak sizes. On the other hand, hummingbirds with greater body mass will have a greater number of interactions within networks. MATERIALS AND METHODS Data collection We used the database from the study by Moreira et al. (2020), which compiled 28 hummingbird-plant interaction networks distributed in the Neotropical region. The networks included in the database were based on studies that met the following criteria: (1) presentation of a basic description of the study area, containing a geographic coordinate; (2) listing of hummingbird species recorded on each plant species; (3) at least five plant species and five hummingbird species listed, totaling at least 10 species; and (4) a minimum of 80% of the hummingbird identified at the species level. For the taxonomic classification of plant species, we used the database of Flora e Funga do Brasil 2020 (https://floradobrasil.jbrj.gov.br/). For hummingbird species, the scientific nomenclature follows the arrangement proposed by the Brazilian Committee of Ornithological Records (Pacheco et al. 2021). For further details regarding the compilation of hummingbird-plant interactions see Moreira et al. (2020). Defining the morphological traits of species All recorded plant species in our database were categorized in two categories: ornithophilous plants and non-ornithophilous plants. For this, we used the database available in Rodríguez-Flores et al. (2019), which characterized botanical families according to the floral morphology of their species. Thus, for botanical families with well-defined floral morphology, the species were easily categorized into ornithophilous and non-ornithophilous plants following the general characteristic of the family (Rodríguez-Flores et al. 2019). However, there are some botanical families with species that have variable floral morphology, being these categorized as “intermediate” by Rodríguez-Flores et al. (2019). For our plant species belonging to families with ‘intermediate’ floral morphology, we conducted additional searches in the literature for studies related to the description of the species of interest. Thus, we consulted the literature to define the floral morphology of the species, considering the corolla shape (tubular, bell-shaped, etc.), corolla size (in centimeters) and corolla color of the plant species. To determine corolla color, flowers were divided into four color categories: white (including all white or pale flowers); yellow (including different shades of yellow); warm colors (including all orange, red, and pink/salmon flowers); and cool colors (including all blue and purple flowers) (Carvalheiro et al. 2014). Flowers with more than one color were classified according to the predominant color (see Carvalheiro et al. 2014). Based on the determined characteristics, it was possible to categorize the different types of flowers of plant species, from flowers with ornithophilous syndrome (i.e., flowers with tubular corollas, larger in size, and with contrasting colors such as orange, red, and violet) to those considered entomophilous – non-ornithophilous plants (Castellanos et al. 2004). For plant species identified at the genus level, it was not possible to accurately determine the pollination syndrome, and such species were excluded from the networks. Networks with more than 5% of plant species for which ornithophilous and non-ornithophilous categorization was not possible were not included in our analyses. Based on these criteria, only 24 out of the hummingbird-plant networks compiled by Moreira et al. (2020), were considered in this study (Fig. 1; Table S1). Additionally, we also estimated the beak size (mm) and body weight (g) for hummingbird species. Beak size is typically measured as the length of the beak from the tip to the base and body weight is commonly measured using a precision scale capable of accurately measuring small weights. We obtained the mean values of these measurements from information available in the literature (e.g., Grantsau 1998). These measurements provide crucial data for understanding morphological variations among hummingbird species and their potential effects on ecological interactions (Araya-Salas et al. 2018; Claudino et al. 2021). page 3 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan Network analyses From the compiled interaction data, we constructed binary matrices with hummingbird species i represented in the columns and plant species j in the rows. The resulting interactions from these matrices yield graphical representations, as species at the same trophic level do not interact with each other (Bascompte and Jordano 2006). Since we evaluated networks extracted from different studies, only presence-absence data could be analyzed. To describe the structure of hummingbirdplant networks, we used the descriptors connectance (C) and modularity (M). Connectance is the proportion of possible interactions that are realized in the network, being usually used to describe the specialization of qualitative bipartite networks because the higher the connectivity, the lower the specialization of the networks (e.g., Rodríguez-Flores et al. 2019). To calculate the network modularity, we used the bipartite modularity index Q (Barber 2007) through the LPAb+ algorithm to detect modules present in the networks (Beckett 2016). For the calculation of these descriptors, we used the bipartite package (Dormann et al. 2008) in the R software (R Core Team 2024). To characterize species-level interactions for hummingbird-plant networks, we calculated the degree (k), specialization (d'), and interaction similarity for each plant species and each hummingbird species present in the matrices. The degree of a species is a measure related to the number of species with which a given species interacts. The d' index is a robust measure of specialization that compares the observed frequency distribution of interactions of a species to the availability of interacting partners (Blüthgen et al. 2006). Additionally, the d' index varies from 'one' for a completely specialized species to 'zero' for a completely generalist species (Blüthgen et al. 2006). The similarity index was used to quantify the similarity between species interactions, for which Jaccard similarity (1 - Jaccard dissimilarity, ranging from 0 to 1) was calculated for plant species and hummingbird species. The bipartite package (Dormann et al. 2008) will be Fig. 1. Distribution of the 24 flower-hummingbird networks analyzed in the study. At this map scale, the overlay networks have been enlarged for better visualization. N page 4 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan used for index calculations. Data analyses To measure the dominance of ornithophilous plant species in the networks, we used the proportion of ornithophilous species relative to the total number of plant species in each network. We employed Generalized Linear Models (GLMs) with Gaussian error distribution (for data with normal distribution) to test the effect of the proportion of ornithophilous plants on the connectivity and modularity of the networks. In these models, the size of the networks (i.e., the total number of interactions between hummingbirds and plants) was used to control for potential effects of species richness on network topology. All statistical analyses were conducted using the R statistical software (R Core Team 2024). To analyze whether network descriptors at the species level (degree, specialization, and similarity) differ between groups of ornithophilous and nonornithophilous plants, Generalized Linear Mixed Models (GLMMs) were employed. In these models, the plant species was used as a random effect variable to control for potential intrinsic differences between species that may affect hummingbird-plant interactions. GLMMs were also used to test the effect of beak size (mm) and body weight (g) on the degree, specialization, and similarity of hummingbird species. In these models, the hummingbird species was used as a random effect variable. For this analysis, were used only hummingbird species for which reliable morphometric data were obtained. All GLMMs were constructed using the lme4 package (Bates et al. 2015). RESULTS In total, the 24 analyzed networks were composed of 34 hummingbird species, 311 plant species, and 1,028 distinct interactions (Table S2). The most frequent hummingbird species in the database were Chionomesa fimbriata, represented in 62.5% of the compiled networks and Chlorostilbon lucidus and Eupetomena macroura present in 54.2% of the networks each. The hummingbird species that interacted with the highest number of plant species were Thalurania glaucopis (n = 139), Phaethornis eurynome (n = 125), and Chlorostilbon lucidus (n = 104). The largest number of plant species recorded in our study (199 species or 63.78%) belonged to the group of ornithophilous plants. Meanwhile, 34.29% (107 species) of species were determined as non-ornithophilous, and 1.92% (six species) were categorized as undetermined. The connectance of the hummingbird-plant networks ranged from 0.21 to 0.57 (mean 0.33 ± SD 0.09). Meanwhile, the modularity of the networks ranged from 0.16 to 0.55, with an average value of 0.36 (± 0.10). There was no effect of the proportion of ornithophilous species on the connectance and modularity of the networks (Table 1). Similarly, the connectance and modularity of the networks were not affected by network size. The similarity of interactions among plant species differs significantly between groups of ornithophilous and non-ornithophilous plant species (χ2 = 7.49, p = 0.006; Table 2). We found higher interaction similarity for ornithophilous plant species compared to nonornithophilous ones (Fig. 2). However, no differences were observed in the degree and specialization of interactions between ornithophilous and nonornithophilous plant species. Our results also show that there are effects of hummingbird body structure on their interactions (Table 3). Beak size positively influenced both the degree (χ2 = 4.086, p = 0.043; Fig. 3) and specialization (χ2 = 14.58, p < 0.001; Fig. 4) of hummingbird species interactions. On the other hand, beak size did not affect interaction similarity. Meanwhile, body weight did not influence any of the analyzed structural parameters. DISCUSSION We did not find an effect of the dominance of ornithophilous plants on network-level topological Table 1. Results of the models (GLMs) showing the effects of the proportion of ornithophilous plant species (%) and network size on the topological descriptors (connectance and modularity) of Neotropical hummingbird-plant networks Response variables Explanatory variables d.f. Sum. Sq. Mean. Sq. F P Network connectance Proportion of ornithophilous plant species 1 0.002 0.002 0.307 0.584 Network size 1 0.012 0.012 1.852 0.187 Network modularity Proportion of ornithophilous plant species 1 0.012 0.012 1.678 0.209 Network size 1 0.015 0.015 2.130 0.159 page 5 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan descriptors (connectance and modularity), but our results show that both plant and hummingbird structural characteristics affect interaction descriptors at the species level in hummingbird-plant networks. For example, the similarity of interactions among plant species differed significantly between plant groups, with ornithophilous plant species showing higher interaction similarity. Additionally, our results show that hummingbird beak length influences specialization, with beak size positively affecting both the degree and specialization of hummingbird species interactions. These results suggest that morphological traits of both plants and hummingbirds can affect the specialization of interactions among these species. The absence of an effect of ornithophilous plant dominance on network topology may be attributed to the high specialization (i.e., low values of connectance and modularity) observed in these networks. This implies that, whether plant communities are dominated by ornithophilous species or not, the networks have low connectivity in the interactions (e.g., Maglianesi et al. 2014; Claudino et al. 2021). The lack of effect regarding the dominance of ornithophilous plants on modularity may be related to the specialization of the Table 2. Results of models (GLMMs) evaluating the effects of plant groups (ornithophilous plants and nonornithophilous plants) on the response variables (degree, specialization, and similarity) of plant species in Neotropical hummingbird-plant networks. The chi-square and P values represent the regression coefficients of the overall model Response variables Model Parameters Degree Randon effects Groups Variance Std.Dev. Plant species 0.896 0.946 Residuals 2.663 1.632 Fixed effects Explanatory variables Chi-square p Plant group 1.131 0.288 Specialization (d’)Randon effects Groups Variance Std.Dev. Plant species 0.001 0.024 Residuals 0.019 0.136 Fixed effects Explanatory variables Chi-square p Plant group 0.004 0.949 Similarity Randon effects Groups Variance Std.Dev. Plant species 0.008 0.089 Residuals 0.030 0.173 Fixed effects Explanatory variables Chi-square p Plant group 7.490 0.006** Fig. 2. Comparison of interaction similarity between groups of ornithophilous and non-ornithophilous plant species in Neotropical hummingbird-plant networks. Fig. 3. Effect of beak size (mm) on the degree of hummingbird species in Neotropical hummingbird-plant networks. page 6 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan ornithophilous plant group, resulting in high interaction similarity across the network. Thus, these intrinsic characteristics of the networks indicate that they have low connectance and low modularity regardless of the proportion of ornithophilous species in the networks. Ornithophilous and non-ornithophilous plants differed significantly in their interaction similarity, with ornithophilous plants sharing a greater number of interactions (i.e., exhibiting higher similarity). According to Santamaría and Rodríguez-Gironés (2007), complementary traits directly affect species interaction. Thus, the set of traits present in the ornithophilous plant group (e.g., diurnal anthesis, high nectar concentrations, flower color, corolla length, and curvature, among others) (Castellanos et al. 2004), which enhance attractiveness to hummingbirds, results in greater visitation frequency by pollinators (Forister et al. 2012). Therefore, hummingbirds tend to visit species of plants that are more morphologically similar, which can optimize resource use efficiency and/or reduce competition (Stiles 1981). Conversely, the lower similarity for non-ornithophilous plants may be related to the variation in functional and morphological traits of plants primarily adapted for insect pollination but occasionally visited by hummingbirds. Our results indicate that certain morphological traits of hummingbirds (e.g., beak length) drive higher specialization and a greater number of interactions, consistent with findings from other studies (e.g., Maglianesi et al. 2014; Claudino et al. 2021). The increased number of interactions among hummingbirds with longer beaks is likely a result of morphological adaptation between the birds' beaks and the flower corolla, as species with longer beaks can interact both with small and larger corollas (Maglianesi et al. 2014). This trait enhances resource use efficiency, allowing hummingbirds to access nectar with less difficulty (Temeles et al. 2009). However, the lack of effect of body mass on the analyzed parameters is consistent with findings in the literature (e.g., Lopez-Segoviano et al. 2018; Marquez-Luna et al. 2019). The body mass of Table 3. Results of models (GLMMs) evaluating the effects of beak size (mm) and body weight (g) on the response variables (degree, specialization, and similarity) of hummingbird species in Neotropical hummingbird-plant networks. The chi-square and P values represent the regression coefficients of the overall model Response variables Models Parameters Degree Randon effects Groups Variance Std.Dev. Hummingbird species 0.841 0.917 Residuals 62.460 7.903 Fixed effects Explanatory variables Chi-square p Beak size (mm) 4.086 0.043* Body weight (g) 0.173 0.678 Specialization (d’)Randon effects Groups Variance Std.Dev. Hummingbird species 0.000 0.012 Residuals 0.024 0.156 Fixed effects Explanatory variables Chi-square p Beak size (mm) 14.587 < 0.001*** Body weight (g) 0.166 0.684 Similarity Randon effects Groups Variance Std.Dev. Hummingbird species 0.012 0.111 Residuals 0.040 0.201 Fixed effects Explanatory variables Chi-square p Beak size (mm) 2.194 0.139 Body weight (g) 0.202 0.653 Fig. 4. Effect of beak size (mm) on the specialization of hummingbird species in Neotropical hummingbird-plant networks. page 7 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan hummingbirds is related to their dominance hierarchy, where larger hummingbirds tend to dominate, excluding smaller species from high-quality energy resources (Marquez-Luna et al. 2019). However, behavioral dynamics were not assessed in this study. The largest number of hummingbird species compiled in our study belongs to the genus Chionomesa. It is worth noting that the genus Chionomesa was recently readopted to group the sister species Chionomesa fimbriata and Chionomesa lactea (see Pacheco et al. 2021), so the discussion presented here is based on publications about the former genus (Amazilia). This genus is composed of hummingbird species with a wide distribution in the Neotropical region. Studies indicate that species in the genus Chionomesa exhibit broad dietary and environmental plasticity, being capable of utilizing a diverse array of floral resources (Feinsinger 1976) and responding favorably to environmental changes and the presence of new resources. The species Thalurania glaucopis interacted with the highest number of plants in the study. This result may be related to the territorial behavior of the species as indicated by Machado and Semir (2006), in a study conducted in Atlantic Forest areas. Among the hummingbird species interacting with the highest number of plant species, the Chlorostilbon lucidus is a species with a wide geographical distribution and diversified diet, considered highly generalist regarding the resources they exploit (Machado 2009). CONCLUSIONS Our findings demonstrate that species-level interaction diversity tends to be more affected by morphological characteristics of plants and hummingbirds than topological descriptors at the network level. Thus, we show that the group of ornithophilous plants, sharing similar morphological traits, exhibits more ecologically similar interactions compared to plants with more variable morphology (non-ornithophilous plants). Additionally, our findings indicate that morphological variation among hummingbird species influences patterns of ecological specialization in the Neotropical region. Specifically, beak size was shown to be the most important trait influencing resource use efficiency, as it had a positive effect on the number and diversity of hummingbird interactions. This study represents the first systematic investigation evaluating the effects of plant characteristics on the specialization of Neotropical hummingbird-plant networks. Studies like this provide important insights into the functional factors shaping plant-pollinator networks. Acknowledgments: The authors are thankful to C.S. Souza, A. Bispo and an anonymous reviewer for their valuable suggestions on the manuscript; and, the Programa de Pós-Graduação em Biodiversidade Animal and Universidade Federal de Goiás for providing logistical support. WSA thanks to CNPq (308928/20229) and FAPEMIG (APQ-00394-18; APQ-03236-22) for the financial support. Authors’ contributions: KCSC and WSA conceived and planned the study; KCSC compiled the database; KCSC, EVDF, and WSA performed data analyses; and KCSC, EVDF, and WSA wrote the manuscript. Competing interests: KCSC, EVDF, and WSA declare that they have no conflict of interest. Availability of data and materials: All of the authors agree with the publication of the data (Supplementary materials (Tables S1 and S2)). Consent for publication: All of the authors agreed to publish the paper. Ethics approval consent to participate: Not applicable. REFERENCES Agostini K, Lopes AV, Machado IC. 2014. Recursos florais. In: Rech AR, Agostini K, Oliveira PE, Machado IC (eds). Biologia da polinização, 1st ed. Editora Projeto Cultural, Rio de Janeiro. Antoniazzi R, Dáttilo W, Rico-Gray V. 2018. A useful guide of main indices and software used for ecological networks studies. In: Dáttilo W, Rico-Gray V (eds) Ecological networks in the tropics, 1st ed. Springer, Cham. doi:10.1007/978-3-319-68228-0_13. Araya-Salas M, Gonzalez-Gomez P, Wojczulanis-Jakubas K, Lopez V, Wright TF. 2018. Spatial memory is as important as weapon and body size for territorial ownership in a lekking hummingbird. Sci Rep 8:1–11. doi:10.1038/s41598-018-20441-x. Barber MJ. 2007. Modularity and community detection in bipartite networks. Phys Rev E 76:066102. doi:10.1103/ PhysRevE.76.066102. Bascompte J, Jordano P. 2006. The structure of plant-animal mutualistic networks. In: Pascual M, Dunne J (eds) Ecological Networks. Oxford University Press, Oxford. Bascompte J, Jordano P. 2013. Mutualistic networks. Princeton University Press, Princeton, USA. Bascompte J, Jordano P, Olesen JM. 2006. Asymmetric coevolutionary networks facilitate biodiversity maintenance. Science 3:431–433. doi:10.1126/science.1123412. Bates D, Machler M, Bolker B, Walker S. 2015. Fitting Linear MixedEffects Models Using lme4. J Stat Softw 67:1–48. doi:10.18637/ jss.v067.i01. Bawa KS. 1990. Plant-pollinator interactions in tropical rain forests. Annu Rev Ecol Evol Syst 21:99–422. doi:10.1146/annurev. es.21.110190.002151. page 8 of 10Zoological Studies 64: 2 (2025) © 2025 Academia Sinica, Taiwan Beckett SJ. 2016. Improved community detection in weighted bipartite networks. R Soc Open Sci 3:40536. doi:10.1098/ rsos.140536. Bender IMA, Kissling WD, Böhning-Gaese K, Hensen I, Kühn I et al. 2017. Functionally specialized birds respond flexibly to seasonal changes in fruit availability. J Anim Ecol 86:800–811. doi:10.1111/1365-2656.12683. Blüthgen N, Menzel F, Blüthgen N. 2006. Measuring specialization in species interaction networks. BMC Ecol 6:9. doi:10.1186/14726785-6-9. Brown JH, Bowers MA. 1985. Community organization in hummingbirds: relationships between morphology and ecology. The Auk 102:251–269. doi:10.2307/4086767. Castellanos MC, Wilson P, Thomson JD. 2004. ‘Anti-bee’ and ‘probird’ changes during the evolution of hummingbird pollination in Penstemon flowers. J Evol Biol 17:876–885. doi:10.1111/j.14209101.2004.00729.x. Carvalheiro LG, Biesmeijer JC, Benadi G, Fründ J, Stang M et al. 2014. The potential for indirect effects between co-flowering plants via shared pollinators depends on resource abundance, accessibility and relatedness. Ecol Lett 17:1389–1399. doi:10.1111/ele.12342. Claudino RM, Antonini Y, Martins C, Beirão MV, Braga EM et al. 2021. Is bigger always better? Neither body size nor aggressive behavior are good predictors to measure the degree of specialization of hummingbird interaction networks in rocky outcrops. BioRXive 1:1–30. doi:10.1101/2021.02.27.433160. Dafni A, Kevan PG, Husband BC. 2005. Practical pollination biology. Environquest Ltd, Ontario, Canada. Dalsgaard B, Magård E, Fjeldså J, González AMM, Rahbek C, Olesen JM, Ollerton J, Alarcón R, Araujo AC, Cotton PA, Lara C, Machado CG, Sazima I, Sazima M, Timmermann A, Watts S, Sandel B, Sutherland WJ, Jens-Christian S. 2011. Specialization in plant-hummingbird networks is associated with species richness, contemporary precipitation and quaternary climatechange velocity. PLoS ONE 6:E25891. doi:10.1371/journal. pone.0025891. Delmas E, Besson M. Brice M-H, Burkle LA, Dalla Riva GV et al. 2009. Ecological Networks - Beyond Food Webs. J Anim Ecol 78:253–265. doi:10.1111/j.1365-2656.2008.01460.x. Dormann CF, Gruber B, Fruend J. 2008. Introducing The Bipartite Package: Analysing Ecological Networks. R News 8:8–11. Dormann CF, Fründ J, Blüthgen N, Gruber B. 2009. Indices, graphs and null models: analyzing bipartite ecological networks. Open J Ecol 2:7–24. doi:10.2174/1874213000902010007. Faegri K, Pijl VDL. 1979. The Principles of Pollination Ecology. Pergamon Press, London, UK. Feinsinger P. 1976. Organization of a Tropical Guild of Nectarivorous Birds. Ecol Monogr 46:257–291. doi:10.2307/1942255. Fenster CB, Armbruster WS, Wilson P, Dudash MR, Thomson JD. 2004. Pollination syndromes and floral specialization. Annu Rev Ecol Evol Syst 35:375–403. doi:10.1146/annurev. ecolsys.34.011802.132347. Forister ML, Dyer LA, Singer MS, Stireman JO, Lill JT. 2012. Revisiting the evolution of ecological specialization, with emphasis on insect–plant interactions. Ecology 93:981–991. doi:10.1890/11-0650.1. Ings TC, Montoya JM, Bascompte J, Blüthgen N, Brown L, Dormann CF, Woodward G. 2009. Review: Ecological networks - beyond food webs. J Anim Ecol 78:253–269. doi:10.1111/j.13652656.2008.01460.x. Junker RR, Blüthgen N, Brehm T, Binkenstein J, Paulus J et al. 2013. Specialization on traits as basis for the niche-breadth of flower visitors and as structuring mechanism of ecological networks. Funct Ecol 27:329–341. doi:10.1111/1365-2435.12005. Landi P, Minoarivelo HO, Brännström Â, Hui C, Dieckmann U. 2018. Complexity and stability of ecological networks: a review of the theory. Pop Ecol 60:319–345. doi:10.1007/s10144-018-0628-3. Las-Casas FMG, Azevedo JSM, Dias FMM. 2012. The community of hummingbirds (Aves: Trochilidae) and the assemblage of flowers in a Caatinga vegetation. Braz J Biol 72:51–58. doi:10.1590/ S1519-69842012000100006. Lopez-Segoviano G, Arenas-Navarro M, Vega E, Arizmend C, Arizmend M. 2018. Hummingbird migration and flowering synchrony in the temperate forests of northwestern Mexico. PeerJ 6:e5131. doi:10.7717/peerj.5131. Machado CG. 2009. Beija-flores (Aves: Trochilidae) e seus recursos florais em uma área de Caatinga da Chapada Diamantina, Bahia, Brasil. Zoologia 26:255–265. doi:10.1590/S198446702009000200008. Machado CG, Semir J. 2006. Fenologia da floração e biologia floral de bromeliáceas ornitófilas de uma área da Mata Atlântica do Sudeste brasileiro. Rev Bras Bot 29:163–174. doi:10.1590/ S0100-84042006000100014. Maglianesi MA, Blüthgen N, Böhning-Gaese K, Schleuning M. 2014. Morphological traits determine specialization and resource use in mutualistic networks. Ecology 95:877–885. doi:10.1890/132261.1. Maglianesi MA, Böhning‐gaese K, Schleuning M. 2015. Different foraging preferences of hummingbirds on artificial and natural flowers reveal mechanisms structuring plant–pollinator interactions. J Anim Ecol 84:655–664. doi:10.1111/13652656.12319. Marquez-Luna U, Corcuera PL, Valverde PL. 2019. Factors affecting the dominance hierarchy dynamics in a hummingbird assemblage. Curr Zool 65:261–268. doi:10.1093/cz/zoy057. Maruyama PK, Vizentin-Bugoni J, Oliveira GM, Oliveira PE, Dalsgaard B. 2014. Morphological and Spatio-Temporal Mismatches Shape a Neotropical Savanna Plant-Hummingbird Network. Biotropica 46:740–747. doi:10.1111/btp.12170. Mendonça LB, Anjos L. 2005. Hummingbirds (Aves, Trochilidae) and their flowers in an urban area of southern Brazil. Rev Bras Zool 22:257–291. doi:10.1590/S0101-81752005000100007. Moreira LT, Falcão LAD, Araújo WSA. 2020. Geographical patterns in the architecture of neotropical flower-visitor networks of hummingbirds and insects. Zool Stud 59:50. doi:10.6620/ ZS.2020.59-50. Olesen JM, Bascompte J, Dupont YL, Jordano P. 2007. The modularity of pollination networks. Proc Natl Acad Sci USA 11:19891–19896. doi:10.1073/pnas.0706375104. Ollerton J. 2017. Pollinator diversity: distribution, ecological function, and conservation. Annu Rev Ecol Evol Syst 48:353–376. doi:10.1146/annurev-ecolsys-110316-022919. Ollerton J, Winfree R, Tarrant S. 2011. How many flowering plants are pollinated by animals? Oikos 120:321–326. doi:10.1111/ j.1600-0706.2010.18644.x. Pacheco JF, Silveira LF, Aleixo A. Agne CE, Bencke et al. 2021. Annotated checklist of the birds of Brazil by the Brazilian Ornithological Records Committee—second edition. Ornithol Res 29:94–105. doi:10.1007/s43388-021-00058-x. Poisot T, Stouffer DB, Kéfi S. 2016. Describe, understand and predict: why do we need networks in ecology? Funct Ecol 30:1878– 1882. doi:10.1111/1365-2435.12799. R Core Team. 2024. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Ratto F, Simmons BI, Spake R, Zamora-Gutierrez V, MacDonald MA et al. 2018. Global importance of vertebrate pollinators for plant reproductive success: a meta-analysis. Front Ecol Environ 16:82–90. doi:10.1002/fee.1763. page 9 of 10Zoological Studies 64: 2 (2025)