Finding the Missing Links: Knowledge Shortfalls in Plant–Hummingbird Interaction Networks Across the Americas

Resumen

Interactions between plants and hummingbirds form a key mutualistic pollination system underlying much of Neotropical biodiversity, and understanding their structure and variability is essential for predicting ecosystem responses to climate change and habitat loss. However, most evidence comes from local interaction networks, generating major knowledge gaps in both species’ geographic coverage (Wallacean shortfall) and documented biotic interactions (Eltonian shortfall). We conducted a continental-scale synthesis of published plant–hummingbird interaction networks across the Americas to evaluate how well current data represent geographic, environmental, taxonomic, and trait dimensions of the Trochilidae and their plant partners. The compiled dataset includes 12,658 recorded interactions involving 214 hummingbird species (~60% of the family) and 1,757 plant species. Network records are strongly concentrated in a few countries and predominantly in warm, humid, tree-covered environments, leaving large geographic regions and seasonal climates underrepresented. Taxonomic coverage is uneven, with Emerald and Hermit clades relatively well represented in published networks, while approximately 40% of all hummingbird species remain entirely unrecorded. Representation of threatened species is particularly limited, with very few Endangered or Critically Endangered species recorded. Cross-network centrality analyses show that a small subset of hummingbird and plant species repeatedly occupy central interactive roles, but these patterns are closely aligned with sampling intensity rather than geographic range size, indicating potential bias in inferred functional importance. By integrating network structure, species traits, and environmental characterization, this study provides a macroecological assessment of current knowledge on plant–hummingbird interactions. Our results identify priority gaps in species and regional coverage and highlight the need for targeted sampling to improve ecological inference and conservation planning for hummingbird pollination systems across the Americas.