LAUSR

Plant functional traits impact the fitness and environmental niche of plants. Major plant functional types have been characterized by their trait spectrum, and the environmental and phylogenetic imprints on traits have advanced several ecological fields. Yet, very few trait data on epiphytes, which represent almost 10% of vascular plants, are available. We collated 76,561 trait observations for 2,882 species of vascular epiphytes and compared these to non‐epiphytic herbs and trees to test hypotheses related to how the epiphytic habit affects traits, and if epiphytes occupy a distinct region in the global trait space. We also compared variation in traits among major groups of epiphytes, and investigated the coordination of traits in epiphytes, ground‐rooted herbs and trees. Epiphytes differ from ground‐rooted plants mainly in traits related to water relations. Unexpectedly, we did not find lower leaf nutrient concentrations, except for nitrogen. Mean photosynthetic rates are much lower than in ground‐rooted plants and lower than expected from the nitrogen concentrations. Trait syndromes clearly distinguish epiphytes from trees and from most non‐epiphytic herbs. Among the three largest epiphytic taxa, orchids differ from bromeliads and ferns mainly by having smaller and more numerous stomata, while ferns differ from bromeliads by having thinner leaves, higher nutrient concentrations, and lower water content and water use efficiency. Trait networks differ among epiphytes, herbs and trees. While all have central nodes represented by SLA and mass‐based photosynthesis, in epiphytes, traits related to plant water relations have stronger connections, and nutrients other than potassium have weaker connections to the remainder of the trait network. Whereas stem‐specific density reflects mechanical support related to plant size in herbs and trees, in epiphytes it mostly reflects water storage and scales with leaf water content. Synthesis. Our findings advance our understanding of epiphyte ecology, but we note that currently mainly leaf traits are available. Important gaps are root, shoot and whole plant, demographic and gas exchange traits. We suggest how future research might use available data and fill data gaps.