Epiphytes are an important component of many forested ecosystems, yet our understanding of epiphyte communities lags far behind that of terrestrial-based plant communities. This discrepancy is exacerbated by the lack of a theoretical context to assess patterns in epiphyte community structure. We attempt to fill this gap by developing an analytical framework to investigate epiphyte assemblages, which we then apply to a data set on epiphyte distributions in a Panamanian rain forest. On a coarse scale, interactions between epiphyte species and host tree species can be viewed as bipartite networks, similar to pollination and seed dispersal networks. On a finer scale, epiphyte communities on individual host trees can be viewed as meta-communities, or suites of local epiphyte communities connected by dispersal. Similar analytical tools are typically employed to investigate species interaction networks and meta-communities, thus providing a unified analytical framework to investigate coarse-scale (network) and fine-scale (meta-community) patterns in epiphyte distributions. Coarse-scale analysis of the Panamanian data set showed that most epiphyte species interacted with fewer host species than expected by chance. Fine-scale analyses showed that epiphyte species richness on individual trees was lower than null model expectations. Therefore, epiphyte distributions were clumped at both scales, perhaps as a result of dispersal limitations. Scale-dependent patterns in epiphyte species composition were observed. Epiphyte-host networks showed evidence of negative co-occurrence patterns, which could arise from adaptations among epiphyte species to avoid competition for host species, while most epiphyte meta-communities were distributed at random. Application of our "meta-network" analytical framework in other locales may help to identify general patterns in the structure of epiphyte assemblages and their variation in space and time.