School of Biological Sciences, Victoria University of Wellington, P.O. Box 600, Wellington, New Zealand.
Ecology. 2010 Feb;91(2):377-85. doi: 10.1890/08-2004.1.
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.
附生植物是许多森林生态系统的重要组成部分，但我们对附生植物群落的理解远远落后于基于陆地的植物群落。这种差异因缺乏评估附生植物群落结构模式的理论背景而加剧。我们试图通过开发一个分析框架来填补这一空白，该框架用于研究附生植物组合，然后将其应用于巴拿马雨林中附生植物分布的数据集。在粗略的尺度上，附生植物物种和宿主树种之间的相互作用可以看作是二部网络，类似于传粉和种子扩散网络。在更精细的尺度上，单个宿主树上的附生植物群落可以看作是元群落，或者是由扩散连接的局部附生植物群落的集合。通常使用类似的分析工具来研究物种相互作用网络和元群落，从而为研究附生植物分布的粗尺度（网络）和细尺度（元群落）模式提供一个统一的分析框架。对巴拿马数据集的粗尺度分析表明，大多数附生植物物种与比预期随机情况下更少的宿主物种相互作用。细尺度分析表明，个别树木上的附生植物物种丰富度低于零模型的预期。因此，附生植物的分布在两个尺度上都是聚集的，这可能是由于扩散限制造成的。观察到了与尺度相关的附生植物物种组成模式。附生植物-宿主网络显示出负共现模式的证据，这可能是由于附生植物物种为避免与宿主物种竞争而适应的结果，而大多数附生植物元群落则随机分布。在其他地方应用我们的“元网络”分析框架可能有助于确定附生植物组合结构的一般模式及其在空间和时间上的变化。
