Data from: Phylogenetic tree shape and the structure of mutualistic networks

Species community composition is known to alter the network of interactions between two trophic levels, potentially affecting its functioning (e.g. plant pollination success) and the stability of communities. Phylogenies vary in shape with regard to the rate of evolutionary change across a tree (influencing tree balance) and variation in the timing of branching events (affecting the distribution of node ages in trees), both of which may influence the structure of species interaction networks. Because related species are likely to share many of the traits that regulate interactions, the shape of phylogenetic trees may provide some insights into the distribution of traits within communities, and hence the likelihood of interaction among species. However, little attention has been paid to the potential effects of changes in phylogenetic diversity (PD) on interaction networks. Phylogenetic diversity is influenced by species diversity within a community, but also how distantly-related the constituent species are from one another. Here, we evaluate the relationship between two important measures of phylogenetic diversity (tree shape and age of nodes) and the structure of plant-pollinator interaction networks using empirical and simulated data. Whereas the former allows us to evaluate patterns in real communities, the latter allows us to evaluate more systematically the relationship between tree shape and network structure under three different models of trait evolution. In empirical networks, less balanced plant phylogenies were associated with lower connectance in interaction networks indicating that communities with the descendants of recent radiations are more diverged and specialized in their partnerships. In simulations, tree balance and the distribution of nodes through time were included in the best models for modularity, and the second best models for connectance and nestedness. In models assuming random evolutionary change through time (i.e., Brownian motion), less balanced trees and trees with nodes near the tips exhibited greater modularity, whereas in models with an early burst of radiation followed by relative stasis (i.e. early-burst models) more balanced trees and trees with nodes near roots had greater modularity. Synthesis: Overall, these results suggest that the shape of phylogenies can influence the structure of plant-pollinator interaction networks. However, the mismatch between simulations and empirical data indicate that no simple model of trait evolution mimics that observed in real communities.

已知物种群落组成会改变两个营养级之间的互作网络，进而可能影响其生态系统功能（如植物传粉成功率）与群落稳定性。系统发育树的形态因进化速率在树内的分布（影响树的平衡性）以及分支事件发生时间的差异（影响树内节点年龄的分布）而有所不同，而这两类特征均可能影响物种互作网络的结构。由于近缘物种往往共享诸多调控互作的性状，系统发育树的形态或可为解析群落内性状的分布、乃至物种间发生互作的可能性提供参考。然而，学界对系统发育多样性（Phylogenetic Diversity, PD）的变化对互作网络的潜在影响关注甚少。系统发育多样性既受群落内物种多样性的影响，也取决于组成物种间的亲缘关系远近。本研究借助实测数据与模拟数据，探究了两类关键的系统发育多样性指标——树的平衡性与节点年龄——与植物-传粉者互作网络结构之间的关联。前者可用于解析真实群落中的分布模式，后者则能让我们在三种不同的性状进化模型下，更系统地探究树的形态与网络结构之间的关系。在实测网络中，平衡性较低的植物系统发育树与互作网络更低的连接度相关联，这表明带有近期辐射演化后裔的群落，其物种间的分化程度更高，且在互作关系中特化程度更强。在模拟实验中，树的平衡性与节点的时间分布被纳入模块化程度预测的最优模型，同时也是连接度与嵌套性预测的次优模型。在假设进化变化随时间随机发生的模型（即布朗运动模型）中，平衡性较低的树以及节点更靠近末端分支的树，其模块化程度更高；而在经历辐射爆发早期后进入相对停滞的模型（即早爆发模型）中，平衡性更高的树以及节点更靠近树根的树，模块化程度则更高。综合来看，本研究结果表明系统发育树的形态可影响植物-传粉者互作网络的结构。但模拟结果与实测数据之间的不匹配，说明不存在能够复刻真实群落中实际情况的简单性状进化模型。