Phylogenetic diversity, functional traits, and tropical forest succession.

Abstract

This dissertation investigated changes in phylogenetic and functional diversity during ecological succession following anthropogenic disturbance in the lowland forests of New Guinea. Plant evolutionary history, functional traits, and interactions with insect herbivores were examined to address questions related to patterns of plant diversity, species coexistence, and possible mechanisms maintaining diversity in tropical forests. Chapter 1 investigated the recovery of forest structure, species diversity, and functional diversity during succession by examining a chronosequence of survey plots located in younger secondary, older secondary, and primary forest. Secondary forests had less biomass and lower species richness compared to primary forests. They also had higher specific leaf area (SLA) and foliar nitrogen and their leaves were more likely to contain latex. By contrast, trees in primary forest had high wood density and foliar carbon, larger seeds, and taller trees with larger crowns. These changes reflect the contrasting life histories of trees in young and old forests and possibly reflect a tradeoff between growth rates and life span that suggests deterministic processes such as environmental filtering or competition may drive species coexistence during succession. Chapter 2 compared plant community phylogenetic structure along a successional gradient in lowland New Guinea based on a chloroplast DNA phylogeny. The sensitivity of results to different methods of phylogenetic branch length estimation was assessed by comparison of equal branch lengths, genetic distance, time-calibration, and a relaxed molecular clock Bayesian estimate. Mean phylogenetic distance among co-occurring trees increased with total basal area per plot, a proxy for forest age. Significant phylogenetic clustering was detected in secondary forest whereas primary forest was significantly over-dispersed relative to null expectations. The sensitivity of these patterns to various methods of branch length estimation and phylogenetic uncertainty was also examined. Power to detect community phylogenetic patterns when equal branch lengths were assumed was weak in comparison to direct molecular and time-calibrated measures of divergence. Inferred change during forest succession was also robust to phylogenetic uncertainty so long as temporal information was incorporated in estimates of divergence. The observed patterns are consistent with processes of environmental filtering during tropical forest succession giving way to other processes in primary forests including density-dependent mortality. Chapter 3 examined the degree to which the abundance of herbivores in a rain forest community is explained by the functional traits of host plants. Per-tree caterpillar and leaf miner abundance was measured together with total leaf biomass (kg), percentage of immature foliage, specific leaf area (cm2 g-1), leaf nitrogen content (% dry mass), and presence of exudates. Apart from leaf nitrogen content, neither plant resources nor herbivore abundance showed evidence of phylogenetic conservatism in our community sample. The plant traits explained only 30% and 16% of variation among individual trees in caterpillar and leaf miner abundance. Leaf biomass was a stronger predictor of herbivore abundance than either resource quality (leaf nitrogen content) or palatability (percent immature foliage, specific leaf area). The primary importance of resource quantity was also observed at the plant species level in analyses of species means and phylogenetic generalized least squares regression. Plant exudates significantly depressed herbivore abundance but apparent convergence among community members evidently weakens the power of plant phylogeny alone to predict herbivore community patterns. Leaf nitrogen content, explaining to some extent caterpillar abundance, provides an alternative example of how a conserved trait at one trophic level can influence community-wide patterns at another.