Organized by:

Helene Muller-Landau, University of Minnesota
 Jérôme Chave, CNRS, Toulouse
The goal of this symposium is to highlight the past and potential future gains for both tropical ecology and theoretical ecology from bridging the often large gap separating these two subdisciplines. Much ecological theory is inspired by the tremendous diversity of tropical forests, yet key assumptions are often unrealistic for tropical communities, making empirical application and testing difficult if not impossible. Further, empirical studies inspired by particular theories are often based on weak linkages to theory, and thus provide at best purely qualitative tests of hypothesized mechanisms. Nevertheless, our understanding of tropical forests has benefited from the development and testing of ecological theory related to the intermediate disturbance hypothesis, the Janzen-Connell model, and the neutral theory, to name just a few. In this symposium, we plan one talk on neutral theory, three talks on mechanisms of diversity maintenance, and two talks on phylogenetic structuring of communities. Each talk will include some theory, some tests of the theory against tropical forest data, and discussion of either how theoretical developments were inspired by or advanced by tropical forests and/or how our understanding of tropical forests was advanced by the theory and comparisons with theoretical predictions.
Liza Comita 1 , Helene Muller-Landau 1 , Richard Condit 2 , S. Joseph Wright 2 and
Stephen Hubbell 3
1
University of Minnesota
2
Smithsonian Tropical Research Institute
3
University of Georgia
Email : comital@si.edu
Theoretical work suggests both stabilizing and equalizing forces can contribute to species coexistence in diverse communities. Equalizing, or neutralizing, processes are those mechanisms that act to minimize fitness differences between species, while stabilizing, or frequency-dependent, processes are those that cause species to limit themselves more than they limit others. The presence of stabilizing processes can be assessed by comparing intra- and interspecific effects on species vital rates. However, stabilizing processes can operate on multiple life stages, therefore integrating intra- and interspecific effects across all life stages is necessary for quantifying the relative contribution of stabilizing processes to species coexistence in ecological communities.Using long-term data on seed dispersal, seedling establishment, and seedling, sapling and tree survival and growth, in conjunction with Bayesian hierarchical models, we quantified the effects of con- and heterospecifics on vital rates across multiple life stages for co-occurring tree species in the moist tropical forest of Barro Colorado Island, Panama. We found that

species are more negatively impacted by conspecifics than heterospecifics, consistent with stabilizing mechanisms. Such effects were found at multiple life stages, but were strongest in smaller size classes. This analysis represents a critical first step in using empirical data to determine the relative importance of stabilizing versus equalizing forces in diverse tropical tree communities.
Keywords : coexistence, density dependence, modelling, Janzen-Connell
Helene Muller-Landau 1 and Frederick R. Adler 2
1
University of Minnesota
2
University of Utah
Email : hmuller@umn.edu
Over 35 years ago, Janzen and Connell suggested that specialized natural enemies concentrate around adult trees, that these enemies elevate seed and seedling mortality near conspecific adults, and that this contributes to the maintenance of tree species diversity in tropical forests by preventing any one species from becoming very common.
While we now have extensive empirical evidence for elevated mortality near conspecific adults and theoretical demonstration that such mortality contributes to diversity maintenance, we lack tools for extrapolating from quantities measured in the field to community-level effects on diversity. Fundamentally, the contribution of spatial Janzen-
Connell effects to diversity maintenance depends on the strength of the resulting stabilizing effect of population-level negative frequency-dependence on the change in reproductive rate with relative abundance. Empirical studies have quantified not wholepopulation effects but local effects, specifically, changes in seed or seedling success as a function of the distance to conspecific adults or the local density of conspecific seeds or adults. Here, we use models to illuminate the links or rather, lack thereof between
Janzen-Connell patterns that are or could be measured in the field and larger-scale influences on population regulation and community diversity maintenance. We first investigate how characteristics of specialized natural enemies and their host plant species influence plant populations and communities in models designed to capture key characteristics of tropical forests, including local resource competition. We then do field sampling within the simulated landscape to see what traditional empirical studies of
Janzen-Connell effects might find, and evaluate the relationship between these observed patterns and actual effects on population regulation and species diversity.We find that the correlations are quite poor, and insofar as they exist, are often in the opposite direction of what might be expected. We close by discussing the implications for interpreting empirical patterns of densitydependence.
Keywords : enemies hypothesis, diversity maintenance, Janzen-Connell effects, densitydependence, seed dispersal

Steven W. Kembel
University of California, Berkeley
Email : skembel@berkeley.edu
The phylogenetic structure of ecological communities may provide insights into the relative importance of processes hypothesized to maintain species diversity, as these processes interact with the evolutionary history of plant niches and leave their signature on the phylogenetic relatedness of co-occurring species. The increasing availability of phylogenetic data for large numbers of species, along with data on the abundance and distribution of tropical trees, have been used to test for evidence of the various processes hypothesized to maintain the exceptional diversity of tropical forests. To date, most comparisons of empirical patterns with theoretical predictions have been based on qualitative models of the relationships between niche evolution, community assembly, and community phylogenetic structure. In this study, I use data from tropical forests on
Barro Colorado Island, Panama, along with simulation studies, to quantify the ability of several proposed measures of phylogenetic diversity to detect evidence of competition and environmental filtering. Measures of community phylogenetic structure differed in their ability to detect processes influencing community assembly, and interactions between species frequencies, sample richnesses, environmental heterogeneity, community assembly processes and niche evolution sometimes made it difficult to attribute patterns of phylogenetic structure to any particular ecological or evolutionary process. However, phylogenetic relatedness varied a great deal among habitats and along environmental gradients within the forests on Barro Colorado Island. In young forests and plateau habitats, trees were more closely related to their neighbors than expected, while trees in swamp and slope habitats were more distantly related to their neighbors than expected. I discuss the potential importance of these results for our understanding of the relative importance of processes such as competition and filtering in structuring tropical forest communities, and the evolution of the niches and habitat associations of tropical trees.
Keywords : community phylogenetic structure, community assembly, competition, environmental filtering, habitat associations
Olivier J. Hardy
Universite Libre de Bruxelles, Belgium
Email : ohardy@ulb.ac.be

Analyzing the phylogenetic structure of natural communities may illuminate the processes governing the assembly and coexistence of species in ecological communities.
Different statistical tools were recently proposed to test such community-wide phylogenetic patterns, based on the phylogenetic clustering or overdispersion of the species in a local community. This provides a much needed framework for addressing oldstanding questions in community ecology as well as the recent debate on community neutrality. The testing procedures are based on (1) a metric measuring the association between phylogenetic distance and species cooccurrence, and (2) a data set randomization scheme providing the distribution of the metric under a given null model.
To verify the reliability of these approaches, I tested their statistical properties against simulated datasets devoid of phylogenetic structure regarding the spatial distribution of species. These datasets were obtained by simulating a locally neutral subdivided community that could include dispersal limitation and/or a phylogenetic signal in the species overall abundances. The consistency of type I error rates of tests based on several metrics combined with different null models was checked. This study shows that the approaches published so far can lead to liberal tests (i.e. tests rejecting the null hypothesis too often) under particular but important conditions. This may occur, for example, when species abundances are distributed non-randomly in space or across the phylogeny.
Although no test was found valid in all conditions, some testing procedures are much more robust than others, in particular using two new metrics. The bad performances of several tests that have been used in previous studies might call for a reexamination of the evidence.
Keywords : community phylogenetic structure, neutral model, null models, randomization tests, type I error rate
C. K. Kelly and M. G . Bowler
University of Oxford
Email : colleen.kelly@zoo.ox.ac.uk
Community structure refers to the number of species in a community and the pattern of distribution of individuals among those species. We use a novel way of representing community structure to show that abundance within closely related pairs of co-occurring tree species in a highly diverse Mexican forest is more equitable than is abundance within more distantly related pairs. This observation falsifies the fundamental assumption of neutral models of community structure, that species are interchangeable, while simultaneously showing the significant impact of non-neutral (stabilizing) processes at the level of the community. Comparison with appropriate models further establishes that species are not interchangeable even within closely related pairs, signifying that relative abundance within such pairs is not simply a function of ecological similarity. The observed patterns instead indicate niche apportionment regulation of relative abundance by competition operating within compartments in which interaction is focused pairwise between congeners but falls away from the phylogenetic structure above the genus level.

Niche processes thus significantly affect community structure through regulating relative abundance in a substantial proportion of species, which in turn potentially enhances community stability. We develop an analytical description of the observed pattern of relative abundances using temporal niche dynamics in which otherwise similar species exploit resource fluctuations over time rather than space, a mechanism by which competitors may stably coexist. One such mechanism of stable coexistence has already been shown to be active in this forest. keywords : community structure; fractional abundance; stable coexistence; hierarchical niches; differential sensitivity storage dynamics; temporal niches
Jérôme Chave
CNRS, Toulouse
Email : chave@cict.fr
Theoretical models of species-rich communities have recently made much progress in relating models to data, beyond simple curve-fitting exercises. In this paper I will present approximate statistical inference methods that can be used when the study model is not as easily analytically tractable as Hubbell's neutral model of biodiversity. These methods are of broad relevance in biostatistics and I will use them to estimate the parameters of a dynamic model of forest community.
Keywords : model, statistical inference, community