Life-history variation and community structure in neotropical rainforest communities: Ecological and phylogenetic influences

Several key questions in evolution and ecology may be answered by the analysis of comparative data sets in which information on the demography, life history, and relative abundance of large numbers of species is jointly analyzed. For example, the detection of strong associations between habitat, soil quality, seed mass, life form, dispersal mode, and other life history traits has shed light on the selective forces (e.g., light availability and disturbance) that influence the evolution of these attributes within and among taxa.

Virtually all investigators of plant communities have used this approach on single community-level data sets compiled by their own research group (Jackson, 1981, Garwood, 1983, Foster and Janson, 1985, Mazer, 1987, Ibarra-Manrr¿quez & Oyama. 1993, Grubb & Metcalfe, 1996 [flora of Queensland, various sites, analyzing pooled data], Rees 1997 [British flora pooled)], Harms and Dalling, 1997, Clark et al., 1998, Hodkinson, et al., 1998 [British flora, pooled]. Relatively few researchers have evaluated data sets including information from more than one community (Hammond & Brown, 1995 (three Neotropical communities), Lord, Westoby & Leishman, 1995 (five temperate floras), Lord et al., 1997 [tropical vs. temperate floras]). In the proposed NCEAS working group, we will bring together a group of community ecologists and evolutionary biologists to use community-level data to address two sets of questions that depend on the joint analysis of multiple data sets. Because these questions (see below) require the analysis of both pooled and separate community-level data, we have selected participants who have independently collected similar data in ecologically similar plant communities.

One important objective of this proposal is to develop statistical approaches and to offer suggestions for analyses that other investigators will find useful for the ecological and evolutionary interpretation of community-level data. We propose to bring to NCEAS investigators representing seven neotropical rainforest locations (Los Tuxtlas, Mexico; Chajul, Mexico; La Selva, Costa Rica; Mabura Hill, French Guyana; BCI, Panama; Tambopata Wildlife Reserve and Manu National Park, Peru). Collectively, data sets currently available from well-monitored one- to fifty-hectare plots in these locations include information on seed size, seedling demography, survivorship, growth rates, growth form, reproductive phenology, dispersal mode, adult abundances, and/or size distributions of hundreds of plant species. Not all data are available for all sites and species, but we have identified a set of questions and hypotheses to which these data are well-suited. Naturally, we fully expect the working group to identify additional theoretical and applied questions to be addressed.

To conduct cross-community studies of ecological and evolutionary patterns, we have chosen to focus on the tree species of Neotropical moist forest communities for several reasons. First, the adaptive significance of life history and functional traits is strongly dependent on environment. For example, the role of seed size in seedling establishment is quite different in open and xeric environments than in the light-limited, mesic forest understory. Similarly, the ecological factors influencing community structure vary across contrasting environments. By focusing on tropical forests, we can assess the generality of adaptive hypotheses across communities, and test evolutionary hypotheses regarding species differentiation in relation to life history under broadly similar conditions. Secondly, in some cases there are strong floristic affinities among Neotropical forests (e.g. the dominant genera or families of trees are shared among sites). This facilitates tests of evolutionary hypotheses because we can extract data on diverse groups of closely related species in different communities. In comparisons of New World and Old World forests, such analyses of closely related species (e.g., congeners) are often difficult or impossible.

The timing is right to develop a joint, concerted effort to analyze cross-community data to detect ecologically and evolutionarily significant patterns and processes. First, a number of research groups have successfully constructed data sets each characterizing hundreds of species, but there currently exists no coordinated data set to allow the comparative analysis of these data across communities. Second, theoretical and empirical work in the analysis of comparative data have increased our understanding of the value of methods for incorporating phylogenetic information into the quantitative study of multivariate data sets. Third, phylogenies describing inter- and intra-familial relationships are now available for many tropical taxa, allowing the use of phylogenetically independent contrasts that can assist in the detection of: (a) traits associated with high rates of diversification, (b) the correlated evolution of multiple traits, and (c) the joint evolution of traits and habitat requirements. Fourth, our list of participants includes several investigators who have not met before. Finally, an NCEAS working group established at this time would capitalize on the temporary residence in the U.S. of two of our participants (Drs. Miguel Martinez-Ramos and Horacio Paz), who offer an exceptional data set from Los Tuxtlas and from Chajul (Mexico) and are currently collaborating with Dr. David Ackerly and Dr. Susan Mazer.