A quantitative framework to estimate the relative importance of environment, spatial variation and patch connectivity in driving community composition.

Perhaps the most widely used quantitative approach in metacommunity ecology is the estimation of the importance of local environment vs. spatial structuring using the variation partitioning framework. Contrary to metapopulation models, however, current empirical studies of metacommunity structure using variation partitioning assume a space-for-dispersal substitution due to the lack of analytical frameworks that incorporate patch connectivity predictors of dispersal dynamics. Here, a method is presented that allows estimating the relative importance of environment, spatial variation and patch connectivity in driving community composition variation within metacommunities. The proposed approach is illustrated by a study designed to understand the factors driving the structure of a soft-bottom marine polychaete metacommunity. Using a standard variation partitioning scheme (i.e. where only environmental and spatial predictors are used), only about 13% of the variation in metacommunity structure was explained. With the connectivity set of predictors, the total amount of explained variation increased up to 51% of the variation. These results highlight the importance of considering predictors of patch connectivity rather than just spatial predictors. Given that information on connectivity can be estimated by commonly available data on species distributions for a number of taxa, the framework presented here can be readily applied to past studies as well, facilitating a more robust evaluation of the factors contributing to metacommunity structure.

Ectoparasites of small-mammals: determinants of community structure in South American savannah.

This study aimed to assess the contribution of hosts characteristics (rodents and marsupials) in the organization of ectoparasite communities present in woodland patches in western central Brazil. We verified the effect of host species, sex, body mass and vertical strata in addition to the role of seasonality on the ectoparasite composition, richness and abundance. The total sampling effort was 22 032 trap-nights equally distributed in 54 woodland patches. Variance partition and principal coordinate analysis were used to verify the existence of significant relationships between response variables and predictors. As expected, host species was the most important variable in ectoparasite community assembly. The composition, richness and abundance of mites and lice were highly influenced by host species, although higher for mites than for lice. Host body mass had a determining role on the richness and abundance of tick species. Vertical stratification and seasonality had weak influence, while the sex of the host had no influence on the organization of these communities. The results are closely related to the evolutionary characteristics of the species involved, as well as with local environmental characteristics of the study area.

Linking wood traits to vital rates in tropical rainforest trees: Insights from comparing sapling and adult wood.

PREMISE OF THE STUDY: Wood density is the top predictor of growth and mortality rates (vital rates) but with modest explanatory power at best. Stronger links to vital rates are expected if wood density is decomposed into its anatomical properties at sapling and adult stages, since saplings and adults differ in wood traits and vital rates. We examined whether anatomical determinants of wood density and strength of the relationship between wood traits and vital rates shift between saplings and adults. METHODS: Using wood segments from near pith (sapling) and near bark (adult) for 20 tree species (three adults each) from Barro Colorado Island, Panama, we quantified wood traits. Vital rates for saplings and adults were obtained from an earlier study. KEY RESULTS: Anatomical predictors of wood density were similar for sapling and adult wood, with wood density variation largely explained by fiber lumen area and fiber wall fraction. In sapling wood only, growth rates decreased with fiber wall fraction and increased with fiber lumen area, while mortality rates increased with vessel area but decreased with fiber wall fraction and vessel density. CONCLUSIONS: Wood traits of sapling trees provide functional insight into the growth-mortality tradeoff. Sapling wood with relatively large fiber lumen area and wide vessels, enabling faster hydraulic transport but less mechanical strength, is associated with fast growth and high mortality. Sapling wood with relatively more fiber wall and many narrow vessels, enabling greater mechanical strength but slower hydraulic transport, is associated with slow growth and low mortality.