Abstract: 1. Leaf-litter ants represent a major component of biodiversity and are excellent bioindicators reflecting the health of terrestrial ecosystems. This study, conducted in an unspoiled forest near the Nouragues Research Station, represents the first inventory of leaf-litter ant diversity conducted in French Guiana, and so can be considered as the baseline dataset for ants in this country. 2. Ants were extracted from the leaf-litter using the Ants of the Leaf Litter Protocol, along an altitudinal gradient at four forest sites, including an inselberg. 3. A total of 196 ant species representing 46 genera distributed over eight subfamilies were collected. Four distinct communities spread over a gradient of diversity were thus identified: the liana forest was the most species-rich (140 species) followed by the forested plateau (102 species), the transition forest (87 species) and the forest at the top of the inselberg (71 species). 4. The discovery of species new to science plus several species recorded for the first time in French Guiana, coupled with the particular context of this area, suggests that the Nouragues Research Station might represent a centre of endemism. Once completed, this leaf-litter ant dataset will contribute greatly to the knowledge of ant biodiversity in French Guiana, and has the potential to progressively become an indispensable tool for country-wide conservation planning programmes.

Abstract: The high spatial variability of soil respiration in tropical rainforests is well evaluated, but influences of biotic factors are not clearly understood. This study underlines the influence of tree species characteristics on soil respiration across a 16-monospecific plot design in a tropical plantation of French Guiana. A large variability of soil CO2 fluxes was observed among plots (i.e. 2.8 to 6.8 μmol m(-2) s(-1)) with the ranking being constant across seasons. There were no significant relationships between soil respiration and soil moisture or soil temperature, neither spatially, nor seasonally. The variability of soil respiration was mainly explained by quantitative factors such as leaf litterfall and basal area. Surprisingly, no significant relationship was observed between soil respiration and root biomass. However, the influence of substrate quality was revealed by a strong relationship between soil respiration and litterfall P (and litterfall N, to a lesser extent).

Abstract: Classifying species by shared life-history traits is important if common ecological response groups are to be identified among different species. We investigated how habitat area, local and regional factors shape plant communities in small isolated closed depressions, and how the species richness is related to the interplay between environmental factors and specific life-history trait combinations. In Central-Western Europe, 169 closed depressions were completely Surveyed for plant presence in two highly contrasted landscapes (forested and open landscapes). All species were clustered into 9 Emergent Groups based oil 10 life-history traits related to plant dispersal, establishment and persistence. Habitat areas were related to species presence using logistic regressions. Most Emergent Groups were more area-dependent in open than in forested landscapes, owing to heterogeneous light levels in forest weakening the species-area relationship. In open landscapes, Floating Hydrophytes were severely underrepresented in very small depressions, owing to the absence of waterfowl poulation. Local environmental and regional factors were related to species richness using Generalized Linear Models. In open landscapes, local environmental factors such as water conductivity or soil productivity are respectively the main predictors. In forested landscapes, the abundance of most Emergent Groups Was better predicted by regional factors, i.e., habitat connectivity and distance to the forest edge. Forested landscapes strongly impeded the closed depressions' colonization by the less mobile Emergent Groups Such as Large-seeded Perennials. (C) 2009 Elsevier Masson SAS. All rights reserved.

Abstract: In the understory of pristine Guianese forests, the myrmecophyte Hirtella physophora almost exclusively shelters colonies of the plant-ant Allomerus decemarticulatus in its leaf pouches. We experimentally tested three non-mutually exclusive hypotheses concerning phenomena that can determine the species specificity of this association throughout the foundation stage of the colonies: (1) interspecific competition results in the overwhelming presence of A. decemarticulatus queens or incipient colonies; (2) exclusion filters prevent other ant species from entering the leaf pouches; and (3) host-recognition influences the choice of founding queens, especially A. decemarticulatus. Neither interspecific competition, nor the purported exclusion filters that we examined play a major role in maintaining the specificity of this association. Unexpectedly, the plant trichomes lining the domatia appear to serve as construction material during claustral foundation rather than as a filter. Finally, A. decemarticulatus queens are able to identify their host plant from a distance through chemical and/or visual cues, which is rarely demonstrated in studies on obligatory ant-plant associations. We discuss the possibility that this specific host-recognition ability could participate in shaping a compartmentalized plant-ant community where direct competition between ant symbionts is limited. (C) 2009 The Linnean Society of London, Biological Journal of the Linnean Society, 2009, 97, 90-97.

Abstract: Though it is recognized that both stochastic and deterministic processes structure all communities, empirical assessments of their relative importance are rare, particularly within any single community. In this paper, we quantify the dynamic effects of dispersal assembly and niche assembly on the seedling layer in a diverse neotropical rain forest. The two theories make divergent predictions regarding the roles of seed arrival and environmental heterogeneity in generating community structure. Put simply, dispersal assembly posits that the stochasticity inherent to seed arrival structures communities, whereas niche assembly suggests that heterogeneity in post-dispersal environmental conditions is more important. We experimentally sowed 15,132 seeds of eight tree species at varying levels of density and diversity. Every six months we censused the seedlings that germinated and assessed the abiotic and biotic conditions of each plot. We assessed the density, diversity, and species composition of three nested subsets of the seedling layer: seedlings germinated from sown seeds, all seedlings germinated between July 2003 and 2004, and all woody seedlings. We partitioned the variance in density and diversity of each subset of the seedling layer into components representing seed-addition treatments and environmental conditions at 6- to 12-month intervals. Seed additions initially explained more variance in the density and diversity than did environmental heterogeneity for seven of eight sown species, but explained little variance in the density or diversity of the entire seedling layer. Species composition was better explained by seed-addition treatments than by environmental heterogeneity for all three subsets and in all time periods. Nevertheless, the variance in community structure explained by seed-addition treatments declined over the two years following germination, presaging shifts in the relative importance of dispersal assembly and niche assembly. Our study quantifies how dispersal assembly and niche assembly may vary among the components of an ecological community and shift dynamically through time.