No abstract

In this study we surveyed all trees g10 cm diameter at breast height within 20 0.1 ha plots in a 3,500 ha forest fragment, surrounded by sugar cane fields, of the Brazilian Atlantic forest to compare tree species assemblages at the forest edge (0-100 m into forest) vs forest interior (>200 m). Plots were perpendicular to the margin. The mean number of tree species was significantly higher in the forest interior (35.4 P SD 7.1 vs 18.4 P SD 4.4). In addition, forest edge differed from interior in the proportion of shade-tolerant, emergent, large/very large-fruited species, and large-seeded Forest edge in the Brazilian Atlantic forest: drastic changes in tree species assemblagesMarcondes A. Oliveira, Alexandre S. Grillo and Marcelo Tabarelli tree species. Among the 134 tree species recorded, 24% were exclusive to the forest edge and 57% to the forest interior. Our results suggest that both the current system of protected areas and archipelagos of small fragments (1) tend to retain only a subset of the original flora, (2) will converge in terms of floristic and ecological composition (biotic simplification and homogenization), and (3) will lose rare and threatened tree species.

No abstract

The role played by abandoned nests of leaf-cutting ants (Atta spp.) as a small-scale disturbance regime that affects plant recruitment, species coexistence and forest regeneration remains poorly investigated. Here we examine whether abandoned nests of Atta cephalotes serve as regeneration niches and operate as particular plant recruitment habitats, favouring forest regeneration after ant activities cease and leading to the establishment of taxonomically/ecologically distinct plant assemblages. Soil properties, canopy openness, light availability and regenerating plant assemblages were evaluated across 18 nests and adjacent control plots in a large remnant of Atlantic Forest in north-east Brazil from December 2004 to December 2005. Surprisingly, nests and control plots exhibited very similar light environments irrespective of nest age, but nest soils exhibited substantial reductions in carbon content (1.45 Ϯ 0.24 vs. 1.79 Ϯ 0.13%) and organic matter (2.50 Ϯ 0.41 vs. 3.08 Ϯ 0.23%), and proved to be much more resistant to penetration (30.57 Ϯ 6.08 vs. 39.48 Ϯ 7.53 mm). Functional signature of regenerating plant assemblages exhibited little variation across both habitat types, as they were dominated by pioneer, small-seeded and vertebrate-dispersed species. However, abandoned nests exhibited less dense, impoverished and more homogeneous regenerating plant assemblages at local and landscape scale; they clearly lacked nest-dependent plant species and represented floristic subsets of the flora inhabiting the undisturbed forest. This recruitment bottleneck was transient in the long term because nest-related effects ameliorated in older nests. Our results suggest that, unlike treefall gaps, abandoned nests represent temporary (relatively long-lasting) islands of unsuitable substrate that reduce plant recruitment, retard forest regeneration, and fail in providing a special regeneration niche able to promote species coexistence and plant diversity.

Impacts of habitat loss and fragmentation on specialist herbivores have been rarely addressed. Here we examine the structure of plant and galling insect assemblages in a fragmented landscape of the Atlantic forest to verify a potential impoverishment of these assemblages mediated by edge effects. Saplings and galling insects were recorded once within a 0.1-ha area at habitat level, covering forest interior stands, forest edges, and small fragments. A total of 1,769 saplings from 219 tree species were recorded across all three habitats, with differences in terms of sapling abundance and species richness. Additionally, edge-affected habitats exhibited reduced richness of both host-plant and galling insects at plot and habitat spatial scale. Attack levels also differed among forest types at habitat spatial scale (21.1% of attacked stems in forest interior, 12.4% in small fragments but only 8.5% in forest edges). Plot ordination resulted in three clearly segregated clusters: one formed by forest interior, one by small fragments, and another formed by edge plots. Finally, the indicator species analysis identified seven and one indicator plant species in forest interior and edge-affected habitats, respectively. Consequently, edge effects lead to formation of distinct taxonomic groups and also an impoverished assemblage of plants and galling insects at multiple spatial scales. The results of the present study indicate that fragmentation-related changes in plant assemblages can have a cascade effects on specialist herbivores. Accordingly, hyperfragmented landscapes may not be able to retain an expressive portion of tropical biodiversity.