Ferns are an important component of the tropical flora, representing about 10% of all vascular plants of tropical forests, constituting the dominant group among the herbaceous vegetation. The aim of this study was to determine whether the ferns communities in two areas of Atlantic Forest, Restinga (FRS) and Lowland Ombrophilous Dense (FTB) forests, located on soils remarkably distinct and separated by a highway, differ in their floristic composition and diversity. These two faces are located in Picinguaba locality, Parque Estadual da Serra do Mar, Ubatuba, São Paulo (23° 31' to 23° 34' S and 45° 02' to 45° 05' W). We delimited seven plots of 20 × 20 m in each area, collected, identified and listed the fern individuals. We performed three methods of multivariate analysis (TWINSPAN, UPGMA and PCO) to verify the existence of floristic groups; calculated the Payandeh index to define the level of aggregation of each species in each of the two areas (FRS and FTB); obtained indicator species for each area; and performed diversity analysis (richness, evenness and heterogeneity). The two areas sampled 3482 individuals, 1366 of which occur in FRS and 2116 in FTB. We found 14 species in FRS and 29 in FTB; these areas shared five species. We found two well-defined floristic blocks (FRS × FTB) in cluster analysis. In the FRS, relatively most species showed aggregated distribution and fewer species showed random distribution in relation to FTB. We found four indicator species for FRS and eight for FTB. The curves resulting from the individual-based rarefaction indicated a strong difference in species richness between the two communities, in favor to FTB. These differences may be related to physical and chemical characteristics of soil (more fertile and less toxic in the FTB), water dynamics (plant species in the FRS are generally subject to water stress) and the sharp break caused by highway that cuts through the park. These aspects could be related to different forms of occupation of niches and to the fact that some species in environments with limited resources and heterogeneously distributed in space, such as the studied Restinga, would eliminate the others.
edaphic parameters; environmental selection; multivariate analysis; species richness
