This study aimed to analyze relationships among the structure of the tree component, edaphic variables and canopy discontinuity along a toposequence in a seasonal upland (hillside) forest in southern Brazil. Soil and vegetation were sampled in 25 plots of 20 × 20 m each. We described the vegetation in terms of structure, richness and diversity, as well as by species distribution patterns. We evaluated canopy continuity, determined sloping and calculated spatial coordinates. We applied partial canonical correspondence analysis (pCCA) to determine whether species distribution correlated with environmental and spatial variables. We identified 1201 individuals belonging to 76 species within 30 families. The species with highest density and frequency were Gymnanthes concolor Spreng., Calyptranthes tricona D.Legrand, Eugenia moraviana O.Berg and Trichilia claussenii DC. The pCCAs indicated significant correlations with environmental and spatial variables. Sand content, boron content and soil density collectively explained 36.17% of the species matrix variation (total inertia), whereas the spatial variables x, y and xy² collectively explained 14.27%. The interaction between environmental and spatial variables explained nearly 4.5%. However, 45.05% remained unexplained, attributed to stochastic variation or unmeasured variables. Terrain morphology and canopy discontinuity had no apparent influence on richness, and changes in species distribution were correlated with sloping, which affects soil features and determines the directional distribution of some species.

O estudo de fenômenos relacionados às variações na estrutura das comunidades vegetais tem cada vez mais envolvido a avaliação do potencial de regeneração das espécies arbóreas, uma vez que a regeneração torna as florestas capazes de se restaurarem após distúrbios naturais ou antrópicos. Neste sentido, objetivou-se avaliar o componente arbóreo regenerante, identificando composição e abundância, grupos funcionais de dispersão, estratificação vertical e necessidades de luz para germinação, além de estimativas de riqueza e diversidade e comparar estes valores com o respectivo componente adulto. O levantamento foi realizado em 20 unidades amostrais de 10 x 10 m, sendo amostrados todos os indivíduos com altura > 0,30 m e diâmetro a altura do solo < 4,7 cm. O levantamento do componente adulto, avaliado em estudo anterior, foi estabelecido em unidades amostrais de 20 x 20 metros, sendo amostrados todos os indivíduos com diâmetro à altura do peito > 4,7 cm. Foram amostrados 1.649 indivíduos em regeneração, pertencentes a 64 espécies, com densidade total estimada em 8.245 ind.ha-1. As espécies com maior densidade absoluta foram Gymnanthes concolor Spreng., Trichilia elegans A.Juss. e Calyptranthes tricona D.Legrand. Uma análise de coordenadas principais indicou nítida separação das unidades amostrais amostradas e um teste de Mantel revelou haver independência entre as matrizes de composição e abundância dos componentes adulto e regenerante (r = 0,19; p = 0,1). Dentre os grupos funcionais, verificou-se que o componente regenerante está mantendo as mesmas proporções observadas para o componente adulto, com maior proporção de espécies zoocóricas, dependentes de luz para germinação e formadoras do dossel florestal. Ao contrário do esperado, houve diminuição da riqueza no componente regenerante (p < 0,001) e manutenção da equabilidade ao longo da área.

Studies related to variation in plant communities have been involved the analysis of regenerative potential of tree species, because saplings may enable forests to restore themselves after natural or human disturbances. The goal of this study was to analyze tree regeneration by identifying composition and abundance, the ecological groups of dispersion, vertical stratification and light requirements for germination and to look for changes in expected richness, diversity and relationships with structure of adult trees. A sapling survey was conducted in 20 sample units of 10 x 10 m where all individual living trees were sampled that had a height > 0.30 m and diameter at soil height < 4.7 cm. Adult trees with a diameter at breast height > 4.7 cm were surveyed in 20 x 20 m sampling units, according to a previous study. We sampled 1,649 living trees (64 species), which generated a total density of 8,245 ind.ha-1. The species with higher absolute density were Gymnanthes concolor Spreng., Trichilia elegans A.Juss. and Calyptranthes tricona D.Legrand. A principal coordinates analysis indicated sharp differences among sampled units and a Mantel test showed no dependence among composition and abundance matrices from adult and regenerative surveys (r = 0.19; p = 0.1). According to the ecological groups, we found that the regenerative component has kept the same proportions of adults, with a higher proportion of zoochorous, light demanding, canopy covering species. Contrary to our initial hypothesis, we found lower expected richness in the regenerative component (p < 0.001) with the same equability in the entire area.