ABSTRACT It is assumed that morphological traits of seedlings reflect different strategies in response to environmental conditions. The ecological significance of this has been widely documented in rainforests, where habitat structure and species interactions play an important role in community assembly. However, in seasonally dry ecosystems, where environmental filtering is expected to strongly influence community structure, this relationship is poorly understood. We investigated this relationship between functional groups of seedlings and life history traits and tested whether functional group predicts the ecological strategies employed by woody species to deal with the stressful conditions in seasonally dry ecosystems. Seedling functional groups, life history traits and traits that reflect ecological strategies for occupying seasonally dry environments were described for twenty-six plant species. Seedlings of species from the Caatinga vegetation exhibited a functional profile different from that observed in rainforests ecosystems. Phanerocotylar-epigeal seedlings were the most frequently observed groups, and had the largest range of ecological strategies related to dealing with seasonally dry environments, while phanerocotylar-hypogeal-reserve seedlings exhibited an increase in frequency with seasonality. We discuss these results in relation to those observed in other tropical forests and their ecological significance in seasonally dry environments.

RESUMO Este estudo foi realizado com o objetivo de investigar a influência da heterogeneidade espacial dos fatores abióticos sobre os padrões espaciais da estrutura da vegetação do componente lenhoso e sua regeneração em uma área de caatinga, no sertão de Pernambuco. O estudo foi realizado em uma área de caatinga localizada no Parque Estadual da Mata da Pimenteira, município de Serra Talhada, Pernambuco. A amostragem foi dividida em dois ambientes, um próximo ao curso d'água com vegetação mais densa (I) e outro a 100 m do curso d'água com vegetação mais aberta (II). Em cada ambiente foram instaladas quatro parcelas permanentes de 20 × 50 m subdivididas em 10 parcelas de 10 × 10 m, nas quais foram medidos os indivíduos vivos com diâmetro do caule ao nível do solo (DNS) ≥ 3 cm e altura total ≥ 1 m. Em um dos vértices de cada parcela foi plotada uma subparcela de 2 × 2 m para medir a altura e o diâmetro dos indivíduos das espécies amostradas no componente lenhoso com DNS entre 0,5 e 2,9 cm (regeneração). Foram calculados os descritores fitossociológicos gerais da comunidade dos diferentes ambientes e avaliada a distribuição de indivíduos em histogramas, tanto do componente lenhoso como para a regeneração. Após verificação da normalidade pelo teste Kolmogorov-Smirnov, os valores dos descritores fitossociológicos por parcela foram comparados entre os ambientes. Nos dados com distribuição normal, foi empregada a análise de variância (ANOVA) de um critério. Dados sem distribuição normal foram analisados pelo teste não paramétrico de Mann-Whitney. Os ambientes apresentaram diferenças significativas quanto ao número de indivíduos (4047,5 versus 3332,5 ind.ha-1) e área basal (23,6 versus 17,6 m2. ha-1), com valores significativamente superiores no ambiente I, padrão oposto ao da regeneração em que o número de indivíduos e a área basal foram maiores no ambiente II (9187,5 versus 10937,5 ind.ha-1) e (1,9 versus 2,5 m2. ha-1), respectivamente. No total foram amostradas 50 espécies (incluindo o componente arbóreo e regenerante), havendo diferenças significativas em termos de diversidade entre os ambientes nos diferentes estratos, embora a similaridade seja superior a 80%. A maioria das variáveis químicas e texturais dos solos não diferiram estatisticamente, com exceção dos teores de Al3+ que apresentou teores superiores no ambiente I. Das demais variáveis edáficas, apenas a pedregosidade diferiu entre os ambientes I e II (3,87 versus 1,47%). A distinção fisionômica da vegetação entre os ambientes é resultado da presença de populações com indivíduos de maiores alturas e diâmetros no ambiente I. A heterogeneidade espacial das variáveis ambientais encontrada nos ambientes estudados, explicou parte das variações das características florístico-estrutural da vegetação.

ABSTRACT This study aimed to investigate the influence of spatial heterogeneity of abiotic factors on the spatial patterns of vegetation structure of the woody component and of its regeneration; it was carried out in an area of "caatinga" located in the Parque Estadual Mata of Pimenteira, municipality of "Serra Talhada", interior of Pernambuco State, Brazil. The sample was divided into two environments, one near the watercourse with denser vegetation (I) and another at 100 meters away from the watercourse with more open vegetation (II). In each environment, four permanent plots of 20×50 m were installed and divided into ten plots of 10×10 m; alive individuals with stem ground level diameter (GD) of ≥3 cm and total height of ≥1 m were measured. In one corner of each plot, a sub-plot of 2×2 m was allocated to measure the individuals' height and diameter of the sampled species in woody component with GD from 0.5 to 2.9 cm (regeneration). Community's general phytosociological descriptors of the different environments were calculated and the individuals' distribution in histograms was evaluated for the woody component as for regeneration. After testing for normality using Kolmogorov-Smirnov test, values of phytosociological descriptors per plot were compared between the environments. Analysis of variance (ANOVA) of a criterion was used for data with normal distribution. Data without normal distribution were analyzed by non-parametric Mann-Whitney test. Environments showed significant differences in the number of individuals (4047.5 versus 3332.5 ind.ha-1) and basal area (23.6 versus 17.6 m2. ha-1) with significantly higher values in environment I; this pattern was opposite to the regenerating pattern in which the number of individuals and basal area were higher in the environment II (9187.5 vs. 10937.5 ind.ha-1 and 1.9 vs. 2.5 m2.ha-1, respectively). In total, fifty species, including woody and regenerating component, were sampled; significant differences in diversity were found between environments in the different strata, although the similarity was more than 80%. Most chemical and textural soil variables did not differ statistically, except for levels of Al3+, which showed higher levels in the environment I. Among other edaphic variables, only the stony variables differed between environments I and II (3.87 versus 1.47%). The vegetation physiognomic distinction between the environments is result from populations' presence of individuals with greater heights and diameters in the environment I. The spatial heterogeneity of environmental variables found in the studied environments explained part of changes in floristic and structural characteristics of the vegetation.

To test the hypothesis that the physiognomy, structure and dynamics of woody species are influenced by the spatial gradient edge-interior, data were collected from these parameters in 2004, 2007 and 2010 in an Atlantic forest fragment. Species were sampled for the Canopy (DS), with circumference > 15 cm to 1.3 m soil and understory (SB) in circumference between ground level> 3 cm and CAP <15 cm on edge-gradient formed within 35 years. In 2004, we assessed three positions in the fragment: edge, intermediate and inner (> 150 m from the forest boundary), and analyzed 1,000 m² per site, divided into plots of 10 m², to measure individuals of higher class. In each plot a sub-plot of 5 m² for sampling understory was installed. Branches were collected from three adult individuals from the species to obtain the density of the wood. The number of individuals (NI), basal area (BA), mortality (TM), recruitment rate (TR), periodic annual diameter increment (IPA), loss rate (P) and gain rate (G) Area and basal turnover rates in number of individuals (TN) and basal area (TA) class tree component in the three locations showed no significant edge effect. At the edge, the understory had lower values of NI and AB and higher TM than in the forest interior. Differences between the strata occur regardless of edge sealing, since young plants present in most individuals in the sample class SB, which are more sensitive than adult trees to changes caused by the creation of the edge. Rates of loss and gain of AB and AB turnover in NI and in SB presented higher values than on the DS, which points to a stratum with larger changes. In DS there were no differences in species richness among environments differs from that observed in SB. The floristic composition of the DS in environments closer to the forest limit was quite distinct from that of the interior, indicating that there has been no recovery of floristic composition. All environments SB had low similarity. The edges of the DS and SB had a higher concentration of individuals in class intervals smaller density values timber. The DS has already recovered biomass, although the edge and intermediate environments have not recovered species composition late. The SB still suffers the edge effect. Possibly, the highest concentration of individuals is in class with lower density wood result on edge.

Analisamos a variação funcional entre espécies decíduas (ED) e sempre verdes (ESV) para compreender as divergências nas estratégias de tolerância e evitação à seca. O estudo foi realizado em um fragmento de floresta tropical sazonalmente seca, localizada no munícipio de Pentecoste (3°47'S, 39°16'W), Ceará, Brasil. Mensuramos 17 traços funcionais foliares em 17 ED e cinco ESV, sendo 12 morfofuncionais, um fenológico e quatro fisiológicos. Verificamos que as ED exibiram maior taxa de fotossíntese por massa (Amassa), menor longevidade foliar (LF) e massa foliar específica (MFE) quando comparadas às sempre verdes. Esses traços foram considerados traços-chaves preditores das estratégias de evitação e tolerância à seca. As ED e ESV apresentaram uma demanda conflitante entre tolerância à seca e taxa fotossintética, pois a LF foi negativamente correlacionada com à Amassa. Embora tenham demonstrado diferenças claras na MFE e LF não observamos diferenças significativas na Aárea e gs, consequentemente, ED e ESV não diferiram na eficiência no uso da água durante o período chuvoso. Apesar da variabilidade substancial dentro do grupo, todas as ED exibem estratégia de evitação à seca enquanto que ESV exibem um conjunto de traços funcionais foliares relacionados a estratégia de tolerância à seca.

We analyzed the functional variation between deciduous and evergreen species to understand the differences across the strategies of avoidance and tolerance to drought. Our study was carried in a fragment of seasonally dry tropical forest located in the city of Pentecoste (3°47‘S, 39°16’W), Ceará, Brazil. We measured 17 functional leaf traits in 17 deciduous and five evergreen species, being 12 morphofunctional, one phenological and four physiological. We found that deciduous species exhibited greater Amass, lower leaf lifespan (LL) and leaf mass per area (LMA) than evergreen species. These traits were considered key predictor traits of avoidance and drought tolerance strategies. Deciduous and evergreen species showed a trade-off between drought tolerance and photosynthetic rate, LL was negatively correlated with Amass. Although we have shown clear differences in LMA and LL not observed significant differences in Aarea and gs, consequently, deciduous species and evergreen species did not find significant differences in WUEi and WUE. Despite substantial variability within the group, all deciduous species fall predominantly in the drought avoidance strategy, while evergreen species display a suite of traits related to drought tolerance.