Structure and phytogeographic relationships of swamp forests of Southeast Brazil

Swamp forests are associated with soils that are saturated or inundated because of a high water table. In Brazil, little is known about the plant ecology of such forests. In this paper, we aimed to describe the phytosociological structure of the tree layer of swamp forests in Restinga de Jurubatiba National Park, in the northern part of the state of Rio de Janeiro, and to evaluate the floristic similarities between these forests and some other possibly related types of vegetation formations in Brazil. The sampling included 84 species, within 62 genera and 34 families. The Shannon diversity index was 3.42, and the Shannon evenness index was 0.77. The forests studied showed an oligarchic structure; Tapirira guianensis, Calophyllum brasiliense and Protium icicariba were the most important species. Oligarchy, or monodominance, and relatively low species richness are the norm in the swamp forests of southeastern Brazil and result from the strong selective character of the saturated/inundated soils. In comparison with local areas of restinga (coastal woodland), Atlantic Forest sensu stricto, other swamp forests and flooded riparian forests, the similarity was low (Jaccard similarity coefficient < 0.25). In addition to the similar ecological conditions, geographic proximity was a key factor determining the patterns of similarity found. Our results indicate that the swamp forests of southeastern Brazil do not represent a distinguishable floristic unit, due to sources of local variation, notably migration and the establishment of adaptive species from neighboring vegetation formations (some 70% of the species surveyed).


Introduction
Swamp forests comprise vegetation that develops on soils saturated or inundated by the water table (WCMC 1992;Scarano 2006).In Brazil, such forests are known by a variety of local names (Dorneles & Waechter 2004).These forests are naturally fragmented and are associated with hydromorphic soils, occurring near springs, on the banks of rivers or lakes and within natural topographic depressions (Ivanauskas et al. 1997;Toniato et al. 1998;Jacomine 2004).They are widely distributed throughout the neotropics and present interfaces with various types of vegetation formations, including different types of forests and grasslands (Teixeira & Assis 2011).
Until recently, little was known about the plant ecology of swamp forests in Brazil.However, in the last 20 years, many local surveys have been conducted in the southeas-tern, southern and central-west regions of Brazil (Teixeira & Assis 2011;Tab. 1).This increase of local floristic data has made it possible to conduct a number of large-scale phytogeographic analyses, in which swamp forests have been compared with flooded riparian forests (Rodrigues & Nave 2004;Silva et al. 2007) or with each other (Teixeira & Assis 2011).However, the authors of such studies have focused their analyses on swamp forests located on the Brazilian Highlands, evaluating only two areas of swamp forest in the coastal plains, where a number of local surveys have already been conducted (Tab.1).
A number of studies have defined the overall structural and floristic patterns of Brazilian swamp forests: low diversity; the predominance of one or a few species-including the flooded-forest specialist species Calophyllum brasiliense Cambess., Symphonia globulifera L.f. and Tabebuia cassinoides (Lam.)DC., in areas of the coastal plain of the state of Rio de Janeiro (Scarano et al. 1997); and flora that is heavily influenced by the surrounding vegetation (e.g., Teixeira & Assis 2011).However, in comparison with other vegetation formations in Brazil, little is known about these wetlands.

Articles
As previously mentioned, the few large-scale comparisons conducted to date have focused mainly on the swamp forests of the Brazilian Highlands.Therefore, there is a need for not only additional local surveys but also for further phytogeographic analyses that include swamp forests on the coastal plain.Within the Atlantic Forest biome, especially in the coastal plains, many swamp forests have been cleared or disturbed by changes in the flooding regime, timber extraction or fire (Ivanauskas et al. 1997;Galvão et al. 2002;Carvalho et al. 2006b;Scarano 2006).This poor conservation status is not surprising, given that the biome as a whole has been reduced to 11.4-16.0% of its original size (Ribeiro et al. 2009).Unfortunately, although the Atlantic Forest is internationally recognized as a biodiversity hotspot (Myers et al. 2000;Mittermeier et al. 2005) and now benefits from a number of conservation initiatives and policies (Aguiar et al. 2003;Rocha et al. 2006), its marginal habitats (sensu Scarano 2002), such as the swamp forests, are not always treated with such care (Scarano 2009, Ribeiro et al. 2011).
Our paper focuses on a particular set of swamp forests in southeast Brazil, namely those that appear on marine sand deposits within areas of restinga (coastal woodland), which are ecosystems associated with the Atlantic Forest (Scarano 2009).These Quaternary deposits are covered by a mosaic of vegetation, ranging from sparse herbaceous to dense forest vegetation (Araujo et al. 1998).
The objective of this study was to describe the phytosociological structure of the tree layer of the swamp forests in southeastern Brazil, as well as to determine the extent to which they are floristically similar to local areas of open restinga, Atlantic Forest sensu stricto, other swamp forests, and flooded riparian forests.We also sought to determine whether those swamp forests presented phytogeographic relationships with the other vegetation formations.

Study area
Our study area was within Restinga de Jurubatiba National Park (22°08'-22°19'S; 41°17'-41°43'W), which is located on the northern coast of the state of Rio de Janei-ro, in southeastern Brazil (Fig. 1).The park encompasses portions of the municipalities of Macaé, Carapebus and Quissamã, covering an area of 14,839 ha, and preserves one of the most important remnants of the restinga ecosystem in Brazil (Esteves 1998;Rocha et al. 2004).
The study area is in the southern portion of an extensive Quaternary coastal plain, the origin of which is closely related to the evolution of the Paraíba do Sul River delta.This portion of the plain is composed mainly of Pleistocene marine sands; Holocene marine deposits are scarce and limited to a narrow strip near what is now the shoreline.The Pleistocene terraces are limited inwards by Tertiary sediments of the Barreiras Formation (Martin et al. 1993).This great coastal plain has been extensively drained and inhabited by humans since the seventeenth century.An extensive network of artificial canals-initially constructed to transport agricultural products and timber but subsequently used in order to increase the amount of arable land-lowered the water table level and dried up many local lagoons and swamps.In addition, levees were built along the Paraíba do Sul River to prevent flooding on part of the plain (Martin et al. 1993;Soffiati 1998).
Restinga de Jurubatiba National Park is also a research station for the Brazilian Long-Term Ecological Research Program (Barbosa et al. 2004;Rocha et al. 2004).In recent years, a number of plant ecology studies have been conducted within the park (Scarano et al. 2004).Mean annual precipitation is 1164 mm and rainfall is strongly seasonal, with monthly averages ranging from 41 mm in winter (June) to 189 mm in summer (December).The mean annual temperature is 22.6°C, with monthly averages ranging from a maximum of 29.7°C, in January, to a minimum of 20.0°C, in July (Henriques et al. 1986).
The swamp forests studied are among some ten types of vegetation formations found in the study area (Araujo et al. 1998) and occur in swales between successive ancient beach ridges, expanding over branches of local lagoons.The swamp forests are subjected to flooding due to a rise in the water table during the rainy season (October-March).The most typical shape of such forests is that of long strips parallel to the coastline, varying in width (but mostly narrow), abruptly or gradually giving way to areas of open restinga on either side (Fig. 1).Although the restinga studied was formed predominantly during the Pleistocene, the conditions did not become conducive to swamp formation until more recently, a result of the silting of lagoons formed by Holocene marine sand damming of waters (Martin et al. 1993).These swamp forests are subjected to two flooding gradients related to variations in topography.First, there is a gradient over the cross section, in which the forest bottom (i.e., the center, which is topographically lower) is more intensely flooded than are the edges, which seldom flood, and there are changes in physico-chemical characteristics of the soil, the hydromorphic conditions resulting in increased accumulation of organic matter (> 2 m thick in the bottom, Araujo et al. 1998).Second, there is a flooding gradient related to the distance from the lagoons.These flooding gradients are responsible for physiognomic and structural variation within short distances (< 100 m).At the canopy level (a height of approximately 20 m), the most common tree species are Tapirira guianensis Aubl., Calophyllum brasiliense and Symphonia globulifera.The understory is typically not very dense.The main understory species are Protium icicariba (DC.)Marchand and Geonoma schottiana Mart., which proliferate along the forest edges and bottoms, respectively.The trees within swamp forest often have slender trunks, with diameters that rarely exceed 50 cm.
According to local inhabitants, these swamp forests were intensively exploited in the past, mainly for the collection of firewood.In addition, some tree species were extracted for their timber or other products, such as hearts of palm from Euterpe edulis Mart.The interviewees reported that the extraction was more pronounced on the firm ground along the forest edges than on the unstable peaty soil of the bottoms.Nevertheless, some of the swamps have been completely deforested.Until recently, there were reports of illegal burning of swamp forests, which during dry periods destroyed a significant area of those forests, mainly due to the difficulties in controlling the burning of dry peat.As a result, the local swamp forests are now a mosaic of different successional stages.

Sampling
The study was conducted between 2002 and 2004 at eight randomly chosen sites of periodically flooded swamp forests associated with different lagoons (Fig. 1).Aiming to include a wide range of environmental conditions, we established three sets of three 4 × 50 m quadrats at each site: one set at each edge and one in the bottom.Therefore, we established a total of 72 quadrats (nine per site), corresponding to a total sampling area of 1.44 ha.Edges and bottoms were distinguished by topographic, edaphic and physiognomic characteristics.We measured diameter at breast height (DBH, 1.3 m above the ground) and sampled only those trees with a DBH  5 cm.For trees that forked or branched below 1.3 m, only the branches with a DBH  5 cm were measured.We used the elongated quadrats in order to match the long, narrow shape of the swamps (the longer side followed the main axis of the swamps).The dimensions of the quadrats were adapted, from Gentry (1982), to our inclusion criterion.
The botanical material was identified on the basis of the specialized literature, by comparison with specimens in the collections of the Herbarium of the Research Institute of the Rio de Janeiro Botanical Garden (code, RB) and, when possible, through consultation with specialists.Voucher specimens were deposited in the RB.Species were classified in accordance with the Angiosperm Phylogeny Group III guidelines (APG III 2009), and the classification of genera by family followed Souza and Lorenzi (2008).

Phytosociological structure
For each species, we calculated the relative values of density, frequency, dominance and importance value (Mueller-Dombois & Ellenberg 1974).The Shannon diversity and evenness indices (H' and J' , respectively), using the natural logarithm, followed Zar (1996).The calculations were performed with the FITOPAC program, version 1.6 (Shepherd 2006).

Phytogeographic relationships
The Jaccard similarity coefficient (Magurran 1988) was used for floristic comparisons between the swamp forests within the study area and other vegetation formations described in previous studies conducted in Brazil.These comparisons included two categories: neighboring vegetation formations (areas of dry restinga and Atlantic Forest); and floodplain forests (other swamp forests and flooded riparian forests).We do not intend, obviously, to include all studies of such vegetation formations, but rather to select those considered appropriate for the establishment of phytogeographic patterns.Phytosociological surveys were the main sources of data (Tab.2).Unidentified species were excluded, as were allochthonous species.According to Mueller-Dombois & Ellenberg (1974), Jaccard similarity coefficients  0.25 indicate floristic similarity between the units compared.The names of all species were reviewed in accordance with the List of Species of the Brazilian Flora (Lista de Espécies da Flora do Brasil 2012) or by consulting specialists.In addition, we calculated the percentages of species within the study area that had also been registered in the same types of vegetation formations used for comparison by the Jaccard similarity coefficient.

Phytogeographic relationships
The swamp forests studied were found to bear little similarity to the types of vegetation formations used for comparison (Tab.2).In most cases, the Jaccard similarity coefficients were lower than 0.25; as expected, the highest values were obtained only when our data were compared with those of two other surveys conducted in swamp forests within our study area (Jaccard, 0.39 and 0.38, respectively).
Our swamp forests were more similar to, or at least as similar as, open restinga formations within the park, such as the so-called Clusia scrub (Jaccard of 0.21-0.17)and Ericaceae scrub (Jaccard, 0.15), as they were to other swamp forests of the coastal plains of the states of Rio de Janeiro, São Paulo and Paraná (Jaccard, 0.18-0.03).Conversely, the swamp forests in our study area presented very low similarity to other swamp forests, including those in the Brazilian Highlands (in the Federal District of Brasilia, the state of Minas Gerais and the state of São Paulo: Jaccard, 0.07-0.03)and those in the coastal plains from the southernmost Brazilian state of Rio Grande do Sul (Jaccard, 0.12-0.05),especially in the southernmost portion of those plains (Taim wetland: Jaccard, 0.05).The Jaccard similarity coefficients for comparisons between our swamp forests and areas of Atlantic Forest sensu stricto in the states of Rio de Janeiro and Espírito Santo or various flooded riparian forests in Brazil were consistently quite low (0.11-0.02 and 0.08-0, respectively).
Despite the low Jaccard similarity coefficients, many of the species surveyed in the swamp forests of Restinga de Jurubatiba National Park were also registered in the types of vegetation formations used for comparison (Tab.3 and 4).Some 70% of the species were shared with neighboring vegetation formations, including open restinga within the park (33.3%) and the Atlantic Forest sensu stricto in the states of Rio de Janeiro and Espírito Santo (53.1%).Considering the floodplain forests, many species were common to other swamp forests (54.3%) and flooded riparian forests (37%).In addition, excluding the flooded-forest specialist species Calophyllum brasiliense, Symphonia globulifera and Tabebuia cassinoides, the great majority (78%) of the species common to other swamp forests also occurred in the Atlantic Forest sensu stricto.In the case of species common to flooded riparian forests, the percentage was also very high (78.6%,Tab. 3).

Phytosociological structure
The results indicated an oligarchic structure in the swamp forests studied.In addition to the known flooded-forest specialist species (Calophyllum brasiliense, Symphonia globulifera and Tabebuia cassinoides), some species common to neighboring areas of dry restinga (Tapirira guianensis and Protium icicariba) and areas of Atlantic Forest (Euterpe edulis  Barros (2000) and Oliveira (2000) conducted in swamp forests within the park were excluded.Barros (2000) and Oliveira (2000) conducted in swamp forests within the park were excluded.
and Geonoma schottiana) showed high densities and are very important to the structure of the forests studied (Tab.3).This indicates the generalist character of these species in relation to the water content of the soil.In a swamp forest, the saturation or inundation of the soil, caused by a rise in the water table, exerts strong selective pressure (Marques et al. 2003;Rocha et al. 2005;Toniato 2006), determining the floristic composition and structure of the forest.The saturated soil restricts the number and abundance of shrub and tree species that are able to establish themselves, whereas it favors occupation by flooded-forest specialists, as well as some generalist species (Marques et al. 2003).A higher degree and longer duration of saturation in the superficial soil layers translates to greater selective pressure.As a result, swamp forests frequently show oligarchy, or monodominance (Araujo et al. 1998;Galvão et al. 2002;Scarano 2006), which are common in habitats subjected to extreme environmental conditions (Richards 1979;Scarano 2002).In fact, species richness, diversity and evenness are lower in swamp forests than in adjacent or nearby forests located on drier soils and thereby spared the direct influence of the water table level (Ivanauskas et al. 1997;Sztutman & Rodrigues 2002;Carvalho et al. 2006b).
In our study area, the species richness and diversity were relatively high when compared with those of other swamp forests (Tab.1), which is attributable to the broad range of environmental conditions included in our sample, especially as regards the phreatic flooding regime (from none up to several months) as a result of small topographic variations.This situation enabled the establishment of species with different ecological requirements-from flooded-forest specialists to generalists, and even species common to the drier soils present in neighboring vegetation formations (see below).However, the considerable differences in abundance of these species led to the low evenness value.
The total density and basal area were low when compared with those reported in other studies using identical or comparable inclusion criteria (Tab.1); even studies using the more rigid inclusion criterion of DBH ≥ 10 cm (Galvão et al. 2002;Carvalho et al. 2006b) indicated values for basal area much higher.Some other swamp forests (IM1, IM2 and IM3) located in areas of restinga, like those studied here, also showed values for basal area that were higher than that found in our study area.These results are possibly related to the history of use of our swamp forests, as mentioned earlier, although some of their equivalents used for comparison have also been subjected to explotation (e.g., PA2, PST and BAT).
The number of dead trees we found represented only a small part of the renewal dynamics of the local swamps, given that our analysis did not include fallen trees that were still alive nor mass mortalities caused by sporadic events.These seem to be particularly applicable to swamp forests.As an example, soon after the end of our survey, an extremely strong wind that lasted for only a few minutes was responsible for a sharp decline in the number of trees at site E.This was attributable to the instability of peaty soils and the shallow root systems of the trees, in response to the proximity of the water table.Trees located along the upper stretches of the forest edge, where there is less accumulation of peat, were less affected.On the other hand, the combination of heavy rains and the fact that the sandbar separating the Preta lagoon from the sea failed to open subjected site C and the neighboring swamps to an exceptional high water level and long-term flooding (2005)(2006), coinciding with the massive death of trees.
These examples illustrate the great fragility of swamp forests, as has been reported (e.g.Scarano et al. 1998;Jacomine 2004).They are particularly sensitive to changes in the flood regime, and when these occur due to human activity and become permanent, in addition to rapid degradation, the forests show no natural recovery (Scarano et al. 1998).In fact, the narrowing of river channels, the construction of embankments and the installation of drainage systems have been responsible for the virtual disappearance of the swamps on the coastal plains in the state of Rio de Janeiro.Fortunately, since the establishment of Restinga de Jurubatiba National Park in 1998, the human activities that were damaging the local swamp forests (cutting, burning and hunting) have diminished and are currently almost nonexistent.

Phytogeographic relationships
The floodplain forests of Brazil (including flooded riparian forests and swamp forests) occur under a variety of ecological conditions-including those related to climate, geology, geomorphology, soil, water, flooding and adjacent floras-which induce variations in their floristic composition, physiognomy, structure and dynamics (Prance 1979;Mantovani 1989;Ivanauskas et al. 1997;Parolin et al. 2004;Rodrigues & Shepherd 2004;Scarano 2006).Studies using multivariate techniques have further indicated that swamp forests differ floristically from flooded riparian forests, mainly due to differences in flooding regimes (Rodrigues & Nave 2004;Silva et al. 2007).Even within a single site, differences in topography, flooding intensity and soil conditions affect the spatial distribution of species and promote phytosociological variation (Keel & Prance 1979;Vilela et al. 2000;Sztutman & Rodrigues 2002;Damasceno-Junior et al. 2005;Scarano 2006;Teixeira et al. 2008).The floristic differences related to this great heterogeneity of ecological conditions were responsible for the low similarity between various Brazilian floodplain forests and the swamp forest studied here, as well as for the relatively high species richness found in our study area, in comparison with other swamps (Tab.1).
Our results indicate that the swamp forests in our study area bear a stronger floristic resemblance to their closest neighbors (areas of dry, open restinga within the same park) and to a few other swamp forests than to the adjacent Atlan-tic Forest sensu stricto in the mountain chains and flooded riparian forests at the regional or national level.Therefore, like ecological conditions, geographic proximity represents a key determinant of floristic composition.In addition, despite the low similarity between our swamp forests and areas of Atlantic Forest sensu stricto, a high proportion (53.1%) of species we found is shared with this forest.Furthermore, in comparison with other swamp forests, those studied here were found to be most similar to those located in the coastal plains of southeast Brazil (in the states of Rio de Janeiro, São Paulo and Paraná), which are geographically and climatically more similar than are swamp forests elsewhere.Moreover, in those coastal plains, the similarity decreased in relation to swamp forests that showed lower species richness, which are often degraded or early successional forests, such as the Passa-Sete and Batuva locations studied by Galvão et al. (2002), or related to specific edaphic conditions, such as those observed in the deep peat forest studied by Sztutman & Rodrigues (2002).All of this leads to the assumption that the swamp forests of southeastern Brazil do not consist of a distinguishable floristic unit.We argue that this is related to potential sources of local variation, notably migration and establishment of adaptive species from neighboring vegetation formations.Topographic variations at the local level and successional stage would control the establishment of these species at a given site.
Our results clearly indicate that the floristic composition of the studied swamp forests is largely related to the surrounding vegetation, i.e., areas of dry restinga and Atlantic Forest (≈70% of the species surveyed are shared with those vegetation formations).A similar pattern has often been reported for other floodplain forests in Brazil (Gibbs & Leitão-Filho 1978;Ivanauskas et al. 1997;Marques et al. 2003;Rodrigues & Shepherd 2004;Carvalho et al. 2006b;Scarano 2006;Teixeira & Assis 2011).This calls attention to the ecological plasticity of some tree species that occur in the surroundings of floodplain forests.Marques et al. (2011) studied the coastal lowland vegetation of southern and southeastern Brazil.Using detrended correspondence analysis, the authors showed that swamp forests did not form a distinct group and tended to be more similar to neighboring terra firme forests and scrublands.This result supports our previous assumption that the swamp forests of southeastern Brazil do not consist of a distinguishable floristic unit.
Another interesting result of the present study is that a high proportion (78%) of the species shared with other swamp forests also occurred in the Atlantic Forest sensu stricto.This clearly indicates that the floristic connection between the swamp forests studied here and their equivalents in southern and southeastern Brazil is primarily attributable to Atlantic Forest species.
The fact that the flora of our study area, like those of other swamp forests (B.C.Kurtz, unpublished data), did not consist predominantly of flooded-forest specialists (even in the more severely flooded bottoms) is at first surprising.Higher plants, like any other aerobic organism, need an adequate supply of oxygen for the performance of vital functions, such as cell division (Hendry & Crawford 1994).However, even terrestrial plants, such as trees, are capable of surviving under conditions of a partial or total lack of oxygen, and there is considerable interspecific variation regarding the length of time they can do so, ranging from hours to years (Crawford 1992).In order to colonize a floodplain habitat, a given plant must first overcome the hazards of dispersal, germination and youth, resorting to biochemical, physiological, morphological or anatomic mechanisms in order to tolerate or avoid the deleterious effects of oxygen deprivation during growth (Scarano 2006).
The wide range of environmental conditions in our study area, particularly in relation to the phreatic flooding regime conditioned by topographic variations, allowed for the establishment of species with different ecological requirements-including generalist species from the neighboring areas of dry restinga or Atlantic Forest (e.g., Tapirira guianensis, Protium icicariba, Euterpe edulis, Geonoma schottiana, Pera glabrata, Sloanea guianensis, Calyptranthes brasiliensis, Pseudobombax grandiflorum, Miconia cinnamomifolia, Simarouba amara, Ficus organensis, Aniba firmula, Xylopia sericea, Myrcia multiflora and Alchornea triplinervia)-and contributed decisively to the high local species richness.Even apparently flood-intolerant species (Jacaranda bracteata, Clusia hilariana and Eugenia excelsa) were established on the seldom-flooded or unflooded patches along the forest edges.In fact, some studies evaluating broad flooding regimes, related to topographic variations, also showed high species richness in Brazilian floodplain forests, including swamps (Rocha et al. 2005;Guedes-Bruni et al. 2006a;Teixeira et al. 2008) and flooded riparian forests (e.g.Carvalho et al. 1995;Vilela et al. 2000).
The ecological succession in swamp forests, which is controlled by the phreatic flooding regime, is also responsible for floristic changes.This can be observed on the coastal plain of southeastern Brazil, where early successional stages, which correspond to permanently flooded sites, are frequently dominated by Tabebuia cassinoides (Gentry 1992;Araujo et al. 1998;Scarano 2006).This species typically forms pure stands (Gentry 1992).However, in the more advanced stages (as in our study area), the soils are subjected only to periodic flooding, which, together with the occurrence of subtle topographic variations, creates a local mosaic of flooded and unflooded patches that varies over the years depending upon the amount of rainfall (see Scarano et al. 1997).As previously mentioned, these conditions promote higher species richness.Therefore, throughout the successional process, the influence of neighboring vegetation formations (areas of Atlantic Forest or restinga, depending on the location) will be progressively stronger.Galvão et al. (2002) showed profound changes in composition, diversity and structure over the course of the successional process of swamp forests in the coastal plains of the state of Paraná.
These changes included a decline in T. cassinoides dominance and an increase in the importance of other tree species, especially Calophyllum brasiliense.
Despite recent advances, knowledge of the ecology of swamp forests in Brazil continues too limited, and there is a need for additional information on their floristic composition, dynamics and phytogeographic relationships.Further local surveys should be encouraged, especially in the northern and northeastern regions, for which there is total lack of quantitative data.Such surveys will enable new phytogeographic analyses including various sets of sites.

Figure 1 .
Figure 1.Map of Restinga de Jurubatiba National Park, in the state of Rio de Janeiro, Brazil, showing coastal lagoons and the Atlantic Ocean (light gray); forest formations (dark gray); and the eight swamp forest study sites (A-H).In the detail, aerial photograph of the southwestern edge of the park, showing the swamp forests associated with the Jurubatiba lagoon (photo: Romulo Campos).

Table 1 .
Quantitative parameters of the tree/shrub-tree layer of swamp forests in southern and southeastern Brazil.

Table 2 .
Jaccard similarity coefficients between the swamp forests studied in Restinga de Jurubatiba National Park, in the state of Rio de Janeiro, and other vegetation formations in Brazil.
* Mean distance to the eight swamp forests studied.

Table 3 .
Tree species (diameter at breast height ≥ 5 cm) surveyed in swamp forests of Restinga de Jurubatiba National Park, in the state of Rio de Janeiro, Brazil, ordered by importance value, phytosociological parameters and occurrence (1/0) in some other vegetation formations in Brazil (based on the surveys presented in Tab. 2).RD -relative density; RDo -relative dominance; RF -relative frequency; IV -importance value (%); ORP -open restinga within the park; ATF -Atlantic Forest sensu stricto (in the states of Rio de Janeiro and Espírito Santo); SWF -swamp forests in southern and southeastern Brazil; FRFflooded riparian forests.
*The surveys of

Table 4 .
Percentage of species in the swamp forests of Restinga de Jurubatiba National Park, in the state of Rio de Janeiro, that were also registered in other vegetation formations in Brazil and the corresponding Jaccard similarity coefficient intervals.
N -number of surveys used in the analysis.*The surveys of