Introduction

The Iguaçu National Park (INP) is located in the Alto Paraná Atlantic Forest Ecoregion, the largest among the 15 ecoregions identified in the Atlantic Forest biome (^{Di Bitetti et al. 2003}). It covered originally 471,204 km² extending from the western side of the Atlantic mountain range, in Brazil, to eastern Paraguay and northeastern Argentina. Its predominant vegetation type is the seasonal semideciduous forest (^{Di Bitetti et al. 2003}). The vegetation cover in this Ecoregion was reduced to 7.8% of its original area, the greatest reduction being observed in Brazil where forest cover declined to only 2.7% of its original area; followed by Paraguay and Argentina, with 13.4% and 50% of remaining vegetation, respectively (^{Di Bitetti et al. 2003}). According to these authors, Argentina preserves the largest continuous forest cover in the Alto Paraná Forest Ecoregion, comprising much of the Misiones Province. In the border region between Brazil and Argentina, the Iguaçu National Park stands out as the largest protected area covering a territory of 185,262 ha (^{Di Bitetti et al. 2003}). The native forest cover in this region constitutes the largest continuous remnant of seasonal semideciduous forest in the context of the Atlantic forest biome (^{Galindo-Leal & Câmara 2005}).

The functional connection between the INP and other protected areas in Argentina and Paraguay as well as in the Brazilian States of Paraná, Santa Catarina, Rio Grande do Sul, São Paulo and Mato Grosso do Sul represents a highly favorable condition for the conservation of theAtlantic Forest biodiversity (^{Di Bitetti et al. 2003}, ^{Galindo-Leal & Câmara 2005}). To this end, some proposals have been discussed, such as the Biodiversity Corridor of the Paraná River (^{Limont et al. 2015}) and the Biodiversity Conservation Landscape for the Alto Paraná Forest Ecoregion (^{Di Bitetti et al. 2003}). These are considered long-term strategies for biodiversity conservation through the maintenance of the ecological process of the biota (^{Di Bitetti et al. 2003}).

An important region for the implementation of those strategies is located between the INP and the southern limit of the permanent preservation area of the Itaipu Lake (PPAIL), in western Paraná State, where the connection between the northern and southern sectors of the Biodiversity Conservation Landscape for the Alto Paraná Forest Ecoregion could be established (^{Di Bitetti et al. 2003}). The PPAIL consists of a 2,900 km long strip, with an average width of 217 m of reforestation or native forest remnants that runs along almost the entire length of the Itaipu Hydroelectric reservoir margins (^{Itaipu Binacional 2016a}).

However, the connection between the INP and the PPAIL is hindered by the eminently agricultural land use in the region and by the busy BR 277 Highway, which separates apart these two protected areas. The map of the Biodiversity Conservation Landscape for the Alto Paraná Forest Ecoregion indicates the establishment of a biological corridor in this region (^{Di Bitetti et al. 2003}). An example of a biological corridor already established there is the Santa Maria corridor consisting of remnants of native vegetation and a 60-m width reforestation strip in the Bonito River Basin (^{Itaipu Binacional 2016b}). Notwithstanding, a wider biological corridor should yet be designed in the region in order to warrant an effective connection between the two sectors of the Biodiversity Conservation Landscape for the Alto Paraná Forest Ecoregion (^{Di Bitetti et al. 2003}). Furthermore, the simple recovery of native forests in the river margins and other permanent preservation areas in small rural properties can greatly enhance the functional connectivity at the landscape scale in the region (^{Seganfredo 2015}).

Either for the establishment of a biological corridor between the INP and the PPAIL, or for the recovery of permanent preservation areas in the region, there will be the necessity of forest restoration (sensu^{Aronson et al. 2011}). Floristic knowledge about native vegetation remnants is critical to the choice of the most suitable species for restoration projects (e.g.^{Monge 2009}); to avoid the use of exotic species, and enable high diversity forest restoration. The current floristic knowledge in the region covers the surroundings of the Santa Maria biological corridor, in the Santa Terezinha de Itaipu municipality (^{Gris 2012}), at an average altitude of 270 m above sea level; and specific areas inside the INP, at altitudes ranging from 150 to 750 m above sea level (^{Souza 2015}).

The present contribution aimed to extend the regional floristic knowledge in the region between the INP and the PPAIL to a 440 km² study area covering an altitudinal range expanding from 200 m to 600 m above sea level in the municipalities of Céu Azul, Matelândia, Medianeira and Serranópolis do Iguaçu. Beyond a floristic list, we aimed to provide useful information compiled from the literature for the application of the list in restoration projects in the study area, such as successional category, dispersal syndrome, conservation status and indication of locally abundant species.

Material and methods

Study area - The study area covers predominantly the municipality of Matelândia and, to a lesser extent, of Céu Azul, Medianeira and Serranópolis do Iguaçu, in Paraná State, Brazil. For methodological convenience, we defined the study area as the polygon circumscribed by the Highways BR 277 and PR 495 and by the northwestern border of the INP, covering an area of 440 km² (figure 1). The average annual temperature in the region is 21.6 ºC, with average annual rainfall of 1,803 mm. Climate is subtropical, Cfa according to Köppen classification system (^{EMBRAPA 2011}). Vegetation in the native forest fragments is seasonal semideciduous forest (^{Di Bitetti et al. 2003}) due to the predominance of phanerophytes, the subtropical climate seasonality and the leaf shedding of part of the tree individuals (^{Instituto Brasileiro de Geografia e Estatística 2012}). Seasonal semideciduous forests of localities inside the INP, near the study area and extending through a similar altitudinal range, were classified as typical and humid submontane forests by ^{Souza (2015)}.

Source: Fundação SOS Mata Atlântica and ^{INPE (2016)}.

Figure 1 Study area, western Paraná State, Brazil. The area consists of a polygon delimitated by BR 277 and PR 495 highways, and by the northwestern limit of the Iguaçu National Park. The localities of botanical records are indicated by stars. 

Floristic survey - We visited 16 native forest fragments located in the study area from October 2013 to December 2015 during field campaigns that had an average duration of 3 days and monthly periodicity. The number of visits to each fragment along the study period varied, being greater in larger and less disturbed ones. As our objective was to describe the floristic composition of the study area as a whole and not of each fragment individually, we did not compiled an individual floristic list of each fragment, so frequency of species across fragments in the area is not available. Fragments were chosen taking into account the sampling of the entire study area with a distribution of sampling effort as homogeneous as possible, and also considering the access authorization by landowners. We also sampled the northwestern edge of the Iguaçu National Park between the limits of the Park and the so called Old Guarapuava Road between Serranópolis do Iguaçu and Céu Azul municipalities, walking the length of this road in that stretch. The forest fragments studied have different disturbance histories and include secondary forests and primary forests with different levels of degradation (degraded primary forest sensu^{Aronson et al. 2011}). The vegetation of the northwestern edge of the INP was considered as secondary forest. The method used for the floristic survey was based on the walking method (^{Filgueiras et al. 1994}), considering only the tree and shrub components and using trails already existent in the forests. The botanical material was collected with the use of a telescopic pruner and all collections were geo-referenced using Global Positioning System (GPS). In addition to the forest fragments and the INP edge, we also collected botanical material along the rural roads in the study area.

We processed the botanical material according to the conventional techniques (^{Fidalgo & Bononi 1989}). Both fertile and vegetative materials were identified by comparison with herbarium specimens deposited in the Forestry Institute of São Paulo State (Dom Bento José Pickel Herbarium -SPSF), and by consulting the specialized literature. The voucher herbarium sheets were deposited in the SPSF Herbarium, with duplicates deposited in the herbarium of the Federal Technological University of Paraná - Campus Medianeira (Figueira Herbarium - FIG). Only fertile materials were deposited in the SPSF herbarium, while 27 vegetative materials, corresponding to 27 species, were deposited in the FIG herbarium. These species were not found in reproductive phase during the study period, but we decided to incorporate the respective voucher to the FIG herbarium considering the absence of herbarium sheets of these species in this recently established collection (^{Cielo-Filho et al. 2016}).

We sorted the species into families and genera according to the APG IV system (^{APG IV 2016}). To check spelling, corrected and accepted names we consulted the List of Species of the Brazilian Flora (^{Brazilian Flora 2020 in construction 2016}). For most species, plant habits classification was obtained in the List of Species of the Brazilian Flora, but we follow ^{Souza et al. (2013)} for palm and fern, although we counted these habits as tree in the results. Verification of endangered species was based on the Red List of Endangered Species of the International Union for Conservation of Nature (^{International Union for Conservation of Nature 2016}) and the Official List of Endangered Species of the Brazilian Flora (^{Brasil 2014}), which was based on the list presented by ^{Martinelli & Moraes (2013)}. We did not use the Red List of Paraná State (^{Paraná 1995}), since this list is out of date. With the aim of supporting future forest restoration projects and the establishment of a biological corridor in the study area, we compiled information concerning the successional category (sensu^{Swaine & Whitmore 1988}) and dispersal syndrome (sensu^{Pijl 1982}) of the identified species from the literature (e.g.^{Lorenzi 1992}, ^{Gandolfi et al. 1995}, ^{Lorenzi 1998}, ^{Silva & Soares-Silva 2000}, ^{Carvalho 2003}, ^{Carpanezzi & Carpanezzi 2006}, ^{Carvalho 2006}, ^{Carvalho 2008}, ^{Lorenzi 2009}, ^{Carvalho 2010}, ^{Barbosa et al. 2015}). We also indicated the native species found in the present contribution which were also registered, as locally abundant species, in phytosociological studies available for the INP and its surroundings (^{Gris 2012}, ^{Souza 2015}). To do this, we considered species occupying the first ten positions in the importance value ranking of a typical submontane forest in the Santa Maria Natural Patrimony Private Reservoir (^{Gris 2012}), and of typical and humid submontane forests in the interior of the INP (^{Souza 2015}).

Results

We performed 940 botanical records of native species, 303 in primary forest and 520 in secondary forest, averaging 51 records per fragment. The remainder 117 records were done on the margins of rural roads. We identified 204 tree and shrub native species from 51 families and 134 genera (table 1). The arboreal habit was represented by 174 species, while the shrubby habit by 96 species. Sixty six species presented both habits. The richest families were Fabaceae (29 species), Myrtaceae (18), Solanaceae (10), Euphorbiaceae, Meliaceae and Rutaceae (nine species each). The richest genera were Eugenia (eight species), Solanum and Zanthoxylum (six species each), Ocotea, Trichilia and Aegiphila (four species each). The most common dispersal syndrome was zoochoric (141 species, 69%) followed by anemocoric (35 species, 17%) and autochoric (28 species, 14%) syndromes. Most (130) species (64%) belonged to the non-pioneer successional category and the 74 remainder (36%) was classified as pioneer. Of the 204 recorded species, seven are threatened due to habitat loss and overexploitation (table 2). Additionally to the native species, we found 15 exotic species, of which eight were recorded inside forest fragments and seven in more anthropogenic environments such as roadsides, gardens and pastures (table 3).

In table 1 we indicated 14 species reported as locally abundant in Typical Submontane Forest and 10 species reported as locally abundant in Humid Submontane Forest, according to phytosociological studies available for the INP and its surroundings (^{Gris 2012}, ^{Souza 2015}). Of the mentioned locally abundant species, only four are pioneer ones: Alchornea triplinervia, Dendropanax cuneatus, Sebastiania brasiliensis and Luehea divaricata.

Discussion

An important aspect of the present floristic survey was the apparently low species richness, when compared to other studies conducted in Atlantic forests at lower latitudes. For instance, surveys of tree and shrub flora in the seasonal semideciduous forest, carried out in protected areas in the municipalities of Paranapanema and Itapeva, both in São Paulo State, registered 349 and 317 species, respectively (^{Cielo-Filho et al. 2009}, ^{Souza et al. 2012}). However, as pointed out by these authors, the high richness in those areas may be related to the presence of Cerrado remnants in the region, which would expand the regional species pool, increasing the richness in remnants of seasonal semideciduous forest. Even without the influence of the Cerrado flora, the number of tree and shrub species (266) reported by ^{Rossetto & Vieira (2013)} for the seasonal semideciduous forest of the Mata dos Godoy State Park, northern Paraná State, was higher than that found in the present study.

It could be suggested that the flora in northern remnants of seasonal semideciduous forest of the Paraná State is richer than in the remnants located at higher latitudes of the State, due to more rigorous climatic conditions. For instance, the number of tree species (197) reported for the Mata dos Godoy State Park, where a comprehensive floristic study in a well conserved forest remnant was carried out (^{Rossetto & Vieira 2013}), was greater than that found in our study (174), as well as in the studies of ^{Souza (2015)} and ^{Gris (2012)} which identified, respectively, 165 and 112 native tree species in the INP and forest fragments nearby. However, when considered together, the floristic data available for the study area and its surroundings - this study, ^{Gris (2012)} and ^{Souza (2015)} - summed up 228 native tree species, what suggests that the tree flora in northern remnants of seasonal semideciduous forest of the Paraná State is not richer than in southern ones.

Another aspect to be considered in these richness comparisons is the differences concerning methods and anthropogenic disturbance histories among study areas. For instance, the use of an inclusion criterion to sampling plant individuals - stem perimeter at breast height equal or more than 15 cm - by ^{Gris (2012)} and ^{Souza (2015)} may have limited the number of species recorded in those studies. This kind of methodological limitation may also have restricted the tree species richness found in other studies, such as in the Rio das Cobras Farm, southwestern Paraná State - 128 species (^{Viani et al. 2011}); in protected areas of the Argentine Province of Misiones: Iguazú National Park and its surroundings -72 species (^{Chediack 2008}) and Osununú Private Reserve - 96 species (^{Velazco et al. 2015}); in protected areas of the Paraguayan side of the area of influence of the Itaipu Lake - 127 species (^{Monge 2009}); and in the São Francisco State Park, northern Paraná State, where 113 tree species were recorded, although in this case the relatively low richness seems to be also related to the intensive practice of selective logging before the protected area had been established (^{Tomé et al. 1999}, ^{Zama et al. 2012}).

It is worth to note that 43 native tree species recorded in the present floristic survey (table 1) were not registered by ^{Gris (2012)} and ^{Souza (2015)} in the INP and its surroundings. This may be a consequence of sampling effect, but also of floristic particularities of the study area in relation to its surroundings. Of that 43 species, 70% are non-pioneer, 72% have zoochoric dispersal syndrome, while the other two dispersal syndromes comprised each, 14% of the remainder species. Comparing these data with the corresponding percentages observed for the whole set of species in this study, there is no evidence of bias concerning successional category or dispersal syndromes for the set of species exclusive of the present research. However, in relation to the distribution of exclusive species among families, there was a highlight for Myrtaceae, with seven of the 18 species found being exclusive. This may be a consequence of the relatively complex taxonomy of this family. Another highlight may be perceived for the distribution of species between habits: 56% of the exclusive species present both arboreal and shrubby habit, while the corresponding value for the whole set of species in this study is of 32%. We suggest that because we did not use an inclusion criterion to sampling plant individuals as the other studies referred above - stem perimeter at breast height equal or more than 15 cm - we attained a better representation of the flora composed by species presenting simultaneously the arboreal and shrubby habits. Furthermore, the overrepresentation of the shrubby-arboreal habit among the exclusive species may be a consequence of the intensive sampling of edge habitats in the present study, as indicated by the greater percentage of shrub-tree-pioneer species in the set exclusive species (9/43 = 21%) than in the whole set of species (29/204 = 14%). Among the exclusive species, there was one threatened, Myrcianthes pungens, found in both primary and secondary forest fragments in the study area, what stress the importance of these remnants to biodiversity conservation.

On the other hand, 54 native tree species were registered by ^{Gris (2012)} and ^{Souza (2015)}, but not by the present study. The data set resulting from the results of those authors complies study sites at altitudes ranging from 150 to 750 m above sea level, while the present study area covers an altitudinal range expanding from 200 m to 600 m above sea level, and 68% of our botanical records were done in study sites located at altitudes between 300 m and 500 m. The weak representation of the regional altitudinal extremes in our data set may help to explain the absence in our species list of the 54 species exclusive of the lists reported by ^{Gris (2012)} and ^{Souza (2015)}. Another possible explanation comes from the landscape context: our study sites are situated in a highly fragmented landscape in which some species may have been lost because of edge effect, dispersal barriers and other process associated with forest fragmentation leading to local species extinction (^{Primack & Rodrigues 2001}).

The distribution of species richness among families in the study area was similar to that found in other seasonal semideciduous forest studies in the Brazilian States of São Paulo (^{Cielo-Filho et al. 2009}, ^{Souza et al. 2012}) and Paraná (^{Viani et al. 2011}, ^{Gris 2012}, ^{Zama et al. 2012}, ^{Rossetto & Vieira 2013}); in the Province of Misiones, Argentina (^{Chediack 2008}, ^{Velazco et al. 2015}); and in the Departments of Canindeyú and Alto Paraná, Paraguay (^{Monge 2009}). However, an interesting feature of the study area was the relatively low representativeness of Lauraceae and the relatively high richness of Solanaceae. This could suggest a particularity of the flora of seasonal semideciduous forest remnants located in higher latitudes of the Paraná State, but, instead, seems to be a consequence of forest fragmentation or sampling effect, as discussed above. For instance, we recorded only six species of Lauraceae, but this number increases to 10 considering the results of floristic surveys nearby (^{Gris 2012}, ^{Souza 2015}), equaling the number found in the Mata dos Godoy State Park in northern Paraná (^{Rossetto & Vieira 2013}). In relation to Solanaceae, the number of tree and shrub species found in the Mata dos Godoy State Park, 19, was even higher than that found in the study area and its surroundings, 15 (this study, ^{Gris 2012}, ^{Souza 2015}). The presence of Fabaceae and Myrtaceae in the first two richness ranking positions is a characteristic shared by other studies of the tree and shrub flora in seasonal semideciduous forests of Argentina, Brazil and Paraguay (^{Cielo-Filho et al. 2009}, ^{Monge 2009}, ^{Souza et al. 2012}, ^{Rossetto & Vieira 2013}, ^{Velazco et al. 2015}).

Despite the above considerations about an apparent lack of particularities in the flora of the study area concerning species richness and families representativeness, a botanist familiarized with the flora of seasonal semideciduous forest remnants located at lower latitudes that came to visit forest remnants in the study area and its surroundings, will still encounter some peculiarities of the tree and shrub flora of these higher latitudes remnants concerning species composition. For instance, she/he will miss some notorious species like, among others: Croton floribundus Spreng, Copaifera langsdorffii Desf., Esenbeckia febrifuga (A.St.-Hil.) A.Juss. ex Mart., Gallesia integrifolia (Spreng.) Harms, Machaerium nyctitans (Vell.) Benth., Piptadenia gonoacantha (Mart.) J.F.Macbr., Zeyheria tuberculosa (Vell.) Bureau ex Verl., Geonoma schottiana Mart. (and Arecaceae in general), Moquiniastrum polymorphum (Less.) G. Sancho (and Asteraceae in general), Miconia theizans (Bonpl.) Cogn. (and Melastomataceae in general).

It must be stressed that some species found by ^{Gris (2012)} in plantings done in the Permanent Preservation Area of Itaipu Lake and in the reforestation strip of the Santa Maria biological corridor may not be native of the region. We highlight the following exotic or probably exotic species registered in ^{Gris (2012)}, but not in the present contribution nor in ^{Souza (2015)}: Annona neosericea H. Rainer, Handroanthus impetiginosus Mattos, Croton floribundus, Manihot grahamii Hook., Anadenanthera peregrina (L.) Speg., Calliandra brevipes Benth., Cassia leptophylla Vogel, Copaifera langsdorffii, Machaerium nyctitans, Piptadenia gonoacantha, Poincianella pluviosa (DC.) L.P. Queiroz. and Eugenia gracillima Kiaersk.. Together with the species explicitly stated as exotic of the Brazilian territory in ^{Gris (2012)}, these species should be avoided in future restoration projects in the region.

Corroborating the ranking of representativeness of dispersal syndromes observed in other studies carried out in seasonal semideciduous forest (^{Silva & Soares-Silva 2000}, ^{Santos & Kinoshita 2003}, ^{Kinoshita et al. 2006}, ^{Zama et al. 2012}) and agreeing with the general pattern observed in tropical forests (^{Howe & Smallwood 1982}), the most frequent dispersal syndrome was the zoochory (69%), followed by anemochory (17%) and autochory (14%). The percentage of zoochoric species that we found in this study was higher than that found in other studies in well preserved seasonal semideciduous forests (^{Silva & Soares-Silva 2000}, ^{Santos & Kinoshita 2003}, ^{Kinoshita et al. 2006}), what shows a good overall representativeness of zoochoric species in our floristic sampling. The same can be inferred for the representativeness of non-pioneer species, 64%, a percentage higher than that found in the Mata dos Godoy State Park, where 55.8% of the species found may be considered as non-pioneers, taking the late secondary and climax categories adopted by ^{Silva & Soares-Silva (2000)} together. The Mata dos Godoy State Park shelters a well preserved seasonal semideciduous forest remnant in advanced successional stage (^{Silva & Soares-Silva 2000}, ^{Rossetto & Vieira 2013}). Given the importance of a good representativeness of non-pioneer and zoochoric species in restoration projects, our study presents a promising species list to support initiatives of this nature in the area studied. In line with this assertion is the inclusion of the shrubby habit in the present floristic survey, since the use of non-tree species in restoration projects have been recently encouraged (^{Durigan et al. 2010}, ^{Barbosa et al. 2015}).

The endangered species recorded in this study denote the overall conservation status of the Atlantic forest, especially with regard to the drastic reduction in vegetation cover (^{Ribeiro et al. 2009}) resulting in significant habitat loss. Another aspect that stands out is the overexploitation of natural resources (timber and palm heart in the case of Euterpe edulis). The most critically threatened species we found in the study area was Araucaria angustifolia, considered "critically endangered" and "endangered", respectively, in the IUCN and MMA consulted lists. It should be noted that this species is rare in the study area, occurring mostly in anthropogenic sectors situated on the edge of the INP, in the strip sampled at the margins of the old Guarapuava road. This could suggest that those specimens were planted there by the first settlers who arrived in the area in the 1950s and 1960s. On the other hand, the presence of adult reproductive individuals of Araucaria angustifolia at the beginning of the colonization period was confirmed by ancient settlers in neighborhood of the Marquezita town, but not for other localities in the study area (N.L. Viapiana, personal communication), indicating that the species is native, but was not widespread in that area. In the highest elevations of the INP, above the altitudinal range of our study area, ^{Souza (2015)} identified a transitional zone between seasonal semideciduous forest and ombrophilous mixed forest, where Araucaria angustifolia is one of the most abundant species. Additionally, Ombrophilous Mixed Forests are also encountered not so far away from our study area (less than 150 km), for instance, in the Argentine Province of Misiones (e.g. in the Cruce Caballero Provincial Park (^{Rios 2006})), and in vast tracts of the Southwest region of Paraná State (^{Castella & Britez 2004}). Notwithstanding, as this species is locally rare and not widespread in our study area, we did not recommend its extensive use in restoration projects in this area. On the other hand, we encouraged the widespread use in such projects of all the other endangered species registered in the present study.

The locally abundant species indicated in table 1 are, in general, more prone to become well established in restoration areas and, so, their use in higher proportions can favor the rapid forest recover. In using the indication of locally abundant species in table 1 to design restoration projects, one should consider if the species were reported as locally abundant in typical or humid submontane forest, and apply this last species category in the restoration of sites located in river margins and depressions of the terrain, where soils are more humid, while the first species category must be reserved to the restoration of sites with drier soil conditions. Some of the locally abundant species can be applied in both situations (table 1). Another consideration refers to the successional category of the locally abundant species indicated, the majority of which, non-pioneer. As restoration plantings, in general, apply a greater proportion of pioneer than non-pioneer species, it would be interesting the indication of additional locally abundant pioneer species, besides the four already mentioned. To do this, we listed the following pioneer species considered as locally abundant through visual assessment by the authors (table 1): Tabernaemontana catharinensis, Trema micrantha, Peltophorum dubium, Bastardiopsis densiflora, Heliocarpus popayanensis, Eugenia pyriformis, Piper aduncum, Myrsine coriacea, Casearia sylvestris, Allophylus edulis, Diatenopteryx sorbifolia, Solanum caavurana, Styrax leprosus, Cecropia pachystachya, Aloysia virgata and Pombalia bigibbosa. Naturally, all species listed in table 1 should be considered in restoration plantings in the study area, the ones not cited as locally abundant should be considered for the achievement of high diversity plantings.

Among the exotic species found in the study area, with the exception of Michelia champaca, Gymnanthemum amygdalinum and Syzygium jambos, all others have confirmed invasive potential, especially Tecoma stans, Leucaena leucocephala and Hovenia dulcis (^{Zenni & Ziller 2011}). Such species constitute serious threats to biodiversity conservation in the region and should be targeted by control actions aimed at the eradication of their populations in natural environments. Additionally, actions aiming to prevent anthropogenic disturbances in the forest remnants should be implemented in order to allows secondary forests in the study area to reach more advanced successional stages, since the susceptibility to invasion by exotic species may be greater in forests of earlier successional stages, as suggested for the invasion by Hovenia dulcis in forests of the upper Uruguay region (^{Lazzarin et al. 2015}).

In conclusion, we reported the floristic composition of an area situated in a strategic region for the conservation of the seasonal semideciduous forest in the context of the Atlantic forest biome. In general, regarding species richness, we did not detect any particularity of the flora of the study area and its surroundings in relation to the flora of seasonal semideciduous forest remnants located at lower latitudes. The same can be said regarding species distribution among families, and frequency distribution of dispersal syndromes and ecological groups, for the flora of the study area. However, some peculiarities emerged due to the lack in the study area of certain tree species commonly found in those remnants. We provided information on dispersal syndromes and successional categories for 204 tree and shrub species and indicated 34 locally abundant and seven threatened ones, in order to aid in the design of restoration projects in the study area. We also identified 15 exotic tree species in the area, of which, 12 have invasive potential, requiring eradication programs.