A bird survey in a transitional area between two major conservation hotspots in southeastern Brazil

Transitional areas may have higher species richness than those located within established environments, as these ecotones provide a variety of resources that facilitate the maintenance of specific fauna. However, they tend to be overlooked in wildlife surveys. In this context, the western region of the state of São Paulo in Brazil is important because it contains areas of ecological transition between 2 major conservation hotspots in Brazil: the Atlantic Forest and the Cerrado. We surveyed the avifauna in a transitional area between these domains, using qualitative and quantitative methods between December 2016 and February 2019. We documented 220 species, of which 7 are endemic to the Atlantic Forest and 4 are endemic to the Cerrado. In addition, we documented 9 species that are threatened at some level (regional, national, or global), demonstrating the high conservation value of this area for birds. Therefore, our results show the potential of transitional areas for maintaining regional biodiversity.


Introduction
Transitional areas between communities tend to receive less attention in biodiversity research than distinct ecosystems. Studies at the global level tend to assess biodiversity across delimited areas to manage resources efficiently (Myers et al. 2000). However, this approach overlooks transitional areas as valuable biodiversity refuges, despite the relevance of such areas as centers of ecological variability and speciation (Schilthuizen 2000).
Transitional areas (or ecotones) are defined as areas of abrupt transition along an environmental gradient, where abiotic and/or biotic factors change quickly at a small geographic scale (Kark and Van Rensburg 2006). The steep gradient between 2 or among more environments exerts a significant influence on local biodiversity. Such areas tend to have higher species richness and abundance, because they support overlapping communities that would normally be restricted to isolated ecosystems (Kunin 1998, Lloyd et al. 2000, Kucherova and Mirkin 2001, Baker et al. 2002, Dangerfield et al. 2003, Kark et al. 2007). In addition, transitional areas may contain unique richness (Odum 1953) because the combined diversity of adjacent sites facilitates the maintenance of rare morphological, genetic, and ecological features of biological taxa (Kark and van Rensburg 2006). In addition, some authors hypothesized that transitional areas are zones where new species could emerge more quickly (Moritz et al. 2000), especially in bird taxa (Kark et al. 2007).
In this context, the southeastern region of Brazil plays an important role in the maintenance of bird diversity because it contains several ecotones of Cerrado and Atlantic Forest ecoregions (São Paulo 2017). Both The city is inserted in a transitional region between Atlantic Forest (green) and Cerrado (yellow). C. Area of JBMB, including the locations of point counts (white dots), and the predominant physiognomies found in JBMB. Each number corresponds to a different point count. White numbers refer to the points located inside the regenerating area ecoregions support high biodiversity and high degree of endemisms, with several species being threatened by the high rates of habitat loss in them (Myers et al. 2000). Consequently, this region could be fundamental for the conservation of both ecoregions biodiversity. Information on bird distribution and community structure in such environments is crucial to conservation and management (Kark and van Rensburg 2006) and fauna inventories are critical for understanding the composition and dynamics of such communities (Santos 2003). Therefore, our work presents a detailed inventory of avian taxa in an important ecotone of southeastern Brazil and emphasizes the occurrence of those species that present high conservation value, such as endemic, threatened, or migratory species.

Methods
The Jardim Botânico Municipal de Bauru (hereafter JBMB) is located at Bauru Municipality (22°19′S, 049°04′W) in the midwestern region of São Paulo state (Fig. 1). According to Alvares et al. (2013), this region climate is defined as Cwa by Köppen classification, with 2 well-defined seasons of hot and wet summers versus cold and dry winters. JBMB covers a total area of 321 ha and is one of three conservation units that integrates the Vargem Limpa-Campo Novo Area of Environmental Protection (Bauru 2000), along with the São Paulo State University Legal Reserve (265 ha) and the Ecological Reserve of "Sociedade Beneficente Enéas Carvalho de Aguiar" (217 ha). Together, these three conservation units encompasses more than 800 ha of native vegetation (Cavassan 2013).
There are different vegetal formations (or phytophysiognomies) in JBMB that belong to the Atlantic Forest and Cerrado domains (Cavassan 2013). For instance, Forest Savanna (which is known locally as Cerradão) occupies 280 ha; Semideciduous Forest (Veloso 1992) occupies 7 ha; and Riparian Semideciduous Forest with permanent fluvial influence (which is known as Semidecidual Flooded Forest) occupies 11 ha (Rodrigues 2000). In addition, the JBMB contains about 15 ha Regenerating Forest Savanna (Weiser 2007) (Fig. 2). All these phytophysiognomies are distributed in an area of more than 300 ha, which makes this remnant one of the most important fragments of high priority for conservation in the midwestern state of São Paulo (Kronka et al. 1998, Durigan et al. 2004. To document avifauna in the JBMB, we used 10-min point counts, which had no sampling distance limit (Blondel et al. 1981) and were adapted to tropical regions (Vielliard and Silva 1990). We sampled 30 points separated by a minimum distance of 200 m from each other, distributing them between the phytophysiognomies of JBMB (Table 1). We surveyed the points twice a month, in the mornings (06:00-10:00) and in the afternoons (15:00-19:00). Sampling was conducted monthly in 2 stages. Stage 1 was conducted from December 2016 to December 2017 in the Forest Savanna, Semideciduous Forest, and Regenerating area. Stage 2 was conducted from February 2018 to February 2019 in the Riparian Forest. In total, we completed 120 h of avian surveys.
Visual records were made using Nikon PROSTAFF 7S 10 × 42 binoculars, while sound records were made with a Marantz PMD-667 recorder coupled to a Yoga HT-81 directional microphone. Photographic records of the species were obtained using a Canon Powershot Superzoom 40× digital camera whenever possible. Pairs of researchers (GSC, RMV, RWG) carried out the observations to avoid observation bias. To complement the bird survey, we recorded sporadic contacts occurring outside sampling period, both within the JBMB and adjacent areas. We did not record the abundance of species, because this survey was descriptive in nature. In addition, we integrated the internal database of JBMB species to our list to complement our own recordings.
Patterns of migratory activity were classified according to Somenzari et al. (2018), with some changes. We considered migratory species (MGT) as those that leave their territory after the reproductive season and return to the same area after a cyclical period. Partially migratory species (MPR) included populations with some migratory individuals, while others remained resident in certain regions. The endemism status of each species was based on Bencke et al. (2006) for the Atlantic Forest and on Silva and Bates (2002) for the Cerrado. The Red List category of each species was defined at a global level based on the International Union for Conservation of Nature (IUCN 2019), at a national level based on Livro Vermelho da Fauna Brasileira Ameaçada de Extinção (ICMBIO/MMA 2018), and at a regional level based on the State Decree No. 63.853 (São Paulo 2018). We also identified species subjected to capture and traffic pressure according to Costa and Monteiro (2016). Taxonomic ordering and nomenclature followed the recommendations of the Brazilian Ornithological Records Committee (Piacentini et al. 2015).

Results
Across all surveys and data sources, we recorded 220 species (Table 2) belonging to 22 orders and 54 families. Thirty-six species (16.4% of the total) were documented by photographs (Figs 3-8). Of the 54 families, 23 belonged to the order Passeriformes and 31 belonged to other orders. We observed 190 species (86.3% of the total) in quantitative surveys and 26 species (11.8% of the total) in qualitative surveys. Just 4 species (1.9% of the total) were only found in the JBMB database. Among non-Passeriformes families, the Accipitridae, Columbidae, and Trochilidae had the highest richness (9 species each, 4.1% of the total), followed by Ardeidae and Rallidae (7 species each, 3.2% of the total). Among the Passeriformes, 30 species represented the Tyrannidae (13.7% of the total) and 27 species represented the Thraupidae (12.4% of the total) families. We recorded 31 species (14.2% of the total) with migratory activity, with 2 species being considered migratory and 29 species being considered partially migratory,. In addition, we documented 36 species (16.5 % of the total) subjected to capture and traffic pressure (Costa and Monteiro 2016). We also documented 2 exotic species, Estrilda astrild (Common Waxbill) and Passer domesticus (House Sparrow). Some species that were documented are considered of conservation relevance. Among the species classified as endemic, we recorded 7 taxa restricted to the Atlantic Forest and 4 restricted to the Cerrado.   Bauru (JBMB), with the scientific name and English names. We also display the method of detection (Quan. = Quantitative survey, Qual. = Qualitative survey, Data. = JBMB databases); the predominant physiognomy (Phys.) where the taxa have been found (SEM = Semideciduous Forest, REG = Regenerating area, SAV = Forest Savanna, WET = Wetlands, UBQ = Ubiquitous); the migratory behavior (Mig.) of each taxa (MGT = migratory species, MPR = partially migratory species); the conservation status at the global (IUCN 2019), national (ICMBIO/MMA 2018) and regional (São Paulo 2018) levels, respectively (LC = least concern, NT = near threatened, VU = vulnerable, EN = endangered); and the taxa that suffers from risk of capture and traffic pressures (Traf.) according to Costa and Monteiro (2016).

Species endemic to the Atlantic Forest
Thalurania glaucopis (Gmelin, 1788). This species inhabits primary forest, forest edges, scrub, and suburban areas, such as parks and gardens (Weller 2018). It is an active nectarivore hummingbird and its presence is associated with the phenology of several plant species. During our surveys, we had only documented this species once, in April 2017, in a Semideciduous Forest area, at point no. 02. (Vieillot, 1818). B. ruficapillus is associated with primary and Gallery Forests, and is usually found in pairs or mixed flocks (Snow and Kirwan 2018). We documented it in a Semideciduous Forest near the Vargem Limpa River. Although there is evidence of population decline (BirdLife International 2018), it is abundant in the Semideciduous Forests of the surveyed area. We recorded it at points no. 02, 03, and 04. (Wied, 1831). The species is mostly associated with Ombrophilous Forest, and is regularly found in well-developed secondary growth and Gallery Forests. In São Paulo state, it is associated with humid forests and preferentially inhabits bamboo thickets and dense growth . In the JBMB, we observed a pair of birds near a nest next to the main trail, which were observed for several months in 2017 at point no. 03. (Wied, 1821). This species inhabits closed and humid areas, and is found in understory branches near the ground up to the midcanopy. A. leucophtalmus plays an important role in the formation of mixed flocks, and it is one of the species responsible for the maintenance of group cohesion (Remsen Jr and Kirwan 2018). It is quite common in Brazil, Paraguay, and Argentina (Sick 1997) and it was observed over several months during our surveys, mostly in the humid areas of the Semideciduous Forest, at points no. 02, 25, and 26. (Wied, 1831). T. poliocephalum inhabits the canopy (Sick 1997) and the edges of humid forests, adjacent clearings, and even gardens (Walther 2018). It forages near the ground and in the canopy of low-growing trees. It is not dependent on   common in Deciduous Forests, buriti (Mauritia flexuosa L.f.) palm grooves, and swampy woodland (Snow and de Juana 2018). It has an abundant population in the JBMB, as it is found from primary forests to regeneration areas.

Todirostrum poliocephalum
It was recorded at points no. 06, 07, 08, and 09. (Temminck, 1823). C. cristatellus inhabits dry Cerrado forests, but it is also associated with riparian woods, forest edges, and clearing areas with sparse vegetation. It is found in many modified landscapes, including eucalyptus plantations, borders, and gardens (Anjos 2018). Although it is considered endemic to the Cerrado domain, the distribution of this species has expanded to areas of Atlantic Forest (Dário et al. 2002). We observed this species over several months on Forest Savanna edges at point no. 21. (Vieillot, 1817). This is a Cerrado-endemic species that possibly also dwells in the Cerrado enclaves in the Pantanal and Caatinga domains (Pacheco 2003). In São Paulo state, it is mainly threatened by fragmentation and the loss of native areas,  which are transformed into pastures and agricultural areas. This species was found in several different months during our surveys, and was mostly observed in areas undergoing regeneration, near to points no. 20 and 21.

Saltatricula atricollis
Threatened species. We found 1 species threatened at a global (IUCN 2019), 2 at a national (ICMBIO/MMA 2018), and 8 at a regional level (São Paulo 2018). (Vieillot, 1816). Globally and nationally Near Threatened, and regionally Vulnerable. This species lives in Tropical Evergreen Forests, palm groves, and gallery woodlands (Collar et al. 2018a). However, in São Paulo state, the native forest vegetation is restricted to small fragments (Nunes and Galetti 2007), decreasing its area of occurrence. P. maracana also has low genetic variability (Craveiro and Miyaki 2000), which further threatens its survival. Natural habitat loss and traffic activities are the main factors increasing its risk of extinction (Juniper and Parr 1998). In the JBMB, we directly observed and heard the species only once, in a wetland near the Vargem Limpa River in August 2016, outside the point counts (22°20′39″S, 049°00′54″W).

Primolius maracana
Amazona aestiva (Linnaeus, 1758). Nationally and regionally Near Threatened. This species inhabits Cerrado, Chacos, and subtropical forests, and is restricted to primary forests with sufficiently tall trees for nesting (Collar et al. 2018b). It has been affected by illegal capture and traffic since the 1980s (BirdLife International 2018) and has suffered population decline due to fragmentation and deforestation (Collar et al. 2018b). It was recorded next to the JBMB main trail, at point no. 02. (Temminck, 1815). Regionally Endangered. The species inhabits areas up to 2500 m above sea level. In tropical lowlands, it shows preference for wet grasslands and Savanna woodland. It is primarily an insectivore, but it also feeds on seeds, fruits, roots, and bulbs when insects are scarce. R. rufescens is considered sedentary and is subject to poaching for its meat. Populations in Brazil are in steady decline due to hunting pressures associated with habitat loss and poisoning by insecticides (Cabot et al. 2019). In the JBMB, we found this species in open grasslands in the regenerating area, especially at points no. 20-23. (Temminck, 1815). Regionally Near Threatened. P. superciliaris occupies a wide range of vegetal physiognomies, from Cerrado to dense and humid forests. It is a large generalist frugivore that forages in flocks of up to 10 individuals, maintaining a sedentary habit. It exhibits no major movement patterns throughout its life (del Hoyo and Kirwan 2018). In the JBMB, the species is abundant in Semideciduous Forest and was recorded at points no. 02-04. (Boddaert, 1783). Regionally Endangered. This species lives in dense vegetation on the margins of water bodies (Bertram and Kirwan 2018), and the loss of riparian habitats is the main threat to its survival (Bressan 2009). There are few records of this species in São Paulo state, which makes its presence in the JBMB extremely valuable. We found one individual in a lake located on the western border of the JBMB, in its limit with the Municipal Zoological Garden of Bauru (22°20′31″S, 049°01′02″W).

Crotophaga major Gmelin, 1788. Regionally Near
Threatened. This species is relatively demanding regarding habitat quality, with few records existing in the state, because it is commonly found in aquatic macrophytes and hanging vegetation above water (Payne and de Juana 2018). The loss of most Riparian Forests due to construction of hydroelectric power plants in São Paulo state led to a significant decrease in its populations. This species was recorded in the JBMB surveys, in which 2 researchers reported the presence of 3 individuals in an area adjacent to Semideciduous Flooded Forest in November 2016 (22°20′40″S, 049°01′04″W). Neopelma pallescens (Lafresnaye, 1853). Regionally Near Threatened. This species is common in dense and tangled vegetation (Snow 2018), especially in Semideciduous and Gallery Forests. It inhabits the midstory stratum of Forest Savanna (1-5 m from the ground) and feeds mainly on insects and fruits (Willis and Oniki 2003). Because the species has narrowly defined habitat preferences, the main threats to it are related to fragmentation and habitat loss in São Paulo state. Therefore, N. pallescens is considered Endangered on the São Paulo Red List (São Paulo 2018). We recorded this species over 2 consecutive months (August and September 2017) in Forest Savanna remnants at points no. 08 and 09.

Bubo virginianus (Gmelin, 1788
Discrimination between congeneric taxa. We recorded some congeneric species that are difficult to differentiate during surveys. Here, we list the diagnostic features that allowed these species to be distinguished, based on Grantsau (2010).

Crypturellus obsoletus, Crypturellus parvirostris, and
Crypturellus tataupa. We identified these species based on song. C. tataupa has a vocalization with trilled notes descending in frequency, which tends to be shorter than the songs of the other two species. C. parvirostris has long songs with whistled notes that are repeated in acceleration, ending in a decline in frequency and with repeated trilled notes at the end of the vocalization. C. obsoletus produces vocalizations with structure similar to C. parvirostris, but the final portion of its vocalization ends abruptly, with no repetition of notes. C. obsoletus was recorded at point no. 02, C. parvirostris was recorded at point no. 12, and C. tataupa was recorded at point no. 16.

Laterallus viridis and Laterallus melanophaius.
Both species were identified by their songs. While L. melanophaius has a fast-trilling vocalization, L. viridis songs have introductory notes, with a slower pace of elements. L. melanophaius was found at point no. 15, while L. viridis was found at point no. 01.
Leptotila verreauxi and Leptotila rufaxilla. L. verreauxi has a crown and neck with a greenish glow, olive-brown upper part, and a pale gray underpart that becomes white on the belly, and vocalizations with 2 low-frequency notes. In contrast, L. rufaxilla has a bluish gray crown, nape, and hindneck with a grayish purple glow, dark-brown upper part, and a greyish pink under part that becomes white on the belly, and low-pitched vocalizations with just 1 descending note. Both species was found at points no. 02-07.
Crotophaga major and Crotophaga ani. While C. ani has a uniform black color from the bill to the tail, C. major has dark-blue underparts and upper parts with a greenish sheen, prominent culmen, and white iris. We found C. major in a flooded field (22°20′40″S, 049°01′04″ W) and C. ani in the open area of Savanna in regeneration, at points 17-24.

Discussion
Our results demonstrate that JBMB have a high species richness, which might be related to its location in the transitional area between the Atlantic Forest and Cerrado domains. We identified 9 species that are subject to some degree of threat and 11 endemic species (7 from Atlantic Forest and 4 from Cerrado). In addition, 31 species exhibited some degree of migratory behavior. When comparing our results to other surveys in Semideciduous Forests, it is evident that the richness of the JBMB is greater than that in the more extensive areas of the interior Atlantic Forests of São Paulo. Donatelli et al. (2007) reported the presence of 65 and 64 species in fragments of 350 and 480 ha of Semideciduous Forest, respectively. We must take into account that their sampling effort was smaller than ours, as they spent 80 h using point counts in secondary-growth forests. However, our study documented higher richness, even when compared to other studies with longer sampling efforts, such as Aleixo and Vielliard (1995), who recorded 97 species in a 251 ha remnant over 235 h using point counts. The elevated richness (220 bird species) in our study could be due to the diverse habitats of the JBMB, which encompasses Semideciduous Forests and other Atlantic forest phytophysiognomies. Moreover, the high degree of anthropogenic activities in the surrounding areas might also enhance species richness by the inclusion of some resilient taxa to disturbed environments, especially in tropical environments (Aleixo 1999, Mulwa et al. 2012, Asefa et al. 2017. Few studies have been conducted in Forest Savanna and have recorded much lower richness, with fewer than 100 species being associated exclusively with this physiognomy (Caverzere et al. 2011). More species were detected in studies that included other Savanna physiognomies, such as grasslands and shrublands. Manica et al. (2010) recorded 160 bird species in 472 ha Cerrado fragment over 95 h using qualitative surveys. However, of all of these species, only 28 were found in Forest Savanna physiognomy. Other studies encompassing more extensive areas over longer sampling periods obtained a similar number of species to our study. For instance, Motta-Júnior et al. (2008) recorded 231 species over 580 h of sampling in 2,300 ha of Cerrado grasslands using point counts. Lucindo et al. (2015) recorded 226 species in 380 h of sampling 2,712 ha mixed physiognomies of Cerrado using transect counts.
Of the non-passerines, the greatest number of species were represented by the families Trochilidae, Accipitridae, and Columbidae. Their presence is associated with the fragmented areas and their conspicuous behavior makes them easy to be detected during sampling (Sick 1997). Columbidae are predominantly granivores, while Accipitridae are carnivores. Such feeding behaviors are favored in open habitats, such as Forest Savanna and the regenerating areas of the JBMB (Chettri et al. 2004, Winkler et al. 2015. The richness of the Trochilidae is indicative of the abundance of floral resources, because species of this family are predominantly nectarivores (Wilman et al. 2014). The presence of several physiognomies in the study area might provide diverse flowering plants at different times of the year, especially in the transition periods between dry and rainy seasons (Pirani et al. 2009). Finally, the presence of a network of wetlands, lakes, and rivers in the study area favored the presence of species dependent on aquatic environments, such as those of the families Ardeidae and Rallidae (7 species each).
Among Passeriformes, Tyrannidae and Thraupidae had more species than the other familes.greatest number of species were represented by the families Both are typical of Neotropical bird communities, and the family Tyrannidae is one of the largest groups in the region Tudor 1994, Sick 1997). This phenomenon might be associated with the generalist habit of several species in this family (Sick 1997), which allows them to access several types of habitats. On the other hand, several species of Thraupidae feed on fruits and act as seed dispersers (Ridgely and Tudor 1989) and are important for maintaining plant diversity. The richness of Thraupidae can also be associated with recent taxonomic changes proposed by Piacentini et al. (2015) which subsumed the formerly separate Emberizidae and Coerebidae as subfamilies within Thraupidae. This fusion increased the number of species of Thraupidae from 99 species to 173 species. Nonetheless, the richness of frugivorous and granivorous birds could indicate the variety of the habitats available in the region.
When comparing our survey to a previous survey on Forest Savanna in the JBMB area, we found 28 species that were not reported by Caverzere et al. (2011). Of these, we recorded 8 endemic species (5 to Atlantic Forest and 3 for Cerrado), and 2 threatened taxa, which are noteworthy records. However, these authors reported 10 species that we did not find during our study. Yet, we observed 8 of these species in other environments near the JBMB. The 2 other species unique to the survey by Caverzere et al. (2011) were Leptodon cayanensis and Hylophilus poicilotis. The absence of records on L. cayanensis in our survey might be related to its tends to occupy a large territory (Brown and Amadon 1989). Its movement across the landscape might make it difficult to record this species in the study area. In contrast, the record of H. poicilotis seems to be questionable, since we recorded only Hylophilus amaurocephalus, a very similar congeneric species. H. amaurocephalus was also recorded by Rangel de Almeida et al. (2010) at the Bauru Ecological Station, with this being the only record of the genus in the vicinity of our survey site. In addition, citizen science databases also show records of only H. amaurocephalus in Bauru (Wikiaves 2019). The nearest record of H. poicilotis is from the city of Campinas, located more than 200 km away from the JBMB (eBird 2019). In comparison, H. amaurocephalus is quite common in the meso-region of Bauru, with records in the nearby cities (eBird 2019). Therefore, we suggest that the record of H. poicilotis in the JBMB is incorrect, because the songs of these 2 species are similar. Because H. poicilotis is endemic to the Atlantic Forest (Bencke et al. 2006), it is not likely to inhabit an area of Forest Savanna, even in a transitional area such as the JBMB.
According to Somenzari et al. (2018), Brazil has 198 migratory species, which corresponds to approximately 10% of all species found in the country. In our surveys, we observed 31 species (14.1%) exhibiting migratory behavior. For species defined as migrants, the records of Chordeiles minor and Chaetura meridionalis are of note. C. minor reproduces in North and Central America, migrates to South America in September, and returns to its locations of origin in April (Cleere 2018). C. meridionalis reproduces in Brazil during the austral summer and migrates north to Panama, Venezuela, and French Guiana in the winter .
Records of endemic species is important because they are negatively affected by habitat fragmentation and loss (Aleixo andVielliard 1995, Anjos 2001). These species can be considered habitat specialists, and their presence in our survey can be an indicative of the quality of forest remnants. In the Semideciduous Forest, we recorded 7 Atlantic Forest endemics. Few endemic species were recorded in our survey compared to other surveys: 33 spp. by Faria et al. (2006), 43 spp. by Ubaid (2009), and 22 spp. by Vianna et al. (2017). Antunes (2007) recorded the disappearance of 9 endemic species over 30 years in a fragment of Semideciduous Forest, while the abundance of another 9 species significantly declined. Alternatively, we recorded 4 Cerrado endemic species, which was more than that recorded by Cavarzere et al. (2011) at the same location; Cavarzere et al. recorded only Antilophia galeata as endemic to the Cerrado. Both Cyanocorax cristatellus and Saltatricula atricollis are Cerrado endemic species that we found exclusively in the regenerating area, which is because they normally inhabit open physiognomies of savanna. Cavarzere et al. (2011) did not found these species because they restricted their survey to closed physiognomies (forest savanna and semideciduous forest). Our study obtained a similar richness of endemic Cerrado taxa when compared to previous studies: 6 spp. by Motta-Júnior et al. (2008), 6 spp. by Manica et al. (2010), and 6 spp. by Lucindo et al. (2015). The presence of such endemic taxa could point to the capacity of the environment to provide sufficient resources across seasons (Antunes 2007) and might be related to the different stages of vegetation regeneration.
The richness of birds recorded in the JBMB might be related to the large area of continuous native forest (803 ha) and the presence of areas at different stages of forest regeneration. The number of species in a given location is associated with habitat heterogeneity, which is an important predictor of bird richness (Stein et al. 2014, Stirnemann et al. 2015 and is important for local community maintenance (Develey et al. 2010). A greater diversity of habitats is related to a greater diversity of resources, which supports a greater number of species. Thus, the regenerating area of the JBMB supports species exclusive to open areas, which dominate such places. Meanwhile, the savanna remnants surrounding the regenerating area could also contribute to increase the number of the species in this location, which could improve the regional richness by landscape supplementation (Dunning et al. 1992). Birds dwelling in closed environments could use the regeneration area to gather supplementary resources in adjacent boundaries (Tubelis et al. 2004, Brotons et al. 2005, which could explain the presence of forest-dwelling species in an open area.
In conclusion, the JBMB is an important area with a richness of bird species higher than other surrounding conservation units. Such richness is attributed to the transition between the Atlantic Forest and Cerrado, as well as to the variety of physiognomies and successional stages in the area favoring the distribution of several taxa within a restricted area. Although we only recorded few endemic and threatened species, these records demonstrate the capacity of the local environment to support species that require specific ecological resources. Our findings also contribute to the knowledge of avian richness in transitional areas, demonstrating the importance of this type of environment for the maintenance of bird diversity.