Amphibians and reptiles from the Parque Nacional da Tijuca, Brazil, one of the world’s largest urban forests

The Parque Nacional da Tijuca in Rio de Janeiro, Brazil, is considered to be one of the world’s largest urban forests, however no systematic inventory of its herpetofauna is available. In the present study, we surveyed the amphibians and reptiles of this park to assess its species composition (including secondary data) and obtain estimates of species richness and abundance. We conducted active searches (460 hours) between January 2013 and December 2015. We identified the taxa endemic to either the Atlantic Forest or Rio de Janeiro state, and verified the conservation status of each species in the international, Brazilian, and state red lists. We also estimated the species richness and sampling sufficiency by rarefaction curves and Bootstrap richness estimator, and analyzed the distribution of the species abundance in Whittaker plots. We recorded 3,288 individuals over 36 months, representing 24 species of amphibians and 25 reptiles. The cumulative species curves, rarefaction, and the richness estimated indicated that sampling effort was adequate. Species abundance adjusted to the log-series model in both amphibians and reptiles. The four most abundant amphibians represented 70% of the individuals recorded in this group, while the two most abundant reptiles represented 60% of the total individuals. The inclusion of the secondary data raised the number of amphibian species to 38, and the number of reptiles to 36. Approximately 80% of the amphibian species and 28% of the reptile species recorded are endemic to the Atlantic Forest, and six of the amphibian species are endemic to Rio de Janeiro state. Six amphibian species and one reptile species are classified under some threat of extinction, and two reptile species were exotic. The considerable diversity of the herpetofauna of the Parque Nacional da Tijuca, which includes endemic and threatened species, reflects the effectiveness of the reforestation of this protected area and emphasizes the importance of its conservation.


Introduction
The Brazilian Atlantic Forest biome is one of the world's biodiversity hotspots (Myers et al. 2000), although it has now been reduced to only around 12% of its original cover as a result of centuries of intense deforestation (Ribeiro et al. 2011, SOS Mata Atlântica, INPE 2018. Approximately 625 amphibian species ) and 300 reptiles (Tozetti et al. 2017) are known to occur in this biome, and new species are still being steadily described (e.g., Cruz et al. 2019, Maciel et al. 2019, Prates et al. 2019. In the state of Rio de Janeiro, 201 species of amphibian have been recorded, including 197 anurans and four caecilians (Dorigo et al. 2018), together with 149 reptiles (Oliveira et al. 2020). This state has a high rate of endemism of both amphibians (Dorigo et al. 2018) and reptiles (Tozetti et al. 2017), and its metropolitan region is considered to be an important area for the conservation of Brazilian biodiversity .
The state capital, Rio de Janeiro, which has a population of more than six million inhabitants, includes one of the world's largest urban parks, the Parque Nacional da Tijuca (PNT), which is the only Brazilian national park located within an urban zone. Like other coastal ranges within the Atlantic Forest domain, the Tijuca massif has a rich network of water courses (Coelho Netto 2005), which favors the occurrence of amphibians. During the 18th and 19th centuries, however, the forest of this area was almost totally destroyed for the production of charcoal, and the establishment of sugarcane and coffee plantations (Rocha et al. 2003, Rocha 2017. As a result, many of the water sources that supplied the city began to dry up (Abreu & Rodrigues 2010, Rocha 2017. It seems likely that fragments of forest persisted in many of the more inaccessible areas within the Tijuca massif, in particular, the highest mountain peaks and the steepest valleys (Coimbra-Filho et al. 1973). In addition to providing seeds of native species for the subsequent reforestation of the area (Coimbra-Filho et al. 1973, Abreu & Rodrigues 2010, these forest remnants probably also provided important refuges for many local animal species (e.g., Rocha et al. 2003). The deforested areas once covered by plantations were almost completely reconstituted by a reforestation project -the first major project of its kind anywhere in the world -which involved the planting of 100,000 trees, beginning in the second half of the 19th century during the Brazilian Empire period. This initiative resulted in the establishment of an exuberant forest, which was converted into a national park in 1961 (Brasil 1961).
In addition to the PNT, two other large parks are found within the urban zone of Rio de Janeiro -the Parque Estadual da Pedra Branca and the Parque Natural Municipal da Serra do Mendanha (PNM Serra do Mendanha). The PNT has a long history of scientific research, first being visited by naturalists during the 19th and 20th centuries. These naturalists described a number of amphibian species based on specimens collected in this area, including Aplastodiscus albofrenatus (Lutz, 1924), Dendrophryniscus brevipollicatus Jiménez de la Espada 1870, Ischnocnema guentheri (Steindachner, 1864), Ischnocnema octavioi (Bokermann, 1965), Phasmahyla guttata (Lutz, 1924), and Scinax trapicheiroi (A. Lutz and B. Lutz, 1954). More recently, herpetofaunal researches in the park included studies of the ecology, natural history or bioacoustics of some taxa (e.g., Heyer 1973, Hepp & Carvalho-e-Silva 2011, Dorigo et al. 2014, Machado et al. 2016, Hepp et al. 2017, Silvae-Souza 2019, Guarabyra et al. 2020. Introductory lists of amphibian species (Carvalho-e- ) and reptiles (Mocelin 2008) were compiled for the management plan of this conservation unit (ICMBIO 2008), although, up to the present time, no systematic inventory has been produced. Given this, we inventoried the herpetofauna of the PNT over a three-year period, using intensive sampling to obtain species composition and estimates of the richness and abundance of the park's amphibians and reptiles. We supplemented our inventory with records of occurrence obtained from herpetological collections and species reintroduction programs in PNT.

Material and Methods
The Parque Nacional da Tijuca (22º55'-23º00' S, 43º11'-43º19' W) is located in the Tijuca massif in the city of Rio de Janeiro, capital of the state of Rio de Janeiro, southeastern Brazil. The park covers a total area of 3,958.38 hectares of dense ombrophilous forest at altitudes between 80 m and 1,021 m above sea level (a.s.l.). The park is divided officially into four sectors (ICMBIO 2008): Sector A -the Tijuca Forest, Sector B -the Carioca Mountains, Sector C -Pedra Bonita/ Pedra da Gávea, and Sector D -Pretos-Forros/Covanca ( Figure 1). The landscape of the park encompasses a range of land uses and vegetation cover, from forest encompassing distinct stages of succession, to open fields, rocky outcrops, as well as infrastructure, such as buildings and roads (Coelho Netto 2005).
The climate is of the Aw type (humid tropical) in the Köppen classification, with intense rains during the austral summer and drier conditions during the winter. Mean monthly temperatures range from 25°C in February to 19°C in June, with an annual mean of approximately 22°C. Annual precipitation is typically between 2,000 and 2,500 mm, although it may reach 3,300 mm in exceptionally rainy years (Coelho Netto 2005, ICMBIO 2008).

Data collection
We searched for amphibians and reptiles in the Sector A of the park over a 36-month period, between January 2013 and December 2015, at altitudes between 400 m and 800 m a.s.l. The searches were based on the Visual Encounter Survey (VES) method of Crump & Scott (1994), and resulted in a total sampling effort of 460 person-hours, with 230 hours of diurnal and 230 hours of nocturnal searches. When conducting an active search, always two researchers moved through the forest searching systematically for amphibians and reptiles in microhabitats such as streams, ponds, bromeliads, the leaf litter, tree trunks, and rocks. The searches were conducted only in a single direction, to avoid the resampling of individuals, thus guaranteeing the independence of the records. All individuals encountered during searches were captured, identified, and most of them released at the capture site. We collected voucher specimens, which we deposited in the herpetological collection of the Museu Nacional (MNRJ), Universidade Federal do Rio de Janeiro (voucher number in Appendix 1). We complemented our inventory with secondary records, which included data from herpetological collections (Appendix 2), and from a faunal reintroduction project for the PNT. We applied the scientific nomenclature proposed by Frost (2020) for the amphibian species. For the reptiles, we used the nomenclature of Costa and Bérnils (2018), which we updated whenever necessary, based on

Data analysis
We calculated rarefaction curves and confidence intervals (95%) with the moment-based estimator of species richness (Colwell et al. 2004) to verify the efficiency of the sampling effort in terms of the number of species recorded during the study for both amphibians and reptiles. For this analysis, one hour of VES constituted a sampling unit, with a total of 460 samples being obtained during the study period as a whole. We considered the convergence of the confidence intervals to zero as an indicator of sampling sufficiency (Colwell et al. 2004). We also estimated the number of species and standard deviation with the Bootstrap richness estimator (Smith & van Belle 1984) with 1,000 runs. Both the rarefaction curves and Bootstrap were calculated in the program EstimateS 8.2.0 (Colwell 2009). We analyzed the distribution of the species abundance, using rank abundance curves or Whittaker plots (Whittaker 1965), which show the ranked log abundance of each species (Krebs 1999). We analyzed these plots in PAST 2.17 (Hammer et al. 2001).
Species endemism in the Atlantic Forest was defined for the amphibians based on Lingnau et al. (2008), Haddad et al. (2016), andFrost (2020), and for reptiles according to Tozetti et al. (2017). We defined the conservation status of each species based on the IUCN Red List (IUCN 2020), the Red List of Threatened Brazilian Fauna (ICMBIO 2018), and the List of the Threatened Fauna of Rio de Janeiro state (Bergallo et al. 2000). The population trend for each species was obtained from IUCN (2020).

Results
We recorded a total of 74 species of herpetofauna (distributed in four orders), being 38 amphibians (37 anurans and one caecilian; Table 1) and 36 reptiles (three chelonians, one amphisbaenian, nine lizards, and 23 snakes; Table 2) at the PNT. Hylidae was the amphibian family with the highest species richness (N = 10 species), followed by Brachycephalidae (N = 5) and Cycloramphidae (N = 5) ( Table  1). With the exception of Gymnophthalmidae, with two species, all lizard families were represented by only a single species. The snake family Dipsadidae was represented by 14 species, and Colubridae by five (Table 2). During our field surveys we recorded a total of 3,288 individuals in Sector A of the Parque Nacional da Tijuca, during the 36 months of the study period, representing a total of 49 species. Most (3,174) of these individuals were amphibians (Table 1), belonging to 24 species of the orders Anura (N = 23 species) and Gymnophiona (N = 1) (Figures 2-4). We recorded only 114 reptiles, although these individuals also represented 25 species (Table 2), distributed in two orders, Chelonia (N = 2) and Squamata (N = 23) ( Figures 5-7). Two of these amphibian species and 11 of the reptiles were recorded opportunistically during the study period (i.e., not during standardized surveys; Tables 1 and 2).
The cumulative curve for the amphibian species reached the asymptote at 213 hours of sampling with a total richness of 22 species, while the curve for the reptiles reached the asymptote at 276 hours,  Table 1. Amphibian species recorded in Parque Nacional da Tijuca in the municipality of Rio de Janeiro, state of Rio de Janeiro, southeastern Brazil, with data on total abundance (Ab; number of records) and relative abundance (RA; percentage of all recorded amphibians) for the species recorded during the Visual Encounter Surveys are given. Other information include: the conservation status of each species based on their category of threat in the international (IUCN 2020), Brazilian (ICMBIO/MMA 2018), and Rio de Janeiro (RJ) state lists (Bergallo et al. 2000); whether they are endemic (End) to the Atlantic Forest (AF) or to Rio de Janeiro state (RJ) or exotic (Ex); their population trend (PT; IUCN 2020); and Sectors of PNT (ICMBIO 2008) where each species was recorded based on primary (P; see Appendix 1) or secondary (S; see Appendix 2) data. * Species recorded in non-standardized encounters. Conservation status: DD = Data Deficient, LC = Least Concern, VU = Vulnerable, EN = Endangered, and PA = Presumably Threatened ("Presumivelmente Ameaçada"). with a total of 14 species, based on a total sampling effort of 460 hours ( Figure 8). The confidence intervals of the rarefaction curves converged to zero for the amphibians, but not for the reptiles, for which the error was estimated as 2.3 species (Figure 8). The species richness estimated by the Bootstrap procedure was similar to that recorded by the VES method for both amphibians (estimated richness = 22.4 species) and reptiles (estimated richness = 15.6 species). The Whittaker plots ( Figure 9) revealed considerable variation in the abundance of the amphibian and reptilian species, which was best adjusted using the log-series model. The most abundant amphibian species (Table  1) were Ischnocnema parva (N = 617 records; 19.4% of all amphibians), Hylodes nasus (N = 601; 18.9%), Ischnocnema guentheri (N = 544; 17.1%), and Crossodactylus gaudichaudii (N = 500; 15.8%). The most abundant reptiles were the lizards Enyalius brasiliensis (N = 55; 48.2% of all reptiles) and Ecpleopus gaudichaudii (N = 15; 13.2%), representing 61.4% of all the reptiles recorded during the surveys ( Table 2). The most abundant snakes were Bothrops jararacussu (N = 8; 7.0% of all reptiles), Thamnodynastes cf. nattereri (N = 7; 6.1%), and Bothrops jararaca (N = 6; 5.3%).
Overall, 30 of the amphibian (ca. 80% of the total) and ten of the reptile species (28%) recorded in the PNT are endemic to the Atlantic Forest biome (Tables 1 and 2). Six of these amphibians -Aplastodiscus albofrenatus, Dendrophryniscus brevipollicatus, Euparkerella brasiliensis, Hylodes nasus, Ischnocnema guentheri, and Scinax trapicheiroi -are considered endemic to the state of Rio de Janeiro. Most of the amphibian and reptile species are listed as Least Concern or Not Threatened in the three lists of conservation status (international, Brazilian, and state). Six amphibian and one reptile species recorded in the PNT are listed under some category of extinction threat. The populations of 15 amphibian species are thought to be declining ( Table 1). Two of the reptile species, the lizard Hemidactylus mabouia and the chelonian Trachemys scripta are exotic (Table 2).

Biodiversity: species richness, composition, and abundance
The results of the present study indicate that the forests of the PNT harbor for a large proportion of the herpetofaunal diversity of the Atlantic Forest of the state of Rio de Janeiro. Considering both the records obtained during our fieldwork and from secondary data, the park's amphibian species richness would amount to at least 38 species, which represent 19% of the amphibians known to occur in the state . When considering the sampled area and the effort employed, the number of amphibian and reptile species recorded during surveys in PNT is consistent with the expected. In fact, the cumulative species curves for both groups reached the asymptote by around the middle of the study period, indicating that sampling effort was adequate in both cases. Similarly, while the confidence intervals of the rarefaction curves for the amphibians converged to zero, indicating sampling sufficiency, those calculated for the reptiles approached the convergence, with reduced estimated error.
In comparison with the PNM Serra do Mendanha − other substantial remnant of dense rainforest in the urban zone of Rio de Janeiro for which systematic medium-term data are available − the results of the present study are broadly consistent. Pontes et al. (2015) recorded 45 amphibians and 39 reptiles at PNM Serra do Mendanha, values closely comparable with the 38 amphibians and 36 reptiles we recorded at PNT. The amphibian species richness recorded in the PNT is also relatively high in comparison with other forested areas in the state of Rio de Janeiro for which data are available (Table 3). While the species richness recorded at the different sites fluctuates considerably, it seems likely that much of this variation is related to differences in survey methods, sampling effort, and the study period (Table 3).
The reptilian species richness recorded in the PNT was also relatively high in comparison with the data from other Atlantic Forest remnants in the state of Rio de Janeiro, as recorded in both short-term (Morro São João, Casimiro de Abreu, N = 9 species; Almeida-Gomes et al . 2008) Vrcibradic et al. (2011) at the Estação Ecológica Estadual do Paraíso, which reached a total of 29 species with the inclusion of secondary data. Hylidae (N = 10) was the most diverse amphibian family in the PNT, followed by Brachycephalidae and Cycloramphidae, with five species. A similar predominance of hylids in the amphibian assemblage has been recorded not only at other Atlantic Forest sites (e.g., Heyer et al. 1990, Almeida-Gomes et al. 2014a, but also at other localities in the Neotropical region (e.g., Toft & Duellman 1979, Duellman 1988, Menin et al. 2017. The high diversity of hylids in South America has been explained by the fact that this region is considered to be the center of origin and diversification of this clade (Wiens et al. 2006).
With 14 species, Dipsadidae was the most diverse reptilian family in the PNT, which is consistent with the pattern typically found for the Squamata in Atlantic Forest remnants in the state of Rio de Janeiro (e.g., Pontes et al. 2009, Martins et al. 2019). The family Dipsadidae, which is widely distributed in the Americas and West Indies (Hedges et al. 2009, Zaher et al. 2009, Grazziotin et al. 2012, includes the vast majority of the snakes of the Neotropical region, with more than 700 species (see Uetz et al. 2019).
The PNT and the PNM Serra do Mendanha (Pontes et al. 2015) shared 25 amphibian species (Sørensen index of similarity = 0.6; Magurran & McGill 2011) and 29 reptiles (= 0.8). Despite these similarities, 13 amphibian species were exclusive to the PNT, while 27 were exclusive to the Serra do Mendanha. Similarly, seven species of reptiles were exclusive to PNT and ten to the Serra do Mendanha. Similarities of the herpetofauna between these two areas would be accounted at least partly because they would have originally been part of a single continuous forest in the past, and can also be indicative of the effectiveness of the reforestation of the Tijuca massif, carried out during the 19th century (Rocha et al. 2003, Rocha 2017. However, the existence of a unique fauna in both areas also reinforces the importance of the preservation of both forests for the protection of the herpetofauna of the Atlantic Forest in general.
The relative abundance of both amphibian and reptilian species in the PNT was adjusted to a log-series model, a pattern typical of communities of medium diversity, showing certain degree of equilibrium between the rare and dominant species (Magurran & McGill 2011). In this model, most species are relatively rare in comparison with the more abundant taxa (Magurran 2004). The four most abundant amphibian species (I. parva, H. nasus, I. guentheri, and C. gaudichaudii) contributed to around 70% of the total records of amphibians obtained in the park. The genus Ischnocnema belongs to the superfamily Brachycephaloidea (sensu Padial et al. 2014), which spawn on the moist forest floor and have direct development, so they do not depend on bodies of water for   (Haddad & Prado 2005, Pombal & Haddad 2007. Species that have direct development tend to be dominant in leaf litter-dwelling anuran communities in Neotropical rainforests (e.g., Crump 1971, Allmon 1991, Donnelly 1994, Rocha et al. 2001, Siqueira et al. 2014. Two stream-dwelling frog species, Hylodes nasus and C. gaudichaudii, were the most abundant species recorded at the PNT. Hylodidae species are generally abundant in small streams in preserved Atlantic Rainforest areas (e.g., Weygoldt 1989, Patto & Pie 2001, Almeida-Gomes et al. 2008, Siqueira et al. 2011b.
The most abundant reptiles in the PNT were Enyalius brasiliensis and Ecpleopus gaudichaudii, and accounted for approximately 60% of the total reptilian abundance recorded in the present study. Enyalius brasiliensis is typical of forested environments, and was relatively abundant in the area in comparison with other forest remnants in the state of Rio de Janeiro, based on short-term (Morro São João; Almeida-Gomes et al. 2008) or mid-term studies (Guapiaçu Ecological Reserve; Almeida-Gomes & Rocha 2014a). Ecpleopus gaudichaudii is a Gymnophtalmidae endemic to the Atlantic Forest, which is typically a major component of local leaf-litter reptile communities (e.g., Dixo & Verdade 2006, Almeida-Gomes & Rocha 2014a, Cruz et al. 2014. Two of the three most abundant snakes recorded during our surveys were members of the family Viperidae (B. jararaca and B. jararacussu), whose representatives are common in areas of mainland Atlantic Forest (Marques & Sazima 2004, Pontes et al. 2009). The Dipsadidae Thamnodynastes cf. nattereri was the second most abundant snake in the PNT, and was also one of the most abundant in the PNM Serra do Mendanha (Pontes et al. 2009).

Conservation: endemism, threat status, and exotic species
Most (ca. 80%) of the amphibian species recorded in the PNT are endemic to the Atlantic Forest . This was somewhat expected, as that biome has a high degree of endemism for amphibians   species recorded are endemic to the Atlantic Forest, although this is also typical for this group (Tozetti et al. 2017). Six of the amphibians are also considered to be endemic to the state of Rio de Janeiro, including two (H. nasus and I. guentheri) that are endemic to the municipality of Rio de Janeiro. Hylodes nasus is known to be restricted to the forests of this municipality, and has, in fact, only been recorded, up to now, in the PNT (Nascimento et al. 2001, Lingnau et al. 2008) and the PNM Serra do Mendanha (Pontes et al. 2015). However, Canedo (2008) has analyzed specimens that may be associated with H. nasus, which were collected in the municipality of Volta Redonda, in the state of Rio de Janeiro, and also in Cachoeiro de Itapemirim, in the state of Espírito Santo, although further research is required to confirm the identity of the taxon. Ischnocnema guentheri is currently considered to be a species complex (Kwet & Solé 2005, Gehara et al. 2013). Based on mitochondrial and nuclear sequences, combined with bioacoustic data, Gehara et al. (2013) identified I. guentheri sensu stricto as a lineage restricted to the PNT, although it seems likely that this form is also present in adjacent areas.
Overall, only six of the species recorded in the PNT (Allobates olfersioides, Cycloramphus eleutherodactylus, Scinax trapicheiroi, Thoropa lutzi, Vitreorana eurygnatha, and V. uranoscopa) are listed under some category of extinction threat (Bergallo et al. 2000, ICMBIO 2018, IUCN 2020, although the populations of 15 species are thought to be declining (IUCN 2020). Of these six species, we recorded only Table 3. The species richness (S) of amphibians recorded in forest remnants in the state of Rio de Janeiro, Brazil. The information provided is: location, municipality, sampling period, total effort employed in the visual survey (VES, in hours), quadrat (m 2 ), and pitfall trap methods (with the bucket volume in liters), other methods or sources used (OMS), and the respective reference. * Information on sampling period or effort was not available. PNM = Parque Natural Municipal, RPPN = Reserva Particular do Patrimônio Natural. In the pitfall trap column, BD = bucketdays, BH = bucket-hours. one, S. trapicheiroi, during our field surveys. However, although S. trapicheiroi is listed as Near Threatened by the IUCN due to its greatly reduced area of occurrence of less than 20,000 km 2 (Rodrigues & Carvalho-e-Silva 2004), it is assigned to the Least Concern category in the Brazilian red list (ICMBIO 2018). Verdade and Rodrigues (2007) allocated three other specific names for geographically widespread populations into synonymy of Allobates olfersioides (Lutz, 1925): A. alagoanus (Bokermann, 1967), A. capixaba (Bokermann, 1967), and A. carioca (Bokermann, 1967. According to this arrangement, A. olfersioides is widely distributed along much of the eastern coast of Brazil, between Rio de Janeiro and the state of Alagoas, 2,000 km to the north. However, Haddad et al. (2013) still considered these disconnected populations as different species: A. olfersioides in the state of Rio de Janeiro, A. capixaba in the states of Espírito Santo, and A. alagoanus in the state of Alagoas. The clarification of whether these populations are distinct species or lineages of the same species is essential (Forti et al. 2017). In their assessment of the conservation status of this amphibian, Haddad et al. (2016) considered only A. carioca to be a synonym of A. olfersioides, which restricted the species distribution to the metropolitan region of Rio de Janeiro. In this case, the reduction of the area of occurrence of the taxon to approximately 7,730 km², the lack of records from some localities, including the PNT, over the past 30 years, and the loss of habitat quality over this period has led to the classification of this species as Vulnerable in both the Brazilian red list (ICMBIO 2018) and that of the IUCN (2020). Thoropa lutzi is classified as Endangered by the IUCN (2020), but as Data Deficient in Brazil (ICMBIO 2018). This species is known to occur only in the states of Rio de Janeiro and Espírito Santo. Populations have not been observed in Rio de Janeiro over the past 30 years, however, including PNT (Sabbag et al. 2018). The other three species, Cycloramphus eleutherodactylus, Vitreorana eurygnatha, and V. uranoscopa, are listed only at the state level (Bergallo et al. 2000), as Presumably Threatened, and only C. eleutherodactylus is listed by the IUCN (2020), as Data Deficient. Cycloramphus eleutherodactylus is known to occur in the Brazilian states of Paraná, São Paulo, Rio de Janeiro, andMinas Gerais (Rossa-Feres et al. 2017, Santos-Pereira et al. 2018), although some studies have reported its disappearance by the beginning of the 20th century (Izecksohn & Carvalho-e-Silva 2001, Eterovick et al. 2005). We did not collect C. eleutherodactylus during fieldwork and it has not been found at PNT since 1972 (Matos 2011). Nevertheless, Lima et al. (2012) concluded that this species may be relatively abundant in some cave environments, and emphasized the importance of including these habitats in amphibian inventories.
Hylodidae species of three genera, Crossodactylus, Hylodes and Megaelosia, are rheophilic and endemic to the Atlantic Forest (Laia & Rocha 2012), where they are typically found in well-preserved forest fragments (e.g., Santos-Pereira et al. 2016, Malagoli et al. 2017). These species are strongly associated with this type of environment, and depend on streams throughout their life cycle (eggs, tadpoles, and adults), being thus vulnerable to the loss of riparian forest (Almeida-Gomes et al. 2014b). These anurans are particularly relevant to conservation initiatives, given that their effective area of occupation is normally restricted to water courses, a minor fraction of the total area of most forests, which makes them especially vulnerable to extinction (Almeida-Gomes et al. 2014b).
Only one of the reptiles recorded in the present study, the tortoise Chelonoidis denticulatus, is considered to be under some threat of extinction, being classified as Vulnerable by the IUCN (2020), although it is considered to be Least Concern in Brazil (ICMBIO 2018). The occurrence of Chelonoidis denticulatus was recorded in the Atlantic Forest of Rio de Janeiro in 1820 by Prince Maximilian zu Wied-Neuwied, who found carapaces of this species on the sandy plains in the north of the municipality of Cabo Frio (Vanzolini 1994, Vargas-Ramírez et al. 2010, Vanzolini & Myers 2015. While C. denticulatus was originally recorded in the Atlantic Forest, it appears to have been eradicated subsequently from the forests of eastern Brazil (Vanzolini & Myers 2015). This species is being reintroduced into the PNT by the Refauna Project (Refauna 2019).
One important question that has yet to be answered regards the original occurrence of the tortoise Chelonoidis carbonarius in the state of Rio de Janeiro. This species was originally described by Spix (1824) from specimens obtained from the region of the Amazon River. Pritchard and Trebbau (1984), Vanzolini (1994), Ernst and Leuteritz (1999) and Vargas-Ramírez et al. (2010) described the geographic distribution of the species in Brazil, and in particular, indicated its absence in the state of Rio de Janeiro. While a number of studies have indicated the presence (e.g., Rocha et al. 2004, Köhler 2008 or possible occurrence (Tozetti et al. 2017) of C. carbonarius in Rio de Janeiro, it is difficult to establish whether this region was part of the original distribution of the species, or whether its occurrence in the region (if confirmed) is due to the deliberate release into the wild of animals raised as pets or for food.
Two of the species of the herpetofauna recorded in the PNT -Hemidactylus mabouia and Trachemys scripta -are exotic invaders. The house gecko, H. mabouia, was originally introduced into Brazil from Africa during the colonial period, and has since expanded its area of distribution considerably in the country's natural environments , Telles et al. 2015, Oliveira et al. 2016, Winck et al. 2017. Hemidactylus mabouia was observed in the built-up areas of the PNT, which is typical of its occurrence at other localities , Oliveira et al. 2016. The presence of this lizard in the vicinity of the park's natural habitats demands special attention due to the potential for the species to become invasive .
The pond slider, Trachemys scripta, was originally found in the southern United States and northern Mexico, but since the 1970s, large numbers have been raised in the United States for the pet trade, and this resulted in the introduction of the species in many places around the world (e.g., Ramsay et al. 2007, van Dijk et al. 2011. The presence of T. scripta in natural environments has negative impacts on the native freshwater turtles (e.g., Cadi & Joly 2004, Hidalgo-Vila et al. 2009). Trachemys scripta has been recorded in disturbed areas within the PNT, probably as the result of intentional release of animals raised as pets. It is still unclear whether this exotic species has established a viable population in the state of Rio de Janeiro (Oliveira et al. 2020).
A number of amphibian and reptile species of interest occur in the PNT, including those endemics to either the Atlantic Forest or to Rio de Janeiro state and/or municipality, those under some threat of extinction, and those with declining populations. The presence of all those taxa reinforces the importance of the park for the conservation of the region's herpetofauna. The considerable diversity of reptiles and amphibians found in the PNT also testifies to the success of the reforestation program of this protected area, which now provides an important refuge for species of both groups. The results of our inventory, which recorded a total of 74 species, constitute the first comprehensive published overview of the composition and abundance of the herpetofauna of the Parque Nacional da Tijuca.

Supplementary Material
The following online material is available for this article: Appendix 1 -Voucher specimens of amphibians and reptiles from the Parque Nacional da Tijuca deposited at the Museu Nacional, Rio de Janeiro (MNRJ).

Author Contributions
Thiago Arnt Dorigo: Substantial contribution in the concept and design of the study; contribution to data collection; contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content. Carla Costa Siqueira: Substantial contribution in the concept and design of the study; contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content. Jane C. F. Oliveira: Substantial contribution in the concept and design of the study; contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content.
Luciana Ardenghi Fusinatto: Substantial contribution in the concept and design of the study; contribution to manuscript preparation; contribution to critical revision, adding intellectual content.
Manuela Santos-Pereira: Substantial contribution in the concept and design of the study; Contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content.
Marlon Almeida-Santos: Substantial contribution in the concept and design of the study; contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content.
Thiago Maia-Carneiro: Contribution to data collection. Caroline N. C. Reis: Contribution to data collection. Carlos Frederico Duarte Rocha: Substantial contribution in the concept and design of the study; contribution to data analysis and interpretation; contribution to manuscript preparation; contribution to critical revision, adding intellectual content.