Herpetofauna of Parque Estadual Altamiro de Moura Pacheco : one of the last remnants of seasonal forest in the core region of the Brazilian Cerrado

Studies in remnants of semi-deciduous seasonal forest of the Cerrado are needed to fill sampling gaps and improve basic knowledge of biodiversity. This study presents data on the herpetofauna of Parque Estadual Altamiro de Moura Pacheco, one of the last protected areas of seasonal forest in the core region of the Brazilian Cerrado. Fourteen sites were sampled between March 2007 and April 2008 using pitfall traps and active searches. A total of 35 anuran species were recorded, most of which were found in hygrophilous environments near forest fragments, open areas and riparian or gallery forest. A total of 29 reptile species were recorded, most of which were found in semi-deciduous forest and riparian or gallery forest. Fourteen of the amphibians and five of the reptiles are endemic to the Cerrado. The richness of amphibians and reptiles found in Parque Estadual Altamiro de Moura Pacheco is regionally representative, with communities typical of forest environments. Key-Words. Semi-deciduous forest; Protected area; Mato Grosso de Goiás; Amphibians; Reptiles.


INTRODUCTION
The Cerrado Biome comprises a mosaic of vegetation in the central region of the South America with dry open areas, forest remnants along streams and rivers, and dense woodlands (Ab'Sáber, 1983;Ribeiro & Walter, 2008).Accelerated destruction of natural landscapes in the Cerrado makes it one of the most threatened biomes of the world (Mittermeier et al., 2004).Current estimates show that only 47% of the Cerrado is covered by its natural vegetation (Beuchle et al., 2015).Since many natural areas have been converted into agriculture and pasture, the creation and maintenance of protected areas is the fastest and most efficient way to conserve species of fauna and flora of the Cerrado (Bensusan, 2006;Françoso et al., 2015).
Recent studies indicate that between 5.5% and 6.5% of original Cerrado vegetation is maintained by protected areas (Mittermeier et al., 2004;Françoso et al., 2015).However, the effectiveness and purposes of these units should be the subject of deeper analyses, since protection efficiency is likely dependent on the type, use restriction, governing agency, size and age of the protected area (Paiva et al., 2015).
Landscapes composed of open and forested remnants create mosaics of heterogeneous habitats throughout the Cerrado (Silva et al., 2006).As a consequence, the Cerrado supports a high level of biological diversity, including 209 amphibian species (Valdujo et al., 2012) and 103 squamate reptiles (Nogueira et al., 2011).A significant part of this biodiversity can be found in protected areas, which may hold high species richness, both at local and regional scales (Nogueira et al., 2009;Morais et al., 2012;Santoro & Brandão, 2014).These areas may be effective for long-term in situ conservation of biodiversity and the maintenance of species' genetic variability (Bensusan, 2006;Luque et al., 2012).Thus, knowledge of the species occurring in protected areas is essential for the development of conservation strategies.
Increases in the number of wildlife inventories has improved knowledge regarding amphibians and reptiles, resulting in the expansion of the area of occurrence (range) of many species and descriptions of new species (e.g., Brasileiro & Haddad, 2015;Vaz-Silva et al., 2015), thus facilitating studies of distribution patterns (e.g., Vasconcelos et al., 2014).However, there are still remains large sampling gaps, most notably in remnants of seasonal semi-deciduous forest in the central region of the Cerrado.This type of forest formation occurs on interfluves and places rich in nutrients, where fallen leaves contribute to increased organic matter in the soil (Ribeiro & Walter, 2008).Due to their high productivity, these forested areas have been exploited for decades and now are restricted to a few small preserved patches (Moura & Bucci, 1981).Currently, between 20% and 60% of the semi-deciduous forest of the Cerrado has been converted into agriculture and pasture areas, and have suffered from wood extraction and other human impacts (Silva et al., 2006).Given this scenario of degradation and loss of natural habitats, studies on biodiversity in semi-deciduous seasonal forest remnants are of extreme importance, both to fill sampling gaps and to improve basic knowledge regarding the conservation status of species.In this context, the present study pres-ents data about the species richness and composition of the herpetofauna of Parque Estadual Altamiro de Moura Pacheco, one of the last protected areas composed of remnants of semi-deciduous seasonal forest in the core region of the Brazilian Cerrado.

Study area
Parque Estadual Altamiro de Moura Pacheco (PEAMP) is a protected area created to preserve one of the last remnants of seasonal and dry forest in the state of Goiás, Brazil (Fig. 1).The total area of PEAMP is about 32.46 km², and the predominant vegetation is seasonal semi-deciduous forest, which covers approximately 75.54% of its extension (Campos & Batista, 2010).This forest vegetation is characterized by an arboreal strata with an average height of between 15 and 25 m and canopy cover ranging from 70 to 95% in the rainy season and 50% or less in the dry season (Ribeiro & Walter, 2008).The vegetation of PEAMP also possesses gallery forests and small patches of woody savannah (Cerrado sensu strictu) and forest-like savannah (Cerradão).It is surrounded by landscapes modified for agricultural activities and urban centers.The main watercourse is the João Leite River, a tributary of the Meia Ponte River and one of the main tributaries of the Paranaíba basin.During field sampling (2007 and 2008), the PEAMP had dams and flooded zones covering approximately 0.03% of its area (Campos & Batista, 2010).This percentage is currently much higher since part of the vegetation was removed for the insertion of the João Leite Reservoir (between 2009 and 2010), which was constructed to supply water to the municipality of Goiânia (Fig. 1).
The phytophysiognomies sampled in PEAMP were classified as: DA = disturbed areas; HF = hygrophilous environments next to forest fragments; HO = hygrophilous environments in open areas; RF = riparian or gallery forest; SF = seasonal semi-deciduous forest; and VE = palm grove marsh (vereda) (Table 1).The climate of the region is characterized tropical (Aw of Köppen), with two well-defined seasons, one rainy and one dry (Peel et al., 2007).The average monthly temperature is 24.77°C, with the highest temperatures in September (26.90°C) and the lowest in June (22.79°C).The average monthly relative humidity is 59.61%, with highest humidity in December (72.72%) and the lowest in August (37.67%).The average monthly rainfall is 137.80 mm, with the greatest rainfall in December (319.68 mm) and the least in August (2.34 mm) (averages obtained for 2007 to 2016; Brasil, 2018).

Data collection
Fieldwork was conducted between March 2007 and April 2008 during six field trips (March, May, July and October 2007;January and April 2008), each lasting ten days for a total of 60 sampling days.We use a combination of two field sampling methods: pitfall trap arrays with drift fences (see Ribeiro-Júnior et al., 2011) and visual and auditory searches (Heyer et al., 1994).Eight sites were selected to set up the pitfall trap arrays (Table 1), with each site containing a single array of four 20-liter buckets buried in the soil in a "Y"-shaped arrangement and interconnected by a 5-meter long, 50-cm high bar-rier.The total sampling effort with this method was approximately 46,080 bucket hours.We conducted active searches along 14 transects of ~1,000 meters in length (Table 1), during morning (09:00h-12:00h), afternoon (15:00h-18:00h) and night (19:00h-22:00h).Each transect was sampled for approximately one hour by two or three researchers.The total effort resulting from active searches was approximately 504 person hours.
Voucher specimens were collected, anesthetized with 5% lidocaine, fixed in 10% formalin and preserved in 70% alcohol.The compilations available in Frost (2018) and Costa & Bérnils (2018) were followed for nomenclature for anurans and reptiles, respectively.Voucher specimens were deposited in Coleção Zoológica da Universidade Federal de Goiás (ZUFG) and Coleção de Herpetologia do Centro de Estudos e Pesquisas Biológicas of the Pontíficia Universidade Católica de Goiás (CEPB), Goiânia, Goiás, Brazil (collection permit #015/2007, process #5601.01383/2007-4).Species of amphibians and reptiles collected on other occasions by other researchers and placed in the collections of CEPB and ZUFG were also considered in the compilation of the species lists (Appendix 1).
Species distribution patterns and the degree of association with the Cerrado were obtained from information available in Valdujo et al. (2012) for amphibians and Nogueira et al. (2011) for reptiles.Species with occurrences in more than two biomes were considered widely distributed.The conservation status of the species was based on the Red List of Threatened Species of the International Union for Conservation of Nature (IUCN, 2017).

Amphibians
We recorded 35 anuran species distributed among seven families (Table 2; Figs.2-4).The richest families were Hylidae and Leptodactylidae, both with 13 species, followed by Microhylidae (three species),  2).The majority of amphibian species were recorded in hygrophilous environments near forest fragments (26 species), open areas (24 species) and riparian or gallery forests (14 species).Twelve species were recorded in seasonal semi-deciduous forest, and nine in disturbed areas and palm grove marsh.Fourteen species are endemic to the Cerrado and four occur in Cerrado and Atlantic Forest, while the remaining 14 species are widely distributed.None of the species are considered at any level of threat with most being classified as Least Concern, while three have yet to be evaluated (Table 2).

Reptiles
We recorded 29 reptile species distributed among 13 families of two orders (Table 3; Figs. 5 and 6).For  Squamata, the richest families were Dipsadidae (10 species) and Boidae (three species); five other families (Mabuyidae, Tropiduridae, Teiidae, Amphisbaenidae and Viperidae) were represented by two species, and four families (Gekkonidae, Dactyloidae, Anguidae and Colubridae) by one species.Testudines were represented by one species of Chelidae and one of Testudinidae (Table 3).Most reptile species were recorded during active searches (27 species), with 22 being exclusively en-countered by this method.We recorded five species using pitfall traps, with no species being exclusively found by this method.The species Amphisbaena anaemariae and Atractus albuquerquei were included from secondary data (specimens housed in scientific collection -CEPB) (Table 3).
The majority of species were recorded in seasonal semi-deciduous forest (20 species) and riparian or gallery forests (16 species).Thirteen species were recorded in None of the species are considered at any level of threat; however, the majority has yet to have their conservation status evaluated (Table 3).

Amphibians
The species richness of amphibians found in PEAMP was similar to that of other localities with a predominance of open vegetation (Kopp et al., 2010;Ramalho et al., 2014;Santoro & Brandão, 2014) and Cerrado forests  (Morais et al., 2012;Melo et al., 2013).This high richness testifies to the importance of the area for species preservation, especially since it is located close to regions of intense urban development.
The predominance of species from the families Hylidae and Leptodactylidade is a common pattern in the Neotropical Region (Duellman, 1999;Wiens et al., 2011), including studies performed close to PEAMP (Campos & Vaz-Silva, 2010;Morais et al., 2012;Santos et al., 2014), thus indicating that the sampling methods allowed the recording of species with different ecological requirements (Lynch, 2005).The active search method was more effective than pitfall traps for anuran sampling.However, complementary sampling methods need to be considered since some fossorial species and/or species with explosive breeding (e.g., D. muelleri) are more difficult to find using observational sampling methods.
Although we recorded many species that are considered habitat generalists (e.g., Boana albopunctata, B. raniceps, Leptodactylus labyrinthicus, L. podicipinus and Physalaemus cuvieri), none of them occurred in all of the sampled environments.The majority of anurans were found in hygrophilous environments, where Boana paranaiba and Physalaemus atim were found exclusively next to forest fragments, while Pseudopaludicola facureae is typical of open areas.Seasonal semi-deciduous, riparian and gallery forests also possessed several exclusive species, such as Rhinella sebbeni, Barycholos ternetzi, Proceratophrys goyana, Odontophrynus cultripes, Leptodactylus mystaceus and Chiasmocleis albopunctata.
Other species seem to depend on waterbodies with forest cover, such as Boana lundii, B. paranaiba and P. atim.The differing preferences among some species for open or forest environments indicates spatial segregation (Gambale et al., 2014;Santos et al., 2014;Dória et al., 2015).
A significant part of the anuran richness found (40%) is endemic to the Cerrado Biome, where some were re-corded exclusively in forest environments (e.g., R. sebbeni, B. ternetzi, P. goyana, O. cultripes, C. albopunctata).Among these, P. atim and R. sebbeni were recently described (Brasileiro & Haddad, 2015;Vaz-Silva et al., 2015), and thus information about their natural history and geographic distribution are lacking, making the assessment of their true conservation status difficult (Morais et al., 2018).Research efforts should be directed towards filling the knowledge gaps for these species, including information about their ecology, natural history and distribution (Guerra et al., 2018).Non-evaluated species represent taxonomic uncertainty (Scinax aff.fuscovarius, Adenomera aff.hylaedactyla and Leptodactylus gr.melanonotus) or are still under evaluation because they have recently been described (e.g., P. facureae).Nonetheless, none of the species are included threat categories.

Reptiles
The reptile richness found in PEAMP was similar to that of other localities with a predominance of forest environments in the core region of the biome (Morais et al., 2012;Oda et al., 2017); however, it can still be considered low when compared to regions with open vegetation types (Recoder & Nogueira, 2007;Valdujo et al., 2009;Recoder et al., 2011).Therefore, although the richness of reptiles is within that expected for Cerrado areas, which varies from 15 to 70 species (Costa et al., 2007), the finding confirms a pattern of greater richness in open environments in the biome (Nogueira et al., 2005(Nogueira et al., , 2009)).
The predominance of species from the family Dipsadidae is a common pattern often for the Neotropical Region (Cadle & Greene, 1993;Costa & Bérnils, 2018).This family includes snakes from different ecological groups (Vidal et al., 2010), and is the richest snake family in the Neotropical Region (Uetz et al., 2018).Other squamate families were represented by few species, cor- roborating what has been found in other Cerrado areas (França & Braz, 2013;Oda et al., 2017;Rios et al., 2017).Interestingly, families composed of species common in open phytophysiognomies of the Cerrado, such as those of Gymnophthalmidae, as well as some other lizard (e.g., Ameivula ocellifera) and snake (e.g., Bothrops pauloensis) species typical of open environments, were not found.
The absence or rarity of these species seems to differen-tiate the structure of the reptile community of PEAMP from other nearby localities with a predominance of open environments.The majority of reptile species was associated with forest environments and many were found exclusively in these environments, including chelonians (e.g., Chelonoidis carbonarius), lizards (e.g., Notomabuya frenata, Ophiodes aff.striatus) and snakes (e.g., Epicrates cras-Table 3. List of reptile species recorded in Parque Estadual Altamiro de Moura Pacheco, Goiás, Brazil, and   sus, Apostolepis albicollaris, Oxyrhopus petolarius digitalis).The snake Erythrolamprus reginae macrosoma was also always associated to sites with waterbodies.By contrast, open environments had lower richness and few exclusive species, such as Amphisbaena alba, Sibynomorphus mikanii mikanii and Taeniophallus occipitalis.This pattern of segregated distribution is well known for the Cerrado biome (Nogueira et al., 2009)  2011), the snake A. albuquerquei should be considered as having wide distribution due to recent records in the Amazon (França et al., 2017).The number of endemic squamate reptiles recorded for PEAMP can be consid-ered low relative to the high levels expected in regions of open high plateaus (Nogueira et al., 2011).None of the reptile species recorded are considered at any level of threat.As one of the last protected areas of seasonal semi-deciduous forest of the Cerrado, PEAMP is very important for biodiversity conservation.The anuran and reptile richness found is regionally representative, with communities typical of forest environments.The communities are composed by endemic species that are mainly adapted to forest areas, habitat-specific and with restricted distributions, and thus are in need of more research efforts to obtain information on geographic distribution, ecology and natural history.Furthermore, since approximately 50% of the points sampled during the present study were flooded during the filling of the João Leite reservoir, the data presented here will be valuable to any future research that assesses the impact of the insertion of the reservoir on the local herpetofauna.

Figure 1 .
Figure 1.Geographic location of Parque Estadual Altamiro de Moura Pacheco in the state of Goiás, central Brazil.Sampling sites 2-8, 11 and 13 are currently flooded by the João Leite Reservoir.
their occurrence, sampling method, and distribution pattern in the Cerrado Biome.Threat status (TS): NA = Not Evaluated, LC = Least Concern.Distribution (DS): END = Endemic, W = Widely distributed, CE = Cerrado, AF = Atlantic Forest.Habitat: DA = Disturbed areas, HF = Hygrophilous environments next to forest fragments, HO = Hygrophilous environments in open areas, RF = riparian or gallery forest, and SF = Seasonal semi-deciduous forest.Method: AS = Active search, SC = Scientific collection, PT = Pitfall trap.
, and demonstrates the lim-itations that open and forest environments represent for the distributions of reptiles.Most reptile species recorded (72%) have wide distributions among Brazilian biomes, while one lizard (Norops brasiliensis), three snakes (E.crassus, A. albicolaris and Bothrops moojeni) and one amphisbaenid (A.anaemariae) are endemic to the Cerrado.Despite being classified as a Cerrado endemic (Nogueira et al.,

Table 1 .
Sampled sites in Parque Estadual Altamiro de Moura Pacheco, Goiás, Brazil, with sampling methods used, geographic coordinates and type of environments sampled.
this method.Twelve species were recorded using pitfall traps, with Dermatonotus muelleri being exclusively sampled by this method (Table