Non-Volant mammals of a remnant of the Atlantic Forest in northeastern Brazil

This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0), which permits reproduction, adaptation, and distribution provided the original author and source are credited. 1 Universidade Federal da Paraíba, Programa de PósGraduação em Ciências Biológicas Zoologia. Cidade Universitária, Castelo Branco III, 58051-900, João Pessoa, PB, Brazil. 2 Universidade Federal de Sergipe, Núcleo de Educação em Ciências Agrárias e da Terra. Rodovia Engenheiro Jorge Neto, km 3, Centro, 49680-000, Nossa Senhora da Gloria, SE, Brazil. 3 Universidade Federal do Sergipe, Programa de PósGraduação em Ecologia e Conservação. Av. Marechal Rondon, s/n, Rosa Elze, 49100-000, São Cristóvão, SE, Brazil. Non-Volant mammals of a remnant of the Atlantic Forest in northeastern Brazil


Introduction
The Brazilian Atlantic Forest is considered a hotspot for biodiversity conservation (Myers et al., 2000;Fisher and Christopher, 2007) due to its species richness, endemism and habitat loss.Recent estimates indicate between 11.4% and 16% of forest remaining, from the original forest cover, mostly in relatively small fragments (Ribeiro et al., 2009).In Brazil, the Atlantic Forest is the second richest biome in terms of mammalian diversity, after the Amazon basin, and around 30% of its species are endemic (Paglia et al., 2012).About 15% of the 732 mammal species found in Brazil are considered to be under some threat of extinction (Chiarello et al., 2008;Paglia et al., 2012;MMA, 2017), of which more than a third are found in the Atlantic Forest (MMA, 2014).
While the Atlantic Forest coincides with the most densely-populated region of Brazil, surprisingly few systematic inventories of its non volant mammalian communities have been conducted, and the vast majority of studies undertaken over the past thirty years have focused on sites in the southern half of the biome.The non volant mammalian faunas recorded at these sites have varied considerably in richness, ranging from nine species (Oliveira et al., 2005;Pardini and Umetsu, 2006) to 56 (Modesto et al., 2008a).
Very few inventories are available for the northern half of the biome (Stevens and Husband, 1998;Oliveira et al., 2005;Percequillo et al., 2007;Silva and Mendes-Pontes, 2008;Chagas et al., 2010;Feijó et al., 2016;Table 1), and one of the primary aims of the present study is to redress this imbalance, especially given the potential importance of regional variation in the composition of the mammalian fauna for the conservation and management of the biome (Beisiegel, 2010).The present study reports a survey of the mammal species of the Mata do Junco Wildlife Refuge, in the northeastern Brazilian state of Sergipe.

Study site
The Mata do Junco Wildlife Refuge (RVS Mata do Junco; RVSMJ) covers a total area of approximately 900 hectares, in the municipality of Capela (Figure 1), state of Sergipe, northeastern Brazil (10º 32' S, 37º 03' W; 10º 34' S, 37º 00' W).One of the aims to the establishment of the     RVSMJ was the protection of the local population of Callicebus coimbrai Kobayashi andLangguth, 1999 (Sergipe, 2007), an endangered (EN) primate species, according to the IUCN Red List (see Veiga et al., 2008), endemic to the Atlantic Forest of Sergipe state and the north shore of Bahia state.The vegetal cover is a semi-deciduous and open forest of slope, with many clearings, surrounded by an anthropogenic matrix, consisting mainly of sugarcane (Saccharum sp.) plantations and cattle pasture.However, the vegetation is well structured, with three main strata -(i) understory with bushes and herbaceous plants; (ii) lower canopy with small trees and lianas; and (iii) main canopy with some emergent trees over 15 m high.The megathermic sub-humid climate is characterized by a rainy winter (April-August) and dry summer, classified as As by Köppen's climate classification (Alvares et al., 2013), with a mean annual temperature of 24.9°C and a mean annual precipitation of 1,372 mm (SEMARH, 2007), in a slope relief ranging from 40 to 120 m above sea level.

Data collection
We conducted the samplings on three days each month, between June 2011 and May 2012.We use complementary approaches to record the presence of cryptic mammal species, both medium-large (adults > 1 kg) or small (adults < 1 kg) species, according to Rocha and Dalponte (2006) and Paglia et al. (2012).The occurrence of medium and large species were verified based on a combination of approaches, as the direct observation, vocalizations, and opportunistic indirect observation of vestiges, such as feces, tracks and footprints, burrows and bones.The tracks were identified based on Becker and Dalponte (1991).Additionally, we considered the presence of some cryptic and rare mammal species, according to the report of few local informers, mainly local residents and landowners close to the Refuge for over 50 years, through non-systematic informal talks.
For small mammal species, we set 36 Sherman type traps (22.5 x 7.8 x 9.2 cm) per night, along three transects in the dense forest, each with six capture points, 50 m apart.Each point included two traps, one set on the ground (substratum), and other in the vegetation (understory), up to 2 m above the ground.We baited the traps with a mixture of peanut candy, sardines, corn and banana, checked in the following morning.
We moved trapped animals to the laboratory, at field station, to the identification and record morphometric parameters, as total body length (mm), caudal length (mm), anterior/posterior foot length (mm) and weight (g).We use coded incisions on the ear to mark the released specimens.Collected specimens were handled in accordance with the Animal Care and Use Committee (Sikes et al., 2011).Species were identified through Gardner (2007), Voss and Jan-sa ( 2009), and Feijó and Langguth (2013).We deposited the voucher specimens (skin and/or skull) of different species in dry conditions, both in the Conservation Biology Laboratory (LBC), and in the mammalian collection of the Federal University of Sergipe (CMUFS), São Cristóvão, Sergipe, Brazil (Appendix 1).

Analyses
We ran the Jackknife 1 species richness estimator and species accumulation curve (observed and estimated) based on 1,000 replications (Colwell and Coddington, 1994), in the software EstimateS 8.0 (Colwell, 2005).
Despite the low richness accessed by traps, estimated species richness according to Jackknife 1 was S = 4.9 species, which indicates that the results of the study were relatively satisfactory, given that 81.3% of this total was recorded.This assumption is supported by the fact that the standard deviations of the two curves overlapped at many points (Figure 2).
The complimentary methods identified 12 species, six of them through direct observation, while other three were registered exclusively through tracks and/or bones (Figure 3).The three remaining additional species, Leopardus sp.Gary 1842, Puma yagouaroundi (É.Geoffory Saint-Hi- lare, 1803) and Galictis cuja (Molina, 1782), were registered through the reports of local residents.Curiously, all the reported species were from the order Carnivora, which may indicate an actual absence of some top predators.

Discussion
The most interesting aspect of the data was the absence of native species of rodents, despite the sampling effort (1,296 trap-nights), although 31 marsupial specimens (representing three different species) have been collected during the sample period.This might indicate some factor limiting the occurrence and/or abundance of rodents at the study site.
It seems unlikely, however, that rodents are absent altogether, and the lack of records in the present study may be due to a combination of factors, including the type of bait (Astúa et al., 2006), the distance from the edge of the forest, and the type of trap (Cáceres et al., 2011;Hice and Velazco, 2013).The recent establishment of the study site as a protected area (Sergipe, 2007) may reinforce the idea of edge effect, associated with historical hunting pressure.Therefore, the maintenance of habitat generalist species is expected (Silva and Pontes, 2008), as well as the recolonization of the area by some species considering their vagility.
Together with the species confirmed in the present study, a total of 30 non-volant mammals are known to oc- cur in the Atlantic Forest of Sergipe (Oliver and Santos, 1991;Stevens and Husband, 1998;Oliveira et al., 2005;Chagas et al., 2010;Rocha et al., 2012).Considering only surveys based on procedures comparable to those applied in the present study, species richness varied considerably, from nine at the Serra de Itabaiana National Park (Oliveira et al., 2005) and Morro Grande Forest Reserve (Pardini and Umetsu, 2006) to 36 at the EPDA-Peti reserve in Minas Gerais (Paglia et al., 2005).A number of factors may have influenced this variation, including zoogeographic limitations (Mares, 1975) and human impacts, as well as the size and age of fragmentation (Brooks and Rylands, 2003) and hunting pressure (Cullen Jr. et al., 2000).This variation reinforces the need for more detailed data from a larger number of sites, as well as continued monitoring in already sampled sites, for a more systematic understanding of the diversity and zoogeography of the mammals of the Atlantic Forest.
The number of recent surveys available for the Atlantic Forest (Table 1) is not only relatively limited, in comparison with the dimensions of this biome, but is also concentrated in the South and Southeastern regions of Brazil (Figure 1).Filling the vast knowledge gaps in the Northeastern region (or even further in the South) is important not only for the understanding of the diversity of mammals within this region in general, but also the local distribution of specific taxa.A number of recent studies have redefined the local ranges of species such as Bradypus torquatus Illiger, 1811 (Chagas et al., 2010) and Marmosops incanus (Rocha et al., 2012).
In relation to the species not found in the RVSMJ, but registered in other areas in the Sergipe state, there are at least three mammal species, which are at risk of local extinction, the rodents Chaetomys subspinosus (Olfers, 1818) and Coendou insidiosus Lichtenstein, 1818 (Oliver and Santos, 1991), and the feline Puma concolor (Linnaeus, 1771) Non-Volant mammals of a remnant of the Atlantic Forest in northeastern Brazil (Chagas et al., 2010).These species have their records just up to the Vaza-Barris river, south portion of the Atlantic Forest in the state.This river may be a geographic barrier to their distribution, or yet those species may have suffered local extinction in other areas.Other species that have their distributions expected to the RVSMJ, as the wild hogs Tayassu pecari (Link, 1795) and Pecari tajacu (Linnaeus, 1758), probably have already been extinguished throughout the state, since there are not even anecdotic records.
Therefore, we highlight the importance of a continued monitoring of the community of mammal species in the RVSMJ.This monitoring may contribute to the understanding the role of ecological succession in the mammal community, as a consequence of the protection of the area.Ultimately, the construction of a vast database on the occurrence and distribution of mammals in the Atlantic Forest will provide an important tool for the development of effective conservation and management strategies.This database is an important management tool, in order to define the priority areas for conservation, development of action plans for endangered species, planning sustainable landscape, and the establishment of new protected areas, as proposed by Costa et al. (2005).

Figure 1 .
Figure 1.Location of the Mata do Junco Wildlife Refuge in Capela, in the northeastern Brazilian state of Sergipe.The numbered points represent other sites with inventories of mammals in the Atlantic Forest (see Table1for localities).The Mata do Junco Wildlife Refuge is the green detail (bottom right).

Figure 2 .
Figure 2. Observed and estimated (Jackknife 1) cumulative species curves (vertical lines = standard deviation) for the community of small mammals of the Mata do Junco Wildlife Refuge, Sergipe, northeastern Brazil, between June 2011 and May 2012.

Table 1 .
Neotropical Biology and ConservationNon-Volant mammals of a remnant of the Atlantic Forest in northeastern Brazil Localities with inventories of mammals in the Atlantic Rainforest.The codes represent the sites in the Figure1.

Table 2 .
Non-volant mammal species recorded at the Mata do Junco Wildlife Refuge in Sergipe, Brazil, between June 2011 and May 2012.C = Captured; O = Observed; V = Vestiges*; R = Reports; S = Carcass.