Bats of Alagoa Grande , a semi-arid area of 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, Laboratório de Mamíferos, Departamento de Sistemática e Ecologia, Centro de Ciências Exatas e da Natureza. Campus I, s/n, Cidade Universitária, Castelo Branco, 58051-900, João Pessoa, Paraíba, Brazil. 2 Universidade Federal da Paraíba, Programa de PósGraduação em Ciências Biológicas, Departamento de Sistemática e Ecologia, Centro de Ciências Exatas e da Natureza. Campus I, s/n, Cidade Universitária, Castelo Branco, 58051-900, João Pessoa, Paraíba, Brazil. 3 Universidade Federal da Paraíba, Departamento de Engenharia e Meio Ambiente, Centro de Ciências Aplicadas e Educação. Campus IV, s/n, Cidade Universitária, 5829700, Rio Tinto, Paraíba, Brazil. Bats of Alagoa Grande, a semi-arid area of Northeastern Brazil


Introduction
Chiroptera is the second most diverse order of mammals, including nearly 22% of the known living mammalian species, with remarkable ecological and trophic diversity (Wilson and Reeder, 2005;Reis et al., 2007).So far, 59 species from eight families are known to occur in the state of Paraíba (Feijó and Langguth, 2011;Ferreira et al., 2013;Nunes et al., 2013;Vilar et al., 2015;Beltrão et al., 2015;Feijó et al., 2015a), but most studies are concentrated in the Atlantic Forest biome.The Caatinga remains poorly sampled (Feijó and Langguth, 2011;Feijó et al., 2016).
The Caatinga is a semi-arid biome characterized by strong seasonality, irregular rainfall, high mean temperatures and long periods of drought (Menezes et al., 2012).It has a long story of scientific negligence (Santos et al., 2011) that has only started to be well-addressed recently.In the last 14 years, the bat richness known from the Caatinga has increased in 20%, currently with 81 species (Oliveira et al., 2003;Paglia et al., 2012;Moratelli and Dias, 2015;Rocha et al., 2015;Feijó et al., 2015a;Feijó et al., 2015b).Recent studies have shown that Caatinga areas can maintain bat assemblages with similar levels of diversity to the moist tropical areas, which highlights the importance to expand the network of protected areas in this threatened biome (Beltrão et al., 2015).
Bat fauna is surveyed mainly by mist-nets set in the ground level (Flaquer et al., 2007), and usually employing six hours of capture per night, right after dusk, corresponding to the first peak of activity to most bat groups (Pedro and Taddei, 2002;Aguiar and Marinho-Filho, 2004).Bat inventories with whole-night sampling effort (12 hours) have not been conducted so far in the Caatinga or even in the northeastern of Brazil (Esbérard and Bergallo, 2005).Herein we report the results of a whole-night bat survey conducted in Alagoa Grande, an area of Caatinga in the state of Paraíba, northeastern Brazil.

Study area
The survey was held in Fazenda Riachão do Progresso (FRP), located in the municipality of Alagoa Grande, state of Paraíba (07º06'01" S; 35º35'57" W), northeastern Brazil.The area is located in the mesoregion of Agreste, within the Caatinga ecosystem, presenting As' climate on Köppen's classification, with the rainy season from March to July (Sudema, 2011).The vegetation is 'Caatinga alta' (sensu Mares et al., 1981), with canopy reaching 12m.The sampling site was a dry forest fragment of scrub-arboreal Caatinga surrounding an artificial lake, which is used by locals for fishing and recreational purposes, as well as water supply.The area is surrounded by pastures destined to cattle raising.

Sampling
We conducted 15 nights of sampling from September/2010 to January/2011, sampling three consecutive nights per month.Nine ground-level mist-nets (7 x 3 m) were opened at 5h p.m. (dusk) and closed at 5h a.m.(dawn), totalizing 12 hours of sampling effort per night.The mist-nets were installed in juxtaposition at the same spots in all sampling nights, following the margin of the lake.Two additional samplings were held by active search in roosts during daytime, to search for species that could be missed by the mist-nets.
Each specimen captured was manually removed from the mist-nets, had the time of capture annotated, the sex determined, anatomic measurements taken using a digital calliper (0.01 mm precision) and their weight using a portable scale (Pesola TM ).Their age class (juvenile, sub-adult or adult) was determined by the calcification of the metacarpal phalanx (Kunz and Anthony, 1982).All captured individuals were euthanized and deposited as voucher specimens at Coleção de Mamíferos of Universidade Federal da Paraíba.All samplings under license #20321-2 granted by Ministério do Meio Ambiente (MMA).

Data Analysis
The sampling effort was calculated according to Straube and Bianconi (2002), multiplying the area of mistnets installed by the amount of sampling hours.We performed a Chao 1 estimator (Chao, 1984) and a species' accumulation curve with 95% confidence intervals that were compared to infer the sampling sufficiency on Es-timateS (Colwellm, 2000).To evaluate the differences on abundance and species richness between the first (5h p.m. to 11h p.m.) and second half of the night (11h p.m. -5h a.m.) (pre-mid and post-mid), the data was screened by a Shapiro-Wilk Test to check normality, and compared by a Wilcoxon rank sum test with continuity correction, analysis performed on R (R Core Team, 2014).

Results
The sampling effort totalized 34.020 h.m 2 .We captured 56 individuals distributed in 5 families, 11 genera and 12 species (Table 1).The family Phyllostomidae comprised the majority of the species registered (N = 6) and the second most abundant species, Artibeus planirostris (Spix, 1823), corresponding to 23% of the captures.The family Emballonuridae was represented by two species.Rhynchonycteris naso (Wied-Neuwied, 1820) was the most abundant species, being represented in 30.7% of the total captures.Only individuals of Glossophaga soricina (Pallas, 1766) were captured in roost samplings.None of the species captured were considered endangered at country and global levels (Chiarello et al., 2004;IUCN, 2014).The estimator Chao 1 showed an average richness of 18 species (Figure 1).
Regarding the comparison to the bat richness and abundance between the two halves of the night (Figure 2), we found that both had the same richness (9 species), and that two species, Carollia perspicillata (Linnaeus, 1758), and Lophostoma brasiliense Peters, 1866, had the same abundances, and T. cirrhosis and G. soricina differed only in one individual between both halves.Rhynchonycteris naso and A. planirostris, presented higher capture rates at the second half of the night.Myotis lavali Moratelli, Peracchi, Dias andOliveira, 2011, Molossus molossus (Pallas, 1766) and Peropteryx.macrotis (Wagner, 1843) were recorded solely in the first half of the night, whereas Artibeus lituratus (Olfers, 1818), Cynomops planirostris (Pe-ters, 1865) and Noctilio leporinus (Linnaeus, 1758) were registered only in the second half.Six of the nine species registered in the second half of the night were recorded in a single capture, including A. lituratus, C. planirostris and N. leporinus.With the exception of Myotis lavali, all species that were exclusive to one of the halves had only a single individual captured.The Wilcoxon test pointed no significant difference between both halves on richness (W = 152, p = 0.08256) but detected differences on abundance (W = 164, p = 0.02959), the second half of the night having 1.8 times the individuals captured (36, compared to 20 for the first half of the night).Analyzing the total captured individuals per night (first, second and third successive nights) in each month of sampling, we found no trend of diminishing in abundance throughout the consecutive nights (Figure 3).

Discussion
Our work recorded a low species richness compared to other surveys in the Caatinga.Novaes and Laurindo (2014) 2008), with similar effort to our study, sampled only seven species for a Caatinga site in Bahia state.The low species richness in the area may be associated to the high level of anthropic landscape alterations, as intense livestock and wood exploitation, and few sparse patches of original vegetation.On the other hand, the difference between the estimated and the observed richness indicates the sampling effort performed was not sufficient, as we surveyed only 67% of the richness predicted by Chao 1.For instance, despite the presence of cattle, frequently attractive to hematophagous bats (Anderson et al., 2012), no Desmodontinae bat was captured in our study.
Five of the nine species recorded in the present work were singletons.Among these species, Peropteryx macrotis, Cynomops planirostris and Noctilio leporinus had between 1 and 4 individuals captured by ground level mist-nets in other studies (Gregorin et al., 2008;Novaes and Laurindo, 2014;Rios et al., 2008;Sá-Neto and Marinho-Filho, 2013;Silva et al., 2004).This result could be related to the inefficiency of the chosen sampling method to capture insectivorous species.Another study in the Caatinga captured 21 individuals of Peropteryx macrotis by sampling daily roosts (Beltrão et al., 2015).
In surveys with continuous sampling effort, Esbérard (2006Esbérard ( , 2009) ) detected a declining trend on capture rates during consecutive sampling nights, and a negative trend on capture rates for common species with continuous effort, suggesting that bats might learn the position of the mist-nets and avoid them.Our results suggest that bats did not learn the position of the mist-nets, as the capture rates only slightly change over the three nights sampled each month.Among the sampled species, we must highlight the unusual high abundance of Rhinchonycteris naso   possibly due to the proximity of mist-nets to a colony of the species, what has also affected the results of the comparison between the two halves of the night: when excluding Rhinchonycteris naso from the statistics, the Wilcoxon test detects no difference between the two halves of the night both for abundance (W = 49, p = 0.4408), and richness (W = 42.5, p = 04023).The final abundance was only detected due to the whole night sampling, as most of the individuals were captured during the second half of the night (Figure 2).Furthermore, Lophostoma brasiliense represents the first record for the Caatinga of the Paraíba state.Records of five insectivorous (Rhinchonycteris naso, Peropteryx macrotis, Molossus molossus, Cynomops planirostris, Myotis lavali) and a piscivorous species (Noctilio leporinus) may be explained by the location of mist-nets, which were set up at the the margin of a perennial body of water.This water body works as a refuge area during the dry season to those species in highly seasonal sites like Caatinga.
The majority of the Neotropical bats have their activity peak immediately before, during or right after sunset when they leave their daytime roosts (Avery, 1986;Erkert, 1978;Catto et al., 1995), which justify the common practice of sampling bats for the first half of the night.However, our data highlighted the importance of a whole-night surveys, mainly in short samplings.Three species (25% of the richness), A. lituratus, N. leporinus and C. planirostris, were sampled only during the second half of the night, as well as the majority (about 64%) of the captures in this study.In contrast, M. lavali, M. molossus and P. macrotis were sampled only during the first half of the night.
Increasing the sampling effort commonly result in higher richness and abundance (Bergallo et al., 2003;Esberard and Bergallo, 2005), what could be achieved by adding more collection days or more hours per night.The first choice allows to detect seasonal variations (food availability, weather oscillations), as well as minimizing the effects of punctual effects (e.g., moonlight intensity, storms, mist).The second choice allows sampling bats both leaving and returning to their daytime roosts, maximizing the chances of capture in different activity periods.However, logistically, 12-hour sampling nights might hinder the practice of diurnal active search in roosts, a practice that allows recording species that normally are not trapped in mist-nets (Voss and Emmons, 1996).Hence, in long term studies, we advocate that it is more recommendable to increase the number of sampling nights, making possible the search for daily roost, whereas, in short-span surveys as ours, the use of wholenight sampling might yield a more realistic scenario of the local bat fauna.
recorded 24 species to Chapada do Araripe in the states of Ceará and Pernambuco employing a comparable amount of mist-net hours.Sá-Neto and Marinho-Filho (2013) recorded 31 species to middle São Francisco River in the Bahia state with 7.6 times the sampling effort used herein.Beltrão et al. (2015) recorded 19 species to RPPN Fazenda Almas in the Paraíba state with 3.663.6h.m 2 sampling effort.In addition, Silva et al. (2004) recorded 16 species to Serra das Almas in the Ceará and Gregorin et al. (2008) recorded 22 species to Serra das Confusões in the state of Piauí.On the other hand, Rios et al. (

Figure 1 .
Figure 1.Bat species' accumulation curve, with 95% confidence intervals and Chao 1 richness estimator in an area of Caatinga in the state of Paraíba, northeastern Brazil.

Figure 2 .
Figure 2. Species abundance of bats before (pre-mid) and after (post-mid) 11h p.m. in an area of Caatinga in Paraíba, northeastern Brazil.

Figure 3 .
Figure 3. Bat abundance distribution by day and month of sampling in an area of Caatinga in Paraíba, northeastern Brazil.

Table 1 .
Kalko et al. (1996)Fazenda Riachão do Progresso, Paraíba, northeastern Brazil, from September 2010 to January 2011, discriminated by abundance and relative abundance between the first and second halves of the night.Feeding guilds are assigned according toKalko et al. (1996), and Collection Number as voucher specimens at Coleção de Mamíferos-UFPB.