AMPHIBIAN CONSERVATION IN THE CAATINGA BIOME AND … · AMPHIBIAN CONSERVATION IN THE CAATINGA BIOME AND SEMIARID REGION OF BRAZIL MILENA CAMARDELLI 1,3 AND MARCELO F. NAPOLI 2 1Programa

AMPHIBIAN CONSERVATION IN THE CAATINGA BIOME ANDSEMIARID REGION OF BRAZIL

MILENA CAMARDELLI1,3

AND MARCELO F. NAPOLI2

1Programa de Pos-Graduacao em Ecologia e Biomonitoramento, Instituto de Biologia, Universidade Federal da Bahia,Rua Barao de Jeremoabo, Campus Universitario de Ondina, 40170-115 Salvador, Bahia, Brazil

2Museu de Zoologia, Departamento de Zoologia, Instituto de Biologia, Universidade Federal da Bahia, Rua Barao deJeremoabo, Campus Universitario de Ondina, 40170-115 Salvador, Bahia, Brazil

ABSTRACT: The Brazilian Ministry of the Environment (Ministerio Do Meio Ambiente, MMA) proposeddefining priority areas for Brazilian biodiversity conservation in 2007, but to date, no definitions of priorityareas for amphibian conservation have been developed for the Caatinga biome or the semiarid region ofBrazil. In this study, we searched for ‘‘hot spots’’ of amphibians in these two regions and assessed whether thepriority areas established by the MMA coincided with those suitable for amphibian conservation. Wedetermined amphibian hot spots by means of three estimates: areas of endemism, areas of high speciesrichness, and areas with species that are threatened, rare, or have very limited distributions. We then assessedthe degree of coincidence between amphibian hot spots and the priority areas of the MMA based on thecurrent conservation units. We analyzed areas of endemism with the use of a parsimony analysis ofendemicity (PAE) on quadrats. The Caatinga biome and semiarid region showed four and six areas ofendemism, respectively, mainly associated with mountainous areas that are covered by isolated forests andpositively correlated with species richness. All areas of endemism coincided with one or more priority areasdefined by the MMA. We identify 15 priority areas for amphibian conservation in the Caatinga biome andsemiarid region, including the creation of new full-protection conservation units.

RESUMO: A definicao de areas prioritarias para a conservacao da biodiversidade brasileira foi propostapelo Ministerio do Meio Ambiente (MMA) em 2007, mas ate o presente momento nao ha estudo metodicoque defina areas prioritarias para a conservacao de anfıbios no bioma Caatinga ou no Semiarido brasileiro.Neste estudo, buscamos por ‘‘hot spots’’ de anfıbios nestes dois polıgonos e verificamos se as areas prioritariasdo MMA coincidem com areas adequadas para a conservacao dos anfıbios. Determinamos os hot spots deanfıbios por meio de tres estimadores: areas de endemismo, areas de alta riqueza de especies e areas comespecies ameacadas, raras e/ou com padroes de distribuicao restrita. Em seguida, acessamos o grau decoincidencia entre hot spots de anfıbios e areas prioritarias do MMA, igualmente considerando as unidadesde conservacao correntemente instaladas. A deteccao de areas de endemismo foi realizada pela Analise deParcimonia de Endemicidade (PAE), usando quadrados. O bioma Caatinga e o Semiarido apresentaram,respectivamente, quatro e seis areas de endemismo, principalmente associadas a areas montanhosas cobertaspor florestas isoladas, e estiveram positivamente correlacionadas a riqueza de especies. Todas as areas deendemismo coincidiram com uma ou mais areas prioritarias definidas pelo MMA. Nos identificamos 15 areasprioritarias para conservacao de anfıbios no bioma Caatinga e Semiarido, alem da criacao de novas unidadesde conservacao de protecao integral no Brasil.

Key words: Amphibians; Complementarity analysis; Parsimony analysis of endemicity; Priority areas;Species richness; Threatened species

THE DETERMINATION of priority areas for theconservation of biological diversity has receivedmuch attention in recent years (Tabarelli andSilva, 2004). The Convention on BiologicalDiversity (CBD), signed during the UN Con-ference on Environment and Development(UNCED) in 1992, aimed to generate guide-lines to reconcile the economic development ofnations with the conservation and sustainableuse of biological resources (Tabarelli and Silva,

2004). To meet these obligations, the BrazilianMinistry of the Environment (Ministerio DoMeio Ambiente, MMA) developed its first‘‘assessment and identification of priority areasand actions for the conservation of biomes’’between 1998 and 2000, and updated it in 2007.The definition of the priority areas was based onthe available information about biodiversity andhuman pressures, in addition to the experienceof researchers, and the degrees of priority weredefined with the use of criteria for biolo-gical richness, areas of endemism, and the distri-butions of endangered species. Furthermore,

3 CORRESPONDENCE: e-mail, [email protected]

Herpetologica, 68(1), 2012, 31–47

E 2012 by The Herpetologists’ League, Inc.

31

the importance of these areas to traditionalcommunities and indigenous peoples and theirvulnerability were also taken into account (MMA,2007).

An important element of subjectivity in thelisting of priority areas proposed by the MMA(2007) is that a sizeable part of the informationused in the delineation of these areas came frompersonal and single observations or from non-methodical or poorly designed studies (M.F.Napoli, member of the Caatinga ThematicGroup of MMA, personal communication,2007). A second important aspect is that thelevel of knowledge and interest in each Brazilianbiome and in the various biological groups thatmake up the Brazilian biodiversity is uneven,which led to an arbitrary and subjectiveselection of priority areas for conservation.

Because of this history, it is necessary todetermine the degree of congruence betweenthe priority areas for Brazilian biodiversityconservation as defined by the MMA (2007)and important areas for conservation of thevarious taxonomic groups defined by biolog-ical studies conducted methodically for thispurpose, such as those aimed at identifyingareas of endemism. Areas of endemisminclude unique biota that result from varioustypes of isolation and that, once identified, canilluminate the natural processes that led totheir origins (Silva, 2008).

Another important question to be answeredis whether the Caatinga biome and theBrazilian semiarid region are identical withrespect to biological diversity, as they largelyoverlap geographically. Nevertheless, a biome,as conceived by the Instituto Brasileiro deGeografia e Estatıstica (IBGE, 2004a), is a setof vegetation types identifiable on a regionalscale, including its allied fauna and flora, withsimilar geoclimatic conditions and a sharedhistory of changes that resulted in biologicaldiversity. On the other hand, the semiaridregion is defined based on Brazilian drylandswith similar climatic characteristics (Ministerioda Integracao Nacional [MIN], 2005): averageannual rainfall, aridity index (the relationshipbetween rainfall and potential evapotranspira-tion), and drought risk. Therefore, the Caatingabiome and the Brazilian semiarid region areconceptually different, and hence have differ-ent areas and boundaries.

With respect to amphibians, there havebeen no methodical studies that definepriority areas for their conservation in theCaatinga biome or in the semiarid region ofBrazil, which points to the urgent need to testwhether the priority areas listed by the MMA(2007) for biomes in the Brazilian territoryinclude the hot spots for amphibians in theseregions. If the congruence is not complete,then it will be necessary to redefine thepriority areas for conservation or to proposenew areas to ensure the conservation of hotspots for amphibians in these areas.

This study aims (1) to identify hot spots ofamphibians in the Caatinga biome and semiaridregion of Brazil by combining areas with highlevels of endemism and species richness withareas that contain species that are threatened,rare, or have very limited distributions; and (2)to determine whether the priority areas definedby the MMA (2007) coincide with areasimportant to amphibian conservation.

MATERIALS AND METHODS

Study Area

The Caatinga is the only biome unique toBrazil and has been recognized as a biogeo-graphic province of the Chacoan subregion(Morrone, 2006). This biome is distributedthroughout the northeast region of Brazil(except for the State of Maranhao) and thenorthern part of the State of Minas Gerais(Fig. 1), covers an area of 852,261 km2, andrepresents 11% of the Brazilian territory(MMA, 2007). The Caatinga biome is charac-terized by semiarid to arid climates (Santosand Tabarelli, 2003) with long dry seasons,irregular rainfall (concentrated in the sum-mer), average annual precipitation of 400 to600 mm (Ab’Saber, 1977), high annualtemperatures (ca. 27uC), and a large temper-ature range (MMA, 2007).

Although the most common landscape inthe Caatinga biome is the steppe savanna,which is composed of thorny and deciduousplants and rivers that are largely intermittentand seasonal (MMA, 2007), the biome is notphysically homogeneous and can be dividedinto smaller ecoregions (Velloso et al., 2002).In this landscape, mountain massifs reachelevations of over 2000 m above sea level

32 HERPETOLOGICA [Vol. 68, No. 1

(masl), creating zones of geomorphologic andclimatic uniqueness (Ab’Saber, 1977) thatharbor rocky mountain fields called ‘‘CamposRupestres’’ (for a description of CampoRupestre, see Rizzini, 1979), riparian forests,and seasonal deciduous and semideciduousforests. Of high biological importance, thesemountainous environments and rough terrainsshow high degrees of endemism (Giaretta andAguiar, 1998; Haddad and Abe, 1999; Nani-wadekar and Vasudevan, 2007) and canpresent barriers to the dispersal of variousphylogenetic groups, in addition to acting ascenters of speciation and refuge (Carnaval,2002; Carnaval and Moritz, 2008).

The semiarid region comprises almost theentire area of the Caatinga biome as well asthe transitional areas between this biome andthe Cerrado and Atlantic Forest biomes(Fig. 1), which makes it slightly more hetero-geneous than the Caatinga biome in terms ofphysical features, although it is similar to theCaatinga biome in terms of topography andclimate. The semiarid region represents11.4% of the Brazilian territory, covers anarea of 969,589 km2, and has recently had itsarea and number of municipalities increasedby MIN (2005) because of changes in itsboundaries.

Database Preparation

For this analysis, we considered onlyspecies with clearly defined descriptions anddiagnoses (e.g., species suspected of consti-tuting species complexes were discarded). Weconsidered only anuran amphibians, becausethe fossorial habits of gymnophionans (Duell-man and Trueb, 1986) invariably leads to theundersampling, and because urodeles do notoccur in the study area (Frost, 2009).

The locations considered in this study werethose included in the coverage areas of theCaatinga biome or the semiarid region (statesof the northeast region of Brazil, except forMaranhao, and part of the State of MinasGerais). First, we searched the literature(articles, books, notes on natural history andgeographical distribution, and dissertationsand theses) for amphibian species that occurin the studied areas. We critically reviewedinformation about species and their locationsbefore entered them into the database. Next,we investigated in loco the following Brazilianherpetological collections: the Museu Nacio-nal (Rio de Janeiro State), the Museu deZoologia da Universidade Estadual de Feirade Santana (Bahia State), the Museu deZoologia da Universidade Federal da Bahia(Bahia State), the Museu de Zoologia da the

FIG. 1.—(A) Topographic map with the anuran sampling locations included in the analysis: 1, Sao Francisco River;2, Jequitinhonha River. (B) Distribution map of the 48 quadrats (operational geographic units) considered for analysiswithin the limits of the Caatinga biome and semiarid region of Brazil.

March 2012] HERPETOLOGICA 33

Universidade de Sao Paulo (Sao Paulo State),and the Laboratorio e Colecao de Herpetolo-gia da Universidade Federal da Paraıba(Paraıba State). In the absence of accurateinformation on the geographic coordinates forthe collection site of any particular specimen,we used the geographic coordinates of therespective municipality. We digitized andmapped the locations of all samples, asdescribed below, to determine whether theywere contained within the boundaries of atleast one of the regions under study. Weexcluded from the analysis samples that wereoutside both polygons.

Mapping and Analysis of Georeferenced Data

We converted the geographical coordinatesof each location into decimal format, digitizedthem in ARCVIEW software version 3.2a(Environmental System Research Institute[ESRI], 2000), and projected them over thepolygons of the Caatinga biome and thesemiarid region. We obtained the following areaboundaries (polygons): the Federal states ofBrazil, from the database Digital Chart of theWorld (ESRI, 1997); the Brazilian biomes, fromthe digital vector map ‘‘Biomas do Brasil,’’ scale1:5,000,000 (IBGE, 2004b); the semiarid region,from the digital vector map ‘‘Redelimitacao doSemi-arido Nordestino,’’ scale 1:298.250 (Se-cretaria de Polıticas de Desenvolvimento Re-gional [SPDR], 2005); the bioclimatic suitabilityof the Brazilian Northeast, from the digital map‘‘Aptidao Bioclimatica no Bioma Caatinga,’’ scale1:300,000 (Companhia Hidroeletrica do SaoFrancisco [CHESF], 2001); the climatic suitabil-ity of the State of Minas Gerais from digitalvector maps of ‘‘Agroclimatic zoning of the Stateof Minas Gerais,’’ scale 1:3,000,000 (Geoproces-samento em Minas Gerais [GEOMINAS], 2009);and the priority areas for biodiversity conserva-tion, from the digital vector map ‘‘Revisao dasareas prioritarias para conservacao da biodiversi-dade (importancia biologica)’’ (MMA, 2007). Weobtained the topographic base of South Americafrom the global digital elevation model (DEM)GTOPO30 (US Geological Survey [USGS],1996).

Determination of Areas of Endemism

Areas were considered endemic regions ifthey had congruent distribution limits of two

or more taxa, in accordance with Platnick(1991). In the search for areas of amphibianendemism, we included all anuran speciesoccurring in the studied area, even if they alsooccurred in other biomes, because we soughtto maximize the search for important areas ofamphibian diversity (areas of endemism, highspecies richness, or distribution of relictualspecies) to contribute to the effective mainte-nance of this diversity in the regions studied.

We identified areas of endemism with theuse of the method of parsimony analysis ofendemicity (PAE; Rosen, 1988), with modifi-cations by Morrone (1994). Parsimony analy-sis of endemicity involves five basic steps: (1)the selection of operational geographic units(OGUs), (2) construction of a matrix ofspecies vs. OGUs, (3) parsimony analysis ofthe data matrix, (4) the identification of theOGUs or groups of OGUs with at least twoendemic species, and (5) mapping the loca-tions where the endemic species were foundwithin each OGU or group of OGUs todelineate the boundaries of each area ofendemism. We defined each OGU as aquadrat (grid) of 1u of latitude by 1u oflongitude (Fig. 1) as proposed by Morroneand Escalante (2002). We removed from theanalysis species present in all quadrats be-cause they did not contribute to discrimina-tion among areas. To avoid the use ofsubsampled quadrats, we used only thosequadrats with one or more geographicalsamples were represented by eight or morespecies. This criterion represented the lowestnumber of species obtained in both publishedand unpublished inventories, with collectionefforts considered adequate (minimum 60person-hr per locality in the rainy season inthe Caatinga biome or semiarid region).

We produced two binary data matrices(absence 5 0, presence 5 1): a matrix of 73anuran species (columns, analogous to traits ina cladistic analysis) vs. 35 quadrats (rows,analogous to taxa in a cladistic analysis) for theCaatinga biome and a matrix of 107 anuranspecies vs. 48 quadrats for the semiaridregion. A hypothetical quadrat containing‘‘0’’ in all columns was added as an outgroupin both matrices. These matrices were con-structed and edited in MESQUITE softwareversion 2.71 (Maddison and Maddison, 2009).


A most-parsimonious geographical cladogramwas sought by heuristic search with the use ofthe Willi Hennig Society edition of TNTsoftware version 1.1 (Goloboff et al., 2008),with an estimated 100 replicas for 10,000randomizations and saving 10 trees perreplication. A strict consensus tree wasestimated from all of the resulting trees,conserving the most robust groupings ofquadrats because the influence of widelydistributed species was minimized (Morrone,1994). In the strict consensus cladogram, weconsidered quadrats or sets of quadratssupported by two or more endemic species(synapomorphies in the cladogram) to beareas of endemism (Morrone, 1994). Next,we mapped the areas of endemism of anuransas described above with the use of ARCVIEWsoftware version 3.2a, having as limits thequadrats in which the endemic species weredistributed. We then superimposed the areasof endemism over the priority areas deter-mined by the MMA (2007) to assess thedegree of coincidence between them. Wetested the null hypothesis of no associationbetween species richness and number ofendemic species in areas of endemism withthe use of Pearson’s correlation analysis (r) onlog-transformed data (Sokal and Rohlf, 1981).We tested the assumptions of normality andhomoscedasticity with the Shapiro–Wilks testand the Levene’s test, respectively, and foundthat they were not violated. We set the a levelof significance to P # 0.05. We carried out thecomputations with the software PAST (Pale-ontological Analysis), version 2.12. Speciesused in the analyses are listed in Appendix 1and quadrats are mapped in Fig. 1.

Determination of Priority Areas

We determined priority areas for amphib-ian conservation in the Caatinga biome andsemiarid region by searching for areas thatcombined endemic species, high to very highanuran species richness, threatened species,and species with restricted geographicaldistributions. We ranked areas from mediumto extremely high importance based on fourcumulative, equally weighted parameters: (1)presence of endemic amphibians, (2) highor very high species richness, (3) presenceof endangered species (sensu International

Union for Conservation of Nature [IUCN],2001), and (4) presence of species known onlyfrom a single locality. By summing parameterswe ranked areas in the following order ofimportance (parameter counts in parenthe-ses): high (1), very high (2–3), and extremelyhigh (4). To break ties within each category ofimportance we conducted a ComplementarityAnalysis (Humphries et al., 1991) consideringonly the previously identified priority areas.This method classifies areas in priority order,considering the most possible biological di-versity and assigning greater importance tothose of higher species richness. The comple-mentarity principle is based on the identifica-tion of a first priority area, which is the areawith the highest number of species. Theremaining areas are ordered according totheir contribution of additional species notfound in the areas of higher priority. Theresidual complement was used to set the firstpriority area, and was computed as thedifference between the total number ofspecies entered in the analysis and thenumber of species present in an area, so thatthe biological diversity of an area is inverselyproportional to its residual complement. Thecumulative percentage, determined by thesum of the percentages of additional species toeach priority area, was used to set prioritiesbetween areas (Mondragon and Morrone,2004).

RESULTS

For the semiarid region, we obtained 255equally parsimonious trees of 383 steps.We obtained six areas of endemism (Fig. 2;Table 1), with quadrats indicated in parenthe-ses: Maranguape (Q1), Jiboia–Timbo (Q34,Q37), Chapada Diamantina (Q36, Q39), LowerJequitinhonha (Q46), Middle Jequitinhonha(Q47), and Upper Jequitinhonha (Q49). Elev-en quadrats (23%) had high (21–30 species) tovery high (31–40 species) species richness(Fig. 3A), from which five quadrats (10%)were also considered to be areas of endemism(high richness Q1, Q39; very high richnessQ34, Q36–37).

For the Caatinga biome, we obtained asingle most-parsimonious tree of 249 steps(Fig. 4). Four areas of endemism resulted fromthis analysis (Fig. 4; Table 1): Maranguape


(Q1), Chapada Diamantina (Q36, Q39), North-ern Minas Gerais (Q43), and Caruaru (Q14).Six quadrats (17%) had high to very highspecies richness (Fig. 3B), of which four (11%)were also considered to be areas of endemism(high richness Q1, Q39; very high richnessQ14, Q36).

In the Caatinga biome and semiarid region,the number of endemic species and speciesrichness were positively correlated (r 5 0.86,n 5 8, P , 0.006), with 60% of the areas ofendemism for amphibians having very high(31–40 species: Q14, Q34, Q36–37; 40%) orhigh species richness (21–30 species: Q1, Q39,20%), and 40% having medium speciesrichness (11–20 species: Q43, Q46–47, Q49).

All areas of endemism for amphibians in thestudied areas were located within one or moreof the priority areas for biodiversity conserva-tion defined by the MMA (2007). However,the degree of protection offered is nothomogeneous (Fig. 5; Table 2). Out of eightareas of endemism, one (12%) has noconservation unit of integral protection(CUIP; see Sistema Nacional de Unidadesde Conservacao da Natureza [SNUC], 2000for definition of the Brazilian categories ofconservation units), four have one CUIP(50%), two have two CUIPs (25%), and one(12%) has four CUIPs (Fig. 5).

The areas of endemism (AE) that presentedthe highest number of CUIPs and plannedactions for the creation of protected areas(MMA, 2007; Fig. 5) were (in descendingorder) Northern Minas Gerais, ChapadaDiamantina, Middle Jequitinhonha, Maran-guape, and Lower Jequitinhonha. The follow-ing areas of endemism allocate threatened (T),near-threatened (NT), or not evaluated (NE)anuran species (sensu IUCN, 2001) restrictedto a specific locality (R) or endemic (E) to theCaatinga biome (CAA) or semiarid region (SAR):Lower Jequitinhonha—Aplastodiscus cavicola(NT) and Xenohyla eugenioi (E/CAA, E/SAR);Maranguape—Adelophryne maranguapensis(T, E/CAA, E/SAR); Middle Jequitinhonha—Aplastodiscus weygoldti (NT), and Pseudisfusca (E/SAR); and Chapada Diamantina—Bokermannohyla juiju (R, E/CAA, E/SAR),Bokermannohyla itapoty (E/CAA, E/SAR),Rupirana cardosoi (NT, E/CAA, E/SAR), andStrabomantis aramunha (E/CAA, E/SAR). TheNorthern Minas Gerais AE did not containanuran species with very limited distributionsthat were endemic to a biome or under somedegree of threat.

For the Caruaru and Upper Jequitinhonhaareas of endemism, there are no currentproposals for creating conservation units

FIG. 2.—(A) Distribution map of the operationalgeographic units (quadrats) identified as areas of endemismfor amphibians in the semiarid region (shaded area). Thedarkened lines delimit the areas of endemism. (B)Informative portion of the strict consensus cladogramobtained from the parsimony analysis of endemicitymethod with the use of 1u 3 1u latitude vs. longitude gridsfor amphibians of the semiarid region. Internal numbersindicate the total number of endemic species supportingareas of endemism (see Table 1 for species list).


(MMA, 2007), and each one contains only asingle CUIP. Moreover, they have no anuranspecies that have very limited distributions,are endemic to a biome, or are under somedegree of threat. The Jiboia–Timbo AE doesnot have any existing CUIPs and only has aproposal for the creation of a conservationunit of undefined category of protection in theSerra da Jiboia (MMA, 2007). The Jiboia–Timbo AE houses a threatened species(subcategory vulnerable, IUCN, 2001), Allo-bates olfersioides (Verdade, 2008), and is thetype locality of Gastrotheca flamma, a treefrog restricted to the Serra da Jiboia that isendemic to the semiarid region and has notyet been assessed by the IUCN (2001). Thenearby Serra do Timbo, not indicated for the

creation of any conservation unit by MMA(2007), is the type locality of Bokermannohylacapra, which has not been assessed by theIUCN (2001), and Phasmahyla timbo, whichhas been evaluated as data deficient. Phasma-hyla timbo is endemic to the semiarid regionand limited to a forest remnant regionallyknown as Mata de Santa Rita in the Serra doTimbo (Cruz et al., 2008).

All areas of high or very high speciesrichness that were not identified as areas ofendemism for anurans in either the Caatingabiome or semiarid region (Q12, Q30, Q38,and Q41) overlap, at least in part, with currentprotected areas. For areas located withinquadrats Q12, Q38, and Q41, the MMA(2007) set the creation of conservation units

TABLE 1.—Anuran species that support areas of endemism of amphibians in the semiarid region (Fig. 2) and Caatingabiome (Fig. 4). OGU 5 operational geographic unit.

OGU Study area Species

1 Semiarid region/Caatinga biome Adelophryne maranguapensisLeptodactylus pustulatus

14 Caatinga biome Agalychnis granulosaIschnocnema vinhaiLithobates palmipesPhyllodytes luteolus

34 Semiarid region Allobates olfersioidesAplastodiscus sibilatusGastrotheca flammaPhyllodytes melanomystaxPhyllodytes wuchereri

37 Semiarid region Aplastodiscus ibirapitangaGastrotheca fissipesPhasmahyla timboScinax strigilatus

(34, 37) Semiarid region Adelophryne pachydactylaBokermannohyla capraFrostius pernambucencisHypsiboas exastis

36 Caatinga biome Bokermannohyla juijuVitreorana eurygnatha

(36, 39) Semiarid region/Caatinga biome Bokermannohyla itapotyRhinella rubescens

43 Caatinga biome Elachistocleis ovalisPseudis bolbodactylaScinax camposseabrai

46 Semiarid region Aplastodiscus cavicolaScinax argyreornatus

47 Semiarid region Aplastodiscus weygoldtiCrossodactylus cyclospinusPseudis fuscaRhinella boulengeriSphaenorhynchus prasinus

49 Semiarid region Bokermannohyla saxicolaChiasmocleis albopunctataHypsiboas lundii


of integral protection as a priority action. Thelatter two quadrats contain three anuranspecies categorized as ‘‘data deficient’’ by theIUCN (2001) and which occur in transitionalareas between the Atlantic Forest and theCaatinga biomes: Q41, Frostius erythro-phthalmus (Angulo, 2008); Q38, Sphaenor-hynchus bromelicola and Dendropsophusnovaisi, both endemic to the semiarid region,and the former restricted to the Municipalityof Maracas, State of Bahia (Peixoto andPimenta, 2004; Silvano and Peixoto, 2004a).Seven quadrats not identified here as areas ofendemism for amphibians and with low tomedium species richness contained someanuran species categorized by the IUCN(2001) as ‘‘threatened’’ (subcategory vulnera-ble) or ‘‘data deficient’’: vulnerable—Q2,Adelophryne baturitensis, restricted to theSerra de Baturite, State of Ceara (Silvano and

Borges-Nojosa, 2004); data deficient—Q6,Q8–9, and Q24, Ceratophrys joazeirensis, en-demic to the Caatinga biome (Skuk and Junca,2004); Q21, Dendropsophus dutrai, in transition-al areas between the coastal Atlantic Forest andthe Caatinga in the states of Sergipe and Alagoas(Silvano and Peixoto, 2004b); Q48, Dendropso-phus novaisi, restricted to the semiarid regionwithin transitional areas between the coastalAtlantic Forest and the Caatinga in the munic-ipalities of Maracas (Bahia State) and Pedra Azul(Minas Gerais State; Peixoto and Pimenta, 2004).

FIG. 4.—(A) Distribution map of the quadrats (opera-tional geographic units) identified as areas of endemismfor amphibians in the Caatinga biome (shaded area) ofBrazil. (B) Informative portion of the strict consensuscladogram obtained from the parsimony analysis ofendemicity using 1u 3 1u latitude vs. longitude grids foramphibians of the Caatinga biome. Internal numbersindicate the total number of endemic species supportingareas of endemism (see Table 1 for species list).

FIG. 3.—Frequency histograms of anuran speciesrichness in the quadrats (operational geographic units):(A) semiarid region and (B) Caatinga biome of Brazil.Classes along the horizontal axis were arbitrarily catego-rized as follows: low species richness, 8–10 species;medium, 11–20; high, 21–30; very high, 31–40. SeeFig. 1 for the geographic distribution map of relatedquadrats (Q).


Out of the six species mentioned above, five(C. joazeirensis, D. dutrai, D. novaisi, F.erythrophthalmus, and S. bromelicola) occur inquadrats with existing CUIPs, and one (A.baturitensis) occurs in areas without CUIP, butwithin areas in which the MMA (2007) has madethe creation of conservation units (a CUIP and anundefined conservation unit) a priority.

The combination of areas of endemism,areas of high to very high anuran speciesrichness, areas with threatened species, andareas with species of a very limited geograph-ical distribution led us to indicate the following15 priority areas for the conservation ofamphibians in the Caatinga biome and semi-arid region (Fig. 1): Serra de Maranguape(Q1), Serra de Baturite (Q2), Serra do Araripe(Q12), Serra or Brejo dos Cavalos (Q14),Agreste Alagoano/Pernambucano (Q21), Jua-zeiro (Q24), Morro do Chapeu (Q30), ChapadaDiamantina (Q36, Q39), Jiboia–Timbo (Serrada Jiboia, Q34; Serra do Timbo, Q37), Maracas(Q38), Serras do Sudoeste da Bahia (Q41),Northern Minas Gerais (Q43), Lower Jequi-tinhonha (Q46), Middle Jequitinhonha (Q47),and Upper Jequitinhonha (Q49). Out of 15priority areas, 2 did not match any of the fourparameters (Juazeiro [Q24] and Agreste Ala-goano/Pernambucano [Q21]), but were main-tained as priority areas for amphibian conser-vation of medium importance because of the

presence of species with very limited geo-graphical distributions.

All categories of importance presented atleast two priority areas (Table 3). As a tie-breaking method, we conducted a comple-mentarity analysis with the use of the cumu-lative percentage of 103 species found withinthe 15 priority areas. As seen in Table 3, thehighest priority area in the first category ofimportance was Jiboia–Timbo, because itpresented the lowest residual complement(50 species). The overall stratification ofcategories of importance was maintained bythe use of the complementarity principle(Table 4). The changes were the positioningof Maranguape (extremely high importance)and Serra de Baturite (very high importance)among areas of high importance, and North-ern Minas Gerais (high importance) amongareas of very high importance.

DISCUSSION

Muller (1973), Kinzey (1982), and Costaet al. (2000) investigated biogeographicalpatterns of terrestrial vertebrates in the Neo-tropics (birds, primates, and mammals, respec-tively) and identified centers and subcenters offaunal dispersal. The Bahia subcenter of faunaldispersal comprises two areas of amphibianendemism identified here as Chapada Dia-mantina (in part of the SAR and CAA) and

FIG. 5.—Distribution of conservation units proposed by the Brazilian Ministry of the Environment (MMA, 2007) forthe semiarid region and Caatinga biome that have already been installed. CU, conservation unit; CUIP, conservationunit of integral protection; CUSU, conservation unit of sustainable use.


TA

BL

E2.

—P

hys

ical

feat

ure

s,ve

geta

tion

char

acte

rist

ics,

hig

hlig

hts

,an

dcu

rren

tst

atu

sof

pro

tect

ion

ofar

eas

ofen

dem

ism

for

amp

hib

ian

s(A

E)

wit

hin

the

pol

ygon

sof

the

Caa

tin

gab

iom

ean

dse

mia

rid

regi

on(F

igs.

2an

d4)

.T

he

pri

orit

yar

eas

for

the

Bra

zilia

nb

iod

iver

sity

con

serv

atio

nw

ere

thos

ed

efin

edb

yth

eB

razi

lian

Min

istr

yof

the

En

viro

nm

ent

(MM

A,2

007)

.Pol

ygon

s(P

O):

AF

,Atl

anti

cF

ores

tb

iom

e;C

AA

,Caa

tin

gab

iom

e;C

ER

,Cer

rad

ob

iom

e;S

AR

,sem

iari

dre

gion

.Oth

erab

bre

viat

ion

s(a

cron

yms

inP

ortu

gues

e):A

PA

,en

viro

nm

enta

lpro

tect

ion

area

;AR

IE,a

rea

ofre

leva

nt

ecol

ogic

alin

tere

st;P

AR

NA

,nat

ion

alp

ark;

PE

,sta

tep

ark;

PM

,mu

nic

ipal

par

k;Q

,qu

adra

t;R

EB

IO,

bio

logi

cal

rese

rve;

RV

S,

wild

life

refu

gear

ea;

UC

,co

nse

rvat

ion

un

it;

UC

PI,

con

serv

atio

nu

nit

ofin

tegr

alp

rote

ctio

n;

UC

US

,co

nse

rvat

ion

un

itof

sust

ain

able

use

.

PO

AE

Ele

vati

onM

ain

saw

sV

eget

atio

nty

pe

Clim

ate

typ

eP

rior

ity

area

s

CA

AS

AR

Mar

angu

ape

(Q1)

Mou

nta

inou

s(c

a.89

0m

)Se

rra

deM

aran

guap

eA

tlant

icF

ores

tre

mna

ntS

ub

hu

mid

trop

ical

UC

PI—

Bot

anic

alP

ark

ofC

eara

(190

ha)

.S

tate

UC

US

(AP

A)—

Lag

amar

do

Cau

ıpe

(188

4h

a),

Est

uar

iod

oR

ioC

eara

(274

4h

a),

and

Ser

rad

aA

rata

nh

a(6

448

ha)

.M

un

icip

alU

CU

S—

AP

AM

aran

guap

eC

AA

Car

uar

u(Q

14)

Mou

ntai

nous

(ca.

650–

1000

m)

Serr

ado

Cav

alo

Sem

idec

iduo

usan

dde

cidu

ous

fore

sts

(Bre

jos

deA

ltitu

de)

Su

bh

um

idtr

opic

alE

xtre

mel

yh

igh

imp

orta

nce

—C

abec

eira

sd

oC

apib

arib

e**

and

Bre

jod

osC

aval

os.

Pro

tect

edar

ea:M

un

icip

alU

C—

Joao

Vas

con

celo

sS

obri

nh

oE

colo

gica

lP

ark

(359

ha)

SA

RA

FJi

boi

a–T

imb

o(Q

34,

Q37

)M

ount

aino

us(c

a.23

0–98

0m

)Se

rra

daJi

boia

and

Serr

ado

Tim

bo

Atla

ntic

For

est

rem

nant

sT

rop

ical

sub

hu

mid

tose

mia

rid

Ver

yh

igh

imp

orta

nce

—M

ilagr

es,S

erra

da

Pio

nei

ra/

Ser

rad

aJi

boi

a,an

dV

ale

do

Jeq

uir

ica;

extr

emel

yh

igh

imp

orta

nce

—S

erra

da

Jib

oia.

**

Pro

tect

edar

eas:

non

eC

AA

SA

RC

hap

ada

Dia

man

tin

a(Q

36,

Q39

)

Mou

ntai

nous

(ca.

200–

2000

m)

Serr

ado

Sinc

ora,

Pic

odo

Bar

bado

and

Pic

oda

sA

lmas

Veg

etat

ion

mos

aic—

sava

nna,

cam

pos

rupe

stre

s(r

ocky

field

s)an

dde

cidu

ous

fore

sts

Tro

pic

alsu

bh

um

idto

sem

iari

d

Hig

him

por

tan

ce—

Par

amir

im;

extr

emel

yh

igh

imp

orta

nce

—M

arim

bu

s/Ir

aqu

ara,

**

PA

RN

AC

hap

ada

Dia

man

tin

a(U

CP

I,15

2,00

0h

a),

Ser

rad

oB

arb

ado,

*S

tate

AR

IEN

asce

nte

sd

oR

iod

eC

onta

s(U

CU

S,

4771

ha)

,an

dm

un

icip

alit

ies

ofA

nd

araı

and

Mu

cuge

.Ad

dit

ion

alp

rote

cted

area

s:S

tate

UC

US

—A

PA

Mar

imb

us/

Iraq

uar

a(1

25,4

00h

a),

and

AP

AS

erra

do

Bar

bad

o(6

3,65

2h

a).

Mu

nic

ipal

UC

PI—

PM

Mu

cuge

and

PM

Ser

rad

asA

lmas

.C

AA

Nor

ther

nM

inas

Ger

ais

(Q43

)F

lat

toge

ntly

undu

latin

g(c

a.45

0–75

0m

)

Ab

sen

tV

eget

atio

nm

osai

c—sa

vann

a,de

cidu

ous

fore

sts

and

trop

ical

rain

fore

st

Tro

pic

alse

mia

rid

Fu

llyin

sert

edin

pri

orit

yar

eas

ofex

trem

ely

hig

him

por

tan

ce—

Alt

o–M

edio

Sao

Fra

nci

sco

Riv

er,

Jap

ore

Riv

erC

orri

dor

,m

un

icip

alit

yof

Mon

talv

ania

,Ja

ıba

Reg

ion

,P

EL

agoa

do

Caj

uei

ro(U

CP

I20

,500

ha)

,P

EM

ata

Sec

a(U

CP

I10

,281

ha)

,P

EV

erd

eG

ran

de

(25,

570

ha)

,R

EB

IOJa

ıba

(UC

PI

6358

ha)

.A

dja

cen

tU

CP

I:P

AR

NA

Cav

ern

asd

oP

eru

acu

.S

AR

AF

Low

erJe

qu

itin

hon

ha

(Q46

)

Mou

nta

inou

s(c

a.70

0–90

0m

)Se

rra

dos

Bar

bado

sA

tlant

icF

ores

tre

mna

nts

Dry

sub

hu

mid

trop

ical

Ver

yh

igh

imp

orta

nce

—V

arze

asd

oJe

qu

itin

hon

ha,

*ex

trem

ely

hig

him

por

tan

ce—

Ser

rad

osB

arb

ados

.*P

rote

cted

area

s:S

tate

UC

PI:

PE

Alt

oC

arir

i(6

151

ha)


Jiboia–Timbo (SAR; Figs. 2, 4; Q36 + Q39,Q34 +37, respectively); the Pernambuco centerof faunal dispersal (Muller, 1973; Costa et al.,2000) comprises the Caruaru area of endemism(in the CAA; Q14 in Fig. 4).

Cracraft (1985) and Goldani et al. (2006)proposed areas of endemism for the SouthAmerican avifauna and for the NeotropicalPlatyrrhini (Primates), respectively, the formerby the method of species–distribution congru-ence (Muller, 1973), and the latter by theParsimony Analysis of Endemicity. Goldaniet al. (2006) used different operational geo-graphic units (OGUs: quadrats, river divisions,and areas determined previously by Cracraft,1985). Cracraft (1985) and Goldani et al. (2006)identified a single large area of endemismencompassing the Caatinga biome and thesemiarid region. However, Cracraft (1985)mentioned that, in the future, some of thepostulated areas would likely be divided intotwo or more areas of endemism, which is inagreement with our results (four areas ofamphibian endemism for the Caatinga biomeand six for the semiarid region). Goldani et al.(2006) showed a strong relationship in thecomposition of platyrrhine primates amongareas of endemism of the Caatinga and AtlanticForest biomes. These findings coincide withour results because the areas of endemism ofamphibians for the Caatinga biome andsemiarid region contained 10 (66.7%) and 26(76.5%) species in common with the AtlanticForest biome, respectively.

In a study of the distribution patterns ofamphibians in 12 natural regions of SouthAmerica, Duellman (1999) found a negativeassociation between species richness and thenumber of endemic species, which, consider-ing the different scales (continental vs. re-gional), does not agree with our results. In theCaatinga biome and semiarid region, thenumber of endemic species and speciesrichness were positively correlated. Positiveassociations between species richness and thenumber of endemic species may indicate theexistence of centers of speciation (Gonsales,2008), which in turn make these areas ofinterest for studying the biogeography of theCaatinga biome and the semiarid region.Furthermore, the combination of high to veryhigh species richness and endemic species in

PO

AE

Ele

vati

onM

ain

saw

sV

eget

atio

nty

pe

Clim

ate

typ

eP

rior

ity

area

s

SA

RA

FM

idd

leJe

qu

itin

hon

ha

(Q47

)

Fla

tten

edlo

wla

nds

tom

ount

aino

us(c

a.80

0m

)

Ab

sen

tV

eget

atio

nm

osai

c—sa

vann

a,C

ampo

sR

upes

tres

(roc

kyfie

lds)

and

Atla

ntic

fore

stre

mna

nts

Tro

pic

alse

mia

rid

Insu

ffic

ient

lykn

own—

Cab

ecei

rado

Rio

doP

eixe

;hig

him

port

ance

—L

imoe

iro

farm

,m

unic

ipal

ityof

Joaı

ma;

*ve

ryhi

ghim

port

ance

—Je

quiti

nhon

ha/

Mat

aE

scur

aC

orri

dor,

Var

zeas

doJe

quiti

nhon

ha*

and

surr

ound

ings

ofR

EB

IOM

ata

Esc

ura;

extr

emel

yhi

gh—

mun

icip

ality

ofB

ande

ira*

*an

dR

EB

IOM

ata

Esc

ura

(UC

PI)

.Add

itio

nal

prot

ecte

dar

eas:

Stat

eU

CP

I:P

EA

ltoC

arir

i(6

151

ha),

and

RV

SM

ata

dos

Mur

iqui

s(2

722

ha).

SA

RC

ER

Up

per

Jeq

uit

inh

onh

a(Q

49)

Mou

ntai

nous

inte

rsec

ted

bypl

ains

and

low

land

s(c

a.50

0–12

00m

)

Ser

raG

eral

Veg

etat

ion

ofhi

ghla

ndfie

lds,

sava

nna

and

shru

bsa

vann

a

Tro

pic

alh

um

idto

dry

sub

hu

mid

Hig

him

port

ance

—m

unic

ipal

ities

ofG

rao

Mog

olan

dF

ranc

isco

Sa;e

xtre

mel

yhi

ghim

port

ance

—Je

quiti

nhon

haR

iver

Val

ley.

Pro

tect

edar

eas:

Stat

eU

CP

I—P

EG

rao

Mog

ol(3

3,32

4ha

)

*In

dic

atio

nto

crea

tion

ofco

nse

rvat

ion

un

it(M

MA

,20

07).

**

Ind

icat

ion

tocr

eati

onof

UC

PI

(MM

A,

2007

).P

riva

tere

serv

esof

nat

ura

lh

erit

age

(RP

PN

)w

ere

not

con

sid

ered

.

TA

BL

E2.

—C

onti

nu

ed.


these areas points to the possible presence ofrare or endangered species not yet known toscience (Brown and Lomolino, 1998).

Silva et al. (2004) identified four areas ofendemism for Passeriformes (Aves) in theAtlantic Forest, three of which include areasof endemism for amphibians identified here(the area of endemism for amphibians and itsrespective quadrats [Q] in the CAA or SAR aregiven in parentheses; Figs. 2 and 4): Costa daBahia (Jiboia–Timbo, Q34, Q37, SAR), Per-nambuco (Caruaru, Q14, CAA), and BahiaCentral (Chapada Diamantina, Q36, Q39,CAA, SAR). Silva and Casteleti (2005) contrib-uted more results to those of Silva et al. (2004)by simultaneously considering butterflies, pri-mates, and birds; they identified eight biogeo-graphic subregions in the Atlantic Forest (fiveareas of endemism and three of transition).These areas of endemism contain six of theareas of amphibian endemism identified here(Figs. 2 and 4): Bahia (Jiboia–Timbo, Q34,Q37, SAR), Brejos Nordestinos (Caruaru, Q14,CAA, SAR; Maranguape, Q1, CAA), andDiamantina (Chapada Diamantina, Q36, Q39,CAA, SAR; Lower Jequitinhonha, Q46, SAR;Middle Jequitinhonha, Q47, SAR). The transi-tion area, Sao Francisco, contains the NorthernMinas Gerais area of endemism (Q43, CAA).

In this study, the areas of endemism identi-fied for amphibians in the Caatinga biome andsemiarid region are found in mountainousregions that may exceed 2000 masl, have highhumidity, and are covered by isolated forests(Table 2). These mountain massifs includeunique ecosystems, such as Campos Rupestres(rocky fields) and seasonal forests. Such envi-ronments provide barriers to the dispersal ofvarious phylogenetic groups and are recognizedas having high degrees of endemism (Giarettaand Aguiar, 1998; Haddad and Abe, 1999;Naniwadekar and Vasudevan, 2007), favoringthe occurrence of geographically restrictedamphibian species (Duellman, 1999; Junca,2005).

The areas of endemism of amphibians alongthe Jequitinhonha River Valley are in atransition region (a mosaic of vegetation types:cerrado, campo cerrado, caatinga, high-altituderocky fields, and Atlantic Forest fragments)between the banks of the Sao Francisco andJequitinhonha rivers. Studies on the areascovered by these rivers (Silva and Straube,1996; Pellegrino et al., 2005) have suggestedthat they may have acted as barriers to geneflow in lizard and bird populations and couldalso explain amphibian endemism in theregion. Moreover, Feio and Caramaschi

TABLE 3.—Priority areas for amphibian conservation at the Caatinga biome and semiarid region. Categories ofimportance were defined upon four parameters of same weight (0, absence; 1, presence): EH 5 extremely high (S 5 4);VH, very high (S 5 3 or 2); H 5 high (S 5 1); M 5 medium (S 5 0). Areas of high and very high species richness: 21–40 species per quadrat (Fig. 3). Threatened species according to IUCN (2001). The priority order within each categoryof importance was based on cumulative percentage obtained from the complementarity analysis (Table 4), and theresidual complement is the difference between the total number of species entered in the analysis (n 5 103) and the

number of species present in an area.

Priorityorder Priority areas Quadrats

Areas ofendemism

Restrictedspecies

Threatenedspecies

High or veryhigh species

richness SumDegree ofimportance

Numberof species

Residualcomplement

1 Jiboia–Timbo 34, 37 1 1 1 1 4 EH 53 502 Maranguape 1 1 1 1 1 4 EH 23 803 Chapada Diamantina 36, 39 1 1 0 1 3 VH 34 694 Caruaru 14 1 0 0 1 2 VH 29 745 Middle Jequitinhonha 47 1 0 1 0 2 VH 16 876 Maracas 38 0 1 0 1 2 VH 33 707 Serra de Baturite 2 0 1 1 0 2 VH 19 848 Northern Minas Gerais 43 1 0 0 0 1 H 20 839 Upper Jequitinhonha 49 1 0 0 0 1 H 17 86

10 Lower Jequitinhonha 46 1 0 0 0 1 H 16 8711 Serras do Sudoeste baiano 41 0 0 0 1 1 H 25 7812 Serra do Araripe 12 0 0 0 1 1 H 24 7913 Morro do Chapeu 30 0 0 0 1 1 H 23 8014 Juazeiro 24 0 0 0 0 0 M 14 8915 Agreste Alagoano/

Pernambucano21 0 0 0 0 0 M 13 90


(2002) pointed to the Jequitinhonha River as apossible disperser of amphibians and reptilesfrom the coast to the interior of Brazil. In fact,the species that are unique to the area ofendemism in the Jequitinhonha River Valley(Fig. 2; Q47) are typical of the coastal AtlanticForest.

The priority areas for biodiversity conser-vation proposed by the MMA (2007) coincid-ed fully with the areas of endemism detectedby the method used here, which is importantfor amphibian conservation in the Caatingabiome and semiarid region. However, thereare some aspects that deserve attentionbecause the degree of importance receivedby each area was not equal.

A more worrying aspect is the presence ofanuran species restricted to areas not protect-ed by CUIPs, such as the Serra da Jiboia(Q34), Serra do Timbo (Q37), and Serra deBaturite (Q2). This fact indicates that there isinsufficient knowledge about species distribu-tions in these regions or that the methodspreviously used to select priority areas fortheir conservation may have been ineffective.The failure to include these areas of ende-mism is serious, as species with restricteddistributions are more susceptible to extinc-tion (Brown and Lomolino, 1998). We em-phasize here the importance of the immediatecreation of at least three conservation units ofintegral protection: two CUIPs in the AEJiboia–Timbo, one in the Serra da Jiboia and

another in the Serra do Timbo, the lattercomprising the Mata de Santa Rita remnant(extremely high importance), and at least oneCUIP in the Serra do Baturite (very highimportance). In addition, we also encouragethe creation of Private Reserves of NaturalHeritage (Reserva Particular do PatrimonioNatural—RPPN) in these three priority areas.

The municipality of Maracas in the State ofBahia deserves special attention because itonly has one conservation unit of integralprotection (Maracas Biological Reserve), andit is very small (ca. 10 ha). This municipality isthe type locality for 11 species of frogs: D.novaisi, D. oliveirai, Phyllodytes tuberculosus,Phyllomedusa nordestina, Physalaemus cica-da, Rhinella jimi, Scinax camposseabrai,Scinax eurydice, S. bromelicola, Trachycepha-lus atlas, and X. eugenioi. The creation ofmore conservation units of integral protectionin Maracas and expansion of the area cur-rently occupied by the Maracas BiologicalReserve are necessary and warrant furtherstudy.

The Caatinga biome and the Braziliansemiarid region are partially different withrespect to the priority areas for amphibianconservation. The geographical boundaries ofthese regions are not coincident, mainly intheir southern limits, where the boundaries ofthe semiarid region include transitional areasbetween the Atlantic Forest biome and theCaatinga biome. These transitional areas

TABLE 4.—Priority areas for amphibian conservation at the Caatinga biome and semiarid region with the analysis ofcomplementarity of Humphries et al. (1991) used as the unique ordination criterion. Categories of importance are

defined in Table 3.

Priorityorder Priority areas

Degree ofimportance Quadrats

Number ofcomplementary species

Accumulativepercentage

1 Jiboia–Timbo EH 34, 37 53 51.52 Chapada Diamantina VH 36, 39 16 67.03 Caruaru VH 14 7 73.84 Northern Minas Gerais H 43 6 79.65 Middle Jequitinhonha VH 47 6 85.46 Maracas VH 38 3 88.37 Upper Jequitinhonha H 49 3 91.38 Maranguape EH 1 2 93.29 Lower Jequitinhonha H 46 2 95.1

10 Serras do Sudoeste baiano H 41 1 96.111 Serra do Araripe H 12 1 97.112 Serra de Baturite VH 2 1 98.113 Juazeiro M 24 1 99.014 Agreste Alagoano/Pernambucano M 21 1 10015 Morro do Chapeu H 30 0 100


differ from the Caatinga biome in speciescomposition and are possibly more influencedby the Atlantic Forest biome. This hypothesisis supported by the presence of four exclusiveareas of endemism composed of five uniqueOGUs in the southern and southeastern limitsof the semiarid region: Jiboia–Timbo (Q34–Q37), Lower Jequitinhonha (Q46), MiddleJequitinhonha (Q47), and Upper Jequitin-honha (Q49). Moreover, when transitionalareas in the southern limits of the semiaridregion are considered, two areas of endemismidentified for the Caatinga biome (Caruaru[Q14] and Northern Minas Gerais [Q43]) areno longer recognized, because five of theirseven exclusive species also occur in OGUsunique to the semiarid region (Q32, Q34,Q37, Q38, Q46, and Q48).

These conflicting results have implicationsfor the conservation efforts for the Caatingabiome and the Brazilian semiarid region.Conservation initiatives for these areas reflectdistinct research and conservation policiesadopted by the Brazilian government. Forexample, the Caatinga Thematic Group at theBrazilian Ministry of Environment (MMA)focuses on identifying priority areas for theCaatinga biome. In contrast, the BiodiversityResearch Program, which is part of theBrazilian Ministry of Science, Technology,and Innovation (MCTI), conducts researchand searches for high-biodiversity areas in thewhole semiarid region. Our results show thatassessing priority areas within each regionalone is not adequate. Rather, combiningresults into a single proposal of priority areasis paramount to safeguard the biodiversity ofboth the semiarid region and Caatinga biomein Brazil.

Finally, it is imperative to make greaterinvestments in faunal inventories, studies ofcommunity and population ecology, and long-term monitoring of the Caatinga biome andsemiarid region to understand the distributionpatterns and areas of amphibian endemismwithin these regions better. There are largegeographical gaps in the sampling of amphib-ians. The presence of patches of relictualforest, rocky fields, high-altitude fields, andmountain massifs that have not yet been fullysurveyed, suggest that there are more geo-graphically restricted amphibian species and,

consequently, areas of endemism than thosefound in this study.

Acknowledgments.—We thank Ulisses Caramaschi,Jose P. Pombal, Jr., Hussam Zaher, Flora A. Junca, andRicardo Rosa for providing space and facilities during thedevelopment of this article and allowing us to examinespecimens under their care; Flora A. Junca and ElizaMaria X. Freire for critically reading and improving thePortuguese version of the manuscript, and Camila A.Souto and Adrian A. Garda for the English version;Angela Zanata and Priscila Camelier for helpful commentsduring the development of the research; Ivan SergioNunes, Gindomar G. Santana, Washington L. S. Vieira,Carolina C. Mello, and Tasso M. Lima for assistanceduring works in the herpetological collections; and theConselho Nacional de Desenvolvimento Cientıfico eTecnologico (CNPq, process 302542/2008–6) for afellowship to Marcelo F. Napoli.

LITERATURE CITED

Ab’Saber, A.N. 1977. Os domınios morfoclimaticos naAmerica do Sul. Primeira aproximacao. Geomorfologia52:1–21.

Angulo, A. 2008. Frostius erythrophthalmus. In IUCN(2009). IUCN red list of threatened species. Version2009.1. Available at http://www.iucnredlist.org. Cited10 November 2009.

Brown, J.H., and M.V. Lomolino. 1998. Biogeography,2nd ed. Sinauer Associates, Inc., Massachusetts, USA.

Carnaval, A.C. 2002. Phylogeography of four frog speciesin forest fragments of northeastern Brazil: a preliminarystudy. Integrative and Comparative Biology 42:913–921.

Carnaval, A.C., and C. Moritz. 2008. Historical climatemodelling predicts patterns of current biodiversity inthe Brazilian Atlantic forest. Journal of Biogeography35:1187–1201.

CHESF (Companhia Hidroeletrica do Sao Francisco).2001. Aptidao Bioclimatica no Bioma Caatinga, PRO-BIO 2001. Available at http://mapas.mma.gov.br/i3geo/datadownload.htm.

Costa, L.P., Y.L. Leite, G.A.B. Fonseca, and M.T.Fonseca. 2000. Biogeography of South American forestmammals: Endemism and diversity in the AtlanticForest. Biotropica 32:872–881.

Cracraft, J. 1985. Historical biogeography and patterns ofdifferentiation within the South America avifauna. Areasof endemism. Ornithological Monographs 36:49–84.

Cruz, C.A.G., M.F. Napoli, and P.M. Fonseca. 2008. Anew species of Phasmahyla Cruz, 1990 (Anura:Hylidae) from the State of Bahia, Brazil. SouthAmerican Journal of Herpetology 3:187–195.

Duellman, W.E. 1999. Distribution patterns of amphib-ians in South America. Pp. 255–328 in W.E. Duellman(Ed.), Patterns of Distribution of Amphibians: A GlobalPerspective. The Johns Hopkins University Press,Maryland, USA.

Duellman, W.E., and L. Trueb. 1986. Biology ofamphibians. McGraw-Hill, New York, USA.

ESRI (Environmental Systems Research Institute). 1997.DCW: Digital Chart of the World. Available at http://www.maproom.psu.edu/dcw.


ESRI (Environmental System Research Institute). 2000.ArcView GIS Version 3.2a. New York, USA.

Feio, R.N., and U. Caramaschi. 2002. Contribuicao aoconhecimento da herpetofauna do nordeste do estadode Minas Gerais, Brasil. Phyllomedusa 1:105–111.

Frost, D.R. 2009. Amphibian species of the world: An onlinereference. Version 5.3 (12 February, 2009). Availableat http://research.amnh.org/herpetology/amphibia. Amer-ican Museum of Natural History, New York, USA.

GEOMINAS (Geoprocessamento em Minas Gerais).2009. Atlas de zoneamento agroclimatico do Estadode Minas Gerais—SEA (1980). Available at http://www.geominas.mg.gov.br.

Giaretta, A.A., and O. Aguiar, Jr. 1998. A new species ofMegaelosia from the Mantiqueira Range, SoutheasternBrazil. Journal of Herpetology 32:80–83.

Goldani, A., G.S. Carvalho, and J.C. Bicca-Marques. 2006.Distribution patterns of Neotropical primates (Platyr-rhini) based on parsimony analysis of endemicity.Brazilian Journal of Biology 66:61–74.

Goloboff, P., J. Farris, and K. Nixon. 2008. TNT: A freeprogram for phylogenetic analysis. Cladistics 24:774–786.

Gonsales, E.M.L. 2008. Diversidade e Conservacao deAnfıbios Anuros no Estado de Santa Catarina, sul doBrasil. Ph.D. Dissertation, Instituto de Biociencias daUniversidade de Sao Paulo, Brazil.

Haddad, C.F.B., and A. Abe. 1999. Repteis e anfıbios.Pp. 19–21. In Workshop Floresta Atlantica e CamposSulinos. Available at http://www.conservation.org.br/publicacoes/files/Sumario.pdf.

Humphries, C.J., R.I. Vane-Wright, and P.H. Williams.1991. Biodiversity reserves: Setting new priorities forthe conservation of wildlife. Park 2:34–38.

IBGE (Instituto Brasileiro de Geografia e Estatıstica).2004a. Biomas do Brasil (1:5.000.000). IBGE–Departa-mento de Cartografia (DECAR), Rio de Janeiro, Brazil.Available at http://mapas.mma.gov.br/i3geo/datadownload.htm.

IBGE (Instituto Brasileiro de Geografia e estatıstica).2004b. Mapa de biomas do Brasil: Primeira aproxima-cao. Ministerio do Meio Ambiente e IBGE, Diretoriade Geociencias, Coordenacao de Recursos Naturais eEstudos Ambientais, Brazil.

IUCN (International Union for Conservation of Nature).2001. The IUCN red list of threatened species.Available at http://www.iucnredlist.org/.

Junca, F.A. 2005. Anfıbios e repteis. Pp. 337–356 in F.A. Junca,L. Funch, and W. Rocha(Eds.), Biodiversidade e Conservacaoda Chapada Diamantina. Ministerio do Meio Ambiente,Distrito Federal, Brazil.

Kinzey, W.G. 1982. Distribution of primates and forestrefuges. Pp. 456–482 in G.T. Prance (Ed.), BiologicalDiversification in the Tropics. Columbia UniversityPress, New York, USA.

Maddison, W.P., and D.R. Maddison. 2009. Mesquite: Amodular system for evolutionary analysis. Version 2.71.Available at http://mesquiteproject.org.

MIN (Ministerio da Integracao Nacional). 2005. Novadelimitacao do Semi-Arido brasileiro. Available at http://www.mi.gov.br/desenvolvimentoregional/publicacoes/delimitacao.asp. Secretaria de Polıtica e Desenvolvi-mento Regional, Distrito Federal, Brazil.

MMA (Ministerio do Meio Ambiente). 2007. Areasprioritarias para conservacao, uso sustentavel e reparticaode benefıcios da biodiversidade brasileira. Atualizacao—Portaria MMA No. 9, de 23 de janeiro de 2007. Ministeriodo Meio Ambiente, Secretaria de Biodiversidade eFlorestas, serie biodiversidade, Distrito Federal, Brazil.

Mondragon, E.A., and J.J. Morrone. 2004. Propuesta deareas para la conservacion de aves de Mexico,empleando herramientas panbiogeograficas e ındicesde complementariedad. Interciencia 29:112–120.

Morrone, J.J. 1994. On the identification of areas ofendemism. Systematic Biology 43:438–441.

Morrone, J.J. 2006. Biogeographic areas and transitionzones of Latin America and Caribbean islands based onpanbiogeographic and cladistic analyses of the entomo-fauna. Annual Review of Entomology 51:467–494.

Morrone, J.J., and T. Escalante. 2002. Parsimony analysisof endemicity (PAE) of Mexican terrestrial mammals atdifferent area units: When size matters. Journal ofBiogeography 29:1095–1104.

Muller, P. 1973. The dispersal centers of terrestrialvertebrates in the Neotropical realm. A study in theevolution of the Neotropical Biota and its nativelandscape. Biogeographica 2:1–250.

Naniwadekar, R., and K. Vasudevan. 2007. Patterns indiversity of anurans along an elevational gradient in theWestern Ghats, south India. Journal of Biogeography34:843–852.

Peixoto, O.L., and B.V.S. Pimenta. 2004. Dendropsophusnovaisi. In IUCN (International Union for Conserva-tion of Nature), IUCN red list of threatened species,Version 2009.1. Available at http://www.iucnredlist.org.

Pellegrino, K.C.M., M.T. Rodrigues, A. Waite, M.Morando, Y. Yonenagayassuda, and J.W. Sites, Jr.2005. Phylogeography and species limits in theGymnodactylus darwinii complex (Gekkonidae, Squa-mata): Genetic structure coincides with river systems inthe Brazilian Atlantic Forest. Biological Journal ofLinnean Society 85:13–26.

Platnick, N.I. 1991. On areas of endemism. AustralianSystematic Botany 4:1–11.

Rizzini, C.T. 1979. Tratado de Fitogeografia do Brasil.HUCITEC/EDUSP, Brazil.

Rosen, B.R. 1988. From fossils to earth history: Appliedhistorical biogeography. Pp. 437–481 in A.A. Myers andP. Giller (Eds.), Analytical Biogeography: An IntegratedApproach to the Study of Animal and Plant Distribu-tions. Chapman and Hall, UK.

Santos, A.M.M., and M. Tabarelli. 2003. Variaveis multiplase desenho de unidade de conservacao: uma praticaurgente para a Caatinga. Pp. 735–776 in I.R. Leal, M.Tabarelli, and J.M.C. Silva (Eds.), Ecologia e conservacaoda Caatinga. Universidade Federal de Pernambuco,Brazil.

Silva, J.M.C., and C.H.M. Casteleti. 2005. Estado dabiodiversidade da Mata Atlantica brasileira. Pp. 43–59in C. Galindo-Leal and I.G. Camara (Eds.), MataAtlantica: Biodiversidade, Ameacas e Perspectivas.Fundacao SOS Mata Atlantica e Conservacao Inter-nacional, Brazil.

Silva, J.M.C., and F.C. Straube. 1996. Systematics andbiogeography of scaled woodcreepers (Aves: Dendro-colaptidae). Studies Neotropical Fauna and Environ-ment 31:3–10.


Silva, J.M.C., M.C. Sousa, and C.H.M. Castelletti. 2004.Areas of endemism for passerine birds in the AtlanticForest, South America. Global Ecology and Biogeog-raphy 13:85–92

Silva, M.B. 2008. Biogeografia de Opilioes Gonyleptidaena Mata Atlantica, com Revisao Sistematica deHernandariinae (Arachnida, Opiliones). Ph.D. Disser-tation, Instituto de Biociencias da Universidade de SaoPaulo, Brazil.

Silvano, D., and D.M. Borges-Nojosa. 2004. Adelophrynebaturitensis. In IUCN (International Union for Conser-vation of Nature). IUCN red list of threatened species.Version 2009.1. Available at http://www.iucnredlist.org.

Silvano, D., and O.L. Peixoto. 2004a. Sphaenorhynchusbromelicola. In IUCN (International Union for Conser-vation of Nature), IUCN red list of threatened species.Version 2009.1. Available at http://www.iucnredlist.org.

Silvano, D., and O.L. Peixoto. 2004b. Dendropsophusdutrai. In IUCN (International Union for Conservationof Nature), IUCN red list of threatened species.Version 2009.1. Available at http://www.iucnredlist.org.

Skuk, G., and F.A. Junca. 2004. Ceratophrys joazeirensis.In IUCN (International Union for Conservation ofNature), IUCN red list of threatened species. Version2009.1. Available at http://www.iucnredlist.org.

SNUC (Sistema Nacional de Unidades de Conservacao daNatureza). 2000. Lei 9.985, de 18 de julho de 2000.Available at http://www.mma.gov.br. Ministerio doMeio Ambiente/Secretaria de Biodiversidade e Flor-estas, Distrito Federal, Brazil.

Sokal, R.R., and F.J. Rohlf. 1981. Biometry: ThePrinciples and Practice of Statistics in BiologicalResearch. W.H. Freeman and Company, New York,USA.

SPDR (Secretaria de Polıticas de DesenvolvimentoRegional). 2005. Redelimitacao do Semi-Arido Nor-destino. Available at http://mapas.mma.gov.br/i3geo/datadownload.htm.

Tabarelli, M., and J.M.C. Silva. 2004. Areas e acoesprioritarias para a conservacao da biodiversidade daCaatinga. Pp. 777–796 in I.R. Leal, M. Tabarelli, andJ.M.C. Silva (Eds.), Ecologia e Conservacao daCaatinga. Universidade Federal de Pernambuco,Brazil.

USGS (US GEOLOGICAL SURVEY). 1996. GTOPO30: Global30 arc second elevation data set. Available at http://www1.gsi.go.jp/geowww/globalmap-gsi/gtopo30/gtopo30.html.

Velloso, A.L., E.V.S.B. Sampaio, and F.G.C. Pareyn. 2002.Ecorregioes—Proposta Para o Bioma Caatinga. Asso-ciacao Plantas do Nordeste/Instituto de ConservacaoAmbiental, The Nature Conservancy do Brasil,Brazil.

Verdade, V. 2008. Allobates olfersioides. In IUCN(International Union for Conservation of Nature),IUCN Red List of Threatened Species. Version2009.2. Available at http://www.iucnredlist.org.

.Accepted: 3 November 2011

.Associate Editor: Christopher Raxworthy

Aromobatidae

Allobates olfersioides

Brachycephalidae

Ischnocnema bilineataIschnocnema vinhai

Bufonidae

Frostius erythrophthalmusFrostius pernambucensisRhinella boulengeriRhinella cruciferRhinella granulosaRhinella hoogmoediRhinella jimiRhinella rubescens

Ceratophryidae

Ceratophrys auritaCeratophrys joazeirensis

Craugastoridae

Haddadus binotatus

Cycloramphidae

Odontophrynus carvalhoiProceratophrys cristicepsProceratophrys renalisProceratophrys schirchiRupirana cardosoiThoropa miliaris

Eleutherodactylidae

Adelophryne baturitensisAdelophryne maranguapensisAdelophryne pachydactyla

Hemiphractidae

Gastrotheca fissipesGastrotheca flamma

Hylidae

Agalychnis granulosaAplastodiscus cavicolaAplastodiscus ibirapitangaAplastodiscus sibilatusAplastodiscus weygoldtiBokermannohyla capraBokermannohyla diamantinaBokermannohyla itapotyBokermannohyla juijuBokermannohyla oxenteBokermannohyla saxicolaCorythomantis greeningiDendropsophus branneriDendropsophus dutraiDendropsophus elegansDendropsophus minutusDendropsophus nanusDendropsophus novaisiDendropsophus oliveiraiDendropsophus soaresiHypsiboas albomarginatus

APPENDIX I.List of species included in the analysis.


Hypsiboas albopunctatusHypsiboas crepitansHypsiboas exastisHypsiboas faberHypsiboas lundiiHypsiboas polytaeniusHypsiboas pombaliHypsiboas ranicepsPhasmahyla timboPhyllodytes luteolusPhyllodytes melanomystaxPhyllodytes tuberculosusPhyllodytes wuchereriPhyllomedusa bahianaPhyllomedusa nordestinaPseudis bolbodactylaPseudis fuscaScinax argyreornatusScinax auratusScinax camposseabraiScinax eurydiceScinax fuscomarginatusScinax fuscovariusScinax nebulosusScinax pachycrusScinax strigilatusSphaenorhynchus bromelicolaSphaenorhynchus prasinusTrachycephalus atlasTrachycephalus mesophaeusTrachycephalus typhoniusVitreorana eurygnathaXenohyla eugenioi

Hylodidae

Crossodactylus cyclospinus

APPENDIX I.Continued.

APPENDIX I.Continued.

Leiuperidae

Physalaemus albifronsPhysalaemus centralisPhysalaemus cicadaPhysalaemus cuvieriPhysalaemus kroyeriPleurodema diplolisterPseudopaludicola mystacalis

Leptodactylidae

Leptodactylus caatingaeLeptodactylus furnariusLeptodactylus fuscusLeptodactylus hylaedactylusLeptodactylus mystaceusLeptodactylus mystacinusLeptodactylus natalensisLeptodactylus latransLeptodactylus podicipinusLeptodactylus pustulatusLeptodactylus syphaxLeptodactylus troglodytesLeptodactylus vastus

Microhylidae

Chiasmocleis albopunctataDermatonotus muelleriElachistocleis ovalisElachistocleis piauiensis

Pipidae

Pipa carvalhoi

Ranidae

Lithobates palmipes

Strabomantidae

Strabomantis aramunha


AMPHIBIAN CONSERVATION IN THE CAATINGA BIOME AND … · AMPHIBIAN CONSERVATION IN THE CAATINGA BIOME AND SEMIARID REGION OF BRAZIL MILENA CAMARDELLI 1,3 AND MARCELO F. NAPOLI 2 1Programa

Documents