BIOLOGICAL CONSERVATION 139 (2007) 37– 46
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biocon
Landscape genetics of Physalaemus cuvieri in Brazilian Cerrado: Correspondence between population structure and patterns of human occupation and habitat loss
Mariana Pires de Campos Tellesa,*, Jose´ Alexandre Felizola Diniz-Filhob, Roge´rio Pereira Bastosb, Thannya Nascimento Soaresb,c, Lorena Dall‘Ara Guimara˜esd, Leoˆncio Pedrosa Limae aLaborato´rio de Gene´tica and Biodiversidade Zootecnia/MCAS/MGene-PROPE, Universidade Cato´lica de Goia´ s (UCG), CP. 86, 74605-010 Goiaˆ nia, GO, Brazil bDepartamento de Biologia Geral, ICB, Universidade Federal de Goia´ s (UFG), CP. 131, 74.001-970 Goiaˆ nia, GO, Brazil cGraduate Program in Agronomy, EA, UFG, CP. 131, 74.001-970 Goiaˆ nia, GO, Brazil dCentro Aplicado de Estudos e Pesquisas em Educac¸a˜ o, UFG, CP. 131, 74.001-970 Goiaˆ nia, GO, Brazil eIbama/MMA, Parque da Chapada das Mesas, Carolina, MA, Brazil
ARTICLE INFO ABSTRACT
Article history: It is now widely recognized that knowledge on population genetic structure is important to Received 11 September 2006 evaluate population viability and persistence or to establish conservation priorities. In this Received in revised form context, species that are locally abundant or widely distributed can be informative on how 27 May 2007 broad scale processes of habitat loss and fragmentation, as those caused by intensive Accepted 2 June 2007 human occupation, affect population genetic structure. In this paper, we analyzed popula- Available online 31 July 2007 tion genetic structure of Physalaemus cuvieri (Amphibia: Leptodactylidae) in the core of the Cerrado biome, in the Goia´s State, Central Brazil, using RAPD molecular markers. Local Keywords: populations are genetically different according to RAPD markers, and an analysis of molec- Physalaemus cuvieri ular variation (AMOVA) revealed a significant interpopulational variance component Genetic distances around 10%. However, these population differentiation patterns are not strongly structured Genetic discontinuities in geographic space, and a Mantel spatial correlogram indicated only a slight significant Habitat fragmentation spatial structure at short geographic distances. These patterns are expected by the ecolog- Human occupation ical and life-history knowledge of the species, leading to a relatively low magnitude of pop- Spatial genetics ulation differentiation coupled with short distance spatial patterns. Moreover, even these Mantel tests weak patterns showed a signature of effects of human occupation and habitat loss on genetic differentiation at regional scale, with discontinuities to gene flow in two particular regions of the State with more intense habitat loss and older human settlement. Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction Earth‘s biodiversity. This becomes crucial by taking into ac- count the high current rates of human population growth One of the great challenges for the 21st century is to develop and energy use, the increasing demand for area and, conse- and implement strategies to avoid loosing a large amount of quently, the high rates of habitat loss, fragmentation and
* Corresponding author: Tel.: +55 6239461709; fax: +55 6239461714. E-mail addresses: [email protected], [email protected] (M.P.C. Telles). 0006-3207/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2007.06.003 38 BIOLOGICAL CONSERVATION 139 (2007) 37– 46
species extinctions (Fahring, 2003; Ewers, 2005). Understand- Filho et al., 2006, 2007), although this figure is almost certainly ing how biological diversity, at multiple hierarchical levels, re- underestimated (see Diniz-Filho et al., 2005; Bini et al., 2006). sponds to human modifications at landscape level is one of At the same time, many studies in amphibian population the challenges for the researchers, and new spatial analyses genetics have been conducted, usually revealing a high level and geographic information systems (GIS) technologies have of population structure, which can be associated with their now an important role in this endeavor (Kidd and Ritchie, ecological and behavioral characteristics, especially the rela- 2006; Fischer and Lindenmayer, 2007). tively low dispersion rates, high habitat fidelity and specificity In the Brazilian Cerrado, recent human occupation has (Jehle et al., 2001; Newman and Squire, 2001; Austin et al., been characterized by a quick expansion from the South 2003; Palo et al., 2003; Palo et al., 2004a; see also Manier and and Southeastern regions of the country with extensive cattle Arnold, 2006 for a comparative evaluation). In Brazil, studies ranching and intensive farming, generating high rates of hab- with a few species of reptiles and amphibians showed effects itat loss and fragmentation (Klink and Moreira, 2002; Oliveira of environmental variation and recent human occupation on and Marquis, 2002; Klink and Machado, 2005; Rangel et al., genetic diversity and population structure, but they were usu- 2007). Studies at broad regional scales are still lacking, but it ally conducted in Atlantic forest region (e.g., Pellegrino et al., is almost certain that this magnitude of anthropic effects 2005; Grazziotin et al., 2006). are creating strong impacts on biodiversity, at ecosystem, At least in some studies revealed that even very recent community and population levels (see Brooks et al., 2002). processes of habitat loss and fragmentation affected genetic The Cerrado is the second largest biome of Brazil (the first diversity in amphibian populations (e.g., Scribner et al., is the Amazon forest) and consists in a mosaic of different 2001; Curtis and Taylor, 2004). Cushman (2006) recently re- habitat types, ranging from open savannas to highly dense viewed the effects of habitat loss and fragmentation in dry forests, and thus harbors an immense floral and faunal amphibians, and called attention to many important issues. diversity. Myers et al. (2000) ranked the Brazilian Cerrado For instance, the first important aspect is that these effects among the 25 global biodiversity conservation ‘‘hotspots’’ ex- cannot be understood at local spatial scales, and analyses at actly because of the high level of plant endemism and high broader geographical extents (i.e., at least regional or land- levels of habitat conversion. Current estimates indicate that scape levels) are necessary to evaluate how complex pro- about 70–80% of the original Cerrado area has been already cesses related to human occupation interact at these scales converted to ‘‘anthropic landscape’’, and only about 6% of this and affect biodiversity. Short term impact of habitat loss remaining area is protected by conservation units and parks. and fragmentation seems to be also strongly linked to life his- Thus, there is urgent need for implementation of conserva- tory characteristics of the species, and thus studies ‘‘...will be tion actions, which must be based on best possible available benefit from integrating large empirical studies with molecu- information on biodiversity patterns (Myers et al., 2000; Cav- lar genetics and simulation modeling’’. alcanti and Joly, 2002). Many studies in conservation genetics have been focused In this context, it is now recognized that knowledge of on endangered species, hoping that a better understanding population genetic structure is important to evaluate popula- of genetic parameters furnishes effective information to in- tion viability and to help establishing significant units for con- crease their persistence. On the other hand, species locally servation (Crozier, 1997; Hitchings and Beebee, 1997; abundant or widely distributed can be informative on how Frankham et al., 2003; Beebee and Rowe, 2004; Frankham, broad scale processes of habitat loss and fragmentation and 2005). Perhaps more importantly, patterns of genetic structure intensive human population affected population genetic within and among populations should be integrated into a structure (Diniz-Filho and Telles, 2006; see also Cushman, more complex evaluation of human occupation patterns 2006; for a more specific discussion regarding amphibian pop- and other environmental data, so that genetic parameters ulations). These approaches are of course not mutually exclu- can be explicitly analyzed on a human occupied landscape sive, but this partition may be useful to ensure how non- (see Manel et al., 2004). This also leads to an evaluation of endangered, highly abundant and widely distributed species how human modifications at landscape level affect genetic are also important targets for applied genetic analyses. diversity. Both possibilities can be integrated into a single In this paper, we analyzed population genetic structure of framework, which has been recently called ‘‘landscape genet- Physalaemus cuvieri (Amphibia: Anura: Leptodactylidae) in the ics’’, analogous to the well-established field of landscape ecol- core of the Cerrado biome, in the Goia´s State, Central Brazil, ogy (Manel et al., 2004; Guillot et al., 2005). It is expected that using RAPD molecular markers. P. cuvieri is a small Leptodac- understanding how human modified landscapes affect (or af- tylidae (about 3 cm of rostrum–anal length), with a broad geo- fected in the recent past) changes in population genetic struc- graphic extent in South America, ranging from Brazilian ture allows a more realistic evaluation of the best possible Northeast up to Paraguay and north of Argentina. Approxi- conservation actions. mately 25% of its geographic range is within the Cerrado Amphibians have been widely used as indicators of habitat biome, and it can be found in many distinct habitats, includ- change, because they are quite sensitive to human induced ing open grassland, flooded savannahs and pastureland. It habitat changes (see Cushman, 2006, for a recent review). At breeds in temporary waterbodies, and the eggs are laid in the same time, there is a general need to understand these ef- foam nests, attached to grass stems at the margin of the pond fects, because amphibian population declines have been re- (see
among local populations will be associated with recent hu- (5U-InvitrogenTM). PCR profiles comprised of a initial denatur- man occupation and habitat loss in Goia´s State. ation of DNA for 3 min at 96 °C, 40 cycles at of 1 min denatur- ation at 92 °C, 1 min annealing at 35 °C, 1 min extension at 2. Methods 72 °C, and one 3 min cycle at 72 °C for final extension. The amplification products were separated on 1.5% agarose gels 2.1. Sampling, DNA extraction, primer selection and stained with ethidium bromide, run in 1X TBE buffer at 120 V RAPD–PCR amplification for 4 h, in groups with 22 samples (individuals) separated by 100 bp DNA ladder (Amersham Pharmacia BiotechTM). Digital The material used for this analyses and estimation of basic images of gels were capture using EDAS120-KODAK and indi- genetic parameters and population structure was described vidual profiles for all gels were scored for the presence (1) or ab- in detail by Telles et al. (2006). In short, it consisted of a total sence (0) of bands (loci), using the 100 bp DNA ladder as a of 214 individuals collected from 18 localities (local popula- reference for aligning bands of different gels to different loci tions hereafter) of Goia´s state, Central Brazil (Fig. 1). Individu- and to minimize ambiguity in the coding procedure. als were dissected in the field and the liver was removed and preserved in 95% ethanol for further DNA extraction, primer 2.2. Statistical analyses selection and RAPD–PCR amplification (Souza et al., 2002). Genomic DNA was extracted using GFX Tissue kit protocol A detailed analysis of patterns of interpopulational genetic (Amersham Biosciences). differentiation was performed by Telles et al. (2006), and only Nine primers with consistent and well amplified bands were the main points will be highlighted here. First, molecular var- used to analyze the full set of DNA samples from the 214 indi- iation obtained by scoring RAPD bands was analyzed using an viduals, after initial tests based on a total of 120 primers. The Analysis of Molecular Variance (AMOVA). The AMOVA allows polymerase chain reaction (PCR) was done in a PCT-100, MJ one to partition multivariate Euclidian distances calculated Research PCR system with a total volume of 20 ll containing between pairs of individuals using RAPD bands into among 9.34 ll of water, 3 ll of genomic DNA (2 ng/ll), 2.6 ll of 10X and within-population variance components (Excoffier et al., PCR Buffer (InvitrogenTM) 2.08 ll of dNTPs (2.5 mM), 0.78 ll 1992; Souza et al., 2002). The proportion of genetic variation of MgCl2 (2.5 mM), 2 ll of primer, 0.2 ll of Taq polymerase among populations is then estimated by the UST statistics
Fig. 1 – Map showing the municipalities of Goia´s State in Central Brazil and the 18 local populations of Physalaemus cuvieri analyzed in this paper. 40 BIOLOGICAL CONSERVATION 139 (2007) 37– 46
(analogous to the classical FST value), an estimate of the We also correlate the genetic distance with average values among-population component of genetic divergence, and of 7 variables of human occupation and habitat loss, mea- the significance of this estimate against the null hypothesis sured at municipality level across the edges established by of no interpopulational variation was tested by 5000 random Delaunay network. First, we evaluated which municipalities permutations using AMOVA V. 1.55 (Excoffier, 1993). of the Goia´s state are situated along each edge, and obtained Since the papers by Sokal and Oden (1978a,b), spatial pat- data for each one. Average values of the socio-economic or terns of genetic variability have been used to investigate fragmentation data (see below) in these municipalities can microevolutionary processes underlying genetic variation then be paired with the genetic distances between pairs of lo- and how the emerging patterns can help establishing con- cal populations linked by the respective edge in the Delaunay servation strategies (see Sokal et al., 1997; Smouse and Peak- network. The first two variables used in the analyses were all, 1999; Diniz-Filho and Telles, 2002, 2006; Escudero et al., geographic distances (G, at log scale) and number of munici- 2003; Epperson, 2003; Pa´lsson, 2004; Telles et al., 2003; Miller, palities along Delaunay network (NM), giving an idea of ‘size’ 2005; Soares et al., in press). Because of the nature of RAPD of the edge. Other socio-economic variables more directly data, we used two different forms of Mantel tests of matrix associated with patterns and processes of human occupation correspondence (Smouse et al., 1986; Manly, 1985, 1997)to at landscape level and used here were average year of crea- evaluate the spatial patterns in genetic distances, estimated tion of municipalities (Y), giving an idea of both age of human by the pairwise AMOVA‘s UST. A first approach is to obtain a settlement and political division of land, average human pop- single standardized Mantel’s Z coefficient, which is actually ulation size in the year 2000 (H, at log scale) and average per a Pearson correlation between pairwise geographic and ge- capita income (I), to furnish an idea of intensity of human im- netic distances. In this paper, we used 5000 random permu- pact and available income. We were also able to obtain two tations to build a null distribution of the standardized variables related more directly to habitat loss, i.e., number Mantel’s Z that allow testing its statistical significance (see of fragments (NF), average area of the fragments (AF). The so- Manly, 1997). However, it is possible to expand this to build cio-economic variables (G,Y,H,I and NM) were obtained from a Mantel correlogram (Diniz-Filho and Telles, 2002; Epper- Brazilian Institute of Statistics and Geographic (IBGE, see son, 2003). In this case, the idea is to compare the genetic
the model is actually the best explanatory model. These val- of the predictors associated with human occupation and hab- ues of wi are usually standardized by their sum across all itat fragmentation, revealed that the best model predicting models evaluated, so they are dependent on the set of models genetic distances includes human population and year of used, and are given then by municipality creation (H + Y)(Table 1). However, the model X with the two variables of habitat fragmentation (AF + NF) ð 1=2DAICÞ ð 1=2DAICÞ wi ¼ e = e : and the one with geographic distances and year of creation i of municipalities (G + Y) have quite close AIC values, being then equivalent, in information content, to the best model.
3. Results Despite this, the wi values for these models, revealing their relative information content (see Burham and Anderson, The AMOVA revealed a significant inter-population variance 2002) are relatively low, around 28%, indicating then a high component (UST = 0.10; P < 0.0002 with 5000 permutations), uncertainty regarding model choice (although the sum of but the Mantel test comparing pairwise UST and geographic these values reveal that there is 86.5% that these are the cor- distances was not significant (r = 0.109; P = 0.219 with 5000 rect models, among those tested). Anyway, it is also impor- permutations). Thus, despite significant population differen- tant to highlight that the R2 of the best model was equal to tiation, there is no global spatial structure in genetic dis- 26%, so that there is a lot of unexplained variation in genetic tances among local populations. Multivariate correlogram distances. showed a slightly significant matrix correlation in the first geographic distance class (r = 0.169; P = 0.042), indicating that genetic similarity among local populations close in geo- 4. Discussion graphic space tends to be slightly larger than expected by chance alone. Matrix correlation then decreases to non-sig- 4.1. Overall patterns of genetic divergence nificant values up to the fifth distance class and then tends to stabilize, so that local populations situated at distances RAPD genetic analysis showed that there is a significant pop- smaller than about 180 km tend to be more similar than ran- ulation structure for P. cuvieri in Goia´s State of Central Brazil, domly selected pairs of local populations (see Diniz-Filho and as found for many other species of anurans worldwide (e.g., Telles, 2002; see Telles et al., 2006 for details). Driscoll, 1998; Baptista, 2001; Marsh and Trenham, 2001; Palo Some of the 45 edges along the Delaunay network have et al., 2004b) and discussed in more detail by Telles et al. high ratios between genetic and geographic distances, and (2006). Although local populations tend to be slightly differ- they successively characterize three regions of the State in ent, this population differentiation is not strongly structured which the discontinuities are concentrated (Fig. 2a). The high- in geographic space, and the Mantel correlogram indicated est 10% ratios is found for the edge between local populations only a slight significant and relatively weak spatial structure 15–4, 15–16, 15–5 and 7–17, in the southeastern of the State. If at short geographic distances. we highlight the 25% highest values of this ratio and showed The combination between significant population differ- that they are found basically in two regions of Goia´s State in entiation (revealed by FST or analogous statistics) and low which discontinuities are concentrated (Fig. 2a). The first is spatial pattern suggests that it may be difficult to find a the southeastern region, with seven local samples with high general explanation for genetic divergence among local ratios for the edges linking sequentially local populations 4, populations across geographic space (Sokal and Oden, 5, 6, 8, 11, 15 and 16. The second is in the northwestern region, 1978a,b; Sokal et al., 1986). Indeed, Pearse and Crandall with high ratios for three edges linking local populations 9, 13 (2004) recently pointed out that, beyond simple estimates and 18. There is an additional high ratio for the edge between of magnitude of population differentiation, it is important local populations 10 and 12, in the northeast region of the to search for more complex spatial and historical processes State. Using the highest 50% ratios, these two groups, espe- affecting genetic divergence in regional scales. P. cuvieri is a cially the southeastern one, are strongly reinforced, with only locally abundant species and, although specific studies on two high ratios in edges connecting the northwestern group dispersal ability of this species are lacking, its wide geo- of discontinuities with northeastern edges (between local graphic distribution and ability to live in distinct habitat populations 10–18) and be connecting the northeastern edge types suggest at least moderate dispersion rates. Thus, this with the entire group of discontinuities in the southeastern combination of life-history characteristics (moderate dis- region (edge between local population 10–17). The southeast- persal reinforced by low genetic drift due to high local ern and northwestern groups of discontinuities correspond, abundances), leads to relatively low levels of population dif- respectively, to areas with highest current population densi- ferentiation and short distance population structured ties in the southeast part of the Cerrado region and of recent (Gillespie, 1998; Barber, 1999; Beebee and Rowe, 2004; Rous- occupation by human populations in the northwest (see Bra- set, 2004), as observed here. sil, 1999; and Soares et al., in press). More broadly, these two Because of the relatively low value of the Mantel test in the groups of discontinuities in genetic variability might be visu- first geographic distance class, it is important to stress that ally associated with distribution of natural vegetation patches even some of the very close local populations might tend to across Goia´s State defined by Ferreira (in press) (Fig. 2b). be independent in respect to genetic variability. This is con- Indeed, a more detailed evaluation of these spatial pat- firmed by mapping the relative genetic distances, revealing terns, obtained using a multiple regression analysis between a set of genetic discontinuities in the region. The most impor- genetic distances along the edges and different combinations tant point is that these discontinuities are not randomly 42 BIOLOGICAL CONSERVATION 139 (2007) 37– 46
Fig. 2 – (a) Results from discontinuity analysis, in which edges with variable sticks and lines show edges with successively high ratios between pairwise genetic relationships and geographic distances between local populations, along the Delaunay network. (b) Map of remaining habitat fragments covering Goia´s State, showing a higher level of habitat loss and fragmentation in southeastern part of the State (see Ferreira et al. (in press)). distributed along the edges of Delaunay network, but instead Thus, beyond microevolutionary and ecological processes, highly aggregated in two main parts of the Goia´s state, corre- human effects may help explaining patterns of genetic dis- sponding to areas densely occupied by human populations. tances in this species. BIOLOGICAL CONSERVATION 139 (2007) 37– 46 43
Table 1 – Results of model selection procedure to explain genetic divergence along Delaunay network, using corrected AIC values (AICC)
* 2 Models AICC Di wi R
1 Full model 40.512 29.421 0.000 0.346 2 G + Y 11.093 0.0025 0.288 0.136 6 NF + AF 11.093 0.0020 0.288 0.138 12 H + Y 11.091 0.0000 0.289 0.263
2 The Di indicates the AICC differences, whereas wi is the probability that each model is the true model. The R is the amount of genetic distance
explained by each combination of predictors. Only the full model and model for which values of Di were smaller than 2 are shown. See text for codes of the predictors in each model.
4.2. Human occupation and habitat loss municipality are involved in the explanation of the groups of discontinuities in the southeast and northwest of Goia´s. The effects of habitat loss and human occupation to the ge- These factors are probably all correlated and, indeed, the netic structure of populations are well known in theoretical southeast region, with the largest cluster of discontinuities, terms, but only recently they started to be investigated empir- was the first region occupied, in the 18th century. This occu- ically (see Fahring, 2003; Cushman, 2006). Even so, studies at pation was accelerated by the expansion of the agricultural regional scales (or larger) are not common, and most analyses frontier in the 1940s and the transference of Brazilian Federal focus of more local differentiation among close fragments or capital (from Rio de Janeiro to Brasilia) in 1960. These com- in the effects on recently established and well-known anthro- bined processes generated a quick process of economic devel- pic barriers to gene flow. In amphibians, a few studies have opment in the region, and as a consequence there was a fast discussed how habitat fragmentation affects abundance (Gray increase in human population size, fragmentation of political et al., 2004) and population genetic structure (Kimberling units (generating a high number of new and recently created et al., 1996; Hitchings and Beebee, 1997; Ficetola and Bernardi, municipalities). As expected, these processes are undoubt- 2004; see also Cushman, 2006). edly associated with high levels of habitat loss and fragmen- Our study showed, at a relatively broad spatial scale, that tation in the region. The same processes occurred a little even recent patterns of human occupation and habitat frag- more recently in the northwest region of the State, in the mentation in the core region of Brazilian Cerrado may have af- southern part of the Araguaia river basin (Ribeiro et al., fected patterns of population genetic structure in P. cuvieri. 1995), and may be then associated with the second group of These effects appear as an additional component in the pat- barriers. terns of population differentiation, beyond those relatively Soares et al. (in press) recently observed similar patterns of weak expected under an ‘‘isolation-by-distance’’ process previ- genetic discontinuity in a Cerrado plant species, and found ously discussed. Although all these effects are relatively low (in that these discontinuities are located in regions characterized terms of a relatively low explanation power of genetic diversity, by the Brazilian government as ‘‘high anthropic pression’’ (the expressed by the r2), they are coherent with expectations based southern of the region, and of Goia´s State) and ’’socio-eco- on a combination of weak ‘‘isolation-by-distance’’ and habitat nomic expansion’’ (the northwestern part) (see Brasil, 1999). fragmentation caused by recent human occupation. Thus, These two groups correspond to the first two groups of dis- our models can help understanding patterns of genetic diver- continuities discussed above. This reinforces that, although sity in this species and to conduct local-scale evaluations that further studies at more local scales are necessary to clarify could potentially clarify how population-level processes, espe- by which population mechanisms the recent processes of hu- cially dispersal, are affected by human-induced landscape man occupation are affecting population structure, these ef- modifications. It is interesting to note that Cushman (2006) call fects are strong enough to be detected in overall, broad attention that species with moderate to high dispersal rates scale exploratory analyses of genetic data. tend to be more affected by habitat loss and fragmentation, and this could help explaining why even a relatively recent pro- 4.3. Concluding remarks cess at landscape level left a signal in genetic population struc- ture of this species. Our analyses showed that patterns of population genetic The evidence towards this explanation is the correspon- structure in P. cuvieiri in Central Brazil are coherent with those dence between genetic distances higher than expected by expected by considering its ecological and life-history, leading geographic distances (discontinuities) and highest values of to a relatively low magnitude of population differentiation socio-economic variables, based on average of municipalities coupled with short distance spatial patterns. Moreover, even along Delaunay network, reflecting simultaneously complex these weak patterns showed a signature of effects of human spatial patterns of human occupation and habitat loss. Be- occupation and habitat loss on genetic differentiation, at re- yond the simple visual correspondence between genetic dis- gional scale. Although this particular species do not require continuities and maps of habitat fragmentation in the specific protection, our analyses are important to allow an region, the regression models selected using AIC suggests evaluation of how humans affect genetic diversity and to dis- that current human population density, characteristics of cuss how alternative methodologies can be used to conserve habitat fragmentation and year since the creation of each this variability. 44 BIOLOGICAL CONSERVATION 139 (2007) 37– 46
Acknowledgements Curtis, J.M.R., Taylor, E.B., 2004. The genetic structure of coastal giant salamanders (Dicamptodon tenebrosus) in a managed This work was partially supported by a PRONEX program of forest. Biological Conservation 115, 45–54. Cushman, S.A., 2006. Effects of habitat loss and fragmentation on the Brazilian agency CNPq / SECTEC-GO (proc. 23234156) for amphibians: A review and prospectus. Biological Conservation establishing conservation priorities in the Brazilian Cerrado. 128, 231–240. Work by MPCT has been supported by Pro´ -Reitoria de Pesqu- Diniz-Filho, J.A.F., Telles, M.P.C., 2002. Spatial autocorrelation and isa and Po´ s-Graduac¸a˜o (PROPE)/UCG, whereas work by JAFDF the identification of operational units for conservation in and RPB has been continuously supported by CNPq productiv- continuous populations. Conservation Biology 16, 924–935. ity fellowships and by the Fundac¸a˜odeAmparoa` Pesquisa da Diniz-Filho, J.A.F., Telles, M.P.C., 2006. Optimization procedures Universidade Federal de Goia´s (FUNAPE/UFG). We thank Luis for establishing reserve networks for biodiversity conservation taking into account population genetic structure. Genetics and Mauricio Bini, Se´rgio Furtado dos Reis, Ka´tia C.M. Pellegrino, Molecular Biology 29, 207–214. Fausto Miziara, Jean Paul Metzger, Kelly Zamudio and anony- Diniz-Filho, J.A.F., Bastos, R.P., Rangel, T.F.L.V.B., Bini, L.M., mous reviewer for many suggestions that improved the ideas Carvalho, P., Silva, R.J., 2005. Macroecological correlates and presented in this paper. spatial patterns of anuran description dates in the Brazilian Cerrado. Global Ecology and Biogeography 14, 469–477. Diniz-Filho, J.A.F., Bini, L.M., Pinto, M.P., Rangel, T.F.L.V.B., Carvalho, P., Bastos, R.P., 2006. Anuran species richness, REFERENCES complementarity and conservation conflicts in Brazilian Cerrado. Acta Oecologica 29, 9–15. Diniz-Filho, J.A.F., Bini, L.M., Pinto, M.P., Rangel, T.F.L.V.B., Austin, J.D., Da´vila, J.A., Lougheed, S.C., Boag, P.T., 2003. Genetic Carvalho, P., Vieira, S.L.V., Bastos, R.P., 2007. Conservation evidence for female-biased dispersal in the bullfrog, Rana biogeography of anurans in Brazilian Cerrado. Biodiversity and catesbeiana (Ranidae). Molecular Ecology 12, Conservation 16, 997–1008. 3165–3172. Driscoll, D.A., 1998. Genetic structure, metapopulation processes Baptista, C.G., 2001. Estrutura gene´tica populacional de and evolution influence the conservation strategies for two Physalaemus cuvieri Fitzinger 1826 (Lissamphibia: endangered frog species. Biological Conservation 83, 43–54. Leptodactilydae) em fragmentos antro´picos e naturais de Duellman, W.E., Trueb, L., 1994. Biology of amphibians. McGraw Cerrado. Unpublished Master Thesis (Animal Biology), Hill, New York. Universidade de Brası´lia, Brası´lia-DF, Brasil. Epperson, B.K., 2003. Geographical Genetics. Princeton University Barber, P.H., 1999. Patterns of gene flow and population genetic Press, New Jersey. structure in the canyon treefrog, Hyla arenicolor. Molecular Escudero, A., Iriondo, J.M., Torres, M.E., 2003. Spatial analysis of Ecology 8, 563–576. genetic diversity as a tool for plant conservation. Biological Beebee, T.J.C., Griffiths, R.A., 2005. The amphibian decline crisis: A Conservation 113, 351–365. watershed for conservation biology? Biological Conservation Ewers, R.M., 2005. Are conservation and development 125, 271–285. compatible? Trends in Ecology and Evolution 20, 159. Beebee, T., Rowe, G., 2004. An Introduction to Molecular Ecology. Excoffier, L., 1993. Analysis of Molecular Variance: software Oxford University Press, Oxford. version 1.55. Genetics and Biometry Laboratory. University of Bini, L.M., Diniz-Filho, J.A.F., Rangel, T.F.V.L.B., Bastos, R.P., Pinto, Genova. M.P., 2006. Challenging Wallacean and Linnean shortfalls: Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of Knowledge gradients and conservation planning in a Molecular Variance inferred from metric distances among biodiversity hotspot. Diversity and Distributions 12, 475–482. DNA haplotypes: application to Human Mitochondrial DNA Brasil, 1999. Ac¸o˜ es priorita´rias para a conservac¸a˜oda restriction data. Genetics 131, 479–491. biodiversidade do Cerrado e Pantanal. MMA, Funatura, Fahring, L., 2003. Effects of fragmentation on biodiversity. Annual Conservation International, Fundac¸a˜o Biodiversitas, UnB. Review of Ecology Evolution and Systematics 34, 487–515. Brası´lia-DF, Brasil. Ferreira, L.G., in press. Conservac¸a˜o da Biodiversidade e Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A.B., sustentabilidade ambiental em Goia´s: Estrate´gias, prioridades Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., e perspectivas. AGMA/AGETOP, Goiaˆnia. Magin, G., Hilton-Taylor, C., 2002. Habitat loss and extinction Ficetola, G.F., Bernardi, F.D., 2004. Amphibians in a human- in the hotspots of biodiversity. Conservation Biology 16, dominated landscape: The community structure is related to 909–923. habitat features and isolation. Biological Conservation 119, Burham, K.P., Anderson, D.R., 2002. Model Selection and 219–230. Multimodel Inference: A Pratical Information-Theoretic Fischer, J., Lindenmayer, D.B., 2007. Landscape modification and Approach, 2nd ed. Springer, New York. habitat fragmentation: a synthesis. Global Ecology and Cavalcanti, R.B., Joly, C.A., 2002. Biodiversity and Conservation Biogeography 16, 265–280. Priorities in the Cerrado Region. In: Oliveira, P.S., Marquis, R.J. Frankham, R., 2005. Genetics and extinction. Biological (Eds.), The Cerrados of Brazil: Ecology and Natural History of a Conservation 126, 131–140. Neotropical Savanna. Columbia University Press, New York, Frankham, R., Ballou, J.D., Briscoe, D.A., 2003. Introduction to pp. 351–367. Conservation Genetics. Cambridge University Press, Colli, G.R., Bastos, R.P., Arau´ jo, A.F.B., 2002. The character and Cambridge. dynamics of Cerrado herpetofauna. In: Oliveira, P.S., Marquis, Gillespie, J.H., 1998. Population genetics: A concise guide. The R.J. (Eds.), The Cerrados of Brazil: Ecology and Natural History Johns Hopkins Press, London. of a Neotropical Savanna. Columbia University Press, New Gray, M.J., Smith, L.M., Brenes, R., 2004. Effects of Agricultural York, pp. 223–239. Cultivation on Demographics of Southern High Plains Crozier, R.H., 1997. Preserving the information content of species: Amphibians. Conservation Biology 18, 1368–1377. Genetic diversity, phylogeny and conservation worth. Annual Grazziotin, F.G., Monzel, M., Echeverrigaray, S., Bonatto, S.L., 2006. Review Ecology and Systematics 28, 243–268. Phylogeography of the Bothrops jararaca complex (Serpentes: BIOLOGICAL CONSERVATION 139 (2007) 37– 46 45
Viperidae): past fragmentation and island colonization in the Palo, J.U., Lesbarreres, D., Schmeller, D.S., Primmer, C.R., Merila¨,J., Brazilian Atlantic forest. Molecular Ecology 15, 3969–3982. 2004a. Microsatellite marker data suggest sex-biased dispersal Guillot, G., Estoup, A., Mortier, F., Cosson, J.F., 2005. A spatial in the common frog Rana temporaria. Molecular Ecology 13, statistical model for landscape genetics. Genetics 170, 2865–2869. 1261–1280. Palo, J.U., Schmeller, D.S., Laurila, A., Primmer, C.R., Kuzmins, S.L., Hitchings, S.P., Beebee, T.J.C., 1997. Genetic substructuring as a Merila¨, J., 2004b. High degree of population subdivision in a result of barriers to gene flow in urban Rana temporaria widespread amphibian. Molecular Ecology 13, 2631–2644. (common frog) population: Implications for biodiversity Pa´lsson, S., 2004. Isolation by distance, based on microsatellite conservation. Heredity 79, 117–127. data, tested with spatial autocorrelation (SPAIDA) and Jehle, R., Arntzen, J.W., Burke, T., Krupa, A.P., Ho¨ dl, W., 2001. The assignment test (SPASSIGN). Molecular Ecology Notes 4, annual number of breeding adults and the effective 143–145. population size of syntopic newts (Triturus cristatus, Triturus Pearse, D.E., Crandall, K.A., 2004. Beyond F-ST: Analysis of marmoratus). Molecular Ecology 10, 839–850. population genetic data for conservation. Conservation Johnson, J.B., Omland, K.S., 2004. Model selection in ecology and Genetics 5, 585–602. evolution. Trends in Ecology and Evolution 19, 101–108. Pellegrino, K.C.M., Rodrigues, M.I., Waite, A.N., Morando, M., Kidd, D.M., Ritchie, M.G., 2006. Phylogeographic information Yassuda, Y.Y., Sites, J.W., 2005. Phylogeography and species systems: putting the geography into phylogeography. Journal limits in the Gymnodactylus darwinii complex (Gekkonidae, of Biogeography 33, 1851–1865. Squamata): genetic structure coincides with river systems in Kimberling, D.N., Ferreira, A.R., Shuster, S.M., Keim, P., 1996. RAPD the Brazilian Atlantic forest. Biological Journal of Linnean marker estimation structure among isolated northern leopard Society 85, 13–26. frog population in the south-western USA. Molecular Ecology Rangel, T.F.L.V.B., Diniz-Filho, J.A.F., Bini, L.M., 2006. Towards an 5, 521–529. integrated computational tool for spatial analysis in Klink, C.A., Machado, R.B., 2005. Conservation of the Brazilian macroecology and biogeography. Global Ecology and Cerrado. Conservation Biology 19, 707–713. Biogeography 15, 321–327. Klink, C.A., Moreira, A.G., 2002. Past and current human Rangel, T.F.L.V.B., Bini, L.M., Diniz-Filho, J.A.F., Pinto, M.P., occupation, and land use. In: Oliveira, P.S., Marquis, R.J. (Eds.), Carvalho, P., Bastos, R.P., 2007. Human development and The Cerrados of Brazil: Ecology and Natural History of a biodiversity conservation in Brazilian Cerrado. Applied Neotropical Savanna. Columbia University Press, New York, Geography 27, 14–27. pp. 69–88. Ribeiro, M.C.L.B., Petrere-Junior, M., Juras, A.A., 1995. Ecological Legendre, P., Legendre, L., 1998. Numerical Ecology. Elsevier, integrity and fisheries ecology of the Araguaia–Tocantins river Amsterdam. basin, brazil. Regulated Rivers: Research and Management 11, Manel, S., Schwartz, M.K., Luikart, G., Taberlet, P., 2003. Landscape 325–350. genetics: Combining landscape ecology and population Richards, S.A., 2005. Testing ecological theory using the genetics. Trends in Ecology and Evolution 18, 189–197. information-theoretic approach: examples and cautionary Manel, S., Bellemain, E., Swenson, J.E., Franc¸ois, O., 2004. results. Ecology 86, 2805–2814. Assumed and inferred spatial structure of populations: The Rohlf, F.J., 1989. NTSYS-Pc: Numerical Taxonomy and Multivariate Scandinavian brown bears revisited. Molecular Ecology 13, Analysis System. Exeter publishers, New York. 1327–1331. Rousset, F., 2004. Genetic Structure and Selection in Subdivided Manier, M.K., Arnold, S.J., 2006. Ecological correlates of population Population. Princeton University Press, New Jersey. genetic structure: A comparative approach using a vertebrate Scribner, K.T., Arntzen, J.W., Cruddace, N., Oldham, R.S., Burke, T., metacommunity. Proceedings of the Royal Society B 273, 2001. Environmental correlates of toad abundance and 3001–3009. population genetic diversity. Biological Conservation 98, Manly, B.F.J., 1985. The Statistics of Natural Selection. Chapman 201–210. and Hall, London. Smouse, P.E., Peakall, R., 1999. Spatial autocorrelation analysis of Manly, B.F.J., 1997. Randomization Bootstrap and Monte Carlo individual multiallele and multilocus genetic structure. methods in biology. Chapman & Hall, London. Heredity 82, 561–573. Mantel, N., 1967. The detection of disease clustering and a Smouse, P.E., Long, J.C., Sokal, R.R., 1986. Multiple regression and generalized regression approach. Cancer Research 27, correlation extensions of the Mantel test of matrix 209–220. correspondence. Systematic Zoology 35, 627–632. Marsh, D.M., Trenham, P.C., 2001. Metapopulation dynamics and Soares, T.N., Chaves, L.J., Telles, M.P.C., Diniz-Filho, J.A.F., Resende, amphibian conservation. Conservation Biology 15, 40–49. L.V., in press. Landscape conservation genetics of Dipteryx Miller, M.P., 2005. Alleles in Space (AIS): Computer software for alata (‘‘baru’’ tree: Fabaceae) from Cerrado region of Central the joint analysis of interindividual spatial and genetic Brazil. Genetica. information. Journal of Heredity 96, 722–724. Sokal, R.R., Oden, N.L., 1978a. Spatial autocorrelation in biology. 1. Myers, N., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.B.A., Methodology. Biological Journal of Linnean Society 10, Kent, J., 2000. Biodiversity hotspots for conservation priorities. 199–228. Nature 403, 853–858. Sokal, R.R., Oden, N.L., 1978b. Spatial autocorrelation in biology. 2. Newman, R.A., Squire, T., 2001. Microsatellite variation and Some biological implications and four applications of fine-scale population structure in the wood frog (Rana evolutionary and ecological interest. Biological Journal of sylvatica). Molecular Ecology 10, 1087–1100. Linnean Society 10, 229–249. Oliveira, P.S., Marquis, R.J., 2002. The Cerrados of Brazil: Ecology Sokal, R.R., Smouse, P.E., Neel, J.V., 1986. The genetic structure of a and Natural History of a Neotropical Savanna. Columbia tribal population, the Yanomama indians XV patterns inferred University Press, New York. by autocorrelation analysis. Genetics 114, 259–287. Palo, J.U., O’Hara, R.B., Laugen, A.T., Laurila, A., Primmers, R., Sokal, R.R., Oden, N.L., Thomson, B.A., 1997. A simulation study of Merila, J., 2003. Latitudinal divergence of common frog (Rana microevolutionary inferences by spatial autocorrelation temporaria) life history traits by natural selection, evidence analysis. Biological Journal of Linnean Society 60, 73–93. from a comparison of molecular and quantitative genetic data. Souza, F.L., Cunha, A.F., Oliveira, M.A., Pereira, G.A.G., Pinheiro, Molecular Ecology 12, 1963–1978. H.P., Reis, S.F., 2002. Partitioning of molecular variation at local 46 BIOLOGICAL CONSERVATION 139 (2007) 37– 46
spatial scales in the vulnerable Neotropical freshwater turtle, Telles, M.P.C., Coelho, A.S.G., Chaves, L.J., Diniz-Filho, J.A.F., Valva, Hydromedusa maximiliani (Testudines, Chelidae): implications F.D., 2003. Genetic diversity and population structure of for the conservation of aquatic organisms in natural Eugenia dysenterica DC (‘‘cagaiteira’’ – Myrtaceae) in Central hierarchical systems. Biological conservation 104, Brazil: Spatial analysis and implications for conservation and 119–126. management. Conservation Genetics 4, 685–695. Stuart, S.N., Chanson, J.S., Cox, N.A., Young, B.E., Rodrigues, A.S.L., Telles, M.P.C., Bastos, R.P., Soares, T.N., Resende, L.V., Diniz-Filho, Fischman, D.L., Waller, R.W., 2004. Status and trends of J.A.F., 2006. RAPD variation and population genetic structure of amphibian declines and extinctions worldwide. Science 306, Physalaemus cuvieri (Anura: Leptodactylidae) in Central Brazil. 1783–1786. Genetica 128, 323–332.