Effects of habitat loss on taxonomic and phylogenetic diversity of understory Rubiaceae in Atlantic forest landscapes E.R. Andrade a,⇑ , J.G. Jardim b , B.A. Santos c , F.P.L. Melo d , D.C. Talora a , D. Faria a , E. Cazetta a a PPG Ecologia e Conservação da Biodiversidade, Laboratório de Ecologia Aplicada à Conservação, Universidade Estadual de Santa Cruz, Rodovia Jorge Amado km 16, 45662-900 Ilhéus, BA, Brazil b Departamento de Botânica, Ecologia e Zoologia, Universidade Federal do Rio Grande do Norte, Campus Universitário Lagoa Nova, 59072-970 Natal, RN, Brazil c Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba, Cidade Universitária, 58051-900 João Pessoa, PB, Brazil d Departamento de Ecologia Aplicada, Universidade Federal de Pernambuco, Avenida Prof. Moraes Rego, 123, Cidade Universitária, 50670-901 Recife, PE, Brazil article info Article history: Received 26 November 2014 Received in revised form 26 March 2015 Accepted 27 March 2015 Available online xxxx Keywords: Deforestation Forest cover Habitat fragmentation Originality Phylodiversity Rare species abstract Habitat loss has been long recognized as a major driver of the current species extinction crisis, but to date few studies have estimated deforestation thresholds above which biodiversity negatively respond to the changes in landscape configuration. In this study, we used a very representative understory plant family (Rubiaceae) to evaluate the effects of forest cover reduction in the landscape on taxonomic and phyloge- netic alpha and beta diversity, species originality, and composition. We selected nine Atlantic forest sites in southern Bahia, Brazil, immersed in landscapes ranging from 9% to 71% of remaining forest cover. We established a 50 Â 100 m plot in each site and identified all the Rubiaceae individuals. Forest cover reduc- tion at the landscape scale is strongly and negatively related to the number of individuals, species, and genera of Rubiaceae. Our model predicts that every 10% decreasing in landscape forest cover results in the loss of about three species and 74 individuals of understory Rubiaceae. Canopy openness at the plot level increased linearly with the decrease of forest cover in the landscape. A trend of linear decrease of the phylogenetic diversity with the loss of forest cover was also observed, increasing the co-occurrence of close relatives in more deforested landscapes. Furthermore, we found a reduction in the presence of spe- cies with higher values of originality along the gradient of forest cover, as a consequence of species rich- ness reduction. Regardless of the geographical distances, areas with similar values of forest cover were different in composition (b-diversity) and phylogenetic relationship (phylogenetic b-diversity), indicating that even highly deforested landscapes contribute to the regional diversity. We concluded that less forested areas are losing species richness and diversity, and as a consequence species originality, and presenting a general trend of phylogenetic impoverishment at the landscape scale. However, regional conservation initiatives should take all landscapes into account, because they all contribute to taxonomic and phylogenetic beta diversity, to ensure the long-term protection of the irreplaceable evolutionary history of Rubiaceae and the numerous organisms it supports. Ó 2015 Elsevier B.V. All rights reserved. 1. Introduction Habitat loss is a major driver of the current rates of species extinction, considered the main global threat to biodiversity, espe- cially in ecosystems with high endemism such as the Atlantic Forest in South America (Tabarelli et al., 2005). The Atlantic Forest from South America leads the world statistics of habitat loss, with over 93% of the original area lost (Galindo-Leal and Câmara, 2005). In Brazil, 11.4–16.0% of the Atlantic Forest still remains, including secondary forest areas and small fragments (Ribeiro et al., 2009), but only 3.5% of natural intact forest (Sloan et al., 2014). This high level of forest conversion results mainly from the accelerated population growth and the consequent increase of anthropogenic actions harmful to the environment. In this context, habitat loss and the synergistic effects of habitat fragmentation adversely affect biodiversity patterns and popula- tion persistence in anthropogenic landscapes (Andrén, 1994; Fahrig, 2003). These landscapes are usually represented by few and small remaining patches (Andrén, 1994), which are subjected to a myriad of modifications due to increasing forest area exposed to edge effects (Laurance et al., 2001; Murcia, 1995; Oliveira et al., http://dx.doi.org/10.1016/j.foreco.2015.03.049 0378-1127/Ó 2015 Elsevier B.V. All rights reserved. ⇑ Corresponding author. Tel.: +55 73 36805524. E-mail address: [email protected](E.R. Andrade). Forest Ecology and Management xxx (2015) xxx–xxx Contents lists available at ScienceDirect Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss on taxonomic and phylogenetic diversity of understory Rubiaceae in Atlantic forest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.03.049
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss on taxonomic and phylogenetic diversity of understory Rubiaceae in Aforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.03.049
E.R. Andrade a,⇑, J.G. Jardim b, B.A. Santos c, F.P.L. Melo d, D.C. Talora a, D. Faria a, E. Cazetta a
a PPG Ecologia e Conservação da Biodiversidade, Laboratório de Ecologia Aplicada à Conservação, Universidade Estadual de Santa Cruz, Rodovia Jorge Amado km 16, 45662-900Ilhéus, BA, Brazilb Departamento de Botânica, Ecologia e Zoologia, Universidade Federal do Rio Grande do Norte, Campus Universitário Lagoa Nova, 59072-970 Natal, RN, Brazilc Departamento de Sistemática e Ecologia, Universidade Federal da Paraíba, Cidade Universitária, 58051-900 João Pessoa, PB, Brazild Departamento de Ecologia Aplicada, Universidade Federal de Pernambuco, Avenida Prof. Moraes Rego, 123, Cidade Universitária, 50670-901 Recife, PE, Brazil
a r t i c l e i n f o
Article history:Received 26 November 2014Received in revised form 26 March 2015Accepted 27 March 2015Available online xxxx
Keywords:DeforestationForest coverHabitat fragmentationOriginalityPhylodiversityRare species
a b s t r a c t
Habitat loss has been long recognized as a major driver of the current species extinction crisis, but to datefew studies have estimated deforestation thresholds above which biodiversity negatively respond to thechanges in landscape configuration. In this study, we used a very representative understory plant family(Rubiaceae) to evaluate the effects of forest cover reduction in the landscape on taxonomic and phyloge-netic alpha and beta diversity, species originality, and composition. We selected nine Atlantic forest sitesin southern Bahia, Brazil, immersed in landscapes ranging from 9% to 71% of remaining forest cover. Weestablished a 50 � 100 m plot in each site and identified all the Rubiaceae individuals. Forest cover reduc-tion at the landscape scale is strongly and negatively related to the number of individuals, species, andgenera of Rubiaceae. Our model predicts that every 10% decreasing in landscape forest cover results inthe loss of about three species and 74 individuals of understory Rubiaceae. Canopy openness at the plotlevel increased linearly with the decrease of forest cover in the landscape. A trend of linear decrease of thephylogenetic diversity with the loss of forest cover was also observed, increasing the co-occurrence ofclose relatives in more deforested landscapes. Furthermore, we found a reduction in the presence of spe-cies with higher values of originality along the gradient of forest cover, as a consequence of species rich-ness reduction. Regardless of the geographical distances, areas with similar values of forest cover weredifferent in composition (b-diversity) and phylogenetic relationship (phylogenetic b-diversity), indicatingthat even highly deforested landscapes contribute to the regional diversity. We concluded that lessforested areas are losing species richness and diversity, and as a consequence species originality, andpresenting a general trend of phylogenetic impoverishment at the landscape scale. However, regionalconservation initiatives should take all landscapes into account, because they all contribute to taxonomicand phylogenetic beta diversity, to ensure the long-term protection of the irreplaceable evolutionaryhistory of Rubiaceae and the numerous organisms it supports.
� 2015 Elsevier B.V. All rights reserved.
1. Introduction
Habitat loss is a major driver of the current rates of speciesextinction, considered the main global threat to biodiversity, espe-cially in ecosystems with high endemism such as the AtlanticForest in South America (Tabarelli et al., 2005). The AtlanticForest from South America leads the world statistics of habitat loss,with over 93% of the original area lost (Galindo-Leal and Câmara,2005). In Brazil, 11.4–16.0% of the Atlantic Forest still remains,
including secondary forest areas and small fragments (Ribeiroet al., 2009), but only 3.5% of natural intact forest (Sloan et al.,2014). This high level of forest conversion results mainly fromthe accelerated population growth and the consequent increaseof anthropogenic actions harmful to the environment.
In this context, habitat loss and the synergistic effects of habitatfragmentation adversely affect biodiversity patterns and popula-tion persistence in anthropogenic landscapes (Andrén, 1994;Fahrig, 2003). These landscapes are usually represented by fewand small remaining patches (Andrén, 1994), which are subjectedto a myriad of modifications due to increasing forest area exposedto edge effects (Laurance et al., 2001; Murcia, 1995; Oliveira et al.,
2 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
2004; Saunders et al., 1991). As a consequence the microclimaticconditions are modified (Murcia, 1995). Besides, habitat loss mightalso change the canopy structure, which in turn modifies theamount of light that reaches the understory (Nicotra et al., 1999),influencing the distribution and persistence of many speciesmainly in this stratum. In Atlantic Forest fragments, there is littleinformation about the understory flora, as most studies focus onlythe tree stratum (Kozera, 2001). Understory species have beenhighly neglected despite the fact that they represent almost two-thirds of the woody plant diversity in tropical forests, playingspecific functions in the plant community (Poulsen and Balslev,1991). In tropical forests, understory plants are also of key impor-tance to the maintenance of many animal pollinators and seed dis-persers (de Souza et al., 2009).
Among the plants in the Neotropics, Rubiaceae is an outstand-ing family, with the fourth highest number of individuals amongall angiosperms (Chiquieri et al., 2004). The family is representedby approximately 13,100 species distributed in 611 genera(Govaerts et al., 2007). Brazil, is one of the numerous hotspots ofthe family diversity in the tropics (Govaerts et al., 2007), harboringapproximately 1396 species in 120 genera (Barbosa et al., 2014). Inaddition, the Rubiaceae family is distinguish in the forest under-story (Guaratini et al., 2008; de Lima et al., 2012; Martini et al.,2008; de Souza et al., 2009), exerting a strong influence on vegeta-tion structure. Several species are important resources for animalsthat feed on pollen and nectar (Castro and Oliveira, 2002; Lopesand Buzato, 2005), and Rubiaceae is appointed as one the mainsources of fleshy fruits to frugivores (Poulin et al., 1999; Snow,1981; Tabarelli et al., 1999). Thus, this family is one of the mostwell-suited to be used in ecological analyses in tropical vegetationdue to its representativeness, fewer taxonomical problems, andrepresentation in all kinds of growth habits (Delprete and Jardim,2012).
Identifying species that are being lost is undoubtedly veryimportant for conservation. However, species identities per se oftenbring little information regarding their function or evolutionaryhistory, which are also components of biodiversity (Cianciarusoet al., 2009; Swenson, 2011). Diversity measures such as phyloge-netic information are increasingly being used to assess biologicalcommunities responses to environmental changes (Helmus et al.,2010). Important advances in this field are helping us to under-stand the effects of fires on community assembly (Cavender-Bares and Reich, 2012; Cianciaruso et al., 2012; Verdú andPausas, 2007), the processes involved in community organizationduring forest regeneration (Letcher, 2010; Letcher et al., 2012),and the outcome of species loss and gain on community phyloge-netic diversity (Arroyo-Rodríguez et al., 2012; Cadotte and Strauss,2011; Santos et al., 2010, 2014).
Species susceptibility to habitat modifications might reducephylogenetic diversity if traits associated to those modificationsare evolutionarily conserved along particular lineages (Cavender-Bares et al., 2004; Webb et al., 2002). A recent study has demon-strated such a phylogenetic trait conservatism in Psychotria, themost common Rubiaceae genus, especially in hydraulic traitsrelated to tolerance to changes in soil moisture (Sedio et al.,2012). If this evolutionary pattern also holds for other traits asso-ciated to vulnerability to deforestation, it is likely that remainingcommunities in more deforested areas are formed by closer rela-tives, being more phylogenetically clustered and poorer than thoseinhabiting less deforested areas (see also Arroyo-Rodríguez et al.,2012; Santos et al., 2010, 2014). Given that the extinction of highlydistinct species from old and species-poor clades results in greaterloss of evolutionary information (Redding et al., 2008; Winter et al.,2013), we also evaluated how forest loss affects the most originalspecies.
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
Our goal was to evaluate how habitat loss at the landscape scalecan affect species diversity of the understory Rubiaceae family. Toaccomplish this goal, we sampled nine forest fragments locatedwithin 16 km2 landscapes ranging from 9% to 71% of remaining for-est cover, a proxy for habitat amount, and assessed not only com-mon descriptors of diversity – such as species composition,richness, and abundance – but also measures of species originality,alpha and beta phylogenetic diversity. We predicted that thereduction of forest cover at the landscape scale would have adetectable negative impact on species diversity, with greaterimpact of rare species in more deforested landscapes. We alsoexpected that forest cover reduction will favor the coexistence ofclosely related species, resulting in decreased phylogenetic diver-sity at both local (alpha) and regional (beta) scales.
2. Material and methods
2.1. Study sites
We conducted this study in Atlantic forest remnants from thesouth of the state of Bahia, northeastern Brazil, in the surroundingsof Una, Mascote, and Belmonte municipalities, where we couldidentify representative patches of the original forests. Thesemunicipalities are located between the Jequitinhonha and theContas rivers, where forest fragments have similar soil types,topography, and floristic composition (Thomas et al., 1998;Thomas, 2003), although human occupation has led to land use dif-ferences among these municipalities.
The Una region contains some of the last remnants of Atlanticforest in northeastern Brazil (Faria et al., 2009). In this region, thereis a Federal Protected Area – the Una Biological Reserve (RebioUNA) – established in 1980. After the incorporation of additionalforest fragments in 2007, it covers an area of 18,500 ha, beingone of the largest blocks of forest in southern Bahia (Schrothet al., 2011). Most Una landscapes comprise a mosaic of differentphysiognomies including mature and secondary forests, rubberand shade cocoa plantations (Faria et al., 2006).
By contrast, the region including Belmonte and Mascote munic-ipalities suffered a different and more pronounced deforestationprocess. In the 1990s, extensive areas of dense rainforest wereburned, yielding a large loss of forest cover due to the expansionof areas for grazing followed by Eucalyptus plantations(Nascimento et al., 2009). Nowadays, most of the regional matrixcomprises mainly open areas of pastures and patches ofEucalyptus crops.
2.2. Sampled landscapes
We used Landsat TM images, obtained on 07/14/2011 in orbits215/70 and 215/71 to map the forest cover in the study area. Themapping focused on these two regions, encompassing 1728 km2
of Una and 2018 km2 of Belmonte region. Due to the difficultiesto accurately map the different components of this mosaic fromLandsat images, and after an intensive field validation process,we considered the landscape forest cover as the sum of all forestcategories, including native forests in different stages of forest suc-cession and also those areas of shade cocoa plantations. We con-structed grids with cells of 16 km2 (hereafter landscapes) toestimate the amount of forest cover. One grid was located in Unaand proximities and another in Belmonte region. The proportionof forest habitats was obtained by manual classification usingIdrisi and ArcGis 9.3.
To ensure rainforest representativeness in our sampling, weexcluded those grid cells dominated by sand vegetation. We also
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx 3
eliminated mountainous, indigenous land, and conflicting regions,and areas with poor or no access. After the exclusion of these cells,and as we needed to sample in landscapes with different forestcover percentages, we performed a stratified sample from pre-cat-egorized classes, and ended up with a gradient of 9–71% (Fig. 1).
Fig. 1. Map of the study areas in south Bahia, Brazil, showing the forest areas (dark grsquares). In detail, each sampled plot (black circles) inside the landscapes with their resplandscape.
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
2.3. Rubiaceae species composition
In each landscape, we identified at least three forest remnantsand randomly selected one of them to place a plot of 50 � 100 min areas of mature or late secondary forest whenever possible. In
ay) and the location of each 4 � 4 km landscapes (ranging from 9% to 71% – blackective forest cover percentages. The ID number represents the identification of each
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
4 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
the nine selected forests, we sampled a total area of 4.5 ha. In eachplot, all understory individuals of Rubiaceae (0.5–4 m tall) weremarked and collected for posterior identification. We collectedfield data between January and December 2012. The botanicalmaterial was included in the reference collection at theUniversidade Estadual de Santa Cruz Herbarium (HUESC).
2.4. Phylogenetic diversity metrics and originality
We constructed a phylogenetic tree based on Bremer andEriksson (2009), using classifications into subfamilies and tribes(Fig. 2). Calibration was done in Phylocom 4.1 (Webb et al.,2008) using dating obtained in Bremer and Eriksson (2009),except for the tribe Gardenieae, that is not considered mono-phyletic. After constructing the time-calibrated phylogeny(Appendix B), we used the COMSTRUCT function of Phylocom 4.1to calculate the following phylogenetic metrics: mean phyloge-netic distance (MPD), mean nearest taxon phylogenetic distance(MNTD), net related index (NRI), and nearest taxon index (NTI).MPD measures the average phylogenetic distance among pairs ofindividuals drawn at random from a sample (including con-specifics); MNTD does the same, but the distance is measured to
foresteddeforested
Fig. 2. Phylogeny of Rubiaceae shrub community (n = 68) in south Bahia, Brazil. Red sqspecies presence on deforested areas (<47%). (For interpretation of the references to col
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
the closest non-conspecific relative (Vamosi and Wilson, 2008;Webb et al., 2008). NRI is a calculation of the effect size of MPDrelative to a null model and indicates whether taxa in a sampleare more phylogenetically clustered (positive NRI) or even (nega-tive NRI) than expected at random. NTI is the standardized effectsize of MNTD and quantifies the extent of terminal clustering(e.g. intrafamilial clustering); positive and negative values of NTIhave the same interpretation of NRI (Vamosi et al., 2009; Webbet al., 2002). We reported MPD and MNTD in millions of yearsand refer to them as metrics of phylogenetic diversity. NRI andNTI are expressed in units of standard deviation, and thereforeshould not be considered a true metric of biological diversity. Werefer to them as metrics of phylogenetic structure (see alsoArroyo-Rodríguez et al., 2012; Santos et al., 2014).
We used the phylogenetic calibrated tree to calculate an origi-nality value for each species, that represents the average rarity ofspecies traits, in other words, the species uniqueness (Pavoineet al., 2005). The species originality was calculated using theEqual-split index, that divides the evolutionary time representedby a branch equally among its daughter branches (Redding andMooers, 2006). The species were ranked based on the originalityvalue that reflects how evolutionary isolated a species is and can
uares indicate species presence on forested areas (<47%) and blue circles indicateour in this figure legend, the reader is referred to the web version of this article.)
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
Fig. 3. (A) number of forest fragments in which a species was recorded and; (B)number of unique species reported in each forest fragment immersed in a givenlandscape with different amounts of forest cover in south Bahia, Brazil.
E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx 5
be used as an indicator of the genetic distinctness (Redding andMooers, 2006). All species that displayed values above the meanEqual-split index were selected as the most originals and theirpresence was graphically represented along the nine forestremnants. We performed permutation tests to evaluate if speciesrichness was related to the number of original species in each rem-nant. We simulated 10,000 communities, each one able to containany of the species presented in all areas, but keeping fixed richnessand abundance values. In each randomization procedure weobtained the number of original species and the p values to inferwhether the observed value were different from the valuesexpected by chance.
2.5. Canopy openness
To measure canopy openness and, consequently, the amount oflight that reaches the understory, we took 10 hemispheric pho-tographs in each plot. We analyzed the photographs using GapLight Analyzer� (GLA, version 2.0), which converts each pho-tograph into a black-and-white picture and counts pixels not cov-ered by vegetation to determine canopy openness (Frazer et al.,1999). We calculated a mean value for each plot evaluated.
2.6. Data analysis
We calculated the Rényi diversity profiles for each one of thenine forest remnants, using the software Past 2.15 (Hammeret al., 2001). The Rényi series allows the comparison of differentdiversity indices, since each alpha value in the diversity profile cor-responds to a different diversity index (a = 0 richness; a = 1Shannon index; a = 2 inverse Simpson index (1/D); and higher avalues approximates to Berger–Parker index). Diversity profilescomprehend a good tool for diversity comparisons (Tóthmérész,1995). In the profile, each community is represented by a lineand in cases where the lines cross along the graphic, the commu-nities are non-comparable because the ranking position dependson indices properties (Tóthmérész, 1995). A smoothed line repre-sents high evenness, because the diversity values (y axis) areslightly influenced by the alpha values. An abrupt decrease alongy axis means low evenness, with rare and/or dominant species.
We performed linear regressions to evaluate the effects of forestcover reduction on total number of individuals, genera, and speciesof understory Rubiaceae and all the phylogenetic indices evaluated(MPD, MNTD, NRI, and NTI). We also performed a linear regressionto evaluate the relationship between the mean canopy opennessand the percentage of forest cover. A linear regression was usedbecause a generalized additive model (GAM) showed a linear rela-tionship between variables. The GAM is a model estimated bysmoothing curves, which allow describing the shape and revealingpossible non-linear relationships between variables, assuming thisstructure is not as rigid as that of a parametric function, such asgeneralized linear models (GLM) (Hastie and Tibshirani, 1986).
We performed a Detrended Corresponce Analysis (DCA) tosearch for a turnover in species composition along the gradientof forest cover, and to infer about the beta diversity patterns. Weevaluated the differences between landscapes by a multivariateanalysis of variance (MANOVA) with the scores of the two axis ofthe DCA ordination. We evaluated the effects of possible geo-graphic influences with a Mantel test between a geographic dis-tance matrix and a species dissimilarity matrix, and also betweena phylogenetic distance matrix that was defined by measuringthe mean phylogenetic distance between species in the plots(MPD). The latter was obtained by the Phylocom 4.1 based onthe mean phylogenetic distance of each pair of species from onelandscape to another (Webb et al., 2008) and it is a measure of phy-logenetic beta-diversity.
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
All the other statistical analyses were carried out in R software(R Core Team, 2013) using vegan (Oksanen et al., 2013), mgcv(Wood, 2006) and ade4 (Dray and Dufour, 2007) packages, withan adopted alpha of 60.05 considered significant.
3. Results
We collected 1857 Rubiaceae individuals in all forests remnantssurveyed, encompassing a total of 21 genera and 68 species(Appendix C). Psychotria was the genus with the highest numberof species (21), followed by Faramea (7) (Appendix C).
Species composition was variable among sampling sites, with22 species (32.4%) present in only one forest remnant, 18 species(26.5%) in two, only six species (8.7%) recorded in six or moreareas, and only one species, Psychotria platypoda, occurring in allsampled sites (Fig. 3A). The fragment in the landscape with thehighest forest cover percentage had the largest number of uniquespecies (N = 5), i.e. species present only in that area (Fig. 3B).From all the unique species, 63.3% were also rare, with one ortwo individuals (i.e. singletons or doubletons, respectively).Almost all areas showed at least one unique species that was alsorare (singleton or doubleton). The only exception was the area with9% of forest cover with no unique species. The number of single-tons ranged from 3 to 16, and the number of doubletons from zeroto five (Appendix A). The number of singletons and doubletons wasnot correlated with forest cover percentage (r2 = �0.12, p = 0.69;r2 = �0.07, p = 0.52, respectively). However, the proportion of rarespecies is being maintained in the landscapes (Appendix A).
Those more forested landscapes (i.e. 71%, 60%, and 47%) showeda higher number of species in the diversity profile (higher valueswhen a = 0) (Fig. 4). Indeed, the fragment located in the landscapewith 71% of forest cover was the most diverse, regardless of thediversity index evaluated. However, we noted that in areas with60% and 47% of forest cover in the surrounding landscape theShannon index (a = 1) decreased, indicating that although richnesswas high, evenness was low in both areas. The area with 15% of
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
Fig. 4. Diversity profile using the Rényi series representing the Rubiaceae assem-blages for each one of the nine forest fragments in Atlantic rainforest in south Bahia,Brazil. Alpha values: a = 0 richness; a = 1 Shannon index; a = 2 inverse Simpsonindex (1/D); and high a values approximates to Berger–Parker index.
Forest cover (%)
0 20 40 60 80
Num
ber o
f gen
era
4
6
8
10
12
14
16
18
Forest cov
0 20 40
Num
ber o
f ind
ivid
uals
0
100
200
300
400
500
600
700
A
y=-65.687+7.428xr²=0.58; p<0.05
y=5.44367+0.14262xr²=0.79; p<0.001
Fig. 5. Linear relationship between the forest cover gradient and the number of; (A) specsouth Bahia, Brazil.
6 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
forest cover in the landscape was the least diverse independentlyof the diversity index evaluated (Fig. 4). The curves representingthe 24% and 35% landscapes were smoother along the profile,which indicates high evenness. Except for areas with 71% (greaterdiversity) and 15% (lower diversity) of forest cover, all remainingfragments were ranked differently depending on the weight givento rare or dominant species.
We found that forest cover reduction results in a linear loss ofindividuals (r2 = 0.58, p < 0.05), genera (r2 = 0.79, p < 0.001), andspecies (r2 = 0.70, p < 0.01) of understory Rubiaceae (Fig. 5). Wealso found a linear but negative relationship between canopyopenness and forest cover (r2 = 0.74, p < 0.01) in which the rem-nants located in forested landscapes presented the lowest valuesof canopy openness.
The DCA ordination showed that in general all landscapes pre-sent high dissimilarity, indicating high beta diversity. The resultssuggested correlation between the percentage of forest cover andthe first axis of DCA (MANOVA; F1,7 = 10.01, p = 0.01). The first axisconcentrates most of the explained variance and indicates a higherdissimilarity of species composition among the more deforestedlandscapes (Fig. 6).
Remnants located in forested areas (>47%) together showedmore exclusive species (n = 29), and in general were well dis-tributed along the phylogeny. Eleven species were found only ondeforested areas (<47%) and would be extinct if these areas were
Forest cover (%)
0 20 40 60 80
Num
ber o
f spe
cies
5
10
15
20
25
30
35
40
er (%)
60 80
B
C
y=7.9106+0.3453xr²=0.70; p<0.01
ies; (B) genera; and (C) individuals of understory Rubiaceae in Atlantic rainforest in
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx 7
completely extirpated (Fig. 2). Forest cover reduction negativelyinfluenced the MPD index (r2 = 0.36, p = 0.050). We did not
DCA-1
-2 -1 0 1 2 3
DC
A-2
-2
-1
0
1
2
8.8%14.7%
24%
28%
35%
40%
47%
60%
71%
Fig. 6. DCA ordination of understory Rubiaceae species from the Atlantic rainforestin south Bahia, Brazil. The diameter of the circles are proportional to the percentageof forest cover and its proximity indicates greater similarity in species compositionbetween areas.
Forest c
0 20 40
MP
D
130
140
150
160
170
180
190
Forest cover (%)
0 20 40 60 80
NTI
-2
-1
0
1
2
3
r²= 0.30 ; p=0.073B
Fig. 7. Relationship between the forest cover gradient and the phylogenetic diversityphylogenetic distance); (B) NTI (nearest taxon index); and (C) NRI (net related index). Redcolour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
observed a linear response between forest cover reduction andthe other phylogenetic indices evaluated (MNTD [r2 = �0.13,p = 0.819], NTI [r2 = 0.30, p = 0.073], NRI [r2 = 0.28, p = 0.080])(Fig. 7). However, it is noteworthy that in general, forested areaspresented negative values of NTI and NRI, and deforested areaspresented positive values, indicating phylogenetic over-dispersionand clustering, respectively (Fig. 7).
We found eleven original species based on the Equal-split index(Fig. 8). The areas located in more forested landscapes showed ahigher number of those original species (Fig. 8). By contrast, inthose fragments surrounded by low percentage of forest coverage(<29%), only one or two original species from the set of 11 speciespreviously selected were found. The results from the permutationtest showed that only three forest remnants (35%, 41%, and 60%)presented more original species than expected by chance(p < 0.05). The number of original species in all other remnantswas influenced by species richness and abundance (9%: p = 0.793;15%: p = 0.584; 24%: p = 0.682; 29%: p = 0.497; 47%: p = 0.086;71%: p = 0.374).
We found no correlation between geographic distances and dif-ferences in species composition (Mantel; Rho = 0.28, p = 0.058).Moreover, the geographical distances were also not correlated withphylogenetic composition (phylogenetic b-diversity). Neither thegeographic proximity nor closeness in percentage of forest coverresulted in similar phylogenetic composition (Mantel; Rho = 0.22,p = 0.203).
over (%)
60 80
Forest cover (%)
0 20 40 60 80
NR
I
-1,0
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
3,0
r²=0.28 ; p= 0.08
A
y = 143.5717 + 0.4676xr²=0.36 ; p=0.05
C
metrics evaluated in Atlantic rainforest in south Bahia, Brazil. (A) MPD (meancircles indicate negative NTI and NRI values. (For interpretation of the references to
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
Fig. 8. Presence of the 11 species with higher values of originality (Equal-splitindex) along the forest cover gradient in south Bahia, Brazil.
8 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
4. Discussion
Our results demonstrated that habitat loss, measured as theamount of forest cover at landscape scale, is strongly and nega-tively related to species richness, abundance, and diversity pat-terns of understory Rubiaceae. Previous studies have consistentlyshown the negative effects of habitat loss on species richness ingeneral (Fahrig, 2003). However, empirical and theoretical evi-dence points to a non-linear relationship between habitat lossand biodiversity metrics, with more pronounced diversity losswhen small and isolated patches dominate the landscape(Andrén, 1994). A disproportional decline in species diversity oran extinction threshold after a certain level of habitat loss is aresult of synergistic effects of habitat reduction per se and itsincreased fragmentation (Andrén, 1994; Ficetola and Denoël,2009; Lindenmayer and Luck, 2005; Pardini et al., 2010). This spe-cies extinction pattern has been found in different plant familiessuch as Sapotaceae (Lima and Mariano-neto, 2013) andMyrtaceae (Rigueira et al., 2013) in anthropogenic landscapes inthe Brazilian Atlantic Forest. By contrast, we reported a linear lossof species and abundance indicating that any amount of forest lostat the landscape has a proportional negative consequence to thediversity patterns of Rubiaceae. Quite simply, our model predictsthat every 10% decreasing in landscape forest cover results in theloss of about three species and 74 individuals of understoryRubiaceae.
We also found that the number of rare species, independently ofspecies identity, was not correlated with the forest cover reductionin the landscape. This result indicates that extinction is mainly dueto habitat loss and its negative effects on species that are sensibleto understory alterations, such as moist loss, independent of spe-cies rarity or abundance. Forest cover reduction increases the meancanopy openness, an indirect measure of light availability tounderstory species at a local scale (Nicotra et al., 1999). Thisincrease in luminosity following habitat loss can alter the microcli-matic conditions and as a consequence might limit the establish-ment of the most sensitive species from the understory that areoften shade-tolerant species (Martini et al., 2008). Therefore, theareas located in more forested landscapes presented more struc-tured canopies that allow the presence of a higher number of spe-cies and individuals. A structured canopy enables a higher habitatheterogeneity that consists in one of the most important factor forthe maintenance of understory species (Enoki and Abe, 2004).
Rubiaceae family plays an important role in tropical forests dueto its high commonness in the understory (Guaratini et al., 2008;de Lima et al., 2012; de Souza et al., 2009), providing importantfood resources to a variety of animals that feed on pollen, nectar,and fruits (Amorim and Oliveira, 2006; Castro and Oliveira, 2002;Melo et al., 2003; Poulin et al., 1999). For this reason, the
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
breakdown of ecological interactions such as pollination and seeddispersal in more deforested areas might also lead to loss ofRubiaceae species. Some species offer high quality resources, richin carbohydrates (Castro and Oliveira, 2002) to ensure pollinationand fruit set as a consequence. These species also present high fruitavailability all over the year (Martin-gajardo and Morellato, 2003).Many Rubiaceae species are dispersed by birds, particularly theunderstory ones (Bremer and Eriksson, 1992). Microclimatic mod-ifications triggered by habitat loss and fragmentation, such as lightincrease and humidity decrease, have negative consequences forunderstory birds (Laps et al., 2003), highlighting the fragility ofthe understory species that might act as frugivores and seed dis-persers and stressing the possible negative consequences for plantdemography.
For instance, Amaioua guianensis is, an ornithocoric species(Amorim and Oliveira, 2006), was only sampled in fragmentslocated in the four most forested landscapes (41%, 47%, 60%, and71%). A possible factor that might limit the presence of this species,in addition to canopy openness, is the absence or reduction ofefficient seed dispersers. This species is also one of the eleventhoriginal Rubiaceae species, which might indicate it has uniquetraits not shared with other individuals from the phylogenetic tree.Moreover, forest cover reduction significantly diminished thenumber of original species in general and, in more deforested land-scapes (<29%), only one or two original species remained. Eventhough our results indicated that this is due to the decay in speciesrichness and abundance, not directly related to forest cover reduc-tion, we reinforce that we are effectively losing genetic features,and species with high conservation values. Since it has been sug-gested that species originality is an important measure to be con-sidered in the definition of priority conservation areas in order topreserve more evolutionary history (Pavoine et al., 2005), we con-sider that, independently of the mechanism this is an importantresult to take into account.
Forest cover reduction also negatively affected local phyloge-netic diversity, suggesting that local extirpation of Rubiaceae spe-cies is occurring in a cluster of the phylogenetic tree. It wasparticularly clear for MPD as 36% of its variability were explainedby the forest cover remaining in the landscape. Although the rela-tionships between forest cover and phylogenetic structure metrics(NRI and NTI) have been not significant, more deforested land-scapes (<29% of forest cover) also show higher positive values ofNRI and NTI when compared with more forested landscape (>29%of forest cover), which showed negative values of NRI and NTI.This trend of increase in phylogenetic clustering with forest reduc-tion suggests that environmental filters operate on conserved traitsthat enable species persistence (Cavender-Bares et al., 2004; Webbet al., 2002). The existence of physiological traits phylogeneticallyconserved was previously showed by Sedio et al. (2013, 2012) forPsychotria genus in Panama. These traits can influence habitat pref-erences and species distribution. Therefore we suggest that habitatloss might act as an environmental filter, modifying forest struc-ture and microhabitats and as a consequence limiting species dis-tribution and increasing phylogenetic clustering. On the otherhand, in more forested landscapes, species are less related andtheir distribution might be structured either by filtering on conver-gent characters or by biotic interactions (Cavender-Bares et al.,2004; Webb et al., 2002).
The more deforested landscapes areas are consequently morephylogenetically clustered, however they are still high dissimilaramong them. The high dissimilarity among forest remnants indi-cates that these areas presented a high b-diversity, including phy-logenetic b-diversity, and reinforces that areas located in similarforest cover percentages or geographically close are diverse in tax-onomic and phylogenetic composition. Thus, all areas contribute to
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx 9
the regional diversity, which is also confirmed by the presence ofrare species in all the sampled forest remnants.
We argue that conservation initiatives will fail in conserving theirreplaceable evolutionary history of Rubiaceae if they focus onlyon the remaining original habitat of conserved landscapes.Rather, a regional approach comprising the entire region of SouthBahia should be adopted to ensure the long-term protection ofthe Rubiaceae taxa and the numerous organisms they support.Similar directions have been indicated for other hyperdiverse trop-ical forest landscapes in Central Amazon (Laurance et al., 2007) andMexico (Hernández-Ruedas et al., 2014) in order to maintain bio-logical diversity and ecosystem services at the regional level.Nevertheless, because species persistence in small fragments tendsto be lower comparatively to those larger forested areas (Gibsonet al., 2013), long-term conservation actions to prevent species lossmight be different in each landscape context.
Our study highlighted that landscape scale habitat loss isstrongly related to significant loss in species diversity and anincrease in phylogenetic clustering for an important group of
Appendix A. Total number of genera, species, individuals, singletongradient of forest cover (%) at the landscape scale in south Bahia, B
Forest cover (%) Number of genera Number of species
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
understory plants. Although the number of original species is alsoreduced in fragments located in more deforested landscapes, lessforested areas are contributing to species and phylogenetic diver-sity to the regional pool. Therefore, all the sampled areas have rarespecies, and those represent not only a contribution to taxonomicdiversity but to increase the potential ecosystem functions in spa-tial scales (Mouillot et al., 2013).
Acknowledgments
The present study is a publication number #8 of the REDESISBIOTA, funded by the Brazilian Council of Science andTechnology (CNPq Proc. 563216/2010-7). We are grateful to thelandowners for allowing us to work on their properties and to allwho helped in the field work (especially Priscila Soares). We alsothank the English editing service funding by PROPP/UniversidadeEstadual de Santa Cruz. ERA received fellowships from FAPESB(BOL 0492/2011) and DF is granted with a CNPq fellowship(307221/2012-1).
s, and doubletons in nine forest fragments located along arazil.
Number of individuals Singletons Doubletons
66 6 568 6 237 8 2
165 8 073 4 1
246 3 3229 16 3633 6 4340 7 4
ased on Bremer and Eriksson (2009), using classifications into
Appendix C. Understory Rubiaceae species found in different landscapes from the south Bahia, Brazil. The percentage of forest coverin the landscape in which the species were found is indicated.
Species 8.8% 14.7% 24.2% 28.8% 34.6% 40.6% 47.3% 60% 70.6%
10 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss on taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.03.049
E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx 11
References
Amorim, F.F.W., Oliveira, P.P.E., 2006. Estrutura sexual e ecologia reprodutiva deAmaioua guianensis Aubl. (Rubiaceae), uma espécie dióica de formaçõesflorestais de cerrado. Rev. Bras. Botânica 29, 353–362.
Andrén, H., 1994. Effects of habitat fragmentation on birds and mammals inlandscapes with different proportions of suitable habitat: a review. Oikos 71,355–366.
Arroyo-Rodríguez, V., Cavender-Bares, J., Escobar, F., Melo, F.P.L., Tabarelli, M.,Santos, B.A., 2012. Maintenance of tree phylogenetic diversity in a highlyfragmented rain forest. J. Ecol. 100, 702–711. http://dx.doi.org/10.1111/j.1365-2745.2011.01952.x.
Barbosa, M.R., Zappi, D., Taylor, C., Cabral, E., Jardim, J.G., Pereira, M.S., Calió, M.F.,Pessoa, M.C.R., Salas, R., Souza, E.B., Di Maio, F.R., Macias, L., Anunciação, E.A. da,Germano Filho, P., Oliveira, J.A., Bruniera, C.P., Gomes, M., Toni, K., 2014.Rubiaceae in Lista de Espécies da Flora do Brasil. Jardim Botânico do Rio deJaneiro. Disponible in: <http://reflora.jbrj.gov.br/jabot/floradobrasil/FB210>.Acess in: 07 Abr. 2014.
Bremer, B., Eriksson, O., 1992. Evolution of fruit characters and dispersal modes inthe tropical family Rubiaceae. Biol. J. Linn. Soc. 47, 79–95. http://dx.doi.org/10.1111/j.1095-8312.1992.tb00657.x.
Bremer, B., Eriksson, T., 2009. Time tree of Rubiaceae: phylogeny and dating thefamily, subfamilies, and tribes. Int. J. Plant Sci. 170, 766–793. http://dx.doi.org/10.1086/599077.
Cadotte, M.W., Strauss, S.Y., 2011. Phylogenetic patterns of colonization andextinction in experimentally assembled plant communities. PLoS ONE 6,e19363. http://dx.doi.org/10.1371/journal.pone.0019363.
Castro, C.C., Oliveira, P.E., 2002. Pollination biology of distylous Rubiaceae in theAtlantic rain forest. SE Brazil. Plant Biol. 4, 640–646. http://dx.doi.org/10.1055/s-2002-35433.
Cavender-Bares, J., Reich, P., 2012. Shocks to the system: community assembly ofthe oak savanna in a 40-year fire frequency experiment. Ecology 93, 52–69.
Cavender-Bares, J., Ackerly, D.D., Baum, D.A., Bazzaz, F.A., 2004. Phylogeneticoverdispersion in Floridian oak communities. Am. Nat. 163, 823–843. http://dx.doi.org/10.1086/386375.
Chiquieri, A., Di Maio, F., Peixoto, A.A.L., 2004. A distribuição geográfica da famíliaRubiaceae Juss. na Flora Brasiliensis de Martius. Rodriguésia 55, 47–57.
Cianciaruso, M.V., Silva, I.A., Batalha, M.A., 2009. Diversidades filogenética efuncional: novas abordagens para a Ecologia de comunidades. Biota Neotrop.9, 0–11.
Cianciaruso, M.V., Silva, I.A., Batalha, M.A., Gaston, K.J., Petchey, O.L., 2012. Theinfluence of fire on phylogenetic and functional structure of woody savannas:moving from species to individuals. Perspect. Plant Ecol. Evol. Syst. 14, 205–216.http://dx.doi.org/10.1016/j.ppees.2011.11.004.
de Lima, R.A.F., de Souza, V.C., Dittrich, V.A.de O., Salino, A., 2012. Composição,diversidade e distribuição geográfica de plantas vasculares de uma FlorestaOmbrófila Densa Atlântica do Sudeste do Brasil. Biota Neotrop. 12,241–249.
de Souza, A., Almeida Jr, E., de Zickel, C., 2009. Riqueza de espécies de sub-bosqueem um fragmento florestal urbano, Pernambuco, Brasil. Rev. Biotemas 22, 57–66.
Delprete, P.G.P., Jardim, J.G., 2012. Systematics, taxonomy and floristics of BrazilianRubiaceae: an overview about the current status and future challenges.Rodriguésia 63, 101–128.
Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagramfor ecologists. J. Stat. Softw. 22 (4), 1–20.
Enoki, T., Abe, A., 2004. Saplings distribution in relation to topography and canopyopenness in an evergreen broad-leaved forest. Plant Ecol. 173, 283–291.
Fahrig, L., 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol.Evol. Syst. 34, 487–515. http://dx.doi.org/10.1146/annurev.ecolsys.34.011802.132419.
Faria, D., Laps, R.R., Baumgarten, J., Cetra, M., 2006. Bat and bird assemblages fromforests and shade cacao plantations in two contrasting landscapes in theAtlantic Forest of Southern Bahia. Brazil. Biodivers. Conserv. 15, 587–612.http://dx.doi.org/10.1007/s10531-005-2089-1.
Faria, D., Mariano-Neto, E., Martini, A.M.Z.A., Ortiz, J.V., Montingelli, R., Rosso, S.,Paciencia, M.L.B., Baumgarten, J., 2009. Forest structure in a mosaic of rainforest
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
sites: the effect of fragmentation and recovery after clear cut. For. Ecol. Manage.257, 2226–2234. http://dx.doi.org/10.1016/j.foreco.2009.02.032.
Ficetola, G.F., Denoël, M., 2009. Ecological thresholds: an assessment of methods toidentify abrupt changes in species-habitat relationships. Ecography (Cop.) 32,1075–1084. http://dx.doi.org/10.1111/j.1600-0587.2009.05571.x.
Frazer, G.W., Canham, C.D., Lertzman, K.P., 1999. Gap Light Analyzer (GLA), Version2.0: imaging software to extract canopy structure and gap light transmissionindices from true-colour fisheye photographs, users manual and programdocumentation. Copyright � 1999: Simon Fraser University, Burnaby, BritishColumbia, and the Institute of Ecosystem Studies, Millbrook, New York.
Galindo-Leal, C., Câmara, I.G., 2005. Mata Atlântica: biodiversidade, ameaças eperspectivas. Fundação SOS Mata Atlântica- Conservação internacional, Belohorizonte, p. 471.
Gibson, L., Lynam, A.J., Bradshaw, C.J.A., He, F., Bickford, D.P., Woodruff, D.S.,Bumrungsri, S., Laurance, W.F., 2013. Near-complete extinction of native smallmammal fauna 25 years after forest fragmentation. Science 341, 1508–1510.http://dx.doi.org/10.1126/science.1240495.
Govaerts, R., Frodin, D.G., Ruhsam, M., Bridson, D.M., Davis, A.P., 2007. Worldchecklist & bibliography of Rubiaceae. The Trustees of the Royal BotanicGardens, Kew.
Guaratini, M., Gomes, E., Tamashiro, J., Rodrigues, R., 2008. Composição florística daReserva Municipal de Santa Genebra, Campinas, SP. Rev. Bras. Botânica 31, 323–337.
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statisticssoftware package for education and data analysis. Palaeontol. Electron. 4, 1–9.
Communities contain closely related species during ecosystem disturbance.Ecol. Lett. 13, 162–174. http://dx.doi.org/10.1111/j.1461-0248.2009.01411.x.
Hernández-Ruedas, M.A., Arroyo-Rodríguez, V., Meave, J.A., Martínez-Ramos, M.,Ibarra-Manríquez, G., Martínez, E., Jamangapé, G., Melo, F.P.L., Santos, B.A.,2014. Conserving tropical tree diversity and forest structure: the value of smallrainforest patches in moderately-managed landscapes. PLoS ONE 9, e98931.http://dx.doi.org/10.1371/journal.pone.0098931.
Kozera, C., 2001. Composição florística e estrutura fitossociológica do estratoherbáceo-subarbustivo em duas áreas de floresta ombrófila densa, Paraná,Brasil. Dissertação de mestrado. Programa de Pós-Graduação em BiologiaVegetal, Instituto de Biologia, Universidade Estadual de Campinas, Campinas.
Laps, R.R., Cordeiro, P.H.C., Kajiwara, D., Ribon, R., Rodrigues, A.A.F., Uejima, A.M.K.,2003. Aves. In: Rambaldi, D.M., Oliveira, D.A.S. (Eds.), Fragmentação deecossistemas: causas, efeitos sobre a diversidade e recomendações depolíticas públicas. Ministério do Meio Ambiente/Secretaria de Biodiversidadee Florestas, Brasília, pp. 153–181.
Laurance, W.F., Williamson, G.B., Delamônica, P., Oliveira, A., Lovejoy, T.E., Gascon,C., Pohl, L., 2001. Effects of a strong drought on Amazonian forest fragments andedges. J. Trop. Ecol. 17, 771–785. http://dx.doi.org/10.1017/S0266467401001596.
Laurance, W.F., Nascimento, H.E.M., Laurance, S.G., Andrade, A., Ewers, R.M., Harms,K.E., Luizão, R.C.C., Ribeiro, J.E., 2007. Habitat fragmentation, variable edgeeffects, and the landscape-divergence hypothesis. PLoS ONE 2, e1017. http://dx.doi.org/10.1371/journal.pone.0001017.
Letcher, S.G., 2010. Phylogenetic structure of angiosperm communities duringtropical forest succession. Proc. Biol. Sci. 277, 97–104. http://dx.doi.org/10.1098/rspb.2009.0865.
Letcher, S.G., Chazdon, R.L., Andrade, A.C.S., Bongers, F., Breugel, M.V., Finegan, B.,Laurance, S.G., Mesquita, R.C.G., Martínez-Ramos, M., Williamson, G.B., 2012.Phylogenetic community structure during succession: evidence from threeNeotropical forest sites. Perspect. Plant Ecol. Evol. Syst. 14, 79–87. http://dx.doi.org/10.1016/j.ppees.2011.09.005.
Lima, M.M., Mariano-neto, E., 2013. Extinction thresholds for Sapotaceae due toforest cover in Atlantic Forest landscapes. For. Ecol. Manage. 312, 260–270.http://dx.doi.org/10.1016/j.foreco.2013.09.003.
Lopes, L., Buzato, S., 2005. Biologia reprodutiva de Psychotria suterella Muell. Arg.(Rubiaceae) e a abordagem de escalas ecológicas para a fenologia de floração efrutificação. Rev. Bras. Botânica 28, 785–795.
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
12 E.R. Andrade et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Martin-gajardo, I.S.A.N., Morellato, L.P.C., 2003. Fenologia de Rubiaceae do sub-bosque em floresta Atlântica no sudeste do Brasil. Rev. Bras. Botânica 26, 299–309.
Martini, A.M.Z., Jardim, J.G., Santos, F.A.M., 2008. Floristic composition and growthhabits of plants in understory, natural, treefall gaps and fire-disturbed areas of atropical forest in southern Bahia, Brazil. In: Thomas, W.W. (Ed.), The AtlanticCoastal Forest of Northeastern Brazil, vol. 100. Memoirs of the New York Bot,Garden, pp. 147–192.
Melo, C., Bento, E.C., Oliveira, P.E., 2003. Frugivory and dispersal of Faramea cyanea(Rubiaceae) in cerrado woody plant formations. Braz. J. Biol. 63, 75–82.
Mouillot, D., Bellwood, D.R., Baraloto, C., Chave, J., Galzin, R., Harmelin-Vivien, M.,Kulbicki, M., Lavergne, S., Lavorel, S., Mouquet, N., Paine, C.E.T., Renaud, J.,Thuiller, W., 2013. Rare species support vulnerable functions in high-diversityecosystems. PLoS Biol. 11, e1001569. http://dx.doi.org/10.1371/journal.pbio.1001569.
Murcia, C., 1995. Edge effects in fragmented forests: implications for conservation.Trends Ecol. Evol. 10, 58–62.
Nascimento, D.M.C., Dominguez, M.J.L., Silva, S.B.de M., 2009. Mudanças naocupação econômica do litoral sul da Bahia: Os exemplos de Belmonte eCanavieiras, Bahia. Rev. Desenbahia 10, 7–28.
Nicotra, A., Chazdon, R., Iriarte, S., 1999. Spatial heterogeneity of light and woodyseedling regeneration in tropical wet forests. Ecology 80, 1908–1926.
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B.,Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2013. Vegan:Community Ecology Package. R package version 2.0-10 <http://CRAN.R-project.org/package=vegan>.
Oliveira, M.A., Grillo, A.S., Tabarelli, M., 2004. Forest edge in the Brazilian Atlanticforest: drastic changes in tree species assemblages. Oryx 38, 389–394. http://dx.doi.org/10.1017/S0030605304000754.
Pardini, R., Bueno, A.D.A., Gardner, T.A., Prado, P.I., Metzger, J.P., 2010. Beyond thefragmentation threshold hypothesis: regime shifts in biodiversity acrossfragmented landscapes. PLoS ONE 5, e13666. http://dx.doi.org/10.1371/journal.pone.0013666.
Pavoine, S., Ollier, S., Dufour, A.-B., 2005. Is the originality of a species measurable?Ecol. Lett. 8, 579–586. http://dx.doi.org/10.1111/j.1461-0248.2005.00752.x.
Poulin, B., Wright, S.J., Lefebvre, G., Calderón, O., 1999. Interspecific synchronyand asynchrony in the fruiting phenologies of congeneric bird-dispersedplants in Panama. J. Trop. Ecol. 15, 213–227. http://dx.doi.org/10.1017/S0266467499000760.
Poulsen, A., Balslev, H., 1991. Abundance and cover of ground herbs in anAmazonian rain forest. J. Veg. Sci. 2, 315–322.
R Core Team, 2013. R: A Language and Environment for Statistical Computing. RFoundation for Statistical Computing, Vienna, Austria <http://www.R-project.org/>.
Redding, D.W., Hartmann, K., Mimoto, A., Bokal, D., Devos, M., Mooers, A.Ø., 2008.Evolutionarily distinctive species often capture more phylogenetic diversitythan expected. J. Theor. Biol. 251, 606–615. http://dx.doi.org/10.1016/j.jtbi.2007.12.006.
Ribeiro, M.C., Metzger, J.P., Martensen, A.C., Ponzoni, F.J., Hirota, M.M., 2009. TheBrazilian Atlantic Forest: how much is left, and how is the remaining forestdistributed? Implications for conservation. Biol. Conserv. 142, 1141–1153.http://dx.doi.org/10.1016/j.biocon.2009.02.021.
Rigueira, D.M.G., Rocha, P.L.B., Mariano-Neto, E., 2013. Forest cover, extinctionthresholds and time lags in woody plants (Myrtaceae) in the Brazilian AtlanticForest: resources for conservation. Biodivers. Conserv. 22, 3141–3163. http://dx.doi.org/10.1007/s10531-013-0575-4.
Santos, B.A., Arroyo-Rodríguez, V., Moreno, C.E., Tabarelli, M., 2010. Edge-relatedloss of tree phylogenetic diversity in the severely fragmented Brazilian Atlanticforest. PLoS ONE 5, e12625. http://dx.doi.org/10.1371/journal.pone.0012625.
Please cite this article in press as: Andrade, E.R., et al. Effects of habitat loss onforest landscapes. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.fore
Santos, B.A., Tabarelli, M., Melo, F.P.L., Camargo, J.L.C., Andrade, A., Laurance, S.G.,Laurance, W.F., 2014. Phylogenetic impoverishment of amazonian treecommunities in an experimentally fragmented forest landscape. PLoS ONE 9,e113109. http://dx.doi.org/10.1371/journal.pone.0113109.
Schroth, G., Faria, D., Araujo, M., Bede, L., Bael, S.A., Cassano, C.R., Oliveira, L.C.,Delabie, J.H.C., 2011. Conservation in tropical landscape mosaics: the case of thecacao landscape of southern Bahia, Brazil. Biodivers. Conserv. 20, 1635–1654.http://dx.doi.org/10.1007/s10531-011-0052-x.
Sedio, B.E., Wright, S.J., Dick, C.W., 2012. Trait evolution and the coexistence of aspecies swarm in the tropical forest understorey. J. Ecol. 100, 1183–1193.http://dx.doi.org/10.1111/j.1365-2745.2012.01993.x.
Sedio, B.E., Paul, J.R., Taylor, C.M., Dick, C.W., 2013. Fine-scale niche structure ofNeotropical forests reflects a legacy of the Great American Biotic Interchange.Nat. Commun. 4, 2317. http://dx.doi.org/10.1038/ncomms3317.
Sloan, S., Jenkins, C.N., Joppa, L.N., Gaveau, D.L.A., Laurance, W.F., 2014. Remainingnatural vegetation in the global biodiversity hotspots. Biol. Conserv. 177, 12–24.http://dx.doi.org/10.1016/j.biocon.2014.05.027.
Snow, D.W., 1981. Tropical frugivorous birds and their food plants: a world survey.Biotropica 13, 1–14.
Swenson, N.G., 2011. The role of evolutionary processes in producing biodiversitypatterns, and the interrelationships between taxonomic, functional andphylogenetic biodiversity. Am. J. Bot. 98, 472–480. http://dx.doi.org/10.3732/ajb.1000289.
Tabarelli, M., Mantovani, W., Peres, C., 1999. Effects of habitat fragmentation onplant guild structure in the montane Atlantic forest of southeastern Brazil. Biol.Conserv. 91, 119–127.
Tabarelli, M., Pinto, L., Silva, J., Hirota, M., Bedê, L., 2005. Challenges andopportunities for biodiversity conservation in the Brazilian Atlantic Forest.Conserv. Biol. 19, 695–700.
Thomas W.W., 2003. Natural vegetation types in southern Bahia. In: Prado P.I.,Landau E.C., Moura R.T., Pinto L.P.S., Fonseca G.A.B., Alger K. (orgs.), Corredor deBiodiversidade da Mata Atlântica do Sul da Bahia. Publicação em CD-ROM,Ilhéus, IESB/CI/CABS/UFMG/UNICAMP.
Thomas, W.M.W., Carvalho, A., Amorim, A., Garrison, J., Arbaláez, A., 1998. Plantendemism in two forests in southern Bahia, Brazil. Biodivers. Conserv. 7, 311–322.
Tóthmérész, B., 1995. Comparison of different methods for diversity ordering. J. Veg.Sci. 6, 283–290.
Vamosi, J.C., Wilson, J.R.U., 2008. Nonrandom extinction leads to elevated loss ofangiosperm evolutionary history. Ecol. Lett. 11, 1047–1053. http://dx.doi.org/10.1111/j.1461-0248.2008.01215.x.
Vamosi, S.M., Heard, S.B., Vamosi, J.C., Webb, C.O., 2009. Emerging patterns in thecomparative analysis of phylogenetic community structure. Mol. Ecol. 18, 572–592. http://dx.doi.org/10.1111/j.1365-294X.2008.04001.x.
Verdú, M., Pausas, J.G., 2007. Fire drives phylogenetic clustering in mediterraneanbasin woody plant communities. J. Ecol. 95, 1316–1323. http://dx.doi.org/10.1111/j.1365-2745.2007.01300.x.
Webb, C.O., Ackerly, D.D., Kembel, S.W., 2008. Phylocom: software for the analysisof phylogenetic community structure and trait evolution. Bioinformatics 24,2098–2100. http://dx.doi.org/10.1093/bioinformatics/btn358.
Winter, M., Devictor, V., Schweiger, O., 2013. Phylogenetic diversity and natureconservation: where are we? Trends Ecol. Evol. 28, 199–204. http://dx.doi.org/10.1016/j.tree.2012.10.015.
Wood, S.N., 2006. Generalized Additive Models: An Introduction with R. Chapmanand Hall/CRC.
taxonomic and phylogenetic diversity of understory Rubiaceae in Atlanticco.2015.03.049
