Contributed Paper

A Phylogenetic Approach to Conserving AmazonianBiodiversityFEDERICO LOPEZ-OSORIO AND DANIEL RAFAEL MIRANDA-ESQUIVEL∗

Laboratorio de Sistematica & Biogeografıa, Universidad Industrial de Santander, A. A. 678 Bucaramanga, Colombia

Abstract: Amazonia is a highly threatened rainforest that encompasses a major proportion of Earth’s

biological diversity. Our main goal was to establish conservation priorities for Amazonia’s areas of endemism

on the basis of measures of evolutionary distinctiveness. We considered two previously identified sets of areas

of endemism. The first set consisted of eight large areas used traditionally in biogeographical studies: Belem,

Tapajos, Xingu, Guiana, Rondonia, Imeri, Inambari, and Napo. The second set consisted of 16 smaller areas

that were subdivisions of the larger areas. We assembled a data set of 50 phylogenies that represented 16

orders and 1715 distributional records. We identified priority conservation areas for the areas of endemism

according to node-based metrics of evolutionary distinctiveness. We contrasted these results with priority

areas identified on the basis of raw species richness and species endemicity. For the larger areas, we identified

Guiana and Inambari as the first- and second-most important areas for conservation. The remaining areas

in this first group scored half (e.g., Napo) or less than Guiana and Inambari on all indices. For the smaller

areas, a subdivision of Guiana (i.e., Guyana and the Brazilian states of Roraima and Amazonas) was at the

top of the ranking and was followed by a subdivision of Inambari (i.e., northwestern portion of Amazonas)

and then another subdivision of Guiana (i.e., Suriname, French Guiana, and the Brazilian state of Amapa).

The distinctiveness-based rankings of the priority of areas correlated directly with those derived from species

richness and species endemicity. Current conservation strategies in Amazonia, although they rely on many

other criteria apart from phylogeny, are focusing on the most important areas for conservation we identified

here.

Keywords: amazonia, biogeography, evolutionary distinctiveness, endemism, prioritization ranking

Un Enfoque Filogenetico para Conservar la Biodiversidad Amazonica

Resumen: La Amazonıa es un bosque lluvioso muy amenazado y que comprende una porcion significativa

de la diversidad biologica de la Tierra. Nuestra meta principal fue establecer prioridades de conservacion para

las areas de endemismo de la Amazonıa con base en medidas de la individualidad evolutiva. Consideramos

dos conjuntos de areas de endemismo identificadas previamente. El primer conjunto consistio de ocho areas

extensas utilizadas tradicionalmente en estudios biogeograficos: Belem, Tapajos, Xingu, Guiana, Rondonia,

Imeri, Inambari y Napo. El segundo consistio de 16 areas mas pequenas que fueron subdivisiones de las

areas mas extensas. Configuramos un conjunto de datos de 50 filogenias que representaron 16 ordenes

y 1715 registros de distribucion. Identificamos areas de prioridad para la conservacion para las areas

de endemismo de acuerdo medidas de la individualidad evolutiva basadas en nodos. Contrastamos estos

resultados con areas prioritarias identificadas con base en la riqueza y endemismo de especies. Para las

areas mas extensas, identificamos a Guiana e Inambari como la primera y segunda area mas importante

para conservacion. Las demas areas en este primer grupo tuvieron valores de la mitad (Napo) o menos que

los de Guiana e Inambari en todos los ındices. Para las areas menores, una subdivision Guiana (Guyana

y los estados brasilenos de Roraima y Amazonas) estuvo arriba en la clasificacion y fue seguida por una

subdivision de Inambari (porcion noroeste de Amazonas) y luego otra subdivision de Guiana (Suriname,

∗Address correspondence to D.R. Miranda-Esquivel, email dmiranda@uis.edu.coPaper submitted June 8, 2009; revised manuscript accepted November 16, 2009

1359Conservation Biology, Volume 24, No. 5, 1359–1366C©2010 Society for Conservation BiologyDOI: 10.1111/j.1523-1739.2010.01482.x

1360 Phylogenetic Prioritization in Amazonia

Guyana Francesa y el estado brasileno de Amapa). Las clasificaciones de la prioridad de areas basadas en

individualidad se correlacionaron directamente con otras derivadas de la riqueza y endemismo de especies.

Las actuales estrategias de conservacion de la Amazonıa, aunque dependen de muchos otros criterios ademas

de la filogenia, estan enfocadas en las areas mas importantes para la conservacion que identificamos en este

trabajo.

Palabras Clave: amazonıa, biogeografia, clasificacion de prioridades, endemismo, individualidad evolutiva

Introduction

Amazonia is a large Tropical rainforest that harbors ahighly diverse biota distributed from the Northern Andesin the west to the northern Brazil Atlantic coast in the eastand to the north and south onto the Guianan and Brazil-ian shields, respectively (Haffer 1969). The numbers ofspecies inhabiting Amazonia are at least 1 million insects,40,000 plants, 3,000 fishes, 1,294 birds, 427 mammals,427 amphibians, and 378 reptiles (WWF 2003; Silva et al.2005a). Amazonian species are not randomly distributed.Birds (Haffer 1978; Cracraft 1985; Bates et al. 1998; Silvaet al. 2005a), mammals (Silva & Oren 1996), frogs (Ron2000), butterflies (Brown 1979), and other taxa have sim-ilar distributional limits (i.e., areas of endemism) (Kresset al. 1998; Morrone 2006).

Species diversity and patterns of endemism in Amazo-nia have been severely affected by rampant deforestation(Laurance et al. 2001, 2004; Fearnside 2005). From 1990to 2003 the rate of forest clearance ranged from 1.1 to2.9 million ha/year; the highest peak was observed in1995 (Laurance 1998; Laurance et al. 2004). Currently,the rate of deforestation is roughly 1.1 million ha/year(INPE 2007). Predominantly, the causes of deforestationin Amazonia have been large-scale development projects,industrial logging, road networks, extensive cattle ranch-ing, high rates of human population growth, fires, andbiofuel agriculture (Nepstad et al. 1999; Cochrane 2001;Laurance & Williamson 2001; Laurance et al. 2001; Fearn-side 2005; Bush & Lovejoy 2007). Altogether, the most im-portant consequences of deforestation in Amazonia havebeen substantial loss of biodiversity, habitat disruption,and global warming (Sioli 1985; Bush 1996; Laurance etal. 2001; Fearnside 2005). Thus, preserving Amazonia isparamount in conservation planning. We addressed thisissue by identifying areas that are of a high conservationpriority with a phylogenetic approach.

Conservation efforts are guided measurements of bio-diversity (Witting & Loeschcke 1995). Purvis and Hec-tor (2000) categorized commonly used measures of bio-diversity as follows into one of three categories: num-ber of species in a site or habitat, distributional even-ness of individuals among species, and estimates of dif-ference (e.g., molecular, morphological) among species.The basic problem with species richness and evennessis that they are driven by species that are common andwidespread and therefore less vulnerable (Brooks et al.

2006; Lamoreux et al. 2006). Using phylogenies as aframework, differences among species have been quan-tified for conservation purposes from two perspectives,evolutionary distinctiveness (ED) and phylogenetic diver-sity (PD).

Evolutionary distinctiveness refers to species-specificmeasures developed to assign scores of evolutionaryvalue on the basis of only the topology of a phylogenetictree (e.g., May 1990; Vane-Wright et al. 1991; Williams etal. 1991) or topology and branch lengths (e.g., Pavoineet al. 2005; Redding & Mooers 2006; Isaac et al. 2007).On the other hand, PD is generally measured for a subsetof species as the sum of the branch lengths of a phyloge-netic tree that spans that subset and the root (Faith 1992a,1992b; Faith & Baker 2006). Redding et al. (2008) identi-fied some of the major differences between ED and PD.Phylogenetic diversity is effective only if all the specieswithin the optimal subset are protected (a very resource-demanding task in some cases). Otherwise, new optimalsubsets are possible. Unlike ED, PD is not species spe-cific and thus does not offer priority species rankings,which are important to conservation approaches suchas the IUCN (International Union for the Conservationof Nature) Red List of Threatened Species. Furthermore,topology is more stable than branch lengths. Increasingthe number of characters or changing the set of char-acters seldom leads to entire shifts in the relationshipsamong species, whereas branch lengths change consid-erably from one set of characters to another and permitconclusions only about the evolution of a given set ofcharacters.

Although in recent years several advances inphylogeny-based conservation metrics have beenachieved (for reviews see Pavoine et al. 2005; Reddinget al. 2008; Schweiger et al. 2008), most of these mea-sures have neglected the role of biotic distribution andbiogeography in the problem of conserving biodiversity(see, however, Moritz & Faith 1998; Rodrigues & Gas-ton 2002; Alvarez-Mondragon & Morrone 2004; Forest etal. 2007). Vane-Wright et al. (1991; see also Williams etal. 1991) were probably the first to couple measures ofED with the biogeographic concept of complementarityof biotas. More recently, Posadas et al. (2001) modifiedVane-Wright et al.’s (1991) approach to consider not onlydistinctiveness and complementarity but also species en-demicity. Including distributional and biogeographicalinformation allows one to, for example, decrease the

Conservation Biology

Volume 24, No. 5, 2010

Lopez-Osorio & Miranda-Esquivel 1361

Figure 1. Eight large areas of

endemism (modified from Bates

et al. [1998] and Silva et al.

[2005a]) and 16 (dashed line)

interfluvial regions in Amazonia

(modified from Silva & Oren

1996). Four areas (Rondonia,

Tapajos, Xingu, and Belem) are

the same in both sets of areas.

conservation value of widespread species and maximizethe number of preserved species with as few areas aspossible (Posadas et al. 2001).

Previous conservation studies in Amazonia have fo-cused on identifying multiple taxonomic groups that havebeen adequately sampled (Heyer et al. 1999) or sites ofhigh species diversity and endemism (Nelson et al. 1990;Kress et al. 1998), and on ranking Amazonian areas ofendemism according to species richness, generic rich-ness, time-based PD measures, and species endemicity(Silva et al. 2005b). Other authors studying loss of Ama-zonian rainforests have used environmental data (Maslinet al. 2005; Malhi et al. 2008; Aragao et al. 2008). Here,our main goal was to establish conservation prioritiesfor Amazonia’s areas of endemism through the use ofVane-Wright et al. (1991) and Posadas et al.’s (2001) EDmeasures, raw species richness, species endemicity, andcomplementarity of biotas.

Methods

Areas of Endemism

Regardless of the scientific need to discover areas of en-demism directly from the data, many biogeographic andconservation studies proceed in the opposite direction:they contrast areas that are already defined and exam-ine the properties of those areas. To evaluate how areasize and geographic limits influence area prioritizationand complementarity, we considered two sets of areasof endemism. The first set encompassed eight large areasdescribed in Bates et al. (1998) and Silva et al. (2005a)(Fig. 1): Belem, Xingu, Tapajos, Rondonia, Guiana, Imeri,Inambari, and Napo. The second set consisted of 16smaller areas defined in Silva and Oren (1996) (Fig. 1). Weidentified this second set of areas with the first three let-

ters of the names of the larger areas followed by a number(e.g., Gui1, Gui2, and Gui3, where each is a subdivisionof Guiana).

Taxa

We selected 50 clades from the literature with speciesor “subspecies” as terminals (Supporting Information).Phylogenies of many Amazonian taxa are scarce; thus,some groups (e.g., plants and insects) were poorly repre-sented in our data set. If a study resulted in trees derivedfrom partitioned and combined data sets, we consideredonly the latter. We selected phylogenies according to thefollowing rules, which rely on the strict concordance be-tween monophyly and the area inhabited. In all cases,for a given tree, t was a subset of terminal nodes thatbelonged to the same species or subspecies. If t was afully resolved monophyletic group and all of its terminalsinhabited the same area of endemism, t was collapsed(i.e., reduced) into a single terminal node. For example,for the tree (a((b(bb))(c(cc)))), where a, b, and c arethree different species (or subspecies) present in areas1, 2, and 3 respectively, the tree was treated as (a(bc)).If t fulfilled the initial condition from this first rule butnot the latter (i.e., each of its terminals was present ina different area), it was not collapsed. Second, if t wasa monophyletic but fully unresolved group, it was col-lapsed into a single terminal node; therefore, if t waspresent in a single area the outcome was the same as forthe first rule, but if t was present in more than one areathe outcome was a widespread terminal node. Third, ift was not monophyletic it was not collapsed. After weapplied these rules, we omitted those trees with morethan one tetratomy or more than two trichotomies (San-martın et al. 2001). We collected distributional data fromthe original papers (locations and coordinates) and fromthe Global Biodiversity Information Facility (GBIF). We

Conservation Biology

Volume 24, No. 5, 2010

1362 Phylogenetic Prioritization in Amazonia

converted such data into decimal degrees (for loca-tions this was possible with the National Geospatial-Intelligence Agency GEONet Names Server). We trans-formed degrees into keyhole markup language (KML)and visualized the KML file in a three-dimensional Earthbrowser to determine presence or absence in a given areaof endemism. The total number of records was 1715.

Prioritization Ranking

Initially, we used raw species richness and species en-demicity to rank priority of areas. We measured endemic-ity as the quotient of endemic and total species. Subse-quently, we applied the area ranking protocol suggestedby Posadas et al. (2001, 2004), which is based on the W

index (Vane-Wright et al. 1991). For any fully bifurcatingtree, each branching results in two sister lineages thatare the same age and have undergone equivalent evolu-tionary change (Krell & Cranston 2004). Commonly, oneof the two lineages originating from a given branchinghas fewer species than the other (Crisp & Cook 2005).The W index considers taxa in such species-poor lineagesto be more unique and vulnerable to lineage extinctionand therefore of higher conservation value. A tree maybe unbalanced—that is, have unequal number of specieson either side of each node—because of misinformativecharacters or biased taxon sampling (Mooers et al. 1995;Stam 2002), among other causes. This node-based W in-dex, although not suitable to capture PD, reliably selectsthe most relictual species (Redding et al. 2008) and thusoffer a valuable approach to preserve biodiversity.

Posadas et al. (2001) proposed the following modifica-tions to the W index. To avoid bias due to different num-bers of taxa when summing the index values obtainedfrom different trees, the W index value of each taxon isscaled by the total, this is Ws index. To include the de-gree of endemicity of a taxon, the W index value of eachtaxon is divided among the number of areas where thetaxon is present, this yields the We index. The outcomeof combining these two modifications is the Wes index.We ranked Amazonian areas of endemism according tothe basic W index and its modifications outlined earlier(Fig. 2). To evaluate the relationship between distinctive-ness and species richness, we correlated the Wes indexto raw species richness.

Finally, we calculated complementarity (measured asthe complement of the Jaccard index of similarity) amongpairs of areas with the aim of evaluating distinctiveness inspecies composition (for a detailed review see Posadas etal. 2001). For all calculations, we used the software Rich-ness (Miranda-Esquivel 2001). These estimates of rich-ness, endemicity, and complementarity are restricted totaxa for which phylogenetic information are available.Thus, they may not have represented accurately the ac-tual features of Amazonia.

Figure 2. Area prioritization according to node-based

indices of evolutionary distinctiveness for three

hypothetical areas: left side, phylogeny of five terminal

taxa, each taxon’s basic distinctiveness weight (W),

total weight (T), and distribution of each taxon

(circles, presence) among three areas (A1–A3); right

side, aggregate weights for each of the three areas

(e.g., area A1 score of 3.3 after summing the basic

distinctiveness weights of taxa inhabiting that area).

To calculate the aggregate area values in column We,

the basic weight of each taxon must be divided among

the number of areas where it is present (e.g., We value

of area A1 is calculated as 1.3/3 + 1/3 + 1/1 = 1.8).

To calculate the area values in the Ws column, the

basic weight of each taxon must be divided by T.

Combining these two modifications of the basic Windex, We and Ws, allows for calculatation of the area

values in column Wes. Figure adapted from Vazquez

and Gittleman (1998).

Results

The most species-rich area was Inambari (103 species).Guiana and Napo were the second- and third-most speciesrich (95 and 76 species, respectively) (Table 1; Fig. 1).For species endemicity, raw W and its modifications, Ws,We, and Wes, the three most important areas were Guiana,Inambari, and Napo (Table 1); the next most importantarea in these rankings (Table 1) was not nearly as valuableas the first three. The Wes index correlated strongly toraw species richness (r = 0.965, p < 0.05). Taking intoaccount that Guiana and Inambari were the largest areas,we also evaluated the relationship between the Wes indexand area size (obtained from Silva et al. 2005a), and theresult was a positive correlation (r = 0.910, p < 0.05).

For the second set of areas, the three richest areaswere Gui2 (58 species), Ina4 (54), and Ron1 (51) (Fig. 1;Table 2). Sites Gui2 and Ron1 had the most endemicspecies (both 23), Gui3 and Tap1 had the second most(both 21), and Ina4 had the third most (19). Raw W in-dex ranked Gui2, Ron1, and Nap1 as top priorities (Fig.1; Table 2). When endemicity was accounted for (We in-dex) rankings changed to Ron1, Gui2, and Gui3 at first,second, and third, respectively. The rankings of Ws andWes measures encompassed the same priority areas. ForWs the priority order was Gui2, Gui3, and Ina2 at first,second, and third, respectively, and for Wes the order

Conservation Biology

Volume 24, No. 5, 2010

Lopez-Osorio & Miranda-Esquivel 1363

Table 1. For the set of eight large areas, raw (W) and standardized (Ws) distinctiveness values, including species endemicity (We and Wes), andspecies richness and endemicity for each area.

Area W We Ws Wes Richness Endemicity

Belem 29.06 12.91 4.08 1.67 21 5Guiana 168.87 114.76 24.57 16.16 95 53Tapajos 67.26 43.50 8.63 4.85 40 22Xingu 38.78 22.12 4.97 2.75 23 9Rondonia 94.05 59.39 7.93 4.28 50 22Inambari 166.84 97.43 23.70 14.79 103 48Imeri 53.31 19.38 5.80 2.71 33 8Napo 130.75 65.52 16.69 8.58 76 28

was Gui2, Ina2, and Gui3, respectively. The Wes and rich-ness (r = 0.803, p < 0.05) were also correlated strongly.For this set of smaller areas, we also evaluated the rela-tionship between Wes and area size, and they were notcorrelated (r = 0.3920, p > 0.05).

For the set of larger areas, comparisons of Inambariand Napo with Xingu yielded the two highest comple-mentarity values, whereas Napo and Inambari were thepair of areas with most species in common (Table 3).For the set of smaller areas, the highest value possiblefor complementarity was the product of various compar-isons with Xin1, and the pair of areas with more sharedspecies were Nap1 and Nap2 (Table 4).

Discussion

Regardless of the set of areas of endemism used, the dis-tinctiveness indices identified partially overlapping pri-ority areas. For example, the Wes rankings identified thetwo areas of highest priority for the set of smaller areasas Gui2 and Ina2, both of which are within the two mostimportant larger areas, Guiana and Inambari. The find-ing that the ED measures we considered seemed partly

resilient to changes in area size is particularly importantfor conservation planning because smaller areas may bemore feasible to conserve, whether from a conceptual orfinancial perspective.

For the set of larger areas, the raw species richness,endemicity, and distinctiveness rankings remained com-pletely stable. Guiana, Inambari, and Napo were alwaysthe three most valuable areas for conservation. So, distinc-tiveness reflected raw species richness and endemicity.

For the set of smaller areas, the distinctiveness indicesyielded unstable rankings. The majority of measures, withthe exception of We, ranked Gui2 as that of highest prior-ity. Richness and endemicity rankings were in agreementwith distinctiveness for the most important area for con-servation, Gui2, but in disagreement for the other areas.

Once one has identified the area of highest priority onthe basis of any attribute—in our case, ED—the secondarea should be selected according to complementarity tomaximize the number of protected species (Vane-Wrightet al. 1991; Posadas et al. 2001). For the set of larger ar-eas, the second-best priority area on the basis of distinc-tiveness was Inambari, but when we considered comple-mentarity (Table 3) the second-best area to conserve wasXingu, which had fewer species in common with Guiana.

Table 2. For the set of 16 smaller areas, raw (W) and standardized (Ws) distinctiveness values, including endemicity (We and Wes) and speciesrichness and endemicity for each area.

Area W We Ws Wes Richness Endemicity

NAP1 90.65 35.11 10.85 4.81 50 10NAP2 88.71 33.05 10.46 3.58 46 12NAP3 30.64 5.16 5.01 0.93 16 0IME1 48.37 15.86 4.80 2.02 30 7IME2 9.32 2.79 0.86 0.33 6 1GUI1 62.27 27.12 6.37 2.36 35 10GUI2 105.64 54.63 18.73 8.96 58 23GUI3 84.03 46.51 14.90 7.10 48 21INA1 36.07 14.24 4.02 1.17 26 6INA2 67.22 38.42 11.68 7.92 40 17INA3 25.45 13.49 4.79 3.31 19 7INA4 84.48 39.21 11.60 5.41 54 19RON1 95.05 59.72 8.14 4.30 51 23TAP1 67.28 42.52 8.71 4.68 40 21XIN1 43.13 23.69 5.56 3.00 24 10BEL1 27.72 10.58 4.14 1.17 20 4

Conservation Biology

Volume 24, No. 5, 2010

1364 Phylogenetic Prioritization in Amazonia

Table 3. For the eight large areas, pairwise complementarity values ranging from 0 (identical species composition) to 1 (no shared or commonspecies).

Belem Guiana Tapajos Xingu Rondonia Inambari Imeri

Guiana 0.916Tapajos 0.849 0.937Xingu 0.952 0.965 0.875Rondonia 0.908 0.957 0.941 0.957Inambari 0.949 0.894 0.956 0.984 0.891Imeri 0.941 0.877 0.942 0.963 0.908 0.876Napo 0.922 0.890 0.964 0.990 0.895 0.807 0.853

We calculated complementarity between a combinationof Guiana and Xingu and the other areas, and Rondoniaranked third. Likewise, for the set of smaller areas (Table4), the best area in the majority of distinctiveness rank-ings, Gui2, was followed by Ime2 on the basis of comple-mentarity, and Xin1 was the area with fewer species incommon with Gui2 and Ime2 combined.

Large reserves (at least 1 million ha) are imperativefor the future of Amazonian biodiversity (Peres 2005;Laurance 2005). Thus, a drawback of our study is thatconservation action does not happen at the spatial scaleswe considered; however, our results should be consid-ered a frame of reference for additional studies of smallergeographical units (e.g., grid-cell-based studies focusingon those areas of endemism we identified here as conser-vation priorities). Such studies, which may yield more-attainable priority areas, currently would be even moredifficult than the approach we took here because ex-haustive surveys of the Amazonian biota, high-quality dis-tributional data, and, if based on evolutionary informa-tion, well-resolved and complete phylogenies would allbe necessary (Mace et al. 2003).

Nevertheless, grid-cell-based studies to set conserva-tion priorities in Amazonia have been conducted. Kresset al. (1998) divided Amazonia into 472 1◦ × 1◦ grid cells

and measured species diversity and species endemism onthe basis of 3991 records that represented various taxo-nomic groups. Kress et al. (1998) included nine of thosecells in the highest category of species diversity (43–67species) and nine in the highest category of concentra-tions of endemism (4–13 endemic species). The majorityof these cells are located within the three most importantareas for conservation according to our results: Guiana,Inambari, and Napo. Kress et al. (1998) emphasized thatgrids with the highest species richness in their study alsocorrespond to many of the most intensively collected ar-eas, and few areas in Amazonia have been thoroughlysampled for most groups (Heyer et al. 1999). Therefore,our results, as in other biodiversity studies of Amazonia,could have been influenced by bias in sampling efforts.

In comparison with the results of previous case studiesin which phylogenetic information was used, our resultswere consistent with those of Silva et al. (2005b), whoapplied two measures devised by Altschul and Lipman(1990) and Faith (1992a, 1992b) to a phylogeny of allNew World primate genera (Schneider 2000). The firstmeasure was clade evolutionary history, a measure equiv-alent to PD calculated with time units, and the secondmeasure was species evolutionary history, measured asthe time from present to the time of last divergence.

Table 4. For the 16 smaller areas, pairwise complementarity values ranging from 0 (identical species composition) to 1 (no shared or commonspecies).

NAP2 NAP3 IME1 GUI2 GUI3 NAP1 INA1 INA2 INA3 INA4 RON1 TAP1 XIN1 BEL1 IME2

NAP3 0.808IME1 0.866 0.757GUI2 0.917 0.928 0.886GUI3 0.944 0.933 0.917 0.767NAP1 0.685 0.842 0.904 0.920 0.946INA1 0.839 0.865 0.880 0.937 0.943 0.866INA2 0.897 0.923 0.906 0.957 0.989 0.902 0.862INA3 0.934 0.971 0.886 0.900 0.953 0.970 0.929 0.843INA4 0.864 0.906 0.880 0.913 0.915 0.818 0.904 0.881 0.942RON1 0.934 0.953 0.905 0.962 0.979 0.902 0.915 0.917 0.971 0.859TAP1 0.951 0.963 0.939 0.957 0.900 0.966 0.952 0.961 0.983 0.932 0.942XIN1 1.000 0.974 0.962 0.962 0.971 0.986 0.980 0.984 1.000 1.000 0.973 0.897BEL1 0.918 0.941 0.958 0.901 0.867 0.939 0.930 0.983 0.917 0.928 0.924 0.846 0.977IME2 0.960 0.842 0.941 0.968 0.962 0.963 0.968 0.955 0.958 0.966 0.982 0.978 1.000 0.960GUI1 0.920 0.891 0.818 0.792 0.831 0.924 0.983 0.944 0.920 0.940 0.964 0.958 0.946 0.900 0.921

Conservation Biology

Volume 24, No. 5, 2010

Lopez-Osorio & Miranda-Esquivel 1365

Silva et al. (2005b) ranked Inambari and Napo as theareas of highest priority with equal values for both mea-sures, and these areas were followed in importance byGuiana and Rondonia. The fact that ED and PD identifiedthe same priority areas suggests that such areas are ade-quately prioritized from two different evolutionary per-spectives and measures. Future phylogeny-based conser-vation studies of Amazonia, however, could evaluate thestability of these results by conducting several analyseswith one major taxonomic group at a time or by creatingrandom subsamples from the total number of phylogeniesassembled.

In Amazonia, two groups of areas, discernible in termsof protected coverage (Silva et al. 2005b), are of con-servation concern. The first group includes Napo, Imeri,and Guiana, which all have more than 40% of their landsacknowledged as protected, and the second group in-cludes Inambari, Rondonia, Tapajos, and Xingu, whichall with 20–40% of their lands protected. In addition,the Brazilian portions of Napo, Imeri, Guiana, and In-ambari have the lowest percentages of area deforestedand could be considered the least vulnerable (Silva et al.2005a). Therefore, despite being based on criteria otherthan phylogeny, current conservation efforts in Amazo-nia are focused on the areas that we identified as the mostimportant areas for conservation.

The Amazonian rainforest is heavily threatened andhence constitutes an obvious target for conservation stud-ies. Study approaches that include evolutionary and bio-geographic information provide an effective frameworkwithin which to set conservation priorities. Guiana, Inam-bari, and Napo are the most important areas for conserva-tion according to node-counting measures of distinctive-ness. Nevertheless, conservation planning is an intricateendeavor, and no particular approach can be considereddefinite because many factors must be considered whendesigning reserves (e.g., spatial arrangement of areas,area size, dispersal capability, environmental variation,and disturbance regimes). Therefore, Amazonia requiresresearch from different disciplines and would highly ben-efit from additional evolutionary, taxonomic, ecological,and environmental studies.

Acknowledgments

J. Arias, I. Garzon-Orduna, J. Wenzel, E. Parada-Vargas,A. Osorio-Guerrero, and three anonymous reviewers pro-vided valuable comments, all of whom helped to greatlyimprove the final version of the manuscript. Nonethe-less, any errors of commission or omission in this paperare our responsibility entirely. This project was fundedby the Division de Investigaciones y Extension, Facultadde Ciencias, Universidad Industrial de Santander (Project5132).

Supporting Information

An overview of Amazonian taxa and case studies fromwhich we assembled 50 phylogenies are available as partof the online article (Appendix S1). The authors are re-sponsible for the content and functionality of these mate-rials. Queries (other than absence of the material) shouldbe directed to the corresponding author.

Literature Cited

Altschul, S. F., and D. J. Lipman. 1990. Equal animals. Nature348:493–494.

Alvarez-Mondragon, E., and J. J. Morrone. 2004. Propuesta de areaspara la conservacion de aves de Mexico, empleando herramien-tas panbiogeograficas e ındices de complementariedad. Interciencia29:112–120.

Aragao, L. E. O. C., Y. Malhi, N. Barbier, A. Lima, Y. Shimabukuro,L. Anderson, and S. Saatchi. 2008. Interactions between rainfall,deforestation and fires during recent years in the Brazilian Ama-zonia. Philosophical Transactions of the Royal Society 363:1779–1785.

Bates, J. M., S. J. Hackett, and J. Cracraft. 1998. Area-relationships in theNeotropical lowlands: an hypothesis based on raw distributions ofPasserine birds. Journal of Biogeography 25:783–793.

Brooks, T. M., R. A. Mittermeier, G. A. B. da Fonseca, J. Gerlach, M.Hoffmann, J. F. Lamoreux, C. G. Mittermeier, J. D. Pilgrim, and A.S. L. Rodrigues. 2006. Global biodiversity conservation priorities.Science 313:58–61.

Brown, K. S. 1979. Ecologia geografia e evolucao nas florestas Neotrop-icais. Universidade Estadual de Campinas, Campinas, Brasil.

Bush, M. B. 1996. Amazonian conservation in a changing world. Biolog-ical Conservation 76:219–228.

Bush, M. B., and T. E. Lovejoy. 2007. Amazonian conservation: pushingthe limits of biogeographical knowledge. Journal of Biogeography34:1291–1293.

Cochrane, M. A. 2001. Synergistic interactions between habitat frag-mentation and fire in evergreen tropical forests. Conservation Biol-ogy 15:1515–1521.

Cracraft, J. 1985. Historical biogeography and patterns of differentiationwithin the South American avifauna: areas of endemism. Ornitho-logical Monographs 36:49–84.

Crisp, M. D., and L. G. Cook. 2005. Do early branching lineages signifyancestral traits? Trends in Ecology & Evolution 20:122–128.

Faith, D. P. 1992a. Conservation evaluation and phylogenetic diversity.Biological Conservation 61:1–10.

Faith, D. P. 1992b. Systematics and conservation: on predicting thefeature diversity of subsets of taxa. Cladistics 8:361–373.

Faith, D. P., and A. M. Baker. 2006. Phylogenetic diversity (PD) andbiodiversity conservation: some bioinformatics challenges. Evolu-tionary Bioinformatics Online PMCID: PMC2674674).

Fearnside, P. M. 2005. Deforestation in Brazilian Amazonia: history,rates, and consequences. Conservation Biology 19:680–688.

Haffer, J. 1969. Speciation in Amazonian forest birds. Science165:131–137.

Haffer, J. 1978. Distribution of Amazon birds. Bonner ZoologischenBeitrage 29:38–78.

Heyer, W. R., et al. 1999. Amazonian biotic data and conservation deci-sions. Environment and Biodiversity 51:372–385.

INPE (Instituto Nacional de Pesquisas Espaciais). 2007. Monitoramentoda floresta Amazonica Brasilera por satelite 2006–2007. INPE, SaoPaulo, Brazil. Available from www.inpe.br (accessed August 30,2008).

Isaac, N. J. B., S. T. Turvey, B. Collen, C. Waterman, and J. E.M. Baillie. 2007. Mammals on the EDGE: conservation priorities

Conservation Biology

Volume 24, No. 5, 2010

1366 Phylogenetic Prioritization in Amazonia

based on threat and phylogeny. Public Library of Science ONE.DOI:10.1371/journal.pone.0000296

Krell, F. T., and P. S. Cranston. 2004. Which side of the tree is morebasal? Systematic Entomology 29:279–281.

Kress, W. J., et al. 1998. Amazonian biodiversity: assessing conserva-tion priorities with taxonomic data. Biodiversity and Conservation7:1577–1587.

Lamoreux, J. F., J. C. Morrison, T. H. Ricketts, D. M. Olson, E. Din-erstein, M. W. McKnight, and H. H. Shugart. 2006. Global tests ofbiodiversity concordance and the importance of endemism. Nature440:212–214.

Laurance, W. F. 1998. A crisis in the making: responses of Amazonianforests to land use and climate change. Trends in Ecology & Evolu-tion 13:411–415.

Laurance, W. F. 2005. When bigger is better: the need for Ama-zonian mega-reserves. Trends in Ecology & Evolution 20:645–648.

Laurance, W. F., and G. B. Williamson. 2001. Positive feedbacks amongforest fragmentation, drought, and climate change in the Amazon.Conservation Biology 15:1529–1535.

Laurance, W. F., M. A. Cochrane, S. Bergen, P. M. Fearnside, P. Delam-onica, C. Barber, S. D’Angelo, and T. Fernandes. 2001. The futureof the Brazilian Amazon. Science 291:438–439.

Laurance, W. F., A. K. M. Albernaz, P. M. Fearnside, H. L. Vasconce-los, and L. V. Ferreira. 2004. Deforestation in Amazonia. Science304:1109.

Mace, G. M., J. L. Gittleman, and A. Purvis. 2003. Preserving the tree oflife. Science 300:1707–1709.

Malhi, Y., J. T. Roberts, R. A. Betts, T. J. Killeen, W. Li, and C. A. Nobre.2008. Climate change, deforestation, and the fate of the Amazon.Science 319:169–172.

Maslin, M., Y. Malhi, O. Phillips, and S. Cowling. 2005. New views on anold forest: assessing the longevity, resilience and future of the Ama-zon rainforest. Transactions of the Institute of British Geographers30:477–499.

May, R. M. 1990. Taxonomy as destiny. Nature 347:129–130.Miranda-Esquivel, D. R. 2001. Richness. http://tux.uis.edu.co/labsist/

software/richness.Mooers, A. O., R. D. M. Page, A. Purvis, and P. H. Harvey. 1995. Phy-

logenetic noise leads to unbalanced cladistic tree reconstructions.Systematic Biology 44:332–342.

Moritz, C., and D. Faith. 1998. Comparative phylogeography and iden-tification of genetically divergent areas for conservation. MolecularEcology 7:419–429.

Morrone, J. J. 2006. Biogeographic areas and transition zones of LatinAmerica and the Caribbean Islands based on panbiogeographic andcladistic analyses of the entomofauna. Annual Review of Entomology51:467–494.

Nelson, B. W., C. A. C. Ferreira, M. F. da Silva, and M. L. Kawasaki.1990. Endemism centres, refugia and botanical collection density inBrazilian Amazonia. Nature 345:714–716.

Nepstad, D. C., et al. 1999. Large-scale impoverishment of Amazonianforests by logging and fire. Nature 398:505–508.

Pavoine, S., S. Ollier, and A.-B. Dufour. 2005. Is the originality of aspecies measurable? Ecology Letters 8:579–586.

Peres, C. A. 2005. Why we need megareserves in amazonia. Conserva-tion Biology 19:728–733.

Posadas, P., D. R. Miranda-Esquivel, and J. V. Crisci. 2001. Using phylo-genetic diversity measures to set priorities in conservation: an exam-ple from southern South America. Conservation Biology 15:1325–1334.

Posadas, P., D. R. Miranda-Esquivel, and J. V. Crisci. 2004. On words,tests, and applications: reply to Faith et al. Conservation Biology18:1–5.

Purvis, A., and A. Hector. 2000. Getting the measure of biodiversity.Nature 405:212–219.

Redding, D. W., K. Hartmann, A. Mimoto, D. Bokal, M. DeVos, and A.O. Mooers. 2008. Evolutionarily distinctive species often capturemore phylogenetic diversity than expected. Journal of TheoreticalBiology 251:606–615.

Redding, D. W., and A. O. Mooers. 2006. Incorporating evolution-ary measures into conservation prioritization. Conservation Biology20:1670–1678.

Rodrigues, A. S. L., and Gaston, K. J. 2002. Maximising phylogenetic di-versity in the selection of networks of conservation areas. BiologicalConservation 105:103–111.

Ron, S. R. 2000. Biogeographic area relationships of lowland neotropicalrainforest based on raw distributions of vertebrate groups. BiologicalJournal of the Linnean Society 71:379–402.

Sanmartın, I., H. Enghoff, and F. Ronquist. 2001. Patterns of animaldispersal, vicariance and diversification in the Holarctic. BiologicalJournal of the Linnean Society 73:345–390.

Schneider, H. 2000. The current status of the New World monkeyphylogeny. Anais da Academia Brasileira de Ciencias 72:165–172.

Schweiger, O., S. Klotz, W. Durka, and I. Kuhn. 2008. A comparativetest of phylogenetic diversity indices. Oecologia 157:485–495.

Silva, J. M. C., and D. C. Oren. 1996. Application of parsimony analysis ofendemicity in amazonian biogeography: an example with primates.Biological Journal of the Linnean Society 59:427–437.

Silva, J. M. C., A. B. Rylands, and G. A. B. da Fonseca. 2005a. Thefate of the Amazonian areas of endemism. Conservation Biology19:689–694.

Silva, J. M. C., A. B. Rylands, J. S. S. Junior, C. Gascon, and G. A. B. daFonseca. 2005b. Primate diversity patterns and their conservationin Amazonia. Pages 337–364 in A. Purvis, J. L. Gittleman, and T.Brooks, editors. Phylogeny and conservation. Cambridge UniversityPress, Cambridge, United Kingdom.

Sioli, H. 1985. The effects of deforestation in Amazonia. The Geograph-ical Journal 151:197–203.

Stam, E. 2002. Does imbalance in phylogenies reflect only bias? Evolu-tion 56:1292–1295.

Vane-Wright, R. I., C. J. Humphries, and P. H. Williams. 1991. What toprotect? systematics and the agony of choice. Biological Conserva-tion 55:235–254.

Vazquez, D. P., and J. L. Gittleman. 1998. Biodiversity conservation:does phylogeny matter? Current Biology 8:R379–R381.

Williams, P. H., C. J. Humphries, and R. I. Vane-Wright. 1991. Measur-ing biodiversity: taxonomic relatedness for conservation priorities.Australian Systematic Botany 4:665–679.

Witting, L., and V. Loeschcke. 1995. The optimization of biodiversityconservation. Biological Conservation 71:205–207.

WWF (World Wildlife Fund). 2003. Safeguarding the Amazon: makingconservation history. World Wildlife Fund, Washington, D.C. Avail-able from www.worldwildlife.org (accessed June 15, 2008).

Conservation Biology

Volume 24, No. 5, 2010