Population Structure, Diversity, and Phylogeography in the Near

bij_1099

Biological Journal of the Linnean Society, 2008, ••, ••–••. With 3 ﬁgures

1 Population structure, diversity, and phylogeography in 2 the near-threatened Eurasian black vultures Aegypius 3 monachus (Falconiformes; Accipitridae) in Europe: 4 insights from microsatellite and mitochondrial

5 DNA variation 22 6 1,2 3 1 1 4 7 N. POULAKAKIS *†, A. ANTONIOU †, G. MANTZIOU , A. PARMAKELIS , T. SKARTSI , 33 8 D. VASILAKIS4, J. ELORRIAGA4, J. DE LA PUENTE5, A. GAVASHELISHVILI6, 9 M. GHASABYAN7, T. KATZNER8, M. MCGRADY8,N.BATBAYAR9, M. FULLER10 and 10 T. NATSAGDORJ9 11 1 12 Natural History Museum of Crete, University of Crete, Heraklion, Greece 44 13 2Yale Institute for Biospheric Studies, Yale University, Newhaven, CT, USA 14 3Department of Genetics and Molecular Biotechnology, Hellenic Centre for Marine Research, 15 Heraklion, Crete, Greece 16 4WWF Greece-Dadia Project, Daia, Souﬂi, Greece 17 5Área de Estudio y Seguimiento de Aves, SEO/BirdLife Melquiades Biencinto, Madrid, Spain 18 6Georgian Center for the Conservation of Wildlife, Tbilisi, The Republic of Georgia 19 7Armenian Society for the Protection of Birds, Yerevan, Armenia 20 8Natural Research, Ltd, Krems, Austria 21 9The Peregrine Fund, World Center for Birds of Prey, Boise, ID, USA 22 10USGS, Forest and Rangeland Ecosystem Science Center, Snake River Field Station, and Boise 23 State University, Boise, ID, USA 24 25 Received 18 February 2008; accepted for publication 21 April 2008

27 The Eurasian black vulture (Aegypius monachus) has experienced a severe decline during the last two centuries 28 and is globally classified as near-threatened. This has led to the extinction of many traditional breeding areas in 29 Europe and resulted in the present patchy distribution (Iberian and Balkan peninsulas) in the Western Palearctic. 30 In the present study, we describe the current genetic status of the European populations using both mitochondrial 31 cytochrome b sequences and nuclear microsatellite markers, comparing with those found in Asia (Mongolia and 32 Caucasus region). Although, mitochondrial (mt)DNA revealed a relatively low genetic variability (haplotype 33 diversity), no evidence of genome-wide genetic erosion exists because nuclear diversity exhibits normal levels and 34 strong differentiation. A highly philopatric dispersal behaviour must be invoked to explain the existence of a clear 35 pattern that revealed by the phylogeographic analysis, which indicates a sharp East–West clinal distribution and 36 an allopatric differentiation. The distribution of mtDNA haplotypes one in the Iberian population and two in 37 Balkan population and the significance divergence at nuclear loci fulfill the definitions of those populations as 38 evolutionary significant units. We discuss how management strategies should aim at the maintenance (or increase) 39 of current genetic variability levels, suggesting that independent conservation plans are urgently required to 40 protect these two breeding European populations from extinction. © 2008 The Linnean Society of London, 41 Biological Journal of the Linnean Society, 2008, ••, ••–••.

42 ADDITIONAL KEYWORDS: conservation genetics – nuclear DNA. 55 43 44 45 *Corresponding author. E-mail: 46 [email protected] 47 †These authors contributed equally to this work.

2 N. POULAKAKIS ET AL.

1 Figure 1. A map of Eurasia showing the location of the sampling areas, the number of samples from each locality, and 2 the present distribution of the species Aegypius monachus. 3 4 INTRODUCTION ing in the present patchy distribution of breeding 39 5 nuclei in Europe (Fig. 1). Extant European popula- 40 6 The Eurasian black vulture (Aegypius monachus)is tions are confined to Spain, the Balkans (Rhodope 41 7 the largest bird of prey (Falconiformes) in the world Mountain), and the Caucasus mountains (Russia, 42 8 (i.e. the unrelated, slightly larger Andean Condor Georgia, Armenia, and Azerbaijan). It is worth noting 43 9 is now affiliated with the Ciconiiformes). By being that there is no evidence (field observations) of 44 10 mainly a scavenger or carrion eater bird with a long contact between Balkan and Iberian populations sub- 45 11 life-span and an extended home-range, vast areas of sequent to the extinction of the intermediate popula- 46 12 suitable habitats are considered indispensable for the tion(s) in 19th Century. Much less information is 47 13 species’ well being (Cramp & Simmons, 1980; Carrete available regarding the status and population trends 48 14 & Donazar, 2005). Its global range extends from of the species in Asia, where the bulk of the global 49 15 the Iberian Peninsula across southern Europe and population resides. It appears that breeding popula- 50 16 through the central Asian plateau to Mongolia and tions are more or less stable in Mongolia (where the 51 17 China. species is described as common) and Pakistan (where 52 18 Listed in Appendix I of the CITES (Convention on it is described as scarce), although fluctuations in 53 19 International Trade of Endangered Species of Wild distribution and breeding success occur. In Kazakh- 54 20 Fauna and Flora) and considered near-threatened stan, however, populations of all vulture species are 55 21 by the IUCN (2006 IUCN Red List of Threatened in severe decline. This trend may be mirrored in 56 22 Species) the Eurasian black vulture has attracted a number of other central Asian countries where 57 23 some conservation attention. It is classified as vulner- populations of both domesticated livestock and wild 58 24 able at European level (BirdLife International, 2006) ungulates have declined greatly in recent years 59 25 and endangered in Greece (Handrinos, 1992). Despite (BirdLife International, 2006). 60 26 the fact that in some regions (e.g. Spain) its popula- Faced with the growing challenge of deriving strat- 61 27 tions are now recovering (Sánchez, 2004), the species egies for salvaging diminishing flora and fauna, con- 62 28 has suffered a serious demographic decline during the servation biologists and ecologists continue to search 63 29 last century, mainly as a result of human pressure for methods that can distinguish unambiguous units 64 30 (i.e. the degradation and destruction of its breeding for conservation purposes (Fraser & Bernatchez, 65 31 habitats, direct persecution and poisoning, the aban- 2001). An understanding of the genetic diversity and 66 32 doning of extensive livestock economy, and the rar- the spatial structure of populations is important for 67 33 efaction of wild ungulate populations) (Hiraldo, 1974; establishing the appropriate scale and subunits for 68 34 Donazar et al., 2002). Many traditional breeding conservation management and minimizing genetic 69 35 areas were lost and the species is now extinct in erosion (Moritz, 1999). Current genetic patterns in a 70 36 France, Italy, Poland, Slovakia, Austria, Croatia, species are shaped both by historical and contempo- 71 37 Yugoslavia, Romania, Moldova, and Cyprus (Cramp & rary factors that affect its biogeography and its 72 38 Simmons, 1980; Meyburg & Meyburg, 1984), result- demography. The relative contributions of historic 73

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 3

1 and contemporary factors in shaping the genetic vultures. Although hair and scat (the most common 57 2 makeup of the species are not easy to unravel, and a non-invasive material) have been used extensively 58 3 combination of several analyses at different temporal for genetic tagging (Taberlet et al., 1997), large-scale 59 4 scales (i.e. haplotype relatedness, demographic his- studies utilizing feathers for individual identiﬁcation 60 5 tory, and population genetics) might be necessary to remain rare (Rudnick et al., 2005). 61 6 describe the geographical structure and investigate 62 7 the historic or contemporary processes that determine MATERIAL AND METHODS 63 8 its origin (Bernatchez, 2001; Godoy et al., 2004). 9 However, there has been no attempt to describe the SAMPLING AND DNA EXTRACTION, AMPLIFICATION, 64 10 geographical distribution of genetic diversity or the AND SEQUENCING 65 11 evolutionary history of A. monachus. Like many In total, 173 samples of A. monachus were collected 66 12 large-sized raptors, A. monachus is difficult to study from wild-caught individuals captured during ﬁeld 67 13 because: (1) it is an endangered species with large ecology projects (Fig. 1). DNA from blood or muscle 68 14 home ranges; (2) it keeps small or limited popula- tissue (N = 135) was extracted with DNeasy Tissue 69 15 tions, whereas some of them are not well monitored Kit (Qiagen) following the manufacturer’s instruction. 70 16 and the available information are limited; (3) its nests DNA from dry skins/bones (N = 12) and naturally 71 17 are located in inaccessible places; and (4) males and molted adult feathers (N = 26) was isolated using 72 18 females are indistinguishable visually, and adults are DNA techniques for museum specimens. The feathers 73 19 sometimes difficult to capture and mark. and skins were washed three times in 1 mL of 10 mM 74 20 To date, few genetic studies have been carried out Tris-HCl (pH 8.0) on a rotary mixer for 24 h per wash 75 21 for large-sized vultures, including bearded vultures to re-hydrate (Austin & Melville, 2006), whereas 76 22 (Gypaetus barbatus) (Negro & Torres, 1999; Gautschi 40 mg of powder bones were incubated in 1 mL of 77 23 et al., 2003; Godoy et al., 2004), griffon vultures proteinase K digestion buffer composed of 10 mM Tris 78 24 (Gyps fulvus) (Le Gouar et al., 2006), Egyptian (pH 8.0), 10 mM NaCl, 50 mM ethylenediaminetet- 79 25 vulture (Neophron percnopterus) (Kretzmann et al., raacetic acid (pH 8.0), 0.5% sodium dodecyl sulphate 80 26 2003), Andean condors (Vultur gryphus) (Hendrickson and1mgmL-1 proteinase K for 1 h at 56 °C under 81 27 et al., 2003), and Old World vultures (Lerner & agitation (Rohland & Hofreiter, 2007). In both cases, 82 28 Mindell, 2005). These studies have shown low levels after centrifugation and washing of the ‘pellet’ with 83

29 of genetic diversity (in G. fulvus, this was quite high 1 mL of ddH2O, we followed used the DNeasy Tissue 84 30 compared to the other species of vultures), strong Kit in accordance with the manufacturer’s instruc- 85 31 differentiation between indigenous and captive popu- tions. A partial sequence of mitochondrial cyt b gene 86 32 lation, and the existence of discernable evolutionary was amplified using the universal primers L14841 87 33 lineage in Europe. and H15149 (Palumbi, 1996) in a polymerase chain 88 34 In the present study, we have applied nuclear reaction (PCR) protocol of denaturation 94 °C for 89 35 multilocus genotypes (eight microsatellites) and mito- 5 min, and 35 amplification cycles of 94 °C for 30 s, 90 36 chondrial cytochrome b (cyt b) sequences to the two 47 °C for 30 s, and 72 °C for 45 s, followed by an 91 37 major remaining breeding populations of Europe extension at 72 °C for 10 min. A negative control was 92 38 (Iberian and Balkan Peninsula). The study aimed to run with each round of PCR. PCR products were 93 39 evaluate whether current genetic diversity of these purified with the NucleoSpin Extract II purification 94 40 populations could have been affected by the species kit (Macherey-Nagel). Single stranded sequencing 95 41 decline by comparing current estimates with those was performed with the primers of PCR, using the 96 42 found in Asia (Mongolia), in which the breeding popu- Big-Dye Terminator (version 3.1) Cycle Sequencing 97 43 lations are more or less stable and the species is kit on an ABI 377 sequencer. PCR fragments were 98 44 described as common. As well as providing insights sequenced in both directions to assure sequence accu- 99 45 into the past and current processes that have shaped racy. Mitochondrial DNA sequences were edited using 100 46 the evolution and genetic structure of A. monachus, SEQUENCHER, version 4.2 (Gene Codes Corpora- 101 47 our objectives are also designed to identify manage- tion) and deposited in GenBank (EF426498–537). 102 48 ment units for conservation of this large bird of prey. 103 49 It known that the ability to genetically profile non- 50 invasively collected samples has proven particularly GENOTYPING 104 51 important for conservation, allowing researchers to We tested 19 microsatellite primer pairs, developed 105 52 genetically ‘tag’ individuals of threatened or endan- for the Griffon Vulture G. fulvus (Mira et al., 2002) 106 53 gered species without capture. Accordingly, we also and the bearded vulture G. barbatus (Gautschi et al., 107 54 used non-invasively collected samples (e.g. feathers 2000). Eight loci proved polymorphic and showed 108 55 and bones) to demonstrate how non-invasive genetic clearly scorable bands in A. monachus. One hundred 109 56 tagging can be of particular use when studying and seventy-three individuals were successfully 110

4 N. POULAKAKIS ET AL.

1 genotyped at these eight microsatellite loci. Thermal whereas a smooth, unimodal shape is typical of 56 2 cycling was performed under the following conditions: expanding populations (Excoffier & Schneider, 1999). 57 3 5 min at 94 °C; a touchdown of 15 cycles of 94 °C for 58 4 1 min, 62 °C for 30 s (with a stepwise decrease of 1 °C 5 at each cycle), 72 °C for 45 s; then 30 amplification MICROSATELLITE ANALYSES 59 6 cycles of 94 °C for 1 min, 47 °C for 30 s, and 72 °C for Departures from Hardy–Weinberg equilibrium and 60 7 45 s, followed by a final extension at 72 °C for 10 min. linkage equilibrium were tested using GENEPOP, 61 8 PCR and loading multiplexes were developed to version 3.4 (Raymond & Rousset, 1995), where the 62 9 reduce the time and cost of genetic analyses (eight significance was evaluated by Fisher exact test 63 10 loci in four PCR reactions and loading into two gel- P-values, applying the Markov chain method (10 000 64 11 lanes). PCR products were analysed in an ABI-377 dememorization). We determined levels of genetic 65

12 sequencer. GENESCAN and GENOTYPER, version differentiation among populations using FST (Weir 66 13 2.0 (Applied Biosystems) were used for the determi- & Cockerham, 1984) in ARLEQUIN, version 3.1 67 14 nation of allele sizes. (Excoffier et al., 2005), where the statistical signiﬁ- 68 cance was assessed by permutation test (10 000 69 15 permutations). 70 16 MITOCHONDTIAL DNA STATISTICAL ANALYSES Genotyping errors, such as non-ampliﬁed alleles, 71 17 The alignment of the cyt b sequences was performed short allele dominance, and scoring of stutter peaks, 72 18 with CLUSTAL X (Thompson et al., 1997). Cyto- were assessed statistically using MICROCHECKER, 73 19 chrome b sequences were translated into amino acids version 2.2.3 (Van Oosterhout et al., 2004). The prob- 74 20 prior to analysis to check for spurious gaps or stop ability of full-sib or unrelated pairs of black vultures, 75

21 codons. Sequence divergences were estimated using P(ID), bearing an identical multilocus genotype, was 76 22 MEGA, version 3.1 (Kumar, Tamura & Nei, 2004). estimated using the software GIMLET, version 1.3.3 77 23 The settings for the DNA substitution model that best (Valiere, 2002) to explore the discrimination power of 78 24 ﬁtted the data were selected by the Akaike Informa- the microsatellite locus combination. For each micro- 79 25 tion Criterion (Akaike, 1974) using the programs satellite locus, we assessed genetic polymorphism by 80

26 MODELTEST, version 3.7 (Posada & Crandall, 1998). calculating the observed number of alleles (Na), allele 81

27 Phylogenetic relationships among cyt b sequences richness (RS), average observed heterozygosity (Ho), 82

28 were estimated using Neighbour-joining (NJ) and gene diversity (HS) across populations, and total gene 83

29 Bayesian inference (BI) methods implemented in diversity (HT) in FSTAT, version 2.9.3 (Goudet, 2001). 84 30 PAUP*4.0b10 (Swofford, 2002) and MRBAYES, Commonly-used population genetic summary statis- 85 31 version 3.1 (Ronquist & Huelsenbeck, 2003), respec- tics, such as the mean number of alleles per locus 86 7 32 tively. BI analysis was run with four chains for 10 (Na), the number of unique alleles (Nua), the observed 87

33 generations and the current tree was saved to ﬁle heterozygosity (Ho), and the unbiased expected het- 88

34 every 100 generations. The robustness of these analy- erozygosity (HE), were computed for each locus and 89 35 ses was assessed using bootstrap method with 1000 population using GENETIX, version 4.05.2 (Belkhir 90 36 replications for NJ and the posterior probabilities for et al., 2001). 91 37 BI as the percentage of samples recovering any par- Each population was tested for heterozygosity 92 38 ticular clade, where probabilities Ն 95% indicate excess to detect recent population bottlenecks. The 93 39 significant support. In all phylogenetic analyses, program BOTTLENECK (Cornuet & Luikart, 1996; 94 40 individuals from two closely-related species of the Piry, Luikart & Cornuet, 1999) was run under the 95 41 same family (Accipitridae) were used as outgroup two-phase model of microsatellite evolution (Dirienzo 96 42 taxa: G. fulvus (X86752) (Seibold & Helbig, 1995) and et al., 1994) with 10% of the infinite allele model and 97 43 G. barbatus (U08943) (Avise, Nelson & Sibley, 1994). 90% of the stepwise mutation model. 98 44 Additionally, a haplotype network, which represents The genetic structure was further examined with 99 45 a set of statistically parsimonious (P Ն 0.95) connec- a Bayesian clustering method and the admixture 100 46 tions between haplotypes, was constructed using the analysis implemented in STRUCTURE, verison 2.2 101 47 program TCS, version 1.21 (Clement, Posada & Cran- (Pritchard, Stephens & Donnelly, 2000). Markov 102 48 dall, 2000). chain Monte Carlo parameters were set on burn-in 103 49 We attempted to gain some insight into the past period of 30 000 and run length of 106 iterations and 104 50 demographic history of A. monachus by calculating each run was repeated five times to ensure conver- 105 51 the mismatch distribution on the pooled populations gence among estimated parameters. The true number 106 52 in ARLEQUIN, version 3.1 (Excoffier, Laval & of clusters (K) was estimated by: (1) using the model 107 53 Schneider, 2005). An irregular and multimodal distri- choice criterion implemented in STRUCTURE that is 108 54 bution of pairwise differences between sequences is the maximal value estimate of posterior probability 109 55 expected in stationary or shrinking populations, of the data for a given K,Pr(X|K) (Pritchard et al., 110

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 5

1 Table 1. Alignment of variable positions found in seven unique haplotypes of cytochrome b sequences and their 2 distribution across seven sampling sites 3 4 Nucleotide position Sampling locality 5 76 00000000112223 Mongolia 8 00013467280560 9 Haplotype 36963200604251 Spain Greece Armenia Kazakhstan Georgia 12 10 11 Hap_1 CCGCTCCTACCATT 8– – – – –– 12 Hap_2 ...T..GG..G... –7 – – – –– 13 Hap_3 .T.T..GG..G... –3 – – – –– 14 Hap_4 ..ATCT.GG...C. –– 5 4 3 –– 15 Hap_5 ..ATC..GGT.CCC –– – – – 42 16 Hap_6 ..ATC..GGT.CC. –– – – – 10– 17 Hap_7 T.ATC..GGT.CC. –– – – – 2– 18 19 Mongolia 1, Erdenesant (Mongolia); Mongolia 2, Khustai (Mongolia). 20 Dots indicate identity with haplotype 1. The total number of individuals from each population with the corresponding 21 haplotype is given. 22 23 2000) and (2) using another ad hoc quantity that is substitution chosen as best ﬁt for our data was the 57 24 based on the second order rate of change of the Tamura–Nei model (Tamura & Nei, 1993) with an 58 25 likelihood function with respect to K (DK) as proposed estimate of invariable sites (I = 0.93). Under this 59 26 by (Evanno, Regnaut & Goudet, 2005). model of evolution, the sequence divergence among 60 27 Analysis of molecular variance (AMOVA) (Excoffier, the specimens of A. monachus were in the range 61 28 Smouse & Quattro, 1992) was performed using ARLE- 0–3.3%, whereas the NJ and Bayesian trees have 62 29 QUIN. F-statistics were used to estimate the propor- similar topologies. The Bayesian 50% majority rule 63 30 tion of genetic variability found among populations consensus tree (lnL =-866.935) is presented in 64

31 (FST), among populations within groups (FSC) and Figure 2A. The seven different haplotypes form four 65

32 among groups (FCT). An AMOVA was run with popu- very well supported allopatric lineages, corresponding 66 33 lations grouped according to the genetic clusters to different geographic regions throughout the Pale- 67 34 found in the structure analysis, and with populations arctic region, where the European populations are 68 35 grouped according to the clades found in the phylo- more closely related to each other. These lineages 69 36 genetic analysis of mitochondrial (mt)DNA structure are the lineage A (Balkan Peninsula, Greek popula- 70 37 analysis. If the ﬁxation index over all loci (Weir & tion), the lineage B (Iberian Peninsula, Spain popu- 71 38 Cockerham, 1984) among populations within a group lation), the lineage C (Caucasian region, Armenia, 72

39 (FSC) departed significantly from zero, then we con- Georgia, Kazakhstan populations), and the lineage D 73 40 sidered that populations in the group should be [(Far Eastern region, Mongolian populations (Khustai 74 41 further subdivided. Significance associated with the and Erdenesant)]. 75 42 fixation index was evaluated through random allelic The shape of the mismatch distribution function for 76 43 permutation procedures (10 000 permutations). the whole data set is multimodal, reflecting the deep 77 44 Correlation analysis between genetic and geo- divergence between the four mitochondrial clades. A 78 45 graphical distances (Mantel test) was calculated with bimodal mismatch distribution is observed in the 79 46 the ISODLE Program implemented in GENEPOP European populations. However, when each clade is 80 476 6 as proposed by Rousset (1997). The statistical sig- analysed separately, the shape of the function is uni- 81 48 nificance of the correlation of these matrices was modal, conforming to the theoretical expectation for a 82 49 assessed with a Mantel randomization test (10 000 growing population (data not shown). 83 50 permutations). The network of mtDNA haplotypes of A. monachus 84 was presented in Figure 2B, where the topology is 85 51 in accordance with the phylogenetic relationships 86 52 RESULTS obtained in BI and NJ analyses. 87 53 MTDNA DATA 88 54 Overall, the 48 sequences defined seven unique hap- MICROSATELLITE DATA 89 55 lotypes (Table 1). Of the 311 sites examined, there Two loci (BV5, BV12) deviated significantly from 90 56 were 14 variable cyt b sites. The model of nucleotide Hardy–Weinberg proportions due to heterozygote 91

6 N. POULAKAKIS ET AL.

1 Figure 2. A, Bayesian 50% majority rule tree of mitochondrial (mt)DNA data based on Tamura–Nei + I model of 2 evolution. Two sequences of Gyps fulvus and Gypaetus barbatus were used as outgroups. Numbers above branches are 3 bootstrap values on Neighbour-joining and posterior probabilities values on Bayesian inference. B, unrooted statistical 4 parsimony network of Eurasian black vulture mtDNA cytochrome b haplotypes. The numbers refer to the numbers of 5 individuals that shared the same haplotype. Haplotypes colours correspond to those in (A). 6 7 deﬁcit (P < 0.05). This was consistent across popula- bounded between 0.01 and 0.0001 (Waits, Luikart & 33 8 tions. Heterozygote deﬁcit could be a sign of the Taberlet, 2001). The four most informative loci are 34 9 population sub-structuring segregating or sampling BV5, Gf11A4, BV16, and BV11. 35 10 artefacts. We assume that the patchy distribution of Diversity estimates varied among microsatellite 36 11 the populations is an important factor contributing to loci (Table 2) and among populations (Table 3). The 37 12 the lack of heterozygosity. However, null alleles, stut- total number of alleles per locus in our sample of 173 38 13 tering signals, or large allelic dropouts could also individuals was in the range 3–20 at the eight loci. 39

14 contribute to ‘false positive’ homozygous patterns. Average gene diversity (HS) among the loci was 0.64 40

15 We examined the latter explanation using MICRO- (range 0.11–0.89), whereas the gene diversity (HT) 41 16 CHECKER. Evidence for scoring error due to stutter- estimates among all populations were in the range 42

17 ing and large allelic dropouts or null alleles was not 0.18–0.91. Mean observed heterozygosity (Ho) ranged 43 18 found. Three pairs of loci (BV5-BV13, BV6-BV20, from 0.36 (Spain) to 0.56 (Greece). The lowest mean 44

19 BV12-Gf3H3) showed some evidence of linkage dis- expected heterozygosity (HE) was found in the popu- 45 20 equilibrium. However, no one pair of loci was consis- lation from Armenia (0.52), whereas the highest was 46 21 tently linked across all samples and therefore normal in the population from Greece (0.72). 47 22 segregation of the eight loci may be assumed. Thus, Two main clusters of populations were identiﬁed 48 23 this locus combination of microsatellite markers is using STRUCTURE by both ad hoc quantities used in 49 24 currently being used to investigate the genetic popu- this study. Plotting the membership coefficient for 50 25 lation structure of Eurasian black vultures. each pre-deﬁned population gave us the structure 51 26 Analysis using GIMLET showed that these loci imprinted in Figure 3. The clusters that comprised 52 27 combined would only produce an identical genotype the uppermost hierarchical level of population struc- 53 28 by chance in the case of full sibs with a probability ture (Evanno et al., 2005) were the Far Eastern 54 29 of < 0.005 (i.e. these markers can even distinguish cluster, consisting of the populations of Mongolia 55 30 siblings with 99.9% probability). General guidelines (Erdenesant and Khustai), and a West Asia – Europe 56 31 for genetic tagging studies suggest using a suite of cluster, consisting of Iberian, Balkan, and Caucasian 57

32 markers that achieve a reasonably low P(ID) and P(IDsib) (Armenia, Kazakhstan, and Georgia) populations. 58

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 7

1 Table 2. A comparison of genetic diversity at eight microsatellite loci in Aegypius monachus 2

3 Locus Na Rs Ho Hs HT FST RST 4 5 BV5 18 3.484 0.377 0.846 0.910 0.117 0.579 6 BV11 14 3.319 0.844 0.838 0.879 0.053 0.316 7 BV12 5 2.108 0.051 0.296 0.367 0.188 0.060 8 BV13 9 2.659 0.696 0.667 0.693 0.079 0.025 9 BV16 20 3.338 0.140 0.891 0.848 0.003 0.000 10 BV20 8 2.868 0.602 0.662 0.792 0.234 0.103 11 Gf3H3 3 1.611 0.026 0.116 0.184 0.442 0.368 12 Gf4A4 20 3.382 0.776 0.816 0.895 0.138 0.310 13

14 FST and RST values were estimated according to Weir & Cockerham (1984) and Goodman (1997), respectively.

15 Na, observed allele number; Rs, allelic richness; HO, observed heterozygosity; HS, gene diversity; HT, overall gene diversity;

16 FST, among population genetic differentiation; RST, gene differentiation accounting for variance in allele size. 17

18 Table 3. Sample size (N), the mean number of alleles per locus (Na), unique alleles (Nua), observed heterozygosity (HO),

19 and Nei’s unbiased expected heterozygosity (HE) for each population based on the eight microsatellite loci 20

21 Population NNa Nua HO HE 22 23 Spain 44 4.00 3 0.3636 0.5950 24 Greece 53 7.625 7 0.5589 0.7259 25 Georgia 3 2.375 0 0.3750 0.5250 26 Armenia 5 3.375 1 0.4000 0.5944 27 Kazakhstan 4 3.25 0 0.4375 0.5833 28 Mongolia 1 (Erdenesant) 52 8.75 12 0.5589 0.6621 29 Mongolia 2 (Khustai) 12 5.5 0 0.5072 0.6520 30 31 32 Table 4. Pairwise comparison matrix of q values among Aegypius monachus populations 33 34 Spain Greece Georgia Armenia Kazakhstan Mongolia 1 35 36 Spain – 37 Greece 0.107* – 38 Georgia 0.144* 0.093* – 39 Armenia 0.093* 0.065* 0.000 – 40 Kazakhstan 0.120* 0.113* 0.011 0.000 – 41 Mongolia 1 0.176* 0.148* 0.103* 0.091* 0.095* – 42 Mongolia 2 0.163* 0.134* 0.079* 0.075* 0.063* 0.000 43 44 *Statistically significant at P = 0.05. 45 46 Overall, 94% of the samples are assigned with a high values. There was no genetic evidence of recent 57 47 degree of certainty (membership coefficient for the bottlenecks in any of the Aegypius populations. 77 58 48 population of origin: P > 0.8). Among the remaining The results of the AMOVA (Table 5) confirmed the 88 59 49 individuals that display lower membership coeffi- presence of phylogeographic structure in our data. The 60 50 cients, only three present an assignment score lower results of this analysis considering all populations as 61 51 than 0.74, probably indicating low levels of gene flow one demonstrated that most of the molecular variation 62 52 and thus high level of genetic diversity among the was distributed within populations (89.68%) rather 63 53 four clusters. than among populations (10.32%) (Table 6), suggest- 99 64 54 Genetic differentiation was examined using mean ing that there are some genetic structures within the 65

55 FST values (0–17.6%; Table 4, where 17 out of 21 group. The AMOVA of the two genetic clusters of 66

56 comparisons deviate signiﬁcantly from zero, the FST structure analysis showed that only 2.86% of the 67

8 N. POULAKAKIS ET AL.

1 Figure 3. Bayesian analysis of the nuclear genetic structure of Aegypius populations based on eight microsatellite loci. 2 Each individual is represented by a thin vertical line, which is partitioned into coloured segments that indicate the 3 individual’s membership in two groups. The populations of individuals are indicated. 4 5 Table 5. Analysis of molecular variance for Aegypius monachus using eight microsatellite loci between: (1) two groups 6 (Far Eastern and West Asia-Europe groups) identiﬁed by STRUCTURE and (2) four groups (Balkan Peninsula: lineage 7 A, Iberian Peninsula: lineage B, Caucasian region: lineage C, and the Far Eastern: lineage D) identiﬁed by phylogenetic 8 analysis of mitochondrial DNA and (3) all populations 9 10 Sum of Percentage 11 Source of variation d.f. squares of variation Statistics P 12

13 (1) Among groups 1 13.435 2.86 FST = 0.112 0

14 Among populations within groups 5 19.338 8.42* FSC = 0.086 0

15 Within populations 339 311.709 88.72* FCT = 0.028 0.068

16 (2) Among groups 3 30.627 12.05* FST = 0.109 0

17 Among populations within groups 3 2.147 0 FSC = 0.017 0.695

18 Within populations 339 311.709 89.03* FCT = 0.12 0.005 19 20 *Signiﬁcant values (P < 0.01). 21 22 variance was attributed to differences among these were still structured within at least one group. An 29 23 clusters and 8.42% of the variance to differences AMOVA was run where the groups were deﬁned based 30 24 among populations within these clusters. In this case, on the phylogenetic lineages of the mtDNA data 31

25 the ﬁxation index among groups (FCT) was not signiﬁ- (Fig. 2A, Balkan Peninsula: Lineage A; Iberian Pen- 32

26 cant (FCT = 0.028, P = 0.068), whereas it was signiﬁ- insula: Lineage B; Caucasian region: Lineage C; and 33

27 cant among populations within groups (FSC = 0.086, Mongolia: Lineage D). This time, the variance among 34 28 P < 0.001). These results suggested that populations groups increased to 12.05% and the variance among 35

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 9

1 Table 6. Analysis of molecular variance (AMOVA) for Aegypius monachus using eight microsatellite loci between: (1) two 2 groups (Far Eastern and West Asia-Europe groups) identiﬁed by STRUCTURE; (2) four groups (Balkan Peninsula: lineage 3 A, Iberian Peninsula: lineage B, Caucasian region: lineage C, and the Far Eastern: lineage D) identiﬁed by phylogenetic 4 analysis of mtDNA; and (3) all populations 5 6 Sum of Percentage 7 Source of variation d.f. squares of variation Statistics P 8

9 (1) Among groups 1 13.435 2.86 FST = 0.112 0

10 Among populations within groups 5 19.338 8.42* FSC = 0.086 0

11 Within populations 339 311.709 88.72* FCT = 0.028 0.068

12 (2) Among groups 3 30.627 12.05* FST = 0.109 0

13 Among populations within groups 3 2.147 0 FSC = 0.017 0.695

14 Within populations 339 311.709 89.03* FCT = 0.12 0.005

15 (3) Among populations 6 32.773 10.32* FST = 0.103 0 16 Within populations 339 311.709 89.68* 17 18 *Signiﬁcant values (P < 0.01).

20 populations within groups decreased to 0%, indicating distribution. This geographical pattern of haplotype 55 21 that none of the populations within each group was distribution (Fig. 2) indicates a sharp East–West 56 22 structured. Based on this group deﬁnition, the ﬁxation clinal distribution of clade proportions in populations 57

23 index among groups was signiﬁcant (FCT = 0.12, ranging from Iberian Peninsula to Asia, which has 58 24 P < 0.001), whereas it was not signiﬁcant among popu- been observed in bearded vultures (G. barbatus) 59

25 lations within groups (FSC = 0.017, P = 0.69). Conse- (Godoy et al., 2004) and the white-tailed eagle 60 26 quently, according to the results of the AMOVA tests, (Haliaeetus albicilla) (Hailer et al., 2006, 2007). 61 27 we considered that the seven sampling populations With only seven haplotypes detected (Table 1), the 62 28 could be assigned into four groups (i.e. that we Eurasian black vulture display a relatively low mito- 63 29 described above), which are consistent with the pro- chondrial variability (haplotype diversity), a phenom- 64 30 duced lineages of BI analysis of mtDNA. enon that has been already reported in other raptor 65 31 species, generally associated with recent demographic 66 crashes: the Mauritius kestrel, Falco punctatus 67 32 DISCUSSION (Groombridge et al., 2000); the Norwegian peregrine 68 33 The results obatined in the present study are unique falcon, Falco peregrinus (Lifjeld et al., 2002); the 69 34 in that they comprise the first population genetic bearded vultures, G. barbatus (Godoy et al., 2004); the 70 35 analysis of the Eurasian black vultures (A. monachus) critically endangered Spanish imperial eagle, Aquila 71 36 in Europe. The analysis of both mitochondrial and adalberti (Martinez-Cruz et al., 2004); the red kite 72 37 microsatellite data allowed us to characterize the (Milvus milvus) (Roques & Negro, 2005); and the 73 38 genetic composition of the current European breeding white-tailed eagle (Haliaeetus albicilla) (Hailer et al., 74 39 populations in terms of both diversity and structure. 2006, 2007). 75 40 The number genetic studies of Eurasian raptors, most Wide-ranging species with high dispersal capabili- 76 41 of them involving critically endangered or threatened ties should have relatively low genetic structuring 77 42 species, have increased in the last decade, inferring (Jones et al., 2004). However, in the present study, 78 43 important issues for the biodiversity and conservation significant structuring of genetic diversity was found 79 44 status of those species (Negro & Torres, 1999; Nesje among most black vulture populations, indicating a 80 45 et al., 2000; Cardia et al., 2002; Vali, 2002; Kretzmann low level of gene flow. Although an analogue pattern 81 46 et al., 2003; Godoy et al., 2004; Martinez-Cruz, Godoy was observed in bearded vultures (G. barbatus) (Godoy 82 47 & Negro, 2004; Helbig et al., 2005a, b; Roques & et al., 2004) and the white-tailed eagle (Haliaeetus 83 48 Negro, 2005; Hailer et al., 2006; Hailer et al., 2007). albicilla) (Hailer et al., 2007), these results contrast 84 49 At the evolutionary level, the phylogenetic analysis with the general pattern of low levels of genetic 85 50 of cyt b sequences reveals the existence of two evolu- differentiation among populations of birds (Crochet, 86 51 tionary lineages in Europe that correspond to the two 2000) and with the results obtained for its sister 87 52 different breeding populations (Iberian and Balkan species, the Egyptian vulture (Kretzmann et al., 2003). 88 53 peninsulas). Haplotypes clustered in two distinct Despite the low mtDNA variability (haplotype 89 54 haplogroups with a predominantly eastern or western diversity), the small number of substitutions (Table 1) 90

10 N. POULAKAKIS ET AL.

1 revealed a clear geographic pattern of population isolation by distance. Thus, nuclear DNA genetic dif- 57 2 structure because each population constitutes a dis- ferentiation, as measured by microsatellite loci, is 58 3 tinct lineage (with the exception of the Caucasian consistent with mtDNA phylogeographical groupings. 59 4 region in which the three populations share the same Birds of prey are vulnerable to extinction because 60 5 haplotype). Moreover, the two European populations most have traditionally been persecuted by humans. 61 6 (Iberian and Balkan) form a sister group, whereas the Based on BirdLife International (2006), A. monachus 62 7 populations from Caucasus are closer to Mongolian has small populations that appear to be suffering 63 8 ones, comprising a group that appear to be more ongoing decline, despite the fact that, in parts of 64 9 distinct from the European populations (Fig. 2). Much Europe, numbers are now increasing (see Introduc- 65 10 of the current haplotype diversity in A. monachus is tion). Focusing on the two major breeding populations 66 11 probably a consequence of this strong within species of Europe, the analysis of mtDNA revealed a low 67 12 phylogeographic structure. One to three different degree of genetic variability (haplotype diversity) 68 13 mtDNA haplotypes were obtained from each of the (Table 1). Although the level of variability in the mito- 69 14 different populations of A. monachus (Table 1), sug- chondrial genome is low, no evidence of genome-wide 70 15 gesting a strong degree of spatial genetic structure genetic erosion exists because nuclear diversity exhi- 71 16 within the species, limited exchange of individuals bits normal levels. Signatures of a genetic bottleneck 72 17 among these populations, and deep divergence in the nuclear genome were not detected, indicating 73 18 between the lineages of A. monachus. In addition, the that the demographic bottleneck suffered during the 74 19 shape of the mismatch distribution for the whole 20th Century was neither critical nor lasting enough 75 20 data set (multimodal) and for the European subset to have an impact on nuclear genetic variation at the 76 21 (bimodal) lead to same conclusion. This pattern might species level. This could be probably due to the long 77 22 have originated by divergence of these lineages in generation time of the raptors, confirming the argu- 78 23 allopatry. However, such a high level of population ment of Hailer et al. (2006) suggesting that the long 79 24 differentiation is specially striking for a bird that generation time of several species (i.e. eagles, turtles, 80 25 usually flies long distances in a single day. A highly large mammals) has acted as an intrinsic buffer 81 26 philopatric dispersal behaviour needs to be invoked to against loss of genetic diversity, leading to a shorter 82 27 clarify this pattern and to explain the existence of a effective time of the experienced bottleneck. 83 28 clear phylogeographical pattern, as in the case of 84 29 Bearded Vultures (Godoy et al., 2004). 30 Although the results from mitochondrial DNA refer CONSERVATION IMPLICATIONS FOR THE EUROPEAN 85 31 only to female dispersal, similar results obtained with BREEDING POPULATIONS 86 32 the eight nuclear microsatellite markers used in the The conservation status of A. monachus has recently 87 33 present study suggest that this is not a sex-specific been a cause for concern. The two main threats to the 88 34 pattern. The use of eight variable microsatellites species are direct mortality caused by humans (either 89 35 was sufficient to reveal a strong genetic structure accidentally or deliberately) and decreasing availabil- 90 36 at a large spatial scale (Table 4). The levels of ity of food. The information obtained in the present 91 37 genetic diversity found in A. monachus populations analysis will be instrumental in refining conserva- 92

38 (Ho = 0.36–0.56) were comparable to values observed tion strategies to protect what remains in Iberian 93 39 in a number of species of endangered vertebrates and Balkan breeding populations of Eurasian black 94 40 (Ciofi et al., 2002) and were similar to those found in vulture. Currently, it is a generally accepted conclu- 95 41 other raptor species (Nesje et al., 2000; Kretzmann sion that the loss of genetic diversity is not desirable 96 42 et al., 2003; Rudnick et al., 2005). because it reduces the ability of species to cope with 97 43 Beside the fact that the structure analysis indicates environmental changes (Frankham, Ballou & Briscoe, 98 44 the presence of two main clusters of populations 2002). 99 45 (Fig. 3) (Far Eastern and West Asia–Europe clusters), Phylogenetic analyses can identify genetically dis- 100 46 the significant hierarchical analyses of genetic tinct lineages worthy of conservation and can also aid 101 47 substructure (AMOVA) supports four distinct groups in setting priorities for conservation efforts. When 102 48 throughout Eurasia. The significance of Mantel test population distinctiveness is evaluated on the basis of 103 49 (P = 0.003) for correlation between genetic distance genetic criteria, it is usually accepted that distinct 104 50 and geographical distance, indicates that the genetic genetic populations are those showing reciprocally 105 51 differentiation observed in nuclear DNA was linked to monophyletic mtDNA alleles and significant diver- 106 52 the phylogenetic groups defined by the mtDNA analy- gence for allele frequencies at nuclear loci (Moritz, 107 53 sis. Consequently, our results support the hypothesis 1994). Based on these criteria, Moritz (1994, 1999) 108 54 that gene flow between the phylogenetic groups is operationally defined evolutionarily significant units 109 55 restricted and that the observed genetic differentia- (ESUs) and management units (MUs) for conserva- 110 56 tion of A. monachus is due to population history and tion, whereas Crandall et al. (2000) argue that the 111

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 11

1 assessment of distinctiveness should appropriately can counteract the effects of inbreeding depression 57 2 incorporate adaptive differences, as evidenced by (Hedrick, 1995; Madsen et al., 1999; Mansfield & 58 3 genetic and ecological data. Land, 2002; Robertson, Karika & Saul, 2006). Trans- 59 4 In Europe, where only two mainly populations locations have been frequently used in the manage- 60 5 exist, the fixation of different mtDNA haplotypes in ment of threatened birds (Komdeur, 1994; Armstrong 61 6 each population (one in Iberian population and two in & McLean, 1995). Despite the success of many con- 62 7 Balkan population) and the significance divergence at ducted programs, translocations are fraught with 63 8 nuclear loci fulfill the definitions of ESUs, even if potential dangers. These include the introduction of 64 9 fixation were to be an epiphenomenon resulting from exotic pathogens and a host of difficulties associated 65 10 anthropogenic reduction and fragmentation of the with acclimating captive raised animals to natural 66 11 species’ distribution. environments (Snyder et al., 1996). Additionally, 67 12 It must be noted that even though there is no rescue effect gene flow or captive propagation could be 68 13 evidence (field observations) of contact between used to raise a population above a demographically 69 14 Balkan and Iberian populations, the two populations critical size, but subsequent reduction or cessation of 70 15 must have been indirectly connected in the past by these factors might be required to enable further 71 16 gene flow through the now extinct Central European adaptation and greater population productivity 72 17 populations. Thus, the hypothesis of recent and (Stockwell, Hendry & Kinnison, 2003). Guidelines 73 18 historical ecological exchangeability (Crandall et al., for translocations emphasize the need for a multi- 74 19 2000) can neither be supported nor rejected due to disciplinary approach by taking into account biologi- 75 20 lack of data. Based on the observation that varying cal, socio-economic and legal requirements (Robertson 76 21 circumstances require differing approaches, Fraser & et al., 2006). 77 22 Bernatchez (2001) provide a framework for manage- The future of the species in Europe depends on the 78 23 ment decisions when ecological data are not available effective implementation of conservation strategies 79 24 or insufficient. Our microsatellite and mitochondrial both in situ and ex situ that must necessarily consider 80 25 data provide strong evidence that historical isolation genetic issues such as those described in the present 81 26 has led to the differential accumulation of mutations study. Consequently, the aim of the conservation 82 27 in both populations, and this, together with the effort would not only be to maintain the habitats that 83 28 importance of genetic drift and geographic subdivi- these ESUs exploit, but also to preserve and/or 84 29 sion, leads us to conclude that the two breeding increase the genetic variation of the European breed- 85 30 populations represent distinct ESUs. Maximum pri- ing populations. 86 31 ority for conservation should be given because on the In summary, the study of genetic variability and 87 32 basis of these, and due to its genetic divergence from population structure should be a key factor in man- 88 33 the Eastern populations, they comprise a major agement and recovery programmes to enable the 89 34 contribution to the total species’ genetic diversity. retention of genetic diversity in endangered or 90 35 It is worth notng that, although we sampled only 12 threatened species. The present study shows that 91 36 individuals from the ‘Caucasus’ region (Armenia mtDNA and microsatellite markers are useful tools 92 37 and Georgia, Kazakhstan), these birds were distinct for performing such investigations in Eurasian black 93 38 from the other European or Mongolian populations, vultures, and that single, non-invasively collected 94 39 whereas the same results were obtained when the vultures feathers and bones yield sufficient quanti- 95 40 analyses were repeated after removing the specimens ties of DNA for both mtDNA and microsatellite 96 41 from Caucasus. genotyping. 97 42 Genetic factors, including loss of diversity and 98 43 inbreeding depression, will increase the extinction ACKNOWLEDGEMENTS 99 44 risks by reducing adaptive potential and decreasing 45 average fitness, respectively (Brook et al., 2002). As The study was conducted in the framework of the 100 46 we describe above, the levels of variability in mtDNA LIFE Nature project ‘Conservation of Birds of Prey in 101 47 are relatively low but there was no evidence of the Dadia Forest Reserve, Greece’, LIFE02/GR/8497, 102 48 genome-wide genetic erosion in the European blank (60% by EU and 40% by WWF Greece), which was 103 49 vulture populations. Consequently, management implemented by WWF Greece. Samples from Spain 104 50 strategies should aim to preserve the extant diversity were collected in the framework of the monitoring 105 51 of the Balkan and Iberian populations. project of the Black Vulture colony in Alto Lozoya 106 52 If the goal of a conservation program is to preserve SBPA supported by ‘Consejería de Medio Ambiente 107 53 genetic distinctiveness and ongoing evolutionary pro- y Ordenación del Territorio de la Comunidad de 108 54 cesses, introductions of birds should be discouraged. A Madrid’ (Peñalara Natural Park). Field work and 109 55 number of studies have demonstrated that artificially sample collection in Mongolia was organized and 110 56 restoring gene flow between isolated populations funded through the Peregrine Fund, Boise, ID, USA. 111

12 N. POULAKAKIS ET AL.

1 REFERENCES sequence repeat loci in human-populations. Proceedings of 59 2 the National Academy of Sciences of the United States of 60 3 Akaike H. 1974. New look at statistical-model identification. America 91: 3166–3170. 61 4 Ieee Transactions on Automatic Control Ac19: 716–723. Donazar JA, Blanco G, Hiraldo F, Soto-Largo E, Oria J. 62 5 Armstrong DP, McLean IG. 1995. New Zealand transloca- 2002. Effects of forestry and other land-use practices on the 63 6 tions: theory and practice. Pacific Conservation Biology 2: conservation of cinereous vultures. Ecological Applications 64 7 39–54. 12: 1445–1456. 65 8 Austin JJ, Melville J. 2006. Incorporating historical Evanno G, Regnaut S, Goudet J. 2005. Detecting the 66 9 museum specimens into molecular systematic and conser- number of clusters of individuals using the software 67 10 vation genetics research. Molecular Ecology Notes 6: 1089– STRUCTURE: a simulation study. Molecular Ecology 14: 68 11 1092. 2611–2620. 69 12 Avise JC, Nelson WS, Sibley CG. 1994. DNA-sequence Excoffier L, Laval G, Schneider S. 2005. Arlequin ver. 3.0: 70 13 support for a close phylogenetic relationship between some an integrated software package for population genetics data 71 14 storks and new-world vultures. Proceedings of the National analysis. Evolutionary Bioinformatics Online 1: 47–50. 72 15 Academy of Sciences of the United States of America 91: Excoffier L, Schneider S. 1999. Why hunter-gatherer popu- 73 16 5173–5177. lations do not show signs of Pleistocene demographic expan- 74 17 Belkhir K, Borsa P, Chikhi L, Raufaste N, Bonhomme F. sions. Proceedings of National Academy of Science of the 75 18 2001. GENETIX, logiciel sous Windows TM pour la géné- United States of America 96: 10597–10602. 76 19 tique des populations: Laboratoire Génome, Populations, Excoffier L, Smouse PE, Quattro JM. 1992. Analysis of 77 20 Interactions CNRS UMR 5000. Montpellier, France: Univer- molecular variance inferred from metric distances among 78 21 sité de Montpellier II. DNA haplotypes – application to human mitochondrial-DNA 79 22 Bernatchez L. 2001. The evolutionary history of brown trout restriction data. Genetics 131: 479–491. 80 23 (Salmo trutta L.) inferred from phylogeographic, nested Frankham R, Ballou BD, Briscoe DA. 2002. Introduction 81 24 clade, and mismatch analyses of mitochondrial DNA varia- to conservation genetics. Cambridge: Cambridge University 82 25 tion. Evolution 55: 351–379. Press. 83 26 BirdLife International. 2006. Species factsheet: Aegypius Fraser DJ, Bernatchez L. 2001. Adaptive evolutionary con- 84 27 monachus. Available at http://www.birdlife.org servation: towards a unified concept for defining conserva- 85 28 Brook BW, Tonkyn DW, Q’Grady JJ, Frankham R. 2002. tion units. Molecular Ecology 10: 2741–2752. 86 29 Contribution of inbreeding to extinction risk in threatened Gautschi B, Jacob G, Negro JJ, Godoy JA, Muller JP, 87 30 species. Conservation Ecology 6: 16. Schmid B. 2003. Analysis of relatedness and determination 88 31 Cardia P, Fraguas B, Pais M, Silva S, Guillemaud T, of the source of founders in the captive bearded vulture, 89 32 Palma L, Cancela ML, Ferrand N. 2002. Análise da Gypaetus barbatus, population. Conservation Genetics 4: 90 33 variação genética de proteínas em populações mediter- 479–490. 91 34 rânicas de águia-perdigueira Hieraaetus fasciatus. Airo 12: Gautschi B, Tenzer I, Muller JP, Schmid B. 2000. 92 35 71–74. Isolation and characterization of microsatellite loci in 93 36 Carrete M, Donazar JA. 2005. Application of central-place the bearded vulture (Gypaetus barbatus) and cross- 94 37 foraging theory shows the importance of Mediterranean amplification in three Old World vulture species. Molecular 95 38 dehesas for the conservation of the cinereous vulture, Ecology 9: 2193–2195. 96 39 Aegypius monachus. Biological Conservation 126: 582–590. Godoy JA, Negro JJ, Hiraldo F, Donazar JA. 2004. 97 40 Ciofi C, Milinkovitch MC, Gibbs JP, Caccone A, Powell Phylogeography, genetic structure and diversity in the 98 41 JR. 2002. Microsatellite analysis of genetic divergence endangered bearded vulture (Gypaetus barbatus,L.)as 99 42 among populations of giant Galápagos tortoises. Molecular revealed by mitochondrial DNA. Molecular Ecology 13: 371– 100 43 Ecology 11: 2265–2283. 390. 101 44 Clement M, Posada D, Crandall KA. 2000. TCS: a com- Goudet J. 2001. FSTAT, a program to estimate and test gene 102 45 puter program to estimate gene genealogies. Molecular diversities and fixation indices (version 2.9.3). Available at 103 46 Ecology 9: 1657–1659. http://www.unil.ch/izea/softwares/fstat.html. 104 47 Cornuet JM, Luikart G. 1996. Description and power analy- Groombridge JJ, Jones CG, Bruford MW, Nichols RA. 105 48 sis of two tests for detecting recent population bottlenecks 2000. Conservation biology – ‘ghost’ alleles of the Mauritius 106 49 from allele frequency data. Genetics 144: 2001–2014. kestrel. Nature 403: 616–616. 107 50 Cramp S, Simmons KEL. 1980. The birds of the western Hailer F, Helander B, Folkestad AO, Ganusevich SA, 108 51 Palearctic. Oxford: Oxford University Press. Garstad S, Hauff P, Koren C, Masterov VB, Nygard T, 109 52 Crandall KA, Bininda-Emonds ORP, Mace GM, Wayne Rudnick JA, Shiraki S, Skarphedinsson K, Volke V, 110 53 RK. 2000. Considering evolutionary processes in conserva- Wille F, Vila C. 2007. Phylogeography of the white-tailed 111 54 tion biology. Trends in Ecology and Evolution 15: 290–295. eagle, a generalist with large dispersal capacity. Journal of 112 55 Crochet PA. 2000. Genetic structure of avian populations – Biogeography 34: 1193–1206. 113 56 allozymes revisited. Molecular Ecology 9: 1463–1469. Hailer F, Helander B, Folkestad AO, Ganusevich SA, 114 57 Dirienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin Garstad S, Hauff P, Koren C, Nygard T, Volke V, Vila 115 58 M, Freimer NB. 1994. Mutational processes of simple- C, Ellegren H. 2006. Bottlenecked but long-lived.high 116

THE GENETIC STRUCTURE OF EURASIAN BLANK VULTURES 13

1 genetic diversity retained in white-tailed eagles upon recov- Madsen T, Shine R, Olsson M, Wittzell H. 1999. Conser- 59 2 ery from population decline. Biology Letters 2: 316–319. vation biology – restoration of an inbred adder population. 60 3 Handrinos G. 1992. Birds. In: Karandrinos M, Legakis A, Nature 402: 34–35. 61 4 eds. The red data book of Greek vertebrates. Athens: Hellenic Mansfield KG, Land ED. 2002. Cryptorchidism in Florida 62 5 Zoology Society and Hellenic Ornithological Society, 125– panthers: prevalence, features, and influence of genetic res- 63 6 243. toration. Journal of Wildlife Diseases 38: 693–698. 64 7 Hedrick PW. 1995. Gene flow and genetic restoration – the Martinez-Cruz B, Godoy JA, Negro JJ. 2004. Population 65 8 Florida panther as a case-study. Conservation Biology 9: genetics after fragmentation: the case of the endangered 66 9 996–1007. Spanish imperial eagle (Aquila adalberti). Molecular 67 10 Helbig AJ, Kocum A, Seibold I, Braun MJ. 2005a. A Ecology 13: 2243–2255. 68 11 multi-gene phylogeny of aquiline eagles (Aves: Accipitri- Meyburg B-U, Meyburg C. 1984. Distribution et statut 69 12 formes) reveals extensive paraphyly at the genus level. actuels du vautour moine Aegypius monachus. Rapin. Med. 70 1310 10 Molecular Phylogenetics and Evolution 35: 147–164. 2: 26–31. 71 14 Helbig AJ, Seibold I, Kocum A, Liebers D, Irwin J, Mira S, Billot C, Guillemaud T, Palma L, Cancela ML. 72 15 Bergmanis U, Meyburg BU, Scheller W, Stubbe M, 2002. Isolation and characterization of polymorphic micro- 73 16 Bensch S. 2005b. Genetic differentiation and hybridization satellite markers in Eurasian vulture Gyps fulvus. Molecu- 74 17 between greater and lesser spotted eagles (Accipitriformes: lar Ecology Notes 2: 557–558. 75 18 Aquila clanga, A-pomarina). Journal of Ornithology 146: Moritz C. 1994. Defining ‘evolutionarily significant units’ for 76 1911 11 226–234. conservation. Trends in Ecology and Evolution 9: 373–375. 77 20 Hendrickson SL, Bleiweiss R, Matheus JC, de Matheus Moritz C. 1999. Conservation units and translocations: strat- 78 21 LS, Jacome NL, Pavez E. 2003. Low genetic variability in egies for conserving evolutionary processes. Hereditas 130: 79 22 the geographically widespread Andean Condor. Condor 105: 217–228. 80 23 1–12. Negro JJ, Torres MJ. 1999. Genetic variability and differ- 81 24 Hiraldo F. 1974. Colonias de cría y censo de los Buitres entiation of two bearded vulture Gypaetus barbatus 82 25 Negros (Aegypius monachus) en España. Naturalia His- populations and implications for reintroduction projects. 83 26 panica 2: 1–31. Biological Conservation 87: 249–254. 84 27 Jones ME, Paetkau D, Geffen E, Moritz C. 2004. Genetic Nesje M, Roed KH, Bell DA, Lindberg P, Lifjeld JT. 2000. 85 28 diversity and population structure of Tasmanian devils, the Microsatellite analysis of population structure and genetic 86 29 largest marsupial carnivore. Molecular Ecology 13: 2197– variability in peregrine falcons (Falco peregrinus). Animal 87 30 2209. Conservation 3: 267–275. 88 31 Komdeur J. 1994. Conserving the seychelles warbler Palumbi SR. 1996. Nucleic acids II: the polymerase chain 89 32 acrocephalus-sechellensis by translocation from Cousin reaction. In: Hillis DM, Moritz C, Mable BK, eds. Molecular 90 33 Island to the Islands of Aride and Cousine. Biological Con- systematics. Sunderland, MA: Sinauer, 205–248. 91 34 servation 67: 143–152. Piry S, Luikart G, Cornuet JM. 1999. BOTTLENECK: 92 35 Kretzmann MB, Capote N, Gautschi B, Godoy JA, A computer program for detecting recent reductions in the 93 36 Donazar JA, Negro JJ. 2003. Genetically distinct island effective population size using allele frequency data. 94 37 populations of the Egyptian vulture (Neophron percnop- Journal of Heredity 90: 502–503. 95 38 terus). Conservation Genetics 4: 697–706. Posada D, Crandall KA. 1998. MODELTEST: testing the 96 39 Kumar S, Tamura K, Nei M. 2004. MEGA3: Integrated model of DNA substitution. Bioinformatics 14: 817–818. 97 40 software for molecular evolutionary genetics analysis and Pritchard JK, Stephens M, Donnelly P. 2000. Inference 98 41 sequence alignment. Briefings in Bioinformatics 5: 150– of population structure using multilocus genotype data. 99 42 163. Genetics 155: 945–959. 100 43 Le Gouar P, Boisselier-Dubayle FR, Samadi MC, Arthur Raymond M, Rousset F. 1995. Genepop (version 1.2) – 101 44 S, Choisy C, Hatzofe JP, Henriquet O, Lecuyer S, population-genetics software for exact tests and ecumeni- 102 45 Tessier P, Susic CG, Sarrazin F. 2006. Genetics of cism. Journal of Heredity 86: 248–249. 103 46 restored populations of griffon vulture Gyps fulvus in Robertson HA, Karika I, Saul EK. 2006. Translocation of 104 47 France and Europe. In: Houston DC, Piper SE, eds. Pro- Rarotonga monarchs Pomarea dimidiata within the South- 105 48 ceedings of the international conference on conservation and ern Cook Islands. Bird Conservation International 16: 197– 106 49 management of vulture populations. Heraklion: Natural 215. 107 50 History Museum of Crete and WWF Greece, 116–126. Rohland N, Hofreiter M. 2007. Comparison and optimiza- 108 51 Lerner HRL, Mindell DP. 2005. Phylogeny of eagles, Old tion of ancient DNA extraction. BioTechniques 42: 343–352. 109 52 World vultures, and other Accipitridae based on nuclear and Ronquist F, Huelsenbeck JP. 2003. MrBayes 3: Bayesian 110 53 mitochondrial DNA. Molecular Phylogenetics and Evolution phylogenetic inference under mixed models. Bioinformatics 111 54 37: 327–346. 19: 1572–1574. 112 55 Lifjeld JT, Bjornstad G, Steen OF, Nesje M. 2002. Roques S, Negro JJ. 2005. MtDNA genetic diversity and 113 56 Reduced genetic variation in Norwegian peregrine falcons population history of a dwindling raptorial bird, the red kite 114 57 Falco peregrinus indicated by minisatellite DNA finger- (Milvus milvus). Biological Conservation 126: 41–50. 115 58 printing. Ibis 144: E19–E26. Rudnick JA, Katzner TE, Bragin EA, Rhodes OE, 116

14 N. POULAKAKIS ET AL.

1 Dewoody JA. 2005. Using naturally shed feathers for Pyrenean brown bear population. Molecular Ecology 6: 869– 28 2 individual identification, genetic parentage analyses, and 876. 29 3 population monitoring in an endangered Eastern imperial Tamura K, Nei M. 1993. Estimation of the number of nucle- 30 4 eagle (Aquila heliaca) population from Kazakhstan. Molecu- otide substitutions in the control region of mitochondrial- 31 5 lar Ecology 14: 2959–2967. DNA in humans and chimpanzees. Molecular Biology and 32 6 Seibold I, Helbig AJ. 1995. Evolutionary history of New and Evolution 10: 512–526. 33 7 Old World vultures inferred from nucleotide sequences of Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, 34 8 the mitochondrial cytochrome b gene. Philosophical Trans- Higgins DG. 1997. The CLUSTALX windows interface: 35 9 actions of the Royal Society of London Series B, Biological flexible strategies for multiple sequence alignment aided by 36 10 Sciences 350: 163–178. quality analysis tools. Nucleic Acids Research 25: 4876– 37 11 Snyder NFR, Derrickson SR, Beissinger SR, Wiley JW, 4882. 38 12 Smith TB, Toone WD, Miller B. 1996. Limitations of Vali U. 2002. Mitochondrial pseudo-control region in old 39 13 captive breeding in endangered species recovery. Conserva- world eagles (genus Aquila). Molecular Ecology 11: 2189– 40 14 tion Biology 10: 338–348. 2194. 41 15 Stockwell CA, Hendry AP, Kinnison MT. 2003. Con- Valiere N. 2002. GIMLET: a computer program for analysing 42 16 temporary evolution meets conservation biology. Trends in genetic individual identification data. Molecular Ecology 43 17 Ecology and Evolution 18: 94–101. Notes 2: 377–379. 44 18 Swofford DL. 2002. PAUP*. Phylogenetic analysis using par- Van Oosterhout C, Hutchinson WF, Wills DPM, Shipley 45 19 simony (*and other methods), Version 4. Sunderland, MA: P. 2004. MICRO-CHECKER: software for identifying and 46 20 Sinauer Associates. correcting genotyping errors in microsatellite data. Molecu- 47 21 Sánchez JJ. 2004. Buitre Negro, Aegypius monachus.In: lar Ecology Notes 4: 535–538. 48 22 Madroño A, González C, Atienza JC, eds. Libro Rojo de Waits LP, Luikart G, Taberlet P. 2001. Estimating the pro- 49 23 las Aves de España. Madrid: Dirección General para la bability of identity among genotypes in natural populations: 50 24 Biodiversidad-SEO/BirdLife. cautions and guidelines. Molecular Ecology 10: 249–256. 51 25 Taberlet P, Camarra JJ, Griffin S, Uhres E, Hanotte O, Weir BS, Cockerham CC. 1984. Estimating F-statistics for 52 26 Waits LP, Dubois-Paganon C, Burke T, Bouvet J. the analysis of population-structure. Evolution 38: 1358– 53 27 1997. Noninvasive genetic tracking of the endangered 1370. 54