Contrasting patterns of mitochondrial and microsatellite genetic ...

5 downloads 0 Views 347KB Size Report
PATRÍCIA H. BRITO. American Museum of Natural History, Ornithology Department, Central Park West at 79th Street, New York, NY 10024, USA. Abstract.
Molecular Ecology (2007) 16, 3423–3437

doi: 10.1111/j.1365-294X.2007.03401.x

Contrasting patterns of mitochondrial and microsatellite genetic structure among Western European populations of tawny owls (Strix aluco)

Blackwell Publishing Ltd

PATRÍCIA H. BRITO American Museum of Natural History, Ornithology Department, Central Park West at 79th Street, New York, NY 10024, USA

Abstract A recent study of mitochondrial phylogeography of tawny owls (Strix aluco) in western Europe suggested that this species survived the Pleistocene glaciations in three allopatric refugia located in Iberia, Italy, and the Balkans, and the latter was likely the predominant source of postglacial colonization of northern Europe. New data from seven microsatellite loci from 184 individual owls distributed among 14 populations were used to assess the genetic congruence between nuclear and mitochondrial DNA (mtDNA) markers. Microsatellites corroborated the major phylogeographical conclusions reached on the basis of the mtDNA sequences, but also showed important differences leading to novel inferences. Microsatellites corroborated the three major refugia and supported the Balkan origin of northern populations. When corrected for differences in effective population size, microsatellites and * = 0.12 mtDNA yielded generally congruent overall estimates of population structure (N ST vs. RST = 0.16); however, there was substantial heterogeneity in the RST among the seven nuclear loci that was not correlated with heterozygosity. Populations representing the Balkans postglacial expansion interact with populations from the other two refugia forming two clines near the Alps and the Pyrenees. In both cases, the apparent position of the contact zones differed substantially between markers due to the genetic composition of populations sampled in northern Italy and Madrid. Microsatellite data did not corroborate the lower genetic diversity of northern, recently populated regions as was found with mtDNA; this discrepancy was taken as evidence for a recent bottleneck recovery. Finally, this study suggests that congruence among genetic markers should be more likely in cases of range expansion into new areas than when populations interact across contact zones. Keywords: clines, genetic structure, isolation by distance, microsatellite, mtDNA; Strix Received 11 March 2007; revision accepted 20 April 2007

Introduction Genetic structure of natural populations is dependent on several demographic and stochastic processes such as mutation, genetic drift, and gene flow (Slatkin 1985). In birds, gene flow is usually the result of juveniles dispersing from their natal areas and, less frequently, adults changing their breeding locations (Johnson & Gaines 1990). Studies of

Correspondence: Patrícia H. Brito, Department of Organismic and Evolutionary Biology, Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, MA 02138, USA. Fax: (617) 4955667. E-mail: [email protected] © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

allozyme frequencies have detected little geographical structure in temperate birds, particularly when compared with other vertebrates (Barrowclough 1983). The use of fast-evolving markers, such as the mitochondrial control region, has indicated much higher levels of genetic differentiation (Zink 1997), and often mitochondrial phylogeographical studies discover clades with significant geographical information (e.g. Wenink et al. 1996; Barrowclough et al. 2004, 2005). Crochet (2000) has suggested that this difference between allozymes and mitochondrial results is due to a difference in effective population size between those markers, and proposed that the same levels of gene flow could produce the distributions of the F-statistics that are usually observed for those markers. Exceptions to

3424 P . H . B R I T O the strong geographical structure estimated with mitochondrial DNA (mtDNA) are common, but these examples are usually interpreted as due to recent range expansions and nonequilibrium populations (e.g. Zink et al. 2003). Microsatellites are popular markers for the study of geographical structure and gene flow because of their high levels of polymorphism and biparental inheritance (Jarne & Lagoda 1996). Microsatellites are codominant and inherited in a Mendelian fashion and, like allozymes, provide an easy way to screen multiple nuclear loci for geographical variation in natural populations. However, analysis of microsatellite data can be challenged by their complex mutation process (Palsbøll et al. 1999; Rubinsztein et al. 1999), by the difficulty to distinguish identical-by-descent from identical-by-state alleles (Estoup et al. 2002), and by their typically high mutation rates that may lead to underestimation of among-population differentiation (Hedrick 1999). Avian studies where microsatellites and mitochondrial results are compared frequently recover a pattern of reduced nuclear genetic structure (e.g. Helbig et al. 2001; Crochet et al. 2003; Eggert et al. 2004), although exceptions have also been reported for both similar levels of genetic structure (Burg & Croxall 2001) and higher levels of nuclear structure (e.g. Johnson et al. 2003). The importance of assessing the congruence among independent genetic markers has been widely discussed in the literature (e.g. Avise 2000). New data sets can lead to independent corroboration (or rejection) of previously established hypotheses, and combined analysis of multiple loci can reduce the variance in the estimated parameters that is due to random effects of sampling and lineage sorting (e.g. Takahata 1989; Hudson 1990). Also, contrasting results across markers, especially when those markers have distinct modes of inheritance, such as microsatellites and mtDNA, may provide new insights that could not be obtained with either type of data alone (Prugnolle & Meeus 2002). Discordant patterns across loci may have different causes. Molecular evolutionary factors associated with the markers of choice, such as the effective population size and the evolutionary rates can have important effects on the ability to detect genetic diversity and geographical structure. For instance, high genetic structure detected with mtDNA is frequently associated with its low effective population size that increases its susceptibility to random effects of drift (e.g. Haavie et al. 2000). Evolutionary rates can also have a strong effect on the ability to detect differentiation. Slow rates of evolution may not produce enough polymorphism to well characterize the existing genetic structure, and rates that are too high may even conceal the signal for high differentiation, a feature that is commonly associated with microsatellites (Hedrick 1999). Molecular incompatibilities between genome groups can also lead to distinct results if they cause differential introgression of

markers across hybrid zones (e.g. Fel-Clair et al. 1998; Payseur et al. 2004). Finally, ecological factors related to the way organisms behave, such as sex-biased dispersal and assortative mating, may also have important effects especially when markers are gender-specific (Prugnolle & Meeus 2002). Discordant patterns in avian studies are usually explained by sex-biased dispersal, which very often corroborates behavioural observations (e.g. male-philopatry and femalebiased dispersal of red grouse, Piertney et al. 2000), but can also reveal unexpected results that contradict current knowledge (e.g. Gibbs et al. 2000). Also, discrepancies between markers are commonly related to the presence of hybrid zones that may lead to the differential introgression of mtDNA and nuclear markers (Hansson et al. 2000), or to a reduced fitness of the hybrid females (avian heterogametic sex) as predicted by the Haldane’s rule (Helbig et al. 2001). In this study, microsatellites are contrasted with results from a previous phylogeographical study based on mitochondrial control region sequences (Brito 2005). The tawny owl (Strix aluco) is a widespread medium-sized owl associated with woodland forest distributed throughout Europe, North Africa, the Middle East and Asia (Cramp 1985). The present study analyses the western European populations of tawny owls (Strix aluco aluco, and Strix aluco sylvatica) whose recent history was shown to have been influenced by the climatic fluctuations of the Late Pleistocene (Brito 2005). The same study showed that western European populations of tawny owls are composed of three major mtDNA lineages that evolved in allopatry in three glacial refugia located in Iberia, southern Italy, and the Balkans (Fig. 1b). The latter was determined to be the sole source of postglacial expansion to northern Europe. Considerable secondary intergradation was also detected in midlatitude populations located in Austria, France, and Spain. This study contrasts results of genetic diversity, genetic structure, and inferences of recent history of European tawny owls estimated from both mitochondrial and microsatellite markers. Although comparing mtDNA with microsatellite results has become a common practice in studies of phylogeography and geographical structure, this study presents a methodological approach that explicitly explores congruence and contrast between microsatellite and mtDNA data, and examines variation among microsatellite loci. General expectations for multilocus phylogeography carried out in geographical areas characterized by the presence of suture zones are drawn.

Materials and methods Sample collection A total of 184 tawny owl samples were collected from 14 populations in western Europe, and sample sizes per © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3425 Fig. 1 Geographical variation of tawny owls across Western Europe, as measured by (a) microsatellite and (b) mitochondrial markers. Histograms show the frequency of the four most influential alleles in the principal component analysis. Histograms’ positions indicate approximate locations of sampled populations. Maps were adapted from Hewitt (1996; Fig. 5), and Hewitt (1999; Fig. 1b), to indicate suitable habitat during the last glacial maximum. Glacial refugia located in Iberia, southern Italy and the Balkans are indicated in green. Dashed line shows the southern limit of the permafrost. Map (b) was adapted from Brito (2005).

population are indicated on Table 1. These populations cover the three southern peninsulas of Iberia (Portugal, Spain-Bilbao, and Spain-Madrid), Italy (northern Italy and Sicily), and the Balkans (Greece) as well as four populations in northern Europe (England, Denmark, Norway, and Finland), three in France (France-north in Lille, France-west in Poitou Charente, and France–southeast in Vaucluse), and one in Austria (Fig. 1). Specific geographical locations of each sample are provided in Brito (2005); with the exception of three samples (FrW30, ItS21, and ItS26), all other individuals were analysed in this study. Fresh tissues (blood, growing contour feathers, or muscle) were collected whenever possible, but for the population in southern Italy (Sicily); six © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

samples were obtained from toe pads of museum study skins that dated from 1982 to 1997. Sampling was done from 2000 to 2002 with the help of local research teams carrying out projects with tawny owl populations, Rehabilitation Centers for Raptors, and Natural History Museums.

Microsatellite genotyping Seven microsatellite loci originally developed for Strix occidentalis (15A6, 1C6, 8G11, 4E10, and 4E10.2; Thode et al. (2002)) and Bubo bubo (Bb111, Bb131, Isaksson & Tegelström 2002) were genotyped. Microsatellite fragments were amplified using a touch-down polymerase chain reaction (PCR)

3426 P . H . B R I T O Table 1 Summary of genetic variation at seven microsatellite loci scored from tawny owl populations in western Europe: Sample size (N), average number of alleles (a),observed and expected heterozygosity (HO and HE, SD is standard deviation); number of alleles per locus (A); and number of private alleles (P), are provided for all populations. Departures from Hardy–Weinberg equilibrium were only found in two cases: France_SE for locus 8G11, and Denmark for locus 4E10

Refugial populations Portugal (Pt) Spain-Madrid (SpM) Spain-Bilbao (SpB) Italy-N (ItN) Italy-Sicily (ItS) Greece (Gr) Midlatitudes France-W (FrW) France-N (FrN) France-SE (FrSE) Austria (Au) Northern populations England (UK) Denmark (Dk) Norway (No) Finland (Fi)

Average over all loci

15A6

8G11

4E10.2

N

a

HO

HE (SD)

A

P

HO

HE

A

P

HO

HE

A

P

HO

HE

20 13 6 13 11 13

6.4 5.9 5.3 6.1 4.3 6.3

0.596 0.565 0.738 0.604 0.633 0.615

0.667 ± 0.113 0.703 ± 0.070 0.697 ± 0.110 0.628 ± 0.122 0.574 ± 0.116 0.640 ± 0.125

6 5 4 5 5 6

2 0 0 0 1 0

0.700 0.692 0.833 0.923 0.818 0.923

0.772 0.766 0.727 0.763 0.788 0.782

5 5 5 5 4 4

0 0 0 0 0 0

0.450 0.615 0.833 0.462 0.727 0.615

0.750 0.714 0.803 0.702 0.740 0.754

13 12 10 9 6 10

0 1 0 0 0 0

0.750 0.769 0.833 0.846 0.818 0.846

0.908 0.899 0.970 0.859 0.831 0.895

9 10 20 20

5.9 5.4 7.4 8.4

0.619 0.543 0.586 0.628

0.669 ± 0.093 0.632 ± 0.100 0.684 ± 0.102 0.665 ± 0.121

6 5 6 6

0 0 0 0

0.889 0.600 0.650 0.600

0.771 0.726 0.745 0.746

3 5 5 8

0 0 0 1

0.333 0.500 0.300 0.800

0.582 0.695 0.765 0.801

10 8 11 11

1 2 2 1

0.778 0.800 0.850 0.850

0.882 0.847 0.899 0.867

11 21 7 10

4.9 7.0 4.6 5.0

0.571 0.618 0.674 0.729

0.591 ± 0.132 0.667 ± 0.104 0.628 ± 0.114 0.701 ± 0.081

5 6 5 5

0 0 0 0

0.818 0.714 1.000 0.800

0.749 0.696 0.835 0.758

3 6 4 4

0 2 0 0

0.455 0.476 0.857 0.900

0.680 0.635 0.780 0.732

8 10 5 7

0 0 0 0

0.818 0.667 0.714 0.800

0.840 0.832 0.670 0.884

N, north; W, west; SE, southeast. Table 1 (horizontal continued) 1C6

Pt SpM SpB ItN ItS Gr FrW FrN FrSE Au UK Dk No Fi

4E10

Bb111

Bb131

A

P

HO

HE

A

P

HO

HE

A

P

HO

HE

A

P

HO

HE

5 6 5 6 3 6 5 4 6 8 6 7 5 6

0 0 0 0 0 0 0 0 0 1 1 0 0 0

0.850 0.615 1.000 0.769 0.818 0.769 0.778 0.700 0.800 0.800 0.909 0.952 0.714 0.900

0.772 0.837 0.849 0.791 0.541 0.763 0.843 0.700 0.746 0.810 0.797 0.813 0.791 0.805

12 7 8 14 6 14 11 11 19 20 8 14 9 9

0 1 0 0 0 1 0 0 0 0 1 0 0 0

0.850 0.727 1.000 0.846 1.000 0.769 0.889 0.800 0.900 0.947 0.818 0.850 1.000 0.800

0.897 0.753 0.924 0.932 0.783 0.935 0.909 0.916 0.939 0.962 0.892 0.932 0.912 0.911

2 3 2 2 1 2 3 3 3 4 2 4 2 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0

0.050 0.308 0.167 0.308 — 0.308 0.333 0.200 0.550 0.350 0.091 0.571 0.143 0.500

0.050 0.335 0.167 0.271 — 0.271 0.386 0.353 0.553 0.422 0.091 0.668 0.143 0.479

3 3 3 2 5 2 3 2 2 2 2 2 2 2

0 0 0 0 3 0 0 0 0 0 0 0 0 0

0.600 0.462 0.500 0.077 0.273 0.077 0.333 0.200 0.150 0.050 0.091 0.095 0.286 0.400

0.526 0.619 0.439 0.077 0.338 0.077 0.307 0.190 0.142 0.050 0.091 0.093 0.264 0.337

protocol designed to reduce nonspecific priming and subsequent arbitrary fragment amplification. Annealing temperatures in the first touch-down cycle were 8 °C above the final annealing temperature, which were 50 °C (15A6, 1C6, 4E10.2, and Bb111), 48 °C (Bb131), and 39 °C (4E10). Otherwise, PCR amplifications followed standard protocols from Brito (2005). HotStart Taq DNA polymerase (QIAGEN) was used in all genomic amplifications. PCRs were confirmed by electrophoresis of 5 µL of PCR product in an

agarose gel stained with ethidium bromide. Samples were diluted depending on the brightness of the PCR bands. Diluted PCR products were combined with an internal labelled size standard (500 ROX, ABI), and loaded for genescan in an ABI 3700 automated sequencer (ABI). Multiplex genescan, when used, were done with microsatellites labelled with different fluorescent dyes. Microsatellites fragments were scored using genotyper version 2.0 (PE Biosystems). All samples were PCR-amplified and scored at least twice, and © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3427 the six samples extracted from toe pads were scored three times to reduce the chances of having null alleles. Due to the recent age of those study skins, it was found that the multiple-tubes approach for accurate genotyping of very small DNA samples by using PCR, as described by Navidi et al. (1992) and Taberlet et al. (1996), was not necessary.

DNA Sequencing Because these microsatellite loci had not been previously tested in tawny owls, each locus was sequenced for selected homozygous individuals to confirm that the amplified repeats were similar to the ones described for the original species. Repeats for each locus were confirmed after sequencing between two and eight individuals representing a minimum of two distinct populations. PCRs were carried out as previously indicated and sequencing was performed using standard procedures described in Brito (2005).

Data analysis micro-checker software (Oosterhout et al. 2004) was used to assess the presence of genotyping errors, such as nonamplified alleles, short allele dominance, and scoring of stutter peaks. Microsatellite genetic diversity was quantified as both absolute (A) and mean number (a) of alleles per locus, observed heterozygosity (HO), unbiased gene diversity (HE), frequency of most common allele (Xc), and number of private alleles; these statistics were estimated using microsatellite toolkit version 3.1 for PC Microsoft Excel (Park 2001). Linkage disequilibrium between loci, and deviations from Hardy–Weinberg genotype frequency equilibrium (HWE) were tested with genepop (Raymond & Rousset 1995). For all these analyses, significance was evaluated by Fisher exact tests where P values were estimated by applying a Markov chain method with the following parameters: 10 000 dememorization; 200 batches; 5000 interactions per batch. Population genetic structure was estimated with RST (Slatkin 1995) using the program rst-calc (Goodman 1997). This statistic is particularly useful for microsatellite data because unlike FST it assumes a stepwise mutation model; also, rst-calc standardizes the variance of allele size prior to calculation, making loci with different variances comparable. Bootstrap (1000 replicates) was used to estimate the 95% confidence intervals for overall RST, and permutation test (1000 permutations) was used to determine if overall RST values across loci are significantly different from zero. Sequential Bonferroni corrections were applied to correct for type I error when appropriate (Rice 1989). structure version 2 (Pritchard et al. 2000) applies a Bayesian method to infer the number, K, of clusters without using prior information of the individual sampling locations. This program distributes individuals among K clusters based on their allelic frequencies, and estimates the posterior © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

probability of the data given each particular K. structure was run for K = 1 to K = 14 clusters. Each run was pursued for 1 million MCMC interactions, with an initial burn-in of 100 000, and an ancestry model that allows for admixture, with the same Alpha (Dirichelet parameter for degree of admixture) for all populations. Five independent simulations were run for each K to assess stability. The final posterior probability of K was computed as suggested by Pritchard et al. 2000), using the runs with highest probability for each K. An additional ad hoc statistic (∆K) was also estimated because it was shown to provide a better predictor of the number of groups (i.e. K) at the uppermost hierarchical level (Evanno et al. 2005). The sensitivity of the final result to specific prior assumptions of alpha and independence of allelic frequencies was also computed. Population structure was further analysed by performing a principal component analysis (PCA) on the allelic frequencies for each population. The data were arcsine square-root-transformed (angular transformation) before analysis and used as input data for the PCA on the covariance matrix. These analyses were performed with sas version 8.02 using the procedure PRINCOMP, and the angular transformation was chosen due to its applicability to percentage and proportion data (Sokal & Rohlf 1995, p. 419). Pairwise estimates of gene flow and geographical distances were then used to infer patterns of isolation by distance as well as to identify possible events of recent population expansion (Slatkin & Maddison 1990; Slatkin 1993). Because sample points are not independent, significance of the correlation coefficient was estimated with Mantel tests (Mantel 1967), 1000 replicates, as performed in genepop (Raymond & Rousset 1995). Nem was estimated from both pairwise RST (microsatellites) and NST (mtDNA), and the program inverse, available from the National Geodetic Survey, US Department of Commerce (www.ngs.noaa.gov), was used to estimate geographical distances in kilometres between pairs of locality coordinates. Isolation–by-distance plots were made for a specific group of populations that, following earlier mtDNA results, reflect the Balkan expansion: Greece, Austria, France-north, England, Denmark, and Finland. The Norway population was excluded from this analysis due to its small sample size (seven individuals). To reflect geography in western Europe, geographical distances were estimated using a stepping-stone model (see Fig. 1).

Results Microsatellite gene diversity The seven microsatellite loci analysed yielded variable levels of genetic diversity within population samples (Table 1). Loci 4E10.2 and 4E10 showed high levels of polymorphism with 5–20 alleles per population and expected heterozygosities

3428 P . H . B R I T O ranging from 0.67 to 0.97. Loci 15A6, 8G11, and 1C6 showed medium levels of polymorphism with 3 –8 alleles per population and expected heterozygosities that ranged from 0.54 and 0.85. Finally, loci Bb111 and Bb131 had low levels of polymorphism with only 1– 5 alleles per population and expected heterozygosities ranging from 0.00 to 0.67. Overall, the mean number of alleles per locus averaged 5.92, ranging from 4.29 in Italy-Sicily to 8.43 in Austria, and expected heterozygosity averaged across loci ranged from 0.57 (Italy-Sicily) to 0.70 (Spain-Madrid). Two pairs of individuals from Denmark had identical alleles for all seven loci, but only one of those pairs shared identical mtDNA haplotypes. No population was fixed for any allele, and private alleles did not account for more than 15% of the allelic frequency in a population; in general, rare alleles were rare everywhere, indicating that the frequency of those alleles did not change geographically. Exact tests of genotypic disequilibrium between loci were not significant suggesting that the loci were independent. Sequential Bonferroni corrections were applied to both linkage disequilibrium and Hardy–Weinberg frequency tests. Departures from Hardy–Weinberg equilibrium were only found in two cases: France-southeast for locus 8G11, and Denmark for locus 4E10. Global tests across populations revealed significant departure from equilibrium frequencies at the same two loci 8G11 and 4E10. Allele frequencies for all seven loci are shown in the Appendix. Microsatellites did not show a pattern of decreasing genetic diversity (HE) with latitude; the four most northern populations (England, Denmark, Norway, and Finland) had genetic diversities that were not distinguishable from the remaining populations (Table 1, Fig. 2). The contrast between mitochondrial and nuclear genetic diversities showed

an apparent weak correlation when all populations are included in the analysis (r2 = 0.23, P = 0.085; Fig. 2). A closer inspection of the data reveals, however, that microsatellites and mtDNA mostly agreed in their assessment of the genetic diversity of refugial and midlatitude populations (r2 = 0.66, P = 0.004; Fig. 2), leaving the northern populations as the cause of the disagreement between markers.

Population genetic structure F-statistics. Estimated levels of population genetic structure varied considerably across microsatellite loci (Fig. 3). RST averaging over all seven loci was 0.16 (P 5, there is probably not a single preferred clustering solution. The posterior probability of K was essentially one for K = 3, and zero for all remaining K. ∆K (Evanno et al. 2005) was computed for all K and also indicated a strong signal for K = 3 (Fig. 4b). Changing the assumptions of ‘equal alpha for each population’ and ‘correlated allele frequencies’ did not change this final result. Although there was a strong signal for K = 3, the proportion of membership of each predefined population in each of the three clusters did not have a simple geographical interpretation. Whereas an Iberian cluster was well differentiated, Iberian populations had on average 80% of their individuals in a single cluster, all other populations were of mixed origin between the two remaining clusters, and these could not be associated either with specific populations or geographical regions.

Multivariate analysis The PCA performed on the population allelic frequencies revealed that the 14 tawny owl populations sampled in western Europe are structured in three groups (Fig. 5) that correspond to the three Pleistocene refugia identified for tawny owls in a previous study (Brito 2005). Iberia, southern Italy (Sicily), and Greece are well differentiated on the first and second principal components, which together explained 39% of the total genetic variance. The total genetic variance explained by the first and second principal components were 25% and 14%, respectively. All midlatitude and northern European populations formed a cluster with Greece. Interestingly, the northern Italy population was more similar to Greece than to southern Italy. The alleles with most influence in the discrimination of the three clusters, as indicated by the first two eigenvectors’ coefficients, were not lineage-specific alleles but rather

Fig. 4 Bayesian inference of the number of clusters (K) of tawny owls. K was estimated using the (a) the posterior probability of the data given each K (five replicates), and (b) the distribution of ∆K.

widespread alleles that had distinct frequencies in the three clusters (e.g. alleles 151 and 155/locus Bb131; allele 129/locus 8G11; and allele 205/locus 4E10; Fig. 1a). Clines. Both mtDNA and microsatellite loci identified a structure of three major groups of tawny owls in southern Europe, and a northward expansion that led to secondary contact in the midlatitudes populations. However, the two sets of markers differed in their geographical placement of the major transition between Balkan genotypes and those of Iberia and Italy (Fig. 6). In a transect from Portugal to Greece, mtDNA showed a steep transition between the populations of Portugal and Madrid, whereas microsatellites placed the contact zone farther east and much closer to the Pyrenees. In the transect between southern Italy and Greece, microsatellites place the contact zone in the south of the Italian peninsula, while mtDNA position it closer to the Alps, between the populations of northern Italy and Austria. Isolation by distance. Pairwise comparison of gene flow estimates (Nem) and geographical distance when plotted on a log–log scale can be used to depict patterns of isolation by distance in equilibrium and nonequilibrium populations (Slatkin 1993). In this study, this approach was applied to © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3431 Fig. 5 Scores of 14 populations of tawny owl on the first two principal components extracted from a matrix of 91 scored alleles from seven microsatellite loci. Populations are represented by different symbols to indicate geography: Iberia (asterisks), Italy (triangles), Greece (square), and nonrefugial populations (circles). Abbreviations are as in Table 1.

Fig. 6 Microsatellite vs. mitochondrial profiles across two clines in the tawny owl. Replacement of (a) Iberian and (b) Italian genomes by Balkans genomes across western Europe as measured by seven nuclear microsatellite loci (black squares) and control region mtDNA sequences (white circles). Microsatellite results are measured by PC1 scores, and mtDNA by the proportion of Balkans haplotypes over (a) Iberian and (b) Italian haplotypes in each population. Error bars indicate standard errors computed using a binomial distribution (mtDNA), and the PC1 eigenvector multiplied by the individual allele frequencies (microsatellites). The distance between population values is proportional to geographical distance. Populations are indicated by their abbreviations (see Table 1 for full names).

a specific group of tawny owl populations that reflects the Balkan expansion (Fig. 7). Scatter plots were congruent between marker classes in showing no detectable pattern of isolation by distance among populations (Mantel’s tests were not significant; P = 0.38 and P = 0.86 for mitochondrial and nuclear markers, respectively).

Avise 2000). Likewise, the lack of concordance can also provide a better understanding of the underlying evolutionary processes by revealing events of sex-biased dispersal, assortative mating, or local effects of selection usually detected by the differential introgression of markers across hybrid zones.

Discussion

Congruent patterns

One of the advantages of looking for concordance across independent loci within a species is that congruent patterns are likely to describe common historical events and not idiosyncrasies of the lineage sorting process (reviewed in

The combined analyses of the seven nuclear microsatellite loci corroborated the major findings of the mitochondrial phylogeography of tawny owls in western Europe (Table 2). Both mitochondrial and nuclear markers showed the

© 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

3432 P . H . B R I T O Fig. 7 Analyses of isolation by distance estimated from mitochondrial (white circles; r2 = 0.01) and nuclear (black circles; r2 = 0.03) markers. Gene flow (Nem) is plotted against geographical distance (Km) on a log-log scale. Population pairs reflect the Balkan expansion, and include Greece, Austria, France-north, England, Denmark, and Finland.

existence of three groups of tawny owl populations that correspond to the three proposed Pleistocene refugia of Iberia, Italy, and Balkans. The Balkans was also confirmed as the main source of postglacial colonization of northern Europe, and two clines were detected with both markers, between Iberia and Balkans, and Italy and the Balkans. Although congruence was found on the major phylogeographical conclusions, historical inferences from analysis of microsatellite data required more assumptions about the markers’ molecular evolution than analysis of mitochondrial data. Microsatellites are highly homoplasious with a mutation process that is not well understood (Palsbøll et al. 1999; Rubinsztein et al. 1999; Weetman et al. 2002) but current analyses assume that similar alleles are identical-by-descent, an assumption that is likely violated in the presence of deep historical structures. The analyses that corroborated the three Pleistocene refugia hypothesis were the Bayesian clustering analysis (structure), and the multivariate analysis (PCA). structure has the advantage of using individuals as the unit of analysis, and it allows the inference of population structure without using the geographical location of sampled individuals. However, in this study, although it has indicated a strong signal for a substructure of three clusters (K = 3) only one cluster could be easily associated to a particular geographical region (Iberia). The failure to unambiguously identify the other two clusters may be related to the fact that differentiation is caused by different frequencies of widespread alleles, and not by lineage specific alleles (Fig. 1a). Similar results have been associated with allele frequencies that vary gradually across geography as in situations of isolationby-distance and hybrid zones (Pritchard & Wen 2004). The PCA was carried out on the population allele frequencies and clearly identified three groups of populations. However,

the historical relationship among these groups is unknown and only assumed to be hierarchical. That is, the corroboration of the three refugia hypothesis with the PCA results assumed that those microsatellite clusters are equivalent to three clades defined by outgroup analysis. In conformity with the mitochondrial data, microsatellite loci provided independent support of a Balkan origin of northern European populations. This conclusion is based on the PCA and on the pairwise RST results that showed northern European populations more similar to the populations sampled in Greece than to Iberian or southern Italy populations. Genetic signatures of range expansion from a Balkan refuge were also identified for both mitochondrial and nuclear data from analyses of isolation by distance (Fig. 7): relatively large values of gene flow were estimated among this group of populations but gene flow was independent of geographical distance, suggesting that populations had recently colonized the area (Slatkin 1993). With time, as populations reach equilibrium between drift and gene flow, patterns of isolation by distance are expected to arise (Wright 1943; Slatkin 1993).

Contrasting patterns Conflicting results between microsatellite and mitochondrial results were found (i) on the quantification of the genetic diversity of northern European populations, (ii) on the estimation of overall geographical structure, and (iii) on the geographical location of the two contact zones detected in western Europe. Genetic diversity estimates of populations sampled in northern Europe differed substantially between mitochondrial and microsatellite markers. The nuclear ‘expected heterozygosity’ and the mitochondrial ‘nucleotide diversity’ © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3433 computed for all 14 populations were weakly correlated, suggesting that microsatellite data could not be used to predict mitochondrial genetic diversity, and vice versa. However, if northern European populations were excluded from the analysis, those statistics became positively and significantly correlated. The lower mitochondrial genetic diversity of northern European populations is congruent with their recent origin from a subset of the original gene pool (Hewitt 1996; Brito 2005). And, the higher levels of genetic diversity estimated with microsatellites for those populations can be explained by these markers’ typically high mutation rate because recovery rate of neutral variation when population size rapidly increases after a sudden reduction is proportional to the reciprocal of the mutation rate (Nei et al. 1975; Lande & Barrowclough 1987). Overall RST (averaged over seven microsatellite loci) was substantially smaller than mitochondrial NST as expected from nuclear markers due to their difference in effective population size (Birky et al. 1983; Chesser & Baker 1996; Crochet 2000). After calibrating NST to its expected result * ), these two if nuclear genes had been sampled ( N ST measures of overall geographical structure became quite similar with the mitochondrial result being now slightly lower than the microsatellite RST; hence, taking into account the lower effective population size of mtDNA, the mitochondrial FST did not substantially differ from that of microsatellites. Discordance between mitochondrial and microsatellite loci was found in the estimated locations of two contact zones of tawny owls in western Europe. Most of the disagreement between mitochondrial and nuclear markers was due to the genetic composition of populations sampled in northern Italy and Madrid whose allelic frequencies made the contact zones shift south to north in Italy and west to east in the Iberian Peninsula. Population genetics theory predicts that hybrid zones tend to move to natural barriers characterized by low density and dispersal (Barton & Hewitt 1985). In Europe, the Pyrenees and the Alps are likely barriers for temperate species such as tawny owls; hence, mitochondrial results are more consistent with theoretical expectations in the location of the Italy– Balkans contact zone, whereas microsatellites recovered more predictable results in the placement of the Iberia– Balkans contact zone. Although, more detailed sampling in the Pyrenees and in the Alps would certainly be required to resolve questions about the shape and central position of these clines, the data in hand are still sufficient to suggest that the two sets of genetic markers place those centres at rather distinct distances from the two major mountain ranges in western Europe. Noncoincident clines such as the ones detected for tawny owls have already been reported in the literature for hybrid zones where cytoplasmic and nuclear markers were both analysed (e.g. Harrison 1990; Ross & Harrison 2002; Morgan-Richards & Wallis 2003; Sequeira et al. 2005). In general, these cline © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

shifts are attributed to the lower probability of mitochondrial genes being linked with nuclear genes under selection (Barton & Jones 1983), or to asymmetrical introgression either due to the displacement of the whole hybrid zone or to asymmetrical mating (Barton 1993). Some degree of discordance across loci is expected due to the stochastic nature of the genetic inheritance. Also, results from microsatellite data reflect the average over several loci, but mtDNA markers are linked in a single linkage unit, which prevents the sampling variance due to the coalescent process to be account for. The lower mitochondrial structure and the different locations of the contact zones are hypotheses that can be further tested with a more extensive population sampling across the two clines and by screening other independent markers.

Contrasting microsatellite loci If PCA is performed on the population allelic frequencies obtained for each locus, then it is possible to observe that the historical signal was not uniform across microsatellite loci. Levels of polymorphism and signatures of population structure varied considerably (Table 2). The congruence between mitochondrial and nuclear markers in the identification of three refugia was obtained in the combined analysis, but with only two of the seven individual microsatellite loci analysed. Also, only three of the seven loci showed an unequivocal signal for the Balkans postglacial expansion to northern Europe. Heterozygosity was not a good indicator of the effectiveness of each microsatellite locus in detecting genetic structure among populations as it might be expected if differentiation were a function of population variability. The two loci with highest genetic diversity (4E10.2 and 4E10) were at the two extremes of the RST scale, and therefore even if some ascertainment bias was committed in the decision to analyse the more polymorphic loci, that bias most likely did not have a large effect on the estimates of geographical structure. Also, because processes like population history, demography, and migration are expected under neutral theory to lead to nearly identical F-statistics (Lewontin & Krakauer 1973), the observed variance among loci was either due to sampling error or to size homoplasy that may have distinct effects in the different loci; only by increasing the number of loci one would increase the likelihood that random effects of error or noise, like size homoplasy, would be counteracted. Finally, this study suggests that congruence among genetic markers is more likely in cases of range expansion into areas not previously occupied (e.g. the Balkans expansion into northern Europe), than when populations interact across contact zones. In the former, population expansion suffers no major obstructions whereas population expansion across contact zones leads to introgression

3434 P . H . B R I T O of one genome group into an area already occupied by another. The formation of hybrid zones with associated cline widths and centres are then dependent on the balance between selection and dispersal (Endler 1977), and therefore may produce different signals across loci (Harrison 1990). This observation has important consequences for multilocus phylogeographical studies carried out in geographical regions that, like western Europe, are characterized by the presence of several suture zones.

Acknowledgements I am very grateful to numerous individuals and institutions that very generously contributed with tawny owl samples or facilitate with the sample collection. A comprehensive list is found in Brito (2005). Without their help this study would have not been possible. I thank G. Barrowclough for advice and support throughout this research, and R. DeSalle for granting me access to his laboratory where all microsatellite genotype was done. G. Barrowclough, R. Zink, J. Groth, S. Edwards, the associated editor (S. Bensch), and three anonymous reviewers provided critical reading of early versions of this manuscript. C. Pomilla provided great help in the lab, and L. Liu offered advice with the statistical analyses. This research received generous support from the Frank M. Chapman Memorial Fund at the American Museum of Natural History, Fundação para a Ciência e Tecnologia, Portugal. This paper is a contribution from the Monell Molecular Laboratory and the Cullman Research Facility in the Department of Ornithology, American Museum of Natural History, and has received generous support from the Lewis B and Dorothy Cullman Program for Molecular Systematics Studies, a joint initiative of The New York Botanical Garden and The American Museum of Natural History.

References Avise JC (2000) Phylogeography. The History and Formation of Species. Harvard University Press, Cambridge, Massachusetts. Barrowclough G (1983) Biochemical studies of microevolutionary processes. In: Perspectives in Ornithology (eds Brush AH, Clark GA Jr), pp. 223–261. Cambridge University Press, New York. Barrowclough GF, Groth JG, Mertz LA, Gutiérrez RJ (2004) Phylogeographic structure, gene flow and species status in blue grouse (Dendragapus obscurus). Molecular Ecology, 13, 1911–1922. Barrowclough GF, Groth JG, Mertz LA, Gutiérrez RJ (2005) Genetic structure, introgression, and a narrow hybrid zone between northern and California spotted owls (Strix occidentalis). Molecular Ecology, 14, 1109 –1120. Barton NH (1993) Why species and subspecies? Current Biology, 3, 797–799. Barton NH, Hewitt GM (1985) Analysis of hybrid zones. Annual Review of Ecology and Systematics, 16, 113 –148. Barton NH, Jones JS (1983) Mitochondrial DNA: new clues about evolution. Nature, 306, 317 – 318. Birky CW Jr, Maruyama T, Fuerst P (1983) An approach to population and evolutionary genetic theory for genes in mitochondria and chloroplasts, and some results. Genetics, 103, 513–527. Brito PH (2005) The influence of Pleistocene glacial refugia on tawny owl genetic diversity and phylogeography in western Europe. Molecular Ecology, 14, 3077 – 3094.

Burg TM, Croxall JP (2001) Global relationships amongst blackbrowed and grey-headed albatrosses: analysis of population structure using mitochondrial DNA and microsatellites. Molecular Ecology, 10, 2647 – 2660. Chesser RK, Baker RJ (1996) Effective sizes and dynamics of uniparentally and diparentally inherited genes. Genetics, 144, 1225 –1235. Cramp S (1985) Strix aluco tawny owl. In: The Birds of the Western Palearctic, pp. 526 – 546. Oxford University Press, Oxford, UK. Crochet PA (2000) Genetic structure of avian populations — allozymes revisited. Molecular Ecology, 9, 1463 –1469. Crochet P-A, Chen JZ, Pons J-M et al. (2003) Genetic differentiation at nuclear and mitochondrial loci among large white-headed gulls: sex-biased interspecific gene flow. Evolution, 57, 2865– 2878. Eggert LS, Mundy NI, Woodruff DS (2004) Population structure of loggerhead shrikes in the California Channel Islands. Molecular Ecology, 13, 2121– 2133. Endler JA (1977) Geographic Variation, Speciation, and Clines. Princeton University Press, Princeton, New Jersey. Estoup A, Jarne P, Cornuet J-M (2002) Homoplasy and mutation model at microsatellite loci and their consequences for population genetics analysis. Molecular Ecology, 11, 1591–1604. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software structure: a simulation study. Molecular Ecology, 14, 2611– 2620. Fel-Clair F, Catalan J, Lenormand T, Britton-Davidian J (1998) Centromeric incompatibilities in the hybrid zone between house mouse sybspecies from Denmark: evidence from patterns of nor activity. Evolution, 52, 592 – 603. Gibbs HL, Dawson RJG, Hobson KA (2000) Limited differentiation in microsatellite DNA variation among northern populations of the yellow warbler: evidence for male-biased gene flow? Molecular Ecology, 9, 2137 – 2147. Goodman SJ (1997) rst calc: a collection of computer programs for calculating unbiased estimates of genetic differentiation and determining their significance for microsatellite data. Molecular Ecology, 6, 881– 885. Haavie J, Saetre G-P, Moum T (2000) Discrepancies in population differentiation at microsatellites, mitochondrial DNA and plumage colour in the pied flycatcher — inferring evolutionary processes. Molecular Ecology, 9, 1137 –1148. Hansson M, Bensch S, Brännström O (2000) Range expansion and the possibility of an emerging contact zone between two subspecies of Chiffchaff Phylloscopus collybita ssp. Journal of Avian Biology, 31, 548 – 558. Harrison RG (1990) Hybrid zones: windows on evolutionary process. In: Oxford Surveys in Evolutionary Biology (eds Futuyma D, Antonovics J), pp. 69 –128. Oxford University Press, Oxford, UK. Hedrick PW (1999) Perspective: highly variable loci and their interpretation in evolution and conservation. Evolution, 53, 313– 318. Helbig A, Salomon M, Bensch S, Seibold I (2001) Male-biased gene flow across an avian hybrid zone: evidence from mitochondrial and microsatellite DNA. Journal of Evolutionary Biology, 14, 277– 287. Hewitt GM (1996) Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society, 58, 247 – 276. Hewitt GM (1999) Post-glacial re-colonization of European biota. Biological Journal of the Linnean Society, 68, 87 –112. © 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3435 Hudson RR (1990) Gene genealogies and the coalescent process. In: Oxford Surveys in Evolutionary Biology (eds Futuyma D, Antonovics J), pp. 1– 44. Oxford University Press, New York. Isaksson M, Tegelström H (2002) Characterization of polymorphic microsatellite markers in a captive population of the eagle owl (Bubo bubo) used for supportive breeding. Molecular Ecology Notes, 2, 91–93. Jarne P, Lagoda PJL (1996) Microsatellites, from molecules to populations and back. Trends in Ecology & Evolution, 11, 424 – 429. Johnson M, Gaines M (1990) Evolution of dispersal: theoretical models and empirical tests using birds and mammals. Annual Review of Ecology and Systematics, 21, 449 – 480. Johnson JA, Toepfer JE, Dunn PO (2003) Contrasting patterns of mitochondrial and microsatellite population structure in fragmented populations of greater prairie-chickens. Molecular Ecology, 12, 3335–3347. Lande R, Barrowclough GF (1987) Effective population size, genetic variation, and their use in population management. In: Viable Populations for Conservation (ed. Soulé ME), pp. 87 –123. Cambridge University Press, New York. Lewontin RC, Krakauer J (1973) Distribution of gene frequency as a test of the theory of gene neutrality of polymorphisms. Genetics, 74, 175–195. Mantel N (1967) The detection of disease clustering and a generalized regression. Cancer Research, 27, 209 – 220. Morgan-Richards M, Wallis GP (2003) A comparison of five hybrid zones of the weta Hemideina thoracica (Orthoptera: Anostostomatidae): degree of cytogenetic differentiation fails to predict zone width. Evolution, 57, 849 – 861. Navidi W, Arnheim N, Waterman MS (1992) A multiple-tubes approach for accurate genotyping of very small DNA samples by using PCR: statistical considerations. American Journal of Human Genetics, 50, 347 – 359. Nei M, Maruyama T, Chakraborty R (1975) The bottleneck effect and genetic variability in populations. Evolution, 29, 1–10. Oosterhout CV, Hutchinson WF, Willis DPM, Shipley P (2004) micro-checker: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes, 4, 535 – 538. Palsbøll PJ, Bérubé M, Jørgensen H (1999) Multiple levels of singlestrand slippage at cetacean tri- and tetranucleotide repeat microsatellite loci. Genetics, 151, 285 – 296. Park SDE (2001) Trypanotolerance in West African cattle and the population genetic effects of selection. PhD Dissertation, University of Dublin, Dublin, Ireland. Payseur BA, Krenz JG, Nachmani MW (2004) Differential patterns of introgression across the X chromosome in a hybrid zone between two species of house mice. Evolution, 58, 2064 – 2078. Piertney SB, MacColl ADC, Bacon PJ et al. (2000) Matrilineal genetic structure and female-mediated gene flow in red grouse (Lagopus scoticus): an analysis using mitochondrial DNA. Evolution, 54, 279–289. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics, 155, 945 – 959. Pritchard JK, Wen W (2004) Documentation for structure Software, Version 2. http://pritch.bsd.uchicago.edu Prugnolle F, Meeus Td (2002) Inferring sex-biased dispersal from population genetic tools: a review. Heredity, 88, 161–165. Raymond M, Rousset F (1995) genepop version 1.2.: population

© 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

genetics software for exact tests and ecumenicism. Journal of Heredity, 86, 248 – 249. Rice WR (1989) Analyzing tables of statistical tests. Evolution, 43, 223 – 225. Ross CL, Harrison RG (2002) A fine-scale spatial analysis of the mosaic hybrid zone between Gryllus firmus and Grillus pennsylvanicus. Evolution, 56, 2296 – 2312. Rubinsztein DC, Amos B, Cooper G (1999) Microsatellite and trinucleotide-repeat evolution: evidence for mutational bias and different rates of evolution in different lineages. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 354, 1095 –1099. Sequeira F, Alexandrino J, Rocha S, Arntzen JW, Ferrand N (2005) Genetic exchange across a hybrid zone within the Iberian endemic golden-striped salamander, Chioglossa lusitanica. Molecular Ecology, 14, 245 – 254. Slatkin M (1985) Gene flow in natural populations. Annual Review of Ecology and Systematics, 16, 393 – 430. Slatkin M (1993) Isolation by distance in equilibrium and nonequilibrium populations. Evolution, 47, 264 – 279. Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139, 457 – 462. Slatkin M, Maddison W (1990) Detecting isolation by distance using phylogenies of genes. Genetics, 126, 249 – 260. Sokal RR, Rohlf FJ (1995) Biometry. The Principles and Practice of Statistics in Biology Research, 3rd edn. W. H. Freeman, New York. Taberlet P, Griffin S, Goossens B et al. (1996) Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Research, 24, 3189 – 3194. Takahata N (1989) Gene genealogy in three related populations: consistency probability between gene and population trees. Genetics, 122, 957 – 966. Thode A, Maltbie M, Hansen L, Green L, Longmire J (2002) Microsatellite markers for the Mexican spotted owl (Strix occidentalis lucida). Molecular Ecology Notes, 2, 446 – 448. Weetman D, Hauser L, Carvalho GR (2002) Reconstruction of microsatellite mutation history reveals a strong and consistent deletion bias in invasive clonal snails, Potamopyrgus antipodarum. Genetics, 162, 813 – 822. Wenink P, Baker A, Rösner H-U, Tilanus M (1996) Global mitochondrial DNA phylogeography of holarctic breeding dunlins (Calidris alpina). Evolution, 50, 318 – 330. Wright S (1943) Isolation by distance. Genetics, 28, 139 –156. Wright S (1951) The genetic structure of populations. Annals of Eugenics, 15, 323 – 354. Wright S (1978) Variability Within and Among Natural Populations. The University of Chicago Press, Chicago. Zink RM (1997) Phylogeographic studies of North American birds. In: Avian Molecular Evolution (ed. Mindell DP), pp. 301– 324. Academic Press, San Diego, California. Zink RM, Drovetski SV, Questiau S et al. (2003) Recent evolutionary history of the bluethroat (Luscinia svecica) across Eurasia. Molecular Ecology, 12, 3069 – 3075.

Patrícia H. Brito is interested in the application of population genetics, phylogeographical, and systematic methods to the study of speciation and hybridization, geographical variation, and gene flow in natural populations of birds.

3436 P . H . B R I T O

Appendix Microsatellite allele frequencies across loci and populations of tawny owl Locus Allele Portugal Spain M Spain B Italy N Italy S Greece France_W France_N France_SE Austria England Denmark Norway Finland Total 15A6

117 121 125 129 133 137 141 145 149 153 8G11 125 129 133 137 141 145 149 153 157 4E10.2 150 155 160 165 170 175 180 185 190 195 200 205 210 215 220 225 230 235 240 245 255 260 270 1C6 96 99 102 105 108 111 114 117 120 123 4E10 160 185 190 195 200 200 205 210 215 220

— — 0.025 0.150 0.175 0.375 0.225 0.050 — — — 0.325 — 0.100 0.300 0.250 0.025 — — — — 0.050 0.025 0.125 0.025 — — — 0.025 0.100 0.125 0.050 0.025 0.200 0.125 0.025 0.100 — — — — — — 0.225 0.175 0.375 0.100 0.125 — — — — — 0.075 0.100 0.025 0.175 0.175 0.225 0.100 0.075 0.075

— — — — 0.192 0.308 0.346 0.077 0.077 — — 0.423 0.038 — 0.308 0.192 0.038 — — 0.038 — 0.077 — — — 0.038 — — 0.077 — 0.115 0.038 0.038 0.154 0.038 0.038 0.269 0.077 — — — — — 0.077 0.077 0.269 0.192 0.192 0.192 — — — 0.045 0.045 0.045 — 0.227 0.227 0.455 0.091 — 0.091

— — — — 0.167 0.500 0.167 0.167 — — — 0.167 0.083 — 0.417 0.167 0.167 — — — — 0.083 0.083 0.167 — 0.083 — — — — 0.083 0.083 0.083 0.083 0.083 — 0.167 — — — — — — 0.250 — 0.250 0.250 0.167 0.083 — — — — 0.083 0.167 — 0.083 0.083 0.167 0.083 0.250 0.083

0.038 — — — — 0.308 0.115 0.346 0.192 — — 0.038 — 0.038 0.346 0.423 0.154 — — — — — — — — — — 0.038 0.077 — 0.038 0.115 0.231 0.154 0.269 0.038 0.038 — — — — — — 0.038 0.077 0.308 0.308 0.192 0.077 — — — — — 0.038 — 0.077 0.077 0.077 0.038 — 0.192

0.318 0.045 — — 0.273 0.136 0.227 — — — — — — 0.091 0.227 0.318 0.364 — — — — — — — — — — — — 0.136 0.227 — 0.273 0.227 0.091 0.045 — — — — — — — 0.045 — 0.591 0.364 — — — — — — — — 0.125 — — 0.375 0.063 0.313 —

0.115 — — — — 0.115 0.308 0.346 0.077 0.038 — — — 0.154 0.308 0.346 0.192 — — — — — — — — — — 0.038 0.038 0.038 0.077 0.077 0.231 0.192 0.115 0.115 — 0.077 — — — — — — 0.038 0.269 0.385 0.077 0.192 0.038 — — — — — — 0.038 0.038 0.038 0.038 0.038 —

0.056 — — — 0.111 0.167 0.444 0.167 0.056 — — — — — — 0.444 0.500 0.056 — — 0.111 0.056 — 0.056 — — — — 0.111 — 0.056 0.056 0.056 0.333 0.056 0.111 — — — — — — — 0.167 — 0.222 0.222 0.167 0.222 — — — — — 0.056 0.056 — — 0.278 — 0.111 —

— — — — 0.050 0.150 0.400 0.350 0.050 — — — — 0.050 0.450 0.100 0.350 0.050 — — — — — — — — — 0.050 — — — 0.150 0.350 0.100 0.100 0.150 — — — — 0.050 0.050 — — — 0.450 0.300 0.200 0.050 — — — — 0.050 — — 0.050 0.050 0.150 0.050 0.250 0.050

0.025 — — — 0.075 0.225 0.375 0.275 0.025 — — 0.025 — 0.150 0.225 0.300 0.300 — — — — 0.050 — — — — — — 0.175 — 0.075 0.150 0.075 0.150 0.150 0.100 0.025 — 0.025 0.025 — — 0.025 — — 0.225 0.050 0.425 0.150 0.125 — — — 0.025 0.075 — 0.050 0.050 — 0.025 0.125 —

0.100 — — — 0.025 0.225 0.225 0.400 0.025 — 0.050 0.025 0.025 0.075 0.225 0.350 0.175 0.075 — — — — — — — — 0.025 0.025 0.025 0.075 0.075 0.300 0.125 0.125 0.100 0.025 0.100 — — — — — — 0.025 0.125 0.300 0.275 0.150 0.050 0.050 — 0.025 — — — — 0.053 0.053 0.026 0.053 0.026 0.079

— — — — — 0.409 0.182 0.273 0.091 0.045 — — — — 0.364 0.227 0.409 — — — — — — — 0.045 — — — 0.045 — 0.045 0.273 0.182 0.273 0.091 0.045 — — — — — — — — — 0.227 0.045 0.318 0.273 0.091 0.045 — — — 0.136 — — — 0.136 — 0.227 —

0.119 — — — 0.071 0.071 0.500 0.214 0.024 — — 0.024 — 0.095 0.571 0.190 0.071 — 0.048 — — — — — — — — — 0.048 0.024 0.071 0.333 0.214 0.048 0.048 0.095 0.024 0.095 — — — — 0.024 — 0.024 0.238 0.095 0.262 0.238 0.119 — — — — — 0.100 0.075 0.075 0.025 0.125 0.100 0.025

0.071 — — — 0.214 0.214 0.214 0.286 — — — — — 0.357 0.286 0.143 0.214 — — — — — — — — — — — — — — 0.143 0.143 0.571 0.071 — 0.071 — — — — — — 0.286 — 0.357 0.214 0.071 0.071 — — — — — — — — — — 0.071 — —

— — — — 0.150 0.150 0.150 0.450 0.100 — — — — — 0.300 0.350 0.300 0.050 — — — — — — — — — — 0.100 0.100 — 0.150 0.150 0.150 0.250 0.100 — — — — — — — 0.050 0.050 0.350 0.150 0.250 0.150 — — — — — — — — — — 0.150 0.050 —

0.063 0.003 0.003 0.016 0.098 0.231 0.291 0.242 0.049 0.005 0.005 0.082 0.008 0.082 0.315 0.277 0.209 0.016 0.005 0.003 0.005 0.022 0.005 0.022 0.005 0.005 0.003 0.011 0.057 0.038 0.076 0.152 0.149 0.177 0.114 0.065 0.057 0.022 0.003 0.003 0.003 0.003 0.005 0.068 0.052 0.310 0.196 0.198 0.125 0.041 0.003 0.003 0.003 0.020 0.042 0.023 0.068 0.068 0.121 0.065 0.099 0.045

© 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

C O N T R A S T I N G P A T T E R N S O F G E N E T I C S T R U C T U R E 3437 Appendix Continued Locus Allele Portugal Spain M Spain B Italy N Italy S Greece France_W France_N France_SE Austria England Denmark Norway Finland Total 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 305 310 315 320 Bb111 200 202 204 206 Bb131 151 155 157 163 165 167

0.050 — — — — — — — — 0.050 — — 0.025 0.025 — — — — — — 0.975 — — 0.025 0.425 0.550 0.025 — — —

— — — — — — — — — — — — — — — — — — — — 0.808 0.154 — 0.038 0.462 0.423 0.115 — — —

— — — — — — — — — — — 0.083 — — — — — — — — 0.917 — 0.083 — 0.750 0.083 0.167 — — —

— — 0.038 0.077 — — — 0.038 0.038 — 0.038 — 0.038 0.115 0.154 0.038 — — — — 0.846 — 0.154 — 0.962 0.038 — — — —

0.063 — — 0.063 — — — — — — — — — — — — — — — — 1.000 — — — 0.818 0.045 — 0.045 0.045 0.045

— — — — 0.038 0.038 — — — 0.115 0.192 0.038 0.077 0.115 — 0.115 0.077 — — 0.038 0.846 — 0.154 — 0.962 0.038 — — — —

*indicates population allele frequencies of 0.000.

© 2007 The Author Journal compilation © 2007 Blackwell Publishing Ltd

0.056 0.056 0.056 — — — — — — — — 0.056 0.167 — — 0.056 — 0.056 — — 0.778 — 0.167 0.056 0.833 0.111 0.056 — — —

0.100 — — — — — — — 0.050 — 0.100 — 0.100 — — — 0.050 — — — 0.800 0.150 0.050 — 0.900 0.100 — — — —

0.175 0.025 0.025 — — 0.050 — 0.025 0.025 0.075 0.025 0.025 0.075 0.025 0.025 0.100 — — 0.025 — 0.625 — 0.175 0.200 0.925 0.075 — — — —

0.079 0.053 0.026 0.026 — — — 0.053 — 0.026 0.053 0.026 0.132 0.053 0.053 0.079 0.053 0.026 0.026 — 0.750 0.025 0.125 0.100 0.975 — 0.025 — — —

— 0.091 — — — — 0.091 — — 0.091 — 0.182 — 0.045 — — — — — — 0.955 0.045 — — 0.955 — 0.045 — — —

0.075 0.075 — 0.075 0.025 — — 0.050 — — 0.150 — 0.075 0.025 — — — — — — 0.429 0.119 0.381 0.071 0.952 — 0.048 — — —

0.071 0.143 0.071 — — — — — 0.286 — — — 0.071 0.071 0.071 — — 0.143 — — 0.929 — — 0.071 0.857 0.143 — — — —

— 0.150 — — — — — — — 0.200 0.100 0.100 0.150 0.050 — — — 0.050 — — 0.650 — — 0.350 0.800 — 0.200 — — —

0.056 0.040 0.014 0.020 0.006 0.008 0.006 0.017 0.020 0.042 0.054 0.031 0.068 0.040 0.023 0.034 0.014 0.014 0.006 0.003 0.780 0.038 0.111 0.071 0.826 0.125 0.041 0.003 0.003 0.003