Drovetskiy, Ann Kessen, Andy Jones, Don Alstad, and Claudia Neuhouser. ... iii. Abstract. I sequenced parts of mitochondrial gene regions for six Eurasian bird ...
UNIVERSITY OF MINNESOTA
This is to certify that I have examined this copy of a doctoral thesis by
Alexandra Pavlova
and found that it is complete and satisfactory in all respects, and that any and all revisions required by the final examining committee have been made.
_______________________________ Robert M. Zink
_______________________________ Date
GRADUATE SCHOOL
Comparative Phylogeography of Eurasian Birds
A THESIS SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY
Alexandra Pavlova
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHYLOSOPHY
Robert M. Zink, Advisor May 2004
Copyright Alexandra Pavlova 2004
Acknowledgments I thank the members of my committee, Robert Zink, Susan Weller, Scott Lanyon, and Andrew Simons for good advice and help with the preparation of this manuscript. I am grateful to my advisor, Robert Zink, whose efforts made it possible for me to become a graduate student, and whose patience, friendship, and persistence helped me to accomplish this work. I offer special thanks to my collaborator, Sievert Rohwer, for finding financial support and providing advice on analyses and comments on manuscripts. I would also like to thank my collaborators, Sergey Drovetski, Evgeniy Koblik, Yaroslav Red’kin, Igor Fadeev, and Evgeniy Nesterov, for collecting birds over many years and for the useful discussions we have had. I am indebted to the many faculty and students who provided support, advice, or comments on manuscripts. These include (but are not limited to) Bob Zink, Sievert Rohwer, Susan Weller, Scott Lanyon, Sergey Drovetskiy, Ann Kessen, Andy Jones, Don Alstad, and Claudia Neuhouser. Special thanks to Rachelle Blackwell-Rago for patiently teaching me lab techniques, and for giving me assistance and advice in the lab. I thank Rachelle, Shannon Farrell, and Mike Westberg for the enormous work they did in the lab. I am grateful to the many people from the Burke Museum, the Bell Museum, the Moscow State University Zoological Museum, and the Moscow State Darwin Museum who participated in collecting trips organized by Sievert Rohwer and Bob Zink. Without all these people this work would never have been accomplished. During the course of this work, tissue samples were obtained from the collections of the Burke Museum of Natural History and Culture of Washington State University, the Zoological Museum of Moscow State University, and the Bell Museum of the University of Minnesota. I am grateful to all of the curators and museum collection personnel that made the specimens available. I offer special thanks to Garret Eddy for financing part of my education and for providing logistic support during multiple expeditions. Lastly, I am grateful to my husband Anton and my children Masha and Nastia for i
constant support, patience, and understanding. I also thank my mother Lyudmila for her help and her illustration of the White Wagtail subspecies. This work was funded through a grant from NSF, by personal investment of Garret Eddy, and through a series of grants from the Dayton and Wilkie Natural History Funds.
ii
Abstract I sequenced parts of mitochondrial gene regions for six Eurasian bird species, which were sampled over vast area, and used variety of methods to explore current population genetics and test several evolutionary models. Yellow and Citrine Wagtails both appear to be paraphyletic with respect to their mtDNA. Yellow and Citrine Wagtails comprise, respectively, three and two lineages which should be considered species. Past population expansion was inferred for the populations of the Yellow Wagtail, whereas the Yamal population of the Citrine Wagtail was found to be evolutionarily stable. For the White Wagtail three mtDNA clades were found, which likely evolved allopatrically in different Pleistocene refugia. These clades were not distributed concordantly with morphological subspecies, suggesting that plumage evolution might be rapid and under effect of strong sexual or natural selection. For the Common Rosefinch incipient speciation in the Caucasus was inferred from the structure of the haplotype tree. The northern populations grouped into northeastern and northwestern groups, with limited gene flow between groups; asymmetric eastward gene flow was detected within each group. Most populations displayed signatures of past population growth. Both the Great Tit and the Willow Tit displayed unstructured haplotype trees, but genetic diversity was higher in the Willow Tit, suggesting higher effective population size. A survey of 34 phylogeographic studies of Eurasian birds revealed four regions of phylogeographic endemism in the southern Eurasia. Three of them, the Caucasus, southern and southeastern Asia, are consistent with areas of suitable habitats during last Pleistocene maximum and probably have played a major role as refugial sites during times of climate cooling. An extreme Pleistocene desert covered central Asia, the fourth region of endemism. It was inferred that endemic clades of central Asia have recently evolved after southward colonization from northern populations. Phylogeographic splits in northern Eurasia were either incongruent or absent in some species, reflecting different post-glacial histories, large geographic distances and the stochastic nature of haplotype trees. Overall it was concluded that avian species differently responded to Pleistocene iii
climatic oscillations. Thus, the Eurasian avifauna consists of species with idiosyncratic histories.
iv
Table of Contents
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .i Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .iii Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vi List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vii Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1: Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Chapter 2: Mitochondrial DNA and plumage evolution in White Wagtails . . . . . . . . . .36 Chapter 3: Phylogeography and population genetics of Common Rosefinch (Carpodacus erythrinus) based on mitochondrial genes . . . . . . . . . . . . . 76 Chapter 4: Mitochondrial phylogeographies of the Great Tit (Parus major) and the Willow Tit (P. montanus) and the role of sampling in inferring population histories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Chapter 5: Comparative phylogeography of Eurasian birds based on mitochondrial DNA sequences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
v
List of Tables
1.1
Traditional taxonomy and scoring of 6 morphological characters for 13 taxa of Motacilla males.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
1.2
Genetic characteristics of Motacilla flava and M. citreola samples. . . . . . . . . . .25
1.3
Number of base pairs sequenced for various gene regions for M. flava and M. citreola.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
1.4
Population pairwise Fst-values for western M. flava samples. . . . . . . . . . . . . . . .28
1.5
Pairwise Fst-values for eastern M. flava samples. . . . . . . . . . . . . . . . . . . . . . . . . 29
1.6
Pairwise Fst-values for M. citreola samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1
Genetic characteristics of 26 geographic samples of M. alba . . . . . . . . . . . . . . . 59
2.2
Population pairwise Fst’s for comparisons between populations of M. alba. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.1
Genetic characteristics of 17 geographic populations of Carpodacus erythrinus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93
3.2
Population pairwise Fst’s for comparisons between populations of C. erythrinus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 96
3.3
Migration rate and population size estimates from Markov Chain Monte Carlo simulations for C. erythrinus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.1
Genetic characteristics of Parus major populations. . . . . . . . . . . . . . . . . . . . . . 125
4.2
Genetic characteristics of Parus montanus populations. . . . . . . . . . . . . . . . . . . 127
4.3
Significant pairwise Fst -values for populations of P. major and P. montanus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.4
Tests for population expansion and selective neutrality for the populations of P. major and P. montanus. . . . . . . . . . . . . . . . . . . . . . . . 130
5.1
Summary of findings from major phylogeographic studies of widely distributed Eurasian birds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
vi
List of Figures
1.1
General location of collecting sites for Motacilla flava and M. citreola. . . . . . . .31
1.2
Maximum likelihood tree showing relationships among major groups of Motacilla flava, M. citreola, and near relatives . . . . . . . . . . . . .32
1.3
Plot of nucleotide diversity and latitude in three groups of M. flava and two groups of M. citreola. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
1.4
Observed and expected mismatch distributions for M. flava and M. citreola. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
1.5
Haplotype network and results of Nested Clade Analysis for northeastern M. flava. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.1
General location of collecting sites for M. alba and phenotypic variation of collected birds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
2.2
Intermediate phenotypes of M. alba. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.3
Maximum parsimony tree for unique male phenotypes of M. alba (A), Maximum Likelihood tree for unique mtDNA haplotypes (B) and Minimum evolution population tree constructed from population pairwise sequence differences (C) . . . . . . . . . . . . . . . . . . . . . . 68
2.4
Plot of nucleotide diversity versus latitude for all localities of M. alba . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.5
Mismatch distributions for northern localities of M. alba . . . . . . . . . . . . . . . . . . 71
2.6
Mismatch distributions for all Krasnodar individuals, only southern Krasnodar haplotypes and all individuals from Primor'e for M. alba . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.7
Haplotype and phenotype distributions of M. alba at several localities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.1
General location of collecting sites for Carpodacus erythrinus . . . . . . . . . . . . . 99
3.2
Maximum Likelihood tree for unique mtDNA haplotypes of C. erythrinus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 vii
3.3
Minimum evolution population tree constructed from population pairwise sequence differences for C. erythrinus . . . . . . . . . . .102
3.4
Mismatch distributions and coalescent likelihood estimates of population size and gene flow for nine populations of C. erythrinus . . . . . . 104
3.5
Plot of nucleotide diversity against latitude for C. erythrinus . . . . . . . . . . . . . . 105
4.1
Collecting localities of P. major and the P. montanus samples . . . . . . . . . . . . . 131
4.2
ML phylogenetic tree of haplotypes for P. major . . . . . . . . . . . . . . . . . . . . . . . 133
4.3
ML phylogenetic tree of haplotypes for P. montanus . . . . . . . . . . . . . . . . . . . . .135
4.4
Nucleotide diversity for P. major and P. montanus populations plotted against North latitude and East longitude . . . . . . . . . . . . . . . . . . . . . . . .137
5.1
UPGMA dendrograms and geographical distribution of clades for (A) Sitta europaea, (B) Picoides tridactylus and (C) Dendrocopos major . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177
5.2
UPGMA dendrograms and geographical distribution of clades for (A) Troglodytes troglodytes and (B) Phylloscopus trochiloides . . . . . . . . . . . .179
5.3
UPGMA dendrogram and geographical distribution of clades for Motacilla flava, M. citreola and M. alba . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
5.4
UPGMA dendrograms and geographical distribution of clades for Parus major (A and B) and P. montanus (C and D) species complexes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
5.5
UPGMA dendrogram and geographical distribution of clades for (A) Motacilla alba and (B) Emberiza schoeniclus. . . . . . . . . . . . . . . . . . . . 185
5.6
UPGMA dendrogram and geographical distribution of clades for (A) Carpodacus erythrinus and (B) Luscinia svecica. . . . . . . . . . . . . . . . . . 187
5.7
General areas of phylogeographic endemism and approximate limits of clade distribution for Eurasian birds overlaid on the reconstruction of vegetation during last glacial maximum. . . . .189
5.8
The frequencies of the most common haplotypes for Eurasian clades arranged from smaller to larger values. . . . . . . . . . . . . . . . .191
5.9
Number of clades detected per species as a function
viii
of the number of sampling localities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 5.10
The frequency of the most common haplotype as a function of sequence length (A) and sample size (B) . . . . . . . . . . . . . . . . .193
ix
Introduction
Phylogeography uses a combination of approaches (coalescent, likelihood, nested clade analysis) along with traditional methods of population genetics to make inferences about the evolutionary history of species from the geographic distribution of genetic variation. Comparative phylogeography links the knowledge about phylogeographic structure of codistributed species into general patterns of regional diversity. In this way, the evolutionary effect of historical events on speciation patterns can be observed. Mitochondrial DNA (mtDNA) is a maternally inherited and fast evolving molecule, and mtDNA lineages are rapidly sorted geographically, yielding intraspecific resolution. For this reason mtDNA is a common genetic marker used in phylogeography. Although patterns of mtDNA variation in some European birds are known, general knowledge about phylogeographic patterns in the Eurasian avifauna is limited. In this work I explore mitochondrial DNA variation of six widespread Eurasian birds and make inferences about their evolutionary history (Chapters 1-4). I investigate how climatic oscillations of Pleistocene shaped phylogeographic structures. Review and synthesis of published studies on 28 avian species permitted an understanding of the history of the Eurasian avifauna (Chapter 5). In Chapter 1, I explore phylogenetic relationships among three monophyletic groups of sequences (haplotypes) of the Yellow Wagtail (Motacilla flava) and two groups of Citrine Wagtail (M. citreola). I use coalescence analysis to test several evolutionary hypothesis and traditional population genetic methods to study current processes in populations of each group. In Chapter 2, I explore patterns of molecular evolution in White Wagtail (Motacilla alba) in conjunction with plumage differentiation. I propose an evolutionary scenario which could lead to the observed patterns of genetic and morphological variation. In Chapter 3, I investigate the evolutionary relationships among groups of geographic localities of Common Rosefinch (Carpodacus erythrinus). I use a coalescent 10
likelihood method to explore contemporary migration patterns between rosefinch populations and other coalescent methods to explore historical population dynamics. In Chapter 4, I analyze phylogeographic structure of Great Tit (Parus major) and Willow Tit (P. montanus). I use my own data and data from other studies to make inferences about species’ history based on samples from complete ranges of both species. In Chapter 5, I compare phylogeographic patterns of mtDNA variation of 28 Eurasian species of birds and define several regions of phylogeographic endemism in southern Eurasia. I investigate how contemporary phylogeographic structure of different species might be affected by the distribution of suitable habitats during the last Pleistocene glacial maximum.
11
Chapter 1
Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history 1.
ABSTRACT.-We conducted phylogeographic analyses of Motacilla flava (Yellow Wagtail) and M. citreola (Citrine Wagtail). We analyzed mitochondrial DNA sequences from 167 M. flava specimens obtained from 17 localities throughout Eurasia and Alaska, and 38 specimens of M. citreola obtained from 7 Eurasian localities. Phylogenetic analysis revealed three clades within traditionally recognized M. flava: Europe and southwestern Asia, northeastern Asia and southeastern Asia. These groups should be considered species, because together they are not monophyletic, and are interspersed with M. citreola, M. cinerea and M. alba. Motacilla citreola also is paraphyletic, consisting of two species-level groups. Northeastern and southeastern groups of M. flava each appear to be sister-taxa to eastern and western groups of M. citreola, respectively. Together these four groups form a clade, whereas the western M. flava group is considerably more distant. Within each of the three groups of M. flava, and the two groups of M. citreola, little phylogeographic structure was detected. Signatures of past population expansion are evident for some populations of M. flava; expansion is more recent in Moscow, Kursk (western group), Yamal and Anabar (northeastern group), and older in Tyva and Vyatka (western group). A history of population stability is inferred for the Yamal population of M. citreola. Nested clade analyses detected contiguous range expansion for southeastern M. flava and restricted gene flow with isolation by distance for northeastern M. flava and eastern M. citreola. 1
The material in this chapter was written in collaboration with R. M. Zink, S. Drovetski, Y. Red'kin, and S. Rohwer and published in the ornithological journal The Auk. (Pavlova A., R. M. Zink, S. V. Drovetski, Y. Red’kin, and S. Rohwer. 2003. Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history. The Auk 120(3):744-758). Copyright 2003 by the American Ornithologists’ Union.
12
Many aspects of population history can be inferred from DNA sequences (Avise 2000; Nee et al. 1996). Phylogenetic analysis of DNA sequences yields a tree, which when superimposed over the geographic distribution of populations, reveals whether the history of populations has been one of isolation, panmixia or some combination thereof. For example, haplotypes from different localities or regions might be reciprocally monophyletic, suggesting a history of isolation. Such a finding often challenges current species limits. Coalescence analyses contribute further information by revealing population increases and the magnitude and direction of gene flow (Hewitt 1996, Templeton 1998). Collectively these analyses constitute phylogeography (Avise 2000). Phylogeographic studies are valuable in showing how populations responded to the Late Pleistocene cycles of glaciation, especially in the northern hemisphere. For example, Merila et al. (1997) compared mtDNA sequences from populations of greenfinch (Carduelis chloris) distributed along a north-south transect in Europe. An unstructured haplotype phylogeny revealed no pattern of historical isolation. However, northward decrease in nucleotide diversity (π) was consistent with post-glacial recolonization, a phenomenon termed leading edge expansion (Hewitt 2000). In another European species, the noctule bat (Nyctalus noctula), no genetic evidence was found that could be interpreted as directional colonization and glacial-induced bottlenecks (Petit et al. 1999). In North America, the few studies available suggest that species have had idiosyncratic Pleistocene histories (Zink 1997; Fry and Zink 1998; Barrowclough et al. 1999). Unfortunately, little is known about broad-scale phylogeographic patterns of vertebrates distributed across Eurasia (Fedorov et al. 1999, Kryukov and Suzuki 2000, Salzburger et al. 2002), the largest expanse of land in the northern hemisphere. Eurasia was not extensively glaciated during the last glacial advance (Würm), unlike North America. However, permafrost covered large northern areas and many habitats were fragmented and displaced southward (Andersen and Borns 1997). We set out to characterize mtDNA phylogeography in a diverse set of Eurasian birds to document patterns of genetic differentiation and to discover genetic consequences of Late Pleistocene climate changes (see Koblik et al. 2000, Rohwer et al. 2001, Zink et al. 13
2002a, b). The Yellow Wagtail (Motacilla flava) is widely distributed throughout the Old World, and has colonized Alaska in the New World (Cramp 1988). Variation within this species has been recognized by the description of two subspecies groups, lutea and flava, each containing numerous subspecies and sometimes treated as separate species (Badyaev et al. 1998, Koblik et al. 2001; Table 1). There are several hybrid zones (Czikeli 1985), and often mixtures of phenotypes associated with different subspecies exist over relatively broad geographic areas. We sequenced parts of three mitochondrial gene regions in 167 individuals of M. flava representing 10 subspecies taken from 16 localities throughout Eurasia and one site in Alaska (Fig. 1.1, Table 2). Phylogenetic analysis of haplotypes revealed three groups of M. flava that were not monophyletic, and which were interspersed with two groups of M. citreola and some outgroups. Therefore we also report here a phylogeographic survey of M. citreola (38 individuals, 7 localities). In addition to exploring discrepancies between the mtDNA gene tree and taxonomic limits, we determined whether groups of populations showed genetic signatures of population and range expansions, similar to those documented for other European species (Hewitt 2000).
METHODS
Motacilla flava and M. citreola specimens were collected during the breeding season. Outgroup taxa (M. clara, M. capensis, M. aguimp, M. cinerea, M. alba, and M. lugens) were represented by up to two individuals. From all specimens a study skin was preserved, and deposited at the Burke Museum, University of Washington, Seattle; the Moscow State University Zoological Museum, Moscow, Russia; or the Bell Museum, University of Minnesota, St. Paul. Tissue samples were stored in lysis buffer or liquid nitrogen. Genomic DNA isolation and purification followed either a modified Chelex (Zink et al. 1998) or phenolchloroform protocol (Hillis et al. 1996). Mitochondrial gene regions (Table 3) were 14
amplified via Polymerase Chain Reaction (PCR) (Saiki et al. 1988). PCR products were cleaned using a Qiaquick PCR Purification Kit (Qiagen). Sequencing reactions were cleaned using standard Sephadex columns and sequenced on an ABI 310 automated sequencer using Dideoxy Terminator kit Protocol. One-stranded automated sequencing was performed for M. flava; 15 sequences were compared in both directions with minimal correction needed (0.4% of uncertain bases were resolved). For M. citreola and outgroups we sequenced both strands of ND3 and Cyt b. Sequences were aligned and edited using Sequencher 3.1.1 (Gene Code Corporation, Inc.). Sequence data have been deposited in Genbank (AF443466-AF443561, AF445463-445558, AF446393-446860). Mitochondrial origin of DNA was supported by sequencing disparate regions, absence of stop-codons and the existence of a large number of haplotypes, all of which are inconsistent with nuclear copies (Zhang and Hewitt 1996). We used Arlequin software (Schneider et al. 2000) to compute π, Fst, Fu’s (1997) test of selective neutrality, mismatch distributions (for localities with greater than 10 individuals), the number of haplotypes (allowing in Arlequin up to 20% missing data per site, and with gaps included), and to perform a Mantel’s (1967) test of pairwise Fst-values vs. geographic distances. Nucleotide diversity provides an index of genetic variability, and can reveal patterns of population expansion. We regressed nucleotide diversity against latitude, expecting smaller values to be in the north because of leading edge expansion (Hewitt 2000). Fst measures the extent of population subdivision taking into account the degree of haplotype differentiation. Fu’s (1997) F-value tests for departure from neutral expectation assuming an infinite-site model without recombination; it is sensitive to population expansions, which generate negative values owing to excess rare haplotypes. The mismatch distribution is the distribution of pairwise base pair differences among individual haplotypes, and its shape provides information about recent changes in population size (Rogers and Harpending 1992). We used Schneider and Excoffier’s (1999) finite-site method, which involves a parametric bootstrap estimate of confidence for the fit of the mismatch distribution to the distribution expected from sudden 15
population expansion. The average number of individuals exchanged per generation among populations (Nm) was computed from pairwise population Fst-values; these values are used as indications of the relative magnitudes of gene flow, not the actual number of individuals moving between populations. Mantel’s tests were used as an approximation of the magnitude of genetic diversity resulting from isolation by distance. These tests compared a matrix of pairwise Fst-values with straight-line geographic distances between samples; the null hypothesis is that the matrices are independent, so that rejection would indicate an isolation-by-distance effect. Using F statistics or an algorithm that assumes that all geographical associations are due to gene flow can yield an estimator of Nm that is biologically misleading (Templeton 1998). We used nested clade analysis (NCA; Templeton et al. 1995) because of its potential to distinguish between gene flow, past fragmentation, and range expansion. NCA uses a haplotype network to test the null hypothesis of no geographical association of haplotypes. TCS v. 1.13 (Clement et al. 2000) was used to construct haplotype networks. Minor ambiguities (closed loops) were resolved using parsimony (by favoring associations of haplotypes from the same locality). Clades were nested according to rules described in Templeton et al. (1987). GeoDis v.2.0 (Posada et al. 2000) with 10000 resampling events was used to calculate the clade distance (Dc, the average geographic distance of haplotypes from the clade to the geographical center of that clade), nested clade distance (Dn, average distance of haplotypes from the clade to the geographic center of all haplotypes at next nesting level), and the difference between interior and tip clades (I-T). We used the inference key of Templeton (1998) to estimate causes of geographical associations of haplotypes. For phylogenetic inference within major geographic groupings, we generated a neighbor-joining tree in PAUP* (Swofford 2000) and estimated the percentage of invariant sites, transition:transversion ratio, base frequencies, and the shape parameter (gamma) on that topology. These values were then input into Paup* and a heuristic ML search was performed. We re-estimated the input parameters on the resultant topology, and used them in a second heuristic search to arrive at an ML topology. We used Paup* to 16
generate maximum parsimony trees (MP) from equally weighted characters, and we constructed strict and 50% majority rule consensus trees. Tree topologies were independent of whether gaps were included. To test species limits, we analyzed a subset of 22 haplotypes including divergent individuals within each taxon, which had been sequenced for additional base pairs (Table 3). We performed branch and bound parsimony searches on these haplotypes with characters equally weighted, excluding missing and ambiguous sites (n = 208; results did not differ if these sites were included). We conducted an ML search following the protocol outlined above, and performed 100 bootstrap replicates. Alternative tree topologies were evaluated with the Shimodaira-Hasegawa (1999) test. We summarized the taxonomic history and states of characters used by taxonomists to classify subspecies (Table 1). These characters, equally weighted, were subjected to maximum parsimony analysis (exhaustive search).
RESULTS TAXONOMIC FINDINGS Both MP and ML analyses (not shown) for all individuals of all taxa resolved three groups of haplotypes currently classified as M. flava, corresponding to Europe and southwestern Asia, northeastern Asia, and southeastern Asia (Fig. 1.1). These three groups did not form a clade. We also found two groups (eastern, western) of M. citreola, which were not sister groups. To explore this result further, we conducted phylogenetic analysis of the restricted set of 22 haplotypes (for 1832 bp), which resolved 46 equally parsimonious trees, the consensus of which (not shown) matched a ML tree (Fig. 1.2) in showing lack of monophyly for the three groups of Motacilla flava and two groups of M. citreola. A MP analysis including gaps revealed 8 trees, which were a subset of the 46 found without gaps. The western clade of M. flava was separated from the other M. flava and M. citreola clades by M. cinerea and M. alba/lugens. Furthermore, none of the trees recovered monophyly of the northeastern and southeastern clades of M. flava because 17
different M. citreola clades were each sister to them (western citreola with southeastern flava, and eastern citreola with northeastern flava). We tested alternative topologies, starting with one of the 46 equally parsimonious trees, which was similar to that shown in Fig. 1.2. Making the two M. citreola groups monophyletic did not result in a significantly worse log likelihood score (S-H test; P = 0.08), nor did forcing the monophyly of the northeastern and southeastern M. flava groups (S-H test, P = 0.08). Rearranging the topology to make western M. flava sister to eastern M. flava plus M. citreola was significantly worse (P = 0.01). A tree with a monophyletic M. flava was significantly worse (P < 0.001). We subsequently treated each of the three groups of M. flava and two groups of M. citreola independently. Lastly, removing M. alba/lugens, primarily white-and-black-plumaged taxa, outside of western M. flava, thereby creating a “yellow” clade (all M. flava, M. citreola and M. cinerea), was not significantly worse (P = 0.61). A ML test did not reject (X2 = 29.8, df = 20, P > 0.05) the assumption of a molecular clock.
PHYLOGEOGRAPHY OF GEOGRAPHIC CLADES
Haplotype trees. —Among the 90 individuals of M. flava from western Eurasia, there were 69 haplotypes. The most common haplotype was found in 6 individuals representing 6 localities. The data set contained 41 parsimony uninformative and 31 informative base positions (including gaps but excluding outgroup sequence). We evaluated over 20,000 equally parsimonious trees (length 117, consistency index (CI) 0.67, rescaled consistence index (RC) 0.45), the consensus of which (not shown) was unstructured; however, 15 of the 17 haplotypes from Vyatka formed a clade (93% of trees) in a 50% majority rule consensus tree (and in the ML tree). One haplotype found in Vyatka was also found in Kursk, Moscow and Vologda. Therefore, this locality apparently exhibits the early stages of evolutionary isolation. Because of the latter result, analyses below were done with and without the Vyatka sample. Subspecies designations 18
did not predict clusters of haplotypes. There were 31 haplotypes among the 57 individuals of M. flava from northeastern Eurasia. The most common haplotype was found in 6 localities and in 16 individuals. The data set contained 17 parsimony uninformative and 9 informative base positions (including gaps but excluding outgroup sequence; 993 bp total). We evaluated over 12,500 equally parsimonious trees (length = 32, CI = 0.88, RC = 0.72), the consensus of which (not shown) was unstructured; however, ca. 50% of haplotypes from Anabar clustered together. Subspecies designations did not predict clusters of haplotypes. There were 17 haplotypes among the 20 individuals of M. flava from the three localities in southeastern Eurasia. The data set contained 13 parsimony uninformative and 13 informative base positions (including gaps but excluding outgroup sequence; 993 bp total). We evaluated 54 equally parsimonious trees (length 35, CI = 0.74, RC = 0.50), the consensus of which was unstructured (not shown), although several basal haplotypes were from Mongolia. Subspecies designations did not predict clusters of haplotypes. Twenty-three haplotypes were found among the 38 individuals of M. citreola. The data set of 1407 base pairs (including gaps) contained 22 parsimony uninformative and 28 informative characters. We obtained over 20,500 equally parsimonious trees (length 62, CI = 0.81, RC = 0.76), the consensus of which (not shown) showed two distinct regional groupings, western (Moscow, Kursk, Tyva; subspecies M. c. werae) and eastern (Mongolia, Anabar, Buryatia, Yamal, Tyva; subspecies M. c. citreola and M. c. quassatrix) (Fig. 1.1). There were nine haplotypes among 13 western and 14 haplotypes among 25 eastern individuals. The most common haplotype was found in eight individuals representing four eastern localities. There was no phylogeographic structure in haplotype trees (not shown) for the eastern or western groups of M. citreola, but two individuals (31 Tyva and 11 Mong) were basal to a polytomy of eastern haplotypes. Genetic diversity.-Fst for the western region of M. flava was 0.13 (P < 0.05) with the sample from Vyatka included, and 0.08 (P < 0.05) without it. Pairwise Fst-values (Table 4) ranged from 0.0 to 0.19, and most comparisons involving Astrakahan’ and Vyatka were significant. Fst for the northeastern group of M. flava was 0.11 (P < 0.05). 19
Pairwise Fst-values (Table 5) ranged from 0.02 to 0.27, and all comparisons involving Anabar were significant. Also, all values between northeastern and southeastern samples were significant. Fst for the southeastern region of M. flava was 0.11 (P < 0.05). Pairwise Fst-values (Table 5) ranged from 0.062 to 0.13. For the eastern M. citreola clade Fst was 0.14 (P < 0.05) and pairwise Fst-values ranged from 0.04 to 0.17. For the western clade Fst was 0.19 (P < 0.05) and pairwise Fstvalues ranged from 0.005 to 0.30 (Table 6). Genetic variability.-Nucleotide diversity (Table 2) for western samples of M. flava ranged from 0.0025 to 0.0047. Several Fu’s F-values (Table 2) were significantly negative, suggesting either population expansion or deviation from selective neutrality. Nucleotide diversity among northeastern samples of M. flava ranged from 0.0005 to 0.0024, with lower values being found in the northeastern most localities (except Alaska). Fu’s F-values were generally negative and significant. Nucleotide diversity for the three southeastern samples of M. flava ranged from 0.0024 to 0.0069 (Table 2); the southernmost sample (from Mongolia) had the highest value. Fu’s F-values were not significant. For M. citreola, nucleotide diversity ranged from 0.0014 to 0.0022 for the eastern clade and 0.0015 to 0.0052 for the western clade. Fu’s F-values were not significant except for the sample from Yamal. No significant associations between nucleotide diversity and latitude were found for each of five studied groups (Fig. 1.3). An overall test was not performed because our samples were taken from five independently evolving groups. Population expansion and gene flow. —For our largest samples of M. flava (Moscow, Vyatka, Tyva, Kursk, Yamal, Anabar), mismatch distributions (Fig. 1.4, a-e, Kursk not shown) were consistent with past population expansion. Nm values for western M. flava (not shown) ranged from 2.2 to very high values indicating gene flow between localities. For northeastern samples of M. flava, Nm values (not shown) ranged from 1.4 to 29 indicating generally high gene flow. No mismatch distributions were computed for southeastern samples; Nm values were large. 20
The mismatch distribution for Yamal population of M. citreola (Fig. 1.4f) differed from the expectation for sudden population expansion (P < 0.05) and was bimodal, which indicates a history of population stability. Nm values within both M. citreola clades were larger than 3.0. Mantel’s tests were not significant for western and northeastern groups of M. flava (tests were not performed on M. citreola and southeastern M. flava because too few sampling localities were available). Nested Clade Analyses.-TCS detected more than 10000 loops in the network of western M. flava haplotypes. It was not feasible to resolve these loops because of too many homoplasious characters (Posada & Crandall 2001). For northeastern M. flava a single haplotype network was constructed. The final nested cladogram comprised 33 0-step clades (haplotypes), 13 1-step clades, four 2-step clades and the entire cladogram (Fig. 1.5). Haplotype 142Anab could have been grouped with one of three clades, 1-2, 1-12 or 1-11. We tested all alternatives and found that although nesting does affect the significance of clades it did not affect the inferences. Clade 1-2 displayed one significant value (Dn is large for 126,129 Alas) but this outcome was inconclusive. Clade 2-0 had significantly structured subclade 1-8, which consisted of three haplotypes from Yamal (Dc and Dn are significantly small), leading to an inference of restricted gene flow with isolation by distance (inference key 1-2-3-4no). The entire cladogram also displayed significant values (Dn is large for 2-0, Dc is small for 2-3, the clade comprising five individuals from Anabar), also suggesting restricted gene flow with isolation by distance (inference key 1-2-3-4no). According to the assumptions of NCA, clade 1-2 is considered ancestral, because its geographical distribution covers all sampled localities (Templeton 1998). Many of tip clades in our analysis of northeastern M. flava have restricted ranges (Fig. 1.5), which is consistent with restricted gene flow. The cladogram (not shown) for southeastern M. flava was nested to a 4-step level and consisted of 17 haplotypes, 15 one-step clades, seven 2-step clades and three 3-step clades. Statistically significant distance values were detected only at two higher-level 3step clades, both of which had haplotypes from all three sampling sites. For a tip clade it 21
was not possible to distinguish between past fragmentation and isolation by distance using the inference key (1-2-3-4-9-10no) owing to inadequate sampling. From these, only one inference of contiguous range expansion could be made (inference key 1-2-11-12no). For southeastern M. flava, a clade consisting of six individuals from Khabarovsk and Sakhalin followed the pattern predicted for ancestral haplotypes. There was no widespread common haplotype in this group, but haplotype 29Khab had the most mutational connections and was considered the root by TCS, suggesting that Khabarovsk was the place from where southeastern M. flava haplotypes expanded (to Mongolia and Sakhalin). However, based on parsimony (tree not shown) five out of eight haplotypes from Mongolia were basal on a strict consensus tree, suggesting expansion from Mongolia. For M. citreola two cladograms were constructed by TCS. Out of ten haplotypes only three were terminal on the network for western M. citreola. Resolving the relationships between haplotypes would be arbitrary, so we did not perform NCA on this group. There were 14 haplotypes, nine 1-step clades, four 2-step clades in a network (not shown) for eastern M. citreola nested to the three-step level. Out of all clades just one 1step clade showed significant distance values. This internal clade consisted of the most common haplotype (eight individuals from four localities), and six haplotypes one mutational step apart from it. The inference key (1-2-11-17-4no) led to restricted gene flow with isolation by distance. The range of this clade covers all sampled localities, but it is most frequent in Tyva (80% of sampled individuals) and Yamal (72%). According to the strict consensus parsimony tree (not shown), haplotypes from Tyva and Mongolia are basal. Thus, either Mongolia or Tyva are the most likely sites of recent origin for this group. Hence, colonization of eastern part of Eurasia by M. citreola went from south to north.
DISCUSSION TAXONOMIC CONCLUSIONS 22
The Yellow Wagtail is a common, well-studied Palearctic species. Our results were surprising in showing that the current taxonomies of M. flava and M. citreola do not reflect the evolutionary history of the mtDNA gene tree. We provided our initial result to G. Voelker, who subsequently confirmed our results (Fig. 1.2) with all relevant outgroups (Voelker 2002). The geographically coherent distribution of the reciprocally monophyletic groups of haplotypes suggests that the mtDNA gene tree is an accurate portrayal of the organismal history of these taxa, as has been shown for many other avian taxa (Moore 1995; see Degnan [1993] for an exception). That is, although the mtDNA tree can misrepresent the organismal tree, it would be illogical to expect such a biased tree to be as geographically structured as that in Figure 2. Therefore, based on the findings of our comprehensive sampling of each taxon, and the analysis by Voelker (2002), taxonomic changes can be inferred from the mtDNA tree. The traditionally classified M. flava consists of more than one species. The western and eastern components should be reclassified as separate species. Based on reciprocal monophyly of mtDNA haplotypes, it is warranted to recognize the northeastern and southeastern taxa as evolutionarily significant units (Moritz 1994), and probably phylogenetic species. If recognized as species, the name Motacilla flava applies to western forms, the northeastern group becomes Motacilla tschutschensis (Gmelin 1789), and the southeastern group is Motacilla taivana (Swinhoe 1863). More rigorous sampling and analyses of all forms are needed to ascertain complete species limits and ranges of each taxon. The taxonomy of M. citreola also requires revision. Our two clades correspond to the eastern subspecies M. c. citreola and M. c. quassatrix (Portenko 1960) and a western form M. c. werae. Where the two forms meet is unclear (Cramp 1988; Dement’ev 1954). However, our sample from Tyva includes individuals from both clades, indicating that this site is situated in a contact zone. Given that the two taxa probably are not sisters (Fig. 1.2), it is likely that this contact is secondary (Cracraft 1989). Our tree (Fig. 1.2) and the high Fst-value (0.88) between the eastern and western groups of M. citreola suggest that 23
two phylogenetic species are involved; these taxa are certainly evolutionarily significant units (Moritz 1994). The western group becomes Motacilla werae (Buturlin 1907) whereas Motacilla citreola remains for eastern populations. Phylogenetic results (Fig. 1.2; Voelker 2002) illustrate that the species-level taxonomy of at least some Palearctic taxa is unreliable. The previous classification of wagtails (Table 1) clearly conflicts with the mtDNA divisions. For example, the two members of the lutea complex (M. f. lutea, M. f. taivana) are in different mtDNA groups. Inspection of the plumage characters (Table 1) used to construct the traditional taxonomy reveals why it conflicts with the mtDNA gene tree. Previous taxonomic groupings were based largely on single characters. When morphological characters (Table 1) are considered simultaneously (parsimony analysis not shown), character conflicts result in an unresolved strict consensus tree. Thus, in fact, the morphological characters on which taxonomists based their subspecies assignments do not support the current taxonomy when analyzed in concert. It follows that the subspecies schemes (Table 1) resulted from differential weighting of morphological features by individual taxonomists. Of interest is the observation that each of the major mtDNA groupings contains similar mixtures of phenotypes. Assuming that the mtDNA tree reflects evolutionary history, this suggests considerable convergence in the external phenotypic characters used by taxonomists in classifying taxa in this group.
POPULATION HISTORY AND GENETIC-PHENOTYPIC EVOLUTION
At the time (18,000 ybp) when the Laurentide Ice sheet extended over much of North America, the landscape of Eurasia differed considerably (see Hewitt 2000). In Eurasia, glaciers covered most of Europe, but east of ca. 1200 E, glaciation was essentially absent. However, permafrost occurred over the non-glaciated landscape, and extended south to 400N, and much farther in central Asia. Thus, although Eurasia was not as extensively glaciated as North America, habitats were strongly displaced southwards (Frenzel et al. 1992). Depending on the degree to which wagtails bred in areas of 24
permafrost, one might posit that the species have recently recolonized the northern reaches of their current Eurasian ranges, following climatic amelioration. Such a recent history leads one to predict the genetic signatures of range and population expansions. As a result of our discovery of multiple independently evolving groups, the sample sizes for each of five studied taxa limited inferences about history of each taxon. Phylogenetic analyses of haplotypes within each of the major clades of M. flava and M. citreola did not recover reciprocally monophyletic groups. Unstructured haplotype trees are an expected consequence of populations or groups of populations having been isolated for less than 2Nef generations (Nef is the inbreeding effective size of the female population), the time required for the evolution of reciprocal monophyly (Avise 2000). However, it is possible that either populations were historically stable and connected by gene flow, or that the current range was only recently reoccupied. Most of the mismatch distributions, Nm values, and the negative Fu’s F-values, are consistent with past population expansions and gene flow. Coalescence theory yields inferences about the relative age of population expansion from the shape of a mismatch distribution (Slatkin and Hudson 1991). Unimodal distributions are indicative of population expansion, and as the mean number of differences between haplotypes increases, the mode of the distribution shifts to the right and the inferred age of population expansion becomes older. If a mismatch distribution is multimodal, a history of population stability can be inferred. For example, the Moscow region was under glacial ice during the last glacial cycle, a recent range expansion is expected. We found that for M. flava populations from Vyatka and Tyva (Fig. 1.4b, c), expansion predates that for Moscow (Fig. 1.4a) and Kursk (not shown). Two northeastern localities Yamal and Anabar (Fig. 1.4d, e) exhibited mismatch distributions similar to those for Moscow and Kursk, indicating relatively recent past population expansions, as expected. Contrary to expectation, the (far northern) Yamal sample of M. citreola showed a signature of stability. Too few samples of M. citreola were available to allow inferences about the histories of the two major clades; it is possible that gene flow from unsampled differentiated populations contributed to a multi-modal mismatch distribution. 25
Although the plot in Figure 3 suggests a northward decline in nucleotide diversity, as found for many European animals (Hewitt 2000), statistical inference is complicated because the data come from five independently evolving groups. No evidence of significant northward expansion was found within any of five studied groups (Fig. 1.3). However, this result could be due to insufficient sampling of the northern Europe and a large part of Siberia (Fig. 1.1). Thus, although it is possible that nucleotide diversity shows a signature of leading edge expansion, further sampling is required. The significant Fst values within groups of M. flava and M. citreola, the tendency for haplotypes from Anabar and Vyatka to cluster together, and limited isolation-bydistance (NCA for northeastern M. flava and eastern M. citreola) suggest that some of these groups of populations are in the early stages of geographic differentiation. This characterization differs from that of widespread bird species found in North America, which are typically less differentiated (Zink 1997). Thus, some but not all (Zink et al. 2002a,b) Eurasian taxa might be more differentiated owing to a less severe glacial history, greater geographic distances, or greater degrees of previous genetic structure having been retained across the last Ice Age. Additional species require study to ascertain which is the general pattern. Our historical scenario for wagtails raises the question of how phenotypic differentiation originated if there has been limited isolation and/or gene flow. Evolution of plumage differentiation (Table 1) could be explained by prior isolation of populations for a time sufficient for morphological differentiation, but not for reciprocal monophyly in mtDNA haplotype trees. Because morphological traits are likely polygenic, they can evolve at a faster rate than single-locus traits such as sequences in the mtDNA genome. Thus, the disparity between mtDNA and morphology could reflect differing evolutionary rates. In addition, broad zones of overlap occur in which phenotypes that characterize subspecies occur at single locations (Table 2), suggesting that previously isolated populations have come into contact. Thus, an alternative explanation would involve spread of phenotypes among major groups via hybridization. In any event, phenotypically defined subspecies are not on independent evolutionary trajectories, as is true for many 26
other birds (Ball and Avise 1992, Zink et al. 2000).
METHODOLOGICAL CONSIDERATIONS Nested-clade-analysis has been used infrequently in avian phylogeography and it is useful to compare inferences based on it to other more commonly used approaches. Methods such as Fst, Mantel’s test, parsimony trees, and mismatch distributions suggested a history of isolation of major groups followed by recent range expansion. Although some of our sample sizes were small, NCA indicated some additional trends, such as isolation by distance, which was not detected by Mantel’s test. This might result because Mantel’s test considers all samples simultaneously, whereas NCA considers (only) significant clades independently. Nonetheless, many assumptions in NCA are tenuous (Knowles and Maddison 2002), which could lead to discrepancies with traditional methods. For example, TCS often identified the most common haplotype as the root haplotype. However, the probability of the most frequent haplotype being oldest is equal to its frequency, which is low in our study (and would decrease with addition of more sequences). In addition, the oldest haplotype is not necessarily the root. An old haplotype might be basal to only one clade of a two-clade haplotype tree, with the real root haplotype having gone extinct. Thus, it remains useful to consider outgroup rooting if possible. In our analysis of southeastern M. flava, it seemed that maximum parsimony was more likely to identify the root than statistical parsimony as implemented in TCS. Thus, both NCA and traditional methods should be considered jointly.
Note added in proof: Alstrom et al. (2003) suggested that there were either two or many species of M. flava, specifically suggesting that many subspecies studied by us might qualify as species. Our data strongly contradict both suggestions, as with broader taxon sampling, we and Voelker (2002) show that three evolutionarily distinct forms exist. Although Alstrom et al. considered M. f. plexa as part of M. f. thunbergi, our study showed that this enlarged taxon is distributed between the western and northeastern
27
groups. Therefore, the taxonomic scheme suggested by Alstrom et al. is inconsistent with our and Voelker’s (2002) mtDNA studies.
ACKNOWLEDGMENTS We thank R. Blackwell-Rago for technical assistance, G. Barrowclough for advice on coalescence analyses, J. Klicka and K. Winker for tissue samples from Alaska, A. Jones for assistance with the morphological analysis, C. S. Wood, B. Schmidt, G. Voelker, D. Banin, A. Andreev, I. Fadeev, E. Nesterov, I. Karagodin, E. Koblik, and V. Sotnikov for logistical help with expeditions and collecting, and especially S. Birks, who subsampled some tissues. We are grateful to G. Eddy for supporting S. D. and funding fieldwork. Additional support came from the NSF (DEB 9707496).
28
LITERATURE CITED Alstrom, P., K. Mild, and B. Zetterstrom. 2003. Pipits and Wagtails of Europe, Asia and North America. Christopher Helm/A and C Black, London. Avise, J. C. 2000. Phylogeography. Harvard Univ. Press, Cambridge Univ. Press, London. Andersen, B. G., and H. W. Borns. 1994 (1997 2nd edition). Ice Age World. Scandinavian University Press. Badyaev, A. V., B. Kessel, and D. D. Gibson. 1998. Yellow Wagtail (Motacilla flava). In The birds of North America, No. 382 (A. Poole and F. Gill, eds.). The birds of North America, Inc., Philadelphia, PA. Ball, R. M., Jr. and J. C. Avise. 1992. Mitochondrial DNA phylogeographic differentiation among avian populations and the evolutionary significance of subspecies. Auk 109: 626-636. Barrowclough, G. F., R. J. Gutiérrez, and J. G. Groth. 1999. Phylogeography of spotted owl (Strix occidentalis) populations based on mitochondrial DNA sequences: gene flow, genetic structure, and a novel biogeographic pattern. Evolution 53:919931. Buturlin S. A. 1907. Ornithologischen Monatsbericht, pg. 190. Clement, M., D. Posada, and K. Crandall. 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology 9:1657-60. Cracraft, J. 1989. Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. In Speciation and its consequences (ed. D. Otte and J. A. Endler), pp. 28-59. Sunderland, MA: Sinauer. Cramp, S. 1988. Handbook of the birds of Europe, the Middle East and North Africa, Vol. 5, Oxford Univ. Press, Oxford. Czikeli, H. 1985. Biochemical genetics in the yellow wagtail complex: conservative avian protein evolution confirmed by isoelectric focusing. Biochemical Systematics and Ecology 12:171-181. 29
Degnan, S. M. 1993. The perils of single gene trees – mitochondrial versus single-coy nuclear DNA variation in white-eyes (Aves: Zosteropidae). Molecular Ecology 2:219-225. Dement’ev, G. P., and N. A. Gladkov [eds]. 1954. Birds of the Soviet Union, Vol. 5. Sovetskaya Nauka, Moscow. Fedorov, V., A. Goropashnaya, G. H. Jarrell, and K. Fredga. 1999. Phylogeographic structure and mitochondrial DNA variation in true lemmings (Lemmus) from the Eurasian arctic. Biological Journal of the Linnean Society 66:357-371. Frenzel, B., M. Pecsi, and A.A. Velichko (eds.). 1992. Atlas of paleoclimates and paleoenvironments of the northern Hemisphere. Hungarian Acad. of Sciences. Budapest. Fry, A. J., and R. M. Zink. 1998. Geographic analysis of nucleotide diversity and song sparrow (Aves: Emberizidae) population history. Molecular Ecology 7:13031313. Fu, Y.-X. 1997. Statistical tests of neutrality against population growth, hitchhiking and background selection. Genetics 147:915-925. Gmelin G. F. 1789. Systema Naturae, & c. Linneai, Caroli. Edito decima tertia, aucta reformata, cura J. F. Gmelin. Hackett, S. J. 1996. Molecular phylogenetics and biogeography of tanagers in the genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution 5:368-382. Hewitt, G. M. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58:247-276. Hewitt, G. M. 2000. The genetic legacy of the Quaternary ice ages. Nature 405:907-913. Hillis, D. M., B. K. Mable, A. Larson, S. K. Davis, and E. A. Zimmer. 1996. Nucleic acids IV: sequencing and cloning. In Molecular systematics (eds. D. M. Hillis, C. Moritz, and B. K. Mable), 2nd ed., pp. 321-381. Sunderland, MA:Sinauer. Knowles, L. L. and W. P. Maddison. 2002. Statistical phylogeography. Molecular Ecology 11:2623-2635. Koblik, E. A., S. Rohwer, S. V. Drovetski, C. S. Wood, A. V. Andreev, D. A. Banin, and 30
V. B. Masterov. 2001. Faunistic records from the eastern regions of Russia. Ornithologia 29:47-57. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Pääbo, F. X. Villablanca, and A. C. Wilson. 1989. Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proceedings of National Academy of Sciences USA 86:6196-6200. Kryukov, A. P., and H. Suzuki. 2000. Phylogeography of carrion, hooded, and jungle crows (Aves, Corvidae) inferred from partial sequencing of the mitochondrial cytochrome b gene. Russian Journal of Genetics 36:922-929. Mantel, N. A. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27:209-220. Merila, J., M. Bjorkland, and A. J. Baker. 1997. Historical demography and present day population structure of the greenfinch Carduelis chloris – an analysis of mtDNA control-region sequences. Evolution 51:946-956. Moore, W. S. 1995. Inferring phylogenies from mtDNA variation: mitochondrial-gene trees versus nuclear-gene trees. Evolution 49:718-726. Moritz, C. 1994. Defining evolutionarily significant units for conservation. Trends in Ecology and Evolution 9:373-375. Nee, S., E. C. Holmes, A. Rambuat, and P. H. Harvey. 1996. Inferring population history from molecular phylogenies. In New uses for new phylogenies (eds. P. H. Harvey, A. J. Leigh Brown, J. Maynard Smith, and S. Nee), pp. 399-411. Oxford Univ. Press, New York. Petit, E., L. Excoffier, and F. Mayer. 1999. No evidence of bottleneck in the postglacial recolonization of Europe by the noctule bat (Nyctalus noctula). Evolution 53:1247-1258. Portenko, L. A. 1973. Birds of the Chukotsk peninsula and Wrangel island, Vol. 2. Nauka, Leningrad. Posada, D., and K. Crandall. 2001. Intraspecific gene genealogies - trees grafting into networks. Trends in Ecology and Evolution 16:37-45. 31
Posada, D., K. Crandall, and A. R. Templeton. 2000. GeoDis: a program for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Molecular Ecology 9:87-488. Rogers, A. R., and H. Harpending. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9:552-569. Rohwer, S., S. V. Drovetski, and C. S. Wood. 2001. Bird specimens in the Burke Museum from Russia and Kazakhstan. Ornithologia 29:260-281. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Ehrlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Salzburger, W., J. Martens, A. A. Nazarenko, Y.-H. Sun, R. Dallinger, and C. Sturmbauer. 2002 Phylogeography of the Eurasian Willow Tit (Parus montanus) based on DNA sequences of the mitochondrial cytochrome b gene. Molecular Phylogenetics and Evolution 24:26-34. Schneider, S., J.-M. Dueffer, D. Roessli, and L. Excoffier. 2000. Arlequin ver. 2.0: A software for population genetic data analysis. Genetics and Biometry laboratory, University of Geneva, Switzerland. URL: anthropologie.unige.ch/arlequin/ Schneider, S., and L. Excoffier. 1999. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. Genetics 152:1070-1089. Shimodaira, H., and M. Hasegawa. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16:11141116. Slatkin, M., and R. R. Hudson. 1991. Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics 129:555562. Swinhoe R. 1863. Proceedings of Zoological Society of London, pg.334. Swofford, D. L. 2000. PAUP*. Phylogenetic analysis using parsimony (*and other 32
methods). Version 4.0b2. Sinauer, Sunderland, MA. Tarr, C. L. 1995. Primers for amplification and determination of mitochondrial controlregion sequences in oscine passerines. Molecular Ecology 4:527-529. Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7:381-397. Templeton, A. R., E. Routman, and C. Phillips. 1995. Separating population structure from population history: a cladistic analysis of the geographical distribution of mitochondrial DNA haplotypes in Tiger Salamander, Ambystoma tigrinum. Genetics 140:767-782. Templeton, A. R., E. Boerwinkle, and C. F. Sing. 1987. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping. I. Basic theory and an analysis of alcohol dehydrogenase activity in Drosophila. Genetics 117:343-351. Voelker, G. 2002. Systematics and historical biogeography of wagtails: dispersal versus vicariance revisited. Condor 104:725-739. Zhang, D-X., and G. M. Hewitt. 1996. Nuclear integrations: challenges for mitochondrial DNA markers. Trends in Ecology and Evolution 11:247-251. Zink, R. M. 1997. Phylogeographic studies of North American birds. In Avian molecular evolution and systematics (ed. D. P. Mindell), pp. 301-324. San Diego: Academic Press. Zink, R. M., G. F. Barrowclough, J. L. Atwood, and R. C. Blackwell-Rago. 2000. Genetics, taxonomy and conservation of the threatened California Gnatcatcher. Conservation Biology 14:1394-1405. Zink, R. M., S. Drovetski, and S. Rohwer. 2002a. Phylogeographic patterns in the Great Spotted Woodpecker (Dendrocopos major) across Eurasia. Journal of Avian Biology 35:175-178. Zink, R. M., S. Rohwer, S. Drovetski, R. C. Blackwell-Rago, and S. L. Farrell. 2002b. Holarctic phylogeography and species limits of Three-toed Woodpeckers. Condor 104:167-170. 33
Zink, R. M., S. J. Weller, and R. C. Blackwell. 1998. Molecular phylogenetics of the avian genus Pipilo and a biogeographic argument for taxonomic uncertainty. Molecular Phylogenetics and Evolution 10:191-201.
34
TABLE 1.1. Traditional taxonomy and scoring of 6 morphological characters for 13 taxa of Motacilla males. Codes: ol=olive, olgn=olive-green, ygn=yellow-green, yol=yellow-olive, lgy=light grey, gy=grey, dgy=dark grey, dgn = dark green, bl=black, y=yellow, w=white, - absent, + present. Taxonomic codes for M. flava: L = M. f. lutea complex, F = M. f. flava complex (1 – flava group, 2 – feldegg group, 3 – thunbergi group; from Cramp [1988]). MtDNA groupings from Figure 2; WE = West, EA = East, NE = Northeast, SE = Southeast.
MtDNA
Taxonomic
Crown
Group
code
flava lutea
WE
L
y
flava flava
WE
F-1
flava beema
WE
flava leucocephala
Back color
Breast
Auricular
Supercilium
Chin color
spots
patch color
ygn
-
y
y
y
gy
yol
-
dgy
w
w
F-1
lgy
yol
-
lgy
w
w
WE
F-1
w
ygn
-
w
-
w
flava feldegg
WE
F-2
bl
olgn
-
bl
-
y
flava thunbergi
WE
F-3
dgy
olgn
+
bl
-
y
flava plexa
NE
F-3
dgy
ol
+
bl
-
y
flava tschutschensis
NE
F-3
gy
ol
+
dgy
w
w
flava taivana
SE
L
ygn
olgn
-
dgn
y
y
flava macronyx
SE
F-3
dgy
olgn
-
dgy
-
w
citreola werae
WE
-
y
gy
-
y
-
y
citreola citreola
EA
-
y
dgy
-
y
-
y
citreola quassatrix
EA
-
y
gy
-
y
-
y
35
TABLE 1.2. Genetic characteristics of Motacilla flava and M. citreola samples; see Figure 1 for locality information. Fu’s F-values presented only if significant (P < 0.05).
Locality
Subspecies
N
Number of
π
Fu’s F
haplotypes Western samples of M. flava Almaty
feldegg
6
6
0.0045
-2.90
Astrakhan’
feldegg, flava
7
7
0.0044
-
Kursk
flava, beema
12
12
0.0033
-8.54
Moscow
flava, thunbergi
18
14
0.0025
-7.29
Tula
flava
5
5
0.0041
-1.90
Vologda
flava
6
5
0.0027
-
Vyatka
beema, flava,
19
19
0.0047
-13.02
17
12
0.0042
-3.53
lutea Tyva
beema, leucocephala
Southeastern samples of M. flava Khabarovsk
taivana
5
5
0.0024
-
Sakhalin
taivana
6
4
0.0029
-
Mongolia
macronyx
9
8
0.0065
-
Northeastern samples of M. flava Alaska
tschu-
6
5
0.0022
-2.26
tschensis Anabar
plexa
10
8
0.0016
-2.97
Anadyr
tschu-
7
7
0.0022
-2.02
tschensis Noyabr’sk
plexa
8
7
0.0024
-3.00
Cherskiy
plexa
8
2
0.0005
-
Yamal
plexa
18
9
0.0016
-3.32
36
TABLE 1.2. Continued.
Locality
Subspecies
N
Number of
π
Fu’s F
haplotypes Western samples of M. citreola Kursk
werae
3
3
0.0024
-
Moscow
werae
2
2
0.0052
-
Tyva
werae,
8
5
0.0015
-
werae X quassatrix Eastern samples of M. citreola Tyva
quassatrix
5
4
0.0014
-
Mongolia
quassatrix
6
3
0.0016
-
Buryatia
citreola
2
2
0.0022
-
Anabar
?
1
1
0
Yamal
citreola
11
7
0.0020
-3.12
37
TABLE 1.3. Number of base pairs sequenced for various gene regions. LCR4 and H1248 (Tarr 1995) amplified the control region (CR), L5215/H5578 (Hackett 1996) amplified ND2, L10702/H11289 (J. G. Groth, pers. comm.) amplified ND3, and L14841 (Kocher et al. 1989)/H15299 (Hackett 1996) amplified Cyt b. NA = not amplified.
Taxa # of individuals
CR
ND2
ND3
Cyt b
Total Base Pairs
M. flava
167
387
352
254
NA
993
M. citreola
38
458
388
561
NA
1407
Motacilla species
22
458
388
561
425
1832
38
TABLE 1.4. Population pairwise Fst-values for western Motacilla flava samples. An asterisk indicates P < 0.05. Kursk
Almaty
Moscow
Astrakhan’
Tula
Vologda
Vyatka
Kursk
0.0
Almaty
0.049
0.0
Moscow
0.0
0.06
0.0
Astrakhan’
0.117*
0.0
0.157*
0.0
Tula
-0.004
0.051
0.016
0.134*
0.0
Vologda
0.037
0.125*
0.021
0.163*
0.006
0.0
Vyatka
0.151*
0.163*
0.186*
0.172*
0.127*
0.171*
0.0
Tyva
0.090*
0.093
0.102*
0.137*
0.098
0.110*
0.184*
39
TABLE 1.5. Pairwise Fst-values for eastern Motacilla flava samples. Note that all values between northeastern and southeastern samples are significant. An asterisk indicates P < 0.05. Sakhalin Sakhalin
Khabarovsk
Mongolia
Cherskiy
Anadyr
Yamal
Noyabr’sk
Alaska
0.0
Khabarovsk
0.128
0.0
Mongolia
0.133*
0.062
0.0
Cherskiy
0.930*
0.939*
0.810*
0.0
Anadyr
0.885*
0.886*
0.767*
0.035
0.0
Yamal
0.914*
0.915*
0.836*
-0.020
0.047
0.0
Noyabr’sk
0.876*
0.875*
0.766*
0.075
0.025
0.056
0.0
Alaska
0.885*
0.888*
0.763*
0.049
0.030
0.032
0.017
0.0
Anabar
0.897*
0.898*
0.797*
0.265*
0.129*
0.23*
0.195*
0.212*
40
TABLE 1.6. Pairwise Fst-values for Motacilla citreola samples. Note that all values between western and eastern samples are significant. An asterisk indicates P < 0.05. Yamal
Buryatia
Mongolia
Tuva-quassatrix
Tuva-werae
Moscow
Yamal
0.0
Buryatia
0.043
0.0
Mongolia
0.048*
0.125
0.0
Tyva-quassatrix
0.148*
0.174
0.104
0.0
Tyva-werae
0.885*
0.908*
0.900*
0.909*
0.0
Moscow
0.861*
0.826
0.869
0.884*
0.296
0.0
Kursk
0.867*
0.859
0.877*
0.891*
0.175
0.005
41
Figure 1.1. General location of collecting sites. Black circles indicate western clade of M. flava, gray circles- northeastern flava, open circles- southeastern flava, black squares- western M. citreola, open squares-eastern citreola.
42
Figure 1.2. Maximum likelihood tree showing relationships among major groups of M. flava, M. citreola, and near relatives, rooted with outgroups. Lab numbers precede locality names. Numbers above branches are percentages of 100 maximum likelihood bootstrap replicates supporting that branch.
58Mongolia 100
M. flava - SE
59Mongolia 89
65Mongolia 4Moscow 93
M. citreola - WE
7Kursk 105Noyabr'sk
100
109Anadyr 100
108Cherskiy 112Yamal
M. flava - NE
126Alaska 84
143Anabar
92
2Yamal 3Yamal 99
95
9Buryatia
M. citreola - EA
24Yamal M. cinerea 100
M. alba 7 M. lugens 31 100
68Kursk 88Moscow
M. flava - WE
163Vyatka 0.005 substitutions/site
43
Figure 1.3. Plot of nucleotide diversity and latitude in three groups of M. flava and two groups of M. citreola. Symbols correspond to those on Fig. 1.1.
0.007
Nucleotide Diversity
0.006 0.005 M.flava-We M. flava-NE
0.004
M. flava-SE 0.003
M. citreola-We M. citreola-EA
0.002 0.001 0 40
45
50
55
60
65
70
75
Degrees North Latitude
44
Figure 1.4. Observed and expected mismatch distributions for M. flava (a. Moscow, b. Vyatka, c. Tyva, d. Yamal, e. Anabar), and M. citreola (f. Yamal).
b. M. flava - Vyatka
a. M. flava - Moscow 0.4 Observed
0.3
Expected
0.35 0.3 Frequency
Frequency
0.4 0.35
0.25 0.2 0.15
0.25 0.2 0.15
0.1
0.1
0.05
0.05
0
0 0
2
4
6
8
10
12
0
2
0.4
0.4
0.35
0.35
0.3
0.3 Frequency
Frequency
6
8
10
12
10
12
10
12
d. M. flava - Yamal
c. M. flava - Tyva
0.25 0.2 0.15
0.25 0.2 0.15
0.1
0.1
0.05
0.05 0
0 0
2
4
6
8
10
0
12
2
e. M. flava - Anabar
4
6
8
f. M. citreola - Yamal
0.4
0.4
0.35
0.35
0.3
0.3 Frequency
Frequency
4
0.25 0.2 0.15
0.25 0.2 0.15
0.1
0.1
0.05
0.05
0
0 0
2
4
6
8
10
Number of pairw ise differences
12
0
2
4
6
8
Number of pairw ise differences
45
Figure 1.5. Haplotype network for northeastern M. flava. Each line in the network represents a single mutation, white circles represent unsampled haplotypes, white rectangles represent zero-step clades (haplotypes), gray rectangles- one-step clades, thick lines indicate two-step clades. Haplotype numbers followed by the locality (see Figure 1): Noya- Noyabr’sk, Anad- Anadyr, Alas- Alaska, Yaml- Yamal, Anab- Anabar, CherCherskiy.
46
Chapter 2
Mitochondrial DNA and plumage evolution in white wagtails 2.
We analyzed sequences of two mitochondrial DNA (mtDNA) gene regions (control region and ND2) totaling 1477 base-pairs from 232 specimens of the white wagtail (Motacilla alba) obtained from 27 localities throughout Eurasia. Although overall haplotype diversity was relatively low (0.79) and the most common haplotype was shared by 45% of individuals, belonging to six subspecies, a high level of population differentiation was detected. The mtDNA tree revealed three clades: (1) most individuals from Krasnodar (belonging to M. a. alba subspecies), (2) all individuals from Almaty and some from Primor'e (belonging to M. a. personata, M. a. lugens and M. a. leucopsis subspecies), and (3) the remaining individuals (representing all localities except Almaty and belonging to all subspecies). We suggest that these three clades represent historically isolated populations that relatively recently came into secondary contact in Krasnodar and Primor'e. The Krasnodar population appeared to receive immigrants from other localities, but distinctive haplotypes from this locality did not appear elsewhere, suggesting asymmetric gene flow. Signatures of recent gene flow between northern populations were detected, and there was no evidence of isolation by distance within the northern group of populations. Mismatch distributions for most localities were consistent with population expansions. We also analyzed 12 male plumage characters from 93 study skins sampled from 24 populations. Phylogenetic trees resulting from separate genetic and morphological analyses were incongruent. None of the six subspecies were reciprocally monophyletic in the mtDNA tree. Plumage evolution seems to be under strong sexual or natural selection, which favors particular phenotypes irrespective of the mitochondrial background. Complicated dispersal events at different evolutionary times could have obscured the effect of an earlier isolation event. The mtDNA tree did not support subspecies or species status for M. a. lugens and M. a. personata, which shared haplotypes with other subspecies of M. alba. Although M. alba was differentiated, recognizing multiple species is unwarranted without additional study.
2
The material in this chapter is part of the collaboration with R. M. Zink, S. Rohwer, E. A. Koblik, Y. A. Red'kin, I. V. Fadeev, and E. V. Nesterov. It is written in the style and format of the Journal of Avian Biology. Submission is pending.
36
The white wagtail (Motacilla alba) is a common passerine bird, whose range covers most of Eurasia, from the British Isles to Chukotka and western Alaska, and from the coast of the Arctic ocean to roughly 36° North latitude in Europe, and to 20° in southeastern Asia (Dement'ev and Gladkov 1954, Cramp 1988, Badyaev et al. 1996). White wagtails occur in a variety of habitats, excluding tall and dense vegetation, from cold regions in high latitudes to arid, hot climates, and are often associated with human settlements (Cramp 1988). Geographic variation of the white wagtail is largely defined by the distribution of black, white and gray in different regions of the body (Stepanyan 2003). Separating the white wagtail into subspecies has been difficult because different forms may hybridize where their ranges meet (Dement'ev and Gladkov 1954, Stepanyan 2003), and because individuals in some subspecies (like M. a. persica) resemble neighboring races (Cramp 1988). Sexual dimorphism exists in some geographical races, which along with the gradual geographic transitions from one form to another, further complicates assigning subspecies status to individuals. The black-backed wagtail (M. a. lugens), which inhabits coastal habitats of Kamchatka, Sakhalin and Primor’e, has recently colonized Alaska; it was given species status by the American Ornithologists' Union (1983). Stepanyan (2003) recognizes M. a. lugens and M. a. personata as species. Cramp (1988) distinguishes four subspecies-groups based on similarity of head pattern of breeding males: 1. M. a. personata and M. a. alboides, 2. M. a. ocularis and M. a. lugens, 3. M. a. alba (including M. a. dukhunensis) and M. a. yarrellii, and 4. M. a. baicalensis and M. a. leucopsis. The first member of each subspecies pair has a gray back and upperparts which differentiates it from the latter member of the pair, which has a black back and upperparts. Further, Cramp (1988) suggests that group 1 and group 2 are related, and that group 3 and group 4 are related, but that groups 1-2 were more “distant” from each other than were groups 3-4. Odeen and Alstrom (2001) assessed the genetic relationships among different subspecies of M. alba by comparing mitochondrial (mt) and nuclear DNA sequences. Their nuclear tree showed no resolution of subspecies, moreover one of two sampled individuals of Motacilla grandis was found in M. alba clade. On the basis of the mtDNA, 37
Odeen and Alstrom (2001) recognized two groups: 1. the alboides group, which included the subspecies M. a. alboides, M. a. leucopsis and M. a. personata, and 2. the alba group, which included M. a. alba, M. a. yarrellii, M. a. baicalensis, M. a. ocularis, M. a. lugens, and M. a. subpersonata. The breeding sites of some of the sampled birds used in Odeen and Alstrom’s (2001) mtDNA study were unknown, because blood from migrating individuals was used. Odeen and Alstrom (2001) assumed that subspecies were monophyletic, but were unable to test this assumption because for most subspecies a single individual was used. In the case of M. a. subpersonata, the two individuals they sampled were not each other's closest relatives. Zink et al. (2003) showed that incorporating multiple individuals and localities often dissolves clades thought to support subspecies, and thus the subspecies status needs further testing. Phenotypic similarities between subspecies may be caused by convergent evolution due to natural or sexual selection (Pavlova et al. 2003). Extensive geographic sampling with multiple individuals taken from the same locality is needed to document the mtDNA pattern of variation, and to test subspecies limits. In our study we made no a priori assumptions about monophyly and used geographic samples of breeding individuals, rather then currently recognized subspecies. Here, we characterize the geographic distribution of variation in plumage phenotypes and mtDNA haplotypes for White wagtails collected across Eurasia during the breeding season. We use the genetic data to infer the recent evolutionary history of the populations and then explore whether this history is reflected in the striking plumage differences seen across this region.
Methods Sampling
A total of 232 white wagtails from 27 populations sampled across Eurasia was collected during breeding season (Fig. 2.1). From almost all specimens a study skin with spread 38
wing was preserved and deposited at the Burke Museum, University of Washington, Seattle; the Moscow State University Zoological Museum, Moscow, Russia; State Darwin Museum, Moscow, Russia; or the Bell Museum, University of Minnesota, St. Paul. Tissue samples were either preserved in 96% ethanol, frozen in liquid nitrogen, or stored in lysis buffer (Longmire et al. 1997). The white-browed wagtail M. maderaspatensis, endemic to the Indian subcontinent, was used as an outgroup based on Voelker (2002). A tissue sample of M. maderaspatensis was loaned to us by American Museum of Natural History (tissue number AMNH 23223, Genbank number AF526470).
Morphological analysis
Motacilla alba dukhunensis occupies the south-eastern European part of Russia and westcentral Asia (Cramp 1988). It differs from M. a. alba by wider white borders on the wing coverts, white tertial edges, and in general has more white on the wing (Stepanyan 2003). However, extent of white area on the wing overlaps considerably between M. a. alba and M. a. dukhunensis. In this paper we refer to both subspecies as M. a. alba. Thus, our sampling included six subspecies (Fig. 2.1, Table 2.1). Only breeding males were used for morphological analysis because subspecies identification of females is difficult in many individuals (Cramp 1988). Study skins of 93 males from 24 populations (housed at the Burke Museum (WA) and Bell Museum (MN)) were scored for 12 plumage characters that have been used to define subspecies (Dement'ev and Gladkov 1954, Cramp 1988): (1) forehead color, (2) crown color, (3) nape color, (4) back color, (5) color of sides of neck, (6) color of ear-coverts, (7) cheek color, (8) chin color, (9) throat color, (10) black stripe through the eye, (11) amount of white on the wing, and (12) depth of black on chest (measured in mm from study skins). Characters 1-9 were scored as white, gray or black, and character 10 as presence or absence. Character 11 was defined as an area of white on the wing divided by the total wing area and was scored from photographs of spread wings using program Scion Image for Windows (available online from Scion corporation at http://www.scioncorp.com/). Three non-overlapping states were assigned to character 11: 39
white area less than 25% of wing (all subspecies except M. a. lugens and its hybrids with M. a. ocularis), white area 25-50% of wing (M. a. lugens and M. a. ocularis hybrids), and white area 50-100% of wing (M. a. lugens). Characters 1-11 were used in a maximum parsimony analysis of unique phenotypes (with M. maderaspatensis as an outgroup). Character 12 (length of black bib on the chest) was omitted from analysis because discrete states could not be assigned to it. Simple linear regression of values for character 12 on geographical coordinates was performed. Character 12 presumably depended on the way the study skin was prepared. Molecular lab methods
Isolation and purification of the DNA was performed using phenol-chloroform protocol (Hillis et al. 1996) or QIAamp Tissue Kit (QIAGEN, Valencia, California). Polymerase Chain Reaction (PCR) (Saiki et al. 1999) with Perkin-Elmer PCR reagents were used to amplify two mitochondrial gene regions. Primers L5215 (Hackett 1996) and H1064 (Drovetski in press) were used to amplify the complete NADH dehydrogenase subunit 2 (ND2) gene, primers LCR4 and H1248 (Tarr 1995) amplified part of the Control Region (CR) gene. For ND2, PCR started with 2.5 min at 95°C, followed by 40 cycles of denaturation at 94°C for 30 sec, annealing at 55°C for 30 sec and extension at 72°C for 1 min, and ended with an extension of 10 min at 72°C. The PCRs for CR were performed with 34 cycles of 1 min at 94°C, 1 min at 50°C, 1 min at 72°C with final extension for 10 min at 72°C. Qiaquick PCR Purification Kit (QIAGEN) was used to clean PCR products. Then either standard Sephadex columns were used to clean sequencing reactions, which were sequenced on ABI 310 automated sequencer using Dideoxy Terminator Kit Protocol, or cleaned PCR fragments were directly sequenced on ABI 3700 automated sequencer using BigDye chemistry (Applied Biosystems). Amplification primers and primers L347 (Drovetski in press) and H5578 (Hackett 1996) were used for sequencing of 1041 base pairs (bp) of ND2, primers LCR4 (Tarr 1995), LCON2 (Zink et al. 1998), 2LCRU (5'-GATGCACTTTGNCCNCATTC-3', designed by R. Blackwell-Rago) and 40
2HCRU (5'-GAATGNGGNCAAAGTGCATC-3', designed by R.B.-R.) were used to obtain 436 bp of control region.
41
Data analysis
Sequences were aligned and edited in Sequencher 3.1.1 (Gene Codes Corporation, Ann Harbor, Michigan). Sequence data have been deposited in GenBank under accession numbers (pending). Mitochondrial origin of sequenced DNA fragments was supported by the absence of stop-codons and the existence of a large number of haplotypes, which are inconsistent with nuclear copies (Zhang and Hewitt 1996). PAUP* (Swofford 2000) was used to generate maximum parsimony trees from equally weighted characters, then strict and 50% majority rule consensus trees were constructed. Maximum likelihood (ML) phylogenetic analyses were performed using PAUP*. ML model and parameters were determined by the Hierarchical Likelihood Ratio Test (hLRT) in Modeltest 3.06 (Posada and Crandall 1998). A likelihood ratio (LR) test was performed to evaluate whether the sequences have been evolving in a clocklike manner. Scores from ML trees with and without a molecular clock enforced (Felsenstein 1981) were compared and the LR was calculated as 2(ln Lclock-ln Lno clock) under the assumption that the LR was χ2 distributed with degrees of freedom (df) equal to the number of taxa minus two (Nei and Kumar 2000). Arlequin software (Schneider et al. 2000) was used to compute the number of haplotypes in population (with 0% allowed missing data per site; there were no gaps in sequences), nucleotide diversity (π), haplotype diversity (h), number of polymorphic sites (S), Fst, τ, θ0 and θ1, mismatch distributions. Tau (τ=2µt) is a relative measure of time since expansion, measured in number of generations ago (t). Theta0=2µN0 and theta1=2µN1, where µ is the mutation rate, N0 and N1 are effective population sizes before and after expansion. We also used Arlequin to perform Tajima's D (Tajima1996) and Fu’s Fs (1997) tests of selective neutrality, analysis of molecular variance (AMOVA), and Mantel’s (1967) test of pairwise Fst values vs. geographic distances. For amongpopulation comparisons we used only populations with sample sizes of four or more individuals. Mismatch distributions were computed for localities with sample sizes greater than 10 individuals. To test the empirical mismatch distribution against a model 42
of sudden expansion we used the generalized non-linear least-squares approach (Schneider and Excoffier 1999) as implemented in Arlequin. We regressed values of nucleotide diversity (π) against latitude and longitude expecting smaller values to be observed on the leading edge of population expansion (Hewitt, 2000). MEGA version 2.1 (Kumar et al. 2001) was used to construct a Minimum Evolution tree from population average pairwise sequence differences corrected for within population variability.
Results Analysis of morphological variation
Characters 1, 2, and 3 did not vary across the subspecies: all of the males in our sample had white foreheads and black crowns and napes. Although the white patch on the wing of M. a. personata differed visually from the two white wing-bars on the wing of M. a. alba, most subspecies (except M. a. lugens) had intermediate values of white area on the wing with overlapping ranges that did not allow discrimination between all subspecies. There were slight significant trends for northern (R2 = 0.1) and western (R2 = 0.32) birds having smaller bibs (P < 0.05). Bib length averaged smaller in M. a. alba than in other subspecies (P < 0.05). Phylogenetic analysis of morphological characters 4-11 showed subspecies to be well-defined with little variation within groups, although several individuals of mixed phenotype could not be assigned to any subspecies. The Tyva sample included M. a. personata, M. a. baicalensis and several intermediate phenotypes (Tsvetkov et al. 2003). One breeding individual looked like M. a. personata, but had the white malar stripe characteristic of winter plumage of M. a. personata (Fig. 2.2A). Another individual resembled M. a. personata with white cheeks (Fig. 2.2B). The third one looked like baicalensis but with a black throat. Phenotypes similar to this latter one were also found in Noyabr’sk (Fig. 2.2C) and Mezen’ among "pure" M. a. alba. Most individuals from Kamchatka possessed intermediate phenotypes between M. a. lugens and M. a. ocularis with varying amounts of black and gray feathers on the back, a mixture 43
of black and white feathers on the chin, and a white wing patch of various sizes (Figs. 2D and E, Koblik et al. 2001). More than one phenotype was also sampled from Krasnoyarsk, Khabarovsk and Primor’e (Fig. 2.1, Table 2.1). From two birds collected in Krasnoyarsk, one was a female M. a. personata. Only a skeleton and a wing were preserved from the other bird, so we could not assign it to subspecies. The wing of this bird differed from a typical wing of M. a. personata, so this bird might belong to either M. a. alba (M. a. dukhunensis) or M. a. baicalensis. There were three individuals of M. a. leucopsis and one M. a. ocularis among four individuals from Khabarovsk. In Primor'e, two individuals of M. a. lugens were collected on the shore of Sea of Japan, and seven M. a. leucopsis from more inland areas. An exhaustive search using the eight parsimony informative characters that describe the 10 unique phenotypes of M. alba yielded a maximum parsimony tree (Fig. 2.3A) (length 13, consistency index (CI) 0.69, rescaled consistency index (RC) 0.53), in which M. a. personata was a sister group to the rest of the phenotypes, M. a. ocularis, M. a. lugens and their hybrids were grouped together, and baicalensis was grouped with M. a. leucopsis. According to the morphological tree (Fig. 2.3A) white chin and black back color evolved twice in two groups of subspecies (one with and one without a black eye stripe).
Molecular phylogenetic analysis
Only 1468 out of 1477 sequenced nucleotides could be used when sites with missing data were excluded from analysis (there were no deletions in the sequences). For the 232 individuals sequenced there were 101 polymorphic sites (ten transversions) resulting in 87 haplotypes, excluding the outgroup. Maximum divergence between any pair of sequences of M. alba was 1.3%, and average divergence between M. alba and M. maderaspatensis was 3% - 3.6%. The most common haplotype was found in all populations except 7 (Norway, 44
Kursk, Kostroma, Sverdlovsk, Almaty, Krasnoyarsk and Cherskiy). This haplotype was shared by 105 individuals (45%) (Table 2.1). The second most common haplotype was shared by 15 individuals from nine populations (Norway, Moscow, Kostroma, Astrakhan', Sverdlovsk, Gorno-Altay, Krasnoyarsk, Khabarovsk and Sakhalin) (Table 2.1). Hierarchical likelihood criteria implemented in Modeltest 3.06 suggested TrN+G (Tamura and Nei 1993) as the most probable model of molecular evolution. Maximum likelihood (ML) analysis with parameters estimated from the data yielded two trees, with three major clades in common (Fig. 2.3B). One clade included 20 out of the 24 individuals sampled from the Krasnodar region near the Caucasus mountains. The second clade, which was sister to the first, included all four individuals sampled from Almaty and 5 of the 9 birds from Primor'e. The third clade comprised the rest of the individuals including the two most common haplotypes. We refer to the first clade as “southern Krasnodar” (SK), to the second clade as “southern Almaty-Primor’e” (AP) and to the third clade as “northern clade” (NC). We also refer to the Krasnodar, Almaty and Primor’e localities as “southern populations” and to the rest of localities as “northern”. Four individuals from Krasnodar and four from Primor'e possessed the haplotypes that belonged to the northern clade. Meanwhile, no haplotypes from the northern clade were found in Almaty, suggesting little or no gene flow between this locality and northern parts of the species’ range. The molecular clock hypothesis was not rejected (-lnL without molecular clock enforced = 3122.53, -lnL with molecular clock = 3082.97; df = 85, P = 0.99). For the MP analysis, the 41 parsimony informative characters resulted in more than 5,000 equally parsimonious trees (length = 170, CI = 0.82, RC = 0.71) whose strict consensus recovered same three ML clades as on the ML tree (Fig. 2.3b). Again, three individuals from Almaty were basal to the rest of Almaty-Primor'e clade. A haplotype from Magadan was basal to the other northern haplotypes (NC clade) in 86 percent of trees. The phenotypic and haplotypic trees (Fig. 2.3A,B) were incongruent. Subspecies 45
did not correspond to monophyletic groups of haplotypes.
46
Genetic variability
Nucleotide diversity (π) (Table 2.1) was high for southern samples: Primor'e (0.0049), Krasnodar (0.0042), and Almaty (0.0034). For northern localities π was low and ranged from 0.0003 in Khabarovsk to 0.0015 in Mezen'. Regression of π on sample size showed independence of these variables (P = 0.77). Overall nucleotide diversity was 0.0026 for all samples, 0.0009 for northern samples only.
Genetic diversity and gene flow
Overall haplotype diversity was 0.79 and ranged from 0.45 in Tyva to 1 in Almaty (Table 2.1). AMOVA showed a high level of population differentiation; 50.8% of molecular variance was due to comparisons among geographical localities (P < 0.05). When samples were divided into three groups (Krasnodar, Almaty with Primor'e and all northern samples pooled together), 75.8% of variance distributed among groups (Fct = 0.76, P < 0.05) and 23.1% within populations. An AMOVA based on subspecies failed to explain genetic variation (Fct = -0.11, P = 0.66). Overall Fst for only northern populations was low (0.07, P < 0.05) with 93% of the variance being accounted for by differences among individuals within populations. Significant pairwise Fst-values (Table 2.2) between geographic localities ranged from 0.02 (Tyva-Magadan) to 0.85 (Tyva-Almaty). All Fst-values involving Krasnodar were large and significant (ranging from 0.5 to 0.71). Comparing Krasnodar with neighboring Astrakhan' yielded an Fst of 0.64, showing a high degree of population differentiation with restricted or no gene flow between these samples. High pairwise Fst values were observed for the Almaty and Primor'e populations, although values for Primor'e-Almaty, Primor'e-Mongolia and Primor'e-Khabarovsk comparisons were not significant, indicating possible gene flow to and from Primor’e. The Primor'e-Murmansk comparison also did not yield a significant Fst value, but given the large geographic distance between these localities, this likely resulted from sharing the common haplotype 47
rather than from gene flow. All pairwise Fst values for the Sverdlovsk population were also large and significant. This implies a history of isolation of the Sverdlovsk population from other populations we sampled. The remaining pairwise Fst values either were not significant or were less than 0.1, indicating generally high levels of gene flow. A minimum evolution tree for populations (Fig. 2.3C) displayed the same pattern as the ML and MP trees: northern populations were grouped together (with Sverdlovsk being on a long branch), and Krasnodar was a sister group to the clade containing the Primor’e and Almaty populations (with Almaty being on a long branch). Mantel’s test was performed only for northern populations because it was obvious from the phylogenetic analysis (Fig. 2.3B) that some other cause than isolation-bydistance was responsible for the genetic structure observed across all localities. No isolation-by distance effect was detected for northern populations (R2 = 0.006, P = 0.74). Regression of nucleotide diversity against latitude yielded an R2 of 0.30 (P < 0.05) (Fig. 2.4). However, statistical significance was due to high values of π for the three southern populations: Krasnodar, Almaty and Primor'e. Considering the fact that the northern samples were all more closely related to each other than to any of the southern samples, the test might not be valid due to violation of independence. When only northern localities were analyzed, there was no association between latitude and nucleotide diversity. Regression of π against longitude for all samples, and northern sites only, did not result in a significant R2. Therefore, we found no signature of leading edge expansion.
Population expansion
Pairwise differences within each northern population were distributed in accordance with the model of sudden population expansion (Fig. 2.5, Table 2.1) and had unimodal distributions (with the exception of Tyva). Both Krasnodar and Primor'e included individuals from two different clades and, therefore, had bimodal mismatch distributions (Fig. 2.6), but these localities still did not differ from a model of sudden expansion using the bootstrap test of Schneider and Excoffier (1999). The mismatch distribution for 48
southern Krasnodar (SK clade) haplotypes (Fig. 2.6) was unimodal and also did not differ significantly from the sudden expansion model. Although southern haplotypes from Krasnodar (SK clade) did display the signature of past population growth, this expansion had to be much older than those in the northern clade (NC), as can be judged from the right shifted mode for the Krasnodar mismatch distribution (Figs. 5 and 6). Estimates of tau (τ) for northern localities ranged from 0.76 in Noyabr'sk to 3.4 in Khabarovsk (Table 2.1), but were not structured geographically.
Tests of selective neutrality
Fu's Fs values (Table 2.1) were not significant for Krasnodar, Almaty, Primor'e, Murmansk, Sverdlovsk, Mongolia and Khabarovsk. For the rest of northern localities Fu's Fs were significant and negative (P < 0.05). This result suggests either deviations from neutral evolution in stable populations or population growth. The values of Tajima's D (Table 2.1) were significantly negative (P < 0.05) for Moscow, Medvedevo, Mezen', Yamal, Noyabr'sk, Tyva, Magadan and Anadyr. Rejection of neutrality suggests that a population has recently passed through a bottleneck and has not yet reached equilibrium.
Discussion Plumage evolution
Our phylogenetic tree constructed from morphological characters (Fig. 2.3A) is mostly consistent with the groups suggested by Cramp (1988), but not in the relationships among groups, because M. a. lugens and M. a. ocularis were not most closely related to M. a. personata. Our tree was not consistent with the two groups of subspecies as suggested by Odeen and Alstrom (2001), because M. a. personata had more similarities with the outgroup, the white-browed wagtail, than with M. a. leucopsis, and M. a. leucopsis was 49
most closely related to M. a. baicalensis. Several convergent evolutionary patterns emerged from our phenotypic tree (Fig. 2.3A): black back color evolves in M. a. lugens and in M. a. leucopsis, making these subspecies distinguishable from M. a. ocularis and M. a. baicalensis correspondingly. Two additional taxa that we did not sample develop black backs: M. a. yarrellii from Britain and Ireland, which looks much like nominate M. a. alba except for the black back, and M. a. alboides from Himalayas, which differs from M. a. personata of central Asia only by its black back. Another character state evolving more than once is white chin color (Fig. 2.3A). Interestingly, both back and chin color change during the prebasic molt: all black-throated subspecies develop white chins and throats in their winter plumage and some of black-backed birds become gray-backed in winter (Alstrom and Mild 2003, Badyaev et al. 1996). It is possible that summer plumage characters primarily function in male-male competition and are of little importance to female mate choice. The more this is true, the more likely will be the occurrence of mixed pairings and the development of hybrid zones where different subspecies make contact. Such interbreeding facilitates high gene flow between populations. Given that up to six subspecies can be present on the same wintering grounds (in India or southeast Asia) (Alstrom and Mild 2003), molting into virtually indistinguishable phenotypes might be adaptive to winter survival in flocks, because this prevents intrasexual aggression (Roskaft and Rohwer 1987).
Incongruence of morphological and molecular trees
Our analysis of mtDNA variation does not support subspecies relationships suggested by Cramp (1988) or Odeen and Alstrom (2001), because none of the subspecies studied by us formed a clade. Among avian species, distinctive plumages are almost always accompanied by fixed differences in mtDNA, but the situation is different at the population and subspecies level. For example, in Yellow wagtails (Motacilla flava) many birds of very different appearances possessed the same mtDNA sequences, whereas in several cases similarly looking birds did not share a recent common ancestor (Pavlova et 50
al. 2003). Although plumages of birds can reflect the evolutionary history of populations (Alstrom and Mild 2003), mtDNA evidence often does not support morphologically defined subspecies as evolutionary units (Ball and Avise 1992, Zink 2004). Thus, mtDNA and morphological patterns can vary independently. There were three clades on the mtDNA gene tree of White wagtails. The northern clade (NC) included individuals from all subspecies, the southern Krasnodar (SK) clade included only birds belonging to M. a. alba subspecies, and the southern Almaty-Primor'e (AP) clade included three phenotypes: M. a. personata, M. a. leucopsis and M. a. lugens (Fig. 2.3B). In some localities individuals from two non-sister clades shared the same phenotypes. In Krasnodar, the birds from southern (SK) mtDNA clade do not share a most recent common ancestor with the individuals belonging to northern (NC) clade, although phenotypically they all belong to M. a. alba subspecies, as do most of the birds from western Eurasia (Fig. 2.7A). Motacilla a. personata was also represented by individuals from two different clades. Therefore, with respect to mtDNA, individuals with M. a. personata phenotype from Almaty that belonged to Almaty-Primor'e (AP) clade in mtDNA tree were not closest relatives to M. a. personata from Gorno-Altay and Tyva, which possessed haplotypes from northern (NC) clade (Fig. 2.7B). There are several alternative explanations for the observed distribution of haplotypes and phenotypes. For example, phenotypically similar members of both clades could have retained ancestral plumage. However, this could only explain phenotypehaplotype distribution of only one of the subspecies. Alternatively, sexual or natural selection could favor a particular phenotype in a region irrespective of the mitochondrial background of individuals. Lastly, sex-biased dispersal, where females are more philopatric than males, could have prevented haplotypes from mixing while facilitating plumage similarity. However, none of these interpretations explain the complicated phenotype-haplotype distributions in Primor'e. Both subspecies represented in our sample from Primor'e had individuals belonging to two mtDNA clades (Fig. 2.7C). Three individuals with M. a. leucopsis phenotype shared the most common haplotype of the northern (NC) clade, whereas the remaining four M. a. leucopsis belonged to the southern 51
Almaty-Primor'e (AP) clade. Two M. a. lugens from Primor'e also belonged to different clades. It is possible that introgression of northern haplotypes to Krasnodar and Primor’e occurs via females that were hatched in the north but stopped “short” to breed in more southern localities. If sexual or natural selection drives evolution of phenotypes, then it should have acted much faster than mtDNA lineage sorting, because none of the subspecies are monophyletic on the mtDNA tree. To conclude, the absence of congruence between morphological and mtDNA variation can be explained by the different time frames of evolutionary history that are recovered with these data sets. It takes approximately 2Nef generations (Nef is inbreeding effective female population size) for mtDNA to reach the stage of reciprocal monophyly, whereas the evolution of morphological characters can take several orders of magnitude less time, because these characters are likely polygenic and under the relentless force of sexual selection (Rohwer 1982, West-Eberhard 1983, Rohwer and Roskaft 1989). Complicated dispersal events at different times in species’ evolution can obscure the effects of earlier isolating events. Field experiments establishing whether rapid color evolution is driven by male-male competition, by female choice, or by other forces are needed.
Phylogeography and Pleistocene history of populations
Diversity in mtDNA was low across the continent with only 38% of haplotypes being found in single individuals and 45% of the individuals from many distant populations sharing the same haplotype. The number of unique haplotypes is low compared to yellow wagtails, where for sequences of about 1000 bp, 77% of haplotypes from the western clade, 54% from the northeastern clade and 85% from the southeastern clade were unique (Pavlova et al. 2003). Insignificant values from Mantel’s test and the presence of two widespread haplotypes suggest a history of recent gene flow between northern sampling sites for white wagtail. Thus, although congeneric, and both consisting of three haplotype groups, white and yellow wagtails experienced differing recent histories. 52
Deep phylogeographic splits are traditionally interpreted as the result of a longterm isolation. When the geographic distributions of the clades overlap, as with white wagtail, at least two alternative interpretations are possible: historical isolation with recent secondary contact, or incomplete lineage sorting. It is unlikely that the mixture of southern and northern haplotypes in Krasnodar (individuals from SK and NC clades) and Primor'e (AP and NC clades) is the result of incomplete lineage sorting, because the three clades are geographically structured with Krasnodar and Primor'e being on the borders of the clade distributions. Our Krasnodar sample includes birds collected at six localities. The four Krasnodar birds with northern haplotypes (from NC) were all collected at sea level near the northeastern shore of the Black sea and the eastern shore of Azov sea (two localities from the northwestern part of Krasnodar region). Whereas some individuals with southern haplotypes (SK clade) were also found in these localities, the other birds with southern haplotypes (SK clade) were collected at higher elevations on the slopes of Caucasus mountains (four localities). No haplotypes from the Krasnodar (SK) clade were found in Astrakhan', even though these populations are less than 900 km apart. According to haplotype tree and pairwise Fst, there seems to be no northward gene flow from Almaty to Tyva or Gorno-Altay. Unfortunately our sample sizes from Almaty and Gorno-Altay were too small to allow strong conclusions and larger samples from these areas could reveal haplotypes that are shared between these populations. The occurrence of the most common haplotype in most northern localities, in Krasnodar and Primor'e suggests movement of genes from northern to southern localities. Therefore, we believe that the best historical scenario for white wagtail is a historical isolation of the southern parts of the range (Krasnodar, Almaty and Primor'e), with Primor'e and Krasnodar being the zones of secondary contact between two clades. Recently Irwin (2002) showed that deep phylogeographic splits may arise without geographic barriers to gene flow. The likelihood of observing such splits increases as the individual dispersal distance and population size increase. Whereas the presence of a geographic barrier (Caucasus) to gene flow between southern Krasnodar (SK) and northern (NC) clades is evident, the cause of the split between Krasnodar (SK) and 53
southeastern localities (Almaty and Primor'e, AP) is not obvious. There are no apparent gaps in white wagtail distribution between these localities. However, regions of poor habitat quality (such as mountain ranges of middle eastern Eurasia) can act as the barrier to high gene flow (Irwin 2002). Therefore this phylogeographic split between two southern clades (SK and AP) may be the result of stochastic processes and not of a geographic barrier. Statistical tests for mismatch distributions have little power against the population growth hypothesis (Ramos-Onsins and Rozas, 2002). Generally, a unimodal distribution of pairwise differences between individuals in each population indicates that the population has been growing at some point in the past, whereas bimodal or ragged distributions suggest stable populations (Rogers and Harpending 1992). Because we believe that the best historical scenario for Primor'e and Krasnodar is recent secondary contact between southern and northern haplotypes, the assumption of the sequence data being distributed according to sudden expansion model was violated for these populations. Therefore, estimates of population parameters are unreliable for Krasnodar and Primor'e. We conclude that Krasnodar population (modern southern Krasnodar (SK) haplotypes) and most of our northern populations have undergone population expansion, with Krasnodar expansion predating expansions on the North. No evidence of bottlenecks was detected by Tajima's D test for the two southern clades (SK and AP) and southern populations from the northern clade (NC): Astrakhan', Sverdlovsk, Baikal (Irkutsk and Buryatia), Mongolia, Sakhalin and Kamchatka. High adaptability of white wagtail to a range of climates could have allowed relatively large populations to survive the maximum cooling of the Late Pleistocene, when most of these sample localities would have been covered with steppe-tundra (Adams 1997). Meanwhile, most of Europe was covered by glacial ice, and polar desert was in Northern Siberia and Sayan mountain region (Adams 1997). Therefore, as evidenced by Tajima's D test, Moscow, Medvedevo, Mezen', Yamal, Noyabr'sk, Tyva, Magadan and Anadyr populations must have colonized their range relatively recently. At the same time the historic Krasnodar population could have survived in large 54
numbers in the warmer climate southeast of the Black Sea, being separated from northern populations by Caucasus glaciation. Our samples from Primor'e come from the zone of Pleistocene grasslands (Adams 1997). Primor'e was isolated from Sakhalin by polar desert at the mouth of the Amur, which could have prevented contact between northern haplotypes (from NC) with haplotypes from Almaty-Primor'e (AP) clade. Meanwhile the Almaty population is situated in the zone of polar desert, so birds must have arrived in this region from more eastern grasslands when the climate became more suitable. Warmer climates of the southeast (contemporary eastern China) may have allowed wagtails to maintain fairly high effective population sizes, which would explain the higher nucleotide diversity in Almaty-Primor'e (AP) clade. We hypothesized that ancestors of modern birds with northern haplotypes (NC) might have survived maximum cooling of Late Pleistocene in the steppe-tundra of Far Eastern part of Eurasia, because more western parts of Siberia have been mostly covered with polar desert (Adams 1997). A haplotype from Magadan was basal to the rest of the northern haplotypes on most of MP trees, thus supporting an eastern origin of the northern clade (NC). However, we did not find a westward decrease in nucleotide diversity in northern populations that would be expected if birds from eastern populations (Kamchatka, Sakhalin, Magadan, Anadyr) dispersed westward. In general, nucleotide diversity in the northern (NC) clade was much lower as in southern (SK and AP) clades. We conclude that white wagtail relatively recently colonized the north of its range, but genetic signature of leading edge expansion (in any direction) has been erased by high gene flow.
Species limits
Although M. a. lugens is the only subspecies found on Sakhalin, both M. a. ocularis and M. a. lugens occur together and hybridize in Kamchatka, exhibiting intermediate values of the characters that are used to distinguish M. a. lugens from M. a. ocularis (large amount of white on the wing and black back) (Koblik et al. 2001, Rohwer et al. 2001). 55
Because these forms share the same haplotypes and extensively hybridize there is no support for classifying M. a. lugens as a species regardless of species concept applied. There is also no support in mtDNA for recognition of M. a. personata as a species. Because all mitochondrial genes are linked and maternally inherited, the mtDNA tree is one of many possible gene trees. It may or may not reflect the true evolutionary history of populations. There are several ways to interpret the phylogenetic tree of mitochondrial haplotypes for white wagtail. If one believes that the three clades represent three historically isolated taxa that just recently came into contact, then three species could be recognized. These clades have narrow zones of secondary contact, just north of the Caucasus and in southern Primor'e. However, these three clades have no clear diagnostic morphological differences. Study of independent sets of data, such as rapidly evolving nuclear DNA regions, is needed to evaluate the validity of the mitochondrial phylogenetic hypothesis. Alternatively, on the basis of high recent gene flow between most northern populations and a widespread haplotype that occurs in most populations including Krasnodar and Primor'e, it can be concluded that M. alba is a single biological species. Lastly, one might also argue that morphological subspecies must be recognized as recently evolved species if the narrow hybrid zones are stable and maintained by selection against hybrids. Extensive morphological study of hybrid zones would be needed to test the latter interpretation. We suggest that the white wagtail is best considered as a single species, given available data.
Acknowledgements We thank the Burke museum for curatorial assistance, S. Farrell and M. Westberg for laboratory assistance. We are grateful to S. Drovetski, D. Banin, I. Karagodin, A. Jones, B. Barber, C. S. Wood, B. Schmidt, R. C. Faucett for logistical help with expeditions and collecting. S. Birks (Burke Museum) subsampled some tissues. S. Drovetski provided useful comments. Special thanks to L. Pavlova who illustrated different subspecies. We are grateful to G. Eddy for funding fieldwork. Additional support came from NSF (DEB 56
9707496) and the Dayton-Wilkie fund (Bell Museum).
References Adams, J. M. 2002. Global land environments since the last interglacial. See http://members.cox.net/quaternary/. Alstrom P. and Mild, K. 2003. Pipits and Wagtails. - Princeton Univ. Press. American Ornithologists' Union. 1983. Check-list of North American birds. 4th ed. - Am. Ornith. Union, Washington, D.C. Badyaev, A. V., Gibson, D. D. and Kessel, B. 1996. White wagtail (Motacilla alba) and black-backed wagtail (Motacilla lugens). In The Birds of North America, No. 236-237 (Poole, A. and Gill, F., eds.). -The Academy of Natural Sciences, Philadelphia, PA, and The American Ornithologists’ Union, Washington, D.C. Ball, R. M., Jr. and Avise, J. C. 1992. Mitochondrial DNA phylogeographic differentiation among avian populations and the evolutionary significance of subspecies. - Auk 109: 626-636. Cramp, S. 1988. Handbook of the birds of Europe, the Middle East and North Africa, Vol. 5. - Oxford Univ. Press, Oxford. Dement’ev, G. P., and N. A. Gladkov [eds]. 1954. Birds of the Soviet Union, Vol. 5. Sovetskaya Nauka, Moscow. Drovetski, S. V., Zink, R. M., Fadeev, I. V., Nesterov, Y. V., Koblik, Y. A., Red’kin, Ya. A., and Rohwer, S. 2003. Mitochondrial phylogeny of Locustella and related genera. - J. Avian Biol. in press. Felsenstein, J. 1981. Evolutionary trees from DNA sequences: A maximum likelihood approach. J. Mol. Evol. 17: 368-376. Fu, Y.-X. 1997. Statistical tests of neutrality against population growth, hitchhiking and background selection. - Genetics 147: 915-925. Hackett, S. J. 1996. Molecular phylogenetics and biogeography of tanagers in the genus Ramphocelus (Aves). - Mole. Phyl. Evol. 5: 368-382. 57
Hewitt, G. M. 2000. The genetic legacy of the Quaternary ice ages. - Nature 405: 907913. Hillis, D. M., Mable, B. K., Larson, A., Davis, S. K. and Zimmer, E. A. 1996. Nucleic acids IV: sequencing and cloning. In Molecular systematics (eds. D. M. Hillis, C. Moritz, and B. K. Mable), 2nd ed., pp. 321-381. - Sunderland, MA:Sinauer. Irwin, D. E. 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution, 56: 2383-2394. Koblik, E. A., Rohwer, S., Drovetski, S. V., Wood, C. S., Andreev, A. V., Banin, D. A. and Masterov, V. B. 2001. Faunistic records from the eastern regions of Russia. Ornithologia 29: 47-57. Kumar, S., Tamura, K., Jakobsen, I. B. and Nei, M. 2001. MEGA2: molecular evolutionary genetics analysis software. - Bioinformatics 17: 1244-1245. Longmire, L., Maltbie M., and Baker, R. J. 1997. Use of “lysis buffer” in DNA isolation and its implication for museum collections. Occasional Papers. - Museum of Texas Tech University, OP-163. Mantel, N. A. 1967. The detection of disease clustering and a generalized regression approach. - Cancer Research 27: 209-220. Nei, M. and Kumar, S. 2000. Molecular Evolution and Phylogenetics. Oxford Univ. Press, Oxford Odeen, A. and Alstrom, P. 2001. Evolution of secondary traits in wagtails (genus Motacilla). In Odeen A. Effects of post-glacial range expansion and population bottlenecks on species richness. Ph.D. thesis. - Uppsala University. Pavlova A., Zink, R. M., Drovetski, S. V., Red'kin, Ya. and Rohwer, S. 2003. Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history. - Auk 120: 744-758. Posada, D. and Crandall, K. A. 1998. Modeltest: Testing the model of DNA substitution. - Bioinformatics 14: 817-818. Ray N., M. Currat, and L. Excoffier. 2003. Intra-deme molecular diversity in specially expanding populations. Mole. Biol. Evol. 20: 76-86. 58
Ramos-Onsins, S. E. and Rozas, J. 2002. Statistical properties of new neutrality tests against population growth. - Mole. Biol. Evol. 19: 2092-2100. Rogers, A. R. and Harpending, H. 1992. Population growth makes waves in the distribution of pairwise genetic differences. - Mole. Biol. Evol. 9: 552-569. Rohwer, S. 1982. The evolution of reliable and unreliable badges of fighting ability. Am. Zool. 22: 531-546. Rohwer, S., Drovetski, S. V. and Wood C. S. 2001. Bird specimens in the Burke Museum from Russia and Kazakhstan. - Ornithologia 29: 260-281. Rohwer, S. and Roskaft, E. 1989. Results of dyeing male yellow-headed blackbirds solid black: implications for the arbitrary identity badge hypothesis. - Behav. Ecol. Sociobiol. 25: 39-48 Roskaft, E. and Rohwer, S. 1987. An experimental study of the function of the red epaulettes and the black body colour of male red-winged blackbird. - Anim. Behav. 35, 4: 1070-1077. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B. and Ehrlich H. A. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. - Science 239: 487-491. Schneider, S., Dueffer, J.-M., Roessli, D. and Excoffier, L. 2000. Arlequin ver. 2.0: A software for population genetic data analysis. - Genetics and Biometry laboratory, University of Geneva, Switzerland. URL: anthropologie.unige.ch/arlequin/ Schneider, S., and Excoffier, L. 1999. Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. - Genetics 152: 1070-1089. Stepanyan, L. S. 2003. Conspectus of the ornithological fauna of Russia and adjacent territories (within the borders of the USSR as a historic region). Academkniga, Moscow, Russia. Swofford, D. L. 2000. PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4.0b2. - Sinauer, Sunderland, MA. Tajima, F. 1996. The amount of DNA polymorphism maintained in finite population 59
when the neutral mutation rate varies among sites. - Genetics 143: 1457-1465. Tamura K., Nei M. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10: 512-526 Tarr, C. L. 1995. Primers for amplification and determination of mitochondrial controlregion sequences in oscine passerines. - Molec. Ecol. 4: 527-529. Tsvetkov, A. V., Red’kin, Ya. A., Koblik, E. A. 2003. On distribution and biology of wagtails in Tuva. - Russian Journal of Ornithology, Express-issue N 229, P. 768787. (In Russian.) Voelker, G. 2002. Systematics and historical biogeography of wagtails: dispersal versus vicariance revisited. - Condor 104: 725-739. West-Eberhard, M. J. 1983. Sexual selection, social competition, and speciation. - Quart. Rev. Biol. 58: 155-183. Zhang, D-X., and Hewitt, G. M. 1996. Nuclear integrations: challenges for mitochondrial DNA markers. - Trends Ecol. Evol. 11: 247-251. Zink, R. M., Drovetski, S., Questiau, S., Fadeev, I. V., Nesterov, E. V., Westberg M. C. and Rohwer, S. 2003. Recent evolutionary history of the bluethroat (Luscinia svecica) across Eurasia. – Mole. Ecol. 12: 3069-3075. Zink, R. M.. 2004. The role of subspecies in obscuring avian biological diversity and misleading conservation policy. – Proc. Roy. Soc. Lond B 271: 561-564. Zink, R. M., Weller, S. J. and Blackwell, R. C. 1998. Molecular phylogenetics of the avian genus Pipilo and a biogeographic argument for taxonomic uncertainty. Mole. Phylo. Evol. 10: 191-201.
60
Table 2.1. Genetic characteristics of 26 geographic samples of Motacilla alba (Fig. 2.1) as calculated by Arlequin with 0% missing data allowed. Nine individuals from Irkutsk and two from Buryatia were pooled. Subspecies were determined by male plumage following Dement'ev and Gladkov (1954). N/s- not significant (for mismatch-distribution is not different from sudden expansion model); * - significant at 0.05 level, unim-unimodal, bimod- bimodal distribution. (Fig. 2.4,2.5); no test, no est - Arlequin was not able
Theta 1
Fu’s Fs
Tajima's D
-
1
-
-
-
-
-
-
-
-
Kursk
51/34
alba
1
0
-
1
-
-
-
-
-
-
-
-
Smolensk
54/34
alba
1
1
-
1
-
-
-
-
-
-
-
-
66/34
alba
4
2
2
3
0.83
6.8
1.28
0
2645
-
n/s
-
24
2
28
15
0.95
42
1.37
3.88
890
n/s
n/s
31
19
13
11
0.63
7.2
1.27
0
3.59
-8.05*
-2.2*
Murmansk Krasnodar
44/37
Moscow
55/37
alba (dukhunensis) alba
pattern
number of
Mismatch
Theta 0
0
nucleotide
1
haplotypes
alba
sites
tau
diversity x104
haplotype diversity
# of polymorphic
common haplotype
# of ind. with most
Sample size
61/10
locality
Norway
Collecting
Subspecies
°East longitude
°North latitude/
to fit the model of sudden expansion.
bimod n/s unim n/s
61
Theta 1
Fu’s Fs
Tajima's D
7
0.69
5.8
1.09
0
2420
unim n/s
-4.72*
-1.8*
Kostroma
58/41
alba
2
0
0
1
0
0
-
-
-
-
-
-
Mezen’
65/44
alba
13
4
12
10
0.92
14.6
no
no
no
unim
est.
est.
est.
no test
-6.82*
-1.8*
10
5
5
5
0.76
8.6
no
no
no
unim
est.
est.
est.
no test
-1.67*
n/s
5
0
2
3
0.7
6.8
1.29
0
1181
-
n/s
n/s
10
6
5
5
0.67
6.8
1.1
0
1048
unim n/s
-2.26*
-1.7*
10
7
3
4
0.53
4
0.76
0
831
unim n/s
-1.96*
-1.6*
Astrakhan’ Sverdlovsk
46/47
56/58
alba (dukhunensis) alba (dukhunensis) alba
pattern
number of
Mismatch
Theta 0
6
nucleotide
9
devo
haplotypes
16
Medve-
sites
tau
diversity x104
haplotype diversity
# of polymorphic
common haplotype
# of ind. with most
Sample size
alba
locality
60/38
Collecting
Subspecies
°East longitude
°North latitude/
Table 2.1. Continued
Yamal
68/68
Noyabr’sk
63/74
Almaty
42/75
personata
4
0
10
4
1
33.8
2.24
3.67
3978
-
n/s
n/s
51/85
personata
2
1
1
2
1
6.8
-
-
-
-
-
-
GornoAltay
(dukhunensis) alba (dukhunensis)
62
Tajima's D -2.0*
-
-
-
1983
unim n/s
-1.93*
n/s
0
5.3
-
n/s
n/s
13.3
0
16
n/s
n/s
3.4
0.77
0
1045
-
n/s
n/s
8.2
1.33
0
1892
unim n/s
-2.74*
n/s
Theta 0
Theta 1
Tyva
50/93
baicalensis,
6
5
0.45
5
2.3
0
0.88
2
0
0
1
0
0
-
-
-
6
5
5
0.71
7.1
1.17
0
4
2
4
3
0.83
13.5
3.4
9
3
15
6
0.89
48.9
4
3
1
2
0.5
12
6
5
6
0.76
hybrids Krasnoyarsk Irkutsk
57/97
dukhunensis
54/104 52/106
Mongolia
47/112
Primor’e
44/132
Sakhalin
n/s
or baicalensis
Buryatia
rovsk
bimod
personata,
9 baicalensis
Khaba-
pattern
tau
12
sites
16
personata,
Mismatch
Fu’s Fs
diversity x104
-2.17*
nucleotide
haplotype diversity
haplotypes
number of
# of polymorphic
common haplotype
# of ind. with most
Sample size
Subspecies
°East longitude
°North latitude/
locality
Collecting
Table 2.1. Continued
51/136 46/141
2 baicalensis leucopsis, lugens leucopsis, ocularis lugens
bimod n/s
63
0.79
9.2
11
5
7
6
0.8
unim n/s
-2.04*
n/s
no
no
no
unim
est.
est.
est.
no test
11
1.78
0
15
pattern
Mismatch
-1.83*
Theta 1
-3.59*
Theta 0
number of
Tajima's D
7
nucleotide
8
haplotypes
6
sites
13
Fu’s Fs
diversity x104
haplotype diversity
# of polymorphic
common haplotype
# of ind. with most
°East longitude
ocularis
tau
chatka
Sample size
Kam-
59/150
Subspecies
Magadan
°North latitude/
locality
Collecting
Table 2.1. Continued
lugens, 52/157
ocularis, hybrids
Anadyr
64/177
ocularis
15
6
11
9
0.85
12.1
1.98
0
1543
unim n/s
-5.0*
-1.81*
Cherskiy
69/158
ocularis
1
0
-
1
-
-
-
-
-
-
-
-
64
Table 2.2. Population pairwise Fst’s for comparisons between populations with sample sizes of four or more individuals. Values are presented only if significant at 0.05 level. Nine individuals from Irkutsk and two from Buryatia were pooled.
Mur Murmansk
-
Krasnodar
0.59
Krd
0.71
Medvedevo
0.67
Mezen'
0.64
Astrakhan'
0.64 0.47
Med
Mez
Ast
Sve
Alm
Tyv
IrBu
Mon
Prim
Kha
Sak
Mag
Kam
0.63
0.07
0.44
0.07
0.07
-
0.56
0.39
0.32
-
0.64
0.5
Noyabr'sk
0.64
0.61
0.72
Noy
-
Yamal
Almaty
Yam
-
Moscow
Sverdlovsk
Mos
0.53
0.84
0.83
0.78
0.75
0.07
0.08
0.6
0.85
-
0.07
0.52
0.8
0.03
0.31
0.68
0.68
Irkutsk+Buryatia
0.64
Mongolia
0.58
Primor'e
0.5
Khabarovsk
0.59
Sakhalin
0.65
Magadan
0.65
0.03
Kamchatka
0.63
0.06
0.07
0.07
Anadyr
0.65
0.05
0.04
0.06
0.03
0.5
0.07
-
0.74
Tyva
0.56
-
0.42
0.42
0.08 0.05
0.06
0.06
0.42
0.8
0.42
0.82
-
0.44
-
0.51
0.43
0.73
0.41
0.79
0.07
0.45
0.79
0.39 0.40
0.44
-
0.06
0.44
-
0.02
0.44
0.05
0.76
0.08
0.42
0.76
0.03
0.44
-
65
Fig. 2.1. General location of collecting sites and phenotypic variation of collected birds. Subspecies distribution differs from Alstrom and Mild (2003) and Stepanyan (2003) because not all sympatric subspecies were sampled at each locality. Motacilla a. dukhunensis (from Krasnodar, Astrakhan’, Sverdlovsk and Noyabr’sk) was pooled with M. a. alba. Of the two birds collected at Krasnoyarsk one was M. a. personata and the other was either M. a. alba or M. a. baicalensis (see text). Intermediate phenotypes are not shown. Subspecies were determined by male plumage only following Dement'ev and Gladkov (1954). Large circles indicate sample sizes of 10 or more individuals.
66
Figure 2.1.
67
Figure 2.2. Intermediate phenotypes of M. alba (Fig. 2.3A): A. Motacilla a. personata with white malar stripe from Tyva. B. Hybrid of M. a. personata and M. a. baicalensis from Tyva. C. Motacilla a. alba with white chin from Noyabr’sk (similar phenotypes are also collected from Buryatia and Mezen’). D and E. Hybrids of M. a. ocularis and M. a. lugens from Kamchatka. D has black back and intermediate size of white wing patch, E has gray back and large white wing patch.
Fig. 2.3. A. Maximum parsimony tree for 10 unique male phenotypes built from eight phenotypic characters of white wagtails sampled from 24 populations. B. Maximum 68
Likelihood tree for 87 unique mtDNA haplotypes. Numbers by branches are bootstrap support from 100 replicates, lab numbers precede locality names of unique haplotypes, localities where the same haplotype was found are listed, numbers in the brackets are number of individuals sharing the same haplotype. AP- Almaty-Primor’e clade, SKSouthern Krasnodar clade, NC- Northern clade. Abbreviations for localities are: AlmaAlmaty, Prim- Primor'e, Krda- Krasnodar, Murm- Murmansk, Mosc- Moscow, BuraBuryatia, Sakh- Sakhalin, Medv- Medvedevo, Meze- Mezen', Irku- Irkutsk, YamlYamal, Maga- Magadan, Astr- Astrakhan', Mong- Mongolia, Noya- Noyabr'sk, KamcKamchatka, Anad- Anadyr, Cher- Cherskiy, Kurs- Kursk, Sver- Sverdlovsk. Motacilla maderaspatensis was used as an outgroup on both trees (A and B). C. Minimum evolution population tree constructed from population pairwise sequence differences corrected for within-population differentiation.
69
Figure 2.3. A.
B.
personata
57 Alma leucopsis 7 Alma 21 Alma 52 Prim AP 96 182 Prim lugens personata x baicalensis (Tyva) 3 Alma 179 Prim 58 215 Prim (N=2) 84 102 Krda alba (western Eurasia including Krasnodar) 105 Krda 116 Krda (N=2) 123 Krda (N=2) 89 alba with white chin 124 Krda 67 Krda SK (Noyabr’sk, Buryatia, Mezen’) 104 Krda white ear, 114 Krda alba white cheek 68 Krda (N=3) baicalensis (Mongolia, Irkutsk, 103 Krda (N=6) Tyva, Buryatia) 127 Krda common haplotype (N=105) leucopsis (Primor’e, Khabarovsk) 6 Murm 213 Mosc white neck 202 Mosc sides 192 Bura 209 Sakh white chin ocularis (Magadan, Khabarovsk, 154 Medv Kamchatka, Anadyr) 169 Medv white throat 175 Medv 158 Meze ocularis x lugens (Kamchatka) 159 Meze black back 176 Meze 177 Meze 214 Prim lugens (Sakhalin, Kamchatka, eye stripe 186 Irku Primor’e) 188 Irku (N=2) 189 Irku large white 142 Tyva, Mosc, Yaml (N=3) wing patch 149 Tyva ocularis x lugens (Kamchatka) all 115 Krda, Medv, Meze (N=3) 84 Maga (N=2) subblack back, 86 Maga species white chin 93 Astr 54 Mong intermediate size 58 Mosc (N=2) of white wing patch 59 Mosc 63 Mosc C. 75 Yaml, Tyva (N=2) 76 Yaml NC 42 Noya 43 Noya 76 49 Noya Krasnodar 12 Maga 51 Maga 220 Kamc Almaty 221 Kamc 239 Anad 228 Anad Primor’e 230 Anad 212 Maga Noyabr’sk 234 Anad 22 Maga Irkutsk + Buryatia 56 227 Anad 233 Anad (N=2) Tyva 68 4 Kamc, Anad (N=2) 34 Kamc, Sakh (N=3) Yamal 238 Anad 45 Astr, Medv (N=2) Medvedevo 160 Meze 161 Meze, Medv (N=3) Mezen’ 164 Meze 165 Meze Magadan 38 Mosc second common hapl (N=15) Anadyr 72 Mosc 121 Krda Murmansk 26 Mong 31 Kamc 16 Cher Moscow 71 Mosc 46 Astr Sakhalin 33 Sakh 198 Sakh (N=2) Kamchatka 14 Kurs 20 Sver Mongolia 1 Sver 148 Tyva, Murm (N=2) Khabarovsk 97 Yaml
personata (Almaty, Gorno-Altay, Tyva)
1
Astrakhan’ Sverdlovsk
83
0.001 substitutions/site
70
Figure 2.4. Plot of nucleotide diversity versus latitude for all localities with sample sizes of four or more individuals. Latitude explains 30% of the variance in nucleotide diversity when all samples are analyzed.
Nucleotide Diversity
0.005
Primor'e Krasnodar
0.004
Almaty 0.003 0.002 0.001 0 40
45
50
55
60
65
70
Degrees North Latitude
71
Fig. 2.5. Mismatch distributions for localities with sample sizes of ten or more individuals for northern localities. On x-axis- number of pairwise differences between sequences, on y-axis- frequency, black dots on graphs indicate observed values, lines show model distribution as calculated by Arlequin (Schneider et al. 2000), vertical dashed line indicates the mean of the distribution.
72
Figure 2.5
73
Fig. 2.6. Mismatch distributions for all Krasnodar individuals, only southern Krasnodar (SK clade) haplotypes and all individuals from Primor'e. Black dots on graph indicate observed values, line- model distribution as calculated by Arlequin (Schneider et al. 2000), dashed line indicates the mean of the distribution.
74
Figure 2.6 0.25
Krasnodar frequency
0.2 0.15 0.1 0.05 0 0
3
6
0.25
12
15
18
Krasnodar southern haplotypes only
0.2
frequency
9
0.15 0.1 0.05 0 0
3
6
9
0.25
12
15
18
Primor'e
frequency
0.2 0.15 0.1 0.05 0 0
3
6
9
12
15
18
number of pairwise differences
75
Fig. 2.7. Haplotype and phenotype distributions of white wagtails at several localities. A. Krasnodar. B. Tyva, Almaty and Gorno-Altay. C. Primor’e. Sample sizes are given under locality names. Black circles indicate unique haplotypes, small white circles indicate haplotypes shared among different localities or subspecies, number inside white circles is number of individuals from the same locality sharing same haplotype; large ovals indicate subspecies, area of their overlap indicate hybrid phenotypes.
76
Figure 2.7
A.
C.
B.
Northern clade (NC)
Northern clade (NC)
alba/ dukhunensis
Primor’e N=9
Northern clade (NC)
lugens 3
leucopsis
6
2 3 2
2
Gorno-Altay N=2
baicalensis 3
Krasnodar N=24
9
personata Southern Krasnodar clade (SK) Almaty N=4 Southern Almaty-Primor’e clade (AP)
Tyva N=16
Southern Almaty-Primor’e clade (AP)
77
Chapter 3
Phylogeography and population genetics of common rosefinch (Carpodacus erythrinus) based on mitochondrial genes 3
Abstract
We analyzed sequences of two mitochondrial DNA (mtDNA) gene regions (control region and ND2) totaling 1536 base-pairs from 186 specimens of the common rosefinch (Carpodacus erythrinus) obtained from 17 Eurasian localities covering most of the species’ distribution. The studied populations possessed a high level of nucleotide diversity, which decreased southward, indicating relatively young age of southernmost populations. A low level of population subdivision (Fst = 0.12) and significant isolation by distance effect were detected; gene flow seemed to be restricted between Krasnodar and the other localities, and between northeastern and northwestern parts of the species range. The Krasnodar population and northeastern group of localities coincide with ranges of subspecies kubanensis and grebnitskii. Possible incipient speciation in Krasnodar was supported by the phylogenetic tree of haplotypes, on which 15 of 19 sampled individuals from Krasnodar formed a clade imbedded in an otherwise geographically unstructured tree. Maximum likelihood coalescent estimates of asymmetric migration rates showed little gene flow from and no gene flow to the Krasnodar population; eastward gene flow within northwestern and northeastern groups of populations was detected. The Irkutsk population appears to generate emigrants to southern localities. A molecular clock hypothesis was rejected. Fu's Fs and Tajima's Dvalues were significantly negative for many studied populations, which might indicate
The material in this chapter is part of the collaboration with R. M. Zink and S. Rohwer. It is written in the style and format of the journal Molecular Ecology. Submission is pending. 76 3
either population growth or selection. Mismatch distributions for most localities indicate past population expansions.
Introduction
The common rosefinch (Carpodacus erythrinus) is a common Eurasian passerine bird, whose range is subdivided into northern and southern parts (Fig. 3.1). The northern range includes mainly temperate continental climatic zones from Northern Europe to Kamchatka and Sakhalin, and the southern range stretches from the southeastern shore of the Black Sea and the Caucasus through the mountain regions of Central Asia (Himalaya and eastern Tibet) to the Taihang Shan range (Northeastern China) (Stepanyan 1990, Cramp and Perrins 1994) (Fig. 3.1). The ranges meet in Dzhungarskiy Alatau and Tarbagatay (mountain ranges of eastern Kazakhstan). Rosefinches inhabit moist scrub types in the western lowlands, drier fields with shrubs in more eastern plains, and more elevated foothill forests, mountain river valleys and alpine meadows in the southern part of the range (Dement'ev and Gladkov 1954). Over the last century, the common rosefinch has undergone westward and northward expansion reaching Sweden in 1930, White Sea coast in 1966, Yugoslavia in 1978, Britain in 1982, and France in 1985 (Cramp and Perrins 1994). The common rosefinch is sexually dimorphic with first year males and females having brownish plumage and adult males having mostly red plumage. Geographic variation of the common rosefinch is reflected in the relative intensity of the red and pink colors, and body size (Dement'ev and Gladkov 1954, Stepanyan 1990, Cramp and Perrins 1994). From four (Dement'ev and Gladkov 1954) to five (Cramp and Perrins 1994, Stepanyan 1990) subspecies are recognized. Subspecies descriptions are quantitative, which causes inconsistencies in descriptions of subspecies' ranges by different authors. For example, following Dement'ev and Gladkov (1954), C. e. grebnitskii is distributed over a large northeastern part of the species’ range (Table 3.1). Cramp and Perrins's (1994) state that grebnitskii is limited to the coastal area of Ohotsk Sea and Kamchatka. Stepanyan (1990) described grebnitskii as restricted to Kamchatka. 77
The evolutionary relationships between birds from different parts of the range are not yet studied. We use mitochondrial DNA sequences to explore the evolutionary history of samples of birds from much of the species’ range.
Materials and Methods
Sampling.-- A total of 186 common rosefinches from 17 geographic localities sampled across Eurasia was collected during the breeding season (Fig. 3.1). Nearly all tissues were vouchered with a study skin with spread wing. Vouchers were deposited at the Burke Museum, University of Washington, Seattle; the Moscow State University Zoological Museum, Moscow, Russia; State Darwin Museum, Moscow, Russia; or the Bell Museum, University of Minnesota, St. Paul. Tissue samples were either preserved in 96% ethanol, frozen in liquid nitrogen, or stored in lysis buffer (Longmire et al. 1997). Great rosefinch (Carpodacus rubicilla), Pallas's rosefinch (C. roseus), purple finch (C. purpureus) and house finch (C. mexicanus) were used as outgroups. One individual per species was used. Tissue samples of great and Pallas’s rosefinches were provided by the Burke Museum, samples of purple and house finches were from the Bell museum. Molecular lab methods.--Isolation and purification of the DNA was performed using phenol-chloroform protocol (Hillis et al. 1996) or QIAamp Tissue Kit (QIAGEN, Valencia, California). Polymerase Chain Reaction (PCR) (Saiki et al. 1988) with PerkinElmer PCR reagents were used to amplify two mitochondrial gene regions. Primers L5215 (Hackett 1996) and H1064 (Drovetski et al. 2004) were used to amplify the complete NADH dehydrogenase subunit 2 (ND2) gene. The cycling parameters were as follows: 2.5 min at 95°C, followed by 40 cycles of denaturation at 94°C for 30 sec, annealing at 55°C for 30 sec and extension at 72°C for 1 min, and ended with an extension of 10 min at 72°C. Primers LCR4 and H1248 (Tarr 1995) amplified part of the Control Region (CR) gene. The PCRs for CR were performed with 34 cycles of 1 min at 94°C, 1 min at 50°C, 1 min at 72°C with final extension for 10 min at 72°C. Qiaquick PCR Purification Kit (QIAGEN) was used to clean PCR products. Cleaned PCR 78
fragments were sequenced on ABI 3700 automated sequencer using BigDye chemistry (Applied Biosystems). Amplification primers and primers L347 (CCATTCCACTTCTGATTCCC, designed by S. V. Drovetski) and H5578 (Hackett 1996) were used for sequencing of 1041 base pairs (bp) of ND2, primers LCR4 (Tarr 1995), LCON2 (Zink et al. 1998), 2LCRU (5'-GATGCACTTTGNCCNCATTC-3', designed by R. Blackwell-Rago) and 2HCRU (5'-GAATGNGGNCAAAGTGCATC-3', designed by R.B.-R.) were used to obtain 497 bp of control region. Data analysis.--Sequences were aligned and edited in Sequencher 3.1.1 (Gene Codes Corporation, Ann Harbor, Michigan) and deposited in GenBank (Accession numbers (pending)). Mitochondrial origin of sequenced DNA fragments was supported by the absence of stop-codons in the coding gene region and the existence of a large number of haplotypes, which are inconsistent with nuclear copies (Zhang and Hewitt 1996). A maximum likelihood (ML) model was determined using the Hierarchical Likelihood Ratio Test (hLRT) in Modeltest 3.06 (Posada and Crandall 1998). PAUP* (Swofford 2000) was used to perform a ML tree search. SEQBOOT version 3.5c (Felsenstein 1993) was used to generate 100 bootstrapped sets of data from which ML trees were constructed in PhyML (Guindon 2003). Then CONSENSE version 3.5c (Felsenstein 1993) was used to analyze the ML bootstrap trees and compute a majority rule consensus tree. A likelihood ratio (LR) test was performed to evaluate whether the sequences have been evolving in a clocklike manner. Scores from ML trees with and without a molecular clock enforced (Felsenstein 1981) were compared and the LR was calculated as 2(ln Lclock-ln Lno clock) under the assumption that the LR was χ2 distributed with degrees of freedom (df) equal to the number of taxa minus two (Nei and Kumar 2000). Arlequin (Schneider et al. 2000) (with 0% allowed missing data per site) was used to compute the number of haplotypes in population, number of polymorphic sites, haplotype diversity (h), nucleotide diversity (π), time since population expansion τ, relative population size before (θ0) and after (θ1) expansion and mismatch distributions. 79
Mismatch distributions were computed for localities with sample sizes greater than 12 individuals; some samples were pooled (see below). To test the empirical mismatch distribution against a model of sudden expansion we used the generalized non-linear least-squares approach (Schneider and Excoffier 1999) as implemented in Arlequin. We also used Arlequin to compute pairwise population Fst for populations with sample sizes of four or more individuals, and to perform Tajima's D (Tajima1996) and Fu’s Fs (1997) tests of selective neutrality, analysis of molecular variance (AMOVA), and Mantel’s (1967) test of pairwise Fst values vs. geographic distances. Geographic distances were calculated with online surface distance calculator (www.wcrl.ars.usda.gov/cec/java/lat-long.htm). Ramos-Onsins and Rozas (2002) suggested a test (R2) for detecting population growth. We used DnaSP version 4.00 (Rozas et al. 2003) to calculate R2.We regressed values of nucleotide diversity (π) against latitude and longitude expecting smaller values to be observed on the leading edge of population expansion (Hewitt, 2000). MEGA version 2.1 (Kumar et al. 2001) was used to construct a Minimum Evolution tree from population average pairwise sequence differences corrected for within population variability. We used Migrate 1.7.3 (Beerli and Felsenstein 1999, 2001) to obtain maximum likelihood non-equilibrium estimates of effective population sizes Θ = 2Nefµ, and rates of migration 2Nefm between every pair of populations, where Nef is effective population size of females, µ is mutation rate per site per generation and Nefm is the number of immigrant females per generation. This method uses a Markov chain Monte Carlo coalescent approach to explore possible genealogies. It relaxes the assumptions of the populations having the same size and symmetric migration rates, which are essential for Fst -derived estimates of Nm, but assumes no drastic increase in population size over time. We ran Migrate on nine populations with sample sizes of 12 and more individuals combining some samples from neighboring localities (see Table 3.3). Two sites with insertions were removed. To preserve degrees of freedom and to decrease the number of parameters being estimated, we used a stepping-stone model (Fig. 3.4), which was mostly 80
consistent with the pairwise Fst matrix: connections between two localities corresponded to insignificant pairwise Fst values for all but Krasnodar-Moscow and KrasnodarSverdlovsk comparisons, both of which yielded significant Fst = 0.3. The following genealogy searching settings were used: 10 short Markov chains each consisting of 50,000 genealogies with sampling every 100 trees, followed by a long Markov chain with 10,000,000 genealogies with sampling every 100 trees. At the beginning of each chain 10,000 trees were discarded so that the next chain is not biased towards the parameters estimated for the previous chain. An adaptive heating scheme with four parallel chains and initial relative temperatures of 1.0, 1.2, 1.5, 3.0 was used. We assumed a sequencing error rate of 0.1%, a transition to transversion ratio of 7, and we estimated base frequencies from the data. Migrate was run five times with different random seed numbers and the Fst -based starting parameters. Estimates of Θ and 2Nfm from five runs were averaged, and then used as initial parameters for the final run, profile likelihoods for all parameters were evaluated at 0.025 and 0.975 percentiles.
Results Molecular phylogenetic analysis.—Three individuals (125 Medv, 136 Medv and 178 Krda) had a single base pair insertion in two sites of the Control region sequence. Of 1536 sequenced nucleotides for ND2 and CR 1524 were analyzed after excluding insertions and sites with missing data. For 186 individuals there were 154 polymorphic sites (142 transitions, 21 transversions, 74 parsimony informative sites) resulting in 148 haplotypes. There was no common haplotype shared by most populations, but five haplotypes were shared by more than three individuals. One haplotype (30 Sver) was shared by total of nine individuals from Medvedevo (lab numbers 128, 129, 131 and 134), Sverdlovsk (30), Almaty (57) and Irkutsk (141, 151, 155). Another haplotype (62 Mong) was shared by seven individuals from Medvedevo (125), Tyva (104, 110), Irkutsk (153), Mongolia (62) and Anadyr’ (183, 195). Haplotype 46 Mong was shared by five individuals from Irkutsk (145, 160, 177), Mongolia (46) and Kamchatka (56). Haplotype 81
80 Maga was found in four individuals from Irkutsk (157), Magadan (80) and Anadyr’ (184, 193). Haplotype 85 Krda was shared by four individuals from Krasnodar (85, 87, 93, 96) (Fig. 3.2). The sequenced part of the control region (3’-end) differed greatly among C. erythrinus and the four outgroups and could not be aligned unambiguously. Carpodacus roseus showed the smallest pairwise distances when compared to C. erythrinus sequences with an uncorrected divergence of 13.5% for ND2 and 9.8% for CR. Maximum divergence between any pair of C. erythrinus sequences was 1.5% for ND2 and 1.6% for CR. We omitted outgroups and used midpoint rooting because the outgroups were too distant to identify correctly the root of C. erythrinus tree. The molecular clock hypothesis was rejected (-lnL without molecular clock enforced = 3929.33, -lnL with molecular clock = 4016.87; df = 146, P = 0.05). Hierarchical likelihood criteria implemented in Modeltest 3.06 suggested TrN+I+G (Tamura and Nei 1993) as the most probable model of molecular evolution. This model assumes 6 substitution types (1, 38.6, 1, 1, 10.9), 4 rate categories, proportion of invariable sites I = 0.747 and gamma distributed rates at variable sites with shape parameter α = 1.0229. Maximum likelihood (ML) analysis with parameters estimated from the data found a tree (Fig. 3.2) with little geographic structure. One clade (clade A on Fig. 3.2) included 15 out of the 19 individuals sampled from Krasnodar, although this clade had low bootstrap support. Another clade (clade B on Fig. 3.2), comprised 18 individuals, only two of which occurred to the west of the Urals. Genetic variability.-- Overall nucleotide diversity (π) was high (0.0044 for all samples combined) and ranged from 0.0024 in Almaty to 0.0056 in Sakhalin (Table 3.1). Regression of π on sample size showed independence of these variables (P = 0.87, data not shown). Contrary to expectation of northward postglacial expansion, nucleotide diversity was higher in the north, than to the south of species' range. Latitude explained 30% of the variability in nucleotide diversity (R2 = 0.30, P = 0.05) (Fig. 3.5). Nucleotide diversity was independent from longitude (P = 0.88). Genetic diversity and population subdivision.-- Overall haplotype diversity was 82
0.99 and ranged from 0.96 in Krasnodar and Medvedevo to 1 in Moscow, Mezen', Sverdlovsk, Almaty, Gorno-Altay, Sakhalin and Kamchatka (Table 3.1). AMOVA showed a relatively low level of population differentiation, with just 12.4% (P < 0.05) of molecular variance due to comparisons among geographical localities. Gene flow.-- Significant pairwise Fst-values (Table 3.2) between geographic localities ranged from 0.03 (Irkutsk- Medvedevo) to 0.42 (Krasnodar- Kamchatka). All Fst-values involving Krasnodar were large and significant (ranging from 0.3 to 0.42). Most of pairwise Fst comparisons involving Kamchatka and Anadyr’ populations were also significant, although Fst-values between Anadyr’ and Irkutsk, Magadan and Kamchatka and between Kamchatka and Magadan were not significant. For Magadan population pairwise comparisons showed significant Fst-values with Medvedevo, Sverdlovsk, Gorno-Altay, and Tyva. This result implies a history of relative isolation of the eastern populations from other parts of species' range. Comparisons between IrkutskMedvedevo, Tyva- Moscow, Tyva- Medvedevo, and Moscow- Gorno-Altay also yielded significant, but low Fst-values (Table 3.2). The remaining pairwise Fst values were not significant, suggesting generally high levels of gene flow. Mantel’s test showed that pairwise population Fst -values were correlated with geographic distances between populations (R2 = 0.2, P < 0.05). Whereas Fst statistics is based on the assumptions of symmetrical gene flow and equal population size and is unable to distinguish between contemporary and historic gene flow, the maximum likelihood approach implemented in Migrate (Beerli and Felsenstein 2001) offers a different perspective, relaxing these assumptions. The maximum likelihood estimates of population sizes and migration rates were generally stable over multiple runs with the exception of two Migrate runs that resulted in infinite estimates of Θ; those two runs were discarded. Maximum likelihood estimates of Θ ranged from 0.0001 in Magadan to 0.02 in Mongolia (Fig. 3.4). When migration rates (Nefm) were overlaid over geographic distribution (Fig. 3.4), several asymmetric migration patterns emerged. The Krasnodar population seemed to send a few emigrants to the Moscow (3.2) and Sverdlovsk (0.5) populations, but does not receive any immigrants 83
from the northern populations. Virtually no gene flow was detected between the neighboring Moscow and Medvedevo-Mezen’ populations. There seem to be more migrants moving eastward than in opposite direction. For example, the Sverdlovsk population receives many immigrants from western populations (32 from Moscow and 19.5 from Medvedevo) and almost none from the east, but it generates more than twice as many emigrants to Irkutsk than to Medvedevo and none to Moscow. However, no gene flow was detected between the Sverdlovsk and the Tyva - Gorno-Altay populations. Asymmetric southward migration from Irkutsk is observed: the estimated numbers of migrants to Tyva - Gorno-Altay and to Mongolia are large (112 and 55 correspondingly), but small in opposite direction (9 and 0). Highly asymmetric estimates are obtained for Anadyr' and Magadan (44.5 from Magadan into Anadyr' and only 0.9 in opposite direction). Finally, two northeastern localities (Magadan and Anadyr') appear to be isolated from the remaining localities: the Irkutsk and the Magadan populations only exchange about one migrant per generation, whereas no gene flow is detected between Magadan and Mongolia. A minimum evolution tree built from among populations pairwise distances corrected for within-population variability (Fig. 3.3) displayed some geographic structure: the Krasnodar population was a sister group to the clade containing all remaining populations and was on a long branch; also eastern populations (Magadan, Kamchatka, Sakhalin, Anadyr’ , Irkutsk) were grouped together. Population expansion.-- When all samples were pooled together, pairwise differences between individuals were distributed in accordance with the model of sudden population expansion (Table 3.1, Fig. 4) and had unimodal distributions. Most of the mismatch distributions for different localities were multimodal, although none of the distributions differed significantly from a model of sudden expansion using the bootstrap test of Schneider and Excoffier (1999). Estimates of tau (τ) ranged from 3.2 in GornoAltay to 8.5 in Moscow (Table 3.1), but were not structured geographically (no estimates were available for Krasnodar). The R2 test detected population growth for the Krasnodar, Moscow, Medvedevo, Sverdlovsk, Gorno-Altay, Tyva, Irkutsk, Mongolia and Magadan 84
populations, but found no deviation from stability for the Mezen’, Almaty, Kamchatka and Anadyr’ samples (Table 3.1). Tests of selective neutrality.-- Fu's Fs values (Table 3.1) were significantly negative for all populations except Mezen', Sakhalin and Kamchatka (P < 0.05), which suggests either deviations from neutral evolution in stable populations or population growth. The values of Tajima's D (Table 3.1) were significantly negative (P < 0.05) for Krasnodar, Medvedevo, Sverdlovsk, Irkutsk and Mongolia. Rejection of neutrality could also come from a population that has recently passed through a bottleneck and has not yet reached equilibrium.
Discussion Gene flow and isolation.-- Although none of the haplotype clades were reciprocally monophyletic with respect to geographic localities or regions, almost 80% of the birds from Krasnodar were closely related to each other and formed a clade, although some birds from Krasnodar possessed haplotypes that did not belong to this clade. Because it takes two times female effective population size (2Nef)generations for mtDNA to reach reciprocal monophyly, this result might indicate, that the Krasnodar population is currently isolated from the main part of the range, but isolation occurred less than 2Nef generations ago. Alternatively, Krasnodar could have been isolated in the past and reached reciprocal monophyly, but is currently receiving immigrants from northern part of the range. We believe that first scenario is more likely, because the Krasnodar clade is imbedded in the haplotype tree structure rather than being a sister clade to other haplotypes, which indicates that the Krasnodar clade is relatively young. The population tree which indeed showed Krasnodar as a sister group to the other populations probably just reflects the fact of the Krasnodar differentiation. Moreover, all pairwise comparisons involving Krasnodar yielded large and significant Fst –values, and non-equilibrium estimates of migration into and out of Krasnodar are close to zero. The southern part of the range includes several mountain ranges; it is connected to the northern part of the 85
species' distribution via a narrow stripe in the Middle Eastern mountain region (Fig. 3.1). The region just north of Krasnodar is uninhabited by common rosefinches, and the geographically closest northern neighbors of the Caucasus mountain population inhabit different habitats (lowlands). Therefore, it appears that the Krasnodar population was founded less than 2Nef generations ago and is currently isolated from the remaining sampled populations. The birds from the Krasnodar population belong to subspecies kubanensis, they show some morphological differentiation from northern birds by being slightly larger and having rosy-red shade of red color (Cramp and Perrins 1994). Therefore, it is possible, that C. e. kubanensis displays the signatures of incipient speciation. Although there were no geographically structured clusters of haplotypes, except for Krasnodar, evidence of some structure emerged from Fst and coalescence analyses. The Anadyr’ , the Magadan and the Kamchatka populations yielded significant pairwise Fst -values when compared with the most of remaining populations, indicating that gene flow is somewhat restricted between northeastern and northwestern parts of the species' range. Maximum likelihood estimates of migration also show very limited or no gene flow between Magadan and more western Irkutsk and Mongolia. Interestingly, this division corresponds to C. e. erythrinus- C. e. grebnitskii subspecies division, as recognized by Cramp and Perrins (1994). Restricted gene flow between two large northern parts of the range may have promoted plumage differentiation. Migration rates for Almaty could not be estimated with confidence due to small sample size. The lack of samples from the southeastern part of the range prevented us from testing whether northern and southern populations are connected by ongoing gene flow. The unidirectional, southward gene flow from Irkutsk to Tyva - Gorno-Altay and to Mongolia and the existence of Krasnodar haplotypes that do not belong to the Krasnodar clade (clade A on Fig. 3.2) suggest possible gene flow from northern to southern populations. Due to the lack of samples from western Europe, we could not test for westward dispersal. However, we detected asymmetric eastward gene flow in northern populations 86
(Fig. 4, Table 3.3). Recently Abdo et al. (2004) demonstrated that while Migrate estimates of Theta are reasonable, the program often underestimates, and sometimes overestimates, the migration rates. They caution against using the confidence intervals around these estimates, because they often do not capture the true value of parameters (Abdo et al. 2004). We ran Migrate six times: in all cases the estimates of Theta were of the same magnitude, and in most cases the estimates of Nefm varied by less than one order of magnitude (Table 3.3). Absence of migration between Altay-Tyva and Sverdlovsk and from Moscow to Medvedevo was consistent over multiple runs, as well as low estimates of Nefm from Moscow to Krasnodar, from Sverdlovsk to Krasnodar, from Medvedevo to Moscow, from Sverdlovsk to Moscow, and from Mongolia to Irkutsk. However, the estimates of Nefm from Krasnodar to Sverdlovsk ranged from 0.3 to 8.4, Nefm from Altay-Tyva to Mongolia ranged from 0 to 7.4, Nefm from Irkutsk to Altay ranged from 5.65 to 12.8 for the five preliminary runs, but was 111.85 for the final run. We do not interpret Nm values as the actual number of females migrating to another locality each generation, but use Nm values as an indicator of migration asymmetry. Our estimates of migration rates appear reasonable. Overall, our results indicate that Carpodacus erythrinus is subdivided into at least three large populations, between which there is a restricted gene flow. These large populations coincide with subspecific groupings of Cramp and Perrins (1994). Population expansion.-- Although many of the individual mismatch distributions were multimodal, a model of sudden population expansion was not rejected for most populations. This result suggests that these populations have undergone major expansion in a past. The mode was between four and six base-pair differences for all mismatch distributions (excluding Krasnodar), which indicates that population expansion was relatively old. Although the R2 test showed that the Mezen’, Almaty, Kamchatka and Anadyr’ populations did not undergo demographic expansions, Fu’s Fs values suggested population growth for Almaty and Anadyr’ (Table 3.1). Ramos-Onsins and Rosas (2002) showed that for large sample sizes, Fs is more powerful than R2 in detecting population 87
growth. Whereas this explains the difference in the results of two tests for Anadyr’, it is still unclear what causes this difference for the Almaty population. Evolutionary history.—Contrary to findings for many Eurasian bird species (Kvist et al. 2001, 2003, Zink et al. 2002, 2003, Holder et al. 1999), a high level of genetic diversity was found in common rosefinch populations. High genetic diversity indicates that rosefinches have had large historical effective population sizes, and polymorphism was not reduced by climatic fluctuations during the Pleistocene Ice Ages. During the glacial maxima (about 18,000 years ago) the large part of Europe was covered by the ice sheet, but the Asia was ice-free, covered by the steppe-like vegetation, or desert shrubs and semideserts in the arid regions (Frenzel et al. 1992). The reconstruction of vegetation based on the pollen records indicates that low bush periglacial tundra was spread over northern part of eastern Eurasia and Alaska (Frenzel et al. 1992). Although the absence of boreal vegetation in northern Eurasia must have displaced the ranges of many avian species, the common rosefinch might have been able to survive maximum cooling without drastic range contraction. Whereas in many species more northerly populations produce reduced allele diversity as the result of northward range expansion (Hewitt, 1996), the opposite pattern was observed for the common rosefinch. Comparisons of the means of mismatch distributions (and, therefore, nucleotide diversity) can give some insights into relative age of populations. Relatively low genetic diversity was found in Almaty, Medvedevo, Tyva, Gorno-Altay and Krasnodar, which might indicate that these populations were founded relatively recently. The likelihood estimates of effective female population size were also small for Medvedevo-Mezen’ and Altay-Tyva. One possible historical scenario, supported by the pattern of nucleotide diversity, might be that Carpodacus erythrinus inhabited northern planes of Eurasia, until some populations became adapted to mountain habitat of Central Asia (the Tyva population of Sayan mountains, Gorno-Altay population of Altay). Then birds dispersed southward through the Alay mountain system (the Almaty population) to Pamir, and northwestern Hymalaya, from where the dispersal became westward (towards Black sea, subspecies kubanensis) and eastward (through Himalaya to the mountains east from Tibet, subspecies roseatus). 88
Better estimates of the rates of molecular evolution of common rosefinch and more sampling will be needed to document and date these inferred historical events with a confidence.
References
Abdo Z, Crandall KA, Joyce P (2004) Evaluating the performance of likelihood methods for detecting population structure and migration. Molecular Ecology, 13, 837-851. Beerli P, Felsenstein J (1999) Maximum-likelihood estimation of migration rates and effective population numbers in two populations using a coalescent approach. Genetics, 152, 763-773. Beerli P, Felsenstein J (2001) Maximum-likelihood estimation of a migration matrix and effective population sizes in n subpopulations by using a coalescent approach. Proceedings of the National Academy of Science, U.S.A., 98, 4563-4568. Cramp S and Perrins CM (1994) Handbook of the birds of Europe, the Middle East and North Africa. Vol. 5, Oxford University Press, Oxford. Dement’ev GP and Gladkov NA [eds] (1954) Birds of the Soviet Union. Vol. 5, Sovetskaya Nauka, Moscow. Drovetski SV, Zink RM, Rohwer S et al. (2004) Complex Biogeographic History of a Holarctic Passerine. Proceedings of the Royal Society of London B, published online. Felsenstein J (1981) Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution, 17, 368-376. Felsenstein J (1993) PHYLIP, Version 3.5. http://evolution.genetics.washington.edu/phylip.html, Seattle, WA. Frenzel B, Pecsi M and Velichko AA [eds.] (1992) Atlas of paleoclimates and paleoenvironments of the northern Hemisphere. Hungarian Academy of Sciences, Budapest. Fu Y-X (1997) Statistical tests of neutrality against population growth, hitchhiking and 89
background selection. Genetics, 147, 915-925. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology, 52, 696-704. Hackett SJ (1996) Molecular phylogenetics and biogeography of tanagers in the genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution, 5, 368-382. 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 (2000) The genetic legacy of the Quaternary ice ages. Nature, 405, 907-913. Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of The Royal Society: Biological Sciences, 359, 183195. Hillis DM, Mable BK, Larson A et al. (1996) Nucleic acids IV: sequencing and cloning. In: Molecular systematics (eds. Hillis DM, Moritz C, Mable BK), 2nd edn, pp. 321-381. Sunderland, MA: Sinauer. Holder K, Montgomerie R and Friesen VL (1999) A test of the glacial refugium hypothesis using patterns of mitochondrial and nuclear DNA sequence variation in Rock Ptarmigan (Lagopus mutus). Evolution, 53, 1936-1950. Kumar S, Tamura K, Jakobsen IB and Nei M (2001). MEGA2: molecular evolutionary genetics analysis software. Bioinformatics, 17, 1244-1245. Kvist L, Martens J, Higuchi H et al. (2003) Evolution and genetic structure of the great tit (Parus major) complex. Proceedings of the Royal Society of London B, 270, 1447-1454. Kvist L, Martens J, Ahola A, Orell M (2001) Phylogeography of a Palaearctic sedentary passerine, the willow tit (Parus montanus). Journal of Evolutionary Biology, 14, 930-941. Longmire L, Maltbie M, and Baker RJ (1997) Use of “lysis buffer” in DNA isolation and its implication for museum collections. Occasional Papers OP-163. Museum of Texas Tech University. Mantel NA (1967) The detection of disease clustering and a generalized regression 90
approach. Cancer Research, 27, 209-220. Nei M and Kumar S (2000) Molecular Evolution and Phylogenetics. Oxford Univ. Press, Oxford. Posada D and Crandall KA (1998) Modeltest: Testing the model of DNA substitution. Bioinformatics, 14, 817-818. Ramos-Onsins SE and Rozas J (2002) Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution, 19, 2092-2100. Rozas J, Sanchez-DelBarrio JC, Messeguer X, and Rozas R (2003) DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics, 19, 2496-2497. Saiki RK, Gelfand DH, Stoffel S et al. (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 239, 487-491. Schneider S, Roessli D and Excoffier L (2000) Arlequin, Version 2.000. A software for population genetic data analysis. http://lgb.unige.ch/arlequin/, Genetics and Biometry laboratory, University of Geneva, Switzerland. Schneider S and Excoffier L (1999) Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. Genetics, 152, 1070-1089. Stepanyan LS (1990) Conspectus of the ornithological fauna of the USSR. Nauka, Moscow, Russia (in Russian). Swofford DL (2000) PAUP*. Phylogenetic analysis using parsimony (*and other methods), Version 4.0b2. Sinauer, Sunderland, MA. Tajima F (1996) The amount of DNA polymorphism maintained in finite population when the neutral mutation rate varies among sites. Genetics, 143, 1457-1465. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution, 10, 512-526. Tarr CL (1995) Primers for amplification and determination of mitochondrial controlregion sequences in oscine passerines. Molecular Ecology, 4, 527-529. 91
Zhang D-X and Hewitt GM (1996) Nuclear integrations: challenges for mitochondrial DNA markers. Trends in Ecology and Evolution, 11, 247-251. Zink RM, Weller SJ and Blackwell RC (1998) Molecular phylogenetics of the avian genus Pipilo and a biogeographic argument for taxonomic uncertainty. Molecular Phylogenetics and Evolution, 10, 191-201. Zink RM, Drovetski S and Rohwer S (2002) Phylogeographic patterns in the Great Spotted Woodpecker (Dendrocopos major) across Eurasia. Journal of Avian Biology, 35, 175-178. Zink RM, Drovetski SV, Questiau S et al. (2003) Recent evolutionary history of the bluethroat (Luscinia svecica) across Eurasia. Molecular Ecology, 12, 3069-3075.
Acknowledgements
We thank the Burke museum for curatorial assistance, S. Farrell and M. Westberg for laboratory assistance. We are grateful to S. Drovetski, D. Banin, I. Karagodin, A. Jones, B. Barber, C. S. Wood, B. Schmidt, R. C. Faucett for logistical help with expeditions and collecting. S. Birks (Burke Museum) subsampled some tissues. We are grateful to G. Eddy for funding fieldwork. Additional support came from NSF (DEB 9707496) and the Dayton-Wilkie fund (Bell Museum).
92
Table 3.1. Genetic characteristics of 17 geographic populations of Carpodacus erythrinus. Letters after subspecies names indicate inconsistencies in subspecies ranges described by different authors: D- Dement'ev and Gladkov (1954), C- Cramp and Perrins (1994), S- Stepanyan (1990). All given values of Fu's Fs, Tajima's D and Ramos-Onsins and Rozas R2 are significant at 0.05 level, no est-
diversity x104
Krasnodar
44/37
kubanensis
19
31
16
0.97
34.4
Moscow
55/37
erythrinus
11
29
11
1
51.5
8.5
0
180
-5.04
n/s
0.1
60/38
erythrinus
19
25
16
0.96
29.5
3.4
1.4
99
-9.95
-1.43
0.07
65/44
erythrinus
4
15
4
1
53.2
-
-
-
n/s
n/s
n/s
56/58
erythrinus
12
32
12
1
45.9
7.3
0
4682
-6.44
-1.52
0.06
7
9
7
1
23.6
4
0
4682
-3.86
n/s
n/s
Medvedevo Mezen’ Sverdlovsk Almaty
42/75
ferghanensis C, S kubanensis D
divercity
number of
nucleotide -
haplotype -
haplotypes 1
sites
-
number of
Sample size 1
°East long
erythrinus
°North lat/
51/34
locality
Kursk
Collecting
Subspecies
polymorphic
Arlequin was not able to fit the model of sudden expansion, n/s- not significant.
Fu’s Fs
Tajim
τ
θ0
θ1
-
-
-
-
-
-
no est
no est
-8.79
-1.62
0.06
no est
a's D
R2
93
GornoAltay
diversity x104
nucleotide
divercity
haplotype
haplotypes
number of
sites
polymorphic
number of
Sample size
Subspecies
°East long
°North lat/
locality
Collecting
Table 3.1. Continued.
τ
θ0
θ1
Fu’s Fs
Tajim a's D
R2
erythrinus S, D 51/85
grading to
7
15
7
1
32.9
3.2
2
1957
-2.99
n/s
0.12
15
23
13
0.98
31.7
4.8
0
7125
-6.78
n/s
0.09
ferghanensis C erythrinus S,
Tyva
50/93
grading to grebnitskii D to ferghanensis C
Krasnoyarsk Irkutsk
57/97
erythrinus
1
-
1
-
-
-
-
-
-
-
-
54/104
erythrinus
33
56
29
0.99
45.3
5.0
2.7
90
-22
-1.83
0.04
20
39
19
0.99
42.7
6.9
0
223
-13.1
-1.62
0.06
1
-
1
-
-
-
-
-
-
-
-
erythrinus grading to Mongolia
47/113
grebnitskii C erythrinus S grebnitskii D
Khabarovsk
51/136
erythrinus S grebnitskii D, C
94
46/141
Magadan
59/150
Kamchatka Anadyr’ All pooled
52/157
64/170
diversity x104
polymorphic
1
56.5
-
-
-
-
-
-
13
12
13
0.99
43.4
4.6
3
76
-5/2
n/s
0.08
5
13
5
1
36.5
-
-
-
n/s
n/s
n/s
15
25
13
0.98
39.6
6.5
0
133
-5.53
n/s
n/s
186
154
148
0.99
43.8
6.27
0.68
106
-24.7
-2.365
0.02
erythrinus S grebnitskii D, C grebnitskii erythrinus S grebnitskii D, C
divercity
number of
nucleotide
3
haplotype
13
haplotypes
3
sites
grebnitskii
number of
Sample size
Sakhalin
Subspecies
°East long
°North lat/
locality
Collecting
Table 3.1. Continued.
τ
θ0
θ1
Fu’s Fs
Tajim a's D
R2
95
Table 3.2. Population pairwise Fst’s for comparisons between populations with sample sizes of four or more individuals. Only significant values are presented.
Krda
Mosc
Medv
Meze
Sver
Alma
Gorn
Tyva
0.125
0.097
Irku
Mong
0.083
0.104
Krasnodar Moscow
0.305
Medvedevo
0.312
Mezen'
0.328
Sverdlovsk
0.295
Almaty
0.324
Gorno-Altay
0.354
0.104
Tyva
0.399
0.088
Irkutsk
0.304
Mongolia
0.309
Magadan
0.346
Kamchatka
0.420
0.109
0.250
0.173
0.163
0.249
0.269
0.198
Anadyr’
0.345
0.059
0.110
0.134
0.081
0.088
0.179
0.127
0.101 0.033
0.106
0.069
0.043
96
Table 3.3. Migration rate and population size estimates from Markov Chain Monte Carlo simulations as implemented in Migrate (Beerli and Felsenstein 2001). Θ = 2Nefµ, where Nef is the female effective population size and µ is the mutation rate per site per generation, 2Nefm is twice the number of migrating females per generation, MoscKurspooled samples from Moscow and Kursk, MedvMez - pooled Medvedevo and Mezen', AltayTyva- pooled Tyva and Gorno-Altay. Runs 1-5 started from Fst-based initial parameter values and differed by random seed number. Average estimates from these five runs were used as initial parameters for the final run. not est.- the parameter assumed to have value of zero. Stepping-stone model and Nefm values from the final run are shown on Fig. 3.4. Parameters
Run1
Run 2
Run3
Run4
Run 5
(sample sizes)
Average
Final
of 5 runs
run
ΘKrasnodar
(19)
0.00267
0.00288
0.00306
0.00369
0.00375
0.00321
0.0046
ΘMoscKurs
(12)
0.00806
0.01164
0.01037
0.00823
0.00816
0.009292
0.0067
ΘMedvMeze
(23)
0.00099
0.00072
0.00065
0.0006
0.00084
0.00076
0.0006
ΘSverdlovsk
(12)
0.00972
0.00948
0.00838
0.01344
0.01063
0.01033
0.0114
ΘAltayTyva
(22)
0.00065
0.00074
0.00047
0.00056
0.00084
0.000652
0.0042
ΘIrkutsk
(33)
0.01051
0.01521
0.00908
0.00839
0.00749
0.010136
0.008
ΘMongolia
(20)
0.01972
0.0147
0.009
0.01134
0.02567
0.016086
0.02
ΘMagadan
(13)
0.0001
0.0001
0.00006
0.00006
0.00007
0.000078
0.0001
ΘAnadyr'
(15)
0.0098
0.01132
0.01589
0.01015
0.00264
0.00996
0.0072
2Nefm Mosc to Krda
0.76
0
0
0.51
0
0.25
0
2Nefm Sverd to Krda
0.12
0
0.15
1.52
0
0.36
0
2Nefm Krda to Mosc
7.8
10.5
6.44
10.32
4.64
7.94
6.32
2Nefm Medv to Mosc
0
1.1
0
1.23
0.66
0.6
0.6
2Nefm Sverd to Mosc
2.67
0
0
0.41
1
0.82
0.3
2Nefm Mosc to Medv
0
0
0
0
0
0
not est.
2Nefm Sver to Medv
10.99
14.77
17.1
11.73
13.28
13.57
24.72
2Nefm Krda to Sver
9.85
1.29
5.66
16.84
0.59
6.85
0.96
2Nefm Mosc to Sver
22.45
32.87
38
47.58
25.61
33.3
65.13
2Nefm Medv to Sver
25.62
29.22
8.10
71.72
29.4
32.82
39.32
2Nefm AlTy to Sver
0
0
0
0
0
0
not est.
97
Table 3.3. Continued. Parameters
Run1
Run 2
Run3
Run4
Run 5
(sample sizes)
Average
Final
of 5 runs
run
2Nefm Irku to Sver
0
2.4
2.42
4.43
0
1.85
1.92
2Nefm Sverd to AlTy
0
0
0
0
0
0
not est.
2Nefm Irku to AlTy
12.36
20.38
11.31
20.02
25.63
17.94
223.72
2Nefm Mong to AlTy
7.50
1.13
5.82
1.73
5.4
4.315
32.47
2Nefm Sver to Irku
76.39
108.65
52.63
31.13
46.02
62.96
56.14
2Nefm AlTy to Irku
14.3
39.16
41.2
14.79
6.8
23.25
17.82
2Nefm Mong to Irku
1.17
0
0
0
0
0.23
0
2Nefm Maga to Irku
1.17
1.62
0
0
0.85
0.73
2.67
2Nefm AlTy to Mong
14.81
0
0
0
0
2.96
0
2Nm Irku to Mong
74.03
69.77
84.64
50.46
234.7
102.72
109.65
2Nm Maga to Mong
0
0
5.07
0
0
1.01
0
2Nm Irku to Maga
2.32
1.79
1.31
1.26
1.25
1.58
2
2Nm Mong to Maga
0
0
0.11
0
0.13
0.05
0
2Nm Anad to Maga
2.59
2.13
2.19
1.71
2.27
2.18
1.88
2Nm Maga to Anad
61.06
108.23
222.09
82.25
41.82
103.09
89.01
98
Figure 3.1. Map of Eurasia showing general location of collecting sites. Shaded area shows distribution of Carpodacus erythrinus (following Stepanyan 1990). Large circles indicate sample sizes of 10 or more individuals, dashed line indicates approximate subspecies boundaries (following Cramp and Perrins 1994).
99
Fig. 3.2. Maximum Likelihood tree for 148 unique mtDNA haplotypes of common rosefinch sampled from 17 populations rooted at midpoint. Thick vertical lines indicate shared haplotypes (see text for localities where the same haplotype was found). Lab numbers precede locality abbreviations of unique haplotypes. numbers in the brackets are number of individuals sharing the same haplotype. Black circles indicate clades with more than 50% ML bootstrap support. Abbreviations for localities are: Alma- Almaty, Prim- Primor'e, Krda- Krasnodar, Murm- Murmansk, Mosc- Moscow, Bura- Buryatia, Sakh- Sakhalin, Medv- Medvedevo, Meze- Mezen', Irku- Irkutsk, Yaml- Yamal, MagaMagadan, Astr- Astrakhan', Mong- Mongolia, Noya- Noyabr'sk, Kamc- Kamchatka, Anad- Anadyr’ , Cher- Cherskiy, Kurs- Kursk, Sver- Sverdlovsk.
100
Figure 3.2 78Maga 2Kurs
170Tyva
7Alma
49Mong
9Kamc
194Anad 165Irku
187Anad 14Alma 83Maga 74Gora 44Mosc 146Irku 171Tyva 99Irku 77Krda 156Irku
98Maga 136Medv 140Irku
6Kamc 26Maga 55Kamc 86Krda 186Anad 191Anad
63Mong 159Irku
1Kamc 5Maga
182Anad
88Krda
Persentage of 100 ML bootstrap replicates
24Khab 185Anad
50-70%
37Sver
71-80% 81-90% 3Gora
4Sakh 21Alma 72Mong 47Alma 73Gora 79Maga 135Meze 169Tyva
60Gora
12Irku 16Irku 70Mong 106Tyva 107Tyva 108Tyva 112Tyva 124Medv
64Mong
132Medv 150Irku
34Sver 53Mong 35Sver
118Medv 164Irku 109Tyva
152Irku
82
B 80Maga (N=4)
48Mong
172Tyva 36Sver 130Medv 181Anad
46Mong (N=5) 18Sakh 23Sakh 81Maga 82Maga 162Irku 190Anad 27Mosc
192Anad 29Mosc 148Irku
66Mong
8Gora 22Mosc 121Medv
117Medv 30Sver (N=9)
31Sver 52Gora 58Alma 61Mong
123Medv 163Irku 32Sver 45Mong 43Mong 68Mong 115Medv 149Irku 20Sver
15Maga 120Meze 54Alma
28Mosc 69Mong
71Mong
114Medv
167Tyva 11Mosc
0.0005 substitutions/site
119Meze
137Irku 180Krda 122Medv 85Krda (N=4) 75Krda
66
A
41Mosc
139Irku 126Medv
62Mong (N=7)
142Irku
173Irku 175Maga
13Sver 138Irku
51Maga 19Sver
25Mosc
143Irku
116Meze
45
168Tyva
38Sver
105Tyva
17Krya
67Mong
10Gora 33Sver
92Krda 91Krda
97Krda
158Irku
147Irku
90Krda 95Krda 178Krda
76Krda
89Krda
94Krda 179Krda
101
Figure 3.3. Minimum evolution population tree constructed from population pairwise sequence differences corrected for within-population differentiation. Only populations with sample sizes of three or more individuals are included.
Magadan Kamchatka Sakhalin Anadyr Irkutsk Moscow Mongolia Mezen Tyva Sverdlovsk Almaty GornoAltay Medvedevo Krasnodar 0.5
102
Fig. 3.4. Mismatch distributions for localities with sample sizes of 12 or more individuals and coalescent likelihood estimates of population size and gene flow for nine populations of common rosefinch (some populations are pooled: Moscow + Kursk, Medvedevo + Mezen', Tyva +Gorno-Altay). On x-axis- number of pairwise differences between sequences, on y-axis- frequency, black dots on graph indicate observed values, linemodel distribution as calculated by Arlequin (Schneider et al. 2000), dashed line indicates the mean of the distribution. None of the mismatch distributions were significantly different from sudden expansion model. Numbers under locality names are likelihood estimates of population size (Θ = 2Nefμ, where is the female effective population size and is the mutation rate per site per generation). Arrows indicate estimates of directional gene flow between localities in units of Nefm (number of migrating females per generation) as calculated by Migrate (Beerli and Felsenstein 2001).
103
Figure 3.4 Anadyr’ 0.0072 Medvedevo + Mezen’ 0.0006 Sverdlovsk 0.0114
Irkutsk 0.0080
Moscow + Kursk
1
0.0067 0.3
1.3
12.4 0
0
0.2
0.0001 1
32.6 3.2 0 0
28.1
0
111.9 0
54.8 8.9
Pooled samples
Tyva + Gorno-Altay 0.0042
0
0
16.2
0.0046
Magadan
19.7
0.5
Krasnodar
44.5 0.9
0 Mongolia 0.0200
10
Fig. 3.5. Plot of nucleotide diversity against latitude for localities with sample sizes of four or more individuals. Latitude explains 30% of the variance in nucleotide diversity.
0.0055
Mezen'
Moscow
Nucleotide diversity
0.005 Irkutsk 0.0045
Sverdlovsk
Mongolia Magadan Anadyr
0.004 Krasnodar
0.0035
Kamchatka Gorno-Altay
Tyva
0.003 0.0025
Medvedevo R2 = 0.30 P = 0.05
Almaty
0.002 40
45
50
55
60
65
70
North latitude
105
Chapter 4
Mitochondrial phylogeographies of the great tit Parus major and the willow tit Parus montanus and the role of sampling in inferring population histories 4. Abstract
Several studies of common Eurasian passerine birds, the great tit (Parus major) and the willow tit (P. montanus), explored population genetics and phylogeography of these species using mainly the mitochondrial control region (CR). However, all of these studies suffered from large gaps in sampling distributions, specifically most of Eastern Europe and Siberia remained unsampled. We added 15 new Eurasian localities greatly extending the coverage of the range for each species. We used part of CR and complete ND2 sequences for 84 great tits and139 willow tits to analyze phylogeographic and population structures of these species and compare our results with those from previous studies. Phylogenetic analysis of haplotypes confirmed the absence of phylogeographic structure within the haplotype trees for both great and willow tits. Genetic diversity was low in the great tit (π = 0.001, h = 0.75) and moderate in the willow tit (π = 0.0023, h = 0.98). Difference in population genetic structure between these species could be attributed to differences in long-term effective population sizes and the relative ages of populations. Our results were consistent with prior hypothesis that the great tit survived the last glacial period in low numbers in an isolated refuge and then recolonized the Eurasian continent. In contrast, the willow tit survived in high numbers during the Pleistocene. Assuming molecular substitution rates are similar in the two tits, then the willow tit colonized its Eurasian range earlier than the great tit. 4
The material in this chapter is part of the collaboration with R. M. Zink and S. Rohwer. It is written in the style and format of the journal Molecular Ecology. Submission is pending.
106
Introduction
Phylogeographic studies add invaluable knowledge about genetic structure of species, help us understand the process of speciation and the general patterns of species’ genetic diversity and influence taxonomic decisions. Although a number of phylogeographic studies appeared during last years, few address species with wide Eurasian or Holarctic distributions (Liebers et al. 2001, Irwin et al. 2001, Zink et al. 2002a, 2002b, 2003, Pavlova et al. 2003, Drovetski et al. 2004, Godoy et al. 2004). Such studies are rare and sampling is often restricted to part of species’ ranges, because collecting is expensive and time-demanding. Inferences about species as a whole based on the limited samples can be incomplete or even erroneous (Zink et al. 2003). In this paper we analyze phylogeographic data representing most of the Eurasian range of two species of Eurasian passerine birds, the great tit (Parus major) and the willow tit (Parus montanus). The great tit is a sedentary bird in most of the southern and central parts of its range. Great tits prefer open mixed and deciduous forests, avoiding dense woodland, and are often found along rivers and around cultivation and settlements (Harrap and Quinn1995). The range of the great tit stretches from Europe and northwestern Africa through northern Asia east to the Pacific, reaching northern Iran, Mongolia and northern China on the south (Harrap and Quinn1995; Fig. 4.1). The species has expanded northward in Europe during past century (Cramp and Perrins 1993). The geographic variation of the great tit (Parus major) is reflected in the amount of yellowish pigment (lipochromes) in the plumage, extent of white in the tail, body size and the shape of the bill (Harrap and Quinn 1995). Genetics of the great tit in the European part of the species’ range was studied by Kvist et al. (1999). In this study, the complete mitochondrial control region (CR) was sequenced for 68 birds from Finland, Sweden, Estonia, Germany, Netherlands and Spain (Fig. 4.1, Table 4.1). Recent demographic expansion was inferred for these populations, and no phylogeographic structure and extensive gene flow were detected. Kvist et al. (1999) hypothesized that the Balkans was a refugium during the last glacial maximum 107
(20,000 years ago) and that great tits subsequently expanded rapidly to northern Europe. More recently, Kvist et al. (2003) examined partial CR sequences of 125 P. major (major group), expanding the initial sampling to Portugal and UK in Europe, three Asian localities (Urals, Amur and Kirghizia), and one African sample (Morocco) (Fig. 4.1, Table 4.1). The P. major (major group) population in Kirghizia was founded in 1960-s, when hundreds individuals from Western Siberia were released into the wild (Formozov et al. 1993). Kvist et al. (2003) found low genetic diversity among individuals and no phylogeographic structuring of haplotypes. At the same time, Amova results showed some populations substructure (Fst = 0.12); moreover, pairwise population Fst -values were large and significant for geographically isolated populations from Kirghizia (range of 0.29-0.54) and UK (range of 0.14-0.4). Kvist et al. (2003) concluded that genetic variation within populations was reduced drastically as the result of population bottleneck during Ice Age related range contractions. Harrap and Quinn (1995) distinguish three major groups among the great tit species complex: major (inhabits Europe, NW Africa and northern Asia and includes 11 races), cinereus (NE Iran, S. Afghanistan, India and Indonesia, includes 13 races) and minor (inhabits SE Russia, Japan and northern SE Asia, includes 9 races). A fourth group, Turkestan tit Parus bokharensis, occurring in lowlands of Central Asia from Turkmenia to Mongolia (Cramp and Perrins 1993), is often included in major group, but is considered a species by Harrap and Quinn (1995). All four groups are considered species by some authors (Stepanyan 1990), a view supported by the molecular study of Kvist et al. (2003). Although some hybridization does occur where group ranges meet, these four groups have had independent evolutionary histories and, therefore, should be considered phylogenetic species (Kvist et al. 2003). In this paper we also refer to Parus major, P. minor and P. bokharensis as species. The willow tit (P. montanus) is a sedentary bird often found in deciduous, coniferous and mixed forests, preferring an alpine environment in the southern part of the species’ range (Cramp and Perrins 1993). The range of willow tit covers a broad zone from northeastern France through central and northern Europe and Russia to the Pacific 108
coast and over the Bering strait into Alaska (Cramp and Perrins 1993, Fig. 4.1). Clinal geographic variation exists among most of the continental subspecies, which differ in song types and coloration (Harrap and Quinn 1995). Three studies have explored genetic variation within the willow tit. Kvist et al. (1998) compared complete mitochondrial control region sequences from Finland and Swedish populations, and concluded that individuals from both localities belong to one panmictic population with high long-term effective population size. Kvist et al. (2001) used partial CR sequences (CR-I), of birds from six European localities (Finland, Sweden, Latvia, Poland, Germany and Alps), six samples from eastern Asia (Japan, Ussuri, Amur, Sakhalin, Magadan and Kamchatka) and one sample from Urals (Fig. 4.1, Table 4.2). Therefore they with the exception of Urals, they lacked samples from eastern Europe and western Asia. Kvist et al. (2001) found no correlation between phylogenetic structure and geographic localities or subspecies of the willow tit, although some interesting patterns of restricted gene flow emerged. Amova showed that populations are subdivided (Fst = 0.15); all pairwise population Fst -values were large and significant for the Japan population (range of 0.23-0.45), as well as most of pairwise Fst -values for the Latvia population (range of 0-0.45). Kvist et al. (2001) concluded that lack of phylogeographic structure was attributed to homogeneous habitat throughout the distribution, lack of geographical barriers, high gene flow and a large long-term effective population size. They suggested that willow tits expanded through Palearctic from southeastern Asia after the last Ice Ages. In the third phylogeographic study of Parus montanus Salzburger et al. (2002) used mitochondrial cytochrome b variation to resolve phylogeographic relationships among several subspecies of the willow tit and the songar tit. Their sampling included Finland, several localities from central Europe and southeastern Asia including Japan, and a single sample from southern Siberia, therefore also leaving a large gap in the northern part of species’ distribution. Twenty-five willow tits from six Eurasian subspecies (baicalensis, borealis, montanus, restructus, rhenanus and sachalinensis) were found to be closely related; all six subspecies shared a haplotype that was distributed 109
over the whole Eurasian continent (including Japan). Salzburger et al. (2002) suggested that the willow tit spread over major parts of Eurasia during Pleistocene and underwent rapid phenotypic divergence in terms of morphology and acoustic traits. Harrap and Quinn (1995) divide The willow tit P. montanus into three subspecies groups: salicarius (northern Eurasia, includes 6 races), kamtschatkensis (Pacific rim in eastern Asia, 4 races) and montanus (mountains of central Europe, with a single race), considering the songar tit P. songarus of Central Asia a separate species. Kvist et al. (2001) and Salzburger et al. (2002) found no structure correlated with subspecies groups within the willow tit on phylogenetic trees constructed from CR and cytochrome b data (respectively); but the songar tit, considered by them subspecies of P. montanus, was found to be polyphyletic. Salzburger et al. (2002) found three distinct lineages in the songar tit: P. (m.) weigoldicus, P. (m.) affinis and P. (m.) songarus; the latter two formed well supported clades on the tree of Kvist et al. (2001). In this paper we consider the willow tit and the three lineages of the songar tit separate species. Therefore, no phylogeographic structure was found for either Great or willow tits. Nonetheless, before any strong statements about the absence of genetic structure in the species are made, more rigorous sampling covering the large unstudied portion of the species’ range is needed. In this study we sequenced part of the CR and complete ND2 gene to analyze geographic distribution of haplotypes and study population structure of the great tit and the willow tit. We included multiple previously unstudied localities, greatly extending the coverage of the range for each species. Thus, more definitive conclusions about phylogeographic structure can be reached. Our goals were to test previous phylogeographic conclusions with more extensive sampling and to compare population genetic parameters derived from different genes in different laboratories.
Materials and methods Sampling For the ingroup, we analyzed 84 individuals of P. major from 15 Eurasian localities and 110
139 individuals of P. montanus from 17 localities (Figure 1, Tables 1, 2). We used 29 individuals of P. minor from Primor’e and four individuals of P. bokharensis from Almaty as outgroups for P. major individuals. We used P. palustris as an outgroup for P. montanus. Birds were collected during breeding season. From all specimens a study skin was preserved, and deposited at the Burke Museum, University of Washington, Seattle; the Moscow State University Zoological Museum, Moscow, Russia; or the Bell Museum, University of Minnesota, St. Paul. Tissue samples were preserved in 96% ethanol or frozen in liquid nitrogen and stored in lysis buffer (Longmire et al. 1997). We combined our partial CR sequences of 84 P. major with the CR sequences of 68 individuals studied by Kvist et al. (1999) (GenBank accession number AF059662 and Appendix 1 in Kvist et al. 1999) to create a second set of sequences that covers larger geographic range. We were unable to merge our CR data with the most recent population studies of Kvist et al. (2001 and 2003), because only 80 base pairs overlap between these and our data sets. Molecular lab methods Isolation and purification of DNA was performed using QIAamp Tissue Kit (QIAGEN, Valencia, California). Polymerase Chain Reaction (PCR) (Saiki et al. 1999) with PerkinElmer PCR reagents were used to amplify two mitochondrial DNA (mtDNA) gene regions. Primers L5215 (Hackett 1996) and H1064 (Drovetski et al. 2004) were used to amplify the complete ND2, primers LCR4 and H1248 (Tarr 1995) amplified part of the Control Region (CR) gene (the 3’ end of CR, or CR-II). PCR conditions are explained in Pavlova et al. (in review). Qiaquick PCR Purification Kit (QIAGEN) was used to clean PCR products. Then cleaned PCR fragments were directly sequenced on ABI 3700 automated sequencer using BigDye chemistry (Applied Biosystems). Amplification primers and primers L347 (Drovetski 2004) and H5578 (Hackett 1996) were used for sequencing of complete ND2 (1041 base pairs (bp)), primers LCR4 (Tarr 1995), LCON2 (Zink et al. 1998), 2LCRU and 2HCRU (Pavlova et al. (in review)) were used to obtain CR-II (626 bp for P. major and 721 bp for P. montanus). Data analysis 111
Sequences were aligned and edited in Sequencher 3.1.1 (Gene Codes Corporation, Ann Harbor, Michigan). Sequence data have been deposited in GenBank (Accession numbers pending). Mitochondrial origin of sequenced DNA fragments was supported by the absence of stop-codons in the coding gene region and the existence of a large number of haplotypes, which are inconsistent with nuclear copies (Zhang and Hewitt 1996). Maximum likelihood (ML) model and parameters were determined by the Hierarchical Likelihood Ratio Test (hLRT) in Modeltest 3.06 (Posada and Crandall 1998). PAUP* (Swofford 2000) and PhyML (Guindon 2003, Auckland University, New Zealand) were used to perform ML tree search. SEQBOOT version 3.5c was used to generate 100 bootstrapped sets of data from which ML trees were constructed in PhyML. Then CONSENSE version 3.5c was used to analyze the ML bootstrap trees and compute majority rule consensus tree. SEQBOOT and CONSENSE are part of PHYLIP 3.5 package (Felsenstein 1993, Washington University, Siettle, WA). To evaluate whether the sequences have been evolving in a clocklike manner a likelihood ratio (LR) test was performed in PAUP*. Scores from ML trees with and without a molecular clock enforced (Felsenstein 1981) were compared and the LR was calculated as 2(ln Lclock-ln Lno clock) under the assumption that the LR was χ2 distributed with degrees of freedom (df) equal to the number of taxa minus two (Nei and Kumar 2000). DnaSP version 4.00 (Rozas et al. 2003) was used to calculate number of haplotypes in population (ambiguous characters and gaps were removed), nucleotide diversity (π), haplotype diversity (h), and to perform R2 test for detecting population growth (Ramos-Onsins and Rozas 2002). We used Arlequin (Schneider et al. 2000) to perform Tajima's D (Tajima1996) and Fu’s Fs (1997) tests of selective neutrality, analysis of molecular variance (AMOVA), and to compute pairwise population mismatch distributions, time since population expansion (τ) and effective population sizes before (θ 0)
and after (θ1) expansion for localities with sample sizes greater than 10 individuals. To
test the empirical mismatch distribution against a model of sudden expansion we used the generalized non-linear least-squares approach (Schneider and Excoffier 1999) as 112
implemented in Arlequin. The Fs test statistic is based on the haplotype frequency distribution conditional the value of Theta, Tajima’s D test is based on the differences between the number of segregating sites and the average number of nucleotide differences. We compared our estimates of nucleotide diversity with Kvist’s et al. (2001, 2003) and regressed values of nucleotide diversity (π) against latitude and longitude for samples of 3 or more individuals expecting smaller values to be observed on the leading edge of population expansion (Hewitt, 2000). Results I. The great tit Parus major.
Molecular phylogenetic analysis For 84 individuals we obtained a total of 1667 nucleotides (1041 base pair (bp) of ND2 and 626 bp of CR-II). There were four insertion/deletion sites in CR. After removing sites with gaps and missing characters 1653 aligned nucleotides remained. Among the 84 individuals there were 34 haplotypes with 47 polymorphic sites, 12 of which were parsimony informative (Table 4.1). Hierarchical likelihood criteria implemented in Modeltest 3.06 suggested TrN+I+G (Tamura and Nei 1993) as the most probable model of sequence evolution. The shape parameter alpha was set at 1.1454, and the proportion of invariable sites was 0.7381. The ML score of the tree found by PhyML was 3237.78. The scores of the ML trees found PAUP* were 3229.26 without a molecular clock and 3290.49 with a clock enforced. The molecular clock hypothesis was rejected (2∆ln L = 122.46, df = 58, P = 1.64x10-6). All 24 haplotypes of P. minor clustered together (this clade was recovered by 100% of ML bootstrapped trees), as well as the two P. bokharensis haplotypes from Almaty. However, the major clade received only 69% bootstrap support and was not recovered by the ML tree found by PAUP*, because the bokharensis clade was imbedded 113
in it making P. major paraphyletic. Phylogeography of the great tit The overall haplotype diversity in P. major was low (0.748) and ranged from zero in Magadan, where all individuals shared most common haplotype, to 1 in Norway, Smolensk and Crimea. The 34 unique haplotypes of P. major were not structured geographically on the ML tree (Fig. 4.2). The most common haplotype (4Norw on Fig. 4.2) was shared by 42 individuals (50% of all sampled individuals) from Norway (lab number 4), Kursk (lab numbers 11, 25), Smolensk (55), Crimea (19), Krasnodar (87, 88), Moscow (36, 37, 40, 45, 46, 96, 98, 101, 103, 104, 106, 108-110), Medvedevo (112, 113), Astrakhan’ (28, 32, 34, 84), Buryatia (7, 8, 122), Mongolia (47, 48), Irkutsk (116, 118, 119) and Magadan (12, 13, 79-83). Three more haplotypes were shared among two or three populations. The haplotype 3Norw was found in three individuals from Norway and Astrakhan’. The haplotype 18Sver was found in five birds from Sverdlovsk, Moscow and Kursk. The haplotype 10Kurs was shared by three individuals from Kursk and Irkutsk (Fig. 4.2).
Combining our CR sequences with those of Kvist et al. (1999) from five European populations resulted in a data set of 152 P. major sequences for 611 bp of CRII. We used our 29 P. minor and four P. bokharensis individuals as outgroups. In the combined data set there were only 23 unique haplotypes because 114 individuals (75% of all sampled P. major) shared the most common haplotype, which was the same for our and Kvist et al’s (1999) sequences. Moreover, all four P. bokharensis individuals shared the same haplotype with four P. major individuals from Krasnodar (1), Germany (1) and Moscow (2). All P. minor sequences formed a clade on the otherwise unstructured tree of haplotypes (not shown), showing deeper divergence from P. major than P. bokharensis. Genetic variability Overall nucleotide diversity (π) was low in P. major (0.001) and ranged from zero in Magadan to 0.002 in Crimea. Regressions of π on latitude showed an insignificant trend towards northern populations having lower π for our and Kvist et al. (2003) data (Fig. 114
4.4A, B). Our data, but not Kvist et al.’s (2003) showed a trend towards eastern birds having lower π (R2 = 0.47, P = 0.06 ) (Fig. 4.4B). Gene flow and population subdivision The total Fst from our ND2 plus CR data was low and not significant (Fst = 0.02, P > 0.05), indicating high gene flow among populations. Few pairwise population Fst values were significant (Table 4.3), with the largest value of 0.28 for Magadan- Buryatia comparison. Population expansion Time since expansion τ estimated from mismatch distribution was 1.2 for the Moscow population and 1.1 for the pooled samples. All values of Tajima’s D, Fu’s Fs were negative and significant for Moscow population and pooled samples indicating either deviations from neutral evolution in stable populations or population growth. The R2 test also detected deviations from the model of stable population for the Moscow population and pooled samples (Table 4.4).
II. The willow tit P. montanus.
Molecular phylogenetic analysis For 139 individuals we obtained a total of 1761 nucleotides (1041 base pair (bp) of ND2 and 720 bp of CR-II). There were six sites with insertions/deletions in CR sequences of the willow tit. A total of 1752 aligned nucleotides were analyzed after removing sites with ambiguous characters and gaps. Of these, 113 sites were polymorphic, and 44 were parsimony informative (Table 4.2). We obtained 99 haplotypes for the 139 individuals. Hierarchical likelihood criteria implemented in Modeltest 3.06 suggested F81+G (Felsenstein 1981) as a most probable model of sequence evolution. The shape parameter alpha was set at 1.63. The ML score of the tree found by PhyML was 4361.44 (Fig. 4.3). The scores of the ML tree found by PAUP* were 4524.54 without a clock and 4601.00 with a clock enforced. The molecular clock hypothesis was rejected (2∆ln L = 152.9076, df = 100, P = 0.0005). The haplotype tree was unstructured. 115
Phylogeography of the willow tit The overall haplotype diversity was high in the willow tit (0.98) and ranged from 0.7 in Kamchatka to 1 in Medvedevo, Mezen’, Noyabr’sk and Tyva+Gorno-Altay (Table 4.2). The most common haplotype (3Khab on Fig. 4.3) was shared by 16 individuals from 8 populations: Moscow (lab number 18), Mezen’ (41), Noyabr’sk (80), Irkutsk (14,107, 116, 138), Buryatia (68, 113, 134), Primor’e (125, 126), Khabarovsk (3), Magadan (50, 79, 88, 92) (Fig. 4.3). The haplotype 29Kamc had a restricted Far Eastern distribution, and was shared by seven individuals from Kamchatka and Anadyr. Four haplotypes were shared among eastern populations. The haplotype 1Bura was found in five birds from Mezen’, Buryatia, Noyabr’sk, Mongolia and Primor’e. The haplotype 2Bura was found in four birds from Buryatia and Primor’e. The haplotype 66Gora was shared by two individuals from Gorno-Altay and Irkutsk. The haplotype 55Mong was shared between three individuals from Mongolia and Khabarovsk. Two more haplotypes were shared among populations. The haplotype 36Medv was found in three birds, one each from Medvedevo, Tyva and Irkutsk. The haplotype 11Mosc was found in one bird from Moscow and one from Mezen’ (Fig. 4.3). Although there were no well supported clades on haplotype tree of P. montanus (Fig. 4.3), existence of some structuring of haplotypes is evident by the presence of the haplotypes (29Kamch, 1Bura, 2Bura, 66Gora, 55Mong) with geographical distribution restricted to eastern Eurasia (Fig. 4.3). Genetic variability Overall nucleotide diversity was moderate in P. montanus (0.0023) and ranged from 0.006 in Anadyr to 0.0032 in Medvedevo. Regressions of π on latitude were not significant for our or Kvist et al. (2001) data (Fig. 4.4C). Longitude explained 59% of the variance of π in our data with eastern birds having lower π (R2 = 0.59, P < 0.05 ), but no such trend was found for the data of Kvist et al. (2001) (Fig. 4.4D). Gene flow and population subdivision The Amova analysis showed that geographical localities explain only 5.3% of the total variation (P < 0.05). However, most pairwise Fst -values involving Anadyr (range from 116
0.2 to 0.43) and Kamchatka (range from 0.1 to 0.22) were large and significant (Table 4.3), indicating restricted gene flow with these populations. Several pairwise Fst -values involving Medvedevo were also significant (Table 4.3). Population expansion Tau ranged from 1.3 in Primor’e to 4.4 in Magadan. With the exception of Magadan, where change of effective population size after expansion was small, the values of τ decreased eastward, indicating that the western expansions are older. Populations with large sample sizes (Moscow, Irkutsk, Buryatia, Anadyr, Primor’e and Magadan) as well as pooled samples displayed significantly negative values of Fu’s Fs. For all of these populations except for Buryatia and Anadyr values of Tajima’s D were also significantly negative. These results might indicate deviations from neutral evolution or population growth (further discussed by Zink (unpublished)). Population expansion is also detected by the R2 test for all of these populations (Table 4.4).
Discussion Compatibility of data sets For the five European populations of the great tit, estimates of nucleotide diversity from the 5’ end of mitochondrial control region (CR-I) (Kvist et al. 2003) were consistently higher (1.64 times on average) than those calculated from the complete control region for the same populations (Kvist et al. 1999) (Table 4.1). This result indicates that CR-I is much more variable than CR-II. However, for the two Scandinavian populations of the willow tit studied by Kvist et al. (1998, 2001) the results were not consistent: πFinland was higher when estimated from CR-I, whereas πSweden was higher when estimated from complete CR sequences. Such inconsistency in the distribution of variation in CR sequences makes it hard to combine the estimates of π obtained from different portions of the gene. Strangely, average estimates of haplotype diversity for the same five European populations of P. major were also 1.13 times higher when estimated from shorter CR-I than from complete CR (Table 4.1). This result is unexpected because the shorter 117
sequences should yield fewer haplotypes. Although our sequences of both Parus species were three times longer than those obtained by Kvist et al. (2001, 2003), values of nucleotide diversity estimated from our data (CR-II plus ND2) were smaller than those calculated from the CR-I data of Kvist et al.(2001, 2003). Average π obtained for P. major from CR-I were 2.5 times higher than average π from CR-II + ND2. In case of P. montanus we were able to compare our estimates of π and h with Kvist et al. (2001) for four eastern populations: Magadan, Kamchatka, Khabarovsk (Amur) and Primor’e (Ussuri), where our and Kvist et al. (2001) samples overlapped (Fig. 4.1, Table 4.2). Estimates of π for these populations were on average 2.9 times higher when estimated from CR-I than those from CR-II + ND2, and 3.2 times higher than those obtained from CR-II alone (Table 4.2). Therefore, the estimate of genetic diversity obtained from different genes can not be compared directly. The great tit In general our results were consistent with those of Kvist et al. (1999, 2003): the haplotype tree (Fig. 4.2) was not structured geographically and the most common haplotype was shared by more than a third of the individuals and distributed over the whole Eurasian continent (Table 4.1). Kvist et al. (2003) suggested that the great tit populations underwent a drastic reduction of the population size as a result of the range contraction during the last glacial maxima. Survival in low numbers in a small southwestern refugium was thought to have been followed by northeastward range expansion and subsequent demographic expansion. Our results (low genetic variation, the absence of phylogeographic structure, trend towards eastern birds having smaller nucleotide diversity, unimodal mismatch distributions and significant R2) support this historical scenario. As evidenced by an insignificant Fst, the great tit is not subdivided, and therefore gene exchange is not restricted among populations (with the exception of the geographically isolated UK and Kirghizia populations, and two eastern populations, Magadan and Irkutsk). Deep sequence divergence between P. major and P. minor (2.8% average clade divergence) and higher genetic diversity in P. minor (Table 4.1) suggest that these 118
species have undergone different histories. It is possible that while P. major was displaced into small refugia (possibly in southern Europe, as suggested by Kvist et al. (2003)), P. minor might have survived Pleistocene climatic oscillations without drastic range and size contractions in southeastern Eurasia, where the climate was more stable. The small sequence divergence between P. major and P. bokharensis (0.8% average clade divergence) and uncertainty of their sister-group relationship (clade of P. bokharensis was nested within P major clade on 30% of the bootstrap trees and on the ML tree found by PAUP) suggest that P. bokharensis might have just recently evolved in Central Asia. The reconstruction of vegetation (Frenzel et al. 1992) suggests that during the Pleistocene cooling this region contained extreme dry desert, which was not suitable for birds. Therefore, Central Asia was recently colonized, possibly from the northern populations. The willow tit Although our phylogenetic tree (Fig. 4.3) was unstructured, the restricted eastern distribution of some haplotypes suggests limited gene flow between eastern and western parts of the species’ range. While Kvist et al. (2001) suggested that the lack of phylogeographic structure within the willow tit is attributed to high gene flow among populations, they also suggest that the presence of gene flow might not be high now, as evidenced by morphological and acoustic differences. Significant pairwise Fst-values that are found for two Far Eastern populations (Anadyr and Kamchatka, subspecies anadyrensis and kamtchatkensis), the population from Japan (subspecies restrictus), Medvedevo and Latvia (both belong to subspecies borealis), suggest that some populations of the willow tit are subdivided and that at least in some cases, genetic differentiation coincides with phenotypic divergence. The absence of monophyly on the phylogenetic tree suggests that population divergence is recent because it takes about 2Nef generations before mtDNA gene trees reach monophyly. Relatively high genetic diversity within the willow tit and large number of haplotypes displayed by our and Kvist et al. (2001) data sets suggest large long-term effective population size and absence of bottlenecks in recent species history. 119
Kvist et al. (2001) reported a westward decrease in nucleotide diversity for salicarius group (that includes most of the species’ distribution, but not Kamchatka, Sakhalin and Japan). The authors concluded that during the last Ice Ages the willow tit occupied large regions of the southeastern Palearctic and colonized the European part of the range from the east. However, our regression analysis of their data did not find this trend to be statistically significant. Moreover, our data show a significant trend in which the eastern birds have lower nucleotide diversity, which might be a trace of past eastward range expansion. Reconstructions of biomes during the last glacial maximum based on pollen and plant macrofossils (Tarasov et al. 1999) suggest that much of northwestern Eurasia was covered by tundra. Although the willow tit typically breeds in forests, it also inhabits woody shrubs and in Norway and Russia extends north to the treeline (Harrap and Quinn 1995). Thus, Pleistocene tundra in northern Eurasia could potentially provide suitable habitats for the willow tit. Substantial genetic diversity of the Willow tit suggests that the species was able to accumulate mutations without drastic range contractions. During climate warming the tits could have expanded eastward.
Conclusions
Although the ranges of the great tit and the Willow tit largely overlap, these species underwent different evolutionary histories. During last glacial maximum the range of the great tit was substantially displaced into small (presumably south-European) refugia, causing a reduction in population size. This was followed by postglacial range and demographic expansion, in which a single haplotype spread around the continent (as predicted by Hewitt 1999). The willow tit, in contrast, did not undergo as severe a population bottleneck, but instead survived in relatively large numbers in the widespread north-Eurasian tundra. The Willow tit also expanded its range eastward after the Ice Ages. Both species displayed signatures of ongoing evolutionary processes. References 120
Cramp S and Perrins CM (1993) Handbook of the birds of Europe, the Middle East and North Africa. Vol. 7, Oxford University Press, Oxford. Drovetski SV, Zink RM, Rohwer S et al. (2004) Complex Biogeographic History of a Holarctic Passerine. Proceedings of the Royal Society of London B, published online. Felsenstein J (1981) Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution, 17, 368-376. Felsenstein J (1993) PHYLIP, Version 3.5. http://evolution.genetics.washington.edu/phylip.html, Seattle, WA. Formozov NA, Kerimov AB and Lopatin VV (1993) New hybridisation zone of the great titmouse and Parus bokharensis in Kazakhstan and relationships in forms of Parus major superspecies. In: Hybridization and the problem of species in vertebrates (ed. Rossomolino OL), pp.118-146. Moscow: Archives of the Zoological Museum Moscow State University. Frenzel B, Pecsi M and Velichko AA [eds.] (1992) Atlas of paleoclimates and paleoenvironments of the northern Hemisphere. Hungarian Academy of Sciences, Budapest. Fu Y-X (1997) Statistical tests of neutrality against population growth, hitchhiking and background selection. Genetics, 147, 915-925. Godoy JA, Negro JJ, Hiraldo F and Donáza JA (2004) Phylogeography, genetic structure and diversity in the endangered bearded vulture (Gypaetus barbatus, L.) as revealed by mitochondrial DNA. Molecular Ecology, 13, 371-390. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology, 52, 696-704. Hackett SJ (1996) Molecular phylogenetics and biogeography of tanagers in the genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution, 5, 368-382. Harrap S, Quinn D (1995) Chickadees, Tits, Nuthatches, and Treecreepers. Princeton University Press, Princeton. Hewitt GM (1999) Post-glacial re-colonization of European biota. Biological Journal of 121
Linnean Society, 68, 87-112. Hewitt GM (2000) The genetic legacy of the Quaternary ice ages. Nature, 405, 907-913. Irwin DE, Bensch S and Price TD (2001) Speciation in a ring. Nature, 409, 333-337. Kvist L, Ruokonen M, Thessing A, Lumme J, Orell M (1998) Mitochondrial control region polymorphism reveal high amount of gene flow in Fennoscandian willow tit (Parus montanus borealis). Hereditas, 128, 133-143. Kvist L, Ruokonen M, Lumme J and Orell M (1999). The colonization history and present-day population structure of the European great tit (Parus major major). Heredity, 82, 495-502. Kvist L, Martens J, Ahola A, Orell M (2001) Phylogeography of a Palaearctic sedentary passerine, the willow tit (Parus montanus). Journal of Evolutionary Biology, 14, 930941. Kvist L, Martens J, Higuchi H et al. (2003) Evolution and genetic structure of the great tit (Parus major) complex. Proceedings of the Royal Society of London B, 270, 14471454. Liebers D, Helbig AJ and De Knijff P (2001) Genetic differentiation and phylogeography of gulls in the Larus cachinnans- fuscus group (Aves: Charadriiformes). Molecular Ecology, 10, 2447-2462. Longmire L, Maltbie M, and Baker RJ (1997) Use of “lysis buffer” in DNA isolation and its implication for museum collections. Occasional Papers OP-163. Museum of Texas Tech University. Nei M and Kumar S (2000) Molecular Evolution and Phylogenetics. Oxford Univ. Press, Oxford. Pavlova A, Zink RM, Drovetski SV, Red'kin Ya, and Rohwer S (2003) Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history. Auk, 120, 744-758. Posada D and Crandall KA (1998) Modeltest: Testing the model of DNA substitution. Bioinformatics, 14, 817-818. Ramos-Onsins SE and Rozas J (2002) Statistical properties of new neutrality tests against 122
population growth. Molecular Biology and Evolution, 19, 2092-2100. Rozas J, Sanchez-DelBarrio JC, Messeguer X, and Rozas R (2003) DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics, 19, 2496-2497. Saiki RK, Gelfand DH, Stoffel S et al. (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 239, 487-491. Salzburger W, Martens J, Nazarenko AA et al. (2002) Phylogeography of the Eurasian willow tit (Parus montanus) based on DNA sequences of the mitochondrial cytochrome b gene. Molecular Phylogenetics and Evolution, 24, 26-34. Schneider S, Roessli D and Excoffier L (2000) Arlequin, Version 2.000. A software for population genetic data analysis. http://lgb.unige.ch/arlequin/, Genetics and Biometry laboratory, University of Geneva, Switzerland. Schneider S and Excoffier L (1999) Estimation of past demographic parameters from the distribution of pairwise differences when the mutation rates vary among sites: application to human mitochondrial DNA. Genetics, 152, 1070-1089. Stepanyan LS (1990) Conspectus of the ornithological fauna of the USSR. Nauka, Moscow, Russia (in Russian). Swofford DL (2000) PAUP*. Phylogenetic analysis using parsimony (*and other methods), Version 4.0b2. Sinauer, Sunderland, MA. Tajima F (1996) The amount of DNA polymorphism maintained in finite population when the neutral mutation rate varies among sites. Genetics, 143, 1457-1465. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution, 10, 512-526. Tarasov PE, Peyron O, Guiot J et al. (1999) Last glacial maximum climate of the former Soviet Union and Mongolia reconstructed from pollen and plant macrofossil data. Climate Dynamics, 15, 227-240. Tarr CL (1995) Primers for amplification and determination of mitochondrial controlregion sequences in oscine passerines. Molecular Ecology, 4, 527-529. 123
Zhang D-X and Hewitt GM (1996) Nuclear integrations: challenges for mitochondrial DNA markers. Trends in Ecology and Evolution, 11, 247-251. Zink RM, Weller SJ and Blackwell RC (1998) Molecular phylogenetics of the avian genus Pipilo and a biogeographic argument for taxonomic uncertainty. Molecular Phylogenetics and Evolution, 10, 191-201. Zink RM, Drovetski S and Rohwer S (2002a) Phylogeographic patterns in the Great Spotted Woodpecker (Dendrocopos major) across Eurasia. Journal of Avian Biology, 35, 175-178. Zink RM, Rohwer S, Drovetski S, Blakwell-Rago RC, Farrell SL (2002b) Holarctic phylogeography and species limits of three-toed woodpeckers. Condor, 104, 167-170. Zink RM, Drovetski SV, Questiau S et al. (2003) Recent evolutionary history of the bluethroat (Luscinia svecica) across Eurasia. Molecular Ecology, 12, 3069-3075.
Acknowledgments We thank the Burke museum for curatorial assistance, S. Farrell and M. Westberg for laboratory assistance. We also thank S. Drovetski, C. S. Wood, B. Schmidt, D. Banin, A. Andreev, I. Fadeev, E. Nesterov, I. Karagodin, E. Koblik, A. Jones, B. Barber, G. Voelker, and V. Sotnikov for logistical help with expeditions and collecting, and S. Birks, who subsampled some tissues. We are grateful to G. Eddy for funding fieldwork. Additional support came from the NSF (DEB 9707496) and Dayton-Wilkie fund.
124
Table 4.1. Genetic characteristics of the great tit Parus major populations (sites with gaps and missing data are excluded) (Figure 4.1). The populations for which most pairwise Fst values were significant are shown in bold font. Asterisks indicate populations, for which two subsets of data are available.
Species
Collecting
sam
number
haplo-
nucleotide
locality
-ple
of haplo-
type
diversity
size
types
diversity
*10-4
Source
P. minor
Primor'e
29
21
0.956
19.8
This study
P.
Almaty
4
2
0.5
6.0
CR+ND2
Norway
2
2
-
-
Kursk
6
7
0.96
15.1
Oryol
2
Smolensk
2
2
-
-
Crimea
3
3
1
20.1
Krasnodar
5
4
0.9
9.6
Moscow
31
15
0.796
11
is shared by
Medvedevo
2
1
-
-
50%
Astrakhan’
8
4
0.75
10.1
individuals
Sverdlovsk
1
-
-
-
Krasnoyarsk
1
-
-
-
Irkutsk
8
5
0.857
11.7
Buryatia
4
2
0.5
12.1
Mongolia
2
1
-
-
Magadan
7
1
0
0
Pooled major
84
34
0.748
10.1
Fst = 0.02, P > 0.05
12 parsimony informative sites;
haplotype
47 polymorphic sites,
most common
(1653 bp used after removing missing data);
P. major
This study: 1667 bp of CR+ ND2
bokharensis
125
Table 4.1. Continued.
type
diversity
size
types
diversity
*10-4
P. major
UK
10
0.844
30
major
Portugal
14
0.945
37.3
Spain
*
9
7
0.917
38.4
Germany *
8
5
0.857
24.7
Netherlands *
6
3
0.6
17.3
Estonia
*
9
6
0.833
26.9
most common
Sweden
*
8
6
0.893
24.7
haplotype
Finland
28
16
0.794
24.3
is shared by
Urals
9
0.917
33.6
36.8%
Amur
15
0.695
27.7
Kirghizia
6
0.933
27.7
Corsica
2
1
-
-
Morocco
1
-
-
-
125
57
0.885
32.5
individuals
Pooled major P. major
Spain
*
9
0.815
24.5
major
Germany
*
8
0.75
14.2
Netherlands *
6
0.5
11.3
most common
Estonia
*
9
0.79
13.2
haplotype
Sweden
*
8
0.781
18.5
is shared by
Finland,
8
0.781
12.7
Finland, Oulu
10
0.78
14.9
Finland,
10
0.9
33.0
68
0.863
18.7
35.3% individuals
Kilpisjarvi
Harjavalta Pooled major
Fst = 0.12
of haplo-
Fst = 0.02
-ple
Source
25 parsimony informative sites;
nucleotide
17 parsim. inform. sites;
haplo-
52 polymorphic sites,
number
52 polymorphic sites,
sam
Kvist et al. 2003: 578 bp of CR (beginning part);
Collecting locality
Kvist et al. 1999: 1189 bp of complete CR;
Species
126
Table 4.2. Genetic characteristics of the willow tit Parus montanus populations (Figure 4.1). The populations for which most pairwise Fst values were significant are shown in bold font. Asterisks indicate populations for which three subsets of data are available.
Collecting
sample
number of
haplotype
nucleotide
locality
size
haplo-
diversity
diversity *10-4
types
haplotype is shared by 11.5% individuals
13
0.989
31.7
Moscow
12
Medvedevo
6
6
1
32.3
Mezen’
8
8
1
31.2
Sverdlovsk
1
1
-
-
Noyabr’sk
5
5
1
19.4
Gorno-Altay
2
9
1
27.9
Tyva
7
Krasnoyarsk
1
1
-
-
Irkutsk
23
21
0.988
21.9
Buryatia
12
8
0.924
12.6
Mongolia
8
7
0.964
23.6
Anadyr
10
5
0.756
6.1
*
13
11
0.974
20.5
Khabarovsk *
7
6
0.952
23.9
Magadan
*
17
14
0.956
20.3
Kamchatka *
5
3
0.7
18.3
139
99
0.980
23.2
*
13
6
0.718
22.7
This study;
Khabarovsk *
7
4
0.714
24.0
720bp of CR-II
Magadan
*
17
9
0.831
21.0
Kamchatka *
5
2
0.4
11.2
Primor’e
Pooled
Fst = 0.053, P < 0.05
common
2
113 polymorphic sites, 44 parsimony informative sites;
most
Smolensk
(1752 bp used after removing missing data);
P. montanus
Source
This study: 1761bp of CR+ ND2
Species
montanus P. montanus
Primor’e
127
Table 4.2. Continued. number of
haplotype
nucleotide
size
haplo-
diversity
diversity
P. montanus
*10-4
Alps
8
6
0.893
76.7
Germany
4
3
0.833
50.8
Poland
2
2
1
67.7
Finland
13
11
0.962
62.0
most
Sweden
12
8
0.892
39.7
common
Latvia
8
6
0.929
38.7
haplotype
Urals
13
12
0.987
67.7
is shared by
Ussuri
*
5
4
0.9
57.5
11.7%
Amur
*
4
4
1
59.2
Magadan
*
7
7
1
80.6
Kamchatka *
7
5
0.857
41.9
Sakhalin
4
4
1
73.3
Japan
7
5
0.905
61.2
94
71
0.981
111.4
62
0.976
67.1
individuals
Pooled
1
montanus P. montanus
Finland
13
13
1
55
Kvist et al.
borealis
Sweden
12
12
1
50
1998
Pooled
25
25
53
1207 bp of
all
Fst = 0.15
types
Source
25 parsimony informative sites;
locality
sample
56 polymorphic sites,
Collecting
Kvist et al. 2001: 592 bp of CR-I;
Species
haplotypes
complete CR;
are unique
52 polym. sites, 13 pars. inform.
1
top row values reported in the Table 1 pg. 932 of Kvist et al. 2001, bottom row – values reported on page
934 of the text (nucl. diversity of 67.1 was calculated by us in DnaSP and was not reported in original paper).
128
Table 4.3. Significant pairwise Fst -values for populations of the great and the willow tits.
Great tit Parus major Irkutsk Kursk
0.09
Magadan
0.14
Astrakhan’
0.10
Moscow
0.17
Krasnodar
Buryatia
0.07
0.28
0.17
Willow tit Parus montanus Medvedevo Moscow+
Anadyr
Kamchatka
0.23
0.1
0.35
0.16
Smolensk Medvedevo
-
Mezen’
0.20
Noyabr’sk
0.43
0.21
Gorno-Altay +
0.28
0.13
0.24
0.12
Buryatia
0.38
0.22
Mongolia
0.35
0.15
Tyva Irkutsk
0.06
Anadyr
0.35
-
Primor’e
0.06
0.26
0.13
0.34
0.16
0.23
0.10
Khabarovsk Magadan
0.06
Kamchatka
0.16
-
129
Table 4.4. Tests for population expansion and selective neutrality for the populations with sample sizes of 10 or more individuals. Asterisks indicate significance at 0.05 level, n/s- not significant, unimod-unimodal, bimod-bimodal, multim-multimodal. mismatch
mismatch
mean
variance
Tau
Theta 0
Theta 1
Mismatch
Tajima’s D
Fu’s Fs
R2
pattern
Great tit Parus major
Moscow
1.9
2.4
1.2
1
8
unimod n/s
-2.34*
-10.7*
0.045*
Pooled major
2.1
3.1
1.1
1.2
12
unimod n/s
-2.65*
-27.8*
0.02*
Willow tit Parus montanus
Moscow
6.0
9.2
3.9
2
4926
multim n/s
-1.85*
-7.1*
0.063*
Irkutsk
4.1
6.1
2.8
0.8
8825
bimod n/s
-2.33*
-17.6*
0.044*
Buryatia
2.4
1.7
2.6
0
4682
unimod n/s
-1.36 n/s
-3.63*
0.098*
Anadyr
1.1
0.6
unimod
-0.94 n/s
-2.1*
0.129*
Primor’e
3.6
6.2
1.3
2.2
5315
multim n/s
-2.02*
-6.15*
0.09*
Magadan
4.0
4.7
4.4
0
28
bimod n/s
-2.09*
-9.18*
0.05*
Pooled montanus
4.3
5.4
3.2
1
323
unimod n/s
-2.57*
-25.7*
0.016*
no estimates
13
Figure 4.1. Collecting localities of the great tit and the willow tit samples: black circles indicate new samples; open circles- samples of Kvist et al. (2003) for the great tit and Kvist et al. (2001) for the willow tit. Shaded area represent species’ world distribution (following Harrap and Quinn 1996).
131
Fig. 4.2. ML phylogenetic tree of haplotypes for the great tit. Parus minor is used as an outgroup. Numbers above the branches are percentages of 100 maximum likelihood bootstrap replicates supporting that branch. Lab numbers precede locality abbreviations. Prim- Primor’e, Alma- Almaty, Mosc- Moscow, Astr-Astrahkan’, Smol- Smolensk, Crim- Crimea, Kurs- Kursk, Krya- Krasnoyarsk, Norw- Norway, Krda- Krasnodar, IrkuIrkutsk, Oryo- Oryol, Sver-Sverdlovsk, Bura-Buryatia.
132
Figure 4.2.
99
0.01 substitutions/site
69
P. minor
P. bokharensis
99M 41M 105M 27A 56S 10Kurs, 115, 143Irku, 114C 26Kur 57M N=3
100
92Pri m 135Pri 138, m 139Prim, 126Pri N=2 m 128Pri 77Pri m 16Pri m 94Pri m 125, 137Prim, m 78, 90, 134Prim, N=2 121Pri N=3 133, m 141Prim, 142Pri N=2 91Pri m 17Pri m m136Pri 131Pri m m15Pri m124Pri 132Pri m 120Pri m m 127Pri m 93Pri 140Prim m
35Alm 1,50,51Alma, a N=3
s 42Mosc 14Mosc 38M 9Krya 20Cri 4Norw-most common haplotype, m 86Krda N=42 117Irku 69Oryo 123Bura 107Mosc 97Mosc 129Irku 130Ir 22Krda 39Mosc 18Sver, 24Kurs, 95,100,102Mosc, 76Krda N=5 23Kur
43, 53M s 111Mosc,
31Ast N=2 70Oryo r 3Norw, 29, 33Astr, N=3
44M
P. major
13
Fig. 4.3. ML phylogenetic tree of haplotypes for the willow tit, rooted with marsh tit P. palustris. The monophyly of the clade was supported by 100% of 100 maximum likelihood bootstrap replicates. Lab numbers precede locality abbreviations, which are the same as for Fig. 4.2. Maga- Magadan, Medv- Medvedevo, Noya- Noyabr’sk, MezeMezen’, Kamc- Kamchatka, Mong- Mongolia, Khab- Khabarovsk, Gora- Gorno-Altay, Anad- Anadyr, also see legend for Fig. 4.2.
134
Figure 4.3. 100
64Mosc 91Maga 98Mosc 103Irku 36Medv, 73Tyva, 115Irku, 141Irku N=3 46Noya 44Meze 39Meze 63Mosc 16Maga 139Irku 4Kam 70Mosc c 109Pri 90Maga m 56Mong 101Smo l 78Tyva 40Meze 69Mosc 32Medv 38Meze 33Medv 31Medv 11Mosc, 42Meze, 131Bura N=2 94Mosc 55, 58Mong, 129Khab, 66Gora, 136Irku, N=3 15Gora N=2 87Maga 17Sver 89Maga 6Khab 121Pri 2, m 67Bura, 52, 124Prim, 59Mong N=4 120Pri 118Irku m 53Pri 128Pri m m77Tyva 122Bura 43Meze 114Irku 105Irku 3Khab-most common haplotype, N=17 0.01 substitutions/site
P. montanus
3Khab-most common haplotype, 123Pri N=17 97Khab, m 96, 75Tyva N=2 48Mong 83Maga 102Irku 19Maga 117Irku 76Tyva 99Mosc 112, 132Bura, 34Medv N=2 85Maga 100Smo 7Irk l 149Anad u 30, 51Kamc, 145, 150-152Anad, 29, 65Kam N=7 9Krya c 148Anad 147Anad 146Anad 144Anad 142Anad 93Maga 119Irku 74Tyva 86Maga 12, 62Mosc, 135Irku N=2 37Meze, 47Noya, 57Mong, 127Prim, 1Bura, 133Irku N=5 108Pri 49Mong m 140Irku 60Mong 137Irku 13Khab 35Medv 72Tyva 106Irku 8Irk 61Bura u 95Khab 84Maga 82Maga 45Noya 130Bura 104Irku 10Maga 71Mosc 54Noya 110Pri m
13
Fig. 4.4. Nucleotide diversity for the great tit (A, B) and the willow tit (C, D) populations plotted against North latitude (A, C) and East longitude (B, D). Closed circles indicate our data, open circles- data of Kvist et al. (2003) for the great tit, and Kvist et al. (2001) for the willow tit (Tables 1, 2). Solid lines are linear regression lines for our data, dashed lines- for data of Kvist et al. (2001, 2003). The values of R2 and corresponding P-values for our data are on lower left of each graph, for the data of Kvist et al.(2001, 2003) are on upper right.
136
Figure 4.4. A. Nucleotide diversity vs latitude P. major R2=0.23 P=0.13
40
50 Our data Kvist et al. 2003
30 20
Our data
R2 =0.35 P =0.12
10
Kvist et al. 2003
0 35
40
nucl. diversity
nucl. diversity
50
B. Nucleotide diversity vs longitude P. major
45
50
55
60
R2 =0.008 P =0.8
40 30 20 R2 =0.47 P =0.06
10 0
65
-20
0
20
North latitude
60
100 Our data Kvist et al. 2001
R2 =0.001 P =0.91
40
Our data
20 Kvist et al. 2001
0 40
45
50
55
60
North latitude
80
100 120 140 160
D. Nucleotide diversity vs longitude P. montanus
65
70
nucl. diversity
nucl. diversity
R2 =0.02 P =0.64
80
60
East longitude
C. Nucleotide diversity vs latitude P. montanus 100
40
R2 =0.04 P =0.53
80 60 40 R2 =0.59 P =0.002
20 0 0
20
40
60
80
100 120 140 160 180
East longitude
13
Chapter 5
Comparative Phylogeography of Eurasian birds based on mitochondrial DNA sequences 5.
Abstract.—Comparison of phylogeographic structure among 28 species of Eurasian birds reveals the existence of several regions of phylogeographic endemism in southern Eurasia. In contrast, phylogeographic splits on the northern part of the continent are either incongruent or absent in some species, reflecting differential post-glacial histories, large geographic distances and the stochastic nature of the haplotype trees. The frequency of the most common haplotype per clade was used as an index of genetic diversity to infer past population dynamics. The clades from southern and southeastern Asia have relatively high genetic diversity, indicating high effective population size and stable population range, consistent with the distribution of suitable habitats during last glacial maxima 18,000- 14,000 years ago. In general, genetic diversity was lower in clades widely distributed throughout Eurasia, indicating past population bottlenecks and recent colonization of the region. However, several widespread Eurasian species (e. g. willow tit, yellow wagtail) possessed high genetic diversity; moreover, for one species (common rosefinch) genetic diversity decreased southwards. These species likely survived the Pleistocene in high numbers in treeless Eurasian steppe-tundra without drastic range contractions. Therefore, Eurasian species of birds responded individually to Pleistocene climatic oscillations. Following the decline of forests the ranges of some species were displaced into multiple refugia in southern European peninsulas, southern mountain regions of Eurasia, southeastern Eurasia and Beringia, while the ranges of other species were fragmented, but stayed sufficiently large to maintain high genetic diversity. Post-Pleistocene recolonization of Eurasia shaped the currently observed mosaic of observed phylogeographic structures.
5
This chapter was written in the style of the journal Evolution. Submission is pending.
138
Study of geographical patterns of haplotype distribution (termed phylogeography by Avise et al. (1987)) along with the application of likelihood and coalescent methods (e.g. Beerli and Felsenstein 1999; Harpending et al. 1998; Rogers 1995; Ray et al. 2003) and traditional methods of population genetics (Wright 1951; Neigel 2002), allow researchers to make evolutionary inferences about populations, such as isolation, bottlenecks, population expansions, gene flow and degree of population differentiation. Comparative phylogeography explains the existence of common phylogeographic patterns among multiple taxa by the similar effects of vicariant events on codistributed taxa. Thus, comparative phylogeography links knowledge from multiple lineages regarding historical and contemporary biogeographic processes at the population and regional scale. The results from this research program is an increased understanding of the process of speciation, the effects of historical events and geographic barriers on species’ diversity, the patterns of regional colonization, and the formation and dynamics of ecological communities (Bermingham and Moritz 1998, Arbogast and Kenagy 2001). Congruent phylogeographic patterns provide evidence for the stability of current species assemblages (Zink 2002). In contrast, incongruent phylogeographic structures of multiple species indicate that species have had differing evolutionary histories. Despite its shortcomings (Ballard and Whitlock 2004), mitochondrial DNA (mtDNA) is widely used to document phylogeographic patterns. As the result of its maternal inheritance (but see Kvist et al. 2003b) and rapid evolution, mtDNA lineages are rapidly sorted geographically, providing good intraspecific resolution (Arbogast and Kenagy 2001). Along with the parallel investigation of other molecular markers and morphological or ecological differentiation within species, mtDNA studies are able to reveal multiple aspects of species’ evolutionary history. Drastic climatic changes during the Pleistocene repeatedly altered the distribution of habitats, resulting in extinction of some taxa and shifts in spatial distribution and population sizes of survivors. In Eurasia, during the maximum cooling of the Pleistocene (about 18,000 years ago), glacial ice covered a large part of western Europe, while the area from northern Eurasia to the southern belt of mountain ranges was treeless steppe139
tundra underlain with permafrost (Hewitt 1999; Stewart and Lister 2001). Fossil evidence from plants and animals suggests that most organisms survived the periods of glacial advances in multiple southern refugia that included Iberia, Italy, the Balkans, the eastern Alps, the southern shores of the Black and Caspian seas, and Beringia (Hewitt 1999; Tribsch and Schonswetter 2003; Seddon et al. 2004). When the climate warmed at about 16,000 years ago, species expanded their range northward (Hewitt 1999). Pollen data suggest that plant and tree species in the east of Europe spread northward earlier than in central and western Europe (Huntley and Birks 1983). Several predictions about the effect of Pleistocene climatic oscillations on species’ genetic structure can be made. Different genetic signatures are expected between species that survived climatic cooling without drastic range contraction, and those species that were forced into southern refugia. The former should possess high effective population size and high genetic diversity. For the latter, the populations with more northerly distribution should have low effective population sizes as the result of multiple displacements of their ranges (Hewitt 1996). This effect should be more pronounced in northwestern Europe, because this area was under the glacial ice during maximum cooling of the Pleistocene. However, a slow northward expansion from southern refugia could maintain most of species’ genetic diversity (Hewitt 1996) and might not be distinguishable from the long-term range occupation. In contrast, rapid long-distance (leptokurtic) colonization from southern refugia with sequential bottlenecking should leave a signature of decreased genetic diversity on a leading front of the expansion (Hewitt 1996). However Stewart and Lister (2001), based on studies of non-analogue assemblages of small mammals (mixture of fossils from mammals of tundra-steppe and of deciduous woodlands), dated to the supposedly treeless Younger Dryas Stadial (11,000-10,000 years ago) in northern Europe, point out the possibility of the existence of cryptic northern refugia, where organisms might have survived in areas of sheltered topography that provide suitable microclimates. Survival in such cryptic refugia would leave the same signature of decreased genetic diversity as the rapid long-distance colonization and provide an alternative, and simpler, explanation to observed pattern of 140
species’ genetic diversity (Cruzan and Templeton 2000). Therefore, distinguishing between two alternative hypotheses (whether or not the species was able to survive in north Eurasian steppe-tundra) is extremely challenging and requires additional evidence, such as palaeo-ecological data. For example, species that are closely associated with trees most likely were not able to survive in a cold and treeless environment, while the coldtolerant species, whose current ranges extend into northern (or mountain) tundra might have been able to survive the climate changes without significant range contraction. Some other predictions about phylogeographic structure of species expanding from southern refugia are homogenizing gene flow over large northern areas due to selection for the dispersal capacity (Liebers and Helbig 2002) and less phylogeographic structure in northern taxa due to their relatively young age (Dynesius and Jansson 2000). The last prediction, however, can not be tested for Eurasian birds, because habitats are not uniformly distributed around the continent and the presence of geographic barriers in southern Eurasia can strongly affect phylogenetic structure of the species. In this work I test some of these predictions using mitochondrial DNA (mtDNA) sequence data for 28 species (or species complexes) of Eurasian birds. I also test whether avian species in Eurasia, as members of a large community assemblage, were similarly affected by geological events and, therefore, show congruent phylogeographic patterns. Rejection of this hypothesis would indicate that present-time avifauna comprise species with different and complex evolutionary histories (Knowles and Maddison 2002). I overlay the patterns of phylogeographic subdivision on the reconstruction of vegetation during the last glacial maximum (18,000-14,000 years ago) (Frenzel et al. 1992; Adams 2002) to find whether the current phylogeographic structure of avian species correlates with distribution of vegetation during maximum cooling and, therefore, can be explained by the Late Pleistocene shifts in species’ ranges.
METHODS I surveyed 34 phylogeographic studies on Eurasian birds (Table 5.1). I compared the patterns of geographical distribution of mtDNA haplotypes and genetic diversity 141
among 28 species with predictions made for postglacial species’ history on the Eurasian continent. For all surveyed species I recorded or calculated from original papers the number of sampling localities, the number of sampled individuals and the number of haplotypes per clade, and the frequency of the most common haplotype for all clades with sample sizes of 9 or more individuals (Table 5.1). Some parts of the mtDNA, such as 5’-end of control region (CR-I), evolve much faster than the others. As a result, genetic diversity estimated from sequences that have been evolving with different rates might not be comparable (Pavlova et al., unpublished). I divided all studies into three groups according to known differences in rates of molecular evolution. Group A: studies based solely on mitochondrial control region (CR) sequences, including rapidly evolving CR-I (average length of sequences 540 base pairs (bp), range 228-1144 bp; N = 18). Group B: studies based on longer (1286 bp on average, range 993-1761bp) sequences, usually parts of multiple mtDNA gene regions (including CR-II, ND2, ND3, ND6 and cyt b; N = 13). Group C: studies based on short sequences of slowly evolving cyt b (up to 336 bp, N = 3). The group C was used only for comparison of the clade distribution because of the small sample size and/or low molecular resolution. I explored geographic distribution of clades for 13 group B species (Dendrocopos major (Zink et al. 2002a), Picoides tridactylus (Zink et al. 2002b), Luscinia svecica (Zink et al. 2003), Motacilla flava, Motacilla citreola (Pavlova et al. 2003), Motacilla alba (Pavlova et al. in review), Carpodacus erythrinus (Pavlova et al. unpublished), Parus major (Pavlova et al. unpublished), Parus montanus (Pavlova et al. unpublished), Emberiza schoeniclus (Zink et al. unpublished), Sitta europaea (Zink et al. unpublished), Troglodytes troglodytes (Drovetski et al. 2003), and Phylloscopus trochiloides (Irwin et al. 2001)). I computed the frequency of haplotype clades for each sampling locality from GenBank sequences or unpublished sequence data. Frequency of the clade per locality was calculates as the number of individuals sampled from this locality and belonging to this clade divided by the total number of individuals sampled from this locality. To visualize geographic distribution of clades I plotted clade frequencies per locality as pie142
diagrams on the species’ distribution maps (Figs. 5.1-5.6). I explored the patterns of nucleotide diversity (π) in these species to summarize patterns of postglacial colonization. I also used small subsets of each of these data sets (up to three individuals per clade) to construct UPGMA phenograms and to compare sequence divergence between individuals from different clades. Due to the subsampling process the depths of the clades shown on the UPGMA trees (Fig. 5.1-5.6) are smaller then those reported in the original papers. For two groups of studies (A and B) I performed linear regression analysis of the frequency of the most common haplotype (MCH) on the length of sequenced fragment, number of sampled individuals and number of sampling localities to ensure that sampling does not influence the frequency of MCH. I also explored how sampling affects tree topology by regressing the number of clades detected per species on number of sampled individuals, haplotypes and localities.
RESULTS Most of the 28 surveyed species and species complexes displayed some degree of geographic structure of their haplotype trees. According to their phylogeographic structure, the surveyed species could be divided into four groups roughly following Avise et al.’s (1987) categories, although placement of some species is somewhat ambiguous (Table 5.1): I. Species with divergent geographically structured clades. Divergent geographically isolated phylogroups were present in eight species (28.6%): Eurasian nuthatch (Fig. 5.1A), three-toed woodpecker (Fig. 5.1B), great spotted woodpecker (Fig. 5.1C), winter wren (Fig. 5.2A), the willow tit species complex (Fig. 5.4C), yellow wagtail, (Fig. 5.3), carrion and hooded crows and rock partridge (Table 5.1). These species have 3 clades on average, with the mean clade divergence of 4.6%. This type of phylogeographic structure is usually interpreted as long-term geographic isolation and suggests survival in multiple refugia due to the loss of suitable habitats during the Pleistocene cooling. Simulations have shown that deep divergences could also be stochastically derived in the absence of a barrier but with limited gene flow (Irwin 2002). 143
II. Species displaying divergent clades with overlapping distribution. Geographically unsorted splits were observed in 15 species (53.6%): greenish warbler (Fig. 5.2B), citrine wagtail (Fig. 5.3), the great tit species complex (Fig. 5.4A, B), white wagtail (Fig. 5.5A), common reed bunting (Fig. 5.5B), bluethroat (Fig. 5.6B), hazel grouse, common raven, dunlin, rock ptarmigan, blue tit, bearded vulture, great reed warbler, grey partridge, and lesser black-backed gull (Table 5.1). The latter four species displayed a gradual west-east change of frequency of the haplotypes belonging to two divergent clades. Category II species have on average 3.2 clades, with the mean divergence of 2.9%. This type of phylogeographic structure is typically explained by secondary contact between allopatrically diverged populations, but it could also arise from incomplete lineage sorting. III. Species with some geographic structure on the phylogenetic tree, but incompletely sorted lineages and small sequence divergence. Incipient divergence was detected in two species (7.1%), common rosefinch (Fig. 5.6A) and red-crested pochard (Table 5.1), and two phylogroups (the western and the northeastern clades of the yellow wagtail). This type of structure can result from limited gene flow between populations with the absence of geographic barriers and might indicate incipient speciation. IV. Species with no geographically structured clades. Three species (10.7%) did not display phylogeographic structure: common chaffinch, Siberian tit, and European greenfinch. The absence of structure in these species could be partially attributed to the limited geographical sampling (Table 5.1). No phylogeographic structure was also found for two widespread Eurasian tits, great tit and willow tit (Fig. 5.4B, D), Eurasian clades of three-toed and great spotted woodpeckers (Fig. 5.1A, B), Holarctic clade of the common raven and the Beringia clade of the rock ptarmigan. Unstructured haplotype trees can be caused by recent range or demographic expansion (e.g. as a result of postglacial recolonization) and usually suggest extensive gene flow (recent or recurrent) between populations. For all species, on average 2.6 clades were detected (range 1-6). This number did not depend on the number of sampled individuals or detected haplotypes (P > 0.05), but 144
was highly correlated with the number of sampling localities (R2 = 0.37, P = 0.0002, Fig. 5.9). The frequency of the most common haplotype (MCH) in each clade was independent from the number of sequenced nucleotides in both groups, but was slightly higher in CR studies (group A) on average (Fig. 5.10A). The frequency of MCH was independent of the number of sampled individuals (in both groups, Fig. 5.10B), or sampling localities (for the group B studies not shown. For group A studies not calculated). If the colonization of Eurasia was rapid, then just a portion of all haplotypes from the refugial population would have been involved in colonization. Under this scenario, one of the haplotypes may have spread over the continent and would now exist in high frequency in many populations (Hewitt 1999). Under the assumption that population expansion is always accompanied by a high frequency of a single haplotype, the frequency of a most common and widespread haplotype in the species can characterize population dynamics, where high frequency (and therefore low genetic diversity) would imply small effective population size and a population bottleneck with following range and demographic expansion, whereas low frequency of the most common haplotype (and therefore high genetic diversity) would indicate long term range stability and high effective population size. For the group B studies, the clades from southern and southeastern Asia tended to have low frequencies of MCH (Fig.8A, Table 5.1), indicating range stability and higher population size. The clades with wide Eurasian distribution tended to have higher frequency of the MCH, indicating range expansion and lower effective population size (Fig. 5.8A, Table 5.1). However, several widespread Eurasian birds possess low frequency of MCH, which suggests that these clades did not undergo drastic range and population size contractions. The widespread clades with low frequency of MCH were observed in common rosefinch and willow tit, the yellow wagtail from western and northeastern Eurasia, northeastern clade of greenish warbler, and northern clade of the bluethroat (Fig. 5.8A). For the group A studies, only two samples (of hazel grouse and Japanese tit) were 145
available from southeastern Asia, and they both had low frequency of MCH (Fig. 5.8B). Among the species with west-east clinal change frequencies of haplotype clades the “western” clades always had higher frequencies of the MCH (Fig. 5.8B). Higher frequency of MCH in western Europe, and thus lower genetic diversity, is consistent with recent postglacial recolonization of western Europe. Low frequencies of MCH are found in Eurasian willow tit, common chaffinch from Europe and North Africa, Eurasian hazel grouse, Nearctic raven, dunlin from northern part of central Siberia, yellow wagtail from Europe and Africa, “eastern” clade of grey partridge from western Europe, and redcrested pochard from eastern Europe and central Asia (Fig. 5.8B, Table1). Therefore, these species do not show a signature of postglacial range expansion. The Sicilian clade of rock partridge in contrast, displayed high frequency of MCH, indicating low effective population size, which might result from a bottleneck. Maximum sequence divergence between individuals from different clades for the group B ranged from 0.2% in unstructured topology of the great tit and 0.7% for the bluethroat to 6.2% for the haplotype tree of the yellow wagtails; for the group A, it ranged from 0.64% in Rock ptarmigan to 6% in the great tit species complex.
DISCUSSION General phylogeographic patterns.—Southern and northern Eurasia display different patterns of phylogeographic structures. Four areas of regional phylogeographic endemism were identified in southern Eurasia: the Caucasus, central Asia, southern Asia and a large region in southeastern Asia (Fig. 5.7). In northern Eurasia phylogeographic structures were incongruent (Fig. 5.7). Southern Eurasia. The Caucasus is an isolated mountain region between Black and Caspian seas. A history of isolation of Caucasian populations is evidenced by endemic clusters of haplotypes in the Caucasus. Clades with distribution restricted to Caucasus were observed for five species (Table 5.1): greenish warbler, winter wren, Eurasian nuthatch, white wagtail and common rosefinch (Figs. 5.1, 5.2, 5.5, 5.6). From all birds surveyed from the 146
Caucasus only two species, great tit and great spotted woodpecker, were not differentiated in this region (Figs. 5.1, 5.4). The Eurasian populations of these two species possessed low genetic diversity. Nucleotide diversity (π) for the populations of great spotted woodpecker ranged from 0.0006 to 0.002 (Zink et al. 2002a). For the great tit π ranged from 0 to 0.002 (group B study, Pavlova et al. unpublished). Low genetic diversity for both species suggests recent colonization of their entire ranges including the Caucasus. Central Asia is comprised of mountain regions of Tian Shan and Pamirs. In central Asia, three species possessed differentiated clades: great tit and willow tit species complexes (Parus bokharensis and P. songarus respectively) and greenish warbler (Table 5.1, Figs.2, 4, 7). Southern Asia includes southern slope of the Himalaya mountain range. In southern Asia, five species had endemic haplotype clades: great tit species complex (Parus cinereus), willow tit species complex (P. affinis and P. weigoldicus), citrine wagtail, white wagtail and winter wren (Table 5.1, Figs. 5.3, 5.4, 5.5, 5.7). Except for the greenish warbler, the extent of the clade distributions and the levels of genetic diversity in central and southern Asia are still unknown, because too few samples are available (Fig. 5.1-5.6). Southeastern Asia includes eastern China, southeastern Russia, Sakhalin, Japan and Kamchatka (Fig. 5.7). In southeastern Asia, endemic clades were observed in eight species: great spotted woodpecker, yellow and white wagtails, great tit species complex (P. minor), winter wren, hazel grouse, carrion crow and common reed bunting (Table 5.1, Figs. 5.1, 5.2, 5.5, 5.7). For most of these species higher genetic diversity was observed in the clades from southeastern Asia, suggesting higher long-term effective population sizes. Northern Eurasia. Unlike southern Eurasia, the phylogeographic lineages in northern Eurasia are rarely concordantly distributed and are mostly unsorted geographically (Fig. 5.7). In some species (Eurasian nuthatch, yellow and citrine wagtails), geographically isolated clades exist in eastern and western parts of the range (Figs. 5.1, 5.3, 5.7). For some species (bearded vulture, grey partridge, great reed warbler, lesser black-backed gull), haplotypes 147
belonging to one clade are more frequent on the west of species’ range, while haplotypes from another clade are more frequent on the east of the range (Table 5.1). In other species (bluethroat and common reed bunting), divergence exists between northern and southern populations (Figs.5, 6, 7). Finally, no phylogeographic structure exists in northern Eurasian clades of some species (three-toed and great spotted woodpeckers, great and willow tits, white wagtail and common raven) (Figs. 5.1, 5.4, 5.5, 5.7). The absence of concordance in the geographic distribution of clades indicates that different species have different post-glacial colonization histories. Clades of many species represent historical populations which diverged in different Pleistocene refugia. When the climate in northern Eurasia became suitable these allopatrically diverged populations expanded their ranges, which now meet in some species (phylogeographic category II), but remain isolated in other species (phylogeographic category I). Northern part of the Eurasian continent differs significantly from its southern part by the absence of mountains and presence of a continuous belt of coniferous forests. The absence of geographical barriers and uniformly distributed habitats facilitates the gene flow among Eurasian localities and explain the absence of complete geographic sorting of the lineages for majority of species (phylogeographic category II). Contrary to findings for small mammals (Hewitt 2004), neither the Ural mountain range nor the large Siberian rivers appeared to be an important barrier to the east-west dispersal, although few phylogeographic splits were found near the Urals (dunlin, yellow and citrine wagtails). This is not surprising considering the agile nature of birds and their ability to disperse over large distances. Population dynamics.--When found in high frequency in many populations, the most common haplotype usually forms a center of a star phylogeny, which is typically interpreted as a range expansion from glacial refugia (e.g. Kvist 2003a). Species possessing low genetic diversity (and high frequency of MCH) are most likely either recent colonizers of their range or they experienced a bottleneck during recent history. On the other hand, high genetic diversity (low frequency of MCH) in species implies that the populations must have occupied their range long enough to accumulate mutations. Except 148
for the European clade of winter wren, the most recent colonizers appeared to be the clades with widespread Eurasian distribution and no structure within clades: the threetoed and great spotted woodpeckers, Eurasian clades of the white wagtail and common reed bunting, and the great tit (Fig. 5.8 A). On the other hand, several widespread Eurasian species (common rosefinch, western Eurasian yellow wagtail, willow tit, the northern clade of the bluethroat and the northeastern clade of the greenish warbler) as well as most clades from southeastern Asia displayed high genetic diversity, which suggests that these populations had high historical effective population size. These species might have been able to survive through the Pleistocene climate oscillations in high numbers without drastic range contractions and, thus, might constitute the core of the north Eurasian avian community. Higher values of the MCH for the group A studies demonstrate strong sampling bias where the populations from eastern Eurasia remain unsampled. Nevertheless, it can be seen, that southeastern clades and clades mainly distributed in the eastern part of the sampled range consistently have lower frequency of the MCH than western clades, suggesting their older age (Fig. 5.8 B). Both groups of studies (A and B) also indicate a possibility of survival of some species in Eastern Europe and Asia during the Pleistocene climatic oscillations. Pleistocene history of Eurasian avifauna.—Although most phylogeographic splits occurred earlier than the last glacial maximum, it was interesting to test if the distribution of refugial forests and grasslands during last Pleistocene glaciation explain the distribution of southern Eurasian areas of phylogeographic endemism. When overlaid on palaeo-reconstruction of vegetation during maximum cooling of the Pleistocene (Frenzel et al. 1992; Adams 2002), the three out of four areas of phylogeographic endemism (the Caucasus, southern Asia and southeastern Asia) were concordant with the Pleistocene forest and grassland refugia (Fig. 5.7). Some endemic clades from the Caucasus probably survived the last glaciation south of the Black Sea, whereas southern and southeastern regions of endemism are on the border of the Pleistocene grasslands and forests. However, the central Asia region is located in the middle of an extreme Pleistocene desert and, therefore, could not play the role of a refugial site. This suggests that the central Asia 149
region is a recently established area of endemism. Three species (species complexes) displayed endemic clusters of haplotype in central Asia. Both tit species, Parus songarus (willow tit complex) and P. bokharensis (great tit complex), have a limited distribution in this region and appear as sister groups to widespread Eurasian P. montanus and P. major, respectively (Fig. 5.4A, C). The central Asian clade of the greenish warbler, the third group endemic for the region, appears as a sister to the clade from western Eurasia (Fig. 5.2B). Concordant northern distributions of sister-lineages to central Asian clades in these three species suggest that central Asian clades might have been recently evolved after colonization from more northern Eurasian populations. These populations could either have survived Pleistocene cooling in treeless steppe-tundra or have colonized northern Eurasia from European refugia. This conclusion is contradictory to recent northward spread of the greenish warblers around the Tibetan Plateau suggested by Irwin et al. (2001), and offers a vicariant, rather than stochastic explanation for the split of the two major lineages of greenish warbler. A similar pattern of recent southward dispersal through central Asia was inferred for the common rosefinch on the basis of asymmetric migration estimates and the southward decrease in nucleotide diversity (Pavlova et al. unpublished). Thus, the mountains of central Asia are probably a region of ongoing differentiation, whereas the other three areas of phylogeographic endemism (Caucasus, Southern Asia and southeastern Asia) likely have played a major role in harboring refugial populations for many species of Eurasian birds during maximum Pleistocene cooling. Allopatric divergence in multiple glacial refugia satisfactorily explains phylogeographic patterns of many species. Southern glacial refugia in the European peninsula explain patterns of genetic variation in some European temperate species (Taberlet et al. 1998, Griswald and Baker 2002, Kvist et al. 2003a). For example, the northward decrease of nucleotide diversity in greenfinch is consistent with postglacial rapid recolonization of Europe (Merila et al. 1997). Southeastern Asia was shown to be an important area of phylogeographic endemism for some small mammals (e.g. Galbreath and Cook 2004). For a diverse range of species distinct phylogenetic clades are described 150
from Beringia (Hewitt 2004). For some Eurasian birds several refugial populations might have been a source for postglacial colonization. For example, white wagtail might have survived in high numbers in southern and southeastern refugia and in a small population in Beringia, and recolonized northern Eurasia from northeastern refugia. This scenario explains the low genetic diversity in northern (π range 0-0.0015) and high in southern (0.0033-0.0049) populations, existence of two areas of secondary contact between three clades (Pavlova et al. in review). Vicariant speciation in multiple refugia also explains several phylogeographic splits in Eurasian nuthatch, winter wren and other species, while low genetic diversity and absence of phylogeographic structure in the great spotted and threetoed woodpeckers and the great tit are consistent with postglacial expansion from a single refugium. Northern populations of several species, however, possess substantial genetic diversity (e.g. in common rosefinch, yellow wagtail, willow tit, bluethroat, hazel grouse, common chaffinch) (Fig. 5.8), indicating that they occupied their range long enough to accumulate mutations. Similarly, phylogeographic study of lemmings (Fedorov et al. 2003) found no support for a refugium in Beringia as a source of postglacial colonization, suggesting a possibility of survival for northwestern populations. Reconstructions based on pollen data suggest that large areas of eastern Europe, Central Siberia, and Kamchatka were covered with steppe-tundra (Frenzel et al. 1992), which may have provided suitable habitats for some species. The plant species responded differently to the Quaternary Ice Ages, and some survived near the edges of European ice-caps (Willis and Niklas 2004). Late Pleistocene polar, mountain and temperate deserts stretched from northern Siberia to the southern mountains (Adams 2002, Fig. 5.7), isolating eastern Asia from western Asia and Europe. A narrow strip of desert might have further separated southeastern and northeastern Asia (which at that time was connected to North America by a land bridge). It is possible therefore, that some avian species survived Pleistocene cooling in northern Eurasian tundra, being vicariantly separated by the extreme desert from their southeastern and northeastern conspecifics. Therefore, large populations might have existed in 151
northwestern, northeastern and southeastern Eurasia during the Pleistocene. This historical scenario explains phylogeographic patterns of several birds: the existence of three lineages in yellow and citrine wagtails and relatively high genetic diversity in their western and southeastern clades, high genetic diversity in northern populations of common rosefinch, bluethroat and willow tit. Age of divergence and rates of molecular evolution.—Despite extremely high stochasticity of coalescence processes (Hudson 1990) and huge variability in the evolutionary rates, sequence divergence is often used in conjunction with the rate of molecular evolution to estimate the time since most recent common ancestor (TMRCA) for groups of individuals or clades. In this way the age of isolation and/or speciation events may be estimated. Here I show why such direct and convenient translation of sequence divergence into evolutionary dates should be avoided. Not only different genes, but also parts of the same mtDNA gene, can evolve with different rates. Comparing the trees A and B in figure 4 it is tempting to conclude that group A is older than group B. Yet these two trees estimate relationships among the same group of species of the great tit complex. The depth of the haplotype tree A (1.38% estimated from the data of Kvist et al. (2003a)) was based on 578bp of the CR-I and was almost three times higher than the depth of tree B (0.48%), which was estimated from 1667bp of two mitochondrial gene regions (ND2 and the CR-II, Pavlova et al. unpublished). Similarly, for the willow tit the estimated depth of the tree from CR-I (2.37%, calculated from the data of Kvist et al. 2001) was 2.8 times higher than the one estimated from the CR-II and ND2 (0.85%, Pavlova et al. unpublished) (Fig. 5.4 C, D). The divergence of the Caucasian clades from their sister clades in five Eurasian species were: 0.41% in common Rosefinch, 1.1% in white wagtail, 3.3% in Eurasian nuthatch, 2.8% in greenish warbler and winter wren, with a mean of 2.04 (Figs 1-6). Although it is tempting to interpret the variance around the mean as different times of population isolation, this variance could also be caused by stochastic differences in the depth of the tree caused by genetic drift (Knowles 2004, Edwards and Beerli 2000), difference in the rates of evolution among the species being compared and difference in evolutionary rates in sequenced gene regions. Therefore, even 152
if the same the sequenced gene regions are used for comparisons between taxa, the other sources of variance should be accounted for before estimating the age of splits. The depth of the trees simulated under the same model of population divergence shown on Fig. 5.1 of Knowles (2004) differed by a factor of 3 for older population divergence and by a factor of 6 for more recent ones. The sequence divergences for Caucasian clades of five Eurasian species differed by a factor of 8. Even if, with other sources of information (such as the structure of the haplotype tree and southward decrease in nucleotide diversity inconsistent with older Pleistocene colonization), a younger age of colonization for the rosefinch can be distinguished from the older events in the other species, the null hypothesis of simultaneous colonization of the Caucasus by wagtail, nuthatch, warbler and wren cannot be rejected. Hence, inferring the relative timing of the evolutionary events by comparing the sequence divergence among taxa is challenging and must be accompanied by the estimation of the confidence intervals (Graur and Martin 2004). It was suggested that stochasticity of the coalescent can be accounted for only if multiple loci are studied (Arbogast et al. 2002, Knowles 2004). However, using external sources of information, some inferences about relative age of splits can be made. Considering current species’ ranges (Figs. 5.1A, 5.2A, 5.B, 5.5A) and the likely contractions of species’ historical ranges (Fig. 5.7), concordant splits of Caucasian lineages in wagtail, nuthatch, warbler and wren can be attributed to divergence in allopatric refugia with a certain confidence. What cannot be concluded is whether these splits have been caused by the same glacial event or by different ones. The rates of mtDNA evolution used in different studies of Eurasian birds varied greatly not only among different gene regions, but also for the CR itself. The rate of 1.6% per million years (MY) estimated by Fleischer et al. (1998) for honeycreepers was used for dating the divergence of cyt b sequences in tits and ravens (Salzburger et al. 2002a, Omland et al. 2000). The rate of 5.5% per MY calibrated for ND2 of Galapagos mockingbirds (Arbogast et al. unpublished data) was used for ND2 sequences of winter wrens (Drovetski et al. 2004). A wide range of rates has been used for the CR sequences: the rate of 2% per MY estimated by Shields and Wilson (1987) for geese was used for 153
dating the splits in small passerines and grey partridge (Kvist et al. 1999, LiukkonenAnttila et al. 2002); a rate of 5% per MY estimated for complete CR of Darwin’s finches (Freeland and Boug 1999) was used for complete CR of the rock ptarmigan (Holder et a.l 2000); rate of 8.5 % per MY calibrated for CR by Liebers and Hebig (2002) from the rate of 1.6% per MY for cyt b was used for dating divergence in the gulls (Liebers and Helbig 2002); a 14.8% per MY rate (calibrated for complete control region from 20.8% per MY for CR-I, estimated by Quinn 1992 for the snow geese) was used in dunlin (Wenink et al. 1996); rates of 20% and 20.8% per MY were used for the bearded vulture (Godoy et al. 2004) and the hazel grouse (Baba et al. 2002), respectively. Therefore depending on the assumptions for the rates of molecular evolution, the estimated age of the same split for the same haplotype tree can vary by a factor of ten. Considering all sources of variance around sequence divergence, variance in molecular evolution in different species, and the variance of the rate estimates, for now it might not be possible to reliably estimate the evolutionary dates. The dates of divergence estimated in phylogeographic literature should not be simply compared among the species without considering all sources of errors around the estimates. The problem of introgression and selection in phylogeographic studies.—Ballard and Whitlock (2004) reviewed arguments against using mtDNA sequences as a reliable basis for inferring evolutionary history of species. In particular, they explained how introgression and selection on mitochondria can become sources of significant bias and emphasized the necessity of using more genetic markers before inferences about species demography can be made. As was pointed out by Ballard and Whitlock (2004), “introgression and incomplete lineage sorting are not barriers to inferences, but are the object of inference themselves”. In four of 28 surveyed species (yellow and citrine wagtails, common raven and blue tit), multiple mitochondrial lineages did not form a monophyletic group that matched current species level taxonomy. Odeen and Bjorklund (2003) inferred the ancient introgression of mtDNA for the yellow and citrine wagtails, because paraphyly of yellow wagtail was supported by a nuclear tree (Z- linked CHD1Z intron), but paraphyly of citrine wagtail was not. If this were a true evolutionary scenario, 154
the inference about evolutionary relationships among mtDNA lineages would not be congruent with the true history of the two species. However, it remains true that all lineages of both wagtails had independent histories for a long time (at least 2Nef generations, where Nef is the effective number of females in a population), which allowed sorting of mitochondrial DNA lineages. Introgression should not affect phylogeographic inferences within mtDNA-defined clades because these clades correspond to groups of geographically isolated or phenotypically distinguishable individuals. In most studies, significant values of Tajima’s D and Fu’s Fs- values were detected for multiple populations. Although these tests may indicate selection, and potentially bias or invalidate the inferences based on coalescence or neutral theory, significantly negative values of Tajima’s D and Fu’s Fs can also characterize expanding populations. It was demonstrated for the great tit and the willow tits (Zink unpublished data) that selection did not greatly affect phylogeographic inferences (Pavlova et al. unpublished). Importance of adequate geographic sampling.—A strong correlation between the number of clades detected per species and the number of sampling localities underlines the importance of complete range sampling for detecting all phylogeographic structure. In many studies just the European part of a species’ ranges were sampled rigorously (but see Baba et al. 2002). Although such sets of data can reveal the absence or existence of phylogeographic structure, no firm conclusions can be made unless additional localities are studied. For example, if phylogeographic structure is found, it is still unclear where the exact location of a split is and how the clades are sorted in the intermediate populations. Yet, many such studies make inferences about species’ evolutionary history. With a single exception (Irwin et al. 2001), no studies of Eurasian birds have sufficiently sampled populations from southern and southeastern Asia. Nevertheless, the study on the greenish warbler (Irwin et al. 2001) showed that this geographic region can be rich in distinct clusters of haplotypes, probably due to the geographic barriers preventing or restricting gene flow between populations. In all studies that examined individuals from 155
southern or southeastern Asia, high genetic divergence from more northerly conspecifics was detected. It is possible that mountain habitats facilitate long-term isolation and subsequent speciation and act as a source of new species in a region. To test this hypothesis more sampling from central and southern Asia is needed. Some studies use genetic data from migrating or wintering individuals identified to subspecies in the field and released (Odeen and Bjorklund 2003). Such samples do not add information to the analysis, but rather pose additional questions. Because the breeding sites of the specimens are unknown, the data are prone to several sources of bias. Subspecies are often based on subjective criteria. Secondly, due to the differential rates of morphological and molecular evolution, species’ phylogeographic structure often does not coincide with subspecific divisions. Incomplete linage sorting or secondary contact between lineages may result in two morphologically identical individuals that possess haplotypes from different clades, whereas the same haplotype can be shared by two or more subspecies (e.g. in white wagtail, Pavlova et al. unpublished). Thus, using specimens sampled away from breeding sites, and not vouchered in museums, is unwise. In conclusion, it is very important to compare morphological features of the organisms with their phylogeographic structure, but one needs to avoid assuming that a subspecies is a monophyletic unit when choosing the sampling localities or analyzing data.
CONCLUSIONS The Eurasian avifauna is a dynamic species assemblage; however, some core species appear to have been stable over long periods of time. Avian species responded differently to the Pleistocene climatic cycles. Although some species were displaced into southern and/or eastern refugia and went through a bottleneck with consecutive range expansion during climate warming, others survived periods of cooling in Eurasian steppetundra. Areas of phylogeographic endemism are not uniformly distributed around the continent: southern mountain regions harbor geographically sorted clades in many species, which possess high genetic diversity, whereas phylogeographic splits in north Eurasian taxa are scattered, indicating relatively recent colonization and ongoing 156
speciation in some species. Southeast-Asian lineages displayed higher genetic diversity, indicating that region was relatively stable during the Ice Ages. As was found for North American birds (Zink 1996), the Eurasian avifauna is comprised of species with different evolutionary histories.
ACKNOWLEDGEMENTS I thank R. Zink for providing me unpublished sequence data and insightful discussions. I also thank A. Kessen and A. Jones for useful comments and A. Polesskiy for help in manuscript preparation.
LITERATURE CITED Adams, J. M. 2002. Global land environments since the last interglacial. See http://members.cox.net/quaternary/. Alstrom, P. and Mild, K. 2003. Pipits and Wagtails. Princeton Univ. Press. Princeton. Arbogast, B. and Kenagy, G. J. 2001. Comparative phylogeography as an integrative approach to historical biogeography. J. Biogeogr. 28:819-825. Arbogast , B. S., Edwards, S. V., Wakeley, J., Beerli, P. and Slowinski, J. B. 2002. Estimating divergence times from molecular data on phylogenetic and population genetic time scales. Annu. Rev. Ecol. Syst. 33:707-740. Avise, J. C., Arnold, J., Ball, R. M., Bermingham, E., Lamb, T., Neigel, J. E., Reeb, C. A. and Saunders, N. C. 1987, Intraspecific phylogeography: The mitochondrial DNA bridge between population genetics and systematics. Ann. Rev. Ecol. Syst. 18:489522. Baba, Y., Fujimaki, Y., Klaus, S., Butorina, O., Drovetskii, S. and Koike, H. 2002. Molecular population phylogeny of the hazel grouse Bonasa bonasia in East Asia inferred from mitochondrial control-region sequences. Wildlife Biol. 8:251-259. Ballard, J. W. O. and Whitlock, M. C. 2004. The incomplete natural history of mitochondria. Mol. Ecol. 13:729-744. 157
Beerli, P. and Felsenstein, J. 1999. Maximum-likelihood estimation of a migration rates and effective population numbers in two populations using a coalescent approach. Genetics. 152:763-773. Bensch, S. and Hasselquist, D. 1999. Phylogeographic population structure of great reed warblers: an analysis of mtDNA control region sequences. Biol. J. Linn. Soc. 66:171185. Bermingham, E. and Moritz, C. 1998. Comparative phylogeography: concepts and applications. Mol. Ecol. 7:367-369. Cramp, S. 1988. The birds of the Western Palearctic. Vol. 5, Oxford University Press, Oxford. Cramp, S. and Perrins, C.M. 1993. The birds of the Western Palearctic. Vol. 7, Oxford University Press, Oxford. ---. 1994. The birds of the Western Palearctic. Vol. 8, Oxford University Press, Oxford. Cruzan, M. B. and Templeton, A. R. 2000. Paleoecology and coalescence: phylogeographic analysis of hypotheses from the fossil record. TREE. 15:491-496. Drovetski, S. V., Zink, R. M., Rohwer, S., Fadeev, I. V., Nesterov, E. V., Karagodin, I., Koblik, E. A. and Red’kin, Y. A. 2004. Complex biogeographic history of a Holarctic passerine. Proc. R. Soc. Lond. B. 271:545-551. Dynesius, M. and Jansson, R. 2000. Evolutionary consequences of changes in species’ geographical distributions driven by Milankovitch climate oscillations. P. Natl. Acad. Sci. USA 97:9115-9120. Edwards, S. V. and Beerli, P. 2000. Perspective: genedivergence, population divergence, and the variance in coalescence time in phylogeographic studies. Evolution 54:18391854. Fedorov, V. B., Goropashnaya, A. V., Jaarola, M. and Cook, A. 2003. Phylogeography of lemmings (Lemmus): no evidence for postglacial colonization of Arctic from the Beringian refugium. Mol. Ecol. 12:725-731. Fleischer, R. C., McIntosh, C. E., and Tarr, C. L. 1998. Evolution on a volcanic conveyor belt: Using phylogenetic reconstructions and K-Ar-based ages of Hawaiian Islands to 158
estimate molecular evolutionary rates. Mol. Ecol. 7:533-545. Flint, V. E., Boehme, R. L., Kostin, Y. V., and Kuznetsov A. A. 1984. Birds of the USSR. Princeton Univ. Press, Princeton. Freeland, J. R., Boag, P. T. 1999. Phylogenetics of Darwin’s finches: paraphyly in the tree-finches, and two divergent lineages in the warbler finch. Auk 116:577-588. Frenzel, B., Pecsi, M. and Velichko, A.A. [eds.] 1992. Atlas of paleoclimates and paleoenvironments of the northern Hemisphere. Hungarian Academy of Sciences, Budapest. Galbreath, K. E. and Cook, J. A. 2004. Genetic consequences of Pleistocene glaciations for the tundra vole (Microtus oeconomus) in Beringia. Mol. Ecol. 13:135-148. Gay, L., Du Rau, D., Mondian-Monval, J.-Y. and Crochet, P.-A. 2004. Phylogeography of a game species: the red-crested pochard (Netta rufina) and consequences for its management. Mol. Ecol. 13:1035-1045. Godoy J. A., Negro, J. J., Hiraldo, F. and Donazar, J. A. 2004. Phylogeography, genetic structure and diversity in the endangered bearded vulture (Gypaetus barbatus, L.) as revealed by mitochondrial DNA. Mol. Ecol. 13:371-390. Graur, D. and Martin, W. 2004. Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Trends. Genet. 20:80-86. Griswold, C. K. and Baker, A. J. 2002. Time to the most recent ancestor and divergence times of populations of common chaffinches (Fringilla coelebs) in Europe and North Africa: insights into Pleistocene refugia and current levels of migration. Evolution. 56:143-153. Harpending, H. C., Batzer, M. A., Gurven, M., Jorde, L. B., Rogers, A. R. and Sherry, S. T. 1998. Genetic traces of ancient demography. Proc. Natl. Acad. Sci. USA 95:19611967. Harrap, S., Quinn, D. 1995. Chickadees, Tits, Nuthatches, and Treecreepers. Princeton University Press, Princeton. Hewitt, G. M. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Biol. J. Linn. Soc. 58:247-276. 159
---. 1999. Post-glacial re-colonization of European biota. Biol. J. Linn. Soc. 68:87-112. ---. 2000. The genetic legacy of the Quaternary ice ages. Nature. 405:907-913. ---. 2004. Genetic consequences of climatic oscillations in the Quaternary. Proc. R. Soc. Lond. B. 359:183-195. Holder, K., Montgomerie, R. and Friesen, V. L. 1999. A test of the glacial refugium hypothesis using patterns of mitochondrial and nuclear DNA sequence variation in rock ptarmigan (Lagopus mutus). Evolution. 53:1936-1950. ---. 2000. Glacial vicariance and historical biogeography of rock ptarmigan (Lagopus mutus) in the Bering region. Mol. Ecol. 9:1265-1278. Hudson, R. 1990. Gene genealogies and the coalescent process. Oxford Surveys in Evol. Biol. 7:1-44. Huntley, B. E. and Birks, H. J. B. 1983. An atlas of past and present pollen maps for Europe 0-13 000 years ago. Cambridge Univ. Press. Irwin, D. E., Bensch, S. and Price, T. D. 2001. Speciation in a ring. Nature. 409:333-337. Irwin, D. E. 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution. 56:2383-2394. Knowles, L. L. 2004. The burgeoning field of statistical phylogeography. J. Evol. Biol. 17:1-10. Knowles, L. L., Maddison, W. P. 2002. Statistical phylogeography. Mol. Ecol. 11:26232635. Kryukov, A. P. and Suzuki, H. 2000. Phylogeography of carrion, hooded, and jungle crows (Aves, Corvidae) inferred from partial sequencing of the mitochondrial DNA cytochrome b gene. Russ. J. Genet+. 36:922-929. Kvist, L., Ruokonen, M., Lumme, J. and Orell, M. 1999. Different population structures in northern and southern populations of the European blue tit (Parus caeruleus). J. Evol. Biol. 12:798-805. Kvist, L., Martens, J., Ahola, A. and Orell, M. 2001. Phylogeography of a Palaearctic sedentary passerine, the willow tit (Parus montanus). J. Evol. Biol. 14:930-941. Kvist, L., Martens, J., Higuchi, H., Nazarenko, A. A., Valchuk, O. P. and Orell, M. 160
2003a. Evolution and genetic structure of the great tit (Parus major) complex. Proc. R. Soc. Lond. B. 270: 1447-1454. Kvist, L., Martens, J., Nazarenko, A. A. and Orell, M. 2003b. Paternal leakage of mitochondrial DNA in the great tit (Parus major). Mol. Biol. Evol. 20:243-247. Liebers, D. and Helbig, A. J. 2002. Phylogeography and colonization history of Lesser Black-backed Gulls (Larus fuscus) as revealed by mtDNA sequences. J. Evol. Biol. 15: 1021-1033. Liukkonen-Anttila, T., Uimaniemi, L., Orell, M. and Lumme, J. 2002. Mitochondrial DNA variation and the phylogeography of the grey partridge (Perdix perdix) in Europe: from Pleistocene history to present day populations. J. Evol. Biol. 15:971-982. McDonald, J. H. and M. Kreitman. 1991. Adaptive protein evolution at the Adh locus in Drosophila. Nature 351:652-654. Merila, J., Bjorklund, M. and Baker, A. J. 1997. Historical demography and present day population structure of the greenfinch, Carduelis chloris – an analysis of mtDNA control-region sequences. Evolution. 51:946-956. Neigel, J. E. 2002. Is Fst obsolete? Conserv. Genet. 3:167-173. Odeen, A. and Bjorklund, M. 2003. Dynamics in the evolution of sexual traits: losses and gains, radiation and convergence in yellow wagtails (Motacilla flava). Mol. Ecol. 12:2113-2130. Omland, K. E., Tarr, C. L., Boarman, W. I., Marzluff, J. M. and Fleischer, R. C. 2000. Cryptic genetic variation and paraphyly in ravens. Proc. R. Soc. Lond. B. 267:24752482. Pavlova, A., Zink, R. M., Drovetski, S. V., Red’kin, Y. and Rohwer, S. 2003. Phylogeographic patterns in Motacilla flava and Motacilla citreola: species limits and population history. Auk. 120:744-758. Quinn, T. W. 1992. The genetic legacy of Mother Goose: Phylogeographic patterns of Lesser Snow Goose Chen caerulecens caerulescens maternal lineages. Mol. Ecol. 1:105-117. Randi, E., Tabarroni, C., Rimondi, S., Lucchini, V. and Sfougaris, A. 2003. 161
Phylogeography of the rock partridge (Alectoris graeca). Mol. Ecol. 12:2201-2214. Ray, N., Currat, M. and Excoffier, L. 2003 Intra-deme molecular diversity in specially expanding populations. Mole. Biol. Evol. 20:76-86. Rogers, A. R. 1995. Genetic evidence for a Pleistocene population explosion. Evolution. 49:608-615. Salzburger, W., Martens, J., Nazarenko, A. A., Sun, Y-H., Dallinger, R. and Sturmbauer, C. 2002a. Phylogeography of the Eurasian Willow Tit (Parus montanus) based on DNA sequences of the mitochondrial cytochrome b gene. Mol. Phylogenet. Evol. 24:26-34. Salzburger, W., Martens, J. and Sturmbauer, C. 2002b. Paraphyly of the Blue Tit (Parus caeruleus) suggested from cytochrome b sequences. Mol. Phylogenet. Evol. 24:19-25. Seddon, J. M., Santucci, F., Reeve, N. and Hewitt, G. M. 2002. Caucasus Mountains divide postulated postglacial colonization routes in the white-breasted hedgehog, Erinaceus concolor. J. Evol. Biol. 15:463-468. Shields, G. F., and Wilson, A. c. 1987. Calibration of mitochondrial DNA evolution in geese. J. Mol. Evol. 24:212-217. Stewart, J. R. and Lister, A. M. 2001. Cryptic northern refugia and the origins of the modern biota. Trends Ecol. Evol. 16:608-613. Taberlet, P., Fumagalli, L., Wust-Saucy, A.G., Cosson, J.F. 1998. Comparative phylogeography and postglacial colonization routes in Europe. Mol Ecol. 7:453-64. Tribsch, A. and Schonswetter, P. 2003. Patterns of endemism and comparative phylogeography confirm palaeo-environmental evidence for Pleistocene refugia in the Eastern Alps. Taxon. 52:477-497. Uimaniemi, L., Orell, M., Kvist, L., Jokimaki, J. and Lumme, J. 2003. Genetic variation of the Siberian tit Parus cinctus populations at the regional level: a mitochondrial sequence analysis. Ecography. 26:98-106. Wenink, P. W., Baker, A. J., Rosner, H-U. Tilanus, M. G. J. 1996. Global mitochondrial DNA phylogeography of holarctic breeding dunlins (Calidris alpina). Evolution. 50:318-330. 162
Willis, K. J. and Niklas, K. J. 2004. The role of Quaternary environmental change in plant macroevolution: the exception of the rule? Proc. R. Soc. Lond. B. 359:159-172. Winkler, H. Christie, D. A. and Nurney, D. 1995. Woodpeckers: A guide to the woodpeckers of the world. Houghton Mifflin Company, Boston. Wright, S. 1951. The genetical structure of populations. Ann. Eugen. 15:323-354. Zink, R. M. 1996. Comparative phylogeography in North American birds. Evolution. 50:308-317. ---. 2002. Methods in comparative phylogeography, and their application to studying evolution in the North American aridlands. Integr. Comp. Biol. 42:953-959. Zink, R. M., Drovetski, S. V. and Rohwer, S. 2002a. Phylogeographic patterns in the great spotted woodpecker Dendrocopos major across Eurasia. J. Avian Biol. 33:175178. Zink, R. M., Rohwer, S., Drovetski, S., Blackwell-Rago, R. C. and Farrell, S. L. 2002b. Holarctic phylogeography and species limits of three-toed woodpeckers. Condor. 104:167-170. Zink, R. M., Drovetski, S. V., Questiau, S., Fadeev, I. V., Nesterov, E. V., Westberg, M. C. and Rohwer S. 2003. Recent evolutionary history of the bluethroat (Luscinia svecica) across Eurasia. Mol. Ecol. 12:3069-3075.
163
TABLE 5.1. Summary of findings from major phylogeographic studies of widely distributed Eurasian birds. Phylogeographic categories roughly correspond to those of Avise et al. (1987) (see results section). References for geographic regions: N- northern, S-southern, Eeastern, W- western etc, C- central. Nu- number of individuals, Nhaps – number of haplotypes, fmc- frequency of the most common
Species (species complex)
Great spotted woodpecker
Phylog. category
haplotype in the clade; underlined values are total for all clades. Seq. diverg.- maximum sequence divergence between individuals.
I (IV)
MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
region 2 clades: Eurasia except SE;
67/ 26 61 /21/ 0.48
Dendrocopos
Three-toed woodpecker
2.9%
Japan
3%3
Fig. 5.1a
I
2 clades:
30/ 17
9 Eurasian and 3 N
(IV)
Eurasia;
21/9/ 0.62
American populations
N America
9/8
1
from Eurasia and
6/ 5
Picoides tridactylus
17 populations
SE Eurasia
major
Source
Fig. 5.1b
1369 bp ND2, ND3, cyt b
1234bp 3.8%1
ND2, ND3, cyt b
Yes: concordant
Zink et al.
subspecific
2002a
division Some: clades distributed
Zink et al.
over different
2002b
continents
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued.
Eurasian nuthatch
I
MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
region 3 clades:
134/ 64
W Europe;
20/ 9/ 0.35
Caucasus;
11/ 4/ 0.45
E and NE Eurasia
Troglodytes troglodytes
Some:
21 Eurasian populations including UK and
1551bp 4.5%1
Japan
Sitta europaea
Winter wren I
Source
103/ 51/ 0.31
ND2, CR-II
Fig. 5.2a
6 clades:
97/ 51
13 populations
W Nearctic;
21/ 13/ 0.38
from N. America
E Nearctic;
27/ 13/ 0.44
and Aleutian
E Asia;
21/ 12/ 0.33
Islands,
Nepal;
1/1
11 populations
Caucasus;
10/ 7/ 0.4
from Eurasia,
Europe
17/ 5/ 0.76
Sakhalin and Japan Fig. 5.2b.
clades correspond to
Zink et al.
subspecies or
unpublished
subspecies groups
5.9%1 2
8%
1041bp
Drovetski et
ND2
al. 2004
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued.
I Yellow wagtail
MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
region species paraphyletic, 3 clades:
167/ 117
populations,
W Eurasia;
90/ 69/ 0.07
1 locality from
(III)
NE Eurasia and Alaska;
57/ 31/ 0.27
Alaska
SE Eurasia.
20/ 17/ 0.15
Fig. 5.4a
Motacilla flava
species paraphyletic, 3 clades:
complex: Motacilla flava
16 Eurasian
(III)
Yellow wagtail
Source
Europe and N Africa; I
104
Pavlova et al.
ND2, ND3,
2003
CR-II
populations from
97/ 19/ 0.29
Europe, Africa and
and 0.28
China, also several
wintering wintering birds)
6.2%
7 breeding
localities of 2 from SE Asia (one is
993 bp 1
individuals identified to subspecies
Some: 2
5.5%
350 bp
2 clades (W and
CR-I
E Palearctic) on nuclear tree
Odeen and Bjorklund 2003
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region
Willow tit
3 clades:
species
Europe and SE Asia
complex Parus
(montanus); I
surveyed area
C Asia (songarus);
105/ 71
Yes: clades
94/ 62/ 0.12 3/ 2
13 populations in Eurasia and Japan
2
4%
Fig. 5.6c
montanus, P. songarus,
China (affinis)
592 bp CR-I
4 clades:
species
Eurasia including Japan
complex
(montanus);
P. weigoldicus
behavior,
Kvist et al. 2001
I
34/ 14 25/ 9
14 localities in
C Asia (songarus);
1/ 1
Europe, C and SE
SE Asia (affinis);
6/ 3
Asia, including Japan
P. songarus P. affinis
species, based on
acoustics
Willow tit
montanus
correspond to
morphology, and
8/ 7
P. affinis
Parus
Source
S Asia (weigoldicus)
2/ 1
6%2
306 bp
Salzburger et
cyt b
al. 2002a
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
Source
region 6 clades with contact in SE Asia and
149/ 101
C Siberia: Greenish warbler Phylloscopus
II
trochiloides
Caucasus;
2/ 2
28 Eurasian
NW Eurasia;
35/ 20/ 0.31
populations
C Asia;
33/ 24/ 0.21
Fig. 5.3.
S Asia;
37/ 23/ 0.27
SE Asia;
5/ 4
NE Asia
37/ 26/ 0.19
Yes: 5.4%1
1200 bp
microsatellites,
Irwin et al.
ND6, CR
morphology,
2001
acoustics
species paraphyletic, 2 clades with contact in
Citrine wagtail Motacilla citreola
II
38/ 23
C Asia:
7 Eurasian populations
W Eurasia ;
13/ 9/ 0.23
N and E Eurasia
25/ 14/ 0.32
Fig. 5.4b
1407 bp, 2%1
ND2, ND3, CR-II
Pavlova et al. 2003
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
Source
region 2 breeding No:
Citrine wagtail
species paraphyletic,
localities: Poland
Motacilla
2 clades:
and India, one
>2000 bp
on nuclear tree
citreola
Europe;
migrating
ND2, CR
species appear
4
India (M. c. calcarata)
monophyletic
individual from
Odeen and Bjorklund 2003
China 4 clades with contact in C Asia and Primor’e:
Great tit
Europe and N Asia
species
(major group);
complex Parus major P. minor
II
125/ 57/ 0.4
clades correspond to
22 Eurasian and
species,
1 African
C Asia (minor);
44/ 29/ 0.16
SE Asia (bokharensis);
3/ 3
P. bokharensis P. cinerea
Yes:
175/ 92
populations Fig. 5.6a
6%2
578 bp
recognized by
Kvist et al.
CR-I
some authors
2003a
based on behavior,
Nepal (cinerea)
3/ 3
morphology, and acoustics
16
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
surveyed area
Source
region 15 Eurasian no structure
Great tit Parus major
84/34/ 0.5
Fig 6b
IV outgroup: Primor’e P. minor
29/ 21/ 0.1
Willow tit Parus
IV
no structure
139 /99/ 0.12
Caucasus and Primor’e:
II
populations
0.5%
ND2,
1%
CR-II
unpublished
0.5%1 0.9%
1761bp
Pavlova et al.
ND2,
unpublished
CR-II
splits in 27 Eurasian
C, S and SE Asia;
9/ 8/ 0.22
localities Fig. 5.5a
203/ 68/ 0.52
Pavlova et al.
No:
20/ 11/ 0.3
C Asia
1667 bp
232/ 87
Caucasus;
Eurasia except
0.2%1
Fig 6b
Fig. 5.6d 3 clades with contact in
Motacilla alba
1 population
17 Eurasian
montanus
White wagtail
populations
1.1%1 1.3%
1477 bp
morphological
ND2,
subspecies are
CR-II
incongruent with mtDNA tree splits
Pavlova et al. unpublished
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 3 incompletely sorted clades:
Common reed bunting Emberiza
Eurasia except II
schoeniclus
Carpodacus
Kamchatka;
III
SE Eurasia;
10/ 9/ 0.2
S Eurasia
9/ 6/ 0.33
populations
1.3%1
Fig. 5.5b
1041 bp
Zink et al.
ND2
unpublished
Some:
clades:
17 Eurasian 187/ 148/ 0.05
Caucasus;
populations Fig. 5.6a
1
0.7% 1.1%
1536 bp ND2, CR-II
mostly E Asia. 2 incompletely sorted
Bluethroat
svecica
13 Eurasian 48/ 17/ 0.46
incompletely sorted
erythrinus
Luscinia
67/ 32
little structure, 2 nested
Common rosefinch
surveyed area
clades: III
Source
Phylogeographic divisions
Pavlova et al.
coincide with
unpublished
some subspecific division
154/ 65
N Eurasia;
59/ 32/ 0.22
S Eurasia
95/ 33/ 0.43
21 Eurasian populations Fig. 5.6b
0.7%1 0.9%
1071 bp CR-II, cyt b
Zink et al. 2003
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 2 clades:
30/ 11 5 Eurasian and
Carrion and hooded crow
surveyed area
N Sakhalin and Eurasia I
Corvus corone
no Primor’e; S Sakhalin, Japan, Kunashir and Primor’e
21/ 7
5 island localities (Sakhalin,
9/ 4
2%
336 bp
Kryukov and
cyt b
Suzuki 2000
Kunashir, Japan)
species paraphyletic, II Common raven Corvus corax
2 clades with contact in
(IV)
72/ 39 for CR
34/ 22/ 0.35
Nearctic
38/ 17/ 0.26
Europe; II
4 Eurasian and 8 North American
Holarctic;
clades:
Calidris alpina
Some:
W Nearctic:
5 incompletely sorted
Dunlin
Source
3
5%
314 bp
differential
CR-I
nuclear
populations
microsatellite
Omland et al. 2000
allele frequency
155/ 39 85/ 20/ 0.52
C Siberia;
29/ 7/ 0.28
E Siberia;
3/ 3
Alaska;
22/ 4/ 0.45
Some: several
15 populations of Eurasia and North America
2.6%
608 bp
congruent
Wenink et al.
CR-I and II
subspecific
1996
divisions
17
Canada
16/ 5/ 0.38
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 2 clades with clinal
Bearded vulture Gypaetus
9 European populations, one
“western” Europe;
100/ 16/ 0.36
from C Asia,
“eastern” Europe, Asia
72/ 34/ 0.19
4 from Africa
4%3
228 bp
Godoy et al.
CR-I
2004
No:
Red-crested
Netta rufina
172/ 50
change of frequency: II
barbatus
pochard
surveyed area
Source
little structure with 2 III
nested clades, both
64/ 44/ 0.3
from W Europe
6 European, 1 C
450 bp
Asian populations
CR-I
microsatellites, introns Yes:
Gay et al. 2004
morphometric 3 incompletely sorted clades, 2 of them with Lesser blackbacked gull Larus fuscus
clinal change of II
272/43
22 populations from N Eurasia
frequency: “western” Eurasia;
162/ 17/ 0.72
“eastern” Eurasia;
106/ 24/ 0.62
E Siberia
3/ 2
from Iceland to Taimyr
1.4%
430 bp
Liebers and
CR-I
Helbig 2002
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 2 incompletely sorted
Hazel grouse Bonasa bonasia
surveyed area
clades:
174/ 62 6 populations from
II Eurasia;
40/ 19/ 0.2
Japan and Primor’e
134/ 43/ 0.07
Eurasia and Japan
1.1%
428 bp
Baba et al.
CR-I
2002
6 incompletely sorted
Rock ptarmigan Lagopus mutus
II
clades: Attu Island; C
26 populations
Aleutian Islands;
across N America
Newfoundland;
154/ 28/ 0.4
Iceland,; Siberia;
and Bering Sea
Some: 1.84% 3
0.64%
676 bp CR-I
region
Rock ptarmigan
(IV)
2 clades:
105/ 13
19 populations
Alaska and Siberia;
38/ 6/ 0.5
from Alaska (3), E Asia (4) and
Lagopus mutus Aleutian Islands
64/ 8/ 0.3 and 0.2
Aleutian Islands (12)
differential frequency of nuclear intron
Holder et al. 1999
haplotypes
Canadian Arctic II
Source
0.64%
1144 bp
Holder et al.
complete CR
2000
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued.
Rock partridge Alectoris
I
graeca
MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 2 clades:
323/ 22
multiple localities
Sicily;
26/ 4/ 0.65
in Alps, Apennines
S Europe
297/ 33/ 0.49
and Pyrenees
2 clades with clinal Grey partridge Perdix perdix
change of frequency: II
surveyed area
227/ 45
“western” W Europe;
164/ 30/ 0.7
“eastern” W Europe
63/ 15/ 0.29
3.1%2
431 bp
Yes:
Randi et al.
CR-I
microsatellites
2003
about 17 populations from Western Europe
3.6%2
390 bp
Liukkonen-
CR-I
Anttila et al.
(68 individuals are
Great reed
2 clades with clinal
warbler
change of frequency:
Acrocephalus
II
arundinaceus
106/ 33
“western” Europe;
76/ 23/ 0.49
“eastern” Europe
30/ 10/ 0.33
Parus caeruleus
I
clades: Europe; Africa and Canary Ils.
2002
farm birds) 6 European breeding localities, 1 African wintering
1.29%2
494 bp
2.2%
CR-II
4 European
species paraphyletic, 2 Blue tit
Source
2 from Canary Archipelago, 1 from Africa
Hasselquist 1999 Yes,
populations, 16/ 5
Bensch and
4.9%3
306 bp cyt b
congruent differences in morphology and vocalization.
Salzburger et al. 2002b
17
Species (species complex)
Phylog. category
TABLE 5.1. Continued. MtDNA
Nu/ Nhaps/ fmc
Number of
Seq.
Number of
Support from
phylogeographic
(per clade)
sampling sites,
diverg.
base pairs/
independent data
sequenced
set?
pattern
region 2 clades with contact in
Blue tit Parus caeruleus
Siberian tit Parus cinctus
E Spain: II
surveyed area
43/ 25 7 European
Barcelona;
2/ 2
Europe
41/ 23. 0.41
no structure
59/ 25/ 0.34
IV
no structure
194/ 18/ 0.61
chloris Common chaffinch Fringilla coelebs
1
Norway (N=56) and Yakutsk (N=3)
Greenfinch Carduelis
populations
1.7%
Finland plus IV
10 European populations
19 populations IV
Source
no structure
190/ 76/ 0.2
from Europe and North Africa
1.5%
996 bp
Kvist et al.
CR-I and II
1999
911 bp
Uimaniemi et
CR-I and II
al. 2003
637 bp
Merila et al.
CR-I and II
1997
275 bp
Griswald and
CR-I
Baker 2002
sequence divergence, calculated from the UPGMA trees (Figs. 5.1-5.6); 2 maximum divergence between clades; 3 average clade divergence.
17
Fig. 5.1. UPGMA dendrograms and geographical distribution of clades for (A) Eurasian nuthatch Sitta europaea (Zink et al. unpublished), (B) three-toed woodpecker Picoides tridactylus (Zink et al. 2002b) and (C) great spotted woodpecker Dendrocopos major (Zink et al. 2002a). Grey areas on the maps indicate species’ ranges (following Cramp and Perrins 1993 for nuthatch, Winkler et al. 1995 for woodpeckers; only the Eurasian part of the range is shown for three-toed woodpecker). Colors of the bars by the clades correspond to the colors of the sampling localities. Abbreviations for general localities for figures 1-7 are as in Table 5.1; Cau- Caucasus, Eur- Europe, Eus- Eurasia, As- Asia.
178
Figure 5.1
179
Fig. 5.2. UPGMA dendrograms and geographical distribution of clades for (A) winter wren Troglodytes troglodytes (Drovetski et al. 2004) and (B) greenish warbler Phylloscopus trochiloides (Irwin et al. 2001). Grey areas on the maps indicate species’ ranges following Cramp (1988), Flint et al. (1984) (winter wren; only Eurasian part of the range is shown) and Irwin et al. 2001 (greenish warbler). Colors of the bars by the clades correspond to the colors of the sampling localities. Pie diagrams indicate localities where more than one haplotype was sampled.
Figure 5.2
A. Winter wren Troglodytes troglodytes
Cau Eup S As E As E Nearc W Nearc 7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
B. Greenish warbler Phylloscopus trochiloides
Cau NW Eus C As S As SE As NE As 7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Percent sequence divergence
180
Fig. 5.3. UPGMA dendrogram for the yellow (Motacilla flava), citrine (M. citreola) and white (M. alba) wagtails (Pavlova et al. 2003). The maps show geographical distribution of the clades for yellow (top) and citrine (bottom) wagtails (see Fig. 5.5 for M. alba). Dashed line on the tree and square on the map indicate the clade of M. citreola calcarata detected by Odeen and Bjorklund (2003); question mark indicates uncertainty of the node placement. Grey areas on the maps show the species’ ranges following Alstrom and Mild (2003). Large circles on the maps indicate samples of Pavlova et al. (2003), small circles indicate breeding samples of M. flava of Odeen and Bjorklund (2003). Colors of the bars by the clades correspond to the colors of the sampling localities. Pie diagram indicates locality where more than one haplotype was sampled.
Figure 5.3
Yellow wagtail Motacilla flava (f) Citrine wagtail M. citreola (c) f
SE As
c
W Eus c
E Eus
f
NE Eus
M. flava
M. alba f ?
7.0
6.0
5.0
4.0
3.0
W Eus
c
2.0
1.0
Percent sequence divergence
S As 0.0
M. citreola
181
Fig. 5.4 UPGMA dendrograms and geographical distribution of clades for great tit Parus major (A and B) and the willow tit P. montanus (C and D) species complexes, constructed from (A and C) CR-I sequences (Kvist et al. 2003a, 2001) (small circles on the maps) and ND2+CR-II (Pavlova et al. unpublished) (large circles on the map). Shaded areas on the maps show distribution of Parus major (top map) and P. montanus (bottom map) following Harrap and Quinn (1995). On the top map black circles indicate samples of P. minor, grey circles- P. bokharensis, and circle with hatching- P. cinereus. On the bottom map black circles indicate samples of the P. songarus, grey circles– P. affinis. Dashed line and the square on the map indicate a clade (single individual of P. (montanus) weigoldicus found by Salzburger et al. (2002a) from cyt b sequences; question mark indicates uncertainty of the node placement. Colors of the bars by the clades correspond to the colors of sampling localities. Pie diagrams indicate localities where more than one haplotype was sampled. Double-headed arrows indicate the difference in the sequence divergence estimates between two sets of data.
182
Figure 5.4
Great tit Parus major species complex
N Eus
A.
C As S As SE As B.
N Eus C As SE As
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Willow tit Parus montanus species complex
C.
N Eus C As S As
?
S As D. N Eus
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Percent sequence divergence
183
Fig. 5.5. UPGMA dendrogram and geographical distribution of clades for (A) white wagtail Motacilla alba (Pavlova et al. in review) and (B) common reed bunting Emberiza schoeniclus (Zink et al. unpublished). The grey areas on the maps show species’ ranges following Alstrom and Mild 2003 (white wagtail), Cramp and Perrins (1994) and Flint et al. (1984) (reed bunting).Colors of the bars by the clades correspond to the colors of sampling localities. Pie diagrams indicate localities where more than one haplotype was sampled.
Figure 5.5
184
Fig. 5.6. UPGMA dendrogram and geographical distribution of clades for (A) the common rosefinch Carpodacus erythrinus (Pavlova et al. unpublished) and (B) the bluethroat Luscinia svecica (Zink et al. 2003). Grey areas on the maps show species’ ranges (following Cramp and Perrins 1994 (rosefinch), Cramp 1988 and Flint et al. 1984 (bluethroat)). Colors of the squares by the clades correspond to the colors of the sampling localities. Pie diagrams indicate localities where more than one haplotype was sampled. Figure 5.6
185
Fig. 5.7. General areas of phylogeographic endemism and approximate limits of clade distribution for Eurasian birds overlaid on the reconstruction of vegetation during last glacial maximum (following Frenzel et al. 1992 and Adams 2002 reconstructions from pollen data). Shapes indicate general areas of phylogeographic endemism: CauCaucasus, C As- central Asia, S As- southern Asia, SE AS- southeastern Asia. Solid lines indicate splits between completely sorted lineages; dashed lines indicate approximate geographic division between geographically unsorted lineages. 1. Phylloscopus trochiloides, Troglodytes troglodytes, Sitta europaea, Motacilla alba, Carpodacus erythrinus. 2. Parus major complex (bokharensis), P. montanus complex (songarus), Phylloscopus trochiloides. 3. Parus major complex (cinereus), P. montanus complex (weigoldicus), Motacilla citreola, Phylloscopus trochiloides, Troglodytes troglodytes. 4. Motacilla flava, Troglodytes troglodytes, Parus montanus complex (affinis), (a) P. major complex (minor), (b) Dendrocopos major, (c) Bonasa bonasia, (d) Corvus corone, (e) Phylloscopus trochiloides, (f) Motacilla alba, (g) Emberiza schoeniclus. 5. Parus caeruleus. 6. Gypaetus barbatus. 7. Alectoris graeca. 8. Acrochephalus arundinaceus. 9. Perdix perdix. 10. Larus fuscus. 11. Luscinia svecica. 12. Sitta europaea, 13. Troglodytes troglodytes. 14. Motacilla flava. 15. M. citreola. 16. Calidris alpina. 17. Phylloscopus trochiloides. 18 Lagopus mutus. 19. Emberiza schoeniclus.
186
Figure 5.7
16 16 17
10
18
13 11
6
5
12 15
9
14
16
b d c
g
8
a
19
7 1
Cau
f 2
C As
17
15
3
4
e
SE As
S As
18
ice sheet
semi-desert
extreme desert and polar desert
steppe-tundra
grassland and forest
B. Frequency of the most common haplotype b mo onas nta ia S nu s N E As G. P. ia ba rba min Eura or sia tus S "E EA F. s co B. b " W ele on Eu ia a bs ra Eu sia E sia ro ur C. pe, N asia c C A o M. . alpi rax N frica ea flav na rc a E NC P. Sib tic pe urop e r rdi e x " , N ia L. mu E" W Africa tus Eu A. aru N. ru Aleu rope f nd t ina ina W ian I Eu s. P. ceus ra cin ctu "E" E sia sN uro G. ba C. c W E pe rba o tus rax urasi "W Hola a rc C. " W E tic al P. pina urasi a ma Ca jo na P. da cae r N rul Eur A. as eu aru nd C. a s Eu ia ina lpin rop ce us a Ala e A. "W ska g L. mu raec " Eur op aS tus e A E C. lask urop a alp e ina Sibe L. C. ch N E ria u f us cu loris rope s" Eu E" ro P. A. N Eu pe pe g r r a a L. rdix " eca sia f us W" S ic cu s " W E ily W" uro N E pe ura sia
P.
C. Frequency of the most common haplotype ery th r in M. fla us E va ur P. P. m W E asia mo u i n nta or S ra sia nu E P. M. s N E Asia tro f ur l a c E. hilo va S as ia ide E A sh oe s nic N E sia lus As M ia S P. tro . alba E As ch ia S L. iloide E As s s v M. e C ia A c it cic a re o N s ia Eu la M. r W P. flava Eu as tro r c h N E E as ia iloi P. d ura tro M. alb es S s ia ch iloi a C As d i S. es N auc a a eu W s u r op Eu s M E. . c itr a ea ras i eo sh Eu a la oe ra T. nic lu E Eu s ia tro s S ras gl ia E S. ody te ur a T. sia e s u tro r g lo opa e E As i d y L. tes a Eu a s S. v eci Cau rope ca eu ca r s E. opa e S Eu us r sh a a s C oe nic au c ia D. lus E as us P. ma jo ura si ma r jor Eu r a a N P . M . a E u s ia ra s tri lb T. dac ty a Eu ia tro r l g lo us E asia u dy tes ra sia Eu r op e
Fig. 5.8. The frequencies of the most common haplotypes for Eurasian clades arranged
from smaller to larger values for a group B of 13 avian species (A); and for a group A of
19 species with CR sequences (B).
Figure 5.8
A.
B. 0.8
0.7
0.6
0.7
Long sequences
0.5
0.4
0.3
0.2
0.1
0
High Ne Long-term residents Low Ne Recent colonizers
0.8
Control Region
0.6
0.5
0.4
0.3
0.2
0.1
0
188
Figure 5.9. Number of clades detected per species as a function of the number of sampling localities. Close geographic localities were lumped for a total of 11 for the data of Randi et al. (2003). The study of Odeen and Bjorklund (2003) was excluded because
Number of clades per species
origin of some specimens was unknown.
7
R2 = 0.37 P = 0.0002
6 5 4 3 2 1 0 0
5
10
15
20
25
30
35
Number of sampling localities
189
Figure 5.10. The frequency of the most common haplotype as a function of sequence length (A) and sample size (B) for two groups of studies.
A. Frequency of most common haplotype
long seq
0.8
CR
0.7
Linear (CR) Linear (long seq)
0.6
R2 =0.02 P = 0.52
0.5
R2 =0.01 P = 0.6
0.4 0.3 0.2 0.1 0 0
500
1000
1500
2000
Base pairs sequenced
Frequency of most common haplotype
B. long seq
0.8
CR
0.7
Linear (long seq)
0.6
Linear (CR)
0.5 R2 =0.06 P = 0.2
0.4 0.3
R2 =0.01 P = 0.54
0.2 0.1 0 0
50
100
150
200
250
300
350
Sample size
190
