Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reproductive isolation in a contact zone between divergent forms of winter wren (Troglodytes troglodytes) Toews, David P. L. 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-0617.pdf [ 5.48MB ]
Metadata
JSON: 831-1.0302422.json
JSON-LD: 831-1.0302422-ld.json
RDF/XML (Pretty): 831-1.0302422-rdf.xml
RDF/JSON: 831-1.0302422-rdf.json
Turtle: 831-1.0302422-turtle.txt
N-Triples: 831-1.0302422-rdf-ntriples.txt
Original Record: 831-1.0302422-source.json
Full Text
831-1.0302422-fulltext.txt
Citation
831-1.0302422.ris

Full Text

R E P R O D U C T I V E I S O L A T I O N IN A C O N T A C T Z O N E B E T W E E N D I V E R G E N T F O R M S OF W I N T E R W R E N {TROGLODYTES TROGLODYTES)  by  D A V I D P.L. T O E W S B.Sc.H, Acadia University, 2005  A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S FOR T H E D E G R E E OF  M A S T E R OF S C I E N C E  in  T H E F A C U L T Y OF G R A D U A T E STUDIES  (Zoology)  T H E U N I V E R S I T Y OF BRITISH C O L U M B I A  August 2007  © David P.L. Toews, 2007  ABSTRACT Geographic variation in vocalizations and genetics of winter wrens (Troglodytes troglodytes) in North America have led to speculation that the eastern (Troglodytes troglodytes hiemalis) and western (T. t. pacificus) subspecies are in fact distinct biological species. To determine whether two regional forms are separate species it is crucial to gather data from an area of overlap between the groups, if such an area exists. To address whether these forms are reproductively isolated, I quantified song characteristics and two types of molecular markers, mitochondrial (ND2) and nuclear (amplified fragment length polymorphisms), in individuals in a recently described overlap area and compared them to those individuals from further east and west. In this overlap area, near Tumbler Ridge, B C , both forms can be found inhabiting neighboring territories and each male wren sings either an eastern or western song, with the differences between these types of song being as distinct as they are in allopatry. The two taxa differ distinctly in mitochondrial D N A and in every case singing type perfectly predicts mitochondrial D N A clade, strongly supporting the hypothesis that the two forms are reproductively isolated where they co-occur and are therefore separate species. Analysis of multilocus nuclear markers supports this result, with only one first generation hybrid individual identified and no evidence of genetic introgression. A n estimate of the initial split between these two species, based on mitochondrial D N A sequence variation, dates to just prior to the Pleistocene, roughly 2.3 million years ago. It is therefore suggested that T.t. pacificus and related western taxa be elevated Troglodytes pacificus with the common name "Pacific wren". I speculate that behavioural isolation due to song evolution is likely an important premating barrier to gene flow in this system. These findings raise the possibility that there are other such morphologically cryptic species pairs in North America.  Ill  TABLE OF CONTENTS Abstract  i'  Table of Contents  iii  List of Tables  iv  List of Figures  •  v  Acknowledgements  vi  Co-Authorship Statement  vii  CHAPTER I - INTRODUCTION  1  C H A P T E R II - Evidence for two species of winter wren {Troglodytes troglodytes and T. pacificus) in North America  15  INTRODUCTION  15  METHODS  17  RESULTS  24  DISCUSSION  26  LITERATURE CITED  42  C H A P T E R III - Patterns of nuclear and mitochondrial genetic variation within and between Troglodytes (t.) pacificus and Troglodytes troglodytes  46  INTRODUCTION  46  METHODS  50  RESULTS  53  DISCUSSION LITERATURE CITED C H A P T E R IV - C O N C L U S I O N  —.55 .64 67  iv  LIST OF TABLES Table 2.1. Eigenvalues, variance explained, and factor loadings of the first three principal components produced in the P C song analysis  34  Table 3.1. A F L P primer combinations resulting in informative polymorphic fragments. ..61 Table 3.2 Eigenvalues and variance explained of the first three principal components produced in the P C analysis of A F L P profiles  .61  Table 3.3 Estimated log probability, variance, and AK for different numbers of clusters (K) from A F L P data  61  Table 3.4 Population pairwise FST values for groups of Troglodytes estimated from A F L P profiles and N D 2 m t D N A sequence  61  V  LIST OF FIGURES Figure 2.1. Breeding distributions of winter wrens (Troglodytes troglodytes) in North America, along with locations of research sites  Figure 2.2. Example song spectrograms from three eastern winter wrens (Troglodytes troglodytes hiemalis) and three western winter wrens (T. t. pacificus)  36  Figure 2.3 Example song spectrograms from six winter wrens in the area of overlap between eastern T. t. hiemalis and western T. t. pacificus wrens, near Tumbler Ridge  37  Figure 2.4 Peak sound frequency distributions differ between T. t. hiemalis and T. t. pacificus songs, both in allopatry and sympatry  38  Figure 2.5 Variation in seven basic song variables in allopatric and sympatric populations of western pacificus and eastern hiemalis winter wrens  39  Figure 2.6 Principle components analysis of individual songs of western (T. t. pacificus) and eastern (T. t. hiemalis) winter wrens are readily distinguished  40  Figure 2.7 Mitochondrial D N A haplotype network o f 16 winter wrens occurring in sympatry in Tumbler Ridge, B C , generated using 1041 bp-of-the N D 2 gene  41  Figure 3.1 Breeding distributions o f winter wrens in North America along with locations of research sites  .....62  Figure 3.2 Individual A F L P Principle Component scores of western (T. (t.) pacificus) and eastern (T. troglodytes) winter wrens  63  vi  ACKNOWLEDGEMENTS I will be forever thankful to all of the collaborators, colleagues, friends and funding agencies that were involved with supporting this project and me. Foremost, thank-you to my supervisor, Darren Irwin, whose guidance and direction made this not only a productive academic endeavor, but also an enjoyable one. Thank-you to my committee members, Dolph Schluter and Sally Aitken, who provided crucial direction to this project and insightful comments on my thesis. Endless thanks to my labmates, Andrew Rush, Katie Kuker and A l a n Brelsford, who went far beyond the call of duty in academic and non-academic support. Thank-you to Jason Weir to Anne Dalziel for indispensable advice in the lab and for comments on this thesis. Thank-you to the Canadian Society of Zoologists for opportunities to share my research and for constructive feedback. A s this would not have been possible without financial support, I am deeply indebted to N S E R C , the University of British Columbia, and the department of Zoology for the opportunity to study in such a welcoming, beautiful, and engaging environment. Thank-you to the South Peace B i r d Atlas Society helpful information regarding distribution of winter wrens in the Tumbler Ridge area. I am grateful to the Macaulay Library at the Cornell Laboratory of Ornithology for providing recordings and the Burke Museum for providing important museum specimens. Finally, I would like to thank my family and friends for endless encouragement and support these past two years - it wouldn't have been possible, or as enjoyable, without you.  Vll  CO-AUTHORSHIP  STATEMENT  A portion of this thesis (chapter 2) was co-authored in collaboration with my supervisor, Dr. Darren Irwin, at the University of British Columbia. Dr. Irwin was the first to identify the unique overlap area described in this chapter and collected the initial genetic samples and recordings. I collected additional genetic samples and recordings from field research during following summers and I performed the molecular and sonographic analyses on all of the samples.  1 C H A P T E R I - INTRODUCTION Under the biological species concept, "species" are defined as groups of interbreeding natural populations that are reproductively isolated from other such groups (Mayr, 1942). Reproductive isolation occurs when the build up of isolating barriers between populations separated for many generations is such that gene flow is either absent or greatly reduced between them (Coyne and Orr, 2004). These isolating barriers can result from segregation spatially, temporally, behaviourally, ecologically, or genetically, or from any combination of these factors. Separated populations w i l l eventually adapt to their different environments and . w i l l gradually accumulate differences over evolutionary time via natural selection. These differences can eventually lead to reproductive isolation upon secondary contact as a byproduct of the adaptation process (Schluter, 2001; Presgraves et al, 2003). Directional selection on beak size and shape in Galapagos finches, for example, is thought to have contributed to reproductive isolation between finches specializing on different food types, as females also use beak shape during mate choice (Grant and Grant, 1997). Additionally, under strong sexual selection, divergence in characteristics involved in mate preference and recognition can lead to varied behavioral and morphological traits involved in mate acquisition, such as flashy displays and plumage, in concert with changes in preference (Coyne and Orr, 2004). Unlike reproductive isolation evolving as a byproduct of adaptation, however, sexual selection can produce reproductive isolation without selection acting upon traits that are involved directly with the natural environment (WestEberhard, 1983). Runaway sexual selection, for example, is thought to have produced the spectacular plumage colourations in closely related birds of paradise, which historically  2 have been under very little predation pressure (Firth and Beehler, 1998). In fact, some forms of reproductive isolation might evolve under sexual selection with very little or no genetic changes between isolated groups, as in instances of rapid cultural evolution (Irwin and Price, 1999). Understanding the relative importance of the drivers of reproductive isolation, namely natural selection, sexual selection, and random genetic drift, has proved difficult for researchers. One reason is that the isolating barriers currently most important in restricting gene flow are not necessarily those important historically during the initial speciation event, which in some taxa may have been acting millions of years before present (Coyne and Orr, 2004). To overcome this obstacle, researchers use a combination of theory, laboratory experiments, and observational studies of natural systems to untangle the important traits and barriers involved in the initial divergence of two varieties from those that persist today. This, it is hoped, w i l l generate a basis from which comparative analysis can be used to determine if, as some researchers suspect, some barriers are consistently more important than others in the process of speciation (Coyne and Orr, 2005). If populations are presently segregated either temporally or spatially then gaining insight into the barriers involved in their isolation, other than geography or time, is difficult using tools currently available. Studying areas where recently diverged populations have ranges or habitats that overlap, however, can be helpful in identifying important isolating barriers in natural systems. For instance, in situ hybrid sterility and inviability can be quantified i f there is evidence of hybridization where divergent populations overlap (Saetre et al, 2002; Naisbit et al, 2002; Coyne and Orr, 2004).  3  Female mate preference, a possible form of pre-mating isolation, can be measured if females are presented with a choice between homo- or heterotypic mate in nature (Hoskin etal., 2005). Portions of the genome under directional selection can be determined by assessing the relative rates of genetic introgression across hybrid zones (Saetre et al., 2002; Rohwer et ai, 2001). Thus, areas of overlap between divergent populations play an important role in our understanding of the evolution and the maintenance of barriers involved in reproductive isolation (Coyne and Orr, 2004). Overlap zones o f closely related taxa in North America are generally considered areas of secondary contact between populations that were previously geographically isolated for a long period of time. There has been vigorous debate over the cause of this initial isolation, however, with glaciations during the Pleistocene being one prominent and contentious explanation (Klicka and Zink, 1997; Weir and Schluter, 2004; Lovette 2005). Glacial advances during the Pleistocene are thought to have promoted fragmentation in the flora and fauna of North America by restricting suitable habitat to a few, small, isolated ice-free refugia (Diamond & Hamilton, 1980). Phylogeographic and biogeographic studies in numerous taxa have found population structure and differentiation consistent with population bottlenecks and expansion during the Pleistocene glacial cycles (Avise, 2000). The impact of such historical allopatric isolation on current biogeographic patterns in North A m e r i c a is illustrated by some of the six major "suture zones" in North America (Remington, 1968). Each of these areas is characterized by narrow zones where one can find overlap areas between divergent populations in many different groups, including plants, insects, mammals, and birds. In some taxa, hybrids are readily formed in  4 these areas, while others at later stages of the speciation process mate assortatively and do not form hybrids. Determining why taxa differ in the extent of reproductive isolation, while sharing a similar biogeographic history, is an area of growing interest and at present an open question (Swenson and Howard, 2004). The diverse life histories, varied biogeography and elaborate mating systems of birds have made avian systems central to the formulation of many speciation theories (Edwards, 2005). Taxa in North America in particular have been intensely studied and provide an excellent system in which to examine the role of the reproductive barriers contributing to species formation. For instance, the suture zone in northwestern North America, centered on the Rocky Mountains of British Columbia and Alberta, allows for a unique opportunity to examine areas where eastern and western avifauna, inhabiting the unbroken chain of boreal forest, can meet. It has long been noted that there are large differences between the avifaunas o f western and eastern North America, both in species composition and between western and eastern forms of the same species (reviewed by Newton, 2003). What is unknown in many of these systems, however, is whether these differences between eastern and western forms represent species level differences, or simply the ends o f a phenotypic continuum. To address these questions properly a through investigation into where and i f these forms meet is required. One potential challenge in resolving these questions is accumulating adequate knowledge of species range boundaries and phylogenetic relationships. Differentiating between phylogenetically relevant traits, which are necessary to infer deep evolutionary relationships, from recently evolved adaptive ones is an ongoing challenge in systematics. Traditionally, avian taxonomists have relied primarily on diagnosable  5 variation in plumage colouration and morphological measurements, mostly in museum specimens, to delineate avian specimens at the species and subspecies level (Pyle, 1997). While documentation of this variation has lead to the designation of many subspecies, the biological significance of such classification, especially in an evolutionary and conservation context, has sometimes been questioned (Zink, 2004). Recent advances in molecular genetics, however, have ushered a new wave of important data allowing for more comprehensive and objective classifications. Phylogenetic reconstructions are now available for most major groupings of birds and researchers are beginning to unravel higher-order relationships between groupings within the neoaves (Ericson et al, 2006). This combination of modern molecular tools with the paleontological and geological record has also generated a wide range of robust molecular clocks (van Tuinen and Hedges, 2001; Arbogast et al, 2006; Weir and Schluter, submitted). A surprising outcome from of the application of these molecular clocks on taxa with deep phylogenetic splits is the general tendency for avian species to retain hybrid viability and fertility for millions o f years following initial population separation (Price and Bouvier, 2002). Thus, this research implicates a central role for pre-mating isolation mechanisms in avian speciation (Edwards, 2005) and has led researchers to reevaluate some avian classifications based solely on traditional methods. Research into the importance of birdsong, for example, as well as other characters involved in mate choice, have not only revealed these characters as an important tool in classifying species, but also suggested that they play an important role in generating reproductive isolation. For instance, the indigo-bird system in A f r i c a is one of the best, i f not only, examples of sympatric speciation in birds, and points to song as the primary isolating mechanism (Sorenson et  al, 2003). Playback experiments between sympatric cultural races of these brood parasites indicate some varieties have become reproductively isolated following invasion of novel hosts. This is thought to have occurred after male fledgling indigo-birds learned and mimicked a novel host's vocalizations, which was then integrated into their own species-specific repertoire (Sorenson et al, 2003). A l o n g with a female preference for males raised by the novel host, Sorenson et al. (2003) believe that song evolution was the important first step involved in the diversification and speciation in this system. A s a learned behavior in many birds, song is subject to rapid cultural evolution in which stochastic variations in copying are spread as individuals learn songs from their parents and/or neighbours (Grant and Grant, 1996; Payne, 1996). In allopatry, as differences gradually accumulate, reproductive isolation may eventually occur upon secondary contact, even i f very little other genetic or phenotypic divergence has occurred. In addition, the process of reinforcement can further act to displace song characters on secondary contact i f hybridization between the two forms produces unfit offspring, selecting for stronger premating isolation (Irwin and Price, 1999). Studies investigating r the role of song in reproductive isolation have found it to be of variable importance: in some systems song correlates strongly to genetic and sometimes morphological divergence (Mundinger, 1982, Martens, 1996, Alstrom and Olsson, 1999, Irwin et  al,.  2001a, b, Packert et al, 2004), while in others it does not (reviewed in Slabbekoorn and Smith, 2002). Interestingly, studies that have incorporated bioacoustic analyses in taxa where there is very little variation in plumage and morphological differences have sometimes revealed cryptic species and have prompted researchers to address vocal variation in studies of reproductive isolation (Irwin et al, 2001a).  7 One of the best examples of a North American avian species where large song type and repertoire differences have been noted between otherwise phenotypically similar subspecies is in the winter wren (Troglodytes troglodytes) species complex. O f the fiftynine species o f wrens, fifty-eight are found exclusively in the N e w World, with the winter wren being the only exception as it also occurs in the O l d World. Subtle geographic variation in plumage has led taxonomists to name more than 34 subspecies worldwide (Hejl et al, 2002, Kroodsma and Brewer, 2005), with 9 recognized within North America. These are grouped into a boreal eastern (T.t. hiemalis, T.t. pullus), western Pacific coast (T.t. pacificus, T.t. salebrosus, T.t. helleri), and a southwestern Alaska island group. Winter wrens are best known for their remarkable song, which, per unit weight, has approximately 10 times the sound power o f a crowing rooster (Brackenbury, 1983) and has been described as the "pinnacle o f song complexity" among songbirds (Kroodsma, 1980). While the winter wren's individual songs have garnered recognition from ornithologists, variation in repertoire size of individuals is also noteworthy. Members o f the subspecies T. t. pacificus in Oregon have a repertoire o f as many as 92 song units that individuals can use to create up to 66 different repeatable song types (Kroodsma, 1980). In contrast, an individual from the T. t. hiemalis subspecies in N e w York has a much simpler repertoire with about 20 song elements used to create only 2 song types (Kroodsma, 1980).. Interestingly, research into habitat preference has not indicated any significant difference between western and eastern North America, as important aspects o f breeding habitat for both subspecies include dead wood, in the form o f fallen logs and coarse  8 woody debris, snags, stumps, slash piles, and very large trees (Hejl et al., 2002). These features are normally associated with old-growth forests in North America and concerns for some populations have been raised because o f the increased scarcity o f this forest as a result o f habitat degradation and destruction. Interestingly, European winter wrens display very little habitat preference, as they are found in almost every habitat from seashore to highest mountain boulder fields (Cramp, 1988). Recent genetic studies o f this system have revealed an unexpected pattern, as variation in mitochondrial D N A (Drovetski et al, 2004) indicates there is a deep split in the entire winter wren complex occurring between a clade corresponding to western North America (71 t. pacificus and other western forms) and a clade corresponding to the rest of the range (Eurasia and eastern North America, including T. t. hiemalis). Combined with patterns o f song variation, this evidence has led to suggestions that winter wrens, which are currently treated as a single species (e.g. Brewer, 2001; Hejl et al., 2002; Kroodsma and Brewer, 2005), may in fact consist o f multiple cryptic species, with the group in western North America being distinct from those in eastern North America and Eurasia (Hejl et al., 2002; Drovetski et al, 2004; Kroodsma, 2005). However, these suggestions are based on comparison o f far west populations with those in eastern North America and almost no research has been done in the interior of the continent. To conclusively address this problem, researchers have called for an exploration o f western Canada to determine i f and where the two major groups o f winter wren in North America, the coastal western and boreal eastern, overlap in their breeding rang (Heijl et al, 2002; Kroodsma, 2005).  9 To address these questions, in chapter 2 I present data from a genetic and sonographic analysis of Troglodytes troglodytes in a recently discovered overlap area between these two forms. B y studying patterns of vocalization and mitochondrial D N A in this area of overlap I determined whether there is significant gene flow between the western, pacificus and eastern, hiemalis, subspecies. Observation of mixed song and mitochondria, for example, would provide good evidence that these forms are hybridizing regularly, exchanging genes, and are unlikely to be strongly reproductively isolated and would not qualify as "good species". In contrast, i f significant covariation between distinct song and genetic patterns were observed in sympatry, then data from this area would support the suggestion that these two subspecies are in fact reproductively isolated and should qualify for full species status. To supplement the genetic data presented in chapter 2, which is based solely on mitochondrial D N A , I present data collected from multilocus nuclear, amplified length polymorphism markers, in chapter 3. These markers have the ability to show genetic introgression where uniparentaly inherited mitochondrial D N A cannot, and this type of analysis is necessary to confirm that the results of chapter 2 are not the result of insufficient sampling. In this chapter I also use these genetic tools to address questions about structuring within the western race o f winter wren. The goal o f analyses of this type is to identify barriers to gene flow in sympatric populations of recently diverged taxa, and it is hoped that this may facilitate our understanding as to the important barriers responsible for reproductive isolation. While no single case study could exhaustively address any of these questions, by combining data with other taxa sharing a similar biogeographic history, a comparative approach may  10  provide the best framework with which to determine if any generalizations can be made about the process of speciation (Coyne and Orr, 2004). B y understanding the processes and mechanisms underlying the formation of new species it is hoped that this will give us a better appreciation for the generation and maintenance of biodiversity and to allow us to better preserve it for the future.  11  LITERATURE CITED: Alstrom, P. and U . Olsson. 1999. The Golden-spectacled Warbler: a complex of sibling species, including a previously undescribed species. Ibis 141:545-568. Arbogast, B. S., Drovetski, S.V., Curry, R.I., Boag, P.T., Seutin, G . , Grant, P.R., Grant, B. R., and D. J . Anderson. 2006. The origin and diversification of Galapagos mockingbirds. Evolution 60:370-382. Avise, J . 2000. Phylogeography: Press, Cambridge, M A .  The History and Formation of Species. Harvard Univ.  Brackenbury, J . 1982. The-structural basis of voice production and its relationship to sound characteristics. In D. F. Kroodsma, E. H. Miller, and H. Ouellet [eds.], Acoustic communication in birds. V o l . 1. Academic Press, N e w York. Brewer, D. 2001. Wrens, Dippers and Thrashers. Yale University Press, N e w Haven, Connecticut. Coyne, J. A . and Orr, H . A . 2004. Speciation. Sinauer, Sunderland, M A . Cramp, S. 1988. Handbook of the birds of Europe, the Middle East and North Africa. The birds of the Western Palearctic. Oxford Univ. Press, Oxford. Diamond, A . and A . Hamilton. 1980. The distribution of forest passerine birds and Quaternary climatic change in Africa. J . Zool, Lond. 191: 379-402. Drovetski, S. V . , Z i n k , R . M . , Rohwer, S., Fadeev, I.V., Nesterov, E.V., Karagodin, I., K o b l i k , E . A . and Y . A . Red'kin. 2004. Complex biogeographic history of a Holarctic passerine. Proceedings of the Royal Society of London B 271:545-551. Edwards, S., K i n g a n , S., Calkin, J., Balakrishnan, C , Jennings, W., Swanson, W. and M . Sorenson. 2005. Speciation in birds: Genes, geohraphy, and sexual selection. Proc. Nat. A c a d . Sci. 102: 6550-6557. Ericson, P . G . , Anderson, C . L . , Britton, T., Elzanowski, A . , Johansson, U.S., Kallersjo, M . , Ohlson, J . L , Parsons, T.J., Zuccon, D. and G . Mayr. 2006. Diversification of Neoaves: integration of molecular sequence data and fossils. Biology Letters 2:543-547. Firth C. and B . Beehler. 1998. The Birds of Paradise. Oxford University Press. Grant, P.R. and B . R . Grant. 1997. Genetics and the origin of bird species. Proc. Natl. Acad. S c i . 94: 7768-7775.  12  Hejl, S. J., Holmes J.A. and D. E. Kroodsma. 2002. Winter Wren (Troglodytes troglodytes). In The Birds of North America, no. 623 (A. Poole and F. G i l l , Eds.). The Birds of North America, Inc., Philadelphia, P A . Hoskin, C., Higgie, M . , McDonald, K., and C . Moritz. 2005. Reinforcement drives rapid allopatric speciation. Nature 437: 1353-1356. Irwin, D. and Price, T. 1999. Sexual imprinting, learning and speciation. Heredity 82: 347-354. Irwin, D. E., Alstrom, P., Olsson, U . and Z . M . Benowitz-Fredericks. 2001a. Cryptic species i n the genus Phylloscopus (Old World leaf warblers). Ibis 143:233-247. Irwin, D. E., Bensch, S. and T . D . Price. 2001b. Speciation in a ring. Nature 409:333-337. Klicka, J . and R. M . Zink. 1997. The importance o f recent ice ages in speciation: a failed paradigm. Science 277:1666-1669. Kroodsma, D. 1980. Winter Wren singing behavior: a pinnacle o f song complexity. Condor 82:357-356. Kroodsma, D. E. and D. Brewer. 2005. Family Troglodytidae (Wrens). Pages 356-447 In Handbook of the Birds o f the World. V o l . 10. Cuckoo-shrikes to Thrushes (J. del Hoyo, A . Elliott, and D. A . Christie, Eds.). L y n x Edicions, Barcelona. Lovette, I. J . 2005. Glacial cycles and the tempo o f avian speciation. Trends in Ecology and Evolution 20:57-59. Martens, J . 1996. Vocalizations and speciation o f Palearctic birds. In Ecology and Evolution of Acoustic Communication in Birds (D. E. Kroodsma and E. H. Miller, Eds.). Academic Press, N e w York. Mayr, E. 1942. Systematics and the Origin of Species. Dover, N e w York. Mundinger, P. C . 1982. Microgeographic and macrogeographic variation in the acquired vocalizations o f birds. In Acoustic Communication in Birds (D. E. Kroodsma and E. H . M i l l e r , Eds.). Academic Press, N e w York. Naisbit, R., Jiggins, C , Linares, M . , Salazar C . and J. Mallet. 2002. Hybrid Sterility, Haldane's Rule and Speciation in Heliconius cydno and H. melpomene. Genetics 161: 1517-1526. Newton, I. 2003. Speciation and Biogeography of Birds. Academic Press, London.  13  Packert M , Martens, J., Sun, Y . H . and M . Veith. 2004. The radiation of the Seicercus burkii complex and its congeners (Aves: Sylviidae): molecular genetics and bioacoustics. Organisms Diversity and Evolution 4:341-364. Payne, R. B. 1996. Ecology and Evolution of Acoustic Communication in Birds, eds. Kroodsma, D. E . & Miller, E. H.Cornell Univ. Press, Ithaca, N Y . Price, T. D. and M . M . Bouvier. 2002. The evolution of FI postzygotic incompatibilities in birds. Evolution 56:2083-2089. Presgraves, D., Balagopalan, L., Abmayr, S. and H . A . Orr. 2003. Adaptive evolution drives divergence o f a hybrid inviability gene between two species of Drosophila. Nature 423: 715-719. Pyle P. 1997. Identification guide to North American birds - part 1. Slate Creek Press, Bolinas, C A . Remington, C . 1968. Suture-zones of hybrid interaction between recently joined biotas. Evolutionary Biology 1:321-428. Rohwer, S., Bermingham, E. and C. Wood. 2001. Plumage and mitochondrial D N A haplotype variation across a moving hybrid zone. Evolution 55: 405-422. Sastre, G.,Borge, T., Lindroos, K., Haavie, J., Sheldon, B., Primmer, C . and A . Syvanen. 2002. Sex chromosome evolution and speciation in Ficedula flycatchers. Proc. R. Soc. L o n d . B . 270: 53-59. Schluter, D. 2001. Ecology and the origin o f species. Trends in Ecology and Evolution 16:372-380. Slabbekoorn, H . and T. B. Smith. 2002. Bird song, ecology and speciation. Philosophical Transactions o f the Royal Society of London B 357:493-503. Sorenson M . D . , Sefc K . M . and R . B . Payne. 2003. Speciation by host switch in brood parasitic indigobirds. Nature 424: 928-931. Swenson, N . and J. Howard. 2004. Do suture zones exist? Evolution 58: 2391-2397. West-Eberhard, M . 1983. Sexual Selection, Social Competition, and Speciation. The Quarterly Review o f Biology 58: 155-183. Weir, J. T. and D. Schluter. 2004. Ice sheets promote speciation in boreal birds. Proc. R. Soc. L o n d . B 271:1881-1887. Weir, J. T. and D. Schluter. 2007. The rate of molecular evolution in birds. Molecular Ecology, submitted.  14  van Tuinen, M . and S. Hedges. 2001. Calibration of A v i a n Molecular Clocks. Molecular Biology and Evolution 18: 206-213. Zink, R. 2004. The role of subspecies in obscuring avian biological diversity and misleading conservation policy. Proc. R. Soc. L o n d B 271: 561-564.  15  CHAPTER II* Evidence for two species of winter wren (Troglodytes troglodytes and T. pacificus) in North America: Distinct song and mitochondria in an overlap zone  INTRODUCTION It has long been noted that there are large differences between the avifaunas o f eastern and western North America, both in species composition and between eastern and western forms o f the same species (reviewed by Newton, 2003). More recently there has been vigorous debate over the causes of these differences, with the importance o f Pleistocene glaciations in causing population subdivision and subsequent speciation being one contentious issue ( K l i c k and Zink, 1997; Weir and Schluter, 2004; Lovette, 2005). One potential challenge in resolving this debate is accumulating adequate knowledge o f species range boundaries in some North American birds. In some taxa, differences have been noted between east coast and west coast populations, but populations in intermediate locations have not been investigated thoroughly enough to determine whether the two coastal forms represent extremes o f a continuum or i f they are in fact members o f distinct biological species, with a contact zone where the two cooccur without interbreeding. Here we investigate one such group that is currently classified as a single species but exhibits notable differences between east coast and west coast populations. Winter wrens (Troglodytes troglodytes) are noteworthy among songbirds both because o f their amazingly long and complex songs (Kroodsma, 1980, 2005; Kroodsma  * A version o f this chapter has been submitted for publication in The Auk. Toews D.P.L. and D . E . Irwin. Evidence for two species o f winter wren (Troglodytes troglodytes and T. (t.) pacificus) i n North America: Distinct song and mitochondria in an overlap zone.  16  and Momose, 1991; V a n Home, 1995) and because they are one of the few passerine species that has a distribution spanning both North America and Eurasia (Brewer 2001; Hejl et al, 2002). Winter wrens display subtle geographic variation in plumage, which has led taxonomists to name more than 44 subspecies worldwide (Hejl et al, 2002; Kroodsma and Brewer, 2005). Research into their vocalizations, however, has revealed marked regional phenotypic differences. Most notably, Kroodsma (1980, 2005) documented large differences in song types and repertoire size between populations in eastern and western North America. Specifically, members of the subspecies T. t. pacificus in Oregon have a repertoire o f as many as 92 song units (or syllables), which can be used to create up to 66 repeatable sequences (defined as song types; Kroodsma, 1980) per bird. The subspecies T. t. hiemalis i n N e w York has a much simpler song with about 20 song elements used to create only 2 song types per bird (Kroodsma, 1980). Songs o f T. t. hiemalis from eastern North America are more similar to songs of Eurasian forms (e.g. T. t. fumigatus in Japan and T. t. troglodytes in Europe) than to those o f T. t. pacificus in western North America (Kroodsma and Momose, 1991). Variation i n mitochondrial D N A parallels these patterns in song (Drovetski et al, 2004), with the deepest divergence in the entire winter wren complex occurring between a western North A m e r i c a clade (i.e. T. t. pacificus and other western forms; see F i g . 2.1) and a clade representing the rest o f the range (i.e. Eurasia and eastern North America, including T. t. hiemalis). These patterns have led to suggestions that winter wrens, which are currently treated as the single species T. troglodytes (e.g. Brewer, 2001; Hejl et al, 2002; Kroodsma and Brewer, 2005), may in fact consist o f multiple cryptic species, with  17 the group in western North America being distinct from those in eastern North America and Eurasia (Hejl et al, 2002; Drovetski et al, 2004; Kroodsma, 2005). When considering whether two regional forms are in fact separate species it is .crucial to gather data from an area of overlap between the groups, i f such an area exists (Irwin et al, 2001a, b; Kroodsma, 2005). Only then can it be determined whether there is 1) gradual change between the traits of the forms, suggesting gene flow between the forms, or 2) a region o f overlap between two forms with distinct differences, suggesting reproductive isolation. The second situation would provide strong evidence that the two forms are separate species under the biological species concept. Prior to the present study, little i f any research on winter wrens had been conducted in the extensive region (spanning more than 2500 km) between eastern North America (from Ontario and Minnesota eastward) and the Pacific coast (e.g. coastal Oregon, Washington, British Columbia, and Alaska). This lack of knowledge has led to calls for " a careful survey o f this wren's vocal behavior from Minnesota west, especially in n. Alberta and throughout British C o l u m b i a , . . . to determine how eastern and western wrens behave i f they m e e t . . ." (Hejl et al, 2002, p. 9). Evidence from this region is essential in determining whether the two forms of winter wren are separate species (Hejl et al, 2002; Kroodsma, 2005).  METHODS The study organism.—Winter wrens, alternatively referred to as Northern wrens (e.g. Kroodsma and Brewer, 2005) or simply wrens (in Eurasia, where they are the only species of wren; e.g. Knightley et al, 1998), have traditionally been classified within the genus Troglodytes, which contains 13 species according to Kroodsma and Brewer (2005). Recently, two studies (Rice et al, 1999; Gomez et al, 2005) have presented molecular  18 data indicating that the Troglodytes genus as currently defined is not a monophyletic group. These studies have shown that winter wrens, traditionally referred to as Troglodytes troglodytes, are the most distantly related of all species within Troglodytes, and have suggested that two other groups, the timberline wren Thryorchilus browni and the four species within the genus Cistothorus, are within the clade defined by all of the Troglodytes. Thus, to make Troglodytes a monophyletic clade, there are two possible solutions. Winter wrens could be placed in their own genus, Nannus, as suggested by Rice et al. (1999) and Gomez et al. (2005). Alternatively, as suggested by Gomez et al. (2005), Troglodytes could be made more inclusive by assigning the Troglodytes genus to the Thryorchilus and Cistothorus genera. We view these studies as important in that they show the current taxonomy likely needs to be changed. However, we think that more work needs to be done to clarify relationships and that it would be premature to change genus names now. Hence, we refer to winter wrens as Troglodytes troglodytes (consistent with the current A . O . U . list), while acknowledging that that name might soon be changed to Nannus troglodytes.  Field research.—During  several trips in A p r i l and M a y 2005 the second author  (DI) conducted an initial survey of variation in Winter wren song in British Columbia and Alberta, with the goal of determining whether there is a gradient in song types across this region or whether two distinct types meet in a contact zone. We first studied allopatric populations o f pacificus (at Gavin Lake, B C ) and hiemalis (at Lesser Slave Lake, A B ) and then traveled between these sites looking for intermediacy or overlap in singing types. We use the term "singing types" to recognize distinct and easily identifiable  19  macrogeographic differences in songs. These qualitative differences primarily include the dominant frequencies o f notes (singers with more lower frequency notes as compared to pacificus singers) and the rhythm o f note delivery (pacificus singers with a staccato note delivery including more trills as compared to hiemalis singers). The term "singing type" should not be confused with the term "song type", which refers to a distinct type o f individual song that has a unique series o f notes and has been used by previous researchers to evaluate individual repertoire size and song complexity (e.g. Kroodsma, 1980; Kroodsma and Momose, 1991; V a n Home, 1995). Each bird can have multiple song types, but each bird belongs to only one singing type (see results). These singing types can be relatively easily distinguished by ear, and our quantitative analysis confirms their distinctiveness (see results). Searching techniques included playback o f both hiemalis recordings (from Lesser Slave Lake) and pacificus recordings (from Gavin Lake) and listening for responses along forest roads and trails in the potential region o f contact. Though wren density in northwestern Alberta and northeastern British Columbia was low, we eventually located four wrens on 9-10 M a y 2005 in the vicinity o f Tumbler Ridge, B C . Three individuals sang songs typical of pacificus from Gavin Lake and one sang songs typical o f hiemalis from Lesser Slave Lake. A more thorough exploration over a 10-day period in late June, 2005 (DI), and 14 days i n mid-May 2006 (DT & DI) was used to determine whether there is interbreeding between the types in this overlap zone. In total, we found approximately 36 male wrens in the Tumbler Ridge area, o f which 16 were caught for temporary study in mist nets using song playbacks. Each was uniquely color-banded so that recordings or other observations taken at different times  20  could be matched to a given bird with certainty. Morphometric measurements and a blood sample (for later genetic analysis) were also taken from captured birds. Twelve of the captured wrens sang pacificus songs, while four sang hiemalis songs. In 2006 only one hiemalis singer was found; the three territories that had been occupied in 2005 by hiemalis singers were occupied in 2006 by non-banded pacificus singers (no individuals banded in 2005 were identified in 2006). The maximum pairwise distance between wrens studied in this area was 44 km, and the minimum was less than 0.2 km (i.e. neighboring territories). Three o f the hiemalis singers had a neighboring pacificus singer observed less than 0.4 k m away (for the fourth hiemalis singer, the closest known neighbor was a pacificus singer 1.4 km away).  Song analysis.—Songs were recorded using an Audio-Technica 815a microphone and a Sony T C D - D 1 0 0 D A T recorder (in 2005) or a Marantz P M D 6 6 0 solid state recorder (in 2006). Recordings ranged in length from three to more than 50 songs. Occasionally, song playbacks were used to induce the birds to sing before recording. While such playbacks undoubtedly affected singing behavior (e.g. song rate and loudness) we are confident as a result of much informal experimentation that they did not noticeably affect the content o f song (e.g. the shape o f syllables, or hiemalis vs. pacificus type) and thus did not affect the conclusions reported herein. To analyze songs quantitatively, we first examined recordings o f wrens from allopatric sites. In addition to our own recordings (hiemalis: five individuals from Lesser Slave Lake, A B (LS); pacificus: two from G a v i n Lake, B C ( G L ) , two from Whistler Interpretive Forest, B C ( W H ) , two from Pacific Spirit Park in Vancouver, B C (PS), and  21  one from Prince George, B C (PG)) we obtained recordings from the Macaulay Library at the Cornell Laboratory o f Ornithology (hiemalis: two from N e w Brunswick (NB) and two from Ontario (ON); pacificus: two from Oregon (OR) and two from California (CA)). We used only these allopatric songs to select the following variables that could be consistently and reliably measured in all allopatric songs. These variables included: Length: the length o f a song (in seconds), as measured visually using on-screen spectrograms in the program Raven (Cornell Lab o f Ornithology, version 1.2.1). Low Freq: The minimum frequency (in kHz) o f a song, measured visually using Raven. High Freq: The maximum frequency (in kHz) a song, measured visually using Raven. Mean Freq: The mean frequency (in kHz) o f a song, measured by starting at the first whole quarter second after the start o f a song, then determining the frequency of sound with the largest amplitude at points in time distributed every 0.25 s throughout the song. These measurements were semi-automated in Raven, using frequency at maximum decibel measurements at a series o f time points imported into the program. We manually removed from the analysis points in time when the individual was not singing (i.e. silence between notes). This analysis enabled an accurate, objective and relatively quick method to determine the mean frequency o f a bird's vocalization. SD Freq: The standard deviation (in kHz) o f the frequencies at time points throughout the song, obtained as described above for " M e a n Freq". This variable quantifies the amount o f variation in frequency o f a bird's song.  22  Percent Blank: To calculate the percentage of time points, as measured above for " M e a n Freq", where the bird was not singing, we divided the number of "blank" points by the total number of points in the song. Trans/Sec: To quantify the temporal pacing o f a bird's song in terms of the rate at which it switches between low and high frequency sounds, we used onscreen spectrograms (in Raven) to manually count the number of times in a song that the fundamental frequency o f sound changes from below 5.5 k H z to above 5.5 k H z (this value was chosen because it is roughly the mid-point in the range o f frequencies in typical winter wren song). The number o f such transitions was then divided by the length of the song to give number of transitions / second.  To ensure consistency, the first author (DT) collected all of the song data. Five songs were chosen at random from the total songs recorded for each individual, from between 10-50 recorded in total, for spectrographic analysis. Songs that contained disruptive background noise or other bird species were excluded from the analysis. One extremely short song was also excluded. We measured songs from nine hiemalis and eleven pacificus allopatric birds (described above) and from the sixteen birds that were captured in the contact zone at Tumbler Ridge. W e then conducted principal components analysis ( P C A ) i n R (R Development Core Team, 2006) using individual means of each bird's songs (five songs per bird, in all cases but one in which four songs were used). Analysis included all seven variables described above. A l l variables except Percent Blank were log-transformed before P C A . Factor loadings from the P C A on individual means were then used to  23  calculate P C scores for each song that was measured. This method o f analysis preserved statistical independence by using each bird once in the P C A that generated factor loadings, but then for graphing purposes allowed us to apply these factor loadings to all of the songs measured. The individual means were also used to test, using A N O V A , whether there were significant differences between allopatric hiemalis, sympatric hiemalis, sympatric pacificus, and allopatric pacificus.  Molecular analysis.—To determine whether singing type was predictive o f genotype, we sequenced the N D 2 mitochondrial D N A (mtDNA) gene for each individual and compared it to previously published hiemalis and pacificus haplotypes occurring in allopatry. A total o f 16 winter wren samples were obtained in Tumbler Ridge from four hiemalis singers and twelve pacificus singers. Blood samples were taken in the field from the brachial vein and stored immediately in S E T Buffer (0.15M N a C l , 0.05M Tris-Cl, 0.001M E D T A , p H 8.0) and left at ambient temperature until returned to the lab and frozen. Total genomic D N A was extracted using a standard phenol-chloroform extraction protocol. Following extraction, the D N A pellet was resuspended i n lOOu.1 o f I X T E (10 m M T r i s - H C l , 1 m M E D T A , p H 8.0) and stored at 4°C. The complete mitochondrial N D 2 gene (1041 bp) was amplified using Invitrogen P C R reagents and Taq polymerase, supplied by N e w England Biolabs. The fragment was amplified with primers L5215 (Hackett, 1996) and H I 0 6 4 (Drovetski et al., 2004) with the following thermal cycling temperature profile: 3 m i n at 95°C, 35 cycles o f 95°C (30 s), 55°C (30 s), and 72°C (30 s), followed by a final extension o f 72°C for 10 min. P C R fragments were sequenced by  .  24  Macrogen Genomics in Seoul, Korea. Sequences contained no indels and were aligned manually using BioEdit sequence editor (Hall, 2005). In addition to our samples, we obtained three sequences from Genbank o f each subspecies from allopatric areas far to the east or west (Drovetski 2004; hiemalis from Ontario: A Y 4 6 0 2 9 1 , A Y 4 6 0 2 9 2 , AY460294;pacificus from Washington: A Y 4 6 0 3 2 3 , AY460330, AY460332). To create a haplotype network we imported the sequence alignment into M E G A 3.1 (Kumar et al, 2004), which calculated the number of nucleotide differences between each of the samples. Due to the small number o f differences within and the large differences between the subspecies, we were able to create haplotype groupings manually. RESULTS Songs recorded at Lesser Slave Lake, A B , Grand Manan Island, N B , and Algonquin, O N (Fig. 2.2), were clearly similar to published songs o f T. t. hiemalis (Kroodsma, 1980; Hejl et al, 2002), whereas songs recorded at Gavin Lake, B C , Whistler Interpretive Forest, B C , Prince George, B C , Vancouver, B C , Corvallis, O R , and Sattley, C A (Fig. 2.2) were clearly similar to published songs of T. t. pacificus (e.g. Kroodsma, 1980; V a n H o m e , 1995; Hejl et al, 2002). Individual wrens at Tumbler Ridge sang either distinctly hiemalis songs or distinctly pacificus songs (Fig. 2.3). In no instance did we observe birds that sang both hiemalis and pacificus types (even during recording sessions on different days, ranging from one day to six weeks apart) and i n no case did individual songs show intermediacy between the two types. Our sampling o f song frequencies (Fig. 2.4) revealed a previously u n q u a l i f i e d difference between hiemalis and pacificus songs. Songs o f hiemalis have a bimodal  25  distribution o f frequencies, with peaks occurring at 4.3kHz and 7.5kHz. This contrasts to pacificus songs, which sing a large portion o f their notes at high frequencies, resulting in a unimodal distribution that peaks at 7.5kHz. These differences between hiemalis and pacificus songs, which can also be seen by examining spectrograms (Fig 2.3), are observed both in sympatry and allopatry. The distinctness o f hiemalis and pacificus songs was confirmed by quantitative analysis. O f the seven basic variables that we measured, three showed highly significant differences between groups (Fig. 2.5; see caption for statistical tests). These include high frequency (Max Freq), mean frequency (Mean Freq), and number of transitions through 5.5kHz (Trans/Sec), which are each higher in pacificus. The differences between hiemalis and pacificus are as strong within the overlap zone as between allopatric populations. Multivariate analysis further illustrates the differences between hiemalis and pacificus. The principal components analysis (Fig. 2.6, Table 2.1) shows no overlap between sympatric hiemalis and pacificus songs, with a noticeable gap between them. Within each taxon, songs differ little between allopatric areas and the contact zone. P C I , which is highly correlated with M a x Freq, M e a n Freq, and Trans/Sec (Table 2.1), is low in hiemalis and high in pacificus (t = -9.4326, d f = 18.878, P < 10" ). 7  Molecular analysis.—Among  the 16 individual winter wrens sampled in sympatry at  Tumbler Ridge, we observed 9 haplotypes that show a striking pattern o f relationships (Fig. 2.7): the 2 haplotypes in the 4 hiemalis singers differ from each other by only a single base pair, and the 7 haplotypes i n the 12 pacificus singers differ from each other by only up to 2 base pairs; in contrast, these pacificus and hiemalis haplogroups differ from  26  each other by 65 or more base pairs. The two most common haplotypes ( A and H , F i g . 2.7) correspond to previously published sequences from allopatric individuals (Drovetski etal, 2004). These common haplotypes differ at 65 of 1041 bp (6.24 %), an amount similar to the divergence found between allopatric pacificus and hiemalis populations by Drovetski et al. (2004). There was a perfect correspondence between singing type and haplogroup; each singer in sympatry belonged to the haplogroup o f allopatric individuals of the same singing type. Thus, in every case singing type was predictive of N D 2 genotype, an association that is highly significant (Fisher's exact test using samples from the contact zone: P = 0.000549). Sequences can be downloaded from Genbank (Accession numbers X X X X X X X X X X [to be determined upon acceptance]). DISCUSSION We have found an area in northeastern British Columbia in which some winter wrens sing only eastern (T.t. hiemalis) song and others sing only western (T. t. pacificus) song, with individuals o f the two types often occupying neighboring territories. Using multivariate analysis, we compared the songs within this overlap area with those from areas to the east and west in which only one of the song types is sung. W e found the differences within the overlap zone as strong as those between allopatric populations. Moreover, we found that each singing type corresponds perfectly to distinct mitochondrial clades. There are two possible explanations for these findings. First, the two forms might have only recently come into contact (e.g. within the last generation) and hence have not yet had a chance to interbreed or exchange singing types through cultural mixing. This  27 possibility is highly unlikely, since the two subspecies that apparently correspond to the singing types have been recorded as occurring in northeastern British Columbia for at least the past half-century (Campbell et al, 1997). In addition, it is highly unlikely that we observed contact within the first generation of contact. Second and more likely, the two types of winter wren in northeastern B C might correspond to distinct species that are reproductively isolated, at least to a high degree. This conclusion is not entirely surprising given the substantial differences in mitochondrial D N A (Drovetski et al, 2004) and song repertoire size (Kroodsma, 1980, 2005) that have been observed between wrens on the east coast and those on the west coast. Songs have been used increasingly in recent years to recognize the existence of morphologically cryptic species of birds (Irwin et al, 2001a; Packert et al, 2004), and this led us initially to ask whether there was gradual variation in winter wren song across North America or whether there was an area in which two distinct singing types were sung in sympatry. Song divergence is not necessarily indicative of genetic divergence. While some studies have found concordance in geographical patterns of acoustic and genetic variation at a small scale (Baker et al, 1982,1984) many others have not (reviewed in Slabbekoorn and Smith, 2002). Macro-scale song variation or "regiolects", however, encompassing extensive subpopulations over a large range, often correlate strongly to genetic and sometimes morphological divergence (Mundinger ,1982; Martens, 1996; Alstrom and Olsson, 1999; Irwin et al, 2001a, b; Packert et al, 2004). Thus, while the presence of regiolects does not indicate a mechanism of isolation, when these variants are maintained in sympatry, such as the overlap of singing types of T. t.hiemalis and T.t.  28 pacificus winter wrens in Tumbler Ridge, they provide strong evidence for reproductive isolation. Intrinsic post-zygotic isolation is not a likely cause of reproductive isolation between such recently diverged sister taxa, as studies have shown that hybrid sterility and inviability generally take much longer to develop than pre-zygotic and extrinsic postzygotic isolation do (Price and Bouvier, 2002). Rather, reproductive isolation between hiemalis and pacific is likely due to either pre-zygotic or extrinsic post-zygotic isolation mechanisms. W e examine each of these possibilities in turn. Pre-mating isolation often plays a major role in maintaining species boundaries in birds (Price, 1998; Irwin et al, 2001a). Dramatic plumage differences, timing of arrival on the breeding grounds, and variation in song characteristics have all been found to contribute to assortative mating among sympatric forms. We have little evidence for plumage differences or arrival time acting as barriers to interbreeding in this case: the two forms differ only in subtle patterns of plumage coloration, most noticeably on the throat, with pacificus being darker than hiemalis (Brewer, 2001; Hejl et al, 2002); we have no evidence for or against differences in arrival times in the overlap zone. Further study is needed to investigate i f either of these factors is acting as barriers to gene flow. However, we believe that differences in song and female response to song may be playing a more important role. L i k e any other trait, song is expected to diverge over time between species for a variety of reasons (Catchpole and Slater, 1995; Irwin, 2000). Females of many species are known to assess male songs when choosing a mate (Catchpole and Slater, 1995; Hasselquist et al, 1996), hence it is likely that song is not only an indicator of  29  reproductive isolation but also plays a role in generating that isolation. The very distinct differences between T.t. hiemalis and T. I. pacificus  song, which can be easily recognized  by a human after moderate training, is surely recognizable to a female winter wren. If females in the overlap area were not choosy between which male songs they preferred, it would be unlikely to observe a strong correlation between song types and m t D N A haplotypes. In fact, we see a perfect correlation between song type and m t D N A haplotype in the overlap area. Extrinsic post-zygotic isolation mechanisms could also contribute to reproductive isolation between T.t. hiemalis  and T.t. pacificus.  One possible source of postzygotic  isolation is seasonal migratory behavior. Two disjunct wintering areas of winter wrens occur in North America: one in the southeastern U.S., where it is thought that T.t. hiemalis mainly winters, and another along the west coast from California through Alaska, where it is thought that T.t. pacificus  mainly winters (Hejl et ai, 2002). Breeding  populations in the interior of the continent are highly migratory. These patterns suggest that the overlap zone between T.t. hiemalis and T.t. pacificus  may correspond to a  migratory divide (Bensch et al, 1999; Ruegg and Smith, 2002; Irwin and Irwin, 2005),-in which two forms breed in sympatry but migrate in different directions to their wintering grounds. Hybrids between the forms might inherit intermediate and likely inferior migratory behavior, likely resulting in selection against hybrids. Thus, differences in migratory behavior can potentially contribute to reproductive isolation, reducing gene flow between the divergent forms and promoting speciation (Helbig, 1991; Bensch et al, 1999; Ruegg and Smith, 2002; Irwin and Irwin, 2005). We suspect that migratory  30  differences contribute to reproductive isolation between T.t. hiemalis and T.t. pacificus, but that song differences play a larger role. Our data lead us to propose that within the currently defined T. troglodytes, the western subspecies, T.t. pacificus, along with other closely related western subspecies (e.g. T. t. salebrosus) should be promoted to the species level designation o f Troglodytes pacificus. We suggest the common name "Pacific wren" for this new species, as that name reflects its scientific name as well as its geographic distribution (although it should be noted that other subspecies o f T. troglodytes inhabit the Pacific coast o f Asia). The eastern subspecies, T.t. hiemalis, and other closely related subspecies (e.g. T. t. pullus), including Old-World forms, should retain the Troglodytes troglodytes species name for now. This includes the European form with the original "pure" trinomial Troglodytes troglodytes troglodytes. We speculate that future work may determine that additional cryptic species may occur with T. troglodytes, as suggested by Drovetski et al. (2004). In particular, it is likely that T. t. hiemalis is phylogenetically distinct from Eurasian forms of T. troglodytes, but testing their distinction under the biological species concept will be difficult as it is unlikely the two encounter each other in a natural setting. To properly date the splitting of the T.t. heimalis and T. (t.) pacificus wrens, a statistically robust phylogenetic tree combined with an accurately calibrated molecular clock is necessary. Drovetski et al. (2004) found a well-supported maximum likelihood tree that dated the split between the two taxa to approximately 1.5 million years before present. This date was based on an N D 2 molecular clock calibrated for Galapagos mockingbirds o f 5.5% sequence divergence per million years (Arbogast et al. 2006). This is at the upper range o f most N D 2 calibrations, which have been estimate i n most m t D N A  31  coding regions to be between 0.01 to 0.025 substitutions per site per lineage per million years (Arbogast et al, 2002). Recent evidence indicates that colonization o f highland areas may be used to calibrate reliable molecular clocks (Weir and Schluter, submitted). This would suggest, for instance, that the montane timblerline wren (Thryorchilus browni) and its lowland relative the rufous-browed wren (Troglodytes rufociliatus; Rice et al, 1999) can be used to calibrate an N D 2 molecular clock within the wrens. These two taxa each differ from their common ancestor by an estimated 0.086 substitutions per site (Toews, unpublished). This speciation event likely occurred prior to the uplift of Cordillera de Talamanca i n Costa Rica and western Panama during the subduction of the Cocos plate beneath the Caribbean plate, approximately 4.5 million years ago (Grafe et al, 2002; Abratis and Worner, 2001). Thus, this calibration results in an estimated rate of sequence divergence o f 3.8% per million years, rather than the 5.5% rate used by Drovetski et al (2004). Applying this rate to the branch lengths reported i n the tree o f Drovetski et al. (2004), we estimate the split between T. t. hiemalis and T. (t.) pacificus at roughly 2.3 million years before present, or slightly prior to the Pleistocene epoch. This estimated divergence date predates all ten o f the divergence dates estimated by Weir and Schluter (2004) for other boreal sister-species pairs, demonstrating that cryptic species can often be quite old compared to species pairs with, obvious plumage differences. The finding that Troglodytes (t.) pacificus and Troglodytes troglodytes hiemalis are distinct biological species, which have likely been evolving independently for millions of years, is especially interesting given their extreme morphological similarity. The two forms differ in plumage only in subtle ways, most noticeably on the throat, with T. (t.)pacificus being darker than T.t. hiemalis (Brewer, 2001; Hejl et al, 2002). In our  32 experience, this difference is xobservable in museum skins but quite challenging to use as a distinguishing feature in the wild, even when examining birds in the hand. The two taxa have broadly overlapping measurements at six morphometric traits (Toews and Irwin, unpublished). In contrast, we have been able to reliably identify individuals as hiemalis or pacificus after hearing just a single song, as confirmed by the quantitative analysis of song and the mitochondrial D N A data. This suggests a practical diagnostic tool differentiate the two species: it is clear from our data that future studies could simply measure song characteristics such as transitions per second or mean frequency to classify birds with high certainty. In addition to their morphological similarity, our field observations also suggest that T. t. hiemalis and T. (t.) pacificus may be quite similar ecologically. Within the contact zone, the two forms seem to have similar habitat preferences, occurring only in mature coniferous forest with much downed wood, a type of habitat that now covers only a small fraction of the landscape. The T. t. hiemalis males that we have observed in the overlap area have occurred near T. (t.) pacificus males, often in abutting territories. In several cases, we have observed 71 (t.)pacificus occupying terrirories that were occupied by- 71 t. hiemalis the year before. Informal playback experiments i n the overlap area also suggest that these two may be competing for resources, as males o f each species respond to songs of the other species about as well as their own. Together, these observations suggest that the two species utilize substantially similar resources. Thus, evidence indicates these species are reproductively isolated while still ecologically similar, perhaps explaining the narrowness o f the contact zone. Formal, more extensive playback  33 experiments are required in both sympatric and in allopatric populations to better understand the competitive and behavioral interactions between these two species. These results raise the question of whether there are other morphologically similar taxa in North America that are currently considered conspecific but in fact are reproductively isolated. Research in central A s i a has revealed the presence of many such cryptic species (reviewed by Irwin et al, 2001a), and further research in the central parts of North America, particularly in the boreal forests of western Canada, may similarly reveal more species pairs. Revealing such cryptic species may have important, implications for the debate about the role of Pleistocene glaciations in speciation. In this case, a taxon that until now was treated as a single species is in fact a species pair that is older than all of the boreal sister species splits examined so far in North America (Weir and Schluter, 2004).  34  Table 2.1. Eigenvalues, variance explained, and factor loadings of the first three principal components produced in the P C analysis (see Figs. 2.5-2.6). Factor loadings are equivalent to simple correlation coefficients between a variable and a principal component.  PCI  PC2  PC3  Eigenvalue  2.712  1.218  1.161  Variance explained  38.7%  17.4%  16.6%  Length  -0.186  0.408  -0.740  Min Freq  -0.025  -0.472  -0.750  Max Freq  0.912  0.127  -0.088  Mean Freq  0.920  0.200  -0.166  SD Freq  -0.150  0.583  -0.077  Percent Blank  -0.328  0.654  0.048  Trans/Sec  0.932  0.070  0.086  Factor loadings:  35  F i g . 2 . 1 . B r e e d i n g distributions o f winter w r e n  (Troglodytes troglodytes) i n N o r t h  A m e r i c a , a l o n g w i t h locations o f research sites. D i s t r i b u t i o n s o f subspecies are indicated w i t h i t a l i c i z e d names a c c o r d i n g to B r e w e r (2001) and H e j l et al. (2002). T h e western group also includes several subspecies o n islands o f f o f A l a s k a (not s h o w n ; see B r e w e r [2001] and H e j l et a l . [2002] for details; but see K r o o d s m a and B r e w e r [2005] for a differing treatment o f western subspecies). R e s e a r c h sites are i n d i c a t e d b y two-letter codes (PS: P a c i f i c S p i r i t Park, V a n c o u v e r , B C ; W H : W h i s t l e r Interpretive Forest, W h i s t l e r , B C ; G L : G a v i n L a k e Forestry Centre, i n the U B C A l e x Fraser R e s e a r c h Forest, northeast o f 150 M i l e H o u s e , B C ; PG: P r i n c e G e o r g e , B C ; TR: T u m b l e r R i d g e , B C ; LS: L e s s e r S l a v e L a k e , A B ) . R e c o r d i n g s obtained f r o m the M a c a u l e y L i b r a r y c a m e from four a d d i t i o n a l sites (OR: C o r v a l l i s , O r e g o n ; C A : Sattley, C a l i f o r n i a ; ON: A l g o n q u i n Park, O n t a r i o ; and NB: G r a n d M a n a n Island, N e w B r u n s w i c k ) .  36  F i g 2.2 E x a m p l e song spectrograms from ( A - C ) three eastern w i n t e r w r e n s  (Troglodytes troglodytes hiemalis) and ( D - F ) three western winter w r e n s (Troglodytes troglodytes pacificus). O n e song is s h o w n for each i n d i v i d u a l f r o m A ) G r a n d M a n a n Island, N e w B r u n s w i c k B ) L e s s e r S l a v e L a k e , A l b e r t a C ) A l g o n q u i n P a r k , O n t a r i o D ) G a v i n L a k e , B r i t i s h C o l u m b i a E ) P r i n c e G e o r g e , B r i t i s h C o l u m b i a and F ) W h i s t l e r , British Columbia  37  C O l D T f N J  CO  ^  T  M  CO LD  f  IN  CO  IC  T  M  00  U)  t  W  CO  T  N  [\J  F i g 2.3 E x a m p l e s o n g spectrograms f r o m s i x w i n t e r wrens i n the area o f o v e r l a p between eastern  hiemalis ( A - C ) and western pacificus ( D - F ) w r e n s , near T u m b l e r R i d g e ,  B C . O n e song is s h o w n for each i n d i v i d u a l . N o t e the s i m i l a r i t y o f each type to allopatric songs (see F i g u r e 2.2).  38  0.12 -j •  Frequency of note (Hz)  Fig. 2.4. Peak sound frequency distributions for hiemalis and pacificus songs, as found both in allopatry and sympatry. Peak sound frequencies were measured every 0.25 seconds throughout songs, resulting in a total sample size o f 1267 for hiemalis songs and 1795 for pacificus songs. Symbols show average frequency distributions for allopatric hiemalis (filled circles), sympatric hiemalis (filled diamonds), allopatric pacificus (open circles), and sympatric pacificus (open diamonds). Lines represent mean frequency between allopatric and sympatric samples at the same note frequency (black, hiemalis; grey, pacificus).  39  o  N  X  0  §  0  O  c  1  u•  0 o  O CD  <D  o o CO  8  0  •  *  CD  1  *  •  O  !  CL  1 1 1  r r / ^  1  /  A"  . < /  «F  #  ^  F i g . 5. Variation in song of allopatric and sympatric populations of western pacificus and eastern hiemalis winter wrens. Seven basic song variables and the first principal component are shown. Each point represents a mean of five songs from a single individual. There are statistically significant differences among groups in three raw variables ( A N O V A s on log-transformed values: Transitions per Second, F- 45.136, d f = 3 and 28, P < 10" ; H i g h Frequency, F= 10.578, df = 3 and 28, P < 0.0001; Mean Frequency, F = 34.476, df = 3 and 28, P < 10" ). P C I , which represents the major axis of variation in the basic variables, also differs significantly among groups ( A N O V A : F = 33.326, d f - 3 and 28, P < 10" ). 10  8  8  40 Higher max. freq., « higher mean freq., and more changes in frequency ™  8° 08  00 o  O o  Song PC1 I  Lower max. freq., lower mean freq., and fewer changes in frequency  CA  OR  VA i WH  GL JPG  TR  LS  ON  NB  Location Fig 2.6 Individual songs of western {pacificus; open symbols) and eastern (hiemalis; closed symbols) winter wrens are readily distinguished by principal components analysis. The analysis summarizes variation in seven variables. The first principal component ( P C I ) accounts for 38.7% o f variance in the entire data set. Each column, ordered left to right by longitude (west to east) represents an individual bird and each point represents a single song, of five, analyzed for each individual. Songs from the sympatric population in Tumbler Ridge (TR; diamonds) are indistinguishable from the respective allopatric songs (all other locations; circles).  41  65 changes  V —#-© s  F i g 2.7 M i t o c h o n d r i a l D N A haplotype n e t w o r k o f 16 winter wrens o c c u r r i n g i n sympatry i n T u m b l e r R i d g e , B C , generated u s i n g 1041 bp o f the N D 2 gene. T h e frequency o f haplotypes is represented by the areas o f the circles (the smallest circles are each representative o f one i n d i v i d u a l ) and the n u m b e r o f nucleotide differences is represented as the n u m b e r o f nodes between the c i r c l e s . N o d e s are c o l o r e d based o n the singing type o f the i n d i v i d u a l  (pacificus singers s h o w n as w h i t e , N = T 2 ; hiemalis singers  as shaded, N = 4 ) . T h e t w o major haplogroups, w h i c h correspond to distinct s i n g i n g types, are separated b y 65 mutations (6.2 % sequence divergence), but haplotypes w i t h i n each group differ f r o m each other b y o n l y 1 or 2 mutations.  42  LITERATURE CITED: Abratis, M . , and G . Worner. 2001. Ridge collision, slab-window formation, and the flux of Pacific asthenosphere into the Caribbean realm. Geology 29:127-130. Alstrom, P. and U . Olsson. 1999. The Golden-spectacled Warbler: a complex of sibling species, including a previously undescribed species. Ibis 141:545-568. Arbogast, B. S., S. V . Drovetski, R. I. Curry, P. T. Boag, G . Seutin, P. R. Grant, Grant, B.R. and D. J. Anderson. 2006. The origin and diversification o f Galapagos mockingbirds. Evolution 60:370-382. Arbogast B. S., Edwards, S.V., Wakeley, J., Beerli, P. and J. B . Slowinski. 2002. Estimating divergence times from molecular data on phylogenetic and population genetic timescales. Annual Review o f Ecology and Systematics 33:707-740. Baker, M . C , Sherman, G . L . , Theimer, T . C . and D. C . Bradly. 1982. Population biology of white-crowned sparrows: residence times and local movements of juveniles. Behavioral Ecology and Sociobiology 11:133-137. Baker, M . C , Cunningham, M . A . and A . D. Thompson. 1984. Cultural and genetic differentiation o f two subspecies o f white-crowned sparrow. Condor 86:359-367. Bensch, S., Andersson, T. and S. Akesson. 1999. Morphological and molecular variation across a migratory divide in w i l l o w warblers, Phylloscopus trochilus. Evolution 53:1925-1935. Brewer, D. 2001. Wrens, Dippers and Thrashers. Yale University Press, N e w Haven, Connecticut. Campbell, R. W . , Dawe, N . K., McTaggart-Cowan, I., Cooper, J . M . , Kaiser, G . W . , M c N a l l , M . C . E., and G . E. J . Smith. 1997. The Birds o f British Columbia. Volume 3. Passerines: Flycatchers through Vireos. U B C Press, Vancouver, British Columbia. Catchpole, C . K., and P. J . B . Slater. 1995. Bird Song: Biological Themes and Variations. Cambridge University Press, Cambridge. Drovetski, S. V . , Zink, R . M . , Rohwer, S., Fadeev, I.V., Nesterov, E . V . , Karagodin, I., K o b l i k , E . A . and Y . A . Red'kin. 2004. Complex biogeographic history o f a Holarctic passerine. Proceedings of the Royal Society o f London B 271:545-551. Gomez, J . E. M . , Barber, B . R., and A . T. Peterson. 2005. Phylogenetic position and generic placement o f the Socorro Wren (Thryomanes sissonii). A u k 122:50-56.  43 Grafe, K., Frisch, W . , V i l l a , I. M . and M . Meschede. 2002. Geodynamic evolution of southern Costa Rica related to low-angle subduction o f the Cocos Ridge: constraints from thermochronology. Tectonophysics 348:187-204. Hackett, S. J. 1996. Molecular phylogenetics and biogeography o f tanagers in the genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution 5:368-382. Hall, T. 2005. BioEdit: biological sequence alignment editor for Windows. [http://www.mbio.ncsu.edu/BioEdit/bioedit.html] North Carolina State University, N C . Hasselquist, D., Bensch, S. and T. von Schantz. 1996. Correlation between male song repertoire, extra-pair paternity and offspring survival i n the great reed warbler. Nature 381:229-232. Helbig, A . J . 1991. Inheritance o f migratory direction in bird species: a cross-breeding experiment with S E - and SW-migrating blackcaps (Sylvia atricapilla). Behavioral Ecology and Sociobiology 28:9-12. Hejl, S. J., Holmes, J . A . and D. E. Kroodsma. 2002. Winter Wren (Troglodytes troglodytes). In The Birds of North America, no. 623 (A. Poole and F. G i l l , Eds.). The Birds o f North America, Inc., Philadelphia, P A . Irwin, D. E . 2000. Song variation i n an avian ring species. Evolution 54:998-1010. Irwin, D. E., Alstrom, P., Olsson, U . and Z. M . Benowitz-Fredericks. 2001a. Cryptic species in the genus Phylloscopus (Old World leaf warblers). Ibis 143:233-247. Irwin, D. E., Bensch, S. and T.D. Price. 2001b. Speciation in a ring. Nature 409:333-337. Irwin, D. E., and J. H . Irwin. 2005. Siberian migratory divides: the role o f seasonal migration i n speciation. In Birds o f Two Worlds: the Ecology and Evolution o f Migratory Birds (R. Greenberg and P. P. Marra, Eds.). Johns Hopkins University Press, Baltimore, Maryland. Klieka, J., and R. M . Zink. 1997. The importance o f recent ice ages in speciation: a failed paradigm. Science 277:1666-1669. Knightley, C , Madge, S. and D. Nurney. 1998. Birds o f Britain and North-West Europe. Yale University Press, N e w Haven, Connecticut. Kroodsma, D. 1980. Winter Wren singing behavior: a pinnacle o f song complexity. Condor 82:357-356. Kroodsma, D. 2005. The Singing Life o f Birds: The A r t and Science o f Listening to Birdsong. Houghton M i f f l i n , N e w York.  44  Kroodsma, D. E., and D. Brewer. 2005. Family Troglodytidae (Wrens). Pages 356-447 In Handbook o f the Birds of the World. V o l . 10. Cuckoo-shrikes to Thrushes (J. del Hoyo, A . Elliott, and D. A . Christie, Eds.). Lynx Edicions, Barcelona. Kroodsma, D. E., and H . Momose. 1991. Songs o f the Japanese population of the Winter Wren (Troglodytes troglodytes). Condor 93:424-432. Kumar, S., Tamura, K . and M . N e i . 2004. M E G A 3: Integrated software for molecular evolutionary genetics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163. Lovette, I. J . 2005. Glacial cycles and the tempo o f avian speciation. Trends in Ecology and Evolution 20:57-59. Martens, J. 1996. Vocalizations and speciation o f Palearctic birds. In Ecology and Evolution o f Acoustic Communication in Birds (D. E. Kroodsma and E. H . Miller, Eds.). Academic Press, N e w York. Mundinger, P. C . 1982. Microgeographic and macrogeographic variation in the acquired vocalizations o f birds. In Acoustic Communication in Birds (D. E. Kroodsma and E. H . M i l l e r , Eds.). Academic Press, N e w York. Newton, I. 2003. Speciation and Biogeography of Birds. Academic Press, London. Packert M . , Martens, J., Sun, Y . H . and M . Veith. 2004. The radiation o f the Seicercus burkii complex and its congeners (Aves: Sylviidae): molecular genetics and bioacoustics. Organisms Diversity and Evolution 4:341-364. Price, T. D. 1998. Sexual selection and natural selection in bird speciation. Philosophical Transactions o f the Royal Society o f London B 353:251-260. Price, T. D., and M . M . Bouvier. 2002. The evolution o f F I postzygotic incompatibilities in birds. Evolution 56:2083-2089. R Development Core Team. 2006. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-proiect.org Rice, N . H., Peterson, A . T. and G . Escalona-Segura. 1999. Phylogenetic patterns in montane Troglodytes wrens. Condor 101:446-451. Ruegg, K . C , and T. B . Smith. 2002. Not as the crow flies: a historical explanation for circuitous migration i n the Swainson's Thrush (Catharus ustulatus). Proceedings of the Royal Society o f London B 269:1375-1381.  45 Slabbekoorn, H., and T. B. Smith. 2002. Bird song, ecology and speciation. Philosophical Transactions of the Royal Society of London B 357:493-503. Van Home, B. 1995. Assessing vocal variety in the Winter Wren, a bird with a complex repertoire. Condor 97:39-49. Weir, J . T., and D. Schluter. 2004. Ice sheets promote speciation in boreal birds. Proceedings o f the Royal Society of London B 271:1881-1887. Weir, J. T., and D. Schluter. 2007. The rate of molecular evolution in birds. Molecular Ecology, submitted.  46  CHAPTER III* Patterns of nuclear and mitochondrial genetic variation within and between Troglodytes (t.) pacificus and Troglodytes troglodytes.  INTRODUCTION Population dynamic and evolutionary studies that are based on variation at a single locus can be misleading, as there are many reasons why the assumption that variation at a single locus is representative of genome-wide variation may be violated (reviewed by Ballard and Whitlock, 2004). Despite these concerns, there has been a flurry of biogeographic studies basing conclusions solely on mitochondrial D N A (mtDNA) variation, which has a non-recombinational and uniparental mode of inheritance and thus acts like a single locus. Subsequently, studies revisiting these systems with data generated from multiple markers often support, but sometimes conflict with conclusions drawn from patterns observed in m t D N A (Irwin et al, 2005; Bensch et al, 2006). In either case, supplementing studies based on a limited number of markers with the addition o f multiple independent markers, especially of nuclear origin, can greatly improve our understanding of taxa with complex evolutionary histories (Hare, 2001; Irwin et al, 2005). Evidence o f hybridization and identification of reproductive isolation between divergent populations can be especially difficult when relying on solely m t D N A . Hybrid individuals, for instance, cannot be diagnosed from a single haploid locus such as m t D N A . In addition, theory suggests that even documentation of deep phylogeographic  * A portion of this chapter is being prepared for submission for publication in Molecular Ecology. Toews D . P . L . Patterns of nuclear and mitochondrial genetic variation within and between Troglodytes pacificus and Troglodytes troglodytes.  47 splits between geographically proximate populations is not necessarily conclusive evidence of reproductive isolation. For example, in a system with continuously distributed populations and a relatively short dispersal distance, the maternal inheritance of mtDNA can create phylogeographic structuring between mtDNA clades even when there are no barriers to gene flow (Irwin, 2002). This sort of phenomenon may have occurred in the greenish warbler system (Phylloscopus trochiloides), a widely distributed and geographically variable ring species in Asia, where a deep mtDNA phylogeographic split was unexpectedly discovered between western Himalayan populations that were previously assumed to be exchanging genes because of a gradient in morphology and song (Irwin et al, 2001). Nuclear markers subsequently indicated that there was in fact gene flow between these populations; this study illustrates the potential error in concluding reproductive isolati.on with evidence gathered only from mtDNA (Irwin et al, 2005). The conclusions forming the basis of the proposed taxonomic split and reproductive isolation of Boreal eastern and coastal western forms of winter wren (Troglodytes troglodytes and T. (t.) pacificus, respectively), elevating T. (t.) pacificus from subspecies to full species status, were based primarily on data from mtDNA and bioacousic variation from limited sampling in a recently discovered overlap area (see chapter 2). There are a number of alternative hypotheses that could not be addressed by this study, however, because of the aforementioned limitations in mtDNA for identifying hybridization and reproductive isolation. Thus, it is possible that these forms are in fact exchanging genes at an appreciable level though not detected through our analyses because a) sampling in the sympatric population was insufficient in detecting hybrids; b)  48 gene flow is being mediated by males in each species, either currently or historically, which would not be detected through analyses o f maternally-inherited m t D N A ; or c) while unlikely, cryptic hybridization could be masked by matrilineally inherited song type. I address these alternatives by utilizing nuclear, amplified fragment length polymorphisms ( A F L P ) markers in the present study. A F L P s are haphazardly distributed dominant nuclear makers that have been used widely to study hybridization in plants (e.g. Ritland & Ritland, 2000) and have recently become more popular i n vertebrate systems (Liu et al, 2000; Bensch et al, 2002a; Bensch et al, 2002b). While they are limited in that they cannot resolve heterozygotes in diploid organisms, as bands are scored as either present or absent, they make up for this deficiency by allowing for a large number o f loci to be analyzed with relatively little effort (Vos, 1995). In addition, A F L P s have the ability to show genetic mixture where m t D N A cannot and are much better suited for detecting hybridization. Therefore, I expect that i f past or present gene flow is in fact significant between these two putative species of winter wren, then heterospecific individuals in the sympatric population w i l l appear more genetically similar than heterospecific individuals from allopatric populations at a large portion o f screened nuclear A F L P loci. I employ the use o f multivariate statistics to illustrate the genetic relatedness o f individual winter wrens and test this hypothesis. The second goal o f this chapter is to examine geographic variation within other closely related subspecies, as the observation that T. t. hiemalis and T. t. pacificus were in fact distinct biological species raises the question to whether other subspecies within this complex exhibit genetic divergence. Many other avian boreal species complexes in North America contain "interior/Rocky Mountain" groups, which are thought to have been  49  isolated during the glacial cycling of the Pleistocene, and many show significant morphological, plumage, acoustic, and genetic differentiation (Weir and Schluter, 2004). Winter wrens are no exception to this biogeographic pattern, and the sampling regime for the present study was designed so that I might address whether the "interior/Rocky mountain" group in this taxon, Troglodytes (troglodytes) salebrosus, showed genetic differentiation from other groups in the complex. T. (t.) salebrosus was first identified by Burleigh (1959) based on slight plumage variation and is described as having medium-brown upperparts with a rufous wash and a throat and breast that is medium to pale brown (slightly paler than T. (t.) pacificus; Pyle 1977). I denote its species name in parentheses, reflecting current A O U listings, to indicate at present it is unclear whether T. (t.) salebrosus is most closely related to the recently named Troglodytes (t.) pacificus group or the T. troglodytes hiemalis group - no known genetic or sonographic studies have been carried out to date on this subspecies. T. (t.) salebrosus has a breeding range that is thought to occur in the south-western interior of British Columbia and north-eastern Oregon, making it geographically most proximate to T. (t.) pacificus, although its distribution has never been formally documented (Phillips 1986). In general, avian species complexes that speciated during the Pleistocene contain "interior/Rocky mountain" that are most closely related to western/coastal groups, although genetic analyses would be needed to address this question in this system (Weir and Schluter, 2004). Therefore, by employing the use of A F L P markers, I quantified the genetic differentiation between pacificus, hiemalis, and salebrosus via two methods: one method that required an a priori designation o f individuals as "salebrosus"''' (pairwise Fstatistics) and one that did not (clustering and assignment tests). This is an important  50  distinction, as differentiating "salebrosus" from pure "pacificus" based on plumage or acoustic traits was unreliable (see results). Thus, in this chapter I 1) use multilocus nuclear A F L P markers to determine whether there has been significant past or present hybridization between T. (t.) pacificus and T. troglodytes, and 2) use a combination of A F L P and m t D N A markers to determine the genetic relatedness o f T. (t.) salebrosus to T. (t.) pacificus and T. t. hiemalis.  METHODS  Sampling A total o f 71 winter wren genetic samples were obtained from the field and from museum specimens (Fig 3.1; the Burke Museum provided the samples [N=5] from W A , M N , N Y ) . Samples from the field were obtained during catch and release study in mist nets following response to song playbacks. Individuals were usually identified by song before capture, using the easily identifiable "singing types" described in chapter 2 to distinguish between T. (t.) pacificus and T. troglodytes. Samples were obtained from across the range o f T. t. hiemalis ( A F L P , n=T5; m t D N A , n=12) and T. (t.) pacificus ( A F L P , n=56, m t D N A = 35) including the sympatric population (TR) described in chapter 2 (see F i g 3.1 for research sites). To sample T. (t.) salebrosus, I traveled to areas in southeastern British Columbia thought to be within breeding range o f this subspecies (n=9; in F i g 3.1 N L , C L , C B , and P N ; D. Irwin provided 4 samples from P N and C L ) .  Molecular Methods  51  Blood from individuals captured in the field was taken from the brachial vein and stored immediately in Seutin Buffer (0.15M N a C l , 0.05M T r i s - C l , 0.001M E D T A , pH 8.0) and left at ambient temperature until returned to the lab and frozen. Total genomic D N A was extracted using a standard phenol-chloroform extraction protocol. Following extraction and ethanol precipitation, the D N A pellet was resuspended in lOOul of I X T E (10 m M T r i s - H C l , 1 m M E D T A , p H 8.0) and stored at 4°C. Our A F L P analysis followed the protocol of Vos et al. (1995) with only minor modifications. D N A was digested with the endonucleases EcoRI and Mse I followed by ligation of the E - and M-adaptors (lOOuM). These fragments pre-amplified using complimentary E - and M-primers. The products from this reaction were diluted 40x,and stored as a stock solution for the selective amplification. Combinations of the E- and M primers (with three additional bases at the 3' end; see Table 1) were used for selective amplification during touchdown P C R in a volume of lOuL. The E-primers were fluorescently labeled with either IR-700 or IR-800 dyes so that reactions could be duplexed with two E-primers and one M-primer. Bands were separated on a L I - C O R 4300 in a 6.5% polyacrylamide gel and the presence or absence of fragments was binary coded (1 or 0) in S A G A v2.0. A l l data analyses were performed using only polymorphic A F L P loci. A portion of the mitochondrial N D 2 gene (954bp) was amplified using Invitrogen P C R reagents and Taq polymerase, supplied by N e w England Biolabs. The fragment was amplified with primers L5215 (Hackett, 1996) and H1064 (Drovetski et al, 2004) with the following thermal cycling temperature profile: 3 min at 95°C, 35 cycles of 95°C (30 s), 55°C (30 s), and 72°C (30 s), followed by a final extension o f 72°C for 10 min. P C R  52 fragments were sequenced by Macrogen Genomics in Seoul, Korea. Sequences contained no inertions/deletions and were aligned manually using BioEdit sequence editor (Hall, 2005).  Statistical Methods Exported binary A F L P genotype data from S A G A were transposed in Excel and imported to J M P I N 4.0 where relationships among taxa were investigated using principal components analysis ( P C A ) . Multivariate analysis is quite amenable to interpreting this type of binary multilocus data and is routinely used in studies investigating the genetic relatedness of individuals (e.g. Irwin et al, 2005; Van Treuren etal, 2005). To compute Fsrstatistics (Wright, 1951), I used Arlequin 3.1 (Excoffier et al, 2005) and grouped samples i n t o ' T . (t.) pacificus'", "T. troglodytes", or "T. (t.) salebrosus" for pairwise comparisons. The current version o f this software package has limited support for dominant markers and requires that A F L P markers be coded as haploid R F L P markers. Under the assumption o f Hardy-Weinberg equilibrium, F^- estimates calculated in this way still provide meaningful values of partitioned genotypic variance, though these are not comparable to the F57 estimates from co-dominant markers, as these are derived from differences in the variance o f allele frequencies. For m t D N A , N D 2 sequence alignment was performed in M E G A 3.1 (Kumar et al, 2004), where editing was required to crop uneven ends. Sequences were then imported into Arlequin, which was used to calculate pairwise Fsr statistics. Structure v2.2 (Pritchard et al, 2000) was used to quantify the occurrence o f hybrid individuals and geographic structuring. Structure implements a model-based  53  clustering method to identify the number o f populations (K) with the highest likelihood i n a given dataset, such that K>1 would indicate genetic structuring (Pritchard et al., 2000). Each population is characterized by a set o f allele frequencies at each locus and individuals are assigned to populations with a given probability (Pritchard et al., 2000). The species level differences between T. (t.) pacificus I T. troglodytes should result in two distinct clusters, with any number o f clusters greater than two indicating structuring within one o f these species. If K>2,1 am especially interested i f the individuals assigned to these clusters correspond to those individuals sampled in the breeding range o f T. (t.) salebrosus. To perform this analysis, I assumed the admixture model and completed multiple runs for values o f K between 1 and 5. Each run consisted o f a 50,0000 step burnin with 50,000 additional cycles, and for each value o f KI ran the parameter set for 10 iterations. To deal with a statistical artifact produced by structure that results in higher likelihoods and variance with larger K, which can make determining the true number of clusters in a dataset problematic, I identified the K value with the largest A/C according to Evanno et al. (2005).  RESULTS A total o f 105 polymorphic A F L P makers were scored using 7 primer pair combinations (Table 3.1)- Visual inspection of A F L P profiles o f Troglodytes troglodytes and T. (t.) pacificus suggested that they differ at a significant portion o f loci. Principal components analysis confirms this observation and shows no overlap between T. troglodytes and T. (t.) pacificus individuals along the first principal component axis, with a noticeable gap between the scores for the two species (Fig. 3.2; see Table 3.2 for  54  eigenvalues and proportion o f variance explained). The first principal component (PCI) explains significantly more variation than P C 2 and P C 3 (20%, 5%, and 4%, respectively) and clearly separates the two species. Individuals from the sympatric population cluster with allopatric individuals o f the same species, and allopatric and sympatric populations within species are essentially indistinguishable. There is, however, a single individual falling between the two species clusters. Preliminary sequencing from a fixed difference in a nuclear intron between the species indicates it is likely a first generation hybrid, as it is heterozygous for this fixed difference (Toews, unpublished). Surprisingly, this individual was not captured in the sympatric population - it was captured in an allopatric western population ( G L in F i g 3.1), which was thought to be exclusive to T. (t.) pacificus. Subsequent pairwise V$T analyses were performed with this individual removed from that dataset. Calculation o f AK using the L(K) output from Structure showed a clear peak at a K-2 (Table 3.3). A l l individuals, with one exception (see below), were assigned to either of these populations population with a high probability by Structure (>90%). Assignments correlate perfectly with our a priori designation o f individuals as either T. troglodytes or T. (t.) pacificus. The hybrid individual was the one exception to this pattern and had an equal probability (45-55%) o f being assigned to either population, consistent with expectation o f a first generation hybrid. Output using this method, then, shows no evidence o f structuring within either o f these species, as would be concluded i f K>2. Calculation o f F s r between T.(t.) salebrosus and T. (t.) pacificus (Table 3.4) detected subtle structuring (F.s7=0.019 P<0.05) using the A F L P markers, although this  55  was much smaller than the differentiation calculated between T. (t.) pacificus and T. troglodytes (F = 0.425, P O . 0 1 ) . ST  The pattern o f geographic structuring in m t D N A is similar to that o f A F L P . m t D N A shows substantial and statistically significant structuring between T. (t.) pacificus and T. troglodytes (Fsr=0.927, P O . 0 1 ) , but no significant structuring between T. (t.) pacificus and T.(t.) salebrosus ( F 5 7 -0.034, P>0.05). The sampled haplotypes included -  the same ones identified in chapter 2 in the sympatric population (see F i g 2.7) as well as a small number o f new ones. None were unique to any o f our geographic localities, and none were associated with any structuring between T. (t.) pacificus and T. (t.) salebrosus.  DISCUSSION In the current study I analyzed nuclear ( A F L P ) and mitochondrial (mtDNA) markers from individuals in the North American winter wren species complex, including individuals from a recently described overlap area between divergent forms within this complex (see chapter 2). The main objective o f this o f this analysis was to address shortcomings o f our previous study, which asserted that Troglodytes troglodytes and Troglodytes (t.) pacificus were distinct biological species by identifying distinct m t D N A and bioacoustic patterns that were maintained in symptary, which was considered strong evidence o f reproductive isolation. Due to a low frequency o f individuals i n the contact area, however, our limited sampling may not have allowed us to detect gene flow or hybridization even i f it was fairly common. This issue is significantly addressed with analyses o f numerous independent nuclear markers ( A F L P s ) because they have the ability to show genetic introgression where a single haploid m t D N A locus cannot.  56  Therefore, i f these forms are not reproductively isolated and gene flow is in fact significant between them, we would expect that heterospecific sympatric individuals would appear more genetically similar than heterospecific allopatric individuals at a significant proportion o f these loci. Thus, small sample sizes can be successfully analyzed with A F L P s to detect the past or present imprint o f hybridization in instances where m t D N A markers cannot. Data from the present study are consistent with the major finding o f chapter 2: that gene flow between T. troglodytes and T. (t.) pacificus is occurring at only an extremely low rate i f at all, and the level of reproductive isolation between the two species is high. Multivariate analysis illustrates this clearly, with the first principal component axis distinctly separating the two species. I found no evidence o f genetic introgression, as heterospecific individuals living in sympatry were as different from each other as heterospecific individuals in allopatry. In addition, because nuclear markers take longer to complete the process o f lineage sorting, the finding that T. troglodytes and T. (t.) pacificus differ at a significant portion of the genome also supports observation that they have not shared a common ancestor for many generations (the m t D N A estimate is 2.3 million years before present; see chapter 2). Thus, I am confident in rejecting the alternative hypotheses that a) sampling didn't allow us to detect significant hybridization in the contact zone b) gene flow is being mediated by males and c) cryptic hybridization was masked by matrilineally inherited song type, as these would all result i n genetic mixing at nuclear loci and would likely be detected by the A F L P analysis. Thus, numerous lines o f evidence ( m t D N A , A F L P , and bioacoustic) suggest it is likely that  57  these two forms of winter wren are strongly reproductively isolated and current taxonomy should reflect this distinction. Interestingly, the analysis did detect a single adult hybrid individual, which had an "eastern" mitochondrial type in addition to a mixed nuclear genotype - there was an equal probability o f assignment (45-55%) to either parent species using A F L P data. It was later identified as a likely first generation hybrid (Fi), as it was heterozygous for a fixed nuclear polymorphism between the species (Toews, unpublished). The location of this sample is surprising as it was taken not from the contact zone, but rather from an allopatric T. (t.) pacificus area, approximately 300km away. This is surprising because i f one were to predict that hybridization was occurring between these species, one might expect the area with the highest probability for detecting hybrids to be the sympatric area. I cannot confidently determine the home territory of this individual, however, as it is possible this individual was still on spring migration (it was caught very early in the breeding season and seemed to be in suboptimal habitat). It is not entirely surprising to find hybrids between divergent avian species, as studies have found that hybrid sterility and inviability generally take much longer to develop than pre-zygotic and extrinsic (i.e. bahavioural or ecological) post-zygotic isolation in avian systems (Price and Bouvier, 2002). It is unclear from the present data i f hybrids at low frequency are able to contribute to the gene pools o f the parental populations, or i f extrinsic factors, such as divergent migratory behaviours or courtship behaviours, have reduced the chances of successful reproduction such that there is no gene flow between the species. There are a number of examples in avian systems where, even in the face o f hybridization, reproductive isolation is considered to be complete.  58 Two species in the well-studied Ficedula system, for example, hybridize at a relatively low frequency, but yet maintain species level differences (in Saetre et. al, 2002). The second goal of this study was to determine i f there is any significant substructuring within the recently named T. (t.) pacificus species. Specifically, I was interested i f there was genetic differentiation between pure T. (t.) pacificus and a related subspecies, Troglodytes (troglodytes) salebrosus. Based on previous records (Hejl et al. 2004) I traveled to areas thought to contain the breeding range for this subspecies and obtained genetic samples from nine individuals in this area (see F i g . 3.1). There were no apparent phenotypic differences (plumage or vocal) between these individuals and other T. (t.) pacificus individuals in the field, so I cannot be completely confident that the individuals in this area were from the T (t). salebrosus subspecies. The results o f the present genetic analysis are generally consistent with these field observations: there was no evidence of sub-structuring within either species when using clustering methods from nuclear makers, without any a priori taxonomic information, and only slight structuring between the subspecies when groupings were defined a priori, although only at nuclear loci. In addition, any genetic structuring that was observed between T. (t.) salebrosus and T. (t.) pacificus was substantially lower than between T. (t.) pacificus and T. troglodytes, and is unlikely to be representative of species level differences. Whether this result is due to insufficient sampling or due to a lack of genetic differentiation between T. (t). salebrosus and T. (t.) pacificus is unclear, however, and should be addressed in the future with more intensive sampling in Oregon and Idaho. Other avian boreal species complexes with both a "west coast/coastal" and an "interior/Rocky M o u n t a i n " group (Weir and Schluter, 2004) show significant genetic  59  differentiation between these groups and usually morphological, plumage, and acoustic differences. The fox sparrow (Passerella iliaca schistacea; Zink 1994), plumbeous vireo (Vireoplumbeus; Johnson 1995), Virginia's warbler (Vermivora virginiae; Brush and Johnson, 1976), Townsend's warbler (Dendroica townsendi; Rohwer et al., 2001), and the red-naped sapsucker (Sphyrapicus nuchalis; Cicero and Johnson, 1995) are all examples o f interior species/subspecies whose superspecies' shares a similar biogeographic pattern with the winter wrens, and all show some level o f phenotypic and genetic differentiation between groups within their respective genus/species. The coastal and interior subspecies o f Hutton's vireo, for example, Verio huttoni huttoni and V. h. stephensi, respectively, show significant genetic differentiation calculated from allozyme variation (FST = 0.614; Cicero and Johnson, 1992) that is much greater than observed between T. (t.) pacificus and T. (t.) salebrosus. Additionally, as our current hypothesis is that song divergence is a likely barrier to gene flow between T. troglodytes and T. (t.) pacificus, we might also expect that songs differ in T. (t.) salebrosus. N o previous recordings are readily available for this subspecies, however, and all individuals heard or captured in the field sang a typical T. (t.) pacificus song. Thus, while more rigorous sampling o f this subspecies should be attempted in future studies, the genetic data suggest it is unlikely that T. (t.) salebrosus exhibits species levels differences, such as those observed between T. (t.) pacificus and T. troglodytes. In addition, because o f its close genetic relatedness to T. (t.) pacificus and distant relatedness to T. troglodytes, I propose that i f it T. (t.) salebrosus is to be considered a subspecies (based on criteria other than genetic uniqueness, which it lacks) it should classified within the T. (t.) pacificus group Troglodytes pacificus salebrosus.  60 In conclusion, the finding that T. (t.) pacificus and T. troglodytes are distinct biological species is especially interesting given their extreme morphological similarity and emphasizes the importance o f supplementing traditional classifications with molecular and bioacoustic data from areas of overlap, i f and where they exist. Our previous reliance on maternally inherited m t D N A markers to resolve reproductive isolation (chapter 2), however, meant that conclusions about recent gene flow and hybridization were limited. The analysis of nuclear markers in sympatric and allopatric individuals in both these species revealed significant differentiation at nuclear loci and supports the conclusion that they are reproductively isolated. Although I documented a low level o f hybridization, it is unlikely this contributes to much, i f any, recent gene flow between these highly divergent species and depends on whether first generation hybrids can reproduce successfully with other hybrids or in backcrossing to either parental species. Future studies should attempt to document the song repertoires of these rare hybrids as well as expand genetic sampling of T. (p.) salebrosus.  61  Table 3.1. A F L P primer combinations resulting in informative polymorphic fragments. Primer combination #1 #2 #3 #4 #5 #6 #7  EcoRl Primer (NNN-3') AGC ACG ACC AGG AAG AGC AAC  Msel Primer (NNN-3')  Number of polymorphic fragments  CAT CAT CAT CAC CAC CAC CAC  20 6 18 16 11 20 14  Table 3.2. Eigenvalues and variance explained o f the first three principal components produced in the P C analysis o f A F L P profiles (see F i g . 2).  PC2  PC3_  4.728 . 4.5%  4.427 4.2%  PCI Eigenvalue Variance explained  20.679 19.7%  Table 3.3. Estimated log probability, variance, and AK at different clusters ( K ) from Troglodytes troglodytes and Troglodytes (t.) pacificus A F L P data. K 1 •2 3 4 5  Ln -4527.92 -3784.74 -3927.05 -3840.04 -3867.96  P(D) 145.78 367.81 777.43 740.76 900.81  AK  523.6 2.5 2.3  -  Table 3.4 Population pairwise differences and F T values for groups o f Troglodytes estimated from A F L P profiles and N D 2 sequence. * indicates pairwise comparisons with PO.01. S  T. t. hiemalis T. (t.) pacificus T. (t.) salebrosus  T. (t.) pacificus  AFLP .  mtDNA  AFLP  0.425* 0.448*  0.927* 0.986*  0.019  mtDNA -0.034  62  F i g 3.1 B r e e d i n g distributions o f winter w r e n  (Troglodytes troglodytes) i n N o r t h  A m e r i c a , a l o n g w i t h locations o f research sites. D i s t r i b u t i o n s o f species/subspecies are indicated w i t h i t a l i c i z e d names a c c o r d i n g to B r e w e r (2001), H e j l et al. (2002) a n d T o e w s and I r w i n  (submitted). T h e western group consists o f Troglodytes (t.)pacificus,  Troglodytes (t.) salebrosus, and a variety o f subspecies o n islands o f f o f A l a s k a (not s h o w n ; see B r e w e r [2001] a n d H e j l et a l . [2002] for details; but see K r o o d s m a a n d B r e w e r [2005] f o r a d i f f e r i n g treatment o f western subspecies). R e s e a r c h sites are indicated b y two-letter codes (PS: P a c i f i c S p i r i t P a r k , V a n c o u v e r , B C ; WH: W h i s t l e r Interpretive Forest, W h i s t l e r , B C ; GL: G a v i n L a k e Forestry Centre, i n the U B C / A l e x Fraser R e s e a r c h Forest, north o f 150 M i l e H o u s e , B C ; PG: P r i n c e G e o r g e , B C ; TR: T u m b l e r R i d g e , B C ; PN: Penticton, B C ; CL: C r i s t i n a L a k e , B C ; NL: N e l s o n , B C ; CB: C r a n b r o o k , B C ; KN: K a n a n a s k i s , A B ; AT, A t h a b a s c a R i v e r , H i n t o n , A B ; LS: Lesser S l a v e L a k e , A B ) . S p e c i m e n s obtained f r o m the B u r k e M u s e u m c a m e f r o m three additional sites (WA, M a z a m a C o u n t y , W A ; NY: M o n r o e C o u n t y , N Y ; MN: Penobscot County, Maine).  63 4 •;  AFLP P C 2  o  o o  o  v  O  o  -2  -0.5  1.5  0.5  2.5  A F L P PCI  Fig 3.2 Individual A F L P Principle Component scores of western (Troglodytes (t.) pacificus; open symbols) and eastern (Troglodytes troglodytes; closed symbols) winter wrens. The analysis summarizes variation in 105 A F L P loci. The first principal component ( P C I ) accounts for 2 0 % of variance in the entire data set and P C 2 accounts for 5%. Individuals from the sympatric population (diamonds; Tumbler Ridge) cluster with allopatric individuals (circles) o f the same species.  64 LITERATURE CITED: Ballard J.O. and M.C.Whitlock. 2004. The incomplete natural history of mitochondria. Molecular Ecology 13: 729-744. Bensch, S., Akesson, S., and D.E. Irwin. 2002a.The use o f A F L P to find an informative S N P : genetic differences across a migratory divide in willow warblers. Molecular Ecology 11: 2359-2366. Bensch, S., Helbig, A . J . , Salomon, M . , and I. Seibold. 2002b. Amplified fragment length polymorphism analysis identifies hybrids between two subspecies of warblers. Molecular Ecology 11: 473-680. Brush, A . H . and N . K. Johnson 1976. The evolution of color differences between Nashville and Virginia's warblers. Condor 78: 412^414. Cicero, N . K. and N . K . Johnson. 1992. Genetic differentiation between populations of Hutton's vireo (Aves: Vieronidae) in disjunct allopatry. Southwestern Naturalist 37: 344-348. Cicero, N . K. and N . K . Johnson. 1995. Speciation in sapsuckers (Sphyrapicus): III. Mitochondrial-DNA sequence divergence at the cytochrome-B locus. A u k 112: 547-563. Drovetski, S. V . , Zink, R . M . , Rohwer, S., Fadeev, I.V., Nesterov, E.V., Karagodin, I., Koblik, E . A . and Y . A . Red'kin. 2004. Complex biogeographic history of a Holarctic passerine. Proceedings of the Royal Society of London B 271:545-551. Evanno, G . , Regnaut, S., and J. Goudet. Detecting the number of clusters o f individuals using the software S T R U C T U R E : a simulation study. Molecular Ecology 14: 2611-2620. Excoffier, L., L a v a l , G . , and S. Schneider. 2005. Arlequin ver. 3.0: A n integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1:47-50. Geoffrey P. J , M c C o r m i c k , M.I., Srinivasan, M . , and J . V . Eagle. 2004. Coral decline threatens fish biodiversity in marine reserves. Proceedings of the Nation Academy of Sciences 101: 8251-8253 Hackett, S. 1996. Molecular Phylogenetics and Biogeography of Tanagers in the Genus Ramphocelus (Aves). Molecular Phylogenetics and Evolution. 5: 368-382. Hall, T. 2005. BioEdit: biological sequence alignment editor for Windows. [http://www.mbio.ncsu.edu/BioEdit/bioedit.html] North Carolina State University, N C .  65  Hare, M . P . 2001. Prospects for nuclear gene phylogeography. Trends in Ecology and Evolution 16:700-706. Hejl, S. J., Holmes J.A. and D. E. Kroodsma. 2002. Winter Wren (Troglodytes troglodytes). In The Birds of North America, no. 623 (A. Poole and F. G i l l , Eds.). The Birds of North America, Inc., Philadelphia, P A . Irwin, D. 2002. Phylogeographic breaks without geographic barriers to geneflow. Evolution 56:2383-2394. Irwin, D. E., Bensch, S. and T . D . Price. 2001b. Speciation in a ring. Nature 409:333-337. Irwin, D.E., Bensch, S., Irwin, J.H. and T.D. Price. 2005. Speciation by distance in a ring. Science 307: 414-415 Johnson, N . K . 1995. Speciation in vireos. I. Macrogeographic patterns o f allozymic variation in the Vireo solitarius complex in the contiguous United States. Condor 97:903-919. Kroodsma, D. 1980. Winter Wren singing behavior: a pinnacle o f song complexity. Condor 82:357-356. Kumar, S., Tamura, K . and M . N e i . 2004. M E G A 3: Integrated software for molecular evolutionary genetics analysis and sequence alignment. Briefings in Bioinformatics 5:150-163. L i u Z., Nicholas A . , L i P., Dunham R A . 2000. Inheritance and usefulness o f A F L P markers in channel catfish (Ictalurus punctatus), blue catfish (/. furcatus), and their F l , F 2 , and backcross hybrids. Molecular and General Genetics 258: 260268. Phillips, A . R . 1986. The Known Birds of North and Middle America. Denver Museum o f Natural History. Price, T. D., and M . M . Bouvier. 2002. The evolution o f F i postzygotic incompatibilities in birds. Evolution 56: 2083-2089. Pyle P. 1997. Identification guide to North American birds - part 1. Slate Creek Press, Bolinas, C A . Pritchard, J . K . , Stephens, M . , and P. Donnelly. 2000. Inference o f Population Structure Using Multilocus Genotype Data. Genetics 155: 945-959.  66 Ritland, C. E. and K. M . Ritland. 2000. D N A fragment markers in plants. Pp. 208-234 in A l l a n J. Baker (ed.), Molecular Methods in Ecology. Blackwell Scientific, Oxford, U . K . Rohwer, S., Bermingham, E., and Wood, C. 2001. Plumage and mitochondrial D N A haplotype variation across a moving hybrid zone. Evolution 55: 405-422. Sastre, G.,Borge, T., Lindroos, K., Haavie, J., Sheldon, B., Primmer, C . and A . Syvanen. 2002. Sex chromosome evolution and speciation in Ficedula flycatchers. Proc. R. Soc. Lond. B . 270: 53-59. Swofford, D. L. 2003. P A U P * . Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Weir, J. T., and D. Schluter. 2004. Ice sheets promote speciation in boreal birds. Proceedings o f the Royal Society o f London B 271:1881 -1887. Wright, S. 1951. The genetical structure of populations. Annals o f Eugenics 15:323-354. Van Treuren, R., Bas, N . , Goossems, P., Jansen, J . , and L. V a n Soest. 2005. Genetic diversity in perennial ryegrass and white clover among old Dutch grasslands as compared to cultivars and nature reserves. Molecular Ecology 14: 39-52. Vos, P., Hagers, R. and Bleeker M . 1995. A F L P : new technique for D N A fingerprinting. Nucleic Acids Research 23: 4407-4414. Zink, R . M . 1994. The geography o f mitochondrial D N A variation, population structure, hybridization, and species limits in the Fox Sparrow (Passerella iliaca). Evolution 48:96-111.  67 C H A P T E R IV - CONCLUSION The main finding from the data presented in this thesis was that two subspecies within the North American winter wren (Troglodytes troglodytes) species complex,  Troglodytes troglodytes pacificus and Troglodytes troglodytes hiemalis, are in fact a cryptic species pair. While this was especially interesting given the morphological similarity of these two species, my sonographic analysis agreed with previous research on the vocalizations of these taxa, suggesting the two have very different songs and have dramatically different repertoires (Kroodsma, 1980). Combining the bioacoustic data with two types of molecular markers (mtDNA and nuclear A F L P ) in sympatric and allopatric populations o f this species, I found no evidence to suggest significant amounts of gene flow between these types. Molecular dating also indicated that the two have not shared a common ancestor since the early Pleistocene, making them one of the oldest boreal sister-species splits known and it is likely that the glacial cycles contributed to or completed the speciation of these two taxa (Lovett, 2005; Weir and Schluter, 2004). Combining these different lines of evidence, we proposed that Troglodytes troglodytes pacificus be elevated to species status, Troglodytes pacificus, and suggested the common name the "Pacific wren". Based on preliminary data suggesting the two are ecologically very similar, and research in other avian systems that suggest premating isolation plays an important role in reproductive isolation in birds (Price and Bouvier, 2002), I predicted that behavioural isolation is a likely barrier to gene flow in this system. In this concluding chapter I would like to expand the rationale for this prediction, propose other potential barriers to gene flow including their likelihood of being present in the winter wren system, outline recent  68 research into the genetic basis o f these barriers, and comment on potential areas o f future research. The preponderance o f prezygotic isolation mechanisms involved in the reproductive isolation o f avian taxa allows for rapid speciation, but also because of the slow development o f intrinsic postzygotic isolation in birds, for continued hybridization between divergent taxa (Grant and Grant 1997; Price and Bouviuer 2002). This means that while detecting the underlying genes and traits responsible for speciation requires sensitive field and lab methods, phenotypes under divergent ecological or sexual selection are likely differentiated at only a small number o f loci (Edwards 2005). Research into the genetic basis o f avian speciation has been greatly advanced with modern theory and molecular tools, and this has lead to exciting developments in our understanding o f the traits involved in this process. Interestingly, speciation i n avian systems proceeds in a fashion unlike other wellstudied systems, which follow closely the predictions o f Haldane's rule. This rule posits that reduced hybrid fitness between incipient species is likely to affect the heterogametic sex (females in birds) before the homogametic sex (Haldane 1922). In Drosophila, for example, male sterility is often the first to evolve followed by male inviability, then female sterility and female inviability, as fruit flies have an X Y sex-determinate system and males are heterogametic (Coyne and Orr, 2004). A v i a n systems, i n contrast, proceed with female sterility evolving first, which is subsequently followed by male sterility, then female inviability and male inviability (Price and Bouvier, 2002). Thus, the appearance of male sterility before the inviability o f the heterogametic sex i n birds (females) has implicated a role for some type o f modulator o f Haldane's rule. W u and Davis (1993)  69  proposed that strong sexual selection on reproductive genes could be a potential candidate as this modulator. Specifically, if there is rapid evolution in genes underlying reproductive traits and displays via strong sexual selection, then this is likely to disproportionately affect males, in opposition to Haldane's rule, and has been proposed as a mechanism to explain the early onset of male sterility. Empirically demonstrating strong sexual selection on reproductive traits by assessing hybrid fitness costs has been difficult in birds, as most hybrids show little evidence o f intrinsic fitness loss. Recent work into post-mating isolation mechanisms in Ficedula flycatchers, however, has supported theory implicating strong sexual selection on reproductive traits (Saetre et al, 2002; Borge, et al, 2005; Backstrom et «/.,2006). This conclusion was drawn from a correlation between reproductive traits occurring on the Z-chromosome, which is thought to be rich in genes coding for secondary sexual characteristics birds, to fitness i n hybrids with hybrid sex chromosomes (Saetre et al, 2002). Fitness estimates showed that all females with a hybrid sex chromosome genotype were sterile, while 2 7 % o f males heterozygous at Z were sterile (Saetre et al, 2002). Subsequent studies in this system found that Z-chomosome genes were unable to move across the hybrid zone, indicating strong selection on genes on this chromosome (Saetre et al, 2002). Without substantial hybridization between T. (t.) pacificus and T. troglodytes it is difficult to apply these types o f methods to determine i f genes on the Z-chromosome were disproportionately under selection while these two forms were isolated. A study comparing genetic variation at Z-linked loci and autosomal loci i n allopatric populations of both species o f Troglodytes could provide a crude linkage map (autosomal or sex-  70 linked), but numerous hybrids and backcrosses would be required to accurately associate genotype with phenotype in this system (Borge, et al., 2005). A more fruitful approach to addressing genes and traits potentially responsible for post-zygotic isolation in a nonmodel system like Troglodytes would be to characterize genes or traits thought to be good candidates for divergent selection. One such trait thought to contribute to isolation between diverging avian taxa, straddling the grey area between an ecological and behavioural trait, is seasonal migratory behaviour. Classic studies into the migratory directionality of hybrid crosses between divergent populations of blackcaps in Europe, for example, which have differing migratory directions, show that first generation hybrids express an intermediate migratory route (Helbig, 1996). Studies also have found that this complex behavioural trait can diverge in a very short period (Berthold et al., 1992) and that it is likely controlled by a few genes o f large effect (Pulido, 2007). This has subsequently led researchers to evaluate migratory divides between taxa with divergent migratory routes as potential sites for speciation, where selection against hybrids with intermediate migratory routes reinforces premating isolation (Irwin and Irwin, 2005). While theory has demonstrated that these dynamics are able to contribute to speciation, empirical evidence from natural systems is scarce. Studies have relied primarily on isotopic signatures of hybrids to determine wintering location, which is not necessarily indicative of migratory route. Studies of flycatchers or w i l l o w warblers indicate that migratory routes of hybrids are not intermediate between parental types, and instead have wintering locations characteristic of one or the other parental types (Bensch et al., 2006; Veen et al., 2007).  71 The two species o f Troglodytes in the present study have disjunct wintering grounds and likely have very different migratory routes, although little research has been done into the migratory behaviour of this taxa. The finding in chapter 3 of a single hybrid individual during spring indicates that at least some after hatch year birds manage to return to the breeding grounds, although it is unclear whether this individual showed any intermediacy or inferiority in migratory behaviour as compared to the parental individuals. Thus, while I suspect that migratory differences contribute to reproductive isolation between T. troglodytes and T. (t.) pacificus, and should be addressed in future work, it is more likely that other factors, such as song, play a larger role. Song has the ability to evolve between diverging populations of birds rapidly because of the stochastic properties of cultural inheritance (Grant and Grant, 1996; Payne 1996). Changes produced v i a the process of "cultural drift", for example, analogous to genetic differences arising from random genetic drift, can change song characteristics over a short period of time. It is thought that dramatically different songs of insular forms of birds, for example, as compared to their mainland relatives, may be a result of the sampling effect from a small number of founders (Baker et al., 2006). In addition to driftlike processes acting on song, habitat-dependent selection has also been found to contribute to differences i n song in isolated populations. A s habitats differ in the density and type of vegetation, song is subject to differing selection pressures in which sounds travel better or degrades as it moves through the physical environment (reviewed in Slabbekorn and Smith, 2002). Lower frequencies, for example, while having less directionality than high frequency sounds, travel better in dense vegetation and are more closely associated with forest-dwelling species (in Slabbekorn and Smith, 2002).  72 The understanding of the genetics of birdsong has advanced rapidly with the evolution of genomic resources and molecular methods. Localization of song production centers in brains o f zebra finches has allowed researchers to begin to unravel some of the genetic architecture involved in this trait. Recent studies show that during times of song plasticity, for example, the gene FoxP2 is upregulated in a region of the songbird brain essential for song learning (Haesler et al, 2004). Mutant forms of this highly conserved gene in humans are also associated with a disorder that affects both the comprehension of language and its production, which was what initially directed researchers to explore it's role in songbirds (Wolfgang et al, 2002). While no correlation in sequence variation has been observed between avian song learners and non-song learners, important expression studies have yet to be performed comparing these two groups (Webb and Zhang, 2005). The differences observed between T. troglodytes and T.(t.) pacificus songs and repertoires are likely due to a combination of cultural and genetic factors. Differences in individual song characteristics, for instance, can probably be attributed to a mixture habitat-dependent selection and female mediated sexual selection. During geographic isolation, eastern and western populations may have had differing selection pressures for frequencies suited to their own habitats, raising or lowering the prevalence of certain frequencies with better or worse transmission properties. In addition, females may have evolved differing preferences for certain notes or song deliveries, maybe preferring more or less trills during singing, for example. Given the differences in repertoire size between T. troglodytes and T. (t.) pacificus it is also likely that there has been strong selection on genes such as FoxP2 and others responsible for song learning in T. (t.) pacificus, with individuals in this species having dramatically more notes and songs in their repertoire.  73 Interesting insights into what proportion of these differences in repertoire size are genetic and learned could be addressed in future studies with lab raised animals, where one could test i f T. troglodytes  individuals have the ability to learn T. (t.) pacificus  type songs and  repertoires and visa versa. Finally, differentiation in habitat preference on the breeding ground could be responsible for the lack of geneflow between these species in secondary contact. A s species are isolated to different habitats they will, over evolutionary time, adapt to those environments (Schluter, 2001). If habitat dependent selection is strong enough, then these isolated populations may have diverged enough in their niches they are unlikely to compete for resources if and when they come into secondary contact. Selection on beak size and shape in Darwin's finches, for example, meant that populations isolated to islands with differing environments and resources no longer competed for resources in secondary contact, and were able to coexist in sympatry (Grant and Grant, 1997). Subsequent molecular work on these finches reared in the lab has discovered the gene  Bmp4 expressed in the cascade responsible for the large beak sizes of magnirostris  Geospiza  and has been proposed as a likely candidate for beak diversity in this system  (Abzhanov et al., 2004). While it is difficult to exhaustively demonstrate ecological similarities between two taxa* I have several lines of evidence suggesting that it is unlikely T. troglodytes T. (t.) pacificus  and  have substantially diverged ecologically. This is primarily based on the  observation that the two prefer similar habitats for territories in the sympatric populations. In fact, I have observed T. (t.) pacificus were occupied by T. troglodytes  individuals inhabiting territories that  during the previous breeding season. This, combined  74 with the observation of interspecific territoriality in response to song playbacks, indicate that while individuals i n sympatry may not be competing for mates, they may still be competing for resources. Future research could further explore the ecological differences between these two species by identifying micro-scale habitat preference or by quantifying isotopic signatures o f individuals in the sympatric populations, to see i f there is any finescale differentiation not detected by our preliminary assessment. In conclusion, the evolutionary dynamics we have observed between divergent forms of winter wren has allowed for a unique perspective on avian contact zones as well as insights into the important processes involved in the evolution and maintenance of biodiversity. More importantly, by integrating our understanding o f the dynamics this contact zone with other songbird species that share similar biogeographic history in a comparative approach, such as the recently discovered overlap areas between Macgillivray's/Mourning warblers and Townsend's/B lack-throated green warblers near the sympatric population o f winter wrens (Irwin and Toews, unpublished data), then this w i l l hopefully one day facilitate generalizations about the process o f speciation. Not only w i l l this enable a better understanding of the^mechanisms involved in reproductive isolation, but also better equips researchers to protect the evolutionary potential of lineages for the future by recognizing the processes that went into their formation.  75  L I T E R A T U R E CITED: Abzhanov, A . , Protas, M . , Grant, B.R., Grant, P.R., and C . Tabin. 2004. Bmp4 and Morphological Variation o f Beaks in Darwin's Finches. Science 305: 1462-1465. Backstrom, N . , Brandstrom, M . , Gustafsson, L., Qvarnstrom, A . , Cheng, FL, and H . Ellegren. 2006. Genetic M a p p i n g in a Natural Population o f Collared Flycatchers (Ficedula  albicollis):  Conserved Synteny but Gene Order Rearrangements on the  A v i a n Z Chromosome. Genetics 174: 377-386. Baker, M . , Baker, M . S . and L. Tilghman. 2006. Differing effects o f isolation on evolution of bird songs: examples from an island-mainland comparison o f three species. Biological Journal o f the Linnean Society 89:331-342. Bensch, S., Bengtsson, G . and S. Akesson. 2006. Patterns o f stable isotope signatures in willow warbler Phylloscopus trochilus feathers collected in Africa. Journal o f A v i a n Biology 37: 323-330. Berthold, P., Helbig, A . , Mohr, G . and U . Querner. 1992. Rapid microevolution of migratory behaviour i n a wild bird species. Nature 360: 668-670. Borge, T., Webster, M . , Andersson, G . and G Saetre. 2005. Contrasting Patterns o f Polymorphism and Divergence on the Z Chromosome and Autosomes in Two Ficedula Flycatcher Species. Genetics 171: 1861-1873. Coyne, J . A . and H . A . Orr. 2004. Speciation.  Sinauer, Sunderland, M A .  Edwards, S., Kingan, S., C a l k i n , J., Balakrishnan,.C, Jennings, W., Swanson, W . and M . Sorenson. 2005. Speciation in birds: Genes, geography, and sexual selection. Proc. Nat. A c a d . S c i . 102: 6550-6557. Grant, B.R. and P.R. Grant. 1996. Cultural Inheritance o f Song and Its Role in the Evolution o f Darwin's Finches. Evolution 50: 2471-2487: Grant, P.R. and B . R . Grant. 1997. Genetics and the origin o f bird species. Proc. Natl. Acad. S c i . 94: 7768-7775. Haesler, S., Wada, K., Nshdejan, A . , Morrisey, E., Lints, T., Jarvis, E. and C . Scharff. 2004. FoxP2 Expression in A v i a n Vocal Learners and Non-Learners. The Journal of Neuroscience 24:3164 -3175. Helbig, A . 1996.Genetic basis, mode o f inheritance and evolutionary changes o f migratory directions i n Paleartic warblers (Aves: Sylviidae). The Journal o f Experimental Biology 199: 49-55.  76  Irwin, D. E. and J. H. Irwin. 2005. Siberian migratory divides: the role o f seasonal migration in speciation. In Birds o f Two Worlds: the Ecology and Evolution o f Migratory Birds (R. Greenberg and P. P. Marra, Eds.). Johns Hopkins University Press, Baltimore, Maryland. Kroodsma, D. 1980. Winter Wren singing behavior: a pinnacle o f song complexity. Condor 82:357-356. Lovette, I. J. 2005. Glacial cycles and the tempo o f avian speciation. Trends in Ecology and Evolution 20:57-59.  Payne, R. B. 1996. Ecology and Evolution of Acoustic Communication in Birds, eds. Kroodsma, D. E. & M i l l e r , E. H.Cornell Univ. Press, Ithaca, N Y . Price, T. D. and M . M . Bouvier. 2002. The evolution o f FI postzygotic incompatibilities in birds. Evolution 56:2083-2089. Pulido, F. 2007. The Genetics and Evolution of A v i a n Migration. Bioscience 57: 165174. Saetre, G . , Borge, T., Lindroos, K., Haavie, J., Sheldon, B., Primmer, C . and A . Syvanen. 2002. Sex chromosome evolution and speciation in Ficedula flycatchers. Proc. R. Soc. Lond. B . 270: 53-59. Schluter, D. 2001. Ecology and the origin of species. Trends in Ecology and Evolution 16: 372-380. Slabbekoorn, H . and T. B . Smith. 2002. Bird song, ecology and speciation. Philosophical Transactions o f the Royal Society o f London B 357:493-503. Webb, D. and J. Zhang. 2005. FoxP2 in Song-Learning Birds and Vocal-Learning Mammals. Journal o f Heredity 96:212-216. Weir, J. T. and D. Schluter. 2004. Ice sheets promote speciation i n boreal birds. Proceedings o f the Royal Society o f London B 271:1881-1887. Wolfgang, E., Przeworski, M . , Fisher, S., L a i , C , Wiebe, V . , Kitano, T., Monaco, A . and S. Paabo. 2002. Molecular evolution o f FoxP2, a gene involved in speed and language. Nature 418: 869-872. W u , C. and A . W . Davis. 1993. Evolution o f Postulating Reproductive Isolation: The Composite Nature o f Haldane's Rule and Its Genetic Bases. American Naturalist 142: 187-212. Veen, T., Svedin, N . , Forsman, K., Hjernquist, M . , Qvarnstrom, A . , Hjernquist, K., Traff J. and Marcel Klaassen. 2007. Does migration of hybrids contribute to post-zygotic  isolation in flycatchers? Proceedings of the Royal Society B 274: 707-712.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0302422/manifest

Comment

Related Items