A Dissertation

Entitled

Evolution, systematics, and phylogeography of Ponto-Caspian gobies

(Benthophilinae: Gobiidae: Teleostei)

By

Matthew E. Neilson

Submitted as partial fulfillment of the requirements for

The Doctor of Philosophy Degree in Biology (Ecology)

______Adviser: Dr. Carol A. Stepien

______Committee Member: Dr. Christine M. Mayer

______Committee Member: Dr. Elliot J. Tramer

______Committee Member: Dr. David J. Jude

______Committee Member: Dr. Juan L. Bouzat

______College of Graduate Studies

The University of Toledo

December 2009 Copyright © 2009

This document is copyrighted material. Under copyright law, no parts of this document

may be reproduced without the expressed permission of the author.

______An Abstract of

Evolution, systematics, and phylogeography of Ponto-Caspian gobies

(Benthophilinae: Gobiidae: Teleostei)

Matthew E. Neilson

Submitted as partial fulfillment of the requirements for

The Doctor of Philosophy Degree in Biology (Ecology)

The University of Toledo

December 2009

The study of biodiversity, at multiple hierarchical levels, provides insight into the evolutionary history of taxa and provides a framework for understanding patterns in ecology. This is especially poignant in invasion biology, where the prevalence of invasiveness in certain taxonomic groups could be related to their evolutionary history.

In this dissertation, I examined the systematics, phylogeography, population genetics, and biogeography of a group of Ponto-Caspian endemic gobies that includes multiple introduced species in Europe and North America. In Chapters 2 and 4 I found highly divergent genetic lineages within two morphologically defined species (Neogobius fluviatilis and Proterorhinus marmoratus) that are widespread throughout the Ponto-

Caspian region. Statistical analyses of morphology identified significant differences according to genetic lineage within each morphospecies, indicating species level divergence among regional taxa. In Chapter 3 I constructed the phylogeny of the Ponto-

iii Caspian gobies, finding broad paraphyly in Neogobius sensu Miller & Vasil'eva (2003), and identifying a novel relationship between the “neogobiin” gobies (Babka,

Mesogobius, Neogobius, Proterorhinus, and Ponticola) and the tadpole gobies

(Anatirostrum, Benthophiloides, Benthophilus, Caspiosoma) as a distinct group highly divergent from other gobiids. I redefined the taxonomy and nomenclature according to the molecular phylogeny, and redescribed the subfamily Benthophilinae to comprise all

Ponto-Caspian endemic gobies. In all three chapters, I estimated divergence times among genetic lineages at multiple taxonomic levels, and found gross concordance between the timing of diversification events within the Ponto-Caspian gobies and major geologic changes in the evolution of the Ponto-Caspian basin. Additionally, in Chapters 2 and 4, I identified potential source locations in the northwestern Black Sea for introduced goby populations in northern/central Europe and North America. This dissertation highlights the use of molecular tools to identify cryptic genetic diversity within morphological similar taxa, as well as the role that geologic evolution of the Ponto-Caspian basin has played in shaping the diversity and evolution of its component fauna.

iv Acknowledgments

Like the proverbial child and the village required for its upbringing, a dissertation is not simply the product of a sole person's effort but results from the combined efforts of a multitude of people who, through their support (academic, financially, or emotionally), help the doctoral student usher to life this new addition to human knowledge. Many have helped me during the course of my dissertation, and my gratitude to them shows no bounds.

First and foremost is my adviser Dr. Carol Stepien, whose chance seating next to me at an American Society of Ichthyologists and Herpetologists social event started me along this journey. Thank you for providing the opportunity, support, and resources to allow me to succeed, and for always encouraging me to do so even when I thought that I could not. Thanks as well go to my advisory committee: Drs. Juan Bouzat, David Jude,

Christine Mayer, and Elliot Tramer. Your helpful comments, thoughts, and criticisms throughout the course of this work are greatly appreciated, and your multiple different viewpoints always challenged me to think differently about this work.

Many thanks go to the former and present students and staff in the Great Lakes

Genetics Laboratory: Diana Brandon, Joshua Brown, Amanda Haponski, Ian Hoffman,

Rachael Lohner, Doug Murphy, Jennifer Ohayon, Lindsey Pierce, Osvaldo Sepulveda-

Villet, and Dr. Rex Strange. All of you have provided valuable discussions and technical assistance, and furnished me with much needed camaraderie and comic relief throughout my laboratory tenure. Thanks go to Ryan Argo, Betsy Bodamer, Kristen DeVanna, Brian

Elkington, Ann Krause, Kerry McKenna, Stacy Philpott, Gwen Tenney, and Mike

Weintraub for enriching my time at the Lake Erie Center and in the department with their

v friendship, and especially to Todd Crail for his friendship and for rapidly dispelling all of my preconceptions about the flora and aquatic fauna of the Midwest.

Numerous people in both North America and Europe contributed samples to this work, including: N. Bogutskaya, V. Boldyrev, J. Brown, L. Corkum, I. Grigorovich, J.

Herler, S. Ibrahimov, H. Jenner, D. Jude, T. Kakareko, J. Kornichuk, V. Kováč, Y. Kvach,

A. Naseka, D. Pratt, S. Rudnicka, M. Sapoto, P. Simonovic, Y. Slynko, A. Smirnov, and

C. Wiesner. Access to museum specimens was graciously provided by N. Bogutskaya, G.

Burgess, J. Lungberg, D. Nelson, C. Wellendorf, and J. Williams. V. Boldyrev and E.

Vasil'eva provided assistance with specimen identification. N. Bogutskaya, W.

Eschmeyer, J. Nelson, and E. Vasil'eva provided valuable and thoughtful discussions about the phylogeny and taxonomy presented in Chapter 3, as well as a whirlwind education on the International Code of Zoological Nomenclature rules. Copious technical assistance and chili peppers were provided by Doug Murphy. I am especially indebted to Patricia Uzmann for administrative assistance, and without whom nothing would really ever get done at the Lake Erie Center.

Goby collections in Michigan and Ohio were performed under the auspices of scientific collecting permits from the Department of Natural Resources of each state.

Several analyses for Chapter 4 were carried out using the resources of the Computational

Biology Service Unit from Cornell University, which is partially funded by Microsoft

Corporation. Funding for this work was provided by a National Science Foundation grant (DEB-0456972 and REU supplement #0620942) to Carol Stepien, and NSF

DeepFin Student Exchange Program award (RCN-0443470 under G. Orti).

Thanks go to my parents and family for their love and support throughout my seemingly endless years in graduate school. Final thanks go to my wife, Laura Bedinger:

vi your constant love and support despite the time and distance between us during this dissertation is awe-inspiring, and I cannot begin to thank you enough. Much love, kudzu.

This dissertation has been brought to you by the letters G, O, B, and Y, and the number 5.

vii Table of Contents

Abstract...... iii

Acknowledgments...... v

Table of Contents...... viii

List of Tables...... xi

List of Figures...... xii

Preface...... xiii

Chapter 1: Introduction...... 1

Systematics, phylogeography, and biogeography...... 1

The Ponto-Caspian region...... 3

The Ponto-Caspian endemic gobies (Gobiidae: Teleostei)...... 5

Objectives ...... 6

Chapter 2: Evolution and phylogeography of the tubenose goby genus Proterorhinus

(Gobiidae: Teleostei): Evidence for new cryptic species...... 8

Abstract...... 8

Introduction...... 9

Methods...... 13

Specimen collection...... 13

Molecular analyses...... 13

Morphological analyses...... 19

Results...... 20

Phylogenetic and phylogeographic patterns...... 20

Morphological analyses...... 23

Discussion...... 25

viii Chapter 3: Escape from the Ponto-Caspian: Evolution and biogeography of an endemic goby species flock (Benthophilinae: Gobiidae: Teleostei)†...... 49

Abstract...... 49

Introduction...... 50

Methods...... 55

Taxon sampling...... 55

DNA analysis...... 56

Phylogenetic analyses...... 57

Divergence time estimation...... 61

Results...... 61

Discussion...... 66

Taxonomic congruency, departures, and nomenclatural changes...... 67

Relationships among Ponto-Caspian endemic gobiid groups...... 69

Higher taxonomic placement of the subfamily Benthophilinae...... 70

Biogeographic patterns...... 71

Conclusion...... 75

Systematic conclusions...... 75

Chapter 4: Historic cryptic speciation and recent colonization of Eurasian monkey gobies

(Neogobius fluviatilis and N. pallasi) revealed by DNA sequences, microsatellites, and morphology ...... 91

Abstract...... 91

Introduction...... 92

Phylogeographic and invasion history...... 95

Objectives and hypotheses...... 96

ix Materials and Methods...... 97

Sample collection, DNA extraction, and amplification...... 97

Phylogeographic analyses...... 99

Population genetic analyses...... 101

Morphological analyses...... 103

Results...... 105

Overall genetic diversity...... 105

Phylogeographic patterns...... 106

Population genetic patterns...... 107

Introduced locations in the Danube and Vistula River basins...... 109

Morphological variation between basins...... 110

Discussion...... 111

Divergence of Black and Caspian Sea lineages...... 111

Introduced populations in the Hron River and Vistula River basin...... 113

Morphological differences between monkey goby lineages...... 114

Conclusions...... 115

Chapter 5: Conclusions and Future Research...... 140

General Conclusions...... 140

Systematics of Ponto-Caspian gobies more complex than expected...... 140

Genetic divergence with little morphological change...... 142

Congruent patterns of genetic divergence across time and taxa...... 144

Future Research...... 146

References...... 149

x List of Tables

Table 2.1 Collection location, salinity (ppt), and number of tubenose goby individuals sampled for molecular analyses...... 32 Table 2.2 Primers pairs used for PCR amplification (including reaction annealing temperatures TA) and DNA sequencing of the tubenose goby...... 33 Table 2.3 Mitochondrial cytochrome b haplotype distribution for populations of the tubenose goby Proterorhinus...... 34 Table 2.4 Cytochrome b sequence divergence among major lineages of Proterorhinus...37 Table 2.5 Age estimates (million years before present) based on pairwise divergence and penalized likelihood (Sanderson, 2002) for major lineages of Proterorhinus...... 37 Table 2.6 Nested clade analysis of Proterorhinus cytochrome b haplotypes...... 38 Table 2.7 Tajima’s (1989) D test for selective neutrality for major lineages of Proterorhinus...... 39 Table 2.8 Morphometrics and meristics of Proterorhinus taxa...... 40 Table 2.9 Summary of principal components analysis of 17 linear measurements from three Proterorhinus lineages...... 41 Table 3.1 Specimen information for species analyzed...... 81 Table 3.2 Summary of maximum parsimony results...... 85 Table 3.3 Divergence times for major lineages/nodes of Benthophilinae...... 86 Table 4.1 Sampling locations of monkey goby populations...... 117 Table 4.2 Bayesian coalescent estimates of clade age for monkey goby clades...... 119

Table 4.3 Pairwise FST comparisons among monkey goby populations...... 120 Table 4.4 Morphometrics and meristics of Black (Neogobius fluviatilis) and Caspian (N. pallasi) Sea basin populations of monkey gobies...... 123 Table 4.5 Summary of principal components analysis of 24 linear measurements from Black and Caspian Sea species of monkey goby...... 125

xi List of Figures

Fig. 2.1 Freshwater tubenose goby...... 42 Fig. 2.2 Collection locations of Proterorhinus specimens...... 43 Fig. 2.3 Maximum likelihood phylogeny of Proterorhinus based on combined sequence data...... 44 Fig. 2.4 Maximum likelihood analysis of Proterorhinus cytochrome b haplotypes...... 45 Fig. 2.5 Statistical parsimony network among Proterorhinus cytochrome b haplotypes..46 Fig. 2.6. Plots of the principal components from PCA of 123 tubenose gobies...... 47 Fig. 3.1 Current range of nominal species of Benthophilinae...... 87 Fig. 3.2 Maximum likelihood phylogeny of the new subfamily Benthophilinae...... 88 Fig. 3.3 Chronogram of Benthophilinae and related Ponto-Caspian gobies...... 89 Fig. 3.4 Combined COI phylogeny of Benthophilinae and other gobies...... 90 Fig. 4.1 Current native and invasive distribution of monkey gobies...... 126 Fig. 4.2 Maximum likelihood phylogenetic analyses of monkey goby...... 127 Fig. 4.3 Statistical parsimony networks of monkey goby cyt b haplotypes...... 129 Fig. 4.4 Isolation by geographic distance among monkey goby populations...... 131 Fig. 4.5 Bayesian STRUCTURE analysis of monkey goby populations...... 133 Fig. 4.6 Spatial principal components analysis of monkey goby allele frequencies...... 134 Fig. 4.7 Plot of PC3 vs. PC2 from a principal components analysis of morphometric data from 141 monkey gobies...... 139

xii Preface

The chapters of this dissertation are organized in chronological order of the completion of each of the individual studies contained within. Chapter 2 has previously been published as:

Neilson ME, Stepien CA (2009) Evolution and phylogeography of the tubenose

goby genus Proterorhinus (Gobiidae: Teleostei): evidence for new cryptic species.

Biological Journal of the Linnean Society, 96, 664-684.

Chapter 3 has previously been published as:

Neilson ME, Stepien CA (2009) Escape from the Ponto-Caspian: evolution and

biogeography of an endemic goby species flock (Benthophilinae: Gobiidae:

Teleostei). Molecular Phylogenetics and Evolution, 52, 84-102.

Chapter 4 is currently in preparation for publication. Each chapter is largely identical to their submitted/published version, with only slight re-wordings.

In addition, the results of the systematic study in Chapter 3 necessitated a large revision of the taxonomy and nomenclature of the Ponto-Caspian endemic gobiids: the taxonomic names used throughout the manuscript reflect this new nomenclature, and the complete description of this revision is contained within Chapter 3.

xiii Chapter One

Introduction

Systematics, phylogeography, and biogeography

The study of the diversification and evolutionary history of life on Earth is one of the most profound and remarkable scientific and intellectual endeavors ever undertaken.

Since the time of Aristotle and the ancient Greeks, humans have attempted to describe, catalog, and explain local, regional, and global patterns of biodiversity. Folk taxonomies exist throughout virtually all cultures worldwide, and many show general concordance with scientific classification schemes (Berlin et al. 1966). This cross-cultural enterprise seems to result from mankind's innate desire to understand the natural world.

The study of biodiversity and the patterns and process that create it underlie several different, yet related scientific fields, each with different aims and scopes.

Systematics is the study of the evolutionary relationships within and among species and higher taxa, and the processes (e.g, speciation and extinction rates) that influence their phylogeny. Throughout the 19th and first half of the 20th centuries, systematists used comparisons of morphological characters to estimate the relationships within and among species. Although the use of morphology still continues today, an increasing number of systematic studies use molecular data, such as DNA sequences, to infer phylogeny among taxa. Molecular data have greatly enhanced our understanding of the phylogeny and

1 evolutionary history of the world's biota: since virtually all life on Earth uses DNA to encode its genetic information, we are able to compare DNA sequences from groups of organisms whose homologous morphological characters are difficult or impossible to define and measure (e.g., eukaryotes, archaens, and eubacteria). The amount of potential information present in molecular data allows us to examine the evolutionary history of

Earth's biodiversity across a wide range of taxonomic levels, from analyses of sister species to the search for the root of the Tree of Life (Avise 2004).

Population genetics lies at the opposite end of the evolutionary scale as systematics: whereas the goal of systematics is to describe the macroevolutionary relationships among species and higher taxa on historical timescales, population genetics investigates the recent, contemporary microevolutionary processes (e.g., selection, drift, gene flow) acting within and among populations, either promoting their cohesion as a single species, or driving separation and eventual speciation (Avise 2004).

Phylogeography (Avise et al. 1987; Avise 2000) sits at the interface of systematics and population genetics, and describes the genealogical linkage of individual alleles within and among closely related species in both a phylogenetic and spatial context.

Although relationships among organisms are a foundation for many other areas, such as ecology, a different approach to the study of biodiversity comes through biogeography, which tries to answer the question: why do some species/groups of plants or animals live in one region and not another? Biogeography attempts to answer this question by investigating past and present geospatial patterns in the distributions of organisms. Historical processes, such as vicariance, continental drift, and dispersal, are used to explain the distribution pattens of co-occurring taxonomic lineages both within

2 and among regions, as these historical processes would influence sympatric lineages in a congruent manner (Riddle & Funk 2004).

Comparative phylogeography (Bermingham & Moritz 1998; Arbogast & Kenagy

2001) provides a strong conceptual link between the realms of historical biogeography and systematics/population genetics. Traditionally, phylogeographic studies are focused on the geographic patterns of genetic diversity within a single species, or among closely related species (Avise 2000). Comparative phylogeography uses the spatial patterns of genetic lineages from multiple co-distributed taxonomic groups to infer general events and processes that shaped their shared history (Riddle & Hafner 2004). By utilizing common genetic patterns across multiple taxa, we can begin to identify the processes

(e.g., vicariance, landscape evolution over geologic time) or events (e.g., the opening and closing of seaways, formation of mountain ranges, environmental fluctuations) that have influenced the biota within a region.

The Ponto-Caspian region

One region of the world of particular interest is the Ponto-Caspian region, which comprises the Azov, Black, and Caspian Seas and their associated river watersheds. The geologic history of the Ponto-Caspian basin extends back to the Oligocene epoch, when the Paratethys Sea was separated from the Tethys Sea due to continental eustatic uplift in

Europe (Rögl 1999). The Paratethys Sea shared many intermittent connections with the

Tethys Sea and the World Ocean throughout the Oligocene up to the mid Miocene, where it began a longer period of isolation and was known as the Sarmatian Sea (see summary in Rögl 1999; Reid & Orlova 2002). In the late Miocene epoch, the Sarmatian Sea split into two separate basins that were precursors to the present-day Black and Caspian Seas.

3 These two primordial Ponto-Caspian basins generally evolved independently, but were also reconnected by a series of transgression and regressions both between basins as well as with the young Mediterranean Sea (Reid & Orlova 2002). Throughout the evolution of the Ponto-Caspian region, a complex series of connections and closures both between basins and with the World Ocean produced a wide range of environmental conditions, including large shifts in both salinity (Reid & Orlova 2002) and oxygen concentration

(Rögl 1999).

A primary driver of interest in the Ponto-Caspian region comes from its fauna: many species from this region constitute a major part of the introduced aquatic taxa in northern/central Europe and North America. Approximately 70% of recently introduced species within the North American Great Lakes come from the Ponto-Caspian region

(Ricciardi & MacIsaac 2000). In addition to the currently established invaders, multiple other Ponto-Caspian species are also predicted to become successfully established if introduced (Ricciardi & Rasmussen 1998; Kolar & Lodge 2002). Despite legislative and regulatory efforts to slow or stop the introduction of invasive species into the Great

Lakes, the rate of introduction is seemingly increasing over time (Ricciardi 2006). This invasional prowess exhibited by Ponto-Caspian species is likely tied to the evolution of the region itself: the geologic and environmental instability characterizing the formation of the contemporary Ponto-Caspian basins likely predisposed many species to invasiveness through the evolution of wide environmental tolerances (e.g., euryhalinity) in multiple taxa (Ricciardi & MacIsaac 2000; Reid & Orlova 2002).

4 The Ponto-Caspian endemic gobies (Gobiidae: Teleostei)

The family Gobiidae is one of the largest fish families, containing approximately

1950 species in 210 genera, and are found in freshwater, brackish, and marine systems worldwide in tropical and temperate zones, with the bulk of diversity found in tropical marine habitats (Nelson 2006). Gobies include some of the smallest and lightest vertebrates in the world (Pandaka pygmaea, Schindleria brevipinguis, and Trimmatom nanus; all are sexually mature at ~10 mm total length), as well as species that scale waterfalls 10,000X their body length during migration from juvenile to adult breeding habitats (Hawaiian amphidromous gobies). Although gobies have a rich taxonomic history, with species descriptions dating back to Linnaeus (1758), basic aspects of their biology and ecology have hindered accurate estimates of the phylogeny and systematics.

Gobies are generally small (most < 100 mm total length) and cryptic, making sampling difficult in a variety of habitats. Gobies (Gobiidae and Gobionellidae) and their closely related families (Eleotridae, Butidae, Odontobutidae, Rhyacichthyidae; sensu Thacker

2009) show a varying degree of morphological reduction and paedomorphism. This reduces the number of potential homologous morphological characters, and compounds difficulties in describing the phylogenetic relationships within and among gobioid groups as well as the placement of gobioids among other acanthomorph groups (Thacker 2003).

The endemic Ponto-Caspian gobies (Anatirostrum, Babka, Benthophiloides,

Benthophilus, Caspiosoma, Mesogobius, Neogobius, Ponticola, and Proterorhinus) represent a unique yet understudied group within the Gobiidae. The majority of the taxonomic and systematic research on the group (Pinchuk 1980; Vasil'eva 1983; Pinchuk

1991; Vasil'eva et al. 1993; Vasil'eva 1999, 2000; Ahnelt & Duchkowitsch 2001, 2004)

5 has been generally descriptive in scope, with few studies hypothesizing relationships among various limited sets of species based on qualitative differences in morphology, and no study using a phylogenetic or cladistic approach to estimate their evolutionary relationships. In addition, the Ponto-Caspian endemic gobies have not been included in recent, large-scale studies of morphological (Pezold 1993) or molecular (Akihito et al.

2000; Thacker 2003, 2009) systematics of the Gobiidae. This paucity of knowledge about the systematics of the endemic Ponto-Caspian gobiids is highlighted by the fact that this group includes multiple invasive species in both Europe and North America. To begin to understand the evolutionary forces behind invasiveness, it is necessary to first understand the evolutionary history of groups containing invasive species.

Objectives

To amend this lack of knowledge about the evolution and diversification of the

Ponto-Caspian endemic gobiids, I conducted three separate studies with overlapping focuses on systematics, phylogeography, population genetics, and biogeography. The specific objectives of this dissertation were:

1. To generate a robust, comprehensive estimate of the evolutionary

relationships among the Ponto-Caspian endemic gobies, with primary focus on

the “neogobiin” group (Babka, Mesogobius, Neogobius, Ponticola, and

Proterorhinus) that contains several invasive species.

2. To identify some of the major biogeographical events, such as geologic

evolution within the Ponto-Caspian basin, that have shaped and driven

diversification within the endemic gobiids.

6 3. To examine population genetic, phylogeographic, and morphological patterns in widespread 'morphospecies' within the Ponto-Caspian basin to identify potential cryptic taxa.

7 Chapter Two

Evolution and phylogeography of the tubenose goby genus Proterorhinus (Gobiidae: Teleostei): Evidence for new cryptic species†

† This manuscript was originally published as: Neilson ME, Stepien CA (2009) Evolution and phylogeography of the tubenose goby genus Proterorhinus (Gobiidae: Teleostei): evidence for new cryptic species. Biological Journal of the Linnean Society, 96, 664-684.

Abstract

Cryptic taxa present unique difficulties in the description of biological diversity, which DNA sequencing approaches often readily resolve. The tubenose goby

Proterorhinus – along with other Ponto-Caspian fauna – has undergone recent Eurasian range expansion, as well as colonized the North American Great Lakes in 1990. I analyzed mitochondrial (cytochrome b [cyt b] and cytochrome c oxidase subunit I [COI]) and nuclear (recombination activating gene 1 [RAG1]) DNA sequences and morphological characters from non-indigenous Great Lakes as well as introduced and native Eurasian populations of Proterorhinus marmoratus (Pallas) sensu lato in order to assess their species identity and biogeographic patterns. Results show marked genetic and morphological divergence that indicates species-level separation between fresh water and marine/brackish lineages, dating to ~3.82-4.30 million years. In addition, freshwater lineages within the Black and Caspian Sea basins show significant genetic and morphological differentiation, corresponding to an estimated 0.92-1.03 million years

8 separation. I describe new evidence to support at least three separate species: the original

P. marmoratus in marine and estuarine habitats within the Black Sea, a freshwater species in the Black Sea basin that was introduced to the North American Great Lakes, and another freshwater species inhabiting the Caspian Sea/Volga River basin. The freshwater tubenose goby in the Black Sea basin originally was described as P. semilunaris (Heckel), and is confirmed here as a valid taxon. The Caspian basin taxon may correspond with P. semipellucidus (Kessler), a putative freshwater species in the Caspian basin that was originally described from a single specimen.

Introduction

Cryptic species are evolutionarily distinct yet morphologically indistinguishable from their relatives, which often precludes correct diagnosis and poses a fundamental problem in describing biological diversity. Morphology traditionally has been the primary means of identifying species; however, other types of information are essential to identify cryptic taxa in the absence of anatomically distinguishing characters (Bickford et al. 2007; Kon et al. 2007). Genetics, behavior, and natural history all have been used to supplement morphology in defining species-level diversity among organisms, with DNA characters revolutionizing systematics in the past 2 decades (Hillis 1998; Johnson et al.

2004; Egge & Simons 2006).

Cryptic divergence, in the absence of apparent morphological separation, has been discovered within a broad range of marine and freshwater taxa (Knowlton 1993, 2000).

Molecular data – such as DNA sequences – often reveal deep genetic divergences within taxa that once were regarded as synonymous, due to little morphological variation

(Knowlton 2000; Gómez et al. 2002; Ślapeta et al. 2006). Several recent introductions of

9 cryptic or previously undescribed species have been detected through genetic analyses of invasive sister taxa, e.g., the green crab Carcinus aestuarii in areas thought to contain its more widespread sister taxon C. maenas (Geller et al. 1997), and introduction to the

Mediterranean Sea of genetically distinct lineages of hardyhead silverside fish

Atherinomorus lacunosus from the Red Sea (Bucciarelli et al. 2002). In cases of invasive or introduced species, misidentification may mask our ability to predict their potential impacts in invaded areas, precluding correct ecological comparisons (Ruiz et al. 1997).

The tubenose goby (Proterorhinus; family Gobiidae), along with members of the closely related genera Apollonia and Neogobius, is part of a growing number of Ponto-

Caspian taxa introduced outside of their native range (Black/Caspian Sea basins, see Fig.

2) into western and central parts of Europe as well as to North America (Ricciardi &

MacIsaac 2000). Proterorhinus was discovered in the North American Great Lakes in

1990 (Jude et al. 1992), attributed to ballast water exchange from transoceanic shipping vessels, and now is limited to the St. Clair and Detroit Rivers, and the western margins of

Lakes Erie, St. Clair, and Superior. In central Europe, Proterorhinus expanded up the

Danube River during the past 2 decades via shipping and the construction of canals

(Ahnelt et al. 1998; Prášek & Jurajda 2005). Naseka et al. (2005) suggested that its recent expansion throughout the Volga River originated from the Don River (through the

Volga-Don Canal) rather than from the lower Volga River and Caspian Sea. To assess the ecology of these introductions and identify source and non-native populations, it is essential to understand their correct species identity.

Proterorhinus was described by Pallas (1814) as Gobius marmoratus from

Sevastopol, Ukraine, and is distinguished from other Ponto-Caspian gobiids by tubular

10 nostrils that extend onto or beyond the upper lip (Fig. 2.1). Tubenose gobies are moderately large gobiids (up to 11 cm total length) with infraorbital neuromast organs in seven transverse rows, and complete oculoscapular and preopercular lateral line canals; scaled nape and breast with cycloid scales, cheek naked; and overall body coloration yellow to light brown with 5-6 irregular dark brown blotches.

Several tubenose goby species were once recognized, including four marine and three freshwater taxa, of which the latter were: Gobius semilunaris Heckel 1837 from the

Maritsa River, Bulgaria; G. rubromaculatus Kriesch 1873 from the Danube River near

Budapest, Hungary; and G. semipellucidus Kessler 1877 from the mouth of the Karasu

River, in Astrabad (Gorgon) Bay, Iran. Smitt (1900) provided the first usage of

Proterorhinus as a subgenus of Gobius, which Berg (1916) elevated to generic status.

Berg (1949) later synonymized all Proterorhinus species as P. marmoratus except for P. semipellucidus, which is currently regarded as a synonym of P. marmoratus (Pinchuk et al. 2003b).

Recently, Stepien et al. (2005) and Stepien & Tumeo (2006) found marked genetic divergence between freshwater and marine specimens of P. marmoratus with mitochondrial (mt) cytochrome (cyt) b sequence data, concluding that they comprise separate species and suggesting P. semilunaris be resurrected for the freshwater clade in the Black Sea basin and North America. Freyhof & Naseka (2007) described a new freshwater species, P. tataricus, from the Chornaya River in the Crimean Peninsula near

Sevastopol, Ukraine using morphological characters; however, no genetic data or modern statistical, phylogeographic, or evolutionary analyses were presented. It is distinguished from P. marmoratus by greater number of mid-lateral scales and wider interorbital

11 distance, and from the Black Sea freshwater P. semilunaris (following Stepien & Tumeo

2006) in having a shorter and deeper head and more second dorsal fin rays. Freyhof &

Naseka (2007) also concluded that “marine” specimens from the Caspian Sea basin likely constituted another separate species, suggesting resurrection of P. nasalis (originally described as G. nasalis by Filippi, 1863).

This study expands on the work of Stepien et al. (2005) and Stepien & Tumeo

(2006) by further investigating the divergence between freshwater and marine

Proterorhinus, with goals of describing their evolutionary and phylogeographic history and to evaluate whether there are additional cryptic taxa. This investigation analyzed

DNA sequence data from two mt genes (cyt b and cytochrome oxidase c subunit I) and a nuclear gene (recombination activating gene 1) to more fully reconstruct phylogenies, estimate divergence times, and describe spatial patterns in genetic variation within and among tubenose gobies from 18 locations throughout much of their known native and introduced ranges. I included locations within the Caspian Sea basin and the Kumo-

Manych Depression (a geologic depression separating the Russian Plain and the northern foothills of the Caucasus Mountains) that were not sampled in prior work, as well as the type locality of P. marmoratus at Sevastopol, Crimea, Ukraine. In addition, I assessed morphological characters of freshwater and marine specimens using multivariate ordination techniques to describe and quantify potential morphological separation. I thus provide the first analysis of the genus Proterorhinus using both multilocus DNA sequences and modern statistical survey of morphology.

12 Methods

Specimen collection

I analyzed specimens obtained by small beach seine or beam trawl from 18 freshwater and marine localities throughout much of the native and introduced range of

Proterorhinus (Fig. 2.2; Table 2.1), with additional museum material used for morphology. Specimens were preserved immediately either in 95% ethanol for molecular analyses or 10% formalin for morphological analyses (following removal of right pectoral fin for genetic study).

Molecular analyses

Genomic DNA was isolated using a Qiagen DNEasy tissue extraction kit

(Valencia, CA) following manufacturer’s protocols. Two mitochondrial (mt) genes

(cytochrome b – cyt b; cytochrome oxidase c subunit I – COI) and a nuclear gene

(recombination activating gene 1 – RAG1) were amplified via the polymerase chain reaction (PCR) using primers listed in Table 2.2. Use of independent loci from both the nuclear and mitochondrial genomes in a combined analysis can aid in the estimation of the true “species tree” by searching for common relationships among lineages in individual gene trees (Avise 2000), maximizing potential inference power. Different genes evolve at different rates, and thus are useful in resolving multiple hierarchical levels within a phylogeny or in estimating separation at different temporal scales

(Quenouille et al. 2004).

PCR amplifications were performed in 25 μL volumes containing: 10 mM Tris-

HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl2 (2.5 mM for COI), 0.001% (w/v) gelatin, 200

13 μM each dNTP, 0.5 μM each primer, 1.5 units Taq polymerase, and ~100 ng (1-3 μL) template DNA. The PCR profile for cyt b and RAG1 included initial denaturation at

94°C for 2 min; 40 cycles of 94°C for 45 s, a gene specific annealing temperature (Table

2) for 30 s, and 60 s extension at 72°C; with a 3 min final extension at 72°C. The cycling profile for COI included initial 3 min denaturation at 94°C; 35 cycles of 94°C for 30 s,

53°C for 30 s, and 72°C for 60 s; with a final 2 min extension at 72°C. PCR reactions were visualized on 1% agarose gels with ethidium bromide, with excess primers and unincorporated nucleotides removed with spin columns (QIAquick PCR Purification Kit,

Qiagen; or QuickStep 2 PCR Purification Kit, Edge Biosystems, Gaithersburg, MD).

Amplicons were sequenced in both directions using dye-labeled terminators on an ABI

3730 (Applied Biosystems, Foster City, CA) genetic analyzer at the Cornell University

Life Sciences Core Laboratories Center. I aligned forward and reverse sequences for each gene per individual with BIOEDIT (Hall 1999).

I employed both phylogenetic and population genetic approaches to evaluate variation in Proterorhinus. Phylogenetic analyses compared variation at three gene loci in P. marmoratus (sensu lato) with representatives of the most closely related genera

(Stepien et al. 2005; Stepien & Tumeo 2006), including the monkey goby Apollonia fluviatilis (Pallas, 1814; formerly Neogobius fluviatilis per Stepien & Tumeo, 2006), round goby A. melanostoma (Pallas, 1814; formerly N. melanostomus per Stepien &

Tumeo, 2006), knout goby Mesogobius batrachocephalus (Pallas,1814), racer goby

Neogobius gymnotrachelus (Kessler 1857), and bighead goby N. kessleri (Günther 1861).

The population genetic approach analyzed variation at the cyt b locus to elucidate fine- scale biogeographic patterns and processes. The phylogenetic approach included a

14 representative from each Proterorhinus population, whereas the population genetic approach included all collected individuals. All sequences were deposited in GenBank

(http://www.ncbi.nlm.nih.gov).

The phylogenetic analyses centered on testing the validity of distinct marine and freshwater Proterorhinus lineages, and evaluated relative support from independent lines of evidence (mt and nuclear DNA) in separate analyses and a total DNA combined evidence approach. Several phylogenetic approaches, having different evolutionary assumptions, were used to evaluate the evolutionary hypothesis (Avise 2004): maximum parsimony (PAUP* 4.0b10; Swofford 2003), maximum likelihood (PHYML v2.4.4;

Guindon & Gascuel 2003), and Bayesian (MRBAYES v3.1.2; Ronquist & Huelsenbeck

2003).

Parsimony analyses were performed with branch and bound searches, and branch support was evaluated via nonparametric bootstrapping (1,000 replications) and decay indices. For likelihood and Bayesian analyses, MODELTEST v3.7 (Posada & Crandall

1998) determined the simplest best-fit model for the dataset using the Bayesian information criterion (Posada & Buckley 2004). For the cyt b data, the best-fit model was HKY+I+G with a shape parameter (α) of 1.1964 and a proportion of invariant sites

(i) of 0.4805; for COI, the best-fit model also was HKY+I+G (α = 1.1987; i = 0.57); and for RAG1, HKY+G (α = 0.0176). Bayesian analyses using Metropolis coupled Markov chain Monte Carlo (MCMCMC) sampling were run for 5 million generations, with sampling every 1,000 generations, to assure convergence of likelihood values. Four separate chains were run in each of two simultaneous analyses, and burn-in period was determined by plotting log likelihood values at each generation to identify the point

15 where values reached stationarity, which occurred after 200,000 generations. I chose a conservative burn-in period of 1 million generations, and discarded prior trees and parameter values. Branch support was calculated from 1,000 bootstrap replications for likelihood analyses and via the posterior probability distribution of clades for Bayesian analyses.

I performed a partition homogeneity test to determine combinability of the three molecular datasets, using reduced numbers of marine and freshwater Proterorhinus (four sequences from each lineage) to minimize phylogenetic noise associated with highly similar sequences (Dolphin et al. 2000) and removing uninformative characters

(Cunningham 1997). No significant incongruence was found among the three genes

(1,000 replicates; P = 0.158): all were combined for simultaneous analysis using search strategies identical to the separate analyses, with GTR+I+G (α = 0.819; i = 0.60) selected as the best-fit model. A single model approach was used for the maximum likelihood analysis, whereas a partitioned mixed-model approach was used for Bayesian analysis.

The models of sequence evolution identified for each individual gene region were assigned using the APPLYTO command, and the appropriate model parameters were estimated for each gene using the UNLINK command.

Population genetic analyses focused on describing fine-scale genetic variation across the current range of Proterorhinus, using cyt b sequences across all individuals. A second maximum likelihood search was performed including all discovered haplotypes, the outgroups from the original phylogenetic analysis, and additional gobioid outgroup taxa (Gobiidae: Amblyopinae – Taenioides limicola [AB021253]; Gobiinae – Gobiosoma bosc [AY848456]; Gobionellinae – Gymnogobius petschiliensis [AY525784];

16 Oxudercinae – Periophthalmus argentilineatus [AB021251]; Eleotridae: Butinae – Butis amboinensis [AB021232]; Eleotrinae – Eleotris fusca [AB021236]; Rhyacichthyidae –

Rhyacichthys aspro [AP004454]). The search strategy (including bootstrap analysis) was identical to that in the initial phylogenetic analyses, except that a slightly more complex model was chosen for this extended cyt b dataset (TrN+I+G; α = 0.9223; i = 0.4706).

To identify boundaries among evolutionary lineages, I used Wiens & Penkrot's

(2002) tree-based method, which evaluates concordance between geography and a haplotype phylogeny to identify gene flow or isolation among populations of a focal species. Deep divergences among geographically discordant lineages can indicate putatively independent ones hidden within a taxon (Wiens & Penkrot 2002). The null hypothesis of lineage monophyly as a chance outcome of random branching processes was tested following Rosenberg (2007). Tajima (1989) D test, implemented in

ARLEQUIN 3.11 (Excoffier et al. 2005), examined whether patterns of variation within

Proterorhinus fit a hypothesis of neutrality.

To estimate divergence times, I calculated pairwise sequence divergence (using the TrN model above) within and among major Proterorhinus lineages in MEGA v4

(Tamura et al. 2007), and using a penalized likelihood approach in R8S 1.71 (Sanderson

2003). I utilized an average pairwise sequence divergence rate for cyt b of 2.05% per

Myr estimated from sister species of Evorthodus gobies by Rocha et al. (2005).

Divergence times were further examined with a penalized likelihood (Sanderson 2002) approach in R8S using an initial age estimate generated under a molecular clock assumption, from which the sequences significantly departed. I conducted a second analysis using penalized likelihood: a semiparametric approach incorporating a roughness

17 penalty to constrain autocorrelation in rate variation between ancestor and descendent branches, with the optimal smoothing parameter determined by cross-validation in R8S.

Divergence time estimates under penalized likelihood require a fixed age for at least one node within the phylogeny. Additional outgroups for the maximum likelihood and age estimation analyses included a node representing the family Gobiidae, set to 53 My for the penalized likelihood analyses as its oldest known fossils date to 51-56 Mya (Bajpai &

Kapur 2004). In addition, Rückert-Ülkümen (2006) described fossil otoliths of

Neogobius as dating to the late Miocene-early Pliocene (~10 Mya). As otoliths of

Apollonia, Neogobius, and Proterorhinus are similar, I thus used a conservative approach and set the age of their most recent common ancestor to 10 My.

Nested clade analysis (NCA) further explored phylogeographic patterns among lineages, testing the association of haplotypes with geographic location (Templeton

1998). A statistical parsimony network was created using TCS v1.21 (Clement et al.

2000); haplotypes were nested following Templeton (1998), and ambiguities or reticulations in the parsimony network were resolved according to Pfenninger & Posada

(2002). For networks that could not be connected with statistical parsimony, the maximum likelihood tree was employed as a guide and grouped individual networks as sister clades at equal nesting levels, with outgroup rooting (using nearest sister group from phylogenetic analyses) determining the tip/interior status of a clade. GEODIS v2.5

(Posada et al. 2000) tested for a significant association of haplotypes and geography, and the inference key of Templeton (2004) was used to identify the likely cause(s) of association for significant clades.

18 Morphological analyses

I quantified morphological variability within and among freshwater and marine tubenose gobies using meristic and mensural characters. All counts and measurements used a Leica MZ-12.5 dissecting microscope, and mensural data (to 0.01 mm) employed vernier calipers. Meristic data included numbers of first dorsal spines, second dorsal and anal fin elements, pectoral fin segmented rays, and lateral scale rows (posterodorsal tip of opercle to base of caudal fin). Measurements were standard length (tip of snout not including lower jaw to midpoint of caudal fin base), head width (maximum width at preopercular margin), head depth (maximum depth at posterior dorsal head margin), head length (tip of snout not including jaw to posterodorsal tip of opercle), eye diameter

(horizontal diameter), snout length (tip not including jaw to anterior eye margin), interorbital width (least distance between left and right orbits), caudal peduncle length

(posterior end of anal fin base to midpoint of caudal fin base), minimum caudal peduncle depth, caudal peduncle width (at minimum depth), pectoral fin length (insertion of longest fin ray to tip), maximum body depth (at anterior margin of first dorsal fin), maximum body width (behind pectoral fin base), pelvic disc length (insertion to posteriormost point), preorbital distance (distance between lip and orbit), and abdomen length (insertion of pelvic fin to vent).

I performed multivariate analyses in R (R Development Core Team 2009) to determine if freshwater and marine lineages of tubenose gobies are morphologically distinguishable. A principal components analysis (PCA) on ln-transformed measurements separated morphological variation into linear combinations of variables that describe overall body size and shape variation among lineages. The first principal

19 component (PC1) primarily describes body size variation, whereas the remaining components encompass body shape variation. I utilized the components that explained

95% of the morphological variance (PC1-4) in further analyses. I tested the hypotheses that freshwater and marine tubenose gobies differ in a) body size and shape, and b) body shape alone, using multivariate analyses of variance (MANOVA), with PC1-4 as dependent variables in the former, and the latter with the shape components alone (PC2-

4). Differences in lineage mean score for each principal component were assessed with

ANCOVA (allometric components) or ANOVA (non-allometric components).

Results

Phylogenetic and phylogeographic patterns

The aligned dataset for the combined three gene regions for 18 marine and freshwater specimens of Proterorhinus, spanning its range, as well as five neogobiin gobies and a gobiin outgroup contained 3,968 bp (cyt b – 1141 bp; COI – 1271 bp; RAG1

– 1556 bp). Cyt b sequences are available as GenBank accession nos. EU331208,

EU444604, EU444607, EU444610-612, EU444618, EU444620-21, EU444624,

EU444626, EU444630, EU444632, EU444636, EU444649, and EU444667-72; COI sequences as EU444673-98; RAG1 sequences as EU444699-724. Nucleotide compositions for each gene are stationary across taxa (χ2 > 5.62; df = 75; p > 0.99).

Phylogenetic analyses of the three-gene dataset are highly congruent among tree- building methods and well-supported (Fig. 2.3). Maximum parsimony of the three-gene dataset found 48 most parsimonious trees of 2,089 steps (CI = 0.648, RI = 0.790, RC =

0.512, HI = 0.352). The majority rule consensus is well-resolved, with high bootstrap

20 support for each species (many at or near 100%) and generally high decay indices.

Marine and freshwater clades of Proterorhinus have 100% bootstrap support and large decay indices (freshwater clade = 60; marine clade = 113). Maximum likelihood and

Bayesian analyses of the combined dataset likewise show a similar topology to the consensus parsimony tree, as well as high support for relationships (100% likelihood bootstrap support; 1.00 posterior probability) among all species as well as the marine and freshwater clades of Proterorhinus (Fig. 2.3). Primary differences among the analyses occur in the terminal branching order of individuals within each of the major lineages; branching order among species as well as among major lineages of Proterorhinus are identical across analyses.

The freshwater Proterorhinus clade comprises two primary lineages: one from the

Caspian Sea basin; the second from freshwater Black Sea basin locations, along with a single specimen from the Kumo-Manych Depression (AMN1 - Fig. 2.3; population U –

Table 1). The freshwater Caspian Sea clade has very high support in all analyses

(parsimony bootstrap [MPBS] = 100, decay index [DI] = 14, likelihood bootstrap

[MLBS] = 100, posterior probability [BPP] = 1.00). The clade from freshwater Black

Sea locations + Kumo-Manych Depression also is highly supported (MPBS = 71, DI = 0,

MLBS = 67, BPP = 1.00), as well as a smaller subclade limited to freshwater Black Sea locations (MPBS = 71, DI = 0, MLBS = 81, BPP = 1.00; Fig. 2.3). Individual analyses of the separate genes (not shown) largely are congruent with the total evidence analyses, again differing primarily in the branching order of the shallowest nodes. In each mitochondrial gene (cyt b and COI), marine and freshwater specimens of Proterorhinus separate into two distinct, robust clades (MPBS > 83, DI > 8, MLBS > 97, BPP = 1.00).

21 Within the freshwater clade, distinct Black and Caspian Sea lineages are resolved with cyt b and are strongly supported (Black Sea lineage – MPBS > 98, DI > 7, MLBS > 98,

BPP = 1.00; Caspian Sea lineage – MPBS > 98, DI > 8, MLBS > 98, BPP = 1.00), whereas only the Caspian Sea lineage is resolved with the COI gene data (MPBS > 99,

DI > 11, MLBS > 96, BPP = 0.94). The relationships among Proterorhinus and other

Ponto-Caspian neogobiins thus are generally congruent with prior studies of neogobiin systematics (Stepien et al. 2005; Stepien & Tumeo 2006): Proterorhinus is closely related to Mesogobius, with a Proterorhinus + Mesogobius clade comprising the sister group to

Neogobius, and Apollonia (a recently elevated genus containing A. fluviatilis and A. melanostoma) as a sister clade to all other neogobiins.

Maximum likelihood analysis of cytochrome b sequences from 151 individuals in 18 population sites (EU444604-EU444666; Fig. 4) show a pattern similar to that of the larger phylogenetic analysis (Fig. 2.3). Tree-based lineage delimitation identifies three phylogenetically and geographically distinct haplotype clades, representing independent evolutionary lineages. Freshwater and marine clades of Proterorhinus show marked phylogenetic divergence with 100% bootstrap support. Within the freshwater clade, among-basin differences (Black Sea, Caspian Sea, Kumo-Manych Depression) also have strong support (>90%). Although there is strong support for the separation of the Black

Sea freshwater clade and the Kumo-Manych haplotype, the paucity of samples (two individuals, one haplotype) from the latter area excludes it from further analyses.

Primary lineages have large ratios of between- to within-lineage molecular divergences, indicating pronounced evolutionary separation (Table 2.4). Tamura-Nei distances among

22 lineages range from 0.039 to 0.173, whereas those within lineages are 0.002-0.007 (Table

2.4). Average divergence between marine and freshwater lineages is 0.167.

Estimated divergence times within and among the marine and freshwater lineages of Proterorhinus (Table 2.5) range from 0.05 Mya (marine lineage) to 6.18 Mya

(divergence of Proterorhinus and Mesogobius). Pairwise divergence estimates among lineages appear generally lower than those using penalized likelihood; however, age estimates from both methods are largely congruent (Table 2.5). Statistical tests of monophyly (Rosenberg 2007) between marine and freshwater lineages as well as between

Black and Caspian basin freshwater lineages are highly significant (P << 0.001 for both tests) indicating that the lineages are distinct taxa.

Nested clade analysis is consistent with tree-based lineage delimitation (Fig. 2.5,

Table 2.6). Proterorhinus lineages are strongly associated with geographic location, reflecting allopatric fragmentation (Fig. 2.5, Table 2.5). Within P. semilunaris (clade 4-3;

Fig. 2.5, Table 2.6), a significantly large interior-tip distance suggests either isolation by distance or long distance dispersal between the Danube and Dnieper River populations.

Range expansion is inferred for P. marmoratus (clade 3-1; Table 2.5, Fig. 2.5), congruent with Tajima’s (1989) D test (Table 2.7).

Morphological analyses

The three major clades of Proterorhinus broadly overlap in morphometric and meristic characters (Table 2.8). Species means differed slightly, however, for several morphometric characters, including: head length, maximum head depth, maximum body depth, and snout length (Table 2.8). These reveal subtle difference in body shape, with the marine lineage being more robust and deep-bodied than the more elongate freshwater

23 lineages. The principal components (PC) analysis further examined difference in body shape; the first four PC explain 95.8% of the variance (PC1 – 90.1%; PC2 – 3.0%; PC3 –

1.4%; PC 4 – 1.3%), and I thus restricted further analyses to these. PC1 primarily described overall size variation, showing high correlation with standard length (r =

-0.971; P < 0.01) and approximately equal loadings on all 17 morphometric variables except interorbital distance and preorbital width (Table 2.9). PC2-4 had low correlation with standard length (|r| ≤ 0.175; P > 0.05 for all components), and represents size- independent shape variation. The strongest influences on PC2-4 were measures of overall head shape (interorbital distance, preorbital width, and snout length) and body shape (body depth, caudal peduncle width and depth). Visual inspection of the principal components showed that although the three clades overlap, there are differences among their mean scores (i.e., different centroids; Figure 2.6). Multivariate analysis of variance

(MANOVA) using both body size and shape information (PC1-4) detected significant difference among the three lineages (Wilks’ λ = 0.416, F8, 234 = 16.128, P << 0.001); body shape information (PC2-4) alone recovered a slightly larger difference (Wilks’ λ = 0.447,

F6, 236 = 19.498, P << 0.001). Univariate ANCOVA for PC1 with SL as the covariate indicated significant differences in lineage mean score (P < 0.001), and Tukey’s HSD post-hoc comparisons identified differences between P. cf semipellucidus and both P. marmoratus and P. semilunaris (P < 0.001 for both comparisons), with no significant difference in mean PC1 score between P. marmoratus and P. semilunaris (P = 0.85).

Univariate ANOVAs for the remaining components discern significant difference among lineages for mean scores on PC2 (P < 0.001) and PC4 (P = 0.015), with no significant difference for PC3 (P = 0.401). Post-hoc comparison tests for PC2 indicate significant

24 differences between P. semilunaris and both P. marmoratus and P. cf semipellucidus (P <

0.001 for both comparisons), with no difference between P. marmoratus and P. cf semipellucidus (P = 0.467). Post-hoc comparisons for PC4 depict a significant difference between P. marmoratus and P. semilunaris (P = 0.019) alone. In general, P. semilunaris has a wider interorbital diameter, wider caudal peduncle, and a longer snout than P. marmoratus and P. cf semipellucidus, whereas P. cf semipellucidus has a shallower body than P. marmoratus and P. semilunaris..

Discussion

The conceptual foundation of what constitutes separate taxa is of fundamental importance in evolutionary studies. Numerous species concepts are described in the literature (Coyne & Orr 2004; de Queiroz 2005), differing primarily in the theoretical definition used to identify biological species and the operational criteria used to delimit them in nature. Whereas most researchers have an intuitive notion of their own personal species concept, few studies explicitly state the criteria or concepts used to recognize distinct taxa. Two species concepts predominate systematic studies: the evolutionary species concept (ESC: sensu Wiley & Mayden 2000) and the phylogenetic species concept (PSC: sensu Mishler & Theriot 2000). The ESC defines a species as a group of organisms with its own independent evolutionary trajectory separate from others in space and time. The PSC defines a species as the least inclusive unit within a phylogenetic classification as evidenced by monophyly. Whereas the ESC provides a sound theoretical definition for what constitutes a species (independent evolutionary lineages), the PSC provides robust operational criteria for delimiting among them (monophyly following

Mishler & Theriot 2000; i.e., diagnosable combinations of characters/synapomorphies, or

25 recent genetic coalescence in other versions of the PSC). I thus use this approach to delimit species of Proterorhinus, coupled with the additional operational threshold to identify species of interspecific variation ≥ 10x mean intraspecific variation (Hebert et al.

2004b).

These results describe three distinct evolutionary lineages of tubenose gobies: a marine lineage comprising P. marmoratus from the Black Sea proper, and two freshwater lineages in the Black (P. semilunaris) and Caspian (P. cf semipellucidus) Sea basins. A potential fourth lineage within the Kumo-Manych Depression also is suggested, which merits further sampling and investigation. Appreciable cytochrome b genetic divergence occurs among the three lineages (Tamura-Nei distance = 0.04 – 0.17), within the range seen for many intrageneric comparisons of fishes (0.01-0.40; Johns & Avise 1998b) as well as in the neogobiin genus Apollonia (Tamura-Nei distance between N. fluviatilis and

N. melanostomus = 0.174; Brown & Stepien 2008), with coincident low divergence within lineages (0.001 - 0.006) similar to intraspecific variation in other gobiids (Harada et al. 2002; Mukai et al. 2004; Brown & Stepien 2008). Monophyly of these lineages is highly significant (P << 0.0001), confirming that they are distinct taxa.

Divergence times among the major lineages of Proterorhinus estimated both with

(pairwise divergence) and without (penalized likelihood) a molecular clock are largely congruent. As the ages specified for the oldest fossils of the Gobiidae and Neogobius likely underestimate their true ages, the approximations for Proterorhinus are minimum estimates. Utilizing estimated divergence times, I reconstructed the evolutionary history of Proterorhinus within the Ponto-Caspian basin in relation to the region’s geologic history. Names for the historic stages of the Black and Caspian Sea basins follow Reid &

26 Orlova (2002). A marine ancestor of Proterorhinus and Mesogobius likely inhabited the

Sarmatian Sea, which encompassed both the Black and Caspian Sea basins ~8-15 Mya during the mid- to late Miocene epoch. At this time, the Pontian Lake-Sea in the Black

Sea basin was connected to a large water body located in the Pannonian Depression

(geologic depression west of the Ponto-Caspian basin and Carpathian Mountains in present day Serbia and Romania), merging their two faunas (Reid & Orlova 2002).

Proterorhinus and Mesogobius then diverged ~4.8-6.2 Mya in the Pontian Lake-Sea in the Black Sea basin, perhaps due to specialization and competition with Pannonian taxa.

The first major division within Proterorhinus occurred ~4.2-4.4 Mya, separating the marine and freshwater taxa. Their early Pliocene divergence likely resulted from decrease in salinity of the Kimmerian Lake-Sea (proto-Black Sea basin), leading to an overall shift from marine to brackish water fauna (Zaitsev & Mamaev 1997; Reid &

Orlova 2002). The separation between the Black and Caspian Sea freshwater lineages occurred during the mid-early Pleistocene epoch (~0.94-1.15 Mya) when freshwater

Proterorhinus moved along with other Black Sea basin fauna from the Gurian Lake-Sea in the Black Sea basin into the Apsheron Lake-Sea in the Caspian Sea basin. This migration occurred across the Kumo-Manych Depression during the Aspheronian transgression: the second major Pleistocene connection between the Black and Caspian

Sea basins (Reid & Orlova 2002; Cristescu et al. 2003). This transbasin connection persisted intermittently for the last ~0.9 Myr due to fluctuating water levels associated with major glacial and interglacial periods (Reid & Orlova 2002), with the most recent connection closing ~9,000 ya. Separation of the waterbodies remaining within the depression after these repeated openings and closures likely led to the divergence of the

27 fourth putative freshwater lineage of Proterorhinus from P. semilunaris ~0.68-0.72 Mya.

Ages estimated for the individual marine, Black Sea freshwater, and Caspian Sea freshwater clades (Tables 2.4-2.5) likely reflect population size changes associated with late Pleistocene glaciations and fluctuations in water levels, indicated by the significant

Tajima’s (1989) D value for P. marmoratus (Table 2.7).

The separation among Black and Caspian Sea lineages of Proterorhinus echoes the pattern seen in numerous Ponto-Caspian organisms. Brown & Stepien (2008) describe a ~1 Myr separation between Black and Caspian Sea lineages of N. melanostomus, whereas an older separation (~4 Mya) characterizes its sister species N. fluviatilis between those basins (Chapter Four). Durand et al. (1999) estimated a

Pliocene divergence between Danube River and Caspian Sea lineages of the chub

Leuciscus cephalus. Cristescu et al. (2003, 2004) found a similar pattern of genetic structure in several Ponto-Caspian crustaceans, with differentiation times between Black and Caspian Sea lineages ranging from ~1 Myr for benthic species and 6-8 Myr for planktonic species. Stepien et al. (2002) and Gelembiuk et al. (2006) estimated more recent separation time (~532,000 and 166,000 years, respectively) between Black and

Caspian Sea populations of Dreissena polymorpha. These repeated patterns of genetic divergence between Black and Caspian Sea basin lineages highlight the role of intermittent basin connectivity/separation in shaping the evolutionary history of Ponto-

Caspian taxa.

Morphological separation of the three major clades of Proterorhinus is not as marked as their genetic divergence: all overlap for individual morphometric variables, yet differ slightly in mean values for several characters (Table 2.8). Some of these characters

28 likely contribute to the significant overall differences in body shape observable among the taxa (Fig. 2.6 and Table 2.9). Although overlapping principal component scores indicate that morphometrics alone cannot classify specimens to individual taxa, a significant MANOVA (and subsequent univariate ANOVAs) for PC1-4 showed that the three major lineages represent three distinct statistical populations, occupying different areas in morphospace. Habitat specific (marine vs. freshwater) morphological differences have been observed in other fishes, including sticklebacks (Gasterosteus aculeatus complex; Walker & Bell 2000) and sea bass (Dicentrarchus labrax; Corti et al.

1996). In addition, significant genetic divergence among morphologically cryptic species is widespread in fishes. For example, Egge & Simons (2006) described the morphologically cryptic North American madtom catfish Noturus maydeni using a combination of karyology, allozymes, and fixed DNA sequence differences. Bonefishes, a circumtropical species complex originally described as a single species (Albula vulpes), contain eight morphologically cryptic species with high degree of genetic divergence

(genetic distance = 5.56-30.6%; Colborn et al. 2001). Quattro et al. (2006) described large phylogenetic separation, as well as differences in vertebral counts, between the widespread scalloped hammerhead (Sphyrna lewini) and a cryptic species in the western

North Atlantic Ocean. Despite large differences in ecology and taxonomy, these three examples all exhibit a common pattern: high degree of genetic divergence in the absence of significant morphological divergence among cryptic taxa.

The combined morphological and molecular separation of geographic lineages of

Proterorhinus indicates that there are at least three separate species of Proterorhinus: P. marmoratus from marine habitats within the Black Sea, P. semilunaris from freshwater

29 drainages in the Black Sea basin and introduced populations in the North American Great

Lakes, and P. cf semipellucidus from the upper Don River and Volga River basin.

Freyhof & Naseka (2007), in a description of a new freshwater species of Proterorhinus

(tataricus) from the Crimean Peninsula, additionally found that Caspian Sea basin tubenose gobies appear distinct from all other species (P. marmoratus, P. semilunaris, and

P. tataricus) and ascribed those populations to P. nasalis. However, the specimens examined by Freyhof & Naseka (2007) were derived primarily from marine regions of the Caspian Sea proper and included only two specimens from the Volga River at

Zam’yany (which also was sampled here). Moreover, they did not incorporate genetic data or employ modern statistical analysis of their morphological data. Given the morphological and genetic differences observed in this study between P. marmoratus and

P. semilunaris, as well as strong salinity differences between the northern and southern

Caspian Sea basins (0-2‰ in the northern reaches near the Volga River delta versus 12-

14‰ in the south Sea), it is likely that the Caspian Sea basins house separate freshwater and “marine” Proterorhinus taxa. Thus, Caspian Sea watersheds presumably contain a freshwater taxon in the Volga River and other northern drainage basins and a second

“marine” taxon in the more saline southern basin (south of the Baku Peninsula,

Azerbaijan). The earliest assigned name for the tubenose goby in the Caspian Sea basin was Gobius nasalis, described by Filippi (1863) from specimens from the Caspian Sea near Baku, Azerbaijan. The only previously described freshwater species of tubenose goby in the Caspian Sea basin was G. semipellucidus, described by Kessler (1877) from a single specimen from the lower Karasu River at Astrabad (Gorgon) Bay, Iran. Although I do not dispute the resurrection of P. nasalis by Freyhof & Naseka (2007) for truly

30 “marine” specimens of Caspian Sea basin tubenose goby, it is likely that the freshwater specimens included in this study constitute yet another separate taxon. This problem thus merits further sampling and investigation using both morphological and molecular techniques, as well as state-of-the-art evolutionary analyses. I thus tentatively identify the freshwater Caspian Sea basin lineage as P. cf semipellucidus.

31 Table 2.1 Collection location, salinity (ppt), and number of tubenose goby individuals sampled for molecular analyses.

Taxon Body of Water Location Salinity N Latitude Longitude Code Proterorhinus marmoratus Dniester River delta near Bilyayivka, Ukraine 0-2 15 46.468333 30.216667 A Sukhyi Estuary near Illichivs'k, Ukraine 14-17 15 46.326700 30.667550 B Odessa Bay Cape Langeron, Ukraine 14-17 13 46.483333 30.755000 C Odessa Bay Cape Malyi Fontan, Ukraine 14-17 9 46.450000 30.766667 D Tyligul Estuary Chervonoukrayinka, Ukraine 17-21 15 46.690000 31.200000 E Black Sea Sevastopol, Ukraine 17-18 23 44.604040 33.540840 F

P. semilunaris Lk Superior St Louis R. estuary, MN, USA 0 9 46.666667 -92.200000 G Clinton River 1 km upstream of Lk St. Clair, MI, USA 0 10 42.594282 -82.803323 H Danube River Dobra, Iron Gate Gorge, Serbia 0 3 44.638100 21.909400 I Dniester River Mohyliv-Podilskii, Ukraine 0 1 48.447860 27.782797 J Kurchurgan Reservoir Hradenytsi, Ukraine 0 5 46.600000 29.986000 K Dnieper River Kiev, Ukraine 0 3 50.490019 30.517685 L Simferopol Reservoir Simferopol, Ukraine 0 8 44.928732 34.149352 M 3 2 P. cf. semipellucidus Rybinsk Reservoir Rybinsk, Russia 0 6 58.362982 38.425064 N Buzuluk River Alexivska, Russia 0 1 50.273600 42.182188 O Volga-Don Canal Karpovska Reservoir, Iliovka, Russia 0 3 48.643269 43.617069 P Chagraiskoye Reservoir Zunda Tolga, Russia 0 3 45.617691 44.211077 Q Volga River Volgograd Reservoir, Volgograd, Russia 0 1 48.870870 44.660139 R Volga River Preshib, Russia 0 3 47.683923 46.509057 S Volga River delta Damchik, Russia 0-6 3 45.788350 47.886953 T

Proterorhinus. sp. Chernozemel'skii Connector 120 km east of Elista, Russia 0 2 46.272008 45.615373 U Table 2.2 Primers pairs used for PCR amplification (including reaction annealing temperatures TA) and DNA sequencing of the tubenose goby.

Gene Primer name Sequence (5'-3') TA Reference (°C) cyt b AJG15 CAAAAACCATCGTTGTAATTCAACT 52 Akihito et al. (2000) H15343goby GGGTTATTAGATCCTGTTTCGTGTAGG this study L15162goby GCTATGTCCTACCATGGGGGCAAATATC 52 " " H5 GAATTYTRGCTTTGGGAG Akihito et al. (2000) COI L6468 GCTCAGCCATTTTACCTGTG 53 Thacker (2003) H7127 ACYTCTGGGTGACCAAAGAATC " " L7059 CCCTGCMGGTGGAGGAGACCC 53 " " H7696 AGGCCTAGGAAGTGTTGAGGGAAG " " RAG1 RAG1F1 CTGAGCTGCAGTCAGTACCATAAGATGT 50 Lopez et al. (2004) R811goby TCATAGCGCTCTAGGTTCTCC this study F709goby CTTATGTCTGCACGCTCTGC 50 " "

RAG1R1 GTGAGTCCTTGTGAGCTTCCATRAAYTT Lopez et al. (2004)

33 Table 2.3 Mitochondrial cytochrome b haplotype distribution for populations of the tubenose goby Proterorhinus. Marine haplotypes

are italic; Black Sea freshwater haplotypes are indicated in bold; Caspian Sea freshwater haplotypes in plain text; Kumo-Manych

Depression haplotypes in bold italic. Population letter codes follow Table 2.1.

Population P. marmoratus P. semilunaris P. cf semipellucidus Haplo- GenBank type accession A B C D E F G H I J K L M N O P Q R S T U 1 EU444604 1 2 EU444605 3 3 EU444606 1 4 EU444607 9 7 5 EU444608 1 1

3 6 EU444609 4 7 EU444610 5 3 1 8 EU444611 1 1 1 3 9 EU444612 2 10 EU444613 1 11 EU444614 1 12 EU444615 1 13 EU444616 1 14 EU444617 1 1 15 EU444618 1 16 EU444619 1 17 EU444620 1 1 3 4 5 1 5 18 EU444621 2 19 EU444622 1 20 EU444623 1 1 21 EU444624 1 1 22 EU444625 2 2 1 23 EU444626 1 24 EU444627 1 25 EU444628 1 26 EU444629 1 1 27 EU444630 2 28 EU444631 1 29 EU444632 1 30 EU444633 1 31 EU444634 2 32 EU444635 1 33 EU444636 2 34 EU444637 6 35 EU444638 7 36 EU444639 1 37 EU444640 1 38 EU444641 1 3

5 39 EU444642 1 40 EU444643 1 41 EU444644 1 42 EU444645 1 1 43 EU444646 1 44 EU444647 1 45 EU444648 1 46 EU444649 6 47 EU444650 1 48 EU444651 1 49 EU444652 1 50 EU444653 1 51 EU444654 1 52 EU444655 1 53 EU444656 1 54 EU444657 1 55 EU444658 1 56 EU444659 1 57 EU444660 1 58 EU444661 1 59 EU444662 1 60 EU444663 1 61 EU444664 1 62 EU444665 1 63 EU444666 1 3 6 Table 2.4 Cytochrome b sequence divergence among major lineages of Proterorhinus, including Tamura-Nei distances within (on diagonal) and among (below diagonal) lineages, and lineage pairwise FST values (above diagonal). * = significant difference following sequential Bonferroni correction (Rice 1989).

P. cf semipellucidus P. semilunaris P. marmoratus P. cf semipellucidus 0.002 0.873* 0.991* P. semilunaris 0.040 0.006 0.980* P. marmoratus 0.160 0.173 0.001

Table 2.5 Age estimates (million years before present) based on pairwise divergence and penalized likelihood (Sanderson, 2002) for major lineages of Proterorhinus from Figure

2.4.

Penalized likelihood Clade Pairwise age age Proterorhinus marmoratus (Black Sea 0.05 0.09 basin marine) Proterorhinus cf semipellucidus (Caspian 0.07 0.13 Sea basin) Proterorhinus semilunaris (Black Sea 0.17 0.26 basin freshwater) Kumo-Manych Depression freshwater 0.68 0.72 Proterorhinus and P. semilunaris Proterorhinus semilunaris and P. cf 0.94 1.15 semipellucidus Freshwater Proterorhinus and P. marm- 4.22 4.40 oratus Proterorhinus and Mesogobius 4.85 6.18

37 Table 2.6 Nested clade analysis of Proterorhinus cytochrome b haplotypes showing significant nesting clades and subclades, clade

dispersion (DC) and displacement (DN) values, inference chain, and the resulting inferred pattern. For clade dispersion and

displacement values, (S) and (L) indicate a significantly small or large value, respectively.

Nest- ing Significant Clade Subclade DC DN Inference chain Inferred pattern Total 4-1 (S) p<0.0001 (S) p<0.0001 1-2-3-4-9-No Allopatric fragmentation among marine and freshwater habitats in the Black 4-2 (L) p<0.0001 and Caspian Sea basins 4-3 (S) p<0.0001 I-T (L) p<0.0001 (L) p<0.0001 4-3 3-3 (L) p=0.0191 1-19-20-No Inadequate geographic sampling between Crimean Peninsula and Danube/Dni- 3-4 (S) p=0.0210 ester Rivers I-T (S) p=0.0182 3

8 3-4 2-6 (L) p=0.0004 1-2-3-5-6-7-8-No Either isolation by distance or long distance dispersal between Danube and 2-9 (S) p<0.0001 (S) p<0.0001 Dnieper River basins I-T (L) p<0.0001 Inadequate geographic sampling between Dniester River and Crimean Penin- 3-3 2-4 (S) p=0.0064 1-19-20-No sula 2-5 (S) p=0.0064 (L) p=0.0064 3-1 2-1 (S) p<0.0001 (S) p<0.0001 1-2-11-Yes Range expansion within northwestern Black Sea 2-2 (S) p<0.0001 (L) p<0.0001 I-T (S) p<0.0001 Table 2.7 Tajima’s (1989) D test for selective neutrality for major lineages of

Proterorhinus. Significantly negative D values (bold) indicate an excess of recent mutations and suggest recent population size expansion.

Lineage n θπ θS Tajima's D P P. cf semipellucidus 20 1.89 2.54 -0.87 0.202 P. semilunaris 54 7.13 8.78 -0.63 0.280 P. marmoratus 75 1.58 6.14 -2.34 0.001

39 Table 2.8 Morphometrics and meristics of Proterorhinus taxa.

P. marmoratus (N = 30) P. semilunaris (N = 76) P. cf semipellucidus (N = 17) Measurement Mean ± SD Range Mean ± SD Range Mean ± SD Range Standard length 47.0 ± 11.7 24.2-66.4 43.7 ± 10.4 17.5-68.0 38.8 ± 7.8 23.0-55.0 % Standard length Caudal peduncle length 14.6 ± 1.7 12.2-19.5 15.6 ± 1.3 11.6-18.5 15.5 ± 2.0 11.8-20.6 Caudal peduncle depth 11.2 ± 1.0 9.1-13.8 10.7 ± 0.9 8.0-12.8 11.3 ± 1.6 7.3-14.2 Caudal peduncle width 4.2 ± 0.8 2.5-5.9 4.5 ± 0.6 3.1-7.1 4.2 ± 0.7 3.0-5.4 Pelvic disc length 21.3 ± 1.8 17.7-24.8 22.5 ± 1.5 18.4-26.5 22.3 ± 1.8 17.4-25.1 Abdomen length 23.7 ± 2.9 18.5-29.8 23.6 ± 2.0 19.0-31.7 23.4 ± 1.8 18.4-25.8 Pectoral fin length 23.6 ± 2.7 18.4-28.0 25.5 ± 2.3 20.3-32.3 23.5 ± 2.0 19.4-26.5 Body depth 21.3 ± 2.3 17.3-25.5 21.1 ± 2.0 15.5-24.7 18.3 ± 2.4 14.5-22.5 Body width 13.9 ± 1.7 9.9-17.1 14.2 ± 1.8 10.3-19.6 12.9 ± 1.8 9.6-15.8 Head length 26.7 ± 1.9 21.0-28.8 29.2 ± 1.2 27.0-31.9 28.5 ± 2.1 23.7-32.6 % Head length 4

0 Head width 68.5 ± 7.1 58.8-92.1 64.5 ± 6.4 52.1-80.1 58.8 ± 4.6 49.3-68.0 Head depth 72.7 ± 7.0 61.3-97.6 65.1 ± 4.9 54.8-76.3 62.0 ± 5.0 52.0-71.2 Eye diameter 22.2 ± 3.1 17.7-29.2 21.7 ± 2.7 17.6-31.4 23.0 ± 2.5 18.8-26.7 Interorbital distance 10.0 ± 2.7 5.9-14.9 12.6 ± 3.0 5.9-21.1 8.1 ± 1.7 5.3-10.9 Snout length 27.5 ± 3.8 21.7-34.3 31.2 ± 4.1 20.6-38.8 27.0 ± 1.7 24.2-30.4 Preorbital width 16.8 ± 2.0 13.1-19.8 16.6 ± 2.4 9.8-24.6 13.9 ± 2.5 9.3-18.1

Meristic Mode Range Mode Range Mode Range scales in lateral series 44 39-49 42 39-53 43 39-47 first dorsal fin spines 6 5-7 6 5-6 6 5-6 second dorsal fin elements 16 15-18 16 14-17 17 15-18 anal fin elements 15 12-16 13 12-15 15 13-16 pectoral-fin segmented rays 15 15-17 15 14-16 15 14-16 Table 2.9 Summary of principal components analysis (eigenvalues and loadings) calculated from 17 linear measurements from three Proterorhinus lineages. Loadings with absolute magnitude greater than 0.3 shown in bold.

PC1 PC2 PC3 PC4 Eigenvalue 1.206 0.219 0.152 0.145 % Eigenvalue 90.1 3.0 1.4 1.3 Standard length -0.212 -0.184 -0.020 0.084 Caudal peduncle length -0.197 -0.104 0.198 0.097 Caudal peduncle depth -0.240 -0.234 0.309 0.086 Caudal peduncle width -0.190 -0.014 0.690 -0.610 Pelvic disc length -0.202 -0.105 -0.192 -0.110 Abdomen length -0.215 -0.255 -0.199 -0.134 Pectoral fin length -0.239 -0.063 -0.261 -0.069 Body depth -0.275 -0.133 0.031 0.338 Body width -0.271 -0.208 0.138 0.178 Head length -0.212 -0.031 -0.135 -0.039 Head width -0.261 -0.081 0.034 0.188 Head depth -0.254 -0.178 -0.008 0.176 Eye diameter -0.139 -0.062 -0.080 -0.035 Interorbital distance -0.399 0.819 0.141 0.264 Snout length -0.283 0.214 -0.375 -0.495 Preorbital width -0.305 -0.008 -0.196 -0.202

41 Fig 2.1 A) Freshwater tubenose goby (Proterorhinus semilunaris), UMMZ 231002, 32.1 mm standard length; B) enlargement of head showing tubular nostrils. Scale bar = 5 mm.

A

B.

42 Fig. 2.2 Collection locations of Proterorhinus specimens in native and invasive European populations. Shaded area indicates current range of Proterorhinus (adapted from Pinchuk et al., 2003a)

43 Fig. 2.3 Maximum likelihood phylogeny of Proterorhinus and other neogobiin outgroups based on total evidence analysis of cyt b, COI, and RAG1 sequence data. Numbers above branches are bootstrap support values from 1,000 pseudoreplications (bold) and decay indices (italic) for maximum parsimony, below branches are likelihood bootstrap support (bold italic) and posterior probability values (plain) for Bayesian analysis.

44 Fig. 2.4 Maximum likelihood analysis of Proterorhinus cytochrome b haplotypes.

Letters represent population codes from Table 2.1. Numbers around branches indicate maximum likelihood bootstrap support (1,000 pseudoreplications).

45 Fig. 2.5 Statistical parsimony network among Proterorhinus cytochrome b haplotypes showing nested clades, highlighting major divisions between habitat type (marine vs. freshwater; clades 4-1 vs. 4-2/4-3), and among major freshwater basins (Black and

Caspian Sea basins, Kumo-Manych Depression; clades 4-2, 4-3 and 4-4). Significant clades from Table 2.6 are shaded and labeled.

46 Fig. 2.6 Plots of the A) second vs. first; and B) fourth vs. second principal components

(PC) from a principal components analysis of 123 tubenose gobies. ● = Proterorhinus marmoratus; ○ = P. semilunaris, □ = P. cf semipellucidus. 95% confidence ellipses and centroids (+) are drawn for each species. A multivariate analysis of variance (MANOVA) using both body size and shape information (PC 1-4), as well as body shape alone detected highly significant differences among taxa (PC 1-5, Wilks’ λ = 0.416, F8, 234 =

16.128, P << 0.001; PC 2-4, Wilks’ λ = 0.447, F6, 236 = 19.498, P << 0.001).

A.

47 B.

48 Chapter Three

Escape from the Ponto-Caspian: Evolution and biogeography of an endemic goby species flock (Benthophilinae: Gobiidae: Teleostei)†

† This manuscript was originally published as: Neilson ME, Stepien CA (2009) Escape from the Ponto-Caspian: evolution and biogeography of an endemic goby species flock

(Benthophilinae: Gobiidae: Teleostei). Molecular Phylogenetics and Evolution, 52, 84-

102.

Abstract

Endemic Ponto-Caspian gobies include a flock of ~24 “neogobiin” species

(containing the nominal genera and subgenera Apollonia, Babka, Neogobius,

Mesogobius, Ponticola, and Proterorhinus; Teleostei: Gobiidae), of which a large proportion (five species; ~21%) recently escaped to invade other freshwater Eurasian systems and the North American Great Lakes. I provided its first comprehensive phylogenetic and biogeographic analysis based on 4,709 bp sequences from two mitochondrial and two nuclear genes with maximum parsimony, likelihood, and Bayesian approaches. I additionally compared its relationships with the tadpole gobies

(Benthophilus and Caspiosoma), which comprise a related endemic Ponto-Caspian gobiid group; along with a variety of postulated relatives and outgroups. Results of all phylogenetic approaches are highly congruent and provide very strong support for recognizing the subfamily Benthophilinae; which encompasses both the “neogobiins” and

49 the tadpole gobies, and genetically diverges from other Gobiidae subfamilies – including

(non-monophyletic) Gobiinae and Gobinellinae. Benthophilinae contains three tribes:

Neogobiini (Neogobius, which is synonymized here with Apollonia; containing the type species N. fluviatilis, along with N. melanostomus and N. caspius), Ponticolini

(containing the genera Mesogobius, Proterorhinus, Babka, and Ponticola - elevating the latter two from subgenera and removing them from the formerly paraphyletic

Neogobius), and Benthophilini (tadpole gobies). Within Ponticolini, Proterorhinus and

Mesogobius comprise the sister clade of the Ponticola and Babka clade. Further work is needed to clarify the interrelationships of the tadpole gobies. Invasiveness is widespread in freshwater and euryhaline taxa of Neogobius, Proterorhinus, Babka, and Ponticola; but not in marine species, Mesogobius, or tadpole gobies.

Introduction

Exotic species pose one of the most serious threats to native ecosystems worldwide (Simberloff & Von Holle 1999; Sax & Gaines 2008) and often present analytical and conceptual challenges – including resolving their taxonomic identity and systematic relationships. As species introductions increase (Cohen & Carlton 1998;

Lockwood et al. 2006), more nonindigenous taxa will originate from poorly known groups lacking identification keys and analysis with modern phylogenetic methodology.

These problems preclude our understanding of fundamental ecological requirements of introduced taxa, including how they adapt to novel habitats and alter the evolutionary trajectory of native ecosystems (Mooney & Cleland 2001), thereby impeding effective management or control. Phylogenetic and biogeographic analyses of DNA sequence data, as accomplished here, thus provide us with the means to identify invasive taxa,

50 elucidate cryptic species, analyze whether congeners and relatives invade in concert, and predict potential new invaders.

For example, the ecology of the North American Great Lakes recently has been restructured by waves of invaders accidentally introduced from ships’ ballast water, primarily from the Eurasian Ponto-Caspian region (including the Aral, Azov, Black, and

Caspian Seas and associated drainages; (Mills et al. 1993; Ricciardi & MacIsaac 2000).

Notable for their ecological effects are the dreissenid zebra and quagga mussels,

Dreissena polymorpha and D. bugensis, which first appeared in the Great Lakes in the mid-1980s via ballast water introduction. Two Ponto-Caspian gobies then entered the

Great Lakes in 1990 (Jude et al. 1992) – the round goby Neogobius melanostomus

(Pallas, 1814) (=Apollonia melanostoma per Stepien and Tumeo, 2006) and the freshwater tubenose goby Proterorhinus semilunaris (Heckel, 1837) (formerly grouped as a single species with the marine P. marmoratus [Pallas, 1814]: Stepien & Tumeo 2006;

Chapter 2). Like the zebra mussel, the round goby spread rapidly throughout all five

Great Lakes (USGS 2003) and is now one of the most abundant benthic fish species

(Jude & DeBoe 1996; Johnson et al. 2005). Its invasion success likely was aided by the prevalence of its native dreissenid mussel prey (Ray and Corkum 1997). Such facilitative interactions among co-evolved invaders may significantly augment the success of invasive communities (Simberloff & Von Holle 1999; Ricciardi & MacIsaac 2000), with widespread ecological consequences – as has occurred with the growing dominance of the dreissenid mussel/round goby benthic community (Vanderploeg et al. 2002).

The round and tubenose gobies are members of an enigmatic group native to the

Ponto-Caspian region containing ~24 species arranged (prior to the present study) in four

51 genera (Apollonia, Mesogobius, Neogobius, and Proterorhinus; Miller 2003a; Stepien &

Tumeo 2006; see Table 3.1), which have been variously termed “neogobiins”. Several taxa also contain putative subspecies divided between the Black and Caspian Sea basins.

This group meets the definition of a species flock sensu Greenwood (1984) – a geographically circumscribed, monophyletic taxon characterized by marked radiation.

The historic endemism and taxonomic diversity of the Ponto-Caspian “neogobiins” are remarkable, and knowledge of their evolutionary history may yield insight on the evolution of species flocks (Johns & Avise 1998a), factors leading to their rapid evolutionary diversification, as well as invasive success in new habitats.

Despite their remarkable radiation, the systematic relationships and placement of

Ponto-Caspian “neogobiin” gobies have been disputed and unclear. This phylogenetic confusion is highlighted by the fact that a large number of the group are invasive (five species; ~ 21%; including the round goby N. melanostomus, monkey goby N. fluviatilis

(= A. fluviatilis) [Pallas, 1814], racer goby N. gymnotrachelus [Kessler, 1857], bighead goby N. kessleri [Gunther, 1861], and freshwater tubenose goby P. semilunaris) in freshwater systems of Eastern/Central Europe and/or North America. For example, N. melanostomus is invasive in both Europe and North America, and in the latter has undergone rapid range expansion since its introduction in 1990 and is implicated in the decline of native Great Lakes fishes (Jude et al. 1995; Corkum et al. 2004).

Detailed investigations of morphology, osteology, and systematics of the Ponto

Caspian “neogobiin” gobies have only recently begun; and relationships of the genus

Neogobius sensu lato with other taxa have been disputed. For example, members of the genera Neogobius sensu lato and Proterorhinus were regarded as subgenera of Gobius by

52 Vasil'eva (1989, 1991, 1999) based on cranial osteology. Birdsong et al. (1988), in a study of vertebral column and median fin osteology, failed to place Neogobius (the sole

Ponto-Caspian representative in their study) in any of their hypothesized genus-groups, whereas Pezold (1993) proposed that this genus may belong to the subfamily

Gobionellinae based on its patterning of infraorbital pores in the cephalic lateral line system (although he did not directly examine any “neogobiin” material). Simonovic

(1999) suggested a close relationship to the subfamily Gobiinae for Neogobius sensu lato and Proterorhinus based on external morphometrics, osteology, and karyology, but did not examine Mesogobius. Ahnelt & Duchkowitsch's (2004) study of postcranial osteology of Proterorhinus placed it along with Neogobius sensu lato in the Gobiinae.

Composition of the genus Neogobius sensu lato and its interrelationships also have been controversial. Miller & Vasil'eva (2003) summarized information, listing the genus as comprising 14 species separated into five subgenera – N. Apollonia (containing only N. melanostomus), N. Neogobius (restricted to N. fluviatilis), N. Eichwaldiella

(containing only N. caspius), N. Babka (= N. gymnotrachelus), and N. Ponticola

(including N. cephalargoides, kessleri, ratan, syrman, etc.). Miller (2003b) elevated the subgenus N. Chasar to generic status (containing a single taxon – C. bathybius) based on increased modal number of dorsal fin rays and differences in the pattern of cephalic sensory papillae; however, this distinction is questionable and the taxon is regarded as belonging in N. Ponticola (E. Vasil’eva, personal communication). Stepien & Tumeo

(2006) elevated Apollonia (including N. melanostomus and N. fluviatilis) to generic status due to its paraphyletic position relative to the other subgenera of Neogobius sensu lato

(i.e., N. Babka and N. Ponticola), based on mitochondrial DNA cytochrome b gene

53 sequences. Moreover, the clade containing Proterorhinus, Mesogobius, Apollonia sensu

Stepien & Tumeo (2006), and Neogobius sensu lato appeared separated from the

Gobiinae tested (Gobius and Zosterissessor; Stepien & Tumeo 2006).

Miller & Vasil'eva (2003) noted that the systematic relationships of the Ponto-

Caspian gobies are poorly understood, provided no hypothesis for their relationships, and expressed the need for a detailed cladistic revision. Although some prior studies examined selected morphological aspects of their relationships and systematics, none investigated relationships of the native Ponto-Caspian gobies as a whole and only those by the Stepien laboratory and one other used a molecular approach (i.e., partial group analyses by (Dougherty et al. 1996; Dillon & Stepien 2001; Stepien & Tumeo 2006);

Chapter 2).

The central goal is to analyze the systematic relationships among Ponto-Caspian

“neogobiin” gobies, and to illuminate some of the factors (biogeographic, evolutionary, or phylogenetic) leading to their diversification that also may augment their success as invasive species. In particular, I investigated the following questions: (1) are the currently recognized species of “neogobiin” gobies valid (i.e., reciprocally monophyletic) taxa? (2) are the current genera valid?, and (3) how was their speciation and diversification shaped by the geologic history of the Ponto-Caspian region? I analyzed the phylogenetic relationships among “neogobiin” gobies, compared with gobiin relatives and outgroup taxa, using DNA sequence data from four gene regions: the mitochondrial

(mt) cytochrome (cyt) b and cytochrome oxidase c subunit I (COI) genes, and the nuclear recombination activating gene 1 (RAG1) and S7 ribosomal protein intron 1 (S7). I included the 19 most prevalent members of ~24 nominal species (Miller & Vasil'eva

54 2003; Freyhof & Naseka 2007; Kovačić & Engin 2008; Chapter 2; see Table 3.1) in the most complete phylogenetic study of the group.

Methods

Taxon sampling

Taxa analyzed in this study, collection locations, and corresponding GenBank accession numbers (http://www.ncbi.nlm.nih.gov) are listed in Table 3.1. Specimens were collected throughout the range of the “neogobiins” within the Ponto-Caspian region

(Fig. 3.1) via small seines, beam/otter trawls, or by hook and line, and include all widely distributed and common taxa. I analyzed all proposed subgenera of Neogobius sensu lato

(= N. Babka, N. Eichwaldiella, and N. Ponticola; Miller & Vasil'eva 2003) and the genera

Apollonia sensu Stepien & Tumeo (2006), Proterorhinus, and Mesogobius; absent taxa either are very rare (e.g., Mesogobius nonultimus), confined to deeper water (N. bathybius), were only recently described (N. rizensis and N. turani; Kovačić & Engin

2008), or have extremely limited distributions (Proterorhinus tataricus; Freyhof &

Naseka 2007). I included seven Gobiinae outgroups (Chromogobius zebratus, Gobius auratus, G. bucchichi, G. fallax, G. niger, Pomatoschistus minutus and Zosterisessor ophiocephalus) that range throughout the Black and Mediterranean Seas and are members of the hypothesized sister lineage of the Ponto-Caspian gobiids (Miller 1990). I also utilized five species of tadpole gobies, including Benthophilus (B. abdurahmanovi,

B. granulosus, B. mahmudbejovi, and B. stellatus) and Caspiosoma caspium. The tadpole gobies constitute a second Ponto-Caspian endemic goby group that is hypothesized to be closely related to “neogobiins”and their relatives (Ahnelt 2003). Specimens were

55 preserved immediately following capture either in 95% ethanol for molecular analyses or in 10% formalin (with removal of right pectoral fin for preservation in 95% ethanol and molecular analyses) for future morphological analyses.

DNA analysis

Genomic DNA was isolated from fin clips or caudal muscle tissue using a Qiagen

DNEasy tissue kit (Valencia, CA) following manufacturer’s protocols. Two mitochondrial genes (cyt b and COI) and 2 nuclear genes (RAG1 and S7) were amplified via the polymerase chain reaction (PCR) using the following primers: cyt b – AJG15

(Akihito et al. 2000), H15343goby (Chapter 2), L15162goby (Chapter 2), and H5

(Akihito et al. 2000); COI – L6486, H7127, L7057, and H7696 (Thacker 2003); RAG1 –

RAG1F1 (Lopez et al. 2004), RAG1-R811goby (Chapter 2), RAG1-F709 (5’-

CTTATGTCTGCACGCTCTGC-3’, this study), and RAG1R1 (Lopez et al. 2004); and

S7 – S7RPEX1F and S7RPEX2R (Chow & Hazama 1998). PCR amplifications were performed in 25 μL volumes containing 10mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM

MgCl2 (2.5 mM for COI), 0.001% (w/v) gelatin, 200 μM each dNTP, 0.5 μM each primer, 1.5 units of Taq polymerase, and ~100 ng (1-3 μL) of template DNA. The PCR profile for cyt b and RAG1 included an initial denaturation of 94°C for 2 min, 40 cycles of 94°C for 45 s, gene specific annealing temperature (cyt b – 52°; RAG1 – 50°; S7 –

60°) for 30 s, and 72°C for 60 s, with a final extension of 72°C for 3 min. The cycling profile for COI included an initial denaturation of 94°C for 3 min, 35 cycles of 94°C for

30 s, 53°C for 30 s, and 72°C for 60 s, with a final extension of 72°C for 2 min. The profile for S7 included an initial denaturation of 94°C for 3 min, 40 cycles of 94°C for 60 s, 60°C for 45 s, and 72°C for 120 s, with a final extension of 72°C for 5 min. PCR

56 reactions were checked on 1% agarose gels stained with ethidium bromide, and excess primers and unincorporated nucleotides were removed from successful reactions with spin column purification kits (QIAquick PCR Purification Kit, Qiagen; or QuickStep 2

PCR Purification Kit, Edge Biosystems, Gaithersburg, MD).

For all genes, amplicons were sequenced in both directions using dye-labeled terminators and PCR primers, and resolved on an ABI 3730 (Applied Biosystems, Foster

City, CA) genetic analyzer at the Cornell University Life Sciences Core Laboratories

Center. Forward and reverse sequences for each gene per individual were aligned by me, a contiguous sequence was created with BIOEDIT (Hall 1999), and sequences for each gene were globally aligned using CLLUSTAL X (cyt b, COI, RAG1; Larkin et al. 2007) or T-COFFEE (S7; Notredame et al. 2000).

Phylogenetic analyses

I used parsimony, likelihood, and Bayesian approaches to reconstruct phylogenies, with PAUP* v4.0b10 (Swofford 2003), PHYML v2.4.4 (Guindon &

Gascuel 2003) and MRBAYES v3.1 (Ronquist & Huelsenbeck 2003), respectively.

Parsimony analyses were performed using unweighted heuristic searches, with starting trees obtained by random addition (100 replications) holding 10 trees per replicate, and tree-bisection-reconnection branch swapping. Branch support was calculated for the inferred branches via non-parametric bootstrapping (2,000 replications).

For likelihood and Bayesian analyses, MODELTEST v3.7 (Posada & Crandall

1998) was employed to determine the simplest best-fit model of evolution for each gene under the Akaike information criterion. For the cyt b data, the best-fit model was

GTR+I+G with a shape parameter (α) = 0.9552 and proportion of invariant sites (i) =

57 0.4627. For the COI gene, the GTR+I+G also was selected (α = 1.1661; i = 0.5770). For the nuclear genes, the best-fit model was TrN+G for RAG1 (α = 0.6191) and TVM+G (α

= 1.1150) for S7.

Bayesian analyses using a Metropolis coupled Markov chain Monte Carlo

(MCMCMC) approach were run for 5 million generations, with sampling every 100 generations, to ensure convergence of likelihood values. Four separate chains were run simultaneously for each analysis, and two analyses were run simultaneously. The burn-in period for the MCMCMC analysis was determined by plotting log likelihood values at each generation to identify the point at which they reached stationarity. In all analyses, stationarity was reached by 50,000 generations; thus, a conservative burn-in period of

500,000 generations was used, and trees and parameter values sampled prior to the burn- in were discarded. Branch support for likelihood analyses was calculated using non- parametric bootstrapping (2,000 replications) and via the posterior probability distribution of clades for Bayesian analyses.

In addition to the separate analyses, I explored the combinability of the four gene regions into a single dataset using several methods. An incongruence length difference

(ILD) test (Farris et al. 1995) in PAUP* (1,000 replications) was employed to determine the congruence of topologies among datasets within a parsimony framework, using heuristic searches with 50 random addition sequences per replicate. As the ILD test is known to be susceptible to noise within datasets, all uninformative characters were removed. Significant incongruence was found among all four genes, as well as within and between mitochondrial and nuclear genomes (P = 0.001 in all tests). To further explore the extent and location of congruence, I calculated partitioned branch support

58 (a.k.a. partitioned Bremer support [PBS]; Baker & DeSalle 1997) for each gene region, and the partition congruence index (Brower 2006) for the four genes combined. Briefly, partitioned branch support determines the contribution of each partition to the total branch support for each branch on a phylogeny, with the sum of PBS for all partitions equaling the total branch support (BS) for each individual branch. The partition congruence index (PCI) incorporates the magnitude of difference between PBS for each partition and the total BS of all partitions combined, and thus summarizes the amount of incongruence among the partitions (Brower 2006). When all partitions are congruent,

PCI is equal to the BS of an individual branch; as the amount of incongruence among partitions increases, PCI decreases linearly and eventually becomes negative at high levels of incongruence. As conflict among the four analyzed genes (i.e., negative values of PBS for one or more genes; PCI values < 0 or substantially less than BS) was limited primarily to intraspecific branches, and deeper interspecific and intergeneric branches were strongly supported, I conducted a second series of phylogenetic analyses using the combined four gene dataset.

Search strategies for the concatenated sequences were identical to those in the separate analyses. MODELTEST selected the TVM+I+G model (α = 0.4680; i = 0.3331) as the best fit model for the combined sequence data. A partitioned mixed-model approach was used for Bayesian analysis of the combined gene regions. Models of sequence evolution identified for each individual gene region were assigned using the

APPLYTO command, and the appropriate model parameters were estimated for each gene using the UNLINK command. Topological differences among the three different

59 analysis methods for the concatenated sequences were tested with a Shimodaira &

Hasegawa (1999) test (10,000 RELL bootstrap replicates) implemented in PAUP*.

To examine the placement of the Ponto-Caspian “neogobiin” gobies among other gobioid fishes, I performed additional phylogenetic analyses. The first was performed using cyt b sequences from the present study as well as additional taxa collected during this study and sequences from GenBank, including: Gobiidae: Amblyopinae – Taenioides limicola Smith, 1964 (AB021253); Gobiinae – Elacatinus macrodon (Beebe & Tee-Van,

1928) (AY846447), Gobiosoma bosc (Lacepède, 1800) (AY848456), Knipowitschia caucasica (Berg, 1916) (FJ526796); Gobionellinae – Acanthogobius flavimanus

(Temminck & Schlegel, 1845) (AB021249), Gymnogobius petschiliensis (Rendahl, 1924)

(AY525784), Rhinogobius giurinus (Rutter, 1897) (AB018997), Tridentiger bifasciatus

Steindachner, 1881 (AB021254); Oxudercinae – Periophthalmus argentilineatus

Valenciennes, 1837 (AB021251); Eleotridae: Butinae – Butis amboinensis (Bleeker,

1853) (AB021232), Ophiocara porocephala (Valenciennes, 1837) (AB021245);

Eleotrinae – Dormitator maculatus (Bloch, 1792) (AB021234), Eleotris fusca (Forster,

1801) (AB021236); Kraemeriidae – Kraemeria cunicularia Rofen, 1958 (AB021250);

Microdesmidae – Gunnellichthys monostigma Smith, 1958 (AB021256); Odontobutidae

– Odontobutis obscura (Temminck & Schlegel, 1845) (AB021243), Odontobutis platycephala Iwata & Jeon, 1985 (DQ010651); Ptereleotridae – Ptereleotris heteroptera

(Bleeker, 1855) (AB021252); Rhyacichthyidae – Rhyacichthys aspro (Valenciennes,

1837) (AP004454). The second analysis was run using COI sequence data from the present study combined with COI sequences obtained by (Thacker 2003) in an analysis of the molecular systematics of gobioid fishes. Phylogenetic analyses for the expanded cyt

60 b and COI datasets were conducted as described above, with the GTR+I+G model (α =

0.600; i = 0.423) chosen as the best-fit model for cyt b, and also as the best-fit model for

COI (α = 0.589; i = 0.547).

Divergence time estimation

To estimate divergence times among major lineages, I used a penalized likelihood approach (Sanderson 2002) implemented in the program R8S 1.71 (Sanderson 2003). An initial age estimate was generated for the extended cyt b ML tree under a molecular clock assumption, from which the sequences significantly departed, and a second analysis was conducted using penalized likelihood with the optimal smoothing parameter (= 3.2) determined by cross-validation in R8S. Divergence time estimates under penalized likelihood require a fixed age for at least one node within the phylogeny. Rückert-

Ülkümen (2006) described fossil otoliths of Neogobius as dating to the late Miocene- early Pliocene (~10 Mya), and Bajpai & Kapur (2004) describe the earliest fossils of

Gobiidae from the early Eocene (51-56 Mya). Thus, I followed Chapter 2 and set the age of the node for Gobiidae to 53 My and the most recent common “neogobiin” ancestor to

10 My, as their otoliths are morphologically similar.

Results

The dataset for the combined four gene regions across 19 “neogobiin” taxa and 12 outgroups comprises 4,709 aligned bp (cyt b – 1,142 bp; COI – 1,271 bp; RAG1 – 1,556 bp; S7 – 740 bp including indels). GenBank accession numbers are EU331173,

EU331175, EU331186, EU331208, EU331225, EU444604, EU444607, EU444610-612,

EU444621, EU444624, EU444626, EU444630, EU444632, EU444636, EU444650,

61 EU444667, EU444668, EU444670, and FJ526747-FJ526795 for cyt b; EU444673-677,

EU444679-680, EU444682-691, EU444694, EU444697-698, and FJ526797-FJ526850 for COI; EU444699-700, EU444702-703, EU444705-706, EU444708-717, EU444720,

EU444723-724, and FJ526851- FJ526904 for RAG1; and FJ526905-FJ526978 for S7

(Table 3.1, 61 individuals). Base composition is stationary across taxa for each gene (χ2 >

31.6; df = 180; P > 0.29).

Phylogenies inferred from the parsimony (MP), likelihood (ML), and Bayesian

(BI) analyses of the combined four gene dataset generally are highly congruent: no statistical significant differences are found among the three topologies (S-H test, P =

0.21), and the ML tree is presented in Fig. 3.2 for clarity. These results reveal a monophyletic clade containing all Ponto-Caspian endemic gobiid taxa – redefined and redescribed in the present study as Benthophilinae Iljin, 1930 – which is the oldest historic subfamilial name. The subfamily Benthophilinae is clearly divergent and distinct from other gobiid taxa, and contains both the former “neogobiins” as well as the tadpole gobies. All genera and all individual nominal species are highly supported (> 95% bootstrap support for MP and ML analyses, > 4 branch support for MP, and > 0.95 posterior probability for BI). Three clades are highly resolved in Benthophilinae, showing that the genus Neogobius sensu lato (containing the subgenera/genera Apollonia sensu Stepien & Tumeo (2006), Neogobius, Babka, and Ponticola; see Table 3.1) is paraphyletic and invalid. Notably, Apollonia sensu Stepien & Tumeo (2006) and the former Neogobius sensu lato are each separated by the genera Proterorhinus and

Mesogobius (Fig. 3.2). The three primary clades of Benthophilinae (corresponding to tribes within the subfamily) are 1) a now-restricted Neogobius (=Apollonia, containing

62 the type species N. fluviatilis) that is monotypic in the tribe Neogobiini, 2) the tadpole gobies, including the genera Benthophilus and Caspiosoma, comprising the tribe

Benthophilini, and 3) a larger clade termed the Ponticolini, which contains Proterorhinus,

Mesogobius, and Iljin's (1927) former subgenera Babka and Ponticola that I here elevate to the level of genera (removing them from Neogobius sensu lato).

The results demonstrated high support for a now-restricted genus Neogobius

(=Apollonia per Stepien & Tumeo 2006), which contains N. fluviatilis (= A. fluviatilis) +

N. melanostomus (= A. melanostoma) + N. caspius (> 83% bootstrap support, 12 branch support, 1.00 posterior probability). The former subgenera Babka (containing B. gymnotrachelus) and Ponticola (comprising Po. cephalargoides, Po. constructor, Po. cyrius, Po. eurycephalus, Po. gorlap, Po. kessleri, Po. platyrostris, Po. ratan, Po. rhodioni, and Po. syrman) are each strongly supported as separate clades (> 97% bootstrap support, > 14 branch support, 1.00 posterior probability), clearly diverge from the other genera, and warrant elevation to generic status. The genera Mesogobius and

Proterorhinus are strongly supported as sister groups (>87% bootstrap, 9 branch support,

1.00 posterior probability), and a Mesogobius + Proterorhinus clade is then the sister clade to Babka and Ponticola (100% bootstrap support, 63 branch support, 1.00 posterior probability).

The two tadpole goby genera Benthophilus and Caspiosoma comprise the tribe

Benthophilini, which constitutes the sister clade to the new tribe Ponticolini (Babka,

Ponticola, Mesogobius, and Proterorhinus). Level of support for this sister relationship varies among analysis methods (high parsimony branch support and posterior probability, lower parsimony bootstrap support, very low likelihood bootstrap support) and represents

63 a short internal branch on the phylogeny. The now-restricted Neogobius (=Apollonia) is strongly supported (> 95% bootstrap, 37 branch support, 1.00 posterior probability) as the sister clade of the Ponticolini + Benthophilini. This result confirms the relationship described by Stepien & Tumeo (2006) and Chapter 2, of a restricted Neogobius

(=Apollonia) as a separate genus from Ponticola and Babka.

Results of expanded cyt b analyses are very similar to the combined four gene analyses (ML topology and support values in Fig. 3.3), with slight placement differences for Ponticola ratan and Po. syrman, and in grouping the Benthophilini (Benthophilus +

Caspiosoma) with the Neogobiini rather than with the Ponticolini. There is strong support for the subfamily Benthophilinae as a clade distinct from the remainder of the

Gobiidae, as well as for generic separation of Neogobius (=Apollonia), Babka, and

Ponticola. Divergence time estimates from the extended cyt b ML tree are reported in

Table 3.3. Separation of the subfamily Benthophilinae from other Gobiinae taxa occurred

~39 million years ago (Mya).

The genera Neogobius, Mesogobius, Proterorhinus, Babka, Ponticola,

Benthophilus, and Caspiosoma) have similar dates of origin, ranging from 4.29-6.25

Mya. Proterorhinus diverged from Mesogobius ~6.18 Mya. Among the remaining genera, the three Neogobius (=Apollonia) species are separated by an estimated 5.47 My,

Proterorhinus species by ~4.29 My, and Babka and Ponticola by ~5.08 My. Within

Ponticola, a split (~4.07 My) into two primary clades occurred: a “kessleri” group containing Po. eurycephalus, Po. gorlap, and Po. kessleri, versus a clade containing the remaining Ponticola species. The “kessleri” group began diverging from one another

~1.37 Mya. Within the second Ponticola clade, Po. ratan and Po. syrman branched off

64 soon after separating from the kessleri group; and the remaining species (Po. cephalargoides, Po. constructor, Po. cyrius, Po. platyrostris, and Po. rhodioni) comprise a second “platyrostris” group that radiated from one another ~1.82 Mya.

Expanded COI analysis trees (Fig. 3.4) are very similar to the combined four gene analyses, and also are generally congruent with the expanded cyt b trees; with all identifying Benthophilinae as a distinct subfamily from other gobiin taxa. The expanded

COI and extended cyt b analyses vary in degree of separation between the Benthophilinae and members of the putative subfamily Gobiinae, as well as in designating its sister taxa.

The expanded COI analysis depicted a clade containing Gobius + Zosterisessor as the sister clade of the Benthophilinae, which are more distantly related in the expanded cyt b analysis. Both expanded analyses have similar support values; including high support for the subfamily Benthophilinae and its component taxa, and low support for most deeper gobiid branches outside of the Benthophilinae (Figs. 3.3-3.4)

Phylogenies inferred from the parsimony (MP), likelihood (ML), and Bayesian

(BI) analyses primarily differed only in the branching order of individual specimens within species. A single exception for high congruent support in the trees occurs in the freshwater tubenose goby Proterorhinus semilunaris. The evolutionary and phylogeographic history of Proterorhinus evaluated in Chapter 2 is very similar to the relationships seen in the present study except for placement of a single individual,

Proterorhinus sp. AMN1. In Chapter 2 and the ML and BI analyses here, AMN1

(collected in the Kumo-Manych Depression – a lowland between the Russian Plain and the northern foothills of the Caucasus Mountains; see Fig. 3.1) groups closely with

Proterorhinus from the Caspian Sea/Volga River clade (Pr. cf semipellucidus); whereas

65 in our current MP analysis, it clusters with the Black Sea freshwater species (Pr. semilunaris). In addition, Pr. semilunaris and Pr. cf semipellucidus are not distinguished as clades in our present MP analysis (Fig. 3.2); and instead form a single large clade with

Pr. semilunaris located basally. This difference likely results from the addition of nuclear

S7 intron data, which were not previously sequenced.

Discussion

Phylogenetic analysis of molecular data for 19 Ponto-Caspian “neogobiin” species yielded a robust phylogeny that generally agrees with prior molecular and morphological data, yet deviates from some earlier morphological hypotheses. The

“neogobiin” and tadpole “benthophilin” gobies together comprise a clade that markedly diverges from other gobiid taxa. I thus resurrected and redescribed the subfamily

Benthophilinae Beling and Iljin 1927 to encompass three tribes; Neogobiini

(Neogobius=Apollonia), Ponticolini (containing Babka, Mesogobius, Ponticola, and

Proterorhinus), and Benthophilini (the tadpole gobies Benthophilus, Caspisoma, etc.)

These results support the primary findings of Stepien & Tumeo (2006; findings #1-3) and

Chapter 2 (#2-3), which (1) distinguish a restricted genus Neogobius (=Apollonia) comprising the monotypic tribe Neogobiini that is differentiated from the remainder of the Benthophilinae (justifying its elevation from subgenus to generic status and recognizing it as a divergent tribe), (2) recognize separate marine and freshwater

Proterorhinus species in the Black and Caspian Sea basins, and (3) resolve a sister relationship between the genera Proterorhinus and Mesogobius.

66 Taxonomic congruency, departures, and nomenclatural changes

Our phylogeny differs from some of the prior morphological hypotheses proposed for Benthophilinae relationships. Firstly, I found that Berg’s (1949) grouping of Iljin’s

(1927) subgenera of Neogobius (Neogobius, Apollonia, Babka, and Ponticola) into a single genus is paraphyletic and invalid. The trees demonstrate clear phylogenetic separation of a restricted genus Neogobius (=Apollonia) in the tribe Neogobiini from the tribe Ponticolini; which includes the newly elevated genera Babka and Ponticola, along with the genera Proterorhinus and Mesogobius. Berg (1949), as first reviser, selected

Neogobius sensu lato as the generic name, and thereby N. fluviatilis became the type species for the genus Neogobius sensu lato. Since the genus name Neogobius must remain with the clade containing N. fluviatilis (W. Eschmeyer, personal communication), the generic name Apollonia is thus synonymized with Neogobius.

In addition, I resolved Neogobius caspius (Eichwald, 1831) as belonging to the now-restricted genus Neogobius (N. fluviatilis and N. melanostomus), which comprises a strongly-supported clade (Figs. 3.2-3.4). Neogobius caspius once was placed in a separate (monotypic) subgenus Eichwaldiella (Whitley, 1930), and later incorrectly was moved (W. Eschmeyer, personal communication) without justification to a monotypic subgenus Neogobius by Miller & Vasil'eva (2003). Its position relative to other

Neogobius/Apollonia species thus was in question prior to this study. Pinchuk (1991) suggested that N. caspius grouped together with N. fluviatilis + N. melanostomus on the basis of mouth size (small in these three species vs. large for taxa now contained in

Ponticola and Babka) as well as tooth size distribution on the dentary, but regarded N. caspius as distinct in the forward position of its anterior and posterior nostrils. Miller &

67 Vasil’eva (2003), in describing Iljin’s (1927) subgenera, presented the diagnostic character of an absent metapterygoid/quadrate bridge as uniting N. fluviatilis and N. melanostomus. The metapterygoid/quadrate bridge likewise is absent in N. caspius

(personal observation), and thus is synapomorphic for a restricted Neogobius clade.

Strong support of the molecular data for this restricted Neogobius (=Apollonia) clade (N. caspius + N. fluviatilis + N. melanostomus), combined with several morphological similarities and the nomenclatural changes described above, leads to the redefinition of a restricted Neogobius (in synonymy with Apollonia) in the tribe Neogobiini.

The molecular phylogenies presented here, as well as in Stepien & Tumeo (2006) and Chapter 2, are congruent in identifying large separation between Ponticola/Babka and Neogobius sensu stricto/Apollonia. In addition, pronounced genetic divergence between Babka and Ponticola (subgenera delineated by Iljin [1927]), along with their morphological separation and autapomorphies (Miller & Vasil'eva 2003), supports their elevation to generic level. The phylogenetic trees presented in this chapter reflect this new nomenclature. The monotypic Babka contains the racer goby B. gymnotrachelus and is the sister species to a strongly-supported monophyletic Ponticola clade, which diverged ~4.51-4.86 Mya (Table 3.4). The Ponticola + Babka clade is the sister group of

Mesogobius + Proterorhinus, with high support; which together form the tribe

Ponticolini. Historically, Babka once was hypothesized to be closely related to the knout goby Mesogobius batrachocephalus (Pallas, 1814) based on early studies of morphology

(Berg 1949) and protein electrophoresis (Dobrovolov et al. 1995), although Vasil'ev &

Grigoryan (1992) concluded that the two were not congeners based on chromosomal morphology; which is further confirmed by their genetic separation in this study.

68 Departures of the phylogeny presented in this Chapter from former systematic hypotheses occur for the clade Ponticola. Notably, Vasil'eva et al. (1993) suggested two distinct groups within Ponticola based on cranial morphometry: one containing Po. gorlap Iljin (1949), Po. kessleri, Po. ratan, and Po. syrman (Nordmann, 1840), and the other encompassing Po. cephalargoides, Po. eurycephalus (Kessler, 1874), Po. platyrostris, and the Caucasian freshwater gobies (Po. constructor [Nordmann, 1840],

Po. cyrius [Kessler, 1874], and Po. rhodioni Vasil'eva and Vasil'ev [1994]). Based on molecular data, Po. ratan is basal to all other Ponticola species; whose branching order differs slightly from that suggested by Vasil'eva et al. (1993). In addition, I resolve two species groups different than those proposed by Vasil'eva et al. (1993): the first group contains Po. eurycephalus, Po. gorlap, and Po. kessleri (designated as the “kessleri” group in Table 3.4); the second comprises Po. cephalargoides, Po. constructor, Po. cyrius, Po. platyrostris, and Po. rhodioni (the “platyrostris” group in Table 3.4).

Relationships among Ponto-Caspian endemic gobiid groups

Inclusion of the tadpole gobies Benthophilus and Caspiosoma is a novel feature of the molecular phylogeny. Although some recent studies have considered their osteology and taxonomy (e.g., Ahnelt et al. 2000; Ahnelt 2003; Boldyrev & Bogutskaya 2007) none of the recent larger-scale studies of goby morphological (Birdsong et al. 1988; Pezold

1993) or molecular (Akihito et al. 2000; Thacker 2003) systematics included any tadpole gobies (e.g., Anatirostrum, Benthophiloides, Benthophilus, and Caspiosoma) or

“neogobiin” taxa (Babka, Mesogobius, Neogobius, Ponticola, and Proterorhinus).

Although the “neogobiins” and tadpole gobies were posited to be sister groups based on

69 shared geography and similar postcranial osteology (Ahnelt 2003), this is the first study to incorporate both in a comprehensive phylogenetic analysis.

All of the analyses strongly support a monophyletic clade comprising the

“neogobiins” and tadpole gobies (Fig 3.2-3.4), for which I resurrected the historic name

Benthophilinae Beling and Iljin 1927, as a subfamily of Gobiidae. Subfamily

Benthophilinae contains three distinctive and divergent clades – designated here as the tribes Benthophilini (the tadpole gobies), Neogobiini (monotypic for the genus

Neogobius), and Ponticolini (Babka, Mesogobius, Ponticola, and Proterorhinus).

Placement of the tribe Benthophilini is inconsistent among some of the trees. All combined sequence data analyses resolve Benthophilini as the sister clade to Ponticolini, however this relationship has mixed support (1.00 posterior probability, < 84% bootstrap support, 13 branch support) and relatively short branch lengths. In the extended cyt b analysis Benthophilini is found as the sister clade to Neogobiini (but with short branch length and no support), whereas in the extended COI analysis Benthophilini again groups with Ponticolini. Additional genetic, morphological, and taxonomic sampling of

Benthophilini is recommended to further resolve its relationships. One potential indication of a close relationship between Neogobiini and Benthophilini is their shared loss of the metapterygoid bridge, in contrast to its presence in the Ponticolini. However, presence of the metapterygoid bridge is widely considered a pleisiomorphic trait within gobiids (Miller 1973).

Higher taxonomic placement of the subfamily Benthophilinae

Relationship of the newly redefined Benthophilinae to other gobiids was contentious prior to this study. Morphological studies placed members of the group

70 either in the Gobiinae (Simonovic 1999; Ahnelt 2003; Ahnelt & Duchkowitsch 2004) or the Gobionellinae (Pezold 1993). The extended cyt b and COI analyses further resolve this question. In both analyses, the Benthophilinae comprises a true taxon, removed from all other gobiin taxa. In addition, both the cyt b and COI analyses suggest a non- monophyletic Gobiinae similar to that found by Thacker (2003 – who did not examine any “neogobiins”) using the mt COI, ND1, and ND2 genes. Analysis of the COI data from the present study combined with Thacker’s (2003) COI data (Fig. 3.4) yields a similar result to the extended cyt b dataset, depicting a clade comprising the

Benthophilinae plus Gobius + Zosterisessor (Gobiinae) nesting as the sister group of one of Thacker’s (2003) Gobiinae clades (IIB). These results yield high support for the subfamily Benthophilinae and its three component tribes, which are the focus of this study, and show less resolution for deeper branches separating other gobioid families.

Further research will be necessary to fully identify the arrangement of the Benthophilinae within the higher-order framework of gobioid systematics, which is not the focus of the present work.

Biogeographic patterns

Analysis of divergence times among lineages of Benthophilinae is in general concordance with major geological events in the Ponto-Caspian basin (Fig. 3.4, Table

3.3). The basin has experienced a tumultuous geological history since the mid-Miocene epoch (~15 Mya), including multiple large sea-level changes and intermittent connection with the World Ocean, and associated inter-basin connections between the Black and

Caspian Sea basins (Kaplan 1995; Reid & Orlova 2002). These fluctuating water levels

71 and connections caused salinity levels within the basins to range 1-30 ppt over the last 5

My (Reid & Orlova 2002), resulting in lineage separations on multiple temporal scales.

The initial separation of the Black and Caspian Sea basins ~5 Mya coincides with the diversification of most Neogobiini + Ponticolini genera (Neogobius, Babka,

Mesogobius, Ponticola, and Proterorhinus), as well as diversification within

Benthophilini (separation of Benthophilus and Caspiosoma; Fig. 3.3). Congruently,

Cristescu et al. (2003, 2004) identified late Miocene divergences (~ 5.0-7.9 Mya) for benthic amphipods, and Audzijonyte et al. 2(008b) found a ~ 5 Mya split between

Paramysis lineages. These divergences within diverse Ponto-Caspian fauna occurred on a similar time scale as large-scale desiccation events in the Mediterranean Sea basin

(Messinian Salinity Crisis ~ 5.9 Mya) and in the eastern Paratethys/early Black Sea basin

~5.5 Mya (Hsü & Giovanoli 1979; Gillet et al. 2007). Desiccation of the Black Sea basin during this period dramatically reduced water levels and increased salinity, enhancing isolation among tributaries within the basin. This led to allopatric separation of taxa residing in these freshwater areas and increased speciation within the more saline basin.

In addition to older divergences within Ponto-Caspian taxa, several recent separation events are identified. Onset of the Pleistocene glaciations created additional fluctuations in water levels within the Ponto-Caspian basin (Reid & Orlova 2002).

Several radiation events occurred 1-2 Mya among the Ponticolini, during the midst of these Pleistocene glacial cycles. Notably, in the Ponticola “platyrostris” species group both Po. cephalargoides and Po. cyrius diverged early ~1.82 Mya, and are distributed at opposite ends of the Ponto-Caspian basin today (northwest Black Sea/Azov Sea and the

Kura River basin flowing into the Caspian Sea, respectively). Ponticola constructor, Po.

72 platyrostris, and Po. rhodioni then separated ~1.3 Mya and are found in the central portion of the Ponto-Caspian basin (marine and freshwater areas of the eastern Black

Sea). A similar distribution pattern occurs within the Ponticola “kessleri” species group; with Po. gorlap occupying marine and freshwater areas of the Caspian Sea basin and Don

River, Po. eurycephalus inhabiting marine areas of the northwest Black Sea and Azov

Sea, and Po. kessleri primarily found in freshwater drainages of the northwest Black Sea

(Dnieper, Dniester and Danube Rivers). The freshwater species of Proterorhinus also occupy an analogous distribution, with Pr. semilunaris occurring in freshwater basins of the northwest Black Sea, Proterorhinus sp. found in the Kumo-Manych Depression

(Don/Manych River basin), and Pr. cf semipellucidus inhabiting the upper and lower

Volga River basin and delta (Neilson and Stepien, 2009). These three species groups demonstrate a congruent biogeographic pattern: initial isolation and separation of a broadly distributed taxon following closure of an interbasin connection ~1.7-2.0 Mya

(Apsheron connection through Kumo-Manych Depression; Kaplan 1995; Reid & Orlova

2002), succeeded by isolation and further radiation within the Black Sea basin due to glacially associated fluctuations in water levels and basin shape. In addition, recent water level transgressions and separations within the Pleistocene coincide with lineage divergences of the two subspecies of Neogobius melanostomus (N. m. melanostomus in the Black Sea and N. m. affinis in the Caspian Sea: see Brown & Stepien 2008; Fig. 3.3).

This pattern of Pleistocene-aged phylogenetic and phylogeographic breaks among

Black/Caspian Sea basins is echoed in a variety of taxa ranging from other fishes (Rutilis frisii; Kotlík et al. 2008) to benthic and planktonic aquatic invertebrates (cladocerans –

73 Cristescu et al. 2003, 2004; dreissenid mussels – Stepien et al. 2002; 2003; Gelembiuk et al. 2006; and mysids – Audzijonyte et al. 2006, 2008a).

Our analysis of divergence times generally is congruent with evolutionary hypotheses proposed for European gobiids, primarily in origins of the “transverse gobies”

(Atlantic-Mediterranean Gobius, Caffrogobius, Nematogobius, Mauligobius, Padogobius, and Zosterisessor) and the “sand gobies” (Economidichthys, Knipowitschia, and

Pomatoschistus; Miller 2003a). Penzo et al. (1998), using portions of the mt 12S and

16S rRNA genes, estimated a separation time of the transverse and sand gobies of ~48

Mya; in the present study using cyt b, I resolved a clade containing Gobius +

Zosterisessor (mostly identical to Penzo et al.’s study) separating from a clade containing the sand gobies Knipowitschia and Pomatoschistus ~35 Mya (Fig. 3.3). In addition,

McKay & Miller (1991) found a close relationship between the sand gobies and western

Pacific gobiids using morphology and isozymes, indicating an earlier separation from the transverse gobies. This association also is seen in this study, with the sand gobies appearing closely related to the western Pacific microdesmids and ptereleotrids (cyt b;

Fig. 3.3) or to other western Pacific gobiids (COI; Fig. 3.4). Although Miller (2003a) groups the Ponto-Caspian Benthophilinae as members of the transverse gobiids, it appears that they diverged much earlier from the transverse + sand goby ancestor, ~42

Mya. Since the Benthophilinae shares the transverse pattern of cephalic neuromasts with other Atlantic-Mediterranean gobiids, the “transverse gobies” group is paraphyletic, retaining this ancestral character trait across multiple evolutionary lineages.

74 Conclusion

The goby subfamily Benthophilinae represents an understudied yet important component of the Ponto-Caspian fish fauna. I presented the most complete phylogenetic and biogeographic study of the group, and clarified outstanding taxonomic issues present for the last 20 years. The Benthophilinae constitutes a unique radiation of gobiid fishes, and is a separate subfamily from the remainder of the Gobiidae. Its evolutionary history has been driven by the dynamic geologic and hydrologic evolution of the Ponto-Caspian basin.

Systematic conclusions

Benthophilinae Beling and Iljin 1927 Type genus Benthophilus Eichwald 1831 Distinguishing features: small to moderate gobiids with infraorbital neuromast organs (comprised of sensory papillae) in 6-7 transverse rows, 4 before and 2-3 above hyomandibular row b, and lacking row a. Dorsal supraorbital rows o showing separation along dorsal midline. Tubular anterior nostrils, lacking process from the rim. Posterior nostril generally near orbit. Uppermost pectoral fin rays contained within membrane. Swimbladder not present. Moderate to large oligoplasmatic eggs; no pelagic larval stage. Benthophilinae can be separated from the Gobiinae (where it was formerly included) by generally increased number of total (≤28) and caudal (18-22) vertebrae. Primarily found in the Azov, Black, and Caspian Sea basins and adjacent river drainages; several species introduced into central and northern Europe and the North American Great Lakes.

Tribe Benthophilini Beling and Iljin 1927 Type genus Benthophilus Eichwald 1831 Distinguishing features: small gobiids with infraorbital neuromast organs in 6-7 rows, 4 before and 2-3 above hyomandibular row b, and lacking row a. Benthophilini can be separated from other members of the Benthophilinae by the combination of complete loss of all head canals, and reduction or complete loss of scales.

Genus Anatirostrum Iljin 1930: 48. Type species Benthophilus profundorum Berg 1927.

Included species Anatirostrum profundorum (Berg 1927). Original name: Benthophilus profundorum Berg 1927.

75 Genus Benthophiloides Beling and Iljin 1927: 309 Synonym: Asra Iljin 1941: 384. Type species Benthophiloides brauneri Beling and Iljin 1927.

Included species: Benthophiloides brauneri Beling and Iljin 1927. Benthophiloides turcomanus (Iljin 1941). Original name: Asra turcomanus Iljin 1941.

Genus Benthophilus Eichwald 1831: 77 Synonyms: Bentophilus Eichwald 1838: 102; Hexacanthus Nordmann 1838: 332; Doliichthys Sauvage 1874: 336. Type species Gobius macrocephalus Pallas 1787

Included species: Benthophilus abdurahmanovi Ragimov 1978. Original name: Benthophilus magistri abdurahmanovi Ragimov 1978. Benthophilus baeri Kessler 1877. Benthophilus casachicus Ragimov 1978. Original name: Benthophilus stellatus casachicus Ragimov 1978, Benthophilus ctenolepidus Kessler 1877. Synonym: Benthophilus magistri lencoranicus Ragimov 1982. Benthophilus durrelli Boldyrev and Bogutskaya 2004. Benthophilus granulosus Kessler 1877. Synonym: Benthophilus squamatus Baer in Lukina 1984. Benthophilus grimmi Kessler 1877. Benthophilus kessleri Berg 1927. Original name: Benthophilus grimmi kessleri Berg 1927. Benthophilus leobergius Berg 1949. Original name: Benthophilus stellatus leobergius Berg 1949. Synonym: Benthophilus aculeatus Baer in Lukina 1984. Benthophilus leptocephalus Kessler 1877. Benthophilus leptorhynchus Kessler 1877. Benthophilus macrocephalus (Pallas 1787). Original name: Gobius macrocephalus Pallas 1787. Synonym: Hexacanthus macrocephalus Nordmann 1838. Benthophilus magistri Iljin 1927. Benthophilus mahmudbejovi Ragimov 1976. Benthophilus nudus Berg 1898. Original name: Benthophilus macrocephalus nudus Berg 1898. Synonym: Benthophilus macrocephalus ponticus Berg 1916. Benthophilus pinchuki Ragimov 1982. Original name: Benthophilus ctenolepidus pinchuki Ragimov 1982. Benthophilus ragimovi Boldyrev and Bogutskaya 2004. Benthophilus spinosus Kessler 1877. Benthophilus stellatus (Sauvage 1874). Original name: Doliichthys stellatus Sauvage 1874. Synonyms: Benthophilus macrocephalus maeotica Kuznetsov 1888; Benthophilus monstrosus Kuznetsov 1888. Benthophilus svetovidovi Pinchuk and Ragimov 1979.

76 Genus Caspiosoma Iljin 1927: 129. Type species Gobiosoma caspium Kessler 1877

Included species: Caspiosoma caspium (Kessler 1877). Original name: Gobiosoma caspium Kessler 1877.

Tribe Neogobiini new tribe, Neilson and Stepien Type genus Neogobius Iljin 1927 Distinguishing features: moderate gobiids with infraorbital neuromast organs in 7 rows, 4 before and 3 above hyomandibular row b, and lacking row a. Neogobiini can be separated from other members of the Benthophilinae by the following characters: head width about equal to depth; metapterygoid bridge absent; dentary with generally small teeth, largest in the outer row.

Genus Neogobius Iljin 1927: 135 Synonyms: Apollonia (subgenus of Gobius) Iljin 1927: 133; Neogobius (subgenus of Gobius) Iljin 1927: 135 Type species: Gobius fluviatilis Pallas 1814

Included species: Neogobius fluviatilis (Pallas 1814). Original name: Gobius fluviatilis Pallas 1814. Synonyms: Gobius sordidus Bennett 1835; Gobius lacteus Nordmann 1840; Gobius stevenii Nordmann 1840; Gobius niger Eichwald 1841 (not of Linnaeus 1758); Gobius fluviatilis nigra Kessler 1859; Gobius fluviatilis pallasi Berg 1916; Gobius caspius Ragimov 1967. Other combination: Apollonia fluviatilis Stepien and Tumeo 2006. Neogobius melanostomus (Pallas 1814). Original name: Gobius melanostomus Pallas 1814. Synonyms: Gobius cephalarges Pallas 1814; Gobius chilo Pallas 1814; Gobius melanio Pallas 1814; Gobius virescens Pallas 1814; Gobius exanthematosus Pallas 1814; Gobius affinis Eichwald 1831; Gobius sulcatus Eichwald 1831; Gobius lugens Nordmann 1840; Gobius grossholzii Steindachner 1894; Gobius marmoratus Antipa 1909. Other combinations: Gobius melanostomus affinis Navozov 1912; Apollonia melanostomus Stepien and Tumeo 2006. Neogobius caspius (Eichwald 1831). Original name: Gobius caspius Eichwald 1831. Other combinations: Gobius (Eichwaldia) caspius Smitt 1900; Neogobius (Eichwaldia) caspius Gaibova 1952.

Tribe Ponticolini new tribe, Neilson and Stepien Type genus Ponticola Iljin 1927 Distinguishing features: moderate gobiids with infraorbital neuromast organs in generally 7 rows, 4 before and 3 above hyomandibular row b, and lacking row a. Ponticolini can be separated from other members of the Benthophilinae by the following characters: metapterygoid bridge present; hyomandibular generally narrow (breadth generally <100% length).

77 Genus Babka Iljin 1927: 132. Synonym: Babka (subgenus of Gobius) Iljin 1927: 132. Type species Gobius gymnotrachelus Kessler 1857.

Included species: Babka gymnotrachelus (Kessler 1857). Original name: Gobius gymnotrachelus Kessler 1857. Synonyms: Gobius macropus De Filippi 1863; Gobius burmeisteri Kessler 1877; Gobius macrophthalmus Kessler 1877; Mesogobius gymnotrachelus otschakovinus Zubovitch 1925. Other combinations: Mesogobius gymnotrachelus Berg 1916; Gobius (Babka) gymnotrachelus Iljin 1927; Gobius (Mesogobius) gymnotrachelus Sözer 1941; Mesogobius gymnotrachelus macrophthalmus Berg 1949; Gobius (Babka) gymnotrachelus gymnotrachelus Bănărescu 1964; Gobius gymnotrachelus macrophthalmus Ragimov 1967; Neogobius gymnotrachelus Miller 1973; Neogobius gymnotrachelus gymnotrachelus Pinchuk 1977; Neogobius gymnotrachelus macrophthalmus Pinchuk 1977.

Genus Mesogobius Bleeker 1874: 317. Synonym: Mesogobius (subgenus of Gobius) Bleeker 1874: 317. Type species Gobius batrachocephalus Pallas 1814

Included species: Mesogobius batrachocephalus (Pallas 1814). Original name: Gobius batrachocephalus Pallas 1814. Synonym: Gobius batrachocephalus borysthenis Pinchuk 1963. Other combinations: Gobius (Mesogobius) batrachocephalus Bleeker 1874; Gobius batrachocephalus batrachocephalus Smitt 1900. Mesogobius nigronotatus (Kessler 1877). Original name: Gobius nigronotatus Kessler 1877. Mesogobius nonultimus (Iljin 1936). Original name: Gobius nonultimus Iljin 1936. Other combination: Mesogobius batrachocephalus nonultimus Miller 1986.

Genus Ponticola Iljin 1927: 134 Synonym: Ponticola (subgenus of Gobius) Iljin 1927: 134 Type species: Gobius ratan Nordmann 1840

Included species: Ponticola bathybius (Kessler 1877). Original name: Gobius bathybius Kessler 1877. Other combinations: Neogobius (Chasar) bathybius Berg 1949; Neogobius fluviatilis pallasi Berg 1949; Gobius (Chasar) bathybius Ragimov 1967a; Gobius bathybius Pinchuk 1976; Neogobius bathybius Pinchuk and Ragimov 1985. Ponticola cephalargoides (Pinchuk 1976). Original name: Neogobius cephalargoides Pinchuk 1976. Synonyms: Gobius syrman Kessler 1859; Gobius constructor Kessler 1874; Gobius cephalarges Chichkoff 1912; Gobius (Ponticola) cephalarges Borcea 1934; Neogobius cephalarges Georghiev, Aleksandrova and Nikolayev 1960; Gobius ratan 1963; Gobius (Ponticola) cephalarges cephalarges Bănărescu 1964; Neogobius ratan Zambriborshch 1968;; Neogobius

78 cephalarges cephalarges Smirnov 1986. Ponticola constructor (Nordmann 1840). Synonyms: Gobius constructor Nordmann 1840; Gobius platyrostris cyrius Kessler 1879; Gobius platyrostris Berg 1916; Gobius platyrostris cyrius Berg 1923; Gobius cephalarges Iljin (1926) 1927; Gobius (Ponticola) platyrostris cyrius Iljin 1927a; Gobius cephalarges constructor Iljin 1927b; Neogobius cephalarges constructor Berg 1949; Neogobius platyrostris constructor Pinchuk 1977. Other combination: Neogobius constructor Vasil’eva and Vasil’ev 1994. Ponticola cyrius (Kessler 1874). Synonyms: Gobius cyrius Kessler 1874; Gobius weidemanni Kessler 1874; Gobius platyrostris cyrius Berg 1916; Gobius constructor Berg 1923; Gobius platyrostris Berg 1923; Gobius cephalarges constructor Iljin 1927; Neogobius cephalarges constructor Berg 1949; Neogobius platyrostris constructor Pinchuk 1977. Other combination: Neogobius cyrius Vasil’eva and Vasil’ev 1994. Ponticola eurycephalus (Kessler 1874). Original name: Gobius eurycephalus Kessler 1874. Synonyms: Gobius cephalarges Nordmann 1840; Gobius platyrostris Ul’janin 1871; Gobius constructor Kessler 1874; Gobius (Ponticola) cephalarges Iljin 1927; Neogobius cephalarges Berg 1949; Gobius cephalarges Pinchuk 1963; Gobius (Ponticola) cephalarges cephalarges Bănărescu 1964; Neogobius platyrostris Georgiev 1966; Neogobius cephalarges Bogachik and Remez 1970; Neogobius platyrostris eurycephalus Pinchuk 1977; Neogobius platyrostris odessicus Pinchuk 1977. Other combinations: Gobius eurycephalus eurycephalus Smitt 1900; Neogobius eurycephalus Miller 1986. Ponticola gorlap (Iljin in Berg 1949). Original name: Neogobius kessleri gorlap Iljin in Berg 1949. Synonyms: Gobius batrachocephalus Eichwald 1841; Gobius kessleri Kessler 1874; Gobius platyrostris cyrius Derzhavin 1926; Gobius cephalarges constructor Derzhavin 1934; Gobius kessleri gorlap Chugunova 1946; Neogobius cephalarges constructor Berg 1949; Neogobius kessleri Oliva 1960; Neogobius iljini Vasil’eva and Vasil’ev 1996. Other combinations: Neogobius (Ponticola) kessleri gorlap Gaibova 1952; Gobius gorlap Iljin 1956. Ponticola kessleri (Günther 1861). Original name: Gobius kessleri Günther 1861. Synonyms: Gobius platyrostris Nordmann 1840; Gobius platycephalus Kessler 1857; Gobius cephalarges Steindachner 1870; Gobius batrachocephalus platycephalus Smitt 1900; Gobius trautvetteri Antipa 1909; Gobius (Ponticola) platyrostris Borcea 1934. Other combinations: Gobius (Ponticola) kessleri Iljin 1927; Neogobius kessleri Berg 1949; Neogobius kessleri kessleri Pinchuk 1977. Ponticola platyrostris (Pallas 1814). Original name: Gobius platyrostris Pallas 1814. Synonyms: Gobius cephalarges platyrostris Smitt 1900; Gobius cephalarges Smirnov 1959. Other combinations: Gobius (Ponticola) platyrostris Iljin 1927; Neogobius platyrostris Berg 1949; Neogobius platyrostris platyrostris Pinchuk 1977. Ponticola ratan (Nordmann 1840). Original name: Gobius ratan Nordmann 1840. Synonyms: Gobius bogdanowi Kessler 1874; Gobius goebelii Kessler 1874; Gobius trautvetteri Kessler 1874. Other combinations: Gobius cephalarges ratan Smitt 1900; Gobius cephalarges bogdanowi Smitt 1900; Gobius cephalarges goebelii Smitt 1900; Gobius rotan Iljin 1927a; Gobius (Ponticola) ratan Iljin

79 1927b; Neogobius ratan Berg 1949; Neogobius ratan goebeli Berg 1949; Neogobius bogdanowi Berg 1949; Gobius ratan goebeli Iljin 1956; Gobius ratan Pinchuk 1963; Neogobius (Ponticola) ratan ratan Bănărescu 1964; Neogobius ratan Zambriborshch 1968; Neogobius ratan ratan Pinchuk 1976. Ponticola rizensis (Kovačić and Engín 2008). Original name: Neogobius rizensis Kovačić and Engín 2008. Ponticola rhodioni (Vasil’eva and Vasil’ev 1994). Original name: Neogobius rhodioni Vasil’eva and Vasil’ev 1994. Synonyms: Gobius constructor Nordmann 1840; Gobius platyrostris cyrius Kessler 1879; Gobius platyrostris Berg 1923; Gobius platyrostris cyrius Berg 1923; Gobius cephalarges Iljin (1926) 1927; Gobius (Ponticola) platyrostris cyrius Iljin 1927a; Gobius cephalarges constructor Iljin 1927b; Neogobius cephalarges constructor Berg 1949; Neogobius platyrostris constructor Pinchuk 1977; Ponticola syrman (Nordmann 1840). Original name: Gobius syrman Nordmann 1840. Synonyms: Gobius trautvetteri Kessler 1859; Gobius eurystomus Kessler 1877; Gobius constructor Borcea 1934. Other combinations: Gobius (Ponticola) syrman Iljin 1927; Neogobius syrman Berg 1949; Neogobius syrman eurystomus Berg 1949; Neogobius (Ponticola) syrman eurystomus Gaibova 1952; Gobius (Ponticola) syrman eurystomus Iljin 1956; Gobius (Ponticola) syrman syrman Bănărescu 1964; Gobius syrman eurystomus Ragimov 1967; Neogobius syrman syrman Smirnov 1986. Ponticola turani (Kovačić and Engín 2008). Original name: Neogobius turani Kovačić and Engín 2008

Genus Proterorhinus Smitt 1900: 544 Synonym: Proterorhinus (subgenus of Gobius) Smitt 1900: 544 Type species Gobius marmoratus Pallas 1814

Included species: Proterorhinus marmoratus (Pallas 1814). Original name: Gobius marmoratus Pallas 1814. Synonyms: Gobius quadricapillus Pallas 1814; Gobius macropterus Nordmann 1840. Proterorhinus nasalis (De Fillipi 1863). Original name: Gobius nasalis De Fillipi 1863. Synonym: Gobius blennioides Kessler 1877. Proterorhinus semilunaris (Heckel 1837). Original name: Gobius semilunaris Heckel 1837. Synonym: Gobius rubromaculatus Kriesch 1873. Proterorhinus cf semipellucidus (Kessler 1877). Synonyms: Gobius semipellucidus Kessler 1877. Proterorhinus tataricus Freyhof and Naseka 2008.

80 Table 3.1 Taxonomic names, geographic origin, GenBank accession numbers, specimen ID, and type species for the genus (*) and

tribe (**) for individuals/taxa analyzed in the present study

GenBank Accession nos. Former taxon and Specimen author Proposed new name Common name Location Latitude Longitude ID cyt b COI RAG1 S7 Tribe Neogobiini Neogobius = Apol- lonia Iljin, 1927 A. fluviatilis = Monkey Goby Danube River, Vilkove, 45.393989 29.586870 AGV7 FJ526749 FJ526804 FJ526858 FJ526913 N. fluviatilis** Ukraine (Pallas, 1814) Sea of Azov, Molochnyi, 46.655616 35.278634 AGV9 FJ526750 FJ526805 FJ526859 FJ526914 Ukraine Ozero Manych, Prujitnoe, 46.016147 43.448435 ANT5 FJ526753 FJ526808 FJ526862 FJ526917 Russia Volga River, Volgograd, 48.870870 44.660139 ALL13 FJ526751 FJ526806 FJ526860 FJ526915 Russia

8 Chernozemelskii Canal, near 46.272008 45.615373 ANG11 FJ526752 FJ526807 FJ526861 FJ526916 1 Elista, Russia

A. melanostoma = Round Goby Dnieper River, Kiev, 50.270000 30.300000 AHC3 EU331208 FJ526799 FJ526853 FJ526908 N. melanostomus Ukraine (Pallas, 1814) Black Sea, Sevastopol, 44.604040 33.540840 AHF8 EU331225 FJ526800 FJ526854 FJ526909 Ukraine Kerch Strait, Kerch, Ukraine 45.358334 36.475834 APC8 EU331173 FJ526803 FJ526857 FJ526912 Volga River, Svetli Yar, 48.484638 44.784676 AMP2 EU331175 FJ526802 FJ526856 FJ526911 Russia Caspian Sea, Nabran, 41.837222 48.620000 AKB1 EU331186 FJ526801 FJ526855 FJ526910 Azerbaijan

N. caspius Caspian Goby Caspian Sea, Nabran, 41.837222 48.620000 APT1 FJ526756 FJ526811 FJ526865 FJ526921 (Eichwald, 1831) Azerbaijan Caspian Sea, Sumgait, 40.600278 49.682222 ALK6 FJ526757 FJ526812 FJ526866 FJ526922 Azerbaijan Tribe Ponticolini Mesogobius Bleeker, 1874 Mesogobius batra- Knout Goby Kanev Reservoir, Kiev, 50.270000 30.300000 AJE3 FJ526755 FJ526810 FJ526864 FJ526920 chocephalus* Ukraine (Pallas, 1814) Black Sea, Odessa, Ukraine 46.470820 30.735090 AGV10 EU444668 EU444697 EU444723 FJ526918 Sea of Azov, Molochnyi, 46.655616 35.278634 AGV11 FJ526754 FJ526809 FJ526863 FJ526919 Ukraine

Neogobius Iljin, Ponticola Iljin, 1927 1927 N. cephalargoides Ponticola cephalarg- Pinchuk's Dniester River Estuary, 46.066667 30.450000 ALC12 FJ526758 FJ526813 FJ526867 FJ526923 Pinchuk, 1976 oides (Pinchuk, 1976) Goby Ukraine Kerch Strait, Kerch, Ukraine 45.358334 36.475834 AGT10 FJ526773 FJ526828 FJ526882 FJ526939 Dniester River Estuary, 46.066667 30.450000 ATW11 FJ526794 FJ526850 FJ526904 FJ526978 Ukraine

N. constructor Po. constructor Constructor Khobi River, Khobi, Geor- 42.317778 41.916111 ATW05 FJ526790 FJ526846 FJ526900 FJ526974 (Nordmann, 1840) (Nordmann, 1840) Goby gia Otap River, Otap, Georgia 42.918333 41.541944 ATW06 FJ526791 FJ526847 FJ526901 FJ526975

N. cyrius Po. cyrius Kura Goby Aragvi River, Tsiteltsopeli, 41.993611 44.760278 ATW03 FJ526788 FJ526844 FJ526898 FJ526972 (Kessler, 1874) (Kessler, 1874) Georgia Ptsa River, Georgia 42.049720 43.728890 ATW04 FJ526789 FJ526845 FJ526899 FJ526973

N. eurycephalus Po. eurycephalus Ginger Goby Cape Langeron, Odessa Bay, 46.483333 30.755000 ALC7 FJ526759 FJ526814 FJ526868 FJ526924 (Kessler, 1874) (Kessler, 1874) Ukraine

8 Sukhyi Estuary, Burlachya 46.326700 30.667550 ALC8 FJ526760 FJ526815 FJ526869 FJ526925 2 Balka, Ukraine Cape Malyi Fontan, Odessa 46.450000 30.766667 ALC9 FJ526761 FJ526816 FJ526870 FJ526926 Bay, Ukraine

N. gorlap Po. gorlap Caspian Big- Karpovska Reservoir, 48.643269 43.617069 AKS4 FJ526762 FJ526817 FJ526871 FJ526927 Iljin, 1949 (Iljin, 1949) head Goby Iliovka, Russia. Caspian Sea, Nabran, 41.837222 48.620000 APT3 FJ526763 FJ526818 FJ526872 FJ526928 Azerbaijan Caspian Sea, Lenkoran, 38.751944 48.868889 APT4 FJ526764 FJ526819 FJ526873 FJ526929 Azerbaijan

N. kessleri Po. kessleri Bighead Goby Danube River, Dobra, Serbia 44.638100 21.909400 APT8 FJ526770 FJ526825 FJ526879 FJ526936 (Günther, 1861) (Günther, 1861) Dniester River, Yampil, 48.235344 28.293024 ALC2 FJ526768 FJ526823 FJ526877 FJ526934 Ukraine Simferopol Reservoir, Sim- 44.921746 34.155719 APT7 FJ526769 FJ526824 FJ526878 FJ526935 feropol, Ukraine

N. platyrostris Po. platyrostris Flatsnout Kerch Strait, Kerch, Ukraine 45.358334 36.475834 AGT7 FJ526771 FJ526826 FJ526880 FJ526937 (Pallas, 1814) (Pallas, 1814) Goby Kerch Strait, Kerch, Ukraine 45.358334 36.475834 AGT9 FJ526771 FJ526827 FJ526881 FJ526938

N. ratan Po. ratan** Ratan Goby Sea of Azov, Ukraine 45.782058 35.487513 ATW07 FJ526792 FJ526848 FJ526902 FJ526976 (Nordmann, 1840) (Nordmann, 1840) Sea of Azov, Ukraine 45.782058 35.487513 ATW08 FJ526793 FJ526849 FJ526903 FJ526977

N. rhodioni Po. rhodioni Rhodion's Vostochnyy Dagomys River, 43.703330 39.688890 ATW01 FJ526786 FJ526842 FJ526896 FJ526970 Vasil'eva and (Vasil'eva and Vasil'ev, Goby Baranovka, Russia Vasil'ev, 1994 1994) Kherota River, Moldovka, 43.464440 39.95333 ATW02 FJ526787 FJ526843 FJ526897 FJ526971 Russia

N. syrman Po. syrman Syrman Goby Danube River, Vilkove, 45.393989 29.586870 AGV4 FJ526774 FJ526829 FJ526883 FJ526940 (Nordmann, 1840) (Nordmann, 1840) Ukraine Danube River, Vilkove, 45.393989 29.586870 AJE8 FJ526775 FJ526830 FJ526884 FJ526941 Ukraine

Neogobius Iljin, Babka Iljin, 1927 1927 Neogobius gymno- Babka Racer Goby Dniester River delta, 46.468333 30.216667 AMU6 FJ526765 FJ526820 FJ526874 FJ526930 trachelus (Kessler, gymnotrachelus* Bilyayivka, Ukraine 1857) (Kessler, 1857) Dnieper River, Kiev, 50.270000 30.300000 AGT1 FJ526766 FJ526821 FJ526875 FJ526931 Ukraine Kanev Reservoir, Kiev, 50.270000 30.300000 AGT3 EU444667 EU444694 EU444720 FJ526932

8 Ukraine 3 Tyligul Estuary, Ukraine 46.470820 30.735090 AGT2 FJ526767 FJ526822 FJ526876 FJ526933

Proterorhinus (Pal- las, 1814) Proterorhinus mar- Marine Tuben- Dniester River delta, 46.468333 30.216667 AME1 EU444621 EU444682 EU444708 FJ526942 moratus* (Pallas, ose Goby Bilyayivka, Ukraine 1814) Cape Langeron, Odessa Bay, 46.483333 30.755000 AMM1 EU444624 EU444687 EU444713 FJ526944 Ukraine Tyligul Estuary, Ukraine 46.690000 31.486783 AMG1 EU444621 EU444684 EU444710 FJ526943 Black Sea, Sevastopol, 44.604040 33.540840 AMR1 EU444621 EU444689 EU444715 FJ526945 Ukraine

Pr. semilunaris Freshwater Lake Superior, MI, USA 46.666667 -92.200000 AOC2 EU444607 EU444690 EU444716 FJ526948 (Heckel, 1837) Tubenose Goby Lake St. Clair, Michigan, 42.594282 -82.803323 AGN1 EU444607 EU444674 EU444700 FJ526949 USA Danube River, Dobra, Serbia 44.638100 21.909400 AKP7 EU444612 EU444677 EU444703 FJ526951 Dniester River, Mohyliv- 48.449428 27.778285 AFE10 EU444604 EU444673 EU444699 FJ526946 Podil'sky, Ukraine Kurchurgan Reservoir, 46.100000 30.200000 AML1 EU444632 EU444686 EU444712 FJ526950 Hradenytsi, Ukraine Cape Malyi Fontan, Odessa 46.450000 30.766667 AMF2 EU444626 EU444683 EU444709 FJ526947 Bay, Ukraine Simferopol Reservoir, Sim- 44.921746 34.155719 AQE1 EU444650 EU444691 EU444717 FJ526952 feropol, Ukraine

Pr. cf semipellu- Volga Tuben- Karpovska Reservoir, 48.643269 43.617069 AKP1 EU444610 EU444675 EU444701 FJ526955 cidus ose Goby Iliovka, Russia. Neilson and Stepi- en, 2009 Chagraiskoye Reservoir, 45.617691 44.211077 AMK1 EU444630 EU444685 EU444711 FJ526954 Zunda Tolga, Russia Volga River, Preshib, Russia 47.683923 46.509057 ALT1 EU444610 EU444679 EU444705 FJ526957 Volga River delta, Russia 45.788350 47.886953 AKP4 EU444611 EU444676 EU444702 FJ526956 Volga River, Volgograd, 48.870870 44.660139 ALU1 EU444611 EU444680 EU444706 FJ526958 Russia

Proterorhinus sp. Chernozemelskii Canal, 46.272008 45.615373 AMN1 EU444636 EU444688 EU444714 FJ526953 Neilson and Stepi- Elista, Russia en, 2009 8 4 Table 3.2 Summary of maximum parsimony results from individual genes and from combined dataset using PAUP* v4b10 (Swofford 2003)

Gene N trees Length CI RI RC HI COI 6 1541 0.447 0.868 0.388 0.553 cyt b 64 1639 0.475 0.882 0.419 0.525 RAG 1 533503 215 0.758 0.950 0.720 0.242 S7 6.3 x 106 367 0.730 0.902 0.659 0.270 combined 54 3879 0.490 0.875 0.428 0.510 CI – consistency index; RI – retention index; RC – rescaled consistency index; HI – homoplasy index.

85 Table 3.3 Divergence times for major lineages/nodes within phylogeny of the new subfamily Benthophilinae, showing ages estimated for the extended cyt b tree (Fig. 3.3) using penalized likelihood in R8S (Sanderson 2003). Nodes representing fixed ages (A and B) and major geologic events in the Ponto-Caspian basin (I-III) are indicated on Fig. 3.3.

Major geologic event in Ponto-Caspian basin (Reid Node Estimated age (Mya) and Orlova, 2002) MRCA of Gobiidae (A; fixed at 53.00) 53.00 MRCA of Neogobius sensu stricto, Babka + Ponticola, and Proterorhinus 10.00 Separation of Ponto-Caspian and Pannonian basins (B; fixed at 10.00) (~12.5-10 Mya; I) Tribe Neogobiini (Neogobius) + Tribe Benthophilini “tadpole gobies” 9.18 (Benthophilus + Caspiosoma) Tribe Ponticolini (Babka + Ponticola + Mesogobius + Proterorhinus) 7.58 Intermittent connections with World Ocean, with intro- gression of marine fauna (8.3-6.4 Mya) Proterorhinus + Mesogobius 6.25 Brief reconnection with Pannonian basin and immigra- tion of endemic Pannonian fauna (6.4-5.8 Mya) Tribe Neogobiini = Neogobius sensu stricto (N. fluviatilis, N. melanostomus, 5.47 Separation of Black and Caspian basins (5.8-5.0 Mya - 8

6 and N. caspius) coincides with Messinian salinity crisis in paleo-Medi- terranean and Black Sea basins; II) Babka (B. gymnotrachelus) + Ponticola (Po. cepharlargoides, Po. con- 5.08 structor, Po. cyrius, Po. eurycephalus, Po. gorlap, Po. kessleri, Po. platyr- Single, large lake in southern Caspian basin (5.2-2.5 ostris, Po. ratan, Po. rhodioni, and Po. syrman) Mya, II) Tribe Benthophilini “tadpole gobies” (Benthophilus + Caspiosoma)) 5.04 marine and freshwater tubenose gobies Proterorhinus 4.29 Ponticola 4.07 Benthophilus 2.17 Black and Caspian basins connected via Kumo-Ma- Ponticola “platyrostris group” (Po. cephalargoides, Po. constructor, Po. 1.82 nych depression, faunal exchange between basins; gla- cyrius, Po. platyrostris, and Po. rhodioni) cially-driven fluctuations in water levels (2.6-0.7 Mya; III) Ponticola “kessleri group” (Po. eurycephalus, Po. gorlap, and Po. kessleri) 1.37 freshwater Proterorhinus 1.18 Fig. 3.1 Current range (excluding introduced range in North America) of nominal species of the new subfamily Benthophilinae

(hatched area; based on Miller 2003a), and locations of taxa sampled in the present study. K-M = Kumo-Manych Depression. 8 7 Fig. 3.2 Maximum likelihood phylogeny (PHYML; Guindon & Gascuel 2003) of the new subfamily Benthophilinae and outgroups based on combined analysis of four gene regions. Numbers at nodes indicate likelihood bootstrap support (2,000 pseudoreplications), with * = 100%.

88 Fig. 3.3 Chronogram for the new subfamily Benthophilinae and related Ponto-Caspian gobies, derived from a penalized likelihood analysis of divergence time (R8S; Sanderson 2003) and maximum likelihood analysis of the extended cyt b dataset. Nodes with fixed ages in divergence time analysis are lettered; numbers around branches indicate support values from phylogenetic analyses of the extended cyt b dataset (likelihood bootstrap). 8 9 Fig. 3.4 Maximum likelihood phylogeny of the new goby subfamily Benthophilinae and other gobioids based on COI sequence data from the present study and from Thacker (2003). Numbers above branches indicate likelihood bootstrap support. In clades spanning multiple families/subfamilies, symbols adjacent to species names indicate familial/subfamilial membership. Pertinent clade described in Thacker (2003) is labeled (IIB).

90 Chapter Four

Historic cryptic speciation and recent colonization of Eurasian monkey gobies (Neogobius fluviatilis and N. pallasi) revealed by DNA sequences, microsatellites, and morphology

Abstract

Species introductions often provide unplanned “natural” experiments to examine ecological and evolutionary theory. Discerning colonization dynamics and pathways is essential for understanding the ecology of invasions, and for their proper management and mitigation. The monkey goby (Neogobius fluviatilis) is one of many Ponto-Caspian species to expand outside its native range into other European freshwater and estuarine habitats in northern/central Europe and is predicted to become successfully established in

North America if introduced. I analyzed the genetic variability, population structure, and divergence of N. fluviatilis across its native range and from introduced locations in the

Vistula and Danube River basins, using nuclear RAG1 and mitochondrial cytochrome b

DNA sequences, 10 nuclear microsatellite loci, and morphology. A large genetic break is identified between taxa in the Black and Caspian Seas, indicating a long-term separation dating to ~3 million years. Based on DNA sequences, microsatellite allele frequencies, and morphology I recognized two separate species of monkey gobies that were formerly identified as subspecies: N. fluviatilis in the Black Sea basin, Don and Volga Rivers, and the Kumo-Manych Depression; and N. pallasi in the Caspian Sea and Volga River delta.

91 Introduced populations of N. fluviatilis in the Vistula and Danube River basins trace their origins to northwest Black Sea populations and have reduced genetic diversity, similar to other introduced Ponto-Caspian gobiids. This study provides a genetic baseline for analyzing potential future introductions of N. fluviatilis and N. pallasi, and highlights the use of molecular tools in characterizing invasions and identifying cryptic species.

Introduction

Species introductions worldwide provide valuable, large-scale 'natural' experiments to test fundamental hypotheses underlying ecological and evolutionary theory (Sax et al. 2007; Blackburn 2008; Prentis et al. 2008). Comparing invasive species with their invasive/non-invasive congeners can yield insights on the broad morphological, genetic, ecological, or life history characteristics associated with success and persistence (Mondor et al. 2007; Dutton & Hofmann 2008; Cincotta et al. 2009). In addition, comparisons among native and invasive populations can provide a finer-grained

'natural' experimental approach to illuminate individual biological or environmental factors that may contribute to invasion success (Shwartz et al. 2009), and how taxa may adapt and react to novel ecological and environmental pressures (Campbell &

Echternacht 2003). Genetic analysis of invasive species may also reveal cryptic diversity within a “single” widespread species (Geller et al. 1997; Andreakis et al. 2007; Chapter

2), prompting reassessment of invasion pathways and taxonomy.

Many of the introduced species in freshwater systems in North America and western/central Europe originate from the Ponto-Caspian region (Azov, Black, and

Caspian Seas and their associated tributaries; Fig. 4.1). The geological and hydrological history of the Ponto-Caspian basins from the mid-Miocene epoch (~15 million years ago;

92 Mya) to the present has regulated the distribution and genetic diversity of its aquatic fauna. Fluctuating water levels associated with Pliocene and Pleistocene glacial cycles

(Reid & Orlova 2002) frequently isolated historic faunas in both the Black and Caspian

Sea basins and their respective watersheds, allowing for local adaptation and division of once widespread taxa into subspecies, sister species, or other distinct genetic lineages in a variety of taxonomic groups. Shifting salinity regimes, due to periodic glacial meltwater influx or saline marine water introgression, further stratified and isolated faunas within the Black and Caspian watersheds. These salinity shifts selected for a high degree of euryhalinity within Ponto-Caspian taxa, which is one of the factors presumed to fuel their success as invasive species (Ricciardi & Rasmussen 1998; Ricciardi & MacIsaac 2000;

Reid & Orlova 2002).

These large-scale environmental changes throughout the Ponto-Caspian region’s history have dramatically shaped the distribution of taxa within its basins. The Caspian basin, with its long history of isolation from both the World Ocean and the Black Sea basin dating to the end of the Miocene epoch (~5 Mya: Dumont 1998; Reid & Orlova

2002), has a high degree of endemism (up to 80% of species: Dumont 1998, 2000).

Audzijonyte et al. (2006) described the phylogeographic patterns of several mysid crustaceans, identifying genetic breaks between the Black and Caspian basins in most widespread species, as well as divergence within the Caspian Sea. Cristescu et al. (2003) found a similar pattern of genetic division among Black and Caspian basins for

Pontogammarus amphipods and onychopod cladocerans. Phylogeographic and phylogenetic studies of fishes also identified distinct genetic lineages endemic to the

93 Black and Caspian basins, which generally corresponded to either sub-species (Brown &

Stepien 2008; Kotlík et al. 2008) or species level separations (Chapters 2, 3).

The monkey goby Neogobius fluviatilis (Teleostei: Gobiidae) is one of a suite of

Ponto-Caspian species to expand and become established outside of its native range during the past 50 years (Fig. 4.1). It is the type species for its genus, which also contains two other species – the highly invasive round goby N. melanostomus that has expanded throughout northern and central European watersheds and invaded the North American

Laurentian Great Lakes, and the Caspian Sea endemic N. caspius. Neogobius fluviatilis is euryhaline and broadly distributed throughout the Ponto-Caspian region, inhabiting both inland freshwater habitats as well as saline waters in the Black and Caspian Seas. It is a moderately large goby (to 195 mm total length) that is generally found in shallow, sandy bottom habitats (Pinchuk et al. 2003a). Like other members of the Ponto-Caspian gobiid species flock (Gobiidae: Benthophilinae; Chapter 3), it lacks a swim bladder, males show territorial nesting behavior, and embryos undergo direct development to hatch as juveniles (Pinchuk et al. 2003a). It is an important forage fish for many commercially important species (e.g., the pike-perches Sander spp., Wels catfish Silurus glanis, and burbot Lota lota), and is one of the more important species taken in commercial and recreational gobiid fisheries in eastern Europe (Pinchuk et al. 2003a).

Neogobius fluviatilis is presently divided into two subspecies: N. f. fluviatilis

(Pallas 1814) in the Black and Azov Seas and associated drainages, and N. f. pallasi

(Berg 1916) in the Caspian Sea basin. The two subspecies are distinguished by the presence of a darker lower band on the first dorsal fin, reduced average numbers of lateral

94 scales and longitudinal scale rows, and a shorter snout in N. f. pallasi (Berg 1949), as well as some reported variations in osteology (Pinchuk et al. 2003a).

Phylogeographic and invasion history

The monkey goby was introduced into the Aral Sea in the 1950s (Baimov 1963) and was first detected outside of its natural range in the Danube River in 1965 (Bănărescu

1970), where it has since continued to expand upstream through Austria and Slovakia

(Ahnelt et al. 1998; Harka & Jakab 2001; Holčík et al. 2003; Jurajda et al. 2005). It was first discovered in the Baltic Sea basin in the Western Bug River in 1997 (Danilkiewicz

1998), moved upstream into the Vistula River basin in Poland (Kostrzewa & Grabowski

2002), and in 2006 was identified in the Aegean Sea watershed (Stefanov et al. 2008). In addition, Kolar & Lodge (2002) predicted that N. fluviatilis would become established and rapidly spread if introduced in the North American Great Lakes, due to its shared life history and ecological characteristics with its invasive congener, the round goby N. melanostomus. The latter spread throughout the Great Lakes and surrounding rivers since its ballast water introduction in ~1990 and is now one of the most abundant benthic fishes in the lower Great Lakes (Johnson et al. 2005; Bergstrom et al. 2008).

Although the monkey goby has shown moderate success in colonization and establishment outside of its native range, it has been less successful as an invasive species than N. melanostomus (Brown & Stepien 2008; 2009). Čápová et al. (2008) examined ontogenetic variability in the external morphology of N. f. fluviatilis from the River Hron,

Slovakia, and found it exhibited a lower degree of allometric growth than two co- occurring invasive relatives, N. melanostomus and the bighead goby Ponticola kessleri.

They concluded that the low degree of allometry and early acquisition of the adult shape

95 phenotype in N. f. fluviatilis would limit its adaptive potential to novel habitats, and thus reduce its rate of spread and potential impact, relative to its invasive relatives that show a larger degree of ontogenetic variation in morphological plasticity. Kakareko et al. (2005) identified both temporal and diet partitioning between introduced populations of N. f. fluviatilis and racer goby Babka gymnotrachelus in the Vistula River, Poland. They found that N. f. fluviatilis was more active during the day and consumed a higher proportion of chironomid larvae than B. gymnotrachelus, which was more active at night and consumed more cladocerans and amphipods (Kakareko et al. 2005). In the

Hungarian section of the Danube River differential habitat use was found among introduced gobiids, with N. f. fluviatilis being more abundant along natural shorelines and gravel or sandy bottom habitats in comparison to N. melanostomus and P. kessleri, which preferred artificial rip-rap habitats (Erős et al. 2005).

Objectives and hypotheses

The primary objective of this study was to investigate the range-wide phylogeographic and population genetic patterns of N. fluviatilis. I used these patterns to evaluate separations of taxa (e.g., distinct species/subspecies in Black and Caspian Sea basins), investigate the role of Ponto-Caspian geologic history in shaping present-day genetic diversity, and to infer potential colonization and spread dynamics in introduced populations. Specifically, I tested the following hypotheses underlying the general phylogeographic history of the monkey goby, and its introduction to the North Sea basin:

1) Populations/lineages within the Black and Caspian Sea basins do/do not comprise distinct taxa,

96 2) Separation of the Black and Caspian Sea populations/lineages does/does not correspond to major hydrogeological events in the evolution of the Ponto-Caspian basin, and

3) The introduced Vistula River population was/was not founded by colonizers originating from the northwestern Black Sea region (e.g, Dniester River).

These hypotheses were tested using both mitochondrial (mt) DNA cytochrome b gene sequences and 10 nuclear microsatellite (µsat) DNA loci, which allow both phylogeographic history of lineages to be inferred (through maternally inherited mtDNA) and fine-scale population genetic relationships to be discerned (via comparisons of biparentally inherited microsatellites with the mtDNA patterns). These genetic data are combined with a survey of morphological variation between basins and with nuclear gene sequences (RAG1) to examine primary clades. Although several studies have examined various aspects of its ecology (Erős et al. 2005; Kakareko et al. 2005; Adamek et al.

2007; Gaygusuz et al. 2007) and morphology (Vasil'eva 1988; Simonović & Nikolic

1995; Čápová et al. 2008), as well as population genetic patterns of several related goby species (Stepien & Tumeo 2006; Ohayon & Stepien 2007; Brown & Stepien 2008, 2009;

Chapter 2), this is the first investigation of the monkey goby's population genetic structure and phylogeography and will provide a genetic baseline for future studies.

Materials and Methods

Sample collection, DNA extraction, and amplification

I analyzed specimens obtained by small beach seine or beam trawl from 18 freshwater and marine localities throughout much of the native range of Neogobius

97 fluviatilis (Fig. 4.1; Table 4.1). Specimens were preserved immediately either in 95% ethanol for molecular analyses or 10% formalin for morphological analyses (following removal of right pectoral fin for genetic study).

Genomic DNA was extracted from fin tissue using Qiagen DNEasy kits (Qiagen,

Inc.; Valencia, CA) using manufacturer’s protocols, stored at 4oC until used for amplification, and archived at -80oC. Amplification and sequencing primers for the mt cytochrome (cyt) b gene were AJG15 (Akihito et al. 2000), H15343 (Chapter 2), H5

(Akihito et al. 2000), and L15066 (Brown & Stepien 2008); primers for RAG1 were

RAG1F1 (Lopez et al. 2004) and R811goby (Chapter 2). All individuals were sequenced for cyt b, whereas a small subset of representative individuals from each major clades identified with cyt b were analyzed for RAG1. PCR amplifications were performed in 25

µL volumes containing 1 unit Taq polymerase, 200 μM each dNTP, 50 mM KCl, 1.5 mM

MgCl2, 10 mM Tris-HCl pH 8.3, 0.5 μM of each primer and ~30 ng of template.

Amplification was conducted on a MJR DYAD thermalcycler (Bio-Rad Laboratories,

Hercules, CA) with an initial 120 sec denaturation at 94oC, followed by 40 cycles of 45 sec at 94oC, a gene specific annealing temperature (cyt b – 52oC; RAG1 – 50oC) for 30 sec, and extension at 72oC for 60 sec; with a final 180 sec extension at 72oC. PCR reactions were checked on 1% agarose gels stained with ethidium bromide, and excess primers and unincorporated nucleotides were removed from successful reactions with spin column purification kits (QIAquick PCR Purification Kit, Qiagen)

Sequencing was performed at the Cornell University Life Sciences Core

Laboratories Center (http://cores.lifesciences.cornell.edu/brcinfo), on Applied Biosystems

Automated 3730 DNA Analyzers (ABI; Foster City, CA). Sequence data were processed

98 and assembled by me, aligned using BIOEDIT v7.0.9 (Hall 1999) and CLUSTAL X v1.8

(Thompson et al. 1997).

Individuals were also analyzed and scored for 13 nuclear microsatellite (µsat) loci, including Ame24, Ame129, and Ame133 from Feldheim et al. (2009); NG70 NG92,

NG115, and NG215 from Vyskočilová et al. (2007); and Nme2, Nme3, Nme4, Nme7,

Nme8, and Nme10 from Dufour et al. (2007). The PCR amplification mixture contained

0.6 units Taq, 50 μM nucleotides, 50 mM KCl, 1.5 mM MgCl2, 10 mM Tris-HCl, 0.5 μM of each primer, and ~30 ng of template in 10 μL. Amplification encompassed a 2 min initial denaturation for 94oC; followed by 35 cycles of 30 sec at 94oC, 1 min annealing at a primer-specific temperature (53oC for cyt b; 50oC for RAG1), and 30 sec at 72oC; with a final 5 min extension at 72oC. Amplification products were diluted 1:50, of which 1 µl was added to 13 µl of an 80:1 formamide/ABI Gene Scan 500 size standard mixture and analyzed on an ABI 3130 Genetic Analyzer using ABI GENEMAPPER v4.0. Output profiles were manually checked to confirm allelic size variants.

Phylogeographic analyses

I estimated the phylogenetic relationships among monkey goby mt and nuclear

DNA sequences using maximum likelihood (ML) with PHYML version 2.4.4 (Guindon

& Gascuel 2003). For the cyt b dataset, ML analyses utilized the GTR+I+Γ model (i =

0.460, α = 0.871) that was selected as the appropriate model using the corrected Akaike information criterion in MODELTEST version 3.7 (Posada & Crandall 1998). ML analyses for the RAG1 sequences used the Tamura-Nei +I model (i=0.872). Nodal support was evaluated using nonparametric bootstrapping (2,000 pseudoreplications).

Two closely related, congeneric species (Chapter 3) were used as outgroups to root the

99 haplotype phylogeny: the round goby N. melanostomus and the Caspian goby N. caspius.

I tested the monophyly of clades on the cyt b haplotype tree under a null hypothesis of random branching processes following Rosenberg (2007).

To evaluate the presence of multiple distinct taxa within N. fluviatilis I used the

Evolutionary Species Concept (Wiley & Mayden, 2000), which defines a species as a group of organisms having its own evolutionary trajectory separate from other such groups in space and time. This concept is similar to the general metapopulation lineage concept of de Queiroz (2005), which underlies most current species concepts and defines species as segments of metapopulation lineages. These two species concepts provide a general, theoretical means for defining species: however, in practice, operational criteria are necessary for species delimitation. I thus further employed the operational criterion of Hebert et al. (2004b) – that interspecific variation is greater or equal to 10x intraspecific variation – to evaluate variation within N. fluviatilis.

Divergence times between the major clades of monkey goby cyt b haplotypes were estimated using the Bayesian Markov Chain Monte Carlo (MCMC) approach implemented in BEAST version 1.4.8 (Drummond et al. 2002; Drummond & Rambaut

2007). MCMC searches were run for 5 million generations with sampling every 1000 generations and the first 500,000 generations discarded as burn-in. The ML tree was used as a starting tree for the MCMC searches, and the model parameters included the

GTR+I+Γ substitution model, an uncorrelated relaxed molecular clock (Drummond et al.

2006), and a coalescent constant size tree prior. Five separate runs (chains) were performed and pooled prior to calculations. I used a cyt b substitution rate of 2.05 % per million years calculated from sister species of Evorthodus gobies across the Isthmus of

100 Panama (Rocha et al. 2005), as used by Brown & Stepien (2008) and Chapter 2. All other model parameters used the default priors. TRACER v1.4 (Rambaut & Drummond

2007) was used to check that likelihood values of the run had reached stationarity and to summarize parameters. BEAST analyses were performed using the computing resources of the Computational Biology Service Unit at Cornell University

(http://cbsuapps.tc.cornell.edu). To examine geographic patterns of cyt b haplotypes, I constructed a statistical parsimony network using TCS v. 1.21 (Clement et al. 2000).

Alternate pathways within the haplotype network were resolved using the criteria outlined in Pfenninger & Posada (2002).

Population genetic analyses

I used ARLEQUIN v3.11 (Excoffier et al. 2005) to identify individual monkey goby mtDNA haplotypes, in comparison with results from other Ponto-Caspian gobies

(Stepien & Tumeo 2006; Chapters 2-3), and to calculate haplotype frequencies. For µsat data, allelic frequencies, conformance to Hardy-Weinberg equilibrium (HWE) expectations, and linkage disequilibrium were evaluated using GENEPOP v4.0.10

(Rousset 2008). Allelic richness was calculated using FSTAT v2.9.3.2 (Goudet 2001).

Levels of significance for HWE and linkage disequilibrium tests were adjusted using standard Bonferroni correction. HWE deviations were tested for heterozygote deficiency or excess, and for the presence of null alleles with MICRO-CHECKER v2.23 (van

Oosterhout et al. 2004).

Genetic data (cyt b and µsat) were analyzed to identify true populations, i.e., samples having significantly divergent gene pools, using the pairwise FST analog θ (Weir

& Cockerham 1984; Weir & Hill 2002) and χ2 contingency tests (Goudet et al. 1996).

101 Fine-scale population relationships have been shown to be better resolved with contingency tests (Balloux & Lugon-Moulin 2002); which are independent from HW assumptions, non-parametric, and less affected by small sample size (Goudet et al. 1996).

The additional use of θ facilitates direct comparisons with other population genetic studies (see Allendorf & Luikart 2007; Stepien et al. 2007). Pairwise comparisons of FST and contingency tests were performed in ARLEQUIN, and table-wide significance levels were adjusted using sequential Bonferroni correction (Rice 1989). I tested for a genetic isolation by geographic distance pattern among populations in both the cyt b and µsat data sets using the Mantel (1967) test with 10,000 permutations performed with the ade4 package (Dray & Dufour 2007) in R v2.9.0 (R Development Core Team 2009)

Both F-statistics and contingency tests employ the population sample as the unit of comparison, whereas the Bayesian model-based approach in STRUCTURE v2.2.3

(Pritchard et al. 2000; Pritchard & Wen 2004) use the individual as the unit, identifying its likelihood of membership in one or more population group regardless of geographic origin. STRUCTURE was used to assign individuals to population groups ranging from

K = 1 to K = N (total sites), with the relative frequency of their predicted group memberships totaling 1.00. Ten independent runs for each K were evaluated, with burn- ins of 100,000 replicates and run lengths of 500,000 (analyses were performed at http://cbsuapps.tc.cornell.edu). I used the ΔK method of Evanno et al. (2005) to determine the most probable value of K.

Finally, I performed multivariate ordination analyses of µsat allele frequencies to further investigate overall patterns and structure among monkey goby populations.

Multivariate ordination of allele frequency data does not have the assumptions of Hardy-

102 Weinburg or linkage equilibrium, and thus can be a useful tool to explore patterns of genetic variation. A spatial principal components analysis (sPCA; Jombart et al. 2008) was used to investigate spatial patterns of genetic structure among monkey goby populations. This technique is an extension of standard centered PCA that simultaneously maximizes both the variation in allele frequencies among individuals and the spatial autocorrelation of allele frequencies (Jombart et al. 2008). The sPCA was performed using the adegenet package (Jombart 2008) in R.

Morphological analyses

I quantified morphological variation within and among Black and Caspian Sea basin monkey gobies using a suite of meristic and mensural characters. Counts and body measurements were made using a dissecting microscope, and morphometric data (to 0.01 mm) employed digital vernier calipers. Meristic data included numbers of lateral scale rows (posterodorsal tip of opercle to caudal fin base), first dorsal spines, second dorsal and anal soft fin rays, and pectoral fin segmented rays. Measurements were standard length (tip of snout not including lower jaw to midpoint of caudal fin base), caudal peduncle length (posterior end of anal fin base to midpoint of caudal fin base), minimum caudal peduncle depth, caudal peduncle width (at minimum depth), tip of snout to first dorsal fin origin, tip of snout to second dorsal fin origin, tip of snout to anal fin origin, snout-vent length (tip of snout to anus), abdomen length (posterior margin of pelvic fin base to anus), length of first dorsal fin base, length of second dorsal fin base, length of anal fin base, maximum body depth (at anterior margin of first dorsal fin), maximum body width (behind pectoral fin base), pectoral fin length (insertion of longest fin ray to tip), pelvic disc length (insertion to posteriormost point), head length (tip of snout not

103 including jaw to posterodorsal tip of opercle), head width (maximum width at preopercular margin), head depth (maximum depth at posterior dorsal head margin), snout length (tip not including jaw to anterior eye margin), eye diameter (horizontal diameter), interorbital width (least distance between left and right orbits), and preorbital width (between lip and anterior margin of orbit). Body and head measurements are reported as percent of standard and head lengths, respectively.

I performed univariate and multivariate analyses in R to determine if Black and

Caspian Sea basin populations of monkey gobies are morphologically distinguishable.

Means of meristic and morphometric proportions were tested for differences among basins using a Student's t-test. Meristic counts were square-root transformed, and morphometric proportions were arcsine square-root transformed, prior to analysis (Sokal

& Rohlf 1995). Significance of t values was adjusted using standard Bonferroni corrections.

Multivariate analysis of morphological variation between basins was performed using a principal components analysis (PCA) on ln-transformed measurements, which separated morphological variation into linear combinations of variables describing overall body size and shape variation among basins. The first principal component (PC1) primarily described body size variation, whereas the remaining components encompassed body shape variation. I utilized the components that explained 95% of the morphological variance (PC1-5) in further analyses. To examine how monkey gobies from each basin group morphologically, I performed discriminant function analyses (DFA) using the historic subspecies assigned to each basin (i.e., fluviatilis vs. pallasi) as the clustering variable. In the first DFA, I used PC 1-5 to incorporate both size and shape; in the second

104 DFA, I utilized only the body shape components (PC 2-5). I tested the hypotheses that

Black and Caspian Sea basin monkey gobies differ in a) body size and shape, and b) body shape alone, using multivariate analyses of variance (MANOVA), with PC 1-5 as dependent variables in the former, and the latter with the shape components alone (PC 2-

5). Differences in basin mean score for each principal component were assessed with

ANCOVA (allometric components; PC 1) or ANOVA (non-allometric components; PC 2-

5).

Results

Overall genetic diversity

I sampled 355 individuals and identified 107 cyt b haplotypes (GenBank accession numbers GQ444336-GQ444442), with 36 (34%) found in multiple individuals and 71 (66%) singletons. Number of haplotypes per location ranged from 1 (sites A, B,

C, N, O) to 16 (site JJ; Sahil, Azerbaijan), with mean number of haplotypes = 4.47 ± 3.67

(Table 4.1). Haplotype diversity per location ranged from 0.00 (sites A, B, C, N, O) to

1.00 (sites I, K, CC, DD, GG, HH, JJ, KK), with mean haplotype diversity = 0.680 ±

0.34. Nucleotide diversity per location ranged from 0.0000 (sites A, B, C, N, O) to

0.0481 (site BB; Volga River Delta), with mean nucleotide diversity per site = 0.0042 ±

0.0081.

Three µsat loci (Ame129, Ame133, and NG70) were unable to be scored in several populations or showed large proportions of null alleles, and thus were excluded from analysis. I identified 185 alleles from the remaining 10 loci, with the number of alleles per locus ranging from 8 (Nme7) to 31 (NG92). A global test for HWE showed no

105 departure from equilibrium expectations (P = 0.51). Individual HWE tests per locus indicated that two loci (NG92 and Nme2) departed from HWE after Bonferroni correction (P < 0.001), due to allele frequencies in a single introduced population (Vistula

River; Table 4.1). As introduced populations generally do not meet the assumptions of

HWE, I included these loci and this population in the remainder of the analyses.

Observed heterozygosity per sampling site ranged from 0.117 (site C; Hron River,

Slovakia) to 0.833 (site DD; Lenkoran, Azerbaijan), with mean heterozygosity = 0.477 ±

0.18 (Table 4.1).

Phylogeographic patterns

Two main clades were identified among the monkey goby cyt b haplotypes (Fig.

4.2A), corresponding to major hydrogeographic areas within the Ponto-Caspian region.

One clade comprised all haplotypes found in the Black Sea basin; including the Azov

Sea, upstream locations of major rivers flowing into the Black Sea (e.g., Dniester,

Dnieper, and Danube Rivers), and introduced Vistula River/Lagoon populations; as well as those found in the Don, Manych, and Volga Rivers, and inland waterbodies in the

Kumo-Manych Depression (a geological lowland located between the Russian Plain and the northern Caucasus Mountains; see Fig. 4.1). The other major clade contained all haplotypes from the Caspian Sea (locations in Azerbaijan) as well as two haplotypes from the Volga River delta. Both major clades had high overall bootstrap support (100%), but little support or resolution among individual haplotypes or geographic areas. Nucleotide divergences (p-distance) within the two clades were 0.008 for the Black Sea clade and

0.007 for the Caspian Sea, with a between-clade divergence of 0.089 (0.081 net divergence). The two clades were separated by 85 fixed nucleotide differences.

106 Statistical tests of monophyly for both the Black Sea and Caspian Sea clades were highly significant (P < 4.6 x 10-32 for both tests), and a test for their reciprocal monophyly also was highly significant (P = 8.1 x 10-33). Bayesian analysis of the divergence of the Black

Sea and Caspian Sea clades estimated that they separated ~3 million year ago (Mya), and the lineages within each clade shared a most recent common ancestor ~ 0.350 Mya

(Table 4.2).

A similar pattern was observed using nuclear RAG1 sequences (Fig. 4.2B): individuals were grouped into two main clades by basin, with a high ratio of between- to within-clade divergence. Within-clade divergence was low, with individuals in the Black

Sea clade separated by hundreds of kilometers (ALE02 – site D; ALL13 – site T; Table

4.1) having identical sequences. This low level of variation in RAG1 sequences reflects this gene's much lower mutation rate than cyt b (Martin 1999).

Statistical parsimony network analysis of cyt b haplotypes revealed additional geographic structuring. Networks for the Black (Fig. 4.3A) and Caspian (Fig. 4.3B) Sea basin haplotypes were not connected, highlighting the large number of fixed differences between the two groups. Within the Black Sea network, haplotypes from the Don,

Manych, and Volga Rivers and Kumo-Manych Depression generally clustered together and separated from the majority of haplotypes found in the Black Sea and their associated rivers (Dniester, Dnieper, and Danube Rivers) and estuaries.

Population genetic patterns

Pairwise comparisons of genetic divergence among sampling locations from both the cyt b and µsat data sets (Table 4.3) displayed broad genetic structure among locations.

The smallest, non-significant values (FST ≈ 0.00 – 0.20) generally occurred among

107 adjacent or nearby locations (e.g., introduced locations A and B in the Vistula River system; sites N – BB in the Volga and Don River basins and waterbodies in the Kumo-

Manych Depression; sites CC – KK along the Caspian Sea coastline of Azerbaijan; see

Table 4.3B). The larger, statistically significant values (FST ≈ 0.20 – 1.00) generally occurred among geographically disparate sites. Pairwise contingency tests among samples (not shown) showed a similar pattern to the pairwise estimates of FST. Both cyt b and μsat data sets for the Black Sea clade population samples corresponded to a genetic isolation by geographic distance pattern (cyt b – rM = 0.26, P < 0.0001, Fig. 4.4A; μsat – rM = 0.54, P < 0.0001, Fig. 4.4B) whereas those from the Caspian Sea clade did not fit the pattern (cyt b – rM = -0.30, P = 0.92; μsat – rM = -0.17, P = 0.79; not shown).

For the Bayesian analysis of population clustering using STRUCTURE, evaluation of ΔK values across the range of K identified K = 4 as encompassing the majority of the genetic structure among individuals (Fig. 4.5). The first cluster contained the two introduced population samples in the Vistula River watershed. The second cluster comprised the primary group for individuals from the northern Black Sea and its associated estuaries and river basins (i.e., Danube, Dniester, and Dnieper Rivers). The third cluster housed individuals from the Don, Volga, and Manych Rivers, as well as those from inland waterbodies in the Kumo-Manych Depression. The fourth comprised individuals from the Caspian Sea (Azerbaijan locations). Individuals from the Volga

River delta generally had equal probabilities of group membership in clusters 3 and 4, reflecting their location at the interface of two hydrographic systems.

The spatial principal components analysis identified an overall global spatial structure in microsatellite allele frequencies (max(t) = 0.049, P < 0.0001, 10000

108 replications; Jombart et al. 2008), and three eigenvalues/eigenvectors with a high degree of spatial autocorrelation were retained (Fig. 4.6A-C). Scores on the first global axis separated individuals from the Caspian Sea (Azerbaijan) and lower Volga River from those in the Black Sea basin, upper Volga and Don River region, and inland waterbodies in the Kumo-Manych Depression (Fig. 4.6A). The second global axis primarily separated individuals from the Volga/Don River basins and Kumo-Manych Depression from those in the Black and Caspian Sea basins (Fig. 4.6B). The third global axis (Fig.

4.6C) generally distinguished individuals from the introduced Vistula River/Lagoon populations from those in the Black Sea basin and associated watershed (i.e., Danube

River drainage).

Introduced locations in the Danube and Vistula River basins

Introduced populations of monkey goby in the upper Danube/Hron River and

Vistula River basin showed strong genetic affinities to populations in the northwestern

Black Sea. The Vistula River and Vistula Lagoon samples were monomorphic for a cyt b haplotype (1; Fig. 4.2 & Table 4.1) that was common across the northwest Black Sea

(Dniester River, site G; Tyligul Estuary, site J; Table 4.1) and is found in high frequency in the Dnieper River (site H; Table 4.1, Fig. 4.1). Values of FST between the two Vistula basin location samples and the Dnieper River sample were not significant (Table 4.3A; sites A, B, and H – below diagonal). The Hron River sample was monomorphic for a haplotype (5; Fig. 4.2, Table 4.1) that also was found in the lower Danube River (site D;

Table 4.1, Fig. 4.1), and these two samples did not significantly differ (Table 4.3A).

Bayesian STRUCTURE analyses (Fig. 4.3) showed a similar pattern to the mtDNA data.

Individuals from introduced localities in the Vistula River basin had their primary

109 membership probability in a cluster distinct from the majority of Black Sea basin samples

(light purple; Fig. 4.5); however, most individuals from the Dnieper River also had a moderate probability of membership in that cluster (light purple). Individuals from the

Hron River were grouped in the cluster containing the majority of individuals from native

Black Sea locations (sites D-J; dark orange; Fig. 4.5). Introduced locations in the Vistula

River system and the Hron River also were allied with the native Black Sea locations in the spatial principal components analysis, with individuals from introduced localities having similar scores as those from native locations on the first two spatial principal components (Fig. 4.6A-B), and the Vistula River and Lagoon individuals separated only on the third (and smallest) significant spatial principal component (Fig. 4.6C).

Morphological variation between basins

Black and Caspian Sea basin populations of monkey gobies broadly overlapped, yet exhibited some significant meristic and morphological character differences (Table

4.4), including: caudal peduncle length, length of anal fin base, body depth, head length, head width, head depth, interorbital width, preorbital width, lateral scale rows, and pectoral fin rays. Principal components (PC) analysis further highlighted differences in body shape between fish from the two basins, with the first five PC explaining 95.8% of the morphological variance (Table 4.5). PC 1 was highly correlated with standard length

(r = -0.968; P<<0.001), with approximately equal loadings on all morphometric variables and thus described size-associated variation. PC 2-5 showed low correlation with standard length (|r| < 0.121; P > 0.141) and represented size-independent body shape variation. The strongest influences on PC 2-5 were measures of head shape (interorbital width) and posterior body shape (caudal peduncle width, body width, length of second

110 dorsal fin base, pelvic fin length; Table 4.4). The first DFA using body size and shape

(PC 1-5) was highly significant (F1, 139 = 58.7, P << 0.001); individuals were classified to the correct taxon 78% of the time. The second DFA using only body shape (PC 2-5) was also highly significant (F1, 139 = 58.5, P << 0.001), and correctly classified 75% of individuals to taxon. Explicit tests of body size and shape differences between Black and

Caspian Sea monkey gobies were performed with multivariate analyses of variance

(MANOVA). The first MANOVA using PC 1-5 detected a highly significant difference

(PC 1-5, Wilks' λ = 0.703, F5, 135 = 11.407, P << 0.001); MANOVA using PC 2-5 also detected highly significant differences in PC scores between basins (PC 2-5, Wilks' λ =

0.704, F4, 136 = 14.303, P << 0.001). Visual inspection of PC scores showed that individuals from the two basins overlapped, but differed in mean component scores (i.e., different centroids; Fig. 4.7). Univariate ANCOVA for PC 1 with standard length as covariate found no difference between fish from the two basins; however, univariate

ANOVAs for the remaining components showed significant differences in their mean scores for PC 2, 3, and 5 (P < 0.03 in all tests).

Discussion

Divergence of Black and Caspian Sea lineages

The present study clearly distinguishes two distinct evolutionary lineages of monkey goby: one in the Black Sea watershed (including the Dnieper, Dniester, and

Danube Rivers) and the Volga-Don-Manych Rivers (including inland waterbodies and canals bounded by these rivers), and the second in the Caspian Sea and Volga River delta.

These two lineages are deeply divergent, with 85 fixed mtDNA cyt b gene substitutions,

111 fixed differences in nuclear RAG1 gene sequences, pronounced differences in microsatellite allelic composition and frequencies, and differences in morphology. These taxa were previously recognized as separate subspecies of monkey goby: N. f. fluviatilis in the Black and Azov Seas, and N. f. pallasi comprising Caspian Sea populations (Berg

1916; Pinchuk et al. 2003). This study supports their taxonomic recognition and, given the pronounced magnitude of their separation, indicates that they should be recognized as distinct species: N. fluviatilis and N. pallasi. Their pronounced separation using multiple types of data (nuclear and mtDNA gene sequences, microsatellite allele frequencies, and morphology) demonstrates that they are independently-evolving lineages and fit commonly accepted species definitions, including the evolutionary species concept

(Wiley & Mayden 2000) and the metapopulation lineage concept (de Queiroz 2005).

Nucleotide divergence between the two monkey goby species is within the range separating species of tubenose gobies Proterorhinus (0.04-0.173; Chapter 2) and other intrageneric comparisons of fishes (0.01-0.40; Johns & Avise 1998). Intraspecific divergences within the two monkey goby species are similar to those found within other gobiid species (0.001-0.021; Harada et al. 2002; Brown & Stepien 2008; Chapters 2-3).

In addition, the ratio of interclade/intraclade variation is similar to the operational criterion of divergence between clades greater or equal to 10x the level of divergence within clades, recommended by Hebert et al. (2004b) to recognize separate species.

Divergence time separating the Black and Caspian Sea species of monkey goby is older than most estimates for other sympatric gobiids. The mean separation and 95% highest posterior density (HPD) of divergence time estimates between the monkey goby species dates within the Pliocene Epoch (~ 5-2 Mya; Reid & Orlova 2002). In Chapter 2,

112 I estimated a divergence date of ~4.3 Mya between marine and freshwater species of

Proterorhinus, and a ~1.0 Mya split between P. semilunaris and P. cf semipellucidus

(freshwater species in the Black and Caspian Sea basins, respectively). Separation time between Black and Caspian Sea lineages of round goby N. melanostomus was estimated at 350,000 years, with intralineage coalescence dating to ~100,000 years by Brown &

Stepien (2008). Species groups of Ponticola gobies spanning the Black and Caspian Sea basins diverged between 1.8-1.3 Mya, whereas the tadpole gobies Benthophilus spp. radiated approximately 2.2 Mya (Chapter 3).

These separation times generally correspond to major geologic and hydrologic fluctuations in the Ponto-Caspian region. The historic Black and Caspian Sea basins have been intermittently separated and connected over the past 5 My (Reid & Orlova

2002), promoting lineage isolation and divergence of faunas during periods of separation, and geographic spread of lineages/faunal exchanges between basins during reunification.

Cristescu et al. (2003, 2004) found a range of divergence times in several Ponto-Caspian crustaceans, with planktonic species showing a late Miocene separation of lineages (~6-8

Mya) and benthic species showing a much more recent Pleistocene separation (~1 Mya).

Cristescu & Hebert (2002) also identified a late Pleistocene radiation in several lineages of Ponto-Caspian onchyopod cladocerans. Estimates of divergences among Black and

Caspian Sea lineages of zebra mussel Dreissena polymorpha ranged from ~530,000 –

160,000 years ago (Stepien et al. 2002; Gelembiuk et al. 2006).

Introduced populations in the Hron River and Vistula River basin

The introductions of N. fluviatilis within the Hron River and the Vistula River system likely originated from the northwestern Black Sea, with the lower Danube River

113 as a source for the Hron River and the Dnieper River as the source for the Vistula River system. Upstream expansion of the monkey goby in the Danube River system has been chronicled over the last ~40 years (Bănărescu 1970; Bíró 1971; Ahnelt et al. 1998;

Kautman 2001; Holčík et al. 2003; Jurajda et al. 2005). Although some range extension is thought to have occurred through natural spread, long upstream dispersal of N. fluviatilis (along with other Ponto-Caspian gobiids) in the Danube River and its tributaries likely resulted from ballast water ship transport (Ahnelt et al. 1998). A similar dispersal scenario likely underlies the introduction into the Vistula River system.

Danilkiewicz (1998) suggested two possible alternatives to explain the presence of N. fluviatilis in the Baltic Sea watershed: range extension via transport from the Dnieper or

Dniester River into the Western Bug River through the Dnieper-Bug Canal, or colonization via postglacial lakes that once linked the headwaters of Black and Baltic Sea basin tributaries. Results from both the cyt b and µsat data support the hypothesis of upstream transport of gobies via shipping vessels from the lower Dnieper River into the

Western Bug River through the Dnieper-Bug Canal. The canal was also identified as the likely introduction pathway for the Baltic Sea basin establishments of the round goby N. melanostomus (Brown & Stepien 2008), the racer goby Babka gymnotrachelus (Ohayon

& Stepien 2007), the mysid shrimp Hemimysis anomala (Audzijonyte et al. 2008), and as a general corridor of invasion from the Black Sea basin to the Baltic Sea basin for multiple macroinvertebrate species (Bij de Vaate et al. 2002).

Morphological differences between monkey goby lineages

Morphological separation between Black and Caspian Sea species of monkey goby is not as pronounced as their genetic separation. However, despite overlap in many

114 morphological and meristic characters, the two lineages differ in mean values for a suite of characters and their overall body shapes. Caudal peduncle length, anal fin base length, body depth, head length, head width, head depth, interorbital width, preorbital width, number of scales in lateral series, and pectoral fin rays significantly differ between the two species (Table 4.3). Multivariate analyses also detected significant differences in overall body shape among individuals from the two basins. Notably, significant discriminant function analyses, MANOVAs, and univariate ANOVA for PC 2, 3, and 5 show that monkey gobies from the two different basins reside in two statistically distinct areas of morphospace. In Chapter 2, I found a similar pattern of morphological and genetic divergence among marine and Black and Caspian Sea freshwater lineages of tubenose goby Proterorhinus spp., with pronounced genetic divergence among lineages accompanied by subtle yet distinct differences in morphology. Although habitat-specific

(e.g., freshwater vs. marine) differences in morphology have been observed in a variety of species (Corti et al. 1996; Walker & Bell 2000), geography appears to be the driving influence of morphological differences between the two monkey goby species. In addition, pronounced genetic divergence is evident within many morphologically

“cryptic” species of fishes (Egge & Simons 2006; Quattro et al. 2006). This study highlights the role of genetic measures, coupled with morphology and other data, as a powerful tool in the description of cryptic biodiversity (Hebert et al. 2004a).

Conclusions

Our study identified a long-term separation of lineages of monkey goby in the

Black and Caspian Sea basins based on pronounced divergences in nuclear and mtDNA sequences, microsatellite allele frequencies, and morphology. These lineages fit widely

115 accepted species definitions, and thus are here recognized as two distinct species. The

Black Sea species, N. fluviatilis, has expanded beyond its historic natural distribution over the past 40 years, primarily due to anthropogenic influences. Two distinct pathways of expansion are identified: the Danube River and the Dnieper-Pripyat-Bug River systems, both of which are major shipping routes and recognized vectors for other introduced species. These results highlight the role of geographic connectivity and isolation in the evolution of Ponto-Caspian fauna, and the influence of recent human- mediated disturbances and changes in shaping present-day population genetic patterns.

116 Table 4.1 Sampling locations of monkey goby populations, with latitude and longitude of collection site, sample size (N), number of

cytochrome b haplotypes (NH), haplotype (h) and nucleotide (π) diversity, number of microsatellite alleles (NA), heterozygosity

measures (HO, observed; HE, expected), FIS, and mean allelic richness per locus (R).

Cytochrome b Microsatellite

Site Designation Body of Water Location Latitude Longitude N NH h π N NA HO HE FIS R Neogobius fluviatilis A Vistula River Torun, Poland 53.006497 18.611708 31 1 0.000 0.000 33 29 0.219 0.275 0.202 1.27 B Vistula Lagoon Tolkmicko, Poland 54.327894 19.526324 22 1 0.000 0.000 24 21 0.248 0.243 -0.021 1.25 C Hron River Slovakia 47.818608 18.740878 13 1 0.000 0.000 20 18 0.117 0.143 0.184 1.15 D Danube River Vilkove, Ukraine 45.391914 29.588268 15 6 0.848 0.002 15 52 0.496 0.540 0.035 1.52 E Dniester Dam Reservoir Makarivka, Ukraine 48.566667 26.750000 5 3 0.800 0.001 12 24 0.207 0.233 0.113 1.22 1

1 F Dniester River delta Bilyayivka, Ukraine 46.468333 30.216667 16 7 0.775 0.005 16 50 0.373 0.457 0.186 1.44 7 G Dniester River Bilgorod, Ukraine 46.229242 30.362880 14 11 0.967 0.006 14 63 0.522 0.581 0.101 1.58 H Dnieper River Dnipryany, Ukraine 46.733333 33.266667 15 3 0.362 0.0003 15 46 0.476 0.478 0.004 1.48 I Khadzhibey Estuary Ukraine 46.650683 30.533750 3 3 1.000 0.0082 4 27 0.432 0.509 0.150 1.47 J Tyligul Estuary Ukraine 46.690000 31.203450 3 2 0.667 0.001 4 32 0.500 0.647 0.227 1.63 K Azov Sea Molochnyi, Ukraine 46.655616 35.278634 4 4 1.000 0.005 4 32 0.474 0.559 0.153 1.55 L Manych River Veselovsky Reservoir, Russia 47.114634 40.791515 - - - - 4 25 0.361 0.463 0.220 1.40 M Rioni River Georgia 42.135000 41.701111 3 2 0.667 0.0006 ------N Ozero Manych Prujitnoje, Russia 46.016147 43.448435 7 1 0.000 0.0000 7 31 0.369 0.439 0.159 1.44 O Tsimlyanska Reservoir Kalach-na-Donu, Russia 48.674471 43.515149 9 1 0.000 0.0000 10 36 0.430 0.444 0.033 1.44 P Volga-Don Canal Iliovka, Russia 48.643269 43.617069 8 3 0.464 0.0004 8 32 0.363 0.384 0.056 1.38 Q Chagraiskoye Reservoir Zunda Tolga, Russia 45.617691 44.211077 5 3 0.700 0.0051 5 36 0.490 0.620 0.210 1.59 R Lake Sarpa Tsagan-Nur, Russia 47.370632 45.208318 15 5 0.476 0.0014 15 43 0.467 0.506 0.078 1.51 S Chernozemel'skii 120 km east of Elista, Russia 46.272008 45.615373 7 5 0.905 0.0016 7 36 0.362 0.502 0.278 1.49 Connector T Volga River Vinovka, Russia 48.870870 44.660139 6 2 0.600 0.0006 6 32 0.400 0.498 0.197 1.49 U Volga River Svetli Yar, Russia 48.484638 44.784676 5 4 0.900 0.0016 5 25 0.460 0.403 -0.143 1.41 V Volga River Stupyno, Russia 48.310313 45.797317 5 3 0.800 0.0010 5 23 0.300 0.310 0.032 1.31 W Volga River Preshib, Russia 47.683923 46.509057 5 2 0.400 0.0003 5 23 0.227 0.304 0.252 1.32 X Volga River Vladimirovka, Russia 47.171956 47.053889 4 2 0.500 0.0004 4 29 0.410 0.474 0.135 1.48 Y Volga River Zam'yaniy, Russia 46.829650 47.600639 6 3 0.600 0.0009 6 29 0.383 0.423 0.095 1.42 Z Volga River Narimanov, Russia 46.601411 47.883446 4 3 0.833 0.0015 4 27 0.410 0.387 -0.061 1.40 AA Volga River Damchik, Russia 45.788350 47.886953 6 3 0.733 0.0010 7 45 0.641 0.672 0.054 1.68 BB Volga River Delta, Branch Obzhorova 46.302213 48.977384 6 3 0.600 0.0481 6 41 0.586 0.686 0.146 1.68 N. pallasi CC Caspian Sea Nabran, Azerbaijan 41.837222 48.620000 4 4 1.000 0.0066 5 51 0.740 0.775 0.045 1.77 DD Caspian Sea Lenkoran, Azerbaijan 38.751944 48.868889 2 2 1.000 0.0088 3 43 0.833 0.817 -0.020 1.82 EE Caspian Sea Niyazabad, Azerbaijan 41.537950 48.924108 8 7 0.964 0.0060 14 76 0.659 0.720 0.084 1.72 FF Caspian Sea Zarat, Azerbaijan 40.888889 49.368889 6 5 0.933 0.0049 6 60 0.683 0.827 0.173 1.81 1

1 GG Caspian Sea Alet, Azerbaijan 39.940000 49.409167 9 9 1.000 0.0038 9 71 0.713 0.759 0.062 1.76 8 HH Caspian Sea Sahil, Azerbaijan 40.217222 49.569167 16 16 1.000 0.0076 19 97 0.676 0.749 0.098 1.75 II Caspian Sea Sumqayit, Azerbaijan 40.600278 49.682222 14 13 0.989 0.0062 15 89 0.716 0.799 0.103 1.80 JJ Caspian Sea Shikh, Azerbaijan 40.304167 49.805000 5 5 1.000 0.0041 5 48 0.761 0.747 -0.018 1.77 KK Caspian Sea Bilgah, Azerbaijan 40.576667 50.030833 13 13 1.000 0.0073 14 75 0.659 0.743 0.113 1.75 Table 4.2 Bayesian coalescent estimates of clade age for monkey goby clades from Fig.

2. Mean age estimates to the most recent common ancestor of each clade are based on a cyt b substitution rate in gobies of 2.05% per million years (Rocha et al. 2005). The 95% highest posterior density (HPD) range is the range of the age parameter space containing

95% of the sampled values. Mya – million years ago.

Clade Mean age (Mya) 95% HPD Black (Neogobius fluviatilis) + Caspian (N. pallasi) Sea clades 3.085 1.806 – 4.365 N. fluviatilis 0.348 0.215 – 0.503 N. pallasi 0.364 0.227 – 0.521

119 Table 4.3 Pairwise FST comparisons among populations of (A) N. fluviatilis and (B) N. pallasi based on cyt b (below diagonal) and

microsatellite (above diagonal) data. N/A = comparison not available due to missing data. * = P < 0.05; ** = significant following

sequential Bonferroni correction (Rice 1989).

A.

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z AA BB

A -- 0.11** 0.74** 0.49** 0.70** 0.58** 0.44** 0.40** 0.48** 0.46** 0.54** 0.57** N/A 0.57** 0.56** 0.58** 0.54** 0.44** 0.52** 0.57** 0.60** 0.64** 0.63** 0.62** 0.62** 0.64** 0.51** 0.57**

B 0.00 -- 0.76** 0.51** 0.72** 0.62** 0.38** 0.36** 0.44* 0.50** 0.60** 0.57** N/A 0.63** 0.58** 0.62** 0.58** 0.43** 0.57** 0.62** 0.67** 0.71** 0.66** 0.69* 0.65** 0.70** 0.54** 0.60**

C 1.00** 1.00** -- 0.51** 0.77** 0.67** 0.40** 0.46** 0.52** 0.65** 0.73** 0.69** N/A 0.76** 0.66** 0.70** 0.71** 0.53** 0.71** 0.73** 0.80** 0.83** 0.78** 0.81** 0.74** 0.82** 0.62** 0.68**

D 0.89** 0.86** 0.32* -- 0.63** 0.22** 0.25** 0.30** 0.24* 0.16* 0.17** 0.37* N/A 0.30** 0.41** 0.41** 0.25** 0.29** 0.23** 0.27** 0.33** 0.37** 0.46** 0.30* 0.45** 0.31* 0.27** 0.31**

1 E 0.96** 0.95** 0.90** 0.49** -- 0.53** 0.41** 0.51** 0.57* 0.59* 0.60* 0.54* N/A 0.65** 0.57** 0.60** 0.53* 0.40** 0.60** 0.61** 0.69** 0.75** 0.69* 0.70* 0.65** 0.71* 0.46** 0.51** 2

0 F 0.72** 0.67** 0.58** 0.43** 0.26* -- 0.36** 0.45** 0.42** 0.30** 0.32** 0.47** N/A 0.36** 0.47** 0.48** 0.26** 0.37** 0.32** 0.35** 0.39** 0.47** 0.52** 0.38** 0.52** 0.42* 0.33** 0.34**

G 0.63** 0.56** 0.54** 0.38** 0.32** 0.02 -- 0.13** 0.03 0.17* 0.32* 0.23* N/A 0.41** 0.35** 0.35** 0.32** 0.18** 0.33** 0.37** 0.44* 0.48* 0.34** 0.44* 0.37** 0.45* 0.24** 0.33**

H 0.12* 0.08 0.98** 0.81** 0.91** 0.60** 0.48** -- 0.12* 0.22* 0.35* 0.13* N/A 0.42** 0.29** 0.34** 0.32** 0.17** 0.32** 0.38** 0.45** 0.48** 0.35** 0.46** 0.35** 0.47* 0.28** 0.37**

I 0.87** 0.83** 0.74* 0.46* 0.54* 0.00 0.00 0.74** -- 0.11 0.32* 0.26* N/A 0.40* 0.33* 0.37* 0.30* 0.17* 0.32* 0.36* 0.46* 0.52* 0.41* 0.47* 0.40* 0.48* 0.23* 0.29*

J 0.92* 0.89* 0.90** 0.71** 0.81* 0.26 0.23* 0.81* 0.09 -- 0.18* 0.26* N/A 0.24* 0.25* 0.26* 0.14* 0.15* 0.16* 0.19* 0.30* 0.35* 0.32* 0.27* 0.30* 0.30* 0.14* 0.18*

K 0.94** 0.92* 0.89* 0.72* 0.81* 0.49* 0.41* 0.85* 0.40* 0.55* -- 0.32 N/A 0.18 0.33 0.36 0.13 0.24 0.10 0.18 0.25 0.30 0.44 0.21 0.39 0.23 0.19 0.22

L N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A -- N/A 0.37* 0.14* 0.21* 0.22* 0.07* 0.27* 0.30* 0.40* 0.46* 0.23* 0.39* 0.22* 0.42* 0.17* 0.27*

M 1.00* 0.99* 0.99* 0.80* 0.91* 0.57* 0.50* 0.96* 0.55* 0.72* 0.48 N/A -- N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

N 1.00** 1.00** 1.00** 0.85* 0.94** 0.66** 0.61** 0.97** 0.74* 0.85* 0.38* N/A 0.98* -- 0.22** 0.22* 0.10* 0.18** 0.05* 0.08* 0.10* 0.04* 0.26* 0.11* 0.26* 0.05* 0.18* 0.22**

O 1.00** 1.00** 1.00** 0.86** 0.95** 0.69** 0.65** 0.98** 0.79* 0.89* 0.47* N/A 0.98* 0.00 -- 0.05* 0.20* 0.07* 0.23** 0.17* 0.29* 0.33* 0.03* 0.29* 0.07* 0.30* 0.15** 0.25**

P 0.99** 0.99** 0.98** 0.84** 0.93** 0.66** 0.61** 0.96** 0.74* 0.84* 0.36* N/A 0.93* 0.00 0.03 -- 0.21* 0.07* 0.26* 0.22* 0.24* 0.32* 0.00 0.28* 0.00 0.27* 0.14* 0.26*

Q 0.92** 0.89** 0.87** 0.71** 0.79** 0.49* 0.42* 0.83** 0.41* 0.54* 0.00 N/A 0.31 0.50* 0.57* 0.49* -- 0.13* 0.05 0.08 0.12* 0.21* 0.25* 0.13* 0.23* 0.17* 0.11* 0.16*

R 0.95** 0.93** 0.93** 0.82** 0.88** 0.67** 0.62** 0.90** 0.73* 0.80* 0.33* N/A 0.82* 0.00 0.00 0.00 0.47* -- 0.13* 0.13* 0.21** 0.25* 0.04 0.21* 0.07* 0.22* 0.09* 0.20**

S 0.97** 0.96** 0.95** 0.81** 0.89** 0.64** 0.59** 0.93** 0.66* 0.77* 0.22 N/A 0.83* 0.10 0.16* 0.03 0.39* 0.04 -- 0.06 0.15* 0.09* 0.30* 0.10* 0.28* 0.11* 0.18* 0.23*

T 0.99** 0.99** 0.99** 0.83** 0.92** 0.65** 0.59** 0.96** 0.69* 0.81* 0.33* N/A 0.92* 0.43 0.51* 0.12 0.45* 0.06 0.00 -- 0.13* 0.15* 0.25* 0.01 0.26* 0.07 0.16* 0.23*

U 0.98** 0.97** 0.96* 0.81** 0.90* 0.62** 0.56** 0.94** 0.63* 0.75* 0.19 N/A 0.84* 0.18* 0.26* 0.05 0.36 0.04 0.00 0.00 -- 0.10* 0.30* 0.05 0.27* 0.03 0.16* 0.27*

V 0.99** 0.98** 0.98** 0.82** 0.91* 0.63** 0.57** 0.95** 0.65* 0.77* 0.27* N/A 0.88* 0.33* 0.42* 0.15 0.40* 0.07 0.03 0.01 0.00 -- 0.38* 0.11* 0.36* 0.00 0.25* 0.31* W 1.00** 0.99** 0.99** 0.83** 0.92** 0.63** 0.57** 0.96** 0.67* 0.80* 0.27* N/A 0.94* 0.07 0.15 0.00 0.40 0.00 0.04 0.28 0.02 0.05 -- 0.36* 0.00 0.34* 0.13* 0.27*

X 1.00** 0.99** 0.99* 0.82** 0.92* 0.61* 0.55* 0.96** 0.63* 0.77* 0.21 N/A 0.93* 0.15 0.25 0.00 0.36 0.00 0.00 0.00 0.00 0.00 0.01 -- 0.33* 0.00 0.17* 0.25*

Y 0.99** 0.98** 0.98** 0.82** 0.91** 0.64** 0.58** 0.95** 0.67* 0.79* 0.24* N/A 0.89* 0.03 0.09 0.00 0.40* 0.00 0.00 0.04 0.00 0.06 0.00 0.00 -- 0.31* 0.15* 0.26*

Z 0.98** 0.98* 0.97* 0.80** 0.90* 0.60* 0.54* 0.94** 0.59* 0.73* 0.13 N/A 0.83* 0.15 0.25 0.10 0.31 0.02 0.02 0.19 0.00 0.13 0.04 0.00 0.00 -- 0.18* 0.26*

AA 0.98** 0.98** 0.97** 0.82** 0.91** 0.63** 0.58** 0.94** 0.67* 0.78* 0.28* N/A 0.88* 0.13 0.20* 0.00 0.41* 0.00 0.00 0.00 0.00 0.00 0.06 0.00 0.00 0.06 -- 0.03

BB 0.70** 0.62** 0.53** 0.51** 0.49** 0.45** 0.41** 0.53* 0.20* 0.23 0.09 N/A 0.14 0.22 0.29 0.24 0.16 0.36 0.21 0.19 0.14 0.15 0.14 0.10 0.18 0.09 0.18 -- 1 2 1 B.

CC DD EE FF GG HH II JJ KK CC -- 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 DD 0.02 -- 0.01 0.00 0.02 0.05 0.00 0.01 0.06* EE 0.12 0.00 -- 0.03 0.03* 0.04* 0.01 0.02 0.03* FF 0.18 0.00 0.03 -- 0.00 0.01 0.00 0.02 0.00 GG 0.00 0.00 0.04 0.14 -- 0.02 0.01 0.05* 0.02 HH 0.01 0.00 0.05 0.17* 0.00 -- 0.04* 0.04* 0.00 II 0.07 0.00 0.00 0.10 0.00 0.01 -- 0.00 0.02* JJ 0.13 0.00 0.03 0.00 0.16* 0.15* 0.08 -- 0.02 KK 0.05 0.00 0.00 0.04 0.01 0.02 0.00 0.03 --

122 Table 4.4 Morphometrics and meristics of Black (Neogobius fluviatilis) and Caspian (N. pallasi) Sea basin populations of monkey gobies. Body and head measurements given as percentage of standard or head length, respectively. * indicates significant difference between taxa using a Student's t-test; ** indicates significance following standard

Bonferroni correction for multiple tests within each group (% standard length, % head length, meristics).

N. fluviatilis N. pallasi (N = 62) (N = 79) Measurement Mean ± SD Range Mean ± SD Range Standard length 84.5 ± 30.9 33.2 - 142.8 80.2 ± 17.3 37.7 - 121 % Standard length Caudal peduncle length 16.3 ± 1.3 13.4 - 20.9 14.9 ± 1.3 11.9 - 17.6 ** Caudal peduncle depth 7.9 ± 0.8 6.2 - 9.7 7.8 ± 0.7 6.6 - 9.6 Caudal peduncle width 4.7 ± 0.9 2.7 - 6.7 4.6 ± 0.8 2.7 - 7 Snout to D1 origin 34.7 ± 2.3 22.4 - 39.4 34.4 ± 1.9 22.3 - 37.6 Snout to D2 origin 50.4 ± 2.4 37.6 - 56.6 50.6 ± 2.1 38 - 55.4 Snout to A1 origin 54.6 ± 3.2 36.5 - 58.8 55.5 ± 3 39.4 - 60.7 Snout-vent length 49.4 ± 4.1 23.4 - 56 50.4 ± 4.1 18 - 56.3 Snout to P2 origin 31.3 ± 1.7 27.6 - 35.3 31.9 ± 2.1 28.6 - 37.3 Abdomen length 19.9 ± 2 16 - 25 20.6 ± 1.9 17.1 - 24.9 * Length of D1 base 16.5 ± 1.3 13.5 - 19.1 17.4 ± 3 14.7 - 34.9 * Length of D2 base 34.2 ± 4.5 6.8 - 39.5 35.4 ± 2.8 32.2 - 51.9 Length of A1 base 30.6 ± 1.9 25.9 - 34 31.5 ± 1.5 27.7 - 36.5 ** Body depth 17.8 ± 2.3 14.4 - 22.8 19.2 ± 2.3 14.8 - 27.4 ** Body width 11.2 ± 1.9 7.3 - 15.8 11.1 ± 1.8 3.6 - 15 Pectoral fin length 24.2 ± 1.6 19.9 - 27.9 24.1 ± 1.4 21.4 - 28.3 Pelvic disc length 20.8 ± 3 2.8 - 25.8 21 ± 1.6 16.8 - 25 Head length 29.4 ± 2.1 25.5 - 34.6 27.9 ± 1.6 24.5 - 31.3 ** % Head length Head width 62.4 ± 11 45.6 - 94.5 73.9 ± 10.7 52.5 - 97.8 ** Head depth 56 ± 6.8 45.7 - 75.5 63.2 ± 9 14.6 - 80.1 ** Snout length 27.2 ± 2.8 20 - 32.5 27.3 ± 2.4 20.6 - 35.4 Eye diameter 20.7 ± 2.6 15.7 - 26.9 21.2 ± 2.2 16.2 - 28.4 Interorbtial width 9.4 ± 3.1 4 - 19.1 11.4 ± 3.1 3.9 - 17 **

123 Preorbital width 14.6 ± 4.2 6.2 - 36.7 16.2 ± 2 10.8 - 20.3 **

Meristic Mode Range Mode Range Scales in lateral series 60 54 - 70 55 47 - 61 ** First dorsal fin spines 6 5 - 7 6 5 - 6 Second dorsal fin rays 16 14 - 16 16 14 - 17 Anal fin rays 14 13 - 15 14 13 - 15 Pectoral fin rays 17 15 - 18 17 15 - 19 **

124 Table 4.5 Summary of principal components analysis (eigenvalues and loadings) calculated from 24 linear measurements from Black and Caspian Sea species of monkey goby. % eigenvalue indicates the total morphological variance explained by each principal component (PC). Loadings with absolute values greater than 0.3 shown in bold.

PC 1 PC 2 PC 3 PC 4 PC 5 Eigenvalue 1.68 0.24 0.20 0.18 0.16 % Eigenvalue 90.7 1.9 1.2 1.1 0.8 Standard Length -0.183 0.126 -0.017 0.018 -0.072 Caudal peduncle length -0.198 0.205 -0.219 -0.051 -0.086 Caudal peduncle depth -0.216 -0.002 -0.147 0.016 0.001 Caudal peduncle width -0.233 -0.198 -0.594 -0.008 0.201 Snout to first dorsal fin origin -0.170 0.141 0.028 -0.020 -0.162 Snout to second dorsal fin origin -0.180 0.122 0.029 0.017 -0.101 Snout to anal fin origin -0.185 0.126 0.053 0.036 -0.117 Snout-vent length -0.186 0.121 0.061 0.041 -0.258 Snout to pelvic disc origin -0.191 0.066 0.039 -0.020 -0.067 Abdomen length -0.194 0.129 0.077 0.079 -0.179 First dorsal fin length -0.192 0.178 0.097 0.113 0.040 Second dorsal fin length -0.184 0.192 0.123 0.313 0.754 Anal fin length -0.191 0.078 -0.047 0.061 -0.007 Body depth -0.221 -0.158 0.005 -0.054 -0.020 Body width -0.231 -0.133 -0.525 -0.054 0.092 Pectoral fin length -0.185 0.109 0.022 -0.008 -0.054 Pelvic fin length -0.160 0.097 0.228 -0.875 0.302 Head length -0.169 0.149 -0.009 -0.020 -0.136 Head width -0.238 -0.172 0.059 -0.114 0.055 Head depth -0.222 -0.187 0.112 -0.030 -0.101 Snout length -0.200 0.159 0.001 -0.037 -0.152 Eye diameter -0.120 0.065 -0.024 -0.040 -0.190 Interorbital width -0.317 -0.726 0.346 0.121 -0.036 Preorbital width -0.252 0.186 0.265 0.261 0.168

125 Fig. 4.1 Current native (stippled) and invasive distribution (shaded) of monkey gobies in the Ponto-Caspian region and

central/northern Europe, respectively, with sampling sites (circles – Neogobius fluviatilis; squares – N. pallasi) for the present study. 1 2 6 Fig. 4.2 Maximum likelihood phylogenetic analyses of monkey goby. A) Cytochrome b;

B) RAG1. Numbers above nodes indicate bootstrap support (% of 2000 pseudoreplications). Trees are rooted using two congeneric species (round goby

Neogobius melanostomus and Caspian goby N. caspius).

A.

127 B.

128 Fig. 4.3 Statistical parsimony networks of monkey goby cyt b haplotypes from the A)

Black (= Neogobius fluviatilis) and B) Caspian Sea (= N. pallasi) basins. Circles are sized according to total observed frequencies. Lines indicated a single mutational step between haplotypes; small, unlabelled circles represent hypothesized unsampled haplotypes. Haplotypes are colored region as in Fig. 4.5.

A.

129 B.

130 Fig. 4.4 Isolation by geographic distance among monkey goby populations. Populations are separated according to clades from Fig. 4.2: ○ = Black Sea clade (Neogobius fluviatilis) populations; Mantel tests for Caspian Sea clade (N. pallasi) populations not significant and not shown. A) Cytochrome b: y = 10.46x – 42.46, R2 = 0.043, P < 0.001;

B) Microsatellites: y = 0.24x – 0.75, R2 = 0.225, P < 0.001.

A.

131 B.

132 Fig. 4.5 Bayesian STRUCTURE version 2.2.3 (Pritchard et al. 2000; Pritchard & Wen 2004) analysis of monkey goby populations

using 10 microsatellite loci. K = 4 was identified as the optimal value of K using the ΔK method of Evanno et al. (2005). Each

individual is represented as a vertical line partitioned into K colored segments representing the individual’s estimated group

membership fractions. Black lines separate different sampling sites, which are labeled above the figure (population codes from Table 1

are labeled below the figure; * indicates introduced locations). Major watersheds are indicated by brackets above populations.

Distribution of cytochrome b haplotypes from the maximum likelihood tree (Fig. 4.2) among populations is shown below figure (solid

line = Black Sea clade haplotypes; dashed line = Caspian Sea clade haplotypes). 1 3 3 Fig. 4.6 Spatial principal components analysis (sPCA; Jombart et al. 2008) of allele frequencies for 355 individuals of monkey goby at 10 microsatellite loci. Solid lines indicate the connection network (Delaunay triangulation) among individuals. Square color/size indicate an individual's score on each spatial principal component. The first global scores (A) separate the Caspian Sea and lower Volga River samples from all others. The second global scores (B) generally separate samples from the Volga and Don

River region from all others. The third global scores (C) primarily separate the introduced Vistula River/Lagoon populations from the remaining western Black

Sea/Danube River populations (large white/black squares on left side of figure).

Autocorrelation statistics (Moran, 1948) for the global scores were I = 0.929, I = 0.858, and I = 0.813, respectively. D) Screeplot of the eigenvalues of sPCA. Retained components are black.

134 A.

135 B.

136 C.

137 D.

138 Fig. 4.7 Plot of the 2nd and 3rd principal components (PC) from a principal components analysis of morphometric data from 141 monkey gobies. ○ = Neogobius fluviatilis; □ =

N. pallasi. A multivariate analysis of variance using both body size and shape information (PC 1-5) as well as body shape alone (PC 2-5) detected highly significant differences among basins (PC 1-5, Wilks’ λ = 0.703, F5, 135 = 11.407, P << 0.001; PC 2-5,

Wilks’ λ = 0.704, F4, 136 = 14.303, P << 0.001). Centroids (black circle and square) and

95% confidence ellipses for the Black (solid line) and Caspian (dashed) Sea basin individuals are given.

139 Chapter Five

Conclusions and Future Research

General Conclusions

This dissertation investigated the systematics, evolutionary history, biogeography, phylogeography, and population genetics of a group of ~20 gobies endemic to the Ponto-

Caspian region. In Chapters 2 and 4, I examined both genetic and morphological variation in a widespread monotypic genus and single species, respectively, and found strong genetic divergence among regional lineages coincident with slight (but significantly different) morphological differences. In Chapter 3, I analyzed the phylogenetic relationships among the Ponto-Caspian endemic gobiids, presented a novel phylogenetic hypothesis for the group, and refined their taxonomy. In all three chapters, I estimated the divergence times among lineages at different hierarchical levels (species, genus, tribe, subfamily) and related these divergences to major events in the geological history of the Ponto-Caspian region. The three studies presented in this dissertation yield several general conclusions and patterns about the systematics, evolution, and phylogeography of the Ponto-Caspian endemic gobiids, which are outlined below.

Systematics of Ponto-Caspian gobies more complex than expected

One of the primary conclusions that arises from the preceding chapters is that the systematics and evolutionary history of the Ponto-Caspian endemic gobiids is more

140 complex than previously estimated. Prior to the present work and preliminary work performed in the Stepien laboratory (Dillon & Stepien 2001; Stepien et al. 2005; Stepien

& Tumeo 2006), no study had presented a phylogenetic hypothesis for the group. The majority of prior taxonomic studies on the “neogobiin” gobies were primarily descriptive in nature, with few studies (Pinchuk 1991; Vasil'eva et al. 1993) presenting a qualitative estimate of species interrelationships and none using phylogenetic methodology.

Vasil'eva et al. (1993) had suggested various species groupings within the genus

Ponticola based on general similarities in cranial morphometry and osteology, which were not recovered in the present study. Instead, two species groups of different composition to those proposed by Vasil'eva et al. (1993) were recovered, both of which were composed of species showing similar patterns of distribution among three regions

(northwest Black Sea/Crimean Peninsula, Azov Sea and eastern Black Sea rivers, and

Don/Volga Rivers/Caspian Sea). In addition, prior to this dissertation and previous work in the Stepien laboratory, no study had suggested a close relationship of Mesogobius and

Proterorhinus. Vasil'eva (1999) described the cranial osteology of Proterorhinus in comparison to Mesogobius, Neogobius, Ponticola, Babka, and Gobius sensu stricto, but only commented on taxonomic level and not systematic relationships. Ahnelt &

Duchkowitsch (2004) studied the postcranial osteology of Proterorhinus in comparison with species of Neogobius, Ponticola, and Babka and suggested a basic evolutionary scenario for the four genera, but they did not examine Mesogobius. The most important, previously unseen complexity, however, was the close relationship of the benthophiline gobies (Anatirostrum, Benthophiloides, Benthophilus, and Caspiosoma) with the remaining “neogobiin” lineages. Although Ahnelt (2003) hypothesized a closer

141 relationship of the benthophiline gobies to “neogobiin” gobies than to other

Mediterranean/Ponto-Caspian gobiids, no explicit relationship was presented. The present work, however, clearly shows that the benthophiline gobies (Benthophilini) and the “neogobiin” gobies (Neogobiini + Ponticolini) are part of a single, larger Ponto-

Caspian species flock, rather than two separate groups as previously thought.

Genetic divergence with little morphological change

Although the benthophiline gobies have undergone an astounding degree of morphological divergence since their separation from their common ancestor with the neogobiine and ponticoline gobies, the primary pattern emerging from Chapters 2 and 4 is one of relative morphological stasis despite large degrees of genetic change. In both of these chapters the amount of variation among genetic lineages ranged from 0.04-0.17, which is well within the range of intrageneric variation in teleost fishes (Johns & Avise

1998b). This large degree of genetic variation among lineages was observed within taxa assumed to be single, widespread species. Multiple freshwater and marine species of

Proterorhinus species had been described from both the Black and Caspian Sea basins before their grouping into Proterorhinus marmoratus by Berg (summarized in Chapter 2); several different forms/varieties had been described Neogobius fluviatilis that were later synonymized, and separate subspecies of N. fluviatilis were described in the Black and

Caspian basins (Pinchuk et al. 2003a). This history of taxonomic revision and synonymy indicates that many authors did not think that that the level of observed morphological variation in both of these groups warranted separation into distinct species. Indeed, although significant morphological differences among genetic lineages were identified for these two groups, these differences were quite subtle.

142 With the increased use of molecular tools in studies of systematics and biodiversity, the results seen in Chapters 2 and 4 of highly divergent, cryptic genetic lineages hidden within widespread morphological “species” is seemingly becoming a dominant pattern. Large genetic breaks within a single widespread morphospecies have been increasingly identified in a wide range of terrestrial and aquatic/marine taxa, including; mosses (Fernandez et al. 2006), pseudoscorpions (Wilcox et al. 1997), spiders

(Starrett & Hedin 2007), butterflies (Hebert et al. 2004a), parasitoid flies (Smith et al.

2006), frogs (Stuart et al. 2006; Elmer et al. 2007), birds (Baker et al. 1995; Olsson et al.

2005), voles (Jaarola & Searle 2004), giraffe (Brown et al. 2007), picoplanktonic alga

(Ślapeta et al. 2006), seaweeds (Andreakis et al. 2007), copepods (Lee 2000), amphipods

(Witt et al. 2006), shrimp (Tsoi et al. 2005), bryozoans (McGovern & Hellberg 2003), rotifers (Gómez et al. 2002), ascidians (Caputi et al. 2007), fishes (Colborn et al. 2001;

Yamazaki et al. 2003; Lima et al. 2005; Egge & Simons 2006; Quattro et al. 2006), and a wide variety of other marine taxa (summarized in Knowlton 1993, 2000). Although molecular identification and genetic tools have been suggested as a way of supplanting traditional morphology-based identification of species (Hebert et al. 2003), genetic information in concert with morphology could result in a more integrative taxonomy. For example, genetic lineages distinguished using DNA sequence data can be used as a priori hypotheses for detecting statistical differences in morphology or for identifying diagnostic characters. By incorporating the new strengths and resolution provided by molecular tools with the rich history of morphology-based systematics and taxonomy, a robust synthetic approach to describing biodiversity is possible (DeSalle et al. 2005;

Schlick-Steiner et al. 2007).

143 Congruent patterns of genetic divergence across time and taxa

The description of biodiversity, although useful and necessary, is simply a single step in discovering and identifying the mechanisms that generate biodiversity. Both historic and contemporary geologic processes that shape the evolution of landscapes are a major factor that can shape biodiversity in multiple lineages simultaneous as well as across multiple time scales, which is the third primary conclusion of this dissertation. In each chapter, the geologic evolution of the Ponto-Caspian basin seemed to play a major role driving divergence across multiple different taxonomic levels (e.g., species, genus, tribe) within the Ponto-Caspian gobies. Estimates of divergence times for various nodes within the phylogenies presented in each chapter generally show correspondence to major events in Ponto-Caspian geological history. Major transgressions and sea level rise generated new interbasin connections allowing taxa to flow between paleobasins, and water regression would sunder these connections causing vicariant separation of the formerly widespread taxa.

For example, separation of freshwater lineages of Proterorhinus occurred ~1 Mya, which coincides with the Apsheronian transgression: a major connection and movement of water and fauna from the proto-Black Sea basin to the proto-Caspian Sea basin (Reid

& Orlova 2002; Cristescu et al. 2003). This date is similar to the ages estimated for the

“kessleri” and “platyrostris” species groups within Ponticola (divergences approximately

1.3 and 1.8 Mya, respectively). These three groups of species, spread across multiple genera, all seemingly represent a common pattern: the common ancestor of each group became widespread across both basins during transgressive periods, subsequently split during regressive periods, and are currently represented by individual species living in

144 distinct regions (e.g., northwestern Black Sea, eastern Black Sea, Caspian Sea basin).

This correspondence of divergence to basin evolution is also evident at higher taxonomic levels: separations occurring between genera and tribes also show gross concordance to cycles of connection and isolation of the early Paratethys/Ponto-Caspian basin to both the

World Ocean and the Pannonian Sea (ancient lake-sea bed located in present day

Romania) approximately 5-9 Mya.

These patterns of genetic divergence in concert with landscape evolution is not observed solely in fishes. Multiple genetic studies of mysids (Cristescu & Hebert 2005;

Audzijonyte et al. 2006, 2008a), cladocerans (Cristescu & Hebert 2002; Cristescu et al.

2003), amphipods (Cristescu et al. 2004; Cristescu & Hebert 2005), and dreissenid mussels (Stepien et al. 2002; Gelembiuk et al. 2006) have all identified divergence among genetic lineages broadly coinciding with major points in the evolution of the

Ponto-Caspian basin. From this diverse group of fishes and arthropods, two major dates of divergence stand out: 1) 4-6 Mya, generally corresponding to the Messinian Salinity

Crisis (Hsü & Giovanoli 1979; Gillet et al. 2007) and the initial separation of the proto-

Black and Caspian Sea basins; and 2) 1-2 Mya, generally corresponding to major transgressions that reconnected the two basins, allowing faunal exchange. Fluctuating environmental conditions throughout these periods would have affected each individual lineage differently according to their ecological tolerances and adaptive potential, causing each to diverge at multiple different points in time and creating a complex overlap of phylogenetic and phylogeographic histories showing general concordance (Soltis et al.

2006).

145 Future Research

Several avenues of potential future research arise from the results of this dissertation. First and foremost is the systematics of the tribe Benthophilini. These species are united by several morphological synapomorphies, including: loss of head canals, tubular anterior nostril lacking a rim process, reduction or loss of scales, and distinct patterns of cephalic free neuromasts (Miller 2003a). Most work on the systematics of benthophiline gobies has been descriptive in scope. (Ragimov 1978;

Ragimov 1982; Ragimov 1985) focused on species and subspecies descriptions for Azov and Caspian taxa. Vasil'eva (1983, 2000) used cranial osteology to comment on the taxonomy of Caspian species of Benthophilus. Pinchuk (1980) showed a close relationship between Benthophiloides and Caspiosoma based on patterns of the cephalic free neuromasts. Ahnelt (2003) investigated the postcranial osteology of Benthophilus and Anatirostrum, and suggested a close relationship to the “neogobiin” gobies (Babka,

Mesogobius, Neogobius, Ponticola, and Proterorhinus in the present study). Boldyrev &

Bogutskaya (2007) revised Benthophilus, including redescriptions of all currently recognized species. Aside from this dissertation, benthophiline gobies have not been included in any recent morphological or molecular study of gobioid systematics (e.g.,

Birdsong et al. 1988; Pezold 1993; Akihito et al. 2000; Thacker 2003). The position of the tadpole gobies relative to the remaining two tribes within the Benthophilinae

(Neogobiini and Ponticolini) is poorly resolved in all phylogenetic analyses in the present work (Figure 3.2, Chapter 3), including short branch lengths and low support values. A phylogenetic analysis of all current benthophiline taxa would further enhance our knowledge of the systematics and evolution of the Ponto-Caspian endemic gobies. This

146 phylogeny would also provide a framework for investigating the tempo and mode of morphological evolution in this extremely divergent group.

A second interesting line of research would be in the comparative phylogeography of gobies in freshwater/brackish systems spanning the Don/Volga River basins and the

Caucasus Isthmus between the Black and Caspian Sea. Comparative phylogeography provides a framework for understanding landscape evolution and the effects of history and geography on species diversity and community structure (Bermingham & Moritz

1998). Three species groups from this region between the Black and Caspian Seas were identified in Chapter 3: the “kessleri” group (Ponticola eurycephalus, P. gorlap, and P. kessleri), the “platyrostris” group (P. cephalargoides, P. constructor, P. cyrius, P. platyrostris, and P. rhodioni), and the freshwater species of Proterorhinus. These three groups show similar distribution patterns and timing of divergences (Table 3.3).

Phylogeographic data from across the ranges of these species would give a more detailed insight into the patterns and timing of their divergences, and would further examine the importance of changes in basin shape and sea level (Reid & Orlova 2002) on taxa within the Ponto-Caspian region. A phylogeographic/population genetic study that includes P. kessleri would be of particular utility as this species is the only Ponto-Caspian goby introduced to northern/central Europe that has yet to be examined using molecular tools, and would contribute to our overall understanding of the genetic patterns of invasions by this group.

A third area of research would be investigating the phylogenetic determinants of community assembly and structure in the gobiid communities in the Ponto-Caspian region. The basic, inherent question in studies of phylogenetic community structure is:

147 are closely related species more or less likely to co-occur with one another, and why?

The ecological factors (e.g., competition, environmental preferences) underlying community structure have been widely studied; however, until the past decade few studies have incorporated the phylogenetic and evolutionary history of the species within the studied communities (Vamosi et al. 2009). Competition between closely related species, which likely have similar and overlapping ecological niches, should reduce the frequency that they co-occur in a particular habitat or area. By incorporating and controlling for environmental factors (Helmus et al. 2007) or morphology (Peres-Neto

2004), the phylogenetic impact on patterns of species co-occurrence can begin to be assessed. For example, are the three individual benthophilin goby tribes (Neogobiini,

Benthophilini, and Ponticolini) phylogenetically clustered or evenly dispersed within the

Black and Caspian Sea basin? This line of research ties together biogeography, evolution, and ecology, and utilizes the phylogenetic hypothesis generated in this dissertation as a springboard into a much different field.

Although there are many potential topics for future research on the Ponto-Caspian endemic gobiids (e.g., statistical analyses of morphology among taxa within each basin, population genetic/phylogeographic studies of each individual taxon), the three ideas outlined are likely the most fruitful and interesting areas that logically follow and would extend the results of this dissertation. These three areas of research would extend our knowledge of the evolution and systematics of the endemic gobies in the region, and yield valuable insight on the landscape, physical, and ecological drivers of species diversity and community structure.

148 References

Ahnelt H (2003) The postcranial skeleton of the benthophiline gobiids Anatirostrum and

Benthophilus (Teleostei: Gobiidae). Folia Zoologica, 52, 213-221.

Ahnelt H, Abdoli A, Naderi M, Coad BW (2000) Anatirostrum profundorum: a rare deep-

water gobiid species from the Caspian Sea. Cybium, 24, 139-159.

Ahnelt H, Bănărescu P, Spolwind R, Harka Á, Waidbacher H (1998) Occurrence and

distribution of three gobiid species (Pisces, Gobiidae) in the middle and upper

Danube region - examples of different dispersal pattern? Biologia (Bratislava),

53, 665-678.

Ahnelt H, Duchkowitsch M (2001) The lateral line system of two Ponto-Caspian gobiid

species (Gobiidae, Teleostei): a comparison. Folia Zoologica, 50, 217-230.

Ahnelt H, Duchkowitsch M (2004) The postcranial skeleton of Proterorhinus

marmoratus with remarks on the relationships of the genus Proterorhinus

(Teleostei: Gobiidae). Journal of Natural History, 38, 913-924.

Akihito, Iwata A, Kobayashi T, Ikeo K, Imanishi T, Ono H et al. (2000) Evolutionary

aspects of gobioid fishes based upon a phylogenetic analysis of mitochondrial

cytochrome b genes. Gene, 259, 5-15.

Andreakis N, Procaccini G, Maggs C, Kooistra WHCF (2007) Phylogeography of the

invasive seaweed Asparagopsis (Bonnemaisoniales, Rhodophyta) reveals cryptic

diversity. Molecular Ecology, 16, 2285-99.

149 Arbogast B, Kenagy GJ (2001) Comparative phylogeography as an integrative approach

to historical biogeography. Journal of Biogeography, 28, 819-825.

Audzijonyte A, Daneliya ME, Mugue N, Väinölä R (2008a) Phylogeny of Paramysis

(Crustacea: Mysida) and the origin of Ponto-Caspian endemic diversity: resolving

power from nuclear protein-coding genes. Molecular Phylogenetics and

Evolution, 46, 738-759.

Audzijonyte A, Daneliya ME, Väinölä R (2006) Comparative phylogeography of Ponto-

Caspian mysid crustaceans: isolation and exchange among dynamic inland sea

basins. Molecular Ecology, 15, 2969-2984.

Audzijonyte A, Wittman KJ, Väinölä R (2008b) Tracing recent invasions of the Ponto-

Caspian mysid shrimp Hemimysis anomala across Europe and to North America

with mitochondrial DNA. Diversity and Distributions, 14, 179-186.

Avise JC (2000) Phylogeography: the history and formation of species. Harvard

University Press, Cambridge, MA.

Avise JC (2004) Molecular markers, natural history and evolution. Sinauer Associates,

Sunderland, MA.

Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE et al. (1987)

Intraspecific phylogeography: the mitochondrial DNA bridge between population

genetics and systematics. Annual Review of Ecology and Systematics, 18, 489-

522.

Baimov UA (1963) On the feeding of the Aral Sea predatory fish in connection with the

introduction of Gobiidae. Voprosy Ikhtiologii, 3, 304-310.

Bajpai S, Kapur VV (2004) Oldest known gobiids from Vastan Lignite Mine (early

150 Eocene), Surat District, Gujarat. Current Science, 87, 433-435.

Baker AJ, Daugherty CH, Colbourne R, McLennan JL (1995) Flightless brown kiwis of

New Zealand possess extremely subdivided population structure and cryptic

species like small mammals. Proceedings of the National Academy of Sciences of

the United States of America, 92, 8254-8258.

Baker RH, DeSalle R (1997) Multiple sources of character information and the

phylogeny of Hawaiian drosophilids. Systematic Biology, 46, 654-73.

Bănărescu P (1970) Die Fische des ponto-kaspischen potamophilen Artenkomplexes und

die karpato-kaukasische Disjunction. Hydrobiologia, 11, 135-141.

Berg LS (1916) Freshwater fishes of the Russian empire. Izd-vo Departamenta

Zemledeliya, Moscow.

Berg LS (1949) Freshwater fishes of the U.S.S.R. and adjacent countries. Izdatel'vesto

Akademii Nauk SSSR, Moscow.

Bergstrom MA, Evrard LM, Mensinger AF (2008) Distribution, abundance, and range of

the round goby, Apollonia melanostoma, in the Duluth-Superior Harbor and St.

Louis River Estuary, 1998-2004. Journal of Great Lakes Research, 34, 535-543.

Berlin B, Breedlove DE, Raven PH (1966) Folk taxonomies and biological classification.

Science, 154, 273-275.

Bermingham E, Moritz C (1998) Comparative phylogeography: concepts and

applications. Molecular Ecology, 7, 367-369.

Bickford D, Lohman DJ, Sodhi NS, Ng PKL, Meier R, Winker K et al. (2007) Cryptic

species as a window on diversity and conservation. Trends in Ecology and

Evolution, 22, 148-155.

151 Birdsong RS, Murdy EO, Pezold FL (1988) A study of the vertebral column and median

fin osteology in gobioid fishes with comments on gobioid relationships. Bulletin

of Marine Science, 42, 174-214.

Blackburn TM (2008) Using aliens to explore how our planet works. Proceedings of the

National Academy of Sciences of the United States of America, 105, 9-10.

Boldyrev VS, Bogutskaya NG (2007) Revision of the tadpole-gobies of the genus

Benthophilus (Teleostei: Gobiidae). Ichthyological Explorations of Freshwaters,

18, 31-96.

Brower AVZ (2006) The how and why of branch support and partitioned branch support,

with a new index to asses partition incongruence. Cladistics: the international

journal of the Willi Hennig Society, 22, 378-386.

Brown DM, Brenneman RA, Koepfli K, Pollinger JP, Milá B, Georgiadis NJ et al. (2007)

Extensive population genetic structure in the giraffe. BMC Biology, 5, 57.

Brown JE, Stepien CA (2008) Ancient divisions, recent expansions: phylogeography and

population genetics of the round goby Apollonia melanostoma. Molecular

Ecology, 17, 2598-2615.

Brown JE, Stepien CA (2009) Invasion genetics of the Eurasian round goby in North

America: tracing sources and spread patterns. Molecular Ecology, 18, 64-79.

Bucciarelli G, Golani D, Bernardi G (2002) Genetic cryptic species as biological

invaders: the case of a Lessepsian fish migrant, the hardyhead silverside

Atherinomorus lacunosus. Journal of Experimental Marine Biology and Ecology,

273, 143-149.

Campbell TS, Echternacht AC (2003) Introduced species as moving targets: changes in

152 body sizes of introduced lizards following experimental introductions and

historical invasions. Biological Invasions, 5, 193-212.

Čápová M, Zlatnická I, Kováč V, Katina S (2008) Ontogenetic variability in the external

morphology of monkey goby, Neogobius fluviatilis (Pallas, 1814) and its

relevance to invasion potential. Hydrobiologia, 607, 17-26.

Caputi L, Andreakis N, Mastrototaro F, Cirino P, Vassillo M, Sordino P (2007) Cryptic

speciation in a model invertebrate chordate. Proceedings of the National Academy

of Sciences of the United States of America, 104, 9364-9369.

Chow S, Hazama K (1998) Universal PCR primers for S7 ribosomal protein gene introns

in fish. Molecular Ecology, 7, 1255-1256.

Cincotta CL, Adams JM, Holzapfel C (2009) Testing the enemy release hypothesis: a

comparison of foliar insect herbivory of the exotic Norway maple (Acer

platanoides L.) and the native sugar maple (A. saccharum L.). Biological

Conservation, 11, 379-388.

Clement M, Posada D, Crandall K (2000) TCS: a computer program to estimate gene

genealogies. Molecular Ecology, 9, 1657-1660.

Cohen AN, Carlton JT (1998) Accelerating invasion rate in a highly invaded estuary.

Science, 279, 555-558.

Colborn J, Crabtree RE, Shaklee JB, Pfeiler E, Bowen BW (2001) The evolutionary

enigma of bonefishes (Albula spp.): cryptic species and ancient separations in a

globally distributed shorefish. Evolution, 55, 807-820.

Corkum LD, Sapota MR, Skora KE (2004) The round goby, Neogobius melanostomus, a

fish invader on both sides of the Atlantic Ocean. Biological Invasions, 6, 173-181.

153 Corti M, Loy A, Cataudella S (1996) Form changes in the sea bass, Dicentrarchus labrax

(Moronidae: Teleostei), after acclimation to freshwater: an analysis using shape

coordinates. Environmental Biology of Fishes, 47, 165-175.

Coyne JA & Orr HA (2004) Speciation. Sinauer Associates, Sunderland, MA.

Cristescu MEA, Hebert PDN (2002) Phylogeny and adaptive radiation in the

Onychopoda (Crustacea, Cladocera): evidence from multiple gene sequences.

Journal of Evolutionary Biology, 15, 838-849.

Cristescu MEA, Hebert PDN (2005) The "Crustacean Seas" - an evolutionary perspective

on the Ponto-Caspian peracarids. Canadian Journal of Fisheries and Aquatic

Sciences, 62, 505-517.

Cristescu MEA, Hebert PDN, Onciu TM (2003) Phylogeography of Ponto-Caspian

crustaceans: a benthic-planktonic comparison. Molecular Ecology, 12, 985-996.

Cristescu MEA, Witt JDS, Grigorovich IA, Hebert PDN, MacIsaac HJ (2004) Dispersal

of the Ponto-Caspian amphipod Echinogammarus ischnus: invasion waves from

the Pleistocene to the present. Heredity, 92, 197-203.

Cunningham CW (1997) Can three incongruence tests predict when data should be

combined. Molecular Biology and Evolution, 14, 733-740.

Danilkiewicz Z (1998) The monkey goby, Neogobius fluviatilis (Pallas, 1811),

Perciformes, Gobiidae – a new species in the ichthyofauna of the Baltic Sea basin.

Fragmenta Faunistica, 41, 269-277. de Queiroz K (2005) Ernst Mayr and the modern concept of species. Proceedings of the

National Academy of Sciences of the USA, 102, 6600-6607.

DeSalle R, Egan MG, Siddall M (2005) The unholy trinity: taxonomy, species

154 delimitation and DNA barcoding. Philosophical Transactions of the Royal Society

B-Biological Sciences, 360, 1905-1916.

Dillon AK, Stepien CA (2001) Genetic and biogeographic relationships of invasive round

(Neogobius melanostomus) and tubenose (Proterorhinus marmoratus) gobies in

the Great Lakes versus Eurasian populations. Journal of Great Lakes Research,

27, 267-280.

Dobrovolov JS, Georghiev TU, Dobrolova SG (1995) Comparative electrophoretic

investigations of the species of the family Gobiidae (Pisces) in the Bulgarian

sector of the Black Sea. Proceedings of the Institute of Fisheries, Varna, 23, 48-

68.

Dolphin K, Belshaw R, Orme CDL, Quicke DLJ (2000) Noise and incongruence:

Interpreting results of the incongruence length difference test. Molecular

Phylogenetics and Evolution, 17, 401-406.

Dougherty JD, Moore WS, Ram JL (1996) Mitochondrial DNA analysis of round goby

(Neogobius melanostomus) and tubenose goby (Proterorhinus marmoratus) in the

Great Lakes basin. Canadian Journal of Fisheries and Aquatic Sciences, 53, 474-

480.

Dufour BA, Hogan TM, Heath DD (2007) Ten polymorphic microsatellite markers in the

invasive round goby (Neogobius melanostomus) and cross-species amplification.

Molecular Ecology Notes, 7, 1205-1207.

Dumont HJ (1998) The Caspian Lake: History, biota, structure, and function. Limnology

and Oceanography, 43, 44-52.

Dumont HJ (2000) Endemism in the Ponto-Caspian fauna, with special emphasis on the

155 Onychopoda (Crustacea). Advances in Ecological Research, 31, 181-196.

Durand JD, Persat H, Bouvet Y (1999) Phylogeography and postglacial dispersion of the

chub (Leuciscus cephalus) in Europe. Molecular Ecology, 8, 989-997.

Dutton JM, Hofmann GE (2008) Spatial and temporal variation in distribution and

protein ubiquitination for Mytilus congeners in the California hybrid zone.

Marine Biology, 154, 1067-1075.

Egge JJD, Simons AM (2006) The challenge of truly cryptic diversity: diagnosis and

description of a new madtom catfish (Ictaluridae: Noturus). Zoologica Scripta, 35,

581-595.

Elmer KR, Dávila JA, Lougheed SC (2007) Cryptic diversity and deep divergence in an

upper Amazonian leaflitter frog, Eleutherodactylus ockendeni. BMC Evolutionary

Biology, 7, 247.

Erős T, Sevcsik A, Tóth B (2005) Abundance and night-time habitat use patterns of

Ponto-Caspian gobiid species (Pisces, Gobiidae) in the littoral zone of the River

Danube, Hungary. Journal of Applied Ichthyology, 21, 350-357.

Excoffier L, Laval G, Schneider S (2005) Arlequin ver. 3.0: an integrated software

package for population genetics data analysis. Evolutionary Bioinformatics

Online, 1, 47-50.

Farris JS, Kallersjo M, Kluge AG, Bult C (1995) Constructing a significance test for

incongruence. Systematic Biology, 44, 570-572.

Feldheim KA, Willink P, Brown JE, Murphy DJ, Neilson ME, Stepien CA (2009)

Microsatellite loci for Ponto-Caspian gobies: markers for assessing exotic

invasions. Molecular Ecology Resources, 9, 639-644.

156 Fernandez CC, Shevock JR, Glazer AN, Thompson JN (2006) Cryptic species within the

cosmopolitan desiccation-tolerant moss Grimmia laevigata. Proceedings of the

National Academy of Sciences of the United States of America, 103, 637-42.

Filippi FD (1863) Nuove o poco note specie di animali vertebrati raccolte in un viaggio

in Persia nell'estate dell'ano 1862. Archivio per la Zoologia, l'Anatomia e la

Fisiologia, Genova, 2, 377-394.

Freyhof J, Naseka AM (2007) Proterorhinus tataricus, a new tubenose goby from

Crimea, Ukraine (Teleostei: Gobiidae). Ichthyological Exploration of

Freshwaters, 18, 325-334.

Gelembiuk GW, May GE, Lee CE (2006) Phylogeography and systematics of zebra

mussels and related species. Molecular Ecology, 15, 1033-1050.

Geller JB, Walton ED, Grosholz ED, Ruiz GM (1997) Cryptic invasions of the crab

Carcinus detected by molecular phylogeography. Molecular Ecology, 6, 901-906.

Gillet H, Lericolais G, Réhault J (2007) Messinian event in the Black Sea: evidence of a

Messinian erosional surface. Marine Geology, 244, 142-165.

Gómez A, Serra M, Carvalho GR, Lunt DH (2002) Speciation in ancient cryptic species

complexes: evidence from the molecular phylogeny of Brachionus plicatilis

(Rotifera). Evolution, 56, 1431-1444.

Greenwood PH (1984) What is a species flock? In: Evolution of fish species flocks (eds.

Echelle AA, Kornfield I), pp. 13-19. University of Maine Press, Orono, ME.

Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large

phylogenies by maximum likelihood. Systematic Biology, 52, 696-704.

Günther A (1861) Catalogue of the acanthopterygian fishes in the collection of the

157 British Museum. 3. Gobiidae, Discoboli, Pediculati, Blenniidae, Labyrinthici,

Mugilidae, Notacanthi. London.

Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis

program for Windows 95/98/NT. Nucleic Acids Symposium Series, 41, 95-98.

Harada S, Jeon S, Kinoshita I, Tanaka M, Nishida M (2002) Phylogenetic relationships of

four species of floating gobies (Gymnogobius) as inferred from partial

mitochondrial cytochrome b sequences. Ichthyological Research, 49, 324-332.

Harka A, Jakab T (2001) Growth and feeding of one year old river goby (Neogobius

fluviatilis) in the Kisköre Reservoir. Halászat, 94, 161-164.

Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003) Biological identifications through

DNA barcodes. Proceedings. Biological sciences / The Royal Society, 270, 313-

21.

Hebert PDN, Penton EH, Burns JM, Janzen DH, Hallwachs W (2004a) Ten species in

one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly

Astraptes fulgerator. Proceedings of the National Academy of Sciences of the

United States of America, 101, 14817.

Hebert PDN, Stoeckle MY, Zemlak TS, Francis CM (2004b) Identification of birds

through DNA barcodes. PLoS Biology, 2, 1657-1663.

Helmus MR, Savage K, Diebel MW, Maxted JT, Ives AR (2007) Separating the

determinants of phylogenetic community structure. Ecology Letters, 10, 917-25.

Hillis DM (1998) Taxonomic sampling, phylogenetic accuracy, and investigator bias.

Systematic Biology, 47, 3-8.

Holčík J, Stráňai I, Andreji J (2003) The further advance of Neogobius fluviatilis (Pallas,

158 1814) (Pisces, Gobiidae) upstream of the Danube. Biologia (Bratislava), 58, 967-

973.

Hsü KJ, Giovanoli F (1979) Messinian event in the Black Sea. Palaeogeography,

Palaeoclimatology, Palaeoecology, 29, 75-93.

Iljin BS (1927) Keys to the gobies (Fam. Gobiidae) of the Sea of Azov and the Black Sea.

Trudy Azovsko-Chernomorskoe Naucho Promyslovoe Ekspedit'sii, 2, 128-143.

Jaarola M, Searle JB (2004) A highly divergent mitochondrial DNA lineage of Microtus

agrestis in southern Europe. Heredity, 92, 228-234.

Johns GC, Avise JC (1998a) Tests for ancient species flocks based on molecular

phylogenetic appraisals of Sebastes rockfishes and other marine fishes. Evolution,

52, 1135-1146.

Johns GC, Avise JC (1998b) A comparative summary of genetic distances in the

vertebrates from the mitochondrial cytochrome b gene. Molecular Biology and

Evolution, 15, 1481-1490.

Johnson JB, Dowling TE, Belk MC (2004) Neglected taxonomy of rare desert fishes:

congruent evidence for two species of leatherside chub. Systematic Biology, 53,

841-855.

Johnson TB, Allen M, Corkum LD, Lee VA (2005) Comparison of methods needed to

estimate population size of round gobies (Neogobius melanostomus) in western

Lake Erie. Journal of Great Lakes Research, 31, 78-86.

Jude DJ, DeBoe SF (1996) Possible impact of gobies and other introduced species on

habitat restoration efforts. Canadian Journal of Fisheries and Aquatic Sciences,

53, 136-141.

159 Jude DJ, Jannsen J, Crawford G (1995) Ecology, distribution, and impact of the newly

introduced round and tubenose gobies on the biota of the St. Clair & Detroit

Rivers. In: The Lake Huron ecosystem: ecology, fisheries and management (eds.

Munawar M, Edsall T, Leach J), pp. 447-460. SPB Academic Publishing,

Amsterdam.

Jude DJ, Reider RH, Smith GR (1992) Establishment of Gobiidae in the Great Lakes

basin. Canadian Journal of Fisheries and Aquatic Sciences, 49, 416-421.

Jurajda P, Černý J, Polačik M, Valová Z, Janáč M, Blažek R et al. (2005) The recent

distribution and abundance of non-native Neogobius fishes in the Slovak section

of the River Danube. Journal of Applied Ichthyology, 21, 319-323.

Kakareko T, Żbikowski J, Żytkowicz J (2005) Diet partitioning in summer of two

syntopic neogobiids from two different habitats of the lower Vistula River,

Poland. Journal of Applied Ichthyology, 21, 292-295.

Kaplan P (1995) The Caspian: it's past, present and future. In: Enclosed seas and large

lakes of Eastern Europe and Middle Asia (ed. Mandych AF), pp. 71-117. SPB

Academic Publishing, Amsterdam.

Kessler KT (1857) Nachträge zur Ichthyologie des südwestlichen Russlands. Bulletin de

la Société Impériale des Naturalistes de Moscou, 30, 453-481.

Kessler KT (1877) The Aralo-Caspian expedition. IV. Fishes of the Aralo-Caspio-Pontine

ichthyological region. St. Petersburg.

Knowlton N (1993) Sibling species in the sea. Annual Review of Ecology and

Systematics, 24, 189-216.

Knowlton N (2000) Molecular genetic analyses of species boundaries in the sea.

160 Hydrobiologia, 420, 73-90.

Kolar CS, Lodge DM (2002) Ecological predictions and risk assessment for alien fishes

in North America. Science, 298, 1233-1236.

Kon T, Yoshino T, Mukai T, Nishida M (2007) DNA sequences identify numerous cryptic

species of the vertebrate: A lesson from the gobioid fish Schindleria. Molecular

Phylogenetics and Evolution, 44, 53-62.

Kostrzewa J, Grabowski M (2002) Monkey goby, Neogobius fluviatilis (Pallas 1811), in

the Vistula River – a phenomenon of Ponto-Caspian Gobiidae invasion. Przegląd

Zoologiczny, 46, 235-242.

Kotlík P, Marková S, Choleva L, Bogutskaya NG, Ekmekçi FG, Ivanova PP (2008)

Divergence with gene flow between Ponto-Caspian refugia in an anadromous

cyprinid Rutilus frisii revealed by multiple gene phylogeography. Molecular

Ecology, 17, 1076-1088.

Kovačić M, Engin S (2008) Two new species of Neogobius from northeastern Turkey.

Cybium, 32, 73-80.

Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H et al.

(2007) Clustal W and Clustal X version 2.0. Bioinformatics (Oxford, England),

23, 2947-2948.

Lee CE (2000) Global phylogeography of a cryptic copepod species complex and

reproductive isolation between genetically proximate "populations". Evolution,

54, 2014-2027.

Lima D, Freitas JEP, Araujo ME, Solé-Cava AM (2005) Genetic detection of a cryptic

species in the frillfin goby Bathygobius soporator. Journal of Experimental

161 Marine Biology and Ecology, 320, 211-223.

Linnaeus C (1758) Systema Naturae, Ed. X.

Lockwood JL, Hoopes MF & Marchetti MP (2006) Invasion ecology. Blackwell

Publishing, Malden, MA.

Lopez JA, Chen WJ, Orti G (2004) Esociform phylogeny. Copeia, 2004, 449-464.

McGovern TM, Hellberg ME (2003) Cryptic species, cryptic endosymbionts, and

geographical variation in chemical defenses in the bryozoan Bugula neritina.

Molecular Ecology, 12, 1207-1215.

McKay SI, Miller PJ (1991) Isozyme criteria in the testing of phyletic relationships

between species of Gobius and related eastern Atlantic-Mediterranean genera

(Teleostei: Gobiidae). Journal of Fish Biology, 39, 291-299.

Miller PJ (1973) The osteology and adaptive features of Rhyacichthys aspro (Teleostei:

Gobioidei) and the classification of gobioid fishes. Journal of Zoology (London),

171, 397-434.

Miller PJ (1990) The endurance of endemism: the Mediterranean freshwater gobies and

their prospects for survival. Journal of Fish Biology, 37, 145-156.

Miller PJ (2003a) Family Gobiidae Risso, 1826. In: The freshwater fishes of Europe, vol.

8(1) (ed. Miller PJ), pp. 157-162. AULA-Verlag, Wiesbaden.

Miller PJ (2003b) Chasar Vasil'eva 1996. In: The freshwater fishes of Europe, vol. 8(2)

(ed. Miller PJ), pp. 94-96. AULA-Verlag, Wiesbaden.

Miller PJ, Vasil'eva ED (2003) Neogobius Iljin, 1927. In: The freshwater fishes of

Europe, vol 8(1) (ed. Miller PJ), pp. 164-171. AULA-Verlag, Wiesbaden.

Mills EL, Leach JH, Carlton JT, Secor CL (1993) Exotic species in the Great Lakes: a

162 history of biotic crises and anthropogenic introductions. Journal of Great Lakes

Research, 19, 1-54.

Mishler BD, Theriot EC (2000) The phylogenetic species concept (sensu Mishler and

Theriot): monophyly, apomorphy, and phylogenetic species concepts. In: Species

concepts and phylogenetic theory: a debate (eds. Wheeler QD, Meier R), pp. 44-

54. Columbia University Press, New York.

Mondor EB, Tremblay MN, Messing RH (2007) Morphological and ecological traits

promoting aphid colonization of the Hawaiian Islands. Biological Conservation,

9, 87-100.

Mooney HA, Cleland EE (2001) The evolutionary impact of invasive species.

Proceedings of the National Academy of Sciences of the United States of America,

98, 5446-5451.

Mukai T, Suzuki T, Nishida M (2004) Genetic and geographical differentiation of

Pandaka gobies in Japan. Ichthyological Research, 51, 222-227.

Naseka AM, Boldyrev VS, Bogutskaya NG, Delitsyn VV (2005) New data on the

historical and expanded range of Proterorhinus marmoratus (Pallas, 1814)

(Teleostei: Gobiidae) in eastern Europe. Journal of Applied Ichthyology, 21, 300-

305.

Nelson JS (2006) Fishes of the world. John Wiley and Sons, Inc, New York, NY, .

Notredame C, Higgins DG, Heringa J (2000) T-Coffee: A novel method for fast and

accurate multiple sequence alignment. Journal of Molecular Biology, 302, 205-

17.

Olsson U, Alström P, Ericson PGP, Sundberg P (2005) Non-monophyletic taxa and

163 cryptic species--Evidence from a molecular phylogeny of leaf-warblers

(Phylloscopus, Aves). Molecular Phylogenetics and Evolution, 36, 261-276.

Pallas PS (1814) Zoographia Rossa-Asiatica, sistens omnium anmalium in extenso

Imperio Rossico et adjacentibus maribus observatorum recensionem, domicilia,

mores, et descriptiones anatomen atque icones plurimorum. St. Petersburg.

Penzo E, Gandolfi G, Bargelloni L, Colombo L, Patarnello T (1998) Messinian salinity

crisis and the origin of freshwater lifestyle in Western Mediterranean gobies.

Molecular Biology and Evolution, 15, 1472-1480.

Peres-Neto PR (2004) Patterns in the co-occurrence of fish species in streams: the role of

site suitability, morphology and phylogeny versus species interactions. Oecologia,

140, 352-360.

Pezold FL (1993) Evidence for a monophyletic Gobiinae. Copeia, 1993, 634-643.

Pfenninger M, Posada D (2002) Phylogeographic history of the land snail Candidula

unifasciata (Helicellinae, Stylommatophora): Fragmentation, corridor migration,

and secondary contact. Evolution, 56, 1776-1788.

Pinchuk VI (1980) The lateral-line system of Caspiosoma caspium (Kessler, 1877) and

the systematic status of Asra turcomanus Iljin, 1941. Journal of Fish Biology, 17,

231-235.

Pinchuk VI (1991) Species groupings in the genus Neogobius (Perciformes). Voprosy

Ikhtiologii, 31, 380-393.

Pinchuk VI, Vasil'eva ED, Vasil'ev VP, Miller PJ (2003a) Neogobius fluviatilis (Pallas,

1814). In: The freshwater fishes of Europe, vol. 8(1) (ed. Miller PJ), pp. 222-252.

AULA-Verlag, Wiesbaden.

164 Pinchuk VI, Vasil'eva ED, Vasil'ev VP, Miller PJ (2003b) Proterorhinus marmoratus

(Pallas, 1814). In: The freshwater fishes of Europe, vol. 8(1) (ed. Miller PJ), pp.

72-93. AULA-Verlag, Wiesbaden.

Posada D, Buckley TR (2004) Model selection and model averaging in phylogenetics:

advantages of the AIC and Bayesian approaches over likelihood ratio tests.

Systematic Biology, 53, 793-808.

Posada D, Crandall KA (1998) Modeltest: testing the model of DNA substitution.

Bioinformatics (Oxford, England), 14, 817-818.

Posada D, Crandall KA, Templeton AR (2000) GeoDis: a program for the cladistic nested

analysis of the geographical distribution of genetic haplotypes. Molecular

Ecology, 9, 487-488.

Prášek V, Jurajda P (2005) Expansion of Proterorhinus marmoratus in the Morava River

basin (Czech Republic, Danube R. watershed). Folia Zoologica, 54, 189-192.

Prentis PJ, Wilson JRU, Dormontt EE, Richardson DM, Lowe AJ (2008) Adaptive

evolution in invasive species. Trends in Plant Science, 13, 288-294.

Quattro JM, Stoner DS, Driggers WB, Anderson CA, Priede KA, Hoppmann EC et al.

(2006) Genetic evidence of cryptic speciation within hammerhead sharks (Genus

Sphyrna). Marine Biology, 148, 1143-1155.

Quenouille B, Bermingham E, Planes S (2004) Molecular systematics of the damselfishes

(Teleostei: Pomacentridae): Bayesian phylogenetic analyses of mitochondrial and

nuclear DNA sequences. Molecular Phylogenetics and Evolution, 31, 66-88.

R Development Core Team (2009) R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria.

165 http://www.R-project.org

Ragimov DB (1978) On the taxonomic status of some species of the genus Benthophilus

(Gobiidae) from the Caspian and Azov Seas. Voprosy Iktiologii, 18, 791-798.

Ragimov DB (1982) New subspecies of Caspian gobies (Gobiidae, Benthophilus).

Zoologichesky Zhurnal, 61, 47-55.

Ragimov DB (1985) Some Caspian species of the genus Benthophilus (Gobiidae).

Voprosy Iktiologii, 25, 937-944.

Reid DF, Orlova MI (2002) Geological and evolutionary underpinnings for the success of

Ponto-Caspian species invasions in the Baltic Sea and North American Great

Lakes. Canadian Journal of Fisheries and Aquatic Sciences, 59, 1144-1158.

Ricciardi A (2006) Patterns of invasion in the Laurentian Great Lakes in relation to

changes in vector activity. Diversity and Distributions, 12, 425-433.

Ricciardi A, MacIsaac HJ (2000) Recent mass invasion of the North American Great

Lakes by Ponto-Caspian species. Trends in Ecology and Evolution, 15, 62-65.

Ricciardi A, Rasmussen JB (1998) Predicting the identity and impact of future biological

invaders: a priority for aquatic resource management. Canadian Journal of

Fisheries and Aquatic Sciences, 55, 1759-1765.

Rice WR (1989) Analyzing tables of statistical tests. Evolution, 43, 223-225.

Riddle BR, Funk V (2004) Phylogeography and diversification. In: Frontiers of

biogeography: new directions in the geography of nature (eds. Lomolino MV,

Heaney LR), pp. 87-91. Sinauer Associates, Sunderland, MA.

Riddle BR, Hafner DJ (2004) The past and future roles of phylogeography in historical

biogeography. In: Frontiers in biogeography: new directions in the geography of

166 nature (eds. Lomolino MV, Heaney LR), pp. 93-110. Sinauer Associates,

Sunderland, MA.

Rocha LA, Robertson DR, Rocha CR, Van Tassell JL, Craig MT, Bowens BW (2005)

Recent invasion of the tropical Atlantic by an Indo-Pacific coral reef fish.

Molecular Ecology, 14, 3921-3928.

Rögl F (1999) Mediterranean and Paratethys: facts and hypotheses of an Oligocene to

Miocene paleogeography (short overview). Geologica Carpathica, 50, 339-349.

Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under

mixed models. Bioinformatics (Oxford, England), 19, 1572-1574.

Rosenberg NA (2007) Statistical tests for taxonomic distinctiveness from observations of

monophyly. Evolution, 61, 317-323.

Rückert-Ülkümen N (2006) Otolithen aus dem Mio-Pliozän von Yalova bei Istanbul,

Türkei. Neues Jahrbuch für Geologie Und Palaeontologie. Monatshefte, 10, 577-

594.

Ruiz GM, Carlton JT, Grosholz ED, Hines AM (1997) Global invasions of marine and

estuarine habitats by non-indigenous species: mechanisms, extent, and

consequences. American Zoologist, 37, 621-632.

Sanderson MJ (2002) Estimating absolute rates of molecular evolution and divergence

times: a penalized likelihood approach. Molecular Biology and Evolution, 19,

101-109.

Sanderson MJ (2003) r8s: inferring absolute rates of molecular evolution and divergence

times in the absence of a molecular clock. Bioinformatics (Oxford, England), 19,

301-302.

167 Sax DF, Gaines SD (2008) Species invasions and extinction: the future of native

biodiversity on islands. Proceedings of the National Academy of Sciences of the

United States of America, 105 Suppl 1, 11490-11497.

Sax DF, Stachowicz JJ, Brown JH, Bruno JF, Dawson MN, Gaines SD et al. (2007)

Ecological and evolutionary insights from species invasions. Trends in Ecology

and Evolution, 22, 465-471.

Schlick-Steiner BC, Seifert B, Stauffer C, Christian E, Crozier RH, Steiner FM (2007)

Without morphology, cryptic species stay in taxonomic crypsis following

discovery. Trends in Ecology and Evolution, 22, 391-392.

Shimodaira H, Hasegawa M (1999) Multiple comparisons of log-likelihoods with

applications to phylogenetic inference. Molecular Biology and Evolution, 16,

1114-1116.

Shwartz A, Strubbe D, Butler CJ, Matthysen E, Kark S (2009) The effect of enemy-

release and climate conditions on invasive birds: a regional test using the rose-

ringed parakeet (Psittacula krameri) as case study. Diversity and Distributions,

15, 310-318.

Simberloff D, Von Holle B (1999) Positive interactions of nonindigenous species:

invasional meltdown? Biological Invasions, 1, 21-32.

Simonovic PD (1999) Phylogenetic relationships of Ponto-Caspian gobies and their

relationship to the Atlantic-Mediterranean Gobiinae. Journal of Fish Biology, 54,

533-555.

Ślapeta J, López-García P, Moreira D (2006) Global dispersal and ancient cryptic species

in the smallest marine eukaryotes. Molecular Biology and Evolution, 23, 23-29.

168 Smith MA, Woodley NE, Janzen DH, Hallwachs W, Hebert PDN (2006) DNA barcodes

reveal cryptic host-specificity within the presumed polyphagous members of a

genus of parasitoid flies (Diptera: Tachinidae). Proceedings of the National

Academy of Sciences of the United States of America, 103, 3657-3662.

Smitt FA (1900) Preliminary notes on the arrangement of the genus Gobius, with an

enumeration of its European species. Öfversigt af Kongliga Vetenskaps-

Akademiens Förhandlingar, 56, 543-555.

Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS (2006) Comparative

phylogeography of unglaciated eastern North America. Molecular Ecology, 15,

4261-4293.

Starrett J, Hedin M (2007) Multilocus genealogies reveal multiple cryptic species and

biogeographical complexity in the California turret spider Antrodiaetus riversi

(Mygalomorphae, Antrodiaetidae). Molecular Ecology, 16, 583-604.

Stefanov T, Vassilev M, Dobrev D (2008) First occurrence of Neogobius fluviatilis

(Pallas, 1814) (Pisces: Gobiidae) in Aegean Sea Watershed. Acta Zoologica

Bulgarica, 60, 331-334.

Stepien CA, Brown JE, Neilson ME, Tumeo MA (2005) Genetic diversity of invasive

species in the Great Lakes versus their Eurasian source populations: insights for

risk analysis. Risk Analysis, 25, 1043-1060.

Stepien CA, Taylor CD, Dabrowska KA (2002) Genetic variability and phylogeographic

patterns of a nonindigenous species invasion: a comparison of exotic versus

native zebra and quagga mussel populations. Journal of Evolutionary Biology, 15,

314-328.

169 Stepien CA, Taylor CD, Grigorovich IA, Shirman SV, Wei R, Korniushin AV et al. (2003)

DNA and systematic analysis of invasive and native dreissenid mussels: is

Dreissena bugensis really D. rostriformis? Aquatic Invaders, 14, 1-10.

Stepien CA, Tumeo MA (2006) Invasion genetics of Ponto-Caspian gobies in the Great

Lakes: a "cryptic" species, absence of founder effects, and comparative risk

analysis. Biological Invasions, 8, 61-78.

Stuart BL, Inger RF, Voris HK (2006) High level of cryptic species diversity revealed by

sympatric lineages of Southeast Asian forest frogs. Biology Letters, 2, 470-474.

Swofford DL (2003) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other

Methods). Sinauer Associates, Sunderland, MA.

Tajima F (1989) Statistical methods of testing the neutral mutation hypothesis by DNA

polymorphism. Genetics, 123, 585-595.

Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular evolutionary genetics

analysis (MEGA) software version 4.0. Molecular Biology and Evolution, 24,

1596-1599.

Templeton AR (1998) Nested clade analyses of phylogeographic data: testing hypotheses

about gene flow and population history. Molecular Ecology, 7, 381-397.

Templeton AR (2004) Statistical phylogeography: methods of evaluating and minimizing

inference errors. Molecular Ecology, 13, 789-809.

Thacker CE (2003) Molecular phylogeny of the gobioid fishes. Molecular Phylogenetics

and Evolution, 26, 354-368.

Thacker CE (2009) Phylogeny of Gobioidei and placement within Acanthomorpha, with

a new classification and investigation of diversification and character evolution.

170 Copeia, 2009, 93-104.

Tsoi KH, Wang ZY, H CK (2005) Genetic divergence between two morphologically

similar varieties of the kuruma shrimp Penaeus japonicus. Marine Biology, 147,

367-379.

USGS (2003) Nonindigenous Aquatic Species Database (http://nas.er.usgs.gov). US

Geological Survey, Reston, VA.

Vamosi SM, Heard SB, Vamosi JC, Webb CO (2009) Emerging patterns in the

comparative analysis of phylogenetic community structure. Molecular Ecology,

18, 572-592.

Vanderploeg HA, Nalepa TF, Jude DJ, MIlls EL, Holeck K, Liebig JR et al. (2002)

Dispersal and emerging ecological impacts of Ponto-Caspian species in the

Laurentian Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences, 59,

1209-1228.

Vasil'ev VP, Grigoryan KA (1992) Karyology of fishes of the family Gobiidae. Voprosy

Ikhtiologii, 32, 27-40.

Vasil'eva ED (1983) Osteological analysis of certain Caspian pugolovkas in connection

with the systematics of the genus Benthophilus Gobiidae. Voprosy Ikhtiologii, 23,

544-556.

Vasil'eva ED (1989) Cranial morphology of the round goby Gobius melanostomus and

syrman goby Gobius syrman in relation to their position in the genus Gobius

sensu lato. Voprosy Ikhtiologii, 29, 186-197.

Vasil'eva ED (1991) The morphology of the skull of goby Gobius gorlap Iljin in

connection with its place in the genus Gobius sensu lato (Gobiidae, Pisces).

171 Byulleten' Moskovskogo Obshchestva Ispytatelei Prirody Otdel Biologicheskii, 96,

36-45.

Vasil'eva ED (1999) Taxonomic status of the tube-nosed goby Gobius marmoratus Pallas

(Gobiidae) based on its skull morphology. Voprosy Ikhtiologii, 39, 139-148.

Vasil'eva ED (2000) Craniological analysis of some species of Benthophilus Gobiidae

and problems of taxonomy and phylogenetic relations of this group. Voprosy

Ikhtiologii, 40, 764-771.

Vasil'eva ED, Vasil'ev VP, Pinchuk VI (1993) Craniological analysis of the goby

subgenus Ponticola Iljin, 1927. III. Comparative morphological study of

Neogobius kessleri, N. ratan, and additional findings on N. syrman relevant to the

diagnosis and content of the subgenus. Voprosy Ikhtiologii, 33, 609-617.

Vyskočilová M, Ondračková M, Šimková A, Martin J (2007) Isolation and

characterization of microsatellites in Neogobius kessleri (Perciformes, Gobiidae)

and cross-species amplification within the family Gobiidae. Molecular Ecology

Notes, 7, 701-704.

Walker JA, Bell MA (2000) Net evolutionary trajectories of body shape evolution within

a microgeographic radiation of threespine sticklebacks (Gasterosteus aculeatus).

Journal of Zoology, London, 252, 293-302.

Wiens JJ, Penkrot TL (2002) Delimiting species based on DNA and morphological

variation and discordant species limits in spiny lizards (Sceloporus). Systematic

Biology, 51, 69-91.

Wilcox TP, Hugg L, Zeh JA, Zeh DW (1997) Mitochondrial DNA sequencing reveals

extreme genetic differentiation in a cryptic species complex of neotropical

172 pseudoscorpions. Molecular Phylogenetics and Evolution, 7, 208-216.

Wiley EO, Mayden RL (2000) The evolutionary species concept. In: Species concepts

and phylogenetic theory: a debate (eds. Wheeler QD, Meier R), pp. 70-89.

Columbia University Press, New York.

Witt JDS, Threloff DL, Hebert PDN (2006) DNA barcoding reveals extraordinary cryptic

diversity in an amphipod genus: implications for desert spring conservation.

Molecular Ecology, 15, 3073-3082.

Yamazaki Y, Goto A, Nishida M (2003) Mitochondrial DNA sequence divergence

between two cryptic species of Lethenteron, with reference to an improved

identification technique. Journal of Fish Biology, 62, 591-609.

Zaitsev Y & Mamaev V (1997) Marine biological diversity in the Black Sea: a study of

change and decline. United Nations Publications, New York.

173