Invasion Dynamics of the Round Goby {Neogobius Meianostomus) in the Trent-Severn Waterway

Invasion dynamics of the round goby {Neogobius meianostomus) in the Trent-Severn Waterway

A Thesis Submitted to the Committee on Graduate Studies in Partial Fulfillment of the Requirements for the Degree of Master of Science in the Faculty of Arts and Science

TRENT UNIVERSITY Peterborough, Ontario, Canada © Copyright by Jake Brownscombe 2011 Environmental and Life Sciences M.Sc. Graduate Program September 2011 Library and Archives Bibliotheque et 1*1 Canada Archives Canada Published Heritage Direction du Branch Patrimoine de I'edition

395 Wellington Street 395, rue Wellington Ottawa ON K1A 0N4 OttawaONK1A0N4 Canada Canada

Your file Votre reference ISBN: 978-0-494-81136-8 Our file Notre reference ISBN: 978-0-494-81136-8

NOTICE: AVIS:

The author has granted a non L'auteur a accorde une licence non exclusive exclusive license allowing Library and permettant a la Bibliotheque et Archives Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par I'lnternet, preter, telecommunication or on the Internet, distribuer et vendre des theses partout dans le loan, distribute and sell theses monde, a des fins commerciales ou autres, sur worldwide, for commercial or non support microforme, papier, electronique et/ou commercial purposes, in microform, autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in this et des droits moraux qui protege cette these. Ni thesis. Neither the thesis nor la these ni des extraits substantiels de celle-ci substantial extracts from it may be ne doivent etre imprimes ou autrement printed or otherwise reproduced reproduits sans son autorisation. without the author's permission.

In compliance with the Canadian Conformement a la loi canadienne sur la Privacy Act some supporting forms protection de la vie privee, quelques may have been removed from this formulaires secondaires ont ete enleves de thesis. cette these.

While these forms may be included Bien que ces formulaires aient inclus dans in the document page count, their la pagination, il n'y aura aucun contenu removal does not represent any loss manquant. of content from the thesis.

1+1 Canada ABSTRACT Invasion dynamics of the round goby (Neogobius meianostomus) in the Trent-Severn Waterway Jake Brownscombe

I studied the invasion dynamics of the round goby (Neogobius meianostomus) in the Trent-Severn Waterway, southeastern Ontario, Canada in 2009 and 2010. The angling removal method was used to sample the pioneering upstream and downstream edges of their range in addition to a longer established area. Edge sites increased drastically in site occupancy, abundance, and goby size over the first summer of occupation. The most rapid upstream expansion occurred during the non-reproductive season (9.1 km). Individuals sampled in their expanded range were small and female biased relative to other range locations. A dispersal model predicts that a detectable round goby population may travel as fast as 9.8 km/year upstream through the Waterway. Tethering experiments indicate that round goby predation rates are lower in a recently invaded area of the Trent River than a longer established area, which suggests that round gobies experience low predation risk during initial stages of invasion.

Keywords: colonization, demographics, diffusion model, dispersal, habitat, invasion dynamics, predation, range expansion, tethering

li Acknowledgements

I would like to extend my gratitude to everyone who provided me with assistance and guidance during my thesis work. To my supervisor, Michael Fox, who has helped me improve as a scientific researcher and writer. Thanks to my supervisory committee, Dave Beresford and Don Mackay for your help with the modeling portion of my thesis. To Lee Gutowsky, for orienting me in the field and helping develop my sampling regime. To my field assistants, Olivia Puckrin, Anna

Rooke, Greg Meisner, and Emily Fobert, for helping during long days of sampling. I also extend my gratitude to my girlfriend, Caitlin Higginson, and my entire family for support. Thanks to Chris Williams and Scott Reid for providing me with fish community data. Also to the Hastings Village Marina, and all the private land owners that allowed us access to sampling sites, and to conduct tethering experiments in the

Trent River in front of their property. Finally, thanks to NSERC for providing funding for this project.

iii Table of contents page ABSTRACT ii Acknowledgments iii List of figures vi List of tables ix Chapter 1: General Introduction 1 Chapter 2: Range expansion, population dynamics, demographics, and habitat use of the round goby (Neogobius meianostomus) in the Trent-Severn Waterway 7 INTRODUCTION 7 METHODS 10 Study location 10 Site selection 13 Sampling 14 Data analysis 15 RESULTS 17 Site occupancy and range expansion 17 Abundance (CPUE) 20 Length distribution 25 Sex ratios 28 Upstream dispersers 28 Habitat 31 DISCUSSION 33 Chapter 3: A diffusion model for upstream dispersal of the round goby in the Trent-Severn Waterway 51 INTRODUCTION 51 METHODS 54 Study area 54

iv Data collection 54 Model 55 RESULTS 58 DISCUSSION 64 Chapter 4: Predation rates of round goby in established and recently invaded areas of the Trent River 69 INTRODUCTION 69 METHODS 72 Study location 72 Site selection 72 Tethering experiment 74 Data analysis 77 RESULTS 78 Predation loss 78 Round goby predators 85 DISCUSSION 85 Chapter 5: General Discussion 98 Future work 102 References 106 Appendix A: Statistical analysis 128 Appendix B: Sample sizes 136

v List of figures page

Figure 1: Sampling sites for round gobies at the center and edges of their range in the Trent-Severn Waterway in May and August of 2009 and 2010 12 Figure 2: Round goby relative abundance (catch/m2) at center of range sites (25) near the point of introduction in the Trent River, sampled using the angling removal method; May and August of 2009 and 2010. For location refer to Figure 1 18 Figure 3: Round goby relative abundance (catch per m2) at sites across the upstream edge of their distribution in the Trent-Severn Waterway in May and August of 2009 and 2010. Star denotes the upstream edge. For location refer to Figure 1 19 Figure 4: Round goby relative abundance (catch per m2) at sites across the downstream edge of their distribution in the Trent-Severn Waterway in May and August of 2010. For location refer to Figure 1 21 Figure 5: Presence/absence (A) and relative abundance (catch per m2) (B) of round gobies at DSE with additional sites extending to the Bay of Quinte in May 2010 sampled using the angling removal method. For scale and location refer to Figure 1 22 Figure 6: Mean round goby catch (per m2) excluding non-catch sites, using the angling removal technique at sites in the upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (+ SE) 23 Figure 7: Mean total length (mm) of round gobies sampled using the angling removal technique at center of range (CORE), upstream edge (USE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (±SE) 26 Figure 8: Length frequency distribution of round gobies in Center of range (CORE), upstream edge (USE), and downstream edge (DSE) sampling sites in the Trent-Severn Waterway; May and August 2009 and 2010 27 Figure 9: Round goby sex ratio in angling sites in CORE, USE, and DSE range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE) 29 Figure 10: The Mean percentage of males and total length (TL) of round gobies in sites at CORE, USE, DSE, Prior USE in 2010 (occupied 2009)

vi and Expanded USE 2010 (beyond edge of range in 2009) range locations; May and August 2009 and 2010 (± SE) 30 Figure 11: Occupancy of round gobies in proportions of rock habitat at upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE) 32 Figure 12: Mean relative abundance (catch/m2, in catch sites) of round gobies in five habitat types (boulder, cobble, gravel, sand, mud) at upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (+ SE) 34 Figure 13: Probability distribution of the gamma function with varying combinations of a and 3 parameters 57 Figure 14: Estimated transit times (months/km) of round gobies from arrival at sites upstream from their edge of range in May 2009. Gamma parameters are a=1.08, p=1.67 60 Figure 15: Mean predicted arrival times (upper and lower 95% confidence limits) of detectable round gobies at locations upstream from their distribution in the Trent-Severn Waterway in May 2009 using a gamma transit time model 61 Figure 16: Probability of upstream movement of round gobies (km/year) from observed range expansion in the Trent-Severn Waterway from May 2009 to August 2010. Black hatched line represents 5 % probability 62 Figure 17: Probability of upstream movement of round gobies (km/year) from upstream model (P=1.67, solid grey line), and a hypothetical downstream model (P=0.84, hatched grey line). Black hatched line represents 5 % probability 63 Figure 18: round goby tethering sites in three habitat types: sandy shoal (SS), shallow rocky shoal (SRS), and a deep rocky shoal at CORE and DSE range locations in the Trent River 73 Figure 19: Tethering device (a) and depiction of a tethered round goby (b), from round goby tethering experiments in the Trent River in June and July of 2010 76 Figure 20: Control variables (round goby length (TL), tether depth, river surface velocity, temperature) (± SE) in round goby tethering experiments in in 3 habitat types (SRS, DRS, SS) at both the CORE (grey bars) and the DSE (white bars) range locations in the Trent River; June -July 2010 79

vn Figure 21: Comparison of mean predation rates of tethered round gobies (± SE) between sites in the area of original introduction (CORE,) and the downstream edge of range expansion (DSE) in three, 2-hour trials conducted in three habitat types in the Trent River; June - July 2010. DRS = deep rocky shoal, SRS = shallow rocky shoal, SS = sandy shoal 81 Figure 22: Relationship between round goby predation rate (%) and predator catch rate (% of predated gobies) from tethering trials at 3 sites at CORE (open circles) and DSE (closed diamonds) range locations of the round goby distribution in the Trent River in 2010 83 Figure 23: Round goby predation rates from tethering trials (black line) and relative abundance (catch per m2) at tethering sites (grey line) at CORE and DSE range locations in the Trent River in 2010. DRS = deep rocky shoal, SRS = shallow rocky shoal, SS = sandy shoal 84 Figure 24: Mean length (TL) of predated and survived round gobies (± SE) from tethering trials in CORE and DSE range locations in the Trent River; June - July 2010 86 Figure 25: Round goby predator species caught on tethered round gobies at CORE and DSE range locations in the Trent River; June - July 2010 87 Figure 26: Relationship between length of round gobies predated on tethers and that of predators captured on tethered round gobies in the Trent River; June-July, 2010. Sites within the two ranges locations are pooled 88

viii List of tables page

Table 1: Results from two-way factorial ANOVA comparing control variables between range locations (CORE and DSE) and habitat types (SRS, DRS, SS) in round goby tethering trials; June - July 2010 80 Table 2: Results from A) two-way factorial ANOVA and B) Newman- Keuls post-hoc analysis comparing predation rates between range locations (CORE, DSE) and habitat types (SRS, DRS, SS) from round goby tethering trials in June and July of 2010. Asterisks indicate p<0.05 82

IX Chapter 1: General Introduction

Invasions by non-indigenous species (NIS) have become more frequent since the 19th century due to human assisted transport, mainly due through international shipping, aquaculture, the pet trade, and intentional introductions (di Castri 1989;

Kraus et al. 1999; Mack et al 2000; Charlebois et al. 1997; Molnar et al 2008). Highly successful invasive species are one of the most prominent threats to global biodiversity due to their negative effects on native species, communities, and ecosystems (Simberloff 1996; Mack et al 2000; Cambray 2003). Their management has also become very economically expensive (Pimentel et al. 1999; Leung et al.

2002). While prevention is the most effective and ideal management strategy, understanding the biological characteristics of NIS can contribute to more efficient and effective management of those already established (Sakai et al. 2001; Simberloff

2003). Invasive species biology also contributes to a framework for predicting high- risk species and vulnerable ecosystems, which is critical for prevention strategies

(Kolar& Lodge 2002).

The Laurentian Great Lakes have been colonized by an estimated 200+ invasive species since the 19th century, mainly due to trans-oceanic ships and their connectivity to other watersheds by canals (Brammeier et al. 2008). Many of these species have become widespread, inducing negative ecological effects (Great Lakes

Commission 2007). Some of these species include the sea lamprey (Petromyzon marinus), the zebra mussel (Dreissena polymorpha), and the round goby (Neogobius meianostomus) (Brown 1989; Ricciardi 2001; Jude et al. 1992). The presence of 2 invasive species can actually facilitate the colonization of additional NIS (Simberloff

& Von Holle 1999; Ricciardi 2001), which combined with human traffic and watershed connectivity, makes the Great Lakes Watershed a hotspot for invasive species.

Gobiidae is the second largest family of teleost fishes in the world, with a widespread distribution in both fresh and saltwater (Robins et al. 1991). Many gobiid species have become invasive in nonindigenous areas throughout the world, causing negative ecological impacts in the United States, Canada, Austria, Australia, and Poland (Brittan et al. 1970; Middleton 1982; Jude et al. 1992; Skora & Stolarski

1993; Wiesner 2005). The round goby is one of these species, which is native to the

Ponto-Caspian region of Eastern Europe, and has become invasive in non-indigenous areas of Europe and in North America (Corkum et al. 2004). The round goby was first discovered in North America in the St. Clair River, Michigan in 1990 (Jude et al.

1992). It has subsequently spread throughout all five Great Lakes and many of their tributaries (Charlebois et al. 1997; Jude 2001; Phillips et al. 2003; Poos et al. 2009;

Raby et al. 2010). Its initial introduction was attributed to the release of infested ballast water from trans-oceanic ships (Hensler & Jude 2007).

The round goby is a relatively small species, reaching a maximum size of approximately 200 mm total length (TL) in North America, which is slightly smaller than in its native range (Charlebois et al. 1997). It is a benthic species that does not have a swim bladder, but has a characteristic fused pelvic fin, which forms a suctorial disk that allows it to remain on substrates in fast flowing water (Jude et al. 3

1992). The round goby shows some affinity for rocky substrates that provide refuge from predators (Ray & Corkum 2001; Eros et al. 2005), but has been found to occupy a wide variety of substrate types in invaded ecosystems (Jude 2001; Johnson et al.

2005a; Taraborelli et al. 2009; Borcherding et al. 2011). It preys upon a wide variety of organisms including invertebrates, crustaceans, fish eggs, and small fishes (Ray &

Corkum 1997; French & Jude 2001; Carman et al. 2006).

The success of the round goby as an invasive species can be attributed to a number of characteristics. It is robust, which is attested by its ability to survive a transoceanic voyage in ballast water (Jude et al. 1992). It has a wide tolerance to temperature, salinity, and oxygen levels (Skora & Stolarski 1993; Charlebois et al.

1997), and is capable of reproducing rapidly, with a very long reproductive season, females spawning multiple times each season, and males guarding nests to protect eggs (Miller 1984; Maclnnis & Corkum 2000; Jude 2001). There is also evidence of highly variable intrapopulation life history characteristics, with individuals at the edge of their population range growing faster, maturing earlier, and exhibiting higher reproductive output (L'avrincikova & Kovac 2007; Kovac et al. 2009;

Gutowsky & Fox in review).

Large round gobies have been found to consume zebra mussels as a large proportion of their diet (Ray & Corkum 1997), and evidence exists that they are reducing numbers of zebra mussels in well-established areas (Raby et al. 2010).

While this is a potential benefit of round goby presence, it is far outweighed by the negative ecological impacts of this invader. One of these impacts is because it preys 4 heavily on zebra mussels and other benthic organisms, which increases contaminant transfer to top predators (Kwon et al. 2006; Hogan et al. 2007). At the root of the round goby's negative ecological impacts is its ability to persist in very high densities, which have been reported up to 133/m2 (Chotkowski & Marsden 1999), but more common maximum estimated densities are 40 to 50/m2 (Charlebois et al.

1997; Jude 2001; Miner & Farver 2004). It has been shown to negatively impact native benthivores such as logperch and mottled sculpin, likely through domination of habitat and resources (Jude et al. 1995; Janssen & Jude 2001). The round goby also consumes the eggs and fry of larger fish species such as lake sturgeon

(Acipenserfulvescens), lake trout (Salvelinus namaycush), and smallmouth bass

(Micropterus dolomieui) (Chotkowski & Marsden 1999; Weimer & Sowinski 1999,

Steinhart et al. 2004a), which may have negative impacts on their populations.

The round goby has been a highly successful invasive species in European and North American freshwater ecosystems, inducing widespread, negative ecological effects. This has prompted researchers to study this species fairly extensively, defining a number of its biological characteristics, including density, habitat use, population and reproductive dynamics, life history traits, trophic effects, genetics, and behaviour. A review of round goby biology and research needs was published in 2001 (Charlebois et al. 2001), many of which have been addressed to some degree today. However there are still a number of characteristics that are poorly understood, many of which are related to dispersal mechanisms. 5

There are a number of characteristics that have been recognized as contributors to the success of the round goby in North American ecosystems, which are related to its ability to acquire habitat and resources in addition to rapid reproduction. Another important component for a successful invasion is often the absence of natural enemies (Colautti et al. 2004; Nentwig 2007). While a wide variety of predators including watersnakes, waterfowl, and many fish species have started feeding on round gobies within the past two decades (Somers et al. 2003;

Jakubas 2004; Johnson et al. 2005b; King et al. 2006; Bur et al. 2008; Campbell et al.

2009; Taraborelli et al. 2010), there is no evidence of the response of predators to the round goby during initial stages of invasion. Low predation risk has not yet been recognized as a contributing factor to the success of the round goby.

Given the highly invasive nature of the round goby, it is a novel species for understanding the biological characteristics of such species in temperate freshwater ecosystems. To contribute to round goby ecology and invasion biology my thesis examines a number of characteristics of the round goby population in the Trent-

Severn Waterway, southeastern Ontario. I addressed many questions regarding dispersal characteristics including: How much range expansion occurs during a year? What time of year does range expansion occur? What are the demographics of individuals at the pioneering edges of range expansion in comparison to an established area? What habitats do round gobies occupy at the very edges of their range? How fast do abundances increase in recently colonized areas? When will round gobies arrive at upstream locations in the Trent-Severn Waterway? How far may individuals disperse in a year? Is predation risk lower to round gobies during 6 initial stages of invasion than in established areas? The results of my research will contribute to the understanding of mechanisms behind the successful colonization and establishment of round gobies in North America, and their dispersal through the

Trent-Severn Waterway. These findings may contribute to more effective management of round goby populations, the prevention of their spread into new areas, and the framework for predicting high-risk invaders and vulnerable ecosystems.

My thesis is composed of five chapters. The objectives of Chapter 2 were to determine the rates of round goby range expansion, site occupancy, and population increase at the pioneering upstream and downstream edges of their range in the

Trent-Severn Waterway. I also compared the dynamics, demographics, and habitat use at range edges to the center of their range in the Trent-Severn Waterway. The objective of Chapter 3 was to develop a dispersal model for the upstream movement of round gobies to predict arrival times of a detectable population at various upstream locations, along with the potential for long distance dispersal events. The objective of Chapter 4 was to determine if predation risk was lower to round gobies occupying a recently invaded area than an established area in the Trent River. The final chapter is a summary of Chapters 2, 3, and 4 with suggestions for future research. 7

Chapter 2: Range expansion, population dynamics, demographics, and habitat use of

the round goby (Neogobius meianostomus) in the Trent-Severn Waterway

INTRODUCTION

Invasions by non-indigenous species (NIS) are generally described by four phases: introduction, establishment, spread, and impact (Vermeij 1996, Sakai et al.

2001; Jerde & Lewis 2007). The spread of NIS is an important aspect of invasions because the degree of their impact is a function of the spatial extent of their range

(Johnson & Padilla 1996; Nentwig 2007). Highly invasive NIS have a combination of characteristics that are advantageous to their success in new environments (Sakai et al. 2001; Olden et al. 2006). Many of these characteristics also contribute to range expansion, which can include aspects of population structure, behavior, life history traits, and habitat use (Mooney & Drake 1989; Shigesada & Kawasaki 1997; Ray &

Corkum 2001). Knowledge of these dispersal characteristics is important for controlling the spread of NIS, and predicting potential high risk species (Sakai et al.

2001; Kolar & Lodge 2002).

The round goby is native to the Ponto-Caspian region of Eastern Europe, and was first discovered in North America in the St. Clair River, Michigan in 1990 (Jude et al. 1992). It has spread rapidly throughout the Great Lakes Watershed, occupying all five Great Lakes by the year 2000, and now a number of their tributaries

(Charlebois et al. 1997; Jude 2001; Phillips et al. 2003; Poos et al. 2009; Raby et al.

2010). The spread of the round goby has been attributed to a combination of further ballast water introductions, bait bucket transfers, and natural population expansion (Charlebois et al. 1997). In the River Danube, Europe, Neogobius spp. have been discovered first in harbours, leading researchers to suspect that some species may be dispersing by attaching to the hulls of ships (hull transport) (Wiesner, 2005). The mechanism of 'natural' range expansion is likely related to its aggressive nature, where large round gobies force smaller individuals from preferred rocky substrates to less ideal habitats, from which they disperse (Ray & Corkum 2001; Johnson et al.

2005a).

Because the round goby is such a successful invader, the characteristics and mechanisms relating to its dispersal are of interest not only for round goby ecology, but for invasion biology as well. The round goby is known to exhibit a high reproductive output, especially at the edges of its range, where individuals mature earlier and allocate a greater amount of energy to reproduction (Gutowsky & Fox in review). Range edges also appear to be comprised of lower densities, a lower proportion of sites containing round gobies, larger individuals, and more male biased sex ratios than in longer established areas of their range in a fluvial ecosystem (Gutowsky et al. 2011; Gutowsky & Fox in press). Previous research has addressed range edges relative to a longer established area, but not the pioneering edges of population fronts. There remain questions about rates of round goby range expansion, the demographics of dispersing individuals, as well as site occupation, abundance and habitat use at pioneering population fronts in fluvial ecosystems.

Habitat can be an important factor limiting the spread of invasive species, one that may dictate their dispersal rate, distribution, and degree of ecosystem 9 disturbance (Lodge 1993; Johnson & Padilla 1996; Shigesada & Kawasaki 1997). The round goby is known to inhabit nearly all habitat types in invaded ecosystems including rock, sand, and mud substrates, and those with macrophyte cover, which is an attribute of its widespread ecological effects (Jude 2001; Johnson et al. 2005a;

Bergstrom et al. 2008; Taraborelli et al. 2009). However, habitat does affect round goby distribution, as it shows some affinity for rocky substrates, inhabits mud substrates less frequently, and wetlands appear most resistant to its invasion (Ray &

Corkum, 2001; Johnson et al. 2005a; Cooper et al. 2007; Young et al. 2010;

Borcherding et al. 2011). While these habitat associations are known, round goby habitat use at pioneering population fronts has not been addressed.

Sampling round goby densities and even relative abundance in all habitat types they occupy has proven difficult with traditional sampling gear. Passive gear

(traps, gill nets) exhibit low capture efficiencies, whereas active gear such as seine nets are limited by depth, and trawls are subject to habitat biases (Johnson et al.

2005a). Electrofishing is also relatively ineffective because the round goby lacks a swim bladder (Charlebois et al. 1997). Visual techniques have been developed for estimating round goby density by snorkeling transects (Ray & Corkum, 2001),

SCUBA diving (Schaner et al. 2009) using a remotely operated vehicle (Johnson et al.

2005a), or a boat mounted camera apparatus (Schaner et al. 2009). These techniques are useful under certain environmental conditions, but are limited by high cost or capture inefficiency in certain habitats occupied by the round goby

(Gutowsky et al. 2011). However, an angling removal method was recently developed specifically to sample the density of round gobies in complex substrates. 10

It is effective at estimating their densities in a range of habitats and depths, including complex rocky substrates that are most problematic for other sampling techniques (Gutowsky et al. 2011).

The objectives of this study were to determine rates of range expansion, population dynamics, demographics, and habitat use of the round goby at the pioneering upstream and downstream population fronts in a fluvial ecosystem. I also compared population dynamics, demographics, and habitat use between range edges and a longer established area. I hypothesized that range edges would be characterized by lower abundance, a lower proportion of occupied sites, larger individuals, and more male biased sex ratios, based on patterns observed in the

Trent-Severn Waterway in 2007/2008 (Gutowsky & Fox in press; Gutowsky & Fox in review). I also hypothesized that round gobies would occupy rocky substrates more often that softer substrates such as sand and mud (Ray and Corkum 2001,

Johnson et al. 2005a), and predicted that this habitat preference would be more pronounced at range edges than at the center.

METHODS

Study location

The Trent-Severn Waterway is a system of connected lakes, rivers, and canals in a 12,550 km2 watershed in Southeastern Ontario (Minns et al. 2004). It is a navigational waterway that contains many dams and locks, connecting Georgian

Bay, Lake Huron to the Bay of Quinte, Lake Ontario. Study locations within the 11

waterway were located in the Trent River, Rice Lake, and the Otonabee River (Fig.

1).

The source of the Trent River is at the eastern end of Rice Lake, stretching 90

km in length to the outlet in the Bay of Quinte (Fig. 1). In the area that center of

range (CORE) sites were sampled (44.28319 ° N, 78.01102° W to 44.32097° N,

77.94218° W) the Trent River has an mean channel depth of 4.4 m and mean width

of 150 m (Navionics 2010). In the area of downstream edge (DSE) sites, the Trent

River has a mean channel depth of 4.8 m and a mean width of 460 m (Navionics

2010). Rice Lake is the furthest downstream of the Kawartha Lakes, located between

the Otonabee River and the Trent River. It has a surface area of 10,018 ha, a mean

depth of 2.6 m, and a mean width of 3204 m (Lester et al. 2004). Upstream edge of

range (USE) sites were located throughout most of Rice Lake, from near the Indian

River, upstream to the Otonabee River (44.21999° N, 78.11313° W to 44.13430° N,

78.24733° W). The Otonabee River commences at the outlet of Katchawanooka Lake and flows 50 km to its mouth in Rice Lake. USE sampling sites were located in the

downstream 12.5 km of the Otonabee (44.20644° N, 78.27860° W to 44.15121° N,

78.22759° W). In this section, the mean channel depth is 5.4 m and the mean width

is 158 m (Navionics 2010).

The round goby was first reported in the Trent River in 2003 at a location

downstream of Lock 18 in the town of Hastings (44.31078° N, 77.952872° W) (Raby

et al. 2010) (Fig. 1). Its introduction was likely due to one or more bait bucket releases offish collected from infested waters. By 2008 the round goby had 44,32033 N

44,2268® H

44,13378 N

15 km , i 7S.13645*W 77,89088* W 77.6412CfW

VH^y^g Figure 1: Sampling sites (dark grey) for round gobies at the center and edges of £•2 their range in the Trent-Severn Waterway in May and August of 2009 and 2010 Lake Ontario 13

expanded its range upstream, with sightings 16 km up the Trent River and into Rice

Lake (44.24136° N, 78.11561° W) (Fig. 1) (OFAH, unpublished data). It had also spread downstream, with sightings 38 km down the Trent River into Percy's Reach

(44.23433° N, 77.79835° W) (Fig. 1). Differences in rates of upstream and

downstream spread are likely related to dispersal mechanisms. Juvenile round

gobies are known to perform vertical migration to the surface at night (Hensler and

Jude 2007; Hayden & Miner 2009). In fluvial systems such as the Trent River, juveniles in the water column would undoubtedly be carried downstream, which

may contribute to downstream expansion.

Site selection

Sampling sites were selected using a random point generator

(www.geomidpoint.com) in a stratified design. The majority of sites were generated

in areas that were predicted to contain the most suitable round goby habitat based

on waterbody characteristics. This was particularly important for locating round

gobies in Rice Lake, which is primarily comprised of mud substrates, but has rocky

shoals near shorelines. Seventy-five sites were sampled at both the upstream and

downstream edges of the round goby range in the Trent-Severn, including the areas

estimated to be their potential future range in 2010. Upstream edge sites were

selected from the most upstream sighting in Rice Lake in 2008,22 km throughout

Rice Lake and 8.5 km upstream in the Otonabee River (Fig. 1). For 2010 an additional 20 sampling sites were generated throughout the Otonabee River, and 4 km further upstream of the 2009 upstream edge sampling area to better 14 characterize round goby range expansion, which was more rapid than expected.

Downstream edge sites were located from 3.5 km upstream of the most downstream

2008 sighting, 13.6 km downstream in the Trent River (Fig. 1). An additional 10 sites were generated at the downstream edge for May 2010 sampling to confirm upstream movement from the Bay of Quinte round goby population in May 2010.

Twenty-five sites were also randomly generated near the center (CORE) of the round goby's range in the Trent River. Sites were located as far as 5.5 km upstream and 1.7 downstream from the point of first detection in the Trent River (44.28329°

N, 78.01055° W to 44.32095° N, 77.94219 ° W) (Fig. 1).

Sampling

Round goby relative abundance (catch/m2) was assessed in May and August of 2009 and 2010 using the angling removal method described in Gutowsky et al.

(2011). This technique involved two people angling within a 2 m2 floating barrier for

20 minutes at each site. The floating barrier was clamped to the side of a 16' Jon boat, which was anchored at both ends to remain on the site. Angling was conducted with micro-light rods and braided fishing line (0.15 mm diameter). Size 20 hooks were baited with a scented plastic maggot imitation, and small weights were attached to the line. All fish caught were identified and round gobies were measured for total length (mm), sexed, and retained until the end of the sampling period. Sex was determined by examining the urogenital papilla, which is triangular on males and rectangular on females (Charlebois et al. 1997). The time of catch was also recorded. All round gobies were returned to the site after sampling was completed. 15

The size of round gobies sampled with this method was restricted mainly to larger than 45 mm, so generally only age 1 and older fish were sampled, which represents the majority of the mature population (Maclnnis & Corkum 2000; Gutowsky et al.

2011; Gutowsky & Fox in review).

Sites were located using a Garmin Oregon 400t Global Positioning System unit (Garmin International Inc., Olathe, KS). Additional parameters measured at each site included substrate composition, depth, current, and temperature. Substrate was quantified at each site by visual estimation of the percentage of rock, sand, and mud within the 2 m2 sampling area using an Aquaview S-series colour underwater video camera (Nature Vision Inc., Brainerd, MN). The dominant rock size was also estimated as boulder (diameter >256 mm), cobble (64-256 mm), or gravel (2-64 mm) (Krumbein & Sloss 1951). Substrate was not sampled in May 2009. In all other sampling dates, substrate composition was visually estimated from roughly 0.5 m above the bottom. Depth was measured using an Eagle Cuda 168 depth sounder

(Eagle Electronics, Catoosa, OK). River velocity was measured by timing the horizontal movement of a float.

Data analysis

While the angling removal method provides the capability of estimating round goby densities (Gutowsky et al. 2011), density estimation would be less accurate at range edges where catches were low at many sites. Therefore catch per m2 was used compare round goby relative abundance between range locations. In any case where data could not be transformed to meet assumptions of normality the 16

appropriate non-parametric test was used (Zar, 2010). In order to examine round

goby abundance in recently invaded areas in comparison to a longer established

area, relative abundance was compared between range locations (USE, CORE, DSE)

and over time (May, August, 2009,2010) using a Nested ANOVA (time nested in

location) and Tukey Unequal N HSD post-hoc analysis. Abundance was rank transformed prior to analysis. Abundance was also compared between range

locations within each sampling date using Kruskal-Wallis tests.

A Nested ANOVA with Tukey Unequal N post-hoc analysis was also used to compare the sizes of round gobies between range locations and over time (time

nested in location). Round goby lengths were analyzed by calculating a mean length

for each site occupied by round gobies; these were rank-transformed prior to

analysis. Mean lengths were also compared between range locations within each

date using Kruskal-Wallis tests. To analyze sex ratios, a ratio was calculated for each

site occupied by round gobies. Sex ratios were compared between range locations

using a Nested ANOVA and Tukey Unequal N HSD post-hoc analysis.

In order to determine the demographics of dispersing individuals, mean

round goby total length (TL), and sex ratios (% male) were calculated for each sampling site and compared between range locations (CORE, DSE, USE), along with

Expanded USE range 2010 (sites occupied in 2010, upstream of August 2009 range

edge) and Prior USE range 2010 (sites occupied in May or August of 2009, resampled in 2010) using Kruskal-Wallis ANOVAs. To determine if round gobies

occupied higher quality habitats at the edges of their range than the center, chi- 17 square goodness of fit tests were used to compare the proportion of sites occupied by one or more round gobies across five proportions of rock substrate between range locations. To address the same question, Kruskal-Wallis tests were used to compare round goby abundance between four substrate types (boulder, cobble, gravel, sand) within each range location. The level of significance for all tests was p<0.05.

RESULTS

Site occupancy and range expansion

Round gobies were present at the center of their range in the Trent-Severn in

84 to 88 % of central sites across the four sampling periods (Fig. 2). At the upstream edge of range in May 2009, round gobies were distributed widely in Rice Lake, but were limited to only a few rocky shoals (Fig. 3). They occupied 13% of the 75 upstream sampling sites, and were detected as far upstream as the mouth of the

Otonabee River. This indicates a 14.2 km upstream advance from 2008 sightings. By

August 2009 round gobies had spread to many sites within their detected upstream range and occupied 32% of sample sites. Very little upstream range expansion occurred during this time period (< 1 km). By May 2010 round gobies were detected in the Otonabee River 9.14 km upstream of the detected population front in August

2009, indicating rapid range expansion over the 8-month period (Fig. 3). They occupied 47% of the 95 sampled upstream sites. By August 2010, changes in round 18

A N

May 2009

August 2009

May 2010

August 2010

5 km

Figure 2: Round goby relative abundance (catch/m2) at center of range sites (n = 25) near the point of introduction in the Trent River, sampled using the angling removal method; May and August of 2009 and 2010. Black dots indicate unoccupied sample sites. For location refer to Figure 1. 19

May 2009

August 2009

May 2010

August 2010

Figure 3: Round goby relative abundance (catch per m2) at sites across the upstream edge of their distribution in the Trent-Severn Waterway in May and August of 2009 and 2010. Black dots indicate unoccupied sample sites. Star denotes the upstream edge. For location refer to Figure 1. 20 goby distribution were similar to those observed over the summer of 2009, with limited range expansion (1.13 km upstream from the May 2010 edge) and a higher site occupancy rate (57%).

At the downstream edge of the round goby's range in May 2009 they were detected as far downstream in the Trent River as Percy's Reach. They occupied 15% of 75 downstream sites and were mainly confined to a large rocky shoal near the start of the reach, with the exception of one individual caught 4 km downstream of any other catch site (Fig. 4). The advance was 5.2 km from previous downstream sightings in the summer of 2008. In May 2010 it appears that round gobies travelling upstream from the Bay of Quinte, Lake Ontario invaded the downstream sampling area (Fig. 5). Therefore the rate of downstream expansion of the Trent River population could not be estimated. In May 2010, round gobies occupied 44% of downstream sampling sites, which increased to 57% in August of 2010.

Abundance fCPUEl

In center of range sites mean round goby catch was relatively consistent over time in comparison with the upstream and downstream edges. It was lowest in May

2009 (3.9/m2), and increased to its maximum in August 2009 (5.8/m2). Mean catch dropped to 5.1/m2 by May 2010, and again to 4.4/m2 by August 2010. At the upstream edge of range in May 2009 round goby relative abundance was very low, with a mean catch (in catch sites) of 1.7/m2 and a maximum of 4/m2 (Figs. 3 and 6).

In August 2009, the mean round goby catch increased to 7.8/m2 with a maximum catch of 17/m2. At the downstream edge of range in May 2009 relative abundance 21

May 2009 -^ tv - >„ it 1\ J AS $r ; fljeu~-- J"

^ August 2009 *V X_, XX. "* or

~~May 2010 \~~

** .x~ A ^ <• i X /* pJX-~er:%.

~~^ August 2010~~

~~5 km~~

Figure 4: Round goby relative abundance (catch per m2) at sites across the downstream edge of their distribution in the Trent-Severn Waterway in May and August of 2010. For location refer to Figure 1. 22

~~• S^f \^-<" K N '"_"" j • "-x- - * C- c / • ° . y' &•' ^~* ^ ''A /'•' -'' 3 :Xi~~

~~1ft ' '/" \ J,- '^ i/ ©/, ' \ -1 \ ; u, ;/ \ i \\ \\~~

~~1/ )1~~

~~0 5 km~~

Figure 5: [A] Presence/absence (large and small black dots, respectively) and (B) relative abundance (catch per m2) of round gobies at DSE with additional sites extending to the Bay of Quinte in May 2010 sampled using the angling removal method. For scale and location refer to Figure 1. 23

~~I May 2009 • August 2009 • May 2010 • August 2010~~

~~USE CORE DSE~~

~~Range location~~

Figure 6: Mean round goby catch (per m2) excluding non-catch sites, using the angling removal technique at sites in the upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE). 24 was also very low, with a mean catch of 1.5/m2 (in catch sites), and a maximum catch of 3/m2 (Figs. 4 and 6). By August mean catch doubled to 3/m2, and the maximum catch was 8/m2.

As hypothesized, mean relative abundance was very low in both upstream and downstream edge sites in May 2009 (Fig. 6). Abundance varied significantly between range locations in May 2009 (Kruskal-Wallis test, H2,46=11.0, p=0.004). It was significantly lower at both the upstream and downstream edges than at the center of range (non-parametric post-hoc analysis, p=0.03, p=0.01, respectively). For the rest of the sampling period, abundance was higher at upstream edge sites than both center of range and downstream edge sites, and downstream edge sites were lower than center of range sites (Fig. 6). In August 2009 downstream edge sites had significantly lower abundance than both center of range and upstream edge range locations (non-parametric post-hoc analysis, p=0.02, p<0.001, respectively). In both

~~May and August of 2010 there was no significant differences in round goby abundance between range locations (Kruskal-Wallis tests, H2,ios=3.36, p=0.19,~~

~~H2,ii5=2.0, p=0.37, respectively). Overall there was a significant difference in round goby abundance among range locations (Nested ANOVA, F2,33i=9.4, p<0.001).~~

~~Downstream edge sites were significantly lower than both center of range and upstream edge range locations (Unequal N HSD, p=0.002, p=0.01). 25~~

~~Length distribution~~

~~In center of range sites the mean TL of round gobies caught was highest in~~

May 2009 (89 mm ± 2.3 SE, Fig. 7). Mean TL decreased in every subsequent sampling period, which reflects a decrease in 80+ mm, and an increase in <70 mm round gobies (Fig. 8). In upstream edge sites in May 2009, the mean TL of round gobies was 79 mm ± 2.6 SE (Fig. 7). Mean TL increased dramatically by August 2009

~~(93 ± 1.7), decreased in May 2010 (76 mm ± 1.1) and remained stable through~~

August 2010 (76 ± 1.2). The peak in mean TL of round gobies in August 2009 was due to the low proportion of <70 mm fish and high proportion of 80 + mm, particularly 100 + mm fish (Fig. 8). A very similar trend was observed in downstream edge sites, where in May 2009 the mean TL of round gobies was 77 mm

~~± 2.8 SE, which increased in August 2009 (87 ± 1.8), decreased in May 2010 (77 ±~~

~~1.2) and remained stable in August 2010 (78 ± 1.2; Fig.7). This peak in mean length was marked by high numbers of in 90+ mm fish (Fig. 8).~~

Overall there was no significant difference in the mean length of round gobies among range locations (Nested ANOVA, F2,322=1.4, p=0.24). In May 2009 there was a significant difference in mean lengths between range locations (Kruskal-Wallis test,

H2,42=10.7, p=0.005). They were significantly larger at the center of their range than at both edge range locations (non-parametric post-hoc analysis, p<0.04 in both cases). There was also a significant difference in mean length among range locations in August 2009 (Kruskal-Wallis test, H2,75=13.9, p=0.001). Upstream edge sites were significantly higher than both center of range and downstream edge sites (non- 26

~~I May 2009 • August 2009 • May 2010 • August 2010 100~~

~~E 85 E H 80~~

~~3 75~~

~~60 USE CORE DSE~~

~~Range location~~

Figure 7: Mean total length (mm) of round gobies sampled using the angling removal technique at center of range (CORE), upstream edge (USE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE). 27

USE CORE DSE 35 30 25 May 20 fl 15 2009 10 s 0 35 30 25 Aug 20 2009 15 9) SO a. E 5 ri n (/• 0 o 35 3? 30 25 May 20 2010 15 "? I 10 5 I , n 0 I t 35 30 25 Aug 20 2010 15 10 F*1 5 0 n LJ 70 80 90 MOO 100

~~Size class (mm)~~

Figure 8: Length frequency distribution of round gobies in center of range (CORE), upstream edge (USE), and downstream edge (DSE) sampling sites in the Trent-Severn Waterway; May and August 2009 and 2010. 28 parametric post-hoc analysis, p<0.001, p=0.047, respectively). There was a significant difference in mean lengths in May 2010 as well (Kruskall-Wallis test,

H2,io4=7.3, p=0.03). Center of range sites were significantly higher than both range edges (non-parametric post-hoc analysis, p<0.05 in both cases). There was no significant differences in mean lengths between range locations in August of 2010

~~(Kruskall-Wallis test, H2,io4=1.3, p=0.53).~~

~~Sex ratios~~

The mean sex ratio of round gobies caught at angling sites followed a similar pattern over time at upstream edge and CORE range locations, which were inconsistent with the downstream edge (Fig. 9). In central sites the sex ratio was the most heavily male biased (63% vs 53 and 57%, in upstream and downstream edges, respectively). There was a significant difference in sex ratios among range locations

~~(Nested ANOVA, F2,282=3.4, p=0.04). Upstream edge sites had a significantly lower proportion of males than central sites (Unequal N HSD, p=0.006).~~

~~Upstream dispersers~~

The round gobies that occupied expanded range sites in 2010, upstream of their detected range edge in 2009 were relatively small (mean TL: 75 mm ± 1.1 SE) in comparison to those occupying other range locations (Fig. 10). There was a significant difference in TL between the expanded upstream edge, upstream edge sites occupied in the previous year (prior upstream edge, resampled in 2010), upstream edge overall, the center of range, and the downstream edge (Kruskal-

~~Wallis test, H4,43i=21.3, p<0.001). Round gobies occupying the expanded upstream 29~~

~~• USE~~

~~May 2009 Aug 2009 May 2010 Aug 2010~~

~~^CORE~~

~~May 2009 Aug 2009 May 2010 Aug 2010~~

~~A DSE~~

~~^# $$•$ M^* *^| ** -^ <&$$• <8$ w&"~~

~~May 2009 Aug 2009 May 2010 Aug 2010~~

~~Date Sampled~~

~~Figure 9: Round goby sex ratio in angling sites in CORE (grey), USE (black), and DSE (white) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE). 30~~

~~Males (%) Mean length (TL) 100 90 90 - 85 80 80 70 2 0i 60 75 3 "i"' "I 50 3 70 3 40 - I" 65 30~~

~~20 i S 60 Expanded Prior USE USE CORE DSE USE 2010 2010~~

~~Range location~~

Figure 10: The mean percentage of males and total length (TL, males and females) of round gobies at CORE, USE, DSE (all dates included), Prior USE (occupied in 2009, resampled in 2010) and Expanded USE 2010 (unoccupied in 2009) range locations; May and August of 2009 and 2010 (±SE) 31 range were smaller than those from the upstream edge, downstream edge, and center of range overall (non-parametric post-hoc analysis, p<0.04 in all cases).

~~However these gobies were not significantly different in size from sites occupied in the previous year in 2010 (p=1.0). These individuals were also female biased (58%)~~

~~(Fig. 10). This can be compared to sites in their prior range in 2010 (occupied in~~

~~2009) at 58% males, which is more similar to downstream edge and center of range sites. There was a significant difference in sex ratio among range locations (One-way~~

ANOVA, F4,379=6.2, p<0.001). There was a significantly lower proportion of males in expanded upstream edge sites in 2010 than in center sites overall, downstream sites overall, and previously occupied upstream edge sites in 2010 (Unequal N HSD, p<0.03 in all cases). Overall, individuals that occupied their expanded upstream range were relatively small and female biased. However a few very large females were sampled in the expanded range in 2010, and the largest female caught in this study (128 mm TL) was located at the most upstream site in May 2010 at the very edge of the detected population front.

~~Habitat~~

~~Round gobies occupied between 88 and 100% of sites with 40% or more rock substrate in all range locations in sampled sites within their detected range (Fig. 11).~~

In sites with 79% or less rock substrate, both range edge locations had lower site occupancy than central sites. There was a significant difference in site occupancy between both edge range locations and the center of range (/2>18.5, p<0.001, df=4, 32

~~I USE OCORE ODSE~~

~~0-19 20-39 40-59 60-79 80-100 Rock habitat (%)~~

Figure 11: Occupancy of round gobies by proportion of rock substrate at upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE). 33 in both cases), and the largest differences occurred in sites with 0 to 19% rock substrate.

Round goby abundance in occupied sites was higher in all three types of rock substrate (boulder, cobble, gravel) than in sand or mud substrates in all three range locations (Fig. 12). Both edge range locations had the highest abundance in gravel dominated substrate, while at the center of range, abundance was highest in boulder substrate. In center of range sites there was no significant difference in round goby abundance between substrate types (Kruskal-Wallis test, H3,62=2.5, p=0.47). Round goby abundance varied significantly between substrate types in upstream edge sites

(Kruskal-Wallis test, H3,i25=48.3, p<0.001). All three types of rock substrate had significantly higher abundance than sand substrate (non-parametric post-hoc analysis, p<0.001 in all cases). Abundance also varied significantly between substrate types in downstream edge sites (Kruskal-Wallis test, H3,io7=18.3, p<0.001).

~~Cobble substrate had significantly higher round goby abundance than sand (non- parametric post-hoc analysis, p=0.002).~~

~~DISCUSSION~~

~~Range edge population dynamics~~

As hypothesized, both the upstream and downstream edges of the round goby's range in the Trent-Severn Waterway were characterized by low site occupancy and low relative abundance during the spring of the first year of 34

~~I USE D CORE DDSE 14~~

~~IN E 12 a 10 re * 8 uai re •o 6 c 3 .Q re 4 v oo re 2 <~~

~~boulder cobble gravel sand mud~~

~~Habitat type~~

Figure 12: Mean relative abundance (catch/m2 in occupied sites) of round gobies in five substrate types (boulder, cobble, gravel, sand, mud) at upstream edge (USE), center of range (CORE), and downstream edge (DSE) range locations in the Trent-Severn Waterway; May and August 2009 and 2010 (± SE). 35 occupation in comparison to sites at the center of their range. Lower abundance in more recently invaded areas has been observed previously in the Trent-Severn round goby population (Raby et al. 2010; Gutowsky et al. 2011), as well as in Lake

~~Ontario, Lake Michigan, and Lake Huron (Bergstrom et al. 2008; Lederer et al. 2008;~~

~~Taraborelli etal. 2009).~~

Over the first summer of occupation, major increases in site occupancy rate and relative abundance of occupied sites were observed in both the upstream and downstream edges of their range. This was particularly true for the upstream edge, where mean CPUE increased almost four fold, peaking in August 2009 at the highest observed in this study at any range location. CPUE doubled at downstream edge sites during the same time period, but did not peak until May 2010. This late peak can be attributed to the coalescence of the Trent-Severn population with the upstream expansion of the Bay of Quinte population. Evidence for this coalescence comes from the sudden, consistent presence, and high abundance of round gobies from the edge of the sample area to the Bay of Quinte (Fig. 5). Such sudden increases in site occupancy and abundance are inconsistent with that typically observed in recently invaded areas, especially at the downstream edge of their range (Fig. 4). If this range expansion was occurring from upstream in the Trent-Severn, it would be expected that a gradient from high to low abundance would be observed at the range edge, as observed at both upstream and downstream edges of their range in this study (Figs. 3 and 4). This was not the case from the downstream edge to the

~~Bay of Quinte (Fig. 5). 36~~

~~By 2010 upstream edge sites decreased significantly in relative abundance, and throughout 2010, all three range locations had similar mean abundance (Fig. 6).~~

~~While recently invaded areas are generally considered to have lower abundance than longer established areas (Raby et al. 2010), during this study period in the~~

Trent-Severn this phenomenon was very brief, and round goby abundance at the edges of its range became fairly homogeneous with a well-established area just one year after first detection (Fig. 6). The rapid increases in abundance observed at range edges were likely attributed to high resource availability due to the low initial goby density, which results in high growth rates, early maturity, and high reproductive output (Raby et al. 2010; Gutowsky & Fox in review). Survival may also be high at range edges due to low predation rates (Chapter 4).

A previous study in Lake Michigan showed that in recently invaded areas, round goby abundance increased rapidly on an annual basis (Clapp et al. 2001). By sampling in both late spring and summer in the Trent-Severn, I have shown that round goby abundance can increase rapidly during the summer of first detection in new areas of this system. The high abundance observed in August 2009 at the upstream edge of their range, and subsequent lower abundance in 2010, similar to a well established area in the system, is suggestive of a peak in abundance during the first summer of occupation in a new area. This peak occurred much faster than in previous examples in the Great Lakes. In Eastern Lake Michigan, round gobies were first detected in 1997 at a very low relative abundance of 0.5 per trawl hour (Clapp et al. 2001). An increase in abundance was observed by 1998 (3.0 and 1.5 per trawl hour) and a large increase in 1999, two years post detection (69 and 38 per trawl 37 hour). In Duluth Harbour, Lake Superior, round gobies were first discovered in

~~1998, and abundance remained very low until 2003 (Bergstrom et al. 2008). A large increase in abundance was observed in 2004, which was six years post detection.~~

~~There are likely differences in biotic and abiotic conditions between the Trent-~~

Severn, Lake Michigan, and Lake Superior, which may have resulted in differences in the rate of round goby population growth. However these differences may also be related to sampling techniques. The previously mentioned studies sampled with bottom trawls, which are often limited to less complex substrates (Johnson et al.

2005a; Bergstrom et al. 2008). These substrates are associated with lower abundance, smaller individuals and a high proportion of females, indicating they are less favourable habitats (Charlebois et al. 1997; Ray & Corkum 2001; Johnson et al.

2005a; Bergstrom et al. 2008). During initial stages of invasion round gobies appear to occupy mainly high quality habitats, and over time individuals occupy habitats of lower quality (Figs. 11 and 12). Differences in sampling techniques may help explain the later abundance increases observed in Lake Michigan and Lake Superior, as trawling targets less ideal habitats, which probably have more delayed increases in abundance than the ideal, complex, rocky habitats sampled in this study using the angling removal method. By sampling with angling in late spring and summer, I have shown that round goby abundance can increase in ideal habitats of recently invaded areas much faster than previously known. 38

~~Range expansion~~

~~Range expansion occurred in similar patterns at both upstream and downstream edges of the round goby's range in the Trent-Severn in 2009 and 2010.~~

~~Little range expansion occurred the primary reproductive seasons (May to August), while large-scale migration occurred during the non-reproductive seasons~~

(September to April). Unfortunately, downstream movement beyond 2009 could not be estimated due to the convergence of the Trent-Severn population with the upstream expanding Bay of Quinte population (Fig. 5).

In May 2009 round gobies had spread 14.2 km upstream and 5.2 km downstream from the area they were known to occupy in 2008. This trend is reversed from typical historical range expansion in the Trent River, where downstream spread had been twice the rate of upstream spread (see Raby et al.

2010). This may be explained by the history of the round goby population in the system. Historically the downstream edge has been characterized by more rapid range expansion and lower abundance than the upstream edge (Raby et al. 2010;

Gutowsky & Fox in press), which could be due to passive movement in river currents by juveniles that migrate to the surface at night (Hensler & Jude 2007). At the downstream edge prior to 2009, range expansion was occurring through a part of the Trent River that contains many high velocity riffle sections above Percy's Reach

~~(Fig. 1). Individuals could have been swept through these sections in the high velocity current, resulting in rapid expansion from 2007 to 2008 (14.2 km)(Fig. 1).~~

~~This may have resulted in a relatively small founder population in Percy's Reach, 39 which may have had low density-dependent pressure for further range expansion.~~

~~This would explain the very low level of downstream expansion by May 2009 (Fig.~~

~~4), and the lower abundance in downstream edge sites than the other two range locations throughout the sampling period (Fig. 6).~~

Very little range expansion occurred from May to August of 2009 at the downstream (< 0.5 km) and upstream edges (< 1 km), or between May and August of 2010 at the upstream edge (*1.1 km). Therefore, it appears that no large-scale upstream movement of adults occurred during the primary reproductive season.

This is not surprising, considering the reproductive characteristics of the round goby. They exhibit high reproductive energy allocation; especially at the edges of their range, a long reproductive season, multiple spawning by females, and nest guarding by males (Miller 1984; Maclnnis & Corkum 2000; Jude 2001; Gutowsky &

Fox in review). It appears that round gobies allocate energy to reproduction more than movement during the late spring/summer months. Some smaller scale movement did occur during these months into sites within their range. These movements may have been density driven by large, aggressive gobies forcing smaller individuals into new areas (Ray & Corkum 2001; Johnson et al. 2005a).

~~In May 2009 round gobies were detected «14 km upstream of 2008 sightings, indicating rapid upstream movement. The same phenomenon was observed by May~~

~~2010, where round gobies had advanced 9.14 km further upstream into the~~

Otonabee River from their detected range in August 2009. Round gobies are known to spread by bait bucket transfer in North America (Raby et al. 2010), and in Europe 40 they have been suspected to spread by attaching to the hulls of vessels (hull transport) (Tsepkin et al. 1992; Sokolov et al. 1994; Weisner 2005). However these types of introductions often result in large gaps between infested zones (Weisner

2005; Raby et al. 2010). The consistent presence and abundance of round gobies from their range edge in August of 2009, throughout the Otonabee River to their expanded range edge in May 2010 is evidence that their dispersal in this study was primarily natural.

The upstream range expansion observed from 2009 to 2010 was rapid considering the low sustained swimming speeds of the round goby (Skora 1996), although it was considerably lower than that observed from 2008 to 2009. The difference in rates of range expansion observed may be related to habitat. From

~~2008 to 2009 round gobies were invading a lentic environment, Rice Lake, while from 2009 to 2010 they were invading a lotic environment, the Otonabee River.~~

~~Upstream expansion into the Otonabee River was likely more energetically costly than movement into Rice Lake.~~

~~The upstream range expansion of the round goby observed from 2008 to~~

2010 was much more rapid than that observed from 2003 to 2008, which suggests that the rate of range expansion is increasing in this system (Figs. 1 and 3). A similar pattern has been observed previously with round gobies in the Illinois Waterway

(Irons et al. 2006). Round goby range expansion is thought to be highly density driven, with large round gobies forcing smaller individuals into previously unoccupied areas (Ray & Corkum 2001, Johnson et al. 2005a). As the round goby 41 population size becomes larger in the Trent-Severn, there are likely a greater number of individuals making large-scale upstream movements that contribute to range expansion.

~~The rates of spread observed in the Trent-Severn Waterway from 2008 to~~

2010 were rapid, but not extraordinary for the round goby. An average of 25 km/year of range expansion was previously observed in round goby moving through Chicago inland waterways from Lake Michigan to the Mississippi River

~~(Steingraeber and Thiel 2000). Round gobies have been collected up to 120 km further downstream of where they were detected the year before in the Illinois~~

~~Waterway (Irons et al. 2006). However, the only previously documented rates of upstream range expansion were much slower, with a maximum 1 km/year from~~

~~1998 to 2003 through Duluth Harbour, Lake Superior, into the St. Louis River~~

(Bergstrom et al. 2008). This is a similar pattern to that of the slow upstream spread of the Trent-Severn population during the first four years of invasion. There are likely a number of differences between the conditions of the Trent-Severn and

~~Duluth Harbour/the St. Louis River, and factors that affect round goby range expansion should be investigated.~~

~~Seasonal migration~~

While round gobies are considered relatively immobile, and often have very small home range sizes (Skora 1996; Wolf & Marsden 1998; Ray & Corkum 2001), it is evident from their movements in the Trent-Severn that they incurred a large-scale seasonal migration upstream into the Otonabee River sometime between autumn 42

~~2009 and early spring 2010. During the winter months, temperatures are very low in Ontario, and at these temperatures round gobies are relatively inactive (Lee &~~

Johnson 2005), so it is unlikely that large-scale movements were occurring during this season. Spring and autumn are more likely the potential seasons for round goby migration. Round gobies are known to migrate into deeper overwintering habitats in the fall, and return to shallow areas in the spring (Knight 1997; Pennuto et al. 2010).

This migration appears to be the primary mechanism by which the round goby expands its range during the early spring and/or fall seasons, while little large scale migration occurs during the late spring/summer. However, it is interesting that the large increases in abundance in edge sites during the summer of 2009 were accompanied by a peak in mean body size, where a large proportion of the sampled population was age 2 or 3 based on length at age data from this system (Gutowsky &

Fox in review). If no migration was occurring during this time period, it would be expected that an increase in abundance would be accompanied by smaller body sizes, as young of year and age 1 fish grew large enough to become incorporated into my size selective sampling method (45+ mm). Because there was a high proportion of large round goby in the samples in August 2009, individuals were likely migrating into edge range locations from closer to the population core. Therefore large scale migrations may have been occurring during the late spring/summer months, but none were detected beyond their population front. 43

~~Center of range: well established population dynamics~~

The round gobies occupying the center of their range showed some indications of a population decline despite the fact that abundance did not vary significantly over time. Over the reproductive season of 2009 mean round goby abundance increased by nearly 50% in center of range sites. An increase in abundance is expected in a stable population during the summer months due to growth and reproduction. Individuals born within the reproductive season may not be included, but summer growth would be expected to recruit many age 1 fish into the sample (Gutowsky & Fox in review). A subsequent 12% decline in abundance was observed by May 2010. This would also be expected, as most temperate fishes exhibit low levels of growth and some mortality over the winter season (Wootton

1998). However, by August 2010 abundance did not rebound as expected, but decreased by an additional 14%. Also, the mean length of CORE gobies declined consistently throughout the sampling period (17% decrease from May 2009 to

~~August 2010). It seems a loss of large round gobies in 2009 preluded a decline in abundance in 2010. A similar trend was observed in Hamilton Harbour, Lake~~

~~Ontario, where a 72% decrease in round goby abundance coincided with a 10% decrease in body length from 2002 to 2008, less than a decade from its detection in~~

Lake Ontario (Mills 2003; Young et al. 2010). In Lake Erie, round goby abundance peaked in 1999, five years after its first detection (Johnson et al. 2005b; Bunnell et al. 2005). Invasive species in general are often characterized by an initial rapid increase in population size, followed by a decline to lower abundance (Parker et al. 44

~~1999; Arin et al. 2005). The round goby may be showing early signs of this decline roughly seven years after its detection in the Trent-Severn.~~

~~Range location demographics~~

~~Round gobies occupying the Trent-Severn Waterway were male biased at all range locations. Round goby populations are commonly male-biased in North~~

~~American ecosystems and in their native range (Corkum et al. 2004; Young et al.~~

2010; Gutowsky & Fox in review). In my study there were significant differences in the sex ratio of round gobies across range locations. Center of range sites were the most heavily male biased (1.7:1), while both edges had a more even distribution of sexes (1.1:1 at the upstream edge and 1.2:1 at the downstream edge). While this finding appears to be inconsistent with that of a previous round goby study in the

Trent-Severn Waterway, in which range edges were more male biased, the difference may be explained by much broader definition of the edge of range in that study. Gutowsky & Fox (in press) defined edge range locations as the most upstream and downstream areas where a moderate abundance of round gobies were located, while the current study examined round gobies at the pioneering edge, in an area invaded less than one year prior to the initiation of the study, which was at very low abundance.

~~The lengths of round gobies sampled in the Trent-Severn varied from 43 to~~

135 mm TL with a mean of 79 mm. These sizes are similar to those found throughout North American ecosystems, which are generally smaller than those found in their native range (Charlebois et al. 1997; Taraborelli et al. 2010; Gutowsky 45

& Fox in press). Contrary to our hypothesis that edges of range would consist of larger individuals based on previous research in the Trent-Severn, range edges consisted of slightly smaller individuals. This again probably relates to varying definitions of range edge from Gutowsky & Fox (in press). Smaller individuals at pioneering range edges may indicate the more mobile members of the population, as previously speculated by Ray & Corkum (2001), Johnson et al. (2005b), and

Bergstrom et al. (2008). Within range locations there was high variability in round goby body size over time. Round gobies were significantly larger in August 2009 than other sample dates at both edge range locations. This was likely due to high resource availability when round goby density was low, contributing to high growth rates at the edge of their range (Raby et al. 2010; Gutowsky & Fox in press) during the first summer of occupation.

~~Demographics of dispersers~~

Round gobies that moved into the Otonabee River prior to May 2010 were female biased at (0.7 M:F), which suggests that a higher proportion of females conducted the large scale migration. Female round gobies are generally smaller than males (Gutowsky & Fox in press), and large round gobies are known to dominate the highest quality refuge, forcing smaller individuals into less complex substrates (Ray

~~& Corkum 2001; Johnson et al. 2005a; Stammler & Corkum 2005; Bergstrom et al.~~

2008). These smaller individuals have been suspected to disperse from these poorer quality habitats into previously unoccupied territory, contributing to range expansion (Ray & Corkum 2001; Johnson et al. 2005a; Bergstrom et al. 2008). 46

Therefore, due their small size, a greater proportion of females may be driven to make large scale movements that expand the round goby's range. Sex specific interactions are poorly understood for the round goby, so this explanation assumes that large male round gobies are aggressive towards small, immature females that do not contribute eggs to their nest.

Individuals that occupied the expanded upstream range area were significantly smaller than those of all three range locations with all sampling dates included, but they were not smaller than previously occupied upstream edge sites in

2010. While there may have been a regional or temporal effect, a greater proportion of small individuals likely conducted the large scale migration, contributing to range expansion. In a previous study, more recently colonized areas of Lake St. Clair and surrounding waterbodies were also comprised of smaller individuals (Ray &

Corkum 2001). As first speculated by Ray & Corkum (2001), the primary mechanism of round goby range expansion is likely intraspecific competition, which forces smaller individuals into less ideal habitats, from which they disperse. My findings are consistent with this mechanism, although large individuals were also found in the expanded upstream range area of the Trent-Severn, and large scale movements are not limited to small round gobies. The movement of these large individuals may be important for range expansion, as the probability of establishment in a new area is higher with adults present (Velez-Espino et al. 2010). 47

~~Habitat~~

As hypothesized, round gobies were more abundant on all types of rocky substrates (boulder, cobble, gravel) than on sand, and very few were caught on mud dominated substrate. Round gobies are known to have an affinity for rocky substrates because they provide complex benthic structure that they use for refuge

(Ray & Corkum 2001; Eros et al. 2005; Johnson et al. 2005a). This affinity was very pronounced in this study, which is likely because the sampling method targeted mainly mature individuals during the reproductive season, and the complex rocky habitats of the Trent-Severn provide ideal nesting sites in addition to refuge (Miller

1984; Maclnnis & Corkum 2000; Ray & Corkum 2001; Gutowsky et al. 2011). At both range edges round goby abundance was highest on gravel dominated substrate, while at the center of range it was highest on boulder. This difference may be related to predation risk, which appears higher in established areas than recently invaded areas of this system (Chapter 4). As predators adapt to feeding on round gobies and predation risk becomes higher, habitat preference may shift to more complex rocky substrates, which likely provide better refuge from predators.

As predicted, round gobies showed a more pronounced habitat preference at the edges of their range than at the center (Figs. 11 and 12). There was significantly higher round goby abundance on rock dominated substrate than sand at both edges of their range, while there was no significant difference between these habitat types at the center. At the edges of their range, gobies mainly occupied substrates with

40% or more rock cover, while at the center of range, gobies occupied substrates 48 with less than 40% rock more often than at the edges. This habitat selectivity reflects their distribution, which is more limited and predictable at range edges. This was especially evident during early stages of invasion in a new area (Figs. 3 and 4).

In a previous study, well-established areas of the Trent River have been found to have higher site occupancy than edge sites (Gutowsky & Fox in press), which appears to reflect habitat selectivity (Figs. 11 and 12). In longer established areas, round gobies are known to occupy more marginal habitats (Johnson et al.

~~2005a; Bergstrom et al. 2008). In a well established area of the Bay of Quinte, Lake~~

~~Ontario, round goby abundance was similar on rock, sand, and mud habitats~~

~~(Taraborelli et al. 2009). It appears that range expansion occurs by individuals moving first into rocky substrates, from which intraspecific competition forces gobies into neighbouring substrates.~~

~~Method limitations~~

The angling removal method was advantageous in that it allowed for standardized and efficient detection of round gobies at a wide range of habitats and depths. Although the major restriction of this method is that it generally samples only round gobies larger than 45 mm, thus limiting the sample to mainly mature individuals (Gutowsky et al. 2011, Gutowsky & Fox in review). Although age 0 individuals were either underrepresented or not sampled at all, the angling method was still very useful for this application because it monitored the spread of the reproducing population, which is of most concern. If a population of immature individuals were to inhabit an area without growing into mature fish to be detected 49 by this method, they may have very short-term ecological effects, but would not contribute to the spread of the population. The only potential negative of using this method to monitor round goby movements is if very small individuals are responsible for range expansion. This means that our detection of round goby movements may have been lagging behind the actual population front.

~~In conclusion, the round goby expanded its range both upstream and downstream in the Trent-Severn Waterway in 2009 and 2010 in similar patterns.~~

During early stages of invasion in May 2009, site occupancy and abundance were very low at both range edges. By August 2009, rapid increases in both abundance and occupancy were observed and abundance was even higher at the upstream edge than at the center of their range. A peak in mean body size at both range edges accompanied this increase in abundance in August 2009, possibly as a result of high resource availability in recently invaded areas. Little range expansion occurred during the late spring/summer seasons, while a large scale migration occurred from

Rice Lake, upstream into the Otonabee River during the fall/winter/early spring seasons. This migration probably occurred in the spring and/or fall, and is likely related to their migration into deeper overwintering habitats. Overall, edge range locations were comprised of smaller individuals and a lower proportion of males than a longer established area at the center of their range in the Trent-Severn.

Individuals that occupied their expanded upstream range in 2010 were female biased and relatively small, which is consistent with the hypothesis that smaller individuals are primarily responsible for range expansion. However large 50 individuals were also found in the expanded upstream range, so range expansion is not attributed solely to small individuals. Habitat preference was more pronounced at range edges, where round goby abundance was significantly higher on rock dominated substrates than sand, which was not the case at the center of their range.

Round gobies also mainly occupy substrates comprised of 40% or more rock cover, while those at the center of their range occupied substrates with less than 40% rock cover more often than at range edges. This habitat preference reflects the more limited distribution of round gobies during initial stages of invasion in a new area, and highlights importance of these habitats for range expansion, from which further colonization of lesser quality habitats occurs. 51

~~Chapter 3 - A diffusion model for upstream dispersal of the round goby in the~~

~~Trent-Severn Waterway~~

~~INTRODUCTION~~

~~Once a non-indigenous species (NIS) becomes established in a new region, range expansion often occurs as individuals disperse from the population core~~

(Shigesada & Kawasaki 1997; Sakai et al. 2001). This is an important aspect of NIS invasions because the spatial extent of their range influences the degree of their impact on recipient ecosystems (Johnson & Padilla; Nentwig 2007). The spread of

NIS is affected by environmental conditions and geographical limitations in relation to species characteristics (Shigesada & Kawasaki 1997). Models that describe the dispersal of organisms are useful for predicting the rate of spread of NIS through ecosystems, and the underlying mechanisms that affect range expansion (Shigesada

~~& Kawasaki 1997; Sharov & Liebhold 1998; Kinezaki et al. 2003).~~

There has been a number of diffusion models developed that vary in complexity to describe the movement of organisms. Some consider just dispersal, while others include population growth, population dynamics, a highly realistic spatial component, long distance dispersal and habitat heterogeneity as model . parameters (Skellam 1951; Matis et al. 1992b; Shigesada et al. 1995; Sharov &

~~Liebhold 1998; Schippers et al. 1996; Faugeras & Maury 1997; Bertolino et al. 2008).~~

More complex models provide a more detailed simulation of the dispersal of organisms and contributing ecological processes, but they can require a large quantity of information on population dynamics and/or ecosystem characteristics. 52

~~Complex models are not always the most useful, as the inclusion of additional model complexity may also contribute to model error (Bunnell 1989).~~

Matis et al. (1992b) developed a gamma transit time model that is useful for modeling the dispersal of organisms when their source, dispersal distances, and arrival times can be estimated. The gamma distribution provides a unimodal curve that can be manipulated to match actual dispersal transit times. This type of model can be used to predict arrival times of a dispersing population at locations of interest

(Matis et al. 1992b). Transit times may also be converted to a time transit distribution to predict dispersal distances per unit time. The tail of this distribution may be useful for predicting the probability of long distance dispersal events that are typically rare, but are often an important component of range expansion

~~(Trakhtenbrot et al. 2005).~~

The round goby (Neogobius meianostomus) has been highly successful in expanding its range in both North American and European ecosystems (Charlebois et al. 1997; Jude 2001; Phillips et al. 2003: Eros et al. 2005; Poos et al. 2009; Raby et al. 2010; Borcherding et al. 2011). Natural population expansion is thought to be driven by large round gobies forcing smaller individuals into less ideal habitats, from which they disperse (Ray & Corkum 2001), and range expansion is likely driven by both short and long distance dispersers (Bronnenhuber et al. 2011).

~~However, round goby spread has been attributed to a combination of further ballast water introductions, bait bucket transfers, and natural population expansion~~

(Charlebois et al. 1997). Therefore it is difficult to determine if observed range 53 expansion is natural, and previously observed rates of range expansion have been highly variable. In the Illinois Waterway, one individual round goby was found 120 km downstream of sightings the year before, although the source of this individual is unknown (Irons et al. 2006). Round goby range expansion was observed at a rate of

~~25 km/year through Chicago inland waterways from Lake Michigan to the~~

~~Mississippi River (Steingraeber & Thiel 2000). The only documented estimate of upstream expansion rate was from Duluth Harbour, Lake Superior into the St. Louis~~

~~River, where a maximum of 1.0 km/year of range expansion was observed~~

~~(Bergstrom et al. 2008). The maximum dispersal distance observed of an individual round goby was 2 km in 218 days from a tagging study (Wolfe & Marsden 1998).~~

The dearth of studies available emphasize that more needs to be known about natural rates of range expansion, and the dispersal capabilities of individual round gobies before we can generalize about the range expansion capabilities of this species.

The Trent-Severn Waterway in central Ontario, Canada includes many lakes and rivers that are the source of highly valued sport fisheries. Round gobies have been expanding their range both upstream and downstream in the Trent-Severn since their introduction to the Trent River in 2003, and by 2008 they were detected up to 16 km upstream into Rice Lake (44.24136° N, 78.11561° W) (Chapter 2). Most of the Trent-Severn is still free of round gobies, but upstream expansion from their source in the Trent River threatens complete infestation. The objective of this research is to develop a diffusion model for the upstream dispersal of round gobies in the Trent-Severn Waterway, and use this model to predict the arrival times of a 54

~~detectable round goby population at various upstream locations, as well as predict the scale and probability of dispersal events.~~

~~METHODS~~

~~Study area~~

~~Bay, Lake Huron to the Bay of Quinte, Lake Ontario. Study locations were located in~~

~~Rice Lake and the Otonabee River (Chapter 2). Sampling sites were located from near their range edge in 2008, 25 km upstream throughout Rice Lake and the lower reach of the Otonabee River (Chapter 2).~~

~~Data Collection~~

~~Round goby distribution was sampled using the angling removal method in~~

May and August of 2009 and 2010 at 75 sites at the upstream edge of their range, including their potential future range in 2010 (Chapter 2). An additional 20 sites were sampled in the Otonabee River in 2010 because range expansion was more rapid than expected. Sites were sampled an additional 4 km upstream to include multiple patches of high quality habitat and further detect range expansion. Sites were randomly selected using a random point generator (geomidpoint.com). Site selection was stratified, generated primarily where rocky substrates were found, 55 which are the most ideal round goby habitats (Charlebois et al. 1997; Johnson et al.

~~2005a; Chapter 2).~~

~~Transit time model~~

~~There were 35 sampling sites that became occupied by round gobies in~~

~~August 2009, May 2010, or August 2010 that were previously unoccupied and upstream of their defined range edge in May of 2009 (Chapter 2). A transit time~~

~~(months/km) was calculated for each of these sites based on their distance from the upstream edge site in May 2009, and the date of their detection.~~

A model of round goby transit times was constructed using a gamma transit time distribution. The model was applied to transit times, apposed to dispersal distances because the data consist of observations of a leading front, instead of movement from a fixed location (Matis et al. 1992b). The gamma distribution was used instead of a simpler distribution such as the Gaussian because it provides a more adequate fit to a wider range of dispersal data (Taylor 1978; Matis et al.

~~1992a). The gamma distribution is also highly tractable, and confidence intervals for arrival times can be calculated from the chi-square distribution (Matis et al. 1992b).~~

~~To produce the model, the inverse gamma function was calculated from randomly generated numbers between 0 and 1 (n=1000), producing a unimodal curve with a probability density function:~~

P(t) = (t/B)«-i[exp(-t/B)]/Br(a) (l) 56 where B is a scale parameter with units of time, a is a unitless shape parameter, and r(a) is the gamma function (Johnson & Kotz 1970; Matis et al. 1992). The gamma function is an extension of the factorial function that generalizes factorials to any real number by shifting the argument down by one (Otto & Day 2007). The distribution from Equation (1) describes the probability that x amount of time will pass before B events occur at a rate of a. As described by Matis et al. (1992b), the mean E(t) and variance V(t) for transit time t were calculated as:

~~E(t) = aB (2) V(t) = ap2 (3)~~

The probability distribution varies in shape based upon a values and in timescale with p values. When a and p are equal the distribution is positively skewed, while a large a (relative to P) increases the mean, producing a normal curve, and a higher p value increases the timescale and therefore the mean of the distribution (Fig. 13).

The model was fit to actual data by varying a and p parameters to achieve similar means and standard deviations of modeled transit times to actual transit times (equations 2 and 3). Transit times for actual and model data were fit into one- month frequency bins (n=10). Model frequencies (n=1000) were downscaled by multiplying bin probabilities by actual data sample size (n=35). A chi square goodness of fit test was used to test model fit by comparing bin frequencies. The significance level was p=0.05. 57

~~Outcome~~

~~Figure 13: Probability distribution of the gamma function with varying combinations of a and p parameters. 58~~

Future arrival times of round gobies at detectable numbers were predicted for various locations upstream in the Trent-Severn Waterway by scaling up the a parameter, where a was multiplied by distances upstream of interest, while P remained constant (Matis et al. 1992b). The 95% confidence intervals for arrival times were approximated from the 2.5 percentiles of the chi-square distribution

~~(Matis et al. 1992b). The chi-square distribution is a special case of gamma when a = v/2 and p = 2.0, where v represents the degrees of freedom (Johnson & Kotz 1970).~~

~~The corresponding chi-square values were multiplied by p/2, to determine the upper and lower 2.5% confidence limits of arrival times.~~

~~In order to estimate the probability of upstream dispersal events in the~~

Trent-Severn, the probability density function of round goby transit times was calculated with Equation 1 using the gamma distribution function, and the inverse of these transit times represented the probability that an individual round goby would move a given distance per year (time transit). The p parameter was reduced by 50% to simulate hypothetical downstream movement because historically, downstream range expansion has been about twice as rapid as upstream expansion in the Trent-

~~Severn Waterway (Fig. 1).~~

~~RESULTS~~

~~The mean transit time for the detectable round goby population front upstream in the Otonabee River in 2009 and 2010 was 1.8 months/km ± 0.31 SE.~~

The modeled transit time distribution appears to be a good fit to actual round goby 59 transit times (Fig. 14), and there was no significant difference between actual and modeled bin frequencies (x29= 11.0, p=0.27).

The model predicts a rate of spread of the population front of 6.7 km/year, but it may be as fast as 9.8 km/year at the 95% confidence interval (Fig. 15). It predicts that round gobies will be detectable in Little Lake in 4.8 years, but may be as early as 3.3 years (September 2012). Similarly, round gobies may be detectable in the Otonabee River at the Trent University campus as soon as 4.3 years (September

~~2013).~~

The probability distribution of round goby dispersal is shown in Fig. 16. The model suggests that there is a 50% probability that some individuals will move 9.3 km upstream/year. There is a 99.5% probability some individuals will move at least

1.3 km upstream/year, and a 25% probability that a highly mobile individual will move as far as 27 km/year. The model was manipulated in timescale by reducing p by 50%, which doubles the transit times, but maintains the shape of the probability distribution (Fig. 17). Under the simplified assumption that downstream movement is twice as rapid as upstream movement, there is a 50% probability that some individuals will move as far as 18.5 km/year. 60

~~factual data n=35~~

~~f\ --- model data n=1000~~

~~T — | —— 1 1 — 1 1——— r 0123456789 10~~

~~Transit time (mo/km)~~

~~Figure 14: Estimated transit times (months/km) of round gobies from arrival at sites upstream from their edge of range in May 2009. Gamma parameters are a=1.08,p=1.67. 61~~

~~0 5 10 km —T r- 78.35607 78.12336~~

Figure 15: Predicted arrival times (upper and lower 95% confidence limits) of detectable round gobies at locations upstream from their distribution in the Trent-Severn Waterway in May 2009 using a gamma transit time model. 62

~~1 0.9 0.8 0.7 0.6 .a 0.5 re .a o 0.4 0.3 0.2 0.1 0 10 20 30 40 50 60 70 80 90 100 110 120 Dispersal (km/year)~~

Figure 16: Probability of upstream movement of round gobies (km/year) from observed range expansion in the Trent-Severn Waterway from May 2009 to August 2010. Black dotted line represents 5% probability. 63

~~S3 re o~~

~~20 40 60 80 100 120 140 160 180 200 220 240 Dispersal (km/year)~~

Figure 17: Probability of upstream movement of round gobies (km/year) from upstream model (J3=1.67, solid grey line), and a hypothetical downstream model (P=0.84, hatched grey line). Black dotted line represents 5% probability. 64

~~DISCUSSION~~

The detectable round goby population front is predicted to travel a mean distance of 6.7 km/year upstream in the Trent-Severn, but may be as rapid as 9.8 km/year based on the 95% confidence interval of the transit time model. Observed upstream range expansion from May 2009 to August 2010 lies within this interval, averaging 8.7 km/year. The mean predicted range expansion is lower than that observed because some sites within round goby range took longer to become occupied, which appears to be related to habitat quality (Chapter 2). Therefore the model incorporates their movements into less ideal habitats as well, and actual arrival times in ideal sites at upstream locations will likely be closer to the upper predicted confidence limits.

The probability distribution for round goby dispersal (time transit) suggests that some individuals will move upstream at least 1.3 km/year, and there is a 25% chance that a highly mobile individual will travel 27 km/year in the Trent-Severn

Waterway (Fig. 16). Because the probability distribution used to predict dispersal distances is asymptotic, it can produce unrealistic estimates, and the energetic limitations to movement of the species must be considered. Round gobies do not exhibit strong sustained swimming abilities (Skora 1996), so the upper confidence limit may be excessively large for upstream movement. This provokes the question of how far individuals can disperse within a given year. Large round gobies have been found to exhibit swimming speeds of up to 20 cm/s at 17 ° C, but only for relatively short durations of time, up to 72 minutes (Hoover et al. 2003). 65

Hypothetically, if a goby could disperse at this speed for 12 hours per day, this individual could travel up to 26 km in a month. However little is known about the dispersal capabilities of highly mobile individuals, and this is an important parameter for predicting round goby population expansion.

The tail of the probability distribution for movement distances produced from the gamma time transit model provides some insight into the scale and probability of long distance dispersal events for round gobies. Genetic characterization of round goby populations suggests that both short and long distance dispersal occurs with this species (Bronnenhuber et al. 2011), and long distance dispersal events may be important components of range expansion. The model developed here suggests that a few individuals may move a much greater distance than the detectable population. A previous modeling exercise suggests that the probability of round goby establishment is still very high under low density conditions (Velez-Espino et al. 2010). Therefore, in a hypothetical management scenario, to implement a barrier for upstream movement of round gobies at Little

~~Lake, the model suggests it would be advisable to do so immediately, despite the fact that the arrival of a detectable population isn't predicted to occur until at least~~

~~September 2012.~~

By scaling down the p parameter, the model retains its shape, but the dispersal distances increase per unit time. For example, p was reduced by 50% to hypothetically represent downstream movement. This was an example of how the model can be easily manipulated to incorporate other variables, which may also be 66 used to account for habitat variability and changes in the rate of range expansion in relation to population size.

When projecting upstream movement, it is assumed that round goby dispersal will follow a similar transit time distribution through upstream areas to that observed in the Otonabee River in 2009 and 2010. This is a simplified assumption because environmental conditions likely affect their movements.

Habitat in particular appears to affect their distribution and abundance, especially at the edges of their range (Chapter 2). Given that round gobies occupy mainly high quality habitats at the edges of their range, it can be reasonably assumed that mobile individuals are in search of such habitats. Therefore dispersal may be more rapid through poorer quality habitats, which is a pattern observed in many ecological dispersion models (Shigesada et al. 1995; Schippers et al. 1996; Kinezaki et al.

2002). Conversely, large areas of poor quality habitat (wetlands) are more resistant to round goby invasion (Cooper et al. 2007; Young et al. 2010). The presence of such an area may reduce its rate of range expansion.

Flow may also be an important factor affecting dispersal. Round gobies do not exhibit high sustained swimming speeds (Skora, 1996), so river currents may inhibit dispersal to some degree. The presence of navigational locks may also temporarily inhibit their movement, although the degree of this effect is unknown.

These factors do not necessarily discount the validity of these model predictions, especially of those regarding the arrival time at Little Lake, because up to this location, the area through which round gobies will disperse is entirely a river 67 ecosystem with no navigational locks. There may be some variability in habitat, although in the lower reaches of the Otonabee River, habitat appears fairly homogenous (personal observation), and may have a negligible effect on round goby dispersal. Additionally, the predictions of this model consider only natural dispersal, while bait bucket, ballast water, and hull transfers may be other factors to consider for round goby spread (Charlebois et al. 1997; Wiesner 2005; Raby et al. 2010).

When estimating long-term dispersal rates it may also be important to consider the invasion timeline and population size. Previous modeling exercises have suggested that range expansion will accelerate once population size becomes sufficiently large (Shigesada et al. 1995; Jerde & Lewis 2007). Models have been developed that incorporate variability in dispersal rates with changes in population size and structure (Shigesada et al. 1995; Yamamura 2002). This type of model may be useful for long term projections of round goby dispersal.

In conclusion, a transit time model predicts that a detectable round goby population will travel at a speed of 6.7 km, but will likely travel 9.8 km/year into ideal habitats. The model estimates that round gobies will be present in detectable numbers in Little Lake in 3.3 years. However, some individuals may travel much further than the detectable population front. These findings provide insight into round goby range expansion in fluvial ecosystems, and uncover some potentially useful information for management of round goby spread. This model is highly manipulable, and may have further applications in predicting movement of round 68 gobies downstream in fluvial ecosystems, through lacustrine ecosystems, the effects of habitat on range expansion, and to other invasive species. 69

~~Chapter 4: Predation rates of round goby in established and recently invaded areas~~

~~of the Trent River~~

~~INTRODUCTION~~

~~Highly successful invasive species are known to proliferate rapidly in introduced regions due to high resource availability and few natural enemies~~

~~(Colautti et al. 2004; Barton 2005; Nentwig 2007). Native predators often exhibit an~~

~~"adaptive" lag period before preying heavily on an exotic species (Carlsson et al.~~

2009). The response of predators depends on whether the exotic prey item is within their feeding capabilities. When the prey item is outside of their capabilities, predators are unable to adapt, which often leads to a decline in predator abundance

~~(Pothoven et al. 2001; Carlsson et al. 2009). When a new prey species is well within its feeding capabilities, a predator can impose biotic resistance to invasion.~~

~~Predation has been suggested as the limiting factor in the spread of invasive fishes in~~

California streams (Baltz & Moyle 1993), and the predatory blue crab (Callinectes sapidus) has been shown to limit the spread of the highly invasive European green crab (Carcinus maenas) in Eastern North America (DeRivera et al. 2005). An exotic prey species may also be within the feeding capabilities of a subset of individuals in the predator population, which can lead to an "adaptive" response over time

~~(Carlsson et al. 2009; Carroll et al. 2005).~~

~~Since its arrival to North American waterbodies in 1990, the round goby~~

(Neogobius meianostomus) has rapidly colonized the Great Lakes and many of their tributaries, displacing a number of small fish species (French & Jude 2001; Balshine 70 et al. 2005; Bergstrom & Mensinger 2009) and becoming a dominant benthic fish. It is now an integral part of aquatic food webs and a dominant source of energy flow to upper trophic levels due to the high density and biomass production of this species

(Johnson et al. 2005b; Campbell et al. 2009). Within the past two decades, many types of North American predators have started to exploit round gobies as a prey source, including waterfowl (Somers et al. 2003; Jakubas 2004; Bur et al. 2008), water snakes (King et al. 2006), and many fish species. Benthivores exploiting round gobies include rock bass (Ambloplites rupestris), burbot (Lota lota), bowfin (Amia calva), freshwater drum (Aplodinotusgrunniens), brown bullhead (Ameiurus nebulosus), and channel catfish (Ictalurus punctatus), along with pelagic and mixed feeders such as smallmouth bass (Micropterus dolomieu) largemouth bass

~~(Micropterus salmoides), yellow perch (Perca flavescens), walleye (Sander vitreus), lake trout (Salvelinus namaycush), northern pike (Esox lucius), and longnose gar~~

~~(Lepisosteus osseus) (Steinhart et al. 2004b; Johnson et al. 2005b; Truemper & Lauer~~

~~2005; Deitrich et al. 2006, Campbell et al. 2009; Taraborelli et al. 2010).~~

There is no question that predators are ultimately preying upon round gobies in North American ecosystems, but their response to round gobies during the initial stages of invasion is unknown. Predators can strongly influence the dynamics and success of invasive species (Carlsson et al. 2009). Given the successful establishment of round gobies in North American ecosystems, the response of predators is clearly too weak to pose a biotic resistance to their invasion. If predation rates are relatively low during initial stages of invasion, this presents a survival advantage for mobile 71 individuals at the edges of their range. Low initial predation risk may be an important contributing factor to their rapid range expansion in invaded ecosystems.

Predation rates are difficult to measure in the field, but can be estimated with tethering experiments. The major assumption of using a tethering experiment is that there is a relationship between the loss of tethered prey and the natural mortality of untethered prey due to predation (Kneib & Scheele 2000). Tethering has been used in the past to assess relative predation risk of many small fish species (Manderson et al. 2004; Diaz et al. 2005; Rypel et al. 2007; Dupuch et al. 2009), and in the case of the round goby, to assess the importance of shelter for survival (Belanger & Corkum

~~2003).~~

~~In this study, I used in-situ tethering to assess the relative predation risk to round gobies in established and recently invaded areas of their occupied range in a~~

Great Lakes tributary. It was hypothesized that round gobies experience a lower predation risk at the edge of their range where predators may be initially naive to their presence, and high predation risk in areas where the invader has been well established, where predators would be expected to have incorporated the new prey type into their diet. I also determined common predators of tethered round gobies, the mean sizes of predated and survived round gobies, and the relationship between predated round goby size and predator size to better understand predation dynamics during range expansion. 72

~~METHODS~~

~~Study location~~

~~The study was conducted in the Trent River in Southeastern Ontario (Chapter~~

~~2). The Trent River has a diverse warmwater fish community and highly valued walleye (Stizostedion vitreum), muskellunge (Esox masquinongy) and bass~~

(Micropterus spp.) sport fisheries. Other obligate and facultative piscivores present in the river and known to prey upon round goby include longnose gar, northern pike, brown bullhead, rock bass, yellow perch, and freshwater drum (OMNR, 2001).

~~Smallmouth and largemouth bass are the dominant predator species in the Trent~~

~~River (OMNR, unpublished data).~~

~~Site selection~~

Sites for tethering experiments were selected at two river locations. Three sites were located in an area where round gobies were well established, near their point of introduction in the Trent River (CORE) (Fig. 18). Round gobies were discovered in this area in 2003 (Raby et al. 2010) so they likely occupied those sites for close to 7 years. Three sites were also located at the downstream edge of their range (DSE) in a recently invaded section of the river. Downstream edge sites were located beyond the edge of round goby distribution in 2008 (unpublished data,

~~OFAH) (Fig. 18), so it is likely that these sites were occupied for less than 2 years.~~

~~At each location, three sites were chosen with comparable habitat types (Fig.~~

~~18). These consisted of shallow rocky shoals (SRS), the most complex habitat with CORE DRS~~

~~Point o» IrstKXSuefton -~~

~~44 37S27 N -~~

~~44 31218 N -~~

~~44 27830'N~~

~~77 92998' W 77 86394 W~~

~~Figure 18: round goby tethering sites in three habitat iVHrt+Xr-J? types: sandy shoal (SS), shallow rocky shoal (SRS), and~~

~~a deep rocky shoal at CORE and DSE range locations in Lake Ontario the Trent River 74 large cobble and boulders; deep rocky shoals (DRS), a less complex habitat with~~

mainly cobble and gravel, and sandy shoals (SS), the least complex, predominantly sand with some gravel and cobble substrate. Each site contained at least 300 m2 of relatively homogenous habitat. Sites were located by scanning the bottom using an

~~Aquaview S-series colour underwater camera (Nature Vision Inc., Brainerd, MN).~~

~~Tethering experiment~~

~~Round gobies were collected for tethering experiments from a rocky littoral site in the Trent River near Hastings, Ontario (44.30771° N, 77.96033° W to~~

44.30855° N, 77.95687° W) (Fig. 18) by a combination of angling and overnight sets with minnow traps, as preliminary trials in 2009 suggested no difference in survival rates between angled and trapped fish. Gobies were collected in the same morning that they were used in tethering trials. They were held in oxygenated water for up to

3 hours before trials commenced, and experienced no mortality between collections and trials. Round gobies used in experiments varied in size from 70 to 123 mm total length with a mean TL of 88 mm (age 1-3, based on previous length at age data in the Trent River) (Gutowsky & Fox in review).

~~Round goby tethering experiments were conducted in June and July of 2010.~~

Trials were conducted on separate days, alternating between range locations with comparable habitat types consecutively to achieve comparable control variables between range locations. Trials were conducted at mid-day between the hours of

~~10:00 and 13:00, and start times varied as much as one hour due to inconsistencies in collection time. 75~~

The tethering apparatus used was tested in 1 m3 holding tanks with 10 round gobies of various sizes for 4 hour time periods. No fish escaped the tether during trials, and no fish experienced mortality or appeared injured in subsequent weeks.

Fifteen preliminary trials were also conducted in the Trent River in study sites prior to the initiation of the study, testing the effectiveness of the tethering method in assessing predation risk, and various derivations of the method in an attempt to reduce predation rates.

Round gobies were attached by 2 m of fluorocarbon fishing line (2.7 kg strength) to a tethering device that consisted of a brick base and a 6-inch bolt with a washer for line attachment (Fig. 19a,b). Gobies were tethered by inserting a size 20 fishing hook below the dorsal fin, following the procedure of Belanger & Corkum

(2003). Tethered gobies were placed in the water from a boat, with the fish lowered first into the water, after which the tethering apparatus was lowered to the substrate by a float line. Tethers were placed at least 10 m apart from each other.

Each trial employed 20 tethered round gobies per site, which were set for 2 hr. At the end of the trial, the float lines were used to retrieve the tethering devices. If the tether line became entangled on the bottom, a snorkeler made the retrieval. Gobies were considered predated if they disappeared, were fatally injured and remained on the hook, or if the goby was missing and a predator was caught on the hook.

~~Captured predators were identified and measured for total length.~~

~~Variables measured in each trial were round goby TL, sex, tether depth, surface water velocity, and temperature. Depth was determined from the length of 76~~

~~A? Ufa'.1 *: b~~

~~Figure 19: Tethering device (a) and depiction of a tethered round goby (b), from round goby tethering experiments in the Trent River in June and July of 2010. 77~~

~~the float line for each tethered fish. Water velocity was estimated by timing the~~

~~horizontal movement of a float, and temperature was measured at <1 m below~~

~~water surface, at one location near the center of the site prior to each trial. The~~

~~relative abundance of untethered round gobies was estimated once per site, at a~~

~~randomly selected location within each site in August 2010 using the angling~~

~~removal method as described in Gutowsky et al. (2011). This method involves~~

~~angling with small baited hooks within a 2 m2 floating barrier for 20 minutes at each site.~~

~~Data analysis~~

~~Two-way factorial Analysis of Variance was used with Newman-Keuls post-~~

~~hoc analysis to test the main hypothesis by comparing predation rates between~~

~~range locations (CORE and DSE) and habitat types (SRS, DRS, SS). The same tests were used to compare control variables (goby length, tether depth, river current,~~

and temperature) between range locations within each habitat type to determine whether conditions were similar between comparable habitat types. Predation rates were arcsine transformed prior to analysis to meet assumptions of normality (Zar

2010). ANCOVA homogeneity of slopes analysis was used to compare the relationship between predation rate and predator catch rate between range locations in order to confirm that any differences in predation rates were due to predation. An independent t-test was used to compare the mean length of predated

and survived round gobies within each range location, and simple regression analysis was used to observe the relationship between round goby total length and 78 predator total length within each range location to determine if predator gape limitations were affecting predation rates at either range location. The level of significance for all tests was p<0.05.

~~RESULTS~~

~~Predation loss~~

Control variables (round goby length, tether depth, river surface velocity, and water temperature) were comparable between range locations in the equivalent habitats (Fig. 20). There were no significant differences between range locations in any control variables (Table 1; p>0.05 in all cases), but tether depth did vary significantly between habitat types (Newman-Keuls, p<0.001).

Predation rates of tethered round gobies were significantly higher at CORE sites than downstream edge sites in all habitat types (Fig. 21, Table 2). Within range locations predation rates were higher on deep and shallow rocky shoals than the sandy shoals. In CORE sites, predation rates were significantly higher on both types of rocky shoals than on the sandy shoal (Newman Keuls, p<0.008 in both cases). The relationship between predation rate and predator catch rate was similar at both range locations (Fig. 22), and there was no significant difference in the slopes of this relationship between CORE and DSE locations (ANCOVA, Fi,i4=0.001, p=0.97).

Differences in predation rates among habitat types followed a similar trend to that of round goby relative abundance within each range location (Fig. 23). There was no significant difference in length between round gobies that survived and were 79

~~I CORE ODSE 100 *V- E 80 4 E I~~

~~60 Q. 3 >• i .Q ai O •a 60 40 »- ? •co 3 o 20 * 1 en~~

~~SRS DRS SS SRS DRS SS~~

~~0.18 30 u 0.16 01 (A 0.14 i 25 E, 0.12 0u r~T*l r*f* 0) 20 i 0.1 i- 3 u ID 0.08 k 15 o 01 ai 0.06 a E 10 > 0.04 01 a> t- 5 > 0.02 0 -1 "—I —•—I SRS DRS SS SRS DRS SS~~

~~Habitat type~~

Figure 20: Control variables (round goby length (TL), tether depth, river surface velocity, temperature) (± SE) in round goby tethering experiments in in 3 habitat types (SRS, DRS, SS) at both the CORE (grey bars) and the DSE (white bars) range locations in the Trent River; June -July 2010. 80

~~Table 1: Results from two-way factorial ANOVA comparing control variables between range locations (CORE and DSE) and habitat types (SRS, DRS, SS) in round goby tethering trials; June - July 2010.~~

~~Goby length F df P~~

~~Range location 0.12 1 0.74~~

~~Habitat type 0.21 2 0.81~~

~~Location x habitat 1.96 2 0.18~~

~~Tether depth F df P~~

~~Range location 4.14 1 0.06~~

Habitat type 43.7 2 <0.001 *

~~Location x habitat 0.37 2 0.70~~

~~River current F df P~~

~~Range location 0.26 1 0.62~~

~~Habitat type 0.61 2 0.56~~

~~Location x habitat 0.54 2 0.60~~

~~Temperature F df P~~

~~Range location 0.82 1 0.38~~

~~Habitat type 0.91 2 0.43~~

~~Location x habitat 0.95 2 0.42 81~~

~~100 -j I CORE DDSE 90 " 80 ~ f IS 70 • £ 60 " re 50 - •*• c o 40 - f 4re3 •a 30 - V O- 20 • 10 - o • DRS SRS SS~~

~~Habitat type~~

Figure 21: Comparison of mean predation rates of tethered round gobies (± SE) between sites in the area of original introduction (CORE, grey bars) and the downstream edge of range expansion (DSE, white bars) in three, 2-hour trials conducted in three habitat types in the Trent River; June - July 2010. DRS = deep rocky shoal, SRS = shallow rocky shoal, SS = sandy shoal. 82

Table 2: Results from A) two-way factorial ANOVA and B) Newman-keuls post hoc analysis comparing predation rates between range locations (CORE, DSE) and habitat types (SRS, DRS, SS) from round goby tethering trials in June and July of 20.10. Asterisks indicate p<0.05.

~~A) Effect df~~

~~Range location 60.5 <0.001~~

~~Habitat type 9.5 0.003~~

~~Location x habitat 1.2 0.33~~

~~Center Center Center Edge Edge Edge~~

~~SRS DRS SS SRS DRS SS~~

~~Center SRS X~~

~~Center DRS NS X~~

~~Center SS * X - -~~

~~Edge SRS * NS X -~~

~~Edge DRS * NS NS X~~

~~Edge SS * * NS NS X 83~~

~~50 ODSE A CORE 45 40 35 30 25 20 15 10 5~~

~~0 ! 1 S I 1 —j— 20 30 40 50 60 70 80 90 100 Predation Rate (%)~~

Figure 22: Relationship between round goby predation rate (%) and predator catch rate (% of predated gobies) from tethering trials at 3 sites at CORE (open circles) and DSE (closed triangles) range locations of the round goby distribution in the Trent River in 2010. Round goby abundance co w3 < fD ON>*»Ji0^ (D N3 W to 1 _l 1 I 1 L__J I cr —i i i i i i i i !•£ c ?0 3 o V) 2 Q- c W 3 3 o (U rt fD era D 3 o 50 3* ^ S- 50 to (U *< p. fD to <-'• M << H GO M a* 2 ° 3, £3 rt r- pa 3 S to 3 <" W W fD rt rt- *-s rt- fD 5' fD to 01 to to H Ji » O- 50 73 N3 to O Q)

~~3" II £fD* CL fD fD "> fD <—> » t3 orq OX) •i 3 <-••~~

~~3 5 3. to o ^ w to to O o O 4f 3 ^ 50 tn m £-£ O S" &3 oo O K 53 ?0 3 Oi-'Mio4itncn^iootDi-» OMMtOJitnfJl-vlOO tD M oIIo m3 « OOOOOOOOOO o o to rv 3 o oooooooo o~~

~~e= tjo o tn (%) ajej uoiiepajd CO 85 predated at either the center of range or downstream edge locations (Fig. 24;~~

~~independent t-tests; t=0.46, p=0.86, df=178; t=0.55, p=0.44, df=178, respectively).~~

~~Round goby predators~~

~~There were a total of 35 predators caught on tethered round gobies in CORE~~

~~sites, and 24 predators caught in downstream edge sites, which is a mean predator~~

catch rate of 26% of tethers with missing gobies. The dominant predator captured was smallmouth bass at both CORE (77%) and DSE (96%) sites (Fig. 25). Rock bass were also captured at both range locations, while brown bullhead, and yellow perch were only caught in CORE sites. There was no significant relationship between

~~round goby size and predator size in either range location (r<0.33, p>0.05 in both~~

~~cases) (Fig. 26).~~

~~DISCUSSION~~

~~Predation rates on round gobies were significantly higher (27%) at sites in an~~

~~area of a well-established round goby population (7 years) than at sites in an area of~~

~~recent invasion (<2 years) in the Trent River (Fig. 21, Table 2). The large, significant~~

difference in tethered predation rates, combined with the comparable relationship between predation rate and predator catch rate at the two range locations (Fig. 22), provides strong evidence that there are differences in predation risk to round gobies

~~inhabiting the two study areas. Because those differences occurred in all three habitat types examined, these results strongly infer that the differences in predation 86~~

~~Figure 24: Mean length (TL) of predated and survived round gobies (± SE) from tethering trials in CORE and DSE range locations in the Trent River; June - July 2010. 87~~

~~100~~

~~90 B CORE N = 35 80 • DSE N = 24 70 vo re •o 60 S3 40 O M "O 30 c o3 cc 20 10 0 smallmouth bass rock bass brown bullhead yellow perch~~

~~Species~~

~~Figure 25: Round goby predators captured on tethered lines in CORE and DSE range locations in the Trent River; June - July 2010. 88~~

~~50 A CORE ODSE 45 o 40 35 o V re 30 •ao* i_ 25 a. si> 20 o 60 15 O cc 10 5 0 7 8 9 10 11 12~~

~~Round goby total length (cm)~~

Figure 26: Relationship between length of round gobies predated on tethers and that of predators captured on tethered round gobies in the Trent River; June- July, 2010. Sites within the two ranges locations are pooled. 89 risk are related to the length of time the species was established in the area and/or its location relative to the edge of population expansion.

While relative predation rates are assumed to reflect natural mortality due to predation (Kneib & Scheele 2000), it is clear that the predation rates of tethered round gobies were unrealistically high, undoubtedly due to their limited mobility.

Although the tethering line was considered long enough for a reasonable amount of movement (2 m), tethered round gobies often became tangled in the complex substrates of the Trent River (personal observation). This is likely the major reason why predation rates were much higher than those found in a previous round goby tethering study (Belanger & Corkum 2003), where the round gobies were tethered in open, shallow (< 1 m), sandy habitats. In addition to habitat complexity, the Trent

River has a very high abundance and diversity of piscivorous fishes, especially of round goby predators such as the smallmouth bass (OMNR, unpublished data), which likely contributed to the high predation rates of round gobies. Lastly, tethered round gobies were placed in fairly deep areas in the Trent River (mean tether depth of 3 m), and depth has been found to be an important factor affecting predation rates

(Rypel et al. 2007). While the rates of predation on the tethers were clearly much higher than those that would occur naturally, artificially high predation rates have been used as relative measure across space in the past. Rypel et al. (2007) observed predation rates of up to 100% during 30-minute trials with tethered mojarra

(Eucinostomus spp.), a small tropical fish in Bahamian tidal creeks, and were able to assess the effect of water depth on relative predation risk. Koh & Menge (2006) also used artificially high predation rates of Lepidoptera imitations by birds in 90 neotropical forests as a relative measure. They defined this type of study as rapid assessment of relative predation rates (Koh & Menge 2006).

Three possible explanations for the CORE - DSE differences in round goby predation rates are that (a) predation rates increase as a function of round goby density (functional response); (b) round gobies support or attract a higher density of predators over time (numerical response); or (c) there are inherent differences in predator communities between the two regions of the Trent River.

Explanation (c) is unlikely given the connectivity and similarity of the two range locations considered. The downstream edge sites were located 35 km downstream of the CORE sites in the Trent River, while tethering sites within the two study areas were as much as 8 km apart from one-another. Fish communities have been strongly correlated with stream characteristics including depth, current, and substrate type (Gorman & Karr 1978). Habitat type was controlled for in the study, and the habitat types tested at both range locations were very similar in all of the aforementioned characteristics (Fig. 20), making any strong differences in predatory fish communities unlikely. Predator species caught on tethered round gobies were also very similar, with smallmouth bass as the dominant predator (Fig.

~~25).~~

~~Invasive species are known to proliferate in new regions due to a lack of natural enemies (Colautti et al. 2004; Nentwig 2007). Based on invasion theory, both~~

(a) and (b) are strong potential explanations for the observed pattern of higher predation risk to gobies in the area where they were longer established. Round goby 91 predators are likely exhibiting a functional and possibly a numerical response to the new prey species, although these two explanations cannot be differentiated in this study. This type of phenomenon has been observed previously, where zebra mussels

~~(Dreissena polymorpha) were preyed upon more heavily 7 years after invasion than~~

~~3 years after invasion in the Mississippi River (Thorp et al. 1998; Bartsch et al.~~

~~2005), but the specific mechanisms behind these predation rates have not been investigated.~~

The significant differences found in predation rates between range locations may be explained by a functional predatory response (a), where round goby predation rates increase in relation their density over time. With a type III functional response, consumption rates increase exponentially over time before saturation occurs (Holling 1959). This accelerated response can result from a learning period, and/or prey switching by predators. Predators will generally target prey species that are most abundant in the environment, and when changes in prey dynamics occur, predators may switch their primary targeted prey. An example of this was illustrated with guppy predation. As fruit fly abundances decreased on the water's surface, guppies switched to prey primarily on the more abundant tubificids on the substrate (Murdoch 1975). This type of change in hunting behaviour may also occur in predators with opportunistic feeding strategies in response to round gobies, where species such as smallmouth bass and yellow perch may begin foraging more often on the benthos in response to high goby densities. 92

~~Round goby abundance was actually higher in rocky habitat tethering sites at~~

~~the DSE than in sites in the CORE area in 2010 (Fig. 23). Edges of range are generally~~

~~of lower abundance than longer established areas (e.g. Raby et al. 2010), and this~~

~~was the case at the DSE in 2009 (Chapter 2), but not in our tethering sites in 2010.~~

~~This was due to a combination of decreasing abundance of round gobies in the CORE~~

area in 2010 from previous years (Chapter 2), while on rocky shoals in the DSE area, gobies occupied the area for one to two years, and abundance became high in these ideal habitats (Chapter 2). Therefore a functional response similar to that observed with guppies appears less likely. However the exponential nature of a type III

~~response may also be due to a learning period. An example of this has been observed~~

~~in pumpkinseed (Lepomisgibbosus), redbreast sunfish (Lepomis auritus) and rock bass (Ambloplites rupestris) populations in Sweden. Lepomids that had been~~

~~exposed to zebra mussels for more than 10 years consumed up to 105% more zebra~~

~~mussels than those exposed for less than 10 years, while previously unexposed~~

individuals consumed almost none (Carlsson & Strayer 2009). Similarly, predatory fish species in North America are likely naive to the presence of round gobies as a prey source during the first few years of their invasion, and within less than a decade they begin to prey upon them more frequently.

There are a number of ways in which predators can respond over time to a novel prey species. Short-term responses result from individual or social learning, while longer term responses include ontological changes in morphology and evolutionary adaptations (Carlsson et al. 2009). In the case of round goby predators,

~~changes in morphology are unnecessary, as predators are already equipped to feed 93~~

~~upon this soft-bodied fish. It is likely that increased feeding on round goby prey is a~~

~~learned response of some type. Predators would first need to recognize round~~

~~gobies as a potential prey source, and then determine how to capture them. Round~~

~~gobies may have different defensive behaviours than typical prey items. They are~~

~~very cryptic; hiding amongst complex structures and in sediments (Ray & Corkum~~

~~2001) and it may take native predators some time to determine the best strategy to~~

~~effectively prey on such a species.~~

~~The ability of predators to associate a new prey item with habitat indicators~~

~~of their presence is important to their feeding success. An empirical example of this~~

can been seen in North American wasps (Polistes spp.) preying upon silver spotted skipper (Epargyreus clarus) larvae, which construct leaf-and-silk shelters for protection. In controlled tests, wasps were able to associate exposed prey items with the habitat indicators of their hidden presence (damaged leaf and silk) and over time, began to prey upon hidden larvae (Weiss et al. 2004). This example is very similar to that of the round goby, as both prey are cryptic species (Ray & Corkum

~~2001; Wiess et al. 2004). Initial exposure of a predator to a round goby would likely~~

occur while a round goby is active. A successful predator may then be able to associate the presence of round gobies with certain habitat indicators such as a large rock crevasse, and begin to prey upon them more frequently by ambushing mobile individuals. Chemical cues may also be an important component of this learning, as a number offish species use chemosensory systems to detect prey via olfaction or taste (Hara 1994). Round gobies are known to secrete pheromones that attract the 94

~~opposite sex (Corkum 2004), which predators may be able to recognize to effectively target them as prey.~~

Round gobies may also be supporting or attracting a greater number of their predators in well-established areas (numerical response). This can occur in the form of increased predator reproduction, or by predator aggregation. Mobile predators can be attracted to areas of high prey density. This has been observed with many predatory fish species, including primarily yellow perch and rock bass in response to emerging lake trout fry (Riley & Marsden 2008). Waterfowl were also found to aggregate in Long Point Bay, Ontario during the zebra mussel invasion (Petrie &

~~Knapton 1999). This may also be the case with round gobies. Their main predator in this study, the smallmouth bass, is known to occupy a wide range of habitat types~~

(Lobb & Orth, 1991), while high densities of round goby appear to be restricted mainly to rocky substrates (Johnson et al. 2005a; Chapter 2). Predators such as smallmouth bass may be aggregating near rocky substrates to exploit round gobies as a prey source.

A reproductive numerical response by predators is another possible explanation. Round gobies occupy a large biomass in invaded ecosystems, and provide a major energy source for predators that exploit them (Steinhart et al.

~~2004b; Johnson et al. 2005b, Taraborelli et al. 2010). Round gobies may be responsible for high growth rates of young-of-year smallmouth bass (Steinhart et al.~~

~~2004b), which has potential to increase their abundance over time. If such a numerical response does occur, the round goby invasion presents an advantage to 95~~

~~predators such as smallmouth bass, and may result in increased abundance of such~~

~~species throughout invaded waters. However, round gobies feed upon the eggs of~~

~~many predatory fish species (Chotkowski & Marsden 1999; Weimer & Sowinski~~

~~1999, Steinhart et al. 2004a), so their effect on their populations is likely a balance of~~

~~increased prey availability and lower reproductive success.~~

~~Native predators are an important factor regulating the long-term dynamics~~

~~of invasive species (Carlsson et al. 2009). Predators that have learned of an~~

~~abundant new prey source have been found to aggregate in areas of high prey~~

~~density, which can lead to the subsequent reduction in the once hyper-abundant~~

~~prey (Petrie & Knapton, 1999). The response of predators to round gobies over time~~

~~seen here in the Trent River and in the Great Lakes (Johnson et al. 2005b;~~

~~Taraborelli et al. 2010) suggests that predation has the potential to contribute to the~~

~~reduction of hyper-abundant round gobies. Similar to other highly successful~~

~~invasive species, round gobies have shown major declines in abundance over time in~~

~~some areas after initially high abundance (Simberloff & Gibbons 2004; Young et al.~~

~~2010). This 'boom and bust' pattern is poorly understood in most cases, but is~~

~~frequently attributed to a response by natural enemies or a reduction in resource~~

availability over time (Simberloff & Gibbons 2004). Predation may be an important biotic factor contributing to the reduction of round goby in North America. This highlights the importance of minimizing overexploitation and maintaining strong predator communities in aquatic ecosystems. 96

~~The driving forces behind the rapid expansion of round goby distribution are~~

~~of interest biologically. Round gobies are generally less abundant at the edges of~~

~~their range than well-established areas (Raby et al. 2010; Taraborelli et al. 2010).~~

~~Therefore movement can be advantageous to individuals and their offspring due to~~

~~less intraspecific competition and higher resource availability (Raby et al. 2010).~~

~~Lower predation rates at the edges of their range are a potential advantage to~~

~~movement for individual fitness, and may be another selective pressure for the~~

~~survival of highly mobile individuals. From an evolutionary perspective this may be~~

~~a contributing factor to the rapid range expansion of the round goby.~~

~~The dominant predator species caught during round goby tethering trials in the Trent River was the smallmouth bass (Fig. 25). It has been well documented that~~

~~smallmouth bass are a common round goby predator in North American ecosystems~~

~~(Steinhart et al. 2004b; Taraborelli et al. 2010), and round gobies have been found to~~

~~comprise up to 75% of their diet (Johnson et al. 2005b). While smallmouth bass feed~~

~~opportunistically a wide range of benthic and pelagic prey items (Vander Zanden &~~

~~Vadeboncoeur 2002), they are known to target primarily crayfish (Probst et al.~~

~~1984). The strong response of smallmouth to round gobies may be related to the~~

~~similarity of round goby and crayfish habitat use in addition to the opportunistic feeding behaviour of the smallmouth bass.~~

There was no significant difference in the mean size of predated and survived round gobies (Fig. 24). There was also no relationship between round goby length and predator length (Fig. 26). These findings suggest that piscivorous fish in the 97

Trent River are not very gape limited in round goby feeding. In a previous tethering study similar sizes of round goby were used, and smaller round gobies were predated more often than larger ones (Belanger & Corkum, 2003), but the dominant predator was the yellow perch, which is gape limited in round goby feeding

(Truemper & Lauer, 2005). The majority of predators in the Trent River were smallmouth bass (Fig. 25) and even young-of-year smallmouth bass have been shown to feed heavily on round gobies (Steinhart et al. 2004b). This further supports the idea that predators in the Trent River are very capable of preying upon round gobies, but are naive to their presence during initial stages of invasion.

~~In conclusion, round gobies experienced lower predation risk in a recently invaded area (« 2 years) than a longer established area («7 years) in the Trent River.~~

~~This suggests that round gobies experience relatively low predation risk during initial stages of invasion due to the naivety of predators to their presence as prey.~~

Low predation risk is often an important component of species invasions (Colautti et al. 2004). This finding provides insight into why round gobies have been so successful in colonizing North American ecosystems, and predation may be a selective pressure for the survival of highly mobile individuals that contribute to range expansion. The response of round goby predators to the presence of round gobies as prey may be functional or numerical, and predators have the potential supply a biotic resistance to the persistence of hyper-abundant round gobies. 98

~~Chapter 5: General Discussion~~

In this thesis I have made contributions to many aspects of round goby ecology and invasion biology. I examined the upstream and downstream range expansion of the round goby in the Trent-Severn Waterway, addressing its rate of range expansion, the time of year expansion occurs, and its population dynamics, demographics, and habitat use at the pioneering edges of its range in comparison to a longer established area. I have also developed a simple transit time dispersal model to describe its upstream expansion through the Trent-Severn Waterway, and

I used this model to predict the arrival times of round goby at various upstream locations, as well as the probability of long distance dispersal events in this highly invasive species. Lastly, I was able to show that round goby predation rates are lower in a recently invaded area than in a longer established area of the Trent River, which suggests that predation risk is lower to round gobies during the initial stage of invasion.

~~Both the upstream and downstream edges of the round goby's range in the~~

Trent-Severn Waterway were characterized by very low site occupancy and relative abundance in May 2009, as these areas were representative of the pioneering detectable population fronts, and based on previous sightings, this was likely the first year of occupation in the majority of sites. Site occupancy and relative abundance increased dramatically in sites within their previously defined range edges over the first summer of occupation, and by August 2009 mean abundance was higher at the upstream edge of their range than the center. Round gobies were 99

~~more selective with their habitat at the edges of their range than at the center, which~~

~~reflects a more limited and predictable distribution in more recently invaded areas.~~

~~In May and August of 2010 both edge range locations were very similar to the center~~

~~in site occupancy and relative abundance, indicating that within a year of~~

~~occupation, population dynamics became relatively homogenous between recently~~

~~invaded and established areas. These findings suggest that management strategies~~

~~for containment of round goby spread should target this species in the spring when~~

~~abundance and distribution are limited and predictable in recently invaded areas.~~

~~However, once detectable gobies occupy an area, abundance increases rapidly and very aggressive measures would be required to deter establishment.~~

~~Round gobies appear to exhibit seasonal variation in dispersal rates, which~~

~~may be related to reproduction. Based on my findings, it appears that during the~~

non-reproductive season, round gobies that are predominantly small and female biased conduct large scale migrations into previously unoccupied areas, expanding the population's range in the Trent-Severn. During the reproductive season very

~~little range expansion occurred, although large mean lengths in edge sites in August~~

~~2009 may indicate that individuals migrated into range edges from closer to the population core. The demographics of individuals that were found in their expanded~~

range in 2010 were likely a product of intraspecific competition, and these findings are consistent with the previously proposed mechanism of range expansion, where large individuals force smaller individuals into less ideal habitats, form which they disperse (Ray & Corkum 2001). A transit time dispersal model predicts that individual round gobies may disperse much more rapidly than the detectable population front advances. Long distance dispersal events are often rare, but can be important components of range expansion (Trakhtenbrot et al. 2005). The question still remains as to how far an individual could disperse within a given year based on energetic limitations.

~~Methods have been developed that may be effective at controlling round goby spread with electric or hydraulic barriers (Charlebois et al. 1997; Hoover et al.~~

2003). I would recommend implementing one of such barriers at the navigational lock at Little Lake, Peterborough immediately to inhibit the upstream spread of round gobies, and ensure that any long distance dispersers do not pass the lock this spring/summer. In order to effectively contain round goby spread, other factors to consider include limiting bait bucket transfers, ballast water transfers, and hull transport, which are potential vectors of round goby dispersal.

Round goby predation rates were significantly lower at the downstream edge of their range in areas that were occupied for less than two years than at the center of their range in areas occupied for close to seven years. Round gobies may experience low predation risk during initial stages of invasion due to the naivety of predators to their presence as a prey item and/or the ability of predators to quickly adapt and exploit this cryptic species. This may be an important component of the successful colonization of round gobies in North America. In retrospect it is clear that our ecosystems were ideal for a round goby invasion. Not only were round gobies equipped to acquire benthic resources more efficiently than native benthivores through aggressive behavior and specialized pharyngeal dentition for crushing highly abundant dreissenids (Ray & Corkum 1997), round gobies also take

~~refuge in substrates, and there are no highly abundant benthic feeding obligate~~

~~piscivores in our ecosystems to supply a natural resistance to their invasion. Instead~~

~~the species that have slowly started to prey upon round gobies are either mixed feeders such smallmouth bass and yellow perch, or facultative benthic piscivores~~

~~such as brown bullhead.~~

~~In longer established areas predators seem to have adapted to exploit round~~

~~gobies as a prey source, which may be a functional and/or numerical response. I~~

~~hypothesize that this response is both functional and numerical to some degree.~~

~~Round gobies are cryptic, and are known to hide in substrates in response to~~

~~predators (Charlebois et al. 1997), so predators probably require a learning period to recognize and effectively prey in this exotic species. Additionally, many predators,~~

~~including the main round goby predator in the Trent River, the smallmouth bass, feed opportunistically both in both the substrate and the water column (Lobb &~~

~~Orth, 1991), and it has been shown that fish predators will switch their primary feeding strategy based on the abundance and location of prey types (Murdoch~~

~~1975). Species such as smallmouth bass likely begin to spend more time feeding on the substrate on rocky shoals where round gobies are abundant. The round goby~~

~~also adds a substantial amount of biomass to aquatic ecosystems (Johnson et al.~~

~~2005b; Campbell et al. 2009), so it may actually contribute to an increase the~~

~~reproductive rates and population size of predators that exploit them. 102~~

~~Future work~~

My research has identified rates and seasonality of range expansion, as well as a number of population characteristics at range edges in comparison to the center of their range. However this study was conducted over a two-year period in specific geographical locations, and continuing similar research tracking range expansion may validate these findings. However, now that the Trent-Severn population has coalesced with the Bay of Quinte population, downstream expansion can no longer be studied in this Waterway. The round goby is a short lived fish species (Charlebois et al. 1997); therefore its population dynamics and demographics may be highly variable across time and space, so longer term studies are important for addressing the dynamics of their range expansion. My findings also suggest that the round goby may be declining in established areas of their range in the Trent-Severn, and further research should continue to monitor their mean abundance and sizes to further address this phenomenon.

My studies also did not incorporate juveniles into the sample. It has been speculated that juveniles are involved in range expansion (Ray & Corkum 2001). If this is the case, the detected population front may have been lagging behind the actual front. Considering that smaller round gobies exhibit slower swimming speeds, and migrate to the surface at night (French & Jude 2001; Hoover et al. 2003), juveniles may be more involved in downstream range expansion than upstream.

Future research should incorporate these individuals into the sampling regime. This could be accomplished by supplementing the angling removal method with minnow 103 trap sampling at the edges of the round gobies range, or by using light traps that are often effective at sampling juvenile fishes (Pierce et al. 2006).

Additional factors to consider are the contribution of human mediated dispersal to range expansion. Ballast water is thought to be the vector for round gobies in many areas of the Great Lakes (Charlebois et al. 1997), and bait buckets were likely the cause of their introduction in the Trent-Severn Waterway (Raby et al.

~~2010). Hull transport is also suspected as an important dispersal mechanism for~~

~~Neogobius spp. in European waters (Tsepkin et al. 1992; Sokolov et al. 1994;~~

Wiesner 2005). While the pattern of range expansion observed during this study appears natural, it cannot be certain that these mechanisms did not contribute to the range expansion observed. Future research should investigate the contribution of these vectors to round goby dispersal, particularly the occurrence of hull transport with the round goby.

By continuing to monitor upstream range expansion, the predictions of the transit time dispersal model could be tested. Validation is an important component of the modeling process (Jorgensen & Bendoricchio 2001), however the curt nature of this research only provided enough time to develop the model. There are a number of other factors that could be included in further developing a round goby dispersal model. The most important of these factors may be habitat because round gobies exhibit advection to certain substrates (Ray & Corkum 2001; Johnson et al.

2005a), especially at population fronts (Chapter 2). Models have been developed that incorporate habitat into dispersal rates (Kinezaki et al. 2002), and implementing such a model for round gobies would be beneficial for more accurate predictions of range expansion, as well as understanding mechanisms that affect dispersal. Another factor to consider is ecosystem type, as it appears that round gobies dispersed through Rice Lake faster than the Otonabee River from 2008 to

~~2010, and round gobies do not exhibit high-sustained swimming speeds (Skora~~

1996), so river currents may inhibit their range expansion to some degree. An additional factor to consider in the Trent-Severn Waterway and other similar systems is the effect of navigational locks on dispersal. In Ontario these locks are operational during the summer months, and the only opportunity for movement exists in the same manner as boats are transferred. Not only does this likely inhibit their dispersal to some degree, but it may also decrease genetic flow, which could have detrimental affects on population growth and therefore, long term dispersal rates (Slatkin et al. 1987).

Incorporating the aforementioned factors into a round goby dispersal model may increase its applicability to other ecosystems, although there are certain advantages to the simplicity of the model I developed. In order to incorporate additional factors such as habitat, more extensive data on round goby dispersal through a wide range of habitat types to incorporate this variable with any accuracy.

~~Incorporating such a variable with limited data may actually add to model error~~

~~(Bunnell 1989), and this should be strongly considered when further developing this model based on relatively short term studies. 105~~

~~Based on the findings of my tethering experiments, it appears that round gobies experience low predation rates during the initial stages of their invasion.~~

However, I do not have predator community data to support that these differences are related to the duration of round goby occupation, apposed to intrinsic differences in the predator communities between the two areas. It may be difficult to resolve this issue because even if I had predatory fish community data, and there were more predators in areas of longer round goby occupation, it could not be determined whether this finding was due to intrinsic community differences, or if round gobies were supporting a greater number of their predators. Further research should examine predatory fish communities during a round goby invasion as a before after control impact (BACI) study, paired with predation risk experiments to address this issue. Such a study could also address whether the response of predators to round gobies as prey is functional and/or numerical, and better defining the timeline of this response. 106

~~References~~

Arin, M., Abades, S. R„ Neill, P. E., Lima, M., & Marquet, P. A. (2005). Spread dynamics of invasive species. Proceedings of the National Academy of Sciences of United States of America 103: 374-378.

~~Balshine, S., Verma, A., Chant, V., & Theysmeyer, T. (2005). Competitive interactions between round gobies and logperch. Journal of Great Lakes Research 31: 68-77.~~

~~Baltz, D.M., & Moyle P.B. (1993). Invasion resistance to introduced species by a native assemblage of California stream fishes. Ecological Society of America 3: 246 -~~

~~255.~~

Barton, D. R., Johnson, R. A., Campbell, L., Petruniak, J., & Patterson, M. (2005). Effects of round gobies (Neogobius meianostomus) on dreissenid mussels and other invertebrates in Eastern Lake Erie, 2002-2004. Journal of Great Lakes Research 31:

~~252-261.~~

~~Bartsch, M.R., Bartsch, L.A., & Gutreuter, S. (2005). Strong effects of predation by fishes on an invasive macroinvertebrate in a large floodplain river. Journal of the~~

~~North American Benthological Society 24:168-77. Belanger, R. M., & Corkum, L. D. (2003). Susceptibility of tethered round gobies~~

~~(Neogobius meianostomus) to predation in habitats with and without shelters.~~

~~Journal of Great Lakes Research 29: 588-593.~~

Bergstrom, M.A., Evrard, L.M., & Mensinger, A.F. (2008). Distribution, abundance, and range of the round goby, Apollina melanostoma, in the Duluth-Superior Harbor and St. Louis River Estuary. Journal of Great Lakes Research 34: 535-543.

Bergstrom, M.A., & Mensinger, A.F. (2009). Interspecific resource competition between the invasive round goby and three native species: logperch, slimy sculpin, and spoonhead sculpin. Transactions of the American Fisheries Society 138:1009-

~~1017.~~

Bertolino, S., Lurz, P.W., Sanderson, R., & Rushton, S.P. (2008). Prediction the spread of the American grey squirrel (Sciurus carolinensis) in Europe: A call for a • coordinated European approach. Biological Conservation 141: 2564-2575.

~~Borcherding, J., Staas, S., Kriiger, S., Ondrackova, M., Slapansky, L.S., & Jurajda, P.~~

~~(2011). Non-native Gobiid species in the lower River Rhine (Germany): recent range extensions and densities. Journal of Applied Ichthyology 27:153-155.~~

~~Brammeier, J., Polls, I., Mackey, S. (2008). 2008 project completion report - preliminary feasibility of ecological separation of the Mississippi River and the Great~~

~~Lakes to prevent the transfer of aquatic invasive species. Great Lakes Fishery~~

~~Commission. Chicago, IL. Brittan, M. R., J. D. Hopkirk, J. D. Connors, & Martin, M. (1970). Explosive spread of the oriental goby Acanthogobius flavimanus in the San Francisco Bay-delta region of~~

~~California. Proceedings of the California Academy of Sciences 38: 207-214.~~

Bronnenhuber, J.E., Dufour, B.A., Higgs, D.M., & Heath, D.D. (2011). Dispersal strategies, secondary range expansion and invasion genetics of the nonindigenous round goby, Neogobius meianostomus, in Great Lakes tributaries. Molecular Ecology

~~DOI:10.1111/j.l365-294X.2011.05030.x~~

~~Brown, J. H. (1989). Patterns, modes and extents of invasions by vertebrates. Pages~~

~~85-109 in J. A. Drake, H. A. Mooney, F. di Castri, R. H. Groves, F. J. Kruger, M.~~

~~Rejmanek, and M. Williamson, editors. Biological invasions: a global perspective.~~

~~Wiley, Chichester, UK.~~

~~Bunnell, D.B., Johnson, T.B., & Knight, C.T. (2005). The impact of introduced round gobies (Neogobius meianostomus) on phosphorus cycling in central Lake Erie.~~

~~Canadian Journal of Aquatic Sciences 62:15-29.~~

~~Bunnell, F.L. (1989). Alchemy and uncertainty: what good are models? General technical report. United States Department of Agriculture.~~

Bur, M. T., Stapanian, M. A., Bernhardt, G., & Turner, M. W. (2008). Fall diets of red- breasted merganser (mergus serrator) and walleye (sander vitreus) in Sandusky Bay and adjacent waters of Western Lake Erie. American Midland Naturalist 159:147-

~~161. 109~~

~~Cambray, J.A. (2003). Impact on indigenous species biodiversity caused by the globalisation of alien recreational freshwater fisheries. Hydrobiologia 500: 217-230.~~

~~Campbell, L. M., Thacker, R., Barton, D., Muir, D. C. G., Greenwood, D., & Hecky, R. E.~~

~~(2009). Re-engineering the eastern Lake Erie littoral food web: The trophic function of non-indigenous Ponto-Caspian species. Journal of Great Lakes Research 35: 224-~~

~~231.~~

~~Carlsson, N.O.L., & Strayer, D.L. (2009). Intraspecific variation in the consumption of exotic prey - a mechanism that increases biotic resistance against invasive species?~~

~~Freshwater Biology 54: 2315-2319.~~

~~Carlsson, N.O.L., Sarnelle, 0., & Strayer, D.L. (2009). Native predators and exotic prey~~

~~- an acquired taste? Frontiers in Ecology and the Environment 7: 525-532.~~

Carman, S.M., Janssen, J., Jude, D.J., & Berg, M.B. (2006). Diel interactions between prey behavior and feeding in an invasive fish, the round goby, in a North American river. Freshwater Biology 51: 742-755.

~~Carroll, S.P., Loye, J.E., & Dingle, H. (2005). And the beak shall inherit - evolution in response to invasion. Ecological Letters 8: 944-51.~~

~~Charlebois, P.M., Corkum, L.D., Jude, D.J., & Knight, C. (2001). The round goby~~

~~(Neogobius meianostomus) invasion: current research and future needs. Journal of~~

~~Great Lakes Research 27: 263-266. 110~~

~~Charlebois, P.M., Marsden, J.E., Geottel, R.G., Wolfe, R.K., Jude, D.J., & Rudnicka, S.~~

~~(1997). The round goby, Neogobius meianostomus (Pallas), a review of European~~

~~And North American literature. Illinois-Indiana Sea Grant Program and Illinois~~

~~Natural History Survey. INHS Special Publication. No. 20.~~

~~Chotkowski, M.A., & Marsden, M.A. (1999). Round goby and mottled sculpin predation on lake trout eggs and fry: field predictions from laboratory experiments.~~

~~Journal of Great Lakes Research 25: 26-35.~~

~~Clapp, D.F., Schneeberger, P.J., Jude, D.J., Madison, G., & Pistis, C. (2001). Monitoring round goby (Neogobius meianostomus) population expansion in Eastern and~~

~~Northern Lake Michigan. Journal of Great Lakes Research 27: 335-341.~~

~~Colautti, R.I., Ricciardi, A., Grigorovich, I.A., & Maclsaac, H.J. (2004). Is invasion success explained by the enemy release hypothesis? Ecological Letters 7: 721-33.~~

Cooper, M.J., Ruetz III, C.R. Uzarski, D.G., & Burton, T.M. (2007). Distribution of round gobies in coastal areas of Lake Michigan: are wetlands resistant to invasion? Journal of Great Lakes Research 33: 303-313.

~~Corkum, L.D. (2004). Pheromone signaling in conservation. Aquatic Conservation:~~

~~Marine and Freshwater Ecosystems 14: 327-331.~~

~~Corkum, L. D., Sapota, M. R., & Skora, K. E. (2004). The round goby, Neogobius meianostomus, a fish invader on both sides of the Atlantic Ocean. Biological Invasions~~

~~6:173-181. Ill~~

~~Diaz, D., Zabala, M., Linares, C, & Hereu, B. (2005). Increased predation of juvenile~~

~~European spiny lobster (Palinurus elephas) in a marine protected area. New Zealand~~

~~Journal of Marine and Freshwater Research 39: 447-453. di Castri, F. (1989). History of biological invasions with emphasis on the Old World.~~

~~Pg. 1-30 in J. Drake, F. di Castri, R. Groves, F. Kruger, H. A. Mooney, M. Rejmanek, and~~

~~M. Williamson, editors. Biological invasions: a global perspective. Wiley, New York,~~

~~USA.~~

~~DeRivera, C.E., Ruiz, G.M., Hines, A.H., & Jivoff, P. (2005). Biotic resistance to invasion: Native predator limits abundance and distribution of an introduced crab.~~

~~Ecological Society of America 86: 3364-3376.~~

~~Dietrich, J. P., Morrison, B. J., & Hoyle, J. A. (2006). Alternative ecological pathways in the eastern Lake Ontario food web-round goby in the diet of lake trout. Journal of~~

~~Great Lakes Research 32: 395-400.~~

~~Eros, T., Sevcsik, A., & Toth, 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.~~

~~Faugeras, B., & Maury, 0. (2007). Modeling fish population movements: From an individual-based representation to an advection-diffusion equation. Journal of~~

~~Theoretical Biology 247: 837-848. 112~~

~~French, J. R. P., & Jude, D. J. (2001). Diets and diet overlap of nonindigenous gobies and small benthic native fishes co-inhabiting the St. Clair River, Michigan. Journal of~~

~~Great Lakes Research 27: 300-311.~~

~~Gorman, O.T., & Karr, J.R. (1978). Habitat structure and stream fish communities.~~

~~Ecology 5: 507-515.~~

~~Gutowsky, L.F., Brownscombe, J.W., & Fox, M.G. (2011). Angling to estimate the density of large round goby (Neogobius melamostomus). Fisheries Research 108: 228-~~

~~231.~~

Gutowsky, L.F., & Fox, M.G. Intra-population variability of life history traits and growth during range expansion of the invasive round goby (Neogobius meianostomus). Fisheries Management and Ecology (in review).

~~Gutowsky, L.F., & Fox, M.G. Occupation, body size, and sex ratio of round goby~~

~~(Neogobius meianostomus) in established and newly invaded areas in an Ontario river. Hydrobiologia (in press).~~

~~Hara, T.J. (1994). Olfaction and gustation in fish: an overview. Acta Physiologica~~

~~Scandinavica 152: 207-217.~~

~~Hayden, T.A., & Miner, J.G. (2009). Rapid dispersal and establishment of a benthic~~

~~Ponto-Caspian goby in Lake Erie: diel vertical migration of early juvenile round goby. Biological Invasions 11:1767-1776. 113~~

~~Hensler, S.R., & Jude D.J. (2007) Diel vertical migration of round goby larvae in the~~

~~Great Lakes. Journal of Great Lakes Research 33: 295-302.~~

~~Hogan, L.A.S., Marschall, E., Folt, C, & Stein, R.A. (2007). How non-native species in~~

~~Lake Erie influence trophic transfer of mercury and lead to top predators. Journal of~~

~~Great Lakes Research 33: 46-61.~~

~~Holling, C. S. (1959). The components of predation as revealed by a study of small~~

~~mammal predation of the European pine sawfly. Canadian Entomologist 91: 293-~~

~~320.~~

~~Hoover, J.J., Adams, R.S, & Killgore, K.J. (2003). Can hydraulic barriers stop the spread of the round goby? U.S. Army Engineer Research and Development Center,~~

~~Vicksburg, MS~~

~~Jakubas, D. (2004). The response of the grey heron to a rapid increase of the round goby. Waterbirds27: 304-307.~~

~~Janssen, J., & Jude, D.J. (2001). Recruitment failure of mottled sculpin Cottus bairdi in~~

~~Calumet Harbor, southern Lake Michigan, induced by the newly introduced round goby, Neogobius meianostomus. Journal of Great Lakes Research 27: 319-328.~~

Johnson, J. H., Ross, R. M., McCullough, R. D., & Mathers, A. (2010). Diet shift of double-crested cormorants in eastern Lake Ontario associated with the expansion of the invasive round goby. Journal of Great Lakes Research 36: 242-247. Johnson, L.E., & Padilla, D.K. (1996). Geographic spread of exotic species: ecological lessons and opportunities from the invasion of the zebra mussel, Dreissena polymorpha. Biological Conservation 78: 23-33.

Johnson, T.B., Allen, M., Corkum, L.D., & Lee, V.A. (2005a). Comparison of methods needed to estimate population size of round gobies (Neogobius meianostomus) in western Lake Erie. Journal of Great Lakes Research 31: 78-86.

~~Johnson, T. B., Bunnell, D. B., & Knight, C. T. (2005b). A potential new energy pathway in central Lake Erie: The round goby connection. Journal of Great Lakes~~

~~Research 31: 238-251.~~

~~Jorgensen, S.E., & Bendoricchio, G. (2001). Fundamentals of ecological modeling.~~

~~Developments in Environmental Modeling 21:19-34.~~

~~Jude, D.J. (2001). Round and tubenose gobies: 10 years with the latest great lakes phantom menace. Dreissena! 11:1-14.~~

~~Jude, D.J., Reider, R.H., & Smith, G.R. (1992). Establishment of the Gobiidae in the~~

~~Great Lakes Basin. Canadian Journal of Fisheries and Aquatic Sciences 49: 416-421.~~

Jude, D.J., Janssen, J., & Crawford, G. (1995). Ecology, distribution, and impact of the newly introduced round and tubenose gobies on the biota of the St. Clair and Detroit rivers. In The Lake Huron Ecosystem: Ecology, Fisheries and Management, M.

~~Munawar, T. Edsall, & J. Leach, eds., pp. 447-460. Amsterdam, Netherlands:~~

~~Ecovision World Monograph Series. 115~~

~~Kinezaki, N., Kawasaki, K. Takasu, F., & Shigesada, N. (2002). Modeling biological invasions into periodically fragmented environments. Theoretical population biology~~

~~64: 291-302.~~

~~King, R.B., Ray, J. M., & Stanford, K. M. (2006). Gorging on gobies: beneficial effects of alien prey on a threatened vertebrate. Canadian Journal of Zoology 84:108-115.~~

~~Kneib, R.T., & Scheele, C.E.H. (2000). Does tethering of mobile prey measure relative predation potential? An empirical test using mummichogs and grass shrimp. Marine~~

~~Ecology Progress Series 198:181-190.~~

Knight, C. (1997). The round goby in the central basin of Lake Erie: range expansion and size-selective predation on zebra mussels. In the Round goby, Neogobius meianostomus (Pallus): a review of European and North American literature, eds.

~~P.M. Charlebois, J.E. Marsden, R.G. Goettel, R.K. Wolfe, D.J. Jude, and S. Rudnicka, pp.~~

~~45-46. Urbana, IL: Illinois Natural History Survey.~~

~~Koh, L.P., & Menge, N.L. (2006). Rapid assessment of Lepidoptera predation rates in neotropical forest fragments. Biotropica 38:132-134.~~

~~Kolar, C.S., & Lodge, D.M. (2002). Predictions and risk assessment for alien fishes in~~

~~North America. Science 298:1233-1236.~~

Kovac, V., Copp, G.H., & Sousa, R.P. (2009). Life-history traits of invasive bighead goby Neogobius kessleri (Gunther, 1861) from the middle Danube River, with a reflection on which goby species may win the competition. Journal of Applied

~~Ichthyology 25: 33-37. Kornis, M.S., & Vander Zanden, M.J. (2010). Forecasting the distribution of the~~

~~invasive round goby (Neogobius meianostomus) in Wisconsin tributaries to Lake~~

~~Michigan. Canadian Journal of Fisheries and Aquatic Sciences 67: 553-562.~~

~~Kraus, F., Campbell, E.W., Allison, A. & Pratt, T.K. (1999) Eleutherodactylus frog~~

~~introductions to Hawaii. Herpetological Review 30: 21-25.~~

~~Krumbein, W.C., & Sloss, L.L. (1951). Stratigraphy and Sedimentation. Freeman, San~~

~~Francisco.~~

~~Kwon, T.D., Fisher, S.W., Kim, G.W., Hwang, H., & Kim, J.E. (2006). Trophic transfer~~

~~and biotransformation of polychlorinated biphenyls in zebra mussel, round goby,~~

~~and smallmouth bass in Lake Erie, USA. Environmental Toxicology and Chemistry 25:~~

~~1068-1078.~~

~~L'avrincikova, M., & Kovac, V. (2007). Invasive round goby Neogobius meianostomus from the Danube mature at small size. Journal of Applied Ichthyology 23: 276-278.~~

~~Lee, V.A., & Johnson, T.B. (2005). Development of a bioenergetics model for the round goby (Neogobius meianostomus). Journal of Great Lakes Research 31:125-134.~~

Lederer, A.M., Janssen, J., Reed, T., & Wolf, A. (2008) Impacts of the introduced round goby (Apollonia melanostoma) on dreissenids (Dreissena polymorpha and Dreissena burgensis) and on macroinvertebrate community between 2003 and 2006 in the littoral zone of Green Bay, Lake Michigan. Journal of Great Lakes Research 34: 690-

~~697. Lester, N.P., Dextrase, A.J., Kushneriuk, R.S., Rawson, M.R., & Ryan, P.A. (2004). Light~~

~~and temperature: key factors affecting walleye abundance and production.~~

~~Transactions of the American Fisheries Society 133: 588-605.~~

~~Leung, B., Lodge, D.M., Finnoff, D., Shogren, J.F., Lewis, M.A., & Lamberti, G. (2002).~~

~~An ounce of prevention or a pound of cure: bioeconomic risk analysis of invasive species. Proceedings of the Royal Society 269: 2407-2413.~~

~~Lobb, M.D., & Orth, D.J. (1991). Habitat use by an assemblage of fish in a large warmwater stream. Transactions of the American Fisheries Society 120: 65-78.~~

~~Lodge, D.M. (1993). Species invasions and deletions: community effects and responses to climate and habitat change. In: Kareiva, P.M., Kingsolver, J.G., & Huey,~~

~~R.B. (eds) Biotic Interactions and Global Change, pp 367-387. Sinauer Associates,~~

~~Sunderland, Massachusetts.~~

Lodge, D. M. (2001). Responses of lake biodiversity to global changes. In Future scenarios of global biodiversity (ed. F. S. Chapin III, O. E. Sala & E. Huber-Sannwald), pp. 277-312. New York: Springer.

~~Maclnnis, A.J., & Corkum, L.D. (2000). Fecundity and reproductive season of the round goby Neogobius meianostomus in the Upper Detroit River. Transactions of the~~

~~America Fisheries Society 129:136-144.~~

~~Mack, R.N., Simberloff, D., Lonsdale, W.M., Evans, H., Clout, M., & Bazzaz, F.A. (2000).~~

~~Biotic invasions: causes, epidemiology, global consequences, and control. Ecological~~

~~Applications 10: 689 - 710. 118~~

~~Manderson, J.P., Pessutti, J., Hilbert, J.G., & Juanes, F. (2004). Shallow water predation risk for a juvenile flatfish (winter flounder; Pseudopleuronectes americanus,~~

~~Walbaum) in a northwest Atlantic estuary. Journal of Experimental Marine Biology and Ecology 304:137-157.~~

~~Matis, J.H., Grant, W.E., Miller, T.H. (1992a). A semi-Markov process model for migration of marine shrimp. Ecological Modelling 60:167-184.~~

~~Matis, J.H., Rubink, W.L., & Makela, M. (1992b). Use of the gamma distribution for predicting arrival times of invading insect populations. Environmental Entomology~~

~~21:436-440.~~

~~Middleton, M. J. (1982). The oriental goby, Acanthogobius flavimanus (Temminck and Schlegel), introduced fish in the coastal waters of New South Wales, Australia.~~

~~Journal of Fish Biology 21: 513-523.~~

~~Miller, P.J. (1984). The tokology of gobioid fishes. In Fish reproduction strategies and tactics, G.W. Potts and R.J. Wooton, eds. pp. 119-153. London: Academic Press.~~

~~Mills, E.L., Casselman, J.M., Dermott, R., Fitzsimons, J.D., Gal, G., Holeck, K.T. (2003).~~

~~Lake Ontario: food web dynamics in a changing ecosystem (1970-2000). Canadian~~

~~Journal of Fisheries and Aquatic Sciences 60: 471-490.~~

~~Miner, J.G., Farver, J.R. (2004). Quantifying round goby dispersal mechanisms: impacts on mitigation efforts. Final Report for Lake Erie Protection Fund Grant, Sea~~

~~Grant 185-02. 119~~

~~Minns, C.K., Moore, J.E., & Seifried, K.E. (2004). Nutrient load and budgets in the Bay of Quinte, Lake Ontario, 1972-2001. Canadian Manuscript Report of Fisheries and~~

~~Aquatic Sciences 2694, Great Lakes Laboratory for Fisheries and Aquatic Sciences,~~

~~Fisheries and Oceans Canada, Burlington, ON.~~

~~Mooney, H.A., & Drake, J.A. (1989). Biological invasions: a SCOPE program overview.~~

~~Biological invasions: a global perspective (ed. by J.A. Drake, H.A. Mooney, F. di Castri,~~

~~R.H. Groves, F.J. Kruger, M. Rejmanek and M. Williamson), pp. 491-508. John Wiley &~~

~~Sons, New York.~~

~~Murdoch, W.W. (1975). Switching in predatory fish. Ecology 56:1094-1105.~~

~~Nentwig, W. (2007). Biological invasions. Springer, Berlin. Pg. 77-90.~~

~~Olden, J.D., Poff, N.L., & Bestgen, KR. (2006). Life-history strategies predict fish invasions and extirpations in the Colorado River basin. Ecological Monographs 76:~~

~~25-40.~~

~~OMNR. (2001). Fish metapopulations in the Trent and Otonabee River portions of the Trent-Severn Waterway. Watershed Science Center, Trent University, ON. CA.~~

~~Otto, S.P., & Day, D. (2007). A biologist's guide to mathematical modeling in ecology and evolution. Princeton University Press, New Jersey.~~

~~Parker, I. M., Simberloff, D., Londsdale, W.M., Goodell, K., Wonham, M., Kareiva, P.M.,~~

~~Williamson, M. H., Von Holle, B., Moyle, P. B., Byers, J. E., & Goldwasser, L. (1999). 120~~

~~Impact: toward a framework for understanding the ecological effects of invaders.~~

~~Biological Invasions 1: 3-19.~~

~~Peirce, R.B, Shroyer, S., Pittman, B., Logsdon, D.E., & Kolander, T.D. (2006).~~

~~Catchability of larval and juvenile northern pike in Quatrefoil light traps. North~~

~~American Journal of Fisheries Management 26: 908-915.~~

~~Pennuto, CM., Krakowiak, P.J., & Janik, C.E. (2010). Seasonal abundance, diet, and~~

~~energy consumption of round gobies (Neogobius meianostomus) in Lake Erie~~

~~tributary streams. Ecology of Freshwater Fish 19: 206-215.~~

~~Petrie, S.A., & Knapton, R.W. (1999). Rapid increase and subsequent decline of zebra~~

~~and quagga mussels in Long Point Bay, Lake Erie: possible influence of waterfowl~~

~~predation. Journal of Great Lakes Research 25: 772-82.~~

~~Phillips, E.C., Washek, M.E., Hertel, A.W, & Niebel, B.M. (2003). The round goby~~

~~(Neogobius meianostomus) in Pennsylvania tributary streams of Lake Erie. Journal of~~

~~Great Lakes Research 29: 34-40.~~

~~Pimentel, D., Lach, L., Zuniga, R. & Morrison, D. (1999) Environmental and economic costs of nonindigenous species in the United States. Bioscience 50: 53-65.~~

~~Pimentel, D., Zuniga, R., & Morrison, D. (2005). Update on the environmental and economic costs associated with alien invasive species in the United States. Ecology and Economics 52: 273-288. 121~~

Poos, M., Dextrase, A.J., Schwalb, A.N., & Ackerman, J.D. (2009). Secondary invasion of the round goby into high diversity Great Lakes tributaries and species at risk hotspots: potential new concerns for endangered freshwater species. Biological

~~Invasions. DOI 10.1007/sl0530-009-9545-x.~~

Pothoven, S.A., Nalepa, T.F., Schneeberger, P.J., & Brandt, S.B. (2001). Changes in diet and body condition of lake whitefish in southern Lake Michigan associated with changes in benthos. North American Journal of Fisheries Management 21: 876-83.

~~Probst, W.E., Rabeni, C.F., Covington, W.G., & Marteney, R.E. (1984). Resource use by stream-dwelling rock bass and smallmouth bass. Transactions of the American~~

~~Fisheries Society 113: 283-294.~~

Raby, G. D., Gutowsky, L. F. G., & Fox, M. G. (2010). Diet composition and consumption rate in round goby (Neogobius meianostomus) in its expansion phase in the Trent River, Ontario. Environmental Biology of Fishes 89:143-150.

~~Ray, W.J., & Corkum, L.D. (1997). Predation of zebra mussels by round gobies,~~

~~Neogobius meianostomus. Environmental Biology of Fishes 50: 267-273.~~

~~Ray, W.J., & Corkum, L.D. (2001). Habitat and site affinity of the round goby. Journal of Great Lakes Research 27: 329-334.~~

~~Ricciardi, A. (2001). Facilitative interactions among aquatic invaders: is an~~

"invasional meltdown" occurring in the Great Lakes? Canadian Journal of Fisheries and Aquatic Sciences 58: 2513-2525. Riley, J.W., & Marsden, J. (2008). Predation pressure on emergent lake trout fry in lake champlain and techniques for assessing lake trout reproduction in deep-water habitats. Master of Science Thesis, University of Vermont. http://hdl.handle.new/123456789/97.

~~Rollo, A., Andraso, G., Janssen, J., & Higgs, D. (2006). Attraction and localization of round goby (Neogobius meianostomus) to conspecific calls. Behaviour 144:1-21.~~

~~Rypel, A.L., Layman, C.A., & Arrington, D.A. (2007). Water depth modifies relative predation risk for a motile fish taxon in Bahamian tidal creeks. Estuaries and Coasts~~

~~30:1-8.~~

~~Sakai, A.K., Allendorf, F.W., Holt, J.S., Lodge, D.M., Molofsky, J., With, K.A., Baughman,~~

~~S., Cabin, R.J., Cohen, J.E., Ellstrand, N.C., McCauley, D.E., O'Neil, P., Parker, I.M.,~~

~~Thompson, J.N., & Weller, S.G. (2001). The population biology of invasive specie.~~

~~Annual Review of Ecology and Systematics 32: 305-332.~~

~~Schaner, T., Fox, M.G., & Taraborelli, A.C. (2009). In inexpensive system of underwater video surveys of demersal fishes. Journal of Great Lakes Research 35:~~

~~317-319.~~

Schippers P., Verboom, J., Knaapen, J.P., & Apledoorn, R.P. (1996). Dispersal and habitat connectivity in complex heterogeneous landscapes: an analysis with a GIS- based random walk model. Ecography 19: 97-106.

~~Sharov, A.A., & Liebhold, A.M. (1998). Model of slowing the spread of gypsy moth~~

~~(lipidoptera: lymantriidae) with a barrier zone. Ecological Applications 8:1170-1179. 123~~

~~Shigesada, N., & Kawasaki, K. (1997). Biological invasions: Theory and practice.~~

~~Oxford University Press p. 61-69, 205-230.~~

~~Shigesada,N., Kawasaki, K., & Takeda, Y. (1995). Modeling stratified diffusion in biological invasions. The American Naturalist 146: 229-251.~~

~~Simberloff, D. (2003). How much information on population biology is needed to manage introduced species? Conservation Biology 17: 83-92.~~

~~Simberloff, D., & Gibbons, L. (2004). Now you see them, now you don't! - population crashes of established introduced species. Biological Invasions 6:161-172.~~

~~Simberloff, D., & Stiling, P. (1996) How risky is biological control? Ecology 77:1965-~~

~~1974.~~

~~Simberloff, D., & Von Holle, B. 1999. Positive interactions of nonindigenous species: invasional meltdown? Biological Invasions 1: 21-32.~~

~~Skellam, J.G. (1951). Random dispersal in theoretical populations. The Nature~~

~~Conservancy, London. 196 - 217.~~

~~Skora K.E. (1996). In Report of the working group on introduction and transfers of marine organisms (WGITMO) ICES Annual Science Conference, Reykjavik, Iceland, pp.~~

~~96-107.~~

~~Skora, K. E., & J. Stolarski. (1993). New fish species in the Gulf of Gdansk, Neogobius sp. [cf. Neogobius meianostomus (Pallas 1811)]. Bulletin of the Sea Fisheries Institute~~

~~1:83. 124~~

~~Slatkin, M (1987). Gene flow and the geographic structure of natural populations.~~

~~Science 236: 787-792.~~

~~Sokolov, L.I., Sokolova, E.L. & Pegasov. V.A. (1994). Ichthyofauna of the Moskva River within Moscow City and some data on its state. Voprosy Ikhtiologii. 35: 364-641. In~~

~~The round goby (Neogobius meianostomus): another unwelcome invader in the~~

~~Mississippi River Basin, eds. Steingraeber, M.T., & Thiel, P.A. U.S. Wildlife Service, WI.~~

~~Somers, C. M., Lozer, M. N., Kjoss, V. A., & Quinn, J. S. (2003). The invasive round goby~~

~~(Neogobius meianostomus) in the diet of nestling double-crested cormorants~~

~~(phalacrocorax auritus) in Hamilton Harbour, Lake Ontario. Journal of Great Lakes~~

~~Research 29: 392-399.~~

~~Stammler, K.L., & Corkum, L.D. (2005). Assessment offish size on shelter choice and~~

~~intraspecific interactions by round gobies Neogobius meianostomus. Environmental~~

~~Biology of Fishes 73:117-123.~~

~~Steinhart, G.B., Marshall, E.A., & Stein, R.A. (2004a). Round goby predation on~~

~~smallmouth bass offspring in nests during simulated catch and release angling.~~

~~Transactions of the American Fisheries Society 133:1100-1106.~~

Steinhart, G. B., Stein, R. A., & Marschall, E. A. (2004b). High growth rate of young-of- the-year smallmouth bass in Lake Erie: A result of the round goby invasion? Journal of Great Lakes Research 30: 381-389. Steingraeber M.T., & Thiel, P.A. (2000). The round goby (Neogobius meianostomus): another unwelcome invader in the Mississippi River Basin. Transactions of the 65th

~~North American Wildlife and Natural Resources Conference, p 328-344.~~

~~Taraborelli, A. C, Fox, M. G., Johnson, T. B, & Schaner, T. (2010). Round goby~~

~~(Neogobius meianostomus) populations structure, biomass, prey consumption and mortality from predation in the Bay of Quinte, Lake Ontario. Journal of Great Lakes~~

~~Research 36: 625-632.~~

~~Taraborelli, A. C, Fox, M. G., Schaner, T., & Johnson, T. B. (2009). Density and habitat use by the round goby (apollonia melanostoma) in the Bay of Quinte, Lake Ontario.~~

~~Journal of Great Lakes Research 35: 266-271.~~

~~Taylor, R.A.J. (1978). The relationship between density and distance of dispersing insects. Ecological Entomology 3: 63-70.~~

~~Thorp, J.H., Delong, M.D., & Casper, A.F. (1998). In situ experiments on predatory regulation of a bivalve mollusk (Dreissena polymorpha) in the Mississippi and Ohio~~

~~Rivers. Freshwater Biology 39: 649-61.~~

~~Trakhtenbrot, A., Ran, N., Perry, G.; & Richardson, D.M. (2005). The importance of long-distance dispersal in biodiversity conservation. Diversity and Distributions 11:~~

~~173-181. 126~~

Truemper, H. A., & Lauer, T. E. (2005). Gape limitation and piscine prey size- selection by yellow perch in the extreme southern area of Lake Michigan, with emphasis on two exotic prey items. Journal of Fish Biology 66:135-149.

~~Tsepkin, E.A., Sokolov, L.I., & Rusalimchik. A.V. (1992). Ecology of the round goby~~

~~Neogobius meianostomus (Pallas), an occasional colonizer of the basin of the Moskva~~

~~River. Biologicheskie Nauki 1: 46-51. In The round goby (Neogobius meianostomus): another unwelcome invader in the Mississippi River Basin, eds. Steingraeber, M.T., &~~

~~Thiel, P.A. U.S. Wildlife Service, WI.~~

~~Vander Zanden, M.J. & Vadeboncoeur, Y. (2002). Fishes as integrators of benthic and pelagic food webs in lakes. Ecology 83: 2152-2161.~~

~~Velez-Espino L.A., Koops M.A., Balshine S. (2010) Invasion dynamics of round goby~~

~~(Neogobius meianostomus) in Hamilton Harbour, Lake Ontario. Biological Invasions~~

~~12: 386-3875.~~

~~Weimer, M.T., & Sowinski, M. (1999). Diet of the round goby (Neogobius meianostomus) in Lake Erie. Dreissena! 10: 7-12.~~

~~Weiss, M.R., Wilson, E.E., & Castellanos, I. (2004). Predatory wasps learn to overcome the shelter defenses of larval prey. Animal behavior 68: 45-54.~~

~~Wiesner, C. (2005). New records of non-indigenous gobies (Neogobius spp.) in the~~

Austrian Danube. Journal of Applied Ichthyology 21: 324-327. Wolf, R.K., & Marsden, J.E. (1998). Tagging methods for the round goby (Neogobius melamostomus). Journal of Great Lakes Research 24: 731-735.

~~Wootten, R.J. (1998). Ecology of teleost fishes. Kluwer, Norwell, MA. pg 123, 238~~

~~Yamamura, K. (2002). Dispersal distance of heterogeneous populations. Population~~

~~Ecology 44: 93-101.~~

~~Young, J. A. M., Marentette, J. R., Gross, C, McDonald, J. I., Verma, A, Marsh-Rollo, S. E.~~

~~(2010). Demography and substrate affinity of the round goby (Neogobius meianostomus) in Hamilton Harbour. Journal of Great Lakes Research 36:115-122.~~

~~Zar, J.H. (2010). Biostatistical Analysis. Pearson Prentice Hall, New Jersey, pg. 262,~~

~~432-437 128~~

~~Appendix A: Abundance stats: Table 1A,B: Nested ANOVA and Unequal N HSD comparing abundance between range locations and over time~~

Univariate Tests of Significance for Var3 (Spreadshei Over-parametenzed model Type III decomposition SS Degr. of MS Effect Freedom Intercept 791045 1 791045 1007 69^0 00000 •Van" 14796 2 7398'. 9 42^0 00010 "Var2"(Var1) 275581 9 _ 3062" 3 JO'' 0 00010 Error 259836* 331, "" 785C ™ Unequal N HSD; variable Var3 (Spreadshee Approximate Probabilities for Post Hoc Test; Error: Between MS = 7850.0. df = 331.00 Varl {1> {2} {3} Cell No. 141.75 154.06 187.43 Cente 0.640191 0 00241 Upstrea 0.640191 0 01064 Downstrea 0 00241 0 01064

~~Table 2A,B,C,D: Kruskal-Wallis tests comparing abundance between range locations in May, August of 2009 and 2010~~

Multiple Compansonsp values(2-tailed); Var3 (Spreadshe ! Independent (grouping) variable: Varl Kruskal-Wallistest: H ( 2, N= 46) =10.96565 p =.0042 Depend. Center Upstream Downstream Var3 R:30.54£ R:18.464 R:16.455 Center 0.02723 0.01436 Upstream 0.02723 1.000001 Downstream 0.01436 1.000001 Multiple Comparisonsp values (2-tailed); Var3 (Spreadshee Independent (grouping) variable: Varl Kruskal-Wallistest: H (2, N= 77) =16 40078 p = 0003 Depend.- Center Upstream Downstream Var3 R:42.523 R:48.862 R:25.019 Center 0.94869. 0 02074' Upstream 0.94869; 0 00023, Downstream 0 02074' 0 00023, Multiple Comparisons p values (2-tailed); Var3 (Spreadshee ! Independent (grouping) variable: Varl Kruskal-Wallistest: H ( 2, N= 105) =3 357702 p =.1866 Depend.: Center Upstream Downstream Var3 R:63.45C R:52.600 R:48.225 Center 0.55484 0.203801 Upstream 0.55484 1.000001 Downstream 0.20380I 1.000001 Multiple Comparisons p values (2-tailed); Var3 (Spreadshee ! Independent (grouping) variable: Varl Kruskal-Wallistest- H ( 2. N= 115) =1.972153 p =.3730 Depend.: Center Upstream Downstream Var3 R:63.40S R:60.217 R:52.366 Center 1.000001 0.65183 Upstream 1.000001 0 77269 Downstream 0.65183: 0 77269. Table 3A: Nested ANOVA and Unequal N HSD comparing length between range locations and over time

Univariate Testsof Significance for Var4 (Spreadshee Over-parameterized model Type III decomposition SS Degr. of MS F P Effect Freedom Intercept 654894' 1 654894 1000.41' 0.00000' •Varl" 1894: 2 9471 1.447 0.23684- "Var2"(Varn 94165: 9 10462S 15.98:0 00000' Error 210787 322 654£

~~Table 4A,B,C,D: Kruskal-Wallis tests comparing abundance between range locations in May, August of 2009 and 2010~~

Multiple Comparisons p values (2-tailed); Var3 (Spreadshee Independent (grouping) variable. Varl Kruskal-Wallistest: H ( 2, N= 42) =10.72483 p =.0047 Depend.: Center Upstream 1 Downstream Var3 R:27 97e R:16.364l R:14.864 Center 0 03506 0 01323- Up stream,. 0 03506, 1 00000! Downstream 0 01323 1.000001 Multiple Comparisons p values (2-tailed); Var3 (Spreadshee Independent (grouping) variable: Varl Kruskal-Wallistest: H ( 2, N= 75) =13.87457 p =.0010 Depend.: Center Upstream Downstream Var3 R:26.90S R:49.685 R:35.250 Center 0.00082 0.55937' Upstream 0 00082' 0 04779 Downstream 0.55937' 004779s Multiple Compansonsp values (2-tailed), Var3 (Spreadshee : Independent (grouping) vanable- Varl Kruskal-Wallistest. H(2, N= 104) =7.321194 p =.0257 Depend.. Center Upstream Downstream Var3 R:69.421 R:48.489 R48.975 Center 0 03361 0 04497 Upstream 0 03361 1.000001 Downstream 0 04497 1.00000' Multiple Comparisons p values (2-tailed); Var3 (Spreadshee Independent (grouping) vanable: Varl Kruskal-Wallistest- H ( 2, N= 104) =1.255297 p =.5338 Depend • Center Upstream Downstream Var3 R:48.071 R: 51.500 R:57.177 Center < 1 00000I, 0.85648 Upstream 1.000001 \ 1.000001 Downstream 0.85648 1.000001 Upstream dispersers: Table 5A,B - One-way ANOVA and Unequal N HSD post hoc analysis comparing sex ratios between Otonabee 2010 (beyond edge of 2009 range), Rice 2010 (occupied 2009), CORE, and DSE sites

Univanate Tests of Significance forVar2 (Spreadsheets) Sigma-restncted parametenzation Effective hypotheas decomposition SS Degr of MS Effect Freedom 121900 1 121900 1085 31 0 00000 2 802* 4 0 700-. 6 23£ 0 00007 42 568! 379 0 112: Unequal N HSD; vanable Var2 (Spreadsheets) Probabilities for Post Hoc Tests Error: Between MS = .11232, df = 379.00 Varl {1} {2} {3} {4} {5} Cell No .73492 .59001 .64909 .43508 .64052 1 Cente f 0.04260 0^46537 0 00027^0.659251 upstrearr 0.04260J 0.78396 OJ9J781 0.95131 downstreart 0.46537"! 0.78396 0J230JI 0.99994' Otonabee 20J 0 00027~\ QJ9\78: 0~"0230U 0 03293I Rice 20id 0.65925I 0.95131~ 0.99994' 0 03293I

Table 6A,B - Kruskal Wallis ANOVA comparing mean sizes between Otonabee 2010 (beyond edge of 2009 range), Rice 2010 (occupied 2009), CORE, USE, and DSE range locations Kruskal-Wallis ANOVA by Ranks; Var3 (Spreadsheetl) Independent (grouping) vanable: Varl Kruskal-Wallistest: H(4, N= 431) =21.27986 p =.0003 Depend.: Code Valid Sum of Var3 N Ranks Center 101 82s 20906.0I Upstream 102 13E 2939Z.5I Downstream 102 117 25972.0I Rice 104 48( 9114.01 Otonabee 10£ 49 7711.51 Multiple Comparisons p values (2-tailed); Var3 (Spreadsheets I Independent (grouping) variable: Varl Kruskal-Wallis test: H ( 4, N= 431) =21.27986 p =.0003 Depend.: (Center jwnstream . Rice I Otonabee Var3 fc254.95 R: 189.88 Rtl5?Jl Centei 0.327861 0.661032 OJ04044J 0.000144 0.327861 1.00000C 1.00000C 0 036756 .JOdwftlream.. 0.66jq32_JL00000C! Jj-000-00? J^L230-37 p.o4044^_i .oboood 1.000000 ^ \ 1JP0000C Otonabee 0 00014^ 0.036756 0 023037 LOOOOOfT 131

~~Habitat: Table 7A - Chi square observed vs expected analysis comparing USE to CORE site occupancy across five levels of rock cover~~

Observed vs. Expected Frequencies (Spreadsheet Chi-Square = 23.66544 df = 4 p < .000093 NOTE: Unequal sumsof obs. & exp. frequencies observed expected O -E (0-E)**2 Case use core /E C: 1 17.948: 46.153! -28.205' 17.2364' C: 2 75.000< 90.909- -15.909! 2.7840! C: 3 89.285", 100.0001 -10.714: 1.1479( C: 4 89.473: 100.0001 -10.526: 1.1080: C: 5 100.0001 88.888! 11.111' 1.3888! Sum 371.7081 425.95HI -54.243:123.6654-

~~Table 8A - Chi square observed vs expected analysis comparing DSE to CORE site occupancy across five levels of rock cover~~

Observed va Expected Frequencies (Spreadsheet Chi-Square = 33.16286 df = 4 p < .000001 NOTE: Unequal sumsof obs. & exp. frequencies observed expected O-E (0-E)**2 Case dse core /E C: 1 11.828( 46.153J -34.325!^ 25.5291 C: 2 71.739; 90.909; -19.17011 4.0423J C: 3 92.000( 100.0001 -8.000C 0.6400( C: 4 87.500T loO.OOoT -12.500( l: 1.5625( C: 5 100.0001 88.888! 11.111) 1.3888! Sum 363.067 425.9511 -62.884^33.16281

~~Table 9A - Chi square observed vs expected analysis comparing DSE to USE site occupancy across five levels of rock cover~~

Observed vs. Expected Frequencies (Spreadsheet Chi-Square = 2.355092 df = 4 p < .670758 NOTE: Unequal sumsof obs. & exp. frequencies observed expected O-E (0-E)**2 Case dse use m C: 1 11.828(j 17.948: -6.1207(2.08726- C: 2 71.739' 75.000(J -3.2608:0.14177; C: 3 92.000C 89.285:, 2.7142<"o~08251- C: 4 87.500(, 89.473: -1.9736J 0.04353' C: 5, 100.0001! 100.0001* 0.0000( 0.000001 Sum 363.0671 371.708s -8.6410: 2.35509: Table 10A - Kruskal-Wallis test comparing round goby abundance between substrate types in CORE range location

Multiple Compansonsp values (2-tailed); catch (Spreadshee Independent (grouping) variable: substrate Kruskal-Wallistest: H ( 3, N= 62) =2.538782 p =.4683 Depend.: boulder cobble gravel 1 sand catch R39.60C R:33.904 R28.55CI R28.00C boulder LOOOOOi 1.00000'' LOOOOOi cobble LOOOOOt ' 1.00000' LOOOOOi gravel 1.000001 ^oodooT " "1.000001 sand 1.000001 LOOOOOi LOOOOOi;

~~Table 11A - Kruskal-Wallis test comparing round goby abundance between substrate types in USE range location~~

Multiple Comparisons p values (2-tailed); catch (Spreadsheel! Independent (grouping) variable: substrate Kru ska I -Wa llistest: H ( 3, N= 125) =48.28736 p =.0000 Depend, boulder cobble gravel sand catch R:92.75C R:86.562 R: 100.96 R:46.101 boulder LOOOOOi; LOOOOOi 0.00311 cobble 1.00000' il.OOOOOi 0.00001 i.poopoi i.oooppl _T„'pjoppoq 0.00311 0.00001^0.00000 ""

~~Table 12A - Kruskal-Wallis test comparing round goby abundance between substrate types in DSE range location~~

Multiple Comparisons p values (2-tailed); catch (Spreadshee Independent (grouping) variable: substrate Kruskal-Wallistest: H ( 3, N= 107) =18.38770 p =.0004 Depend.: boulder cobble gravel sand catch R:63.472 R:72.921 R:74.333 R:43.812 boulder 1.00000' ^1-Q.oPQQj,J) . 10541' cobble LOOOOOi' l!,oppbqi JIJD0198, gravel LOOOOOi LOOOOOi ^0j"2748j sand 0.10541' 0.00198 O.A 2748- 133

~~Chapter 4:~~

Table 13A - Two-way ANOVAs of round goby length (A), tether depth (B), river current (C), and temperature (C) comparing range locations (CORE and DSE) and habitat types (SRS, DRS, SS). Var 1 = range location, Var 2 = habitat.

~~Univariate Tests of Significance for Var3 (Spreadsheetl i Sigma-restricted parameterization Effective hypothesis decomposition~~

~~Effect ltdori Jit* mOSm. 137410.2 , 11 137410.C- 14267.3£ 0.OOO00C I* Mi 1.1 f 1.1j ~ 0.12 0.738432~~

~~4.0 2; y 2.0J 021 0.81390C ^ ...... ^,., _._. 196018306j 115.6 12 9.6~~

B Univariate Tests of Significance for Var3 (Spreadsheet Sig ma-restricted parameterization Effective hypothesis decomposition IP9MK 1 Effect IX** r-ij 'lit irjefeept; 1772.506! 1 1772.506; 6090.707, O.OOOOOC ' 1 .206J 1 1206J 4.146 0.064437 £}* :3- ! 25.437J '"' Z .12T71S;"" .43,704 0,00000c ~ 0.2171 2 O.IO9] 6.373; 0.69633E Bffori 3.492: 12 0.291 i

Univariate Tests of Significance for Var3 (Spreadsheetl Sig ma-restricted parameterization Effective hypothesis decomposition ISS Degr|pf; *If IP Effect Freed'onl Intercept 0.30026J 1 0.30026^374.8787, O-OOOOQC 'Varl' op 0020T ito.0002W" 02552 6.622617 Var2" q.ooop7£: 2! 0.00048S "I 0.610? P.55891J "Vart^Vara" 0.000867^ 2| 0.000434 0.5AU 0.595494 Error 0.00961 a 12l 0.000801'

~~D Univariate Tests of Significance for Var3 (Spreadsheetl | Sig ma-restricted parameterization Effective hypothesis decomposition~~

Effect tntelgept * 9135.01^ 1,9135.01-1 4474.29^, O.OOOOOC "Var1» 1,681 1^ 1-681 0.823 0,38212£ "Var2" *Var1"*'Var2" " 3.861 ^ . " "~ '"£ " '"lill" CL946"0.415582 Error 24.50C 12! 2.042 ". Table 14A: Two-way ANOVA of predation rates comparing range locations (CORE and DSE) and habitat types (SRS, DRS, SS) used for tethering sites in the Trent River in June and July 2010.

Univariate Tests of Significance for Var3 (Spreadsheetl) Sig ma-restricted parameterization Effective hypothesis decomposition *SSS Degr. of Effect Freedom m> intgffiii HEEL 9.07343S 1^07343|JM.677jLg.0000pC aatfjj 0.620746; 1 0,620741., 6a523|0p0p00£ 0.19440^1 2" 0.097202 9.4774J 0.003394 0.025092; 2; 0.012546 1.2233 0.32848C Irer 0.12307E 12 0.010256

~~Table 15A: Post hoc analysis of predation rates comparing range locations (CORE and DSE) and habitat types (SRS, DRS, SS) from tethering experiments in the Trent~~

~~River in June and July of 2010.~~

Newman-Keuls test; variable Var3 (Spreadsheetl) Approximate Probabilities for Post Hoc Tests Error: Between MS = .01026, df = 12.000 HP- Var2 {1} p?yip| Cell No. 1,0210 liS &701 Centej SRS P,5ir456[0.0p5976 P^Pj011^^.00JJ\WJjOOp257 Cjentej DRSI 0J51145C '"o;60789^ 0.00208J pppn2iPP003K Centei SS 0.005976 01)67894; 0.259867 0/118033?0.02682E Edge ..SRS PPQ113C 04p20^i" 025986V ]"!".'. ^816jp£p.l2121J Edge DRS] 0F0?106 0.66?72S 0.18033C 0.81680S; ™J 0.179596 Edge SS 0.000257 0.000352] 0.02682S! 0.1212It 0.179596; 135

~~Table 16A: Test of homogeneity of variances in the relationship between predation~~

~~rate and predator catch rate, comparing range locations (CORE and DSE) in round~~

~~goby tethering experiments in the Trent River in June and July 2010.~~

Univariate Tests of Significance for Var3 (Spreadshee Sigma-restncted parameterization Effective hypothesis decomposition SS 1 Degr. of MS F P Effect 1 Freedom Intercept 0.03£i 1 0.035<* 0.00045 0.98325. "Varl" 17.282 1 17.281" 0.21999 0.64626. "Var2" 299.94: 1 299.942 3.81827,0.07097. "Var1"*"Var2" 0.08£l 1 0.085- 0 0010810.97421. Error 1099.76'l 14 78.554:

~~Table 17A: t-test of mean size of predated and survived round gobies from tethering~~

~~experiments at CORE (A) and DSE (B) range locations in the Trent River in June and~~

~~July of 2010.~~

T-test for Independent Samples (Spreadsheetl) A Note Vanableswere treated as independent samples Mean | Mean t-value df P Valid N Valid N Std.Oev. Std Dev F-ratio P Group 1 vs Group . Group 11 Group 2 Group 1 Group 2 Group 1 Group 2 Variances Variances Varl vs. Var2 87 6594 88 5476 -0 46475' 178 0 64267 138 42 10 8024 10 9838 103385- 0 85906'

T-test for Independent Samples (Spreadsheets) B Note Vanableswere treated asindependent samples Mean Mean t-value df P Valid N Valid N Std.Oev, Std.Dev.j F-ratlo I p Group 1 vs Group ! Group 1 Group 2 Group 1 Group 2 Group 1 Group 2 i Variances] Variances Van vs. Var2 87 7093i! 86 9468 § 0 54857I 178 0 58398 86 94 8 90851 9 67133.S 117859i] 0 44274

~~Table 18A: Simple regression analysis of the relationship between round goby size~~

~~and predator size from tethering experiments at CORE (A) and DSE (B) range~~

locations in the Trent River in June and July of 2010. A Test of SS Whole Model vs. SS Residual (Spreadsheetl 12) Dependnl Multiple Multiple Adjusted SS df MS SS df MS F P Vanable R R* R» Model Model Model Residual Residual Residual Varl 0 32212 0 10376 0 07660)1 177 919 1 177 919 1536 76' 33 46 5684' 3 82059' 0 05914 B Test of SS Whole Model vs SS Readual (Spreadsheetl 12) Dependnl Multiple Multiple Adjusted SS df MS SS df MS F P Vanable R R» R> Model Model Model Residual Residual Residual Varl 0 18898 0 03571 -0 00811 32 7278 1 32 7278 883 631' 22 40 1650 0 81483 0 37647 Appendix B: Sample sizes Chapter 2: Table 1A - Abundance comparisons over time within range locations May-09 Aug-09 May-10 Aug-10 CORE 21 22 20 21 USE 14 29 45 53 DSE 11 26 40 41

~~Table 2A - Size comparisons May-09 Aug-09 May-10 Aug-10 CORE 20 22 19 21 USE 11 27 45 52 DSE 11 26 40 40~~

~~Table 3A - Overall sex ratio comparison between range locations CORE DSE USE Otonabee Rice 2010 2010 83 84 124 39 49~~

~~Table 4A - Sex ratio comparisons over time within range locations May-09 Aug-09 May-10 Aug-10 CORE 20 22 20 21 DSE 8 24 26 26 USE 9 27 41 47 137~~

~~Table 5A - Site occupancy across five proportions of rock substrate 0-10 20-30 40-50 60-70 80-100 CORE 13 11 19 23 9 DSE 93 46 25 16 17 USE 117 60 28 19 16~~

~~Table 6A - Abundance in substrate types boulder cobble gravel sand mud CORE 6 26 10 21 1 DSE 18 19 5 64 0 USE 8 24 14 79 0~~