INVESTIGATION OF INTER- AND INTRASPECIFIC GENETIC VARIABILITY OF

EUHRYCHIOPSIS LECONTEI, A BIOLOGICAL CONTROL AGENT FOR THE

MANAGEMENT OF EURASIAN WATERMILFOIL

A Dissertation

Presented to

The Graduate Faculty of The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Lara Diane Roketenetz

May, 2015

INVESTIGATION OF INTER- AND INTRASPECIFIC GENETIC VARIABILITY OF

EUHRYCHIOPSIS LECONTEI, A BIOLOGICAL CONTROL AGENT FOR THE

MANAGEMENT OF EURASIAN WATERMILFOIL

Lara Diane Roketenetz

Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Stephen C. Weeks Dr. Monte E. Turner

______Committee Member Dean of the College Dr. Randall J. Mitchell Dr. Chand Midha

______Committee Member Interim Dean of the Graduate School Dr. Shanon Donnelly Dr. Rex D. Ramsier

______Committee Member Date Dr. R. Joel Duff

______Committee Member Dr. Francisco Moore

ii

ABSTRACT

Euhrychiopsis lecontei (milfoil weevils) have been used as a native, augmentative biological control agent in the management of the invasive aquatic weed, Myriophyllum spicatum (Eurasian watermilfoil), since the 1990’s. Although much research has been conducted on E. lecontei’s life history and effectiveness as a biological control agent, detailed genetic characterization of the weevil and its close relatives has been lacking in the literature. The current studies sought to fill this gap in knowledge by collecting molecular genetic data for this important native insect. Characterization of interspecific diversity, via the examination of mitochondrial DNA of E. lecontei and other members of the tribe Phytobiini, suggest that the North American members of this group are sister species (E. lecontei and Parenthis vestitus) as are European members (Eubrychius velutus and Phytobius leucogaster) of the group. This study therefore supports the hypothesis that this group of milfoil specialists shares a recent common ancestor that colonized aquatic habitats and utilized one or more Myriophyllum species as a host. Intraspecific studies utilizing mitochondrial and microsatellite DNA of E. lecontei were also performed. These data indicate that while the species shows some geographic structuring on a continental scale, population-level studies indicate high levels of panmixia with relatively high rates of inbreeding. Since the continued management of Eurasian watermilfoil represents a large economic and ecological burden across the United States

iii and Canada, studies such as these are of particular value in determining if E. lecontei has suitable viability as a native biological control agent for this invasive plant.

iv

DEDICATION

This work is dedicated to my parents, Dan and Diane Roketenetz, without whom this journey would not have been possible.

v

ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Stephen Weeks for the huge amount of support, both academically and personally, that he has given me during my time at the University of Akron. I would not have been able to do this without him and I am grateful for his guidance.

I would also like to thank the current members of my committee: Dr. Randall

Mitchell, Dr. R. Joel Duff, Dr. Shanon Donnelly, and Dr. Paco Moore as well as the previous members of my committee: Dr. Jessica Hopkins and Dr. Kevin Butler. Thanks for the innumerable meetings and discussions that made this research possible.

These projects would not have been achievable without the amazing support of friends and colleagues along the way. Special thanks to my fellow Integrated Bioscience students, particularly, Dr. Hope Ball, Dr. Timothy Astrop, Dr. Heath Garris, Dr. Rafael

Maia, Dr. Alyssa Stark, Alissa Calabrese, Scott Thomas, and Ashley Wain, whose assistance, friendship, humor, and guidance kept me sane over the years. Field and laboratory assistants are too numerous to count, but several people deserve to be mentioned by name including: Robert Canning, Colin Cassin, Annmarie Abeyesekera,

Jeannine Molina Abel, and Camila Faria. I would also like to thank Dr. Eric Sager (Trent

University) and Dr. Ryan Thum (Grand Valley State University) for their advice and support, as well as for the use of their facilities and staff. Dustin Wcisel was a rock star vi in the GVSU/AWRI sequencing lab who not only gave me confidence in my molecular techniques, but a place to stay in Muskegon, Michigan. Dr. Charles Vossbrinck

(Connecticut Agricultural Experiment Station), Dr. Elisabeth Gross (University of

Konstanz), Dr. Peter Sprick (Curculio Institute), and Nathan Harms (United States Army

Corps of Engineers) provided extremely valuable information in either sequence or specimen data that made much of this project possible.

EnviroScience, Inc. had the vision to provide funding for these projects and the adventures I had along the way were the icing on the cake. Special thanks to Marty

Hilovsky, Cortney Marquette, Sarah Lomske Walters, Rebecca McMenamin, Nancy

Cushing and Kyle Borrowman. I am so grateful for your expertise and friendship.

EnviroScience Inc. summer interns, especially, Israel Merkle, Kyle Vogel, Laura

Brutscher, Justin Richardson, Anh Tran, Kim Sage, and Michelle Gorrie made field days fun and living away from home more than bearable.

I want to thank the University of Akron Department of Biology for the valuable teaching experience with special thanks to Dr. Amy Hollingsworth and Ashley Ramer for their pep talks and patience. Additional funding was provided by Choose Ohio First

Tiered Mentoring and Bioinformatics Scholarship programs, University of Akron

Graduate School Government, and North American Lake Management Society.

Finally, to my friends and family outside of academia (and especially to my best friends and partners-in-crime, Josh and Emmy) – thank you for providing me with the balance, support, friendship, love, and encouragement I needed to finish this endeavor.

vii

TABLE OF CONTENTS Page

LIST OF TABLES ...... x

LIST OF FIGURES ...... xii

CHAPTER

I. MOLECULAR CHARACTERIZATION OF INTERSPECIFIC DIFFERENTIATION OF MILFOIL WEEVILS (CURCULIONIDAE, CEUTORHYNCHINAE, PHYTOBII)...... 1

Introduction ...... 1

Methods and Materials ...... 6

DNA Extraction, Amplification, and Sequencing ...... 8

Analyses ...... 11

Results ...... 13

Discussion ...... 17

Implications for Management of Eurasian watermilfoil ...... 20

Conclusions and Future Studies ...... 23

II. PHYLOGEOGRAPHIC CHARACTERIZATION OF GENETIC VARIATION IN EUHRYCHIOPSIS LECONTEI THROUGHOUT NORTH AMERICA ...... 25

Introduction ...... 25

Methods and Materials...... 30

Results ...... 34

viii Discussion ...... 43

Implications for Management of Eurasian watermilfoil ...... 50

Conclusions and Future Studies ...... 51

III. DETERMINATION OF POTENTIAL HOST RANGE SHOFT AND SYMPATRIC SPECIATION OF EUHRYCHIOPSIS LECONTEI FEEDING ON AN ANCESTRAL HOST (MYRIOPHYLLUM SIBIRICUM), VERSUS A NOVEL, INTRODUCED HOST (MYRIOPHYLLUM SPICATUM) ...... 53

Introduction ...... 53

Materials & Methods ...... 57

Results ...... 61

Discussion ...... 66

Implications for Management of Eurasian watermilfoil ...... 74

Conclusions and Future Studies ...... 75

LITERATURE CITED ...... 77

APPENDICES ...... 92

APPENDIX A BAYESIAN ANALYSIS FOR FOUR MEMBERS OF THE TRIBE PHYTOBIINI ...... 93

APPENDIX B 18S ANALYSIS FOR FOUR MEMBERS OF THE TRIBE PHYTOBIINI ...... 97

ix LIST OF TABLES

Table Page

1.1 Morphological Features and Ecological Associations of Milfoil weevils. Information summarized from: Anderson, 2002; Blatchley and Leng, 1916; Buckingham and Bennett, 1981; Buckingham and Passoa, 1984; Colonnelli, 2004; Newman et al., 2006; Solarz and Newman, 2001...... 9

1.2 Collection locations and sample sizes for each milfoil weevil species ...... 10

1.3 Outgroups obtained from GenBank for phylogenetic analysis. All accession numbers correspond to COI sequences unless otherwise noted ...... 13

1.4 Estimates of average evolutionary divergence over sequence pairs within and between groups. The number of base differences per sequence from averaging over all sequence pairs within (on diagonal) and between (below diagonal) each group are shown. Standard error estimate(s) for the average number of base pair differences are shown in parentheses and were obtained by a bootstrap procedure (50 replicates). The analysis involved 21 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. All ambiguous positions were removed for each sequence pair. There were a total of 991 positions in the final dataset. Evolutionary analyses were conducted in MEGA5. Estimates of divergence (Ma = Million years ago) are also reported (above diagonal) using the average number of base pair differences and the standard insect molecular clock of 2.3% change per 1 million years. EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus...... 16

2.1 Euhrychiopsis lecontei collection locality data and sample size ...... 32

2.2 Polymorphism summary statistics for total data set based on 984 bp of Euhrychiopsis lecontei COI mtDNA ...... 34

x

2.3 List of waterbodies (location), associated haplotypes, sample size of each Haplotype/waterbody (N) and clade assignation within each population ...... 38

2.4 SAMOVA results for K=2 groups ...... 41

2.5 Polymorphism summary statistics for Clades A and B based on 984 bp of Euhrychiopsis lecontei COI mtDNA ...... 42

3.1 Location and sample sizes of waterbodies. The centroid of each waterbody is reported for location ...... 58

3.2 Distance (km) between sites ...... 59

3.3 Microsatellite markers used in this study. Locus name, GenBank accession numbers, and primer sequences obtained from Vossbrinck et al. (2010) ...... 60

3.4 Hierarchical Analysis from six microsatellite markers for six populations within the Kawartha Lakes region ...... 62

3.5 Pairwise Fst (below the diagonal) and migration rates per generation (Nm; above the diagonal) for the six populations within the Kawartha Lakes. ∞ = infinite migrants per generation between sites ...... 62

3.6 Summary of results from three separate Θ (Fst) analyses for Stony Lake. EWM = Eurasian watermilfoil; NWM = Northern watermilfoil...... 63

3.7 Summary of results from two separate Fst analyses for Big Bald Lake Site 1 and Stony Lake Site 4. EWM = Eurasian watermilfoil; NWM = Northern watermilfoil ...... 65 LIST OF FIGURES

Figure Page

1.1 Molecular phylogenetic analysis by Maximum Likelihood method. The evolutionary history was inferred by using the Maximum Likelihood method based on the General Time Reversible model. The tree with the highest log likelihood (-4799.3426) is shown. The percentage of trees (>50%) in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter = 1.0153)]. The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 48.0902% of sites). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 30 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 991 positions in the final dataset. Evolutionary analyses were conducted in MEGA5. EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus ...... 15

2.1 Molecular phylogenetic analysis by Maximum Likelihood method. The evolutionary history of a portion of COI mtDNA from 67 individuals of Euhrychiopsis lecontei was inferred using the Maximum Likelihood method based on the Tamura 3- parameter model (Tamura, 1992). The tree with the highest log likelihood (-3338.7135) is shown. The percentage of trees (> 50%) in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among xii

sites [5 categories (+G, parameter = 0.1597)]. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 43 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 986 positions in the final dataset. Evolutionary analyses were conducted in MEGA5. Designation of Clade A is marked with a green box and designation of Clade B is marked with a red box. Yellow stars indicate >99.0% similarity to known specimens of E. lecontei in the BOLD data set ...... 36

2.2 Haplotype Network of 39 haplotypes inferred from 984 base pairs of mtDNA COI gene for 67 individuals of Euhrychiopsis lecontei. Sample sizes ranged from one to four individuals per population. Black circles indicate missing intermediate haplotypes. Size of the yellow circles correlates to the number of individuals that share that haplotype. Red numbers correlate to the polymorphic positions in the data set. See Tables 2.1 and 2.2 for exact locations of the haplotypes. Clade A is designated by a green box and Clade B is designated by a red box...... 40

2.3 Barrier to gene flow predicted by raw genetic data and geographic locations in Alleles in Space (Miller, 2005) is indicated by the purple line. Clades designations are represented by green (Clade A) and red (Clade B) circles ...... 43

2.4 Proposed glacial refugia for Clades A (green) and B (red). Colored arrows depict the possible colonization routes of Euhrychiopsis lecontei into the distinct geographic areas delineated by the genetic barrier ...... 49

3.1 Locations of the seven sampling sites within the Kawartha Lakes, Ontario, Canada ...... 59

A1.1 Bayesian phylogenetic relationship of the four milfoil weevil species. EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus...... 96

B1.1 Molecular phylogenetic analysis of the partial sequence of the 18S gene for the four target species and outgroups by Maximum Likelihood method. The evolutionary history was inferred by using the Maximum Likelihood method based on the Kimura 2-parameter model. The tree with the highest log likelihood (-3238.8113) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the

xiii branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter = 0.0500)]. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 25 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 1,749 positions in the final dataset. Evolutionary analyses were conducted in MEGA5. EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus ...... 99

xiv CHAPTER I

MOLECULAR CHARACTERIZATION OF INTERSPECIFIC DIFFERENTIATION

OF MILFOIL WEEVILS (CURCULIONIDAE, CEUTORHYNCHINAE, PHYTOBIINI)

Introduction

Invasive species are non-indigenous plants or animals that become established in a natural area and pose risks to the existing community (Czarapata, 2005). The intentional or accidental introduction and establishment of invasive species into various habitats can cause dramatic loss of local biodiversity (Madsen et al., 1991), change in overall species composition (Creed, 2000), and loss of recreational value (Smith and

Barko, 1990). The control of such species represents a substantial economic burden for governments and management agencies (Knight, 2001; Pimentel et al., 2005). According to Pimentel et al. (2005) economic losses (i.e., induced damages, loss of recreational uses and control measures) due to invasive aquatic weeds is approximately $110 million per year, although the ultimate costs associated with invasive species can be quite difficult to determine based on the complexity of compiling and analyzing such data (Lovell et al.,

2006).

1 In North America, Eurasian watermilfoil (Myriophyllum spicatum L.) is an invasive, submersed aquatic plant found throughout most of the United States and portions of Canada (Creed, 1998). Its native range is restricted to Europe, Asia, and

Northern Africa (Czarapata, 2005). Reed (1977) reports that it likely appeared in the

Chesapeake Bay area as early as the 1880’s; however, the first well-documented introduction occurred in the 1940’s in a fish pond outside of Washington, D.C. (Couch and Nelson, 1985; Sheldon and Creed, 1995). Couch and Nelson (1985) hypothesize that this introduction likely came at the hands of government officials. Since then, it has become one of the most noxious aquatic weeds in North America (Smith and Barko,

1990) and has now spread into four Canadian provinces (Quebec, New Brunswick,

British Columbia and Ontario; USDA Plants Database) and all but one state (Wyoming) in the contiguous US (Creed, 2000; Jacobs and Mangold, 2009; Jacono and Richerson

2010; Madsen, 2005; USDA Plants Database).

Eurasian watermilfoil is a strong competitor due to its rapid growth rate, its ability to reproduce vegetatively, and its ability to tolerate a wide variety of environmental conditions (Sheldon and Creed, 1995). Eurasian watermilfoil is able to begin photosynthesizing at low water temperatures (~15°C; Smith and Barko, 1990) and therefore can grow rapidly in the spring, forming a dense canopy that excludes many native macrophytes that begin to grow later in the season (Madsen et al., 1991). Upon reaching the surface, the plant produces inflorescent flower spikes which rise above the water (Smith and Barko, 1990). Eurasian watermilfoil’s dense floating mats can reduce diversity and abundance of native macrophytes (Madsen et. al., 1991), as well as

2 associated invertebrate and fish populations (Keast, 1984). The growth and senescence of thick mats of Eurasian watermilfoil degrades water quality by increasing water temperature (Unmuth et al., 2000) and depleting dissolved oxygen (Honnell et al., 1992).

Historic management techniques of this invasive plant have included chemical, mechanical, physical and biological methods (Richardson, 2008; Sheldon, 1997; Sheldon and O’Bryan, 1996). Phytophagous insects are often utilized as classical biocontrol agents. Spencer and Lekić (1974) reported 25 different insect species that fed on

Eurasian watermilfoil from a portion of the plant’s native range. Since the early 1990’s, researchers have been investigating the ability of the milfoil weevil (Euhrychiopsis lecontei Dietz, 1896) to control populations of Eurasian watermilfoil (Creed and Sheldon,

1995; Sheldon and Creed, 1995; Sheldon and Creed, 2003; Tamayo and Grue, 2004).

Euhrychiopsis lecontei are distributed throughout the northern continental United States and portions of Canada (Creed, 1998). It is believed that northern watermilfoil

(Myriophyllum sibiricum) is this weevil’s historic host (Creed and Sheldon, 1994). All life stages (egg, larva, pupa, and adult) of E. lecontei are dependent upon watermilfoil species, indicating that this weevil is an obligate aquatic species (Solarz and Newman,

2001).

Euhrychiopsis lecontei have been used commercially for Eurasian watermilfoil control in lake management projects (EnviroScience Inc., Stow, OH), through augmentation programs, where these weevils are reared in mass numbers and purposefully introduced into Eurasian watermilfoil infested waterbodies. Several aspects of the milfoil weevil’s ecology and life history make it an ideal candidate for

3 consideration as a biocontrol agent for Eurasian watermilfoil: 1) it is native to North

America; 2) it exhibits host specificity for milfoil species (Solarz and Newman, 1996); 3) it successfully feeds and oviposits on the invasive Eurasian watermilfoil (Solarz and

Newman, 1996); 4) it consumes milfoil during all stages of development (Solarz and

Newman, 2001); and 5) it affects physiological functioning of Eurasian watermilfoil through primary and secondary pathways (Creed and Sheldon, 1995). By preventing M. spicatum from monopolizing resources, E. lecontei could play an important role in maintaining community structure and species richness (Creed, 2000).

Most research conducted to date on this species has focused on behavior, ecology, host-specificity, and life-cycle development of Euhrychiopsis lecontei (Newman, 2004).

Additional studies have been conducted on abundance, distribution and effectiveness of

E. lecontei in a specific lake or lakes (Alwin et al., 2010; Jester et al., 2000; Newman and

Biesboer, 2000; Newman and Inglis, 2009; Parsons et al., 2011). It has been found that the effectiveness of natural populations of E. lecontei in controlling Eurasian watermilfoil has varied across lakes (Creed and Sheldon, 1995; Lillie, 2000; Newman and Biesboer,

2000), and the cause of such variation has not been well established. Natural E. lecontei populations in the wild range from about 0.5 – 1.5 weevils per stem in most lakes

(Newman et al., 2001), and it is unlikely that these densities are high enough to control a widespread infestation of Eurasian watermilfoil. As previously stated, in order to successfully augment E. lecontei populations to the density required to control infestations of milfoil, these weevils must be artificially reared in a laboratory and released into affected waterbodies.

4

There are a few other species of weevil that utilize Myriophyllum in at least one portion of their life cycle or as a food source (Colonnelli, 2004). Therefore, sequence data from milfoil specialists were analyzed in order to propose a phylogenetic relationship between them, because these weevils could represent other important potential biocontrol agents. The target species of this investigation are Eubrychius velutus Beck 1817,

Euhrychiopsis lecontei, Parenthis vestitus Dietz 1896, and Phytobius leucogaster

Marsham 1802. Their hypothesized ancestral hosts include several different species of

Myriophyllum; specific host relationships are discussed below (Table 1.1).

The fact that very few groups of weevils are known to utilize aquatic habitats

(Korotyaev, 2008) raises some interesting evolutionary questions. There are a number of questions about the proper taxonomic condition of the genera mentioned above in addition to their relationships with each other. Do these weevil species share a recent common ancestor that also invaded an aquatic ecosystem and utilized a Myriophyllum species as a host, or conversely, has Myriophyllum been colonized independently by multiple unrelated weevils on separate occasions? The former scenario hypothesizes that host expansions/shifts (and eventually possible speciation) are more common when the shift is to a plant closely related to the ancestral host, when the novel host is one that is utilized by a close relative, or the novel host is one that was used by an ancestor of the insect (Janz and Nylin, 2007). The latter scenario, however, hypothesizes that, although uncommon, weevils may be able to colonize aquatic habitats more readily than previously suggested (Korotyaev, 2008).

5 To date, there has been a limited amount of genetic research conducted on these insects (Solarz, 1998). To test these hypotheses, molecular data from combined partial sequences of Cytochrome Oxidase I (COI) and Cytochrome b (CtyB) mitochondrial

DNA were obtained and analyzed. The current dataset represents some of the first gene- based molecular sequences collected for these important weevil taxa. The use of these genes as standards in insect molecular systematics is advocated in Caterino et al. (2005) because they are widely used and can therefore be utilized in broader insect phylogenetic studies.

Methods and Materials

The tribe Phytobiini (Subfamily Ceutorhynchinae) comprises nine genera

(Colonnelli, 2004) and includes all of the species of interest for this study. Eubrychius,

Euhrychiopsis and Parenthis are monotypic genera. Phytobius has four described species

(Colonnelli, 2004), however it is also reported as a monotypic genus within North

America (Anderson, 2002).

Eubrychius velutus are distributed throughout Europe and Asia (Colonelli, 2004;

Newman et al. 2006) and do not occur in North America (Anderson, 2002). All life stages are found on submersed portions of the plant, typically within the upper 5 cm of the plant (Newman et al., 2006). Adults overwinter in a near-shore terrestrial habitat

(Blatchley and Leng, 1916; Newman et al., 2006).

Euhrychiopsis lecontei are distributed throughout the northern continental United

States and portions of Canada (Creed, 1998). Their ancestral host is Myriophyllum

6 sibiricum (Northern watermilfoil), but they are commonly found in association with M. spicatum (Creed and Sheldon, 1994). Similar to Eubrychius, this weevil is an obligate aquatic species (except for its overwintering stage) and is found on submersed portions of the plant (Solarz and Newman, 2001). However unlike Eubrychius, they are typically found within the top 20 cm of the plant (Newman et al., 2006). Euhrychiopsis lecontei has been characterized as being very closely related, although distinct, from E. velutus

(Creed and Sheldon, 1994; Tamayo et al., 1999; Newman et al., 2006).

The distribution of Phytobius leucogaster is holarctic (Buckingham and Bennett,

1981; Colonnelli, 2004). Host plants include Myriophyllum sibiricum, M. spicatum and

M. verticillatum. Adults are found on the flowers and fly freely between flowers. Adults can survive forced submersion for 8 – 24 hrs (Buckingham and Bennett, 1981) indicating that they are semi-aquatic in nature.

Parenthis vestitus are distributed in North America (Colonnelli, 2004). It has been associated with Myriophyllum heterophyllum (Colonnelli, 2004), a native milfoil species that is becoming a noxious weed in some areas of North America (USDA Plants

Database) and M. brasiliense (Colonnelli, 2004), a non-native milfoil that can also become invasive (USDA Plants Database). Adults are found on emergent flowers and fly freely (Buckingham and Passoa, 1984), similar to Phytobius leucogaster. Brief periods

(~5 min.) of submersion have been observed in the lab (N. Harms, pers. comm.).

Although little is reported on this weevil in the primary literature, Blachley and Leng

(1916) hypothesized that Parenthis was semi-aquatic, due to its morphological similarities to Phytobius spp. Buckingham and Passoa (1984) also refer to them as semi-

7 aquatic. Basic morphological characteristics, range, host association and ecotype for each species of interest are described in Table 1.1.

DNA Extraction, Amplification, and Sequencing

Euhrychiopsis lecontei specimens were collected from various locations in the

United States and Canada. Parenthis vestitus specimens were collected from one locality in the United States and were confirmed by Dr. Robert Anderson of the Canadian

Museum of Nature. Phytobius leucogaster specimens were collected from Ontario,

Canada. Eubrychius velutus specimens were collected in Germany. Collection localities and sample sizes are summarized in Table 1.2.

8

Table 1.1. Morphological Features and Ecological Associations of Milfoil weevils. Information summarized from: Anderson, 2002; Blatchley and Leng, 1916; Buckingham and Bennett, 1981; Buckingham and Passoa, 1984; Colonnelli, 2004; Newman et al., 2006; Solarz and Newman, 2001.

Species Eubrychius Euhrychiopsis Parenthis Phytobius velutus lecontei vestitus leucogaster

Size (mm) 2.2-2.5 2 – 3 2.5 2.5 – 3 Color Black with Brownish black Black with Dark gray with greenish- with yellowish intermixed white spot on yellow stripes stripes on elytra scales of white elytra and with a grayish and whitish and gray. white/yellow yellow underbelly White spot on underbelly underside elytra and grayish white underbelly Distribution Europe and Northern US and United States Holarctic Northern Asia Canada and Canada Myriophyllum M. M. sibiricum; M. M. M. spicatum; M. Host heterophyllum; spicatum heterophyllum sibiricum; M. association(s) M. spicatum; verticillatum M. verticillatum; M. elatinoides Portion of Submersed Submersed Emergent Emergent flowers plant where meristems meristems flowers typically (within top (within top 20 found 5cm of plant) cm of plant) Larval Meristem; Meristem; mine Unknown Within ovaries Feeding externally on through lower and buds; leaves and parts of stem externally on upper stems flowers and stems Pupal Externally on Within stem, 20 Unknown In excavation of Chamber side of stem cm or more submersed stem below top of plant Aquatic status Aquatic Aquatic Semi-aquatic Semi-aquatic

9

Table 1.2. Collection locations and sample sizes for each milfoil weevil species. Sample Species Size Collection Locations Latitude Longitude University of Konstanz Eubrychius Limnological Institute Ponds, velutus 4 Konstanz, Germany 47.694902 9.193276 Hannoversche Treue, 2 Salzgittter-Engerode, Germany 52.082302 10.375972 Euhrychiopsis Auburn Lake, Minnesota , lecontei 1 USA 44.864751 -93.680275 1 Burke Lake, Washington, USA 47.133589 -119.91836 Christmas Lake, Minnesota, 1 USA 44.898056 -93.541389 1 Lake Eligo, Vermont, USA 44.599586 -72.356315 1 Lake Ovid, Michigan, USA 42.93841 -84.410806 Parenthis vestitus 4 Honey Creek, Oklahoma, USA 34.423433 -97.153524 Phytobius leucogaster 3 Long Lake, Ontario, Canada 46.407637 -81.000187 3 Bethel Lake, Ontario, Canada 46.471812 -80.95856

Weevils are small and highly sclerotized. To facilitate DNA extraction, the specimens were placed live into ATL lysis buffer upon collection and were ground with a melted pipette tip directly in a 1.5 mL centrifuge tube and stored at room temperature until DNA extraction. Some weevil samples collected remotely were preserved in water at -20°C. DNA was extracted with the DNeasy® Blood and Tissue kit (Qiagen®) following the protocol for purification of total DNA from insects (DNeasy® Blood and

Tissue Handbook, 2006). Amplification of the partial mtDNA Cytochrome Oxidase COI

(LCO/C1 primer pair) sequence was conducted using the standard primers LCO 5’ –

GGTCAACAAATCATAAAGATATTGG –3’ and C1N2776 5’ –

GGATAATCAGAATATCGTCGAGG – 3’ or primer pairs C1-J-2183 Jerry 5’ –

CAACATTTATTTTGATTTTTTGG –3’ and L2-N-3014 Pat 5’ –

10

TCCAATGCACTAATCTGCCATATTA – 3’. Amplification of the partial mtDNA

Cytochrome B (CytB) sequence was conducted using the primers CB-J-10933 CB1 5’ –

TATGTACTACCATGAGGACAAATATC – 3’ and CB-N-11367 CB2 5’ –

ATTACACCTCCTAATTTATTAGGAAT – 3’.

PCR amplification was conducted with 1µL DNA extract in a 25 µL reaction mixture using Promega HotStart Taq. Amplification conditions for COI (LCO/C1 primer pair) included an initial denaturation period of 2 min at 94°C, followed by 30 cycles of

94°C denaturation (30s) 45°C annealing (30s) and 72°C extension (60s) with a final extension of 10 minutes. Amplification conditions for COI (Jerry/Pat primer pair) included an initial denaturation period of 3 min at 94°C, followed by 30 cycles of 94°C denaturation (30s) 50°C annealing (30s) and 72°C extension (60s) with a final extension of 5 minutes. Amplification conditions for CytB (CB1/CB2 primer pair) included an initial denaturation period of 3 min at 94°C, followed by 30 cycles of 94°C denaturation

(30s) 46°C annealing (30s) and 72°C extension (60s) with a final extension of 5 minutes.

Purification and sequencing of the samples was completed by Annis Water Resources

Institute (Muskegon, Michigan) using the same primers used in PCR amplification.

Analyses

Forward and reverse sequences for each gene were assembled, manually reviewed and edited with the software package Sequencher v.4.6®. Sequences were imported into

MEGA 5.2 (Tamura et al., 2011) and aligned with the built-in automated aligner

CLUSTALW. The COI partial sequences obtained from the C1/ LCO primer pair and the

11

Jerry/Pat primer pair were trimmed to an overlap region consisting of 526 base pairs. The

CytB partial sequences were trimmed to an overlap region consisting of 465 base pairs. A concatenated data set was assembled consisting of individuals that were sequenced for both mtDNA genes. The concatenated data set consisted of 991 base pairs. MEGA 5.2 assigned the General Time Reversible with gamma distribution and proportion of invariant sites (GTR+G+I) as the appropriate evolutionary rate model for the data under both Bayesian Information Criterion (BIC) and Akaike Information Criterion, corrected

(AICc). A Maximum Likelihood Tree with 500 bootstraps was constructed in MEGA

5.2. Estimates of average evolutionary divergence over sequence pairs within and between groups were also conducted in MEGA 5.2.

To identify outgroups for phylogeny reconstruction, sequences of closely related weevil species were initially identified by an NCBI BLAST search in GenBank with one of the newly derived COI sequences of Euhrychiopsis lecontei. Several outgroups from

Curculionidae and in the study organisms’ same subfamily (Ceutorhynchinae) or from closely related subfamilies (based on phylogeny in McKenna et al., 2009) were included in the data set. A presumably evolutionarily distant outgroup was chosen from family

Chrysomelidae. The subfamilies (or family for the distant outgroup) and GenBank ascension numbers for the outgroups are listed in Table 1.3. Only COI data were used for the outgroups, unless otherwise noted in Table 1.3 because CytB sequence data was only available for Curculio glandium. Outgroups were aligned with the ingroup data set in

MEGA 5.2 with CLUSTALW and trimmed to the appropriate region. Missing data for

12 the outgroups was indicated in the alignment as unknowns and treated as phylogenetically unimportant.

Additional analyses are included in Appendix A and Appendix B. Appendix A includes a Bayesian Metropolis–coupled Markov Chain Monte Carlo analysis

(10,000,000 generations) of a dataset consisting of 4,164 base pairs (including missing data) for four genes (COI, CytB, Elongation factor-1-alpha, and 18S). Appendix B includes a Maximum Likelihood analysis for 18S and incorporates individuals from the subfamily Bagoninae (genus Bagous), one of the only other weevil groups in

Curculionidae known to have colonized aquatic habitats.

Table 1.3. Outgroups obtained from GenBank for phylogenetic analysis. All accession numbers correspond to COI sequences unless otherwise noted. Species Subfamily (or Family*) GenBank Ascension

Number

Ceutorhynchus erysimi Ceutorhynchinae DQ155881.1 Ceutorhyncus neglectus Ceutorhynchinae DQ058697.1 Ceutorhynchus rapae Ceutorhynchinae DQ058702.1 Ceutorhynchus scrobicollis Ceutorhynchinae EU551172.1 Ceutorhynchus Ceutorhynchinae DQ058696.1 subpubescens Curculio dentipes Curculionidae AB367610.1 Curculio glandium Curculionidae AY327711.1 (COI); AY327748.1(CytB) Sitona discoideus Entiminae EF118298.1 Cassida rubiginosa Chrysomelidae* AM283218.1

Results

Based on analysis of the concatenated data set of mtDNA genes, COI and CytB, consisting of 991 base pairs, all individuals were appropriately grouped with other individuals of their genus with 100% bootstrap support for each species (Fig. 1.1). The 13

European/Holarctic species (Eubrychius velutus, EV, and Phytobius leucogaster, PL) are sister species with 97% bootstrap support and the North American species (Euhrychiopsis lecontei, EL, and Parenthis vesititus, PL) are sister species with 100% bootstrap support.

Based on the branch lengths, E. velutus is more distantly related to P. leucogaster than E. lecontei is to P. vesititus. The four species of interest from tribe Phytobiini form a monophyletic group (99% bootstrap support), and are grouped with 59% bootstrap support with outgroups that represent other members of the subfamily Ceutorhynchinae

(Fig. 1.1). The average number of pairwise differences within and between the study organism groups is reported in Table 1.4. Based on these average differences and using a standard insect molecular clock for mtDNA (Brower, 1994), an estimate of divergence time is also included in Table 1.4. Estimated divergence time between the North

American species is 3.11 Ma. Estimated divergence time between the European/Holarctic species is 4.68 Ma. Eubrychius velutus appears to have split from the North American species between 6.22 – 5.95 Ma. Phytobius leucogaster appears to have split from the

North American species between 5.26 – 4.93 Ma.

14

15

Figure 1.1. Molecular phylogenetic analysis by Maximum Likelihood method. The evolutionary history was inferred by using the Maximum Likelihood method based on the General Time Reversible model. The tree with the highest log likelihood (-4799.3426) is shown. The percentage of trees (>50%) in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter = 1.0153)]. The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 48.0902% of sites). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 30 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 991 positions in the final dataset. Evolutionary analyses were conducted in MEGA5 (Tamura et al., 2011). EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus.

Table 1.4. Estimates of average evolutionary divergence over sequence pairs within and between groups. The number of base differences per sequence from averaging over all sequence pairs within (on diagonal) and between (below diagonal) each group are shown. Standard error estimate(s) for the average number of base pair differences are shown in parentheses and were obtained by a bootstrap procedure (50 replicates). The analysis involved 21 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. All ambiguous positions were removed for each sequence pair. There were a total of 991 positions in the final dataset. Evolutionary analyses were conducted in MEGA5 (Tamura et al., 2011). Estimates of divergence (Ma = Million years ago) are also reported (above diagonal) using the average number of base pair differences and the standard insect molecular clock of 2.3% change per 1 million years (Brower, 1994). EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus.

EV EL PV PL EV 2.33 (0.74) 5.95 Ma 6.22 Ma 4.68 Ma EL 135.63 (10.22) 10.60 (2.04) 3.11 Ma 4.93 Ma PV 141.92 (10.08) 70.80 (8.50) 7.50 (1.80) 5.26 Ma PL 106.78 (9.23) 112.40 (8.85) 120.33 (10.23) 10.27 (2.11)

16

Discussion

The family Curculionidae (“true weevils”) is the most speciose of any weevil family, with approximately 51,000 species described, representing over 80% of all weevil species (Oberprieler et al., 2007). It is therefore no surprise that Curculionidae systematics have long been considered messy, complex and frustrating (Colonnelli, 2004;

Hundsdoerfer et al., 2009; Marvaldi et al., 2002; McKenna et al., 2009; Oberprieler et al., 2007). The emergence of molecular techniques over the last few decades has led to newly hypothesized phylogenetic relationships among the subfamilies (Hundsdoerfer et al., 2009; Marvaldi et al., 2002; McKenna et al., 2009). These new phylogenies do not agree in all details, nor do they even agree regarding the number of recognized subfamilies noted earlier within Curculionidae.

Despite phylogenetic uncertainty in the broader family, the subfamily

Ceutorhynchinae has been one of the best taxonomically defined groups of weevils

(Korotyaev, 2008) and several keys utilizing their morphological characters have been published (e.g., Colonnelli, 2004; Thompson, 1992). A characteristic of many species within the Ceutorhynchinae is the adaptation to aquatic habitats, a condition found in only a few other subfamilies within the Curculionidae, most notably in Bagoinae

(Korotyaev, 2008). Korotyaev (2008) hypothesized that key morphological adaptations for swimming and the ability to live in aquatic environments may have arisen independently in several tribes of Ceutorhynchinae. Interestingly, the divergence times within the Phytobiini weevils (3 - 6 Ma) correspond to the estimated diversification of

17

Bagoninae (~5 - 10 Ma) reported in McKenna et al. (2009), suggesting that the adaptation to water may have occurred at roughly the same time.

The current study sheds important light on the relationships of four members of the tribe Phytobiini. The phylogenetic relationship of the four target species (Eubrychius velutus, Euhrychiopsis lecontei, Parenthis vestitus and Phytobius leucogaster) in

Ceutorhynchinae supports the hypothesis that this group of milfoil specialists shares a recent common ancestor that originally colonized aquatic habitats and utilized one or more Myriophyllum species as a host. This conclusion was not unexpected as phytophagous insects are thought to shift or expand their host range when a new potential host is closely related to an ancestral host (Janz and Nylin, 2007). Subsequent speciation of these weevils could have therefore been facilitated by host shifts because utilization of different host plants can rapidly lead to host race formation within a population (Drès and

Mallet, 2002).

The occurrence of host-adapted sympatric host races among other insects (e.g., the apple maggot fly) provides support to the idea that full speciation can occur through host shifts (Bush, 1969). Drès and Mallet (2002) hypothesize that this occurs in a step- wise fashion. First, spatially or temporally mediated assortative mating occurs because hosts are grouped in space and time. Therefore, individuals of a population that are attracted to a novel host would be more likely to mate with each other than with those that remain associated with the historical host. Second, once genetic separation between the host races occurs, hybrid disadvantage can reinforce the separation of the host races and rapidly lead to sympatric speciation through reproductive isolation (Wright, 1982).

18

Alternatively or in conjunction with host shift speciation, geographic isolation of maternal lineages of these species appears to have occurred because the two sets of sister species share overlapping ranges (Table 1.1). These species separate into two distinct clades (Fig. 1.1) that are mostly geographically disjunct (although Phytobius leucogaster does occur in North America), but display similarities in the partitioning of the resource

(Myriophyllum spp.) within each clade. For example, each sister pair has one species that utilizes the submersed portion of the plant (i.e., Euhrychiopsis and Eubrychius) as well as one species that utilizes the emersed flower of the plant (i.e., Parenthis and Phytobius).

In utilizing these different portions of the plant, the weevils are not in direct competition for food and oviposition sites with its sister species.

There appear to be other examples of convergent evolution of traits between

Eubrychius and Euhrychiopsis, as well as between Phytobius and Parenthis. For example, both Eubrychius velutus and Euhrychiopsis lecontei are brownish in color with yellowish/greenish stripes and utilize the submersed portion of the plant. This patternation appears to afford these two species some level of protection through camouflage. The varying light and dark stripes on the elytra mimic the leaflet pattern of

Myriophyllum, and therefore may make them harder to distinguish from the plant by their predators. Parenthis vestitus and Phytobius leucogaster are both active flyers and utilize the emersed flower portions of Myriophyllum. These two species also have similar patternation to one another, with a white spot on their gray elytra. Although not as clear as the camouflage benefit in Eubrychius and Euhrychiopsis, this coloration may function similarly in protecting against predators while the weevil is on the flower of the plant. In

19 essence, although Eubrychius velutus and Phytobius leucogaster are sister species, they exhibit differences in morphology, host use and ecological requirements that are mirrored in the sister pair Euhrychiopsis lecontei and Parenthis vestitus (Table 1.1).

Implications for Management of Eurasian watermilfoil

Many species of Myriophyllum are now considered invasive outside of their native ranges (e.g., Myriophyllum spicatum in North America; M. heterophyllum in

Eurasia), and the potential of these milfoil weevil specialists to control invasive

Myriophyllum populations is of great importance in determining appropriate management strategies. Three of the four weevil species are documented to utilize more than one species of Myriophyllum as host plants (Table 1.1) and some of these host associations

(e.g., Euhrychiopsis lecontei on M. spicatum) are relatively recent (Colonnelli, 2004;

Creed and Sheldon, 1995; Sheldon and Creed, 1995; Sheldon and Creed, 2003; Tamayo and Grue, 2004). Moreover, there also appears to be overlap in the host associations of these milfoil weevils (Table 1.1; i.e., three species utilize M. spicatum and M. sibiricum).

Futuyma (2000) hypothesized that closely-related plants are more often likely to cause adaptation than more distantly-related plants because they are more likely to display underlying genetic variation in feeding response. The author further states that when a population encounters a new food source, “rapid, often substantial genetic responses” are likely to occur (Futuyma, 2000). This would suggest that these four species of weevil may, in fact, continue to diverge, if assortative mating and pre-mating isolation

20 mechanisms are reinforced through genetic adaptations to the varying food sources.

Eventually, this could lead to more effective biological control agents for invasive species of Myriophyllum.

Currently, of the four species of interest, only Euhrychiopsis lecontei is being actively used as an augmentative biological control agent for Myriophyllum spicatum.

The success of biological control programs for invasive plants can increase where more than one agent is released (Denoth et al., 2002). The close phylogenetic relationship of the four weevil species and their varying impacts on Myriophyllum spp. suggest that if used together as a suite of biological control agents, there would be greater potential for effective control of M. spicatum or other invasive species of Myriophyllum.

Newman et al., 2006 suggested that Eubrychius velutus would be a less effective control agent on invasive populations of M. spicatum because the larval damage from E. velutus is limited to the upper 5 cm of the plant and their larvae do not hollow out the stem for feeding or pupation. Conversely, E. lecontei larvae mine through a large portion of the upper stem of M. spicatum which destroys nearly all of the lacunae which act as reservoirs for carbon dioxide (Sheldon and Creed, 1995). The destruction of the lacunae, combined with the loss of carbon dioxide, has a dramatic effect on plant buoyancy (Creed et al., 1992; Creed and Sheldon, 1995). Myriophyllum plants infested with weevil larvae also maintain open wounds and are more prone to pathogenic attack by bacteria and fungi

(Creed, 2000). Both E. velutus and E. lecontei eggs are laid on the apical portion of the plant stem and the larvae’s primary food source is the meristematic tissue of the watermilfoil (Sheldon and O’Bryan, 1996 Newman et al., 2006). Therefore, herbivory by

21

E. velutus and E. lecontei larvae destroy meristematic tissue, suppressing the production of root biomass and inhibiting the translocation of sugars to plant roots (Creed and

Sheldon, 1995). Thus, the suppression of new growth may be more detrimental to the milfoil than the actual loss of plant tissue through herbivory (Sheldon and Creed, 1995).

The destruction of meristematic tissue (i.e., the growing tip) may also reduce the viability of asexual fragments of M. spicatum as this plant can expand rapidly within a waterbody through vegetative reproduction via stem fragmentation (Madsen et al., 1988).

Myriophyllum spicatum also has the ability to reproduce sexually (Coble and

Vance, 1987; Martin and Valentine, 2014). An early study by Coble and Vance (1987) concluded that Myriophyllum spicatum had a high potential for seed production but a low

(>50%) rate of germination and high mortality of seedlings under typical environmental conditions. However, hybridization between M. spicatum and M. sibiricum has been documented (Moody and Les, 2002; 2007), so ample evidence exists that sexual reproduction in these plants is not inconsequential. Myriophyllum spp. are reported to be more invasive in introduced ranges as a direct result of hybridization and hybrid vigor

(Moody and Les, 2002; LaRue et al., 2012).

Phytobius leucogaster has been reported as directly consuming ovaries and flowers of Myriophyllum spicatum (Buckingham and Bennett, 1981), effectively physically reducing the ability of the plant to reproduce sexually and set seed. Although there is no specific data reported for Parenthis vestitus in the primary literature, the convergence of certain features between P. leucogaster and P. vestitus (e.g., actively flying, semi-aquatic nature) could suggest that further investigation would reveal that P.

22 vestitus also utilizes Myriophyllum flowers for oviposition and larval development and would exert similar constraints on the seed set of the plant. This could be an important insight, as Swope and Parker (2012) determined that bud herbivory rather than direct seed predation elicited the most effective control of a terrestrial invasive plant.

Biological control of sexual reproduction in Myriophyllum populations could also reduce the occurrence of hybridization events. This may be especially relevant from a management perspective because hybrid populations of M. spicatum x M. sibiricum have been shown to be more resistant to the effects of a commonly used herbicide in traditional management techniques than populations of M. spicatum (LaRue et al., 2012).

Using more than one of these weevils could lead to more effective augmentative biocontrol programs because several different aspects of the plant’s ecology would be targeted by these milfoil specialists. For example, Euhrychiopsis targets the actively growing meristems as well as mines the lower portion of the stems, disrupting carbohydrate storage to the roots and opening wounds in the stem that can allow pathogens to attack the plant. If a secondary biocontrol agent, such as Phytobius, was also introduced in high numbers through augmentative means, emersed flowers of

Myriophyllum would be attacked later in the growing season, thereby limiting the potential of the plant to set seed or to form hybrids through sexual reproduction.

Conclusions and Future Studies

In conclusion, phylogenetic and evolutionary relationships between milfoil weevil specialists have been identified. These have clearly supported the single origin

23 hypothesis of aquatic adaptation in this group. The phylogenetic analysis could be expanded by sequencing more genes for more individuals of the tribe Phytobiini. This would allow for integration of Phytobiini into larger weevil phylogenetic studies. For example, a recent study was able to utilize 15 – 20 kilobases of mitochondrial genome for a large number of weevil species via high through-put sequencing and produced a robust phylogenetic analysis (Gillett et al., 2014).

Further investigation into the plausibility and suitability of augmentative biocontrol programs which utilize more than one species of these weevils is also warranted. Only Euhrychiopsis lecontei is currently being used in augmentative rearing facilities. Feasibility studies on rearing of the other three weevil species that are milfoil specialists could be conducted to determine whether or not the addition of another biocontrol agent is economically and biologically feasible. Since the continued management of Eurasian watermilfoil represents a large economic and ecological burden across the United States and Canada, additional studies into these weevil milfoil specialists are of particular importance.

24

CHAPTER II

PHYLOGEOGRAPHIC CHARACTERIZATION OF GENETIC VARIATION IN

EUHRYCHIOPSIS LECONTEI THROUGHOUT NORTH AMERICA.

Introduction

The term “biological diversity” was first popularized by Thomas Lovejoy in 1980 as a way to characterize the vast amount of variation observed in the world’s biota (Soulé and Wilcox, 1980). The term was later shortened to “biodiversity” and was rapidly adopted by conservation biologists and the general public as a way of describing and delineating both the variety of life on earth as well as the pressures influencing the survival of this variable biota (Wilson and Peter, 1988). However, the definition of biodiversity itself has long been a complicated issue, with little consistency across disciplines (DeLong, 1996). For example, biodiversity, as defined by the United States

Congress Office of Technology Assessment (1987, p. 3), consists of the “variety and variability” of three major biological subdivisions: genetic, species and ecosystem.

DeLong (1996, p. 745) suggests that this definition lacks important functional and operational facets of biodiversity and therefore suggests the following all-encompassing definition:

25 Biodiversity is a state or attribute of a site or area and specifically refers to the variety within and among living organisms, assemblages of living organisms, biotic communities, and biotic processes, whether naturally occurring or modified by humans. Biodiversity can be measured in terms of genetic diversity and the identity and number of different types of species, assemblages of species, biotic communities, and biotic processes, and the amount (e.g., abundance, biomass, cover, rate) and structure of each. It can be observed and measured at any spatial scale ranging from microsites and habitat patches to the entire biosphere.

Genetic diversity is often studied in terms of evolutionary biology (Hughes et al.,

2008). It has been suggested that the “information content,” and therefore the evolutionary potential, contained in an organism’s DNA should be a major focus of biodiversity preservation (Crozier, 1997; Gugerli et al., 2008). Hughes et al. (2008) also suggest that intraspecific genetic diversity can have important effects on populations, communities and ecosystems, such that this genetic diversity may be as integral to ecosystem functioning as is species diversity. Therefore, the maintenance of the genetic diversity within species is clearly becoming an important consideration for conservation biologists (Allendorf et al., 2010; Avise, 2010).

The ability to measure some form of genetic diversity in a comparable manner within a species across their range (i.e., from a variety of geographic locations) can shed light on how such diversity arose, and perhaps elucidate the necessary steps required to maintain that biodiversity (Purvis and Hector, 2000). A methodology to measure genetic diversity in an accurate, cost-effective manner was first proposed by Herbert et al., in

2003. This method is called Genetic (or DNA) barcoding, a technique where a certain portion of the genome is sequenced and then compared to a database of reference

26 sequences to confirm species-level identifications (Herbert et al., 2003; Ascunce et al.,

2009; Hajibabaei et al., 2007). DNA barcoding allows for quick identification and assessment of intraspecific variation, particularly in a time where there appears to be an increasing shortage of taxonomic experts (Adler and Foottit, 2009). This technique can also reveal the occurrence of cryptic species, which can have profound implications for not only global biodiversity estimates, but also for ecosystem functioning (Adler and

Foottit, 2009; Bickford et al., 2006). It has also been proposed that this methodology can provide valuable insight into complex phylogenetic and population genetics studies as more sequences are added to the reference libraries (Hajibabaei et al., 2007) and can function as a “global bioidentification system for animals” (Herbert et al., 2003).

DNA barcoding utilizes a portion of the mitochondrial gene Cytochrome Oxidase

I (COI) to measure genetic diversity within and across animal species (Caterino et al.,

2000; Herbert et al., 2003). Some studies which investigated genetic structure of various organisms utilized the number of differences in base pairs within the COI gene sequence among populations to define and assign haplotypes to each population (Waits et al.,

1998; Cognato et al., 2003). Haplotypes can differ by as little as one base pair and one population may contain more than one haplotype (Waits et al., 1998; Cognato et al.,

2003).

Once haplotypes are assigned to the various populations, Geographic Information

Systems (GIS) can be used to examine the data in a geographic context. A rapidly- growing field of study that utilizes GIS (i.e., “phylogeography”) addresses the interaction between genetic variation, phylogenetic relationships and the environmental variables

27 associated with geography (Manel et al., 2003; Wang, 2010). Intraspecific phylogenetics can delineate geographic subspecies, greatly contributing to insights about regional biodiversity (Hewitt, 2004). The use of GIS data and spatial statistics to analyze geographic patterns of genetic diversity represents an integrative approach utilizing principles and tools from both biology and geography.

From a biodiversity perspective, the class Insecta represents 58% – 67% of all described species on earth (Adler and Footit, 2009). However, it is widely accepted that insect diversity is poorly understood and the actual number of insect species on earth is still being debated (Sabrosky, 1958; Stork, 1993). Insects are extremely important in ecosystem functioning (Adler and Foottit, 2009; Weisser and Siemann, 2004) and provide over $57 billion/year to the United States alone in terms of ecological services

(Losey and Vaughan, 2006). Some of these important functions include: pollination, improvement of soil health, waste management, biocontrol of weeds and other insect pests, medical uses, food sources for many other animals – the list goes on and on. E.O.

Wilson (1985) famously coined the phrase that “If insects were to vanish, the environment would collapse into chaos”.

Of increasing concern, then, is that the estimates of diversity and ecological services do not typically include incipient species, morphologically-indistinguishable species, or cryptic species (Bickford, et al., 2006; Adler and Footit, 2009). An obvious conclusion is that the Insecta would therefore constitute an extremely important group of organisms to target with DNA barcoding and with phylogeographic studies (Adler and

Foottit, 2009) to not only increase our knowledge of insect biodiversity but also because

28 improper identification of economically important insects (e.g., those used in biological control or for medicines) can have dire consequences on ecological or human health

(Bickford et al., 2006).

Beetles (Coleoptera) are the most biologically diverse group of described insects and the Curculionidae, the “true” weevils, exceed over 51,000 described species, making up at least 17% of known Coleopterans (Bouchard, et al., 2009). Weevils are phytophagous insects and feed on virtually all types of plants (Oberprieler, et al., 2007).

DNA barcoding has become an essential tool in correctly identifying weevils that are considered pests of many economically important plant species, such as the avocado stem weevil (Engstrand et al., 2010), citrus root weevil (Ascunce et al., 2009), and boll weevil

(Barr et al., 2013). Within the Curculionidae, only a few subfamilies (i.e., Bagoinae,

Brachycerinae, Ceutorhynchinae, Curculioninae) are recognized as having species that utilize aquatic or semi-aquatic plants as hosts, and most of these species are poorly studied (Center et al., 2002). Some of these weevils utilize wetland plants that are important food for wildlife (e.g., Listronotus spp. feed on arrowheads; Tanysphyrus spp. feed on duckweeds; Rhinocus spp. feed on smartweeds; Center et al., 2002). Other weevil species seem to have potential as biological control agents for invasive aquatic plants (e.g., Bagous spp. feed on Hydrilla; Neohydronomus spp. feed on waterlettuce;

Neochetina feed on waterhyacinth; Center et al., 2002). The simple system of DNA barcoding would be useful for adding to our knowledge of the biodiversity of aquatic weevils as well as for correctly identifying weevil biocontrol agents that are used in the management of invasive aquatic plants.

29 The primary purpose of this project was to ascertain the genetic diversity of natural populations of Euhrychiopsis lecontei. This was accomplished by analyzing COI sequence data that have been collected for E. lecontei throughout their native range. GIS and spatial statistics were then used to assess which forces, if any, may be influencing the genetic structuring of E. lecontei populations. This information allowed an assessment of the diversity of an important biological control agent that is currently being used to manage an invasive aquatic weed, Myriophyllum spicatum (Eurasian watermilfoil).

Methods and Materials

Euhrychiopsis lecontei were sampled from 35 populations throughout the U.S.

(Table 2.1.) Sample sizes ranged from 1-4 individuals per population for a total of 67 individuals.

Specimens were placed live into ATL lysis buffer and were ground with a modified pipette tip directly in a 1.5 mL centrifuge tube and stored at room temperature until DNA extraction. Some weevil samples collected remotely were preserved in water at -20°C. DNA was extracted with a DNeasy® Blood and Tissue kit (Qiagen®) following the protocol for purification of total DNA from insects (DNeasy® Blood and

Tissue Handbook, 2006).

Amplification of the partial mtDNA Cytochrome Oxidase COI subunit 1 sequence

(~1085bp) was conducted using the standard primers LCO 5’ –

GGTCAACAAATCATAAAGATATTGG –3’ and C1N2776 5’ –

GGATAATCAGAATATCGTCGAGG – 3’. PCR amplification was conducted with 1µL

30

DNA extract in a 25 µL reaction mixture using Promega HotStart Taq. Amplification conditions included an initial denaturation period of 2 min at 94°C , followed by ten -1°C stepdown cycles beginning at 55°C and ending at 45°C, then 30 cycles of 94°C denaturation (30s), 45°C annealing (30s) and 72°C extension (60s) with a final extension of 10 minutes. Purification and sequencing of the samples were completed by Annis

Water Resources Institute (Muskegon, Michigan) using the same primers used in PCR amplification. Forward and reverse sequences were assembled, manually reviewed and edited with the software package Sequencher®. Sequences were imported into MEGA5

(Tamura et al., 2011) and aligned with the built-in automated aligner CLUSTALW.

Sequences were trimmed to the shared 986 base pairs for the purposes of determining the haplotypes and to ensure that the sequence was properly coding for amino acids.

Sequences were then uploaded to DnaSP v.5.10.01, a program which analyzes polymorphism data between sequences and outputs the haplotype file (Rozas et al.,

2010). Based on the haplotype designations, a Maximum Likelihood Tree with 500 bootstraps was constructed in MEGA 5.2 (Tamura et al., 2011). Outgroups for the tree were chosen from previous studies (Parenthis vestitus was determined to be the sister species of Euhrychiopsis lecontei) or from GenBank from the subfamily

Ceutorhynchinae, to which E. lecontei belongs. GenBank accession numbers for the outgroups are Ceutorhynchus neglectus (DQ058697.1), C. erysimi (DQ058698.1), and C. gallorhenanus (DQ058700.1). Each haplotype sequence was input into the online identification system for the Barcode of Life Database (BOLD) to verify percent similarity with known specimens in the database

31

Table 2.1. Euhrychiopsis lecontei collection locality data and sample size. Waterbody State/Province Latitude Longitude Sample Size (n) Auburn Lake Minnesota 43°35'56.55"N 88°12'23.51"W 4 Big Bald Lake Ontario Canada 44°34'38.15"N 78° 23' 31.10"W 2 Burke Lake Washington 47°8'0.92"N 119°55'6.10"W 1 Cedarville Bay Michigan 45°59'36.44"N 84°21'3.69"W 3 Chain Lake Michigan 45°52'40.09"N 84°45'17.92"W 2 Chalk Hill, Menominee 45°29’31.42"N 87°47’54.36"W 1 River Michigan Christmas Lake Minnesota 44°53'53"N 93°32'29"W 3 Clark Fork Driftyards Idaho 48°10'36.65"N 116°14'7.29"W 1 Clear Lake Ontario Canada 46° 14' 12.67"N 81° 45' 34.55"W 1 Crystal Lake Idaho 48°10'35.85"N 116°52'45.03"W 2 Fairfield Pond Vermont 44°51'14.28"N 72°59'25.89"W 1 Indian Lake Connecticut 41°55'1.56"N 73°29'45.94"W 1 Kingsford Res., 45°49'18.22"N 88°08'07.79"W 2 Menominee River Michigan Lake Canadis Ohio 41°8'53.71"N 81°26'50.27"W 2 Lake Carroll Illinois 42°10' 49.62"N 89° 52' 53.69"W 1 Lake Eligo Vermont 44°35'58.51"N 72°21'22.73"W 3 Lake Hodgson Ohio 41°7'56.06"N 81°17'18.09"W 1 Lake Ovid Michigan 42°56'18.28"N 84°24'38.90"W 3 Lake Scugog Ontario Canada 44° 33' 44.43"N 78°8'39.25"W 2 Little Bearskin Wisconsin 45°42' 37.57"N 89°42'4.74"W 3 Luna Lake Ohio 40°55'10.68"N 81°37'6.88"W 1 McDill Pond Wisconsin 44°30'6.66"N 89°32'56.62"W 4 McFarlane Lake Ontario Canada 46°24'57.91"N 80°57'47.34"W 1 Michigamme Falls, 45°57'59.47"N 88°12'29.37"W 2 Menominee River Michigan Minocqua Lake Wisconsin 45° 52' 26.17"N 89°41'40.97"W 2 North Lake Wisconsin 43° 9' 12.07"N 88° 22' 44.12"W 2 Osoyoos Lake Washington 48°58'36.29"N 119°26'19.201"W 1 Otter Lake Minnesota 44°53'18.28"N 94°24'32.89"W 4 Peavy Falls, Menominee River Michigan 46°00'12.76"N 88°12’37.54"W 1 Pigeon Lake Ontario Canada 44°33'25.94"N 78° 30' 0.63"W 1 Richard Lake Ontario Canada 46°26'14.47"N 80°54'57.86"W 2 Spring Lake Wisconsin 44°0'39.92"N 89°9'37.48"W 3 Tripp Lake Wisconsin 42°49'34.51"N 88°43'8.96"W 2 White Rapids, Menominee River Michigan 45°32'12.35"N 87°48'16.41"W 1 Williams Lake Wisconsin 43°45'44.09"N 89°22'31.88"W 3

(http://www.boldsystems.org/index.php/IDS_OpenIdEngine; Ratnasingham and Herbert,

2007). The haplotype file was also subsequently analyzed with Network v.4.6.1.2 which 32 allowed for the creation of a haplotype network through the median joining calculation

(Bandelt et al., 1999; fluxus-engineering.com).

Location coordinates (x,y values; latitude, longitude) for the samples were input into ArcGIS v.10 and a distance matrix was created. In all cases, the centroid of the waterbody where the sample was collected was used. A Mantel test was conducted in the program Alleles in Space (Miller, 2005). Alleles in Space also was used to determine the genetic barrier(s) across the landscape using Monmonier’s Maximum Difference

Algorithm.

A spatial analysis of molecular variance (SAMOVA) was conducted in

SAMOVA 1.0 (Dupanloup et al., 2002). This program uses a simulated annealing process repeated 100 times to determine which groups of populations are maximally genetically differentiated from each other, without previous assumptions regarding which individuals are assigned to each group (Eble et al., 2011). The program was run for K=2

– 10 groups to determine the most robust groupings of populations. The smallest K that gives the highest percent of variation explained by the proposed sample groupings (Fct), that is not significantly different from a higher percentage of variation described by a higher K, was utilized (Eble et al., 2011). A likelihood ratio test was utilized to determine the grouping value (K) that explained the greatest variation in the data. Following the assignation of population groups, Tajima D’s statistic was calculated for each population grouping in DnaSP v.5.10.01 (Rozas et al., 2010) to test for signs of recent population expansion. Mismatch distributions were calculated for the whole data set and for each population grouping independently in DnaSP v.5.10.01 9 (Rozas et al., 2010) to obtain

33 tau (τ), which functions as a unit of mutational time. These results were used in a program that estimates time since divergence from substitution rates (http://www.uni- graz.at/zoowww/mismatchcalc/).

Results

For the 67 individuals of Euhrychiopsis lecontei sampled from 35 populations, 39 haplotypes were detected with an average number of nucleotide differences of 8.83

(Table 2.2). 984 of the 986 base pair sites (missing data were excluded) were utilized in the analysis conducted by DnaSP v.5 (Librado and Rozas, 2009). Haplotype diversity was high, whereas nucleotide diversity was low (Table 2.2). Tajima’s D statistic was negative, but not significant (Table 2.2). Data were calculated in DnaSP v.5 (Librado and

Rozas, 2009); and MEGA 5.2 (Tamura et al., 2011).

Table 2.2. Polymorphism summary statistics for total data set based on 984 bp of Euhrychiopsis lecontei COI mtDNA. No. of No. of Average No. of Haplotype Nucleotide Tajima’s D; p- samples haplotypes Nucleotide Diversity diversity value Differences (k); (h); SD (π); SD SE 67 39 8.834; 1.695 0.958; 0.00898; -0.40230; 0.014 0.00075 p>0.10

Of the 984 base pair sites, 48 were polymorphic. 16 of these polymorphisms were singleton mutations with two variants, and 32 mutations were parsimoniously informative with two variants. No transversions were detected in the data set. The translated protein code (MEGA5; Tamura et al., 2011) revealed 328 amino acids in the final data set. One

34 non-synonymous substitution was detected in the translated protein code. A base pair change at position 331 in the sequence data (C → T) resulted in three haplotypes (H33,

H34, and H38) having a serine (TCA) instead of a proline (CCA) at amino acid position

111.

The maximum likelihood tree for the haplotype data set indicates that all haplotypes of the sampled Euhrychiopsis lecontei form a monophyletic group with 98% bootstrap support (Fig. 2.1). Within E. lecontei, most branch lengths (which measure the number of substitutions per site) ranged from 0.0 to <0.0004. Based on a branch length of 0.008 and a 92% bootstrap support, the tree indicates that there are two major groupings of haplotypes: Clade A contains 27 haplotypes made up of 49 individuals from

26 populations; Clade B contains 12 haplotypes made up of 18 individuals from 9 populations (Table 2.3). Clade A is composed of haplotypes: H1–H3, H6, H10–H14,

H16–H17, H19–H21, H23, H27–H37 and H39 (Fig. 2.1 – green box; Table 2.3). H37 is slightly separated from the other individuals in Clade A by a branch length of 0.006 and

72% bootstrap support (Fig. 2.1). Clade B is composed of haplotypes: H4–H5, H7–H9,

H15, H18, H22, H24–H26, and H38 (Fig. 2.1 – red box; Table 2.3).

35

Haplotype 1 Haplotype 12 Haplotype 32 Haplotype 36

6 Haplotype 3 Haplotype 33 Haplotype 2

31 Haplotype 19 Haplotype 21 22 31 Haplotype 27 Haplotype 30 Haplotype 28 Haplotype 39 Haplotype 35 Haplotype 20 Haplotype 23 25 60 Haplotype 11 Haplotype 13 Haplotype 17 20 Haplotype 29

72 Haplotype 10 Haplotype 31 Haplotype 16 92 34 Haplotype 6 91 Haplotype 14 Haplotype 34 Haplotype 37

87 Haplotype 5 27 Haplotype 38 98 Haplotype 18 Haplotype 4 64 Haplotype 26 Haplotype 22 Haplotype 8 100 Haplotype 24 Haplotype 15 40 Haplotype 25 37 Haplotype 7 Haplotype 9 Parenthis vestitus Ceutorhynchus erysimi Ceutorhynchus neglectus 37 Ceutorhynchus gallorhenanus

0.05

36

Figure 2.1. Molecular Phylogenetic analysis by Maximum Likelihood method. The evolutionary history of a portion of COI mtDNA from 67 individuals of Euhrychiopsis lecontei was inferred using the Maximum Likelihood method based on the Tamura 3-parameter model (Tamura, 1992). The tree with the highest log likelihood (- 3338.7135) is shown. The percentage of trees (> 50%) in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter = 0.1597)]. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 43 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 986 positions in the final dataset. Evolutionary analyses were conducted in MEGA5 (Tamura et al., 2011). Designation of Clade A is marked with a green box and designation of Clade B is marked with a red box. Yellow stars indicate >99.0% similarity to known specimens of E. lecontei in the BOLD data set.

37

Table 2.3. List of waterbodies (location), associated haplotypes, sample size of each Haplotype/waterbody (N) and clade assignation within each population. Waterbody (Location) Haplotypes Clade Assignation based (Number of Individuals) on ML Tree, Haplotype Network and SAMOVA results Auburn Lake (MN) H1(1), H2(1), H3(1) A Big Bald Lake (Ont.) H4(1), H5(1) B Burke Lake (WA) H6(1) A Cedarville Bay (MI) H7(2), H8(1) B Chain Lake (MI) H9(2) B Chalk Hill Reservoir (MI) H10(1) A Christmas Lake (MN) H11(1), H12(1), H13(1) A Clark Fork Driftyards (ID) H14(1) A Clear Lake (Ont.) H15(1) B Crystal Lake (ID) H6(1), H16(1) A Fairfield Pond (VT) H17(1) A Indian Lake (CT) H18(1) B Kingsford Reservoir (MI) H12(2) A Lake Canadis (OH) H19(1), H20(1) A Lake Carroll (IL) H21(1) A Lake Eligo (VT) H22(3) B Lake Hodgson (OH) H20(1) A Lake Ovid (MI) H7(1), H24(1), H25(1) B Lake Scugog (Ont.) H4(1), H26(1) B Little Bearskin (WI) H12(2), H27(1) A Luna Lake (OH) H28(1) A McDill Pond (WI) H21(1), H29(1), H30(1), H31(1) A McFarlane Lake (Ont.) H28(1) A Michigamme Falls Reservoir (MI) H32(1), H33(1) A Minocqua Lake (WI) H23(1) A North Lake (WI) H27(1); H28(1) A Osoyoos Lake (WA) H34(1) A Otter Lake (MN) H12(1), H21(1), H35(1); H36(1) A Peavy Falls Reservoir (MI) H37(1) A Pigeon Lake (Ont.) H38(1) B Richard Lake (Ont.) H28(2) A Spring Lake (WI) H27(1), H28(2) A Tripp Lake (WI) H28(1), H39(1) A White Rapids Reservoir (MI) H12(1) A Williams Lake (WI) H28(3) A

All haplotype sequences were uploaded to BOLD, which reported percent similarity to known specimens of E. lecontei ranging from 97.45% – 99.81%. BOLD only positively confirms species identity for individuals with >99% similarity.

Haplotypes H6, H14, H37, and H39 were confirmed as Euhrychiopsis lecontei. 38

The haplotype network shows two main haplotype groupings, which exactly match the designations of Clades A and B from the ML tree, and are separated by 10 base pair changes (Fig. 2.2). One haplotype (H37) is separated from Clade A by 9 base pair changes (Fig. 2.2). Therefore, this haplotype is appears to constitute a third type that is distinct from the other two clades.

With K=2, the SAMOVA results independently support the separation of the

Clades A and B based on the population locations (Table 2.3; Table 2.4). Although higher Fct’s were obtained with increasing K, the difference between K=2 and K=3 was not significant based on a Likelihood Ratio Test (p>0.75). Therefore, the majority of the variation is between groups (Clade A and Clade B; Table 2.4).

39

Figure 2.2. Haplotype Network of 39 haplotypes inferred from 984 base pairs of mtDNA COI gene for 67 individuals of Euhrychiopsis lecontei. Sample sizes ranged from one to four individuals per population. Black circles indicate missing intermediate haplotypes. Size of the yellow circles correlates to the number of individuals that share that haplotype. Red numbers correlate to the polymorphic positions in the data set. See Tables 2.1 and 2.2 for exact locations of the haplotypes. Clade A is designated by a green box and Clade B is designated by a red box.

40 Table 2.4. SAMOVA results for K=2 groups. Source of Degrees of % of variation Fixation p-value variation freedom indices Between 1 77.6 Fct = 0.77629 p<0.0001 clades Among lakes 33 10.6 Fsc= 0.47223 p<0.0001 within clades Within lakes 32 11.8 Fst = 0.88193 p<0.0001

Of the 984 sites that were used in the analysis, the number of variable sites within

Clade A was 37; 23 of these polymorphisms were singleton mutations with two variants and 14 mutations were parsimoniously informative with two variants. Within Clade B,

16 sites were variable; seven of these polymorphisms were singleton mutations with two variants and nine were parsimoniously informative with two variants. Between clades, there were two fixed differences: 30 mutations that were polymorphic in Clade A but monomorphic in Clade B, nine mutations that were polymorphic in Clade B but monomorphic in Clade A, and seven shared mutations. The average number of nucleotide differences between populations was 16.7. Haplotype diversity within each clade was high, while nucleotide diversity was low (Table 2.5). The average number of nucleotide differences, the haplotype diversity and the nucleotide diversity within each

Clade were similar to each other (Table 2.5). Both Clade A and Clade B had negative

Tajima D values, but was statistically significant only for Clade A (Table 2.5).

The Mantel test conducted in Alleles in Space (Miller, 2005) indicates that there was a positive (r=0.223) and significant (p=0.002) correlation between geographic and

41 genetic distances from 1000 replicates performed. Therefore, individuals that are closer to each other geographically are more likely to be more closely related genetically.

Table 2.5. Polymorphism summary statistics for Clades A and B based on 984 bp of Euhrychiopsis lecontei COI mtDNA. Clade No. of No. of Average No. Haplotype Nucleotide Tajima’s D; samples haplotypes of Diversity diversity p-value Nucleotide (h); SD (π); SD Differences (k); SE A 49 27 3.561; 0.823 0.930; 0.00362; -1.94; 0.024 0.00051 p < 0.05 B 18 12 3.889; 1.118 0.948; 0.00395: -0.63; 0.001 0.00075 p > 0.10 Data were calculated in DnaSP v.5 (Librado and Rozas, 2009); and MEGA 5.2 (Tamura et al., 2011).

Through Monmonier’s Maximum Difference Algorithm, Alleles in Space

(Miller, 2005) predicted the geographic location of a barrier to gene flow. The predicted barrier exactly matches the geographic separation already determined by the partitioning of Clades A and B through the SAMOVA analysis (Fig. 2.3).

Mismatch distributions calculated for Clade A gave a τ = 1.578 and for Clade B, τ =

3.220. Using divergence estimates of 2.3%/million years (Brower, 1994) as the lower limit and 3.54%/million years (Papadopoulou et al., 2010) as the upper limit, and a generation time of 0.25 years (Newman et al., 2001 reports 3 – 6 generations per year), the online mismatch calculator (http://www.uni- graz.at/zoowww/mismatchcalc/mmc1.php) predicts that Clade A’s time since expansion is approximately 69,582 years ago and Clade B’s time since expansion is approximately

141,987 years ago.

42

Figure 2.3. Barrier to gene flow predicted by raw genetic data and geographic locations in Alleles in Space (Miller, 2005) is indicated by the purple line. Clades designations are represented by green (Clade A) and red (Clade B) circles.

Discussion

The primary purpose of this study was to conduct broad-scale genetic sampling of an aquatic weevil, Euhrychiopsis lecontei that is currently being used as a biocontrol agent for the invasive weed, Myriophyllum spicatum (Eurasian watermilfoil) and to determine what phylogeographic patterns exist, if any. This is not only specifically important for management implications (discussed below) but also to add to our knowledge of a rather understudied species of aquatic weevil. Because freshwater biodiversity is currently under intense stress from human activities, those organisms that rely on freshwater ecosystems are more vulnerable to eventual extinction (Dudgeon et al.,

2005). A recent study by Pimm et al. (2014) discusses the importance of classification of relatively unknown species, as more accurate biodiversity estimates worldwide are imperative for conservation purposes. Increasing documentation of poorly studied aquatic species, such as E. lecontei, may be even more relevant today because extinction rates for 43 freshwater species with low dispersal rates may be climbing past the point of no return as a consequence of habitat modification and pollution (Pimm et al., 2014).

Euhrychiopsis lecontei’s ancestral host is Myriophyllum sibiricum (Northern watermilfoil) and these weevils are distributed throughout the northern United States and southern Canadian provinces (Creed, 1998; Newman, 2004). Euhrychiopsis lecontei populations are characterized as temporally and spatially patchy (Creed and Sheldon,

1994; Tamayo et al., 2000) and their exact range is questionable. For example, E. lecontei were only recently found in California (Cline et al., 2013) and are still reported as absent from Montana, Wyoming, North Dakota and South Dakota (Creed, 1998).

Accurately determining their distribution is difficult because both the weevil’s small size and its relatively low population densities make field sampling difficult. Therefore, perceived absence of this species in an area could simply be the result of limited sampling (i.e., lack of sampling effort, inaccessibility to sites where this species actually resides, or inappropriate timing of sampling efforts). Alternatively, absence of E. lecontei in a certain area could reflect an actual gap in the distribution resulting from an artifact of unsuitable habitat imposed by environmental or ecological parameters.

However, the range of E. lecontei may very well be expanding with the spread of the invasive Eurasian milfoil into more lakes (Jester et al., 2000). Because of sampling and taxonomic identification difficulties, having a relatively quick and accurate method of determining if an individual belongs to E. lecontei or is a cryptic species would be useful.

DNA barcoding is such a useful tool, both for biological control reasons but also allowing for greater understanding of this species and its distribution.

44 The maximum likelihood tree (Fig. 2.1) reveals the monophyly of Euhrychiopsis lecontei with 98% bootstrap support and lends credibility to the hypothesis that all 67 individuals sampled were the same species, since the intraspecific distances in this species are less than the intraspecific distances within the outgroup (Ceutorhynchnus). As a group, haplotype diversity (Hd, Nei, 1987) was high, approaching 1, which indicates that most individuals within the data set have their own unique haplotype (Table 2.2;

Adams and Villablanca, 2007). Nucleotide diversity (π) was low (Table 2.2). Nucleotide diversity tends to change more slowly than haplotype diversity (Adams and Villablanca,

2007). However, the combination of high haplotype diversity and low nucleotide diversity can suggest that the population as a whole has undergone a relatively recent range expansion from a small remnant population (Avise, 2000; Jong et al., 2011). This is supported, although not significantly, by the negative Tajima D statistic (Table 2.2). Fu and Li’s D and T test statistics (data not reported) were also negative but not significant.

However, the maximum likelihood tree (Fig. 2.1), the haplotype network (Fig.

2.2), the SAMOVA results (Table 2.4), and the barrier to gene flow (Fig. 2.3) clearly demonstrate that there are genetic and geographic structuring within the species, revealing two major Clades (A and B). Clade B appears to be more closely related to the outgroups and is therefore slightly ancestral to Clade A with 92% bootstrap support (Fig.

2.1).

Comparison of the 39 haplotype sequences to The Barcode of Life Database

(BOLD) reported similarities to three known specimens of Euhrychiopsis lecontei (two from Manitoba, Canada and one from Alaska) ranging from 97.45% – 99.81%. It is

45 clear, based on the placement of the haplotypes (H6, H14, H37, H39) that were >99.0% similar within the maximum likelihood tree (Fig. 2.1), that all haplotypes included in

Clade A can be confirmed as E. lecontei. Several haplotypes within Clade B (H4, H7,

H8, H9, H15, H24, H25) were >98% similar to the three known specimens of E. lecontei in BOLD. Although they cannot be positively identified as E. lecontei based on this methodology, it is likely that increased deposition of known specimens of this species into BOLD from throughout their range would allow for positive confirmation of the members of Clade B through BOLD as well. It should be noted that H37 (Peavy Falls,

Menonminee River, MI), the “outlier” in Figs. 2.2 and 2.3, is one of the individuals with

>99% similarity to the known specimens in BOLD. Since this individual is so closely related to individuals from Manitoba, Canada, this suggests that there may be more geographic structuring within E. lecontei than this study revealed, which could be confirmed with increased sample size from more disjunct locations (especially northern) throughout their range.

The Mantel test results reveal that there is a positive and statistically significant correlation between genetic differentiation and geographic distance, supporting the hypothesis for some geographic structuring occurring in this species. This result was expected because Euhrychiopsis lecontei tend to have rather limited dispersal capabilities. Newman et al. (2001) report that weevils only tend to disperse by flight in spring and fall when they are moving between the lake and their on-shore overwintering habitats. Furthermore, the authors state that overwintering E. lecontei were generally only found one to two meters from shore, although some specimens were found 20

46 meters from shore. Therefore, unless there are many lakes in close proximity, it is relatively unlikely that E. lecontei are dispersing between lakes during their fall/spring flights. Rather, overwintering E. lecontei will often return to the same lake where they were born, and perhaps even to the same milfoil bed. In mid-summer, flight muscles are extremely reduced, making dispersal during the breeding season extremely limited

(Newman et al., 2001).

The result of the Mantel test is also congruent with this weevil species’(and its host plant’s) ecologies, not only because of E. lecontei’s limited dispersal capability but also because one primary mode by which Eurasian watermilfoil tends to spread is vegetatively through viable fragments after autofragmentation or mechanical fragmentation (Madsen and Smith, 1997). Fragments of Eurasian watermilfoil are often able to establish new populations once they settle onto the substrate after being dispersed by water flow or boat traffic, especially between closely situated waterbodies (Madsen and Smith, 1997). Since E. lecontei’s life cycle is tied so intimately with this plant, there is a great chance that viable fragments of Eurasian watermilfoil that move between waterbodies may be transporting eggs, larvae, pupae or adults of E. lecontei between these same waterbodies. Thus, even with E. lecontei’s limited dispersal capabilities through swimming or flight, they are likely able to achieve intermediate dispersal between hydrologically-linked waterbodies by “hitching a ride” on Eurasian watermilfoil fragments. However, it should be noted that even populations separated by a great geographic distance (e.g., those in Washington/Idaho vs. those in Minnesota/Wisconsin; see Tables 2.1 and 2.3; Fig. 2.3) fall within the same clade, suggesting that there are other

47 evolutionary forces, such as past glaciation events, playing a role in the geographic structuring of this weevil.

When examining the demographic history of each clade independently (Table

2.5), similar (but not identical) demographic histories can be inferred. Both clades have high Hd and low π, indicating that each clade has independently undergone a relatively recent range expansion from small remnant populations (Avise, 2000; Jong et al., 2011).

When comparing the Hd and π values from each clade, Hd is significantly different between Clade A and Clade B, but π is only marginally significantly different (Clade A does not overlap the mean of Clade B, but Clade B does overlap the mean for Clade A) based on 95% confidence intervals. Furthermore, although both Clades A and B have negative Tajima’s D values, they are only significantly negative for Clade A. Negative

Tajima D’s can indicate that the population has recently undergone a demographic expansion (Tajima, 1989). Several recent studies note that a star-like pattern in a haplotype network (see H12 and associated star pattern in Clade A; Fig. 2.2) can be indicative of a bottleneck and recent population expansion (Allcock and Strugnell, 2012;

Ludt et al., 2012; Ley and Hardy, 2014).

Together, these results suggest that Clade A has undergone a more recent demographic expansion than Clade B. This is supported by predictions from the online mismatch calculator (http://www.uni-graz.at/zoowww/mismatchcalc/mmc1.php) which indicated that the time since expansion for Clade A was approximately 69,582 years ago and the time since expansion for Clade B was over two times longer (~141,987 years ago). These dates fall within the Wisconsin glaciation (Clade A) and Illinoian glaciation

48

(Clade B) of the Laurentide icesheet (Richmond and Fullerton, 1986). The extents of the glacial maxima for these two glaciation events are shown in Figure 2.4.

The cumulative information from the maximum likelihood tree (Fig. 2.1), Hd, π,

Tajima’s D (Table 2.5) and the barrier to gene flow (Fig. 2.3) suggests that these groups have been isolated, potentially through several glacial and interglacial periods, and survived in isolated glacial refugia, as suggested in other similar studies of the phylogeographic patterns in another weevil species (acorn weevils, see Aoki et al., 2008 and 2009). The geographic locations of Clades A and B also support this conclusion, as

Figure 2.4. Proposed glacial refugia for Clades A (green) and B (red). Colored arrows depict the possible colonization routes of E. lecontei into the distinct geographic areas delineated by the genetic barrier determined by Alleles in Space (Miller, 2005).

the proposed glacial refugia indicated in Fig. 2.4 align perfectly with two of the ten proposed North American glacial refugia for terrestrial plants and animals (Beatty and

Provan, 2010). This study highlights the importance of such glacial refugia in maintaining geographic structure even in aquatic organisms. 49

Implications for Management of Eurasian watermilfoil

This project was also important in confirming that the species being utilized in different biocontrol rearing operations is correctly identified as Euhrychiopsis lecontei, the intended biocontrol agent for Myriophyllum spicatum (Eurasian watermilfoil).

Incorrectly identified specimens used in such biological control programs would likely be unsuccessful, since such biocontrol programs rely on host specificity of the insect.

It is also crucial to recognize the potential genetic diversity within the native populations of E. lecontei because any significant geographically-structured genetic variation between natural milfoil weevil populations may affect the development of successful and acceptable augmentative biocontrol programs for Eurasian watermilfoil.

Any genetic differentiation into distinct geographic areas will require careful choice of source populations for particular augmentation projects. Since only a few source populations are being utilized for rearing E. lecontei, (namely, Michigan weevils for eastern North America projects, Idaho weevils for western U.S. projects and Ontario,

Canada weevils for Ontario projects), human-mediated, long-distance dispersal of organisms (as opposed to typical short-distance natural dispersal) can have dramatic effects on genetic diversity (Trakhtenbrot et al., 2005) through increased gene flow from disparate populations.

Commercial companies that provide Euhrychiopsis lecontei for biological control projects have encountered regulatory constraints where state or local agencies have dictated that source populations for a specific project must come from the lake where the laboratory-reared biocontrol agents will be deployed. However, based on the information

50 garnered in this project, it does not appear necessary to limit source collections for rearing populations in this way. In broad geographic areas (i.e., Clade A vs. Clade B), many individuals from independent populations (even across state or country boundaries) share the same haplotype (e.g., Hap12) and haplotypes can vary by as little as one base pair (e.g., Hap12 and Hap27; Fig. 2.2). Thus, requiring that source E. lecontei populations come from the same lake as where the project will be conducted is unwarranted based on the relatively widespread regional genetic similarity of this organism, even though ecological differences might occur.

Conclusions and Future Studies

Multiple lines of evidence support the hypothesis that all individuals sampled from a broad geographic region throughout Euhrychiopsis lecontei’s range are one species. Based on the analysis by the Barcode of Life database, no cryptic species were discovered through this study, although distinct geographic and genetic structuring was revealed, likely due to current geographic isolation of the sampled waterbodies, as well as the locations of historic glacial refugia. Further work should include more geographically widespread sampling throughout the U.S. and Canada and specifically targeted in areas where E. lecontei has not yet been recorded (i.e., Montana, Wyoming,

North Dakota, South Dakota), not only for the increase in species-specific knowledge about this important weevil taxon, but also to potentially reveal more complex geographic structuring. Additionally, because baseline genetic data has now been recorded for natural populations of E. lecontei, studies should be conducted on

51 augmented populations to verify that rearing and deploying the insect has not degraded the genetic variability, and potentially its effectiveness, as a biocontrol agent (Mackauer,

1976).

52

CHAPTER III

DETERMINATION OF POTENTIAL HOST RANGE SHIFT AND SYMPATRIC

SPECIATION OF EUHRYCHIOPSIS LECONTEI FEEDING ON AN ANCESTRAL

HOST (MYRIOPHYLLUM SIBIRICUM) VERSUS A NOVEL, INTRODUCED HOST

(MYRIOPHYLLUM SPICATUM).

Introduction

The formation of new species is the essence of evolution (Darwin, 1859).

However, since Darwin proposed the theory of natural selection as the driving force behind speciation, biologists have argued how speciation occurs and what actually constitutes a species. Ultimately, speciation is an adaptive process wherein barriers to gene flow lead to reproductive isolation of closely-related populations, ultimately leading to genetic divergence (Bush, 1975). Many evolutionary biologists were adamant that all speciation must occur through allopatry, based on some geographic boundary that leads to reproductive isolation (Via, 2001; Mallet, 2010).

Sympatric speciation (speciation with no geographic boundaries), on the other hand, has historically been a controversial topic among biologists as a viable mode of

53 speciation because evidence that gene flow could be eliminated in overlapping populations was scant (Via, 2001; Jiggins, 2006; Bolnick and Fitzpatrick, 2007; Nosil

2008; Mallet et al., 2009). Sympatric speciation is a result of intrinsic factors which reduce gene flow and lead to reproductive isolation within the range of the species in patchy environments (Bush, 1975; Via, 2001).

The formation of host races appears to be a major mechanism for sympatric speciation (Drès and Mallet, 2002). Bush (1975) proposed that the formation of host races is an evolutionary stepping stone on the way to full speciation. The existence of host races within a population still allows for appreciable gene flow between the races

(Bush, 1975). Once a stable host race is formed, a reduction in gene flow and further reproductive isolation can eventually lead to complete reproductive isolation of the races and the formation of distinct sister species (Via, 2001).

The existence of host races has been documented in several different species of phytophagous insects (Camara, 1997; Futuyma, 2000). Camara (1997) hypothesized that because specialization in phytophagous insects is so prevalent, specialization as an attribute of this group is stable over evolutionary time. Essentially, host shifts must either be relatively uncommon, or those species that incorporate a new food source into their diet must then quickly diverge from their ancestral population to become specialists on the novel host. Otherwise, there would be far fewer specialist phytophagous insects and many more generalists that could feed on a variety of food sources. Bush (1975) asserted that “it is difficult to believe that all of the greater than 500,000 specialist phytophagous insects speciated allopatrically”. Within the past few decades, more

54 supporting evidence has accumulated that sympatric speciation can indeed occur through host range expansions and host shifts (Bush, 1994; Berlocher and Feder, 2002; Bolnick and Fitzpatrick, 2007). Therefore, this is an important and burgeoning field of study for evolutionary biologists.

Much of the pioneering research on host shifts was conducted on Rhagoletis

(Diptera, Tephritidae), the apple maggot fly (Bush, 1969; Feder et al., 1988). Apple maggot flies are native to North American and their ancestral host is the native hawthorn

(Crataegus spp.), but reports of the flies attacking domesticated, introduced apples date as far as back the 1860’s (Bush, 1969). Mate recognition and mate choice are closely correlated to host plant choice, suggesting pre-mating isolating mechanisms (Feder et al.,

1988). Furthermore, there appears to be a temporal separation of the host races based on when the fruits are ripe – early emerging adults reproduce on apple fruits whereas later emerging adults reproduce on hawthorn trees (Bush 1969, Feder et al., 1988). The genetic variation needed in the parent population to switch to a new host may be present before the new host is ever encountered, and therefore host shifts may be most likely when the ancestral and new hosts occur together within the dispersal capabilities of the insect (Bush, 1975). Feder et al. (1988) found early evidence of genetic differentiation between the two host races of Rhagoletis pomonella which suggested that there was non- random mating in sympatric populations.

Phytophagous insects, such as Euhrychiopsis lecontei (milfoil weevil), are ideal candidates for additional investigations of host range expansions/shifts for several reasons. They are host specialists that rely on a Myriophyllum spp. (or milfoil) for all

55 their needs (e.g., food, courtship, mating, oviposition and development) and they exhibit high host fidelity (Solarz and Newman, 1996). This weevil is native to North America and is distributed throughout the upper contiguous United States and southern Canada

(Creed, 1998). Euhrychiopsis lecontei’s ancestral host is Myriophyllum sibiricum

(Northern watermilfoil; Creed and Sheldon, 1994); however, it has been utilized as a biological control for M. spicatum (Eurasian watermilfoil) for approximately two decades

(Creed and Sheldon, 1995; Sheldon and Creed, 1995; Sheldon and Creed, 2003;

Newman, 2004; Tamayo and Grue, 2004).

Host expansions/shifts are more common when the shift is to a plant closely related to the ancestral host or when the novel host is one that is utilized by a close relative (Janz and Nylin, 2007). Accordingly, based on plant characteristics that favor host range expansion, it is not surprising that milfoil weevils have exhibited a host-range expansion to Eurasian watermilfoil for several reasons: Myriophyllum sibiricum and M. spicatum (the novel host) are closely related plants and are morphologically very similar

(Aiken et al., 1979). Additionally, M. spicatum is the host of the Eubrychius velutus

(European milfoil weevil), which is hypothesized to be closely related to Euhrychiopsis lecontei (Newman, 2004), thus indicating that E. lecontei may have the necessary traits to utilize M. spicatum effectively. Based on recent molecular findings regarding the confirmation of a close phylogenetic relationship between Eubrychius velutus and

Euhrychiopsis lecontei (Roketenetz, unpubl.), E. lecontei’s host shift to M. spicatum might have been easily predicted (Janz and Nylin, 2007).

56

Because evidence supports the fact that host range expansions and shifts can lead to sympatric speciation through genetic divergence (e.g., Scheffer and Hawthorne, 2007), further investigation into whether or not this can be documented in populations of

Euhrychiopsis lecontei is merited. Since the host expansion onto Myriophyllum spicatum has occurred in a relatively recent time frame (Myriophyllum spicatum was first documented in the US in the 1940’s and has been spreading to new locations ever since;

Smith and Barko, 1990), it may be possible to show host-associated genetic divergence, which would support early evidence for sympatric speciation of this weevil feeding on various host plants. To determine these more recent genetic changes within a species, microsatellite analyses are the appropriate methodology because these genomic regions are known to change quickly and therefore can provide evidence of recent population divergence (Selkoe and Toonen, 2006).

Therefore, the purpose of this study was to investigate, through the use of six microsatellite markers, whether natural populations of Euhrychiopsis lecontei are genetically diverging in lakes where they have access to both the novel host, Eurasian watermilfoil (Myriophyllum spicatum), and their ancestral host, Northern watermilfoil

(M. sibiricum).

Methods and Materials

Euhrychiopsis lecontei were sampled from waterbodies within the Kawartha

Lakes region of the Trent-Severn waterway north of Peterborough in Ontario, Canada. In

2010 and 2011, a total of 275 individuals were collected from three lakes (Table 3.1; Fig.

57

3.1). Individuals were collected from each sampled site from either Northern watermilfoil

(NWM; Myriophyllum sibiricum), Eurasian watermilfoil (EWM; Myriophyllum spicatum), or both, and kept in separately labeled containers identifying the host plant.

Plants were identified in the field by visual characteristics (Madsen et al., 2009). Big

Bald Lake had one site, designated as B1; Pigeon Lake had two sites, designated as P1 and P2; and Stony Lake had 4 sites, designated as S1 – S4. B1 and S4 contained both

NWM and EWM; P1, S1 and S2 contained only EWM; and P2 and S3 contained only

NWM. Distance between sites is reported in Table 3.2. Therefore, all lakes, but not all sites within each lake, contained both the ancestral host (M. sibiricum) and the novel host

(M. spicatum).

Specimens were placed live into ATL lysis buffer and ground with a modified pipette tip directly in a 1.5 mL centrifuge tube and stored at room temperature until DNA extraction. DNA was extracted with a DNeasy® Blood and Tissue kit (Qiagen®) following the protocol for purification of total DNA from insects (DNeasy® Blood and

Tissue Handbook, 2006).

Table 3.1. Location and sample sizes of waterbodies. The centroid of each waterbody is reported for location. Waterbody Latitude Longitude # of Total # of weevils Total # of weevils sites from NWM from EWM Big Bald 44.575955 -78.387719 1 23 10 Lake

Pigeon Lake 44.490298 -78.498269 2 27 12

Stony Lake 44.551984 -78.132287 4 54 149

58

Figure 3.1. Locations of the seven sampling sites within the Kawartha Lakes, Ontario, Canada.

Table 3.2. Distance (km) between sites. Big Bald B1 Pigeon P1 Pigeon P2 Stony S1 Stony S2 Stony S3 Big Bald B1 Pigeon P1 6.5 Pigeon P2 8.3 2.0 Stony S1 18.7 25.3 27.0 Stony S2 18.8 25.3 27.1 1.4 Stony S3 22.5 28.9 30.7 4.2 3.7 Stony S4 21.6 28.2 29.9 2.9 3.4 2.7

Six microsatellite markers (Table 3.3; Vossbrinck et al., 2010) were utilized to investigate whether or not populations of Euhrychiopsis lecontei in the three sampled lakes were genetically differentiated by host plant. Vossbrinck et al. (2010) suggested 59 that none of the markers used herein contained null alleles. PCR amplification was conducted with 1µL DNA extract in a 15 µL reaction mixture using Promega HotStart

Taq. Amplification conditions included an initial denaturation period of 2 min at 94°C , followed by ten -1°C stepdown cycles beginning at 55°C and ending at 45°C, then 30 cycles of 94°C denaturation (30s), 45°C annealing (30s) and 72°C extension (60s) with a final extension of 10 minutes. Fluorescently-tagged primer products were optimized and sized using a LIZ500 size standard at Annis Water Resources Institute (Muskegon,

Michigan). GeneMapper v.3.7 and Peak Scanner v. 1.0 (Applied Biosystems, ©2003 and

©2006, respectively) were used to score fragment size by eye for each marker independently.

Table 3.3. Microsatellite markers used in this study. Locus name, GenBank accession numbers, and primer sequences obtained from Vossbrinck et al. (2010). Locus GenBank Primer Sequences Fragment # of size Name Accession Size range variants Number found GA2 HM237046 F: CCCTGGGATTCGTTTACCTT 257 – 275 7 R: CCACATAGATCACACCATATACCTTT GB7 HM237047 F: CGTCAGAATGAAGTGCTGGA 171 – 183 4 R:TATAAGCGGGCTCAGGTGAT GC2 HM237048 F: TTTTTGCTTGTCCTGTTTTTG 215 – 245 7 R: AAGCAAGGCTTTCCAGATCA GC5 HM237049 F: ATTGACTTGTGTAAATACGATTGTG 167 – 195 8 R: ATGATGATTATCCCGGACCA GC6 HM237050 F: GCACATATTTGTTGAAACTTGCAT 192 – 208 6 R: GGGACTGTATCAACGACCTACA GC7 HM237051 F: TCGCTCCTTTACAAAATCTTCA 260 – 275 6 R:GCCGTCCTTGTCTCACCTAT

60

Following scoring, GDA v. 1.0 (Genetic Data Analysis; Lewis and Zaykin, 2001) was used to evaluate hierarchical relationships of host plant and lake. FSTAT v. 2.9.3.2

(Goudet, 2001) was used to calculate Fst and the associated p-values for various scenarios.

Results

All individuals (n=275) from the six populations (Big Bald Eurasian watermilfoil, n = 10; Big Bald Northern watermilfoil, n=23; Pigeon Eurasian watermilfoil, n=12; Pigeon Northern watermilfoil, n= 27; Stony Eurasian watermilfoil, n=149; Stony Northern watermilfoil, n=54) were analyzed in a Weir and Cockerham

(1984) hierarchical Fst analysis in GDA v. 1.0. Summary results from this analysis including Θs (i.e., Fst for host plants) and Θp (i.e., Fst for lakes), the inbreeding coefficient across the total population (F), the inbreeding coefficient among subpopulations (f), and the 95% confidence interval bounds are provided in Table 3.4. Using the pooled data of each host plant from each lake, this analysis determined that there was significant differentiation among host plants (Θs = 0.025) and lakes (Θp = 0.019), suggesting that there was indeed genetic differentiation between milfoil weevils utilizing the different host plants.

Pairwise Fst and relative migration rates per generation (Nm) between these six populations were also calculated (Table 3.5). All of the pairwise Fst’s were low (i.e., close to 0) and relative migration rates were high (i.e., > 1), suggesting that these six populations are relatively homogenous. Additionally, the highest migration rates (i.e., ∞) do not appear to be solely related to geographic distance between sites (see Table 3.2;

61 e.g., Stony NWM sites and Big Bald EWM and NWM sites still had a relative migration rate per generation of ∞.)

Table 3.4. Hierarchical Analysis from six microsatellite markers for six populations within the Kawartha Lakes region.

Coefficient Value 95% CI 95% CI lower lower bound bound Θlake 0.019 0.003 0.035

Θhost 0.025 0.011 0.044

F 0.130 0.033 0.234

F 0.108 0.016 0.209

Table 3.5. Pairwise Fst (below the diagonal) and migration rates per generation (Nm; above the diagonal) for the six populations within the Kawartha Lakes. ∞ = infinite migrants per generation between sites. BigBald- BigBald- Pigeon- Pigeon- Stony- Stony- EWM NWM EWM NWM EWM NWM BigBald-EWM ∞ 25 10 17 ∞ BigBald-NWM -0.0074 ∞ ∞ 12 ∞ Pigeon-EWM 0.0099 -0.0136 ∞ 6 19 Pigeon-NWM 0.0244 -0.0079 -0.0265 6 15 Stony-EWM 0.0145 0.0207 0.0401 0.0417 18 Stony-NWM -0.0113 -0.0002 0.0133 0.0167 0.0138

Because the initial hierarchical analysis (Table 3.4) was pooled from multiple sites within each lake (especially from Stony Lake, which had the highest total sample size), and because not all sites had both host plants, further analyses were conducted to determine whether site-to-site variation might be confounded with the estimates of Θs in the above comparison. Three additional Fst analyses were conducted for Stony Lake

62

(Table 3.6), not only because it has the highest sample size of any lake studied but also because it has sites that have only Eurasian watermilfoil (S1 and S2); a site with only

Northern watermilfoil (S3); and a site with both host plants at the same location (S4). The first analysis (Analysis A) assessed the overall Fst for the three populations within Stony

Lake that contained Eurasian watermilfoil (S1, n = 54; S2, n = 61; S4, n = 34 – Eurasian watermilfoil individuals only); the second (Analysis B) assessed overall Fst for the two populations within Stony Lake that contained Northern watermilfoil (S3, n = 40; S4, n =

14 – Northern watermilfoil individuals only). The third analysis (Analysis C) pooled all

Stony Northern watermilfoil individuals (n = 54; from sites S3 and S4) and all Stony

Eurasian watermilfoil individuals (n = 149; from S1, S2, and S3) to determine the Fst between different hosts.

Table 3.6. Summary of results from three separate Θ (Fst) analyses for Stony Lake. EWM = Eurasian watermilfoil; NWM = Northern watermilfoil. Analysis Θ 95% F 95% f 95% Confidence Confidence Confidence Interval bounds for Interval Interval f bounds for bounds for F Θ Stony 0.006 0.001 – 0.047 -0.006 – 0.107 0.041 -0.010 – 0.099 EWM sites 0.012 only (A) Stony 0.055 0.007 – 0.205 0.019 – 0.403 0.159 0.000 – 0.331 NWM sites 0.113 only (B) Stony 0.014 0.005 – 0.095 0.012 – 0.192 0.082 0.001 – 0.175 EWM vs. 0.024 NWM (C)

The genetic differentiation for Northern watermilfoil sites (Analysis B; S3 and

S4) was as high as the genetic differentiation for either the Eurasian watermilfoil sites

63

(Analysis A; S1, S2, and S4) or the genetic differentiation from the analysis that pooled the Eurasian watermilfoil individuals from Stony Lake and compared them to Northern watermilfoil individuals (Analysis C; all results reported in Table 3.6). Thus, the within- milfoil comparisons (Analyses A & B) indicate that there are substantial site-to-site genetic differences, at least among the weevils within Stony Lake (Table 3.6), which may have confounded the genetic differentiation between host plants with site-to-site variation within lakes reported in Table 3.4.

Because of the possibility of confounding site-to-site differentiation with between host plant differentiation in the above analyses, a final set of analyses was conducted for only the sampled sites that contained both Eurasian watermilfoil and Northern watermilfoil at the same geographic location (Table 3.7). These sites are Big Bald 1 (B1) and Stony 4 (S4). B1 contained 10 individuals from Eurasian watermilfoil and 23 individuals from Northern watermilfoil; S4 contained 34 individuals from Eurasian watermilfoil and 14 individuals from Northern watermilfoil. A hierarchical Fst analysis was conducted on these data in GDA (Table 3.7, Analysis A) to examine the relative influence of hosts and lakes on genetic differentiation for these two sites. Because of the reduced sample size of this subset analysis, the hierarchical analysis was not powerful enough to find either Fst to be significantly different than 0 (Table 3.7). To increase the power of the Fst test, equal samples of weevils were taken from each site within lakes and the Lake factor was dropped from the analysis, to concentrate on the main effect of host plant differentiation (Table 3.7, Analysis B). With this reduced model, the genetic differentiation between host plants was significant (Table 3.7). Because of the above

64 noted potential for confounding site-to-site differentiation with host plant differentiation, the comparisons in Table 3.7 are our best estimates of the true genetic differentiation between hosts plants for these weevils.

Table 3.7. Summary of results from two separate Fst analyses for Big Bald Lake Site 1 and Stony Lake Site 4. EWM = Eurasian watermilfoil; NWM = Northern watermilfoil.

Analysis Θlake 95% CI Θhost 95% CI F 95% CI f 95% bounds bounds bounds CI for ΘP for ΘS for F bounds for f Hierarchical - -0.035 – 0.011 -0.002 – 0.163 0.037 – 0.154 0.015 – Analysis of B1 0.012 0.010 0.026 0.284 0.284 and S4 (A) Pooled analysis N/A N/A 0.019 0.001 – 0.208 0.081 – 0.193 0.078 – of NWM and 0.040 0.355 0.327 EWM individuals from B1 and S4 (same sample size) (B)

The inbreeding coefficients among subpopulations (f) were also calculated for all of the analyses. Results indicate that f ranged from 0.041 – 0.193 (Tables 3.4, 3.6, and

3.7). The inbreeding coefficient was highest (0.193) in the analysis of the pooled, reduced data set from individuals on Eurasian watermilfoil and Northern watermilfoil sampled from sites B1 and S4 (Table 3.7, Analysis B).

65 Discussion

Bush (1975) suggests that pre-mating reproductive isolating mechanisms are likely to occur in phytophagous insects, such as Euhrychiopsis lecontei that are using more than one food source, because of the use of hosts by specialists for all or a portion of their life cycle, specifically, courtship and mating. Weevil researchers have documented that adult weevils prefer to feed, mate, and oviposit on the plant species in which they completed development (Solarz and Newman, 2001; Tamayo and Grue,

2004) potentially setting up the pre-mating isolation mechanisms necessary for a host shift (Bush, 1975) for populations developing on the two species of Myriophyllum.

Early in his research, Bush (1969) proposed that populations can become rapidly adapted to a new host when the ancestral host population declines in abundance. Importantly, the ancestral host (Myriophyllum sibiricum) of Euhrychiopsis lecontei is decreasing in abundance relative to the increased abundance of the potential new food source (M. spicatum; Boylen et al., 1999; Madsen et al., 1991) because M. sibiricum does not display the same invasive tendencies of M. spicatum (Aiken et al., 1979; Madsen and

Boylen, 1988). At the same time, the novel host, M. spicatum, is increasing its range and is now found in almost all of the contiguous United States as well as in Canada (Creed,

1998). If females can locate, recognize, and utilize the new plant for oviposition (or even if this happens by accident) and larva can find, recognize, utilize, and survive to reproduce on the new host plant, a host expansion and the possibility of a new host race formation with the weevils is not far-fetched (Bowers et al., 1992). Once pre-mating isolating mechanisms are in place, a reduction in gene flow between populations using

66 various hosts in the same geographic region can result in sympatric speciation (Bush,

1975; Via, 2001).

Competition for new oviposition sites for milfoil weevils may be a driving factor in the early stages of their host range expansion because the milfoil weevils prefer to oviposit on the apical meristems of the plant (Sheldon and O’Bryan, 1996), and these specific portions of the plant can therefore represent a limited resource. This pattern was seen as a primary step in speciation of coral gobies (Munday et al., 2004) because competition for coral space is especially strong in gobies. Thus any gobies capable of exploiting a new, unoccupied piece of coral could likely have higher fitness than those competing for space, even if the new habitat is not ideal. Teixeira and Zucoloto (2012) determined that bean weevils’ can exhibit behavioral oviposition plasticity when the amount of available resources changes. Similarly for E. lecontei, if oviposition sites on the ancestral host are declining simply because Myriophyllum sibiricum is declining throughout its range (J. Bissell, CMNH, pers. comm.), E. lecontei that are able to oviposit and have larvae complete development on the novel host, M. spicatum, will have higher fitness than those individuals that are competing for space on the limited ancestral host.

Myriophyllum spicatum also has the tendency to begin growing earlier in the season

(Madsen et al., 1991), thus being readily available when Euhrychiopsis lecontei returns to the water after overwintering (Newman et al., 2001; Reeves et al., 2009). This, coupled with similarities in M. spicatum and M. sibiricum’s morphologies (Madsen et al., 2009), makes it plausible that oviposition on the novel host initially happens by accident.

67

Conversely, the weevils may possess a genetic propensity, through an exaptation (Janz and Nylin, 2007), to include the novel host in its oviposition repertoire.

Because of the of the specific biology of these weevils, combined with the host dynamics of the invasive species they consume, I sought to investigate whether evidence of genetic differentiation in this weevil species could be detected through the use of microsatellite-based genetic data. Studies of host shifts in phytophagous insects are of great interest to evolutionary biologists because these types of studies can detect signals of sympatric speciation due to ecological niche partitioning (Drès and Mallet, 2002;

Rundle and Nosil, 2005; Fitzpatrick et al., 2008; Matsubayashi et al., 2010).

The results of this study do not strongly support the hypothesis that Euhrychiopsis lecontei have diverged genetically into cryptic sister species within the Kawartha Lakes that are utilizing different host plants. However, the results of the first comparison of microsatellite data for the six populations (Big Bald NWM, Big Bald EWM, Pigeon

NWM, Pigeon EWM, Stony NWM, Stony EWM) demonstrate that there is a weak, but significant, signal from host plants in the hierarchical analysis (Table 3.4), suggesting that these weevils may be in the earliest stages of reproductive isolation based on host fidelity.

However, it is important to note that this initial analysis (Table 3.4) may have overestimated the genetic differentiation due to host plant association. Because of the potential of confounding site-to-site variation revealed in the second analysis (Table 3.6), the best estimate of genetic differentiation due to host plant association is represented by the analyses conducted on the two sites that contained both host plants in the same

68 geographic location (BB1 and S4). The analysis for this reduced data set conducted for these two sites (Table 3.7; Analysis B) indicate that there is a significant signal of genetic differentiation due to host plant association. Although the signal is weak (i.e., Θs is close to 0), it nonetheless suggests some genetic differentiation between weevils on differing host plants, and thus warrants some explanation and comparison to similar studies.

Wright (1982) hypothesized that speciation of closely related taxa could occur due to the presence of one or more vacant ecological niches. Similar to the results found in this study, Hendry et al. (2000) showed Fst values (i.e., Fst = 0.025) between two populations of wild salmon that inhabited different, but not geographically distinct, habitats within the same portion of a river system (i.e., mid-river residents vs. shoreline residents) and determined that this level of genetic differentiation suggested that there was evidence of early reproductive isolation, and therefore the potential for sympatric speciation, for these fish. Although this study focused on habitat rather than host plant, the authors determined that the niche differentiation in these fishes was substantial enough to cause pre-mating isolation (Hendry et al., 2000), which is one important way that sympatric speciation through host shift has been shown to occur (Bush, 1975). Upon its arrival and subsequent establishment in a lake system, Eurasian watermilfoil represents a type of vacant ecological niche for the milfoil weevils to utilize. The Fst value (0.019) obtained in my study for the two sites with both host plants at the same geographic location (Table 3.7, Analysis B) is slightly less than the Hendry et al. (2000) study but nonetheless indicates that some pre-mating isolation may have occurred, possibly due to host plant specialization.

69

Low Fst and high Nm (i.e., Table 3.4) are generally interpreted to indicate that there are little to no barriers to gene flow between populations, and thus that the compared groups represent a panmictic population with low overall genetic differentiation and high relative migration rates. However, Drès and Mallet (2002) stress the importance of recognizing that Fst and the associated Nm values do not actually provide a true value for m, the actual gene flow between populations. Recently diverged populations may, in fact, show Fst’s far below 1 and Nm’s far above 0, while still exhibiting little to no actual gene flow between populations (Drès and Mallet, 2002).

Other studies regarding genetic differentiation in insects show similar ranges for Fst for host races. Importantly, McPheron et al. (1988) concluded host-associated genetic differentiation in apple maggot flies feeding on apples or hawthorns with a hierarchically- determined Fst of 0.012. Emelianov et al. (1995) showed Fst values of 0.065 for larch budmoths between their two different hosts (larch and pine). These reported values are similar to the Fst value found for milfoil weevils on different hosts at the same geographic location (Table 3.7; Analysis B), supporting the conclusion that milfoil weevils are in the early stages of host race formation.

The explanation for the slight signal associated with host plants found in this study may be explained by the Myriophyllum spp.’s, as well as Euhrychiopsis lecontei’s, life histories. Studies conducted by Bush (1969) and Feder et al. (1988), support the assertion that apple maggot flies (Rhagoletis pomonella) have diverged into two separate species, feeding on apple and hawthorn, specifically because of temporal isolation due to the difference in timing of the ripening period of the fruits. The temporal separation of

70 the ripening of apple and hawthorn is approximately 3-4 weeks (Feder et al., 2003).

Therefore, R. pomonella that develop in apple fruits have different selection pressures than those that develop in hawthorn fruits, effectively setting the stage for reproductive isolation leading to sympatric speciation (Feder and Filchak, 1999). A similar separation, although much shorter in scale, occurs in timing between the growth of M. spicatum

(Eurasian watermilfoil) and M. sibiricum (Northern watermilfoil) populations.

Populations of Eurasian watermilfoil tend to grow rapidly in the spring, forming a dense canopy that excludes many native macrophytes (Madsen et al., 1991), including Northern watermilfoil. Eurasian watermilfoil’s explosive growth rate in the spring occurs at water temperatures of about 15C (Smith and Barko, 1990). This temperature corresponds to when milfoil weevils are returning to the water to breed (Newman et al., 2001). If the first milfoil weevils to enter the water in the spring only have the ability to locate

Eurasian watermilfoil, host-associated genetic differentiation may be expected.

Populations of Euhrychiopsis lecontei can exhibit an increase in the rate of reproduction when feeding occurs on Eurasian watermilfoil as opposed to on Northern watermilfoil (Lillie and Helsel, 1997). Also, although weevils reared on Eurasian watermilfoil preferred to oviposit on Eurasian watermilfoil rather than on their native host of Northern watermilfoil, the opposite was not the case, as weevils reared on

Northern watermilfoil showed no oviposition preference (Solarz and Newman, 1996).

This suggests that any gene flow between the individuals feeding on the different hosts may be unidirectional, since weevils reared on Eurasian watermilfoil do not tend to return to Northern watermilfoil for feeding, mating and development. Therefore, the low level

71 of genetic differentiation detected due to host plant association may again be a product of the temporal separation of the weevils utilizing the different hosts, which could establish pre-mating isolating mechanisms.

Although the main purpose of this study was to determine if there was a genetic signal indicating differentiation between individuals using different host plants, the analyses conducted shed light on an important factor in Euhrychiopsis lecontei’s life history within these lakes: the inbreeding coefficients from the first analysis of the six populations from the three study lakes (Table 3.4), and the third analysis of the two sites with both host species present (Table 3.7) showed high levels of inbreeding (>10%), likely brother-sister matings, for this species within the sampled lakes. Once E. lecontei returns to the water after overwintering, they tend to swim to the nearest milfoil plant for feeding and breeding activities (Newman et al., 2001). They are notably poor swimmers

(Reeves and Lorch, 2009) and therefore do not tend to move between distantly-located plants on their own accord. Females can lay multiple eggs on one meristem (Sheldon and

O’Brien, 1996), and since development most often takes place on the same plant where the eggs have been laid (Sheldon and O’Brien, 1996), the chances of mating with a sibling are increased. These high levels of inbreeding among subpopulations suggest that milfoil weevils may indeed develop the necessary pre-mating isolating mechanisms required for host race formation and subsequent sympatric speciation since they are more likely to continue mating with other individuals in the same subpopulation (i.e., plant patch) over several generations per summer.

72

Several aspects of the milfoil weevil and host plant ecologies could have skewed the results and overestimated the amount of genetic differentiation from host plant, and therefore should be investigated more deeply in future studies. Since 3-6 generations of

E. lecontei are typically produced each year (Newman et al., 2001), the timing in the emergence of the various host plants may not be disparate enough to cause pre-mating isolating mechanisms that last throughout the growing season and that would ultimately reduce the overall gene flow between the weevils feeding on the various species of milfoil. Additionally, Newman et al. (2001) reported that approximately 75% of female weevils collected from the soil in the fall and spring have already mated. Therefore, the small level of host-associated genetic differentiation observed could be a remnant of any small amount of population structure from the preceding year, if the sampled weevils were collected primarily from the first wave of adults entering the water in the spring.

Finally, Eurasian watermilfoil has the ability to rapidly spread throughout connected waterbodies through auto-fragmentation (Smith and Barko, 1990) as well as through mechanical fragmentation through human-mediated activities such as boating (Johnstone et al., 1985). Therefore, milfoil weevils that are attached to this plant may easily be transported throughout the Kawartha Lakes on floating fragments of Eurasian milfoil, which would contribute to the high rates of migration observed between hosts and lakes.

Since milfoil weevils, on their own, are typically quite poor dispersers (Newman et al.,

2001; Reeves and Lorch, 2009), this characteristic of the host plant to disperse widely may have contributed to the lower Fst and higher Nm values that were observed in this study. Importantly, based on the results of this study, it seems warranted to collect large

73 sample sizes of adult weevils throughout the growing season to determine if periodic temporal isolation, rather than lack of gene flow, is contributing to the small amount of genetic differentiation via host that was observed in this study.

Implications for Management of Eurasian watermilfoil

The natural populations of Euhrychiopsis lecontei in the Kawartha Lakes are characterized by high levels of inbreeding, which could affect these populations’ ability to effectively control nuisance populations of Myriophyllum spicatum if inbreeding depression limits the effective population size or the long-term sustainability of the weevils in the wild (Allendorf and Lundquist, 2003). Milfoil weevils reared in the lab for augmentative biological control purposes are generally collected from a variety of lakes within a wide geographic range, and not typically from the lakes where they will eventually be deployed. This suggests that introducing lab-reared milfoil weevils into lakes, such as the Kawartha Lakes, may help to decrease high levels of inbreeding by diversifying the baseline genetic stock that will then be present in the lake system. This artificial (i.e., laboratory facilitated) increase in genetic diversity may therefore raise the effectiveness of the milfoil weevil as a biological control agent in the management of

Eurasian watermilfoil (Hopper et al., 1993).

Furthermore, if Euhrychiopsis lecontei has genetic variation that can lead to adaptability on closely related species of milfoil, there may be reason to use the milfoil weevil as a biocontrol agent for other noxious species of milfoil (western watermilfoil –

Myriophyllum hippuroides and variable-leaved watermilfoil – M. heterophyllum). As

74

G.L. Bush emphasized in his 1975 paper on apple maggot flies, the need for “applied evolutionary biology” in understanding speciation mechanisms is essential in order to develop “realistic pest control programs”.

Conclusions and Future Studies

Although the results of this study do not support the hypothesis that

Euhrychiopsis lecontei feeding on different species of Myriophyllum within the Kawartha

Lakes are genetically differentiated enough to warrant the designation of cryptic sister species in the process of sympatric speciation, the data do suggest that there is early genetic differentiation via host plant association. These data are comparable to the levels of host-associated genetic differentiation in other studies that have been inferred as indicative of the early stages of host race formation (McPheron et al.,1988; Emelianov et al., 1995; Hendry et al., 2000). The few sites that contained both milfoil species in the same area combined with site-to-site genetic differentiation within lakes weakened the power to detect genetic differentiation in this study by reducing the total sites that could be compared. Due to the nature of the host plant ecologies, it was difficult to find patches of Eurasian watermilfoil and Northern watermilfoil in the same geographic location.

Therefore, future studies should focus on finding and collecting weevils from both

Eurasian and Northern watermilfoil at the same geographic location over a broader geographic range. Although this scenario is becoming increasingly difficult because of the increase of Eurasian watermilfoil and the decline of Northern watermilfoil, an increase in the number of sites where both plants are sympatric and an increase in the

75 number of weevils collected from the various milfoil species would likely give a clearer picture as to whether or not there is stronger evidence of host-associated genetic divergence in this weevil species.

76

LITERATURE CITED

Adams, M.S., and F.X. Villablanca. 2007. Consequences of a genetic bottleneck in California condors: a mitochondrial DNA perspective. Pages 35-55 in: Mee, A., and L. S. Hall (editors), California condors in the 21st century. Series in Ornithology No. 2. Cambridge, Massachusetts, and Washington, D. C.: Nuttall Ornithological Club & American Ornithologists' Union.

Aiken, S. G., P. R. Newroth, and I. Wile. 1979. The biology of Canadian weeds: 34. Myriophyllum spicatum L. Canadian Journal of Plant Science 59(1): 201-215.

Allcock, A.L., and J.M. Strugnell. 2012). Southern Ocean diversity: new paradigms from molecular ecology. Trends in Ecology and Evolution, 27(9), 520-528.

Allendorf, F. W., P.A. Hohenlohe, and G. Luikart. 2010. Genomics and the future of conservation genetics. Nature Reviews Genetics, 11(10), 697-709.

Alwin, Thomas G., M.G. Fox, and K. Spence Cheruvelil. 2010. Estimating Lake-wide Watermilfoil Weevil (Euhrychiopsis lecontei) Density: The Roles of Quadrat Size, Sample Size, and Effort. Journal of Aquatic Plant Management 48: 96–102.

Allendorf, F.W. and L.L. Lundquist. 2003. Introduction: population biology, evolution, and control of invasive species. Conservation Biology, 17(1), 24-30.

Anderson R.S. 2002. Family 131 Curculionidae. In American Beetles: Polyphaga: Scarabaeoidea through Curculionoidea (Vol. 2). Arnett, R. H., Thomas, M. C., Skelley, P.E., and J.H. Frank (Eds.). CRC Press. pp 722 – 815.

Aoki, K., M. Kato and N. Murakami. 2008. Glacial bottleneck and postglacial recolonization of a seed parasitic weevil, Curculio hilgendorfi, inferred from mitochondrial DNA variation. Molecular Ecology, 17(14), 3276-3289.

Aoki, K., M. Kato, and N. Murakami. 2009. Phylogeographical patterns of a generalist acorn weevil: insight into the biogeographical history of broadleaved deciduous and evergreen forests. BMC Evolutionary Biology, 9(1), 103.

Ascunce, M.S., H.N. Nigg, and A. Clark. 2009. Molecular identification of the economically important invasive citrus root weevil Diaprepes abbreviatus (Coleoptera: Curculionidae). Florida Entomologist, 92(1), 167-171 77

Avise, J.C. 2000. Phylogeography: the history and formation of species. Harvard University Press.

Avise, J. C. 2010. Perspective: conservation genetics enters the genomics era. Conservation Genetics, 11(2), 665-669.

Bandelt, H.J., P. Forster, and A. Röhl. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution, 16(1), 37-48.

Barr, N., R. Ruiz-Arce, O. Obregón, R.D. Leon, N. Foster, C. Reuter, T. Boratynski, and D. Vacek. 2013. Molecular diagnosis of populational variants of Anthonomus grandis (Coleoptera: Curculionidae) in North America. Journal of Economic Entomology, 106(1), 437-449.

Beatty, G. E., and J.I.M. Provan. 2010. Refugial persistence and postglacial recolonization of North America by the cold‐tolerant herbaceous plant Orthilia secunda. Molecular Ecology, 19(22), 5009-5021.

Berlocher, S. H., and J.L. Feder, J. L. 2002. Sympatric speciation in phytophagous insects: moving beyond controversy? Annual Review of Entomology, 47(1), 773- 815.

Bickford, D., D.J. Lohman, N.S. Sodhi, P.K. Ng, R. Meier, K. Winker, K., K.K. Ingram and I. Das. 2006. Cryptic species as a window on diversity and conservation. Trends in Ecology & Evolution, 22(3), 148-155.

Blatchley, W. S., and C.W. Leng. 1916. Rhynchophora or weevils of north eastern America. The Nature Publishing Company.

Bolnick, D.I., and B.M. Fitzpatrick. 2007. Sympatric speciation: models and empirical evidence. Annual Review of Ecology, Evolution, and Systematics. 459-487.

Bouchard, P., V.V. Grebennikov, A.B. Smith, and H. Douglas. 2009. Biodiversity of Coleoptera. Insect Biodiversity: Science and Society, 265-301.

Bowers, M.D., N.E. Stamp and S.K. Collinge. 1992. Early stage of host range expansion by a specialist herbivore, Euphydryas phaeton (Nymphalidae). Ecology 73: 526- 536.

Boylen, C.W., L.W. Eichler, and J.D. Madsen. 1999. Loss of native aquatic plant species in a community dominated by Eurasian watermilfoil. In Biology, Ecology and Management of Aquatic Plants (pp. 207-211). Springer Netherlands.

78

Brower, A.V. 1994. Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the National Academy of Sciences, 91(14): 6491-6495.

Buckingham, G.R. and C.A. Bennett. 1981. Laboratory biology and behavior of Litodacylus leucogaster, a Ceutorhychine weevil that feeds on watermilfoils. Annals of the Entomological Society of America 74: 451-458.

Buckingham, G. and S. Passoa. 1984. Flight muscle and egg development in waterhyacinth weevils. Proceedings VI International Symposium on the Biological Control of Weeds 19: 497-510.

Bush, G.L. 1969. Sympatric host race formation and speciation in frugivorous flies of the genus Rhagoletis (Diptera, Tephritidae). Evolution : 237-251.

Bush, G.L. 1975. Modes of animal speciation. Annual Review of Ecology and Systematics, 6(1): 339-364.

Bush, G.L. 1994. Sympatric speciation in animals: new wine in old bottles. Trends in Ecology & Evolution 9(8): 285-288.

Camara, M.D. 1997. A recent host range expansion in Junonia coenia Hübner (Nymphalidae): oviposition preference, survival, growth, and chemical defense. Evolution: 51: 873-884.

Caterino, M.S., S. Cho and F.A.H. Sperling. 2000. The current state of insect molecular systematic: a thriving Tower of Babel. Annual Review of Entomology 45: 1–54.

Center, T.D., F. A. Dray, Jr., G.P. Jubinsky, and M.J. Grodowitz. 2002. Insects and Other Arthropods That Feed on Aquatic and Wetland Plants. U.S. Department of Agriculture, Agricultural Research Service, Technical Bulletin 1870. 200 pp.

Cline, A. R., B. Villegas, M.J. Pitcairn, and C.W. O'Brien. 2013. The Status of Euhrychiopsis lecontei (Dietz)(Coleoptera: Curculionidae) in California, with Notes on Other Weevils Associated with Milfoil. The Coleopterists Bulletin, 67(2), 75-80.

Coble, T.A. and B.D. Vance. 1987. Seed germination in Myriophyllum spicatum L. Journal of Aquatic Plant Management, 25(1), 8-10.

Cognato, A.I., A.D. Harlin and M.L. Fisher. 2003. Genetic structure among Pinyon Pine Beetle Populations (Scolytinae: Ips confuses). Environmental Entomology 32: 1262–1270.

79

Colonnelli, E. 2004. Catalogue of Ceutorhynchinae of the world, with a key to genera. (Insecta: Coleoptera: Curculionidae). Argania editio.

Couch, R. and E. Nelson. 1985. Myriophyllum spicatum in North America. Pp. 8-18 in L.W.J. Anderson, ed. Proceedings of the First International Symposium on Watermilfoil (Myriophyllum spicatum) and related Haloragaceae species, July 23- 24, 1985, Vancouver, British Columbia. The Aquatic Plant Management Society, Inc., Vicksburg, Mississippi.

Creed, R.P., Jr. 1998. A biogeographic perspective on Eurasian watermilfoil declines: additional evidence for the role of herbivorous weevils in promoting declines? Journal of Aquatic Plant Management 36: 16-22.

Creed, R.P, Jr. 2000. Is there a new keystone species in North American lakes and rivers? Oikos 91: 405–408.

Creed, Jr., R.P. and S.P. Sheldon. 1994. Aquatic weevils (Coleoptera: Curculionidae) associated with northern watermilfoil (Myriophyllum sibiricum) in Alberta, Canada. Entomological News 105: 98-102.

Creed, R.P., Jr. and S.P. Sheldon. 1995. Weevils and watermilfoil: Did a North American herbivore cause the decline of an exotic plant? Ecological Applications 5: 1113-1121.

Creed, Jr., R.P., S.P. Sheldon and D.M. Cheek. 1992. The effect of herbivore feeding on the buoyancy of Eurasian watermilfoil. Journal of Aquatic Plant Management 30: 75-76.

Crozier, R.H. 1997. Preserving the Information Content of Species: Genetic Diversity, Phylogeny, and Conservation Worth. Annual Review of Ecology and Systematics 28: 243–268.DeLong, Jr., D.C. 1996. Defining Biodiversity. Wildlife Society Bulletin 24, 738-749.

Czarapata, E.J. 2005. Invasive plants of the upper Midwest: an illustrated guide to their identification and control. Madison, WI: The University of Wisconsin Press.

Darwin, C.R. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray.

De Jong, M.A., N. Wahlberg, M. Van Eijk, P.M. Brakefield and B.J. Zwaan. 2011. Mitochondrial DNA signature for range-wide populations of Bicyclus anynana suggests a rapid expansion from recent refugia. PLOS ONE, 6(6), e21385, 1–5.

80

Denoth, M., L. Frid, and J.H. Myers. 2002. Multiple agents in biological control: improving the odds? Biological Control 24(1), 20-30.

Drès, M., and J. Mallet. 2002. Host races in plant–feeding insects and their importance in sympatric speciation. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 357(1420): 471-492.

Dudgeon, D., A.H. Arthington, M.O. Gessner, Z.I. Kawabata, D.J. Knowler, C. Lévêque, R.J. Naiman, A-H. Prieur-Richard, D. Soto, L. Melanie, J. Stiassny and C.A. Sullivan. 2005. Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews, 81(2), 163–182.

Dupanloup, I., S. Schneider, and L. Excoffier. 2002. A simulated annealing approach to define the genetic structure of populations. Molecular Ecology, 11(12), 2571- 2581.

Eble, J.A., R.J. Toonen, L. Sorenson, L.V. Basch, Y.P. Papastamatiou, and B.W. Bowen. 2011. Escaping paradise: larval export from Hawaii in an Indo-Pacific reef fish, the yellow tang Zebrasoma flavescens. Marine Ecology Progress Series, 428, 245-258.

Emelianov, I., J. Mallet, and W. Baltensweiler. 1995. Genetic differentiation in Zeiraphera diniana (Lepidoptera: Tortricidae, the larch budmoth): polymorphism, host races, or sibling species. Heredity, 75(4), 16-424.

Engstrand, R.C., J. Cibrián Tovar, A. Cibrián-Jaramillo, and S.O. Kolokotronis. 2010. Genetic variation in avocado stem weevils Copturus aguacatae (Coleoptera: Curculionidae) in Mexico. Mitochondrial DNA, 21(S1), 38-43.

Feder, J L., C.A. Chilcote, G.L. Bush. 1988. Genetic differentiation between sympatric host races of the apple maggot fly. Nature 336(3): 61-64.

Feder, J. L., and K.E. Filchak. 1999. It’s about time: the evidence for host plant-mediated selection in the apple maggot fly, Rhagoletis pomonella, and its implications for fitness trade-offs in phytophagous insects. In Proceedings of the 10th International Symposium on Insect-Plant Relationships (pp. 211-225). Springer Netherlands.

Feder, J.L., J.B. Roethele, K.E. Filchak, J. Niedbalski, and J. Romero-Severson. 2003. Evidence for inversions related to sympatric host race formation in the apple maggot fly, Rhagoletis pomonella (Diptera: Tephritidae). Genetics 163: 939-953.

Fitzpatrick, B. M., J.A. Fordyce, and S. Gavrilets. 2008. What, if anything, is sympatric speciation? Journal of Evolutionary Biology, 21(6): 1452-1459.

81

Foottit, R. G., and P.H. Adler. (Eds.). 2009. Introduction. Insect biodiversity: science and society. John Wiley & Sons. pp 1-6.

Futuyma, D.J. 2000. Potential evolution of host range in herbivorous insects. In: VanDriesche, R.G., T.A. Heard, A. McClay, R. Reardon (Eds.), Proceedings of Session: Host-Specificity Testing of Exotic Arthropod Biological Control Agents—The Biological Basis for Improvement in Safety. USDA Forest Service Forest Health Technology Enterprise Team, Morgantown, pp. 42–53.

Gillett, C. P., A. Crampton-Platt, M.J. Timmermans, B. Jordal, B.C. Emerson, and A.P. Vogler (2014). Bulk de novo mitogenome assembly from pooled total DNA elucidates the phylogeny of weevils (Coleoptera: Curculionoidea). Molecular biology and evolution, msu154.

Goudet, J. 2001. FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3.2). Available from http://www.unil.ch/izea/softwares/fstat.html.

Gugerli F., T. Englisch, H. Niklfeld, A. Tribsch, Z. Mirek, M. Ronikier, N.E. Zimmermann, R. Holderegger, and P. Taberlet. 2008. IntraBioDiv Consortium. Relationships among levels of biodiversity and the relevance of intraspecific diversity in conservation – a project synopsis. Perspectives in Plant Ecology, Evolution and Systematics 10: 259–281.

Hajibabaei, M., G.A. Singer, P.D. Hebert, and D.A. Hickey. 2007. DNA barcoding: how it complements taxonomy, molecular phylogenetics and population genetics. TRENDS in Genetics, 23(4), 167-172.

Hasegawa, M., H. Kishino and T.A. Yano. 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution 22(2): 160-174.

Hebert, P.D., S. Ratnasingham, and J.R. de Waard. 2003. Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270(Suppl 1), S96-S99.

Hendry, A.P., J.K. Wenburg, P. Bentzen, E.C. Volk, and T.P. Quinn. 2000. Rapid evolution of reproductive isolation in the wild: evidence from introduced salmon. Science, 290(5491), 516-518.

Hewitt, G.M. 2004. The structure of biodiversity–insights from molecular phylogeography. Frontiers in Zoology, 1(4), 1-16.

82

Honnell, D.R., J.D. Madsen, and R.M. Smart. 1992. "Effect of Aquatic Plants on Water Quality in Pond Ecosystems," Proceedings, 26th Annual Meeting, Aquatic Plant Control Research Program, Miscellaneous Paper A-92-2, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, pp 30–34.

Hopper, K.R., R.T. Roush, and W. Powell. 1993. Management of genetics of biological- control introductions. Annual Review of Entomology, 38(1), 27-51.

Hughes, A.R., B.D. Inouye, M.T.J. Johnson, N. Underwood and M. Vellend. 2008. Ecological consequences of genetic diversity. Ecology Letters 11: 609–623.

Hundsdoerfer, A. K., J. Rheinheimer, and M. Wink. 2009. Towards the phylogeny of the Curculionoidea (Coleoptera): Reconstructions from mitochondrial and nuclear ribosomal DNA sequences. Zoologischer Anzeiger-A Journal of Comparative Zoology, 248(1), 9-31.

Jacobs, J. and J. Mangold. 2009. Ecology and Management of Eurasian Watermilfoil (Myriophyllum spicatum L.). Invasive Species Technical Note No. MT-23. United States Department of Agriculture Natural Resources Conservation Service.

Jacono C.C. and M.M. Richerson. 2010. Myriophyllum spicatum. USGS Nonindigenous Aquatic Species Database, Gainesville, FL. http://nas.er.usgs.gov/queries/factsheet.aspx?SpeciesID=237

Janz, N. and S. Nylin. 2007. The oscillation hypothesis of host plant-range and speciation. In: Specialization, speciation and radiation: the evolutionary biology of herbivorous insects: 203-215. University of California Press, Berkeley, CA.

Jester, L.L., M.A. Bozek, D.R. Helsel, and S.P. Sheldon. 2000. Euhrychiopsis lecontei distribution, abundance, and experimental augmentations for Eurasian watermilfoil control in Wisconsin lakes. Journal of Aquatic Plant Management 38: 88–97.

Jiggins, C.D. 2006. Sympatric speciation: why the controversy? Current Biology 16(9): R333-R334.

Johnstone, I.M., B.T. Coffey and C. Howard-Williams. 1985. The role of recreational boat traffic in interlake dispersal of macrophytes: a New Zealand case study. Journal of Environmental Management 20: 263–279.

Keast, A. 1984. The introduced aquatic macrophyte, Myriophyllum spicatum, as habitat for fish and their invertebrate prey. Canadian Journal of Zoology 62: 1289–1303.

83

Kimura M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16:111-120.

Knight, J. 2001. Alien versus predator. Nature 412: 115–116.

Korotyaev, B.A. 2008. Geographical distribution of the weevil subfamily Ceutorhynchinae (Coleoptera, Curculionidae). Entomological Review, 88(8): 928–947.

Lanfear, R., B. Calcott, S.Y. Ho, and S. Guindon. 2012. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution 29(6): 1695-1701.

LaRue, E.A., M.P. Zuellig, M.D. Netherland, M.A. Heilman and R.A. Thum. 2013. Hybrid watermilfoil lineages are more invasive and less sensitive to a commonly used herbicide than their exotic parent (Eurasian watermilfoil). Evolutionary applications, 6(3), 462-471.

Lewis, P.O., and D. Zaykin. 2001. Genetic Data Analysis: Computer program for the analysis of allelic data. Version 1.0 (d16c). Free program distributed by the authors over the internet from http://lewis.eeb.uconn.edu/lewishome/software.html

Ley, A.C. and O.J. Hardy. 2014. Contrasting patterns of gene flow between sister plant species in the understory of African moist forests–The case of sympatric and parapatric Marantaceae species. Molecular phylogenetics and evolution, 77, 264- 274.

Librado P. and J. Rozas. 2009. DnaSP ver. 5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics. 25:1451–1452.

Lillie R.A. 2000. Temporal and spatial changes in milfoil distribution and biomass associated with weevils in Fish Lake, WI. Journal of Aquatic Plant Management 38: 98–104.

Lillie, R. and D. Helsel. 1997. A native weevil attacks Eurasian watermilfoil. Findings Bureau of Research WDNR 40.

Losey, J. E. and M. Vaughan. 2006. The economic value of ecological services provided by insects. Bioscience, 56(4), 311-323.

84

Lovell, S.J., S.F. Stone, and L. Fernandez. 2006. The Economic Impacts of Aquatic Invasive Species: A Review of the Literature. Agricultural and Resource Economics Review 35(1): 195–208.

Ludt, W. B., M.A. Bernal, B.W. Bowen, and L.A. Rocha. 2012. Living in the past: phylogeography and population histories of Indo-Pacific wrasses (genus Halichoeres) in shallow lagoons versus outer reef slopes. PloS one, 7(6), e38042.

Mackauer, M. (1976). Genetic problems in the production of biological control agents. Annual Review of Entomology, 21(1), 369-385.

Madsen, J.D. 2005. Eurasian watermilfoil invasions and management across the United States. The Journal of Marine Education 21: 21–26.

Madsen, J.D. and Boylen, C.W. 1988. The physiological ecology of Eurasian watermilfoil (Myriophyllum spicatum) and a native macrophyte (Potamogeton praelongus): depth distribution of biomass and photosynthesis. Rensselear Fresh Water Res. Inst. Rep. 88-5, Rensselear Polytechnic Institute, Troy, NY.

Madsen, J. D. and D.H. Smith. 1997. Vegetative spread of Eurasian watermilfoil colonies. Journal of Aquatic Plant Management, 35, 63-68.

Madsen, J.D., L.W. Eichler, and C.W. Boylen. 1988. Vegetative Spread of Eurasian Watermilfoil in Lake George, New York. Journal of Aquatic Plant Management 35: 63–68.

Madsen, J.D., J.W. Sutherland, J.A. Bloomfield, L.W. Eichler and C.W. Boylen. 1991. The decline of native vegetation under dense Eurasian watermilfoil canopies. Journal of Aquatic Plant Management 29: 94–99.

Madsen, J.D., J. Cheshier, V. Phuntumart, R. Thum, and M. Welch. 2009. Eurasian watermilfoil survey of three reservoirs in the Lower Clark Fork River, Montana: II. Taxonomic analysis of native and nonnative watermilfoils. Geosystems Research Institute GRI Report # 5035.Mallet, J., Meyer A.,, Nosil, P. , and J.L. Feder. 2009. Space, sympatry and speciation. Journal of evolutionary biology, 22(11): 2332-2341.

Mallet, J., A. Meyer, P. Nosil and J.L. Feder. 2009. Space, sympatry and speciation. Journal of Evolutionary Biology 22: 2332-2341.

Mallet, J. 2010. Group selection and the development of the biological species concept. Philosophical Transactions of the Royal Society 365: 1853-1863.

85

Manel S., M.K. Schwartz, G. Luikart and P. Taberlet. 2003. Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology and Evolution 18: 189–197.

Martin, C.W. and J.F. Valentine. 2014. Sexual and asexual reproductive strategies of invasive Eurasian milfoil (Myriophyllum spicatum) in estuarine environments. Hydrobiologia, 727(1), 177-184.

Marvaldi, A.E., A.S. Sequeira, C.W. O'Brien, and B.D. Farrell. (2002). Molecular and morphological phylogenetics of weevils (Coleoptera, Curculionoidea): do niche shifts accompany diversification? Systematic biology, 51(5), 761-785.

Matsubayashi, K.W., I. Ohshima, and P. Nosil. 2010. Ecological speciation in phytophagous insects. Entomologia Experimentalis et Applicata, 134(1), 1-27.

McKenna, D.D., A.S. Sequeira, A.E. Marvaldi and B.D. Farrell. 2009. Temporal lags and overlap in the diversification of weevils and flowering plants. Proceedings of the National Academy of Sciences, 106(17), 7083-7088.

McPheron, B. A., D.C. Smith, and S.H. Berlocher. 1988. Genetic differences between host races of Rhagoletis pomonella. Nature, 336(6194), 64-66.

Miller, M.P. 2005. Alleles In Space (AIS): computer software for the joint analysis of inter-individual spatial and genetic information. Journal of Heredity, 96(6), 722- 724.

Miller, M. A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. InGateway Computing Environments Workshop (GCE), pp. 1-8. IEEE.

Moody, M.L. and D.H. Les. 2002. Evidence of hybridity in invasive watermilfoil (Myriophyllum) populations. Proceedings of the National Academy of Sciences, 99(23), 14867-14871.

Moody, M.L. and D.H. Les. 2007. Geographic distribution and genotypic composition of invasive hybrid watermilfoil (Myriophyllum spicatum × M. sibiricum) populations in North America. Biological Invasions, 9(5), 559-570.

Munday, P.L., L. van Herwerden and C.L. Dudgeon. 2004. Evidence for sympatric speciation by host shift in the sea. Current Biology 14: 1498-1504.

Newman, R.M. 2004. Invited review: Biological control of Eurasian watermilfoil by aquatic insects: basic insights from an applied problem. Archiv fuer hydrobiologie, 159(2): 145-184.

86

Newman, R.M. and D.D. Biesboer. 2000. A decline of Eurasian watermilfoil in Minnesota associated with the milfoil weevil, Euhrychiopsis lecontei. Journal of Aquatic Plant Management. 38: 105–111.

Newman, R.M., E.M. Gross, W. Wimmer and P. Sprick. 2006. Life history and developmental performance of the Eurasian milfoil weevil, Eubrychius velutus (Coleoptera: Curculionidae). The Coleopterists Bulletin 60: 170–176.

Newman, R.M., and W.G. Inglis. 2009. Distribution and Abundance of the Milfoil Weevil, Euhrychiopsis lecontei, in Lake Minnetonka and Relation to Milfoil Harvesting. Journal of Aquatic Plant Management 47: 21-25.

Newman, R.M., D.W. Ragsdale, A. Milles, and C. Oien. 2001. Overwinter habitat and the relationship of overwinter to in-lake densities of the milfoil weevil (Euhrychiopsis lecontei), a Eurasian watermilfoil biological control agent. Journal of Aquatic Plant Mangement 39: 63–67.

Nosil, P. 2008. Speciation with gene flow could be common. Molecular Ecology17(9): 2103-2106.

Oberprieler, R.G., A.E. Marvaldi, and R.S Anderson. 2007. Weevils, weevils, weevils everywhere. Zootaxa, 1668: 491-520.

Papadopoulou, A., I. Anastasiou, and A.P. Vogler. 2010. Revisiting the insect mitochondrial molecular clock: the mid-Aegean trench calibration. Molecular Biology and Evolution, 27(7), 1659-1672.

Parsons, J.K., G.E. Marx, and M. Divens. 2011. A study of Eurasian watermilfoil, macroinvertebrates and fish in a Washington lake. Journal of Aquatic Plant Management 49: 71-82.

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

Pimm, S. L., C.N. Jenkins, R. Abell, T.M. Brooks, J.L. Gittleman, L.N. Joppa, P.H. Raven, C.M. Roberts, and J.O. Sexton. 2014. The biodiversity of species and their rates of extinction, distribution, and protection. Science, 344(6187), 1246752.

Purvis, A., and A. Hector. 2000. Getting the measure of biodiversity. Nature, 405(6783), 212-219. Ratnasingham, S., and P.D. Hebert. 2007. BOLD: The Barcode of Life Data System (http://www. barcodinglife. org). Molecular ecology notes, 7(3), 355-364.

87

Reed, C. F. 1977. History and distribution of Eurasian watermilfoil in the United States and Canada. Phytologia 36: 417-436.

Reeves, J. L. and P.D. Lorch. 2009. Visual plant differentiation by the milfoil weevil, Euhrychiopsis lecontei Dietz (Coleoptera: Curculionidae). Journal of Insect Behavior, 22(6), 473-476.

Reeves, J.L., P.D. Lorch, and M.W. Kershner. 2009. Vision is important for plant location by the phytophagous aquatic specialist Euhrychiopsis lecontei Dietz (Coleoptera: Curculionidae). Journal of Insect Behavior, 22(1), 54-64.

Richardson, R.J. 2008.Aquatic Plant Management and the Impact of Emerging Herbicide Resistance Issues. Weed Technology 22: 8–15.

Richmond, G.M., and D.S. Fullerton. 1986. Summation of Quaternary glaciations in the United States of America. Quaternary Science Reviews, 5, 183-196.

Ronquist, F., M. Teslenko, P. van der Mark, D.L. Ayres, A. Darling, S. Höhna, B. Larget, L. Liu, M.A. Suchard, and J.P. Huelsenbeck. 2012. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61(3): 539-542.

Rundle, H.D., and P. Nosil. 2005. Ecological speciation. Ecology letters, 8(3), 336-352.

Sabrosky, C.W. 1953. How many insects are there? Systematic Zoology, 2(1), 31-36.

Scheffer, S.J., and D.J. Hawthorne. 2007. Molecular evidence of host‐associated genetic divergence in the holly leafminer Phytomyza glabricola (Diptera: Agromyzidae): apparent discordance among marker systems. Molecular ecology, 16(13), 2627- 2637.

Selkoe, K.A., and R.J. Toonen. 2006. Microsatellites for ecologists: a practical guide to using and evaluating microsatellite markers. Ecology letters, 9(5), 615-629.

Sheldon, S.P. 1997. Investigations on the potential use of an aquatic weevil to control Eurasian watermilfoil. Journal of Lake and Reservoir Management 13: 79–88.

Sheldon S.P. and R.P. Creed, Jr. 1995. Use of a native insect as a biological control for an introduced weed. Ecological Applications 5: 1122-1132.

Sheldon S.P. and R.P. Creed, Jr. 2003. The effect of a native biological control agent for Eurasian watermilfoil on six North American watermilfoils. Aquatic Botany 76: 259-265.

88

Sheldon S.P. and L.M. O'Bryan. 1996. Life history of the weevil Euhrychiopsis lecontei, a potential biological control agent of Eurasian watermilfoil. Entomological News 107: 16–22.

Smith C.S. and J.W. Barko. 1990. Ecology of Eurasian watermilfoil. Journal of Aquatic Plant Management 28: 55-64.

Solarz, S. L. 1998. Genetic and environmental effects on preference and performance traits of the milfoil weevil Euhrychiopsis lecontei Dietz, Ph.D. Dissertation, University of Minnesota, St. Paul, MN.

Solarz, S. L., and R.M. Newman. 1996. Oviposition specificity and behavior of the watermilfoil specialist Euhrychiopsis lecontei. Oecologia 106: 337-344.

Solarz, S.L. and R.M. Newman. 2001. Variation in hostplant preference and performance by the milfoil weevil, Euhrychiopsis lecontei Dietz, exposed to native and exotic watermilfoils. Oecologia 126: 66-75.

Soulé, M.E., and B.A. Wilcox. 1980. Conservation Biology: An Evolutionary-Ecological Approach. Sinauer Associates, Sunderland, Massachusetts.

Spencer, N.R. and M. Lekić. 1974. Prospects for Biological Control of Eurasian Watermilfoil. Weed Science 22: 401–404.

Stork, N. E. 1993. How many species are there? Biodiversity & Conservation, 2(3), 215- 232.

Swope, S.M. and I.M. Parker. 2012. Complex interactions among biocontrol agents, pollinators, and an invasive weed: a structural equation modeling approach. Ecological Applications, 22(8), 2122-2134.

Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics, 123(3), 585-595.

Tamayo, M., C.E. Grue and K. Hamel. 2000. The relationship between water quality, watermilfoil frequency and weevil distribution in the state of Washington. Journal of Aquatic Plant Management 38: 112–116.

Tamayo M. and C.E. Grue. 2004. Developmental performance of the milfoil weevil (Coleoptera: Curculionidae) on watermilfoils in Washington State. Environmental Entomology 33: 872-880.

89

Tamayo, M., C.W. O’Brien, R.P. Creed, Jr., C.E. Grue and K. Hamel. 1999. Distribution and classification of aquatic weevils (Coleoptera: Curculionidae) in the genus Euhrychiopsis in Washington state. Entomological News 110: 103-112.

Tamura, K. 1992. Estimation of the number of nucleotide substitutions when there are strong transition-transversion and G+ C-content biases. Molecular Biology and Evolution, 9(4), 678-687.

Tamura K., D. Peterson, N. Peterson, G. Stecher, M. Nei and S. Kumar. 2011. MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Molecular Biology and Evolution 28: 2731-2739.

Teixeira, I.R.V., and F.S. Zucoloto. 2003. Seed suitability and oviposition behaviour of wild and selected populations of Zabrotes subfasciatus (Boheman)(Coleoptera: Bruchidae) on different hosts. Journal of Stored Products Research, 39(2), 131- 140.

Thompson, R.T. 1992. Observations on the morphology and classification of weevils (Coleoptera, Curculionoidea) with a key to major groups. Journal of Natural History, 26(4), 835-891.

Trakhtenbrot, A., R. Nathan, G. Perry, G., and D.M. Richardson. 2005. The importance of long‐distance dispersal in biodiversity conservation. Diversity and Distributions, 11(2), 173-181.

Unmuth, J.M.L., R.A. Lillie, D.S. Dreikosen, D.W. Marshall, 2000. Influence of Dense Growth of Eurasian Watermilfoil on Lake Water Temperature and Dissolved Oxygen, Journal of Freshwater Ecology 15: 497–503.

U.S. Congress, Office of Technology Assessment. 1987. Technologies to Maintain Biological Diversity, OTA-F-330. Washington, DC: U.S. Government Printing Office.

USDA Plants Database http://plants.usda.gov/java/

Via, S. 2001. Sympatric speciation in animals: the ugly duckling grows up. Trends in Ecology and Evolution 16: 381-390.

Vossbrinck, C., J.C. White, Jason; G.J. Bugbee, K. Prapayotin-Riveros, M. Marko, R. Thum, E. LaRue, N. Havill. 2010. Isolation of microsatellite markers for the watermilfoil weevil Euhrychiopsis lecontei. Molecular Ecology Resources. Unpubl. http://biomath.trinity.edu/manuscripts/10-6/mer-10-0222.pdf

90

Waits, L.P., S.L. Talbot, R.H. Ward and G.F. Shields. 1998. Mitochondrial DNA phylogeography of the North American Brown Bear and implications for conservation. Conservation Biology 12: 408–417.

Wang, I.J. 2010. Recognizing the temporal distinctions between landscape genetics and phylogeography. Molecular Ecology 19: 2605–2608.

Weir, B.S., and C.C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution, 1358-1370.

Weisser, W.W., and E. Siemann. 2004. Insects and ecosystem function (pp. 3-19). Heidelberg: Springer.

Wilson, E.O. and F. M.Peter. 1988. Biodiversity, National Academy Press.

Wright, S. 1982a. Character change, speciation, and the higher taxa. Evolution 36: 427- 443.

Wright, S. 1982b. The shifting balance theory and macroevolution. Annual Review of Genetics, 16(1): 1-20.

91

APPENDICES

92

APPENDIX A

BAYESIAN ANALYSIS FOR FOUR MEMBERS OF THE TRIBE PHYTOBIINI

A separate Bayesian analysis of the four target species for four genes was conducted for all sequenced individuals (Euhrychiopsis lecontei, n = 79; Parenthis vestitus, n = 4; Phytobius leucogaster, n = 12; Eubrychius velutus, n = 17). This analysis was conducted on a concatenated data set of the partial sequences of Cytochrome

Oxidase I (COI) and Cytochrome b (CtyB) mitochondrial DNA, elongation factor-1 alpha

(EF1-α) nuclear DNA, and the 18S ribosomal RNA genes.

Amplification of the partial COI and CytB sequences were conducted using the primers described in the Methods and Materials section of this chapter. Amplification of the partial EF1-α sequence for E. lecontei was conducted using the primers EF1F 5’ –

TGGTGAATTTGAGGCTGGTATCT – 3’ and EF1R 5’ –

TAGGTGGGTTGTTCTTGGAGTCA – 3’. Amplification of the partial 18S sequence was conducted using primers 18SUF 5’– CGAATTCAACCTGGTTGATCCTGCCAGT

– 3’ and 18SUR 5’ – CCGGATCCTGATCCTTCTGCAGGTTCACCTAC – 3’.

Amplification conditions for COI and CytB are identical to those described in the

Methods and Materials section of this chapter. Amplification conditions for Ef1-α

(EF1F/EF1R primer pair) included an initial denaturation period of 3 min at 94°C,

93 followed by 30 cycles of 94°C denaturation (30s) 47°C annealing (30s) and 72°C extension (60s) with a final extension of 5 minutes. Amplification conditions for 18S

(primer pair) included an initial denaturation period of 3 min at 94°C , followed by 30 cycles of 94°C denaturation (30s) 46°C annealing (30s) and 72°C extension (60s) with a final extension of 5 minutes.

Following the alignment of each individual gene, the sequences for all 4 genes were concatenated in Mesquite v. 2.75 (Maddison and Maddison, 2010). The order of concatenation was COI (1080 base pairs), Ef1-α (605 base pairs), CytB (465 base pairs) and 18S (1750 base pairs). The total length of the concatenated data set was 4,164 base pairs, including missing data and gaps. Since every individual was not sequenced for every gene, missing data was indicated in the alignment as unknown (i.e., ?). Real gaps, where insertions/deletions were present, were included as phylogenetically important in the analyses. Suitable sequences of closely related weevil species were initially identified by an NCBI BLAST search in GenBank with one of the newly derived sequences of Euhrychiopsis lecontei to assign appropriate outgroups. One outgroup was chosen from the Subfamily Ceutorynchinae (Ceutorhynchus scrobicollis) and one from the Subfamily Curculioninae (Curculio glandium). These species had the best overall coverage of the genes included in the concatenated gene alignment. GenBank Ascension numbers for C. scrobicollis are: EU551172.1 (COI); JX091299.1 (EF1-α); and

JX091171.1 (18S). CytB sequence was not available for C. scrobicollis. GenBank

Ascension numbers for C. glandium are: HQ165403.1 (COI); AY327674.1 (EF1-α);

AY327749.1 (CytB) and AJ850003.1 (18S).

94

The concatenated data set was exported for analyses into PartitionFinder v.1.1.1

(Lanfear et al., 2012). PartitionFinder assigned the following evolutionary rate models to the data: the Hasegawa-Kishino-Yano Model (Hasegawa et. al. 1985) with gamma distribution and proportion of invariant sites (HKY+I+G) to the mtDNA COI and CytB regions; the Symmetrical Model with invariant sites (SYM+I) to the EF1-α region; and the Kimura 2-Parameter Model (Kimura, 1980) with invariant sites (K80+I) to the 18S region. Data was uploaded to Cipres Science Gateway (Miller et al., 2010) and analyzed in MrBayes 3.2.2 (Ronquist et al., 2011) on XSEDE, an online supercomputer. A

Bayesian Metropolis –coupled Markov-Chain Monte-Carlo analysis of 10,000,000 generations was performed with two sets of four chains (three cold and one hot chain per set). The chains were sampled every 1000th generation with convergence of the lines reaching less than the recommended 0.01 average standard deviation of split frequencies after the 2,650,000 generation. The average standard deviation of split frequencies after

10,000,000 generations was 0.00541. The first 2,500,000 generations were discarded as a burn-in phase. The output trace files were also viewed in Tracer v1.6

(http://tree.bio.ed.ac.uk/software/tracer/) to visually check for convergence. The final tree with posterior probabilities was visualized in FigTree v1.4

(http://tree.bio.ed.ac.uk/software/figtree/). Figure A1.1 shows the phylogenetic relationships of the four weevil species based on the concatenated data set.

The results from this analysis support the results from the Maximum Likelihood analysis reported in the Results section of Chapter 1 (Figure 1.1). The monophyly of the four members of the tribe Phytobiini is supported with 100% posterior probability.

95

Euhrychiopsis lecontei and Parenthis vestitus are sister species with 100% posterior probability and Eubrychius velutus and Phytobius leucogaster are sister species with 96% posterior probability.

Figure A1.1. Bayesian phylogenetic relationship of the four milfoil weevil species. EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus.

96

APPENDIX B

18S ANALYSIS FOR FOUR MEMBERS OF THE TRIBE PHYTOBIINI

A separate analysis was conducted for the partial sequence of the 18S ribosomal

RNA gene (1749 bp) in order to examine the relationship between the four target species

(Euhrychiopsis lecontei, n = 6; Parenthis vesititus, n =1; Phytobius leucogaster, n = 5;

Eubrychius velutus, n = 6) and members of the genus, Bagous, one of the only other weevil groups in Curculionidae known to have colonized aquatic habitats. Forward and reverse sequences for each gene were assembled, manually reviewed and edited with the software package Sequencher v.4.6®. Sequences were imported into MEGA 5.2 (Tamura et al., 2011) and aligned with the built-in automated aligner CLUSTALW.

Amplification of the partial 18S sequence was conducted using primers 18SUF

5’– CGAATTCAACCTGGTTGATCCTGCCAGT – 3’ and 18SUR 5’ –

CCGGATCCTGATCCTTCTGCAGGTTCACCTAC – 3’. Amplification conditions for

18S (primer pair) included an initial denaturation period of 3 min at 94°C , followed by

30 cycles of 94°C denaturation (30s) 55°C annealing (30s) and 72°C extension (60s) with a final extension of 5 minutes.

Real gaps, where insertions/deletions were present, were included as phylogenetically important in the analyses. Suitable sequences of closely related weevil species in Curculionidae were initially identified by an NCBI BLAST search in GenBank 97 with one of the newly derived sequences of Euhrychiopsis lecontei to assign appropriate outgroups. One outgroup was chosen from the subfamily Ceutorynchinae

(Ceutorhynchus scrobicollis; GenBank ascension number: JX091171.1) and one from the subfamily Curculioninae (Curculio glandium; GenBank ascension number: AJ850003.1).

Other outgroups from the subfamily Bagoinae (genus Bagous), are also included in the analyses. A distant outgroup, Cassida rubiginosa (Coleoptera, Chrysomelidae), was also chosen for inclusion in this analysis. GenBank ascension numbers for these individuals

(Bagous spp. and C. rubiginosa) are included next to the species names in Figure B1.1.

The results of this analysis support the phylogenetic findings reported from the COI phylogeny in this chapter. Euhrychiopsis lecontei and Parenthis vestitus are reported as a signal lineage in this analysis, indicating that they are closely related, and

Phytobius leucogaster and Eubrychius velutus are reported as sister species. They represent a separate clade from the Bagous spp. individuals, supporting the conclusion that the ability to live in aquatic environments arose independently in these different subfamilies of Curculionidae.

98

EL1 EL18 EL39 EL54 EL57 EL8 PV3 89 PL1 PL11 PL12 83 PL2 PL4

99 EV14 EV15 EV2 87 EV7 63 EV5 EV8 OUT1 CS OUT2 CG AF389066.1 Bagous myriophyllae AJ849994.1 Bagous puncticollis 98 AJ849993.1 Bagous limosus 62 FJ867738.1| Bagous americanus AY748124.1 Cassida rubiginosa

0.005 Figure B1.1. Molecular phylogenetic analysis of the partial sequence of the 18S gene for the four target species and outgroups by Maximum Likelihood method. The evolutionary history was inferred by using the Maximum Likelihood method based on the Kimura 2-parameter model (Kimura, 1980). The tree with the highest log likelihood (-3238.8113) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites [5 categories (+G, parameter = 0.0500)]. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 25 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. There were a total of 1,749 positions in the final dataset. Evolutionary analyses were conducted in MEGA5 (Tamura et al., 2011). EV=Eubrychius velutus; PL=Phytobius leucogaster; EL=Euhrychiopsis lecontei; PV=Parenthis vestitus.

99